Roman black glass

peter cosyns

Jewels and tableware made of black-appearing glass were popular in the Roman Empire. Compositional changes (due to modifications in glassmaking technology and use of raw materials) over the period considered (1st - 5th century AD) have been investigated on a large number of samples originating from various archaeological excavations in Europe, Northern Africa and the Near East. In the course of this work, over 400 samples of Roman glass, the greatest part of them deeply colored glass fragments, were embedded into acrylic resin and mechanically ground and polished in order to obtain flat surfaces of unaltered glass. The samples were analyzed with SEM-EDX and the quantification was performed by using a standard-less ZAF software. The trace elements contained in a selection of glass samples were determined via LA-ICP-MS. The data collected in this study show that from about 150 AD a change in the black glass production process occurred, involving coloration of raw glass made with iron in the secondary workshops. Furthermore, from the 4th century AD on we can observe a change in the type of raw glass used, while the coloring process was maintained. The main aim of this paper is to provide glass scholars with the analysis results, as reference and comparison for further studies.

Recent excavations in the Roman and late Roman city of Serdica (present day Sofia, Bulgaria) prove that a diversity of urban production activities has taken place there, including secondary glass working. A ceramic sherd with a thick layer of apparently black opaque glass adhering to its inside has been found in an uncertain context, but most likely dating from the late Roman period of the site. This find demonstrates the technology of secondary glass melting and adding iron oxide for colour modification in a wheelturned ceramic vessel originally made as a household ware. To our knowledge, this evidence is one of the few direct examples for black glass making known so far. The fragment was analysed by OM and SEM-EDS, and LA-ICP-MS. The results demonstrate that the glass has a proper Roman composition with an excessive amount of iron oxide (ca. 9wt%) which gives it its distinct colour. A detailed study of the microstructure of the material allows an understanding of the way in which the iron has been added to the batch, with small iron oxide particles seen in the glass at the bottom of the crucible. The interpretation proposed here is that these particles are unlike naturally occurring material but seem to be flakes of hammer scale. It can be assumed that such specific black glass was used for shaping small jewellery, such as beads and/or bangles. An attempt is made here to discuss that technology in comparison with a similar and contemporary find from another region of the Balkans.

1413 glass fragments were found at the settlement of Yurta -Stroyno during the three years of excavation and one year of the surface survey. Most of the retrieved glass was highly fragmented, which is a result of the fragility of the material itself as well as of its deposition in secondary contexts. The glass collection from Yurta -Stroyno includes a wide range of vessels and glass items, such as personal ornaments and window panes. For the vessels, high quality glass was used, resulting in thin -walled fragments made of translucent, colourless glass, with a minimum of impurities. glass was decorated by wheel -cutting, mould blowing and applied threads. During the surface survey, a group of specific glass fragments was found, later identified as waste from glass production. these include threads, drops, moils, heat -melted fragments and fragments of raw glass, with the latter counting 81 pieces with total weight of 189 g. The retrieved glass material was first typologically classified and preliminary dated based on comparattive data from other settlements, further, the method of XRF analysis was chosen to complement the data needed to interpret the finds. The analysed set of samples represents a selection of different glass objects (89 pcs.), and fragments of the raw glass and the production waste (including production indicators; 28 pcs.)

The production, distribution and consumption of black glass in the Roman Empire during the 1st - 5th century AD An archaeological, archaeometric and historical approach Peter Cosyns Proefschrift ingediend tot het behalen van de academische graad van doctor in de Archeologie en de Kunstwetenschappen, Vakgroep Kunstwetenschappen en Archeologie Faculteit Letteren en Wijsbegeerte, Vrije Universiteit Brussel promotor: Prof. dr. Karin Nys (VUB) co-promotor: Prof. dr. Koen Janssens (UA) Brussel 2011 Table of contents ACKNOWLEDGEMENTS .............................................................................................................................................. vii INTRODUCTION ...................................................................................................................................................... ix 1. The delimitation of the studied material ................................................................................... xi 2. The chronological delimitations ............................................................................................... xii 3. The geographical delimitation ................................................................................................. xiii 4. Delimitation of the research topic ........................................................................................... xiii 5. The delimitation of black-appearing glass or the colours of black ...................................... xiii 6. Methodology ............................................................................................................................. xiv 7. Previous research on the various issues of Roman black glass ....................................... xviii ABBREVIATIONS .................................................................................................................................................... xxi Chapter 1 Defining ‘black’ and the perception of colour ........................................................................... 1 1.1 Introduction ------------------------------------------------------------------------------------------------------------------------ 1 1.2 Colour perception in Antiquity ------------------------------------------------------------------------------------------------- 1 1.3 Linguistic and psychological approach to colour perception ---------------------------------------------------------- 2 1.4 Colour perception in archaeology -------------------------------------------------------------------------------------------- 3 1.5 Proposed application ------------------------------------------------------------------------------------------------------------ 4 1.6 Initial observations --------------------------------------------------------------------------------------------------------------- 7 1.7 Conclusion ------------------------------------------------------------------------------------------------------------------------11 Chapter 2 Technology................................................................................................................................. 13 2.1 Introduction -----------------------------------------------------------------------------------------------------------------------13 2.2 The production -------------------------------------------------------------------------------------------------------------------13 2.2.1 The raw materials ........................................................................................................................ 13 2.2.2 The workshops ............................................................................................................................ 15 2.2.3 Built-in basins and crucibles ........................................................................................................ 16 2.2.4 The tools ...................................................................................................................................... 16 2.2.5 The production waste and recycling ............................................................................................ 19 2.3 The Artefacts ---------------------------------------------------------------------------------------------------------------------19 2.3.1 Vessels ........................................................................................................................................ 19 2.3.2 Jewellery ...................................................................................................................................... 26 2.3.3 Architectural decoration ............................................................................................................... 31 2.3.4 Counters ...................................................................................................................................... 32 2.4 Comparative survey of the various colours of black glass in relation to the applied technology ------------34 2.4.1 Production ................................................................................................................................... 34 2.4.2 Vessels ........................................................................................................................................ 35 2.4.3 Jewellery ...................................................................................................................................... 38 Chapter 3 Chrono-Typology ....................................................................................................................... 45 3.1 Introduction -----------------------------------------------------------------------------------------------------------------------45 3.2 Vessels -----------------------------------------------------------------------------------------------------------------------------46 3.2.1 Introduction .................................................................................................................................. 46 3.2.2 Cast vessels ................................................................................................................................ 47 3.2.3 Free-blown vessels ...................................................................................................................... 51 3.2.4 Mould-blown vessels ................................................................................................................... 66 3.2.5 Rod-formed vessels ..................................................................................................................... 70 3.2.6 Summary ..................................................................................................................................... 77 3.3 Jewellery ---------------------------------------------------------------------------------------------------------------------------80 3.3.1 Arm rings ..................................................................................................................................... 80 3.3.2 Finger rings .................................................................................................................................. 93 3.3.3 Beads ........................................................................................................................................ 105 3.3.4 Pendants ................................................................................................................................... 116 3.3.5 Gems ......................................................................................................................................... 121 3.3.6 Hairpins ..................................................................................................................................... 124 3.4 Architectural decoration ------------------------------------------------------------------------------------------------------ 126 3.4.1 Tesserae .................................................................................................................................... 126 3.4.2 Twisted rods .............................................................................................................................. 127 3.4.3 Bichrome plaques ...................................................................................................................... 130 3.4.4 Inlays or intarsia......................................................................................................................... 131 3.5 Counters ------------------------------------------------------------------------------------------------------------------------- 131 Chapter 4 Chronology............................................................................................................................... 137 4.1 Introduction --------------------------------------------------------------------------------------------------------------------- 137 4.2 Production ----------------------------------------------------------------------------------------------------------------------- 137 4.2.1 Primary production..................................................................................................................... 137 4.2.2 Secondary production ................................................................................................................ 138 4.2.3 Recycling ................................................................................................................................... 140 4.3 Vessels --------------------------------------------------------------------------------------------------------------------------- 140 4.3.1 Period I ...................................................................................................................................... 142 4.3.2 Period II ..................................................................................................................................... 143 4.3.3 Period III .................................................................................................................................... 146 4.3.4 Period IV .................................................................................................................................... 146 4.4 Jewellery ------------------------------------------------------------------------------------------------------------------------ 147 4.4.1 Arm rings.................................................................................................................................... 147 4.4.2 Finger rings ................................................................................................................................ 155 4.4.3 Beads ......................................................................................................................................... 157 4.4.4 Pendants .................................................................................................................................... 158 4.4.5 Gems ......................................................................................................................................... 160 4.4.6 Hairpins ...................................................................................................................................... 162 4.5 Architectural Decoration ----------------------------------------------------------------------------------------------------- 162 4.5.1 Tesserae .................................................................................................................................... 163 4.5.2 Twisted rods............................................................................................................................... 164 4.5.3 Bichrome plaques ...................................................................................................................... 165 4.5.4 Inlays or intarsia ......................................................................................................................... 165 4.5.5 Written sources .......................................................................................................................... 165 4.6 Counters ------------------------------------------------------------------------------------------------------------------------ 166 4.7 Conclusion ---------------------------------------------------------------------------------------------------------------------- 170 4.7.1 Period I....................................................................................................................................... 170 4.7.2 Period II...................................................................................................................................... 171 4.7.3 Period III..................................................................................................................................... 172 4.7.4 Period IV .................................................................................................................................... 172 Chapter 5 Context analysis ...................................................................................................................... 175 5.1 Introduction --------------------------------------------------------------------------------------------------------------------- 175 5.2 Civil contexts ------------------------------------------------------------------------------------------------------------------- 179 5.2.1 Domestic context ....................................................................................................................... 185 5.2.2 Funerary context ........................................................................................................................ 192 5.2.3 Ritual context ............................................................................................................................. 203 5.3 Military contexts --------------------------------------------------------------------------------------------------------------- 206 5.3.1 Funerary context ........................................................................................................................ 209 5.3.2 Ritual context ............................................................................................................................. 209 5.3.3 Public contexts ........................................................................................................................... 210 5.3.4 Domestic contexts ...................................................................................................................... 210 5.4 Other ----------------------------------------------------------------------------------------------------------------------------- 216 5.4.1 Caves ......................................................................................................................................... 216 5.4.2 Rivers ......................................................................................................................................... 217 5.5 Conclusion ---------------------------------------------------------------------------------------------------------------------- 218 Chapter 6 Production, Distribution and Trade ........................................................................................ 221 6.1 Introduction --------------------------------------------------------------------------------------------------------------------- 221 6.2 Production ---------------------------------------------------------------------------------------------------------------------- 222 6.2.1 Organisation of the production of coloured glass ....................................................................... 224 6.2.2 Analysis of secondary workshops .............................................................................................. 227 6.3 Distribution and trade -------------------------------------------------------------------------------------------------------- 234 6.3.1 Vessels ...................................................................................................................................... 234 6.3.2 Jewellery .................................................................................................................................... 241 6.3.3 Architectural decoration ............................................................................................................. 263 6.3.4 Counters .................................................................................................................................... 263 6.4 Conclusion ---------------------------------------------------------------------------------------------------------------------- 266 Chapter 7 Use and function ...................................................................................................................... 271 7.1 Introduction --------------------------------------------------------------------------------------------------------------------- 271 7.2 Reflections on the use and meaning of colours in Roman society ----------------------------------------------- 272 7.3 The vessels --------------------------------------------------------------------------------------------------------------------- 273 7.3.1 Cast vessels............................................................................................................................... 274 7.3.2 Free-blown vessels .................................................................................................................... 274 7.3.3 Rod-formed vessels ................................................................................................................... 277 7.4 Jewellery ------------------------------------------------------------------------------------------------------------------------ 278 7.4.1 Arm rings.................................................................................................................................... 279 7.4.2 Finger rings ................................................................................................................................ 284 7.4.3 Beads ......................................................................................................................................... 285 7.4.4 Pendants .................................................................................................................................... 288 7.4.5 Gems ......................................................................................................................................... 290 7.4.6 Hairpins ...................................................................................................................................... 295 7.5 Architectural decoration ----------------------------------------------------------------------------------------------------- 296 7.5.1 Tesserae .................................................................................................................................... 296 7.5.2 Rods........................................................................................................................................... 297 7.6 Counters ------------------------------------------------------------------------------------------------------------------------ 299 7.7 Conclusion ---------------------------------------------------------------------------------------------------------------------- 301 Chapter 8 Chemical analysis .................................................................................................................... 305 8.1 Introduction --------------------------------------------------------------------------------------------------------------------- 305 8.2 Methodology ------------------------------------------------------------------------------------------------------------------- 318 8.2.1 SEM-EDX................................................................................................................................... 318 8.2.2 LA-ICP-MS................................................................................................................................. 320 8.3 Studied material --------------------------------------------------------------------------------------------------------------- 322 8.4 Results --------------------------------------------------------------------------------------------------------------------------- 325 8.4.1 SEM-EDX .................................................................................................................................. 325 8.4.2 LA-ICP-MS................................................................................................................................. 333 8.5 Archaeological Relevance -------------------------------------------------------------------------------------------------- 336 8.5.1 Augst vs Avenches .................................................................................................................... 336 8.5.2 Deviant material ......................................................................................................................... 338 8.6 Conclusion ---------------------------------------------------------------------------------------------------------------------- 340 8.7 Addendum ----------------------------------------------------------------------------------------------------------------------- 343 8.7.1 p-XRF and µ-XRF ...................................................................................................................... 343 8.7.2 Portable X-Ray Fluorescence (p-XRF) ...................................................................................... 343 8.7.3 Micro X-Ray Fluorescence (µ-XRF) ........................................................................................... 346 8.7.4 Raman spectroscopy ................................................................................................................. 351 Chapter 9 Optical analysis ....................................................................................................................... 357 9.1 Introduction --------------------------------------------------------------------------------------------------------------------- 357 9.2 Absorption spectroscopy on Roman glass artefacts ----------------------------------------------------------------- 358 9.3 Methodology -------------------------------------------------------------------------------------------------------------------- 363 9.4 UV-Vis-NIR Spectroscopy results on Roman artefacts in strongly-coloured glass appearing black ----- 370 9.5 Use of photonics and its efficiency in relation to archaeological questions ------------------------------------- 392 9.6 Conclusion ---------------------------------------------------------------------------------------------------------------------- 399 Chapter 10 Historical Analysis ................................................................................................................... 404 10.1 Introduction --------------------------------------------------------------------------------------------------------------------- 404 10.2 Case study I: Macro-level --------------------------------------------------------------------------------------------------- 406 10.3 Case Study II: Meso-level --------------------------------------------------------------------------------------------------- 410 10.4 Case-study III: Micro level --------------------------------------------------------------------------------------------------- 418 10.5 Conclusion ---------------------------------------------------------------------------------------------------------------------- 420 General Conclusion ......................................................................................................................................... 422 THE CATALOGUE ................................................................................................................................................... 429 APPENDICES ................................................................................................................................................... 435 LIST OF REFERENCES ............................................................................................................................................. 489 LIST OF FIGURES ................................................................................................................................................... 511 LIST OF TABLES ................................................................................................................................................... 519 ACKNOWLEDGEMENTS I would like to express my deep gratitude towards the staff of the Vrije Universiteit Brussel (VUB) who enabled me with a grant from the Research Council (OZR 1247 BOF) to write this dissertation within a research project on black glass, and by enabling the implementation of this research to encourage the collaboration between research groups with expertise in different research disciplines through the funding of the Horizontal Orchestrated Action project HOA15, under the supervision of Prof. Dr Hugo Thienpont (TONA), Prof. Dr Karin Nys (SKAR/MARI) and Prof. Dr Herman Terryn (MEMS). My most profound appreciation, however, goes to my supervisor Prof. Dr Karin NYS. From the very start, she believed in me and my work. Although I am quite sure that I drove her to great despair with my line of action, she never seized to coach and to encourage me to complete this work. Further I wish to thank every member of the research team MARI and all the others of the ‗Vakgroep Kunstwetenschappen en Archeologie‘ for being interested in my work and for encouraging me to accomplish this research. A special note is essential concerning the generating of the Chapters 8 and 9. Chapter 8 has only been made possible thanks to a partnership with the Department of Chemistry, Universiteit Antwerpen (UA – MiTAC) of Prof. Dr Koen Janssens and his team at the Centre for Micro- and Trace AnalysisDr Olivier Schalm, Dr Veerle Vander Linden and Drs Simone Cagnofor the time and energy they spend in the chemical analyses, but also for their significant input to achieve the here presented results. Yet it is noteworthy to add the substantial contributions by Dr Bernard Gratuze (CNRS Orléans, France) for having executed the initial LA-ICP-MS analysis. I am highly indebted to the participation of the Department of Applied Physics and Photonics (VUB – TONA) of Prof. Dr Hugo Thienpont and the outstanding collaboration with his assistant Dr Wendy Meulebroeck, whose contribution by means of absorption spectrometry analysis was of considerable importance. Without her zealous scientific commitment and her fascination for archaeology, Chapter 9 could not be written. I would like to thank curators, curator assistants and archaeologists for allowing me to study theirprimarily unpublishedmaterial and especially everybody who granted me permission to sample material for analyses: In Belgium: KMKG-MRAH, Brussels Provinciaal Gallo-Romeins Museum - PGRM, Tongeren ARON, Tongeren Musée Archéologique, Namur Grand Curtius, Liège Erfgoedcel, Tienen VIOE, Brussels Universiteit Gent - UGent, Ghent Provinciaal Archeologisch Museum - PAM, Velzeke Stedelijke Musea, Brugge Université Libre de Bruxelles - ULB, Brussels Université Catholique de Louvain – UCL-CRAN, Louvain-la-Neuve Ministère de la Région Wallonne – MRW, Jambes Musée du Malgré-Tout, Treignes Archeologisch Museum, Grobbendonk formerly Katholieke Universiteit Leuven – KUL, Leuven Antwerpse Vereniging voor Romeinse Archeologie – AVRA, Kontich Anne Cahen-Delhaye; Eric Gubel; Cécile Evers; Natasja Massar; Luc Limme; Dirk Huyghe; Claire Massart Guido Creemers; Bart Demarsin; Else Hartoch Petra Driesen; Elke Wesemael Jean-Louis Antoine Constantin Chariot; Jean-Paul Philippart; Jean-Luc Schütz Staf Thomas; Tom De Bruyne Rica Annaert; Marleen Martens; Alain Vanderhoeven; Sophie Vanhoutte Roald Docter; Wim Declercq; Partick Monsieur Marc Rogge; Kurt Braeckman Hubert De Witte Eugène Warmenbol Fabienne Vilvorder; Raymond Brulet Jean Plumier; Frédéric Hanut; Olivier Vrielynck Pierre Cattelain Danielle Horemans Arnold Provoost Rik Verbeeck And abroad: Vienna and Magdalensberg, Austria formerly Rheinisches Landesmuseum Bonn, Germany Römermuseum Augst, Switzerland formerly Musée Romain d‘Avenches, Switzerland Schweitzerisches Nationalmuseum Zürich, Switzerland Barbara Czurda-Ruth; Eleny Kaudelka-Schindler Anna-Barbara Follmann-Schulz Sylvia Fünfschilling; Sandra Ammann Anne de Pyre-Gysel Heidi Amrein vii University of Koper, Slovenia; formerly Proskrajina Musej Celje, Slovenia Musée des Antiquités Nationales, M.A.N., SaintGermain-en-Laye, France CNRS, Aix-en-Provence, France CNRS-Lattes, France CNRS, Maison de l‘Orient Lyon, France Israel Antiquities Authority - IAA Jerusalem, Israel Department of Antquities, Nicosia formerly LAARC and University College London, UCL, London, UK Petrie Museum for Egyptian Archaeology, London, UK formerly English Heritage, Portsmouth, UK formerly Durham University, UK Cardiff University, UK Colchester Archaeological Museum, UK Ashmolean Museum, Oxford, UK Gemeentelijke Archeologische Dienst, Nijmegen, Netherlands Museum Het Valkhof, Nijmegen, Netherlands Irena Lazar Hélène Chew Danièle Foy Michel Bats Marie-Dominique Nenna Yael Gorin-Rosen Pavlos Flourentzos; Despo Pilides John Shepherd Stephen Quirke Justine Bayley Jennifer Price Ian Freestone Paul Sealey Helen Whitehouse Peter Vanden Broecke; Harry van Enckvort Louis Swinkels; Annemie Koster I wish to thank in particular for all their help, support, discussions, ideas, advice, scholarly attitude and/or contributions to my research over the past years: Dirk Pauwels, Eugène Warmenbol, Frédérique Hanut, Souen Fontaine, Jennifer Price, Danièle Foy, Florence Doyen, Jean-Marc Doyen, Bernard Gratuze, Justine Bayley, Yael Goren-Rosen, Marie-Dominique Nenna, Sylvia Fünfschilling, Jean-Marc Doyen, and Alexis Wielemans. The team members of doctoral researchers at the VUB university department SKAR (Sara Adriaenssens, Tamara Berghmans, Barbara Borgers, Danielle Caluwé, Jan Coenaerts, Pieter-Jan Deckers, David Dewilde, Marie Geraerts, Ariane Jacobs, Nelleke Teughels, Hilde Wouters). This PhD is no exception to the rule: it alwaysalwaystakes longer than originally planned. The decision to write in an international language was often regretted and certainly did not advance a smooth wind-up. Stratton and Sarah were so kind to review a few chapters for which I am utterly grateful. I claim all responsibility for the many remaining mistakes and I solemnly apologise to native speakers for the desecration of their beautiful but complicated language. And it would have taken even longer if not for the logistic support from Tinny and Koen, Ilse and Peter, Jonas. They offered me their homes as a welcome and necessary seclusion from my ever alive and active offspring. Marianne was always willing to help out were possible, which she did. In no time Bruno conjured the very professional cover. I hope this does not get readers‘ hopes up too high... Without a doubt, for years my mother‘s phone calls would always end with an enquiry about the progress of my work. She, and my father, were and will remain a great support to me. Yet, I am most indebted to Ann Hasendonckx. During this giant‘s labour I was not much of a companion. Although at times physically present, my mind often strayed to Roman black glass. She spent lots of time alone with the children. Her criticism on me and my work was mostly fierce but true. Nevertheless she supported me every single day during these last tough years. Finally, our two little ‗kakkewieten‘, Martha and Emiel kept me going. They, and maybe you too, would have preferred a book on princes or wild animals, but unfortunately my topic is merely black glass. viii INTRODUCTION The original idea for having undertaken this research on Roman black glass arose from the poor availability of comparative material when studying the glass finds from the 1996-2002 excavations of the Roman site at TienenGrijpenveld (BE). The scattered availability of related information clearly illustrates that black glass artefacts occur regularly on Roman sites albeit the number of artefacts remains, in most cases, limited. Despite the ubiquitous occurrence of Roman black glass artefacts, these objects have been mainly discussed in the margin of entire glass assemblages, barely receiving attention as a subject of research. Furthermore, the problem of the colour perception, the black appearance of an object, was neglected. On the other hand, the main research focus within the world of glass researchers remained up till now limited to vessels and production. The pervasive disinterest in other functional types resulted in the reduced number of publications with regard to black glass artefacts, the majority of which consists of jewellery. Reading through these publications resulted in several observations: 1) that black glass artefacts have been produced all through the Roman imperial period; 2) that black glass was used to make the whole range of consumer goods, with only a few exceptions such as window glass; 3) that its distribution covers the entire Roman Empire; 4) that all over the empire a great uniformity in the shape prevails, at most showing minor variations in the decoration; and 5) that some commodities demonstrate a regional production and consumption. Because the study of a rather scantily occurring material on excavation sites like black glass only can give interesting results when taking into account a wide-ranging geographical area, like the Roman Empire, and a long chronological period, like the entire Roman imperial period, we therefore chose to cope with the black glass through a vertical study approach instead of a horizontal one. Hence, the main objectives of this manuscript were to understand the different operating processes influencing Roman glass production in general and consumption of Roman black glass within the Roman economy, considering the globalizing character of glass production and the local or (inter)regional character of glass consumption. Present observation necessitates considering whether these black glass artefacts are commodities of 1) local; 2) regional; or 3) long-distance trade. The empire-wide use of similar objects indicates an aspect of uniformity within the Roman material culture but also results in questioning on how to understand this omnipresence of analogous objects. The fact that a specific type of object was a local commodity but produced all over the Roman Empire in a similar way or that a commodity coming from one single production centre was traded all over the Empire demonstrates that all through the imperial period, the Roman material culture was overall mainstream up to the tiniest artefact, whatever the prevailing economical organization. We therefore set up a large-scale status quaestionis in the attempt to recognize workshops that produced black glass artefacts, or regions where workshops should be identified, to better understand the impact of its production. But when uniformity of shape is preponderant, the stylistic features appear everywhere analogous and not distinctive. Hence, technological aspects can be crucial to recognize idiosyncrasies, given that every workshop should have had its own specific technological know-how. The small technical features within the production process of the glass artefact may reveal the fingerprint of a glass blower or generation of glass blowers from one workshop on the one hand or the technological tradition of a workshop on the other hand. The technological expertise of a workshop can also become recognizable through the identification of different recipes and production techniques used to make glass appear black. For that reason, we sampled a wide range of material from various places of the Roman Empire and from every phase within the Roman imperial period. This was also necessary to assess whether the glass workshops used imported black glass to make artefacts or whether the glass workshops themselves coloured ‗naturally coloured‘ or decolourized glass. Two approaches will thus verify whether it is possible to detect different workshops and if such detection can offer new insights on the production, distribution and use of Roman glass. The main aim of this study on Roman black glass has been to focus the greatest scientific assets possible on a minor segment within the Roman glass studies by putting forward a holistic study. Hence, the interdisciplinary character of the research is executed with an interaction between the archaeological, the archaeometric and the historical disciplines, verifying if this vertical analysis is supplying new information to the study of Roman glass. This vertical analysis comprises a diachronic study on the composition, production, function, and distribution of material from the entire Roman Empire. So we can ascertain geographical differences based on synchronic variations within the entire Empire as well as chronological variations by means of diachronic variations within a regional or supra-regional area. Possibly it also will help to refine dating as well as interpreting the archaeological context, given that the obtained knowledge can provide a better insight into the social and economic aspects of Roman society. Such approach is perhaps not new, but needs greater valorisation because ―only rarely does one find an explicit and focused critique of the various sources and approaches employed, as well as their integration, especially when archaeometric analyses are included in the equation‖ (Martinón-Torres, Rehren 2008, 13). ix Various theoretical models on how the Roman economy was organized have been proposed through centuries. The aim of this research is to verify whether a minor segment within the Roman material culture can add fresh information to this discourse and provide modification proposals. The study of Roman black glass covers a wide range of material that has been produced and consumed within the entirety of the Roman Empire and this during the first five centuries AD. Considering on the one hand the immensity of the studied area, the long-term period, and the large extent of the discussed materialabout production, vessels, jewellery, architectural decoration, gaming pieces, and, on the other hand, the holistic approach of the research, it was neither possible to be exhaustive nor possible to process in detail the many discussed issues. If the here-presented research on Roman black glass does aim to examine one particular issue, it is regionalism versus globalization, showing a Roman material culture that confronts regional and globalizing consumption. Given that we in the first place wanted to provide an overall picture of the production, distribution and consumption of all commodity types diagnosed as black glass, we verified the studied material on the basis of chronological and/or regional idiosyncrasies. Hence, the focus of the research lay in assessing as many approaches and parameters as possible, in order to improve the knowledge on Roman black glass, but above all to demonstrate that a holistic approach to a minor segment within the material culture can have a great effect on the understanding of the glass production and consumption in the Roman Empire. Although we initially designed an arrangement of the different issues per commodity type to obtain comprehensible units, we became convinced that an arrangement of topics provides the reader a more convenient overview of the assembled data on Roman black glass. We therefore turned the structure inside out by upgrading the chapter sections to individual chapters and downgrading the initial chapters to the level of chapter sections. The dissertation is subdivided into four parts:     The first seven chapters that form Part I give a detailed overview of the compiled dataset from an archaeological approach. In separate chapters are discussed the colour perception in antiquity (Chapter 1), the techniques utilized to make black glass artefacts (Chapter 2), the classification of the various commodities in black glass (Chapter 3), the typo-chronology (Chapter 4), the provenance of these artefacts (Chapter 5), the production and consumption (Chapter 6), and finally the use and meaning of the various commodities in black glass (Chapter 7). The two chapters with the archaeometric approach, representing Part II, discuss various archaeological issues on the basis of chemical analyses (Chapter 8) and optical analysis (Chapter 9). The selection of various types of commodities from different areas and periods made it possible to characterize geographical and chronological groups of black glass that can work as a tool to solve archaeological queries. Part III consists of a historical approach (Chapter 10) integrating the analysis results from both previous approaches into specific cultural and socio-economical questions examined on three different levels: microlevel, meso-level and macro-level. The black glass artefacts have been set into a wider and more general view of the Roman society during the imperial period. The catalogue constitutes Part IV. This card index gives an account as detailed as possible of every single black glass artefact recorded depending on the availability of information per object and within the feasibility of the project. The dataset worked out in Access 2007 has been put on CD-ROM to suit the qualifications for a user-friendly and fast consultation. Yet, extended lists of the finds are on hand in the tables and appendices of various chapters. The dataset is supported by a comprehensive card index of every single piece on the supplied CD-ROM. The source of the general discussion varies throughout the subsequent chapters. That in Chapters 1, 2, and 7 reflects what is available in literature on the respective topics. That in Chapters 4-6 is based on the catalogued material in the database and that in Chapter 3 has been developed from both sources. The general discussion in Chapters 8–9 is built up from the analysis results. The final discussion in Chapter 10 compares the compiled datasets of all previous chapters in comparison with what is available in literature. Concerning the established distribution patterns, we had to take different parameters into consideration. It is important to note that the actual observed concentration zones are, like always, the result of different factors not related to the real trade distribution but rather due to the availability of the archaeological and bibliographical records. The conspicuous concentrations observed can be explained by intensive excavations (e.g., Karanis [EG]; Colchester [UK]; Pompeii [IT], but also by the nature of the archaeological site (e.g., cemetery) or its historical x circumstances (eruption of Mt. Vesuvius). We therefore had to bear in mind these factors when dealing with each type of material. 1. The delimitation of the studied material The initial aim was to generate an exhaustive study, but due to practical reasons this could not be carried out fully. First of all, we encountered various limitations when recording published material. The colour of the glass artefacts is sometimes omitted from the description. Therefore, we took into consideration that we possibly th overlooked a number of published objects in black glass. Sometimes the opposite happensespecially in 19 th century and early 20 century excavation reportsso that we might notice the discovery of black glass artefacts but without any description, illustration or measurements. Fortunately, we received from various institutions in Belgium and abroad the opportunity to study their collections. As these collections included unpublished material, we are aware that we had to take into account an unknown amount of material that remains unknown to us. For practical reasons, we had to make a selection of museums and archaeological depots we planned to attend. Not only was it impossible to gain permission to visit all museums with a collection of Roman glass, but we also have been confronted on several occasions with selected museums being inaccessible due to renovation. For instance, the Kelsey Museum of Archaeology of the University of Michigan in Ann Arbor, Michigan (USA), with large glass collections from various Egyptian sites such as Karanis, was closed for researchers until April 2010. The large quantities and high diversity of black glass material made inevitable the choice of material to accomplish our research aims. Most information accumulated from published material concerns the glass production, the vessel assemblages and the bracelets. Consequently, we could discuss extensively on these topics, but jewellery other than bracelets has been considered as well. We therefore could focus on the various jewellery types. More restricted is the treatise on the architectural decoration material (e.g., cakes, rods and tesserae) and the counters. Different reasons made these functional types receive only little attention. The specificity of the architectural decoration material and the ubiquitous counters makes it more convenient to consider separately each material category in view of the entire range of colours wherein both categories have been produced. The discussion of both categories, however, includes guidelines for more penetrating future 1 investigations. The very rare examples of glass sculpture meant that we did not consider this category. Some categories or unique pieces have been excluded from this present study for various practical reasons:  the exceptional occurrence and little information about the date and provenance makes it impossible to classify correctly some specific artefacts Firstly, there is the cylindrical bucket (situla) with silver handle of the San Marco treasure (Harden et al. 1987, 220-221, no. 122; Brown 1991, 85, no. 1). This unique piece has only been excluded because it turns out to be impossible to classify it with certainty. Cast in deeply violet glass appearing black, the vessel would perfectly fit in st the overview of 1 century cast vessels in a monochrome deeply coloured glass. The bucket, however, is dated th th on stylistic grounds in the 4 5 centuries AD, considering the wheel-cut technique and the figurative design (Albizatti 1923, 51-63; Fremersdorf 1951, 24; Harden, Toynbee 1959, 201). The few known equivalents are the th rd early Hellenistic cast situla now in the Metropolitan Museum of Art, New York dated in the late 4 –early 3 century BC (Lightfoot 2003, 19-20, figs. 3-5) and the 3rd century AD diatretum-situla also in the San Marco Treasure (Hahnloser 1971, no. 13; Brown 1991, 85, fig.1a). It is not unrealistic that the piece was made in the Hellenistic period and that it survived for centuries (as it has done until now!) before it got decorated with a wheelcut scenery depicting Dionysus together with Satyrs and Maenads. The present silver handle on its turn is dated 7th century AD (Brown 1991, 85, no. 1). An answer will probably only be reached when chemical analysis can be done to compare its composition. Another example is the fragment of a horse leg of a life-size statue now in the collections of the British Museum, London. Depending on the author, the piece is said to have been made of deep purple glass appearing black (Harden et al. 1987, 220-221, no. 122) or made of obsidian (Grose 1989, 342). A chemical analysis can establish the used material as there is a difference between both materials [see Chapter 8]. Apparently, the piece has been analysed on its composition and is said to confirm Grose‘s statement that the statue was made of obsidian, but 2 the analysis has not been published so far. 1 Concerning the sculpture in black glass/obsidian, we more or less are relying on the brief notes by Pliny the Elder (Naturalis Historia XXXVI, 196197). 2 Personal communication by Jennifer Price 2009. xi Also not integrated, although much more eye-catching, are the cameo-vases and the cameo-plaques, because incorporating them would only create confusion and the extensive discussions would lead us too far from the present one (Goldstein et al. 1982; Harden et al. 1987, 53-84, no. 29-36; Journal of Glass Studies 32, 1990). The huge interest in this very specific commodity category had already led to plenty of detailed discussions and publications. These lavishly decorated vessels are customarily discussed with their true hue leaving little discussion within this research. The bichrome cameo vases however cannot be excluded from the discussion on strongly-coloured glass vessels appearing black, and therefore we will refer to this material on various occasions. This luxurious class shows the masterful craftsmanship of the 1st century AD glass workers creating a black-andwhite contrast by combining an opaque white decoration on a poorly translucent deep blue and violet glass.  the material is characteristic for the pre-imperial period or post-Roman period Some artefact categories have been removed at some point from the database because of their inconsistent date vis-à-vis the here-discussed material. One of these is a group of black appearing bangles in deep purple, blue and brown glass occurring on early Roman sites in the north-western provinces are in fact re-used late La Tène bangles which have been made in a different technology and received a different shape and decoration uncharacteristic for Roman jewellery. Another group of bracelets omitted from the database is classified by Maud Spear as type B4 (Spaer 1988, 57). This type of bracelet, made of deep purple glass appearing black, is characterized by five to seven lengthwise ribbings and a very clear circular widening at the seam. Present in th th ample museum collections, this type of glass bangle is generally dated late Roman–early Byzantine (5 7 century AD), and its provenance is situated in the east Mediterranean, although most pieces have no provenance. th th Maud Spaer says that this type was particularly fashionable in Syria during the 6 7 century AD (Spaer 2001,199-200, no. 449). Coincidently a jeweller‘s shop close to the entrance of a church has been recently excavated at Kala‘at Seman (SY) where were retrieved a large number of bracelets, including type B4, together th 3 with a number of coins dating to the mid 7 century AD. The shop seems to have been suddenly abandoned after a collapse, making it a very well-datable closed context. Considering the dated context and the nonstandard features for Roman glass banglesthe use of deep purple glass; the wide, flat closed shape with one end creating a circular bezelwe believe this to be an early Byzantineearly Islamic type we did not find in any late Roman context. Another category that has been omitted due to its principally post-Roman dating is the so-called linen smoother or slick-stone, a solid hemispherical glass cake with rounded edge resembling a gigantic plano4 convex counter. Its top surface is convex and its bottom surface is concave, repeatedly showing a large pontil 5 scar in the centre. The diameter varies between 65-95 mm, and the height is in the order of 20-40 mm, whereas the weight averages between 250-300 grams. This very recognizable type of artefact in black-appearing glass is th th perhaps characteristic for the 9 to 13 century, but a number of pieces also occur in earlier Carolingian contexts (Haevernick, Haberey 1963; Isings 1980, 233, fig.156; Macquet 1990; Mainman, Rogers 2000, 2534, fig.1236, nos. 6592-6594; Gratuze et al. 2003b; Munier 2009, 68, fig.6; Isings 2009). But some of these semi-circular blocks seem to appear already in Merovingian contexts, like the two pieces in the Musée Archéologique of Namur (respectively inv. no. 1788 and inv. no. A2776) one from a Frankish tomb at Saint-Denis-Nausipont (BE) (s.n. 1862, 223) and another from Flavion (BE) (unpublished). The production of these hemispherical glass lumps thus appears to precede even the Carolingian era, probably starting from the Merovingian period onwards, with the stemmed pieces from Augst (CH) (Rütti 1991, 165-166, fig.105-106) as Roman forerunners. The two pieces from the late Roman military camp at Oudenburg (BE) (unpublished) probably have to be considered medieval artefacts, since both pieces are not from well-dated Roman contexts but from the thick black earth layer covering the entire site. Besides the fact that the massive pieces are commonly considered late medieval and postmedieval objects, the material is only regarded as appearing black due to its thickness. Because this type of artefact lasts for many centuries and had a vast distribution area, it is apparent that chemical analysis can help identify chronological and regional clusters (Gratuze et al. 2003b; Bayley 2009; Wedepohl 2009). 2. The chronological delimitations Artefacts in deeply coloured glass appearing black have been produced and distributed all along the Roman imperial period for five centuries and even beyond. Within the East, black glass remained in use all through the 3 personal communication by Danièle Foy 2009 Chemical analysis already provided evidence to the contrary that ―linen smoothers‖ functioned as cakes of raw glass metal to be processed in the batch on the basis of a too-high concentration of impurities or the presence of elements not present in other glass artefacts (Foy 2004, 32; Gratuze et al. 2003b, 101107). 5 It appears that these so-called linen smoothers were produced by gathering a mass of hot glass from the batch using a large metal rod followed by pressing the hot glass in a hemispherical or slightly conical shape. That the metal rod has been swivelled round is clearly visible from the spiralling creases at the convex under side. 4 xii early Byzantine period up to Islamic times until today, while in the West its use only lasted in the Merovingian and th Anglo-Saxon cultures for a short period of time, although again from the 17 century onwards for the production of wine and liquor bottles. When sorting out glass artefacts that appear black, it becomes clear that black glass occurs from the very early times of glass production until today, as well as during the pre- and post-Roman period. It is clear that the production and use of black glass was not restricted to the Roman period only. 3. The geographical delimitation To understand best the distribution and trade of the black glass artefacts, we took into consideration the objects from all countries that coincide with the former Roman Empire, not limited by artificial modern borders. The validity of this more general archaeology is, for instance, well explained by Michel Feugère with his TPC-team at the CNRS Lattes (2008, 26-28), who describes the approach of ‗Technique-Production-Consommation‘ as: …difficile en effet de se limiter à une période ou à un territoire, tant il est vrai que les frontières seraient, sur une telle thématique, impossibles à fixer ou à respecter. Tel quel, notre thème d‘étude possède cependant une vraie légitimité, que cette contribution me donne l‘occasion d‘analyser en détail. (Feugère 2008, 21) The studied area is defined as largely as possible, encompassing 32 actual countries around the Mediterranean and on the European continent, reaching from Portugal to Syria and from Egypt to Scotland. Conscious that a fully exhaustive study for such a vast region was not realistic to accomplish within the given term, we therefore decided to work on specific items about black glass. We sampled material from different areas to obtain an overview on the production and use of black glass within the Roman Empire. We also checked as far as possible the occurrence of black glass items from beyond the empire because some commodities are considered to be non-Roman. 4. Delimitation of the research topic Besides the long date range and vast area of distribution the project planned to consider, we were aware of the enormous difficulties of assembling all material from all different functional types from all countries of the former Roman Empire. In a first stage, we checked all possible publications within our reach and started to go to museums and archaeological depots to study the unpublished material. Some striking aspects were visible at once: (1) several regions are not providing material as abundantly compared to others, which is partly due to the limited excavation activities on sites or layers of the concerned period and partly due to the limited excavation results published; (2) the concentrations of specific material are featuring chronological and/or geographical assemblages. Soon a large variety of material emerged, including vessels, a wide variety of jewellery (e.g. bracelets, beads, pendants, finger rings and hairpins), but also architectural decoration material and counters. For practical reasons, we narrowed the core of the studied material to vessels and jewellery, even though the architectural decoration material and the counters were, to a certain degree, taken into consideration.    5. Within the extensive range of use of black glass by the Romans, the existence of some material type will only be mentioned, but not extensively, like the glass gems imitating nicolo, the mosaic tesserae and other architectural decoration material, and the counters. The main material discussed here are the vessels (cast, free-blown, mould-blown, rod-formed); jewellery (bracelets, beads, finger rings, pendants). A comprehensive overview of the black glass material from the provinces Gallia Belgica and Germania Inferior is provided and an outline of that from Britannia and Germania Superior with a literature study and as much material as possible. The delimitation of black-appearing glass or the colours of black The studied corpus groups diverse objects having the main feature of being deliberately deeply coloured to appear black, and this black aspect is the result of the density of the colouring and the opacity of the fabric. We assumed it was necessary to explain how we looked at the very deeply coloured Roman glass in the first place. Additionally, we described what sort of glass artefacts we considered to be black even though the glass matrix consists of a particular hue, and we excluded artefacts that we categorized as black-appearing. Therefore, we considered within the introductory Chapter 1 a discourse on the colour(s of) black with respect to the Roman xiii glass. Further, we have described and discussed the black-appearing glasses with an archaeometric approach looking at their chemical composition [see Chapter 8] and their colour properties through optical analysis [see Chapter 9]. Different chemical analysis methods were used to define different groups within the so-called black glass compositions, discriminating among different periods and different geographical zones of production [see Chapter 8]. In parallel, an optical investigation by means of absorption spectroscopy was employed to define systematically the delimitations of the black appearance of glass [see Chapter 9], making possible operating in an objective way with the homogeneity of the considered glass material with a black appearance. The subjective character of the colour perception affects specifically our heuristic research. Probably by tradition, the terminology used to define the colour aspect of a glass object appearing black varies from author to author and with the category of artefacts. Some publications give the subjective evaluation by limiting the description as ―black glass‖, while others only report the detected colour by means of light in transmission, without defining in what way the piece appears black. While it is commonly known that black and white are not to be considered as coloursthe former absorbs all colours of the colour spectrum, while the latter reflects themboth tones had serious meaning in antiquity and consequently a great impact on the material culture. Nevertheless, in reality socalled black and white glass will always have one dominant tinge. Through the absorption of all reflective light, black glass is so densely coloured that it is usually said to be opaque (Bayley 1999, 90; 92) [see Chapters 1; 9]. The investigating eye is therefore the best medium to discriminate the true colour of black glass, particularly with the assistance of a strong light. When one uses light in transmission, one dominant colour will become visible which makes it possible always to determine the true colour of black-appearing glass, such as green, brown, purple, blue or even red. Similarly, white and colourless glass will always contain a greenish, bluish, yellowish or pinkish tinge. These prevailing tinges are the result of their particular chemical composition owing to different behaviours of the (de-)colouring process and thus are of great interest to the archaeologist interested in the production of glass and its distribution pattern. Within the study of glass in general and archaeological glass in particular, the colour description plays an important role. It has long been standard knowledge that the colours of Roman glass and even the tinges within almost colourless glasses are inextricably bound with chronology. However, this colour chronology is subject to certain conditions. The colour description gives a certain idea of the chemical composition, particularly when it is 6 combined with the typological determination of the artefact. However, these colour observations can only serve as a guiding principle for explaining the typological analysis and always need verification from analysis of the chemical composition. The description of glass colours is rather complex because the observation of the colour varies depending on who determines the object and under what conditions it was done. Not only the individual perception of the colour is biased, but also the terms used to describe colours are subjective. It is therefore essential that the black appearance or the will to make the glass appear black should be considered in selecting the studied material. If we keep in mind the intentional black colouring and not the colour itself and its eventual opacity, we can start inquiring about the particularity of this specific group of objects. Are we dealing with a tinge or do the black glass artefacts have to be seen as a particular group that surpasses the strict framework of hues? In other words, do the types and shapes in so-called black glass coincide with those in all other colours, or do they show a characteristic typology? Can we detect different concentration zones within the distribution of these so-called black glass artefacts vis-à-vis whether the same material in intensely coloured, natural coloured or decolourized glass, and do we find them within similar contexts? Is there a difference in distribution when taking the functions into account (e.g. vessels, bracelets, finger rings, beads)? Can we distinguish concentration zones when observing systematically the used colourants to make glass appear black, or in other words, are the socalled black artefacts indifferently brown, green, blue or purple or can we discriminate particular distributions or chronological attributions or subcategories inside each category? 6. Methodology In the first place it was essential to define the criteria to select black glass and the samples. The discussion above makes clear that black (glass) in fact never occurs, causing a primary problem in the methodology for researching something that in se does not exist but was definitely used and referred to by Romans. The fundamental question here is a priori to define when a piece of glass can/may be described as black. Although the problem is primarily to create a scientifically-based, objective arrangement, it is equally important to check to what extent it is possible to understand the Roman perception of colours and its meaning. In addition to the problem within the research on 6 Recently presented work on the selected glass from the military camp of South Shields, UK provides evidence for production groups depending on the selection of glass colours (Jackson, Price forthcoming Annales of 18th AIHV conference Thessaloniki-2009). xiv the colour of black glass is that of the sometimes carelessly established material description. In particular concerning the jewellery, we are aware of the problem caused by describing black glass material mistakenly, e.g. in jet, obsidian or even bronze. But an inspection on all published material in jet, obsidian… in national and international museum collections and archaeological depots was beyond the scope of this project. The used methodology distinguishes three approaches: (a) an archaeological approach, (b) an archaeometric approach and (c) a historical approach. (a) Archaeological approach This approach defines the research criteria for the study on Roman black glass and provides a framework (Figure 1) wherein the research will be carried out to come to: Technology Context Typology Date Typo-chronology Figure 1: Schematic view how to reach a typo-chronology a. An analysis of the chrono-typology The technologies used to make the wide variety of black glass artefacts have been the basis of the typology and have been linked with the dated contexts wherein such materials were retrieved. This analytical research will result in a chrono-typology within each function (vessels, bracelets, beads, rings and gaming pieces), which will be compared to morphological parallels in the different materials. b. Contextual analysis One of the additional aspects of the archaeological research consists in identifying how (and consequently also why) artefacts got left behind at a specific place. Important therefore is to read correctly the archaeological feature, which is the last archaeological context wherein small finds became part of. The different aspects of use, function and meaning of every type of the acquired chrono-typology have been approached by means of a context analysis (military, burial, ritual, domestic or workshop). The obtained data have been interpreted in particular from an historical approach to get a better understanding of consumption patterns and social, economic and cultural aspects. c. Analysis of production, distribution and trade of black glass Determining the provenance of each object is important to set out distribution maps of the vessels as well as of the jewellery (bracelets, finger rings, beads) and gaming pieces. It provides a deeper understanding of the local consumption patterns and the regional and interregional distribution and trade in the whole Roman Empire during the imperial period. The set up of distribution maps helps to define whether local, regional and/or interregional distributions are identifiable by means of similarities or differences for specific commodities (relation within one civitas; between several civitates within one province; between different provinces). xv Was the market supplied through a few workshops responsible for a long-distance distribution, or was the production of black glass artefacts carried out by a wide range of local production centres for local and regional distribution? In the first case, we merely can deduce a chronological significance for the various types of material. Concerning the second possibility, we should check whether regional differences can be observed and if they evolve differently over time. In the last situation, the set up of a typology will not be useful unless there has been uniformity in production in all workshops or in those of a specific region. When everywhere the stylistic features are analogous, then technological aspects become crucial, given that every workshop should have had its own specific technological know-how. This expertise can be linked to the different recipes used, and therefore we sampled a wide range of material from various places of the Roman empire and from every phase within the Roman imperial period. Small technical features within the production process of the glass artefact could also reveal the fingerprint of a glass blower or generation of glass blowers from one workshop. So there are two approaches that will verify if it is possible to detect different workshops. Therefore, the aim of the here-presented research questions on Roman black glass has been to apply the greatest range of scientific assets possible to a minor segment within Roman glass studies by putting forward a holistic study. We set up a large-scale status quaestionis of this uncommon or perhaps even peculiar type of glass artefacts within the Roman society in the hope of developing new insights on the production, distribution and use of Roman glass. The three above-mentioned goals on the selected material can only be acquired in the first place by means of publications but also by consulting the national and international collections/depots, their unpublished archaeological reports and their archaeological diaries. The heuristic part requires a series of criteria for the selection of black glass and for the selection of samples to set up a databank of published objects. A digital cardindex in Access 2007 was worked out to enable an identical analytical registration for all material from every site, with attention for measurements, colour, shape, decoration, technology, preservation, conservation and provenance/context [see Part IV]. The digital cardindex was designed to integrate drawings and photographs of the object. The analytic research involved the registration of the research material reaching the selection criteria of black glass (e.g., description [see above], photograph, drawing and occasionally sampling) within Belgian and a number of important European museum collections/archaeological depots. It is important to place the studied material within the Roman glass tradition and compare it with similar shapes in non-black glass from different regions in the Roman Empire and throughout the imperial period. This will help in the study of consumer patterns by looking for the chronological and geographical differentiation in the use of different materials for similar shape and/or decoration pattern. To accomplish a correct overview on the consumer patterns of black glass artefacts, it is important to compare the Roman black glass in a similar way with comparative shapes in materials other than glass. Also is taken into account a comparison between the black glass of Roman times with that of the pre-Roman period and the post-Roman period. (b) Archaeometric approach This approach gives valuable information concerning the colour, date and origin of black glass. The acquired data helped to solve the archaeological and historical questions by means of cluster analysis, seriation or other statistical methods. Only diagnostic material were taken into account for sampling. Black glass fragments which are undeterminable were not sampled for physico-chemical analysis as well as most pieces without context information. a. Determine what makes green, blue, purple or brown glass so dark and/or opaque that the coloured glass appears optically black. But when can we start speaking of so-called black glass? To understand what black means or stands for, we had to select a system to define black. Besides inquiring what black is, it is even more important to unravel and understand how the Romans did see all this. When dealing with black glass, the identification of its colour remains very important as glass in all kinds of colours can appear black. The method used to come to reasonable results was acquired through optical spectroscopy (UV-Vis-NIR) by Dr ir. Wendy Meulebroeck at the Department of Applied Physics of the Vrije Universiteit Brussel (VUB – TONA). b. Determine what black glass is made of and define how the Romans made glass appear black. To fully understand what black glass was made of in Roman times, we were assisted by Prof. Dr Koen Janssens of the MiTAC at the University of Antwerp, under whose supervision Dr Olivier Schalm, Dr Veerle Van der Linden and Drs Simone Cagno analysed the selected samples by means of electron probe micro analysis (SEM-EDX) at the University of Antwerp and using laser ablation inductively coupled plasma micro spectrometry (LA-ICP-MS) at the xvi University of Warsaw (Poland) and the University of Ghent (UGent). These physico-chemical analyses form the basis to (a) define clusters by means of plotting the chemical composition into graphics; (b) detect clusters which are to be linked with glass workshops (determining the origin); and (c) detect clusters which are to be linked with a specific period (determining the date). The obtained results are to be checked with other results of national and international natural science research, but a selection of samples measured by the Antwerp team with LA-ICP-MS was checked by Dr Bernard Gratuze at the CNRS Centre Babelon, Orléans (France) to see if the results corroborated. c. Acquire insight in the recognition of different productions, in time as in space. Can we come to a geographical and chronological subdivision of the differently coloured glass appearing black by sampling all kinds of well chosen samples? To what extent is there an evolution definable within the Roman period itself? Are there any differences observable between Roman black glass and that of previous or later periods? The performed physico-chemical analyses on selected samples included an intensive course at the University of Antwerp to initiate Peter Cosyns into learning all different stages of sampling, embedding in resin, preparation for the measuring and the measuring of the samples with the JEOL 6400. The latter was always done under supervision of Dr Olivier Schalm or Drs Simone Cagno for a number of the resins, but most have been measured by them or Dr Veerle Van der Linden. The calculations of the analysis results were done under the supervision of Prof. Dr Koen Janssens. By plotting the obtained data into graphics, we were able to compare and deduce the acquired results to understand the characteristic material-technical features of the analysed black glass samples. Although it was also the aim to determine if the production of black glass was intentional or coincidential by defining how long the material was heated, how frequently and at what temperature the material was heated, we 7 have not been able to arrange techno-morphological analysis by using vibration spectrometry. Defining the thermic history of black glass would give very important information for discerning glass workshops and production periods, but partly due to time limitations this investigation has been omitted from this research. (c) Historical approach This approach concentrates on the social, economical, cultural and religious aspects of the use of black glass by means of written sources combined with the obtained results from the archaeological and archaeometric approach. After a review of the literature (e.g. ancient sources, historical and archaeological publications), it is necessary to integrate the archaeological and archaeometric data with the acquired data from the study of literature to acquire a better knowledge of the social, the economic and the cultural impact of black glass in Roman society. a. Socio-cultural and religious aspects (function, meaning and acculturation processes) For what purpose did the Romans make objects in black glass? Can it all be interpreted as purely functional (tableware, hairpins, counters) or ornamental (bracelets, finger rings, pendants, tesserae), or are there some specific objects made for a particular reason with a special significance attributed to them? We tried to define why a specific object has a specific shape and check if black glass objects were used for specific rituals within a specific social group within a specific period, integrating morphological parallels. We made an attempt to determine what meaning Romans gave to the wearing of black glass jewellery and the use of black (glass) vessels. b. Socio-economic aspects (production and consumption) We hoped to determine which mechanisms within the Roman society were responsible for the emergence of nd rd black glass vessels in the second half of the 2 century AD and the disappearance in the second half of the 3 century AD, while black glass remained in production for jewellery. In connection with this, we needed to th determine the reasons for the vanishing of the production of bracelets in western Europe in the first half of the 5 century AD, while in the Byzantine Empire and later in the Islamic world the production remained stable. Important in the research of Roman black glass is the detection of the eventuality of chronological and geographical differentiation within the variety of black colours by use of distinct colouring agents. Combining 7 This technique can detect how black glass is made and is to be executed at the CNRS of the Université de Lyon 1, under the supervision of Prof. dr. Bernard Champagnon, will define the thermic history of specific selected samples to determine whether or not the production of Roman black glass was intentional. xvii typology with chromatography and chemical composition is a necessary approach in the challenge to solve fundamental questions: Were the different black colours produced in different periods and/or different regions? What are the technical aspects with a link to chronology, distribution patterns and origin of the material? Publications referring to black glass objects, unfortunately, do not always distinguish its true colour from its black appearance by only pointing out its black appearance or by identifying the hue without mentioning that the object appears black. When choosing a research topic on a subject such as the Roman black glass, one of the questions that were raised spontaneously dealt with the issue of how the subject is to be approached. Restricting the research to a minor fraction within the glass researcha material group representing not more than a small part in the archaeological findsmight give reason to some scholars to raise doubts regarding the value of the research if not objections to the aim of it. While black glass is scarcely attested on excavations, it is found on most sites all over the Roman Empire, and st th this within contexts from the 1 century AD up to the 5 century AD. Consequently, this vast research domain required a different approach than is usually the case. To date, glass studies ain general consider all glass items from one specific site or region and ultimately are limited to one specific period or function. Mostly without any limitation concerning the colour, this type of research is to be seen as a study with a horizontal approach. When the research queries are focussing on a limited range of material with a specific restriction in colour or function but covering a vast area and a long-lasting period, only a thin section is taken into consideration, so that we can call this type of research a vertical approach. The primary aim of this research is to verify to what extent this characteristic study can enhance glass research in general. Some fundamental questions leaped to the forefront: How did the Romans produce black glass? Were different recipes used to reach black glass? Can productions be discerned clustering different regions and/or different periods? Is there a difference between vessels and jewellery? What material is idiosyncratic for a certain region and/or period? Can the context give a comprehensive insight of the social, economical and cultural function of the black glass artefacts? These and many other questions can only be answered comprehensively with a multidisciplinary approach, combining the archaeological data with archaeometric analyses and by examining the historical facts. The archaeological approach resulted in accumulating information on different aspects like technology, typology, context and date. These aspects are interrelated since the typology is the initial group subdivided into main groups by means of the used technology. These main groups can enclose an amount of variants based on decoration. This methodological approach is not always used in previous researches like in the study of the preIslamic bracelets from Palestine (Spaer 1988) where the typology is based on the opposite way. This approach resulted in a subdivision independent of any chronological grounds, as not the decoration but the technology seems to have an important role. However, to be useful for archaeologists confronted with the material, it is vital to study the dated contexts to come to a functional and valuable chrono-typology. The archaeometric approach is of essential assistance to the archaeological approach by pinpointing the chemical composition of all kinds of material from different geographical areas and consecutive periods. The historical approach resulted in collecting information on glass in general and on black glass material in particular by ancient writers and work from present-day historians on Roman economics and socio-cultural issues. 7. Previous research on the various issues of Roman black glass A thorough study of the Roman black glass can add value within the study of the material culture of Roman Antiquity. Therefore, it is surprising that at present no overall study on Roman black glass has been carried out. Nevertheless, this type of glass is integrated in general studies on Roman glass (e.g., Vessberg 1952; Isings 1957; von Saldern 1980) and in archaeological reports (e.g., Harden 1936; Van den Hurk 1973; Rütti 1991; Cool, Price 1995). The information is in general limited to a description of the artefacts and of their context. The technology, chemical composition, origin, function and meaning of black glass objects are only described sporadically and very generally. Some publications give an insight on a specific aspect of Roman black glass, but with limitations in time and space: Isings (1964) gave an small overview of the then-known black glass vessels st from the 1 -–3rd century AD in the western part of the Roman Empire; De Witte (1977) discussed the typology, chronology, function and technology of Roman black glass bracelets from Belgium in his thesis; Spaer (1988) mapped out a typology on pre-Islamic glass bracelets from Israel; and finally recent research, on the basis of a xviii chrono-typological study of the black glass vessels in northern Gaul, pointed out the unique opportunities black glass can provide archaeological contexts as a ‗type fossil‘ (Cosyns, Hanut 2005). Bracelets The research on Roman glass bracelets has remained rather marginal until now. We see that the scarce information at everyone‘s disposal makes this material remain unpublished or only mentioned in a very limited sense. The rarity of this material as such on Roman sites and the absence of an in-depth study and an userfriendly manual prevents this type of material from being known by archaeologists. This becomes obvious when reading through general publications on the small finds of large archaeological sites. Sometimes it is easy to verify doubts because of the proximity of the museum or archaeological depot where the object is deposited. We know now that the so-called jet bracelets from the late Roman cemetery of Oudenburg (BE) which are exhibited in the Archaeological Museum of Bruges (Mertens, Van Impe 1971, 128, no. 100: 7-8, pl. XXXII: 5-6; 208, no. 194: 3, pl. XL: 4) and those from a number of tombs in Amiens and now in the Archaeological Museum of Amiens (Dilly, Mahéo 1997, 120; 123) are made of black glass. Far away material and indistinct publications are less easily checked. For instance, it is unclear whether some bracelets of the Roman cemeteries from Pécs (HU), ancient Sopianae, are not rather of black glass than of jet (Fülep 1977, 29, grave R/151-163, pl. 14: 2; 43, grave R/227, pl. 26: 4; 55, grave R/307, pl. 39: 4) and if mentioned as being of glass, it is not always clear from which colour of glass they are produced (Fülep 1977, 50, grave R/275: 8; 55, grave R/307: 1). The main studies on glass bracelets concentrated in the first place on the pre-Roman Celtic bracelets (Haevernick 1960; Peddemors 1975; Gebhard 1989a-b; Venclová 1990; Karwowski, 2004; Wagner 2006) and to a lesser degree on the Islamic and Byzantine material (Spaer 1992; Shindo 1996; 2001; Boulogne 2007; 2008). On Roman glass bracelets, several studies were discussed local or regional material (Kilbride-Jones 1938; Stevenson 1956; De Witte 1977; Spaer 1988; Riha 1990; Wagner 2006) but an attempt at synthesis was never undertaken to come to a convergent classification system with a corresponding chronology. Beads Compared to the pre-Roman glass beads (Haevernick 1960; Guido 1978; Zepezauer 1989; 1992; Venclová 1990; Wagner 2006) and early medieval glass beads (Guido 1999; Destexhe 2003; Brugmann 2004), we see rather little interest in the research of Roman glass beads (Guido 1978, Swift 2000; 2003). This type of jewellery is, within the here-presented work, almost entirely restricted to the comprehensive publications on Roman cemeteries where all material is discussed (Keller 1971, 87-94; Mertens, Van Impe 1971; Vanvinckenroye 1984; Riha 1990; Chéhab 1985-1986). Gems Similar to the intaglios in semiprecious and precious stones, those in glass can be plain or decorated, small and large, and in different colours, while shapes seem to remain limited [see Chapter 3]. Restricting this survey to black glass gemstones only is considering hardly more than a limited group of the wide range of gemstones used for insertion, above all in finger rings. We are aware that we are excluding a large part of the glass intaglios, made entirely of or partly of black glass. This commodity group is so vast that it is very suitable for a separate research. Accordingly, we only wanted to verify to what extent this study can be of added value to the entire body of knowledge regarding Roman black glass production and consumption. A wide range of detailed studies on the engraved gems from the different Roman provinces in the West are at hand, making available a comprehensive comparative dataset: for the British Isles (Henig 1978²), France (Guiraud 1988; 1996; 2008), Belgium (Sas 1992; 1993), the Netherlands (Maaskant-Kleibrink 1978), Italy (Sena Chiesa 1966; 1978), Germany (Henkel 1913; Krug 1975; 1977; 1978; 1980; 1995; Platz-Horster 1984; 1987; 1994), Switzerland (Vollenweider 1979) and Austria (Zwierlein-Diehl 1973; 1979). Unfortunately, it appears that the study of gemstones, thus far, has been focussed on the intaglios in various (semi)precious stones and in particular on the rendered image, whereas it is only too obvious now that the majority of the gems found in the archaeological record has been produced by moulding glass [see Chapter 1]. In particular, those made of blue8 on-black glass imitating gemstones in nicolo seem to form the bulk of the gemstones (Sas 1993). It is for that reason all the more strange to observe that the study of so-called inferior, bulk-produced artefacts in cheaper 8 It is interesting to see that in antiquity imitations of (semi)precious stones were common and that nobody seemed to notice these except for taking a disparaging opinion towards the people who dared to fake genuine gemstones with glass (Pliny the Elder, Naturalis Historia XXXVII, 197). xix materials and of rather mediocre workmanship never received the same attention as the more precious counterparts decorated with comparatively higher-quality workmanship. This lack of interest has previously been noticed too when considering the technological part of the glass gemstones: Die moderne Literatur zu Gemmen berührt die Frage der Glaspasten-Herstellung kaum, und auch die umfangreiche Spezialliteratur zu Glas lässt diese Frage beiseite. (Krug 1978, 486-487) The importance of these less luxurious objects with inferior decoration that remains certain is that glass gems are not unique pieces. In contrast with the (semi)precious stones, engraved by hand and thus considered exclusive pieces, the cast or moulded glass gems are characterised by the possibility of serial production. Such sets can give interesting information: 1) on the distribution pattern and thus the trade radius of a specific piece of jewellery; 2) on the date range of production and of its (re)use. Counters Due to a lack of synthesis or any in-depth research on glass counters, thus far, only a narrow view of this topic is available. Considering only those counters made in a deeply coloured glass appearing black, it is pretty obvious that the here-presented contribution cannot be considered as a comprehensive study. The black glass counters should be considered together with those made in other coloured glass metals. Additionally, we have to emphasise that within the scope of the present research, we were unable to work exhaustively on this type of artefact in black glass. The heuristic research would have taken us too far from the main issue of the research. Aware of the restricted approach, we are, nonetheless, convinced that interesting new information on the matter has been supplied here by investigating separately the standard features, and likewise the other material in black glass (e.g., the technology, the dimensions, the contexts, the typology, the distribution). A first step has been to check what shapes were made in (black) glass. Secondly, we verified whether the different types can be linked to chronological and/or geographical models. Further, we have related the technological aspects to the chemical composition, to the various colours of the glass metals used to produce black glass counters, and to the ancient literature referring to the production of glass counters. xx ABBREVIATIONS Countries Listed are the countries where Roman black glass artefacts have been retrieved. The codes are based on official international codes for the representation of names of countries ISO 3166. This standard is published and maintained by the ‗International Organization for Standardization‘ (ISO) at Geneva (CH) and defines the names of countries and country codes. More specifically is made use of part 2 (ISO 3166-2) country subdivision codes. Online information: http://www.iso.org/iso/english_country_names_and_code_elements http://en.wikipedia.org/wiki/ISO_3166-2 Table 1: Country subdivision codes alphabetical per code AL Albania AT Austria BE Belgium BG Bulgaria BS Bosnia CH Switzerland CY Cyprus CZ Czech Republic DE Germany DZ Algeria EG Egypt ES Spain FR France GR Greece HR Croatia HU Hungaria IL Israel IT Italy JO Jordan LB Lebanon LU Grand Duchy of Luxemburg LY Libya MA Morocco MK Macedonia MN Montenegro NL Netherlands PT Portugal RO Rumania SB Serbia SI Slovenia SY Syria TN Tunisia TR Turkey UA Ukraine UK United Kingdom AL DZ AT BE BS BG HR CY CZ EG FR DE LU GR HU IL IT JO LB LY MK MN MA NL PT RO SB SI ES CH SY TN TR UA UK alphabetical per country Albania Algeria Austria Belgium Bosnia Bulgaria Croatia Cyprus Czech Republic Egypt France Germany Grand Duchy of Luxemburg Greece Hungaria Israel Italy Jordan Lebanon Libya Macedonia Montenegro Morocco Netherlands Portugal Rumania Serbia Slovenia Spain Switzerland Syria Tunisia Turkey Ukraine United Kingdom Measurements       d. h. w. l. th. max. pres. diameter height width length thickness (of the wall for a vessel) maximum preserved xxi For the vessels we use max.pres.l. and max.pres.w. when the orientation of the fragment is unclear, and max.pres.h. and max.pres.w. when the orientation of the fragment is known; th. defines the thickness of any part (rim; wall; foot; handle) of the vessel; d. defines the diameter of rim; foot; body (maximal diameter). For the beads we use the same codes as for the arm rings and the finger rings:  d. (in) = interior diameter or opening; d. (ex) = exterior diameter or top surface; d. (section) = diameter of an O-shaped section;  h. = distance from basal (inner) surface to top (outer) surface or height of the section (equal to thickness (th.) when discussing vessels); egularly confused with width for jewellery;  w. = distance from the left side to the right side or the width of the section but regularly defined as ‗length‘ = l.);  l. = maximum preserved length of a bead, arm ring or finger ring (i.e. max.pres.l.). Dating system     I, II, III, IV, V A, B a, b, c, d 1,2,3 century half a century quarter century third century To simplify the reading and writing of dates, we adopted the abridged dating system introducing Roman numbers for each century, the capitals A and B to define respectively the first and second half of a century and the minuscules a, b, c and d to specify the successive quarters of a century. st Example: ‗Ic-IIB AD‘ = ‗the third quarter of the 1 century AD to the second half of the 2 150/200 AD nd century AD‘ = 50/75- This shortening system is especially used in the tables and in continuous explanations on dated finds or dated contexts, but in the main sections of the text the classical full description will be kept. When a source is giving a clearly defined date of the context wherein the black glass object was retrieved, the dates are given between brackets. Publications The applied abbreviations of the main typologies in use on Roman glass vessels: I = Isings AR = Augusta Raurica T = Trier AV = Avenches/Aventicum Clasina Isings (1957) Beat Rütti (1991) Karin Goethert-Polaschek (1977) Françoise Bonnet-Borel (1997) xxii PART I THE ARCHAEOLOGICAL APPROACH Chapter 1 DEFINING ‘BLACK’ AND THE PERCEPTION OF COLOUR Colour is one of the most prominent features of human experience, but has often been ignored or overlooked in archaeological research… Colour awareness and colour sensitivity must however be an integral part of many archaeological analysis concerned with the development and nature of human cognition. (Viewpoint in Cambridge Archaeological Journal 9:1 (1999), 109) 1.1 Introduction When dealing with black glass, we are confronted with a heterogeneous visual interpretation of the material. Therefore, it is essential to articulate our reasoning about the visual examination of colours in general and of glass hues in particular. This approach determined our definition of black glass and our classification of certain artefacts as ‗black‘ glass described by others as deeply coloured glass. 1.2 Colour perception in Antiquity When thinking of black glass artefacts, we have to bear in mind not only the limited perception of colours by the naked eye but also the variation in the perceptions of each human being reading colours individually. ‗Black‘ is a saturation of any colour by absorbing all light and reflecting any, making the hue of the material, e.g., the glass matrix, not (directly) visible to the naked eye. This was noticed already in Antiquity by Theophrastus or Strato, th rd Aristotle‘s successors at the Peripatetic School in Athens (late 4 or early 3 century BC), to whom Aristotle‘s 9 book on colours, peri chromatoon is attributed: … black appears to us in three ways. In the first, that which is not seen is, generally speaking, black naturally (for any light from such things is reflected as black); or secondly, black is that from which no light is conveyed to the eyes; for that which is not seen, when the surrounding region is seen, gives an impression of black. Thirdly, all things appear black of the kind from which a very small amount of light is reflected. (Aristotle, On colours I [transl. Hett, W. S., 1963, 5]) We may thus assume that three degrees of black were distinguished in Hellenistic times and in all probability also 10 during Roman times. Firstly, the passage proposes a number of things to be black that in nature reflect black light, most likely corresponding to shiny black materials. Secondly, black objects are considered as not reflecting light at all, presumably referring to matte/dull black materials. Thirdly are those items reflecting light very poorly, thus making perceptible the true colour of the object. Most interesting is that in the two latter cases, the explanations have been translated as ‗gives an impression of black‘ and ‗appear black‘. Put in plain words, these two descriptive passages mean that the author(s) of de coloribus considered a certain category of objects as black, even though they must have observed these objects appearing differently from what they viewed as true black, whether it involved a more greyish tinge or a specific colour. The use of the words υαντασία (‗impression‘) and υαίνεται and υαίνονται (‗appearing‘) indicate the knowledge in Antiquity that (a degree of) very deeply coloured materials were defined as black. Furthermore, we have to be aware that in Antiquity, the complexity of colour perception was already recognized (Bruno 1977, 47-51). The people in Antiquity, for instance, knew that the colour of an object could vary due to the lighting conditions, the degree of opacity-transparency, and the physical aspects of colour perception. We do not see any of the colours pure as they really are, but all are mixed with others; or if not mixed with any other colour they are mixed with rays of light and with shadows, and so they appear different and not as they are. Consequently things appear different according to whether they are seen in shadow or in sunlight, in a hard or a soft light, and according to the angle at which they are seen and in accordance with other differences as well. (Aristotle, On colours III, 793b, 12-19 [transl. Hett, W. S., 1963, 17]) So that all colours are a mixture of three things, the light, the medium through which the light is seen, such as water and air, and thirdly, the colours forming the ground, from which the light happens to be reflected. (Aristotle, On colours III, 793b, 34-35; 794a, 1-3 [transl. Hett, W.S., 1963, 19]) 9 This tractate was also translated into Latin: de coloribus. It is, however, essential to take into account a possible discrepancy between this theoretical description of various gradations of black and its visual identification from the individual perception by common people at that time. 10 1 Finally, it is essential to mention that Romans like Pliny the Elder actually gave huge importance to the technical discourse on colours. Providing a multitude of very precise information on the pigments in, for example, metals, he failed to give much attention to the colours in artistic works unless it were within a moral discourse (Naas 11 2006). The reason is that in Antiquity, there were no rules nor parameters established to categorize the colours with basic colour terms (Brécolaki 2006). 1.3 Linguistic and psychological approach to colour perception From the papers of the conferences ‗Farbe. Material-Zeichen-Symbol‘ held in 1983 in Berlin (Kurzrock (ed.) 1983), on colour semantics and colour symbolism, and ‗Progress in Colour Studies‘ held in 2004 in Glasgow (Pitchford, Biggam (eds.) 2006a-b), on colour perception and colour categorisation, it is clear that the discussion on colour is a hot topic in linguistics and a pertinent issue in psychology. During the Glasgow meeting, two aspects were discussed in more detail: 1) the issue of categorizing the colours, which is mainly an anthropological debate focussing on variance correlated to age, gender and ethnicity (Bornstein 2006) 2) the issue of naming these ordered categories, which is chiefly a linguistic aspect, although the anthropological discourse is never far away, seeing the interest of ethnographical research (Berlin, Kay 1999) A milestone in a better understanding of the medium of colour is the anthropological work by Brent Berlin and 12 Paul Kay. Studying twenty different languages, they verified the terminology used to designate the colours by paying attention to the hue and brightness. This was made possible by plotting the results on a Munsell colour chart. Clusters were formed, and after analysis of the results, seven linguistic stages could be determined incorporating 22 combinations based on 11 perceptual categories encoded in the basic colour terms (Berlin, Kay 1999, 2-3; 2245, tab.1, fig.3): Stage I: Stage II: Stage III: Stage IV: Stage V: Stage VI: Stage VII: 2 terms (white AND black) 3 terms (+ red) 4 terms (+ (a) green OR (b) yellow) 5 terms (+ yellow AND green) 6 terms (+ blue) 7 terms (+ brown) 8 or more terms (+ purple; pink; orange; grey OR combinations of these) Brent and Kay also demonstrated that a linguistic group in, for instance, Stage I can perceive more colours than black and white. However, these languages ignore any basic colour term to designate those hues as they apparently give importance to the degree of brightness distinguishing ‗brilliant‘ and ‗dull‘ colours (Berlin, Kay 1999, 23-25). Languages within Stage II use the term ‗dark‘ to designate black, including all dark colours, except when it is a ‗warm‘ colour, such as brown, yellow, or red. Remarkably, languages in this stage, for instance, consider blood as ‗black‘ and not as ‗red‘ (Berlin, Kay 1999, 25-28). Within the same line, we can refer to the Greek µασρος, meaning ‗black‘, which is used on Cyprus to designate dark brown objects while other brown objects are called κόκκινός, meaning ‗red‘ (Berlin, Kay 1999, 28). Conversely, the word µασρος is also used in Greek to describe the so-called red wine, which is in fact characterised by a very dark purple colour. Within Stage III, a basic colour term for green or yellow is introduced, while blue and violet are termed ‗black‘. Important to these observations, nonetheless, is that besides the use of basic colour terms in a language, a number of particular colours can be named after the colour of an object, after the origin of the material, or simply by evoking a situation to name a colour for which no basic term exists. In Latin, for instance, no basic colour term existed for blue. But this colour was described in several ways: aeri similem, meaning ‗likewise the sky‘, or caerulea, referring to ‗cerulean blue‘. The blue pigment is labelled with words such as indicum, referring to ‗indigo‘ pigment balls that were supposed to come from India, and lapis lazuli, referring to the blue stone. Conversely, Vincent Bruno (1977, 85) refers to a 11 In Antiquity not only a dress code was applicable in Roman society and culturally embedded, the use of colours and colour combinations seems to have been socially stratified and gender fixed. This issue however could not be worked out any further within the limitations of this project, in particular because it is not explicitly relatable to the consumption of black glass. Though, we touched this matter when some examples are described within the chapter on use and function of black glass commodities [see Chapter 7]. 12 Although its high impact on researchers dealing with colour and colour perception since its publication in 1969, criticism grows for the last decade (Saunders, van Brakel 1997; Saunders 2000). 2 passage in Pliny‘s Naturalis Historia (XXXV, 46) describing the use of the Indian blue pigment indigo for painting with an initial black (nigrum) aspect that turns into purplish blue (purpurae caeruleique) when diluted with water. Also Vitruvius describes various methods to obtain black, one of which is gained from burning dregs of wine that can result in a blue pigment, indicum. what makes that black pigments have been used to produce various blue tones in painting. The use of the finer wines will allow us to imitate not only black but indigo. (Vitruvius, de architectura X, 4 [transl. Granger, F., 1962, 123]) The Romans thus not only used the words niger or atramentum to indicate ‗black‘, but also to mean ‗darkened‘ or 13 ‗darkener‘ like the meaning of the Greek word μέλας (Bruno 1977, 84). Furthermore the quotes by Vitruvius and 14 Pliny make clear that the terms for black in Latin have also been employed to refer to a blue hue. Vincent Bruno explains that there must have been no necessity for the Roman writers like for instance Pliny the Elder and Cicero to use a specific term to refer to blue as a pigment (Bruno 1977, 73-87). It is, of course, not our intention to evaluate here any further the significance of the presence and absence of specific colour terms in Latin and the use of colours in the Roman material culture. We therefore wish to refer to the various papers in the aforementioned proceedings of the Berlin and Glasgow conferences and to the work by Berlin and Kay, but also to recent work debating colours in Antiquity (Bruno 1977; Kurzrock (ed.) 1983; Berlin, Kay 1999; Cleland, Stears (eds.) 2004; Rouveret et al. (eds.) 2006). Hence, the linguistic and psychological issues are very complex themes that would bring us too far from the archaeological discussion, especially when encountering the specific difficulties and research challenges encompassing colour perception and colour categorisation. We therefore have to be conscious of the biased assumption on how to look at Roman glass hues due to the current determination of colours influenced by the scientific colour spectrum and digitisation (Hunter, Harold 1987) [see Chapter 9]. 1.4 Colour perception in archaeology Various classification systems have already been adopted to categorize Roman glass in the hope of establishing an objective method to quantify the finds from archaeological contexts. The system most widely adopted by ® others dealing with Roman archaeological glass is the PANTONE colour chart by Letraset, introduced by Beat Rütti when he considered the glass from Vitudurum (Rütti 1988, 13) and Augst/Kaiseraugst (Rütti 1991, 109; 344, Taf.220). Rütti discerned 76 colour variations to describe the Roman glass vessels from Augst/Kaiseraugst which can be grouped into seven main hues: 9 blues; 20 blue-greens; 26 greens; 5 yellows (of which 1 is orange); 7 15 browns; 6 purples (of which 1 is pink) and white (Rütti 1991, 110). These very detailed subdivisions, however, are primarily useful for establishing possible clusters that coincide with particular chemical compositions. The motivation behind over-subdividing glass hues is the need to determine regional and chronological sets of vessel shapes which have been produced with a distinct type of raw glass. Similarly Rupert Gebhard observed after studying the material from Manching (DE) that the 21 different tinges of blue glass used for the production of late La Tène arm rings correspond with specific chronological periods (Gebhard 1989b, 99-102). Other methods were proposed earlier. Ludwig Berber used in his work on the glass from Vindonissa (CH) a chart that was initially set up to describe pencil colours (Berger 1960, 96). Marianne Stern used the Munsell colour chart when examining various large glass collections (Stern 1977; Stern, Schlick-Nolte 1994; Stern 2001). AnnaBarbara Follmann-Schultz made use of the colour terms employed by stamp collectors (Follmann-Schultz 1988, 2). Despite these efforts, the usefulness of these colour charts is limited to opaque colours providing their hue by reflection of light. Most Roman glass, on the other hand, is transparent or translucent, meaning that the thicker the glass is, the more the intensity or value of the colour increases (see below). Hence, two vessels with different wall thicknesses but produced from glass of the same batch will show a colour shift. This explains the tendency not to name the variety in glass colours narrowly but rather by using a subjective though consequent system with broad 13 Niger stands for ‗dark‘ while atrantum stands for ‗black as soot‘ – respectively referring to a tinge and a hue. To describe the blackness of glass the ancient writers used obsianum, cf. ‗vitrum obsianum‘ (Pliny) - referring to the resemblance of pitch-black obsidian. 14 Vincent Bruno, using a passage in Aristotle‘s ‗de coloribus‘, explains in a very lucid way that this was also the case for the Greek word μέλας (Bruno 1977, 92-93). 15 It is surprising that black is prominently absent within Beat Rütti‘s colour chart on glass vessels of Augst/Kaiseraugst, while 5 pieces of a total of 5,067 or 0,1% have been designated as black (Rütti 1991, 109, tab.18). Furthermore, after checking the catalogue, 51 pieces described as ‗schwarz erscheinend‘ have been excluded from his colour list. When taking into account all excluded black-appearing pieces, the ratio opposed to all glass fragments recorded then increases 10 times up to 1%. This ratio is the same as that for the black glass from the cemetery of En Chaplix in Avenches (CH), containing 88 so-called black pieces out of 9,989 items. With 0,88% of the total, it consists of 23% of all the strongly-coloured objects, seeing that this group covers 4% of the total amount of glass objects (Martin-Pruvot 1999, 175, fig. 342). 3 descriptive terms as explained by Hilary Cool and Jennifer Price in their publication on the glass vessels from Colchester (UK) (Cool, Price 1995, 7-8). They distinguish four categories: 1) strongly-coloured glass; 2) lightly tinted glass; 3) colourless or decolourised glass; 4) blue-green or ‗naturally coloured‘ glass. The black glass fragments recorded for Colchester are described as strongly-coloured pieces appearing black, with the description mentioning the true hue of the glass unless not visible. This is also the case in the publications on the glass from Augst/Kaiseraugst (CH) mentioning ‗eine dicht gefärbte, meist olivgrüne oder braune Glasmasse‘ (Rütti 1991, 110) and Avenches (CH) stating ‗le ―noir‖, c‘est à dire du vert olive très foncé ou brun vert foncé à peine translucide d‘apparence noire opaque‘ (Bonnet Borel 1997; Martin-Pruvot 1999, 175-176, note 17). When looking more specifically to definitions given to black glass one immediately understands that the subjectivity on the perception of the colour black and the interpretation of its definitionin particular in relation with the transparency of glasscan easily induce confusion and thus inaccuracies in the colour designation. The wide-ranging concept of black glass and subjectivity towards its perception has been well described by Willy th th Van den Bossche when discussing glass bottles of the modern ages between the 16 and mid 19 century: (1) A popular term for a glass bottle colour of dark green, dark brown or any other colour that is so dark as to appear almost black. (2) A bottle term used since the 16th century for dark brown or dark green glass resulting from the inclusion of oxides of iron, manganese and sulphur combined with the use of a reducing furnace atmosphere (e.g. from coal, peat…). When increasing the melting temperature and the residence time the colour changes from brown to olive-green and light green. Black glass was first made in Lorraine, North-east France, and Southern Belgium in the early 16th century, where it is called ‗verre noir‘… (Van den Bossche 2001, 392) He for instance mentions a glass workshop in the north of France that was called ‗verrerie de verre noir‘ while only dark (olive) green to dark brown glass bottles were produced by the workshops around Darney (FR) (Van den Bossche 2001, 52; 310). Another definition concerns the study on early medieval Anglo-Saxon glass beads found in England (Guido 1999). Coloured glass is a complex subject as many different factors affect the final appearance of the glass. Among the most important are the impurities it contains (which may have been added accidentally or deliberately), the furnace conditions under which it was founded (such as the time for which it was heated, the temperature it reached and the furnace atmosphere, which can be oxidising or reducing) and its bulk composition. (Bayley 1999, 89) 1.5 Proposed application The perception of colours is an individual and thus extremely subjective observation with sometimes very arbitrary 16 and even biased judgments. In search for a definition to categorize black glass artefacts, it is important to realize what should be understood under colour. It is essential to recognize that our modern perception of colours has changed from that in Antiquity. Nowadays, colour is subdivided into three different aspects: 1) hue; 2) saturation; 3) value. Hue refers to the position of the colour in the colour spectrum, e.g., yellow, red, and blue [see Chapter 9]. Saturation is also called chroma and represents the purity of the colour. The higher the number, the purer the colour, with 0 for neutral grey. Finally, value stands for the intensity of a colour, referring to how light or dark the colour is from white (10) to black (0). Today we use a very elaborate categorisation method to label colours where the ‗hue‘ of a colour is essential. In Roman times, no classification system for colours was used (Brécoulaki 2006), and it appears as if then the ‗saturation‘ and the ‗value‘ of a colour were much more important. On the other hand, it is equally essential to realize that various reasons are the basis for why black glass material lay within a twilight zone: The techniques of production, inherently connected with the types of object, make researchers describe some functional types more easily as black (e.g., counters, bracelets, and beads) and others as deeply coloured (e.g., vessels, particularly those produced by the free-blown technique); 16 Under variable circumstances. the same observer might even attribute various colours to one and the same object (e.g., daylight vs. artificial light; morning vs. evening; sunny vs. cloudy; white light vs. yellow light). 4 The difference of perception among researchers dealing with black glass objects causes discrepancies in colour categorisation. Some researchers do not pay attention to the true colour and speak purely of black. Others, when perceiving the true colourbasically by using strong light in transmissionignore the black appearance of the object. A third group, however, combines both features by giving the true colour of the glass matrix and by mentioning that the artefact appears black. In literature, it can be observed that some vessel types are constantly regarded as black or black-appearing even if the true colour of the glass matrix is quite easily recognizable (Figure 2a-c), while others are barely considered black-appearing (Figure 2d-e). The salient detail that we noticed is that the first group consists of artefacts in a green glass, whereas those in blue, brown, and purple glass basically fall into the second group. Figure 2: a) Carchesium of Heel (NL); b+c) Bulbous cup of Heerlen (NL) (photographs by author, by courtesy of Limburgmuseum, Venlo); d+e) Bulbous jar (picture d) from Foy, Nenna 2001, 85, fig.93; picture e) from Harden et al. 1987, 111, fig.44) nd rd This means that the late 2 –early 3 century AD vessels in very dark green glass are easily accepted as blackappearing because everybody assumes that the purpose was to produce black-looking artefacts. More debate st and disagreement exists when, for instance, the famous cameo-glass vessels from the 1 century AD are taken into account. In a barely discernible very dark blue to violet glass, the cameo-glass vessels are, in fact, blackappearing objects, making the true colour only detectable with the help of a light source in transmission. Yet they are nearly always described as deep blue translucent glass (Harden et al. 1987, 54). However, the intention to contrast this deep blue glass with opaque white glass of the superimposed figurative scenes clearly supports the description of the deep blue glass as black-appearing. While this issue remains a matter of debate for the cameoglass, it is clear for the fragment of a foreleg of a life-size horse statue in the British Museum or the numerous black tesserae from Les Houis near Sainte Menehould [cat.no.1411], although they were made from blue, bluegreen, blue-grey, green and purple glass (Cosyns 2009, 93) [see Chapter 3]. The true colour of black glass tesserae is not recognized, particularly not when the tesserae are embedded within a mosaic. On the other hand, artists have been using the tesserae to show elements or aspects that are commonly deemed as black. 5 Furthermore, the life-size horse statue is always recognized as that of a black horse whether it is produced in 17 deeply purple glass or in obsidian [cat.no.2836] (Harden et al. 1987, 28, no. 6). Bichrome artefacts incorporating white opaque glass cause another dilemma. The reflection of light due to the underlying marvered white opaque glass makes the black glass become visible in such areas, as the counter from Cortil-Noirmont (BE) (Lefrancq 1989, B13, no. 30) [cat.no.228] or the globular cup from Heerlen (NL) [cat.no.2898] clearly demonstrate (Figure 3). When the true hue of the black-appearing glass is made visible, a translucent coloured glass can be observed. Considering these pieces as black, makes defining black glass objects difficult. Figure 3: a) Counter from Cortil-Noirmont (BE) (left part reconstructed); b) Globular cup from Heerlen (NL) (photographs by author, a) by courtesy of KMKG-MRAH, Brussels; b) by courtesy of RMO, Leiden) Besides the two possibilities of logical observations dividing these strongly-coloured glasses into a ‗black‘ group and a coloured group, it is important to mention the black appearance of the glass objects even if the colour matrix is observable. However, Justine Bayley says that opaque black glass which is obtained by adding iron to the batch is, in fact, neither opaque nor black (Bayley 1999, 90). This pseudo-opacity is best demonstrated by the large quantities of glass counters in various poorly translucent colours. Sometimes the true hue becomes only visible after the flaking of a chip, as with the black glass counter from Tienen (BE) [cat.no.531] (Cosyns et al. 2006b, fig.1b). The observed hue, with the use of a strong light in transmission, shows that sometimes counters combining various translucent glass colours have been produced as demonstrated with for instance the two pieces from Nijmegen (NL) and Oudenburg (BE) (Figure 4a-b) [cat.nos.4466 and 465]. Figure 4: a) Black glass counter from Nijmegen (NL) with light in reflection (left) and light in transmission (right); b) Black glass counter from Oudenburg (BE) exposed to white light in transmission (photographs by author, a) by courtesy of GAD, Nijmegen; b) by courtesy of VIOE, Brussels) To cut this Gordian knot and bring to a close the discussion, we define Roman black glass when it is perceived as black by the naked eye as well as when it appears black when placed on a white paper and only light in reflection is used. Yet we are aware that the study of colours involves more than defining colour categorisation by outlining colour terms and colour perception. It is also important to discuss, even though only briefly, the use of black glass in view of colour symbolism [see Chapter 7]. 17 A debate exists between those who regard the piece as being made from glass (Harden 1987, 28, no. 6) and those who claim the piece is, in fact, made of obsidian (Grose 1989, 342). At the latest AFAV-meeting in Fréjus (November 2009), Jennifer Price informed us that unpublished analysis results pointed to obsidian. 6 1.6 Initial observations A more careful look at the 4475 inventoried entries demonstrates the mass of black glass artefacts can be 18 confined within seven distinctive colour groupsblue, blue-green, brown, green, grey, purple and red. However, the opacity of the glasscaused by the intensity of the colouring agents within the glass matrix [see Chapters 8 and 9]makes it not always possible to identify the glass hue of black glass artefacts within first view (Figure 5; 19 Table 2). Table 2: Overview of the different colours of Roman black glass colours appearance true hue quantity unknown unknown 9 black unknown 3175 black blue 49 black purple 123 black brown 72 black green 873 black blue-green 143 black red 26 black grey 5 4475 % 0,2 71,0 1,1 2,7 1,6 19,5 3,2 0,6 0,1 100 colours of Roman black glass 1% 3% 2% black unknown black blue 19% black purple 3% 4% black brown black green black blue-green 71% 1% 0% black red black grey Figure 5: Pie-chart with different hues of Roman black appearing glass artefacts 18 Within the chapters 8 and 9respectively on ‗chemical analysis‘ and ‗photonics‘we elaborate on the archaeological relevance concerning the use on the different glass hues. 19 A more thorough examination of the material will support the identification of the glass hue either by electronic binocular or after pinching of a flake from the artefact. 7 colours of Roman black glass black blue black purple 68% 11% 13% black brown black green black blue-green black red 9% 2% 0% 4% black grey 6% Figure 6: Pie-chart with different hues of Roman black appearing glass artefacts (excluding the unidentified group) When excluding the black glass artefacts with unidentified hue it is clear that the green tinted artefacts are most frequent with more than 2/3 of the recognized items (Figure 6). This group includes a wide range of different greens without distinction. The second largest group, blue-green, is connected to the green groupas will be discussed in the here aforementioned chapterscovering together 79% of all black appearing glass artefacts. The two remaining groups of some importance are purple (9%) and brown (6%). The reason why black appearing artefacts in blue glass remain marginal is partly due to the easiness of discerning the blue hue as has been made clear previously, but also because most vessels in black appearing blue glasslike for instance the cameo glass vesselshave been systematically excluded from this project [see introduction]. The artefacts in so-called red glass seem to have this colour superficially due to an extreme oxidation during the manufacturing process and probably need to be incorporated in the green group [see Chapters 8 and 9]. Nearly all red glass entries originate from the late Roman glass workshop at Les Houis nearby Sainte Menehould (FR) [cat.nos.1415; 1428; 1429; 1431; 1433; 1486; 1535; 1539; 1562; 1576; 1586; 1589-1592; 1594; 1596-1599; 1601; 1603] (Cosyns 2009, 10, fig.5). The amount of artefacts in grey glass is negligible but may cover important information on recycling by adding obsidian to the batch (Cagno et al. 2010) [see Chapter 8]. Another issue considered within the following chapters is the use of applied decoration in one or more contrasting 20 colours (Table 3). The majority of black appearing glass artefacts are monochrome and remain ‗undecorated‘, but the number of entries decorated with a contrasting colour is not insignificant (Figure 7a). Most decorated artefacts are bichrome (88%) meaning artefacts with only one extra colour contrasting with the black appearing core (Figure 7b-c). Table 3: Categorization of decorated artefacts (n = quantity) colour % colour n n unknown 9 0,2 unknown 9 undecorated 3845 85,9 monochrome 3845 decorated 621 13,9 bichrome 541 polychrome 80 4475 100 TOTAL 4475 TOTAL % 0,2 85,9 12,1 1,7 100 decorated glass black core + 1 colour black core + 2 colours black core + 3 colours black core + 4 colours TOTAL n 541 64 14 2 621 % 87,1 10,3 2,3 0,3 100 20 The here discussed numbers do not consider the decorated monochrome artefacts by tooling, casting or pressing. Also excluded here are the artefacts with applied decoration in black glass similar to the core. 8 coloration of Roman black glass coloration of Roman black glass 2% 14% 12% monochrome undecorated bichrome decorated polychrome 86% 86% decorated Roman black glass black core + 1 colour 10% 3% 2% black core + 2 colours black core + 3 colours 87% black core + 4 colours 1% Figure 7a-c: Pie-chart with the ratios between undecorated vs. decorated black appearing glass artefacts (excluding the unidentified group) Figures 8a-b and Table 4 do not provide additional information of the over-all impression of monochrome artefacts as it is clear that the ratios of the range of glass hues applied is nearly identical to that of the general overview in Figures 5-6 and Table 2. Conversely a quantification of all black artefacts with coloured decoration can give a clearer idea on the use of the glass hue applied and its combinations (Figure 8c-d; Table 4). About 2/3 (62,1%) of the applied decoration on bichrome artefacts is done with blue glass (including pale blue, turquoise and ultramarine blue) (Figure 8c; Table 4). Together with white, yellow and redthe other most popular applied colours on black appearing glass artefacts, correspondingly 20,5% (1/5), 7,0% and 2,6%the group of blue tinted glass encompasses nearly all bichrome artefacts. Black glass artefacts show in only very few occasions a decoration in greenish yellow, green, brown or guilded. When looking to the polychrome material it is clear that the four most popular hues display all possible combinations when combined in pairs. Polychrome artefacts combining three or more colours show to be of negligeable quantity but white and blue are the most applied glass hues, respectively with seven and six combinations out of a total of eight. Further verifications of the database reveals particular colour combinations of applied glass trails on a specific hue of the core (Tables 5-6). The overview of the combinations is incomplete seeing the large amount of entries listed in Table 4 of which the glass hue of the core remains unidentified. The listing of specific colour combinations between core and applied decoration in the Tables 5 and 6 might be evidence for possible workshop practices and demonstrate the necessity to check the various colours used in Roman glass workshops vis-à-vis applied technology, commodity type and shape, and the various periods within a five century lasting era. The use of opaque white is predominantly applied and not related to any particular core glass hue. This occurrence seems very logical seeing that a black-and-white contrast is the most coherent contrast. The four major glass hues of applied decorationblue, white, yellow, red - are present on green glass artefacts; yellow is absent on blue and blue-green glass objects as well as on brown pieces; blue is absent on brown, purple and red glass. Only artefacts with a core in brown or green glass have received a polychrome decoration and always demonstrate combinations with white. The observation that brown core glass is known combining white with yellow illustrates bichrome artefacts with a brown core and a yellow decoration might have existed. 9 Table 4: List of colours and colour combinations (n = quantity) monochrome % monochrome n (excluding undefined hue) undefined n % bichrome + 1 colour % unknown 36 6,7 45 3,6 blue 336 62,1 n 2607 67,8 45 1,2 blue brown 67 1,7 brown 67 5,4 greenish yellow 1 0,2 purple 102 2,7 purple 102 8,2 red 14 2,6 blue-green 142 3,7 blue-green 142 11,5 white 111 20,5 red 26 0,7 red 26 2,1 yellow 38 7,0 grey 3 0,1 grey 3 0,2 green 3 0,6 853 22,2 green 1 0,2 1 0,2 541 100,1 n % 1 50,0 1 50,0 2 100,0 blue green 853 68,9 guilded brown 3845 100,1 n % blue/red 28 43,8 white/blue/pink 1 blue-green/red 1 1,6 white/blue/red 4 7,1 white/blue/ 28,6 red/green white/yellow/blue/red white/yellow 4 6,2 3 21,4 white/blue 9 14,1 white/yellow/ blue white/yellow/ 2 14,3 3 21,4 1 7,1 14 99,9 TOTAL polychrome + 2 colours polychrome + 3 colours white/red 15 23,4 green white/yellow/red white/green 1 1,6 yellow/blue/red yellow/blue 3 4,7 yellow/blue-green 1 1,6 yellow/red 1 1,6 yellow/green 1 1,6 TOTAL 64 100,2 1238 99,9 n % monochrome black glass polychrome + 4 colours monochrome black glass 4% 5% undefined blue blue brown 0% brown 22% 8% 68% red 2% 69% grey 0% green 1% 0% unknown 7% 7% blue 1% 2% 5% blue/red blue-green/red white/yellow red white yellow white/blue 44% 23% yellow/blue guilded brown white/red white/green green 62% grey polychrome black glass (2 colours) greenish yellow 20% red green 2% 2% 0% blue-green 12% blue-green bichrome black glass 3% 0% purple purple 1% 2% 4% 2% 1% yellow/blue-green 14% 6% yellow/red 1% yellow/green Figure 8a-d: Pie-charts defining the ratios of glass hues utilized for the core of the artefact (a + b) and for the applied decoration (c + d) 10 Table 5: Overview of colour combinations between core and decoration of bichrome artefacts hue of the black appearing artefacts hue(s) of the applied decoration blue blue; white blue-green blue brown white green blue; white; red; yellow grey white purple white; yellow red Table 6: Overview of colour combinations between core and decoration of polychrome artefacts hue of the black appearing artefacts hue(s) of the applied decoration blue blue-green brown white/yellow green white/blue; white/red grey purple red - 1.7 Conclusion Summarizing the relevance of these colour studies to the discussion on Roman black glass, we have to highlight the complexity of the colour perception of glass in general. This is attributable to the heterogeneous visual interpretation of the material and to the variety of ways to categorize Roman glass artefacts. Hence, it has been essential to formulate a reliable method for the visual evaluation of colours. We explained our definition of black glass and the basis of our decision to include a category of ‗black‘ glass artefacts elsewhere described as deeply coloured glass. The individual approach of colour, in casu black, is one of the main reasons why the research on Roman black glass presented here could never develop into an exhaustive corpus. The linguistic and psychological studies on colours are focused on the understanding of colour perception by the human being and the use of colour semantics within a linguistic group. However, the development of an unbiased categorisation system is critical to enable comparative studies on glass assemblages on various levels: 1) between successive excavations on the same site; 2) between consecutive periods on the same site; 3) between a range of sites of the same period. Furthermore, each glass hue is related to a particular chemical composition as well as to specific furnace conditions (Biek, Bayley 1979, 14; Henderson 1985; Brill 1988; Bayley 1999). The listings of specific glass hues and colour combinations between core and applied decoration, which is likely to be linked with particular workshop practices, will be demonstrated in possible relation to technology [see Chapter 2], typology [see Chapter 3], chronology [see Chapter 4] and distribution [see Chapter 6]. Consequently, a colour categorisation can be valuable to distinguish different workshop assemblages [see Chapters 8 and 9] because the recognized workshops can be related to regional customs (tradition) or believes (ritual). In Chapter 7 we discuss the use of colours and their meaning in Roman society with a special focus on the ‗colour‘ black. A pertinent question that remains is whether an unbiased method exists to categorize black glass or, more precisely, deeply coloured glass that appears black in order to enable an impartial quantitative study. To verify how to categorize the Roman black glass artefacts, we carried out optical analysis using absorption spectroscopy; making available quantifiable data for comparative studies [see Chapter 9]. We defined Roman strongly-coloured glass as black: 1) when no colour is observable but the glass remains black even illuminated by a light source in transmission, as well as 2) when it appears black to the naked eye when placed on a white paper and illuminated only by light in reflection. 11 12 Chapter 2 TECHNOLOGY 2.1 Introduction In this chapter we focussed on the glass technologies which were used in the production of black glass artefacts during the entire Roman imperial period. We prescribed this focus primarily because this aspect is here utilized as the basis for establishing a classification of glass artefacts, thus enabling better archaeological processing. Another reason is that this facet is essential to better oversee and understand the various glass assemblages to recognize possible chronological and geographical idiosyncrasies. The intention was by no means to bring a full discussion on Ancient glass technology, but neither was it to provide a simple synopsis. Hence, the information formulated herewhich can be partly seen as a glossarywill rather be an endorsement of what has become general knowledge. However, we assume that within the limits of this overview, the chapter includes information that supplements the current knowledge on various aspects of Roman glass production which will adjust some prevailing assumptions. We based the terminology adopted here on only a few of the very many publications covering technological 21 aspects in Roman glass production. For instance, ‗Glass of the Caesars‘ (Harden et al. 1987) provides a comprehensive description of the variety of techniques used to produce glass in Roman times. In addition, Marianne Stern gives comprehensive outlines on technical terms related to vessel glass production on various occasions (Stern 1995, 19-44; 1999; 2001). In their handbook on Romano-British glass vessels, Jennifer Price and Sally Cottam provide valuable information about manufacturing, characteristic vessel parts, and decoration techniques (Price, Cottam 1998, 10-39). Maud Spaer‘s impressive work includes technological aspects concerning glass jewellery and other small glass finds (Spaer 2001). This has been combined with information by Birte Brugmann, who studies Anglo-Saxon glass beads (Brugmann 2004). Because these publications give very comprehensive descriptions and illustrations, we only aimed in this chapter to provide information and explanation on the technological aspects to better understand the production techniques employed by the Romans for the manufacturing of the various types of artefacts in black glass within an evolution period of five centuries. The challenge of this chapter is to provide particular chronological and/or geographical boundaries helpful in dating and/or in establishing the distribution pattern of these various artefact types‘ technological features. 2.2 The production Seeing that the production of black glass was not different from that of glass in any other hue, the same categories of archaeological evidence have to be considered as those that are recognized when examining the glass production in Roman imperial times and glass working in general: raw materials, furnaces, crucibles, tools, and glass waste from the production process. 2.2.1 The raw materials st nd Ancient writers in the 1 and early 2 centuries ADe.g. Strabo (64 BC-20 AD), Pliny the Elder (23/24-79 AD), Flavus Josephus (37/38-post 100 AD), Tacitus (c.55-115/120 AD) reported on the raw materials necessary to 22 produce raw glass and explained the location of sand suitable for making glass: Then one comes to Ptolemais, a large city, in earlier times named Ake [Akko]… Between Ake and Tyre there is a sandy beach, which produces the sand used in making glass. Now the sand, it is said, is not fused here, but is carried to Sidon and there melted and cast. Some say that the Sidonian, among others, have the glass-sand that is adapted to fusing, though others say that any sand anywhere can be fused. I heard at Alexandria from the glass-workers that there was in Aegypt a kind of vitreous earth without which many-coloured and costly designs could not be executed, just as elsewhere different countries require different mixtures; and at Rome, also, it is said that many discoveries are made both for 21 For practical reasons, we focused on publications in English, but we are well aware of the existence of equal standard works in, for instance, French, German, Greek, and Italian. 22 More references on glass and glass working in Ancient literature are available in Trowbridge 1930; Humphrey et al. 1998, 375-380; Stern 1999, Whitehouse 1999; Stern 2007. 13 producing the colours and for facility in manufacture, as, for example, in the case of glass ware, where one can buy a glass beaker or drinking-cup for a copper [obol]. (Strabo, Geographica XVI, 2.25 [transl. Jones, H.L., 1961, Strabo VII, 271-273]) That part of Syria which is known as Phoenicia and borders on Judea contains a swamp called Candebia amid the lower slopes of Mount Carmel. This is supposed to be the source of the River Belus, which after traversing a distance of 5 miles flows into the sea near the colony of Ptolemais. … The river is muddy and flows in a deep channel, revealing its sands only when the tide ebbs. For it is not until they have been tossed by the waves and cleansed of impurities that they glisten. Moreover, it is only at that moment, when they are thought to be affected by the sharp, astringent properties of the brine, that they become fit for use. The beach stretches for not more than half a mile, and yet for many centuries the production of glass depended on this area alone. … Now, however, in Italy too a white sand which forms in the River Volturno is found along 6 miles of the seashore between Cuma and Literno where it is the finest grained, and reduced to powder in a mortar or mill. Then it is mixed with three parts of soda, either by weight or by measure, and after being fused is taken in its molten state to other furnaces. There it forms a lump known in Greek as sand-soda [hammonitrum]. This is again melted and forms a lump of clear colourless glass. Nowadays sand is similarly blended also in the Gallic and Spanish provinces. (Pliny the Elder, Naturalis Historia XXXVI, LXV.190-194 [transl. Eichholz, D.E., 1962, Pliny X, 149-155]) At a distance of about two furlongs [stadia] from the town [Ptolemais] runs the diminutive river Beleus. On its bank stands the tomb of Memnon, and close to it is a very remarkable region, a hundred cubits in extent. It consists of a circular basin which produces vitreous sand. Numerous boats put in to this spot and empty the basin of its sand, whereupon it is filled up again by the action of the winds, which, as if by design, drift into it the common sand outside, the latter being all promptly converted by this mine into vitreous matter. But the phenomenon which, to my mind, is even more remarkable, is that the excess particles of glass which overflow from the cavity become ordinary sand as before. Such are the curious properties of this spot. (Flavus Josephus, Bello Iudaeica II, 189-191 [transl. Thackeray, H.S.J., 1967, 397]) The river Belus also empties into the Jewish Sea; around its mouth a kind of sand is gathered, which when mixed with soda is fused into glass. The beach is of moderate size, but it furnishes an inexhaustible supply. (Tacitus, Historiae V, 7 [transl. Moore, C.H., 19989, Tacitus III, 187-189]) Not having the intention to start a textual analysis between the available ancient sources, it is clear that most Roman writers remain rather brief and that they, most likely, got their information from a thus far unknown older source. The location of the areas discussed by the ancient writers where the Romans extracted the sand is still a 23 matter of discussion. However, Pliny the Elder also talks about the extraction of sand in Italy to produce glass, and mentions similar activities in Spain and France (if we consider it equal to Gaul). Possibly in Classical and st Hellenistic times the Belus sand dominated the glass production, while in the second half of the 1 century AD new sand sources suitable for the production of glass were discovered all over the Roman Empire. Sand provided the necessary silica for the creation of a glassy substance, although quartz pebbles or obsidian also provided silica. Silica from quite pure sand was the basic network former of Roman glass, but a flux or network modifier was required to lower considerably the melting point of the silica. This was necessary because the melting point 24 of sand in Roman times was too high to be technically attainable and subsequently not profitable. The standard flux used in Roman times was soda, a mineral alkali extracted from natron and trona, although an alternative could be supplied through vegetable alkali such as potash from plant ash. Egyptian natron and trona from Wadi el-Natrun in the western Nile Delta between Alexandria and Cairo were the classic soda-providers in Roman th times and this up to about the 8 century AD (Décobert 2003; Nenna 2008a). A third essential element in Roman glass is lime which is used as a stabilizing agent against water infiltration. Strangely the deliberate adding of a stabilizer to the batch such as lime has never been mentioned by ancient writers and consequently its presence is interpreted as a coincidental adding through the use of calcareous sand. On the whole it is usually acknowledged that black glass is obtained by adding a strong colorant to the batch such as iron (green, brown), manganese (purple) or cobalt (blue). How black glass is made in Roman times and how we have to look at its production and distribution is fully discussed in the appropriate chapters [see Chapters 6 and 8]. In Chapter 8 is demonstrated via chemical analysis that both ‗naturally coloured‘ and decolourized raw glass have been used to produce black glass. To decolourize the ‗naturally coloured‘ raw glass antimony oxide and manganese oxide are known to have been added to the batch in Roman times. Also discussed more in detail within Chapter 8 is the regional and chronological variability of applied oxides responsible for the black hue of ‗naturally coloured‘ or decolourized raw glass, and in what kind of concentrations these colorants were added. In 23 We can refer to recent and on-going research focussing on the location and characterisation of these sands: Israel (Brill 1988 264-269; Freestone, Gorin-Rosen, Hughes 2000), Italy (Silvestri et al. 2006; 2008), and Spain (personal communication Degryse). It is generally acknowledged that the Belus River is to be located in the north of present day Israel between Haifa and Akko, and more exactly in the area of modern Nahr Na‘man (Thackeray 1967, 397, note a). Question is whether Brill‘s samples from the Belus river bedding were taken at a correct location, if it can be determined where sand was extracted during the Roman imperial period. 24 Pure silica has a melting point of 1410 °C (periodical table of Mendeleyev). Furthermore, experimental archaeology has pointed out that heating up to c. 1000 °C is quite easy, but once above it the amount of combustion and necessary time to acquire higher temperatures increase exponentially (Nicholson, Jackson 2000, 19) [see also Chapter 8]. 14 addition it will be made clear that the commonly available glass used for production of black glass artefacts during the Roman imperial period was a silica-soda-lime glass. Accordingly recognized characteristic compositions such 25 as the late Roman so-called HIMT glass are important to develop the trade and distribution patterns on the various types of black glass artefacts. 2.2.2 The workshops The production of glass in Roman times involves two types of workshops. Firstly there are the primary glass workshops – mainly located in the south-eastern Mediterranean [see Chapter 6] – where raw glass was produced from raw materials. In a second stage, the chunks of raw glass were distributed all over the Roman Empire to 26 secondary glass workshops, where artefacts were produced by re-melting the raw glass together with cullet. Unfortunately, unambiguous artefactual evidence is generally lacking where the production of black glass artefacts is concerned. In most cases, the furnaces have been largely destroyed over the course of time or on purpose when the area changed use. On the other hand, evidence of Roman glass production is difficult to identify and to interpret when present. This is perfectly described through a well-reasoned comparison on the availability and degree of fragmentation of glass vessels in various contexts: There is a marked contrast between the wide range of vessels and objects, which survive complete in burials or, more frequently, are found in fragments on settlements, and the ephemeral and episodic nature of the evidence for the production processes by which they were formed. (Price 2002, 81) In recent times Marie-Dominique Nenna stated that archaeological evidence for Roman primary glass workshops is nearly non-existent (Nenna 2008a; Nenna 2008b, 61-62), and contrasts greatly with the massive quantity of sites where secondary glass workshops have been found (Foy, Sennequier 1991; Sternini 1995; Amrein 2001; Foy, Nenna 2003, 40-60; Nenna 2008a, 132-136). An ensuing assumption is that the production of glass artefacts could have taken place in nearly any settlement throughout the entire imperial period. In relation to this concern it is worth mentioning Jean Andreau who most recently drew attention to the difficulty in determining the distribution patterns of glass vessels and to discern the trade routes that result due to the very high homogeneity all over the Roman Empire (Andreau 2010, 114-115). Les formes de tous ces objets sont très homogènes d‘une région à l‘autre et même d‘une moitié de la Méditerranée à l‘autre. (Andreau 2010, 114) We have to stress that all raw glass hitherto excavated at primary glass workshop is ‗naturally coloured‘ or decolourized and that deeply coloured raw glass including black appearing raw glass has not been retrieved yet on such sites. Conversely, various secondary workshops yielded pre-fabricated lumps of deeply coloured glass similar in shape and size to the rough blocks of raw glass found in the shipload of shipwrecks all over the Mediterranean (Foy, Nenna 2001, 24-39; Nenna 2008a, 130-131). The homogeneous chemical composition of 16 sampled chunks of differently coloured raw glass from the glass workshops of La Montée de la Butte at Lyon (FR) (Robin 2008, 43) demonstrates that the chunks are all made of the same type of raw glass which must have been imported [see Chapter 8]. Although a number of secondary workshops contain chunks and/or cullet of deeply coloured glass, none of the known workshops hold available evidences to testify the colouring of imported decolourized or ‗naturally coloured‘ raw glass was done locally. In view of that, it seems excluded that the rough 27 blocks of black glass were locally produced in a secondary workshop. Where these glass blocks were produced 28 is uncertain. To this point it was impossible to determine where and how the black colouring of ‗naturally coloured‘ and decolourized raw glass has taken place. As mentioned in the previous paragraph the chemical analyses, discussed in Chapter 8, will help to broaden our understanding on the regional and chronological variability and to increase our knowledge on the distribution patterns. In Chapter 6 is verified how from the archaeological data we can discern distribution and consumption patterns on the various types of artefacts in 25 HIMT [High Iron Manganese Titanium]: HIMT-glass is produced by using less pure sand with much higher concentrations of iron oxide, manganese oxide and titanium oxide, present as impurities [see Chapter 8]. 26 Cullet refers to recycled broken glass vessels as well as production waste such as knock-offs, moils, and misshapen artefacts to lower the melting temperature of the raw glass. Cullet only becomes a significant alternative to raw glass from the medieval period onwards as can be seen in the Serçe Limanı shipwreck (Bass et al. 2009). 27 In Chapter 8 are discussed the results from the chemical analysis on the irregular chunks of black glass from the glass workshops of ‗La Montée de la Butte‘ at Lyon [cat.no.1224-1225] (Robin 2008) and of Les Houis nearby Sainte Menehould (FR) [cat.no.1570] (Cosyns 2009) secondary workshops respectively dated mid 1st–early 2nd century AD and 3rd–5th century AD. 28 In Chapter 6 are considered various models on the organization of glass production in the Roman Empire and evaluated against the available information from the black glass data. In Chapter 10 the archaeological and archaeometric evidences are put side by side to discuss the organization of black glass production and consumption in the Roman Empire. 15 black glass. In addition we verified in Chapter 6 whether these blocks were imported from primary workshops or from specialized intermediate workshops in order to supply the local market? The secondary workshops known to have processed black glass are considered more in detail in Chapter 6 when discussing the production of black glass. Important to take into account is that secondary workshops never processed solely black glass, on the contrary it only formed a (minor) part of the assemblage. For this reason it is impossible to verify in what proportion a workshop processed glass in various hues. Nearly all secondary workshops are located in an urban centre even though sometimes in its periphery. The exception is the workshop of Les Houis nearby Sainte Menehould (FR) that is a prominent rural production centre within the then very dynamic Argonne region. Because evidence of clear structures such as furnaces is quasi nil we, however, identified the presence of workshops on the basis of characteristic distribution patterns [see Chapter 6]. 2.2.3 Built-in basins and crucibles In secondary workshops, the crushed raw glass chunks were melted in built-in basins or in crucibles. It is impossible to discuss the matter here in detail as this topic is a research on its own (see for further reading Sternini 1995, 73-82; Fischer 2009, 92-106). Generally speaking the built-in basins can be considered st nd characteristic for the early imperial period (1 and 2 centuries AD), whereas the crucibles were typical from the st late Roman period even though crucibles have been reported on 1 century sites too (Foy, Nenna 2001, 64). In relation to the workshops where black glass was utilized, the workshop ‗Äussere Reben‘ from Kaiseraugst (CH), nd rd dated mid 2 –early 3 century AD, yielded crucibles likewise the workshop ‗Palais Kesselstatt‘ at Trier (DE) th which is dated second half 4 century AD [see Chapter 6]. The built-in basins have never been the issue of a systematic study, although a few examples have been excavated, such as those at ‗La Manutention no. 1‘ at Lyon (FR) (Foy, Nenna 2001, 42-43), and at the Legionary camp at Bonn (DE) (Follmann-Schulz 1991, 36-37). None of the known built-in basins contained traces of black glass processing. The number of fragments of red fired clay with black glass on one side yielded at ‗Äussere Reben‘ workshop in Kaiseraugst [cat.nos.3688-3740] (Fischer 2009) might be considered as remains of (a) builtin basin(s), but can equally be the remnants of the furnace opening where the glassblower spilled glass while gathering his gob of hot glass from the molten pool whether it be a crucible or a built-in basin [see Chapter 6]. Numerous crucibles are known and are consistently thick-walled open vessels with a glass layer inside the vessel. The pottery used as crucibles was, in fact, ordinary cooking ware made from refractory clay so that these receptacles could stand the very high temperatures (Bayley 1992). Some late Roman examples from Trier (DE) show antique repairs of cracked walls by inserting a thick layer of pumice [cat.nos.2069-2070; 2073] [see Chapter 6]. The use of plain wares results in a wide variety of shapes due to the regionalism of the cooking ware production [see Chapter 6]. Hence, the pottery shapes of the crucibles can help determining the date of the workshop [see Chapter 4]. 2.2.4 The tools The tools used to manufacture black glass artefacts appear to be similar to those used to produce artefacts in other glass hues, and can thus be divided into the traditional categories: moulds and stamps; blowing pipe and pontil rod; pincers and shears; and marver. 2.2.4.1 Moulds and stamps The difference between a mould and a stamp is that the latter is pushed-in into the glass, whereas the former shapes the glass by pouring or blowing the hot gob of glass into the mould shape. In the production of glass vessel the application of two different production techniquescasting and mouldblowingmade use of moulds [see below page 22]. The recognized mouldseither open or closed and consisting of one piece or comprising multiple sectionscan usually be attached to either casting or mouldblowing. Casting was done in late Republic and early imperial times by using open moulds in one single piece st [see Chapter 6]. This assertion is attested by a number of characteristic features on the 1 century AD cast vessels in monochrome deeply coloured glass: 1) small and/or shallow shapes; 2) plain and uncomplicated shapes; 3) an absence of any seam. Mould-blowing, on the other hand, was performed in Roman times by using 16 closed moulds considering the mould-blown vessels are high and have a closed shape. The moulds employed for mould-blowing consisted, in contrast to those used for casting, of two or more sections. The number of mould 29 segments used can be determined from the presence of ridges on the outer surface of the mould-blown object. A seam appears when inflating the hot gob of glass and forming a small space between the different segments. Roman glassworkers appear to have adopted black glass infrequently in the production of mould-made vessels [see Chapter 3]. The use of stamps in the production of glass vessels is limited to the creation of medallions to decorate some specific vessel shapes in imitation of metal ware [see Chapters 3]. Stamps were mainly used in the production of glass jewellery to mould the decoration swiftly. Seeing the smallness of some types of jewellery the use of moulds cannot be excluded in the production of jewellery in black glass [see Chapter 3]. 2.2.4.2 Blowing pipe and pontil rod a. b. c. Figure 9: a) Irregular circular mark from a massive punty gob; b) Thick ring-shaped mark from a hollow punty gob or ring pontil; c) Thin irregular annular scar from the moil of the blowing pipe (photographs by author) Both blowing pipes and pontil rods are related to glass blowing and more particularly the blowing of vessels. With a blowing pipe, the glassmaker gathers the hot glass from the crucibles in the furnace to blow a vessel. To finish the object, a pontil rod or punty is used. The massive iron rod is preheated before being dipped in the hot glass. When the vessel was cracked off from the blowing pipe the gob of the punty got attached to the base to finishing off the vessel by rendering the rim and eventual handle(s) as well as to apply glass trailed decoration. When the vessel is finished, it is knocked off from the punty and placed in the annealing chamber. Interestingly, knocking off the object from the pontil leaves a scar on the bottom surface. It not only can be used to estimate the diameter of the pontil rod but also determine the type of pontil. On the basis of the scar size and shape it can be recognized in the first place whether a pontil rod or a blowing pipe was used and secondly how the glass end on the pontil rod was shaped (Figure 9a-b). The pontil mark is a distinctive feature that can help determining regional and/or 29 When the mould-blown shape has angles as, for example, the square bottles no seam is observable, even though five mould segments were used (four sides and one base). Sometimes the type of mould results in a clear seam like for instance on the grape-shaped amphoriskoi [see Chapter 3]. 17 chronological idiosyncrasies. This issue is discussed more in detail when examining some specific vessel shapes to establish the fingerprint of a glass workshop and glassblower [see Chapter 6]. Roman sites seldom yield glass working tools, though a 380 mm long iron blowing pipe fragment with a maximum diameter of 16 mm was retrieved at the glass workshop at Les Houis near Sainte Menehould (Chew 1989, no. 36; rd th Foy, Nenna 2001, 77, no. 59), a 3 5 century AD glass house where black glass objects were produced [see Chapter 6]. It consists of an ordinary tube made by enfolding a 3 to 4 mm thick sheet of iron. The full length and variety in diameters of the Roman blowing pipes as well as of the pontil rods are unknown due to the lack of evidence. 2.2.4.3 Pincers and shears Various jacks (types of pincers) and shears were used during the modelling of an artefact under productionwhether it was a vessel, bracelet, finger ring, bead or pendantand in applying a decorative design onto the object. The use of jacks in the production of vessels helped in modelling the foot, handles and mouth, or to apply a constriction at the neck for instance. From several pieces of production waste it is clear that jacks of various size and shape must have been used in Roman glass workshops. Some glass waste with pincer imprints are very pointed and narrow, but others appear to be wider. a. b. c. d Figure 10a-d: Various uses of jacks and pincers a) to model a base-ring [left: cat.no.595; right: cat.no.414]; b) to apply glass trails [Kaiseraugst - cat.no.3721]; c) to model handles [Liberchies – top: cat.no.413 ; bottom: cat.no.407]; d) reconstruction of how pincers were utilized to apply glass trails on an artefact (drawings and pictures by author, by courtesy of PAM, Velzeke; Musée Archéologique, Nivelles; Augusta Raurica, Augst; CRAN, Louvain-la-Neuve) 18 Traces of pincers used for shaping the foot sometimes remained visible, such as on the base fragments of carchesia from Velzeke (BE) [cat.no.595] and Liberchies (BE) [cat.no.414] (Figure 10a). The oblique incisions are the result of a clockwise rotation of the blowing-pipe when modelling the base-ring in a smooth movement. Two pieces of production waste from the Kaiseraugst workshop [cat.nos.3712; 3721] excavated between 1978 and 1980 are very probable what remained after applying a trailed decoration on vessels (Figure 10d) as regularly 30 attested on both extremities of the body of globular cups (Form IIB4-5) and finally ending up as waste. The Velzeke piece, measuring 43 mm in length and 20 mm in width, shows the print of a wide, pointed pincer over a length of 32 mm, measuring 8,5 mm at the edge and tapering towards a rounded point of about 4 mm width (Figure 10b). The trail-end of c. 3,5 mm diameter is the remnant of its drawn-out/stretched application giving the idea of the thickness of the applied trails. Pincers were also used to model the handles; this is for instance clearly observable on the two handle fragments of a skyphos from Liberchies (BE) [cat.nos.407; 413] (Figure 10c). The importance of pincers and shears in the production of arm rings is inherently connected to the rod-formed rd th technique employed during the 3 and 4 century AD. The numerous types of open bracelets with ends tooled into various shapes are the optimum effect of using pincers in the production [see Chapter 3]. 2.2.4.4 Marver A marver is a sort of working table on which the inflated hot glass or gather, also called paraison, can be rolled into shape during the blowing process. Another function is to roll-in colouring elements were placed on the marver, such as specks, or directly applied to the object, such as applied trails. A flat stone slab, for instance in marble, was most likely used in Roman times, but one in metal is not excluded. When the artefact is anymore reheated in the furnace the thermal shock the hot glass undergoes when touching the colder surface result in characteristic chill marks. A metallic surface results in irregular concentric chill marks (Figure 17), while a stone marver leaves a pitted surface on the glass object, as can be seen on the exterior of square bottles (Figure 14). Marvering could also have been done by using a wooden tool. A wooden marver is most likely used to shape the small polygonal bottles seeing the irregular shape of each side and the smooth surface. A wooden tool was possibly also used to marver the decoration on the late Roman counters with polychrome decoration [see Chapter 3]. 2.2.5 The production waste and recycling The characteristic production waste included: 1) moils: the glass parts between the blowing pipe and the cracked-off vessel detached from the blowing pipe 2) knock-offs: the glass remaining on the blowing pipe or pontil rod from the blowing pipe or the pontil rod. In Roman times, all this waste material together with the malformed objects were usually recycled as cullet, 31 leaving few indicative bits and pieces for the archaeologists. Accordingly, it is clear that not much diagnostic material is available to identify a glass workshop where black glass was produced or where black glass working was done to manufacture various types of artefacts, whether it were vessels, jewellery, counters, architectural decoration material, or sculpture. 2.3 The Artefacts 2.3.1 Vessels Various techniques were used in the production of black glass vessels. Three main techniques have been distinguished: casting, blowing, and rod-formed. Within the group of blown vessel glass, free-blown vessels constituted by far the largest group, but a number of mould-blown vessels show that this technique was also applied in the production of black-appearing glass (Table 7). 30 However, after the application of a handle, the remaining glass was also discarded as waste. Such activities demonstrating the abundance of recycling in Roman times – in particular in the late Roman period – are discussed in chapters 9 and 11 in perspective to the research results from very recent chemical analyses on late Romano-British glass and the hypotheses inferred from it regarding the organization of glass production in the Roman Empire (Foster, Jackson 2010). 31 19 Table 7: The ratio of different techniques utilized in the production of black glass vessels technique quantity percentage cast vessels 86 13,2 free-blown vessels 421 64,7 mould-blown vessels 24 3,7 rod-formed vessels 48 7,4 undefined 72 11,1 TOTAL 651 100,1 2.3.1.1 Casting This technique enfolded any heat-forming technique by which monochrome and polychrome glass were moulded, whether it was by mould-pressing or by sagging. Single-piece open moulds as well as multipartite closed moulds were used to carry out mould-pressing (Stern 1995, 29-30, fig.15). Sagging involved the reheating of a glass disc in the furnace by putting it on top of a mould so that it slowly sagged over or into the mould (Grose 1989, 33, fig.8-9). The most common sagged shapes were the linear-cut bowls and the ribbed bowls or so-called ‗pillar-moulded bowl‘. This technique resulted in a different surface property, with a shiny external surface and a dull inner surface. This was due to the fire-polishing of the top surface in the furnace and the contact of the underside with the mould. The technique also resulted in an irregular rim that got grounded. Very few examples of sagged monochrome black glass vessels are known. Some pieces have been found in Liberchies (BE) [cat.nos.417; 425] as well as in Augst (CH) [cat.no.3474]. They all reveal their true colour when a strong light beam is used, as when an art photographer takes pictures for an exhibition catalogue. Some interesting pieces are the ribbed bowl in the Toledo Museum of Art, Toledo (Ohio,USA) (Grose 1989, 266, no. 240) and the bucket at the San Marco Treasury, Venice (IT) (Harden et al. 1987, 220-221, no. 122). Most cast black glass vessels were mould-pressed within a single-piece open mould implying the use of a pressing object to compress with force the hot glass into a shape. By using a presser, the inner side became smooth, but with an uneven wall thickness; this method contrasted with the creation of mould-blown objects using air as the compressing force, which resulted in a vessel with consistent wall thickness, thus allowing a negative 32 relief on the inside of the moulded vessel. This presser could be made of stone or clay, but most were probably made of metale.g., copper, bronze, or iron. It seems that no evidence has as yet been found, other than the design pattern in the centre, as David Grose explains in a more detailed discussion of these types of manufacturing (Grose 1989, 31-33). 2.3.1.2 Blowing 33 Characteristic to blowing glass was the use of a blowing pipe in iron that was not only used to inflate air in the paraison to start modelling the glass bulb but particularly to gather the hot glass from the very blazing furnace, keep enough distance from the hot glass in the furnace tank. Based on information from for instance the glass workshop at Les Houis nearby Sainte Menehould the blowing pipe was made of tubular folded and hammered iron sheet (Chew 1989, no.36; Foy, Nenna 2001, 77, no.59). Based on the knowledge from Renaissance period onwards the glassblower‘s tools remained basically unchanged until today. Because we assume that already from the invention of glassblowing the perfect instruments for manual blowing must have been developed from the Roman imperial period. Taking into account this assumption we may suppose that the Roman glassblowers used blowing pipes of about 1,2m to 1,5m long and with a diameter ranging between 15-25 mm. o Free-blowing Characteristic to the free-blowing technique was the gradual inflating, reheating, and tooling of the paraison to end up with the desired shape: 1) pre-heating of the gathering-end of an iron blowing pipe until it reached the furnace temperature and consequently that of the hot glass 32 Only angular-shaped vessels show a difference, as the spherical inflated bulb in a square, rectangle, or polygon has to expand further to the corners, while the bulb already touches the mould. Hence, the wall thickness will diminish towards the corners. 33 We do not want to ignore the proposal by Marianne Stern arguing for the use of blowing pipes in clay (Stern 1995, 39-42, in particular 40-41; Stern 1999, 446-447, fig. 8; Stern 2005, 17-18), but no archaeological indications are hitherto available to suggest this method was employed by the Romans. We do have knowledge of such a ceramic tool from excavations in Sagalassos (Lauwers et al. 2005, 28, fig. 4; Lauwers et al. 2007, 194-198), but the use of this 50 mm long and 9 mm in diameter thick rod fragment stuck in a chunk of glass remains rather vague. This could have been used in the preparation of the glass melt as stirring rod, in the production of glass beads, or in glassblowing as a pontil rod. However, it could not have been used as a blowing pipe, since it is not hollow. 20 2) 3) 4) 5) gathering hot glass (if necessary repeatedly in several times running until the required glass mass [was] reached to blow the desired vessel); modelling the paraison by: a. blowing a small air bubble in the paraison; b. swirling the blowing pipe to elongate the paraison to the required length; c. modelling manually the inflated paraison into the desired shape by using various tools such as a marver, jacks, and a mailloche; knocking-off the vessel from the blowing pipe; eventually the use of a pontil rod to finish the vessel. This technique thus resided in manipulating freely the blowing pipe and the gathered paraison, resulting in specific shapes. The fingerprint of a glass worker linked to a workshop makes it possible to characterize idiosyncrasies related to a well-defined period and/or region of production. Some publications show visually the consecutive stages in the production of various types of free-blown glass vessels (Tait 1995, 213-241; Dily, Mahéo 1997, 21-27). Their purpose was to explain in a clear but general way the various aspects of glassblowing technique. The challenge here has not been the chaîne opératoire as such, but rather improving our understanding of particular regional and/or chronological assemblages by studying in detail the variation of explicit technical features on a particular type of vessel shape in black glass [see Chapter 6]. o Mould-blowing The mould-blowing technique covered the same actions as the free-blowing technique. The only difference compared with free-blowing lay within step 3 (modelling the paraison). The final stage involved a modelling of the inflated paraison into the desired shape by using a mould. From the mould-blown vessels can be discerned two different types of moulds: 1) plain circular ones (eventually with concentric circles); 2) angular moulds or circular moulds with a more complex design. The former mould-type allows the glassworker to turn around the glass blower‘s pipe while blowing the vessel in the mould, whereas the later mould type consists of blowing very briefly but powerfully without moving the blower‘s pipe. The use of a single, bi- or multipart mould made possible to produce ad infinitum identical objects. Differences between two objects from the same mould could occur in the modelling of handles, rim, and neck, because these parts were freely shaped after taking out the inflated paraison from the mould. This is noticeable from the absence of a seam on neck, rim, and handles on all Roman mouldblown vessels. It also makes clear that the moulds just reached up to the shoulder of the vessels, possibly meaning that closed moulds were unusual or not used at all in Roman times. Detailed work concerning the mouldblowing technique became recently available (Stern 1995; Foy, Nenna 2006a-b; Fontaine-Hodiamont (ed.) 2010). We therefore refrain from a detailed discussion on all features concerning mould-blowing, instead we refer to these reference works. The moulds into which glass recipients were blown were mainly made from limestonemainly of marble, but 34 Roman glassblowers also must have used moulds in metal and clay. The type of vessel likely influenced the choice of the mould material. Square bottles seem to have been blown singularly in limestone moulds built up from five piecesa bottom piece and four side pieceslike the examples known from Augst (CH), Avenches (CH), Bonn (DE), Cologne (DE), Lyon (FR) (Rütti 1991). A typical feature of mould-blown vessels is the flashing or seam caused by the imperfectly closed mould segments. This is not visible on the square and rectangular bottles because the mould segments joined at the corners and stopped at the start of the shoulder. This means that the shoulder, neck, and rim were not mouldblown. This would only have been possible with an open mould. Other forms such as juglets and amphoriskoi in the shape of shells, animals, human heads, or fruit show a clear seam up to the rim, providing evidence of the number of mould segments used. In black glass, not many mould-blown examples have been known up till now. This is striking in comparison with the loads of mould-blown vessels in ‗naturally coloured‘ and decolourised glass, and in contrast with the relatively large quantity of free-blown black glass vessels. Besides some head flasks, the only two other shapes recorded are the grape-shaped amphoriskos and the prismatic bottle [see Chapter 3, more particularly 3.2.4. mould-blown vessels]. Given the very poor occurrence of mould-blown vessels in black glass, it seems very likely that this technique was barely adopted during Roman times to produce black glass vessels. The paucity in number can 34 Most likely in bronze as those surviving from early Byzantine and early Islamic times. 21 possibly be explained by the technical complexity of blowing vessels in black glass. The usual working time for hot glasscharacterized by a low iron concentrationaverages around 20 seconds from the moment of removal 35 from the batch. Present-day glassblowers state that the glass metal becomes ‗shorter‘ the more iron is added to the batch. Technically speaking, this means that the higher the iron content (FeO and/or Fe 2O3), the shorter the workability of the hot glass, with a limit that can result in an unworkable batch. An incident at the glasshouse ‗Het glaspaleis‘ in Lommel (BE) where a part of the steel covering of the furnace tank fell into the batch resulted in a black glass metal that became unusable because the too-short working time ended in a hot glass that could no 36 nd rd longer be inflated. The highly elevated concentrations of iron oxide (up to 14 %) in use during the late 2 and 3 centuries AD in the north-western provinces of the Roman Empire must have resulted in a drastically lower working time for the black glass metal. Therefore, this technical aspect must have had its consequence in the shaping and blowing of free-blown vessels too, perhaps explaining a part of the reason why these vessels demonstrate harmonic observed proportions which seem to be connected to the golden section giving an object aesthetically pleasing proportions [see Chapter 3, more particularly 3.2.3. Free-blown vessels]. 2.3.1.3 Rod-formed 37 The rod-formed technique consisted of applying hot glass around a metal rod or mandrel and then shaping the hot glass in a specific form. The insides of rod-formed vessels took the shape and size of the metal rod and were therefore cylindrical. Tapering inner sides were created by later manipulations with pincers when the vessel was removed from the rod to form the rim. Hence, the only resemblance the technique bore to the core-formed technique was that neither involved blowing. The shapes of the inside and the wall thickness were substantially different from each other. The pre-Roman core-formed vessels, produced around a pre-shaped core in clay, therefore did not need much glass to obtain a particular shape, while large amounts of glass were necessary to produce rod-formed bi-conical and globular-shaped vessels around a bare rod. Accordingly the inside of coreformed vessels corresponded with the vessel shape, whereas that of rod-formed vessels took the shape of the metal rod. Only very few rod-formed vessels were undecorated. On the whole, they were decorated with wide ribbings, either vertical or twisted, with pushed-in decoration or with applied opaque glass trails in zigzags [see Chapter 3]. The production of rod-formed vessels was very coarse, giving these recipients a very clumsy appearance. The technique was not restricted to black glass only. Black glass accounts for the large majority of rod-formed vessels, but ultramarine and pale blue glass were other commonly used glass colours for the production of rod-formed vessels. Many of these coloured vessels have a large quantity of air bubbles. Limited to the production of toilet ware, i.e., unguentaria, generally assumed to be kohl tubes and wide-mouthed jars, the technique is idiosyncratic for the later Roman-early Byzantine period in the Levant and Egypt [see Chapter 3, more particularly 3.2.5. Rod-formed vessels]. 2.3.1.4 Decoration It appears that a large set of decoration techniques, if not all modi operandi, have been employed on the vessels produced in black glass. The plain vessels are the most generally seen, but the decoration is very diverse: applied glass trails, abraded concentric lines, engraved figurative motifs, indented, marvered, painted, and so forth. Vessels of the same shapesuch as the small bulbous cup (Form IIB.4/5) were embellished using various decoration techniques or simply left undecorated [see Chapter 3]. o Applied glass trails The applied glass trails could be executed in the same black glass as the body, but regularly contrasting opaque colours like white, turquoise blue, yellow, or red were also used. Depending on the production technique, different colours and decoration patterns were favoured. The free-blown vessels had no opaque red trails and the applied glass trail consisted of one concentric line on the shoulder and one on the lower side of the body, or, when all over the body, with an upward festoon pattern. The former had no marvered trails, whereas the festooned trails of the latter were always marvered. The trailed decoration on rod-formed vessels was mostly applied in zigzags with 35 Personal communication by Danny Theys, a professional glassblower of the glasshouse ‗Theys-Miseur‘ in Holsbeek/Leuven, Belgium. Personal communication by Jeroen Maes, head of the glasshouse ‗Het glaspaleis‘ in Lommel, Belgium. It is not clear whether the surface of the rod was enveloped with a coating layer to facilitate the removal of the mandrel from the finished product. Seeing the clean inner surface of these vessels clay is excluded. 36 37 22 contrasting colours such as opaque yellow, white, and turquoise blue glass, whereas black-on-black was never used. Occasionally, more than one trail was applied. In those cases, different colours were combined. o Wheel-cutting st The abraded concentric lines occurred mainly on 1 century AD free-blown vessels such as the so-called Hofheim cups (Isings Form 12 – see Chapter 3, Form IB.1) [see Chapter 3]. These wheel-cut concentric lines were designed in the same position as the applied trails below the rim (on the shoulder if there was one) and at the lower part of the body (generally coinciding with the largest diameter of the vessel). The other decoration on black glass vessels were engraved figurative motifs on the body. The bulbous cup in the Römisches Germanischen Museums in Cologne (DE) [cat.no.1840] is decorated with several registers of floral motifs representing ivy leaves alternating with undulations. Even if the authenticity of the decoration has been questioned, we agree with Fremersdorf‘s view that it is genuine (Fremersdorf 1958, 42-43, Taf.72). Real wheel-cut vessels in black glass are unknown thus far, except for the bucket in the San Marco treasure in Venice which was made from a deep-purple glass that appears black (Harden et al. 1987, 220-221, no.122) th [cat.no.2827]. The cylindrical bucket with silver handle is dated 4 century AD on the stylistic basis of the wheelcut design and technique, although the vessel shape, the cast technique and the poorly translucent deep purple glass are not common for this period. Another wheel-cut piece in deep purple glass appearing black is the fragment in the Petrie Museum for Egyptian Archaeology, London (unpublished) [cat.no.931] (Figure 11) or the fragment in Corning Museum of glass which is also said to be from Egypt [cat.no.996]. Figure 11: Wheel-cut fragment in the Petrie Museum for Egyptian Archaeology, London (photograph by author, by courtesy of Petrie Museum) o Marvering 38 st Marvered decoration was limited to free-blown vessels and was typical for the 1 century AD black glass vessels. This technique consisted of applying small particles of coloured glass in the hot glass by first rolling these small bits onto the hot glass when not yet inflated (Figure 12). After reheating the paraison, the coloured glass particles were marvered into the hot glass by rolling it again on the marver (mainly a smooth plate of marble). During the blowing process, the coloured glass particles enlarged according to the repeated blowing to achieve the required shape (Figure 2d-e). Although opaque white glass was used principally, it also occurs in combination with opaque yellow and opaque pale blue glass. A variant of this technique was to roll on a marver the applied white opaque glass trails and pull it afterwards up and/or downward to obtain a festooned or feathered pattern such as the bulbous cups (Isings Form 94 – see Chapter 3, Form IIB.4 and 5) and the larger jugs (Isings Form 54 – see Chapter 3, Form IIB.12) and amphoriskoi (Isings Form 15 – see Chapter 3, Form IIB.14). 38 This decoration technique was also used on late Roman and early Byzantine jewellery in the eastern Mediterranean. 23 Figure 12: Vessel with rolled-in glass particles in a contrasting colour to be marvered from the Canal Bianco necropolis, Adria (IT) (photograph taken from Barovier et al. 2003, 231, no.3; cat.no.2586) o Pressed-in decoration The true pressed-in decoration was only applied to rod-formed vessels. Probably this was due to the great wall thickness of the body. The stamps were circular in shape and were equal to the stamps used on the discoid beads and pendants (see below). The stamps sometimes modified the original shape of the vessel. A piece in Grand Curtius, Liège has three pressed-in decorations on a regular distance, thus changing the globular body into a triangular shape. Some vessels had a stamped decoration. This consisted mainly of an applied medallion in relief placed at the start of a handle on the body of a vessel. When the handle was vertical, there was only one medallion applied; when the handle was horizontal, a medallion was attached at either side. o Tooling By using pincers or rods, various decorative patterns could be obtained, such as a dented surface. A pin caused a circular indent, whereas a flat tool was needed to obtain elongated indents. Indented decoration was regularly applied on free-blown cups and goblets to give it a polygonal shape with rounded edges [see Chapter 3 – Form IIB.4-5]. The indents were circular or elongated. The dented decoration was created by pushing in the body with a pincer or another pointed tool for the circular indents, and a wide metal tool for the elongated indents. The former was typically applied on cups with a globular body, while the elongated dents were applied on taller vessels with a cylindrical shape. Sometimes indents and applied glass trails were used in combination on one single vessel. This ornamental pattern was not restricted to black glass vessels. Similar vessels with identical decoration are known to have been produced in colourless and ‗naturally coloured‘ blue-green glass. Indented decoration seemingly st was not applied on the 1 century AD black glass vessels although the technique was already adopted in other glass hues (Isings 1957, Forms 32; 35). A variant on these vessels were shallow bowls with scalloped edge showing a wavy body and rim caused by the indented upper part of the body until the rim [see Chapter 3 - Form IIB.7] like the one from the tumulus III of Esch (NL) [cat.no.2880]. Occasionally, tooling was applied on the late Romanearly Byzantine rod-formed vessels of the Levant, as seen on the rod-formed vessels with large and coarsely twisted ribbings [see Chapter 3, more particularly 3.2.5. Rod-formed vessels]. o Moulding Mould-blowing in Roman times implies the blowing of the prepared paraison in a mould to obtain vessels continuously in an identical shape with an identical size and an identical design. A main marker is the very characteristic seam. Yet not all shapes needed to be blown in such a way that seams are inevitable. Plain cylindrical vessels have the advantage to allow the inflating of the paraison while rotating the blowing pipe so that a seam can be avoided. According to us there must be a number of vessels mould-blown in a rotating way that these forms are in general described as being free-blown. For instance the cylindrical cup from Tongeren [cat.no.562] (Vanderhoeven 1962, 23, no.15; Cosyns et al. 2006a, 31-34, figs.1-2) is to us an example of rotated mould-blown vessels [see Chapter 3, Form IC.1]. Some moulded effects, conversely, are obtained during a short moment to finish the free-blown shape of a specific vessel type. But is it correct to call the applying of a decoration by using a mould when free-blowing glass as mould-blowing? For instance, the carchesia or carinated 24 beakers (Isings Form 36b – see Chapter 3, Form IIB.1) have a very characteristic carination just above the foot, which cannot be obtained by free-blowing only or achieved by simple tooling. We rather would like to suggest that the Romans used for the shaping of this form a wide ring or open mould. By pushing the spherical lower part of the inflated hot glass against a wide ring, a part of the convex surface was flattened, thus creating this specific carination (Figure 13). The pinprick pitted surface of the flattened lower body part of various carchesia, like for instance, the one of Heel (NL) [cat.no.2883] (unpublished) (Figure 14) or Grobbendonk [cat.no.291] (unpublished) is due to thermal shock when the hot glass touched a much colder surface. It is not clear what material the mould was made of. A tripod with metal ring may well have been used, but based on the observed chill marks, an open cylindrical mould in stone was most likely. Similar pitted surfaces are observable on the walls of mould-blown vessels such as the ubiquitous square bottles, of which several mould segments are known in marble and in terracotta (Rütti 1991, 163-164, fig.103; Hochuli-Gysel 1993, 87, fig.5-6; Sternini 1995, 90-95, figs. 129-136, 141142; Foy, Nenna 2001, 51). The opening of the ring-mould used to blow the carchesium of Heel must have been c. 60 mm in diameter, while that of Grobbendonk c. 50 mm. This pitted surface is not attested on all carchesia. We may, therefore, attribute this technical feature as an element to help recognize the fingerprint of a glassblower and, consequently, the production of a specific workshop [see Chapter 6]. Figure 13: Reconstruction of possible production process (drawing by author) Figure 14: Pitted surface due to a thermal shock on bottom of the Heel carchesium (NL) [cat.no.2883] (photograph by author, by courtesy of Limburgmuseum, Venlo) 25 2.3.2 2.3.2.1 Jewellery Arm rings In the Celtic world, the knowledge required to produce glass bracelets already existed for three centuries before the Romans conquered most of their territories. The Celtic way of producing has already been discussed in detail on various occasions (Haevernick 1952, 212-215; Haevernick 1960, 23-32; Kunkel 1961; Korfmann 1966; 39 Küçükerman 1988, 30; Wagner 2006, 32-34; Gérardin et al. 2010). Briefly, the Celtic way consisted of gathering some glass from the furnace by means of a metal rod and preparing an annular object by first perforating and afterwards by swirling it around until the required shape was obtained, and/or by rolling it over a cone in terracotta or stone to give the bracelet the desired diameter. Possibly, both actions were done consecutively but then definitely ending with the cone, given the very characteristic section of Celtic glass bangles, which is not straight and pure semi-circular. Firstly, the slanting section made the inner diameter different from one end to the other. Secondly, the end with the narrower diameter is pointed, whereas the other end is much thicker and rounder. A partial process of this gathering technique was also used to produce the glass beads in the late Iron Age (Küçükerman 1988, 30). The Romans, and, in fact, the entire Mediterranean world were not acquainted with glass arm rings [see Chapter 3]. Surprisingly, when they became fashionable in the Roman world, a proper way of producing glass bracelets was introduced instead of adopting the Celtic technique. As an alternative to the gathering of a small gob of hot glass with a metal rod and the use of a cone, the Romans used prefabricated glass canes to produce bracelets. This rod-formed technique is called here rod-made to contrast the technique of shaping rod-formed vessels on a metal rod. The rod-made technique was also adopted for the production of most other glass jewelleryfinger rings, beads, pendants, and hairpinsbut in particular for the production of utensils and architectural decoration material. How these rods were made is difficult to proof but most likely the Romans produced glass canesplain and twisted, monochrome and polychromeas nowadays is still the case in artisanal glass workshops in India by drawing a glass wire after having wrapped the hot gob of glass around a fixed pin (Kock, Sode 1995, 21). Most interesting is that the production of twisted glass rods in present-day northern India are made by drawing a quadrangular gob of hot glass. The variations in density of the twists resulted from the fastness of turning the metal rod while drawing the glass into a cable (Kock, Sode 1995, 16). To obtain rods with a uniform thickness the complete gob of glass needed to have the right temperature and keep it as well during the entire wire drawing process that could last for hundreds of metres (Kock, Sode 1995, 16). After cooling the cables were cut into shorter rods of a specific length. The twining of glass rods needs to be considered as the result of deliberate movements during the production process. Considering that a clockwise movement of the wrist is the least demanding the S-twined rods must be the work of right-handers whereas the Z-twined rods must have been made by left-handers. The mixed occurrence on the secondary glass workshop of Les Houis nearby Sainte Menehould (FR) is evidence for the unsuitability of this feature to detect idiosyncrasies. 39 The recent discussion on the making of Celtic glass bangles by Simone Wick (2008) simply needs to be ignored, as the described experiment is full of flaws in the argument. We assume an interpretation of observed technical features on archaeological material that is as objective as possible is more appropriate than high-tech interpretations without base. Our investigations can be supported by ethnographical research on the technological knowledge of artisans working with glass under rather basic conditions, but with a more correct technological and socio-economic impetus (Kock, Sode 1995). 26 Figure 15: (left) Thin elongated pointed elliptical scar of a punty on apex of the top surface of an open elliptical bracelet from Poitiers-Vienne (FR) [cat.no.1345] (drawing by author, by courtesy of MAN, Saint-Germain-en-Laye); (right) Closed, circular bracelet with clear seam from Kaiseraugst (CH) [cat.no.3687] (drawing taken from Riha 1990, pl.74:2930) This rod technique also made it possible to produce easily open bracelets by using pincers, jacks and scissors. Concerning the closed bracelets, it resulted in the characteristic seam where the two ends of the glass rod joined (Figure 15-right). It is obvious that fragmentary bracelets might miss the part with the seam, but when present, the seam is easy to recognize. Some intact bracelets even though show very high qualitative workmanship making it sometimes difficult to discern the seam. The use of glass canes in diverse thicknesses has been observed for the making of open bracelets. Another technological element detected on a number of open and closed bracelets is the presence of a thin elongated pointed elliptical scar centrally on the apex of the top surface like for instance on the piece from cemetery "la Pierre Levée des Dunes" at Poitiers-Vienne (FR) [cat.no.1345] (Figure 15-left). Considering such occurrence on pieces of the secondary glass workshop at Les Houis nearby Sainte Menehould we believe the mark is the result from a distinct action within the chaîne opératoire of the bracelet production caused by using a small punty or pontil rod or the scar of a broken off glass gem or applied beaded decoration and consequently a decorative remainder. A more careful look at this feature might be useful to detect possible finger prints of different workshops, as it can have to do with the making of specific types of bracelets or point to the technological routine of a specific workshop. Or is it a generally introduced technique utilized by a particular generation glassworkers within a wider region independent from the workshops modelling bracelets? Looking more careful at the Roman black glass bangles, a range of features point at the use of various techniques that are reflected in the various sections observed: 1) Bracelets with an O-shaped section were made with circular rods and must have been made without the use of a support when modelling the bangle. Open and closed bracelets were produced, and the latter show a clear seam; 2) Bracelets with a broad D-shaped section probably were made with the same circular rods as those for the bracelets with O-shaped sections, but flattened at the inner side by using a support. Open and closed bracelets were produced, and the latter show a clear seam; 3) Bracelets with a very wide and flattened D-shaped section probably were made with circular rods too, but have been strongly flattened by being pressed with a indented tool on a flat support. This would explain the very broad width and the limited height. Only open bracelets were produced; 4) Bracelets with a narrow D-shaped section have not been made with rods but must have been made using the La Tène technique, implying the gathering of a gob of hot glass with a metal or ceramic rod and pulling it over a conical tool in metal, stone, or clay until the required shape was reached. Only closed bracelets were produced without leaving a seam. Black glass bracelets with an angular shaped sectiontriangular and quadrangulardo occur but are very rare. Even though there was no opportunity for a close look at the technical particularities of this unusual material, we 27 presume these bangle types were rod-made and not gathered and rolled over a cone as was usual in the late Iron Age. Most likely a mould or other tool must have been used to give the glass an angular shape. It is the challenge in the following chapters to verify whether this is due to chronological, regional, or purely 40 practical reasons (read: mechanical and technical). Therefore, we based the classification of the bracelets in the first place on the technical aspect. Only bracelets of type 1, 2 and 4 using only a few techniques: 41 show undecorated examples, but when decorated, a wide variety occurs by - Twisting rods with pincers only were used on bracelets of type 1, but pincers were also used for creating the spatula-shaped ends of open bracelets; the twisting was done clockwise (S-twined) or anti-clockwise (Z-twined); - Glass trails in one or more colours applied before twisting only were used on bracelets of type 1; - Jacks were used on bracelets of type 2, 3 and 4; - Combs were used on bracelets of type 2 and 3; - Stamps were only used on bracelets of type 4; - Polychrome mottled decoration was only used on bracelets of type 4; - The application of a glass gem is presumed on all types of bracelets. 2.3.2.2 Finger rings Finger rings in glass are, in contrast to the bracelets, typical for the Mediterranean period and totally absent in the Celtic world. Worked out in only a small number of colours, black glass appears predominantly to have been used. The same classification per section type can be made parallel to the bracelets, and similar decoration techniques have been adopted by twisting, applying glass trails, and using stamps or jacks [see Chapter 3]. 2.3.2.3 Beads ...any relatively small object, irrespective of shape, color or material, that has an opening (or openings) for stringing. (Spaer 2001, 43) The majority of all glass beads were rod-made (Guido 1978; Spaer 2001), yet a considerable number of black 42 glass bead types were produced by folding and pressing. Particular decoration techniques were used on specific types of beads. The annular, globular and cylindrical beads basically were decorated with applied glass trails and dots. This decoration was principally in one or two colours, but sometimes even three different colours were combined. Normally, they were not marvered, showing the glass threads laying in relief on the core of the bead [see Chapter 3]. o Rod-formed Two technical movements‘ variations can be distinguished. To make a bead there is on the one hand the gathering of a small amount of hot gob of glass by means of a metal rod, called a pontil or punty. On the other hand the term can refer to the aforementioned rod-made technique characteristic to Roman bangles [see here above in this chapter Bracelets] implying the use of prefabricated glass rods to manufacture beads. With both techniques the hot gob of glass is wound around another rod, also called a mandrel. o Folded and pressed Pressed beads were made in two different ways: 1) by pressing the hot glass in a mould, or 2) by pressing a stamp in the glass mass, usually when on a marver. This technique is very old and was already in use in the Bronze Agee.g., the Mycenaean cast beads (Pini 1981, 48; Demakopoulou 1988, 219, nos.203-204). A very characteristic bead in black glass obtained by folding the hot glass and pressing a stamp in the glass mass is the ribbed spacer better known as the ‗Trilobitenperle‘ (Haevernick 1983) [see Chapter 3]. Some confusion 40 Our classification system is different from that established by Maud Spaer (1988), whose taxonomy of the Palestinian pre-Roman glass bracelets is developed on the basis of decoration [see Chapter 3]. 41 From Chapter 3 onwards these will be named correspondingly Type A; Type B; Type D; Type U as adopted in the catalogue database. 42 For further reading on technical aspects in bead manufacturing we refer to Spaer 2001, 4456. 28 exists apparently about the correct use of the word ‗Trilobitenperle‘, as it is sometimes used to describe any folded-pressed beads with a double perforation such as the theatre-mask beads from Cesenate (IT) (Fadini, Montevecchi 2001, 51). A detailed examination of this type of bead demonstrates they were made by applying black-appearing glass on a working table and by putting two thin rods next to each other before folding the glass mass (Figure 16). Yet, elsewhere the technique is described differently as ‗rod-formed and furnace wound‘ (Spaer 2001, 76). An undecorated example from Tielrode (BE) [cat.no.529] not only shows the bead was folded, but the flat basal surface also contains chill marks, irregular concentric circles caused by the thermal shock of the hot glass on a cold(er) metallic surface (Figure 17). Figure 16: Possible reconstruction of production process (drawing by author) Figure 17: Basal side of the so-called ‗Trilobitenperle‘ from Tielrode (BE) [cat.no.529] showing concentric chill marks (photograph by Marc De Meirelaer, by courtesy of Museum Van Bogaert-Wauters, Hamme) The ribbed decoration sometimes was simply applied by using a sharp utensil in metal, but basically this was done by pressing a stamp onto the top surface. Finally, the two rods were removed. The production method explains the presence of a clear seam at one side up to the perforation and in between the two perforations. This manufacturing feature illustrates that the folding of the glass mass was done around the second rod. It provides 43 additional information useful in recognizing the difference between the so-called ‗Trilobitenperlen‘ in jet and those in black glass (Table 8), even though a trained eye can immediately distinguish both materials. We noticed that ribbed spacers described as being made from jet sometimes show a clear seam on the drawing (Keller 1971, pl.47:6). Considering that there is absolutely no reason for the jet-carver to have cut an imaginary seam on the bead, it shows that the draughtsman looked more carefully at the artefact than did the archaeologist. For that reason, we are persuaded that Erwin Keller erroneously described at least one of the ribbed spacers from the cemetery of Göggingen as being made of jet, but presumably all are made of black glass (Keller 1971, 87, 232 (10), fig.27:5, pl.9:15-18) [cat.nos.1888-1890]. Other indicative aspects demonstrate the difference between ‗Trilobitenperlen‘ in black glass and those in jet: 1) these beads remain very rare in Britain, the region emblematic for jet objects; and 2) those in jet show very distinctive features: 43 That a wide range of coals have been used during Roman times to produce artefacts, e.g., Kimmeridge shale, durain, etc we are owing to Lindsay Allason-Jones (see for further reading Allason-Jones 2001 and 2005). However, since we are not focussing on this issue, we describe them all as made from jet. 29 Table 8: Differences between the so-called ‗Trilobitenperlen‘ in jet and black glass feature jet black glass surface colour texture decoration pattern proportions shape smooth but cracked solid opaque brownish black soft and shiny very symmetrical well-proportioned nearly semi-circular (resembling half a melon bead) smooth with iridescence poorly translucent to opaque greenish black hard + air bubbles and inclusions asymmetrical disproportioned irregular and flat The pressed decoration implies the use of small moulds or stamps. Although no such moulds/stamps have been recorded yet, they most probably were made in baked clay, as is demonstrated for the glass gems [see below in this chapter], or metal. The use of such moulds/stamps resulted in beads with an identical decoration pattern. Hence, a very detailed study of this type of bead carries with it a very interesting research potential in dating the various stamps, but in particular the distribution impact of each stamp because these so-called ‗Trilobitenperlen‘ are widely spread all along the Rhine and Danube limes and beyond [see Chapter 6]. Well-dated contexts can also lead to specific dating of the various decoration patterns. 2.3.2.4 Pendants Pendants, unlike beads, have apertures placed off-center to facilitate suspension, as their name suggests. … The distinction between beads and pendants is not always clear-cut. (Spaer 2001, 151) Pendants can be considered as beads with a suspension loop, whether it has been applied separately or as a perforation within the bead itself, whether it is protruding or fully integrated within the shape of the object. The various techniques used to produce pendants in glass during the Roman imperial period were identical to those used for the beads. Casting, mould-pressing or stamping, rod-forming, gathering and drawing, and combinations of different techniques were adopted to produce pendants in black glass. Yet it is clear that a number of techniques were never or rarely used to produce pendants in black glass, while other techniques were adopted in abundance. The various decoration techniques applied on pendants are identical to those that were used for beads, but the main style was the applying of opaque, coloured glass trails spiralling and in zigzag. The used colours were white, yellow, turquoise, and red. Another method was the stamped or mould-pressed technique, but, as with the matching beads, this technique was mostly executed on brownish-yellow and ultramarine blue coloured glass. It was adopted to produce discoid beads and pendants. A shape that does not exist as a bead is the vessel-shaped pendant. All vessel-shaped pendants were rodformed by gathering and trailing. The decoration sometimes was open worked by trailing (Spaer 2001, 171-172, fig.79-80), but this variant does not occur in black glass. 2.3.2.5 Gems Usually gemstones were engraved in different styles dependent on the period of manufacture [see Chapter 5] and probably also in relation to the skills of the engraver. In contrast with the gemstones in semiprecious stones, the glass gems appear on the whole to have been moulded, yet high-quality cutting or engraving do also occur on a number of glass gemstones (Platz-Horster 1987, Taf. 43, no. 211) [cat.no.2176]. Sometimes a combined technology can be noticed in the decoration, as can be seen on the Xanten gemstone no. 250 with trireme (PlatzHorster 1987, 139, Taf. 50, no. 250) [cat.no.2159]. The ubiquitous gemstone using black glass was the so-called pseudo-nicolo gem having a bottom layer in black glass and a topping in opaque blue glass, into which the decoration was pressed. However, gemstones in monochrome black glass or with an opaque white strip in the centre also were made, all in imitation of the very popular but expensive or rare semiprecious stones. The gemstones were mainly oval in shape but were also circular, and they had, in general, an abraded sloping edge. o Moulding – casting – pressing No clear factual evidence is thus far at hand to give a full and correct picture of the chaîne opératoire of glass gemstones during the Roman imperial period. No workshop has been excavated specialized in the production of glass gems. However, given the sometimes partial impressed image or the perfect resemblance in size and decoration, it is clearly deducible from the material itself that most glass gemstones were moulded. It has been 30 reported that the material in which the moulds were made remains unknown. Wax and plaster are excluded as unusable, whereas stone should be considered as most appropriate to be used as a mould; stone is hard as well as workable and heat resistant (Krug 1978, 486). An example of a mould to cast gemstones in glass is said to have been found representing in slightly high relief a standing woman looking to the left with the right arm resting on a pillar (Ogden 1982, 129). It is not clear whether the piece can be excluded as a clay sealing, but if the piece was used as a mould to cast glass gemstones, not only the size of the gem but also its figurative motif should be very recognizable. Such evaluation is far beyond the scope of this study seeing the amount of gemstones present in the large museum collections worldwide. Careful examination of the gemstones provided information on how the manufacturing process of glass gemstones must have been. Glass gems were cast into a mould and by pressing of the gob of hot glass to obtain a flat rear (Krug 1978, 486-487). The technique could also imply a more complex course of action as for instance the layered gemstones combining two or three layers in different colours. The most frequent types are the blueon-black gemstones also called pseudo-nicolo gems in imitation of the semiprecious nicolo, and the white striped black stone in imitation of the semiprecious sardonyx. o Polishing When the gemstone was shaped, it had to be polished in order to remove the rough edges and any irregularities to obtain a natural character similar to that of the semiprecious gemstones. Mainly of oval shape, circular ones do occur, like two undecorated pieces one from Augst (CH) [cat.no.3389] and another from Kaiseraugst (CH) [cat.no.3655] (both unpublished). The edges of both pieces were bevelled but were left with a very rough top surface, while all pseudo-nicolo gemstones had a very neatly cut and polished edge [see Chapter 3]. May we interpret both pieces as examples of a provisional stage in the manufacture of gemstones? o Engraving The engraving of glass gemstones was done by drilling in a similar way, as attested on intaglios in semiprecious and precious stones. The various styles already discerned can also be recognized on glass gemstones, for instance, the golden finger ring with black glass stone in the Rheinishes Landesmuseum Bonn (Platz-Horster 1984, 87, Taf. 22, no. 78) which is decorated with an engraved grape in front of a hare in a very realistic style. 2.3.2.6 Hairpins Hairpins in glass had typical features: 1) they were relatively short (60-70 mm, although longer ones do occur sporadically); 2) the diameter of the shaft thickened from point to pinhead, but just below the pinhead a constriction created a sort of shoulder; and 3) most of the time the pinhead remained a plain globular end although sometimes flattened, narrowed or shaped into a cone. It was correctly observed by Donald Harden when looking at the hairpins of the cemetery at Lankhills that these objects could not be cast, but must have been drawn and tooled (Clarke 1979, 315). The constriction below the head makes clear that hairpins in glass were made by holding a gob of hot glass in the middle with a tool while the bottom part was pulled down to a point with the other hand or by a second person. On rare occasions, hairpins have been recorded with a shaft in iron decorated with a head in (black) glass (Harden 1979, 249; Cosyns 2009, 90). 2.3.3 2.3.3.1 Architectural decoration Rods Rods were used for ornamenting wall and vault mosaics, in particular for bordering polychrome panels but also as stirring rods [see Chapter 7]. Both types were made by pulling hot glass into trails with pincers and subsequently by twisting them, while both types are not fully contemporaneous [see Chapter 4]. The earliest rods were used for architectural decoration and have a rather complex production process. 2.3.3.2 Tesserae Tesserae are small blocks of inlay with an irregular cubic or triangular shape (Figure 18). Normally in stone, the glass tesserae have also been regularly recorded in various opaque and poorly translucent colours. Based on the black glass tesserae from the glass workshop at Les Houis nearby Sainte Menehould (FR) we determined smallsized (4-5 mm side) or medium-sized (8-10 mm side), weighing between 1 and 2 grams. It is noteworthy that 31 Frank Sear mentions the use of ―exceptionally small‖ glass tesserae in the octagon of Nero‘s Domus Aurea of 3-4 mm side and 5-6 mm side (Sear 1977, 92, no.62). The tesserae were apparently made in multiple ways. Chunks of raw glass were used to chop off bits and pieces, as the presence of chunks of raw glass in various colours in early Byzantine churches shows, e.g., at Bir 44 Messaoudi, Carthage (TN) (unpublished), Itanos (GR). However, normally discoid glass cakes, i.e., round and oval flat pieces with rounded edges, are correlated with the production of tesserae (Rütti 1991; Foy 2007; Marii, Rehren 2009). That glass cakes have been made for the production of tesserae is again most obvious from the finds in early Byzantine churches, as from the Petra church (JO) (Marii, Rehren 2009, 296). Chunks of raw glass and handy glass cakes were thus concurrently utilized. Figure 18: Tesserae in various shapes and sizes from Les Houis, near Sainte Menehould (FR) (drawings by author, by courtesy of MAN, Saint-Germain-en-Laye) 2.3.3.3 Inlays Inlays are very well known in various polychrome panels, like, for instance, the hexagonal tile from Pompeii (IT) (Beretta et al. 2006, 293, no. 3.19) or the fish from the Revetment Panel at the Corning Museum of Glass (Glass of the Caesars, 31, no. 9). Other mosaic plaques consist of inlays combining polychrome and monochrome parts, such as in the ‗Birds in Thicket‘ at the Corning Museum of Glass (Glass of the Caesars, 32-33, no. 10), or are made from monochrome sheets, such as the Panel with Thomas presumed to be from the Faiyum (EG) at the Corning Museum of Glass (Glass of the Caesars, 34, no. 11) or the remarkable panels from Kenchreai near Corinth (GR). Opus sectile panels were flat sheets that were cut into shape or rather sawn, as grinded edges are not observed, but the panels were also cast in moulds. Inlays and mosaics in opus sectile integrated polychrome and 45 monochrome glass sheets in all sorts of colours, but parts in black glass never seem to have been used. For four years we searched in vain for opus sectile panels made of black glass. Recently, it emerged from the Roman 46 collections of the villa of Lucius Verus, but we have not been able to study the material in detail. Even though the ratio between the panels in black glass and those in other colours is negligible, it is now proven to have been used and more will appear. 2.3.4 Counters In whatever colour the glass counters were made, they all had the distinct plano-convex shape featuring a convex top side, a flat bottom, and a rounded edge [see Chapter 3]. Only on rare occasions has an unevenly pinched off edge been noticed; for instance, on the Oudenburg piece [cat.no.465]. Was the piece intentionally tooled, or is it only due to long use? Concerning the method by which glass counters were made, they are generally described as ‗formed by dropping a small blob of hot glass on to a flat surface‘ (Price 1995, 129) or as ‗made by placing small ―spoonfuls‖ of molten glass to set on a bed of sand‘ provoking ‗a typical sand-roughened underside‘ (Crummy et al. 2007, 186). However, when examining the bottom side of the counters in black glass, two distinct surfaces are noticeable, implying different techniques: the first group has a flat and smooth bottom side (Figure 19a), whereas the second group has a more irregular, pitted surface (Figure 19b). Jennifer Price has already noticed these two groups within a set of 87 glass counters from the fortress at Usk (UK) (Price 1995, 129). 44 Presentation by Coutsinas, ‘Le materiel en verre de la Cité d‘Itanos‘, at the glass conference of the Association Internationale pour l‘Histoire du Verre in Thessalonica, Greece (September 2009). 45 The Panel of Thomas incorporates opaque and translucent glass such as red, yellow, dark blue, turquoise, pink, brown, green, white, purple, and gold sandwich glass (Glass of the Caesars, 34, no. 11). 46 With ‗Glass technology colours, forms and shaping in the 2nd century opus sectile glass materials from the villa of Lucius Verus in Rome‘ Marco Verita communicated on the opus sectile panels from the Gorga collection, Rome at the glass conference of the Association Internationale pour l‘Histoire du Verre in Thessalonica, Greece (September 2009). 32 a. b. c. Figure 19: Two counters from the castellum Oudenburg (BE) a) with a flat and smooth lower surface [cat.no.465]; b) with a slightly irregular, pitted surface [cat.no.466]; c) counter from a re-fused glass sherd from Utrecht (NL) [cat.no.2989] (photographs by author, a-b) courtesy by VIOE, Brussels; c) courtesy by ACVU, Amsterdam) She describes the first group, consisting of more than 20 pieces, with little or no signs of wear, as having shiny surfaces and slightly irregular, pitted basal surfaces. The remaining pieces, forming the second group, have dull, slightly worn, or heavily used surfaces and flat, smooth basal surfaces (Price 1995, 129). Presumably, this variation of the basal surface is due to different technological methods. The irregular, pitted surface on the glass counters can be explained as the result of a sudden thermal shock that the hot and viscous glass underwent when touching a cold(er) working surface (an aspect that also can be observed on mould-blown vessels). The absence of such chill marks on the counters with a straight and very smooth base surface has to be interpreted as the fusion effect by remelting shaped pieces, placed in the furnace on a smooth heat-resistant plate that gradually took the same temperature as the shaped glass sherds. This latter technique is described by Pliny the Elder, who explains the production of black glass counters: Pieces of broken glass can, when heated to a moderate temperature, be stuck together, but that is all. They can never again be completely melted except into globules separate from each other, as happens in the making of the glass pebbles (calculi) that are sometimes nicknamed ‗eyeballs‘ (oculi) and in some cases have a variety of colours arranged in several different patterns. (Pliny the Elder, Naturalis Historia, XXXVI, LXVII. 199 [translation by Eichholz, D.E., 1962, Pliny X, 157]) An example that illustrates best the re-fusion of broken glass into calculi comes from recent excavations at Utrecht (NL) [cat.no.2989], showing at the basal side not only a flat and smooth surface but also a corner of the presumably square sherd that has not melted fully because of a restrained heating temperature as mentioned by Pliny (Figure 19c). That explains also the not-fully-round shape of the counter, as the rounded edge did not reach the bottom to swallow the fused sherd. Strange here is the preservation condition, without any sign of iridescence such as that seen on other re-fused counters. Most probably, the re-fusing of glass sherds into plano-convex calculi or oculi affected the stability of the glass metal. Less resistant to water infiltration, metal oxides dissolved, creating an iridescent corrosion layer at the surface. This resulted in a diluted composition and made such material ill-suited for chemical analysis, or at least the changed conditions should be taken into consideration when choosing to do so. Since the Utrecht counter has remained entirely intact, no sampling has been carried out. Future chemical analysis on the fused counters will determine whether the used recipe is related to the preservation conditions. This would then imply the fused counters with flat and smooth basal surfaces are dividable into two subgroups. A more in-depth research can provide more information in discerning the chronological and geographical variability of glass counters. Within the last sentence of the passage by Pliny, we construe the making of polychrome counters, or, in other words, the remelting into counters of fragments from polychrome vessels. The emerald green glass with yellow circles is prevalent in vessels, in particular shallow bowls like the ones from a tomb at Hollogne-aux-Pierres (BE) (Fontaine-Hodiamont 1994). Such counters have been recorded all over the Roman Empire, e.g., Egypt (Harden 47 1936, 294; Cooney 1976, nos. 831; 843; 846; 855), Hungary (Fülep 1977, 32, grave R/170:5, pl.17:15). Counters in other polychrome glass such as those combining translucent brown with opaque white or translucent purple with opaque yellow are also equally known from vessels. Keeping in mind the remark by Pliny the Elder, it is very likely that broken polychrome vessels were recycled into counters. If so, we can easily imagine a similar application with vessels in black-appearing glass. This hypothesis put into perspective the assessment in the 47 The piece was found together with a set of undecorated bone counters and probably therefore erroneously determined as a yellow and green painted bone counter. 33 attempt to come to a comprehensive overview of black glass so that the impact of black glass production within Roman society can be better understood. While counters were mainly monochrome, bi-chrome and polychrome counters have also been recorded. A purple or olive to bottle green hue can be observed when using light in transmission on counters in blackappearing glass. Sometimes the true hue becomes only visible when a chip has flaked, as is the case with the black glass counter from Tienen (BE) [cat.no.531] (Cosyns et al. 2006b, fig.1b). Some, coincidentally, seem to be made from various translucent glass colours, like a black counter from Nijmegen (NL) (unpublished) [cat.no.4466] (Figure 4a) ranging from yellowish green over pale green to dark bottle green glass. Others are partly in colourless and dark purple glass, like the counter found in the late Roman castellum of Oudenburg (BE) (unpublished) [cat.no.465] (Figure 4b) or the counter from Grobbendonk (BE) incorporating pale blue-green glass (unpublished) [cat.no.290]. The only convincing explanation for this phenomenon seems to be that these fused counters were made by putting various sherds of recycled vessel fragments on top of each other to reach a thickness of 6-7 mm. Some black glass counters have a coloured decoration, making visible the true hue of black- appearing glass. So are the pieces from the very rich tumulus of Cortil-Noirmont (BE) (Lefrancq 1989, B13, no. 30) [cat.no.228] with wavy concentric circles of opaque white glass and purple in the centre, clearly dark purple glass. Considering these pieces as black creates a dilemma regarding how to define the black glass objects. Should we then not integrate in the corpus the large number of bichrome white marbled ribbed bowls in dark purple glass appearing black? The purple hue of the glass utilized to produce these objects becomes visible due to the marvered glass, resulting in layers of translucent black-purple glass covering parts of the white opaque glass. The effect is a reflection of light making the black glass become clearly purple in these areas (Figure 3a). This is also observable on the black-appearing vessels such as the globular cup with white opaque festooned decoration pattern (Figure 3b). Finally, some black counters are reported to have tiny opaque red dots on their convex surfaces, as reported for several counters from the Roman fortress at Usk (UK) (Price 1995, 129-130, nos. 1; 5). Jennifer Price suggested an accidental cause rather than an intentional action to apply a decoration. The rare presence of red opaque spots at random on various black glass vessels confirms the unintentional purpose and is to be explained as minor defects by oxidation of iron or copper/bronze particle on the glass surface [see Chapter 8]. 2.4 Comparative survey of the various colours of black glass in relation to the applied technology 2.4.1 Production 48 A large part of the 204 entries related to production remain indefinite (Table 9). The only waster of a cast object is restricted to a sole piece in green glass. Of the chipped material roughly half is of indefinite hue (14 ex.), and except for two pieces in brown glass and one in grey glass all others are of green (7 ex.) or blue-green (6 ex.) hue. Approximately 10% of the registered production material shows tooling marks. Of three pieces no distinct glass hue could be discerned; one piece from the glass workshop ‗Äussere Reben‘ at Kaiseraugst (CH) is of purple glass [cat.no.3692]; all other are either green (15 ex.) or blue-green (3 ex.). Nearly 1/5 of the production material is molten glass from either a built batch or a crucible. Only six pieces have an indefinite hue; nine entries are of green glass; and the remaining 23 ex. have a blue-green hue. Both re-fused pieces are made out of blue glass and appears to be bichrome. 48 Due to practical reasons we have omitted the architectural decoration and counters, but within the different following chapters we briefly attend to the matter and its impact on production and distribution on these commodities. 34 Table 9: Overview of various types of black glass production material technique quantity percentage indefinite/unknown 111 54,4 cast 1 0,5 chipped 30 14,7 tooled 22 10,8 molten 38 18,6 re-fused 2 1,0 204 100,0 TOTAL 2.4.2 Vessels From the 651 entries related to vessels 580 have been selected for an evaluation of the applied technology in relation to the used glass hue (Table 10). The 71 omitted pieces were removed because: 1) we could not define the production technique (59 ex.), either because we had no detailed information on the item or because the piece was too fragmentary; 2) six entries are pressed or stamped appliqués applied on a vessel as medallion to adorn the start of a handle. Normally such decorative items are applied on free-blown vessels but as there is no certainty we excluded these from the quantitative comparison; 3) the six tooled items are also excluded from the evaluation. These are all handle fragments which are most likely from blown vessels, but too indefinite to be able to determine the adopted technique. Table 10: Overview of various adopted techniques in black glass vessel production vessels shapes individuals cast 13 86 free-blown 38 421 mould-blown 3 24 rod-formed 3 48 57 579 TOTAL The majority of the Roman glass vessels appearing black has been free-blown and embraces 2/3 of the entire vessel production in black glass (Figure 20). Together with the casting technique 90% is covered. Rod-forming and mould-blowing represent two much smaller assemblages. But while the rod-formed vessels were restricted to a specific consumption market within a limited distribution area from only the late Roman period [see Chapters 3; 4 and 6], mould-blowing seems a much lesser practised technique for the production of black glass vessels in comparison with the contemporaneous mould-blown vessels in ‗naturally‘ coloured and decolourized glass, an empire-wide production technique which has been applied on a very large scale. Most likely this observation is to be related with the specificity of the technological properties which are caused by the composition of (black) glass matrix. We therefore considered important to look at the ratios of the various glass hues defined for the production of black appearing glass. vessels 5% 5% 23% cast free-blown mould-blown 67% rod-formed Figure 20: Ratios of the various production techniques for black glass vessels The largest group of the total vessel production is made in green tinted glass (Table 11) and when omitting the pieces of undefined glass hue the vessels in green glass comprise 2/3 of the totality and 3/4 when the blue-green glass is involved (Figure 21). The blue glass vessels are poorly represented, but should be considered at about 10% like the brown and purple glass hue seeing that the black appearing cameo-glass vessels have not been taken into consideration here. 35 Table 11: Overview of various adopted glass hues in the production of black appearing vessels vessels quantity percentage undefined 188 28,9 green 307 47,2 blue-green 44 6,8 blue 11 1,7 brown 40 6,1 purple 60 9,2 red 1 0,1 651 100 TOTAL vessels 0% 6% 2% 0% UNDEFINED GREEN 9% 29% 7% vessels 13% BLUE-GREEN 2% BLUE BLUE-GREEN 9% BLUE 10% 66% BROWN 47% GREEN BROWN PURPLE PURPLE RED RED Figure 21: Ratios of the various production techniques for black glass vessels in relation to the adopted glass hue Compared to the general view, the glass hue of the cast vessels remained undefined to a much larger extent (Table 12). This is due to technological aspects as cast artefacts are much thicker than for instance the freeblown vessels. Blue-green and red glasses have not been recorded to cast black appearing vessels and the segment of the blue and brown hues remain very poor. When omitting the objects with undefined hue the green glass artefacts cover about 1/2 of all cast vessels (Figure 22). Striking however is the much higher ratio of purple cast vessels compared to the general viewcorrespondingly 33% and 13%. It is very tempting to believe at this stage that a number of glass workshops (eventually in a particular region) used solely purple glass to cast vessels, while others used green glass, but it could equally be a chronological feature, where the cast vessels in one hue are younger/older than those in the other hue. Table 12: Overview of various adopted glass hues in the production of black appearing cast vessels cast vessels quantity percentage undefined green blue-green blue brown purple red TOTAL 36 26 0 2 5 17 0 86 42,4 30,6 0,0 2,4 5,9 18,8 0,0 100 cast vessels 0% cast vessels 0% UNDEFINED GREEN GREEN 19% 6% 42% 2% 0% BLUE-GREEN BLUE-GREEN 33% 53% BROWN 31% BLUE BLUE PURPLE PURPLE RED BROWN 10% 4% 0% RED Figure 22: Ratios of the cast black glass vessels in relation to the adopted glass hue 36 By far the largest group of free-blown vessels is made of green glass (Table 13) and when the records with undefined hue are put aside the green glass covers about 3/4 of all free-blown vessels (Figure 23). Most other groups represent each not more than 10%. Due to the very large number of records it is obvious that the obtained ratios more or less correspond to that from all the vessels (Figure 21). It is however a striking coincidence that the black appearance of the much thinner free-blown vessels is in general more easily accepted for the artefacts made of green glass even if the green hue is as observable as the other hues. In the following chapters we will discuss what might the possible reasons be behind these ratios as it can be connected with the technological know-how or the economical factors in choosing for specific recipes to produce more adequate very dark coloured glass to manufacture black appearing vessels. Table 13: Overview of various adopted glass hues in the production of black appearing free-blown vessels free-blown vessels quantity percentage undefined green blue-green blue brown purple red TOTAL 88 241 26 6 27 32 1 421 21,3 57,0 6,1 1,4 6,4 7,6 0,2 100 free-blown vessels 6% 2% 6% free-blown vessels 10% 0% 8% 0% UNDEFINED 21% GREEN GREEN 8% BLUE-GREEN 2% BLUE-GREEN BLUE 8% BLUE BROWN BROWN 57% 72% PURPLE PURPLE RED RED Figure 23: Ratios of the free-blown black glass vessels in relation to the adopted glass hue The number of mould-blown vessels in black appearing glass remains so far very low and is therefore not fully representative (Table 14). Striking fact is the high proportion of blue-green glass (Figure 24). But here again we have to take into account that the black aspect of most mould-blown vessels made in very deeply blue, brown or purple glass is mainly ignored in contrast to the vessels in green and blue-green glass. On the other hand we have to preview what will be discussed on production [see Chapter 6] as all blue-green records come from Augst/Kaiseraugst (CH). The use of a particular glass hue is here related to a workshop. Table 14: Overview of various adopted glass hues in the production of black appearing mould-blown vessels mould-blown vessels quantity percentage undefined green blue-green blue brown purple red TOTAL 9 7 5 0 1 2 0 24 34,8 30,4 21,7 0,0 4,3 8,7 0,0 100 37 mould-blown vessels mould-blown vessels 0% 0% UNDEFINED 4% 0% 9% GREEN GREEN 35% 22% BLUE-GREEN 7% 0% 13% BLUE-GREEN 47% BLUE BLUE BROWN 33% BROWN PURPLE 30% PURPLE RED RED Figure 24: Ratios of the mould-blown black glass vessels in relation to the adopted glass hue The overview of the colour ratios for rod-formed vessels is not representative at all, even though more records are available. The glass hue of 3/4 of the material remains undefined due to the wall-thickness of the rod-formed vessels (Table 15). When the records with undefined hue are dropped, brown seems to be the most frequent glass hue (Figure 25). Table 15: Overview of various adopted glass hues in the production of black appearing rod-formed vessels rod-formed vessels quantity percentage undefined green blue-green blue brown purple red TOTAL 36 3 1 2 6 0 0 48 rod-formed vessels 0% 0% 2% 4% 75,0 6,3 2,1 4,2 12,5 0,0 0,0 100 rod-formed vessels 0% 0% UNDEFINED GREEN GREEN 13% BLUE-GREEN 25% BLUE-GREEN 6% BLUE BLUE 75% 50% 8% BROWN PURPLE 17% RED BROWN PURPLE RED Figure 25: Ratios of the rod-formed black glass vessels in relation to the adopted glass hue 2.4.3 Jewellery This group encompasses a wide range of jewellery types to adorn body and dress and is by far the largest group of black glass artefacts (Table 16). Seeing the variety of techniques adopted to manufacture a wide range of jewellery types the quantified and percentage ratios have been plotted in two different views in order to obtain supplementary information. The first two stacked bar charts show the quantity and percentage composition of the different commodity types for each technique (Figures 26a-b). The two following stacked bar charts show the same amount of pieces but giving an idea of the ratios of the different techniques applied per commodity type (in terms of quantity and percentage) (Figures 26c-d). The plots show that a technique has been applied for the production of a specific commodity type, such as for instance the folding for beadmaking and the gatheringswirling in the bangle production. This is possibly the result of specifically introduced (innovative?) techniques to produce particular artefact types within a particular region or period of the Roman Empire. Accordingly the analysis results on technology of black appearing glass artefacts hold information to define idiosyncrasies within the Roman imperial glass production and consumption. 38 Table 16: Quantitative overview of the various jewellery types for each technique gathered-swirled jewellery undefined cast rod-made rod-formed cone-rolled arm rings 133 0 669 36 613 finger rings 0 1 1 202 0 beads 6 6 24 472 0 pendant 0 17 3 113 0 gems 0 339 0 0 0 hairpins 0 0 42 1 0 139 363 739 824 613 TOTAL + folded total 0 0 295 0 0 0 295 1451 204 803 133 339 43 2973 jewellery folded arm rings gathered-swirled + cone-rolled finger rings rod-formed beads pendant rod-made gems cast hairpins undefined 0 100 200 300 400 500 600 700 800 900 jewellery folded arm rings gathered-swirled + cone-rolled finger rings rod-formed beads pendant rod-made gems cast hairpins undefined 0% 20% 40% 60% 80% 100% jewellery hairpins undefined gems cast pendant rod-made rod-formed beads gathered-swirled + cone-rolled finger rings folded arm rings 0 200 400 600 800 1000 1200 1400 1600 39 jewellery hairpins undefined gems cast pendant rod-made rod-formed beads gathered-swirled + cone-rolled finger rings folded arm rings 0% 20% 40% 60% 80% 100% Figure 26:a-b) Stacked bar charts of the different commodity types per technique (in terms of quantity and percentage); c-d) Stacked bar charts of the different techniques applied per commodity type (in terms of quantity and percentage) Because of the diversity of commodity types jewellery encompass we discuss the applied glass hue for each jewellery commodity seperately. It is important to observe that the very high amount of artefacts with undefined hue is observed for all jewellery commodity types up to 98% of the recorded items (Tables 16-22). A more detailed examination of the material should be considered in order to come to a more correct analysis of the variety of applied glass hues and their ratios. 2.4.3.1 Arm rings Green glass is mostly used for the production of bangles and when taking into account the blue-green glass items 91% of the black appearing is covered (Figure 27). Concerning the six so-called red pieces it is important to notify that this hue is only observed on small fragments from the glass workshop of Les Houis nearby Sainte Menehould (FR) (Cosyns 2009) [see Chapter 6] and never within the main consumer‘s market. An additional argument is that reddish brown to purplish brown colour is, on the basis of close examination, the surface colour only. It is therefore more likely to consider these pieces as production waste. There is no clear connection between the applied glass hue and the adopted manufacturing technique because the rod-made and gathered-swirled bangles show similar ratios as showing in the general pie-chart of Figure 27 (right). In the following chapters we verified whether the concerned glass hues are dependable to specific types or production/consumption regions [see respectively Chapter 3 and 6]. All fragments of gathered-swirled bangles in purple glass are possibly of late La Tène bangles which occur in early Roman contexts in the north-western provinces. Table 17: Quantitative overview of the various glass hues applied for arm rings arm rings quantity percentage undefined green blue-green blue brown purple red grey TOTAL 1021 359 31 11 7 15 6 1 1451 70,4 24,7 2,1 0,8 0,5 1,0 0,4 0,1 99,9 40 1% 1% 1% 2% bracelets 3% UNDEFINED 0% 0% GREEN 3% 2% bracelets 1% 0% GREEN BLUE-GREEN 25% BLUE-GREEN 7% BLUE BLUE BROWN BROWN PURPLE 70% 84% RED GREY PURPLE RED Figure 27: Pie-charts of the colour ratios of the bracelets; (left) with undefined hue incorporated;(right) with defined hues only 2.4.3.2 Finger rings Striking is that, except for one piece in dark blue glass, all identified black appearing finger rings are made of a green glass (Table 18, Figure 28). Table 18: Quantitative overview of the various glass hues applied for finger rings finger rings quantity percentage undefined green blue-green blue brown purple red TOTAL 166 37 0 1 0 0 0 204 0% 1% 0% finger rings finger rings 0% 0% UNDEFINED 18% 81,4 18,1 0,0 0,5 0,0 0,0 0,0 100 3% 0%0% 0% 0% GREEN GREEN BLUE-GREEN BLUE 81% BLUE BROWN 97% PURPLE RED Figure 28: Pie-charts of the colour ratios of the finger rings; (left) with undefined hue incorporated (right) with defined hues only 2.4.3.3 Beads Seeing the very small amount of defined glass hues the obtained ratios are possibly not fully representative for a generalizing idea on the applied glass colours in the manufacturing process of black appearing beads (Table 19, Figure 29). But similar to the bracelets is the green glass predominant and have other glass colours been used to manufacture black appearing beads. There is apparently no correlation between the applied manufacturing technique and the used glass colours, but the folded spacer-beads are in contrast to the cast and rod-formed beads for the most part known in green glass and in a lesser degree in blue-green and brown glass. 41 Table 19: Quantitative overview of the various glass hues applied for beads beads quantity percentage undefined green blue-green blue brown purple red grey TOTAL 731 51 2 9 2 5 1 2 803 1% 0% 7% 91,0 6,4 0,2 1,1 0,2 0,6 0,1 0,2 100 beads 0% 1% 0% 0% beads 1% UNDEFINED 3% GREEN 3% GREEN BLUE-GREEN 7% BLUE-GREEN BLUE BLUE 12% BROWN BROWN 3% PURPLE 71% PURPLE 91% RED RED GREY GREY Figure 29: Pie-charts of the colour ratios of the beads; (left) with undefined hue incorporated; (right) with defined hues only 2.4.3.4 Pendants The representativeness of the here presented overview is insignificant seeing the low number of pendants with recognized hues. It is therefore of no use to link the applied manufacturing technique with the used glass colours. The only clear element is that besides the ubiquitous green glass the black appearing pendants have been made of blue and purple glass (Table 20, Figure 30). Table 20: Quantitative overview of the various glass hues applied for pendants pendants quantity percentage undefined green blue-green blue brown purple red TOTAL 121 8 0 1 0 3 0 133 pendants 1% 0% 0% 2% 0% 91,0 6,0 0,0 0,8 0,0 2,3 0,0 100 pendants 0% UNDEFINED GREEN 6% GREEN 25% BLUE-GREEN BLUE BLUE BROWN 91% PURPLE 0% 67% 8% PURPLE 0% RED Figure 30: Pie-charts of the colour ratios of the pendants; (left) with undefined hue incorporated (right) with defined hues only 42 2.4.3.5 Gems The comparative survey of the gems is impossible because nearly all remain of undefined colour (Table 21). The pie-chart at the right in Figure 31 is only a reflection of 2% of the totality recorded. The items with identified colour are all monochrome cast pieces. Nearly all undefined pieces are the bichrome nicolo-paste gems where an opaque blue glass is topping a black glass layer. The difficulty in observing the true glass hue is because the pieces seldom show cracks or are still sealed within a finger ring. A closer examination is however indispensable in order to enable the colour observations of the pieces. Within the chapters on archaeometric approach we verified whether the hue of the so-called nicolo-paste gems can be determined through chemical analysis and optical spectroscopy. Table 21: Quantitative overview of the various glass hues applied for gems gems quantity percentage undefined green blue-green blue brown purple red TOTAL 331 5 0 1 1 1 0 339 0% 0% 0% 2% 97,6 1,5 0,0 0,3 0,3 0,3 0,0 100 gems 0% 0% gems 0% UNDEFINED 13% GREEN GREEN BLUE-GREEN 13% BLUE BLUE 12% BROWN 98% 62% BROWN PURPLE 0% PURPLE RED Figure 31: Pie-charts of the colour ratios of the gems; (left) with undefined hue incorporated; (right) with defined hues only 2.4.3.6 Hairpins The lower ratio of undefined hue for the hairpins compared to the other commodity types is mainly due to the closer examination of most items we were able to execute at the Musée des Antiquités Nationales (MAN), SaintGermain-en-Laye (FR) and at the Centre de Recherche Archéologique Nationale (CRAN) of the Université Catholique de Louvain-la-Neuve (UCL), Louvain-la-Neuve (BE). Likewise the other jewel types most hairpins are of green glass (Figure 32, Table 22). Striking is that blue-green and brown glass are the sole alternatives seeing the absence of any other glass colour. We will verify in the following chapters whether the recorded hairpins are idiosyncratic to a region and/or period and whether one or more workshops can be identified that produced such material [see Chapters 3 and 6]. 43 Table 22: Quantitative overview of the various glass hues applied for hairpins hairpins quantity percentage undefined green blue-green blue brown purple red TOTAL 16 23 2 0 2 0 0 43 5% 5% 0% 37,2 53,5 4,7 0,0 4,7 0,0 0,0 100 hairpins hairpins 0% 0% 0% UNDEFINED 0% 0% GREEN 7% 8% 37% GREEN BLUE-GREEN BLUE-GREEN BROWN 53% BROWN 85% Figure 32: Pie-charts of the colour ratios of the hairpins; (left) with undefined hue incorporated; (right) with defined hues only 44 Chapter 3 CHRONO-TYPOLOGY If your types do not survive a critical examination, your whole work may lose its meaning. (Crawford 1921, 147) 3.1 Introduction A classification of black glass artefacts was not the goal of the research, but it is essential for an enhanced analysis of all levels within the chaîne opératoire to better understand the consumption of black glass artefacts. Aiming for a useful quantifiable set of types to enable valuable analytical processing a workable labelling system 49 was necessary. This results in a primary workable classifying tool for identification of material groups to establish chronological and regional idiosyncratic subgroups. A classification system of archaeological artefacts always takes into account the external propertiese.g., colour, shape, size and decoration. With the abovementioned phrase by Osbert G.S. Crawford in mind we gave priority to the external properties related to technology to come to a workable categorization of the different functional types in black glass. The variability of this issue seemed most suitable to indicate a regionalism of production and/or consumption (provenance of the object), or a production within a well-defined period (dating of the object). Hence, a taxonomy built on technological features can be particularly helpful to the archaeologist. Conversely, decoration patterns barely illustrate more than the variations of a specific type generated from an applied technology in use within a particular period and/or region. However, decoration techniques are taken into account as they appear idiosyncratic to a certain region, providing information on the success of particular variants of one type in civita(te)s or provincia(e) [see Chapter 6]. It is true that a typology for archaeological artefacts can only turn out to be a useful classification to archaeologists and historians if the division of classes and types are well-considered and definite (Hill, Evans 1972, 232). Besides, the old sore of any classification system is its transience, as stated in the past (Klejn 1982). A vivid polemic clearly exists on evaluating the effectiveness of a typology, but the most logical and interesting criterion for evaluation is its usefulness in predicting correlations or making visible correlations between types and with other artefacts found through excavations (Klejn 1982, 29). We are however fully aware of the limitations to take into consideration solely those artefacts made of black glass in particular considering the large amount of artefacts produced in other glass hues than black glass. We therefore checked all available classification systems for their usefulness within the wide diversity of functional categories of black glass artefacts and proposed new classification systems where necessary. New discoveries from very recent and present-day excavations, particularly from areas where such material has been scarce, have put constant pressure on the taxonomies. The wide variety of black glass artefacts necessitated looking at each functional type (vessels, bracelets, finger rings, beads, etc.) individually to come to a structured overview of typological classifications useful throughout the archaeological approach. Particularly in regard to the bracelets the formulation of a useful new classification system for the benefit of the field worker is challenging. A similar sorting has been done on the finger rings, although the much lower amount of material 50 made it less profitable. But it was certainly not the purpose here to set up random classifications just to produce listings of different shapes in black glass. Where available, we verified the validity of existing typologies and tried to make use of them to integrate the different black glass artefact shapes. For the vessels, the problem has never been that explicit, as we can incorporate them into the wealth of existing classifications and catalogues (Isings 1957; Goethert-Polaschek 1977; Rütti 1991; Cool, Price 1995; Bonnet-Borel 1997; Price, Cottam 1998). Not only limited to the late Roman period, the black glass beads and pendants can give interesting information on the possible continuity of the production and consumption of glass beads and pendants all through the early medieval period, whether it is Byzantine or Merovingian/Anglo-Saxon/.... For the hairpins and counters, an attempt has been made to formulate a limited typological classification, insofar as this was possible. The black glass gems, rods and tesserae have also been reviewed, but a typology is hardly manageable and senseless, due to the very 51 limited variety of types or the specific character of the material. 49 Inevitably we therefore choose to tag each form type with numerical categories. We nevertheless decided to extend this somewhat impersonal system by nominating each form with a description so that the reader can find swiftly for what stands these intangible numbers. Further specifications on the implemented classification system have been discussed per functional type. 50 Compared to the arm rings which were nearly all produced in black glass, many finger rings were produced in hues other than black. 51 Only a more thorough study of the gems, the rods and the tesserae in other glass hues will demonstrate whether it is meaningful to classify these functional types. 45 In Chapter 6 the location and organization of the workshops producing black glass artefacts are discussed to better understand their distribution and consumption. Our account of the workshops responsible for black glass production, therefore, only focused on furnaces, crucibles, tools, production waste, etc., within the issue of the consumption of black glass finished products. Even though the chronology will only be discussed in detail in the next chapter, we considered useful in this chapter to describe already the various shapes of each functional type in chronological order. The five centuries concerning this diachronic study are roughly defined in four consecutive phases of 100 to 150 years: Period I Period II Period III Period IV 1st–mid 2nd century AD mid 2nd–mid 3rd century AD mid 3rd–mid 4th century AD mid 4th–5th century AD 0-150 AD 150-250 AD 250-350 AD 350-500 AD The identified periods are relative phases in history that are not strictly linked to the traditional historical development of the Roman Empire, but are stages in the evolution of a small segment within Roman material culture. The four chronological phases observe the general knowledge on the emergence and disappearance of archaeological glass artefacts established by prior works (e.g., Isings 1957; Riha 1990; Rütti 1991; Cool, Price 1995; Bonnet-Borel 1997; Price, Cottam 1998; Martin-Pruvot 1999). A more thorough discussion implementing the black glass material from dated contexts is presented in a subsequent chapter aiming for a refinement of the dating of this type of material [see Chapter 4]. The underlying objective of this chapter was to establish a chrono-typology providing suitable information to the archaeologist and the museum curator confronted with Roman black glass artefacts. We are persuaded that every type of black glass artefact can grow into a fossil guide. When the typologies developed on the basis of technological and stylistic features are combined with contextual analysis, they can provide a dating criterion. When typologies are combined with distribution maps, it creates the possibility of observing regional patterns and influences among regions. Establishing sequences of types that are idiosyncratic for a specific region and/or period is the most important advantage in building up classification types. The categorization is based on objective observations on the various shapes within each functional type (vessels, jewellery, and counters). Bearing in mind Ian Hodder‘s statement that any scheme of archaeological theory is in itself subject to theoretical critique (Hodder 2002, 80), and Andrew Jones‘ discussion on the effect of cultural conditioning on the description and categorization of archaeological objects (Jones 2002, 3-6), the chrono-typologies presented here are inherently open to question and improvement. 3.2 Vessels 3.2.1 Introduction st rd A first synopsis encompassing 10 shapes discussed 29 black glass vessels dated between the 1 and the 3 century AD (Isings 1964, 176-179). For the most part, these were from the Benelux and the German Rhineland, but they included some Italian examples from the Gulf of Naples. This paper basically was a reassessment of what Clasina Isings already discussed earlier in stray notes in her major publication ―Roman glass from dated finds‖ (Isings 1957, Forms 1; 14; 25; 26; 32; 36b; 39; 42d; 88b; 94). Unfortunately, some vessels were erroneously defined: - The carchesium from Blehen (BE) is made of pale blue glass; 52 The black fragment of an indented beaker from Mettet (BE) is not of the early Form 32 The small jug from Schaarbeek (BE) has been defined as Form 88bi.e., jugs with trefoil rim and without base-ring. Conversely, the Schaarbeek piece has no trefoil rim and has a base-ring. Furthermore, Form 88b jugs have ovoid or globular bodies, whereas the Schaarbeek jug has a carinated body like a carchesium (Isings Form 36b and here Form IIB.1). Even though it concerns a unique piece, at present it 52 The Mettet piecediscussed here as a Form IIB.5 variantrather resembles the metalescent beaker type Gose 199 what corresponds with Isings Form 94. 46 is considered a separate type, Form IIB.2, whereas the characteristic Isings Form 88b jugs are described as Form IIB.3. When only considering the free-blown vessels of Period II from within the north-western provinces, 14 vessel shapes have been recorded from 41 sites (Cosyns, Hanut 2005), and since then other sites have provided more examples and more shapes have emerged (Cosyns et al. 2006a). Upon examining the Period I vessels in the western provinces, we attached to the list 8 cast vessel shapes and 13 free-blown shapes (Cosyns, Fontaine 2009). A number of common shapes that were so far not known to have been produced in black glass – such as the aryballos, the grape-shaped amphoriskos, the ink-well, and the square bottle – have recently been incorporated in the database, mainly supplied by unpublished material from Nijmegen (NL) and Augst/Kaiseraugst (CH). The here proposed classification of vessels is not to be considered as a typology but rather a listing of the different shapes sorted per technology or per period according to the major general works on glass (Morin-Jean 1913; Isings 1957; Goethert-Polaschek 1977; Rütti 1991; Price, Cottam 1998; Pirling, Siepen 2006, 239-304). The subdivision per technological class is done by using capital lettersA (cast), B (free-blown), C (mould-blown), and D (rod-formed). To come to a clear reading, a further subdivision is done per consecutive period by assigning a Roman numeral (I, II, III or IV), depending on the period corresponding with the form, as explained in the general introduction. Because of practical reasons and the continuous accumulation of material we were not able here to present lists per function, making that cups, dishes, plates, jugs and flasks are not presented one after the other. We are aware that this amalgamation gives a somewhat chaotic overview that needs reviewing. 3.2.2 Cast vessels Cast vessels in black-appearing glass appear to be solely manufactured in Period I and nearly all during the JulioClaudian dynasty [see Chapter 4]. We were able to define 13 different shapes which can be nine can be catalogued as cups and bowls (forms 1a; 2; 3; 4b; 5; 6; 11-15); dishes (forms 1b; 4a; 10); plates or stands (forms 7; 8); and a boat-shaped pyxis (form 9). These were generally made of a green glass, but purple, brown and blue coloured glass was also used to produce black glass vessels. The resemblances in shape and in production technique remind of the contemporaneous stone vessels and in some extend also the colour (obsidian vessels) (Gasparri 1999) what raises questions such as: Were the stone vessels and those in cast and wheel-cut glass made in the same workshops or even by the same workmen? Or were all workshops working separately with a 53 particular material but using the same techniques? 53 We have to notice the equal similarities in shape with the metal ware (silver and bronze) and the pottery (particularly the terra sigillata). The study of skeuomorphism can help to verify the possible implications of the consumers‘ mentality in approaching specific shapes in different materials and its impact on the Roman material culture. Such approach needs an extensive research we were unfortunately not able to fulfil within the present research project. 47 1a. 1b. 5. 2. 6. 10. 3. 4a. 7. 11. 4b. 8. 12. 9. 13. 14. Figure 33: Period I cast vessel shapes in strongly coloured glass appearing black (Forms IA) 3.2.2.1 Form IA.1: Shallow hemispherical cup or bowl without foot [Isings 1/18; AR 3; Trier 5-6; Cosyns, Fontaine 2009, 1] The hemispherical cups and bowls in black-appearing glass conform to the examples in other monochrome vividly coloured glass. The black examples are made from purple or olive-green glass. This open hemispherical vessel shape is shallow, with a height averaging between 45-50 mm. Both variants, derived from the (rim) diameter, occur in black glass (Figure 33.1a-b): a) cups with a small diameter around 80 mm (Isings Form 1), like a piece from House II, Pompeii (IT) with a diameter of 82 mm and a height of 45 mm, and b) cups with a large diameter (Isings Form 18), like, for instance, the pieces from House VI, Pompeii (IT) with a diameter of 200 mm and a height of 50 mm, and from Avenches (CH) with a diameter of 160 mm (Amrein 2001, 71, pl. 7, no. 145). The shape of variant a and that of the free-blown hemispherical cups (see Form IB.11), or the so-called ‗Hofheim‘ cups, are very much alike, but the cast pieces have a thicker wall and are more shallow. Normally having a convex base without a foot, these vessels do have an annular base-ring on rare occasions, as, for instance, the Pompeian piece from House VI [cat.no.2765] (Beretta et al. 2006, 211, no. 1.28). Most often, the body is decorated with wheel-cut concentric lines. The Pompeian cup from House II [cat.no.2767] (Beretta et al. 2006, 207, no. 1.17) has a wheel-cut line on the outer side just under the rim, which is cut and ground straight. The piece from House VI, on the other hand, has two wheel-cut parallel concentric lines on the outer side halfway the bodyone deep and one superficial – and one on the inner side just below the rim. The rim is cut straight but ground round. 3.2.2.2 Form IA.2: Shallow bilobated cup with base-ring [Isings 2; AR 6.1/6.2; Trier17/7; Cosyns, Fontaine 2009, 2] A convex body with strong constriction slightly below the rim gives the cup a bilobated shape (Figure 33.2). Other characteristics are a horizontal or obliquely everted rim and a short diagonal base-ring. Known in polychrome glass, this vessel type is also manufactured in monochrome strongly coloured opaque glasse.g., white, red, deep blue, deep green and black-appearing glass. Only olive-green to brownish glass is utilized to manufacture black-appearing vessels. The shape has a noted counterpart in terra sigillata (Dragendorff 27) and in silverware. 48 The examples in black glass are mainly fragmentary, but they are identical to those in other glass hues, with a height ranging around 40-50 mm and a foot diameter of 35-40 mm. The ratio between the rim diameter and the height is about 3:1, whereas that between the foot diameter and the height is about 1:1. The rim diameter of this form averages between 90 mm and 140 mm (Trier 17), but pieces with a wider rim diameter (Trier 7) from 160 mm up to 230 mm occur occasionally (Price, Cottam 1998, 49). 3.2.2.3 Form IA.3: Shallow hemispherical cup or bowl with base-ring [Isings 20; AR 8; Trier 18; Cosyns, Fontaine 2009, 3] This shallow, open shape is a bowl on an annular foot with a hemispherical body and a thickened rim bulging outwards (Figure 33.3). Mainly reported as black, the hue of the glass is apparently always olive-green when detected, e.g., Magdalensberg (AT) (Czurda-Ruth 1998, no.433; 2001, 426, no.1; 2004, 266; 293, nos. 40-41). The diameter of the foot is about the same as the height of the vessel, while the diameter of the rim is 2 to 3 times the height. The average size of this vessel type ranges around 40 mm (h.), 115 mm (d. rim), and 40 mm (d. foot). The black glass example from Adria (IT) [cat.no.2587] is somewhat larger, measuring 58 mm (h.), 118 mm (d. rim), and 58 mm (d. foot) (Bonomi 1996, 155, no. 345; Diani 1998, 34). The hemispherical bowl from Richebourg (FR) [cat.no.A4491] (Arveiller-Dulong et al. 2003, 147, fig.2) is higher in proportion to the diameter of the rim and of the foot, resulting in a narrower bowl with a more vertical side. The piece in the Newark Museum is 33 mm high with a rim diameter of c. 70 mm and a base diameter of 30 mm (Auth 1976, 193, no. 297) [cat.no.A4492]. This vessel shape is also produced in the blowing technique (Isings Form 42), in the majority by using blue-green glass. 3.2.2.4 Form IA.4: Shallow cup or dish with rounded rim and basal ridge [Isings 22; AR 9.1; Trier 20; Cosyns, Fontaine 2009, 4] This shape is a shallow cup or dish, depending on the small or large size of the diameter (Figure 33.4). Characteristic are the projecting rounded rim and basal ridge, and the short cylindrical body averaging 25-30 mm in height, although some have a small diameter and are proportionately taller. The side can sometimes be slightly convex but is in general slightly tapering downward, like the example from the Gulf of Fos (FR) [cat.no.A4476], measuring 27 mm (h.), 65 mm (d. rim), and 61 mm (d. foot) (Foy, Nenna 2003, 237, no. 27). The cups have a proportion between the diameter and the height of 2:1, whereas that of the dishes is 6:1. Very popular in various strongly coloured monochrome glass, the black-appearing examples recorded are all in a green, olive-green or brownish glass. This shape shows moulded concentric circles and/or a raised disc with a central dot on the base, mainly at the outside but sometimes at the inside. This vessel shape is also produced in the blowing technique, in the majority by using blue-green glass (Isings Form 41a for the cups and Form 48 for the dishes). 3.2.2.5 Form IA.5: Shallow hemispherical bowl without foot [Isings 25; Cosyns, Fontaine 2009, 5] The baseless bowl with two vertical handles from Pompeii (IT) (Figure 33.5) is decorated with inlays in gold, silver, bronze and cornelian (Isings 1957, 39-40), and is reminiscent of the obsidian skyphoi and patera with semi54 precious incrustations from Castellamare di Stabiae (IT) (Haevernick 1963). This shallow hemispherical cup in black appearing glass is a unique piece of 57 mm high with a wide rim diameter of 193 mm. Another pristine piece made in black appearing purple glass must have been the vessel said to have come from in or around the imperial villa of Tiberius on Capri (IT) and now in the Metropolitan Museum of Arts, New York (Winfield 1949, 59; Goldstein et al. 1982, 102) [cat.no.A4695]. The rim fragment of this dish or plate which is decorated with marine creatures in opaque white and translucent pinikish glass has an estimated diameter of 527 mm. 3.2.2.6 Form IA.6: Cylindrical cup with vertical base-ring and winged handles [Isings 39; Trier 29a; Cosyns, Fontaine 2009, 6] Skyphoi are sometimes made in strongly coloured glass but are mainly made in colourless glass. Up to now, one single example of a cast skyphos in black appearing glass has been recorded. Coming from Pompeii (IT), it formed part of the antiquarium (no. 771/4) (Isings 1957, 56). Typical features are the cylindrical body with vertical 54 Not being able to study the material on location, we incorporated the piece in the list on behalf of Clasina Isings‘ description reporting the piece is made of black glass even though it is not excluded that the vessel is made of obsidian. 49 rim, flat base with a vertical base-ring, and winged handles (Figure 33.6). Some pieces are said to be in obsidian like for instance those from the villa of Boscoreale (IT) but it is not impossible that they are made of black appearing glass (Cagno et al. 2010, 9-10). Skyphoi in black glass are also manufactured with the free-blowing technique (see Form IB.7). 3.2.2.7 Form IA.7: Tripod table [Cosyns, Fontaine 2009, Form 19] This vessel shape is a tripod with a quadrangular or circular tablet (Figure 33.7). This shape needs to be considered together with the plates of Form IA.8, seeing that small body fragments of this type can easily be confused with the other. The legs are not rendered with decoration, but the similarity to the silver examples with lion‘s paws is apparent as discussed by Michel Feugère on the tripod table from Cruzy (FR) (Feugère 1987, 62). 3.2.2.8 Form IA.8: Rectangular tray or plate [Cosyns, Fontaine 2009, Form 20] Flat, rectangular plates decorated with outsplayed rim and two handles (Figure 33.8) are well known from various sites in the Mediterranean in strongly coloured glass, while only two have hitherto been reported as appearing blacki.e., the piece from Adria (IT) [cat.no.2584] (Bonomi 1996, 203, no. 452) and the fragment from the urban villa of ‗de Hondstraat‘ in Tongeren (BE) [cat.no.565] (unpublished). Both pieces are made from a deep purple glass. 3.2.2.9 Form IA.9: Boat-shaped vessel Boat-shaped vessels are known in various monochrome strongly coloured glasses (Harden et al. 1987, no. 24) (Figure 33.9). The one from Santa Elena di Melma, Silea (IT) in deep purple glass has a black appearance (Casagrande, Ceselin 2003, no. 244, Barovier et al. 2003, 71, no.8) [cat.no.2805] as does the one at the British Museum (Tait 1995, no.69) [cat.no.A4591]. Characteristic are the four small triangular-shaped feet, quite centrally positioned, and the wheel-cut decoration. The prow of the Silea piece is broken off, but other examples demonstrate it to be pointed and only slightly gradient, whereas the stern is rounded and erected. 3.2.2.10 Form IA.10: Wide rimmed bowl or shallow plate with base-ring [AR 15/24; Trier 25; AV V23/32] These bowls have in general an almost horizontal, wide rim, shallow convex body and high vertical base-ring (Figure 33.10). A piece from Avenches (CH) [cat.no.3611] in dark olive-green glass appearing black has a diameter of 200 mm, a horizontal rim of 12 mm width and 2,5-3 mm thickness with a thicker rounded edge and a wall thickness of 1,5-2 mm (Bonnet-Borel 1997, 23, pl.5, no. 23). Because the bowls determined as AR 24 by Beat Rütti (1991, II, 245, pl.33; 251, pl.39) and AV V32 (Bonnet-Borel 1997, 23; 25) show minor differences we consider the incomplete piece from Bahnasa (EG) that is now in the Petrie Museum for Egyptian Archaeology, London [cat.no.812] (unpublished) rather as a variant with its somewhat anomalous wide everted rim and low or absent constriction. These bowls and plates have generally been cast in polychrome or in deeply coloured and decolourized monochrome glass. Decolourized glass is mainly used for the production of the well-known variant with faceted decoration on rim and body (elliptical and circular). 3.2.2.11 Form IA.11: Shallow dish with wide overhanging rim [AR 16; Trier 23] Only one example of this shallow dish with wide overhanging rim has been recorded in black- appearing glass [cat.no.4145] (Figure 33.11). It is the London piece reported by Jennifer Price and Sally Cottam as a very dull brown glass (Price, Cottam 1998, 56, fig.13:a). This shape is particularly known in colourless glass and can be linked to a wide variety of colourless cast vessels, which often have a faceted decoration. 3.2.2.12 Form IA.12: Shallow bowl with high base-ring This shallow bowl or small saucer has a high foot within the flat or slightly concave wide base and the straight everted body has a rounded-rim edge (Figure 33.12). These bowls are less than 20 mm in height and have a rim diameter between 100-120 mm. The only piece recorded in black appearing glass comes from Ribnica (HR) [cat.no.3173] (unpublished) and is made in a greenish brown glass. The vessel with high diagonal base-ring has an estimated rim diameter of 120 mm, a height of 15 mm, and a wall thickness varying between 2,5 mm and 3,0 50 mm. A similar but polychrome piece has been reported from Piercebridge (UK) (Price, Cottam 1998, 55, fig.12:c; pl.1:4) which is made of very dark purple glass decorated with opaque white roundels. 3.2.2.13 Form IA.13: Deep conical bowl with base-ring Comparable to the Form IA.12 but with high conical body, occasionally decorated with an abraded concentric line below the rim or an applied glass trail (Figure 33.13). The base-ring can start from the junction of the body and the base or begin within a wide base. The rim diameter ranges between 150 and 200 mm and the height between 40 and 60 mm. Similar vessels have been produced in decolorized glass, like for instance the one from Segontium (UK) (Price, Cottam 1998, 55, fig.12:b) and resemble the early terra sigillata cups Form Haltern 10A (Von Schnurbein 1982, II, pl.63, nos.1528-1530). The fragment from Augst/Kaiseraugst (CH) [cat.no.3585] is made of a deep green glass and has an estimated rim diameter of 180 mm. The London piece [cat.no.4146] is too fragmentary to determine its rim diameter but with a height of 35-40 mm without footstand and an abraded concentric line just below the rim the piece with horizontal base and straight everted body can be classified as a deep conical bowl with base-ring. The London piece is thick-walled with a wall thickness ranging between 2,6-3,3 mm whereas the Augst piece is very thin-walled with 0,9 mm. 3.2.2.14 Form IA.14: Ribbed bowl [Isings 3; AR 2; Trier 3] The ribbed bowl or so-called pillar-moulded bowl is a thick-walled shallow hemispherical bowl without foot very distinctive with parallel protruding ribbings tapering towards the base (Figure 33.14). We have attached the ribbed bowls with the vessels made with the cast technique for practical reasons even though we are aware of the very convincing hypothesis that the Romans probably have produced this vessel shape rather by pinching and sagging a discoid plaque of hot glass. The pieces in black glass show no difference from the ubiquitous vessel shape produced in other glass hues. Most of the ribbed bowls are made in the ‗naturally coloured‘ transparent pale blue and pale blue-green glass but strongly coloured ribbed bowls, polychrome and monochrome, are also commonly present on early Roman sites. The major publications have not described black-appearing examples thus far, but the occurrence of two monochrome pieces at Liberchies (BE) [cat.nos.417; 425] and one monochrome piece at Augst (CH) [cat.no.3474] demonstrate that this is probably due to the perception of colours. The ones from Liberchies are made of deep greenish brown glass, and the Augst piece is from deep blue glass. 3.2.3 Free-blown vessels Free-blown vessels made of black appearing glass have been produced in the successive Periods I and II. Within Period I 19 shapes have been cataloguedcups/bowls (4 types); goblets/beakers (2 types) unguentaria (2 types); nd rd jugs (4 types); lids (1 type). The majority of the black glass vessels from Period II (mid 2 –mid 3 century AD) are more or less limited to tableware. The entire gamut of vessels in black glass remains more or less exclusively for 55 drinking use: beakers, cups, bowls, jugs and jars, though, toilet waree.g., unguentaria and aryballoiand storage vesselse.g., square bottles and the so-called inkwellshave been recorded as well. Still, nearly. Thus far, 20 free-blown forms in a monochrome black glass have been distinguished. In great contrast to that stands the total absence of the cast vessels and the near non-appearance of mould-blown vessel shapes. Besides the plain vessels, some shapesForms IIB.4, IIB.5, IIB.12 and IIB.14comprise examples with a decorative pattern created either by applying a glass trail horizontally on the shoulder and/or on the lower part of the body, or by working out the trails into a festoon pattern (―Federmusterzier‖ in German). The trails of the former style remain unmarvered, whereas for the latter, the trails are always marvered into the core glass and combed upward, vertically or obliquely. The horizontally applied trails on the shoulder and/or at the lower part of the body can be in the same black colour as the body or in a contrasting white or blue opaque glass. Festoon patterns are always made from opaque white glass trails. The applied glass trail decoration creates a sharp contrastive effect with the black glass of the body. These opulently decorated vessels stand in great contrast with the very sober shapes such as the carchesium (Form IIB.1) or the globular cup (Forms IIB.4 and 5). The appearance of the decorated black glass vessels might be 55 The here discussed solid-bodied ‗unguentaria‘ are rod-formed and made from black-appearing brown glass [see Chapter 2 and this chapter 3.2.5. rod-formed vessels]. They seem however closer to the so-called linen-smoothers or ―Glättsteine mit Stangengriff‖ (Rütti 1991, I, 164-166) than to most solid bodied toilet wareidiosyncratic for the southeast Mediterranean (Harden 1936, 272; 275-277; Cosyns et al. 2006a, 35, tab.1, fig.4), which is nearly always free-blown. Furthermore the dark aspect of the latter group is due to the thickness of deep blue-green to green glass used to produce solid-bodied vessel. 51 seen as part of a trend towards more elaborately worked glass vessels, or as an attempt to catch the attention of consumers not very much attracted to the elementary forms of sober tableware. Hence, it is not surprising to find these decorated vessels in richly furnished tombs [see Chapter 4]. Many examples with trailed decoration date nd rd back to the late 2 –mid 3 century AD and are typical for the north-western European continent. The applied glass trail on these vessels consists of a concentric line on the shoulder and on the lower side of the body. No 56 opaque red trails have been reported for this period and in this region. The red colour is typical for rod-formed th th vessels of the Levant in Period IV from the late 4 to 6 century AD. Period I 1. 2. 8. 9. 16. 3. 4. 10. 11. 17. 5. 6. 12. 18. 7. 13. 14. 15. 19. Figure 34: First century AD free-blown vessel shapes in strongly coloured glass appearing black (Forms IB) 3.2.3.1 Form IB.1: Hemispherical cup [Isings 12; AR 34; Trier 30; Cosyns, Fontaine 2009, Form 7] Well known in blue-green and various deeply coloured glasses, this type of hemispherical cup is generally known as a ‗Hofheim‘ cup, from the Roman military camp in the German Rhine region where it was recorded for the first time. This drinking vessel has a flattened or slightly concave base without a foot (Figure 34.1). The upper part of the body can be cylindrical or tapering towards the rim. The rim is cut and ground, and the body is decorated with abraded concentric lines. The intact Pompeian piece from House I, of Lesbianus [cat.no.2769] is 61 mm high and has a diameter of 76 mm. Its rim is cut and ground obliquely inward. Just below the rim, the outer side is adorned with one shallow abraded concentric circle, whereas on the lower body, three were applied, one fine line on either side of a wide line (Beretta et al. 2006, 229, no.2.3). The almost complete cup from Polhov Gradec (SL) is 85 mm high (Ložar 1938, 93, fig.16; Lazar 2003, 35; Steps into the past 2006, 109, no. 3, fig.83). The piece from Augst (CH) is made of deep purple glass (Rütti 1991, 53, pl. 48, no. 1104), whereas the one from Olbia (FR) is blown in a deep brown glass (Fontaine 2006, no. 130). The Pompeian cup is said to be of a deep blue glass. Enamelling 57 has hitherto not been recorded on ‗Hofheim‘ cups in black glass (Rütti 1988, 46-52). 56 Red opaque glass trails are only known on jewellery in the North-western provinces of the Roman Empire. Enamelling is known from the amphoriskos in black appearing green glass from Kertch (UKR) and now in the collections of the Hermitage Museum, St. Petersburg (Kunina 1997). 57 52 3.2.3.2 Form IB.2: Small globular jug or flask [Isings 14; Cosyns, Fontaine 2009, Form 8] This small jug or flask has a globular body with a concave base without a foot, and one vertical handle from shoulder to halfway down a narrow cylindrical neck (Figure 34.2). The rim is diagonally folded outward, up, and back inward. The pieces from Locarno (CH) (Simonett 1941, 149, fig.128, no. 18; Isings 1957, 31), and Montebelluna (IT) [cat.no.2761] (Casagrande, Ceselin 2003, no. 147) are plain. However, the body can be ribbed or decorated with applied glass trails in a contrasting colour, like, for instance, the intact piece of the former Constable-Maxwell collection, which has 14 spiral trails in unmarvered opaque white glass from the base to the shoulder (Smith 1957, no. 57). The Montebelluna jug is 80 mm high and has a maximum diameter of 66 mm. 3.2.3.3 Form IB.3: Large globular unguentaria [Isings 16; AR 141; Cosyns, Fontaine 2009, Form 9] Unguentaria with globular bodies, flat bases, and constrictions at the lower end of their necks are mainly known in blue-green glass but also have been recorded in deeply coloured glass (Figure 34.3). Only two of them are recorded as appearing black: one piece came from a mid 1st century AD context in Este (IT) [cat.no.2743] st (Zampieri 1998, 34, no. 6), and the other was found in a 1 century AD level of ‗Hanghaus I‘ in Ephesus (TR) [cat.no.3896] (Czurda-Ruth 2007, 90, no. 247). Both pieces are made from deep purple glass. 3.2.3.4 Form IB.4: Small globular or pear-shaped unguentaria [Isings 26a; AR 132; Trier 70b; Cosyns, Fontaine 2009, Form 10] This vessel is an unguentarium, featuring a globular or pear-shaped body and a short, cylindrical neck (Figure 34.4). The only known black piece so far comes from a mid 1st century AD context in Este (IT) and is now in the Museo Atestino [cat.no.2744] (Isings 1957, 40). 3.2.3.5 Form IB.5: Cantharus on stemmed foot [Isings 36a-b/38; AR 91/96; Cosyns, Fontaine 2009, Form 11] This drinking vessel, with or without small applied handles, has a knobbed stem (Isings 36a), like the piece from Pompeii (IT) (Cibi e Sapori a Pompei, 58, no. 45) [cat.no.2770], or a short-stemmed foot (Isings Form 38), like the piece in deep purple glass from el-Ashmunein (EG) and now in the Louvre Museum (Arveiller-Dulong, Nenna 2005, no., 765) [cat.no.856] (Figure 34.5). 3.2.3.6 Form IB.6: Modiolus [Isings 37; AR 90; Cosyns-Fontaine 2009, Form 12] This high cup with a straight body tapering towards the foot has a characteristic small, round handle (Figure 34.6). One single piece, coming from Avenches (CH), is known in olive-green glass appearing black and said to come from a 40-70 AD context (Amrein 2001, 74, pl. 9, no. 160) [cat.no.3600], but due its fragmented state it is not excluded that the piece formed part of a cylindrical cup with concave neck and applied ribbing just below the everted rounded rim [see Form IIB.17]. 3.2.3.7 Form IB.7: Skyphos [Isings 39; Trier 29a; Cosyns, Fontaine 2009, Form 6b] These cylindrical cups on base-ring with two vertical handles modelled by pincers are similar to those made by casting (Figure 34.7). At present, only handle fragmentsall in black/green glassare known, coming from 58 northwestern sites such as one from Waldorf (DE) (unpublished) [cat.no.2107] and two Liberchies (BE) (Liberchies VI) [cat.nos.407; 413]. 3.2.3.8 Form IB.8: Jug with ovoid body and decorative handle [Isings 54; Cosyns, Fontaine 2009, Form 13] This is a jug with an ovoid body, long cylindrical neck, and a vertical handle with stamped decoration at the base in a contrasting colour (Figure 34.8). Two examples in black-appearing glass have been recorded so far. One jug in deep purple glass was retrieved from the Period I cemetery at St-Paul-Trois-Châteaux (FR) (Foy, Nenna 2001, 58 Personal communication with Anna-Barbara Follmann-Schultz 53 126, no. 161-13) [cat.no.A4483], and another in deep blue glass comes from Pompeii (IT) (Beretta et al. 2006, 217, no. 1.47) [cat.no.2772]. 3.2.3.9 Form IB.9: Beaked jug with vertical handle [Isings 56b; Cosyns, Fontaine 2009, Form 14] This vessel is a jug with a beaked rim, ovoid body tapering towards the concave base without foot and vertical handle with thumb rest (Figure 34.9). Examples of this type of jug include one in deep blue glass in the Louvre museum and reportedly coming from Italy (Arveiller, Nenna 2005, 47, no. 51), and another in deep purple glass from Pompeii (IT) (Maiuri 1971, 131, no. 119b). 3.2.3.10 Form IB.10: Bulbous jar with vertical handles [Cosyns, Fontaine 2009, Form 15] This vessel is an urn with a folded rim and, when with handles, short vertical handles in a contrasting colour (Figure 34.10). Two examples in black-appearing glass have been recorded, one from Murviel-les-Montpellier (FR) in purple glass (Foy, Nenna 2001, 128, no. 162-4) [cat.no.A4480], and another from Nîmes (FR) (Foy, Nenna 2001, 206, no 374) [cat.no.A4479]. 3.2.3.11 Form IB.11: Bulbous jar with horizontal handles [Isings 64; Cosyns, Fontaine 2009, Form 16] This jar has a spherical body, a short neck, and an outward-folded rim with a groove on the inside to hold a lid (Figure 34.11). Characteristic are the horizontal handles. The handle on the Dürres-Durrachium fragment is made from opaque white glass and is decorated at the base with a medallion appliqué of 27 mm in diameter representing a female theatre mask (Tartari 2004, 33; 68, tab.XIX-V69:3). 3.2.3.12 Form IB.12: Hemispherical bowl with tubular rim and base-ring [Isings 69b; MJ 85; Cosyns, Fontaine 2009, Form 17] This vessel is a bowl with hemispherical body decorated with a tubular rim and tubular foot ring (Figure 34.12). One black piece comes from tomb 34 of the cemetery at Canal Bianco (IT), which is ascribed to the first half of st the 1 century AD (Bonomi 1996, 164, no. 372; Diani 1998, 34). 3.2.3.13 Form IB.13: Shallow hemispherical cup with vertical folded rim This vessel is a shallow hemispherical cup with a pronounced, outward-folded, vertical rim and a concave base without a foot (Figure 34.13). Only two examples have been attributed to this shape so far. One piece is the small cup from tomb 1 of the cemetery at Montebelluna (IT), decorated on the shoulder with 7 spirally wound trails in 59 opaque white glass (Casagrande, Ceselin 2003, 134, no. 169) [cat.no.2759]. The other piece comes from tomb 147 of the Canal Bianco cemetery at Adria (IT) (Bonomi 1996, 127, no. 290) [cat.no.2586]. This cup is more stretched out into a conical shape, and it is accordingly higher and more slender, but it seems to have been produced in the same way and with the same amount of hot glass. The rim is a bit more everted, and its decoration consists of unmarvered thick grit in opaque white glass on the lower body. Both pieces are made of purple glass. Examples in other glass hues demonstrate that the characteristic shape of the rim is a fingerprint of one particular glassblower and thus the product of one single glass workshop. The unmarvered decoration of grit in a contrasting colour should also be considered a feature of this glassblower/glass workshop (Harden et al. 1987, no. 43). 3.2.3.14 Form IB.14: Jug with high vertical handle [Isings 13] This form type incorporates several variants of jugs with a high vertical handle only recorded thus far in northern Italy, but which may be considered as the forerunners of the globular- bodied (Isings Form 52) and triangularnd bodied (Isings Form 55) jugs [see Form IB.20] known from Flavian times to the mid 2 century AD, and mainly 60 produced in ‗naturally coloured‘ glass. The jugs can have a somewhat cylindrical tapering downwards 59 This piece awkwardly has been linked to Isings Form 68, which is a toilet vessel with a horizontal folded rim and neck produced in blue-green and pale blue glass. 60 Compared to their successors, the early types apparently are not optic blown. 54 [cat.no.2758] or a biconical body with pronounced carination [cat.no.2762] (Casagrande, Ceselin 2003, nos. 140; 142) (Figure 34.14). All are made of very dark purple glass and described as such, but are not considered black61 appearing. Both examples come from the cemetery at Montebelluna (IT) and have a concave base or an open base-ring, a tall narrow cylindrical neck, and a small, diagonal, outward-folded rim. This type of jug is about 140150 mm high, and the maximum diameter ranges around 80-90 mm for the cylindrical variant and around 115-150 mm for the carinated variant. 3.2.3.15 Form IB.15: Domed lid with knob [Isings 66a; Cosyns, Fontaine 2009, Form 18] This shape is a domed lid with a knob handle in the centre (Figure 34.15). Although it is not a vessel per se, a lid is considered together with the vessel with which it forms one set – e.g., the cast jar on a stemmed foot at the Archaeological Museum of Turin [cat.no.2824], the jar with globular body (Form IB.11) from Nîmes (FR) (Foy, 62 Nenna 2001, 206, no. 373) [cat.no.A4478], and the cantharus from Vaison-la-Romaine (FR) (Foy, Nenna 2001, 173, no. 270) [cat.no.A4481]. Lids are mainly produced in blue-green glass, but the pieces from Nîmes and Vaison-la-Romaine are made of very deep purple glass appearing black. The knob is known in various shapes. Clasina Isings described four different types (Isings 1957, 85-86). The lid at the Turin museum is a knobbed lid (variant c) with abraded decoration, while the two French lids are domed lids with a knob on top (variant a). The height of a lid varies between 55 and 75 mm, and the diameter varies according to the dimensions of the vessel for which it is produced. The lid from Nîmes is 75 mm high and 136 mm in diameter. 3.2.3.16 Form IB.16: Shallow dish/bowl with rounded rim [Isings 42-47] This shallow dish/bowl is a blown variant of the cast Isings Form 5 which is characterized by a horizontally outward folded rounded rim (Figure 34.16). The only piece in black appearing glass catalogued so far is a fragment in deep blue glass at the Petrie Museum of Egyptian Archaeology, London which apparently comes from Abydos (EG) [cat.no.794]. Its diameter is 165 mm and the preserved height is c.30 mm showing only the curve of the convex body. It appears to be a rather shallow bowl that must approach 40-50 mm. Considering the absence of the base it is impossible to determine whether the vessel was with or without foot; and if with foot, whether its base consisted of a tubular base-ring or a solid one. Closest comparative shapes are the shallow bowl (Isings Form 42) and the dish on base-ring (Isings Form 47). 3.2.3.17 Form IB.17: Large bulbous jar with wide horizontally folded rim [Isings 67a; Trier 147] This vessel shape is mainly known in ‗naturally coloured‘ glass from burials using the glass vessel as cremation urn. The known examples are made in either very dark blue or in deep purple glass, like those from Rochefleur (FR) [cat.no.4144] and Santelpido (IT) [cat.no.4144] (Figure 34.17). The black appearing vessels are generally decorated with large opaque white spots, which are marvered and clearly enlarged through blowing like for instance the pieces from Arles (FR) [cat.no.4484] and Apt (FR) [cat.no.A4485]. We believe that the conical piece with unmarvered white opaque glass specks on a black appearing unfinished bulbous jar from the cemetery of Canal Bianco at Adria (Figure 12) is an uncompleted vessel that probably was meant to be finished as the other large jars with globular body, wide horizontally outward folded rim and a marvered body with large white elliptical spots. 3.2.3.18 Form IB.18: Unguentarium with tall neck [Isings 8/28b] 63 The only recorded piece was found at Drapers's site, London (UK) [cat.no.4144] (unpublished ) and is made of deep purple glass. It has a conical body and a tall cylindrical neck bulging downwards with a constriction at the shoulder (Figure 34.18). The maximum preserved height of the piece is 140 mm. Another smaller intact vessel originates from Limassol (CY) [cat.no.777], has a height of 123 mm and a spiralling trail decoration in opaque white glass. 61 Apparently Italian researchers agree to consider only glass as black when its true hue cannot be discerned with the naked eye, while the black appearance is swept away from the moment a glimpse of the glass hue can be recognized. 62 Because the lid has a larger diameter than the cantharus and is made of a strongly coloured purple glass that appears black, it is said not to be from one and the same set (Foy, Nenna 2001, 173, no. 270). 63 Personal communication from John Shepherd - 2008. 55 3.2.3.19 Form IB.19: High and slender beaker with folded base-ring [Isings 34; AR37/70] This tall cylindrical vessel shape has a convex body tapering towards the foot and a folded tubular base-ring with pushed-in base (Figure 34.19). The rim is everted and knocked-off. Sometimes the body may be decorated with concentric circles. The few pieces so far recorded come from Swiss sites such as Augst [cat.nos.3536; 3554] and Courroux. Confusion however remains to dating the shape to Period I, II or III because a base fragment was found in the glass workshop of Äussere Reben at Kaiseraugst (CH) [cat.no.3762], which was active between c.125-225 AD (Fischer 2009). We therefore believe that the fragment was erroneously incorporated into the late rd th 3 -4 century AD vessel shape of high conical beakers on folded base-ring (Isings Form 109; AR70; Trier 59) (Rütti 1991, no.1459). This permits us to assume a possible production in Period I or in Period II. The determination of the chemical composition of the vessel fragments would enable us to easily ascribe the pieces to either the early period (low iron content) or the younger period of the glass workshop (Cosyns et al. 2006a; Van der Linden et al. 2009) [see Chapter 8]. Period II Figure 35: Typology of free-blown shapes (drawings nos. 1-14 from Cosyns, Hanut 2005, fig.2; nos. 15-20 drawn by author; no.19 from Isings 1957) 3.2.3.20 Form IIB.1: Wide-mouthed beaker with carinated body and base-ring or carchesium [Isings 36b; Trier 31; Morin-Jean 6; Cosyns, Hanut 2005 Form 1] This carchesium shape features a wide everted mouth of more or less the same size as the carination at the lower part of the slightly concave cylindrical body (Figure 35.1). Some carchesia, such as those from Amay (Lehance, Willems 1987, 52, fig.4, no. 9), Wancennes (Mignot 1984, 224-227, fig.37, no. 1), and Esch IV (van den 56 Hurk 1975, 82, fig.19) have a slender cylindrical shape with a height-rim diameter of 99-65 mm [cat.no.2882]. Others, like the three from grave 176 in Cutry (Liéger 1997, 148, fig.28, nos. 3-5), are more compact and have a body that is rather concave. The height of these vessels averages 90-100 mm. Different is the intact piece from Boulogne (FR) which is 100 mm high and has a rim diameter of 109 mm (Canut 1991, 66, pl. 18) [cat.no.1028]. The decoration on some beakers consists of applied glass threads, in the same colour as the vessel, around the top of the body and just below the carinated body, like the example from Esch IV. A fragment from Colchester can be determined as such a type of vessel (Cool, Price 1995, 93, no. 615). This very thin-walled vessel type (0,6-1,5 mm) has a fire-rounded rim edge or a short inward folded rim. The base-ring is never tubular, but solid. Usually pinched out [see Chapter 2], it is sometimes made from an applied trail. Also typical is the concave base with central pushed in cone and a pontil mark, which is either small massive circular with a diameter of about 8-15 mm or large thin annular with a diameter ranging between 20-25 mm (Figure 9) correspondingly related to the use of a massive punty and a blowing pipe [see Chapter 2]. 3.2.3.21 Form IIB.2: Spouted jug with carinated body and base-ring [Isings 88-variant; Trier 129b; Cosyns, Hanut 2005 Form 2] Only one example of this type is known and comes from a grave in Schaarbeek (Hanut 1999, 8-9, fig.6). This 105 mm high vessel is a hybrid shape composed of elements from Forms IIB.1 and IIB.3. The general profile is that of a carinated beaker with pulled-out spout and ribbon handle attached to the rim edge with a thumb rest (Figure 35.2), and therefore mistakenly classified by Clasina Isings as a Form 88b with globular body (Isings 1957, 105). A glass thread in the same colour as the body is applied around the top of the body and just below the carinated body. The closest parallel is the spouted jug (Trier 129b) from tomb 154 of the St. Matthias cemetery at Trier (DE). This somewhat smaller example78 mm heightis made of black-and-red glass with an applied base-ring and rim, a vertical handle, and a spiralling trail around the neck in opaque white glass (Goethert-Polaschek 1977, 221th 222, no.1334, pl.72) [cat.no.A4618]. Although the jug from Trier is dated in the 4 century AD in relation to the general date of the form Trier 129a, the tomb has not revealed other burial gifts to confirm this late date. Seeing that the jug is more comparable to the Schaarbeek jug than to the St. Matthias cemetery yielded burial from the st 1 century AD (Goethert-Polaschek 1977, 267-327) we assume more acceptable to consider the vessel a Period II product. 3.2.3.22 Form IIB.3: One-handled spouted jug with globular body [Isings 88b; Trier 116c; Cosyns, Hanut 2005 Form 3] This type is very rare in black glass and is usually produced in blue-green, greenish or colourless glass. The globular body has no foot but a concave base with central kick (Figure 35.3). The vertical handle stands from shoulder to rim opposite to the spout and is finished with a slender thumbrest. Examples with a handle in an angle with the spout are only known in naturally coloured or decolourized glass. The example of grave 176 of the cemetery at Cutry (FR) was found in association with another small jug of the same type but in dark green glass (Liéger 1997, 148, fig.28, no. 2) and three black glass carchesia, Form IIB.1 (see above). 3.2.3.23 Form IIB.4: Small bulbous cup with short concave neck, wide-mouthed rim and base-ring [Isings 94; AR 104.1; Trier 37; AV V90; Cosyns, Hanut 2005 Form 4] This shape (Figure 35.4) and the next are very similar but encompass characteristic variations concerning dimensions and ratios. These bulbous cups are mainly blown in blue-green glass, but are known in black glass especially in the Rhine region as far south as present-day SwitzerlandAugst (Rütti 1991, nos. 1997; 1999-2003) and Avenches (Martin Pruvot 1999, 199, nos. 1258; 1259; 1261 and 230, no. 1490, pl. 149). An intact Belgian example comes from Thier-Laurent, Bois-et-Borsu, not far from Clavier-Vervoz [cat.no.199] showing a finishing touch by abrasion of the rim and applied trails on shoulder and foot (Figure 36 bottom). The glass hue and particularly the sharply angled rim, shoulder ridge and pronounced high base-ring obtained by abrasion, stands in sharp contrast with the much lesser elaborate free-blown counterpart from the villa of Lauw (BE) (Figure 36 top). The Bois-et-Borsu cup is therefore a unique piece as it has to be considered a marked example of glass vessel intending to evoke metal ware, whether it be silver or bronze, but equally because the vessel demonstrates that the more clumsy made black glass cups are to be considered as well as an imitation of metal ware. 57 Figure 36: (top) Base fragment of a small bulbous cup (Form II.B.4) from the Roman villa at Lauw, Belgium; (bottom) Mouldblown globular beaker from Bois-et-Borsu (photograph by author, by courtesy of PGRM Tongeren (top) by courtesy of KMKG (bottom)) This type of vessel shape is known decorated in various ways: o o o o 3.2.3.24 with a marvered festoon pattern in opaque white glass like for instance the example from a stone box grave at Heerlen (NL) (Brouwer 1991, 47, no. 16) (Figures 2b-c; 3b) [cat.no.2898]; with applied glass trails on shoulder and lower part of the body like the pieces from Lauw (BE) (unpublished) [cat.no.334] (Figure 36 top) and Liège (BE) (Halbertsma 2006, 31, no. 20) [cat.no.429] but mainly from Swiss sites such as Avenches, Augst and Kaiseraugst (CH) (Rütti 1991, 102, nos. 2006-2012) [cat.nos.3757; 3525]; from Altbachtal, Trier (DE) comes the lower part of a monochrome cup in a very dark, almost opaque, amber yellow glass appearing black (Goethert-Polaschek 1977, 45, no. 135, fig.16); with indented decoration likewise the fragment from the villa of Mettet (BE) [cat.no.449] showing elongated indents and an applied trail on one end (normally at shoulder or lower body); with engraved decoration like for instance the piece from Luxemburgerstrasse at Cologne (DE) [cat.no.1840] showing a decoration pattern of vine and ivy leaves; chevrons and 'lotus'-motif and curls in 8 registers. Form IIB.5: Small bulbous beakers with vertical neck [Isings 94; AR 104; Trier 38; Cosyns, Hanut 2005 Form 5] These vessels have a short or high vertical neck, thickened rim and base-ring with small diameter compared to the larger preceding form type (Figure 35.5). They are known in monochrome black glass and likewise the previous shape also decorated in various ways, in particular with a marvered festoon pattern, applied glass trails, and indentations. The first variant have opaque white marvered festoon-patterned decoration like for instance the beaker from 64 Rheinbach-Flerzheim (DE) [cat.no.1999] , a rim fragment and a base fragment from Avenches (CH) (Martin 64 Personal communication from Anna-Barbara Follmann-Schulz – 2005. 58 Pruvot 1999, 230, no. 1490, pl. 149) and various pieces from ash pits 2/3 of the barrow at Siesbach (DE) (Abegg 1989, 209, nos. 312, 319-26). A second variant on this type of bulbous cup is decorated with an applied glass coil on the shoulder and on the lower part of the body just above the foot like for instance the Ernesto Wolf beaker (Stern 2001, 193 no. 84). Presumed from Cologne, the vessel is 73 mm high and has a single thread in opaque white glass applied on the shoulder and just above the base-ring (Figure 37). A similar bulbous beaker in pale bluish glass retrieved from nd tomb 24 in Amay, Belgium is attributed to the last quarter of the 2 century AD (Willems, Amand 1960, 761; 766, fig.4: D15). Its dimensions are 70 mm (height), 85 mm (maximum diameter of the body), and 56 mm (diameter rim). Figure 37: Small bulbous cup in Württembergisches Landesmuseum Stuttgart, formerly in the Ernesto Wolf Collection (illustration from Stern 2001, 193, no. 84) Figure 38: (left) Faulquemont, France (photograph by Hubert Cabart); (right) MET 17.194.166, Metropolitan Museum of Art, New York (photograph by Chris Lightfoot, courtesy by Metropolitan Museum of Art, New York) The third variant (Isings Form 94/Trier 36) is a bulbous cup with an indented decorated body (Figure 38). Mainly with a vertical rim, short or high, some examples have an outsplayed rim. These indented small cups can have a coiled decoration similar to variant 2 on the shoulder and on the lower part of the body, just above the foot. The example in the Metropolitan Museum of Art, New York [cat.no.4436] has four deep circular indents and a horizontally applied glass thread in blue glass on the shoulder and on the lower part of the body just above the foot (Figure 38 right). Recently, a bulbous cup with indented body has been found in Faulquemont (FR) (Cabart, 2005, 20, fig.4) (Figure 38 left). This variant is also known in other colours of glass such as a colourless example from Arlon (BE) (Bertrang 1954, 96) and a blue-green piece from Trier (DE) (Goethert-Polaschek 1977, 44, pl. 35, no. 134). The latter cup has a bulbous body decorated with six indents, and a rounded vertical rim. Because it is very plausible that all these identity vessels are the fingerprint of one single glass blower, or at least attributable to one specific glass workshop, we may assume that black glass vessels formed only a part of the entire production of one glass workshop. These drinking vessels are clearly inspired on cups in metalescent ware. Besides a chromatic relation, also a morphological one can observe between small bulbous cups in black glass and those in metalescent ware from 2 the Mosel area around Trier (DE), such as the vessel shape Niederbieber 33a-c/Gose 209 (Oelmann 1976 , 4059 42). The black glass cup show the same variants noticeable in Niederbieber 33, where variant A of is plain and undecorated, variant B is plain but with applied trails and variant c has indents as well as applied trails. The long indented decoration and applied glass trail on the fragment from the villa of Mettet [cat.no.449] corresponds with the beakers type Gose 206 but are in fact not more than a taller variant of Gose 209. Several other shapes in black glass can be linked with the metalescent ware like for instance type Gose 199 (Form IIB.5); Niederbieber 31a (Form IIB.4-5). 3.2.3.25 Form IIB.6: Bulbous cup with a very thick flattened base [Isings 96-variant; AR 40; Cosyns, Hanut 2005 Form 6] In prior work we erroneously recognized this shape as an Isings Form 94-variant (Cosyns, Hanut 2005, 115, form 6), but the hemispherical body with a very thick flattened base without a foot, slightly outsplayed rim with a ground edge, and wheel-cut grooves or ridges on the body is rather to be connected with the colourless and palecoloured hemispherical bowls (Isings Form 96b1), and consequently with Beat Rütti‘s AR60.1B, in particular nos. nd th 1317-1352 (Rütti 1991, II, 68-69, pls.58-62) (Figure 35.6). This shape is dated from the mid 2 –to the 4 century AD and thus fits the general Period II date of the black glass vessels. An intact example of such a bulbous cup was discovered in barrow III in Esch-Kollenberg (NL) which is dated at about 170-180 AD (Van den Hurk 1973, 224-225, no. III, 36, fig.67) [cat.no.2881]. It is 76 mm high, and its greatest diameter is 97 mm. Another, although incomplete, piece is known from Augst (CH) (Rütti 1991, II, no. 1186) [cat.no.3540]. Furthermore, we are convinced that another fragment from Augst/Kaiseraugst (CH) should be incorporated into this type instead of type AR 37 (Rütti 1991, II, no. 1133, Taf. 49) [cat.no.3536]. We suggest that the small rim fragment should be drawn with a stronger inclination. Recently, a new example was found during excavations at Faulquemont (FR) nd rd and coming from a late 2 –early 3 century AD context (Cabart, 2005, 20, fig.4) [cat.no.1073]. 3.2.3.26 Form IIB.7 Wide-mouthed bowls with a scalloped edge and base-ring [Isings 42d; Cosyns, Hanut 2005 Form 7] The indented body up to the rim of such shallow open vessels generate a scalloped edge (Figure 35.7). Thus far only the piece from barrow III at Esch-Kollenberg (NL) (Van den Hurk 1973, 225, no. III, 37, fig.69) is known in black glass. This extremely rare shape (Isings 1957, 59) is also known in colourless glass, e.g., one from barrow 65 III in Champion (BE), near Namur (Del Marmol 1851, 57), and a recently excavated piece in Jabbeke (BE). The fragmentation degree of the latter piece was very high, but partial restoration provided enough evidence for its attribution as a wide-mouthed bowl with a scalloped edge, and with a spiralling coil just below the rim. The vessel nd rd comes from a late 2 –early 3 century AD cremation tomb, which formed part of a rural cemetery of about 40 graves. Finally, a third example has been reported by Jorge Alarçao (1967, 5) in his account on the glass vessels from Jerumenha (PT). 3.2.3.27 Form IIB.8: Hemispherical shallow bowl with a thickened rim and base-ring [Gellep 717; Cosyns, Hanut Form 8] This shallow open vessel form (Figure 35.8) is unique and only known in black glass from grave 3815, KrefeldGellep (Pirling 1997, 36, no. 3815,2). Recently, Renate Pirling (2006, 299) confirmed that the Krefeld-Gellep piece rd remains a unique copy thus far in glass but that it resembles a pottery vessel shape which is typical for the 3 century AD. 3.2.3.28 Form IIB.9: Shallow saucepan (trulla) with flat handle [Isings 75a; Cosyns, Hanut 2005 Form 9] Trullae or saucepans with flat handle have a cylindrical body with thickened, vertical rim, and usually show a thin base-ring (Figure 35.9). The only example registered in black glass comes from sarcophagus II in Stein (Isings 1971, 25, no. 78). Most examples of this type are greenish or colourless, and similar glass vessels are known with a snake-thread decoration. 65 Personal communication from Yann Hollevoet – 2009. The artefact has erroneously been published as an Isings 32 beaker (Hollevoet, Hillewaert 2009, 94). 60 3.2.3.29 Form IIB.10: Bulbous flask with a high narrow neck, funnel-mouth, and concave base [Isings 92; Cosyns, Hanut 2005 Form 10] Bulbous flasks have a high narrow neck with wide everted mouth and short inward-folded rim, and concave base (Figure 35.10). They have been made with as well as without handles, and with as well as without constriction at the base of the neck. The example from sarcophagus II in Stein is the only one without handles thus far recorded in black glass (Isings 1971, 11-12, no. 19). This type of jug with two vertical handles is equally unique, and is only known from an intact piece at the Musée National d‘Art et Histoire, Luxemburg [cat.no.2866] (unpublished) (Figure 39). Similar to the previous shape is the funnel-shaped rim with an inward- folded edge, the narrow cylindrical neck without constriction at its base, and the two simple vertical handles applied on the shoulder in a simple blob and attached to the neck below the rim. However, no thumb-rest or loop has been modelled, and the handles have an O-shaped section. Above all, the body is globular, with only a slightly concave, flattened base. A certain parallel can be seen in the jug from Arsimont (BE) in ‗naturally‘ coloured glass (Mariën 1983, 112) although the piece has only one handle that is attached at about the middle of the neck. This vessel is dated in nd the third quarter of the 2 century AD, and measures 200 mm with a maximum diameter of 131 mm. Figure 39: a two-handled jug from unknown provenance (LU) (photograph by Jeannine Geyssant by courtesy of MNAH, Luxemburg) 3.2.3.30 Form IIB.11: Ovoid one-handled jug with base-ring [Isings 120-variant; Cosyns, Hanut 2005 Form 11] This ovoid one-handled jug has a concave neck and funnel-shaped mouth with rounded rim and base-ring (Figure 35.11). The jug from barrow VI at Esch (NL) (van den Hurk 1977, 120, no. VI, 25) is the only one known in black glass [cat.no.2879]. It has a spiral coil of similar glass wound up many times around the top of the neck. The piece measures 123 mm (h.), and has a maximum diameter of body of 77 mm. A rim fragment with a vertical handle comes from Straten, Bevingen (BE) (Wesemael, et al. 2004) [cat.no.519]. The upper part of the neck is decorated with an applied coil. The decoration and the handle are made from the same dark olive-green glass as the core. 3.2.3.31 Form IIB.12: Ovoid one-handled jug with a rounded or spouted mouth [Isings 54; Cosyns, Hanut 2005 Form 12] These one-handled jugs have a rounded or spouted mouth and a base-ring (Figure 35.12). The handle bears a vertical scalloped thumb-rest at the top and, in some cases, has an applied medallion on the body just below the start of the handle. Such medallions have a rather limited range of depictions: - a theatre mask like the piece in the Benaki Museum (Clairmont 1977, 12) and two in the Bible Land Museum (Bianchi (ed.) 2002, 303, GR-39-40); a lion‘s head like the piece in the Corning Museum of Glass [cat.no.4445] or the one in the British Museum (Cooney 1976, 2, no. 8); a human head (perhaps the Medusa or Bacchus) like the one on the jug from Weilerwist-Hausweiler (Follmann-Schulz 1992, 21-22, no. 10) [cat.no.2108]. 61 The black glass jugs recorded so far are all decorated with marvered opaque white trails in upright festoons, vertically as at Nijmegen (Isings 1964, 176, fig.4, no. 4), or the cemetery of the Luxemburgerstrasse, Cologne (Isings 1964, 176, fig.5), and oblique like the Weilerwist-Hausweiler piece. 3.2.3.32 Form IIB.13: Ovoid beaker with plain or indented body [Morin-Jean 113; AR 102-variant; Cosyns, Hanut 2005 Form 13] This type of ovoid beaker incorporates a number of variants (Figure 35.13): 1) plain or indented body, 2) outsplayed or slightly tapering cylindrical rim ending with a (fire-)rounded edge, and 3) base-ring or a concave base without foot. Characteristic of this type of drinking vessel is the absence of a shoulder and neck. The oval body of the beaker carries on through to the fire-rounded rim. We have to extend this type given the fact that diverse variants have come to light besides the baseless beaker with indents from Eschweiler-Hastenrath (DE) (Cosyns and Hanut 2005). The best parallel is Form 113 of Morin-Jean (1913), whereas Isings (1957) and Goethert-Polaschek (1977) do not provide good parallels for this type of ovoid beaker. However, if further comparisons should be made, the shape can be related in some way to the AR102 pieces (Rütti 1991, Taf. 88, nos. 1962-1973) with nos. 1962-1969 having a more inward rim, and nos. 1970-1973 with an outsplayed rim featuring an S-shaped profile. The latter variant shows continuity in production, since the vessel shape resembles the much later-dated Goethert-Polaschek Form 62a and Isings Form 131. Table 23: List of ovoid beakers, Form IIB.13 cat.no. site body not catalogued not catalogued not catalogued 1876 not catalogued not catalogued 2975 not catalogued not catalogued not catalogued foot colour h. (mm) reference Fécamp (FR) indented base-ring blue-green 78 Lisieux (FR) plain base-ring 87 Neuville-lesDieppe (FR) EschweilerHastenrath (DE) Maaseik (BE) plain base-ring indented concave plain base-ring dark bottle green66 almost colourless black/olive green blue-green Lyons-la-Forêt (FR) Nijmegen (NL) applied trails base-ring colourless 109 Sennequier 1985, 63-64, no.38 Sennequier 1985, 63, no.36. Sennequier 1985, 63, no.37. Follmann-Schulz 1992, 8788, no. 47 Janssens 1977, 29, no.12, pl.XII Vanpeene 2007, 27, fig.1:8 plain base-ring 128 Isings 1964, 176, no. 4 Brognon (FR) applied trails base-ring black/bottle green colourless 132 Vitrum 1990, 163, no.183 Cologne (DE) applied trails base-ring colourless 155 Doppelfeld 1966, no. 125 Maaseik (BE) applied trails base-ring blue-green 183 Janssens 1977, 29, no.12, pl.XII 90 90 95 An overview on this vessel shape shows that the shape was also produced in naturally coloured and decolourized glass (Table 23). The intact example with an inward rim from a burial in Nijmegen (NL) which is stored at the Rijksmuseum voor Oudheden, Leiden [cat.no.2975] shows a close link to the metalescent ware Niederbieber 30a (Oelmann 1976²) and Trier Form 3 – Group 43 (Symonds 1992, 54-55, fig.35: 639-656). The Nijmegen variant is also known in naturally coloured glass such as the ones from the cemeteries of Maaseik (BE) and Monceau-surSambre (BE), as well as in colourless glass such as the beakers with a snake-thread decoration from, for instance, the villa of Brognon (FR), Lyons-la-Forêt (FR), and Cologne (DE). Only one example with an S-profile in black glass has been recorded thus far: the Eschweiler-Hastenrath (DE) beaker with an indented body and concave base is full of air bubbles. Two other more or less similar beakers come from Neuville-les-Dieppe (FR) and Fécamp (FR) and are made from almost colourless and blue-green glass. These two pieces have an applied glass trail in the same colour as the body just under the belly and under the rim. Geneviève Sennequier (1985, 63-64) mentions other examples mainly from Normandye.g., Lillebonne and Pîtres. There are similarities with Form IIB.5. The three pieces in the collections of the Musées Départementaux de Seine-Maritime, Rouen have a base-ring pinched from the paraison. The base-ring of the Nijmegen beaker appears to be secondarily applied and well-shaped by pinching obliquely with an instrument (Figure 10a). 66 We have not been able to verify whether this ‗verre vert bouteille très foncé‘ should not be considered as black-appearing glass. 62 3.2.3.33 Form IIB.14: Amphoriskos or wide mouthed two-handled jug [Isings 15; Cosyns, Hanut 2005 Form 14] Such amphoriskoi or two-handled jugs with an elongated ovoid body and base-ring, wide neck, and out-turned tubular rim (Figure 35.14) are normally made in blue-green glass or in strongly coloured glass, particularly dark ultramarine blue and opaque turquoise. These vessels are generally dated pre-Flavian (Isings 1957, 32-34; Price, Cottam 1998, 147-148); however, a unique example in black glass with opaque white marvered festoons was rd found in a 3 century grave at Elsdorf-Esch (DE) (Gaitzsch 1999, 77-79; Cosyns, Hanut 2005, 116) [cat.no.1869] [see Chapters 4 and 7]. 3.2.3.34 Form IIB.15: Jug/carafe with one or two vertical handles The jug has a less elongated ovoid body compared to the previous form and a much narrower cylindrical neck without constriction at the base (Figure 35.15). Only one single example of this type of two-handled jug or amphoriskos has been recorded so far in black glass and is unknown in other glass colours to us. This unique piece comes from the villa of Bassenge (BE), and though excavated in 1888, it remained unnoticed in the large collections at Grand Curtius, Liège due to its fragmentary preservation [cat.no.193] (unpublished). Even though the fragmentary preservation (41 fragments) the archaeological completeness of the piece made possible a reconstruction drawing of the entire vessel. The concave base does not show a pontil mark. The general form reminds also the bulbous flask Form IIB.10 but the Bassenge jug has a lower part that is differently shaped – the more elongated body tapers to an open base-ring. The rim is funnel-shaped with an inward-folded edge. The vertical handles are wide, flat, and plain, applied on the shoulder in a simple blob and attached to the neck under the rim ending in a folded thumb-rest. The shape and manufacturing technique of the rim and the handle(s) remind these of the grape-shaped amphoriskoi [see this chapter 3.2.4. Mould-blown vessels] The fragments resulted in two large parts that could not be assembled, though the drawing shows that from a logical connection between the spherical-shaped shoulder and body, the original height should be estimated at c.160 mm. The maximum diameter of the body is 78 mm and that of the foot is 41,5 mm. The neck is about 50 mm high and has a diameter of 15 mm. The funnel-shaped rim has a maximum diameter of 45 mm. The plain handles are 2 mm thick and 14 mm wide. The only good parallel we could trace is the jug found during the A27 excavations at Fishbourne (Price, Cottam 1996, 167; 180, fig.6.29, no.84). Although no glass hue is mentioned we presume it is not black appearing, otherwise the authors would have referred to it. The given dimensions correspond very closely although the piece is higher (estimated height of 180 mm), the body more slender (60 mm diameter) and the neck wider (20 mm diameter). In some way there is resemblance with the reconstructed shape 75 from Avenches (CH) (Martin-Pruvot 1999, 225, no. 1424, pl. 146). This two-handled jug in decolourized glass comes from a c. 160-180 AD cremation tomb at the cemetery ‗En Chaplix‘. This equally unique piece from Avenches, however, has a constriction at the base of the neck, and much longer and double ribbed handles – 67 suggesting a rather conical body. 3.2.3.35 Form IIB.16: Shallow cup with horizontally folded rim [Isings 42-variant; AR 109.2; AV V100] These shallow cups are a variant of the Isings Form 42a cups showing a different rim but otherwise with similar characteristics such as the outward bended horizontal rim, the convex body, the wide base-ring and the concave base with pinched in centre (Figure 35.16). Mainly produced in colourless and ‗naturally coloured‘ glass, some in black glass occur in the Swiss area such as Avenches (CH) (Martin Pruvot 1999, 201, nos 1274-1275, 194, no. 1217) [cat.nos.3602; 3621-23] and Colchester (UK) (Cool, Price, 1995, 99-100, no. 694) [cat.no.4026]. The basering of the Avenches examples is massive and pinched out with a jack, while the examples from Augst/Kaiseraugst show a folded tubular base-ring. 3.2.3.36 Form IIB.17: Cylindrical cup with fire-rounded rim and trailed decoration [Isings 85; AR 98.2; AV V84] Mainly in colourless glass this very common type of cylindrical cup is a variant on the ubiquitous undecorated cylindrical cups with double base-ring (Isings Form 85b) (Figure 35.17). The decorated type in black glass has been only attested in Augst/Kaiseraugst (CH) and Avenches (CH). Characteristic is the applied trail below the firerounded vertical or slightly everted rim and just above the strong curve at the lower part of the body. The trails are 67 This is probably the reason why Chantal Martin-Pruvot compared the piece with the tall necked jug with conical body, Isings Form 55 (MartinPruvot 1999, 225226). 63 of the same glass hue as the body. The base is flat or slightly concave and has an applied base-ring. Some show a trailed ring at the centre around a pontil mark. 3.2.3.37 Form IIB.18: ‗Inkwell‘-pyxis [Isings 77; AR 177; Trier 161] Characteristic to these ‗Inkwell‘-pyxides is the slightly convex cylindrical body (sometimes tapering to the top), the 68 horizontally folded foot and shoulder, and the inward-sloping rim with rounded edge (Figure 35.18). The known 69 examples in various museum collections are all of unknown provenance. A black piece from the former Ernesto Wolf collection is now in the Landesmuseum Württemberg, Stuttgart (Honroth et al. 2007, 108, fig.138 no.175), and a second piece was acquired by the Corning Museum of Glass (Whitehouse, 1997, 199, no. 348) – formerly in the Ray Winfield Smith collection (Smith 1957, 55, no. 64). Similar vessels are known in other glass hues. Various examples in blue glass with a white marvered wispy decoration have been reported in the RömischGermanischen Museum in Cologne (La Baume 1973, D13, pl.18:1), the Newark Museum (Auth 1976, 119, no. 146), and from the former Constable-Maxwell collection (Constable-Maxwell 1979, 43, nos. 44-45). The piece from the Toledo Museum, Ohio is similar to the ones above but green instead of blue. At the National Museum of Scotland and at the Newark Museum are examples in a translucent bluish glass (Lightfoot 2007, 115, no. 271; Auth 1976, 119, no. 145). Most of them have a lid, and even though most of them are probably from different makers, the shape of the vessel supposes a lid to cover the pyxis. The shape is flat-domed or more or less conical ending in a bulbous point. The height of the so-called inkwell pyxides with horizontally outward-folded foot and shoulder averages between 53 and 70 mm and the diameter between 35 and 71 mm, with nearly similar size for rim and foot. Other so-called ‗inkwell‘ shapes produced in Roman times, as described by Clasina Isings (1957, Form 77), can have a cylindrical or polygonal body without the folded base and shoulder or concave top, and be equipped with (mainly three) dolphin-shaped handles. It is very likely that the small handle fragments in black/green glass from Augst are handles of such a shape as that of an inkwell-pyxis and not of an aryballos. 3.2.3.38 Form IIB.19: Animal-shaped vessel [Isings 95] Only two pieces have been recorded in black glass: A dolphin-shaped vessel from a rich cremation burial in Lillebonne (Arveiller-Dulong et al. 2003, 153, fig.21) [cat.no.1214] and a boar-shaped vessel from Cologne (Isings 1957, 112) [cat.no.1849] (Figure 35.19). The Isings Form 95 is however not restricted to only animal-shaped vessels but comprises any ―queer bottle in a fancy shape‖ (Isings 1957, 112) distinguishing shapes representing a fish (form 95a), a pig (form 95b) and a gladiator‘s helmet (form 95c) (see also Glass of the Caesars 1989, 134136, nos.63-65). Other shapes can be perfectly added to this form type such as the unique sandal-shaped vessel found in 1971 in Cologne (DE) (Glass of the Caesars 1989, 137-138, no.66). These artefacts have been freeblown and shaped by pinching corporal elements and further completed by applying glass trails for rendering details. 3.2.3.39 Form IIB.20: Aryballos with globular body [Isings 61; AR 151; Trier 135; AV V 125-129] This small bulbous flask is characterized by a short cylindrical neck with wide horizontal outward folded rim; two proportionally large vertical dolphin-shaped handles; a concave base with a clear pontil mark (Figure 35.20). Likewise the peculiar examples of the mould-blown square bottles in black glass from Augst/Kaiseraugst (CH) [see further in this chapter], is the aryballos very rarely attested in black glass. We have three possible fragments from Augst/Kaiseraugst (CH) of which the elaborate handle fragment is coming from the glass workshop ‗Äussere Reben‘ in Kaiseraugst [cat.no.3778] [see Chapter 6] and a bottom fragment with a maximum diameter below 50 mm [cat.no.3753]. This piece has a very large but thin pontil mark of about 20 mm [cat.no.3553]. The majority of this bulbous bath flask has been made in ‗naturally coloured‘ blue-green to pale blue glass but decolourized glass has been applied regularly. Hitherto few remarks can be made as we have no material at hand to tell whether the black aryballoi were always plain or also was decorated with applied glass trails or show wheel-cut decoration, but aryballoi with annular shaped body are discussed in Period III and the grape-shaped aryballos from Cologne is discussed further below in this chapter [see 3.2.4. mould-blown vessels IIC.3]. 68 Also polygonal inkwells are attested (see Simon-Hiernard 2000, 383-384). Only the base fragment with horizontally outward-folded base-ring, which is presumably part of such an inkwell pyxis was excavated in Heerlen (NL) [cat.no.2901]. 69 64 Period III 1. 2. 3. Figure 40: Typology of Period III free-blown shapes (drawings taken from Pirling, Siepen 2006 (1-2); Dilly, Mahéo 1997 (3)) 3.2.3.40 Form IIIB.1: Convex cup with out-turned fire-rounded rim [Isings 96b2b; AR 61; Gellep 189] This small cup has no foot, a slightly pushed-in base, a rounded body with pulled-out knobs, and a rim with a rounded edge slightly turned outward (Figure 40.1). The piece from Tongeren (BE), now in the Grand Curtius, Liège, is made of a deep green glass and has 9 protruding knobs at the lower part of the body (Vanderhoeven 1958, 59, no.60) [cat.no.563]. Particular are the horizontally pinched protrusions whereas these are generally vertically pinched. Generally in colourless glass this shape is also known in naturally coloured glass. Also less common is the spiralling coil just under the rim, winded seven times around. A similar cup with a height of 63 mm and a rim diameter of 115 mm in yellowish-green glass with 11 knobs comes th from the late 4 century tomb 2 of the villa HA382 at Hambacher Forst (DE) (Gaitzsch et al. 2003, 195, fig.72). The Tongeren beaker has somewhat smaller dimensions: 47 mm height; 72 mm rim diameter. The shape is also known in a larger size with higher vertical walls, type Gellep 515 (Pirling, Siepen 2006, 245). The Pieces from nd Augst/Kaiseraugst (CH) are apparently much earlier dated between the second half of the 2 century AD and the rd third quarter of the 3 century AD (Rütti 1991, II, 70-71, pl.63). We nonetheless assume this variant of the Isings Form 96 of later date than the Form IIB.6 with thick base and decorated with abraded contentric lines, not excluding that such vessels in black glass were produced at the end of Period II. 3.2.3.41 Form IIIB.2: Unguent bottle with flattened circular body [Isings 101-variant – Trier 80 – Gellep 526] Unguentaria with a flattened circular body (Figure 40.2) are not a frequently recorded shape, although they are known from all over the Roman Empiree.g., Egypt, Syria, Cyprus, and the Lower Rhine region like KrefeldGellep (DE), Cologne (DE), Trier (DE), and Strasburg (FR) (Pirling 2006, 272). This vessel shape is essentially known in decolorized, or a pale-coloured glass, even though several examples are produced in a deeply coloured glass like, for instance, the deep blue one from Strasburg (FR) (Arveiller-Dulong, Arveiller 1985, 138, no. 303), or the deep purple and deep blue ones from tombs 227 and 327 from the Jakobstrasse in Cologne (DE) (Friedhoff 1991, 283, Taf. 95:3/12; 329, Taf. 115:2). They appear in two sizes: small ones with a height averaging between 50-60 mm, and medium-sized ones with a height of around 90 mm. The only one known in black glass comes from tomb 5530 of the Krefeld-Gellep cemetery and is, with a height of 92 mm, medium-sized (Pirling 2003, 69, no. 5530:4, Taf. 6:3a-b; 133:2). Very exceptionally there is one large example of 129 mm known from Mainz (DE) in pale green glass (GoethertPolaschek 1977, 137, no. 766). Unguentaria with flattened circular bodies very rarely have handles like the one in pale bluish-green glass from Trier-Biewererstrasse, with two complex multiple dolphin-shaped handles in ultramarine blue glass (Goethert-Polaschek 1977, 137, no. 770). The constriction at the lower end of the cylindrical neck and the slightly concave, pushed-in base makes the similarity with the very widespread contemporaneous unguentaria with bulbous bodies (Trier 79, Isings Form 101) striking. More salient are the comparable dimensions and the existence of equally small-, medium- and large-sized shapes. 3.2.3.42 Form IIIB.3: Aryballos with annular body [Trier 139; Morin-Jean 36] Most are in very pale coloured or decolourized glass but an intact piece from Amiens (FR) is made in black appearing brown glass (Dilly, Mahéo 1997, 120, pl.14, no.257) [cat.no.1006] (Figure 40.3). These vessels are rather small. The Amiens aryballos measures 74 mm high and the body has a maximum diameter of 67 mm and the rim diameter is 30 mm, what makes is slightly larger than the piece in decolourized glass from Trier with a height of 68 mm and a rim diameter of 29 mm (Goethert-Polaschek 1977, 232, no.1389). This vessel shape has a 65 flattened base but examples show four small applied foot knobs, e.g. Trier (Goethert-Polaschek 1977, 232, no. 1389). 3.2.4 Mould-blown vessels The number of vessel shapes generated from the mould-blown technique is rather insignificant. Those recorded so far remain restricted to a few examples of specific shapes, with the grape-shaped amphoriskos and aryballos, and the square bottle being the most frequent shapes. This is in part due to the technical difficulty of blowing vessels in moulds with an iron-rich glass matrix, which shortens the time of workability significantly [see Chapters 2 and 8]. The only mould-blown vessel shape in black glass of Period I noticed thus far is cylindrical cup without nd rd foot stand [see Form IC.1]. In the later 2 to mid 3 century AD, mould-blown vessels essentially occur in the rd th north-western provinces. No example is hitherto identified dating in the later 3 or 4 century AD. Only in the th th early Byzantine period(5 7 centuries AD) were vessels blown again in a black-appearing glass. SyroPalestinian workshops produced at that time mould-blown polygonal flasks and jars in dark to very dark purple and brown glasses, with some having a pretty black appearance (Stern 1995, 256-257, no. 173; Newby 2008, 136, no. 34; 142, no. 37; 208-210, nos. 65-66). 1. 6. 2. 7. 3. 4. 5. 8. Figure 41: Typology of mould-blown vessels (1-5 = Forms IC; 6-7 = Forms IIC; 8 = Form IVC) (all drawings by author; photograph (3) taken from Whitehouse 2001, 65, no.536) Period I 3.2.4.1 Form IC.1: Cylindrical cup with everted rim [AR 38; Trier 39; AV V 47] Various features are characteristic to the cup: 1) the outward-bended rim with cut and ground edge, 2) the wheelcut concentric lines on the body, and 3) a somewhat profiled, thickened base without foot (Figure 41.1). This shape is to our opinion incorrectly described as a free-blown shape because this vessel shape can only be obtained by rotating the blowing pipe during the mould-blown process. The fact that the vessel shows no seam does not exclude the use of a mould. The vessels of this type are perfectly cylindrical with a perfectly constant everted rim and a steady horizontally cut edge. The vessel has a particular curve that only can be obtained from a mould. Mainly attested in decolourized or ‗naturally coloured‘ glass, some examples in black glass are known like for instance the intact piece from the Tongeren area (BE) [cat.no.562] (Vanderhoeven 1962, 23, no.15; Cosyns et al. 2006a, 31-34, fig.1 [top]; fig.2) and a base fragment in a black/brown-green glass comes from the cemetery ―En Chaplix‖ at Avenches (CH) [cat.no.3617] (Bonnet-Borel 1997, 29, no. AV V 47; Martin-Pruvot 1999, 188, no.1181). The Tongeren cup has two parallel wheel-cut lines on the lower part just above the curving and another pair of lines on the upper part just below the rim. Though being cylindrical its shape has cubic dimensions with a height 66 of 74 mm, the diameter of the body is 72,5 mm and that of the rim 75 mm. The Avenches-fragment has a body diameter of 65 mm. This type of drinking vessel is rather made in a colourless or pale-coloured glass. Significant is that none of the 30 pieces attached to this type from Augst/Kaiseraugst are made in black glass (Rütti 1991, 55-56, nos. 1148-1177), nor are the examples reported by Jennifer Price and Sally Cotham on the Romano-British vessels (1998, 94-95). The date range is situated in the second century AD up to the late Antonine period. 3.2.4.2 Form IC.2: Truncated conical cup [Isings 29/31 variant] This particular form type is hitherto only represented by a unique but fragmented piece (max. pres. h.:49 mm) from excavations within London (UK) [cat.no.4142] (unpublished), with no parallel to our knowledge in any other glass hue. Besides its resemblance to the cups of Isings Form 29, there is also close parallel with the mouldblown truncated conical beakers (Isings Form 31). The fragmented pieces make it impossible to reconstruct the entire vessel shape, but from what remains can be assumed the shape of a conical cup with a horizontal ribbed cylindrical body (max.d.: 50 mm) tapering towards the foot and with a wide truncated annular ring (d.: 45 mm) (Figure 41.2). Due to the restrictions of the casting technique and the characteristic features of the vessel we believe this rather thin-walled vessel shape is mould-blown. Furthermore the sharp relief of the concentric ribbings and the dotted disc at the centre of the foot the vessel was most likely blown in a metal mould using a rotating movement and possibly abraded to resemble metal ware. 3.2.4.3 Form IC.3: Sports cups Athlete and circus cups are usually made of transparent pale ‗naturally coloured‘ or decolorized glass. Not only is no more than one single fragment known in black appearing glass, the piece is also unique as David Whitehouse (2001, 65, no.536) comments on the piece: ―The fragment differs from most first century sports cups in shape, color, and depth of relief. ... The wall is unusually thick. Despite the anomalous character of the fragment and its pristine condition, I see no reason to doubt its antiquity.‖ The fragment appears that of a hemispherical cup with the representation of an athlete and a palm leaf separated by a vertical border in one register [cat.no.4455] 70 (Figure 41.3). 3.2.4.4 Form IC.4: Hexagonal bottles with high relief Hexagonal bottles have been thoroughly studied by Marianne Stern (1995, 74-86; 113-148). She distinguished three classes based on the mould seams on the bottom of the vessel. Furthermore she identified the use of clay st moulds and metal moulds (Stern 1995, 75). All classes are dated in the 1 century AD. Characteristic are the short cylindrical neck, small horizontal rim folded out – up – inward, shoulder and belly of the body also contain mould-blown design, correspondingly festoons and triangular leaves or narrow gadroons (less frequent) (Figure 41.4) The only catalogued piece in a black appearing glass is one from the Toledo Museum of Art, which is made of a dark brown glass (Stern 1995, 133-135, no.40) [cat.no.A4614] (Figure 42 - left). The piece was blown in a threepart mould construction type MCTIV (Stern 1995, 133). Based on the representation of a pomegranate, grapes and a cedar cone, the high relief is determined as Fruit Type Series B1 (Stern 1995, 78-81; 133-134). Marianne Stern also mentions that the festooned floral motif on the shoulder is more elaborate than of Series A and that the belly has gadroons instead of the more popular triangular leaves. The piece has an average size with its 85 mm height and a base diameter of 23 mm. 3.2.4.5 Form IC.5: Double-headed bottle without handles A tall cylindrical neck, a horizontal rim folded outward – up – inward and a body composed of two human faces back to back are the main features of this form (Figure 41.5). The black appearing piece from Toledo Museum of Art is apparently made of a deep reddish purple (Stern 1995, 223, no.142) [cat.no.A4613] (Figure 42 - right). The body consists of two Medusa heads from two different moulds clearly revealing the seam of two mould 70 A fragmented circus cup from Augst/Kaiseraugst (CH)is made of very dark brown transparent glass (unpublished). 67 71 segments. The flat base is circular and has a base mark showing a central disk within a square with concave sides. The piece is 105 mm high and the diameter of the base 37 mm. Figure 42: (left) The hexagonal bottle (Form IB.4) and (right) the double head-shaped bottle (Form IB.5) from the Toledo Museum of Art (photographs taken from Stern 1995, colour plates 7; 23). Period II 3.2.4.6 Form IIC.1: Grape-shaped amphoriskos [Morin-Jean 131; Isings 91a; Trier 138] 72 This shape is mainly known from examples in transparent pale blue-green to pale blue glass (Simon-Hiernard 2000, 363-370); however, some have been blown in deeply coloured glasses, like the piece from the rich tumulus at Vorsen/Frésin (BE) (Schuermans 1864, 256-261) or the fragments from Trier (DE) (Goethert-Polaschek 1977, 231-232, nos.1386-1387). The amphoriskoi of Period I (Isings Form 78e) have smaller dimensions and lack handles, but thus far none have been recorded in a black-appearing glass. For Period II, the grape-shaped amphoriskoi appear in various sizes and workmanship (Figures 43-44). The variant with a beaded stem (Isings Form 91b) is unknown in blackappearing glass. Two groups can be distinguished: the large-sized type of about 170-180 mm high, and the smaller-sized type ranging between 100-110 mm in height. The distinction related to the size and shape of the grapes and the pattern it provides can possibly be linked to different glass workshops using different moulds. The grapes of the Nijmegen (NL) amphoriskos look thick and spherical and are placed in an irregular, rather naturalistic pattern (Figures 41.6 and 43). The amphoriskoi of Bologna (IT) (Figure 44) and Wederath (DE) are much smaller and not very protruding, showing a fairly rigid symmetrical and stylized pattern. This variety in grapes is most probably due to the use of different materials to produce the moulds, i.e., the use of clay results in naturalistic grapes and the use of bronze results in stylized grapes. Nevertheless, the various stylistic elaborations and idiosyncratic technological features must be regarded as specific for particular glass workshops in different regions and/or during successive generations of glass blowers. Various smaller and larger moulds can be recognized to produce the grape-shaped amphoriskoi with naturalistic or stylized grapes, but a thorough synthesis on this issue has not been completed (Simon-Hiernard 2000, 363-370; Arveiller forthcoming). However small grape-shaped amphoriskoi with naturalistic grapes do exist, but all of them recorded so far were made from clear, transparent glass. On the other hand, no large grape-shaped amphoriskos with stylized grapes has been recorded. 71 Marianne Stern remarks that though the facial features are similar, one face is smaller than the other. The grape-shaped vessel from Myres‘ excavations on Cyprus and now in the Metropolitan Museum of Art, New York (Karageorghis 2000, 287, no. 470), is more a small jar and thus a totally different shape with a similar decoration pattern produced with a similar technique. The jar is squat only 48 mm highand has no handles. The ovoid body has small pastille-shaped grapes reminiscent of those on the ovoid bottles with a stylized grape pattern as discussed by Marianne Stern (1995, 191-195, nos. 120-128), although these have a double-sized body compared to the Cypriot piece in the Metropolitan Museum of Art. Instead of finishing the piece by swaying the blowing pipe to obtain a long cylindrical neck, the rim has been outsplayed and folded inward. 72 68 Since the amphoriskoi from Wederath (DE) are very fragmentary, no true comparison can be done with the Nijmegen piece. However, the dissimilarity in shape of the grapes displays an important formal distinguishing feature for the amphoriskoi from both sites. The Nijmegen amphoriskos has very naturalistic, thick, bulging grapes that are close to each other and show no specific pattern, whereas the piece from Wederath (DE) has small, round, slightly protruding knobs in a diagonal pattern with lots of space between each grape. Noticeable is the piece from Bologna (IT) (Figure 44) (Meconcelli-Notarianni 1979, no. 213) with a similar pattern. Figure 43: The grape-shaped amphoriskos from Nijmegen-Hees (Photograph by the author, courtesy of the Museum Het Valkhof, Nijmegen) Figure 44: The grape-shaped amphoriskos from Bologna, Italy(illustration taken from Meconcelli Notarianni 1979, no. 213) In the collections of the former Museum Wallraff-Richartz in Cologne (DE) was a grape-shaped amphoriskos with an applied yellow opaque glass trail on the rim, reported to have been excavated during the laying of the drains and sewers in Cologne (‗städtischen Kanalisierungsarbeiten‘, s.n. 1908, 389). It is not clear what type of mould was used. 69 3.2.4.7 Form IIC.2: Prismatic bottles [Isings 50] Prismatic bottles are usally square bottles but also include rectangular and octagonal (Figure 41.7). This ubiquitous vessel shape can be considered as a fossil guide in Roman archaeology, as are the ribbed bowls. Peculiar are those square bottles which are not mould-blown in pale blue glass or the naturally coloured bluegreen glass. Until recently, no example of this type in a black-appearing deeply coloured glass was listed. Within the material from Augst/Kaiseraugst (CH) not incorporated in Beat Rütti‘s work (1991) were found six fragments of this type in a very deeply blue-green glass appearing black [cat.nos.3462;3493-94;3521;3532;3551;3777] (Figure 45). From the remains cannot be said whether the bottles had a square body, a rectangular or octagonal one and whether the body was large or small. No mould fragments from the ‗Äussere Reben‘-workshop have been found or the base fragment of a prismatic bottle in black glass to determine the base-mark(s). Figure 45: Two fragments of mould-blown prismatic bottles from the ‗Aussere Reben‘ workshop at Kaiseraugst (CH) [cat.no.3777 (left); cat.no.3462 (right)] (photographs by the author, by courtesy of the Römermuseum, Augst) Period IV 3.2.4.8 Form IVC.1: Double-headed jug Although the apparent popularity of using deeply coloured glass for head-shaped vessels (Stern 1995, 201-246) we recorded only one single piece in black appearing glass that can be ascribed to Period IV (Figure 41.8). Probably many more have been overlooked. The Corning piece (Whitehouse 2001, 161, no.1174) [cat.no.4457] not only shows characteristic features attested on other non-black appearing examples: a) the straight funnelshaped mouth with knocked-off rim; b) the vertical handle tooled at two places to form a semicircular shape with pinched-out thumb pieces, wrapped around the neck to form a ring; c) the so-called knobby hair dress, similar to the small stylized grapes attested on a group of grape-shaped amphoriskoi (Figure 44). The dimensions of the Corning piece also correspond to the small size of this vessel shape in other glass hues. But with a height of 95 mm the piece is relatively tall. The rim diameter of 36 mm ranges within the general size of 34-38 mm. In accordance with Marianne Stern‘s observations on identical material in the collections of the Toledo Museum of Art we are tempted to assume that the faces of this specific vessel shape represent childlike faces, one serious and one smiling (Stern 1995, 241-242, nos.161-163). 3.2.5 Rod-formed vessels The various vessel shapes that have been produced using the rod-formed technique are very idiosyncratic except for the squat, large, triangular, solid-body vessels with tall necks (see Form IID.1). The others represent only a reduced number of shapes that is to be situated within a relatively short date range (Period IV) [see Chapter 4], confined to a well-demarcated regional production and distribution (Levant and Egypt) [see Chapter 6]. The two major shapes discernable within those in black glass are the tall but slender tubular vessels (Form IVD:1) and the wide-mouthed squat jars (Forms IVD.2-4). Within the limits of this present study on black glass, it was not possible to set up an elaborate typology because similar material in other glass hues was not taken into consideration, nor able to execute an exhaustive research. We have tried, however, to give a first clear overview on this issue by categorizing each rod-formed vessel shape occurring in black glass on the basis of distinctive features (Table 24): 70 Table 24: Overview of features used to categorize the rod-formed vessels the diversity of decoration  plain  with applied glass thread decoration  with ribbed body o twisted o straight/vertical ribbings with stamped/pressed-in decoration the bottom  with (small) pad foot  without foot the body  hemispherical  with conical body  with biconical body the handles  with handles o vertical o horizontal  without handles Table 25: Types of decoration of the rod-formed vessel shapes known in black glass type plain Form IID.1: Large conical solid body vessel with tall neck IID.2: Large conical solid body vessel with short Form neck IVD.1: Tall and slender tubular vessel Form Form IVD.2: Squat miniature jar Form IVD.3: Truncated conical jar Form IVD.4: Biconical jar 3.2.5.1 x - ribbed applied glass trails x x x - x x x x x stamped x - Form IID.1: Tall candlestick unguentaria with a triangular solid body [Isings 82b2/Harden Class XIII, A] Figure 46: Rod-formed unguentaria with triangular shaped solid body and tall neck from Elkab (EG) (drawing by F. Roloux, by courtesy of the MRAH-KMKG, Brussels; photograph by author, by courtesy of the MRAH-KMKG, Brussels) Based on the observations of the four pieces from Elkab (EG) [cat.nos.868-871] (Figure 46) we clearly establish a form of rod-formed unguentaria. This vessel type is generally acknowledged to be manufactured by blowing as can be observed on comparative material (Harden, 1936, 260, nos. 815-82; Auth 1976, 115, Arveiller, Nenna 2005, 251). The rotated solid glass mass and the non-inflated volume limited to the neck, which is drawn out to a tall cylinder 73 resulting into glass with large elongated air bubbles are striking. The blown solid-based unguentaria are made in strongly blue-green to green glass of mediocre quality (Arveiller, Nenna 2005, 251), whereas these rod-formed ones are apparently made in a dark greenish-brown glass. Related to the thin-walled blown candlestick 73 The dark coloured glass and the thick glass mass full of air bubbles makes one think of poorly heated glass remains used to produce these plump vessels. 71 unguentaria (Isings Form 82b2) in pale ‗naturally coloured‘ glass this vessel shape is integrated in the Period II chrono-typology and not in Period IV, the phase featuring rod-formed vessels [see Chapter 4]. However, where the latter type is ubiquitous for the entire Roman Empire, the blown solid-based pieces remain distinctive to Egypt and the southern Levant (Cosyns et al. 2006, tab.1) [see Chapter 6]. The four pieces from Elkab (EG) (Figure 46) are made from black-appearing deep greenish-brown glass and are 74 thus far the only pieces recorded to have been rod-formed. Because the Louvre Museum possesses a piece in dark green glass with a very similar in shape is to our estimation erroneously described as blown (Arveiller, Nenna 2005, 263, no.792) we assume that a reassessment of the material is required. None of the three base fragments fit the rim fragment but according to the measurements of the preserved height we can estimate the total height at about 100 mm. With a maximum diameter of the body varying between 54 mm and 56 mm the diameter of the three bases range between 50 and 60 mm equalling the dimensions of large solid-based unguentaria. Two unguentaria have a faint constriction at the base of the neck and one is without a constriction. Those with constriction seem to have been turned in a mould to obtain a neat conical shape whereas the one without a constriction has a more concave shaped body and therefore shaped in a free movement. If not flat, the basal surface is concave and repeatedly shows a large pontil scar in the centre. The presence of a pontil mark on the base is logical as the pulling of the neckdemonstrated by the large vertically elongated air bubbles in the glass matrix of the neckand the shaping and folding of the rim necessitated the use of a pontil rod. The external diameter of the neck varies between 17-20 mm and the wall thickness ranges between 3,5-4,5 mm. With an opening of 14 mm and the inner diameter of c. 10 mm restricted to the length of the neck ending at the start of the body these vessels have a very limited usable volume compared to the size of the vessel and the mass of glass used. 3.2.5.2 Form IID.2: Squad candlestick unguentaria with a triangular solid body [Isings 82b2variant/Harden Class XIII,C ] Figure 47: squad candlestick unguentarium with solid body (photograph by courtesy of the Yale University Art Gallery, New Haven) Only one piece of this variant has been inventoried as black appearing glass. The piece at Yale University Art Gallery (unpublished) [cat.no.A4696] is unique and refined in its kind as it is well executed with a marvered upward festooned decoration of glass trails alternating in opaque red and white glass (Figure 47). Straight at neck and rim this festooned pattern is somewhat twisted at the body due to a rotational movement during the manufacturing process. With a height and a base diameter of 70 mm the piece is a fairly large squad unguentarium. 74 Only two out of four unguentaria from Elkab are published in Capart, Fouilles d‘Elkab (3, 1954, 112, nos. 6768): the rim fragment (E 7776d) is erroneously published as E7776a. 72 3.2.5.3 1 Form IVD.1: Tall, slender balsamaria 3. 2. Figure 48: Rod-formed alabastra with twisted body without foot (left) or with foot (right) (drawings by author, by courtesy of Grand Curtius, Liège) 75 The type is purely known as rod-formed vessels even if similarities in blown vessels can be found for the tubular vessels [see Chapter 8]. A range of variants is noticed according to the decoration pattern along with the presence or absence of specific elements such as feet or handles (Figure 48). The dimensions of the ribbed vessels, straight or twisted, range between 90 and 140 mm in height and the body has a maximum diameter ranges between 20-25 mm while the rim diameter can vary between 15 and 25 mm. Numerous museum collections have examples in black glass but also in ultramarine blue glass (Hentrich, von Saldern 1974, 256, no. 750; Auth 1976, 147, no.189; 225, no.490; Whitehouse 2003, 47-48; Israeli 2003, 294, no. 391-392). Other glass hues such as opaque turquoise (Stern 2001, 186, no.78) and greenish glass (Smith, 157-159, no. 321) were much lesser utilized to make rod-formed balsamaria. We were not able to record balsamaria other than the four variants as illustrated in Figure 48: with a rounded base and without foot (variant 1); those with a pad-shaped foot applied around the rounded bottom (variants 2-4) can be distinghuished on the basis of its decoration technique: variant 2 has a body with twisted ribbings; variant 3 has straight vertical or poorly twisted ribbings wrapped with a spiralling trail in opaque glass hue (generally white or blue, but yellow and red arer not excluded); variant 4 has no ribbings but only applied glass trail decoration. We differentiate two subcategories for variant 2 based on some external features connected to the manufacturing technique and thus possibly related to technical knowledge of two different glassworkers/workshops. Subtype A has a high vertical rim and large handles whereas subtype B has a short rounded rim and squat vertical handles. An example of variant 3 balsamaria is the vessel B2403 in the Grand Curtius, Liège with three wide, poorly twisted, vertical ribbings decorated with an opaque turquoise glass trail spiralling 12 times anticlockwise from the shoulder down to just above the disk base, which apparently was added before the trailed decoration was applied (Figure 48:3). The piece has a thick rounded rim with constriction and two vertical handles from rim to shoulder. They were added after the decoration was applied. The vessel is 95,5 mm high and has a maximum diameter of 21,7 mm at the body (unpublished) [cat.no.4446]. An example of variant 4 is the balsamarium in the Israel Museum, Jerusalem (Brosh 2003, 381, no. 516) [cat.no.A4615]. The vessel is decorated by marvering- applied glass trails on a plain shape in zigzags or spiralling on various parts of the body. The example has five spirally wound trails in opaque red glass alternating four times with a single zigzag trail in opaque blue glass. Strangely, this vessel of 120 mm in height and with a diameter of 14 mm is described as being blown and dated in the early Islamic period. In every respect, the piece is similar to the other rod-formed balsamaria: its pad-shaped base, its tall and slender body, its vertical handles from shoulder to rim which protrude slightly above it, and the high, uneven rounded rim and should therefore be considered 75 It was, however, quite surprising to find a tall and slender rod-formed balsamarium from the Dobkin collection in the Israel Museum, Jerusalem described as a blown elongated bottle dated 7th8th century AD (Brosh 2003, 381, no. 516) [cat.no.A4615]. 73 another vessel type of late Roman production. The absence of well-dated finds maintains however such conflicting date ranges [see Chapter 5]. 3.2.5.4 Form IVD.2: Wide-mouthed hemispherical squat jar The most common shape of squat jars is the wide-mouthed jar with hemispherical body comprising various subtypes. Instead of creating an inconvenient number of incoherent types we categorized the wide-mouthed squat jars within three main subclasses based on the decorative style and formal aspects (Figure 49): Variant 1 features an incised decoration showing a vertical or twisted ribbed body; Variant 2 features a glass trail decoration applied in zigzag just below the shoulder in a contrasting opaque glass hue (white, yellow, red, blue); Variant 3 features a stamped decoration making the globular body got flattened sides by means of three large circular stamps. We differentiated the subcategories on the basis of the absence of both handles and foot (a); the 76 presence of handles (b); the presence of both handles and foot (c). Such subdivision is only useful to enable a quantification of the different variations that were produced in view of a study of the production and distribution of such idiosyncratic commodities. The minor variations present in the wide variety of squat hemispherical jars speaks in favour of one or a very few workshops responsible for the production during a relatively short period (one or two generations). Seeing that nearly all recorded material is from unknown provenance makes a more thorough assessment ineffective. Only an in depth appraisal on this commodity type including those in all other glass hues as well as an evaluation of the contemporaneous free-blown toilet ware plus good archaeological finds can enhance our knowledge on the consumption of these vessels and help discerning the idiosyncrasies related to glass workshops. Almost all examples, independent of the subtypes of the squat vessel shape, are about 40–50 mm in height and have a more or less equal diameter (Table 26). In analogy to the tall and slender balsamaria, this generally black glass vessel shape is also known in different glass hues, e.g., turquoise, ultramarine blue, dark green, and a wide variety of decoration techniques. It is striking that plain jars have not been recorded so far and that the various types of decoration never occur combined (Table 25). The vessels are thus decorated with ribbings, or with applied glass trails, or with a stamped pattern (Table 26). The vessel type can occur with or without a foot, but especially the hemispherical squat jar has no foot. When present it always consists of a pad-shaped disc that due to its small size and irregular shape demonstrates to miss its raison d'être. The handles were added separately when the vessel was completed, and are always applied from the shoulder to the rim and in most cases the handles rise above the rim and show tooling with a pincer. Table 26: List of rod-formed squat jars (Form IVD.2) type ribbed 4443 4449 IVD.2:1b IVD.2:1b vertically diagonally - - 38 47 max. Ø body 36 44 4453 IVD.2:1b diagonally - - 46 4460 IVD.2:1b diagonally - - 4468 IVD.2:1b vertically - 4470 IVD.2:1b diagonally 4465 IVD.2:2a - 4441 IVD.2:2b 4459 4448 4451 4442 cat.no. 76 applied trails Ø rim Ø base 30 31 - Israeli 2003, 294, no.393 unpublished 45 31 - Whitehouse 2001, 47, no. 959 50 55 30 - - 47 37 27 - greenish white - 40 40 33 - Hentrich, von Saldern 1974, 188, no.278 Hayes 1975, 121-122, no.479, pl.31 Auth 1976, 147, no.190 - 39 ? 30 - Stern, 2001, no.81 - white - 46 36 25 - Arveiller-Dulong, Nenna 2005, no.1156 IVD.2:2b - white - 39 46-52 ? - Whitehouse 2001, 51, no.969 IVD.2:2c IVD.2:3b IVD.2:3c - white - indefinite lion 47 46 62 43 46 42 36 35 27 13 18 unpublished unpublished Israeli 2003, no.394 stamped h. references A logical fourth alternative should comprise vessels with a foot but without handles. 74 Figure 49: The various subtypes of squat jars with their various decorations (drawings by author, by courtesy of Grand Curtius, Liège; photographs taken from Arveiller, Nenna 2005, no. 1156 and Israeli 2003, no.394) o With ribbed body So far only the baseless hemispherical aryballos-type with handles (Form IVD.2:1b) (Figure 49) and the more slender amphoriskos-type with pad base (Form IVD.3:1) (Figure 51) show examples with ribbed decoration appearing in two variants (Table 26). Some have straight vertical ribbings on the body such as the piece now at the Bible Lands Museum in Jerusalem (Schlick-Nolte 2002, 105, no. V-69) demonstrating that the hot gob became tough when clearly pinched in with a jack. However, the main ribbed design is slightly twisted clockwise with a S-twined pattern. The intact piece now in the Corning Museum of Glass (Whitehouse 2001, 47, no. 959) [cat.no.4453] has 10 ribbings twisted clockwise from base to shoulder. o With applied glass thread decoration Generally, the unmarvered trailed decoration is limited to the shoulder, even if the applied decoration can be as low as halfway down the body or down to the carination for those with a biconical body. The glass trails are only in opaque hues, mainly in white and red or both colours combined. On a few occasions, the rod-formed jars have a trailed decoration in yellow or turquoise glass. The trails always show a zigzag pattern. They can be short or wide, covering a large surface, and they can be sharp or undulating. The applied colours used and the geometric patterns obtained are likewise those on the contemporaneous black glass beads and pendants, although spiralling trailed decorations are fully absent here. The old collections of RGM Cologne encompass a Form IVD.2:2c piece is 48 mm high and has a thick, discshaped foot; two vertical handles (one missing); and an oval body, which is decorated on the shoulder with an applied red opaque glass thread in zigzag (La Baume 1973, B14, pl. 6:3) [cat.no.2105]. Both variants within subtype 3 have an applied glass trail decoration. A piece with a foot and decorated on the shoulder with a short zigzag in opaque white glass is kept in the Grand Curtius, Liège (Figure 49c middle). A piece without a foot in the Corning Museum of Glass collections has a similar unmarvered decoration on the shoulder. Two other fragments of biconical jars coming from Egypt and now at the Flinders Petrie Museum of 75 Egyptian Archaeology, London are decorated with several parallel concentric zigzags alternating in yellow and turquoise glass [cat.nos.930; 993]. o With pressed decoration Squat jars with stamped decoration are rather uncommon, although they perfectly fit within the category of the squat jars with a vertical or obliquely ribbed globular body. A piece with a stamped imprint from the former Eliahu Dobkin collection is now in the Israel Museum, Jerusalem (Israeli, 2003, 295, no. 394) [cat.no.4442] (Figure 49c right). The jar has a pad base; an irregular, globular body; vertical handles from shoulder to rim which partly protrude above it; and a thick, rounded rim with constriction below it. Three circular impressions make the vessel triangular- shaped. This triangular shape is emphasized by applying a vertical handle in each of the three corners. Even though the provenance of the piece at the Grand Curtius, Liège [cat.no.4451] remains unknown, due to its 77 idiosyncratic technique and shape, its resemblance to the Israel Museum piece, and its acquisition source , the Curtius piece most likely originates from somewhere in the Syro-Palestine Levant (Lebanon, Israel, Jordan, Syria). The Curtius jar has no foot, and the design of the stamp is poorly rendered and thus less legible. However, it probably represents an animal looking to the left (towards the rim) (Figure 50). Hence, it is not unthinkable that the Curtius piece has also a stamped lion similar to the ones on the Dobkin piece in the Israel Museum. Nevertheless, a different stamp must have been used, seeing the difference in stamp-diameter; those on the Dobkin piece are c. 16 mm, whereas those on the Curtius piece are c. 14 mm. A bad imprint from the same stamp on the Curtius piece is therefore excluded. The variations in lions stamped on the medallion pendants imply that assorted stamps from one or diverse die cutters were in circulation [see within this chapter 3.3.4.5 the discoid pressed pendants]. The Dobkin piece (with foot) is 64 mm high and has a maximum width of 42 mm, whereas the Curtius piece (without foot) is 39 mm high with a maximum width of 41,5 mm. The foot of the Dobkin piece is 8 mm high and has a diameter of 21 mm. 78 Figure 50: Squat globular jar with pressed decoration (Form IVD.2:3) (photograph and drawings by author, by courtesy of Grand Curtius, Liège) 3.2.5.5 Form IVD.3: Truncated conical jar The jars with a truncated conical body tapering downwards are less common and are only known with a pad base and handles (Figure 51 top). This shape is decorated with large twisted ribbing (variant 1) or with applied short zigzag trail just below the shoulder (variant 2). No conical jars with stamped decoration have been recorded so far. The rather small vertical handles can protrude in a similar way as the balsamaria with twisted decoration. 77 Armand Baar bought the piece in Lebanon in the 1920s before it entered the Grand Curtius. Seeing that it is quite probable that the rod-formed vessel got discovered wherever it entered the antique market a Levantine provenance is highly acceptable. 78 Comparable material in Zahn 1929. Sammlung Baurat Schiller nos. 334-335, pl.16; JGS 7, 1965, 28, fig.3. 76 Figure 51: (top) Truncated conical (Form IVD.3); (bottom) biconical jar (Form IVD.4) Table 27: List of rod-formed conical and biconical jars (Forms IVD.3-4) applied cat.no. type ribbed stamped h. trails 3878 IVD.3 white 68 4450 4463 IVD.3 IVD.3 - 930 IVD.4 - 4452 4471 4472 4462 IVD.4 IVD.4 IVD.4 IVD.4 - 3.2.5.6 white turquoise + white turquoise; yellow white white white red + white max. Ø body 53 Ø rim 26 Ø base 12 references Riemenschneider 1989, 173-174, pl.2:7 unpublished Stern, 2001, no.80 - 62 51 42 31 35 31 20 - 50 50 38 30 unpublished - 68 57 50 45 35 33 33 36 30 25 22 30 20 14 18 17 unpublished Auth 1976, 226, no.492 Auth 1976, 226, no.493 Stern, 2001, no.79 Form IVD.4: Biconical jar Also the jars with a biconical body are less common and are only known with a pad base (Figure 51 bottom). The biconical-shaped jars differ in size from the other shapes, being higher, 50–60 mm, and having a more slender diameter of about 35 mm (Table 27). The decorative style is restricted to applied decoration (mainly bichrome) by means of glass trails in zigzag, single or crosswise by using two different glass hues (white+blue; yellow+blue; white+red). No plain biconical jar or examples with stamped decoration have been reported to this point. Only the jars with biconical and conical body tapering downwards have handles applied in the reverse way, from rim to shoulder. Overall, these rod-formed vessels have two vertical handles, but examples with three handles or no handles also occur. The vertical handles have been applied in two opposed manners: from the rim towards the shoulder, and from shoulder to rim like those on the ribbed-bodied squat jars and tall slender balsamaria. Examples of this type are the biconical jar of the Grand Curtius (Figure 51 bottom) and the more squat version from the former Ernesto Wolf collection and now in the Landesmuseum Württemberg, Stuttgart (DE). They are decorated on the shoulder with applied glass trails in red-on-white crossing zigzag remaining unmarvered (Stern 2001, 187, no. 79). 3.2.6 Summary In total, we could record 650 pieces of vessels, matching with 57 vessel shapes: 85 ex. are corresponding with 15 cast shapes, 423 ex. with 36 free-blown shapes, 24 ex. with 3 mould-blown shapes and 48 ex. with 5 rod-formed 77 79 shapes. Although 207 pieces remained undefined, the amount of diagnostic material assigned to a typological shape is sufficient to draw different conclusions. Ultimately, the list might end up with any known shape or even include singular shapes that only exist in black glass. The relevance of making such a separate typology is for practical reasons to enable quantifications and comparisons between the different vessel shapes in black appearing glass, used techniques and different periods (Tables 28-29). Table 29 clearly demonstrates the dominating presence of two types of free-blown vessels of Period II. The majority of shapes however are known 80 from only one or a few more finds. Most intact vessels are the thick-walled rod-formed vessels from Period IV (Forms IVD) and the carinated beakers from Period II (Form IIB.1). Table 28: chronological overview of the different techniques of black glass vessel production Period I Period II Period III Period IV cast free-blown mould-blown rod-formed The glass vessels appearing black are known in a large set of shapes, but basically black glass is used to produce tableware, even though storage and toilet ware also have been recorded. In comparison with the forms nd st produced during Period II (mid 2 mid 3rd century AD), a larger variety of shapes is observable for the 1 century AD vessels. Besides the drinking and pouring vessels, crockeryplates, dishes, and bowlswas also produced in a deeply coloured glass appearing black in the latter period. Whether cast or blown, the glass vessel shapes appearing black are similar to those in deeply coloured glass. The repertoire of cast vessels consists of shallow hemispherical bowls, type Isings 1 (Form IA.1), which is a form that can be traced back to the late Hellenistic linear-cut bowls, usually manufactured in deep blue, purple or amber brown glass (Grose, 1991). Forms IA.2-4 correspond to the other deeply coloured monochrome glass vessels similar to the Samian ware, such as the bilobated cups and dishes, type Dragendorff 27 (Form IA.2); the hemispherical cup with base-ring, type Dragendorff 40 (Form IA.3); and the small cylindrical cup with a pushed-in base, type Dragendorff 22-23 (Form IA.4). The production of the cups with double handles (Form IA.5) and the skyphoi (Form IA.6a-b) are produced both by the cast as well as the blown technique. These two shapes are also known from abundantly decorated examples with figurative and geometrical scenes and incrustations in gold, silver, bronze and cornelian. The other cast shapes are equally known in deeply coloured glass, such as the less common tablets (Form IA.7) and plates (Form IA.8), or the omnipresent ribbed bowl (Form IA.9), which was also known in clear-coloured and deeply coloured polychrome glass. Unlike these latter shapes, the small dish with overhanging rim (Form IA.10) has been largely produced in colourless glass. The 1st century AD blown vessels are commonly simple items scarcely or without any decoration, such as the cups with a cut rim and abraded concentric lines, Form IB.1 (Isings Form 12); the unguentaria, Forms IB.2-4 (Isings Forms 14; 16; 26); the canthari, Form IB.5 (Isings Forms 36/40); the modioli, Form IB.6 (Isings Form 37); the jugs, Form IB.9 (Isings Form 56b); the urns with arcuated handles, Form IB.11 (Isings Form 64); or the cups with a broad collared rim, Form IB.12 (Isings Form 69). The aesthetic potential from other forms got exploited by using white opaque glass in various ways, contrasting sharply with the black. It is the case for the jugs from Pompeii and Saint-Paul-Trois-Châteaux with white opaque handles that end in medallions with pressed figures, Form IB.18 (Isings Form 54) or the urns from Murviel-les Montpellier and Nîmes with vertical handles in white opaque glass, Form IB.20. This play of black and white colours is also attested on Period II material but then through the application of glass trails in various patterns (see further). The morphological shapes and ornamental details of these objects cannot be attributed to the use of black glass, but to the intent to imitate metalware, as did all monochrome vessels during Period I. Perhaps black glass compared to other dark-coloured glass best satisfied the well-known Roman preference for metalware, since the polishing of black glass renders a more convincingly metalescent aspect than any other glass hue. Various cast 79 Of 72 pieces the production technique could not be determined. A simple deduction gives a ratio 11 pieces per shape, but a closer look shows that some shapes are only represented by one single piece while other shapes are abundantly present (Form IIB.1 = 139 pieces). 80 78 shapes, like Forms IA.2-4, are on the other hand usually seen as imitations of Samian ware, but since they appear to precede the clay models (Stern 1994, 328, no.99), these glass vessels must have copied models in silver and bronze which were also the models for the Samian ware. In a previous publication (Cosyns, Hanut 2005, 113) we stated that all vessel shapes dating from Period II are tableware and exclusively used for drinking: beakers, cups, bowls, jugs and jars. We identified at that time 14 different free-blown vessel forms in a monochrome black glass. To this group, however, needs to be added seven other shapes which are of different use than tableware. In great contrast to that stands the total absence of the cast vessels and the near non-appearance of mould-blown vessel shapes. It‘s possible that new shapes might occur as a result of new field work, but mostly those news finds will fit in this classification. A large number of the cast shapes can be related to the contemporaneous ceramic repertoire, more specifically terra sigillata. Further research on the analogy between black glass vessel shapes of Period II and metallescent ware could provide promising results. The Period IV-vessel production in black glass metal appears to have been restricted to the rod-formed technique and limited to the Levant [see Chapter 6]. For the late Roman/early Byzantine period – Period IV – we have no knowledge of free-blown vessels having been produced in black glass, but black appearing mould blown vessels th th in deeply coloured purple or brown glass characteristic to the 6 -7 century AD probably were already in circulation in the late Roman period. In a first attempt we focussed on the Western distribution (Cosyns, Hanut 2005; Cosyns et al. 2006a; Cosyns, Fontaine 2009). The deeply coloured vessels appearing black from the Eastern provinces still await systematic investigation. Missing a detailed corpus incorporating a chronological and distribution study [respectively Chapters 4 and 6], we can, nevertheless, make some tentative deductions from the examples registered (Table 29): Black-appearing vessels were distributed within the entire Roman Empire; 1) 2) 3) black-appearing vessels were made in all available techniques and styles; particular vessel shapes in black-appearing glass were produced in specific periods of the entire imperial period; black-appearing vessel shapes mainly remained unique pieces and were only on rare occasions massproduced. Period II Period I Table 29: Frequency list of the different shapes in black glass recorded per period and per technique (n = quantity) cast [A] free-blown [B] mould-blown [C] rod-formed [D] shape shape shape shape n n n n Form IA.1 8 Form IB.1 5 Form IC.1 2 Form IA.2 10 Form IB.2 3 Form IC.2 1 Form IA.3 7 Form IB.3 5 Form IC.3 1 Form IA.4 11 Form IB.4 3 Form IC.4 1 Form IA.5 1 Form IB.5 6 Form IC.5 1 Form IA.6 1 Form IB.6 1 Form IA.7 18 Form IB.7 3 Form IA.8 10 Form IB.8 3 Form IA.9 2 Form IB.9 2 Form IA.10 2 Form IB.10 3 Form IA.11 1 Form IB.11 4 Form IA.12 1 Form IB.12 1 Form IA.13 1 Form IB.13 2 Form IA.14 6 Form IB.14 2 Form IB.15 2 Form IB.16 1 Form IB.17 3 Form IB.18 1 Form IB.19 2 139 Form IIB.1 Form IIC.1 10 Form IID.1 4 Form IIB.2 1 Form IIC.2 8 Form IID.2 1 Form IIB.3 2 62 Form IIB.4 Form IIB.5 11 79 Period III Form IIB.6 Form IIB.7 Form IIB.8 Form IIB.9 Form IIB.10 Form IIB.11 Form IIB.12 Form IIB.13 Form IIB.14 Form IIB.15 Form IIB.16 Form IIB.17 Form IIB.18 Form IIB.19 Form IIB.20 Form IIIB.1 Form IIIB.2 Form IIIB.3 3 1 2 1 2 6 1 3 2 1 5 11 2 2 4 1 1 2 Period IV Form IVC.1 1 Form IVD.1 Form IVD.2 Form IVD.3 Form IVD.4 5 28 3 6 3.3 Jewellery The Roman jewellery made of black glass involves primarily bracelets, finger rings and beads, but also incorporates pendants, gems and hair pins. We primarily based the classification of the different jewellery types on some major works like that of the vessels. Concerning the bracelets we however elaborated a typology to answer two dominant questions is this study: 1) whether the artefacts in the entire empire are to be regarded either as a homogeneous group of analogous material or as a more heterogeneous group of homologous material [see Chapter 6]; 2) whether the black glass jewellery was only customary for the later imperial period or a typical Roman commodity in use during almost the entire imperial period, as Emilie Riha‘s research presents on the Augst/Kaiseraugst material (1990, 66, tab.82) [see Chapter 4]. The classification of the finger rings is arranged according to the principles used for the typology of the bracelets. 3.3.1 3.3.1.1 Arm rings Preceding research The study of Roman bracelets is a relatively new and barely examined artefact type within the material studies 81 resulting in limited and regular flawed identifications. The first to touch on the subject discussed the very idiosyncratic bangles of the British Isles (Kilbride-Jones 1938; Stevenson 1956; 1976) a class which is nearly totally absent on the Continent (van Lith 1977; Isings 2009) or dealt with it in the margin of late Iron Age glass 81 It is striking that even glass specialist have difficulties to distinguish Period II-III black glass bangle fragments from handles of Period I vessels (Foy 2010a, 324, nos.570-571). 80 bangles (Haevernick 1960). limited finds: 82 More elaborate studies on the classification of Roman glass bangles, focussed on From a present day country – Belgium (De Witte 1977), Israel (Spaer 1988); From a specific region – the Upper Rhine region (Wagner 2006); From a particular site – Augst/Kaiseraugst (CH) (Riha 1990). These publications excel in drawing up each time a typology based on the peculiarities of the available material demonstrating the local and regional occurrence of the studied material. But only the integration of the material from all areas into one discussion makes possible to recognize regional idiosyncrasies. Only recently some publications attempted to emphasize the importance of enlarging the study area of this particular commodity type to the entire Roman Empire, considering the omnipresence of black glass bracelets recorded in all countries that formed part of the Roman Empire (Cosyns 2003b; Cosyns 2004, 15; Cosyns et al. 2006a). The Roman glass bracelets are, in general, distinct from the late Iron Age in terms of 1) the colour(s) of glass used; 2) the applied technique [see Chapter 2]; and 3) the shapes and patterned design. During the late Iron Age, 83 the applied glass hues were translucent blue, purple, brown, blue-green, green and colourless glass. Conversely relatively few Roman glass bangles were made of clear translucent coloured glass in Roman times, 84 showing a preponderance of strongly coloured glass appearing opaque black. The glass that makes the object appear black is made from a poorly translucent glass. When using a strong light source in transmission, the observed colours of these so-called black glasses are mainly bottle-green or olive-green, but blue-green also is present. In very rare cases, the bangle was made of an amber brown glass. The use of deeply purple-coloured glass for the production of bracelets is only noticed within the eastern Mediterranean and it was popular only from th th the early Byzantine period onwards (5 6 century AD). Although these bracelets are nearly always monochrome, some uncommon types are bichrome or even polychrome when decorated with applied glass trails or speckled with spots of opaque glass. The introduction of the rod-made technique in Roman times resulted in specific shapes distinct from those fashioned in the late Iron Age [see Chapter 2]. Characteristic to the la Tène bangles is the total absence of open shapes and the frequent use of applied glass zigzag trails in opaque white or yellow glass. In Roman times open shapes regularly occur and are a typical result of the rod-made technique. Applied glass trails on Roman glass bangles are in contrast rather unusual but when applied nearly always red, blue and white opaque glass is used. Yellow opaque glass, regularly used to decorate the la Tène bangles, seems to have been (re-)introduced at the end of the late Roman period [see Chapter 4]. The research on the Romano-British glass bangles is in a sense the initial study on the Roman glass bracelets (Kilbride-Jones 1938; Stevenson 1956; 1976; Price 1988). They are not related to the late la Tène production and very different from the continental Roman productions. Clearly dissimilar to the continental and Mediterranean material these glass bangles are characterized by a very regional production and consumption. The RomanBritish glass bangles are markedly polychrome; featuring various gaudy colour combinations on a core in bluegreen or white glass and in a lesser extent in black glass; whereas the Roman glass bangles on the continent and in the Mediterranean remain sober in monochrome black glass. Three main categories have been recognized in the Romano-British glass bangles by distinction of three features: 1) section, 2) used colour and 3) decoration pattern, where the latter two features have been used to create subtypes separately and in combination (Kilbride-Jones 1938, 366-367). Further research however made clear that this classification is not able to integrate a number of pieces (Price 1988) imposing a reassessment of the material 85 in general and the current classification system in particular. o Type 1: Dated IB AD Type 1 bracelets are characterised by a wide D-shaped section; obliquely applied glass trails or cables eventually in combination with dots. They appear mainly in a translucent, yellowish-green glass but purplish glass has been 82 Thea Haevernick discussed only eleven pieces of Roman glass bangles all together without distinction as variant A of the late La Tène type 17, Armringe mit eingestochener Verzierung (‗bracelets with incised decoration‘) (Haevernick 1960, 65-66; 209-210, Taf. 14). 83 The very late La Tène material of type Haevernick 3 is regularly of such poor translucent glass they are black appearing. This is mainly the case for the bangles in purple glass and in a lesser degree for the blue and brown glass. 84 Bangles in coloured and decolorized glass are essentially noticed in the eastern Mediterranean such as Israel (Spaer 1988) and Lebanon (Chéhab 1985-1986). 85 Based on the chemical analysis results of some material from London (UK) we assume the study of Romano-British bangles most important within the issue of local glass production and consumption during the Roman imperial period [see Chapter 8]. 81 reported too. The decoration is mostly a combination of opaque yellow and blue glass, but white and green do also occur. o Type 2: Dated IB – IIB/IIIA AD Type 2 bracelets are characterised by a medium-sized triangular shaped section; mainly in translucent blue and blue-green glass; lengthwise applied twisted bichrome glass cables (mainly combining blue and white opaque glass) the high amount of variations can be grouped into two subtypes: 1) with several twisted glass cables applied parallel lengthwise and marvered; 2: with a single twisted glass cable at the centre of the top surface, applied lengthwise and not marvered; sometimes also oval or spiral blobs are applied aside or on top of the cables or bands. o Type 3: Dated I - IV AD This group of bracelets consists of ten sub-types on the basis of the glass hue (A-J) each comprising a number of subcategories based on variations in section and decorative pattern (Kilbride-Jones 1938, 376-389); characterised by a medium sized angular D-shaped section that can show multiple variations; [glass hue] mainly in opaque glassyellow (subtypes B; E) or white (subtype A; C-D)and few in translucent glass - decolourized (subtypes F-G), yellowish-green (subtypes H) and blue (subtypes I-J); [decoration] plain (subtypes A-B) or decorated with scrolls or pothooks marvered glass trail in contrasting colours (subtypes C-I). Various patterns of applied glass threads in various colours on top of a mainly triangular section in monochrome translucent blue, blue-green, opaque white, yellow or black glass are identifiable. The wide variety of colour combinations and application patterns in decorating the bangles created a large amount of sub-types which are totally absent on the Continent. Consequently, the formulated typology on Romano-British glass bangles is totally unworkable in the determination of the glass bangles from the rest of the Roman Empire. However, we briefly discuss the Romano-British bangles that have been made in black-appearing glass all through the different chapters to compare the idiosyncrasy of this commodity vis-à-vis the continental bangles, showing not only an isolated regional distribution but also an unrelated chronology. However, a separate research on this particular topic is needed to verify whether these discriminated peculiarities can be linked to one or more workshops. In 1977, Hubert De Witte catalogued the pre-Roman and Roman bracelets from Belgium for his master‘s thesis at the University of Ghent (De Witte 1977), primarily to compare the obtained results on pre-Roman bangles with the then-recently published results from the Netherlands (Peddemors 1975). However, he also incorporated a detailed, highly structured and workable typology on Roman glass bangles by dividing the studied material into major categories in accordance with the sectionin relation to the used technologyand by sub-dividing each type into variants on the basis of the various decoration patterns. All limitations in his typology are due to the restricted area he studied and to the fact that over the last 30 years, much more material has come to light. Unfortunately, his work never got published and remained more or less unused, creating the necessity for others update and fine-tune this workable typology. Hubert De Witte divided all Roman glass bangles firstly into two key divisionsi.e., open (a) and closed (b) bracelets (Figure 52). The open bracelets are elliptical, with a space between both rod ends, while the closed bracelets are circular, with a seam that connects both rod ends or seamless when gathered, swirled and cone rolled. He further identified four major categories (A-D) based on the section of the bracelete.g., O-shaped, Dshaped, and the used technologythe typical Roman technique of using glass rods or the Iron Age technique of gathering some glass with a metal rod [see Chapter 2]. The subdivisions of each category are based on the applied decoration. For the rod-made bracelets with an O-shaped section, i.e., type A, De Witte identified four sub-categories, which are equally present in both open and closed bracelets: type aA1/bA1 bracelets are plain; type Aa2/Ab2 bracelets have wide twists and type aA3/bA3 small twists; and type aA4/bA4 have alternating plain zones and twisted decoration. The rod-made bracelets in Belgium with a D-shaped section appear to be open when plain (aB1) and closed when decorated with pressed-in decoration by means of an indented spatula (type bB1). The rod-formed bracelets with lengthwise ribbings are only known as open specimens: type C1 has five ribbings and type C2 has more than five ribbings. The seamless bracelets with small D-shaped sections are only known as closed bracelets. 82 Figure 52: Typology of Roman glass bracelets set up by Hubert De Witte (taken from De Witte 1977, 21) In the late 1980s, Maud Spaer worked out a typology on the pre-Islamic glass bracelets from Palestine (Spaer rd th 1988). For that material she has discussed dates from the 3 to the 7 century AD, which coincide with the late Roman and early Byzantine period. The typology developed by Spaer is done the opposite way compared to the one by De Witte. Her major categories are based on decoration, whereas further subdivisions were executed according to the section, which is related to the technology. Interestingly the Roman bangles from Palestine Maud Spaer discusses are all closed, demonstrating that the open bangles are atypical for the south-eastern Levant. Type A includes the monochrome undecorated bracelets: Type A1 pieces have an O-shaped section and have a seam. Type A2 pieces have a D-shaped section (with a rounded [a] or flattened [b] outer surface). Type B includes tooled or moulded and monochrome bracelets and is Spaer‘s major group, consisting of six categories on the basis of the decoration: Type B1 pieces can be seamed as well as seamless and have facetted ornamentation. Type B2 pieces can be seamless and seamed and their decoration consists of vertical ribbings. Spaer subdivides this type into four subcategories on the base of different parameters, i.e., B2a is seamed; B2b is seamless; B2c has three lengthwise ribbings; B2d has a ‗bulging inside section‘. Type B3 pieces are mainly seamless and have diagonal ribbings. Type B4 pieces are mainly seamed and their typical features are the lengthwise ribbings, three to six, and without (a) or with (b) bezel. Type B5 pieces have a D-shaped section, are decorated with round or other protuberances, and are mainly seamless. Different subcategories were identified: bracelets with one to three rows of round or ovoid protuberances pressed or tooled (a); one row of large protuberances, stamped singly (b); and Cshaped protuberances with uncertain technique. Type B6 pieces are seamless and are characterised by stamped symbolic motifs; they can be Jewish (a); Christian (b); pagan (c); or uncertain (d). 83 Type C bracelets are seamed, have a circular section and are spirally twisted: Type C1 pieces are monochrome and can be densely (a) or loosely (b) twisted. Type C2 pieces are multicoloured with the application of glass threads in contrasting colour(s). Type D bracelets are multicoloured in all kinds of variants beside those with twisted decoration, i.e., Type C2. This type is more seen as an Islamic type; however, some few variants with a D-shaped section do occur in the late Roman-early Byzantine period too: Type D1 is decorated with slightly protruding coloured specks. In 1990, Emilie Riha published her study of the Roman jewellery from Augst and Kaiseraugst. In one chapter, she discusses the bracelets (Riha 1990, 52-66) with only the last three pages of it describing the different types of glass bracelets recorded from the excavations in Augst and Kaiseraugst (Riha 1990, 64-66). She discerned five types of bracelets. However, based on intermittent criteria (colour, section, and decoration), the classification remains somewhat indefinite as well as limited due to making a classification on material from only one site. The types established by Emilie Riha are as follows: Type 3.30 bracelets have a wide D-shaped section and coincide more or less with type C by Hubert De Witte by incorporating all bracelets with lengthwise ribbings. However, some bracelets of De Witte‘s type B are also included (Riha 1990, no. 625). Type 3.31 integrates a large part of De Witte‘s type A and concerns all kinds of bracelets with an O-shaped section. Riha has subdivided this type into five sub-categories: 3.31.1: with twisted decoration but in a (non-black) translucent glass 3.31.2: with twisted decoration in black glass 3.31.3: with a twisted decoration but in combination with a different coloured glass (this type is not taken separately by De Witte, but in our extended typology it corresponds with A6) 3.31.4: with wide twisted decoration (they match with De Witte‘s type A2) 3.31.5: having alternating twisted decoration and plain zones (these bracelets correspond with De Witte‘s type A4) Type 3.32 are open bracelets decorated with applied ribbings giving a roughly star-shaped section (these bracelets fit with De Witte‘s type A2) Type 3.33 are plain open bracelets with an O-shaped section with a diameter wider than 8 mm (these bracelets correspond with De Witte‘s type A1) Type 3.34 are closed bracelets with a D-shaped or O-shaped section with a diameter smaller than 5 mm (the Dshaped bracelets correspond with De Witte‘s type D and the O-shaped bracelets with type A1) Reassessing the pre-Roman and Roman glass bracelets from Belgium, we became conscious that there was a necessity for an update of the typology conceived by De Witte, as time and extensive excavations involved the predictable additions and adaptations. The preliminary results (Cosyns 2003a; Cosyns 2004) required an improved discussion and reassessment within this chapter [see below in this chapter]. Recently, Heiko Wagner published the results of his PhD research on La Tène beads and bracelets from the upper Rhine region, incorporating also the Roman material from Luxemburg, the northeast of France, and the Rhine region in Germany (Wagner 2006, 155-171; 311-321). Besides using Riha‘s typology (Types 30; 32; 31.2; 31; 31.3; 34), he distinguished many more types (in italics and between brackets a concordance with our typology established here appears below as Wagner‘s subdivisions are not taken into consideration in Table 30): (1) Römische Armringe mit D-förmigem Querschnitt, schräg gekerbt (= Type B3) (2) Römische Armringe mit D-förmigem Querschnitt, mit einem kammartigen Werkzeug schräg eingekerbt (= Type B3) (3) Römische Armringe mit D-förmigem Querschnitt, mit einem kammartigen Werkzeug oder Rädchen quer gekerbt (= Type B2) 84 (4) Römische Armringe mit ovalen (augenförmigen) Eindrücken, mit einem hohlen, röhrenartigen Werkzeug eingestochen (= Type B5) (5) Zweirippige römische Armringe ohne Querkerbung (= Type C4) (6) Mehrrippige römische Armringe ohne Querkerbung (= Type C) (7) Zweirippige Armringe (= Type C4) (8) Dreirippige Armringe mit quer gekerbter Mittelrippe (= Type C3) (9) Armringe mit drei gleich hohen Rippen, alle Rippen mit einem mehrzinkigen Gerät quer gekerbt (= Type C3) (10) Armringe mit fünf glatten Rippen (= Type C1) (11) Dreirippige römische Armringe mit Querkerbung (= Type C3) (12) Mehrrippige römische Armringe mit Querkerbung (= Type C) (13) Mehrrippige römische Armringe mit Zinkengerät quer gekerbt (mindestens drei Längsrippen) (= Type C) (14) Fünfrippige Armringe mit quer gekerbter Mittelrippe (= Type C1) (15) Fünfrippige Armringe, Mittelrippe mit einem mehrzinkigen Werkzeug quer gekerbt (= Type C1) (16) Mehrrippige Armringe (ca. 6 Rippen) (= Type C2) (17) Armring mit unbekannter Rippenanzahl (zwei Rippen erhalten, Ansätze von zwei weiteren Rippen) (= Type C) (18) Armring mit vierkantigen Querschnitt (= uncategorized – Type U1) (19) Römische Armringe mit dreieckigem Querschnitt (= uncategorized – Type U2) (20) Tordierte römische Glasarmringe (ohne Unterscheidung in Varianten) (= Type A2-6) (21) Offene, rundstabige Armringe (= Type A) (22) Geschlossene, rundstabige Armringe (= Type A) (23) Armringe mit ehemals rundem oder ovalen Querschnitt (offen oder geschlossen?) (= Type A) (24) Römische Armringe mit rundem oder ovalem Querschnitt (unverziert), offen (= Type A1) (25) Römische Armringe mit rundem Querschnitt offen oder geschlossen (= Type A) (26) Römische Armringe mit rundem Querschnitt geschlossen (= Type A) This list of Roman glass bracelet types proposed by Heiko Wagner shows that different criteria in subdividing the bracelets have been mixed up, similar to Emilie Riha‘s work. According to us Heiko Wagner therefore has developed a high number of irrelevant types. A number of types are featuring very vague characteristics creating irrelevant categories comprising only one, two or at the utmost three pieces that can be reduced to for instance different variants of one single general category (see, for example, nos. 21-26 in the list above). That's why his subdivisions are too far-fetched in differentiations and consequently not useful. The difficulty in using such a sorting is the absence of particular technological or decorative features, to come to a methodological classification with archaeological relevance. This makes it more problematic to verify the significance of each specific type with regard to chronology, distribution and use. 3.3.1.2 Proposed typology We are aware that the abovementioned criticism in the introduction amplifies the restrictions in mapping out a typology rather than its use, indicating the limitations of any proposed typology, not in the least ours. Nevertheless, a typology can be useful when it is workable vis-à-vis the parameters important to archaeological research, e.g., chronology, technology, or distribution. Therefore, we decided not to make use of the subdivisions by Maud Spaer (1988), Emilie Riha (1990) or Heiko Wagner (2006) as their typologies are not suitable for chronological purposes or remain too general by using too disparate categories. A concordance list of all previously established typologies with our own will make easier conversion to the various classification systems (Figure 53; Table 30). Not to start from scratch, we have developed a typology that is based on the typology conceived by Hubert De Witte in 1977 (Figure 52). Hence, our typology too is in the first place based on technology. Accordingly, the section of the bracelet is essential. Four major classes could be discriminated, of which the first three were manufactured with an analogous technology, by using glass rods. The bracelets of the fourth class are shaped by gathering hot glass from the batch [see Chapter 2]. The fifth remains uncategorized as it is not clear from the few examples recorded so far what the specific technological characteristics are to produce these bangles with quadrangular and triangular section. Each of the five major categories discerned have been subdivided on the basis of the various decoration patterns. Nevertheless, we also tried to simplify De Witte‘s classification model. For example, the open/closed feature of bracelets, which seemed essential to him, is not considered imperative because there is no chronological difference between the open and closed types [see Chapter 4]. Nonetheless, we continued to take this visual peculiarity into consideration as it appears to be a regional feature and thus it has to be linked to local production with regional distribution range [see Chapter 6]. We therefore took notice in the data base whether the piece could be determined as being part of an open (a) or a closed (b) arm ring. On the 85 other hand, we added many other types essentially by including the Mediterranean material but also from all kinds of newly registered material since the late 1970s found in north-western Europe. variable section according to the number of lengthwise ribbings Type Type Type A B C A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3 A4 B4 C4 D4 A5 B5 A6 Type D D5 D6 D7 D8 Figure 53: Typology of the black glass bracelets (drawings by the author) Table 30: Concordance list of the proposed typologies on Roman glass bracelets. [* our taxonomy is a final elaboration of our previously proposed classification system (Cosyns 2003; 2004) which is based on the typology proposed by Hubert De Witte (1977)] Cosyns Features De Witte Spaer Riha 2009* 1977 1988 1990 O-shaped section Type A open and/or closed seam A1 plain Aa1/Ab1 A1 3.33; 3.34 A2 wide twists Aa2/Ab2 C1b 3.31.1/2/4; 3.32 A3 fine twists Aa3/Ab3 C1a 3.31.1/2 A4 alternating plain-twisted Aa4/Ab4 not incorporated 3.31.5 decoration A5 alternating fine-wide twists not incorporated not incorporated not incorporated A6 twisted decoration with applied not incorporated C2 3.31.3 glass thread(s) D-shaped section (wide) Type B open and/or closed seam B1 plain Ba1 A2 3.30 B2 notched with indented spatula Bb1 not incorporated 3.30 B3 vertically or obliquely ribbed not incorporated B2 not incorporated B4 impressed circular or oval not incorporated B5 not incorporated protuberances B5 impressed complex not incorporated not incorporated not incorporated 86 Type C C1 C2 C3 C4 Type D D1 D2 D3 D4 D5 D6 D7 D8 Uncategorized types U1 U2  D-shaped section (very wide and flattened) open 5 lengthwise ribbings Ca1 6 or more lengthwise ribbings Ca2 3 lengthwise ribbings not incorporated 4 lengthwise ribbings not incorporated D-shaped section (narrow) closed seamless plain D vertically or obliquely ribbed not incorporated impressed theatre mask not incorporated polygonal shape not incorporated stamped buttons not incorporated ‗crumbed‘ decoration (with not incorporated specks) in-line dotted decoration not incorporated impressed complex not incorporated angular-shaped section ? ? quadrangular-shaped section not incorporated triangular shaped section not incorporated not incorporated not incorporated B2c not incorporated 3.30 3.30 3.30 3.30 A2 B2 not incorporated B1;B6a-d B5a; B5b D1 3.34 not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated not incorporated Type A The main feature of the type A bracelets is the O-shaped section made from a rod. The closed bracelets have a constant section-diameter and are circular shaped, whereas the open ones have a section-diameter narrowing towards the ends and are elliptical shaped. The rod ends of the open bracelet can be simply (fire-)rounded and straigth as if these ends were cut by means of jacks, but others are spatula-shaped by using pincers – in rare occasions showing an elaborate variation, such as the piece from Nijmegen (NL) with a vertical crest on top of the spatula-shaped end (unpublished) [cat.no.2968] (Figure 54). This type comprises an extended range of variations (Figure 53): o Type A1: undecorated bangles This ubiquitous type has been recorded from Belgium to Israel. All of them are made in black-appearing glass. Closed bangles have been found everywhere, whereas the open bangles appear to be rather regionally confined to the north-western provinces [see Chapter 6]. o Type A2: wide twisted bangles The rods used to manufacture this type of bangle contained several round ribbings all around at regular distances, resulting in twisted ribbings by turning around the rod when hot. Depending on the direction, the twists are Stwined or Z-twined. All known examples are made in black-appearing glass. o Type A3: fine twisted bangles Similar to Type A2, the used rods contained several round ribbings all around at regular distances, but they probably were smaller and applied closer to each other. Depending on the direction, the twists are S-twined or Ztwined. All known examples are made in black-appearing glass; 87 Figure 54: Open bracelet of type A4 from Nijmegen showing with an unusual crested spatula-shaped end (drawing by the author, by courtesy of the Gemeentelijke Archeologische Dienst, Nijmegen) o Type A4: bangles with alternately plain and twisted zones Bracelets of this type have mainly 4 to 6 fine twists. A piece from Oudenburg (BE) [cat.no.473] and one from Trier (DE) [cat.no.2039] by contrast have 2 wide twists alternating with a plain zone. Open, elliptical as well as closed, circular variants occur, but the majority appears to be open, elliptical. The fragment from Matagne-la-Grande (BE) has a very flat and round spatula-shaped end (Cattelain, Paridaens 2009, 111, fig.50:3) [cat.no.434], whereas the end of Nijmegen piece [cat.no.2968] is elaborate (Figure 54). o Type A5: bangles with alternating fine and wide twisted decoration This type of bangle combines in one single piece the wide twists of Type A2 and the fine twists of Type A3. With an alternation of every 3 to 6 twists only very small pieces can be confound with Type A2 or A3 fragments. o Type A6: twisted bangles with applied glass trails This type of bracelet is similar to Type A2 (with wide twists) or to Type A3 (with fine twists) but includes a decoration of applied twisted trails of opaque glass in a contrasting pattern of every such-and-such twists (mainly 86 every two or four twists). Normally, opaque white or a pale yellow is used, but red and blue also occur. On rare occasions, combinations of white, yellow, red and blue have been applied. Patterns can be seen from some waste material found at les Houis demonstrating trail patterns in blue/red/white; blue/white/red (Cosyns 2009), and a piece from the sanctuary of Bastendorf (LU) (Wagner 2006, 311, no. R3) [cat.no.2856] showing a combination of 1 red/3 yellow/1 turquoise lines. The mass of small fragments of this bracelet type that has been retrieved at Les Houis nearby Sainte Menehould (FR) as well as production waste with a reddish-brown surface indicate that Type A6 bracelets have been produced in the workshop(s) of Les Houis. On rare occasions, bracelets have decorations combining different colours: white, blue and red. Due to the rather limited material known, no further models can be classified so far. Remarkably, the bichrome and polychrome decoration with applied glass threads in opaque white, yellow, red and blue glass is limited to those bangles with an O-shaped section. It is notable that these polychrome-decorated black glass bangles have been very rarely observed on excavations. Is this observation indicative of its unpopularity or very brief prevailing fashion, or is it an indicator of the very partial knowledge of the late Roman 87 material culture we have at this time from archaeological evidence? Finally we have to acknowledge that a number of bangles catalogued as Type A2 bangles are somewhat deviant featuring three to seven protruding and unwinded lengthwise ribbings showing an O-shaped section with three or more protrusions close to each other or a starlike section with protrusions all around the O-shaped section. A closer look reveals that they appear to be malformed or discarded unfinished Type A2-6 pieces and for that 86 The pale yellow hue can be coincidental because of alterations occurring under specific furnace conditions, in particular when the glassworker did not maintain full control over the temperature of the glass melt. Opaque white and yellow glass show the same chemical composition, making it difficult to determine whether the obtained colour is intentional or coincidental (Gérardin et al. 2010, 8). This aspect is in the opposite way attested on the late La Tène glass bracelets featuring a zigzag in yellow opaque glass that sometimes turns out very pale yellow or even opaque white. 87 Large amounts of similar material from so-called Islamic Egypt in various museum collections, e.g. Petrie Museum for Egyptian Archaeology and British Museum in London, might indicate that this material is perhaps erroneously considered post-Roman. 88 88 reason they should be considered as production waste. We assume that the pieces reveal the production technique of all sorts of ‗Type A‘-bangles with twisted ribbings used by Roman glassworkers in the north-western provinces, possibly from one glass workshop. Those with three ribbings close to each other were to form the Type A4 bangles with alternating plain and twisted zones. Those with some space between large ribbings were to create Type A2 bangles with wide twistse.g. the piece from Oudenburg (BE) [cat.no.478]. Type A3 bangles were made of rods with ribbings close to each other and those with alternating close ribbings and space in between generated Type A5 bangles. Strangely some pieces appear to be perfectly well finished products and were found in a clear consumers‘ context far from any production area such as the piece from Binnenmaas (NL) (Holland 30-6, 1998, 356, fig.4) [cat.no.A4490]. It is thus not excluded that we have to deal with a separate type that seems to have been produced in Augst/Kaiseraugst (CH) or in any case that was rather popular within the area. Because the variations within type A demonstrate an analogous tooling we may assume that bracelets of Types A1-A6 could have been made in one single workshop. The challenge here has been to verify whether specific types were produced in specific workshops and whether different glass workshops producing similar black glass bracelets can be distinguished [see Chapter 6].  Type B This type of bracelet is characterized by a wide D-shaped section with a width ranging above 10 mm but rarely exceeding 20 mm. No example until now is preserved completely, but based on the widening towards the middle and the rather elliptical shape we can conclude that these bracelets are mostly, if not all, open ones. Less common are the closed ones with circular shape. Characteristic to this variation in shape is the occurrence of a seam connecting the two ends of the glass rod [see Chapter 2]. o o Type B1: undecorated bangles Type B2: wide transversally lobed decoration by notching with indented spatula The vertical lines of dashes pressed with comb-like tool with pointed teeth create squarish lobes decreasing towards the end when the bracelet is open. Based on the number of dashes generating the indented lines different comb-like tools have been applied with 6 to more than 10 teeth. Sometimes the used tool is too small to cover the entire width of the bangle leaving two, partly superposed short indented lines. Sometimes the tool is clearly wider than the artefact. o Type B3: diagonally incised lobes or protruding ribbings Type B3 bangles have similar notchings with indented spatula as Type B2 bangles but obliquely applied so that the formed ribbings are also at an angle. o Type B4: decorated bracelets with impressed circular or oval protuberances Type B4 pieces are decorated with round, oval or C-shaped protuberances stamped, pressed or tooled in three or four rows. o Type B5: bangles decorated with complex pressed patterns Only few have been recorded so far. One is coming from a burial of the early Anglo-Saxon cemetery at Orpington (UK) (Tester 1968, 125-150) [cat.no.4170] and another is coming from Chessel Down (UK) (White 1988, 108, no.3, fig.54.2) [cat.no.3976]. The Orpington piece is unique showing continuous single band with complex design which consists of a four-petalled ‗flower‘ pressed between two parallel pairs of lines of six or seven dashes pressed with comb-like implements with pointed teeth.  Type C This type of bracelet has a wide and flat D-shaped section and characteristic are the lengthwise ribbings. This type is only known as open bracelet with an elliptical shape and tapering towards the ends. 88 It is not excluded that some pieces are misshapen and discarded rods to form Type C bangles such as the pieces from Éprave (BE) [cat.no.270]; Liberchies (BE) [cat.no.369] and Cologne-Müngersdorf (DE) [cat.no.1837]. 89 o Type C1 has five lengthwise ribbings and exists in different subcategories: (a) undecorated; (b) decorated with transverse indents on the central ribbing; (c) decorated with transverse indents on two or more ribbings. o Type C2 has six lengthwise ribbings or more and exists in different subcategorie: (a) undecorated; (b) decorated with transverse indents over the entire width except the outer ribbings; (c) decorated with transverse indents over the entire width. o Type C3 has three lengthwise ribbings and exists in different subcategories: (a) undecorated; (b) decorated with transverse indents on the central ribbing; (c) decorated with wide pinched in transverse ribbings on all ribbings. o Type C4 has four lengthwise ribbings Only recognized a few times in a fragmentary state we do not know whether this bangle type exists in different subcategories  Type D This type of bracelet has a small D-shaped section with a width ranging below 10 mm and averaging around 7-8 mm. Its width and height are slightly irregular, and the different edges show an oblique shape, resulting in a different interior diameter on both sides. The more pointed side has a smaller inner diameter compared to the other side, which is rounded. This is the result of rolling down the bracelet over a cone to reach the desired diameter. Another difference with the plain type B bracelets is the technique used to make these bracelets. While the type B1 bangles are open or seamed when closed, the type D arm rings are definitely closed and seamless [see Chapter 2]. o o Type D1: undecorated bracelets Type D2: bangles with incised ribbings These bracelets can be decorated with vertical (a), diagonal (b), or crescent shaped (c) incisions; o Type D3: pressed or stamped figurative decoration. This type of bracelets is decorated with one or more impressed, grotesque faces, theatre masks or Negroid faces, but it is not excluded that sooner or later black glass bangles will be excavated with other representations such as the depiction of the Medusa head. Because identical types of impressed decoration are present on the double perforated pressed beads [see in this chapter 3.3.3. Beads]; o Type D4 has a stamped decoration mainly resulting in a polygonal shape. (a) plain facets, normally straight but sometimes concave; (b) decorated with an abstract motif such as the waffled pattern on a bracelet from Kaiseraugst (CH) [cat.no.3677]; (c) decorated with a figurative motif. A popular ornamentation described by Maud Spaer when discussing the Palestine material is the stamped symbolic motif (Spaer type B6): Jewish (1); Christian (2); pagan (3); uncertain (4). o Type D5: bracelets with stamped buttons Bracelets with a stamped decoration have one or multiple rows of buttons which could be close to each other or more distant from each other. The buttons are in most cases round but oval shaped buttons sometimes occur. In general the buttons form a simple relief popping up smoothly from the bangle surface, but the bangles with single button line regularly show the buttons are pushed in the surface left in a circular sink mark. 90 o Type D6: bracelets with ‗crumbed‘ decoration Bracelets with ‗crumbed‘ decoration show a polychrome mottled upper surface with marvered or unmarvered droplets of opaque glass in white, yellow, red and blue; o Type D7: bracelets with applied dotted decoration Such bangles are decorated at the top with unmarvered dots in opaque coloured glass placed on a line. Two examples come from the late Roman layers at the amphitheatre of Metz (FR), one with white dots at an 89 interval of 23 mm [cat.no.1270] and one with blue dots. o  Type D8: bracelets with a pressed/tooled decoration showing a complex design Uncategorized This type of bracelet is technologically speaking linked with types B and C bangles as they are also made by using glass rods leaving a seam when the artefact was made a closed bangleexcluding a pre-Roman production. The characteristic feature is that only undecorated ones have been produced; besides the angular section that is either quadrangular (variant 1) or triangular (variant 2), or the B-shaped section (variant 3). o o o Type U1 are characterized by a quadrangular section coinciding with Heiko Wagner‘s type 18 (Armring mit vierkantigen Querschnitt) (Wagner 2006) Type U2 bangles feature a triangular section, coinciding Thea Haevernick‘s type 2 (Haevernick 1960) and Heiko Wagner‘s type 19 (Römische Armringe mit dreieckigem Querschnitt) (Wagner 2006) Type U3 comprises the bracelets with two lengthwise ribbings and therefore comparable to the late La Tène shape described by Thea Haevernick as type 7d (Haevernick 1960). Because not many examples have been retrieved so far and the close link with the late La Tène bangles it is not excluded that these bracelets are the earliest glass bangle types of the Roman imperial period. At least these seem to be artefacts responsive towards the traditional wishes of the local consumer of Celtic origin to fulfil an established custom that survived within the newly created Gallo-Roman society. 3.3.1.3 Summary 90 About 1 on 10 pieces of the 1453 black glass arm rings recorded could not be assigned to a specific type due to their fragmentary preservation or because the publication contained insufficient information to define the type (135 ex.). Of the defined 1318 ex. about 1/3 (426 ex.) are from type A and nearly half of the totality (613 ex.) are from type D. Type B and C bangles, respectively 127 ex. and 106 ex. contribute to a somewhat equal share of nearly 10 % (Figure 55). From Table 31 and the pie charts in Figure 55 can be deduced that Type D1 bangles are predominant and that other types are rarely attested in black glass. Table 31: List with amounts of Roman glass bracelets per type and colour (n = quantity) Type % Type % Type n n n % A1 A2 A2/3 A3 A4 A5 A6 134 89 33 86 32 25 27 31,5 20,9 7,7 20,2 7,5 5,9 6,3 B1 B2 B3 B4 B5 indefinite 64 40 5 5 10 3 50,4 31,5 3,9 3,9 7,9 2,4 C1 C2 C3 C4 indefinite 37 28 38 2 1 34,9 26,4 35,8 1,9 0,9 subtotal 426 100 subtotal 127 100 subtotal 106 99,9 Type D1 D2 D3 D4 D5 D6 D7 D8 subtotal n % 427 119 9 16 20 12 7 3 613 69,7 19,4 1,5 2,6 3,3 2,0 1,1 0,5 100,1 89 With only one dot the fragment was too small to identify the interval between dots. Only 34 non-black appearing ‗continental‘ bracelets in opaque and translucent glass have been recorded. In order to obtain a more correct picture of the non-black appearing glass bracelets produced in the Roman Empire a more thorough investigation is necessary to verify possible regional differentiation or even peculiarities, such as the mainly polychrome Romano-British glass bangles using various clear translucent and opaque colours. 90 91 The Roman glass bangles show a wide range of diameter sizes, but when reducing it into size categories two distinct sizes occur: 1) the small bangles with a diameter below 60 mm and 2) the large bangles with a diameter above 60 mm up to 95 mm. indefinite other 3% 9% A 30% D 42% C 7% C 8% B 9% B5 8% Type A A6 6% A5 6% A4 8% B2 32% A2 21% B 10% Type B indefinite 2% B4 B3 4% 4% A1 31% A3 20% A 34% D 48% B1 50% A2/3 8% C4 2% indefinite 1% C1 35% C3 36% C2 26% D3 1% D6 D7 D5 2% 1% D8 D4 3% 1% 3% D2 19% D1 70% Type C Type D Figure 55: Pie-charts demonstrating a distribution between the different types and between the different variants per type 92 3.3.2 Finger rings 3.3.2.1 Preceding researchThe first studies on Roman finger rings are sometimes not more than systematic compilations of the then-known material from a specific region, museum collection or specific material (Henkel 1913; Hagen 1937; Guiraud 1989; Popović 1992). However, none of these publications focussed on glass finger rings. It is evident that this is due to their less frequent occurrence. Hence, the first classifications of finger rings can barely be called useful typological taxonomies for those in glass, seeing that attention has been paid on the whole to the finger rings in metal – precious (gold and silver) and non-precious (bronze and iron). Firstly, in 1913 Friederich Henkel classified 60 finger rings in glass from the Rhine region into five large classes, and recognized three different sub-categories within class 2 (Henkel 1913, 154-158) (Table 32): 1) 2) 3) 4) 5) Schlichte Reife; Mit einer die Platte andeutenden Schwellung; a. Mit glatter Außenseite; (= plain) b. Mit schräg oder quer gestellten Rillen; c. Mit Buckeln an der Außenseite; Der Reif ist zur Bildung einer Platte abgeflacht; Die Platte ist als Scheibe an dem einen Ende eines Stäbchens gebildet und das Gegenende unter der Platte angeschmolzen; Die Platte ist von einer besonders gebildeten Auflage überhöht. Table 32: List of different classes of glass finger rings per colour with % in black glass (based on the catalogue of Henkel 1913, 154-158) black blue blue-green green brown white TOTAL % Class 1 2 2 2 2 2 10 20 Class 2 20 2 1 2 1 26 77 Class 2a 12 2 2 16 75 Class 2b 3 1 1 5 60 Class 2c 5 5 100 Class 3 4 1 5 80 Class 4 3 1 1 1 6 1 13 23 Class 5 1 1 3 1 6 0 TOTAL 29 7 4 6 12 2 60 48 Henkel‘s class 1 includes all seamless plain rings with a regular O-shaped or D-shaped section. Class 2 is more complex, embracing various sub-categories. With a similar section as those of class 1, the contrast is their irregular shape with a bulge featuring a seam. Variant a is ordinarily plain, variant b has oblique or lateral incisions, and variant c is characterised by one or more rows of protuberances. The third class comprises two variants of plain rings with a bezel: flattened (a) or pressed in (b). Class 4 consists of a thick circular to elliptical bezel applied on an open or roughly closed ring made from a narrow rod. Henkel‘s last class is somewhat the trash bin and encloses all kind of variants not fitting in his classification system. One variant is similar to class 3a but has a gem in a contrasting (opaque) colour set in the flattened bezel (Henkel 1913, 157-158, nos. 1751-1754). Another ring (no. 1755), from a very thick rod in colourless to brownish glass with an opaque yellow spiralling trail all around, is of a type that is repeatedly registered in the Mediterranean, e.g., Cyprus and Italy. This variant never occurred in 91 black glass. From the 60 pieces recorded by Friederich Henkel in 1913, 29 ex. or 48% are made in black92 appearing glass. 2 When attempting to use existing typologies on Roman finger rings (Henig 1978 , Guiraud 1989; Popović 1992) to build up one useful to classify those produced in (black) glass, we faced some difficulties. The finger ring types from Hélène Guiraud‘s typology does not incorporate all the known types we have recognized in (black) glass (Figure 56): Guiraud 1989 types 1; 2; 3; 4; 8; 9 (Table 33). Besides the various types in black glass not incorporated in Guiraud‘s typology, some specific types correspond to various black glass types and subtypese.g. Guiraud type 4f = all type C black glass finger rings. 91 The sixtieth glass ring in Henkel‘s catalogue should be excluded since it is a golden ring in which an eye-shaped gem is set. The piece has been counted as the white piece in class 5 of table 31 but does not interfere with the percentage calculation of the black glass examples. 92 The other 33 pieces incorporate 3 ex. in dark blue, 1 ex. in blue and 3 ex. in pale blue glass; 3 ex. in blue-green and 2 ex. in yellowish green glass; 1 ex. in green, 5 ex. in pale green and 2 ex. in greyish green glass; 4 ex. in brown, 2 ex. in pale brown, 3 ex. in yellowish brown and 1 ex. in so-called purple-brown; 2 ex. in so-called white glass. All these various colour specifications can be grouped into larger hue categories: blue (7 ex.), blue-green (6 ex.), green (7 ex.), brown (11 ex.), colourless (2 ex.) 93 a TYPE 9 b Figure 56: Typology proposed by Helène Guiraud based on the material from France (taken from Guiraud 1989, figs. 9, 11; 21; 26; 32; 37; 41; 45; 47) Table 33: Concordance table of the established typologies black glass typology Guiraud 1989 type A1:1 type 8 type A2:1 type 2g type A2:3 type 2b type A2:4 type 2b type A3:1 type 4d type A3:2 type 3c type A4:1 type 2f;3e type B2:1 type 2i type B6 type 9 type C1:1 type 4f type C2:1 type 4f type C3:1 type 4f 94 3.3.2.2 Proposed typology The taxonomy is a provisional classification that will need adjustments when including the glass finger rings in other hues than black. The challenge is to make a classification system that fits all Roman finger rings in glass, but the key test is to verify whether a taxonomy of the black glass finger rings can generate valuable information on the chronology and distribution of its production and consumption. The proposed typology is based on 204 finger rings in black glass mainly assembled from publications. Besides an exhaustive research on the finger rings in black glass, the classification system needs verification from the finger rings in all other glass hues. This work in progress thus cannot be considered exhaustive. The classification criteria are identical to those presented in the preceding part on the bracelets. Based on the technical aspects, we first have taken into account the section and secondly the type of decoration. We discerned three main form types: 1) type A = with a thin circular or semicircular section; 2) type B = with a broad semi-circular or flattened D-shaped section; 3) type C = with an elliptical to round section. The finger rings from type C demonstrate a distinct difference from the first two types, which are fairly similar variations in forms, only differing in size. Both groups appear to be characteristic for a specific period, i.e. types A and B for Period III and type C for Period IV [see Chapter 4]. Figure 57: Proposed typology of Roman black glass finger rings (drawings by the author) We catalogued 203 finger rings in black glass. 26 ex. could not be assigned to a specific type due to its fragmentary preservation or for lack of information from publications. Of the 176 ex. remaining about 2/3 belong to type A. Type B and C finger rings are much less frequent. When looking per type it is clear that type A1 is predominant and that some types are nearly or totally absent in black glass (Figure 58, Table 34). 95 Table 34: List of finger rings in black glass (n = quantity) type n % type n % type n % A1 A2 A3 A4 A5 indefinite 66 20 14 2 3 0 37,5 11,4 8,0 1,1 1,7 0 13 0 5 11 4 4 0 7,4 0,0 2,8 6,3 2,3 2,3 0 C1 C2 C3 indefinite 6 12 3 13 3,4 6,8 1,7 7,4 subtotal 105 59,7 B1 B2 B3 B4 B5 B6 indefinite subtotal 37 21,1 subtotal 34 19,3 TOTAL 176 black glass finger rings 66 20 14 A1 A2 A3 13 2 3 A4 A5 11 5 0 B1 B2 B3 B4 4 4 6 B5 B6 C1 10 3 C2 C3 black glass finger rings C3 C2 2% 7% C1 4% B6 B5 2% 2% B4 B3 6% 3% B2 0% C 7% A1 38% A3 A2 B1 8% 11% 7% A5 A4 2% 1% Figure 58: Column-chart and pie-chart correspondingly showing the ratios in terms of quantity (top) and percentage (bottom) of the different types of black glass finger rings 3.3.2.2.1 Type A: Closed circular ring with narrow D-shaped or O-shaped section Table 35: List of type A finger rings in black glass A1 A2 n % n % n variant 1 variant 2 variant 3 variant 4 variant 5 variant 6 variant 7 TOTAL 40 8 2 5 1 2 1 59 67,8 13,6 3,4 8,5 1,7 3,4 1,7 100 14 2 2 1 19 73,7 10,5 10,5 5,3 100 7 6 1 14 A3 % 50 42,9 7,1 100 n 1 1 2 A4 % 50 50 100 n 2 2 A5 % 0 100 100 96 type A finger rings 2% 2% A1 15% A2 A3 20% 61% A4 A5 Figure 59: Pie-chart displaying the ratios of the various type A finger rings  Type A1: plain annular This type of finger ring encompasses various subtypes but characteristic to all is the perfect circular shape and the most widespread shape (Figure 59; Table 35). The different variants recognized for this type of ring comprise: (A1:1) undecorated; (A1:2) lateral/transverse incised pattern; (A1:3) oblique incised pattern; (A1:4) applied glass trail in zigzag or wavy; (A1:5) mottled decoration with applied specks in various colours; (A1:6) with an applied glass gemstone (A1:7) with a relief decoration type A:1 finger rings 2% 3% 2% A1:1 3% 8% A1:2 A1:3 A1:4 14% 68% A1:5 A1:6 A1:7 Figure 60: Pie-chart displaying the ratios of the various subtypes of the simple annular finger ring The majority of the 59 type A1 entries in black glass are plain (variant 1) (Figure 60; Table 35). We have to acknowledge that it is uncertain whether some pieces are part of a fully plain finger ring due to their degree of fragmentation. It is therefore possible that they could be fragments of finger rings with a discoid bezel (type C1). Most finger rings of this type in black glass show a D-shaped section and a seam which is sometimes only recognizable from a bulge due to the merging of the overlapping rod ends. o Type A1:2: annular with lateral/transverse incised pattern The number of incisions is not fixed as it is – just like for the quality of the surface finishing – mainly affected by the ability and concentration of the glassworker, not forgetting that the glass composition and the furnace conditions are also important in relation to the workability of the glass. Nevertheless, groups can be discerned showing the deliberate idea to produce finger rings with very many, some or just a few incisions as well as pinching the incisions only superficially or very deeply. 97 o Type A1:3: annular with oblique incised pattern This finger ring variant includes two pieces from Augst/Kaiseraugst (CH) having a great number of deep obliquely incisions close to each other in a more or less twisted design (Riha 1990, nos. 285 and 2997) [cat.nos.3648; 3404]. A fragment in the Archaeological Museum at Nivelles and reportedly from Liberchies (BE) has the variability in incisions which is most probably due to the disparity in malleability when incising the hot glass. We then can assume incisions from hot to cold condition: a so-called twisted effect; deeply incised; superficially incised. The inner diameter is rather small, between 15-18 mm. The height ranges at 4-5 mm and the width is slightly larger. o Type A1:4: with applied zigzag This subtype of finger ring is characterised by an applied zigzag on the top surface of the ring except for the bezel. An example with a single zigzag in opaque red glass comes from Trier (DE) (Loeschke 1925, fig.2:5; Wagner 2006, no. R88) [cat.no.2049]. Other examples have been excavated at the late Roman site of the archbishop‘s palace at Sens (FR) (Arveiller-Dulong 1994) [cat.no.1648]. o Type A1:5: with applied specks in various colours Only one such finger ring with mottled decoration has been catalogued so far but the type must have been more widespread considering the regular occurrence of similarly decorated arm rings and the so-called ‗crumb‘ beads and pendants. The piece at the National Museum of Damascus (SY) [cat.no.3862] is of unknown provenance (Zouhdi 1974, 94-96, a). o Type A1:6: with applied gemstone This variant of type A1 finger rings features a small pad-like glass gemstone. It is not excluded that the exemples with an elliptical scar on the apex of the ring are due to the broken off gem. o Type A1:7: with a relief decoration Only one piece is hitherto catalogued with a relief decoration all around the ring on the crest of the top surface. The piece of unknown provenance, now at the Musée Rolin in Autun (Vitrum 1990, 105-106, no.76), has a beaded decoration on the crest, 17 droplets in total, and a tapering diameter. Probably this is the result of the rodformed production technique.  Type A2: with flattened surface as bezel Three variants within this type of finger rings can be recognized (Figure 57): (A2:1) with plain bezel and plain ring; (A2:2) with plain bezel and incised ring; (A2:3) with a mounted small monochrome gemstone in stone or in glass; (A2:4) with a mounted bichrome glass gemstone by superposing two glass layers in contrasting colours. The finger rings from Medizinische Klinik, Bonn (DE) all have a plain bezel. The finger ring from tomb 162 of the southwestern cemetery of Tongeren (BE) has a simple, small stone set in (Vanvinckenroye 1984, II, pl. 91, no. 162:4) [cat.no.551]. Very few finger rings of variant b were produced in black glass. The majority of this subtype is made from a translucent or transparent deep brown, yellowish brown or colourless glass with a gemstone in opaque turquoise, white or yellow glass, but mainly combining two contrasting colours (Ravagnan 1994, nos. 510). Variant (b) is a well-known type in metal with a gemstone in glass or a semi-precious stone set in centrally on the flattened bezel (Riha 1990, Typ 1 Variante 5).  Type A3: with multiple roundels on ring bezel These finger rings mainly have three roundels of the same diameter (A3:1) or one larger central roundel (A3:2), but a variant with four applied roundels on the flattened ring bezel (A3:3) as for instance the piece from Mont Berny at Saint-Étienne-Roilaye (FR) (unpublished) [cat.no.1629]. An intact piece of variant (A3:1) has been retrieved from the Insula 42-treasure in Augst (CH) (Riha 1990, 118; 192, no. 2996, Taf. 79). The finger ring has an inner diameter of 16,5 mm and a rather irregular O-shaped section showing a clear seam at the inner side [see Chapter 2]. The three roundels are here of similar size, lying close beside each other. The roundels of the Augst 98 piece are pinched out from the ring, but they equally can be manufactured separately and affixed on the ring. The latter type is typical for type B3 finger rings with a wide flattened D-shaped section.  Type A4: with three circular roundels on ring bezel Such finger rings have two sloping sides creating a pointed bezel whereon a glass gem can be applied. One variant has a plain ring and the other has an incised decoration on the sloping parts (Figure 57).  Type A5: with B-shaped section These finger rings with a narrow double D section have been recognized in three different variants (Figure 57): (A5:1) plain; (A5:2) with an applied gemstone; (A5:3) with a flattened bezel. The last variant is very similar to the rd th multiple finger rings in jet and gold which are characteristic for the 3 and early 4 century AD. 3.3.2.2.2 Type B: closed circular ring with wider and flat D-shaped section Six subcategories have been discerned of the 37 finger rings with wide, flattened D-shaped section catalogued, with the types B1 and B4 as the most common shapes (Figure 61; Table 36). Table 36: List of type B finger rings in black glass B1 B2 n % n % n variant 1 variant 2 variant 3 variant 4 variant 5 variant 6 TOTAL 2 5 4 2 0 0 13 15,4 38,5 30,8 15,4 0 0 100,1 0 0 0 3 2 5 0 B3 % n 60 40 100 B4 % 11 11 100 100 n 2 1 1 4 B5 % 50 25 25 100 n 4 4 B6 % 100 100 type B finger rings 12% B1 6% 37% B2 B3 B4 B5 33% 12% 0% B6 Figure 61: Pie-chart displaying the ratios of the various type B finger rings  Type B1: plain annular o Type B1:1: undecorated This category of finger rings differs only from type A1 by its much wider section. However, the height is similar to that of type A1 examples, making the plain section look like a flattened D with rounded edges. An example is known from Sézanne (FR) (Cabart et al. 1994) [cat.no.1673]. o Type B1:2: with vertical/lateral incised pattern The relatively small inner diameter ranging around 16-18 mm is due to its specific use as finger ring. Seeing that adult men have a ring finger with a diameter 20 mm these narrow finger rings were most likely worn by girls and women with small hands. The proportion between the width and the height appears correlated to roughly 2:1. The finger rings with vertical incised pattern are rather small and very flat compared to those with oblique and 99 crosswise incised pattern. The so-called piece from Mainz (DE) (Henkel 1913, no.1723) and one of the finger rings from Dalheim (Henkel 1913, no.1724) have respectively width of 4,5-5,5 mm and 4,5-7 mm, and a height of 2,5-3 mm and 2,5-3,5 mm. The number of incisions is not fixed, as it is mainly affected by the ability of the glassworker and his concentration, not forgetting that the glass composition and the furnace conditions are also important in relation to the workability of the glass as well as the size of the artefact. Nevertheless, groups can be discerned showing the intent to produce finger rings with very many incisions, some or just a few as well as pinching the incisions only superficially or very deeply. The so-called Mainz piece has a great number of superficial lateral incisions, whereas the Dalheim pieces have a deeply incised decoration, with two lateral and eight oblique incisions, respectively. o Type B1:3: with oblique incised pattern The piece from Niederbronn-les-Bains (FR) is also pinched obliquely, but in a way it appears twisted (Wagner 2006, no.R31). The variability in incisions is most probably due to the disparity in temperature of the hot glass when applying the incisions. We may then assume incisions applied in hot to cold condition: 1) a so-called twisted effect, 2) the deeply incised, and 3) the superficially incised. The pieces from Dalheim (LU) (Henkel 1913, no.1726) with 5 mm (w.) and 3,5-3,7 mm (h.); Rheingönheim (DE) (Wagner 2006, no.R34) with 4,5-6,5 mm (w.) and 3,2-3,9 mm (h.); Niederbronn-les-Bains (FR) with 7-7,8 (w.) and 3,7-4,3 (h.) are larger than the vertically incised finger rings, though the opening diameter is similar ranging between 16 and 18 mm. An additional variance of this type of finger ring is the occurrence of an applied glass gem in a contrasting colour, for the most part depicting the Medusa. An intact piece in the Israel Museum, Jerusalem has a gem in opaque white glass with the winged head of the Gorgoneion (Spaer 2001, 210, no. 489). The widening of the finger ring under the gem is caused by pressing the gem when affixing it on the ring. A similar but very fragmented piece comes from Augst (CH) (Riha 1990, no. 65) but is also known in onyx (Riha 1990, no. 64). o Type B1:4: with incised lozenge pattern These finger rings have a crosswise incised pattern resulting in a lozenge pattern. This variant is very rare and only reported from Mainz (DE) (Wagner 2006, no.R21) [cat.no.1944] and Cirencester (UK) (Charlesworth 1982, no.312, fig.69) [cat.no.4005]. With a width of 6,5-7,8 mm and a height of 3,8-4,3 mm the Mainz piece demonstrates that the wider the finger rings, the higher. This premise also applies to the abovementioned types B1:2 and B1:3. Apparantly the glass worker deliberately took a specific glass mass from the batch according to the type of finger ring he was producing. o Type B1:5: with applied zigzag No piece of this type in black glass has been recorded so far. o Type B1:6: with applied specks in various colours We have not catalogued black glass finger rings with wide, flattened D-shaped section and decorated with colourful specks because this type of decoration is generally considered as early medieval (early Byzantine/early Islamic/Merovingian/…). Nonetheless we estimate it as fitting the proposed classification as besides the type A1:5 finger rings we have noticed a similar design on bracelets as well as on annular and globular beads and poppyth th head pendants from the Period IV onwards (end 4 early 5 century AD) and thus to be regarded as late Roman commodities [see Chapter 4].  Type B2: with flattened bezel This type of finger ring is similar to type A2:1 except for the much broader section. An intact piece is in the archaeological museum of Zara in Croatia and comes from one of the cemeteries in the Zara area – Zara, Nona or Asseria (Fadic 1998, no. 232). The finger ring has an inner diameter of 22-26 mm. The oval bezel is concave, as if it were made to be filled up with something, as can be seen on the even more massive glass finger rings 100 from Hellenistic times (Stern 2001, 368, no. 202). A similar piece in the Museo Vetraria, Murano but coming from 93 the same Zara area is made of an opaque dark red glass (Ravagnan 1994, no. 2).  Type B3: with three circular roundels on ring bezel These finger rings have three elliptical or circular roundels of the same diameter (a) or a larger central roundel (b). Two major differences discriminate type B3 finger rings from those of type A3: 1) a flattened D-shaped section instead of an O-shaped one and 2) the roundels are not small and pinched out from the glass mass of the ring but large and manufactured separately and affixed later on the ring. Two examples have been retrieved from the socalled Zara necropoli in Croatia and are now in the Museo Vetraria, Murano (Ravagnan 1994, nos. 17-18) [cat.nos.769-770]. The three roundels on both pieces are equally sized, but contrary to those of type A3, the discs are laying apart from each other. The space between the three roundels is most probably due to a technical aspect. The inner diameters of the Zara rings are 23 mm and 25 mm respectively.  Type B4: with impressed protuberances This type of finger ring corresponds with Henkel‘s class 2c (Henkel 1913, 155-156). Characteristic to this type of finger ring are the more or less parallel lengthwise rows of round bulges all over the surface. Variations occur in the number of rows: mainly, there are three rows due to the dimensions of this type of artefact, but two like a piece from Augst (CH) (Riha 1990, no.287) and four rows also occur regularly. Contrasting to most finger rings with only a slightly uneven width and height, the Wiesbaden piece (Henkel 1913, 156, no. 1730) [cat.no.2101] has a clear narrower part of 7 mm compared to the 11 mm wide other half. The result is that the narrow part only has two rows, while the wider part bears four rows. Normally bottle green to brownish olive green glass is used to produce black appearing finger rings but the piece from Matagne-la-Grande (BE) is said to be of a dark blue glass and that from Mazières-en-Mauges (FR) (Simon 2000, 157-158, no. 89:2) [cat.no. 1265] made in dark purple glass. Most type B4 finger rings match the general opening diameter of 16-18 mm but some are larger with an opening ranging between 18-20 mm and 20-22 mm like two examples from Saint-Dié-des-Vosges (FR) (Wagner 2006, no.R70-71) [cat.nos.1377-1393]. The finger rings with parallel rows of round bulges are one of the popular categories in black glass with 11 ex. recorded. At present, only three variations have been recorded. The resemblance with bronze finger rings with protuberances is striking. The absence in other metals and in jet as well as in other glass hues speaks in favour of an imitation of bronze rings in black glass. This finger ring type must have been produced in one or more glass workshops or jeweller‘s shops somewhere in the north-western provinces because all recorded material is retrieved from sites within Gallia Belgica, Germania Inferior and Germania Superior with a concentration in Germania Superior [see Chapter 6].  Type B5: with complex impressed design Within this group, we divide the finger rings according to the stamp/mould used to create a specific decoration pattern. The three recorded variants show much resemblance in design, but they remain unique models. o Type B5:1: radiant sun The sole intact piece known until now came from the subaquatic site of Trou de Han at Han-sur-Lesse (BE) [cat.no.304] (Figure 62 left and middle). Its inner diameter is 16,5 mm, its height 3,0-3,5 mm and width 6,9 mm. The finger ring has a quite complex design, including three lengthwise ribbings cut into four metopes by four pairs of lateral ribbings. Three of these metopes show the same dimensions, and in the central one there is an image of a radiant sun with 9 rays. A similar design is attested on a fragmented piece in black appearing dark blue glass that was retrieved during the excavations at Ermin Street in Wanborough (Monk 2001, 171, fig. 67, no.144) [cat.no.A4619] (Figure 62 right). 93 We were not able to check whether the glass was produced from a deep red glass or of a black glass with an opaque brownish-red oxidation layer as we discussed for some bracelet fragments from the glass workshop Les Houis nearby Sainte-Menehould. 101 Figure 62: (left + middle) Finger ring from Trou de Han (drawing by the author, courtesy of Musée du Monde Souterrain, Hansur-Lesse); (right) finger ring fragment from Wanborough (drawing taken from Monk 2001, fig. 67) o Type B5:2: floral motif The unique example within this subcategory, housed in the Großherzogliche Kabinetts-Museum, Darmstadt (DE), is of unknown provenance [cat.no.2083]. The ring has an inner diameter of 17-18 mm (as most other finger rings), a height of 3,5 mm and a width of 8 mm. Similar to the Trou de Han piece are the three lengthwise ribbings and the lateral ribbings creating metopes in which a figurative motif is stamped. But instead of four pairs of lateral ribbings there are two groups of four loose ribbings. Whereas the Trou de Han piece shows an underside with just the lengthwise ribbings covering roughly half of the surface, this is only about one-third of the surface on the Darmstadt piece. The three sections have comparative dimensions dividing the front side of the ring in three equal metopes. What is depicted appears unclear, but Henkel suggests a floral motif (Henkel 1913, 156, no. 1732). Despite the differences with the ‗radiant sun‘-pieces from Trou de Han and Wanborough we assume that all three finger rings were made by the same jeweller-glassworker. The very distinct related features of both pieces compared to the very dissimilar other types of finger rings in glass are to us a convincing argument. o Type B5:3: non-figurative design The piece from the Shepton Mallet 1990 excavations at Fosse Lane (UK) show a complex geometric motif with two combined lozenges contours with a lozenge-shaped cut-out in the centre. Between the lozenges are on either side two V-shaped motifs with a central triangular cut-out (Price, Cottam 2001, 204, no.26) [cat.no.4053].  Type B6: closed, polygonal rings Characteristic to this shape is the polygonal shape of the outer surface while the inner surface preserves a circular shape. This polygonal shape of glass finger ringse.g., hexagonal, octagonalis achieved by flattening out the external surface, thus causing a number of angles. No particular tool is needed to achieve this result, but intentionally manufactured stamps are not excluded, considering the waffled decoration on, for instance, the polygonal black glass bangle (Type D4) from Kaiseraugst (CH) [cat.no.3677] and the finger rings of subtype B6. No octagonal finger ring in black glass has yet been recorded, but seeing they exist in gold, silver and bronze as well as in jet, we may assume they have been produced in (black) glass too. The angles can be sharp or rounded, whereas the straight sides can be flat, hollowed or faceted. The pieces from Sinsin (BE) (Warmenbol 1984, 8, fig.6) [cat.no.516] (Figure 63) and Augst (CH) (Riha 1990, no. 286) [cat.no.3326] are plain and have six sides with rounded edges and thicker angles. This is due to the use of a tool to flatten the sides into a polygonal shape. Polygonal finger rings in (black) glass are a very infrequent shape and none holds a kind of decoration that corresponds with the much more widespread equivalent in bronze and silver. Our type B6 corresponds to Guiraud type 9 and is said to be very uncommon in Gaulabout 3% of all finger rings (Guiraud 1989, 196-197). In Popović‘s list of Balkan material, the type is totally absent (Popović 1992, 13). Figure 63: Plain hexagonal finger ring in dark green glass appearing black from Trou del Leuve at Sinsin (BE) (drawing taken from Warmenbol 1984, 8, fig.6) 102 3.3.2.2.3 Type C: with a discoid bezel This class of finger rings provides some difficulties in categorizing the entire group with a discoid bezel. Firstly, three categories on the basis of two technical aspects can be discriminated (Figure 64). The first technique consists of pinching one rod end into a discoid bezel and attaching it to the other end (1). The second technique consists of affixing a separately produced discoid bezel onto a closed (2) or an open ring (3). In theory, both categories can be subdivided in at least three times three subcategories according to the specificities of the bezel: 1) plain bezels; 2) those with stamped decoration and 3) those with incised decoration, and these of the ring: 1) plain; 2) with applied zigzag; 3) with twisted decoration (Figure 57). In reality, we recorded only some combinations created in black glass. Nevertheless, we have to take all parameters into account as other types not occurring in black glass may well have been produced in other glass hues. Within the catalogue we used the technological aspect creating three categories of type C-finger rings and on the basis of the decoration of the ring each can be divided into three subcategories. The bezel is not considered here. finger rings with discoid bezel (type C) discoid bezel pinched from one end plain ring applied trails in zigzag or wavy pattern twisted ring plain discoid bezel plain discoid bezel plain discoid bezel stamped discoid bezel stamped discoid bezel stamped discoid bezel incised discoid bezel incised discoid bezel incised discoid bezel discoid bezel affixed on closed ring discoid bezel affixed on open ring … … Figure 64: Theoretical classification of Type C finger rings Table 37: List of type C finger rings in black glass decorated with protuberances (dimensions in mm) C1 C2 C3 n % n % n % variant 1 6 100 9 66,7 3 100 variant 2 0 0 0 0 0 0 variant 3 0 0 3 33,3 0 0 TOTAL 6 100 12 100 3 100 103 type C finger rings 11% 33% C1 C2 56% C3 Figure 65: Pie-chart displaying the ratios of the various type C finger rings  Type C1: plain ring The largest category within the type C finger rings are those with a plain ring, whether the bezel is pinched out from one end or produced separately and applied afterwards on an open or closed ring (Figure 65; Table 37). Besides black glass, this type of finger ring is particularly produced in a translucent or transparent yellow to orange glass. The resemblance to the medallion pendants and pressed beads is strong. This type of finger ring is also known from examples in metal and described by Emilie Riha as ‗Typ 8 – Variante 2‘ (Riha 1990, 35, nos. 134-135). The rings can have an inner diameter from 17 to 20 mm. The rod is about 3-4 mm wide and high and the diameter of the bezel ranges from 10-12 mm. o Variant 1: with undecorated gem Few finger rings with discoid bezels show no decoration. Because the gem has been pressed, a flat tool must have been used and not a stamp. o Variant 2: with stamped gem Some finger rings roundels on the ring bezel are decorated with an impressed decoration – a symbol, a text or a figurative element. A finger ring with chrismon-monogram in a very dark blue glass comes from Matagne-laGrande, Belgium (Sas, Thoen 2002). Figurative elements can be diverse, but portraits are mainly represented. Single as well as double portraits occur, like, for instance, the finger ring from Saint-Dié-des-Vosges (FR) (Wagner 2006, no. R72) [cat.no.1378]. Only one example in black glass holding an inscription has been recorded. The inscription on the finger ring from Petronell – Carnuntum (AT) [cat.no.143] reads "UTERE FELIX" (translated: wear this in prosperity). o Variant 3: with incised gem The difference from both previous types is that here the gem has an incised decoration. This is a rather unusual ornamental style and is, so far, only known in black glass from a piece from Zara (HR) and now in the Museo Vetraria in Murano (Ravagnan 1994, no. 11) [cat.no.768].  Type C2: with applied zigzag An intact piece comes from the amphitheatre in Metz (FR), showing an applied wavy trail decoration in opaque white glass on the ring up to the discoid bezel [cat.no.1273] (pers. comm. Hubert Cabart). The bezel is formed by pinching one end (variant a) and has an impressed decoration. The depiction is unrecognizable, but the occurrence of two analogous, amorphous embossments side by side most likely refers to the popular double portrait. This motive was very fashionable on all kinds of glass jewellery during the late Roman period [see 3.3.3 beads].  Type C3: with twisted ring Finger rings of this type are characterised by a separately prepared bezel applied on a twisted ring. The fine or wide twists can be S-twined as well as Z-twined. The black glass piece from Augst (CH) is built up from an S- 104 94 twined ring in fine twists (Riha 1990, no. 283). The bezel is lost, excluding the determining of the variant, but most probably the missing part of the finger ring was a separate discoid bezel. Another example from Augst in a colourless glass with fine opaque yellow glass trail twisted clockwise and with thicker section has lost its bezel too, but is clearly an open ring that must have held an affixed bezel (Riha 1990, no. 284, Taf. 14). 3.3.3 Beads The study of glass beads remained until recently limited to either the large group of beads comprising a wide range of materials or to the entire group of glass material and consequently never received proper support. The few prior outlines executed on Roman glass beads resulted however in the necessity to create valuable and 95 effective typologies (Haevernick 1983; Guido 1978; Spaer 2001; Mandruzatto 2008). But most important for future research on the Roman glass beads is to apply the established typologies to improve our knowledge on the glass bead production and consumption in order to better understand the use and function of this commodity type within the Roman material culture (Swift 2000). To come to new readings of the heap of bead production and consumption in the Roman imperial period it is sometimes essential to adjust parameters instead of continuously demanding to develop the elaboration of the established parameters. Most commonly in use is the classification established by Margaret Guido (1978, 91-102), even though it has a whole range of shapes missing that appear frequently on the continent, as she took into account only the beads from the British Isles. Nevertheless, the glass beads designated in Guido‘s typology gives a clear impression of the Roman bead production characteristic for the first three centuries AD. Figure 66: Classification of the Roman glass beads from the British Isles (after Margaret Guido 1978, fig.37); Encircled with full red line the types known in black glass; encircled with dotted red line the types in black glass recorded on the continent From the 22 types she distinguished (Figure 66) only three types appear in black glass in Britain: the small segmented beads (types 1-2); cylindrical beads (type 5); small biconical beads (types 12-13). Four other types are also known in black glass on the continent: the elongated cylindrical beads (type 4); the barrel-shaped beads (type 16); the discoid shaped flattened beads (type 17a) and the so-called ‗melon‘ beads (types 21-22). It is apparent that a regional Romano-British production can only be recognized when considering the continental distribution and consumption patterns. An accurate and consequent approach of the beads can therefore result in 94 Although this piece is appearing black in Emilie Riha‘s publication, the colour of the glass is described as ‗grau-gelblich‘ (yellowish grey), probably swayed by the iridescence. 95 Essential within the glass bead research are the studies on contemporaneous material from outside the Roman Empire (TempelmannMaczyńska 1985) and on early Medieval glass beads, such as the Anglo-Saxon beads (Guido 1999; Bruggmann 2004). 105 a better insight of the bead production and help to recognize glass workshops on a local, regional, inter-regional and eventual empire-wide level. Margaret Guido more or less limited her description of the Roman glass beads to the monochrome beads (Guido 1978). As a result Guido‘s typology represents for the most part those produced in a monochrome translucent emerald green, ultramarine blue or colourless glass, the latter alternatively gold-sandwiched. Beads produced in black glass remain rather marginal for Britannia, resulting in Guido‘s study in a limited number of black glass shapes taken into consideration. Moreover, by excluding the bichrome and polychrome beads, characteristic for th th the (the second half of) the 4 century and 5 century AD [see Chapter 4], a large part of the Roman glass bead production got disconnected from the entire discussion. Remarkably these colourful beads do get integrated in her other comprehensive publication dealing with early medieval Anglo-Saxon glass beads (Guido 1999, 170-190). The annular, globular and barrel-shaped beads of this period appear in different sizes and are usually decorated with a single or double zigzag in a white, yellow, red or turquoise blue opaque glass trail. Similar beads have been th th retrieved in various late Roman contexts of the late 4 –early 5 century AD all over the Roman Empire such as Trier (DE) (Cosyns 2009), Augst/Kaiseraugst (CH) (Riha 1990, 77-93), Horbat Qastra (IL) (unpublished). The fact that these late Roman beads frequently appear in Anglo-Saxon, Merovingian, and Longobard tombs in the west and in so-called early Byzantine tombs in the Levant adds to the difficult chronology. Hence, we cannot simply th assign cultural significance to the political and military shifts occurring at the beginning of the 5 century AD [see Chapter 4]. Furthermore, the presence of polychrome black beads is noticed in late Roman contexts within the entire Mediterranean and specifically in the eastern Mediterranean, an area where Barbarian people such as Anglo-Saxons or Merovingians were unquestionably fully absent. Therefore, we believe we have to exclude these late Roman/early Byzantine glass beads as imports from the so-called barbaric areas. What‘s more, the negligible quantity of (polychrome) black glass beads within so-called Barbaricum beyond the Rhine and Danube (Tempelmann-Maczyńska 1985) seems rather to point to an inverse import coming from the Mediterranean area [see Chapter 6]. th th As a result Guido‘s overview on the Romano-British glass beads misses those in black glass dated 4 and 5 century AD. Obviously this is because they are nearly exclusively known from Continental sites. The most idiosyncratic ones are the so-called spacer beads with their characteristic double perforation. They were manufactured in various monochrome glass coloursi.e., brownish yellow, ultramarine blue and deep green appearing black. Most characteristic spacer beads are the ribbed ones in black-appearing glass (the so-called Trilobitenperlen), but also portrait-beads, theatre mask–beads as well as various other types have been produced in black glass. The exclusion of the double perforated spacer beads in Guido‘s work is most likely due to the striking absence of this type of glass bead in Britannia. All recorded spacer-beads so far are solely known in jet (Allason-Jones 1999) what reflects the success of the locally extracted jet and shale. Besides these monochrome black glass beads, there is a multitude of bi- or polychrome decorated black glass beads with one contrasting colour, or with two different glass colours combined. Rarely, three different colours have been applied. In most cases, the glass trails have been applied in an opaque white glass, but also a large number have been decorated with yellow opaque glass trails and/or dots. To a lesser extent, opaque turquoise blue and vivid red to brownish red glass has been used. Some of the Roman black glass beads have been incorporated in the typology on contemporaneous so-called ‗Barbaric‘ beads from Central Europe by Magdalena Tempelmann-Maczyńska (1985). Various beads were made in black glass. From the 387 types recognized by her (TM), about 65 are made from black glass, of which 19 can be regarded as being of Roman production:  [monochrome, plain] TM11 (small cylindrical); TM27 (micro cylindrical); TM40 (small annular); TM69 (small biconical);  [monochrome, incised] TM103d, 167 and 168 (melon);  [polychrome] TM198e (globular ‗crumb‘); TM276 (large annular with cross-cutting wavy trails); TM293b-c (large cylindrical with parallel wavy trails); TM300; TM301 (large annular); TM302 and 303 (large squat cylindrical); TM330-331 (elongated barrel-shaped); TM332 (juglet pendant); TM348 (large elongated cylindrical); TM380c (large annular). Considering the massive category and the huge variety of shapes, glass hues, decorative design and techniques applied it is hard not to contribute to the already existing confusion and come to useful quantifications. In order to keep an overview of the black glass beads produced during the Roman imperial period we set up a chart of the different types of glass beads known in black glass (Figure 67). We created two main classes on the basis of the applied technique: (a) the rod-formed beads made by wrapping hot glass around a metallic mandrel; (b) the folded and pressed spacer-beads with the very characteristic features of a double perforation and one flattened 106 side. A second level defines the various shapes. A third level considers the variety of decorative styles applied. Finally we take into account that each class was produced in different sizes: <15 mm (small); 15-25 mm (medium); >25 mm (large). The small beads encompass also the micro-beads with a diameter <8 mm. Beads Spacer-beads Rod-formed Annular beads undecorated Eight-shaped beads 'Trilobitenperlen‘ with wavy pattern plain with zigzag pattern vertically ribbed with crossing zigzag pattern with dots in between two interwoven glass trails with mottled pattern horizontally ribbed Theatre mask beads Lion head beads Globular beads Cylindrical beads 'Melon' beads large Frog beads Helmeted head beads Medusa head beads small Barrel-shaped beads Biconical beads Portrait beads See detailed subdivision figure 71 Discoid beads Figure 67: Chart based on the different types of black glass beads (by author) 107 3.3.3.1 Rod-formed beads Only few types of the wide-ranging variety of rod-formed beads have been produced in black glass. We defined six different classes of rod-formed beads in black glass (Guido 1999, 170, type 2). Like everyone else, we separated the so-called melon-beads but in fact they are a decorative subcategory derived from tooling annular or globular beads with an incised decoration occurring already from the early Iron Age. This separate approach has grown historically under influence of the ubiquitous blue to turquoise quartzceramic ‗melon‘ beads. The cylindrical, conical and biconical beads appeared newly in Roman times. The predominant differentiation of annular, globular and barrel-shaped bead categories is the ratio lengthdiameter: 1. 2. 3. length-diameter ratio ≤1:2 length-diameter ratio c. 1:1 length-diameter ratio 1,5:1 to 2:1 annular beads: globular beads: barrel-shaped beads: Concerning the cylindrical, biconical and discoid beads, the most important aspect is the shape: 4. cylindrical beads: a straight perforated side and a constant diameter from one end to the other; short cylindrical = when the length-diameter ratio is equal to globular beads and long cylindrical = when equal to barrel-shaped beads. 5. biconical beads 6. discoid beads : flattened round or elliptical we cannot speak of the diameter but talk of the thickness or height of a bead instead The diameter of the perforations remains in general relatively small, with a ratio versus the maximum external diameter of 1:5 but going up to 1:2 for a number of annular beads. We catalogued 799 beads of which 502 are rod-formed and 297 are folded and pressed spacer-beads. Besides the main annular, globular and barrel-shaped beads, as the most classical bead shapes, the cylindrical and in a lower extend also the biconical beads are the most frequent rod-formed black glass beads (Figure 68; Table 38). The discoid-shaped beads occur rarely in black glass and can be considered on the same level as the types incorporated within the miscellaneous group. The ‗melon‘ beads can be divided into 15 ex. annular and 19 ex. globular (15 ex. could not be assigned to a specific shape being short of information). Table 38: Overview of the rod-formed beads in black glass rod-formed beads n % annular beads globular beads barrel-shaped beads biconical beads cylindrical beads ‗melon‘ beads discoid-shaped beads miscellaneous undefined TOTAL discoidshaped beads 1% 112 68 39 22 68 49 5 11 128 502 22,3 13,5 7,8 4,4 13,5 9,8 1,0 2,2 25,5 100,0 rod-formed beads miscellaneous 3% ‗melon‘ beads 13% cylindrical beads 18% annular beads 30% globular beads 18% biconical beads 6% barrel-shaped beads 11% Figure 68: Pie-chart showing the ratios of the different bead classes in black glass 108  Annular beads Because all black glass annular beads are made with the rod-formed production technique the size and the decoration were looked at first to classify the 127 entries. The annular beads can be subdivided in four major classes: 1) plain; 2) ribbed or so-called ‗melon‘ beads [see below]; 3) applied glass trail; 4) applied specks or socalled ‗crumb‘ beads. These can occur in different sizes i.e. micro (2 ex or 1,6%); small (54 ex or 42,5%); 96 medium (32 ex or 25,2%); large (12 or 9,5%). Table 39 demonstrates most annular beads in black glass are plain but when decorated annular beads mainly have applied glass threads in a wavy or zigzag pattern, sometimes with a dot in between two interwoven glass trails. 97 We distinguish six a. b. c. d. e. f. varieties in decoration: undecorated (Guido 1999, type 2i) with wavy pattern (Guido 1999, type 2v) with zigzag pattern (Guido 1999, type 2vi) with crossing zigzag pattern (Guido 1999, type 2viii-ix) with dots in between two interwoven glass trails (so-called ‗eye bead‘) (Guido 1999, type 2viii) with mottled pattern (so-called ‗crumb bead‘) (Guido 1999, type 2x) Table 39: Overview of annular beads in black glass (n = quantity) variant a variant b variant c variant d variant e variant f ‘melon beads’ TOTAL n % micro small medium large indefinite 10 7,9 6 2 2 - 12 9,5 3 4 5 15 11,8 13 1 1 - 127 100 2 54 32 12 27  54 42,5 2 17 13 3 19 7 5,5 3 4 - 27 21,3 12 7 5 3 2 1,6 1 1 - Globular beads The globular beads in black glass occur in various decorative categories similar to the annular beads: a. b. c. d. e. f. undecorated (Guido 1999, type 2i) with wavy pattern with zigzag pattern (Guido 1999, type 2xv) with crossing zigzag pattern (Guido 1999, type 2viii-ix) with dots in between two interwoven glass trails (so-called ‗eye bead‘) (Guido 1999, type 2viii-ix) with mottled pattern (so-called ‗crumb bead‘) (Guido 1999, type 2xi) Margaret Guido (1999, 170) kept the undecorated annular and globular beads together within type 2i. At first sight there is no reason why these should not be identified as two separate shapes. Table 40 demonstrates that the 87 globular beads recorded are for the most part undecorated as the annular counterparts. A quite considerable part of the globular beads is composed of the so-called melon beads [see below]. Most globular beads are small sized but no useful conclusions can be proposed from such general classifications. Nevertheless there is a potential to discern chronological idiosyncrasies as is the case for the ‗melon beads‘. Besides the abovementioned six variants we catalogued two segmented globular beads, both from Nor‘Nour (UK) (Dudley 1967, nos. 95a-b) [cat.nos.4166-4167] and one faceted globular bead from Conimbriga (PT) (Alarcao 1976, 210; 213, no.298, pl. XLVI) [cat.no.3090]. It needs verification whether the glass hue of the applied glass trails relates directly to a specific pattern such as short, regular zigzags and large zigzags applied either in a smooth movement or a coarse motion and all possible combinations of the above. Colour combinations and pattern variations possibly comprise regional styles and chronological evolutions but in the same way the creation of existing differentiations were subjected to sociocultural practice and belief. 96 97 The size of 27 annular beads (or 21,3 %) remained unknown when cataloguing them. We excluded here the incised decoration technique as we discussed the annular ‗melon‘ beads with a separate class. 109 Table 40: Overview of globular beads in black glass (n = quantity) variant a variant b variant c variant d n % micro small medium large indefinite  39 44,8 1 22 5 3 8 2 2,3 1 1 - 14 16,1 7 1 2 4 0 0 - variant e variant f ‘melon beads’ TOTAL 1 0,9 1 - 8 9,2 4 3 1 20 23,0 6 6 8 - 87 100 1 41 15 14 13 Cylindrical beads Although different from the annular and globular beads the cylindrical (black) glass beads also occur in various decorative categories: a. b. c. d. e. f. undecorated festooned spiralling pattern (Konrad 1997 form 14a) with zigzag pattern with crossing zigzag pattern with dots in between two interwoven glass trails (so-called ‗eye bead‘) spiralling pattern We inventoried 66 cylindrical beads (Table 41). The cylindrical beads are present in Margaret Guido‘s overview of forms (1978, 96, types 4-5) which categorizes the material according to their size: the long cylindrical beads (type 4) and short cylindrical beads (type 5) (Figure 66:4-5). Most are undecorated (variant a), but a number of elongated cylindrical beads are decorated with applied spiralling trail in opaque white, yellow, blue or red glass – mainly festooned (variant b) – or zigzag in the centre flanked at the ends with a short spiral (variant c) and consequently resembling barrel-shaped beads [see below]. Others show a crossed zigzag pattern with eyes within the lozenges (variant e). There are no cylindrical beads inventoried with mottled specks or so-called ‗crumb‘ pattern. Table 41: Overview of cylindrical beads in black glass (n = quantity) variant a variant b variant c variant d variant e variant f TOTAL n % micro small medium large indefinite 4 6,1 0 0 0 4 0 7 10,6 0 1 0 1 5 66 100 0 14 5 25 18  37 56,1 0 10 4 10 13 13 19,7 0 0 0 13 0 4 6,1 0 3 0 1 0 1 1,5 0 0 1 0 0 ‗Melon‘ beads This is the most common and oldest name for ribbed beads with longitudinal grooves and lobes, and supposedly derived from their resemblance to a creased melon. But this type of bead (Guido 1978 type 21-22; Guido 1999 type 2iii; Riha Type 11.1.2) should better be incorporated under the annular or globular beads with incised/tooled decoration (Figure 66:21-22). Gustavus Eisen prefers to call these beads ‗lotus‘-shaped as it ―harmonizes better with the amuletic nature of the beads‖ while names as ―segmental beads‖; ―ribbed beads‖; ―pulley beads‖; ―nasturtium seed beads‖ all refer to the form of the beads, ―without any consideration of their nature‖ (Eisen 1930, 20) [see Chapter 7]. 98 Generally known in a pale blue to turquoise quartz-ceramic material , similar beads in glass regularly occur too on Roman sites, and this in two sizes: large (with a diameter above 25 mm) and small (with a diameter below 20 mm). Fifty entries deal about the so-called ‗melon beads‘. Already a good number have been discussed when considering the annular (15 ex.) and globular beads (20 ex.) (Tables 39-40). The shape of another 15 ex. remains 98 Mostly these type of beads are described as being made from frit, glass paste or faience, but we choose to use the English translation of the German ‗Kwartskeramik‘ (Busz, Gercke (eds) 1999). 110 uncertain, one of which is in fact a ‗crumb‘ bead mottled with colourful glass specks. The small annular ones all remain plain monochrome black beads, whereas those with an applied spiralling decoration with trails in white, yellow or red opaque glass are medium or large globular ones (Table 42). Both categories should be considered separately, not only for their differences in external properties, but particularly because the large bichrome beads feature early Roman beads of a probably late La Tène derivation during the Julio-Claudian period and thus characteristic for Period I. Conversely, the small-sized plain melon beads form part of the late Roman glass jewellery from only Period II onwards up to Period IV [see Chapter 4]. Variant a = plain annular Variant b = annular with spiralling decoration Variant c = plain globular Variant d = globular with spiralling decoration Table 42: Overview of melon beads in black glass (n = quantity) variant a variant b variant c variant d indefinite TOTAL n % micro small medium large indefinite 50 100 0 26 9 12 3  15 30 13 1 1 - 0 0 - 6 12 5 1 - 14 28 1 5 8 - 15 30 7 2 3 3 Barrel-shaped beads Characteristic to the barrel-shaped beads is the convex body tapering towards the ends with a length-diameter ratio of 2:1. Margeret Guido discerned two categories of barrel-shaped beads in Roman times: the short and the elongated: (Figure 66:16) (Guido 1978 type 16). The short barrel-shaped beads have a length that more or less equals the diameter of the bead. These squat beads can occur in very large dimensions, like the one in a brownish-red glass in the British Museum, London and presumably with Egyptian provenance, with a length of 51 mm and a maximum diameter of 31 mm (Cooney 1976, 33, no. 320). The length-diameter of elongated barrelshaped beads hrates to 5:1. Both classes have identical decorations with applied glass trails in a contrasting colour, generally opaque white, but opaque yellow, opaque red and opaque turquoise are also utilized. Based on the decorative pattern we recognized six variants (Table 43): a. b. c. d. e. f. undecorated with zigzag pattern with crossing zigzag pattern with dots in between two interwoven glass trails (so-called ‗eye bead‘) spiralling pattern with mottled pattern (so-called ‗crumb bead‘) The design of the decorated beads is characterised by spiralling coils at each end and a zigzag motive in the centre. The central design demonstrates two variants. Variant 1 consists of only one wide zigzag in the centre, and variant 2 is composed of a wide cross-cutting double zigzag. These classes match with types 330 and 331 of group XXII proposed by Magdalena Tempelmann-Maczyńska (1985). One third of the barrel-shaped beads is plain (variant a) and more than half is large sized (Table 43). All other decorative patterns are only represented by a small number of pieces. Table 43: Overview of barrel-shaped beads in black glass (n = quantity) variant a variant b variant c variant d variant e n % micro small medium large indefinite 15 34,9 0 2 4 7 2 3 7,0 0 0 0 3 0 5 11,6 0 0 1 4 0 5 11,6 0 2 1 1 1 2 4,6 0 0 0 2 0 variant f indefinite TOTAL 3 7,0 0 0 1 1 1 10 23,3 0 0 0 6 3 43 100 0 4 7 24 7 111  Biconical beads In general, biconical beads are small or very small (micro beads) below 2,5-3 mm diameter and length and mainly undecorated (Guido 1978, types 12-13) (Figure 66), but rare examples of decorated biconical beads with blurred trail decoration have been reported as the one now at the Corning Museum of Glass (Whitehouse 2001, 59, no.987) [cat.no.4464]. Very popular in ultramarine blue translucent glass, the biconical beads in opaqueappearing black glass only occur occasionally.  Discoid beads Discoid beads are monochrome flat-pressed circular beads (Guido 1978 type 17a variant) with a small perforation not in the centre of the disc but all through the flat side (Figure 66:17 left). Many were retrieved from the late Roman cemetery site of Castra, Israel (unpublished) [cat.nos.2391; 2396], but most pieces were unfortunately in such a bad conservation condition that it is unknown whether all of them were made from black glass or whether this is due to a thick blackish weathering crust. A necklace conserved in the archaeological museum of Ancona (IT) consists of a set of 20 perfectly well99 preserved discoid beads in a deeply brown glass appearing black on a golden chain (Figure 69) [cat.no.A4606]. Two pieces come from Augst/Kaiseraugst (CH) and are catalogued as ‗Linsenförmige Perle‘ (Riha 1990, nos. 1348-1349) [cat.nos.3344; 3653]. No. 1348 is described as being in ultramarine blue glass but is in reality made in a black-appearing olive-green glass. The discoid beads are all small sized with a diameter that ranges between 12-13 mm and their thickness between 3-5 mm. One bead from Arles [cat.no.1014] is more elongated giving it a more ellipsoidal shape (Foy 2010a, 494, no.1049). Figure 69: Roman necklace combining a golden chain with black glass discoid beads (photograph by author, by courtesy of the Archaeological Museum of Ancona)  Miscellaneous On rare occasions we recorded isolated shapes proving that the Roman glassworker/jeweller was very creative in producing all sorts of beads using all sorts of coloured glass and varying in shape, size, applied technique. We limit ourselves in summing up the various shapes in Table 44. Table 44: Overview of the miscellaneous shapes of beads rarely occurring in black glass (n = quantity) cat.no. shape monochrome bichrome polychrome size n small 643 conical 2 x 4380 medium 1250;1253 plano-convex 2 x large 1057; 1914;1920 1048;3145 2382 Period II-IV IV quadrangular 3 x - - large II-IV diamond-faceted polyhedron triangular-shaped annular 2 1 x x - - small small II-III III-IV 99 Unfortunately, there was no opportunity to study this material during our visit at the museum, and we are not aware of any publication discussing or describing this item in more detail. 112 3.3.3.2 Spacer-beads beads Various types of pressed beads have been recorded (Figure 70). Characteristic but not necessarily present, are two lateral suspension holes, hence the name double perforated pressed beads. Only the discoid beads have one single, very narrow perforation. 1 2 3 4 illustration of lion‘s head bead not available 5 6 7 8 Figure 70: overview of Roman cast and pressed beads produced in black glass (3: drawings by author;1: taken from Van den Hurk 1980, fig.6:m; 2; 5-8: taken from Spaer 2001, 76)  Eight-shaped beads This type of bead has a flat top and bottom surface with a bevelled edge, resulting in a trapezoidal section with a double perforation, one in either circle, similar to the more ubiquitous Trilobitenperlen discussed below (Figure 70:1). To my knowledge only one site, Esch-Hoogkeiteren in the Netherlands, has yielded such type of beads (Van den Hurk 1980, 374-378, no. VII:5:m, fig.6) [cat.no.2878]. Within grave VII were found 14 of these eight-shaped beads in black glass, forming part of a 240 mm necklace in combination with at least 69 other beads in different 100 shapes small globular (47 ex.); long biconical (9 ex.); saddle-roof (7 ex.); small segmented (3 ex.); small biconical (1 ex.). The latter is in gold and the remaining in various coloured glassestranslucent green (11 ex.), blue (2 ex.), yellow (1 ex.); opaque white (12 ex.), blue (16 ex.), terracotta (12 ex.), turquoise (2 ex.), dark brown (7 ex.) and 3 beads in so-called opaque silver-coloured glass.  ‗Trilobitenperlen‘ Despite an initial solid synthesis by Thea Haevernick (1983) demonstrating that this type of ribbed bead with double perforation is widespread, the known information remains somehow rarely exploited. 100 Van den Hurk (1980, 374375) mentions 81 complete beads and 2 damaged ones, bringing the total to 83 pieces. Counting the numbers of the various shapes and assorted glass colours, it appears that the two damaged beads were not taken into account. 113 The 339 recorded ‗Trilobitenperlen‘ of the 190 entries include two main variants on the base of decoration: the beads with ribbings parallel to the double perforation can be ‗plain‘ or transversally incised, creating a lattice or checker pattern (Figure 70:2). The name was given because of its resemblance to trilobite fossils but some scholars unfortunately denominate all double perforated beads as Trilobitenperlen (Haevernick 1983, 265-276), including those decorated with a theatre mask, a portrait or a Medusa head. Others describe such beads as ribbed and checker-ribbed spacers (Spaer 2001, 76, nos. 56-57). The so-called Trilobitenperlen correspond with Riha 11.27 (Riha 1990) and Mandruzatto type XIII.A (Mandruzatto 2008). The ribbings parallel to the perforations are very regularly rendered by means of a stamp, whereas the transverse lines are incised and consequently result in an irregular, checkered pattern. This can be very clearly noticed on the plain Trilobitenperle and the four beads with transverse incisions at the Bible Lands Museum, Jerusalem (Bernheimer 2002, 301, no. GR-36a-e). The shape may be square or rectangular, but always shows rounded corners (Figure 70:2). The ‗Trilobitenperlen‘ are medium to large sized with a width ranging between 20 and 30 mm and a length ranging between 10 and 20 mm. The perforations of 2-3 mm are applied in the middle close to each other at the longest side, but sometimes ‗Trilobitenperlen‘ have perforations at the shortest side. A seam visible along the perforated sides displays the folding of the hot gob of glass on a marver and obtained a ribbed design by pressing a tool or mould. Generally the spacer-beads are single finds but sometimes these beads were found in a set. The combined length of the set of beads and the contextual information from burial archaeology demonstrates that sets of 6 to 12 beads were used as bracelets and the unique set of 31 ‗Trilobitenperlen‘ were used as necklace [see Chapter 7]. Also the 10 beads from the Vemania castellum treasure at Isny (DE) were part of a necklace in combination with golden beads and a large set of short cylindrical jet beads (Garbsch 1971, 137, fig.30) [cat.no.1892].  ‗Theatre mask‘ beads Theater mask beads feature a double perforation like the previously discussed Trilobitenperlen. The pressed-in decoration representing a theatre mask has already been reported on black glass bracelets of type D3. Two main style variants are noticeable: 1) the tragic or severe-looking theatre masks and 2) the comical-looking theatre masks (Figure 70:3). The irregularly-shaped example from the Bible Lands Museum (Bernheimer 2002, 300-301, no. GR-34) is decorated with a circular stamp showing a grotesque theatre mask that refers to the head of Medusa. The similar-looking piece in the British Museum is stamped in oblique with a circular stamp of c. 20 mm diameter; due to the narrowness of the bead, the stamped image is not complete (Cooney 1976, 163, no. 1879) and eight others are known to have been part of the former G. Sangiorgi collection and sold at Christie‘s New York in 1999. The nine beads with identical theatre masks in severe style from a bracelet of the late Roman cemetery at Pécs – Sopianae (HU) are pressed with a rectangular stamp (Fülep 1984, 90).  ‗Lion‘s head‘ beads Some double perforated beads have one side decorated with a pressed lion‘s head (Figure 70:4). These beads are flat and discoid. Martin-von-Wagner-Museum, Würzburg (Haevernick 1983, 275) and Museum Warshau (Filarska 1962, 66, pl. II,7; Haevernick 1983, 276) show the use of different matrices to have a lion‘s head impressed.  ‗Frog‘ beads The representation on one of the pressed discoid beads in the Israel Museum, Jerusalem [cat.no.2524] depicts a frog together with an inscription in Greek (‗ZOHN‘) with two letters at either side (Figure 70:5). The provenance of the bead of the former Dobkin collection is unknown. The piece is 17 mm wide, 14 mm long and 6 mm high. The diameter of the perforations is 2 mm (Spaer 2001, 76, no. 58). Only very occasionally in black glass, this representation seems to occur much more regularly in transparent yellowish glass and on discoid pendants. The late Roman cemetery of Keszthely-Dobogó (HU) produced a similar bead in transparent yellowish glass from tomb 110 (Ságí 1981, 74, fig.56:2) and a pendant showing the same representation and inscription in yellow glass around the neck of a child in tomb 56 (Ságí 1981, fig.13b; Gesztelyi 2001, 240, fig.5;7).  ‗Helmeted head‘ beads Only one discoid bead with the stamped representation of a helmeted head has been recorded so far and it is of unknown provenance (Figure 70:6). The piece is now in the Israel Museum, Jerusalem and is 16 mm wide, 12 mm long and 6 mm high. The diameter of the perforations is 2 mm (Spaer 2001, 76, no. 59). 114  ‗Medusa head‘ beads This type of discoid double perforated pressed bead is adorned with a frontal face of the snake-headed Gorgoneion (Figure 70:7). Various seals can be recognized from the differences on some examples now in the British Museum, London and in Israel Museum and Bible Lands Museum, Jerusalem. The head is pressed at right angles to the double perforation. This type of bead has been recorded from various places in the eastern Mediterranean.  ‗Portrait‘ beads Different types have been identified: single and double portraits, man and women, looking to the left or to the right or in rare occasions en face (Figures 70:8; 71). The portraits are, for the most part, so small that characteristic details are absent. Therefore, it is mostly impossible to recognize the representation of specific persons. Portrait beads Double portrait (male + female) Single portrait Male Female Looking to the left Looking to the left Looking to the right Looking to the right Male left looking right towards female Triple portrait Male right looking left towards female Figure 71: Organization chart of portrait beads o Single male portrait The representation of the male portrait is always in profile and can be represented looking to the left or to the right. Bearded and beardless portraits do occur. o Single female portrait The representation of a female portrait can be en face or in profile, and when in profile, she can be represented looking to the left or to the right. A number of the single portrait beads represent a human figure with undetermined sex. Although we tend to categorize these beads for practical reasons as single portraits with undetermined sex it is obvious that this type is no particular type as such but rather the remainder of indefinite single portraits. o Double portrait There is a large variety discernable. The portraits differ with regard to the position of the female in relation to the male; the features of each portrait (male having a round beard, a pointed beard or beardless, and female having a specific hair dress); and the presence or absence of an inscription. 115 The beads are mainly rectangular with two rounded sides, but one piece in the Israel Museum, Jerusalem [cat.no.2518] is nicely circular (d.:14-15 mm). The piece shows the male on the left, looking right to the female. o Triple portrait On few occasions a triple portrait can occur. So far it consists of male representations two (younger) man (generally beardless) at the right looking left to an older person mainly with beard. This sort of representations also occurs on lead seals such as those from northern British sites – e.g. Southshields (UK) representing Septimius Severus and his two sons Caracalla and Geta (Allason-Jones, Miket 1984, 327) [see Chapter 7].  Miscellaneous Various other designs encountered on spacer-beads only occur in a negligible amount at present and therefore incorporated within an assorted category. For instance the bead with the representation of an imperial bust with a winged Victoria in front from Aquileia (IT) (Mandruzzato 2008, 162, B6) 3.3.4 Pendants A good number of pendants have been manufactured in the same way as beads. Pendants can even be regarded as beads with a suspension loop instead of the centrally perforated body which matches those pendant types with some bead typese.g., pressed discoid (spacer) beads and pendants, yet some have been specifically produced as pendants. The presented list of types known in black appearing glass gives a good sequence of all Roman pendant types circulating throughout the entire empire but cannot be considered comprehensive (Figure 72). A more general overview including all other glass hues is necessary considering some pendant types are solely produced in black glass and sometimes black is completely excluded. The majority of the recorded types are of late Roman date, more particularly from Period IV [see Chapter 4]. PENDANTS bulbous body and gooseneck suspension loop 'poppy-head‘ shaped barrel-shaped jug-shaped vase-shaped discoid-shaped Figure 72: Schematic view of the various Roman pendant types distinguished in black glass (drawings by author or taken from Spaer 2001) 116 The initial classifications remained very generale.g. per material, glass pendants were classified within one single category, type B1 (Eisen, Kouchakji 1927) and can hardly be considered typologies useful for archaeology. Within the brief period of the study of glass pendants the research of this commodity evolved drastically. Dan Barag provided a detailed classification of the discoid pendants distinguishing three different chronological types on the basis of the suspension loop‘s shape and subdividing each type into six categories on the basis of the pressed decoration (Barag 2001, 173-175). In particular the Israeli material from the Israel Museum collections and the Borowski collection, both in Jerusalem has received great interest (Spaer 2001; Barag 2001; 2002). The discoid pendants with pressed decoration, labelled stamped pendants, have been discussed by Dan Barag (Barag 2001, 173-176; 179-185, nos. 355-407), whereas Maud Spaer has looked at all others including the vessel-shaped pendantsjar pendants (Spaer 2001, 171; 177, nos. 332-338), jug-shaped pendants (Spaer 2001, 171-173; 178, nos. 339-354), the simple looped pendants (Spaer 2001, 176, nos. 408-422) as well as a miscellaneous set (Spaer 2001, 186-190, nos. 423-436). The late Roman vase-shaped and other glass pendants from the Egyptian antiquities collection in the British Museum, on the other hand, have been catalogued in a jumble with the amulets together with Pharaohnic material (Cooney 1976, 23, no. 241; 25, no. 263; 26, no. 269). 3.3.4.1 Pendants with bulbous body and gooseneck suspension loop This type of pendant is made in one single movement from a very small droplet of glass with one extremity stretched out and turned into a loop with the end flattened on the bulbous droplet, giving the bead a gooseneck loop, with the other extremity broken off (Figure 72). This pendant type has very rarely been attested to date. Besides the piece in black glass from Tienen (unpublished), we have only recorded one other in a deep blue glass (Pantone 541) coming from Liberchies (BE) (Vilvorder 1997, 148, fig.101:27); one from Augst (CH) in ‗naturally coloured‘ glass (Riha 1990, 161, no. 1343, Taf. 39:1343); and one in unspecified glass hue from the cemetery "dell'Università Cattolica" close to S.Ambrogio basilica at Milan (IT) (Palumbo 1999, 29, fig.4:1). 101 Despite the small size of these pendants (Table 45) there is a very striking resemblance to the so-called ‗poppy-head‘-shaped pendants of the late Roman-early Byzantine period [see below 3.2.4.2. ‗Poppy-head‘102 shaped pendant]. Probably therefore Emilie Riha catalogues the Augst piece as ‗Achterförmige Perlen‘ and describes it together with the simple poppy-head shaped pendants as ‗Doppelperlen mit Aufhängöse über einem kugeligen Teil‘ (Riha 1990, 91, 11.26). It is unclear what the relevance is to consider both pendant shapes in one and same group, but comparing the piece with the simple poppy-head shaped pendants she created confusion. Accordingly, following Riha‘s explanations, Antonella Palumbo describes the bead from Milan (IT) as ―vago a forma di ‗otto‘‖ (Palumbo 1999, 29). Table 45: Dimensions (in mm) of pendants with bulbous body and gooseneck suspension loop max. diameter max. height cat.no. site colour (bulbous body) (suspension loop included) 537 not catalogued not catalogued not catalogued 3.3.4.2 Tienen (BE) Liberchies (BE) Augst (CH) Milan (IT) black blue blue-green unknown 7 6 9 5,5 12 10 12 9,5 ‗Poppy-head‘ pendants These pendants are generally described as jar pendants (Spaer 2001, 177, nos. 332-336; 338; Whitehouse 2003, 50, no. 967), but since they have no rim or handles, we prefer to distinguish them from vase-shaped pendants 103 which are completed with one or two handles and an explicit rim (Spaer 2001, 177, no. 337). Others prefer to speak about basket-shaped pendants ‗en forme de panier‘ (vitrum 1990, 128, no.129:2). We favour the nomenclature of ‗poppy-head‘-shaped seeing the resemblance with the seed pod of a poppy hanging downwards. Occurring in various decorative ways, they all are clearly built up similarly with an annular suspension loop attached to a globular body and a discoid pad base added at the bottom. The body could remain (a) plain (Spaer 2001, 177, no. 332-334); (b) decorated with tooled vertical ribbings (Spaer 2001, 177, no. 338), comparable to the so-called ‗melon‘ beads); (c) decorated with applied wavy or zigzag line (Spaer 2001, 170, fig.76 right; 177, no. 101 We are aware of the fact that the database confronts the dimensions of only four pieces, which makes the present deduction necessitates further corroboration. 102 The term ‗eight-shaped bead‘ is taken over from Keller 1971, 92. 103 It appears odd and over-reaching to interpret a rather undefined shape as a specific vessel shape. 117 335); or (d) speckled with spots in one or various opaque glass hues (Spaer 2001, 177, no. 336), comparable to the so-called ‗crumb‘ beads (Figure 73). Sometimes the body is flattened and appears manufactured with very swift movements giving it a rather clumsy appearance. A number show a handle that is not applied separately but made in one movement from the glass gob of the body. These similar-looking pieces imply one specific handling method, making them most probably attributable to one workshop or to several contemporaneous workshops. a. b. c. d. Figure 73: Diversity of the poppy-head-shaped pendants: a. plain; b. ribbed; c. wavy/zigzag line; d. mottled (drawing by author) 3.3.4.3 Barrel-shaped pendants The barrel-shaped pendants are similar to the barrel-shaped beads with generally one large central suspension loop, e.g. the piece from the inhumation burial at Colchester (UK) (Crummy, 1983, 35, fig.37, no.1504) [cat.no.4023], but occansionally three loopse.g. the piece from Aquileia (IT) (Mandruzatto 2008, no.158) (Figure 74). The body is in general solid but some have a central perforated body like the one from the necklace found within the inhumation tomb 544 of the cemetery at Bregenz (AT) (Konrad 1997, 76; 220, pl.32:C) [cat.no.20]. The body shows a great variety in shapesstraight cylindrical; barrel-shaped; biconicaland sizefrom short to very elongatedbut all are very characteristically decorated with applied glass trails in one or two contrasting opaque hues (white; yellow; red; blue). The design can be solely a spirally twined trail at both ends leaving a plain space in between or showing a simple or double crosswise zigzag pattern. Less common are the one sided zigzag from one end to the other like the Aquileia piece (Figure 74). Figure 74: Diversity of the barrel-shaped pendants: top: single looped – left: elongated (Trier); middle+right: short (Cyprus; Colchester); bottom: multiple looped - Aquileia (top: drawings by author; Colchester taken from Crummy 1983, no.1504; photograph taken from Mandruzatto 2008) 3.3.4.4 Jug-shaped pendants The jug-shaped pendants are hollow due to the rod-formed technique by wrapping hot glass around the mandrel [see Chapter 2]. Despite the great diversity of jug-shaped pendants, these were all catalogued by Horace Beck (1927, 33) as class B15. They can be very plain or elaborately worked out variants, such as the example from tomb 1127 of the large cemetery of Pritzier (DE) with a tall cylindrical neck built-up from winding a coil eight times, an oval-shaped body decorated with a zigzag in opaque white glass, a small discoid foot and one long vertical handle from rim to shoulder (Tempelmann-Maczyńska 1985, 57; Taf. 9:332) [cat.no.A4600] (Figure 75). The body can be tall or more truncated and have a globular, cylindrical or biconical body. All have a pad base and one vertical handle from shoulder to rim. Sometimes the glass trail is further wound around the rim or applied separately in a contrasting colour to accentuate the rim. Interestingly, the open-worked jug-shaped pendants have not been recorded in black glass and are nearly only produced in blue glass like the aryballos-shaped one in the British Museum (Cooney 1976, 26, no. 269). Some remain plain or have a mottled decoration, but the majority of the jug-shaped pendants have a trailed decoration by applying and marvering opaque blue (turquoise), white, yellow or red glass in a zigzag or wavy pattern or spiralling around the body. None have been attested so far with a ribbed body. 118 Figure 75: Diversity of the jug-shaped pendants (all drawings by autho, except left taken from Tempelmann-Maczyńska 1985; other) 3.3.4.5 Vase pendants Only one of the seven jar pendants discussed by Maud Spaer, is, according to us, a vase pendant (Spaer 2001, 177, no. 337) (Figure 72). It has a ribbed globular body, an everted rim, two small vertical handles from shoulder to rim, a discoid pad base and an annular suspension loop. A very similar-looking piece in the Israel Museum, Jerusalem is the ‗amphora‘ bead (Spaer 2001, 75, no. 50), showing no discoid foot or annular suspension loop. The central perforation and the absence of the foot and suspension loop proofs the piece must have been deliberately produced as a bead. Its resemblance in shape and manufacturing technique means that the vase beads and pendants were probably made in the same workshop. Its resemblance to the cantharus-headed jet pins tempts us to determine the shape of the vase as that of a cantharus [see Chapter 7]. 3.3.4.6 Pressed discoid pendants Discoid pendants with pressed-in decoration are very similar to the discoid beads with impressed decoration [see 3.3.3.2] (Figure 72) and must be seen as contemporaneous productions from the same workshop, leaving open the discussion whether only one or more workshops produced similar material. Different from the beads is the presence of a vertical suspension loop instead of a single or double perforation through the middle. Given our restriction to black-appearing glass artefacts, we only discuss a very limited number of the discoid pressed pendants and beads (Table 46). A full treatise would allow us to identify the various stamps employed with the intention of recognizing regional distributions and chronological different periods. Of course, such an investigation should evaluate whether the specific stamps could not have been reproduced frequently from one patrix, subsequently becoming prevalent in different contemporaneous workshops. Table 46: Ratio of various coloured pressed discoid pendants in large collections (YG = yellowish green; YB = yellowish brown; P = purple; G = green; B-G = blue-green; B = blue) black YG YB P G B-G B collection total n % n % n % n % n % n % n % reference Corning Museum (USA) 38 2 5,3 8 21 17 44,7 1 2,6 1 2,6 5 13,2 4 10,6 Whitehouse 2003, 13-26. Borowski Collection (IL) 40 2 5 4 10 26 65 - - 2 5 2 5 3 7,5 Barag 2002, 307-321. Israel Museum (IL) 56 2 3,6 3 5,3 33 59 2 3,6 1 1,8 6 11 8 14,3 Spaer 2001, 179-185. Dan Barag distinguished three different types of discoid pendants on the basis of the suspension loop‘s shape and the axis of the stamp featuring three overlapping consecutive periods (Barag 2001, 173-175): o Type A (second half 4thearly 5th century AD) The discoid pendants of this type are the result of careful workmanship. They have a suspension loop which is flattened at the front. The design is always impressed with great care in a vertical axis with the suspension loop. o Type B (late 4th5th century AD) Compared to the type A pendants, type B pendants are of more inferior workmanship. They have a suspension loop which is more protruding, being rather rounded, convex and annular. The design is impressed less carefully, mainly stamped along a vertical axis, while some are stamped within a horizontal axis. 119 o Type C (second half 5 first half 6 century AD) th th This type of discoid pendant is of mediocre workmanship, having an irregular shape and size. Its suspension loop is large and annular. The design is impressed nonchalantly along no particular axis.  Category 1: Classical mythology and pantheon The classical Olympian gods are lacking. The generally occurring representations depict Victory or Nikè, Hercules, Sol Invictus, Romulus and Remus or the Gorgo Medusa. Typical for type A, they sometimes crop up on type B and C pendants. One piece of type B in black glass depicting a helmeted figure wearing a paludamentum, probably representing Minerva/Athena, is now in the collections of the Bible Lands Museum, Jerusalem (Barag 2002, 3104, no. LA-18) [cat.no.2555]. The bust is represented in profile facing the left and in front of her the cornucopia. The diameter of the impression is 13 mm, and the pendant is 19 mm wide and 23 mm high. Because the suspension loop is in a vertical axis at the bottom of the representation, the stamped figure is hanging upside-down when worn.  Category 2: Biblical themes All illustrations are of Old Testament passages (Adam and Eve, the Sacrifice of Isaac, Daniel in the Lions‘ Den, and the Good Shepherd) and occur on all three types. One piece of type C in black glass depicting Daniel in the Lions‘ Den is now in the collections of the Bible Lands Museum, Jerusalem (Barag 2002, 310-311, no. LA-5) [cat.no.2549]. Daniel as an orans flanked by two lions and rendered in a stylistic style within a circle is stamped in ‗a 10 o‘clock axis‘. The two seated lions are oriented away from Daniel, but their heads are turned towards him. The diameter of the impression is 14 mm, and the pendant is 15 mm wide and 18 mm high.  Category 3: Jewish symbols Typical is the menorah (the seven-armed candlestick) on a tripod, the shofar (the ram‘s horn), the lulav (palm branch) and the etrog (the citron). Jewish symbols occur on all three types. One piece of type C in black glass with reddish-brown streaks depicting a menorah is in the collections of the Israel Museum, Jerusalem (Spaer 2001, 181, no. 373) [cat.no.A4593]. The diameter of the impression is 11 mm, and the pendant is 21 mm wide and 27 mm high.  Category 4: Christian symbols The Christian symbols feature the Chi-rho, sometimes combined with an alpha-omega, the cross and stylite saints such as Saint Symeon. Dan Barag curiously adds the double portrait pendants to the Christian symbols by considering these confronting busts of opposite sexes representing married couples as a well-known image in the th tradition of 4 century AD early Christian art (Barag 2001, 174). Christian symbols occur on all three types. One 104 piece of type C in black glass depicting a square cross within a circle stamped in ‗a 10 o‘clock axis‘ is in the collections of the Israel Museum, Jerusalem (Spaer 2001, 181, no. 375) [cat.no.A4594]. The diameter of the impression is 20 mm, and the pendant is 21 mm wide and 25 mm high.  Category 5: Faunal motifs The lion is the largest group within the category of faunal motifs specifically but also of all categories in general. Walking to the left or to the right with a frontal face or in profile, the lion regularly is represented with a star/sun and a crescent moon. On few occasions, the lion is shown with one paw on a bull‘s head. Other regularly depicted animals are the frog, bull, goat, and hen. Sometimes the representation is executed stylistically. Faunal motifs occur on all three types. One piece of type C in black glass depicting a lion walking to the left with the head in profile is in the collections of the Israel Museum, Jerusalem (Spaer 2001, 183, no. 393) [cat.no.A4595]. The diameter of the impression is 15 mm, and the pendant is 17 mm wide and 25 mm high. 104 The colour is described as an opaque matrix with greyish surface. 120  Category 6: Miscellaneous motifs Dan Barag puts all other motifs not suitable for the above-mentioned five categories within category 6. It incorporates floral motifs, the seven-rayed star, and a pair of sandals. Probably these can become separate categories in the future. We propose to take the portrait pendants as a separate category just as for the finger rings, beads and the gems. 3.3.5 Gems Like numismatic studies, for centuries the study of gemstones has been a specialized theme within Roman material culture studies. The study of gemstones not only discusses the material and shape or size, but also the rendering (technique, style), the represented theme (iconography; iconology) and the carrier wherein the gems were set. Considering the high amount of loose stones from unknown provenance an attempt to provide a full 105 overview on the gemstones made of black glass seems quite pointless within this diachronic research. Despite the awareness that the gemstones in black glass represent a major issue by itself, we wanted to present general but useful information on a particular segment within the consumption of black glass during the Roman imperial 106 period. Because a number of gem types have been made in black glass we wish however to discuss very briefly the different recorded types. Only few different gemstone types have been discerned that were made of black glass or in combination with glass in a contrasting colour: Plain conical gems, discoid pressed decorated gems; blueon-black discoid gems with bevelled edge imitating nicolo (elliptical and round); black and white-banded discoid gems, imitating onyx (elliptical and round) Gemstones in glass are for the most part loose finds, anymore set-in in the object it adorned. But when still inserted various gem types made of black have been recorded depending on the artefact type. Because particular types are related to specific uses, we made here already a distinction between gems in brooches, gems in finger rings and gems in helmets while the function and use of the gems will only be discussed in a later chapter [see Chapter 7]. All gems have been produced by casting; are round or elliptical; have a flat or convex top and basal surface; are made in monochrome or bichrome glass. The brooches are adorned with the inset of a disc-shaped gem (conical or plano-convex), whereas the most frequent shape of gem for the finger ring is elliptical with two flat or curved sides. No square, rectangular or polygonal gems in (black) glass are known to our knowledge. Elongated oval gems in glass are not only fairly rare, they are only known for the early Roman imperial period in the form of the well-known three-coloured banded gems in opaque white-emerald green-ultramarine blue glass and never in black glass. 1. 2. 3. 4. Figure 76: Overview of black glass gemstones: 1) plain conical; 2) plano-convex; 3) small elliptical; 4) large round 3.3.5.1 Plain conical gems A particular use of black glass is the setting in the centre of disc-shaped and oval gilded bronze brooches. The glass gem in the centre of these round or elliptical discoid fibulae is a small plain cone in a very deeply coloured 105 We started to inventory as much as possible black glass gemstones in the data-base, but seeing the enormous amount of this artefact category remained underdeveloped within our research so that on the whole we had to remain very brief. See introduction for references on the issue pointing to the large amount of works published so far but this would have let us too far from the main research objectives. 106 The gems appear to have received more exclusive value in contrast to other jewellery categories made of black glass, such as arm rings, finger rings and beads. It would be very useful to focus in a further research project on the glass gemstones from the different areas within a specific region such as the north-western provinces. The gemmatologists rather focus on the precious and semi-precious stones or on the rendering and the interpretation of the figurative scene. Interesting about the glass gemstones is the technological aspect of impressed decoration when casting the pieces. Seriation is possible and distribution patterns of gems with a particular image from the same mould can determine the extent of popularity of a workshop/jeweller. 121 glass appearing black with a diameter averaging around 20-25 mm and a height of 10-15 mm (Figure 76:1). The gems are mainly disregarded from any discussion because the attention is fully drawn at the brooches (Böhme 1972, Type 45; Riha 1979, type 3.17; Bayley, Butcher 2004, T270-271). Most common are the plain conical gemstones in monochrome black appearing green glass set in a circular discoid brooches (T270) with 41 ex. in comparison with 18 ex. set in elliptical discoid brooches (T271) (Bayley, Butcher 2004, type T270-271) (Table 130). Probably many more have been excavated so far as the conical glass gemstone is sometimes defined as being of stone (Vrielynck, 2007, 54). This very specific use of the conical gemstones makes it indeed possible to ascribe loose finds to discoid brooches like the one from the military fort at Richborough (UK) (Bayley, Butcher 2004, 134, no.388). Numerous examples of this specific brooch are known from Britannia, though examples of these Punzverzierte ovale Scheibenfibel mit Glaseinlage were also attested on the Continent along the Rhine region up to Augst/Kaiseraugst (CH) (Riha 1979) and beyond the Roman Empire into Barbaricum [see Chapter 6]. 3.3.5.2 Plano-convex gems In contrast to the generally simplified and abstracted design of plain conical gems centrally set in discoid fibulae type T270/271, stands a very elaborate piece from the Roman cemetery of Wancennes (BE) (Bequet 1883, 368369; Mignot 1984, 223-224, fig.36-18; Cosyns in press) and now in the Musée Archéologique of Namur [cat.no.617]. The gem in question is a plano-convex circular gem with a diameter of 18 mm decorated with a cast or pressed image of a human face (Figure 76:2). It adorns the centre of the discoid gilded bronze brooch, which measures 37 mm in diameter. The gemstone, in a very dark bottle green glass with a black appearance, most likely represents the head of Medusa obtained by mould pressing. The outer side of the bronze plate is decorated in repoussé-technique with a row of eyes flanked by a row of pearls and an inscription ‗PERSEUS CONCIDERA[T] ° CAPUD GORG[ON]IS‘ (trans.: ‗Perseus knocked off the head of the Gorgon‘) [see Chapter 7]. A similar small plano-convex gem with a similar mould-pressed human face comes from cremation tomb 80 from the cemetery of Wanzoul (BE) (Destexhe 1989, 76-77, pl. 37:2) [cat.no.624]. This piece –with a diameter of 12 mm and a maximum thickness of 5 mm– is somewhat smaller than the Wancennes piece, but the representation of the face of Medusa with thick cheeks within a beaded circle is similar. Although no brooch has been reported as a burial gift, it is very probable that this gem was also set in a discoid gilded bronze brooch. No clear date has been provided for the Wancennes piece, but the Wanzoul burial context has been dated between the end of the nd rd 2 and first half of the 3 century AD. A much larger piece with a diameter of 27 mm, now at the Bible Lands Museum, Jeursalem (Bernheimer 2002, 283, no.GR-3), clearly demonstrates a human face on the roundel in blue glass depicting a Medusa-head, with wings in the hair dress and a lolling tongue. Seeing the resemblance with nd the abovementioned pieces we believe that the piece – which is considered a Hellenistic fabric of the 2 century BC– is dated much too early. Similar pieces are recorded in the museums of Munich and London (Brandt et al. 1972, nos. 3522-23; Walters, 1926, no.3541). A pale greenish blue bead in the Israel Museum, Jerusalem is decorated with a comparable face which is described as a grotesque face or possibly Bes (Spaer 2001, 74, no.44). Four small discoid bronze plates from a late Roman tomb 212 at Pécs (HU), ancient Sopianae, represent the head of Medusa in front within a beaded circle in repoussé (Fülep 1977, 40, pl. 32:3), not only demonstrating the omnipresence of this type of imagery but also of its occurrence on ornamental elements of male dress. A second type of plano-convex shaped gems is illustrated by the small piece in deep purple glass from the villa of Valkenburg (NL) and now in the Limburgmuseum at Venlo [cat.no.2991]. The gem appears similar to the small plano-convex counters with round edge. Another small plano-convex gem in black glass with engraved decoration is the one still set in the golden finger ring (Guiraud type 2c) from the excavations on the Hunerberg at Nijmegen (NL) (Van Enckevort 2008) [cat.no.A4617]. 3.3.5.3 Small elliptical gems The elliptical gemstones with convex or flat surfaces are made in a wide range of material as well as in various glass hues (Figure 76:3). In the introduction we refered to the wide assortment of publications available on intaglios, including wide-ranging reference lists. We consequently limited the account of this functional type by providing only the most important observations. Working out this vast category of gemmata properly would be an entire research project in itself. 2 We use the typology built by Erika Zwierlein-Diel and John Boardman as presented by Martin Henig (1978 , 35, fig.1) taking into account all sorts of gemstones, including those in glass (Figure 77). Two major groups have been discerned within the gem shapes type C: gems with convex top surface; 2) type F: gems with flat top surface. Each type is subdivided on the basis of the section variations into seven and eight subtypes respectively, taking 122 into consideration that every C subtype can demonstrate three variants in the convexity of the top surface – ―a‖ for weakly convex, ―b‖ for moderately convex, and ―c‖ for strongly convex. Curved Flat Figure 77: Overview of gem shapes (illustration taken from Henig 1978², fig.1; after E. Zwierlein-Diel & J. Boardman) The black glass gems are only known of with a flat surface (type F). The nearly total absence of black glass gems with a convex surface cannot fully be explained from a technological point of view. Since glass gems are made by pressing glass in a mould [see Chapter 2] convex surfaces can be obtained as easily as flat ones. Given that the gemstones in glass used on finger rings are to be considered as imitations of semiprecious stones, it is important to know that nicolo, onyx and other layered stones are for the most part cut in type F2, F4 and F4a and in F3 when the stones were thick enough (Sas 1993, 114). While F2 and F3 have a fully bevelled edge, F4 and F4a show a carination halfway through the section, becoming straight or conversely bevelled. Typological studies on the different semiprecious stones popular during the Roman imperial period show that the convex gems in semiprecious stones are idiosyncratic for the Republic period and the very early imperial period (Sas 1993, 114). The layered nicolo gems are solely cut in the types F2, F3, F4 and F4a, respectively flat conical, high conical, biconical and conical with straight sides for the bottom layer (Sas 1993, 116, fig.6). This information, however, needs to be taken with some reservation as a large part of the pseudo-nicolo intaglios in glass could not be examined fully due to the fact that they are still mounted in a ring (Table 47). The flat topside in all cases means that the glass intaglios are all of type F. Table 47: Number of gemstones from Belgium related to finger rings (taken from Sas 1993, 126) intaglio intaglio in setting intaglio in ring ring 99 ex. 4 ex. 102 ex. 16 ex. 44,8% 1,8% 46,2% 7,2% total 221 ex. 100% The catalogue of Belgian material set up in 1993 by Kathy Sas includes 89 intaglios in glass, of which 56 pieces (62,9%) are imitations of nicolo while only 20 pieces are made from nicolo (Sas 1993, 129, fig.22).  Nicolo imitations Bichrome blue-on-black glass gems are two-layered glass gems combining a support in black glass covered with a layer in bluish glass. This gem type with a double layered glass is known in literature as ‗nicolo paste‘ and ‗pseudo-nicolo‘ and was mass-produced imitating nicolo-onyx, a popular semiprecious stone in Roman times. We have observed that some authors speak about a layer of light blue glass topping a layer in dark blue glass and not of one in black-appearing glass although the opacity of the used deeply coloured glass makes it impossible to discern the true hue. Because we were unable to verify these pieces, it is uncertain whether we should speak of two dissimilar technological varieties or whether it is only due to the discrepancy in colour perception as explained in Chapter 1. There seems at least to be a consensus about considering these intaglios as an imitation of nicolo stones. 123 All blue-on-black glass gemstones have a flat surface (Henig type F) and bevelled edge. The bevelled edge is either short (Henig variants 2 and 4) or wide (Henig variant 3). When the gem with short bevelled edge is still setin the original ring it is impossible to determine whether we have to deal with variant 2 or 4. For this reason the combined description F2/4 occurs in the database. Although particular attention is drawn to the section of these gemstones by Martin Henig (1978²), the differences in shape have to be taken into account too. Emilie Riha explains that the bichrome glass gemstones imitating nicolo have been divided into two large groups by Gertrud Platz-Horster (Riha 1990, 22-23). The first group has, in fact, a third glass layer in between the blue and the black layer. The blue glass of the first group has a relatively thick layer in vivid deep blue glass, whereas the second group only consists of a thin and dull greyish blue top surface. Furthermore, it is suggested that the second group comprises only the thin gemstones, while the first group is represented by the thicker ones. A thorough assessment of this material from a technological point of view, including the assessment of the chemical composition probably would provide interesting new information for a better reading of this commodity within the Roman material culture.  White-banded onyx imitations The white-on-black glass gemstones in glass are commonly of relatively large dimensions compared to the semiprecious stones. Examples of black glass gems with a white band through the centre of the top surface are rarely recorded. An intact gem, now in the Fitzwilliam Museum, Cambridge (UK), measures 18,5 x 9,5 mm and is adorned with an engraved scene representing the shepherd Faustulus who finds Romulus and Remus suckling a wolf (Glass at Fitzwilliam 1978, 31, no. 49b). Another black glass gemstone imitating white-banded onyx is located in the Limburgmuseum, Venlo (NL), apparently from a villa rustica at Valkenburg [cat.no.2991]. The loose flat gem with bevelled edge (Henig type F2) measures 12,5 x 9,3 mm and is 2,7-3,1 mm thick. A bead-edged pastoral scene is represented with a bearded man standing towards the left and carrying a lamb on his left arm, while a dog in front of the shepherd is jumping towards the lamb. The scene is pushed in and is of very high quality.  Black onyx imitations A number of small sized gems are made of monochrome black glass and have an elliptical shape with curved surfaces (Henig type C1). In combination with the engraved design, the applied decorative style and the depicted themes these glass gems appear similar to the early imperial gemstones in semiprecious black onyx. These monochrome black glass gems can be considered pseudo-onyx stones; one of the many imitations of the more expensive semiprecious gems produced on a large scale Pliny the Elder complained about. 3.3.5.4 Large round and polygonal gems Alongside the small and elliptical shaped pseudo-nicolo gems, a small group are large and circular in shape (Figure 76:4), clearly demonstrating an affinity with a polygonal shape, like the octagonal gemstones on the Berkasovo-helmet (Mackensen 2009, 289-294, Taf.25) [cat.no.3108]. The decorated helmet from Berkasovo (SB) in gilded silversheet contains eight large intaglios in blue-on-black glass. The dimensions of the Berkasovo gems have not been provided, probably because they are still set on top of the helmet next to the crest of the helmet (Figure 132). The other large round gems have a thickness of 3,5-5 mm and a bevelled edge similar to the abovementioned small elliptical gems type F2 but measure 20-25 mm in diameter, with 15-16 mm for the blue top layer. 3.3.6 Hairpins Roman hairpins from archaeological contexts are above all manufactured in bone or bronze, but silver and even golden hairpins were also relatively commonly used. In particular, in contexts of the late Roman period hairpins were more recurrent when prevailing hairstyles led to their frequent use (Crummy 1983, 19-30; Riha 1990, 95114; Stutzinger 1995; Allason-Jones 1999; Rodet-Belarbi, Van Ossel 2003, 337-340). Hairpins in lead, iron, jet/shale and glass do occur as well, but appear to cover only a fraction of this functional type, as can be observed 107 for Augst/Kaiseraugst (CH) (Riha 1990, 96, tab.122). The reduced number of hairpins in glass makes it less 108 suitable to develop a separate typology; this is also the case for jet/shale hairpins. Various typologies are 107 108 Emilie Riha mentions a share of 0,6% of all hairpins from Augst and Kaiseraugst that are made in iron, lead, jet or glass. Ninna Crummy distinguished only two types of hairpins in jet: 1) with spherical head; 2) with faceted head (Crummy 1983, 2728, fig. 24). 124 available covering the hairpins in general (Harden 1979, 315-316) or focused on specific materials such as bone and bronze (Crummy 1983; Riha 1990). Besides the glass pins produced in opaque black glass, pins in translucent to transparent brown, bluish-green and pale green glass also occur. Some of the pins in black glass show streaks in very vivid red. Most glass hairpins have a short shaft of 50 to 80 mm in length and a large head of 8 to 10 mm in diameter. The shaft of black glass pins has a maximum diameter of 4 to 5 mm, thinner than the similar hairpins in jet/shale/cannel coal. The very standardized measurements mean that no classification can be built up from it. Compared to the pins in jet/shale/cannel coal, the (black) glass pins fall within the thick short type. The shaft is plain and cylindrical, tapering slowly towards a pointed end and a constriction just below the head. The most characteristic feature is the shape of the pinhead. The head can be (a) spherical, (b) conical, (c) discoid (flattened sphere) or (d) elliptical (narrowed sphere) (Figure 78). a b c d Figure 78: Glass hairpins with a) spherical head; b) conical head; c) discoid head; d) elliptical head On rare occasions only the head is made from (black) glass and attached to an iron pin, as attested in Oudenburg (BE) [cat.no.471], Trou de Han (BE) [cat.no.309;311;313], and Augst/Kaiseraugst (CH) (Riha 1990, 184, nos.2784-2787) [cat.no.A4592]. Les Houis (FR) [cat.no.1473-1475;1477-1485]. That the Romans produced iron hairpins with a glass head has been attested earlier at the late Roman cemetery of Lankhills at Winchester (UK) 109 (Harden 1979, 249). Similar are those in bronze and silver mainly decorated with an emerald green polygonal bead. An intact bronze pin with plano-convex head in black glass comes from the villa at Shakenoak near Wilcote (UK) (Harden 1971, 106, no. 152, fig.45:69) [cat.no.4201] (Figure 79). The total length of the piece is 79 mm. The glass head is 5 mm high and 9 mm in diameter. Pushing the bronze wire into the hot glass flattened the glass blob on the underside and slightly enlarged the diameter. Hitherto only one intact example in black glass has been recorded. The piece comes from the glass workshop of Les Houis nearby Sainte Menehould (FR) and measures 67,5 mm in length with a maximum diameter of 4,5 mm for the shaft and 8,8 mm for the head (Cosyns 2009) [cat.no.1474]. Figure 79: Bronze hairpin with black glass head from Shakenoak (UK) (taken from Harden 1971, fig.45:69 – scale 1:1) 109 The large plano-convex and small globular pieces holding the remains of two metallic pins from Roman sites as, for instance, Augst/Kaiseraugst (CH) and Trou de Han (BE) are intrusive modern material. These pieces are not Roman hairpins but 18th–19th century buttons. The two pins are the remains of a loop to attach the button on a piece of clothing. 125 3.4 Architectural decoration The glass architectural decoration material can be separated in various categories, i.e. tesserae, twisted rods, bichrome plaques and inlays or intarsia (Grose 1989, 356-358; Pasqualucci 1999, 213). No classification on the architectural decoration material seems compulsory since each category is not very assorted. Only the tesserae show a variety of shapes. Classifying these tesserae shapes however seems useless because all variants have been made in whenever period, whatever colour, whatever material, and wherever place. The rods, on the other hand, have been produced only in glass, during a confined period and in one way: twisted. These are known in monochrome and bichrome glass. The bichrome plaques form a somewhat strange category because they can be considered as unfinished products to produce twisted rods as well as they have been used as inlays [see Chapter 7]. 3.4.1 Tesserae Working out a taxonomy on glass tesserae implicitly leads to a first classification on the basis of their colours (Fiori 2001, 303-5), but as we are here only dealing with black-appearing glass, we can skip this categorization. Regarding the black glass tesserae, we recognized differences in size and shape which are equally apparent in other colours (Figure 80). Figure 80: Selection of black glass tesserae shapes from the site at Les Houis, nearby Sainte Menehould (FR) in Musée des Antiquités Nationales (M.A.N.), Saint-Germain-en-Laye (photograph by the author, courtesy of M.A.N.) Three groups can be distinguished when concerning the size: 1) 2) 3) small-sized tesserae – below 5 mm side; medium-sized tesserae – between 5–10 mm side; large-sized tesserae – above 10 mm side. Within the three above-mentioned sizes, some specific shapes turn up repeatedly: 1) 2) 3) 4) cubic shaped – squares on six sides; pyramidal shaped – triangles on two sides, and rectangles on three sides; rectangular shaped – squares on two sides and rectangles on four sides; trapezoidal shaped – trapezoids at two sides, squares and/or rectangles on four sides. In relation to the black glass tesserae, we always bear in mind that they can be made from various coloured glasse.g., green, brown, purple, blue-green (Figures 81; 82), and consequently from different recipes. The initial idea was logically that each recipe was characteristic for a chronological phase and/or regional demarcation pointing to a specific workshop or a group of workshops. But how can we interpret the presence of various kinds of black glass colours in one single workshop, as in Les Houis near Sainte Menehould (FR)? Is it that the workshop enjoyed a long-lasting and successful tesserae production in a wide range of colours, persisting through subsequent evolutions within Roman glass technology? Or did the glass workshop at Les Houis not include the colouring of glass, making the workshop depend on the available or attainable supply from various 126 specialized glass centres involved with the production of coloured glass cakes? In that case, we then have to interpret the several glass cakes retrieved from Les Houis as imported raw material and Les Houis as only a local workshop within an extensive chaîne opératoire to provide mosaicists with material. Figure 81: Set of tesserae from the site at Les Houis, nearby Sainte Menehould (FR) in Grand Curtius Museum, Liège (BE) since the late 19th century (photograph by author, courtesy of Grand Curtius Liège) Figure 82: Set of 455 black tesserae (Inv.no. 83401-B) from the site at Les Houis, nearby Sainte Menehould (France) in Musée des Antiquités Nationales (M.A.N.), Saint-Germain-en-Laye (France) (photograph by the author, courtesy of M.A.N.) 3.4.2 Twisted rods The number of recorded rods that are made of black appearing glass remained rather small compared to the occurrence of glass rods in other hues. A more detailed study probably would increase the number, but we can question whether this accumulation would contribute to the understanding of this commodity, its variations, its 127 110 evolution, its technique as well as its use and function? An excellent set of material, R1609, in the collections of the Royal Museums of Art and History of Brussels (De Meester de Ravestein 1884², 463) includes 111 pieces of which 81 are twisted rods in various colours (Table 48; Figure 83). But other large glass collection contains similar materiale.g. Toledo Museum and Corning Museum (USA) (Grose 1989, 349; 358; 370); the Gorga collection in Rome (IT) (Pasqualucci 1999); Landesmuseum Württemberg, Stuttgart (Honroth et al. 2007, 136-138). The rods are plainly twisted (S- and Z-twined) or decorated with a twining opaque white or yellow glass trail every few twists. These spiral trails can be very wide easily > 5 mm and sometimes even wider than the space between two twists. The opaque yellow glass trails are solely applied on rods in decolourized rods. All others have opaque white spiralling glass trails. Mainly in a translucent deeply coloured glass, some appear black to the naked eye, even with the help of a strong light source. Only thin sections via sampling were able to reveal the true colour. Table 48: List of various glass colours of twisted rods from R1609 at the KMKG-MRAH, Brussels colour quantity % so-called black purple brown blue (from pale to deep coloured) blue-green white opaque colourless + yellow TOTAL 10 10 2 35 4 3 17 81 12,3 12,3 2,5 43,2 4,9 3,7 21,0 99,9 The material is always fragmented with a preserved length ranging between 20-90 mm, but generally lesser than 60 mm and as far as we know at most reaching 110 mm like a fragment in Toledo Museum (USA) (Grose 1989, 370, no.670a). One can wonder whether these rods were initially much longere.g. the length of the side of the polychrome mosaic panel. It is likely that bits and pieces were embedded in the plaster, but a more close study should provide the answer to this issue. The diameter of these twisted rods range between 5-10 mm with a preponderance between 7-8 mm. The twisted rods were in particular used during the early Julio-Claudian period [see Chapter 4] to border the polychrome wall and vault mosaic panels in nymphaea and balnea together with rows of particular seashell types [see Chapter 5] (Sear 1977; Grose 1989, 358, fig.170). As far as we know only twisted rods were used to frame the central polychrome mosaic panel. It is unclear whether the Romans also used the plain rods for bordering their mosaic panels and whether they are also encountered in the archaeological findings as such. Figure 83: An overview of the R1609 material from the KMKG-MRAH, Brussels (photograph by the author, courtesy by KMKGMRAH) 110 We expect in museum collections a far larger amount of pieces that are mainly from former private collections built up in the late 19th and early 20th century. Information on provenance and context is missing in nearly all cases because the material got acquired primarily from the antique market. 128 Addendum: The so-called stirring rods Considering the technology and shape, we thought it best to insert here the discussion of the so-called stirring rods as they resemble the twisted rods used in the wall and vault mosaics discussed above. Even though this type of material is normally discussed with the vessels (Isings Form 79), we have to acknowledge that these rods are not recipients but rather utensils and received already multiple functions [see Chapter 7]. Within the scope of this discussion, it is evident that the function of small twisted fragments cannot easily be ascertained. It is for instance not clear whether the 30 mm small fragment at the Musées de Poitiers (FR) (Simon-Hiernard 2000, 377, no. 350) [cat.no.1737] is a fragment of a so-called stirring rod or if it has been used as architectural decoration material to border a wall mosaic. David F. Grose (1989, 358) differentiates both applications on the basis of the intensity of the twists which are tighter on the architectural rods and that these lack the different ends characteristic to the twisted rod utensils. Essentially produced in transparent pale bluish and pale yellowish green glass, only a small number of them are produced in a very dark glass. Seeing the very few strongly coloured and black-appearing examples we could record (13 ex.), it is totally beside the point to discuss here any classification. Once more we have to admit that only when all rods in whatsoever colour are taken into account a clear-cut overview of variations in the decoration of the shaft and the ends can be obtained. It is however possible to generate a framework wherein this commodity type can be subdivided on the basis of technology and decoration (Figure 84): 1) simply drawn and remaining plain or 2) drawn and twisted; dividable on the basis of the range of finishing at both ends. RODS plain decorated (twisted) S-twined (clockwise) Z-twined (anti-clockwise) S-/Z-twined (partly clockwise; partly anti-clockwise) top end loop figurative decoration (vase; bird; face) bottom end discoid pointed ... Figure 84: Chart of the rod varieties based on technology 129 Figure 85: Stirring rod from tomb 272 at Amathous (CY) (drawing by the author, by courtesy of the District Museum of Limassol) The Amathous rod (Figure 85) [cat.no.776] one of the few black-appearing piecesis the complete but broken rod from tomb 272 from the cemetery of Amathous (CY) (Tytgat 1992, 85, no. 692, pl. XIX). It is a simply twisted rod like two other examples coming from Cyprus and now at the Fitzwilliam Museum (Glass at Fitzwilliam 1978, 50, no. 96). But while the two Fitzwilliam rods are S-twined, the Amathous rod is Z-twined. Another example is now in the collections of the Glass Museum at Murano (IT) (Ravagnan 1994, 119, no.222). The Murano rod, on the other hand, is partly S-twined and partly Z-twined, as is the rod of the former Ernesto Wolf collection and now in the Landesmuseum Württemberg, Stuttgart (Stern, Schlick-Nolte 1994, 396-397, no. 228). Frequently, these rods are decorated with an applied glass trail twisting around every few twists (as, for instance, in the two Fitzwilliam pieces), although both the Amathous and the Murano rod have no applied decoration. Concerning the shape, it is clear that the Amathous piece with one looped end and a flattened extremity is the most common type of the stirring rods (Isings Form 79). Nevertheless, a wide variety is apparent. Some are discoid flattened on either extremity (Calvi 1968, no.124, pl.2:4; Hentrich, von Saldern 1974, 214, no. 620; Auth 1976, 229, no. 522) or have a loop combined with a vase (Spaer 2001, 264, no.631) or with a bird (Constable-Maxwell 1979, 116-117, nos. 199; 202). The Murano rod, on the other hand, has two pointed extremities and the Aquileia rod has one pointed end and a spatula-shaped bended end. The lengths of these rods demonstrate range widely, between 110 and 260 mm, with the average around 190 mm. The Amathous rod is a small example, with a length of 112 mm and a shaft diameter of 5,5 mm (Table 49). The discoid end has a diameter of 10 mm and the ring is 25 mm in diameter. Table 49: List of so-called stirring rods (dimensions in mm) cat.no. site hue 776 Amathous (CY) so-called black not catalogued Cyprus blue not catalogued Cyprus blue 771 Zara (HR) purple 2725 Aquileia (IT) so-called black 3104 3.4.3 Torre de Ares (PT) blue length 112 193 162 210 169 reference Tytgat 1992, 85, no.692, pl.XIX Glass at Fitzwilliam 1978, 50, no.96 Glass at Fitzwilliam 1978, 50, no.96 Ravagnan 1994, 119, no.222 Calvi 1968, no.122, pl.2:3 http://www.matriznet.imcip.pt/ipm/MWBINT/MWBINT00.asp Bichrome plaques Likewise the twisted rods large museum collections host bichrome plaques combining various translucent glass hues with opaque white glass. The large set of architectural decoration material in KMKG-MRAH R1609 collected by Emile de Meester de Ravestein during his diplomatic mission at the Vatican in de 1850‘s. Similar material is available in most large museum collections. Likewise the twisted rods the bichrome plaques are generally in a blue glass combining opaque white. None has been made in colourless glass or is combining other than opaque white. The other glass colours are dark green, dark brown and black. The bichrome plaques are sometimes grooved lengthwise and show quadrangular patterning of the opaque white ribbons, generally in the middle, but sometimes in the corners (Figure 86).The bichrome plaques are much larger conserved with a length ranging between 20-135 mm, and likewise the twisted rods most are below 60 mm. The rectangular plaques have a width that ranges between 9-20 mm and a height that ranges between 2,4-5,2 mm. A more detailed study of these artefact types would provide in a certain patterning and a categories to better evaluate these objects. 130 Figure 86: An overview of the R1609 rectangular bichrome plaques from the KMKG-MRAH, Brussels (photograph by the author, courtesy by KMKG-MRAH, Brussels) 3.4.4 Inlays or intarsia We cannot say much on this material as none have been studied in detail. A very promising position holds the 111 material from the villa of Lucius Verus that is since recently under study. Another study has been reported on opus sectile material from Ostia (IT) (Verità et al. 2008). Most interesting concerning the inlays in black glass is the material from Porta Marina where mosaics in opus sectile have been excavated using black (2 ex.), white and yellow monochrome pieces. Important to this study is that the analytical results demonstrate that different techniques of coloration have been used pointing to different provenance of the glasses and accordingly produced in various glassmaking centres. 3.5 Counters Counters have been made in various materialsbone, ivory, stone, ceramics, amberand sometimes even 112 combining different materials. Most emblematic for the glass counters is the discoid shape with a plano113 convex section and rounded edge. Most glass counters average between 15 and 22 mm in diameter, but the 114 size can range from below 10 mm to over 40 mm in diameter. The height of the counters is quite standard no matter what the size is, ranging essentially between 6 and 8 mm but counters with a height less than 5 mm or up to 10 mm do occur (Harden 1936, 292-294; Price 1995, 129). Besides the typical round, counters a fairly number have an elongated to ovoid shape, while quadrangular and irregular shaped counters are rather rare. The flat bottom surface can be smooth or rough. Polychrome glass counters do occur, although they are mainly made in a monochrome glass hue, undecorated or decorated with dots in one or more contrasting colours. Despite the various publications discussing glass counters as regular finds when discussing the small finds of an excavation report, no systematic research has been carried out so far to investigate in detail the possible significance of the variations in size, shape, colour, and production technique for a chronological and geographical consumption patterning. Regularly has been referred to ‗type 9‘ in Véronique Arveiller-Dulong‘s work on the Roman glass in the collections of the Musée Archéologique of Strasburg (FR) (Arveiller-Dulong, Arveiller 1985, 57-58) to nominate typologically the counters (Simon-Hiernard 2000, 387-388). But this is based on a misapprehension since the publication of this rich glass collection consists of a subdivision into chapters per functional type due to the limited space at her 111 First results were presented at the 18th congress of the AIHV in Thessaloniki 2009 mentioning a quantity of inlay plaques in monochrome black glass: L. Saguì, P. Santopadre, M. Verità, Glass technology, colours, forms and shaping in the 2nd century opus sectile glass materials from the villa of Lucius Verus in Rome. 112 Six counters from tomb 100 of the late Roman cemetery of the castellum at Oudenburg combine glass with bone. Two glass layersa transparent blue-green glass layer on top of a white opaque oneare superimposed upon a base of bone (Mertens, Van Impe 1971, 72, nos. 2-9, pl. XI, 2-9). 113 Counters different from the classic plano-convex shapefor instance, the conically shaped onesoccur rarely and are largely limited to materials other than glass. Furthermore, the flat-topped conical counter in black glass topped with whitish glass, that has been reported from the late Roman castellum at Liberchies (BE) (Mertens, Brulet 1974, 86, no. 10, fig. 35:10), could well be a gem normally set in a discoid brooch [see 3.3.5. Gems]. A similar piececombining blue and whitecomes from tomb 781 of the late Roman cemetery of Krefeld-Gellep (D) (Pirling 1966, 97, no.5, pl.67:37). 114 The counter from the villa of Valkenburg (NL) shows extreme dimensions, with a diameter of 41,5 mm and a height of 10,6 mm [cat.no.2992]. 131 disposal. Where one chapter deals with all window panes, another chapter considers all glass counters, whatever their size, shape, production technique or colour. This approach, by neglecting chronological and/or geographical differences, does not provide a useful classification for archaeologists. Despite the detailed specifications on the Strasburg material, it is meaningless to refer to it as a typology but was also not meant to be one. Ironically enough, author herself mentions the lack of a classic typology for this type of glass artefact (Arveiller-Dulong, Arveiller 1985, 57). At the same time, a paper on the glass from the Roman fort of Vindolanda (UK) was published by Jennifer Price (Price 1985, 206-214). She discussed the glass counters by dividing them into two categories: the small planoconvex pieces with a diameter between 10 and 20 mm and the large ones with a diameter between 27 and 32 mm. Difficulty only starts when counters are recorded from other sites which are too large to be attached to the small counters and not large enough to be classified under the large counters. We therefore thought useful to introduce the term medium-sized counters in our catalogue for those with a diameter between 20 and 27 mm. Furthermore, we found it more appropriate to use only the term counter, no matter what size the piece has, so as not to confuse them with these plano-convex pieces [see Chapter 7]. The range of factors makes it challenging to create a taxonomy on glass counters (Figure 87), especially when 115 trying to create a chrono-typology and in search of possible regional variations. Size: Small Medium large Colour: Monochrome Bichrome Polychrome Decoration: Undecorated Dotted Stripes Production technique: Fused Poured/cast Planoconvex shape Shape: Round Oval Quadrangular Irregular Figure 87: Schematic view on the various features used to classify Roman glass counters 115 Such an endeavour necessitates, in fact, a more in-depth study including all glass counters. Due to the topic of the research, we limited ourselves to look at the monochrome black counters and excluded from this evaluation all other monochrome as well as the polychrome counters. Consequently, colour is here not used as a classification element. The late Roman counters in opaque white, turquoise and black glass with a dotted decoration on the top surface have not been catalogued as polychrome counters, but as monochrome counters with a contrasting monochrome or bichrome decoration. 132 We therefore subdivided the plano-convex counters in monochrome black glass depending on:  116 size :     small= d.: below 20 mm; medium = d.: 20 - 27 mm; large = d.: above 27 mm. shape:      round = when perpendicular measurements show a discrepancy of < 3 mm oval = when perpendicular measurements show a discrepancy of ≥ 3 mm quadrangular = square with rounded corners irregular = when the circumference is asymmetrical production technique    117 [see Chapter 2]: fusing of glass sherds as described by Pliny the Elder = flat and smooth basal surface by pouring a drop of glass on a marver = irregularly pitted and rough basal surface. 118 ; decoration:   undecorated decorated - monochrome dotted (opaque white, opaque red or translucent blue) bichrome dotted (opaque red and translucent blue) heart-shaped The monochrome counters can be small, medium and large in size, and circular, oval, quadrangular and irregular in shape with normally a rounded edge, although pieces with a sharp or grounded edge can occur. Looking at the dimensions of all recorded black glass counters, a correlation is observable between the diameter and the height (Figure 88). The larger the diameter is, the greater the height. ROMAN BLACK GLASS COUNTERS 11 10 9 8 HEIGHT 7 6 5 4 5 10 15 20 25 30 35 40 45 DIAMETER Figure 88: Plot of the dimensions of Roman black glass counters (in mm) 116 In the database we named the counters with a diameter below 12 mm as very small and those with a diameter above 35 mm as very large. This feature could not be verified systematically in the database because we based our information on mostly publications. From the 710 pieces inventoried [see below] we determined 60 counters with a rough basal surface (cast) and 23 with a flat, smooth surface (fused), correspondingly 72,3% and 27,7%. Because the 83 pieces that could be determined only represent 11,7% of the total amount of the catalogued black glass counters no significant conclusions can be proposed from this feature. 118 We sampled some black glass counters for chemical analysis [see Chapter 8], but an extensive sampling is recommended to verify to what extend the fused counters have been made from one piece or from a mix of glass scrap (e.g. the counters from Nijmegen (NL) and Oudenburg (BE) (see Chapter 1; Figure 3a-b) [cat.nos.4466 and 465], and whether the Roman glassworker made use of production waste or only recycled old glass. 117 133 ROMAN BLACK GLASS COUNTERS 11 10 9 8 HEIGHT plain counters 7 decorated counters 6 5 4 5 10 15 20 25 30 35 40 45 DIAMETER Figure 89: Plot opposing the dimensions of the plain against the decorated counters in black glass (in mm) When isolating the decorated counters in the plot, a slight but clear difference can be noticed between the plain counters and the decorated ones (Figure 89). The fact that the decorated counters are shorter yet larger in diameter is almost certainly due to the production process these counters underwent. They most likely had similar dimensions as the plain ones initially, since they were produced in the same way. But the coloured glass drops, applied as decoration, had to be marvered into the core glass, making the hot glass more flat and accordingly also larger. The inventory of the black glass counters provided clear data making it possible to formulate conclusions regarding their shape and size. Unfortunately, there has been no opportunity to verify the production technique systematically, so we cannot make any definite conclusions. We observed, however, that most large counters have a flat and smooth basal surface, implying a fusing technique, but this has been noticed on small and medium counters as well. The majority of all black glass counters is round, while only a few are oval, quadrangular or irregular (Table 50). From the 710 pieces within 574 entries 103 entries or 141 pieces could not be inventoried correctly within the present research. From the 569 pieces that could be catalogued in 420 entries 518 ex. or 91 % is round (29 ex. 119 large sized; 119 ex. medium sized; 327 ex. small sized; 11 ex. very small sized ); 39 ex. or 6.7 % is oval (6 ex. medium sized; 33 ex. small sized); 3 ex. or 0,5 % is quadrangular; 9 ex. or 1,6 % is irregular. When looking to the size we observe that 2/3 of all recorded black glass counters is small sized and 1/4 is medium sized. Large and very small counters represent only a minor segment (Table 50). Finally, checking the decoration, we observed that from the 710 black glass counters 635 ex. or 89,4 % is undecorated and that 75 ex. or 10,6 % is decorated with dots in one or two contrasting colours, i.e. opaque white, red, and turquoise or deep blue (Table 51). While the majority of the undecorated black glass counters are small sized and in a lesser degree medium sized, the opposite ratio can be observed for the decorated black glass counters (Table 51). Different clusters appear when plotting the dimensions of the black glass counters (Figure 89). We discuss in the next chapters the possible connections with chronological periods and regional areas [see Chapters 4 and 6]. 119 It was not verifiable how many of the 23 counters from the Corbridge Hoard (UK) and the 9 counters from tomb 6352 of the cemetery at Krefeld-Gellep (D) are small and medium sized. 134 Table 50: List of black glass counters per size and shape (* see note 81) size round oval quadrangular irregular large medium small very small TOTAL 29 119* 327* 11 486* (518) 0 6 33 0 39 0 0 3 0 3 1 1 7 0 9 total % 30 126* 370* 11 537* (569) 5,6 23,5 68,9 2,0 100 Table 51: List of monochrome black glass counters plain vs dotted decoration (* see note 81) size undecorated decorated total % large medium small very small indefinite TOTAL % 30 74* 350* 11 138 603 88,9 0 52 20 0 3 75 11,1 30 126* 370* 11 141 678 100 4,4 18,6 54,5 1,6 20,8 99,9 Aware of considering only the counters in black glass we observed that the dotted counters are nearly always medium-sized with a diameter that never exceeds 23 mm but some examples are small-sized as for instance some pieces from Amiens [cat.nos.1001-1002] (Table 52). 120 121 Counters with dotted decoration are only known in black and white glass. The observed decoration demonstrates a common pattern with a central dot encircled by four equally sized dots applied crosswise. Irregularities in design, however, do occur, even within one set: 1) 2) 3) 4) 5) 6) 7) 8) only a central dot (Lankhills: 1 black 50f); with 2 dots on either side of the central one (Lankhills: 1 black 50g); with only 3 dots surrounding the central one (Nijmegen: 1 black, 2 white; Amiens: 1 white; Lankhills 50h); with 5 dots surrounding the central one (Lankhills: 1 white 50c); with 6 dots surrounding the central one (Lankhills: 1 black 50l); with 8 small dots around a large central dot (Amiens: 1 white); with a second red dot (Nijmegen: 1 black); with unbalanced pattern (Nijmegen: 1 black). The abovementioned list demonstrates that one set of counters is sometimes very heterogeneous, as in the set of 26 glass counters from tomb 51 of the late Roman cemetery of Lankhills, Winchester (UK) (Clarke 1979, 251-254, fig.69:50). The decoration consists of dots in one or two different colours, generally in opaque red and translucent blue glass, but a pattern with opaque white dots is known too. Usually the central dot in opaque red glass is surrounded by dots in translucent or opaque blue glass, but sometimes the opposite is attested (Lankhills) or all the dots are in opaque red glass (Amiens). Interestingly, all ―white‖ counters from Amiens (FR) and some black ones are decorated with only red dots, whereas only one counter of the 14 preserved white ones from Nijmegen has only red dots. It appears that dotted counters have so far nearly never been recorded with solely opaque blue glass decoration; a single example is known from tomb 51 in Lankhills (UK). Another striking feature is that the red glass is the same everywhere, whereas two different blue hues were used depending on the colour of the core (Figure 90). Black counters have opaque ultramarine blue dots, while the white ones have translucent turquoise blue dots. Both types can occur within the same set. Only rarely are ultramarine dots attested on white counterse.g., one of the 15 pieces from tomb 49 at the Hunerberg cemetery of Nijmegen (NL) (van Enckevort 2008). The opaque red spots on the counters from the Roman fortress at Usk (UK) have to be considered as unintentionally, but rather due to circumstances during the production process as Jennifer Price already assumed (Price 1995, 129). 120 Also the Lullingstone counters are included, even though reported to be of brown glass. It is thus not excluded that these counters are not at all black in opposition to all other dark-coloured dotted counters. 121 Besides the counters in bright white glass, we consider the turquoise-tinged counters also as opaque glass. 135 Table 52: The dimensions of two sets of decorated gaming counters Amiens (FR) – rue Just Haüy, tomb 4 Nijmegen (NL) – tomb 39 black ―white‖ black ―white‖ d. h. d. h. d. h. d. 21 6 19-20 5 20,8-21,7 6,4 21,8 20-21 6 19-21 5 21,8-22,3 6,4 21,5 20 5,5 19 5 17,8-19,1 6,6 20,8-22,6 19 5 21 6 20,7 6,2 20 6 18-21 6 21,5-22 6,2 20,5-21,3 20 5 21 7 21,5-22,3 6,1 20,3 20 5 19,5-21 6 17,5-17,8 5,8 18,9-20 18 6 21 6 17,6-18 5,8 23 20 5,5 18-19 6,5 21,5-22,8 5,9 19 6 19 6 20,6-21,1 6,5 22 19 5 17-19 5,5 21,1-21,7 6,2 21,7 20 5 20-21 6 21-21,2 6,1 21 17-20 6 20-22 6 21,3-21,7 6,5 19 15 6 21 5 20,8-22,7 6,1 17,1 16 6 21,2-21,8 6,0 17-18 6 12 6 h. 6,3 6 6,7 6, 9 5,4 5,8 6,2 6,1 6,4 7 6,2 6 5,9 6,5 Figure 90: Variety of blue glass used to apply a dotted decoration on counters 136 Chapter 4 CHRONOLOGY 4.1 Introduction The black glass material has been categorised in the previous chapter within a roughly outlined chronological subdivision of the Roman imperial period to facilitate the analysis of the wide variety of material covering a period of five centuries (Table 53, left column). This rough phasing into four wide-ranging periods of 100-150 years is based on the prevailing general chronologies from similar material in other glass hues. The purpose of this chapter is to refine the previously defined phases into a more elaborate timetable (Table 53, right); this will be accomplished by comparing the artefact types (sometimes wrongly or, in most cases, very generally dated) with the different artefact types from dated contexts. Additionally, we will verify in this chapter whether the obtained information on black glass artefacts from dated contexts can add value to the chronology of Roman glass in general. The primary and initially most useful result from thorough material studies is the supply of good dating material for archaeological research. However, a good dating of specific commodity types is based on the evaluation of dated finds from archaeological contexts. Hence, this chapter only verifies the availability of well-defined strata and contexts as tools to refine the dating of the various types of glass artefacts and establishes typo-chronologies of specific material categories to help better determine the successive layers of an archaeological site with the aim of understanding its development. Table 53: detailed timetable of the Roman imperial period EARLY EMPIRE 10 BC–70 AD 10 BC–35 AD 35–70 AD Flavian–early Antonine period 70–150 AD 70–100 AD 100–150 AD late Antonine–Severan period 150–230 AD 150-190 AD 190-230 AD ‗Crisis‘ or intermediate period 230–280 AD 230-250 AD 250-280/5 AD Julio–Claudian period PERIOD I (I–IIA AD) PERIOD II (IIB–IIId AD) PERIOD III (IIId–IVA AD) PERIOD IV (IVB–V AD) Tetrarch–Constantinian period ‗Decline‘ period LATE EMPIRE 280/5–350/60 AD 280/5–310 AD 310–350/60 AD 350/60-480 AD 350/60–410 AD 410–480 AD AugustanTiberian period ClaudianNeronian period Flavian period early Antonine period late Antonine period Severan period ‗anarchy‘ period ‗regional power‘ period Tetrarchs Constantinian Renaissance ‗challenging‘ period ‗collapse‘ period To keep this chapter comprehensible, each functional type is first discussed separately in chronological order. The availability of the necessary dated contexts is verified to obtain a more precise dating of every class of black glass artefacts. Conversely, a chronological overview of black glass production and consumptionconsidering all artefact types throughout the Roman imperial periodis given in the interim conclusion of this chapter. Prior to summarising the data, we have to realise the rather limited availability of well-dated contexts. 4.2 Production 4.2.1 Primary production We have to remain brief on this topic because there is, hitherto, no archaeological evidence available indicating production of black glass in primary glass workshops during the Roman imperial period. This is not at all 137 122 surprising, given the nearly total absence of primary glass workshops dating within the first five centuries AD. The total absence of deeply coloured glass at all known primary glass workshops makes us believe that in ancient times these production sites only produced ‗naturally coloured‘ and/or decolourised raw glass. This view corroborates the, thus far, first and sole clear-cut archaeological evidence of primary glass production in Roman imperial times, that of Beni Salama at Wadi el-Natrun (EG), where ‗naturally coloured‘ as well as decolourised raw 123 glass was produced. Based on the pottery, the activity period of this large workshop has been dated in the st nd 1 2 centuries AD (Nenna 2008a, 127; Nenna 2008b) [see Chapter 6]. Most likely, specialised intermediate workshops and/or secondary workshops were responsible for the production of black glass [see Chapters 6 and 8]. This assumption corroborates the present-day knowledge on shiploads retrieved from shipwrecks all over the Mediterranean containing blocks of raw glass to supply the secondary workshops, which never yielded pieces of 124 black glass. For that reason, we are not able to establish the origin of the two rough blocks of black glass found st at the 1 century AD secondary workshop at ‗La Montée de la Butte‘ in Lyon (FR). It is not excluded that both pieces came from a distant primary workshop in the south-eastern Mediterranean. Hereafter and in Chapter 8 we will discuss these two blocks in more detail. 4.2.2 4.2.2.1 Secondary production Period I To this point, no known Period I glass workshop has yielded direct evidence for the production of black glass artefacts. Only indirect evidence of some black glass production is provided by two secondary glass workshops: 125 126 ‗Derrière la Tour‘ at Avenches (CH) and ‗La Montée de la Butte‘ at Lyon (FR) , dated in the mid 1st century AD (40-70 AD) and the mid 1stearly 2nd century AD (40-90/110 AD), respectively. The Avenches workshop yielded a large amount of production waste showing the production of various free-blown vessel shapes in deeply coloured glass sometimes with black appearance (mainly purple) (Amrein 2001), but no vessels in blackappearing glass as such have been reported to our knowledge. Also, the Lyon workshop delivered no more than indirect evidence. The use of black glass is only supported through the two chunks of black glass without any 127 relationship to particular form types in black glass (Robin 2008). This lack of evidence makes it impossible to determine what has been produced in black glass, even if looking at what the Lyon workshop has produced in other glass huese.g. stirring rods, unguentariaengenders an idea of what might have been produced in black glass. It is clear that both workshops produced glass artefacts in various glass hues and appear to have 128 specialised in the production of vessels. 4.2.2.2 Period II Two secondary glass workshops were excavated in 1974-1978 in Regio 17B/C at ‗Äussere Reben‘ Kaiseraugst (CH) (Rütti 1991, 151) that have been recently the subject of an in-depth study (Fischer 2009) [see Chapter 6]. Only the workshop of Regio 17B, south-west of the road junction, gives clear information on the setting up of a workshop where black glass was used to produce vessel shapes (Fischer 2009). The urban workshop consists of one large square room wherein the structures of 14 successive furnaces have been identified (Fischer 2009, 4770). Based on the pottery as well as some other archaeological small finds, the workshop must have been in use nd rd during the 2 early 3 century AD, broadly dated, covering an activity period from 80/90 to c. 250 AD but with nd rd some reservation considered active from the second quarter of the 2 to the early 3 century ADi.e. c. 125210/220 AD (Fischer 2009, 27). Interesting here is that the Regio 17C workshop at the other side of the road is considered to have been active from c. 220 to 260/270 AD and did not yield any black glass. This absence of black glass in the Regio 17C workshop corroborates the late Antonine and Severan date of the Period II black 122 Most evidence has come from the early Byzantineearly Islamic period in the Levant (late 4th9th century AD) (Gorin-Rosen 2000; Jennings et al. 2001) 123 The site, still under excavation, is part of a larger ongoing research project directed by Marie-Dominique Nenna to determine the primary glass production in the western Nile Delta during the Roman imperial period (Nenna 2007, 127-130) 124 Other deeply coloured chunks of raw glass have been retrieved, such as lumps of dark blue glass from the shipwreck Sanguinaires A (FR) dating from the early Hellenistic period (Alfonsi, Gandolfo 1997; Foy, Nenna 2001, 24, fig.3) and the block of deep bluish glass from the 1st2nd century AD site at ‗Place Jules-Verne‘ in Marseille (FR) (Foy, Nenna 2001, 25, fig.6) which show that in Hellenistic and early imperial times, deeply-coloured chunks of raw glass were commodities of long-distance trade; likewise, the ‗naturally coloured‘ and decolourised glass. 125 The Avenches workshop was excavated in 19891990 and the structures as well as the material received a meticulous assessment by Heidi Amrein (2001). 126 The Lyon workshop was excavated in 2000-2001 (Motte, Martin 2003) and was recently re-evaluated by Laudine Robin (2008). 127 We thank Laudine Robin for drawing our attention to these blocks, but also by supplying us a sample of both blocks for chemical analysis, which generated striking results [see Chapter 8]. 128 The chemical analysis opens a more interesting prospect to the discussion on the type of glass workshop and what was produced seeing that both greyish-brown blocks give a composition that resembles that of obsidian [see Chapter 8]. 138 glass vessels (Cosyns, Hanut 2005) [see this chapter, 4.3.2. Vessels Period II]. Did the glass workshop of Regio 17B shift to the other side of the road for a new start? The black glass material from the Regio 17B workshop mainly consists of production waste, formless bits and pieces of batch and small chunks of raw glass. A part of the assemblage produced in the Regio 17B workshop 129 can be determined on the basis of artefact deposits from the last (two) generation(s) of active glassworkers. 130 The workshop produced at least aryballoi (Form IIB.20) [cat.no.3778] ; cylindrical cups with base-ring and 131 applied glass trail below the rim (Form IIB.17) [cat.nos.3755-3763] ; and an indeterminate form type with narrow 132 folded foot and concave kick on the bottom (Form IIB.21). In subsequent chapters, we will verify on the basis of distribution patterning [Chapter 6] and chemical analysis [Chapter 8] what other black glass artefacts have been produced in Augusta Raurica and what might be attributed to the Regio 17B workshop. 4.2.2.3 Period III Various workshops active in this period that manufactured black glass artefacts have been located within the Argonne region in the north of France: Les Houis nearby Sainte Menehould; Lavoye-La Clairière; Froidosth th Berthancourt [see Chapter 6]. These workshops are generally dated in the 4 5 century AD with a production th heyday assumed in the 4 century AD and are thus also active in Period IV. This dating is mainly based on the rd pottery determination by George Chenet (1925). The period of activity has since then been extended to the 3 century AD (Chew 1989, 57; Foy, Sennequier 1991, 57; Gazenbeek, Van der Leeuw 2003, 291). Unfortunately, the sites are not well-excavated, nor has the archaeological material been intensively studied or published. An adjustment is, however, to be considered on the basis of the collected black glass artefacts that have been produced in these Argonne workshops. More or less similar material, the production of these secondary glass rd workshops in the Argonne must have started somewhere in the 3 century AD and most likely already somewhere in the first half of the century, seeing the Period II-vessel material such as that of Form IIB.1 [see this chapter, 4.3. Vessels and 4.4.1. Bracelets; and Chapter 6]. The vessel fragments can be considered waste material, but they were also collected to be used as cullet (Chew 1989, 57). In any case, the recycling of broken glass implies the accumulation of contemporaneously prevailing artefacts. Typically done with this sort of material it is anymore successful when material ran out of production and turned into a rather marginal consumption item or heirloom. Only intensive field work combined with an exhaustive study of all the archaeological artefacts within the different museums as well as the archive material collected by Dr. Meunier and his son-in-law George Chenet and the work from some other contemporaneous workers (Colson 1903; Barthélemy 1904; Mauget 1903; 1904; 1907; Strohm 1920) will entail the opportunity to acquire a more comprehensive image of the Argonne workshops. It appears that the three Argonne workshops at least produced jewellery and tesserae in black glass. Concerning the jewellery, we have already observed that there is a preponderance of bracelets and that finger rings and hairpins have also been produced, whereas beads and pendants are completely absent (Cosyns 2009). The rd th bracelet types are mainly 3 early 4 century AD types A, B and C (A1-6; B1-2; 4 and C1;3), whereas only 2 th th pieces of the later 4 early 5 century AD type D (D1) have been catalogued. The finger ring types demonstrate rd th th th a similar pattern with only the 3 early 4 century AD types A and B, whereas the later 4 early 5 century AD type C is not reported in the Argonne workshops. Neither are the typical Period IV beads and pendants as present in the workshop of Palais Kesselstatt at Trier (DE) (Cosyns 2009). The black glass material from the Argonne workshops demonstrates that the study of this idiosyncratic material provides more detailed dating evidence to define the active period of the Argonne workshops more accurately than the dated archaeological contexts and should offer a clearer view on the date of the artefacts manufactured in these workshops. 4.2.2.4 Period IV The glass workshop from Palais Kesselstatt in Trier (DE) is dated in the last third of the 4th century AD (GoethertPolaschek 1984, 165-166, no. 64). Various types of jewellery have been produced, but in contrast to the Argonne workshops we recorded a preponderance of finger rings (type C) and the presence of beads and pendants, 129 This hypothetical assertion proposes that the black glass production in Augusta Raurica can be dated between 190/200 and 210/220 AD (one generation) or between 170/180 and 210/220 AD (two generations). 130 Two other fragments of black glass aryballoi were excavated elsewhere in Augst/Kaiseraugst [cat.nos.3553; 3753]. 131 Six other fragments of these cylindrical cups come from various areas in Augst/Kaiseraugst [cat.nos.3490-91; 3495; 3504; 3743; 3773] and three are reported from Avenches [cat.nos.3594; 3600; 3603]. 132 Because the workshop was a vessel-producing centre, we may assume that the presence of the single elongated cylindrical bead [cat.no.3642] was coincidentally lost by the wearer and does not constitute evidence for bead making. It is generally accepted that the different glass productionsraw glass, vessels, window panes and jewellerywere executed in different workshops because each production necessitated different installations, equipment and skills. 139 whereas the bracelets are attested by only the Period IV type D1 (Cosyns 2009). The beads (in particular, annular, globular and cylindrical beads) and the pendants (in particular, vessel-shaped and cylindrical) show decorations with applied trails in contrasting opaque colours such as white, yellow, red and turquoise blue in zigzag patterns, th th characteristic from the later 4 century AD but lasting to about the mid-6 century AD. The Trier workshop and the abundance of polychrome black glass beads and pendants demonstrate that their production was done in Roman glass workshops, by Roman glass workers and for Roman customers, and not in Barbaricum or by the Migration People [see Chapter 6]. 4.2.3 Recycling No direct archaeological evidence of recycling black glass have been recorded so far; consequently, dated contexts are excluded from consideration here. On the other hand, Pliny the Elder describes the production of the so-called oculi or calculi (Pliny the Elder, Naturalis Historia, XXXVI, 199) [see Chapter 2], and from chemical analyses there is good evidence indicating the recycling of glass in the production of black glass artefacts. 4.3 Vessels The chronology of the vessels is described according to the typological subdivisions in four consecutive periods (I-IV) and per technologye.g. cast (A), free-blown (B), mould-blown (C), and rod-formed (D). Because of the wide variety of vessel shapes, we thought it useful to convert this initial classification into a more functional arrangementi.e. tableware, toilet ware and storage ware (Tables 54-55). The next step was to verify the accumulated material from dated contexts with the general dates of identical material in other glass hues given in 133 the prevalent publications (Table 56). Table 54: Number of shapes per functional types per period Period I Period II Period III Period IV tableware 24 16 2 1 toilet ware 4 4 2 2 storage ware 4 2 total 40 25 4 3 total 41 9 6 56 Table 55: Number of shapes per functional type per period and per technique (A= cast; B= free-blown; C= mould-blown; D= rodformed) Period I Period II Period III Period IV total technique A B C D A B C D A B C D A B C D tableware 11 10 3 15 2 2 1 38 toilet ware 1 2 1 1 2 1 2 2 8 storage ware 1 3 1 1 6 total 13 15 4 17 5 1 4 1 2 52 Table 56: List of entries per applied technique per period Period I Period II cast 80 (97) free-blown 49 (61) 274 (275) mould-blown 5 16 rod-formed 4 indefinite 14 7 total 148 (177) 302 (303) Period III 4 4 Period IV 1 43 44 indefinite 7 104 (106) 3 1 42 (43) 157 (160) total 87 (104) 424 (439) 31 48 63 (64) 651 (684) These tables show a predominant presence of black glass vessels in Periods I and II and a near- absence of shapes overall in the two succeeding periods. Black glass toilet ware and storage ware do not show the great diversity in shapes as is available in tableware. This is possibly due to the fact that toilet ware and storage ware were made as containers for specific consumables, whereas the tableware demonstrates the individuality of the 133 We based the chronological mainframe of the here-discussed form types in black glass on what has been presented in Isings 1957; GoethertPolaschek 1977; Rütti 1991; Cool, Price 1995; Bonnet-Borel 1997; Price, Cottam 1998; Martin Pruvot 1999; Arveiller-Dulong, Nenna 2005. 140 consumer. The number of shapes per function in Table 54 clearly shows that black glass tableware can be considered commonly produced consumption goods in Periods I and II. Conversely, Table 56 demonstrates that the true consumption of black glass vessels is to be considered much shorter during both defined periods of 150 years. It seems more likely that the success of black glass vessels was a result of a trend that lasted for a period of somewhat less than 50 years, roughly equivalent to two generations of glassworkers/consumers. It would be not at all surprising if rod-formed toilet ware of Period IV from well-dated finds would generate a the also lasted for about 50 years. Table 56 displays that about one-third of all recorded vessels have remained indefinite. This has two reasons: the high degree of fragmentation or the lack of clear information available/obtained on the fragment. IV III PERIOD II PERIOD I Table 57: Chronological overview of all vessel types produced in black-appearing glass [see Chapter 3] PERIOD I PERIOD II PERIOD III PERIOD IV SHAPE 0 50 100 150 200 250 300 350 400 450 FORM IA.1 FORM IA.2 FORM IA.3 FORM IA.4 FORM IA.5 FORM IA.6 FORM IA.7 FORM IA.8 FORM IA.9 FORM IA.10 FORM IA.11 FORM IA.12 FORM IA.13 FORM IA.14 FORM IB.1 FORM IB.2 FORM IB.3 FORM IB.4 FORM IB.5 FORM IB.6 FORM IB.7 FORM IB.8 FORM IB.9 FORM IB.10 FORM IB.11 FORM IB.12 FORM IB.13 FORM IC.1 FORM IC.2 FORM IC.3 FORM IC.4 FORM IIB.1 FORM IIB.2 FORM IIB.3 FORM IIB.4 FORM IIB.5 FORM IIB.6 FORM IIB.7 FORM IIB.8 FORM IIB.9 FORM IIB.10 FORM IIB.11 FORM IIB.12 FORM IIB.13 FORM IIB.14 FORM IIB.15 FORM IIB.16 FORM IIB.17 FORM IIB.18 FORM IIB.19 FORM IIC.1 FORM IIC.2 FORM IID.1 FORM IID.2 FORM IIIB.1 FORM IIIB.2 FORM IVD.1 FORM IVD.2 From a technological point of view, we may conclude that the vessels cast in black-appearing glass are characteristic of Period I. Those produced by the mould-blown technique were produced in Period I and II but remained uncommon, while it was a very popular technique to manufacture vessel shapes in ‗naturally coloured‘ 141 and decolourised glass. Rod-formed vessels have been produced from Period II, but they are idiosyncratic to 134 Period IV and can easily be distinguished from the Period II rod-formed unguentaria. The free-blown vessels are long-lasting but appear mainly in Periods I and II (Tables 56-57). 4.3.1 Period I Based on prior investigations (Grose 1989; 1991; Stern, Schlick-Nolte 1994; Cool, Price 1995), it has been demonstrated reliable to estimate the chronology of Roman glass vessels to a certain degree by looking at its colour. ‗Naturally coloured‘ blue-green glass is clearly attributable to the first three centuries AD, gradually making rd room during the later 3 century AD for the greenish and yellowish-greenish ‗naturally coloured‘ glasses, so th th characteristic for the 4 and early 5 century AD. Finds from dated contexts can be very helpful to refine more the date of specific vessel shapes. Even though it is not specifically due to the presence of black glass vessels, the dated contexts refine the dating of, for instance, the rectangular tray, Form IA.8. Not merely of Julio-Claudian date (Stern, Schlick-Nolte 1994, 326, no.98), good st examples clearly indicate this particular shape to the middle of the 1 century AD. From Fréjus (FR) are st fragments dated in the first half of the 1 century AD (Price 1985, 68, note 23; Price, Cottam 2009); a deposit at st Pasaje Cobos (ES) dated mid-1 century AD (Price 2004); and deposit VII at Cosa (IT) dated 50-55 AD (Grose st 1991, 10, Cosa no.6). Cast vessels in black glass seem thus characteristic of the first half of the 1 century AD. But dated contexts can be misleading; the wide-rimmed bowl/plate with a base-ring, Form IA.10A (Figure 33.10) piece from Avenches (CH) came from a dated context ranging between 150-300 AD, but these bowls are in st rd general dated from the last quarter of the 1 century AD up to the second quarter of the 3 century AD (Bonnet Borel 1997, 23). The Egyptian piece in the Petrie Museum of Egyptian Archaeology, considered a variant but close to AR 24.1 by Beat Rütti (1991) and AV V32 (Bonnet Borel 1997), has received roughly the same date range, correspondingly 70/100-250 AD and 150-250 AD. We, however, assume this vessel shape to be a Flavian or Antonine production, which we will prove by means of the results from chemical analysis [see Chapter 8]. A chemical analysis is also necessary to categorise the cylindrical cup with everted rim, Form IC.1 (Figure 41.1), as a Period I-shape, seeing the very wide-ranging chronology going from Neronian times up to the third quarter of rd the 3 century AD (1991, I, 44; II, 55-56, nos.1148-1177; Bonnet Borel 1997, 29). Although the context of the cup in the Provinciaal Gallo-Romeins Museum Tongeren remains unknown, this supposedly local discovery from in or st around Tongeren (BE) was dated on a stylistic basis by Michel Vanderhoeven in the 1 century AD (Vanderhoeven 1962, 23, no.15). The Avenches cup in black glass came from a cremation tomb dated in the third nd quarter of the 2 century, c. 160-180 AD (Bonnet Borel 1997, 29, no. 47; Martin Pruvot 1999, 188, no. 1181). A nd production around the middle of the 2 century AD is therefore not excluded, but it leaves open for speculation a correct date. Many archaeological sites in the Gulf of Naples are repeatedly used as dated contexts because the Vesuvius eruption in 79 AD made the area a gigantic closed context. Also, several black glass vessel shapes have come from, for instance, Pompeii and Boscoreale, which predate 79 AD, such as a baseless, shallow, handled cup, Isings Form 25, with the inside decorated with inlay in gold leaf, silver, bronze and cornelian found in Pompeii (Isings 1957, 39-40; Isings 1964, 27). Unfortunately, such contexts only give an idea how early certain shapes were already in use, as it is quite possible that the shape was still produced (long) after 79 AD. On the other hand, the absence of certain shapes does not prove that the shape was only produced after 79 AD. The hemispherical cup with cut rim and abraded concentric lines, Form IB.1, is definitely the morphological blown equivalent of the shallower cast Form I.1a. Both shapes appear more or less in contemporaneous dated contexts st within the period in which deeply coloured monochrome vessels were produced, i.e. roughly the mid-1 century AD including the second and third quarters or 30/40-70 AD. Nonetheless, we are persuaded that we have to consider the Hofheim cup as a later production evolved from the cast hemispherical cups. The much faster blowing technique had been introduced for the production of these shallow cups as a reaction to the growing consumer market. This necessitated a swifter and less complex technique to achieve similar vessel shapes. Furthermore, the very elaborate cast technique must have been technological knowledge monopolised by few glass workshops (in the eastern Mediterranean?), whereas all over the empire new glass workshops arose. In the secondary workshops at Lyon (Motte, Martin 2003; Robin 2008) and Avenches (Amrein 2001), the cast technique 134 The determination of the period of this vessel type is not based on well-dated archaeological contexts but only on the result of some logical reasoning in relative dating. In the same region, this technique appeared already in Period II for the production of toilet ware, but there is no indication for continuity. 142 seems not to have been applied at all. It is an economical pattern that when (glass) manufactories multiply, the (glass) production expands exponentially, and the consumption market subsequently will be flooded with (glass) products. As a result of the law of supply and demand, the prices drop, and for that reason more and more people can afford these products (in glass). It is thus not surprising that the cast technique and the more valued decolourised or specifically coloured glass was maintained much longer for tableware than for toilet ware and containers which were produced in the ordinary ‗naturally coloured‘ glass by means of the very fast and flexible free-blowing technique. Another economical aspect that was a response to the increasing demand was the improvement of the production capacity from the same amount of glass by free-blowing very thin-walled vessels compared to the thick-walled cast vessels. Taking into consideration that the mould-blown technique only came into use from the Claudian period (Ib AD) onwards (Stern 1995) means that the mould-blowing technique was a relatively late innovation in the Roman glass industry. The introduction of this technique happened at the time the production of cast vessels in monochrome strongly-coloured glasses was successful and very fashionable. Because this trend lasted to about the late Neronian period (Ic AD), we may assume that the production of mould-blown vessels in monochrome strongly-coloured glasses only persisted for about one generation. Unfortunately, the available material in blackappearing glass cannot confirm this assumption due to the absence of well-dated contexts. On the whole, only very few mould-blown examples in black glass are known so far and contrast with the relatively large quantity of free-blown black glass vessels. The very poor occurrence of mould-blown vessels in black glass cannot be attributed to the fact that this technique was barely adopted in Period I because loads of mould-blown vessels in ‗naturally coloured‘ and decolourised glass were produced at that time. Another element of evidence to early date the free-blown and mould-blown vessels in monochrome black (and most likely other deeply-coloured) glass in Period I is the absence of the pontil mark. The introduction of the pontil st rod is assumed in Flavian times (Id AD) (Stern 1999; Israeli 2005). Hence, it is not surprising that 1 century AD vessels blown in black glass do not have a pontil mark. The conclusion is that the blown vessels in black glass in st the 1 century AD were not in production when the glassblowers got acquainted with this technique. If this occurred, two options can be suggested: the pontil rod got introduced earlier, or black glass vessels were also produced in Flavian times. st The pieces at hand from dated contexts demonstrate a clear distribution during the mid 1 century AD, between st 40-70/80 AD. It appears that around the middle of the 1 century AD, the black glass vessels were cast for the most part, while free-blown vessel shapes only occur sporadically (Cosyns, Fontaine 2009), and that the cast vessels in black glass got slowly outcompeted by the free-blown vessels. For instance, Magdalensberg (AT), with a terminus ante quem at c. 45 AD, yielded cast vessels in black glass but none were made with the blown technique. Conversely, even the earliest forms of cast vessels such as Form IA.1 make it impossible to propose an early date for the production end, given the rather late terminus ante quem of the finds from, for instance, Pompeii (IT) with the eruption of 79 AD. Furthermore, there seems to be a sort of continuity in the production of cast vessels in black glass, given the Flavian-Antonine shapes such as the small shallow dishes with overhanging rim (Forms IA.11) and everted rim (Forms IA.12) (Rütti 1991, 42; Price, Cottam 1998; Martin Pruvot 1999). Also the mould-blown cylindrical cup with cut rim and abraded concentric lines (Form IC.1) can be considered a Period st nd I form type of the late 1 to mid-2 century AD. It is unclear whether the base fragment from tomb 114 at Courroux (CH) is to be placed in this same time span or whether it is a Period II vessel since it cannot be linked to a definite shape. On the one hand, the burial contained an As from Domitian minted after 81 AD (Martin-Kilcher st nd 1976, 197, pl.50:D2), and the cemetery of Courroux is generally dated 1 –early 2 century AD. Similar base fragments in black glass have been recorded in Augst, one of which was found in the Regio 17B workshop 135 ‗Äuserre Reben‘, dated c. 125–210/220 AD (Fischer 2009, 27). Besides this marginal continuation in the use of black glass for the production of vessels during the Flavian and early Antonine period, it is apparent that the production of the black glass vessels corroborates the chronological evolution of vessels in other monochrome st deeply-coloured glass with its peak in the mid-1 century AD. 4.3.2 Period II nd rd During the 2 and the 3 century AD, the casting technique went out of use and the vessel production in black glass mainly concentrated on the free-blown technique, although a whole range of vessels were mould-blown. 135 Chemical analysis is advised to verify whether this form type (pro forma described as Form IIB.21) contains a high iron oxide concentration characteristic for the post-150 AD period or a low iron oxide concentration characteristic for the pre-150 AD period (Van der Linden et al. 2009). 143 This technique was apparently uncommonly used to blow black glass objects. The grape-shaped amphoriskos, Form IIC.1, from Nijmegen (Isings, van Lith 1992, 17, fig.19) is similar to the blue-green ones from Germania inferior, in the Dutch Rhine region, down to Aquitania in southwest France and the purple one from the tumulus of nd rd Vorsen (BE), and it can be dated in the second half of the 2 or the first half of the 3 century AD on the basis of simple stylistic correspondence (Simon-Hiernard 2000). In Northern Gaul within the provinces Gallia Belgica and Germania Inferior, free-blown black glass vessels do not nd appear in closed contexts earlier than the mid-2 century AD. The earliest dated contexts containing black glass nd vessels in the north-western provinces appear during the third quarter of the 2 century AD. Approximately 90% of all free-blown black glass vessels date between 180-230 AD, while no context gives a date before 150 AD or after 270/280 AD (Cosyns, Hanut 2005, fig.3). However, the use of black glass remained th th popular for the production of jewellery and lasted until the end of the late Roman Empire (end 4 early 5 century AD). The use of black glass for making free-blown vessels coincides with the expansion of the production of glass nd tableware from the last quarter of the 2 century AD. Nevertheless, vessels in this ‗colour‘ remain very scarce; for example, Augst only inventoried 0,1% of 5118 glass vessels in black glass (Rütti 1991, 109). nd Free-blown black glass vessels do not appear in closed contexts earlier than the mid-2 century AD in the northwestern provinces (Cosyns, Hanut 2005, 116). The earliest dated contexts only appear in the third quarter of the nd 2 century AD. Besides the carinated beaker or carchesium (Form 1/Isings Form 36b) from the chronological homogeneous horizon dated between 170/180 and 230/240 AD, we have the cylindrical beakers with outsplayed rim (AR nd 38/Trier 39) which are supposed to date from the second half of the 2 century AD, although none of the known examples came from a well-dated context. The carinated beakers (Form 1/Isings Form 36b) are contemporaneous with the snake-thread-decorated glass vessels which are supposed to have come from the Cologne workshops. Both are almost equally widespread in Germania inferior, but also other black glass vessel shapes were found in association with snake-thread glass tableware, such as in Esch IV, the Netherlands (van den Hurk 1975, 79-80, nos. IV:11; IV:12) or in the grave of Elsdorf-Esch, Germany (Gaitzsch 1999, 79). The three jars (Form IIB.4/Isings Form 94) registered among the burnt material from pits 2/3 of the Siesbach barrow, Germany, belong to a context which was dated by dendrochronology in the years 173-174 AD (Abegg 1989, 227). The study of the necropolis ‗En Chaplix‘ of Avenches (CH) gives some well-dated structures containing Form IIB.4 (Martin-Pruvot 1999, nos. 1254-1262). By the Flavian period, vessels in black glass or in any other deeply coloured glass were not a popular commodity nd anymore, and as a consequence they were taken out of production in the glass workshops. By the mid of the 2 century AD, there was within the north-western provinces, including Gallia Belgica, Germania inferior and Germania superior, a revival of the consumption of black-appearing glass vessels. Some contexts, however, throw doubt upon the absence of black glass vessels from the Flavian period to the Antonine period. Together with the undefined vessel shape in black glass from tomb 114 of the Courroux cemetery in Switzerland was found an As minted under the reign of Domitian (Martin-Kilcher 1976, 180, no. 5, pl. 31B:5). Given the difficulty of nd rd situating the base fragment within the late 2 early 3 century AD typology in combination with this Flavian coin, st nd some might be tempted to date the burial within the late 1 early 2 century AD. Some recent numismatic rd studies, however, demonstrated that Domitian coins also appear in small coin deposits dated in the early 3 century AD, as in Treignes (BE), Altbachtal (DE) and various sites in the civitas Remorum (FR) (Doyen 1980). Hence, it is not excluded or even more realistic that this free-blown base fragment in deep green, black-appearing nd rd glass is to be dated in the second half of the 2 century or even early 3 century AD. In the case of the Courroux piece, a chemical analysis can give a decisive answer about the true dating of the vessel since a low iron content st nd rd would place it in the 1 century AD and a high iron content would indicate the late 2 early 3 century AD [see Chapter 9]. The carinated beaker, Form IIB.1, is generally dated half a century to a century earlier than the carchesia in black st nd glass. Carinated beakers, all in colourless glass, are dated at the end of the 1 to 2 century ADe.g. in the Yale University Art Gallery (Matheson 1980, 41-42, nos.112-113); and in the former collection of Erwin Oppenländer (Hentrich, von Saldern 1974, 204, no.573a). Without provenance, they do not come from dated contexts and might thus have been dated too early. Presumably relying on this supposition, Fremersdorf dated the black glass st carchesia from Cologne (DE) in the 1 century AD (Fremersdorf 1959); likewise, Morin-Jean for those known from Boulogne-sur-Mer and Abbeville (Morin-Jean 1913). 144 Relying on these eminent but also superseded reference works and typologies can result in the inaccurate dating 136 of newly published material. It has been recently demonstrated that these carinated beakers are very well nd rd dated in the late 2 to the early 3 century AD from mainly Belgian, Dutch and German contexts (Cosyns, Hanut 2005, tab.1). One of the small bulbous cups with short concave neck, Form IIB.4, was found in an inhumation tomb in Avenches (CH) and dated after 180-200 AD (Martin Pruvot 1999, 199, no. 1261). From the other fragments, no clear context is known. The bulbous cup with outsplayed rim and flat base, Form IIB.6, has very rare parallels in other colours. This fairly rd uncommon shape is generally dated from FlavianTrajan times up to the middle of the 3 century AD (Rütti 1991). The early start is mainly due to an example from Augst which is dated on the basis of ceramics between 70-150 AD (Rütti 1991, 57, no. 1186). In contrast, the bulbous cup from barrow III in Esch is dated between 160-170/80 (Cosyns, Hanut 2005, 117, tab.1) and the recently retrieved piece from Faulquemont, France (Cabart, 2005, 20, fig.4) endorses this later date. The jug with ovoid body, Form IIB.11, from barrow VI at Esch (NL) is dated (van den Hurk 1977, 120, no. VI, 25) [cat.no.2879]. The rim fragment with vertical handle from a cellar dump at Straten (BE) [cat.no.519] was found together with a sestertius depicting Faustina the Younger (161-175 AD), wife of Marcus Aurelius, which might be used as a terminus post quem to date the context after 161 AD. Based on the ceramic finds, a slightly later date nd rd should be proposed somewhere around the end of the 2 century or the early 3 century AD (Wesemael, Hensen, Doucet, 2004). Both fragments have thus to be seen as early examples of ovoid jugs with one handle, rd since most known dated contexts of this vessel shape in other coloured glass are situated between the 3 century th AD and the early 5 century AD (Isings 1957, 149-152). Although the indented beaker, Form IIB.13, from Eschweiler-Hastenrath (DE) is dated 3 4 century AD (Follmann-Schulz 1992, 87-88, no. 47), we have to notice that the beaker entered the Rheinishes Landesmuseum st th Bonn after a purchase in 1890 and that the burial gifts have to be dated between the 1 and the 4 century AD. So probably we have here a mix of material from different contexts. Since this shape clearly imitates pottery rd shapes from the 3 century AD, and the vessels in black glass in the north-western provinces went out of rd circulation in the third quarter of the 3 century AD (Cosyns, Hanut 2005), we may assume that the Eschweilerrd Hastenrath beaker is an early 3 century AD vessel. rd th The black glass beaker with oval-shaped body and without neck from Nijmegen (NL) presently at the RMO, Leiden [cat.no.2975] is similar to the dark olive-green counterpart from the cremation tomb 287 of the southrd western cemetery at Tongeren (BE), which is generally dated 3 century AD by the author (Vanvinckenroye 1984, 144, pl.125:2). This form type is also reported by Beat Rütti for Augst/Kaiseraugst (CH), who linked it with cylindrical cups on base-ring type AR98 (Form IIB.17). The contexts from Augst/Kaiseraugst all fall within the range of 150 and 270/5 AD (Rütti 1991, I, 50). The shape of the rod-formed toilet bottles with solid body, Forms IID.1-2 show typological correspondence with the thin-walled, free-blown candlestick unguentaria with a short triangular body (Isings Form 82a2/b2) in ‗naturally coloured‘ glass and the solid body unguentaria (Harden Class XIII) in deep green glass. Hence, both techniques must have been used contemporaneously. This very close resemblance even allows dating the form type in the rd early 3 century AD, as is shown from stratigraphical analysis of different Roman forts in the Egyptian southeastern desert between Koptos in the Nile Valley and Myos Hormos and Berenikè on the Red Sea coast (Brun 2003). The well-dated military forts in the eastern desert of Egypt between Coptos and Quseir al-Qadim show the blown toilet bottles with solid-based body only occur in the stratified layers of Period II (Brun 2003, 383-387). The Elkab examples [cat.no.868-871] are rather similar to those from the fortress of Khashm al-Minayh–Didymoi (EG) rd which are dated in the first half of the 3 century AD (Brun 2003, 385, fig.9:7) so that we may assume the rodformed counterparts from Elkab have been produced in the Severan period. The type was preceded by more nd squad types of unguentaria with solid base in the second half of the 2 century AD, as shown by the examples nd from the fort of Al-Zarqâ–Maximianon (EG) and demonstrating an increasing use from the last third of the 2 century AD (Brun 2003, 383-385, fig.8:9-10). Analogous material from other Egyptian sites at the Red Seae.g. nd Quseir al-Qadim (Meyer 1992, 30); Berenike (Nicholson 2000, 206)is dated in the 2 century AD. The material 136 For instance, the publication of Abel Liéger on the Gallo-Roman cemetery of Cutry (Meurthe-et-Moselle) in the northeast of France reads about the three carinated beakers in black glass from tomb 176: ―le très riche ensemble de la T.176 … datable de la fin fin du Ier siècle‖ (Liéger 1997, 66). Although he does not give any reference on the ground dating these pieces so early, he possibly looked at Morin Jean and other classic references for a non-glass specialist. 145 nd from the fort at Umm Balad–Kainé Latomia (EG), which is dated around the mid of the 2 century AD, however lacks such unguentaria but contains taller unguentaria with a high and more slender conical body (Isings Form 82b1) (Brun 2003, 383, fig.6:8). Based on the presented information the unprovenanced Yale piece nd [cat.no.A4696] (Figure 47) can be dated in the second half of the 2 century AD. 4.3.3 Period III rd By the mid-3 century AD, the black glass carchesium went out of production, but examples regularly occur in rd contexts of the third quarter of the 3 century AD such as the late Roman hill-top watchtower at Nismes (BE) with a short occupation around 255-265 AD (Doyen 1980) [cat.no.459] or the fragments from a cellar at the vicus of Vieux-Virton (BE) with a fill dating 250-260 AD [cat.no.614]. Also other vessel forms remained in circulation up to rd the second half of the 3 century AD, such as the 280 AD-dated sarcophagus II in Stein (NL) yielding a shallow saucepan Form IIB.9 (Isings 1971, 25, no.78) [cat.no.2987] and the bulbous flask (Form IIB.10) (Isings 1971, 1112, no.19) [cat.no.2986] or the 230-270 AD-dated grave of Elsdorf-Esch (DE) comprising the amphoriskos (Form IIB.14) [cat.no.1869] (Gaitzsch 1999, 79). As these objects seem to be unique pieces, one might think these objects were heirlooms at that time. Striking is the gradual disappearance in burial contexts of blown vessels in rd black glass from the mid 3 century AD onwards. For instance, none have been recorded in the late cemeteries of Tongeren (Vanvinckenroye 1984), Oudenburg (Mertens, Van Impe 1971), Cologne-Jakobstrasse (Friedhoff 1991). The unguentarium with flattened body, Form IIIB.2, from Tomb 5530 at the Krefeld-Gellep cemetery is dated midth 4 century AD (Pirling, Siepen 2006, 272) [cat.no.1929]. The two other examples from the same site but made in rd th a pale yellowish green glass are from tombs dated in the transition from the 3 to the 4 century AD and the first th half of the 4 century AD, correspondingly tomb Kr9 and tomb 3198 (Pirling, Siepen 2006, 272). The five pieces th from Trier (DE) came from one single burial of the St. Matthias cemetery, tomb 112, which is dated in the 4 century AD (Goethert-Polaschek 1977, 137). The deep blue one from Strasburg (FR) is dated in the second half rd of the 3 century AD (Arveiller-Dulong, Arveiller 1985, 138, no.303). This vessel shape from the Jakobstrasse rd tombs in Cologne (DE) is known in decolourised glass from tomb 28, which is dated in the first half of the 3 th century AD and in deeply coloured glass from tombs 227 and 327, which are both dated mid 4 century AD nd (Friedhoff 1991, 283, Taf.95:3/12; 329, Taf.115:2). The piece from Cologne dated by Fritz Fremersdorf in the 2 century AD (Fremersdorf 1958) is most probably dated too early. Taking into consideration the above-discussed rd th dated contexts, we can infer a mid 3 –mid 4 century AD production date of the unguentaria with flattened circular body. Bearing in mind that the use of black glass in the vessel production faded away by the middle of the rd 3 century AD implies that the black glass piece from Krefeld-Gellep can be regarded as an early production. 4.3.4 Period IV The only Period IV vessels in black glass recorded up till now are rod-formed. No precise chronology of the tall and slender balsamarium, Form IVD.1, and the squat jar or amphoriskos, Form IVD.2, is hitherto available, as no publication provides well-dated contexts on the rod-formed vessels. Hence, it is not surprising that, depending on th th 137 the author, the proposed date range swings between the 4 and 9 century AD. Such a long-lasting use would be an exceptional case, meaning that the production and consumption of the coarsely shaped rod-formed vessels did not change from the late Roman period well into the Islamic period. Furthermore, all rod-formed vessels can be considered as one homogeneous group, as all types and subtypes are all made in a similar fashion and in a very limited number of shapes by using one single technique [see Chapters 2 and 3: 3.2.5]. However, a more th th clear-cut date range into the 4 –5 centuries AD has been proposed by Birgit Schlick-Nolte (2002, 104) and confirmed by Yael Israeli, who has demonstrated connections between the stamped lions on the body of squat jar th no.394 (Israeli, 2003, 295, no.394) and those on the stamped discoid pendants dated in the 4 century AD (Barag 2001, 173-176; Barag, 2002). A more accurate date will, however, only be possible when newly excavated parallels come to light from dated contexts. 137 Confusion can occur as deviating dates have been given erroneously when referring to general publications with outdated dates. For instance, Ulrike Riemenschneider (1989, 176, pl.2:7, no.7) gave a 5th4th century BC date to a rod-formed jar (Form IVD2 variant 5) in the old collections of the Altenessen museum based on Pol Fossing (1940) and Peter La Baume (1973). 146 4.4 Jewellery 4.4.1 Arm rings The available studies dealing with the dating of the different types of arm rings fall short. Thea Haevernick (1960) did not provide any chronological overview at all. The first constructive basis was offered by Hubert De Witte (1977, 46-51), and well-dated material from Belgium made it possible to date sharply the Type D1 bangles within th th a date range of about 50 years, from the mid 4 to early 5 century AD. The three other types, Types A-C, are th dated with not many hard data showing a much longer date range between 125 AD and the early 4 century AD. Taking into account that his study was restricted to solely Belgian material makes a re-evaluation necessary in order to verify whether the Belgian chronology corroborates identical material of dated contexts from elsewhere in the Roman Empire. A beneficial addition was supplied by Maud Spaer (1988) with the Palestinian material, but rd th she gives mainly general dates between the 3 and 5 century AD. Another important contribution helping to outline the chronology of the black glass bracelets is Emilie Riha‘s work on the material from the Swiss sites Augst and Kaiseraugst (1990, 66, tab. 82). The data from several contexts show, albeit contradictorily, longth lasting date ranges from Flavian times to the end of the 4 century AD, corresponding to the entire Roman st presence at many sites in the north-western provinces. This sometimes results in a very early dating in the late 1 nd and 2 century AD, which clearly conflicts with the dated finds from other sites. The work by Heiko Wagner (2006, 157-160) discussed the Roman glass bracelets from the upper Rhine region but did not contribute to the chronology debate. Neither did my initial work on the Roman black glass bracelets (Cosyns 2004), because it only intended to draw attention to the importance of reassessing the material on a vaster area covering the entire Roman Empire without adding new information to the issue. The here-proposed chrono-typology of the Roman black glass bracelets confronts the available dating from the aforementioned studies with material collected from additional dated finds. It is of no use to sum up all inventoried material because many of the inventoried artefacts come from unknown provenance, undated contexts, not welldatable contexts, or not yet studied or published contexts, or they remained undefined due to the unavailability of adequate information to determine the artefact or due to the fragmentation of the object. Hence, we take into account only those examples giving the possibility to bear out and when possible even sharpen the hithertoassumed date ranges. 138 The bracelets are discussed per category and subtype (Table 30; Figure 53) [see Chapter 3] to keep an overview on the eventual typo-chronological differences within the production or use of the Roman glass bracelets. In the assessment of the chronology of the black glass arm rings, we considered the open and closed bracelets 139 jointly because the open variant is a regional commodity contemporaneous with the closed bracelets [see Chapters 4 and 6] that have been produced in one single glass workshop such as Les Houis nearby Sainte Menehould (FR) [see Chapter 6]. 4.4.1.1 Type A: Arm rings with O-shaped section [open or closed with a seam] From the 425 inventoried Type A bangles, the majority have come from unspecified contexts or from undated contexts. Although 155 entries have received a date, only a few have come from well-dated archaeological finds (Table 58). The black glass jewellery from the workshops in the Argonne region like Lavoye (FR) and Sainte Menehould (FR) rd th are not recorded in Table 58. Their date situated in the wide date range between 3 and 5 century AD is solely based on related ceramic finds. o Type A1: plain rd Two fragments from Braives (BE) were retrieved together with pottery dated in the second half of the 3 century AD [cat.nos.217-218]. A piece from Liberchies (BE) came from the destruction layer of a building with an nd rd rd occupation date in the 2 3 century AD [cat.no.399]. A 3 century AD burial in Xanten (DE) yielded an intact open bracelet with spatula-shaped ends [cat.no.2172]. 138 139 For a detailed overview of the different types, we refer to where the full typology of bracelets is discussed in detail. The open variant has been solely attested in the north-western provinces [see Chapter 6]. 147 Table 58: List of type A bangles from dated finds cat.no. type site 3468 A1 Augst (CH) 217 A1 Braives (BE) 218 A1 Braives (BE) 399 A1 Liberchies (BE) 2172 A1 Xanten (DE) 1260 A1 Maule (FR) 4120 A1 London (UK) 4293 A1 Wanborough (UK) 2194 A1 Intercisa (HU) 2204 A1 Intercisa (HU) 2844;2846 A1 Amman (JO) 1352;1360 A2 Reims (FR) 1794 A2 Bad Wimpfen (DE) 2855 A2 Bastendorf (LU) 2324;2329;2332 A2 Gezer (IL) 2319 A2/3 El Bassa (IL) 404 A3 Liberchies (BE) 367 A3 Liberchies (BE) 398 A3 Liberchies (BE) 357;358 A3 Liberchies (BE) 500 A3 Roly (BE) 163 A4 Aiseau-Presles (BE) 434 A4 Matagne-la-Grande (BE) 606 A4 Vervoz –Clavier (BE) 554 A4 Tongeren (BE) 3365 A4 Augst (CH) 3379 A4 Augst (CH) 3423 A4 Augst (CH) 3650 A4 Kaiseraugst (CH) 1238 A5 Maule (FR) 590 A5 Vechmaal (BE) 625 A5 Waudrez (BE) 604 A6 Vervoz –Clavier (BE) 2436 A6 Jerusalem (IL) 2856 A6 Bastendorf (LU) 3643 A6 Kaiseraugst (CH) date IIIb-IIId AD IIIB-IVA AD IIIB AD II-III AD III AD IV AD III AD IV-V AD IVa AD IVa AD III AD III AD IIIA AD II-III AD IV-VI AD IVd AD II AD IId-IIIa AD IIIA AD IId-IIIa AD IIB AD IIIB-IVa AD IVa AD IIIB-Va AD III AD Id-II AD Id AD IId-IIIA AD IIId-IVA AD IV AD IIIc AD (c.275) IIIc AD IIIB-Va AD III AD II-III AD III-IVA AD reference Riha 1990, no.675 Vilvorder 1993, Braives 5, 243 De Witte 1977, II:58, no.30 De Witte 1977, II:58, no.103 s.n. 1950, 155-157, fig.12 Arveiler 2006, 156, no.6 Barber, Bowsher 2000, 148 Monk 2001, 171, fig.67, no.143 Vágó, Bóna 1976, 62, no.1050, Taf.15, 1050:2 Vágó, Bóna 1976, 103, no.1239, Taf.24, 1239:8 Harding, 1950, 81-94, nos.359; 375 personal communication Hubert Cabart Hoffmann 2002, 427, no. W387 Wagner 2006, 311, no.R2 Macalister, 1912, pls.94:11-12; pl.104:19 Iliffe 1934, 81-91, pl.24:1 Graff 1962; De Witte 1977, II:58, no.94 De Witte 1977, II:58, no.109 De Witte 1977, II:58, no.110 De Witte 1977, II:58, no.111-111bis De Witte 1977, II:58, no.140 Lefrancq 1988, 27, no.PG31:3, fig.3:6 Cattelain, Paridaens 2009, 111, no.3, fig.50:3 Witvrouw, Witvrouw 1976, 178, fig.40:17 Mertens, Vanvinckenroye 1975, 51 Riha 1990, no.661 Riha 1990, no.666 Riha 1990, no.665 Riha 1990, no.664 Arveiler 2006, 159, no.38.2 unpublished Hanut, Capers 2003, 86-90 Witvrouw, Witvrouw 1976, 178, fig.40:16 Rahmani, 1967, 89, fig.18,2, pl.24B Wagner 2006, 311, no. R3 unpublished rd From the above-mentioned dated contexts, we can propose a 3 century AD date for the type A1 bracelets. o Type A2: decorated with large twists The Palestine material is dated somewhat later in the 4 6 century AD. This is a relative chronology based on the finds from long-lasting, large family burials in use over generations, regularly attested in the eastern Mediterranean. th th st We wish to mention that the very early dates from Augst (CH) in the second half of the 1 century AD (Riha 1990, st nd no.670) and late 1 2 century AD (Riha 1990, nos.668; 672) are definitely deviant from all other dates. Most likely a reassessment of the context material might offer a different look on the date, yet we have to remark that the shape is rather nonstandard and interpreted as waste material of misshapen arm rings with wide twists. It is reasonable to think that this deviation in shape is related to an early production. However, chemical analysis on the piece defines a clear high iron oxide ratio, whereas Period I black glass features low iron content [see Chapter 8]. It is therefore more correct to think that this early date is linked to the excavation methods adopted, not noticing that the early level was at a certain point cut by a much later waste pit filled with production waste from a workshop in Augst. o Type A3: decorated with fine twists The layer wherein a fragment of open bracelet from Bon Villers at Liberchies [cat.no.404] was found is said to nd have contained pottery of the 2 century AD, but the context circumstances remain vague and therefore are not 148 very reliable (Graff 1962). Another piece from Liberchiesa closed type this timecomes from the fill of a cellar nd rd including pottery of the late 2 –early 3 century AD (Claes, Millau 1962) [cat.no.367]. A second piece of closed bracelet from Liberchies was retrieved from the levelling stratum of a building that has been dated in the first half rd of the 3 century AD (Claes 1978) [cat.no.398].Two non-joining fragments [cat.nos.357-358], most likely from one single bracelet, were found in the fill of a cellar. In the same stratum, a coin minted under Pertinax (193 AD) and nd rd pottery of the late 2 –early 3 century AD (De Witte 1977, I:48, II:46-47, nos.111-111bis) were retrieved. The piece of closed bracelet from the south corner of the inner courtyard of the villa 'la Crayellerie' at Roly [cat.no.500] nd has been dated in the second half of the 2 century AD on the basis of the excavated pottery and a sesterce 140 struck under Antoninus Pius (138-161 AD). o Type A4: with alternating plain zones and twisted decoration Few examples come from dated contexts, but we have to be careful to adopt, for instance, the much too early dated contexts of Augst pointing to a date as early as the Flavian period. Either the subsequent levels have not been carefully discerned and/or excavated, or the material is intrusive. The Kaiseraugst fragment dated in the last rd th quarter of the 3 to the mid 4 century AD corroborate better the generally proposed time-span. The Augst nd rd fragment [cat.no.3424] from a context dated between the end of the 2 and the mid 3 century AD can be considered an early production. The piece from the southern portico adjacent to building D at Clavier-Vervoz is rd th dated between the mid 3 to the early 5 century AD. o Type A5: The villa ‗Zouwveld‘ at Vechmaal, dated 2 3 century AD, was destroyed by fire under the Frankish incursions. The piece from the destruction layer of the bath complex is dated around 275 AD (unpublished). The Waudrez piece was found in a well that was re-used as a ritual pit showing an assemblage dated in the third quarter of the rd 3 century AD (Hanut, Capers 2003, 86-90). nd o rd Type A6: with twisted decoration and applied Little material of Type A6 comes from dated contexts, but none gives a very precise date range. The piece from rd th the southern portico adjacent to building D at Clavier-Vervoz is dated between the mid-3 to the early 5 century nd rd AD. The piece from the Bastendorf sanctuary is dated 2 3 century AD and is therefore the earliest dated context, but we have to bear in mind that the piece comes from a disturbed layer. The pieces from Jason‘s Tomb rd at Jerusalem are dated in the 3 century AD. The fragment from Kaiseraugst comes from an unknown context in rd th Regio 18 and is dated 3 mid 4 century AD. 4.4.1.2 Type B: Arm rings with a wide D-shaped section, plain or with ribbings across [open or closed with a seam] Table 59: List of type B bracelets from dated contexts cat.no. type site date reference Nahal Rakafot (IL) Jebel Yofeh, Amman (JO) Ramat-Rachel, Jerusalem (IL) Karm al Shaikh (Rockefeller Museum site), Jerusalem (IL) III AD III AD III-IV AD Spaer 1988, 54 Harding 1950, 81-94 Spaer 1988, 54 III-IV AD Spaer 1988, 54 B1 Mount of Olives, Jerusalem (IL) III-IV AD Spaer 1988, 54 B1 B2 B2var B2 Dourbes (BE) Augst (CH) Augst (CH) Lamyatt Beacon (UK) IIIB-Va AD IId-IIIA AD IIIB AD IV AD 236 B2 Dourbes (BE) IIIB-Va AD 297 B2 Hamois, Emptinne (BE) IId-IIIA AD 1351; 1353 B2 Reims (FR) III AD 1017 1029 B3 B4 Arles (FR) Carhaix (FR) III-IVA AD III-IVa AD 1033 B5 Chartres (FR) IIIc AD De Witte 1977, part II, 21, no.34, pl.IX:2 Riha 1990, no.625 Riha 1990, no.629 Leech 1986, 293, fig.24, no.14 De Witte 1977, part II, 21, no.35, pl.IX:3A-B Van Ossel, Defgnée 2001, 163, fig.142:5 Personal communication by Hubert Cabart Foy 2010a, 492, no.1041 Labaune, Le Cloirec 2008, 28, fig.8 unpublished (personal communication by Stéphane Willerval 2008-4) 2479-80 2838-43; 2845; 2847 2481 2443-44; 2451; 2459; 2463; 2466; 2468 2438; 2449; 2452-53; 2456; 2462; 2464 244 3390 3405 A4596 B1 B1 B1 B1 140 The bracelet fragment was found below tuff blocks in the south corner of the large courtyard which are connected with the second building phase (De Witte 1977, I:48; II:58, no.140). 149 From the 128 catalogued Type B bangles, only 36 come from dated contexts (Table 59). o Type B1: arm rings with wide D-shaped section, plain The contextual attribution of the Dourbes material remains perhaps unknown, but on the basis of coins and terra sigillata with roller stamp decoration, the occupation of the hill fort 'La Roche à Lomme' began during the second rd th half of the 3 century AD and lasted until the early 5 century AD (Brulet 1974; 2008, 575-576). o Type B2: arm rings with wide D-shaped section, indented decoration nd th The material corroborates the date range of the other bangle types falling within the late 2 to 4 century AD with the piece from the villa of Champion at Hamois (BE) and a piece from Augst (CH) coming from the earliest dated context, coinciding with Period II. The piece from Lamyatt Beacon (UK) came from the youngest context, situated th within the 4 century AD. The Palestine material has received a more general date range between both extremes (Spaer 1988, 55). o Type B3: arm rings with wide D-shaped section, incised decoration Only the piece from the excavations of the circus at Arles (FR) is dated. The level wherein it was retrieved, layer rd th 124, is dated 3 first half 4 century AD (Foy 2010a, 492, no.1041). o Type B4: arm rings with wide D-shaped section, impressed decoration (protuberances) Only the piece from the excavations of the circus at Arles (FR) is dated. The level wherein it was retrieved, layer rd th 124, is dated 3 first half of the 4 century AD (Foy 2010a, 492, no.1041). o Type B5: arm rings with wide D-shaped section, impressed decoration (complex design) The sole piece from a dated context is the bangle fragment from Chartres (FR) which was found in a waste pit with material dating before 260-280 AD but sealed somewhere in 280/285 AD, with a layer dated on the basis of 141 the coins. 4.4.1.3 Type C: Arm rings with a wide D-shaped section, with lengthwise ribbings [open] From the 106 entries, about half came from undated contexts, and of the 52 dated finds, most came from the rd Argonne workshops Lavoye (FR) and Sainte Menehould (FR), giving a general date range between the 3 and th the 5 century AD. Only 12 pieces remain useful, showing a coherent date range within Periods II and III for the making of the idiosyncratic Type C bangles [see Chapter 6], so that the production and consumption of this variant was less long-lasting compared to Type A and B bangles (Table 60). Table 60: List of type C-bracelets from dated contexts cat.no. type site date 1053 C1 Dehlingen (FR) IIIc AD 3341 C1 Augst (CH) III AD 3672 C1 Kaiseraugst (CH) IIB-IIIA AD 394 C2 Liberchies (BE) II-III AD 498 C2 Roly (BE) IIB-IIIA AD 524 C2 Taviers (BE) IId-IIIA AD 566 C2 Tourinnes-St-Lambert (BE) IIc AD 1874 C2 Epfach (DE) IVB AD 3428 C2 Augst (CH) IId-III AD 298 C3 Hamois, Emptinne (BE) IIIB-IVA AD 1359 C3 Reims (FR) III AD 3357 C3 Augst (CH) IB-III AD o reference Wagner 2006, 311, no.R4 Riha 1990, no.628 unpublished De Witte 1977, II:50-51, no.122 De Witte 1977, part II, 57-58, no.141, pl.XXVIII:5A-B Claes 1954, 244, fig.4a; De Witte 1977, II:61-62, no.150 De Waele, Hanut 2009 Werner 1969, 180, pl.38:31; 50:8 unpublished Van Ossel, Defgnée 2001, 163, fig.142:4 personal communication by Hubert Cabart Riha 1990, no.631 Type C1: with 5 lengthwise ribbings The only dated find with clear-cut date comes from the villa rustica of Gurtelbach at Dehlingen (FR). Dating the rd material in the third quarter of the 3 century AD, Heiko Wagner (2006, 311, no.R4) does not state precisely the 141 Personal communication by Stéphane Willerval, 2008-4. 150 context information. The pieces from Augst/Kaiseraugst (CH) match with the more general date range, going from nd rd the second half of the 2 through the 3 century AD. o Type C2: with 6 or more lengthwise ribbings nd rd A piece from Liberchies (BE) was found in the fill of a cellar which contained 2 and 3 century AD material (De Witte 1977, 50-51) [cat.no.394]. The piece from the villa 'la Crayellerie' at Roly (BE) [cat.no.498] was retrieved nd rd from a waste pit beside the villa. The fill can be dated last quarter 2 first half 3 century AD on the basis of the presence of mortaria fragments in terra sigillata of type Dragendorff 45 with a small spout in the shape of a lion nd rd head. The context wherein the fragment from Taviers [cat.no.524] was found is dated late 2 first half 3 century AD based on the pottery, despite the presence of some earlier coins: one struck under Hadrian (117-138 AD) and one under Marcus Aurelius (161-180 AD) (Claes 1954, 244; De Witte 1977, II, 62). The piece from the fill of ‗Raum A‘ in the warehouse (Magazingebäude) of the hill fort Lorenzberg at Epfach (DE) th is dated in the second half of the 4 century AD. The late date is because of the presence of a coin struck by Constantius II between 346 and 350 AD (Werner 1969, 180). The bangle fragment is most likely to be interpreted as residual from the area where earth was taken on the site to fill up the room. o Type C3: with 3 lengthwise ribbings It is not possible to find out whether the very long date ranges of the Augst material are due to the residual nd rd occurrence of early material within late 2 3 century AD levels or due to intrusive material within the second half st nd of the 1 first half of the 2 century AD because a later pit was cut through the level without notice of the rd archaeologist. The other two dated finds are from ‗Rue Desteuque‘ at Reims (FR)3 century ADand from the refill layer of room 8 in the residential building of the villa of Champion at Hamois-Emptinne (BE)second half rd th 3 first half 4 century AD. 4.4.1.4 Type D: Arm rings with a narrow D-shaped section, plain or decorated [open or closed with a seam] A large number of the 614 ex. of Type D arm rings originates from a context which is only very generally dated or not dated, contributing to a confused view on the chronological delimitation of their manufacturing. We only present a selection of the accumulated datable material in Table 61 and refer to the catalogue for a full overview. Despite the fact that most of the materials from dated contexts are Type D1 bracelets, all Type D bangle variants show a homogeneous chronological delimitation in Period III and Period IV (Table 61). Unlike the three preceding bangle types (A-C) which are characteristic for Period III we believe that type D bangles are typical products of Period IV. o Type D1: plain The late Roman hill-fort at DourbesLa Roche à Lomme, where 17 fragments of this type were retrieved [cat.nos.235;236-43;245-53], is dated on the basis of coins and Samian ware decorated with roller stamp in the th th th early 4 to the end of the 4 /early 5 century AD (Brulet 1974; 2008, 575-576). The late Roman hill-fort NismesLa Roche Trouée, not far from Dourbes, yielded one piece [cat.no.456] and its context is dated mid th th 4 early 5 century AD on the basis of a set of four coins, two of which were struck between 388-402 AD (De th Witte 1977, II:52-53). Hubert De Witte, however, maintains a general 4 century date (De Witte 1977, I:50). Similar dated contexts are known from the other side of the river Meuse, at Éprave, with two fragments of this th 142 type of bracelet [cat.nos.263;268] dated in the 4 century AD on the basis of coins and sigillata decorated with roller stamp (Mertens, Rémy 1973, 38-39, fig.24:19-20). Not far from Éprave, four intact bracelets were found in the late Roman cemetery within the Roman bathhouse of Furfooz-Hauterecenne [cat.nos.278-81]. Two were th found in tomb 2, which is dated at the end of the 4 century AD (Nenquin 1953, 88) and the other two came from th tomb 22, which is roughly dated 4 century AD (Nenquin 1953, 98), whereas Dasnoy dates the tomb in the last th third of the 4 century AD (Dasnoy 1969-70). The fragments from Namur-Grognon [cat.nos.451-53] come from a th fill of a cellar, which is roughly dated 4 century AD by Hubert De Witte (1977, I:49). A date range of the context somewhere in the last quarter of that century can be proposed seeing the presence of a coin struck under Magnus Maximus (383-388 AD). This later date most likely should corroborate the sigillata when a more detailed study of the different roller stamp decoration patterns will be carried out. 142 The coins were struck under Constantius II (337-350 AD), Constantine I (335-341 AD), Constantine II, Constans (346-350 AD), Magnetius (350353 AD) and Arcadius (383-402 AD). 151 The two complete Type D1-bangles made in yellowish green glass and in very dark purple, nearly blackappearing glass [cat.nos.517-18] from tomb C of the Spontin cemetery were found together with coins of which the youngest was minted during the reign of Arcadius (383-408 AD). Based on this terminus ante quem, André th th Dasnoy dated the tomb between 383 AD and the very end of the 4 or the first decennium of the 5 century AD th (Dasnoy 1966, 202). All pieces of type D1 found in Belgium within a dated context fall into the 4 century and the th very beginning of the 5 century AD, but the dated contexts of Spontin and Furfooz can markedly refine the th delineation to the mid 4 century AD or even towards the last quarter of the century. In Belgium, no clearly dated th contexts of the 5 century AD are available that contained type D1 bracelets. It is important to mention that a lot of recent publications rely on the dating of the burials in the balneum at Furfooz-Hauterecenne (BE) by Nenquin (1953). Table 61: List of type D bracelets from dated contexts cat.no. type site 14 D1 Bregenz (AT) 235; 241-43; 245-53 D1 Dourbes (BE) 237-39 D1 Dourbes (BE) date IVB AD IIIB-Va AD IIIB-IV AD reference Konrad 1997, 71-72; 235, pl.51:D1 De Witte 1977, nos. 40-51 De Witte 1977, nos. 36-38 240 D1 Dourbes (BE) IIIB-IVA AD Brulet 1974, 35, no. C2; De Witte 1977, no. 39 263;268 D1 Éprave (BE) IVb-IVc AD 278-81 451-53 456 D1 D1 D1 Furfooz (BE) Namur (BE) Nismes (BE) IVd-Va AD IVd AD IVB-Va AD 479;482-83 D1 Oudenburg (BE) IV AD 517-18 2015-16 1982-86 3667 757-61 1234-35 1643;1645;164953;1655;165758;1660;1662;1664-65 1755 A4607 2312 2313 2315 2321 2320 2421 2477-78 A4623 A4624 673 2756 2850 2837 D1 D1 D1 D1 D1 D1 Spontin (BE) Tawern (DE) Niederzier (DE) Kaiseraugst (CH) Štrbinci (HR) Maule (FR) IVd-Va AD IVB AD IVB-Va AD IVB-Va AD IVb-IVc AD IV AD Mertens, Remy 1973, 20-22; 38-39, figs. 24:1920; De Witte 1977, II, 27, nos.54-55, pls.XLI:6-7 Nenquin 1953, 74-88De Witte 1977, nos. 63-66 De Witte 1977, nos. 125-127 De Witte 1977, no. 128 Mertens, Van Impe 1971, 126-128, nos.7-8, pls.XXXII:5-6; 208-209, no.3, pl.LX:4 Dasnoy 1969-1970, 176-178, fig.4 Trierer Zeitschrift 15, 1940, 77, fig.40 Gaitzsch et al. 2003, 194-196 unpublished Jelinčić 2007, 217 Arveiler-Dulong 2006, 156, no.6; 11 D1 Sens (FR) IVd - Va AD Arveiller-Dulong et al. 1994, 184-186, nos.60-66 D1 D1 D1 D1 D1 D1 D2 D1 D1 D1 D1 D2 D2 D2 D2 Vireux-Molhain (FR) Canterbury (UK) Ain Yabrud (IL) Al Tarshiha (IL) Askalon (IL) El Bassa (IL) El Bassa (IL) Horshat Tal (IL) Mezad Tamar (IL) Tarshiha (IL) Tell Malkhata (IL) Virovitica (HR) Milan (IT) Tyr (LB) Amman (JO) IVB AD Va AD IV-V AD IV-V AD IV-V AD IVd AD IVd AD V-VII AD III-VI AD IV-V AD IV-V AD IIIA AD IVb-IVB AD IVa-c AD III AD 2326-28;2330-31;2335 D2 Gezer (IL) IV-VI AD 3207 3678 3788 2454 3677 3657 2445 2455 2461 D2 D2 D3 D4 D4 D5 D6 D6 D6 Talavera la Nueva (ES) Kaiseraugst (CH) IIId-IV AD IV(B) AD IV-Va AD III-IV AD IIId-IVA AD IVB-Va AD III AD III-IV AD III-IV AD Lémant 1985, 56, fig.60:17 Bennett 1980, 406-410 Husseini 1938, 54-55, pl.6,1 Iliffe 1934, 9-16 Iliffe 1933, 182-183, pl.48,4 Iliffe 1934, 81-91, pl.24:1-2 Iliffe 1934, 81-91, pl.24:2 Spaer 1988, 54 Erdmann 1977, 111, nos.932-933 Spaer 1988, 54 Spaer 1988, 54 Jelinčić 2007, 216-218, pl.1-3 Palumbo 1999, 29-30, fig.3 Chéhab 1985, II, 26 Harding, 1950, 81-94, no.345 Macalister 1912, pls.94:8;102:12-13;109:20; 119:25 Torrecilla Aznar 2004, 333, fig.4:14 Riha 1990, no.627 Fünfschilling, Hedinger 2008, no.3384 Spaer 1988, 54 unpublished unpublished Rahmani 1967, 89, fig.18,3, pl.24B Spaer, 1988, 54 Bagatti, Milik 1958, 160, fig.37:32 152 Pfyn (CH) Ketef Hinnom, Jerusalem (IL) Kaiseraugst (CH) Kaiseraugst (CH) Jerusalem (IL) Jerusalem (IL) Jerusalem (IL) Early dated material comes from, for instance, building 5 of the Aldenhovener Platte settlement at Aldenhoven rd th (DE) [cat.no.1761] with a dated context in the 3 century AD. A child‘s tomb dated in the second half of the 4 century AD was found near Tawern (DE) and contained two bracelets [cat.no.2015]. Other examples are known from Niederzier (DE) in the small cemetery in connection with the villa Hambach 382 (Gaitzsch et al. 2003). From tomb 1 were retrieved four bracelets [cat.nos.1983-86] and from tomb 5 one single bracelet [cat.no.1982]. Both th th tombs are dated from the second half of the 4 century to the very beginning of the 5 century AD (Gaitzsch et al. 2003, 194-196). th George Chenet mentions the discovery of a small bracelet in a context dated in the second half of the 4 century AD on the basis of the present pottery without referring to its provenance (Chenet 1941, 21, pl.1:25). o Type D2: with incised decoration, vertically or diagonally The bulk of the 119 Type D2 bangles came from dated finds. The 43 dated entries show a homogeneous date rd th range from the 3 to the 6 century AD (Table 61). The earliest pieces came from Kiskorija South at Virovitica (HR) (Jelinčić 2007, 216-218) and the burial of Jebel rd rd Yofeh at Amman (JO) (Harding, 1950, 81-94), dated in the first half of the 3 century AD and the 3 century AD, rd th respectively. A piece dated last quarter 3 4 century AD came from the villa 'El Saucedo' at Talavera la Nueva rd th (ES) (Torrecilla Aznar 2004, 333, no.14). Others, mainly from Jerusalem (IL) are dated 3 4 century AD (Baramki, 1932, 3-9; Bagatti, Milik 1958, 160), as is the material from Croatia (Jelinčić 2007, 216-218), although this might be because the author used Maud Spaer‘s chronology. o Type D3: with stamped decoration representing grotesque theatre masks None of the material comes from dated contexts, but considering the resemblance within the Type D assemblage rd th we may assume that it is of late Roman date, at least 3 5 centuries AD. Taking into account the general chronology of the black glass bangles, and in particular that of Type D bangles, we believe that the piece from the st site ‗Mariana‘ at Lucciana on Corsica (FR) [cat.no.1221] was erroneously dated 1 century AD (Foy, Nenna 2001, 95, no.120). o Type D4: tooled decoration with facetted surface th th The material receives a date range from the 4 to 5 century AD, but Maud Spaer (2001, 200-201) attributed this th th bangle type to the end of the 4 5 century AD. On the basis of a general date of the material from three excavation sites within Jerusalem (IL), some vaster lapses of time are provided: the YMCA excavations are dated th th 4 6 century AD, whereas the material from the Ketef Hinnom excavations and the Temple wall excavations are rd th dated 3 4 century AD (Spaer 1988, 55). o Type D5: applied decoration with simple line of dots on the crest None of the material has come from dated finds to our knowledge, but, for instance, Maud Spaer (1988, 56-57) rd th dated the Palestine material in the 3 4 century AD. But seeing the general chronology of the Type D bangles and the fact that similar material was retrieved in Aquileia (IT)first destroyed by the Visigoths in 401 AD and then again in 408 AD, making many inhabitants flee to the lagoon; the final destruction was inflicted by Attila in th th 451 ADthe production of this bangle type is rather to be situated almost certainly in the 4 mid-5 century AD. o Type D6: applied decoration with varicoloured specks (so-called ‗crumb‘ bangles) The only dated finds come from Jerusalem (IL), showing a d ate range from the 3 rd material from Jason‘s tomb is dated 3 century AD (Rahmani, 1967). o rd to 4 th century AD. The Type D7: applied decoration with simple line of dots on the crest The Braga material related to the Fujacal workshop is dated in the 4 5 century AD (da Cruz 2009) [cat.nos.3040;3064]. Also, the two pieces from the amphitheatre at Metz (FR) are dated in the second half of the th th 143 4 to the 5 century AD on the basis of the archaeological material. th 143 th Personal communication by Hubert Cabart. 153 o Type D8: pressed decoration with complex decoration pattern No dated context information is available on the Type D8 bangles from Aquileia (IT) [see comment on Type D5 above]. 4.4.1.5 The Romano-British bangles We have not included the Romano-British bangles in the assessment because we are convinced that their chronology needs to be treated separately due to their atypical character within the Roman imperial material culture. The Romano-British bangles are very idiosyncratic commodities; they do not fit the general typology based on continental material, nor do they correspond with the continental chronology. The Romano-British 144 contexts concerned are for the most part dated in the first two centuries AD ; however, similar material seems to rd th come from 3 and 4 century AD contexts (Kilbride-Jones 1938; Stevenson 1956; Price 2002) [see Chapter 3]. 4.4.1.6 Summary The evaluation of such a wide variety of types and subtypes from dated contexts necessitates a schematic overview to provide a clear idea of the chrono-typology of black glass bangles (Table 62). Instead of incorporating all dated material, implying the presence of imperfectly dated materialresulting in a useless all-embracing Period I-IV date rangewe only propose here an expurgated version of the most likely date range of each subtype, which is in the first place based on the well-dated material but stands in relation to the date ranges of the black appearing vessel glass, and the main trend developments of black appearing jewellery in glass and jet from late Antonine-Severan period. Table 62 generates some general views on the black glass bracelets: 1) absence in Period I; 2) appearing from Period II onwards; 3) continuity all through the Roman imperial period and beyond; 4) simultaneous occurrence of Types A, B and C in Periods II and III; 5) idiosyncrasy of Type D bangles in Period 4. Compared to the other th bracelet categories, the Type D-bracelets prove to be a well-delineated class emerging in the 4 century AD and particularly popular in the Rhine and Danube provinces, even though the late Roman type of glass bracelets show an empire-wide occurrence [see Chapter 6]. Table 62: Chronological overview of the black glass arm rings PERIOD I PERIOD II SHAPE 0 50 100 150 PERIOD III 200 250 PERIOD IV 300 350 400 450 TYPE A1 TYPE A2 TYPE A3 TYPE A4 TYPE A5 TYPE A6 TYPE B1 TYPE B2 TYPE B3 TYPE B4 TYPE B5 TYPE C1 TYPE C2 TYPE C3 TYPE D1 TYPE D2 TYPE D3 TYPE D4 TYPE D5 TYPE D6 TYPE D7 TYPE D8 144 The first Continental piece of Romano-British glass bangle is known from a ‗period 2/3‘ context in the Roman fort at Valkenburg Z.H. (NL) that coincides with a date range of c. 4769 AD (van Lith 1977). Another comes from the very recently published material on the glass from the early medieval town Dorestad (NL) (Isings 2009). 154 Yet we have to admit that the available information on the dated finds is rather unsatisfactory. Some material appears to be systematically dated too early, such as the material from Augst/Kaiseraugst (CH). Additionally, the material from the late Roman glass workshops in the Argonne region, Sainte Menehould–Les Houis and Lavoye, only have an assumed date, as comprehensive and accurate excavations have not been executed. The earliest nd dated Roman black glass bangles have come from mixed 2 century contexts, but in Augst/Kaiseraugst the st oldest pieces are said to have come from contexts dating to the late 1 century AD (Riha 1990, 66, tab.82) and are consequently the earliest testimonies. However, the proposed context dates of the Augst/Kaiseraugst material are inconsistent with all other known contexts containing Roman black glass bangles. Furthermore, it is already th th clear that a number of contexts are intermingled due to not very careful excavations in the 19 and early 20 century in particular (Fünfschilling 2005). Other sites with conflicting dated contexts appear also to be the result of mixed fills comprising material from different periods. Additionally, there is always the risk that archaeologists sometimes take over too naively or too dogmatically a formulated typo-chronology when facing forms and categories of small finds with which they are unfamiliar. The here-presented typologies are set up to enable quantifications and cluster groups geographically and chronologically. Also, it is important to take notice of site-bounded restrictions or of possible inaccuracies incorporated in the archaeological publication. A striking example is the type A4 bracelet fragment from the Limes st nd castellum Miltenberg-Altstadt [cat.no.1949], which is dated mid 1 –mid 2 century AD (Beckmann 2004, 96). Based on the typo-chronology from Augst and Kaiseraugst (Riha 1990, 64, tab.82), the bracelet fragment has been used to date the occupation layer wherein the artefact was found to the early occupation period at the Limes castellum Miltenburg-Altstadt. In contrast, the author tried to identify the beginnings of the fort occupation based on the glass artefacts. 4.4.2 Finger rings From the 204 recorded finger rings, only 66 entries are dated finds, of which 55 could be defined typologically (Tables 63-64). Table 63: Overview of the ratios per finger ring shape and subtype variant 1 variant 2 variant 3 variant 4 type A 55 22 14 2 type B 11 3 11 type C 6 8 1 - 4.4.2.1 variant 5 2 1 - variant 6 4 - total 95 30 15 % 67,4 22,0 10,6 Type A: Finger rings with narrow O-shaped or D-shaped section, plain or decorated o Type A1 Not all 31 dated pieces of type A1 are dated finds from a well-defined context. An amount of the material is dated on the basis of the general occupation date of the site, such as the material from the glass workshop(s) of Les rd th Houis nearby Sainte Menehould (FR), which yielded material from 3 5 century AD, or by comparing the items with dated parallel objects in other materials. All in all, the available material displayed in Table 64 demonstrates a rd th quite homogenous date range running from the 3 century AD to the early 5 century AD. The earliest material came from the jewellery (work)shop close to the northern gate of the legionary camp at Bonn (Platz-Horster 1984, 11-16; 38-48, pls. 4-7, nos. 11-31), whereas the latest material is that from ‗la cour de l‘archevêché‘ at Sens (Arveiller-Dulong et al. 1994, 186) and that from the Period IV glass workshop of Palais Kesselstatt at Trier (DE) [cat.no.2049]. The dated contexts from Augst (CH), wherein black glass material was found, sometimes present st rd incoherent date ranges from the late 1 to the end of the 3 century AD (Riha 1990, 137, no.292) [cat.nos.3414;3427]an observation that is noticed previously for other commodities. The site, however, also yielded very precisely dated contexts such as the treasure from insula 42 in regio 1 of the colonia Augusta Raurica, which is dated between 250 and 280 AD. Though the youngest coins were struck under Decius in 249251 AD, it has been proposed that the treasure got concealed after 260 AD and most likely during the crisis period featuring the incursions of Germanic people in 270-275 AD (Toma ević-Buck1980; Riha 1990, 118). o Type A2 Only 10 dated entries have been compiled, but compared to type A1 these provide a much more defined date rd th range, falling in the 3 4 centuries (Table 64). Only one entry from a not-yet studied context at Augst (CH) 155 provides a conflicting preliminary date range between 90 and 130 AD (unpublished). A plain example came from rd the ‗Insula 42‘ treasure of Augst (CH) and thus is dated from the third quarter of the 3 century AD. Table 64: List of finger rings from dated finds cat.no. type site type A date 1659;1667 1796-97;1804; 1809-10;1814; 3414;3427 1823-24;1830 3404 A1:1 A1:1 A1:1 Sens (FR) Bonn (DE) Augst (CH) IVd-Va AD III AD Id-III AD A1:3 Augst (CH) IIIB AD 1150 A1:4 Laneuvelotte (FR) IV AD 1648 A1:4 Sens (FR) IVd-Va AD 2049 A1:4 Trier (DE) IV AD 1811;1813;1817 A2:1 Bonn (DE) III AD 3376 A2:1 Augst (CH) IIIB AD 3398 3410 3646 1464;1476 551 1079 3401 3917 type B A2:1 A2:1 A2:1 A2:2 A2:3 A3:2 A3:2 A3:2 Augst (CH) Augst (CH) Kaiseraugst (CH) Sainte-Menehould (FR) Tongeren (BE) Forêt d'Argonne (FR) Augst (CH) Birdoswald (UK) Id-IIb AD IIIB-IVA AD IV AD III-IV AD IIIA AD III-IV AD IId-IIIb AD IVB AD Loeschke 1925, 338, fig.2, no.10; Wagner 2006, 320, R88 Platz-Horster 1984, 11-16; 38-48 Toma ević-Buck1980, fig.3; Riha 1990, 192, no.2996 unpublished Riha 1990, 137, no.290 unpublished unpublished Vanvinckenroye 1984, 138, no.272:4, pl.119 unpublished unpublished Wilmott et al. 1997, 283, fig.194, no.85 23 3648 169 type C B1:3 B1:2 B4:1 Bregenz (AT) Kaiseraugst (CH) Aiseau-Presles (BE) IVA AD IVB AD III-IVa AD Konrad 1997, 88; 210, fig.12:15, pl.17:D3 Riha 1990, 137, no.285 Lefrancq 1988, 27, no.PG33, fig.3:8 437 1647 1654,1661 2040 C1:1 C2:3 C2:1 C3:1 Matagne-la-Grande Sens (FR) Sens (FR) Trier (DE) IVa AD IVd-Va AD IVd-Va AD IV AD Rober 1983 Arveiller-Dulong et al. 1994, RAE 45, 186, no.75, fig.8 :75 Arveiller-Dulong et al. 1994, RAE 45, 186, no.75, fig.8 :751913, 157,no.1747 Henkel o reference Arveiller-Dulong et al. 1994, 186, nos.71-72, figs.8:71-72 Platz-Horster 1984, 11-16; 38-48 Riha 1990, 137, no.292 Toma ević-Buck1980, fig.3; Riha 1990, 192, no.2996 Cabart, Bull. AFAV 2008, 77, fig,1:1383.8 Arveiller-Dulong et al. 1994, RAE 45, 186, no.75, fig.8:75 Type A3 Only three items have been catalogued, all three giving a different but coherent date range (Table 64). The earliest context comes from Augst (CH) and dates back to the Severan period (unpublished). The youngest context is that of Building 197 at Birdoswald (UK). The piece was found in the floor layer dated in Phase e of th Period 5 that corresponds to the backfill of the sub-floor after the south horreum changed use by the end of the 4 th or at the start of the 5 century AD (Wilmot 1997, 202; 283). It is thus likely that the fragment should be dated earlier, as it is clearly a residual deposit. o Type A4 None of the two pieces come from dated finds. The proposed date of the piece from Autun (FR) or its vicinity is based on its resemblance to Helène Guiraud‘s type 2f finger rings in gold, silver, bronze and iron, which are dated nd rd 2 3 century AD (Vitrum 1990, 104, no.73; Guiraud 1989, 181). o Type A5 None of the two pieces come from dated finds. The proposed date of 3 4 century AD for the piece from Arles (FR) is presumably based on style features relative to dated objects in other materials (Foy 2010a, 496, no.1055). rd 4.4.2.2 th Type B: Finger rings with wide D-shaped section, plain or decorated It is of no use to discuss here separately all six recognised variants in analogy of type A because we have catalogued only three pieces from dated finds (Table 64). The one from the cave ‗Trou des Nutons‘ at Aiseaurd th Presles (BE) is dated 3 early 4 century AD (Mariën 1970; Lefrancq 1988). Another dated find is a type B1 156 finger ring with diagonal ribbings dated between 300/10 and 330/40 AD on the basis of the burial gifts from tomb 401-402, a double inhumation burial of children, from the late Roman cemetery at Bregenz (AT) (Konrad 1997, 88; 210). The third piece came from the lower city of the colonia Augusta Raurica at Kaiseraugst (CH) and has th been dated in the second half of the 4 century AD on the basis of the retrieved coins (Riha 1990, 137, no.285). Maud Spaer perhaps situates these finger rings in the late Romanearly Byzantine period, proposing a date from th th rd th the 4 to the 6 century AD. The few datable finds point to a concurrence with the type A finger rings in the 3 4 century AD. 4.4.2.3 Type C: Finger rings with narrow O-shaped or D-shaped section, plain or decorated with applied or tooled bezel The most characteristic dated context is Palais Kesselstatt in Trier, a Period IV glass workshop with numerous type C finger rings with different technologies. All dated finds of this type point towards a production period in the th th 4 and 5 century AD (Table 64). We can conclude that black glass finger rings are idiosyncratic to the late Roman imperial period. No finger rings in black glass have been recorded from Period I or Period II contexts. The emergence of finger rings made of black glass is to be placed in Period III and disappears again during Period IV. It is very likely that type A and B nd finger rings were already produced in Period II from the very end of the 2 century AD and remained in th th th production up to about the mid-4 century AD, whereas the type C finger rings are common for the 4 and 5 century AD. This advance in time is mainly presumed on the basis of the occurrence of similar shapes in other materials such as jet, which point to a production under Severan rule. We therefore assume that we catalogued the very fragmented material of type A from Sens (FR) wrongly and that they should be determined as fragments rd th of the later type C finger rings in jet, which were popular during Severan rule and all through the 3 –4 century AD. Even though the fragmentation of the pieces cannot provide a definite answer, the homogeneity of the assemblage supports this assumption. 4.4.3 Beads No reference book treating Roman glass beads currently exists, despite their very frequent occurrence, especially in Roman tombs. The few studies carried out can be regarded merely as major stepping stones in setting up a typo-chronology on glass beads of the entire empire. Unfortunately, their elaboration has remained chronologically and geographically very confined. Whether research has been done on material from a museum collection (Spaer 2001), a specific site (Riha 1990), a specific region (Wagner 2006) or a more vast area (Guido 1978; Tempelmann-Maczyńska 1985) the Roman (black) glass beads represented only a (small) part within these studies. This also concerns the overview given by Ellen Swift, who studied the late Roman dress accessories of a more vast region, i.e. Great Britain, the Rhine region and the Danube region, so that the types of glass beads and the covered time-span remained limited (Swift 2000). To come to a suitable chronology on the black glass beads, we verified as much as possible these publications with the dated finds compiled in our database. It is clear that on the more northern European Continent as well as in Britain local glassworkers continued to produce glass beads, faithful to their Celtic traditions, throughout the early Roman imperial period since it took several generations to have the Romanised material culture adopted. Only in the later Roman period glass did beads regain importance and become a very popular material in the jewellery production in generale.g. bracelets, gems, pendants, finger rings and beads. The start date for the production of these ornamental glass items th th remains undetermined. It is generally accepted that the glass beads were manufactured in the 4 and the early 5 century AD, but this date range seems to be conservative, not taking its use into account from the outset until its disuse. The only beads in black glass from Period I contexts are the large ‗melon‘ beads with opaque white spiralling trails. st Most pieces come from mid 1 century AD contexts. A piece from Tienen (BE) came from a TiberianClaudian context dated 30-50 AD (unpublished) [cat.no.536]. Beat Rütti dates the piece from Oberwinterthur (CH) somewhat later within the Claudian-Neronian period (Rütti 1988, 96, no.1935). The various pieces from Augst (CH) are mainly from non dated contexts or from contexts with a large date range. However, two melon beadsone of which with a dark blue glass matrixoriginating from well dated contexts, respectively 30-70 AD and 0-50 AD corroborate with the ones from Tienen and Oberwinterthur (Riha 1990, 157, nos. 1146; 1149) [cat.no.3352]. The recent study of the late Augustan military camps of Oberaden, Haltern and Anreppen in the Upper Rhine-region demonstrates however that the wearing of these large globular ‗melon‘ beads with white spiralling trail can be dated much earlier (Tremmel 2006, 235-285). One piece was found in the large Anreppencamp at the Upper Lippe, a tributary of the Rhine, is dated very sharply within the late Augustan period as it was 157 build within the winter of 4/5 AD and abondonned following Varrus‘ defeat in 9 AD. Also the piece from the Haltern fort is of early date, as it is considered to be build in 12 BC and abandoned by 16 AD (Rudnick 2006, 29). Bettina Tremmel (2006, 235-237) considers the Haltern material as late Augustan likewise the very well dated Anreppen camp on the basis of the correspondence of material and excludes early contexts seeing the Oberaden camp, which is dated 11-8/7 BC, only yielded a markedly small amount of glass. nd Pressed beads are present from the later 2 century AD onwards, such as the example from grave VII of EschHoogkeiteren (NL): a necklace containing 14 double-perforated, eight-shaped beads in black glass (Van den Hurk 1980, 374-378, no.VII:5:m, fig.6) [cat.no.2878]. Nevertheless, double-perforated pressed beads in black glass are th th idiosyncratic for the later Roman period, particularly for the 4 and 5 century AD. A large number of beads made of black glass dating from the late Roman period, i.e. 4 5 century AD, provoke some confusion because these are on the one hand also attested in early medieval contexts in the Weste.g. Anglo-Saxons, Merovingians, Longobards and Alamanni. Conversely, these are also attested in the Eastern Mediterranean, such as Cyprus, Israel and Egypt, regions not influenced by the Migration People. th th These beads are generally attributed to the Migration People (Guido 1999). A perfect example for this confusion is the large annular bead type with applied trails in white or yellow opaque glass. In Chapter 4, we have discussed the context of a number of rod-formed bead types in black glass accumulated in Western Europe which are th th assigned to late Roman contexts dating from the mid-4 to the early 5 century AD as well as to contexts from the th th Migration People dating to the 5 and up to the mid-6 century AD. These black glass beads with applied trails and/or small dots in opaque glass are equally present in the so-called early Byzantine period in the Levant and rd th Egypt, roughly dated from the end of the 3 to the early 7 century AD. In Chapter 6 on production, distribution and trade, the issue of the consumption pattern of this type of material is examined more in detail to argue for a production and consumption of glass jewellerybracelets, beads and pendantsall over the Roman Empire from th th the mid-4 to the mid-6 century AD. th From the mid-4 century AD onwards, the black glass beads with applied trails and/or small dots in opaque glass start to appear and continue to be produced, typically in annular, globular or barrel shapes. Within the margin of these facts, it is important to mention the association that can be made with the appearance of a new type of 145 glass, i.e. HIMT-glass , which is supposed to have been produced in Egypt [see Chapter 8]. A typical shape is the elongated barrel-shaped bead with applied glass trails rendering spirals at the ends and a single or cross-cutting double zigzag in the middle. Although it is clear that this type of bead is typical from the late th Roman period onwards as far back as the 4 century AD, it is equally known from early Byzantine contexts in the eastern Mediterranean and in the newly created realms by the Migration Peoplee.g. Anglo-Saxon, Merovingian and Alamanni. This means that sometimes an erroneous discourse of archaeological contexts can result in deviant conclusions when establishing a chronology of artefacts. The example found in 2001 in Comacchio (IT) is th th dated late 6 –early 7 century AD on the basis of the very delimited dating of the eastern cemetery of the S. Maria in Pado Vetere church and other examples from elsewhere (Corti 2007, 75-76). The point is not that we question the dating of the cemetery but that these elongated barrel-shaped beads are said to be characteristic for th th the later 6 and early 7 century. Besides the possibility of re-use of ancient beads and other objects as burial gifts in the tombs of Migration People, we have to be aware that the beads show a kind of continuity of late th th Roman taste and trends all through the 5 century and even up to the mid-6 century AD. Within Chapter 7 on use and function, we will explore this issue with the aim of seeing whether this material can provide evidence for sustained Roman traditions in north-western Europe under the rule of the Migration People. Some well-dated contexts are known for the ‗Trilobitenperlen‘. The castellum Vemania at Isny (DE) revealed a treasure containing a necklace made of ten ‗Trilobitenperlen‘ in black glass in combination with over 50 cylindrical beads in jet, and which is dated around 305 AD (Garbsch, Kos 1988, 34) [cat.no.1892]. 4.4.4 Pendants rd th From the available information the black glass pendants can be dated in the 3 to the 5 century AD illustrating continuity from Period II to IV but most idiosyncratic for Period IV (Table 65). 145 HIMT-glass stands for a type of glass with a high iron manganese and titanium content considered characteristic for Egypt, and stands in contrast to the until-then predominant Levantine glass. 158 Table 65: List of sites with black glass pendants from dated finds cat.no. site type date reference 2475 4024 20 2390 2600 3177 3205 4023 21 1226 Makher (IL) Colchester (UK) Bregenz (AT) Horbat Qastra (IL) Aquileia (IT) Alcalá de Henares (ES) Talavera la Nueva (ES) Colchester (UK) Bregenz (AT) Marseille (FR) ‗poppy-head‘-shaped ‗poppy-head‘-shaped barrel-shaped barrel-shaped barrel-shaped barrel-shaped barrel-shaped barrel-shaped jug-shaped jug-shaped IV AD IVb-IVd AD IVA AD III-V AD IV AD IV AD IIId-IV AD IVb-IVd AD IVc AD Vd Spaer 2001, 170, fig.76 right Crummy 1983, 35, fig.37, no.1501 Konrad 1997, 76; 220, pl.32:C unpublished Mandruzzato 2008, 79, no.152 Complutum 1998, 256, no.202 Torrecilla Aznar 2004, 333, fig.4:5 Crummy 1983, 35, fig.37, no.1504 Konrad 1997, 76; 230, no1:c, pls.48:E1; 91:4 Foy 2010b, 307, fig3:4 1666 Sens (FR) jug-shaped IVd-Va AD Arveiller-Dulong et al. 1994, 189, no.97, fig.9:97 2056 2972 A4633 A4658 A4635 2333 2429 2474 Trier (DE) Nijmegen (NL) Sagvar (HU) Ptuj (SL) Askalon (IL) Gezer (IL) Jalame (IL) Makher (IL) jug-shaped jug-shaped jug-shaped jug-shaped jug-shaped jug-shaped jug-shaped jug-shaped IVd-Va AD IV AD IVB-Va AD IVB-Va AD IVb-Va AD IV-VI AD IV AD IV AD Loeschke 1925, 338, fig.1:2 unpublished Stern 1977, 112 Stern 1977, 112 Stern 1977, 112 Macalister 1912, 377-378, pl.119, no.9 Weinberg 1988, 230, no.22 Spaer 2001, 170, fig.76 left 4.4.4.1 Pendant with bulbous body and gooseneck suspension loop Only three examples of this type of pendant have been recorded so far, and only the one from Tienen [cat.no.537] is made of black glass. Its context remained until now unstudied, and thus no clear dating of the object is possible. On the other hand, the Tienen-Grijpenveld site has a significant decline after the third quarter of the 3rd century AD and does not enclose contexts later than the beginning of the following century. The pendant in blue-green rd th glass from Augst (CH) is dated in the second half of the 3 first half of the 4 century AD (Riha 1990, 91; 161, no.1343). The one from a burial at cemetery "dell'Università Cattolica" close to S.Ambrogio basilica in Milan (IT) is th to be dated from the second third of the 4 century AD seeing the occurrence of 4 follis struck under Constantinus th th giving a terminus post quem of 337 AD, and an oil lamp of 4 –5 century type (Palumbo 1999, 29). The one from Liberchies (BE) in pale greyish blue glass comes from the destruction layer covering the latrines and adjacent nd rd buildings north of the bath complex in sector G is dated late 2 –3 century AD (Brulet 1997, 30). Seeing the very striking resemblance to the so-called poppy-head-shaped pendants of the late Romanearly Byzantine period, we are tempted to consider the pendant with bulbous body and gooseneck suspension loop as the forerunner of the so-called poppy-head-shaped pendants [see Chapter 3]. In connection with the available context information, we rd th may assume that the bulbous pendant with gooseneck suspension loop was produced in the 3 early 4 century AD. 4.4.4.2 ‗Poppy-head‘-shaped pendants Only two out of twenty recorded ‗poppy-head‘-shaped pendants come from dated finds (Table 65). The burial gift th from Locus A-2023 of the cemetery at Makher (IL) is dated in the 4 century AD (Spaer 2001, 170) and the one from tomb 1 of the cemetery at Butt Road, Colchester (UK) is dated between 325 and 400 AD (Crummy 1983, 35). rd th Others have been given a more informal date range either between the late 3 and 4 century AD (Spaer 2001 th th 177) or between the late 4 and 6 century AD (Spaer 2001, 185). 4.4.4.3 Barrel-shaped pendants From the 19 recorded barrel-shaped pendants we noticed 6 entries from dated finds (Table 65). A homogenous rd th date range between the late 3 to the end of the 4 century AD is to be proposed. The wider date range for the Horbat Qastra piece is due to the fact that the material is not yet studied in detail. 4.4.4.4 Jug-shaped pendants Only few pieces of the 50 entries are known to have come from dated contexts but the dated finds show a very th th homogeneous date range from the 4 to 6 century AD with a concentration for Period IV (Table 65). 159 A more in depth study of the rod-formed pendants entails the prospect to detect a more precise date range for the different types and subtypes comparable to the results from Dan Barag‘s study on the pressed discoid pendants briefly presented here below. 4.4.4.5 Pressed discoid pendants Dan Barag not only worked out a typology on the discoid pendants with pressed decoration, but he also gave date ranges for the three types discriminated on the basis of the suspension loop‘s shape which more or less subdivided each type into six categories on the basis of the pressed decoration (Barag 2001, 173-175). The three types seem to switch in time with Type A as the earliest style falling within the first third of Period IV (Table 66). The Type B discoid pendants more or less coincides the second third of Period IV, whereas Type C is th characteristic for the last third of Period IV and beyond up to the mid 6 century AD: Table 66: Chronological bar chart of the types of discoid pressed pendants (based on Barag 2001) 300 AD 350 AD 400 AD 450 AD 500 AD Type A Type B Type C 550 AD 600 AD Table 67 shows the various figurative decoration categories per variant of pressed discoid pendant established by th th Barag (2001). The decorative repertoire of Type A pendants, dated from the second half of the 4 early 5 century AD, is mainly that of classical mythology and the pantheon (category 1) and that of faunal motifs th th (category 5). Somewhat later but partly contemporaneous are the pendants of Type B dated late 4 mid-5 century AD). This type of pressed pendants is rare, but most are decorated with faunal motifs (only a few have th th been reported of Type C from the second half of the 5 first half of the 6 century AD). It is very possible seeing the quite overlapping occurrence of the three shapes and six styles that each shape is not very much a product of its time but rather a (technological) feature of a specific workshop and thus discriminates three different glass workshops where similar types of jewellery were produced during Period IV with workshop A being the most successful and workshop B the least thriving. Table 67: Overview of the decorative categories per type of pressed discoid pendant (Barag 2001) type A type B IVB-Va AD IVd-VA AD type C VB-VIA AD total Barag‘s Category 1: Classical mythology and pantheon Barag‘s Category 2: Biblical themes Barag‘s Category 3: Jewish symbols Barag‘s Category 4: Christian symbols Barag‘s Category 5: Faunal motifs Barag‘s Category 6: Miscellaneous motifs portraits TOTAL 2 1 7 11 2 21 21 7 5 9 30 5 5 82 4.4.5 20 4 3 2 19 3 5 54 1 1 1 5 7 Gems None of the recorded pieces came from a well-dated context except for the one from the Iron Age hill fort of Hod Hill (UK) [cat.no.4069] that was used by the Roman armies during the invasion of Britannia under Claudian rule. The Romans built a camp in 43 AD within the captured hill fort and remained active to about 50 AD. The second st quarter of the 1 century AD date of the circular gem in black glass with the representation of a rudder with tiller, corn-ear and cornucopia attached (Henig 1978², 236, no.405) only confirms the generally accepted date for the monochrome black glass gems. The gems in monochrome black glass known as imitations of semi-precious stones in, for instance, hematite or onyx are generally dated in the Julio-Claudian Period and therefore ascribed to Period I. Characteristic for these Period I gemstones are the circular shape and the convex top surface, contrasting with the gemstones from following periods which are of elliptical shape and flat surfaces. The early dated golden finger ring with black glass intaglio in the Rheinishes Landesmuseum Bonn is decorated with a grape in front of a hare in a very realistic style, which is regarded, like the finger ring, as being in use during the late Republican period–Augustan reign (Platz-Horster 1984, 87, Taf.22, no.78). The dating of gemstones can be done on the basis of various external aspects such as shape and engraved style. ‗Gemmatologists‘ have confirmed the interesting results a stylistic approach offers for the development of a typo160 chronology (Table 68). However, using stylistic criteria to date the glass gemstones is not very secure because this type of jewellery remained in use for a long period, while early genuine stones seemed to have been utilised as patrix to make a matrix for the production of one or a whole set of gemstones as attested from the material of the jeweller‘s shop of Medizinische Klinik close to the military camp of Bonn (DE) (Platz-Horster 1984) [see Chapter 3]. For instance the four pseudo-nicolo gemstones loose or still set in its original finger ring from the late nd rd Roman sanctuary at Liberchies (BE) [cat.nos.A4601-4] are stillistically dated (2 -)3 century AD (Sas 2008, 118rd 119, nos.1;3-5) but the sanctuary was built either at the very end of the 3 century AD (c.285-290 AD) or within th the first decades of the 4 century AD (310-320 AD) (Brulet 2008, 34-35). Table 68: General typo-chronology of the engraving styles applied on intaglios proposed by Kathy Sas (Sas 1993, 119-122) chronology typology 1st–first half 2nd century AD classical modelled style 1st–2nd century AD classical linear style 2nd century AD classical simplified style 2nd–first half 3rd century AD plain rigid style 3rd century AD incoherent style o Conical gems in monochrome black glass rd Oval and circular bronze brooches (types T270 and T271) with conical gems in black glass are generally dated 3 century AD (Riha 1979; Jones 1996, 181-182). An elliptical example with a conical gem in dark blue glass is rd th known from Fishbourne (UK) and comes from a late 3 or early 4 century AD context (Cunliffe 1971, 106, fig.40, no. 43). Those from the Roman fort at Caernarfon (UK) (Casey et al. 1993, 166, no. 13) and Wanborough (UK) th . are dated 4 century AD (Butcher 2001, 64-65, fig.26, nos. 137-138). The round disc-brooch enclosing a conical gem described to be of ‗dark green‘ glass was found in inhumation tomb 162 of the south-western cemetery of th Tongeren which is dated in the third quarter of the 4 century AD based on the other burial gifts (Vanvinckenroye 1984, 96-97, pl. 91:162). It included an As by Magnentius minted in Trier in 350 AD, three vessels in pottery, and a hemispherical cup in pale green glass type Isings Form 96 (Vanvinckenroye 1984, 97; 211). rd Oval and circular bronze brooches with inlaid glass gems can thus be dated between mid-3 century AD and the th th 5 century AD, being most popular in the 4 century (Allason-Jones, Miket 1984, 115). o Blue-on-black glass paste gems imitating nicolo Nicolo was used from the very early imperial period onwards. Emilie Riha explains that the two stylistic groups discerned by Gertrud Platz-Horster are related to two distinct periods. The first groupi.e. the thick gems with a vivid deep blue glass on the top surfaceis proposed to be an early imperial production, and she can supply six st examples, of which five come from a dated context assigned to the 1 century AD (Riha 1990, 22, nos.11;19; 50; 51; 55). The second group she calls a late imperial productioni.e. the thin gems with a dull greyish-blue top nd rd surfaceeven though the dating falls within the second half of the 2 to mid 3 centuries AD (Riha 1990, 23; nos.3; 20; 33; 37; 38; 40; 48). The golden finger ring with oval gemstone in blue-on-black layered glass paste from the Roman treasure of nd Regensburg-Kumpfmühl (DE) [cat.no.1998] is dated in the late third quarter of the 2 century AD with a closing date of 166 AD (Overbeck 2000, 64-65). While the nicolo gemstones became popular from the beginning of the nd 2 century AD onwards (Sas 1993, 110), we cannot find earlier imitations in glass of this semiprecious stone than the second half of that same century. It is also not quite clear when this glass imitation went out of use. Recently, a late production centre of glass paste gemstones imitating nicolo has been indicated in Braga (PT), which has th been dated in the 4 century AD (da Cruz 2009, 99). A P3/F4-gemstone with a Concordia-motif has been th th reported from the late 4 early 5 century AD tomb 6010 of the cemetery ‗La Hache‘ at Cutry (FR) (Boulanger et al. 2009, 27). When taking into account archaeological evidence, we see that well-dated contexts give sometimes divergent th dates. A perfect example is tomb 83 of the cemetery at Abbeville, which is dated late 4 century AD [cat.no.A4576]. The inhumation tomb of an adult woman yielded a golden finger ring of type Guiraud 4a (Guiraud th nd rd 1988, 188, fig.27), datable in the 4 century AD, while the inserted pseudo-nicolo glass gemstone is dated 2 3 century AD (Guiraud 1998, 134-135, no.14). The Roman large, round, discoid, pseudo-nicolo gems are mostly reported as medieval artefacts, either 8 10 th th century AD (Gandert 1956) or 11 13 century AD (Schulze-Dörrlamm 1990). In the medieval period these soth th 161 called ‗Alsengemmen‘ altered to a totally different use and meaning by having Christian motifs. We observed that they were re-used as adornment in Christian reliquaries and as Christian jewellery (Gandert 1956; SchulzeDörrlamm 1990). It is, however, clear from a good number of loose finds within unmistakable (late) Roman contexts in Conimbriga (PT), Augst (CH), Kaiseraugst (CH) and Sens (FR) [see Chapter 5] as well as the th unmistakable Roman helmet from Berkasovo (SB), which is dated in the last third of the 4th - start 5 century AD based on characteristics of the helmet type Deurne/Berkasovo 1 and the regularly connected 'Zwiebelknopffibel' (Mackensen 2009, 288-294), that the so-called Alsengemmen are to be regarded as re-used late Roman artefacts th th of the 4 -5 century AD. 4.4.6 Hairpins We have to acknowledge that only very few black glass hairpins have come from dated contexts. From the 43 entries, we have mainly generally dated contexts on the basis of the overall material from the Argonne area23 rd th ex. from the 3 4 century AD workshop(s) of Les Houis nearby Sainte Menehould and 9 ex. from an unknown 146 context at Forêt d‘Argonne. The only piece from a well-dated context is the intact composite one from the villa of Shakenoak Farm nearby Wilcote (UK) [cat.no.4201], which came from an unsealed deposit south of Building B of later-3rd century AD date (Harden 1971, Shakenoak II, 106, no.152, fig.45:69). Another hairpin with iron pinshaft (now lost) was found during the 1963-1968 excavations of the cemetery 'Pousse Motte', tomb 664, at Maule rd th (FR). The inhumation tomb is perhaps not dated but all burials are to be situated in the 3 8 century AD (Arveiller-Dulong 2006, 159, no.41). The hairpin with globular head from the legionary fortress at Oudenburg is from an undated context but fits rd th perfectly the assumed date of this type of commodity in the 3 4 century AD because the fortress was built th around 200 AD and remained in use until the beginning of the 5 century AD. Also the two hairpins from th Liberchies are to be dated in the 4 century AD (unpublished) [cat.nos.372-373]. th The black glass hairpin with faceted head from Braga (PT) is linked to the 4 century AD workshop of Fujacal, although the piece has been lost and no further information is provided (da Cruz 2009, 99; 102, fig.4, no. 5.1) [cat.no.3003]. Another hairpin with polyhedric head [cat.no.1027] came from one of the three sarcophagi of the ‗cimetière des Capucins‘ at Beauvais (FR) excavated in 1860-1861. The drawing of the pin is closer to one in jet than in black glass, but due to the vagueness of the notes in the ‗manuscrit Mathon‘ a re-assessment by Richard nd rd Schuler could only determine the bulk of the material to be from the 2 3 century AD without being able to attribute the material to one of the three inhumation burials (Schuler 1995, 62, fig.19:3). When comparing the black glass hairpins with those in other glass hues (Crummy 1983, 28; Allason-Jones, Miket 1984, 275) and with the closely related jet hairpins (Allason-Jones 1999), it appears that no (black) glass hairpins st nd have been recorded from 1 2 century AD contexts. Furthermore, when taking into consideration all the rd th available information, we can hitherto only generally date the black glass hairpins in the 3 4 century AD, with most likely Period III as the production heyday of it. 4.5 Architectural Decoration We already mentioned earlier that the study of glass architectural decoration material should be studied jointly with the material in all glass hues. Seeing the complexity of such study, it was not within reach of this project to manage it. We therefore limited the section to a discussion that is related to accessible past work. For the ease of the survey, we discuss the black glass architectural decoration material per category, i.e. tesserae, twisted rods, bichrome and trichrome plaques, and inlays or intarsia. Consequently, the different categories can be referred to one and the same context. Apart from the archaeological data, the written sources can also provide information on the date of using glass as architectural decoration elements. 146 The Shakenoak pin is not fully of glass but made of an iron pin that is decorated with a globular pinhead in black-appearing glass [see Chapter 3]. 162 4.5.1 Tesserae The largest part of mosaics including black glass tesserae that have been recorded are based on the study by Frank Sear and demonstrate a use of black glass tesserae all through the Roman imperial period (Sear 1977) (Table 69). To come to a good dating system of the glass tesserae, the material from dated contexts should be assessed in order to verify whether shape, size and hue can already give clues. However, apart from external properties, the chemical composition of the applied glass is needed to help establish the dating of the loose finds from stratigraphic contexts to facilitate the determining of the date of archaeological levels. Table 69: List of sites with black glass tesserae cat.no. site 2815 Sperlonga (IT), ‗Villa of Tiberius‘ 2778 Rome (IT), ‗Villa dei Centroni‘ 2779 Rome (IT), ‗Domus Aurea‘ 2818 Tivoli (IT), ‗Villa of Brutus‘ 2732 Baia (IT), ‗Baths of Mercury‘ 3893 Tunis (TN), ‗Odeon Hill House‘ 2764 Ostia (IT), Isola Sacra (Tomb 88) 2816 Tivoli (IT), ‗Villa of Hadrian‘ 2817 Tivoli (IT), ‗Villa of Hadrian‘ 3631 Baden-Stadthofquelle (CH) 2853 Leptis Magna (LY), ‗Baths of Hadrian‘ 2780 Rome (IT), Stadium of Domitian 2777 Rome (IT), Via S. Basillo 2019 Trier (DE) 2020 Trier (DE), ‗S.Barbara Baths‘ 1411Sainte Menehould-Les Huois (FR) 12 1625 Saint-Emilion (FR), ‗villa du Palat‘ context grotto natatio nymphaeum nymphaeum fountain arcosolium nymphaeum serapeum nymphaeum frigidarium exedra fountain workshop octagonal basin date I AD IA AD Ic AD Id AD II AD II AD IIA AD IIA AD IIA AD IIB AD IIB AD IId-IIIa AD IIIB AD IIIa-c AD IIId-IV AD III-IV AD IVd-V AD reference Sear 1977, 64-66, no.24 Sear 1977, 67-68, no.28 Sear 1977, 90-92, no.61 Sear 1977, 96-97, no.74 Sear 1977, 114-115, no.109 Sear 1977, 163, no.222 Sear 1977, 101-103, no.86 Sear 1977, 110, no.99 Sear 1977, 111-112, no.104 Sear 1977, 176, no.269 Sear 1977, 148-149, no.177 Sear 1977, 125, no.132 Sear 1977, 128, no.136 Sear 1977, 173, no.260 Sear 1977, 172, no.256 unpublished Monturet, et al. 1980, 76 Glass was used for architectural decoration as early as the late Bronze Age cultures in Egypt and the Levant/Mesopotamia. In the Roman world, already from the late Republic-early imperial period glass was well incorporated into wall and vault mosaics (Sear 1977), but only a limited number of black glass tesserae come from Period I contexts. Although there are plenty of examples from Augustan times showing the use of glass tesserae in a variety of colours, its use becomes much more popular from Neronian times when the use of twisted glass rods and tesserae in Egyptian blue disappears (Sear 1977) (Table 70). Table 70: Timetable of various architectural decoration materials in Period I (according to Sear 1977) material type 100 BC 70 BC 30 BC 0 35 AD 50 AD Egyptian blue Twisted glass rods Glass tesserae 79 AD Black glass tesserae were at that time used in combination with shells and tesserae in Egyptian blue, white stone and red glass as attested from the grotto of the villa of Tiberius at Sperlonga (IT) (Sear 1977, 64). Although the construction is dated late AugustanTiberian, consecutive building phases up to the NeronianFlavian period are believed when the use of glass tesserae increased (Sear 1977, 65). Seventy mosaic fragments from the ‗Villa dei Centroni‘ at the via Anagnina in Rome contain black glass tesserae besides mosaic stones in green, yellow, blue and red glass as well as twisted glass rods (Sear 1977, 67). The villa is considered contemporaneous with the 147 earliest phase of the theatre of Marcellus on the basis of the yellow bricks that were utilised. Frank Sear, st however, considers it more opportune to date it to the first half of the 1 century AD because of the use of the glass mosaics (Sear 1977, 68). The vault of the nymphaeum in Nero‘s ‗Domus Aurea‘ at Rome is decorated with four circular mosaic panels and in the centre an octagonal mosaic panel. The badly preserved octagon represents two men, Odysseus and the Cyclop Polyphemus, which are on the whole composed of black and green glass tesserae (Sear 1977, 90, no.61, pls.35; 36:1). The works started in 64 AD, but the ‗Domus Aurea‘ was not finished after Nero‘s death in 68 AD. His direct successors lived in it and finished the house, but already Vespasian lived only sporadically in it. Trajan made it inaccessible by using it as foundations for a bath complex he built on top of 147 Build in 23 BC by Augustus in commemoration of Marcellus. 163 st it. Accordingly, Frank Sear dates the mosaic in the third quarter of the 1 century AD, between 64 and the early 70s AD (Sear 1977, 92, no.61). The black glass tesserae from the mosaics of the Flavian barrel-vaults at the so-called ‗Villa of Brutus‘ in Tivoli are said to be ―particularly large and shapeless‖ and ―none too evenly cut‖ (Sear 1977, 97, no.74), contrasting with the tesserae from Neronian contexts, which are well-cut and of very small size (probably to produce very delicately detailed figurative panels). Glass tesserae have been found in large numbers on various sites built or rebuilt under Hadrian‘s rule. Yet only few included tesserae in black glass, such as the nymphaeum and the Serapeum at Hadrian‘s villa in Tivoli (Sear 1977, 96-97; 110, nos.74;99); the so-called ‗baths of Mercury‘ at Baia (Sear 1977, 114-115, no.109) and the arcosolium mosaic from tomb 88 of the Isola Sacra cemetery at the Gate to Rome in Ostia, although the latter nd could equally date somewhat later in the first half of the 2 century AD (Sear 1977, 101-103, no.86). The black glass tesserae from Period II contexts so far recorded include, for instance, the upper storey of the exedra at the Stadium of Domitian on the Palatine in Rome. The exedra is considered an addition from Severan nd rd times and thus dating at the end of the 2 first third of the 3 century AD (Sear 1977, 125, no.132). The fountain rd in the house under the Via San Basilio in Rome is dated in the second half of the 3 century AD. th The black glass tesserae from the 5 century AD building plan of the villa du Palat at Saint-Émilion (FR) and from th a fill in room IV (context 1035) of the early Christian basilica from Sidi Jididi (TN), which is dated in the 7 century AD on the basis of ceramics (Foy 2004, 328-329, fig.194), demonstrate that black glass continued to be in use for architectural decoration material. 4.5.2 Twisted rods Most twisted rods available in large museum collections (Brussels; Stuttgart; Corning; Toledo) were acquired in th the 19 century in Italy. Luckily, a good number are still in situ such as in Pompeii (IT). According to Frank Sear‘s thorough study on wall and vault mosaics, twisted glass rods were solely incorporated in Roman architecture as st decorative elements during the first half of the 1 century AD (Sear 1977) (Table 71). Earlier finds of twisted rods are, however, known but consist only of loose finds. This means that these finds could have been produced for other purposes than for architectural decoration, such as utensilse.g. stirring rods, needles and pinsor handles of vessels [see Chapter 3]. The recent find of twisted rods from a dump of a workshop in Jerusalem (IL) is dated in the first half or the middle st 148 of the 1 century BC (Israeli 2005, 54) and predates Frank Sear‘s assertion for at least 50 years. Not wanting to put in question the integrity of the scientific work, a later date for the Jerusalem context can still be proposed because the date of the fill in the bath complex can be considered a terminus ante quem, whereas the sealing by a road built under Herod‘s reign between 37 and 4 BC (Israeli 2005, 54) gives an ultimate terminus post quem at 4 BC. From this approach, the Jerusalem context would then corroborate the general introduction of glass in 149 Roman architecture under Augustus. A glass workshop that produced twisted glass rods is ‗La Manutention st no.2‘ at Lyon and dates from the middle of the 1 century AD (Foy, Nenna 2001, 77, no.71). The youngest st contexts containing fragments of twisted rods are the late 1 century AD deposit of the ‗Centrum‘ in the rue du 150 Maréchal-de-Lattre-de-Tassigny at Amiens (FR) (Bayard, Massy 1983, 325; Dilly, Mahéo 1997, 17) and those st nd from the glass workshop ‗La Monté de la Butte‘ at Lyon dated mid-1 early 2 century AD (Motte, Martin 2003). The widespread production of twisted rods exclusively focussed on the consumption of utensils from the Flavian period, which possibly influenced the change from opaque coloured glass to transparent glass. Although the colours of the numerous fragments of twisted glass rods used as architectural element in the Aquileia bath complex are not described, it is worth mentioning it because the complex was initially dated to the st nd end of the 1 beginning of the 2 century AD. Conversely, Frank Sear considers the bath complex much earlier and based on stylistic groundsin particular due to the presence of twisted glass rods, shells and tesserae in sost called Egyptian bluehe dates the mosaics in the early 1 century AD (Sear 1977, 100, no.82). 148 149 150 It is impossible to determine whether the twisted rods were production waste of architectural decoration elements. On the other hand, the introduction of twisted rods in Sear‘s chronology probably needs to start several decennia earlier. Taking into account the other glass material, the deposit can, in all probability, be considered a later dump. 164 Most black glass rods from datable contexts came from sites within the Gulf of Naples, which were submerged by volcanic ash after the Vesuvius eruption of 79 AD. Most material came from fountains (nymphaea) [see Chapter 5], with Pompeii as provider of most examples: the nymphaeum of ‗Casa del Torello‘ (House V-I-7) is dated in the st second half of the 1 century AD on the basis of the use of brick construction (Sear 1977, 61, no.16). Frank Sear, however, points to the archaising technique used, as at that time it was common to cut glass tesserae. He therefore suggests that the concerned mosaic is an older one that has been cut out and reinstalleda practice 151 that has been recognised in the nymphaeum of Hadrian‘s villa in Tivolli (Sear 1977, 50-52, no.8) and that of Nero‘s ‗domus aurea‘ in Rome (Sear 1977, 90-92, no.61). Based on the masonry the columbarium of the freedman Pomponius Hylas is dated late Tiberian (Sear 1977, 66). The fountain niche in the nymphaeum of a villa on Cape Posillipo near Naples containing twisted rods made of st black glass with an opaque yellow trail is dated at the middle of the 1 century AD (Sear 1977, 77, no.39, pl.23:3). Table 71: List of sites with black glass twisted rods cat.no. site A4584 Pompeii (IT), ‗Casa del Torello‘ A4586 Rome (IT), columbarium of Pomponius Hylas A4587 Rome (IT), ‗Villa dei Centroni‘ A4588 Cape Posillipo, near Naples (IT), villa A4585 Aquileia (IT), baths A4583 Amiens (FR) 4.5.3 context nymphaeum columbarium natatio nymphaeum tepidarium (?) dump date Ic AD Ib AD IA AD mid I AD IA/Id-IIa AD Id AD reference Sear 1977, 61, no.16 Sear 1977, 65-66, no.25 Sear 1977, 67-68, no.28 Sear 1977, 77, no.39 Sear 1977, 100, no.82 Dilly, Mahéo 1997, 17 Bichrome plaques There is no possibility of evaluating the date of the bichrome plaques in black glass because the only material we th have at hand is from old museum collections acquired in Italy in the 19 century and without any knowledge of their provenance such as the material in the Corning Museum (Grose 1989) and the Toledo Museum (Goldstein 1979). According to the combined occurrence with twisted rods, we assume that the bichrome plaques are contemporaneous with the twisted rods and thus to be situated in Period I. 4.5.4 Inlays or intarsia Intarsia are mainly considered late Roman, dated in the 3 4 century AD (Oliver 2001), but the occurrence of (black) glass intarsia in the villa of Lucius Verus at Rome (IT) [cat.no.2802] has recently indicated that this was nd 152 already in use in the 2 century AD. The black glass loose elements of opus sectile from Aquileia (IT) [cat.no.2588] are of unknown provenance (Mandruzzato 2008, 155, fig.5). rd 4.5.5 th Written sources The available archaeological data, as discussed here above, provide no more than a relative chronology which is based upon general stylistic aspects and on the presence or absence of specific architectural decoration material, such as pumice, various types of shells, tesserae in so-called Egyptian blue or twisted glass rods. Strict specifications of the formal aspects or used technology of each artefact category have not been set up systematically, so an evaluation of the external properties of this kind of commodity resulting in chronological subdivisions could not be done here. We are consequently restricted to no more than an impression about when the use of glass elements as architectural decoration in general got introduced, became popular and went out of fashion. Written sources can contribute to a better dating of the architectural decoration material in glass, such as a comment on mosaics by Pliny the Elder dealing with its introduction: Mosaics came into use as early as Sulla‘s regime. At all events, there exists even to-day one made of very small cubes which he installed in the temple of Fortune at Palestrina. After that, ordinary tessellated floors were driven from the ground level and 151 It consists of the remains of a late Republican villa dated 4030 BC (Sear 1977, 5052), that got incorporated into the later structures of the early 2nd century AD. 152 This was done by Marco Verita during the 18th conference of the AIHV in Thessaloniki 2009: Sagui L., Santopadre P., Verita M. Glass technology, colours, forms and shaping in the 2nd century opus sectile glass materials from the villa of Lucius Verus in Rome. 165 found a new home in vaulted ceilings, being now made of glass. Here to we have a recent invention. At any rate, Agrippa, in the baths that he built at Rome, painted the terra-cotta work of the hot rooms in encaustic and decorated the rest with whitewash, although he would certainly have built vaults of glass if such a device had already been invented or else had been extended from the walls of a stage, such as Scaurus which we have described, to vaulted ceilings. (Pliny the Elder, Naturalis Historia XXXVI, LXIV.189 [transl. Eichholz, D.E., 1962, Pliny X, 149]) Seeing that Sulla was dictator from 82 to 79 BC, we may assume from Pliny‘s passage that mosaics got st introduced in Italy, i.e. in Roman architecture, by the end of the first quarter of the 1 century BC. Though Pliny states that glass mosaics were still absent in the buildings erected by Agrippa, we may deduce from Pliny‘s report that between 27 and 12 BCthe period of Agrippa‘s active building policyit was unusual to decorate public buildings with wall and vault mosaics by using glass tesserae. This assertion proposes the incorporation of glass as a new elemental material in the architectural decoration under Augustus‘ rule. The oldest archaeological examples, discussed by Frank Sear, where architectural decorations in which glass was used also date back to the reign of Augustus: the nymphaeum at the ‗Via degli Annibaldi‘ in Rome is dated 30-20 BC (Sear 1977, 54-56, no.13); and the nymphaeum at the ‗Casa dell‘Ancora nera‘ (House VI.X.7) in Pompeii (Sear 1977, 56-58, no.14); the ‗Grotta di Paris‘ at San Vittorino, near Tivoli (Sear 1977, 58-60, no.15); the cryptoporticus at ‗Via XX st 153 Settembre‘ in Rome (Sear 1977, 62-63, no.19) are all three dated at the end of the 1 century BC. Significant is 154 that glass tesserae were absent in late Republican nymphaea like the fountain at Hadrian‘s villa in Tivoli (IT) , which is dated 40-30 BC and only got covered with tesserae in yellow, green and blue glass in a second stage, supposedly under Tiberius‘ reign (Sear 1977, 51). The available archaeological cases challenge Pliny‘s statement, seeing that the introduction of glass for the architectural decoration in private buildings corresponds with the building activities of Agrippa. At this moment it is unclear what reason should be given to this difference. If the glass elements in private buildings got implemented earlier, than the ensuing question is why it took longer for public buildings. A different paragraph by Pliny the Elder explains the possible reason why it took several generations before glass tesserae became a widespread architectural decoration material: I shall not allow these two birds of a feather, two Gaiuses or two Neros as you please, to enjoy unchallenged even renown such as this; and so I shall show that even their madness was outdone by the resources of a private individual, Marcus Scaurus., whose aedileship may perhaps have done more than anything to undermine morality, … As aedile he constructed the greatest of all works ever made by man, a work that surpassed not merely those erected for a limited period but those to last for ever. This was his theatre, which had a stage arranged in three storeys with 360 columns; and this, if you please, in a community that had not tolerated the presence of six columns of Hymettus marble without reviling a leading citizen. The lowest storey of the stage was of marble, and the middle one of glass (an extravagance unparalleled even in later times), while the top storey was made of gilded planks. (Pliny the Elder, Naturalis Historia XXXVI, XXIV.113-114 [transl. Eichholz, D.E., 1962, Pliny X, 89]) It appears from this paragraph that in the late Republic, glass was considered excessively luxurious and that the 155 156 use of glass in the theatre of Scaurus built in 58 BC was symbolising the spoil of sobriety, one of the Roman st virtues. Probably because of the depreciation of glass and its common availability by the end of the 1 century BC, it was regarded as an extravagance. Agrippa, as representative of the emperor Augustus, was bound to the imperial propaganda honouring Roman virtues and traditions. For that reason, it is very likely that the public buildings resulting from Agrippa‘s building policy had to remain sober and simple. 4.6 Counters Plano-convex counters appear from Augustan times in polychrome as well as monochrome glass and should be considered characteristic of the Roman imperial period. The Roman imperial plano-convex counters in various opaque colours including yellow, white, blue and green contrast with the more conical-shaped counters in a clear translucent or transparent glass which are characteristic for the Hellenistic period in the Mediterranean. Black glass counters seem however rarely occurring in this early stage of the Roman imperial period. Evidence for this st st assumption are burials from the later 1 century BC and the early 1 century AD containing glass counters lack 153 For ‗Casa dell‘Ancora nera‘ it is stated that despite the use of glass discs and broken pieces of glass vessels, glass tesserae were totally absent (Sear 1977, 58). The use of broken glass in wall veneer is also attested in ‗Casa del Granduca‘ (House VII.IV.56) in Pompeii (Sear 1977, 67, no.27). 154 Frank Sear mentions only the presence of tesserae in Egyptian blue that have been erroneously considered as weathered blue glass tesserae (Sear 1977, 49). 155 David Grose considers correctly that the so-called glass columns are most likely to be interpreted as decorative elements in glass affixed on wood (Grose 1989, 357). 156 Marcus Aemilius Scaurus was in 58 BC aedile and organized the Aedilician Games after having defeated Arteas III, king of the Nabateans. 166 st counters in black-appearing glass, as can be demonstrated by the late 1 century BC grave at Weldwyn Garden City (UK) (Harden 1967, 14-16; Price 1995, 129; Cool, Philo 1998, 190). The total absence of black glass st counters is also observed at the pseudo sanctuary of Cybele at Lyon (FR) which is dated between the mid 1 st century BC and the first decennies of the 1 century AD (Desbat 2003, 402-403). Other well-dated British examples provide evidence for an evolution in the glass counters production and consumption. By the Claudian invasion, this variety of colours disappeared and made place for sets of so-called black-and-white counters such st as the ‗The Warrior‘s burial‘ BF64 at Colchester-Stanway (UK) [cat.no.A4582] from about the mid-1 century AD. This rich tomb yielded an entire gaming set including eleven black/blue counters, nine in opaque white glass and a gaming board from maple wood with brass handles and other fittings (Crummy, P. et al. 2007, 186-188; 217, fig.89). The earliest set with black and white glass counters we recorded so far is that from tomb 34 of the Canal st Bianco cemetery at Adria (IT), which is dated in the second quarter of the 1 century AD (Bonomi 1996, XXI:15) [cat.no.A4621]. The wide variety of coloured glassesmonochrome and polychromemost likely evolved within the TiberianClaudian period into sets of dark coloured and white counters, and likewise the huge numbers of black glass counters from the Tiberian period. To date, it remains unclear how to interpret this change from vivid coloured counters to black-and-white counters because Britain was at that time not yet incorporated into the Roman Empire. The presence on the continent of counters in similar vivid colours from, for instance, Nijmegen (NL) (unpublished) makes clear that it has to be interpreted as a chronological fact and not a regional 157 difference. At first sight it looks as if the external properties of the black glass countersthe different shapes (oval, round, quadrangular and irregular) and the different dimensions (small, medium, large) do occur jointly in one and the same context, as the Wanborough counters illustrate (Monk 2001, 172). This would then suggest that in all three periods a mix of small and large counters is present. Accordingly, no datable system is deducible from the external properties, making this approach seem useless for chronological purposes. For instance, Caerleon st yielded 48 counters, of which half came from dated contexts. Five are dated in the last quarter of the 1 century nd AD, 13 from the early 2 century AD and six from c. 160-230 AD (Brewer 1986, 156). Where the Continental nd rd st nd contexts concerned mainly date from the 2 and 3 century AD, those in Britannia are to be dated 1 and 2 st century AD. The 1 century military camps on the Rhine, e.g. Neuss-Novaesium (DE) (van Lith 1994, 311-314, nos.766-795) and Asberg-Asciburgium (DE), for instance, yielded only small-sized counters with a diameter below or equal to 20 mm (Table 72). Considering the limited number of inventoried counters as well as the huge number of material in other glass hues excluded, it is impossible to establish here a chronology on counters. The results from the available dated material we verified, however, demonstrate that a classification on the basis of the characteristic features such as size (small/medium/large) (Table 73); applied technique (cast/fused) (Table 74); decoration (Table 75); and glass hue (Table 76) can be significant to set up a workable typo-chronology on counters. First of all, we have to make two observations in respect to the 574 entries concerning the counters: 1) the bulk of dated counters is from Period I contexts, while a large group remain of indefinite date (Figure 91); 2) the majority of the counters is small-sized and the amount of large counters is little, but the size and shape of a good number of counters could not be catalogued (Table 72). 3% 3% Period I (0-150 AD) 38% 56% Period I (0-150 AD) 7% Period II (150-250 AD) Period II (150-250 AD) Period III (250-350 AD) Period III (250-350 AD) Period IV (350-500 AD) indefinite 1% Period IV (350-500 AD) 87% 3% 2% Figure 91: Ratios of black glass counters per period from dated contexts 157 It thus cannot be the result of a difference in taste by the British local nobility before the conquest, who might have neglected the black-andwhite counters as an imported product in favour of those in vivid colours. 167 Table 72: Quantification of the counters per size for each period size diameter (in mm) quantity % small medium large indefinite TOTAL < 20 20-25 > 25 - 354 94 25 103 576 61,5 16,3 4,3 17,9 100 Because the set of glass counters from inhumation tomb 6352 at Krefeld-Gellep (DE) (Pirling, Siepen 2006, 599600, no.6352:12, pl.105-106) [cat.no.1904] and that from the Corbridge Hoard (Allason-Jones, Bishop 1988, 82, fig.99:276) [cat.no.4033] include small- as well as medium-sized counters in black glass, correspondingly 9 and 23 ex., they were split up and thus incorporated in the small and medium numbers. The Krefeld-Gellep tomb is th dated to the first half of the 4 century AD by means of the burial gifts, whereas the Corbridge Hoard is dated to nd the second quarter of the 2 century AD, 122-138 AD on the basis of the militaria present in the wooden cist. Table 73 displays the amount of counters for each period per size category. Not negligible is that more than half of the inventoried small counters, 190 items or 55,4%, have no context information available or are from not precisely dated/undatable contexts. The bulk of the dated small counters came from Period I contexts with 71,4% and when taking into account a more general dating encompassing Periods I and II, 93,2% of all dated smallsized black glass counters were produced in the early Roman period. Even if half of the medium-sized black glass counters have no further context information, two thirds of the dated medium-sized black glass counters are datable in the second half of the Roman imperial period. Large counters have come from a too-small number of dated contexts, but are rarely recorded in late Roman contexts. Table 73: Quantification of the counters per size for each period period small medium Period I (0-150 AD) Period II (150-250 AD) Period III (250-350 AD) Period IV (350-500 AD) indefinite TOTAL 105 9 5 3 190 312 23 2 15 352 7 1 1 1 48 58 large 2 4 0 1 16 23 4 29 1 92 1 0 1 25 Table 74 demonstrates the amount of counters for each period according to the adopted technique, cast or fused. The adopted manufacturing technique could be determined for only 86 records or not more than 15% of all inventoried black glass counters; with 61 items cast (70,9%) 51 small; 9 medium; 1 largeand 25 fused (29,1%)23 small; 1 medium; 1 large. The number of determined pieces is hitherto too small to make any conclusion and both techniques are limited to Period I or remain indefinite. Table 74: Quantification of the counters per used technique for each period period cast fused Period I (0-150 AD) Period II (150-250 AD) Period III (250-350 AD) Period IV (350-500 AD) indefinite TOTAL 30 0 0 1 22 53 3 1 4 61 7 0 0 0 13 20 3 0 1 25 In Table 75, the monochrome and polychrome counters are subdivided in relation to the subsequent periods. Fifty-one entries or 8,9% are bichrome or polychrome decorated, showing one or more coloured dots in opaque white, red or blue or combining two colours [see Chapter 3]. All dated bi-/polychrome counters are of late Roman period except for one early piece from Olbia (FR) (Fontaine 2006, 346, no.410). The counter in deep purple glass with a white central dot [cat.no.1317] came from stratigraphic layer 5 within the south of insula VI, which is dated 50-80 AD. 168 Table 75: Quantification of the monochrome and polychrome counters per period period monochrome polychrome Period I (0-150 AD) Period II (150-250 AD) Period III (250-350 AD) Period IV (350-500 AD) indefinite TOTAL 189 15 3 3 270 480 1 0 4 3 8 16 28 4 19 531 0 27 0 43 Table 76 shows that a more extensive colour analysis of the black glass counters has to be done in order to be able to verify the occurrence in each period of the different glass hues that can be applied to manufacture blackappearing glass. We bear in mind that additionally chemical analysis is required because each optical hue might comprise various tinges which are due to a different chemical composition [see Chapter 8]. Table 76: Quantification of the counters per glass hue for each period period indefinite green blue-green Period I (0-150 AD) 176 4 2 22 1 1 Period II (150-250 AD) 13 0 0 Period III (250-350 AD) 7 0 0 30 Period IV (350-500 AD) 6 0 0 indefinite 248 14 4 TOTAL 502 19 7 st purple 2 3 1 1 0 17 24 blue 1 1 0 0 3 5 brown 4 1 0 0 1 0 8 14 other 1 0 0 0 2 3 nd Nearly all 1 -2 century AD counters fall within a cluster situated above the red line ratioThe large counters from nd rd dated contexts within Belgium were found in tumuli of the mid-2 early 3 century AD. We assume that this category of counters, made by the fusion technique, is probably characteristic of Period II, but this could not be clearly demonstrated and necessitates a more exhaustive re-assessment to verify whether the proposed assumption can be validated. th The idiosyncratic 4 century AD bichrome or polychrome counters decorated with one or more dots in opaque white, blue and/or red glass are seemingly medium-sized (Figure 92). Because these counters were finished by marvering the colourful applied dots we believe that the monochrome counters with the same diameter-heigth ratio were possibly also produced in Period III-IV. ROMAN BLACK GLASS COUNTERS 11 10 9 8 I-IIA AD HEIGHT IIB-IIIA AD plain counters 7 decorated counters 6 IV AD 5 4 5 10 15 20 25 30 35 40 45 DIAMETER Figure 92: Plot of the black glass counters according to the dimensions 169 In view of the larger diameter vis-à-vis the height of the fused large counters of Period II, it needs verification whether the flattened counters below the red line are all corresponding to one single production technique, i.e. fusion, or whether the diameter-heigth ratio of the decorated counters only coincidences as a result of the marvering. On the other hand the counters above the red line should be checked if each and every one corroborates the poored/cast technique. If so the production technique is indicative for the dating of the manufacturing of the plano-convex counters. Despite the fact that a re-evaluation of all glass counters from dated contexts regardless of the glass hue is necessary to reach a more refined chronology of the glass counters, we value the above-mentioned evaluation as indicative to the potential of assessing a plain commodity such as counters. 4.7 Conclusion We here sum up briefly the acquired information of the chronological analysis and present the periodicity of the black glass production and consumption during the entire Roman imperial period according to the previously defined four periods. It is, however, essential to realize that the production of black-appearing glass is as old as the invention of glass production itself and thus not a Roman invention achieved on the eve of the imperial period. Conversely, we can state that the Roman black glass commodities are idiosyncratic with a consumption that is affected by cyclic movements of consumption behaviour in fashion and applied manufacturing technique appears, showing in Period I the continuity of the casting technique during the AugustanTiberian period and the general 158 application of the glassblowing technique under Claudian rule. A chronological overview of Roman black glass material can therefore be used as index fossil for field work and material studies. One of the aims within this chapter was thus to define the chronology on the use of various deeply coloured glasses appearing black. A clear-cut pattern was perhaps not within reach of this research due to the wide variety of glass hues that have been used all through the Roman imperial period to produce artefacts in black glass. Anyhow, a rough estimation was possible, demonstrating that during Period I black glass artefacts have been produced in all possible huesi.e. blue, purple, green and brown glass. The black glass in use in Periods II and III is seemingly restricted to green and blue-green glass in various tinges. During Period IV, yellowish-brown or so-called olive-green glass becomes prevalent, but by the end of this period brown and purple glasses reemerged in the Levant to become customary in subsequent centuries in the east Mediterranean. In view of these observations, we had optical analyses undertaken on a set of samples. Nevertheless, we studied to the best of our ability the various hues of black glass artefacts in view of local or regional production centres and a (inter)regional trade and distribution [see Chapter 6]. Such approach can provide additional information on the consumption of black glass commodities because the production of idiosyncratic commodity types in a specific hue can be linked to a local or regional production, so that it is possible to characterise the assemblages of each workshop. It is, however, incontestable that more information on the assemblages of glass workshops producing black glass artefacts are needed to assign more explicitly the glass hues to a certain period (and a certain area). But the complexity in classifying the use of the various black glass hues is also related to the regionalism of production as well as to the type of material manufactured [see Chapter 6]. 4.7.1 Period I From the late Hellenistic/late Republic period, black-appearing deep blue and deep purple glass was already accepted in the production of the cast negroid-head pendants, and the core-formed and cast bichrome and 159 polychrome vessels. The popular bichrome and polychrome cast vessels under Augustan rule seem to have faded away during Tiberius‘ reign and got replaced by a consumption (and thus also a production) of monochrome strongly-coloured glass. There is no hard evidence available to prove to what extent the polychrome and monochrome deeply-coloured vessels occurred contemporaneously. There are, however, almost no contexts of Augustan-Tiberian date that have yielded monochrome black glass vessels. Magdalensberg (AT) represents one of the few exceptions, unless all pieces have to be dated to its occupation end in 40-45 AD under early Claudian reign. Albeit the same vessel shapes remained en vogue. Subsequently, the growing implementation of 158 Strikingly, there is about a century between the assumed invention of glassblowing and the integration of the technique in the secondary workshops in the production of glass vessels. 159 This material has not been studied within the scope of our research. 170 the free-blown technique at secondary glass workshops all over the empire during the Claudian period resulted in new vessel shapes related to the new technique, even if a good number of shapes were derived from cast shapes [see Chapter 10]. The monochrome deeply-coloured vessels come into vogue during the Tiberian and Claudian period and occur fairly frequent in contexts up to the reign of Nero, when decolourised glass became the new fashion besides pale blue glass. Despite the fact that the monochrome deeply-coloured vessels went out of use by the end of Nero‘s reign, production probably stopped only gradually in Flavian times. Sites in the Gulf of Naples, like Pompeii, show that this material was in use up till the late 70s AD. On the other hand, there are only few Flavian contexts recognised as containing black glass vessels and nearly none of post-Flavian times. The rich tumulus from Riemst (BE) containing a shallow hemispherical cup (Form IA.1) is dated in Hadrian‘s reign around 130 AD (de Schaetzen 1950; Vanderhoeven 1976), but its fragmentary state and very bad conservation condition with a thick weathering crust on either side as well as on the fractured surfaces indicates the shallow bowl must have been used for a long period of time and almost certainly has to be regarded as an heirloom. It can therefore nd be assumed that no vessels in monochrome deeply-coloured glass were produced in the 2 century AD but nd sporadically can be retrieved from contexts dating in the first half of the 2 century AD. The monochrome black glass vessels, like all other monochrome strongly-coloured glass vessels, were 160 idiosyncratic for the period c.30-c.70 AD with cast vessels for the Tiberianearly Claudian period and freest blown vessels during the late ClaudianNeronian period. The black glass vessels from 1 century AD-dated st contexts corroborate perfectly this mid 1 century AD colour feature in the Roman glass production (Cosyns, Fontaine 2009). For this reason, we divided Period I into two sub-periods: a) the Julio-Claudian period (or preFlavian period) from the start of our era to c.70 AD, comprising the black glass production and consumption; and b) the Flavian–early Antonine period from c.70 to c.150 AD, when black glass appears to have been neglected in the secondary workshops. The black glass jewellery in Period I was restricted to only large melon beads with a twisted coil in opaque white glass. Limited to the northern frontier zone, this type of bead resembled many late Iron Age beads and can be considered an example of the continuity of Celtic traditional glass bead making. This type of bead was contemporaneous with the vessel production within a date range of c.30-c.70 AD. Late Hellenistic/late Republic black glass pendants have not been recorded in Period I contexts. During nearly the entire JulioClaudian period, the architectural decoration material was limited to twisted rods bordering polychrome wall and vault mosaics. Only from the Neronian times, when mosaicists started to leave aside the artistic style polychrome wall panels with tesserae in a variety of vivid glass colours bordered with shells and twisted glass rods, did the black glass tesserae make their appearance. The black glass counters seem to have emerged only by the Claudian period and were characterised by 1) a small diameter (below 20 mm) or medium-sized with a diameter above 20 mm but mainly below 23 mm; 2) a poorly translucent glass so densely coloured that it appears opaque; and 3) a plain monochrome surface. Summarising, we can say that during Period I the use of black glass remained limited to the manufacturing of cast and free-blown vessels, ‗melon‘ beads, architectural decoration material and small counters, with a distinct change of the material culture under the rule of the emperor Nero. This (r)evolution with regard to technique, style and fashion was probably due to the re-organisation (and re-location?) of the glass production [see Chapter 10]. 4.7.2 Period II We have given Period II a time span of 100 years from 150-250 AD, roughly corroborating the revival of blackappearing glass vessels which were produced by a new technique(Cosyns et al. 2006a; Van der Linden et al. 2009) [see Chapter 8]. Different from the empire-wide distribution of cast vessels in Period I, this production was characterised by a regional distribution limited to the north-western provinces [see Chapter 6]. The vessels were 161 free-blown, mould-blown and in rare occasions rod-formed. Cast black glass vessels are missing in Period II. 160 Nevertheless, cast vessels also occur sporadically in late 1st and early 2nd century-dated contexts [see Chapter 5]. We catalogued from the early 3rd century AD a few distinctive plain unguentarium shapes commonly known as candlestick-unguentaria in a very deep brownish or greenish glass appearing black by means of the rod-formed technique which is typical for Period IV. Generally produced in a pale bluish or blue-green glass and by using the free-blown technique, this type of toilet ware has a characteristic tall cylindrical neck and low conical body. Typical is the solid body base with only a slight part of the tall neck useful as a receptacle [see Part IA; 9. Use and meaning]. 161 171 Where the Period I vessels mainly imitated shapes in metal and stone, the vessel shapes of Period II were rather inspired by metal, although there is also pottery showing equivalent shapes. Similar to the previous period, this period does not show continuity in production. From the dated finds, it can be deduced that during the first quarter of this period, black glass vessels either were not produced or were produced on a very small scale, because the earliest vessels are dated around 170/180 AD. It is only from the late reign of Marcus Aurelius onwards that black-appearing glass vessels became relatively fashionable again, and this all through the Severan period and continuing all through the subsequent intermediate period, which coincided with the so-called ‗crisis‘ period between 230 and 250/280 AD. We decided to take the entire Late Antonine period (150-190 AD) as one block because the black glass jewellery only started booming from the Severan period (190230 AD) and likewise the success of black-coloured jewellery made from jet became idiosyncratic for the entire late Roman period. The intermediate period and the start of Period III are partly overlapping because there is a lack of good dated finds from this period. Furthermore, there is not much difference between the assemblages of the later Period II and Period III. During this period the plano-convex counters are medium and large sized and made of monochrome black glass. Summarising, it is clear that besides the vessels, this period was characterised by a definite use of black glass for making architectural decoration material (tesserae and intarsia); beads and gems and counters. It is presumably the time when the bangle and finger ring production started. 4.7.3 Period III We labelled this period as the Tetrarch–Constantinian period (250/80-350/60 AD), although the denomination does not include the first quarter of this period and the last decennia. It is characterised as the continuing crisis period of the intermediate period within Period II. During this period, a wide variety of bracelets and finger rings were produced. The architectural decoration material in black glass during this period was limited to tesserae and discoid glass cakes that are assumed to have been the raw material from which tesserae were knocked off [see Chapter 7]. The architectural decoration material was limited to black glass tesserae to accentuate the figurative representations. Whether it be floral, faunal or human figures, its use is limited to small lines, and thus only small quantities were needed. Medium-sized polychrome counters were idiosyncratic to the Constantinian phase, although it is not excluded that th the production of such counters lasted until the early 5 century AD. 4.7.4 Period IV The material of Period IV (350/60-500 AD) contrasts drastically from that of the previous period in technique and style, as will be demonstrated with the chemical approach also in applied raw glass [see Chapter 6]. In view of these features, we assume that the rod-formed vessels confined to the Levant and Egypt were made during Period IV, showing again a regional distribution, but now, with regard to technique. But compared with the Period II counterparts, these were very crudely made and did not resemble (a) blown vessel shape(s). It is highly possible that these rod-formed vessels were made in a workshop of jewellery artisans and not in a workshop where vessels were free- and mould-blown because they look as if they were made by non-specialists. For the western part of the Roman Empire, Period IV can be divided into two phases. The first phase lasted until the Roman armies abandoned the Roman frontier fortification systems in Britannia and the Rhine-region between 406 and 410 AD. The second phase, called the ‗decline‘, lasted until 480 AD. Although the Roman political role and particularly its military presence disappeared in large parts of north-western th Europe in the first decade of the 5 century AD, we have to be aware that the Roman(ised) population did not adopt directly the dress code of the newcomerse.g. the wearing of glass jewellery. Generations must have passed before a full transformation took place. While this is a very acceptable theory, we are aware that it is a difficult matter to provide good evidence and for that reason it remains purely hypothetical. Most common during the early Byzantine period and in the Eastern Mediterranean, crumbed beads and bracelets cropped up in the West under the Migration Peoplese.g. in Merovingian and Anglo-Saxon burials. An intact example in the 172 collections of the Israel Museum is without provenance, but dated mid-Islamic as it is compared with a similar one th th from a 13 -15 century context. th Rod-formed black bead types came into circulation around the mid 4 century AD and continued to be used up to th the mid-6 century AD. Hitherto, it has been unclear whether we may speak of 1) the continuity of production of such beads during the start of the early Medieval period in the Western part, 2) whether we have to consider this as the effect of import from the prevailing (eastern) Roman Empire, 3) or whether it is only the result of a very intense ‗re-use‘ of old beads [see Chapter 7]. Within the stage of research, it is more appropriate to speak of a continuity in the use of the Roman material culture in the West for more than one century. For instance, the annular, globular or barrel-shaped beads and the pendants in black glass with applied glass trails or small dots. Within Chapters 5 and 6 on chronology and on trade and distribution, respectively, the issues on the dating and on the consumption pattern will be examined more in detail to argue for a production and consumption of glass th th jewellerybracelets, beads and pendantsall over the Roman Empire from the mid-4 to the mid-6 century AD. 173 174 Chapter 5 CONTEXT ANALYSIS There is a marked contrast between the wide range of vessels and objects, which survive complete in burials or, more frequently, are found in fragments on settlements, and the ephemeral and episodic nature of the evidence for the production processes by which they were formed. (Price 2002, 81) 5.1 Introduction A context analysis on archaeological artefacts can be carried out on various levels, but depending on the tackled issues, specific contextual information is compulsory. We therefore assumed it would be helpful to explain the limitations, methodology and aim of this chapter. Various constraints hampered an in-depth context analysis within the scope of this project: 1) the great disparity of artefact types from a vast area and a long-lasting era; 2) the rather limited availability of detailed context 162 information from museum material and old excavations ; and 3) the difficulty in acquiring information on the context circumstances of well excavated but unpublished material. By linking the specific functional types and explicit typological shapes to well-defined context types, explicit information on the consumer groups of the various black glass artefact types can be made available. We therefore verified whether the contextual evaluation can point to consumption patterning within a social dimension by connecting specific functional assemblages to social group, distinguished by social role, gender, age, income status, etc. Hence, the challenge of this chapter was in the first place to find out to what extent a general context analysis can help refine an assessment of the consumption pattern of black glass commodities throughout the Roman Empire and provide an insight on the social significance of the various artefact types in black glass throughout the Roman imperial period. Is it, for instance, possible to characterize commodity types as purely civilian or military? Is typically military material present at civil settlements and vice versa, and what can be the explanation? Can context analysis help in connecting particular commodity types to gender, age, social class, or religious persuasion? Were some commodity types intended for purely domestic use, or were they used in rituals or as burial gifts? The potential regional use of specific artefact types is examined in the following chapter, which discusses production and distribution [see Chapter 6], while the use and function of the various material groups in black glass is discussed subsequently [see Chapter 7]. The context analysis, here presented, contributes to our understanding of general social aspects of the consumption patterning of the black glass artefacts, while the subsequent chapters respectively focus on the socio-economic and the socio-cultural meaning of these consumer goods within Roman society. The applied context types used to carry out the context analysis are those defined in the fields ‗site‘, ‗site character‘, and ‗context‘ of the Access database [see Part IV – DVD]. We differentiated the character of the sites 163 to discriminate the material from civil and military settlements. This enabled us to verify whether the use of specific artefact types is ubiquitous or restricted to purely military or civil complexes. The important difference between the types is that the civil settlements can be regarded as self-regulating, autonomously developing, socio-economical habitats involving a local population that reacts instinctively to economical growth as well as to crises and disasters. Military settlements, on the other hand, are well-organized but temporarily and artificially constructed habitats involving a mainly male adult population of non-native origin. Politically regulated military camps were created on purpose to counter military threats and crises, whereas the civil settlements grow organically, essentially depending on economical rules. Figure 93 visualizes schematically the framework of this chapter: The objects given the predicate ‗civil settlement‘ come from non-military settlements, including rural settlements (villae), small towns, and large towns); 162 The rod-formed tall balsamaria and squat jars, characteristic to Period IV, mainly come from the antique market circuit, leaving blank their provenance. For that reason, no context analysis could be undertaken on this category of material. 163 Alternative contexts have been considered such as caves and rivers since these cannot be labelled as civil or military. 175   The artefacts from a ‗military settlement‘ have been recovered from various types of military contexts (including burgi, castella, and castra); A limited number of inventoried pieces have been retrieved from caves or dredged from rivers. Because of the particularity of these contexts, we decided to consider them separate from the civil and military settlements. When looking at the various context circumstances, we distinguish two main classes: the domestic context 164 covering all that comes from the dwelling (including workshops ) and the public context (including ritual deposits and burial gifts). Subsequently, the various functional categories from civil and military contexts have each been subdivided into four context types:     The material labelled as ‗domestic‘ can come from any household context, such as a ‗dwelling‘e.g. villa, barracks, ‗ditch‘, ‗waste pit‘, or ‗well‘, but also the material assigned to a ‗glass workshop‘ is categorized as domestic; The ‗public‘ context involves all public spaces and buildings such as bath complexes, fountains, libraries, and horrea, but also streets and fora; The artefacts recorded with ‗burial‘ come from a funerary context, including inhumation and cremation burials; The pieces inventoried as ‗ritual‘ come from a religious context, including sanctuaries, temples, and ritual deposits. CONTEXT ANALYSIS Civil Military contexts contexts Domestic/artisanal Domestic/artisanal * PER PERIOD * PER PERIOD Public Public * PER PERIOD * PER PERIOD Burial Burial * PER PERIOD * PER PERIOD Ritual Ritual * PER PERIOD * PER PERIOD Figure 93: Scheme of the subdivisions of this chapter To gain a clearer insight into the consumption patterning of the various artefact types in black glass and its consumers, we needed to further subdivide the wide range of commodity categories. Seeing the continuously developing society in the Roman Empire and the inherently coupled changing material culture, we maintained the subdivision of four phases established earlier. This would allow us to verify the existence of an evolution in consumer behaviour: 164 We assume that the glass workshops have to be considered within the domestic environment, as is, for instance, well demonstrated by the glass workshop of Kaiseraugst (CH) (Fischer 2009). 176 Period I : Period II: Period III: Period IV: st nd 1 –mid-2 century AD nd rd mid-2 –mid-3 century AD rd th mid-3 –mid-4 century AD th th mid-4 –5 century AD An initial verification of the entire database, without distinguishing the various commodity categories, shows that 1161 ex. of the 4475 inventoried pieces, or approximately 25,9%, are of unknown provenance, meaning that for 165 three quarters of the inventoried artefacts, there is useful provenance information available (Table 77). This is mainly owing to material from museum collections worldwide which either come from the antiquarian circuit or th th from the 19 century and early 20 century excavations. We believe, however, that valuable information can be gained from the finding circumstances of a good number of these unprovenanced artefacts. Because such a timeconsuming task could not be executed within the reach of this project, the here-proposed numbers are only reference values which need re-evaluation. Table 77: Overview of number of sites and number of items per context (* omitting the artefacts from unknown provenance and the additionally recorded material – see introduction) sites % %* entries % %(*) ratio quantity/site civil settlements 356 53,1 84,6 2773 62,0 83,7 7,8 military settlements 53 7,9 12,6 497 11,1 15,0 9,4 caves 6 0,9 1,4 29 0,6 0,9 4,8 rivers 6 0,9 1,4 15 0,3 0,5 2,5 unknown 250 37,3 1161 25,9 TOTAL 671/421* 100,1 100,0 4475/3314* 99,9 100,1 6,7 (AV); 6,1 (MD) The two simple column charts at the top of Figure 94 displays plainly that most of the material has come from civil settlements but also that the context of a huge amount of entries remains undefined. The two general pie charts at the centre of Figure 94, including all catalogued artefacts in relation to the number of sites (left) and the number of entries (right), misrepresent the results caused by the impact of the entries with unknown context. We therefore made a second set of pie charts excluding the artefacts from unknown provenance. The obtained picture of the range of intensity per context makes clear that a Roman black glass artefacts are only a minor occurrence in military contexts (12,6% of the sites; 15% of the entries); however, the amount of artefacts is higher, with a ratio of 9,4 items per site compared to the average of 6,7–, whereas the number of artefacts from caves and rivers is negligible (Table 77). Further in this chapter, we discuss more in detail the ratios of the various commodities within each specific context group. The first impression from the pie charts, however, is that most black glass artefacts come from civil settlements and that this observation becomes even stronger when excluding the material from unknown contexts (Figure 94 bottom right). 165 For 153 records, the country origin was not known; for 509 pieces, the find place was not known; for 1563 pieces, more detailed specifications about their contextual provenance was lacking or was unavailable during the research. This does not mean that all remaining records held detailed contextual data. 177 356 400 350 300 250 200 150 100 50 0 250 53 civil settlements military settlements 6 caves 6 rivers unknown sites 2773 3000 2500 2000 1161 1500 497 1000 15 29 500 0 civil settlements military settlements caves rivers unknown entries sites sites 1% 1% civil settlements civil settlements military settlements 37% 53% military settlements 13% caves caves rivers rivers unknown 85% 8% 1% 1% entries entries 1% 0% 1% 26% 11% 62% civil settlements military settlements caves rivers unknown 0% 15% civil settlements military settlements caves 84% rivers Figure 94: Ratios of sites and entries in relation to the settlement character (incorporating or excluding the items from unknown provenance) 178 5.2 Civil contexts To better understand the general consumption patterns of black glass artefacts within all non-military settlements, the material from civil contexts has been tackled in two different ways: firstly, by distinguishing four main contextual circumstancesdomestic (+ artisanal), public, ritual, burial; secondly, by subdividing all material into rural settlementsvillae and small towns, i.e. viciand urban settlementslarge towns such as municipia and coloniae. The purpose of considering the different context circumstances of the various commodities in relation to both subdivisions is to end up with ratios to be able to verify and refine the presented general charts in Figure 94. When classifying the black glass artefacts from civil settlements according to the proposed four major context circumstances, we have to bear in mind that 1432 entries, or about half of the recorded material, possess no detailed provenance information (Table 78). When leaving these out of consideration, it is clear that only little material has come from public and ritual contexts and that most material has come from contexts with a domestic or a burial character (Figure 95). 166 Table 78: Overview of black glass artefacts per context (* incorporates the items related to secondary glass workshop s settlement context sites % entries % ratio domestic* 68 19,1 576 20,8 8,1 public 17 4,8 83 3,0 4,8 ritual 27 7,6 63 2,3 2,3 burial 166 46,6 619 22,3 3,7 unknown 141 39,6 1432 51,7 10,4 TOTAL 356 100,0 2773 100,0 7,8 ) It is perhaps not surprising that the largest amount of black glass artefacts has come from burials. One of the reasons is that the Roman tradition dictated the offering of burial gifts for the deceased in the afterlife and that the consecrated character of the area kept most cemeteries unharmed. Another reason is that the intensive th excavations of Roman cemeteries since the 19 century account for this high number of sites that delivered black glass artefacts. The high number of black glass artefacts from domestic contexts provides a surprisingly elevated ratio. The explanation is the addition of the material from glass workshops (386 entries for only 10 sites). When excluding this from the count, the dwellings without artisanal activities deliver only 166 entries from 58 sites, resulting in a ratio of 2,9 items per sitesomething very similar to the ritual context and quite close to the public and burial contexts (Table 78). At first sight, we noticed the occurrence of multiple examples of black glass artefacts in various types of civil funerary contexts. Although it was not possible to verify within the present research whether these objects were gifts from burials at a villa, vicus, municipium or colonia, such an approach would provide interesting information about patterns of distribution and consumption in general and about socio-cultural aspects in particular. Finetuning of the available data is also necessary seeing that the here-presented evaluation disregards the various functional types as well as a chronological sorting, something we aimed for below. 166 The 197 items of production waste and chunks of raw glass that are mainly related to workshops are left out of consideration here because this material cannot be integrated in the quantifications of particular commodities. The material comes from clear workshops (Avenches (CH); Lyon (FR); Les Houis near Sainte Menehould (FR); Trier (DE); Braga (PT); Kaiseraugst (CH)); assumed ones (Lavoye (FR); Augst (CH)); or the provenance has an unclear character (Luni (IT); Nijmegen (NL); Celje (SL)). 179 sites sites domestic 16% unknown 34% public 4% domestic 24% ritual 6% public 6% burial 60% ritual 10% burial 40% entries entries domestic 21% unknown 52% public 3% domestic 43% burial 46% ritual 2% burial 22% ritual 5% public 6% Figure 95: Pie-charts showing the ratio of black glass artefacts from the various Roman civil contexts When refining the subdivision of the material from the database per commodity class in relation to the provenance context, it is possible to look at the numbers in two different ways: per commodity class or per context category (Table 79). To present a better visualization of the accumulated data, we inserted double chartsi.e. including and excluding the material from unknown provenanceto see the impact of the amount that is excluded and to give a more accurate ratio of the distribution pattern for each commodity (Figures 96; 97). Table 79: Overview of the various commodity classes per context type commodity classes Vessels Jewellery Architectural decoration Counters Production TOTAL 180 domestic n % 115 22,5 251 14,5 9 33,3 16 5,4 185 93,9 576 20,8 public n % 12 2,3 45 2,6 15 55,6 9 3,0 2 1,0 83 3,0 civil settlements ritual burial n % n % 7 1,4 97 18,9 50 2,9 462 26,6 0 0,0 1 3,7 6 2,0 57 19,1 0 0,0 0 0,0 63 2,3 617 22,3 unknown n % 281 54,9 929 53,5 2 7,4 210 70,5 10 5,1 1432 51,7 total 512 1737 27 298 197 2771 1000 900 800 700 vessels 600 jewellery 500 architectural decoration 400 counters 300 production 200 100 0 domestic public ritual burial unknown Figure 96: Column chart showing the quantity of the various commodities per context category The column chart in Figure 96 displays clearly that most black glass artefacts have come from burial and domestic contexts, when excluding all material from unknown provenance, whereas the amount of material from public and ritual contexts is rather marginal. This is also well reflected in the pie charts of Figure 97, but, for instance, the conspicuously large amount of architectural decoration material coming from public buildings makes it necessary to consider the proposed numbers and ratios with some prudence. We therefore discuss hereafter the material per context category. Nevertheless, some brief supplementary conclusions can be deduced from the pie charts. A majority of the vessels have come from domestic and burial contexts. Jewellery and counters, on the contrary, have come from burial contexts and to a lesser degree from domestic environments. Nearly all production material has come from domestic contexts. Yet the presence of production material does not implicitly point to the presence of a workshop at the finding spot. The structure wherein such material has been found could have been a post-destruction layer, a back-filling that functioned as the foundation for a new building, or it could have been purely residual. In total, 298 recorded black glass counters have come from civil settlements, of which 210 items or nearly three quarters of the total have no further provenance information. Of those with more context details, 16 ex. come from domestic areas, 9 ex. from public buildings and places, 6 ex. from ritual deposits and 57 ex. from funerary contexts (Figure 97). This means that, leaving out of consideration those from unknown contexts, two thirds of the black glass counters from civil settlements have come from burials. When analysing more closely the 1737 recorded pieces of black glass jewellery that have come from civil settlement contexts, we see that 251 ex. or 14,5% comes from a domestic milieu, 45 ex. or 2,6% from public contexts, 50 ex. or 2,9% from ritual deposits, 462 ex. or 26,6% from burials, and 929 ex. or 53,5% from unknown find circumstances (Table 80). Table 80: Overview of ratios of the different types of jewellery per period from civil settlement contexts (n = quantity) civil settlements jewellery domestic public ritual burial unknown % % % % % n n n n n bracelets 126 15,0 30 3,6 35 4,2 151 18,0 499 59,3 finger rings 45 31,5 8 5,6 4 2,8 18 12,6 68 47,6 beads 18 3,2 2 0,4 4 0,7 260 46,4 276 49,3 pendants 4 6,5 1 1,6 1 1,6 19 30,6 37 59,7 gems 34 33,3 4 3,9 6 5,9 12 11,8 46 45,1 hairpins 24 82,8 0 0,0 0 0,0 2 6,9 3 10,3 TOTAL 251 14,5 45 2,6 50 2,9 462 26,6 929 53,5 total 841 143 560 62 102 29 1737 181 vessels vessels domestic 23% burial 42% public 2% unknown 55% burial 19% ritual 1% ritual 3% jewellery public 5% jewellery domestic 14% public 3% ritual 3% unknown 53% domestic 31% burial 57% burial 27% architectural decoration ritual 0% burial 4% unknown 7% domestic 33% counters domestic 5% burial 4% ritual 0% domestic 36% public 60% public 3% counters domestic 18% ritual 2% burial 19% burial 65% unknown 71% production 1% public ritual 6% 6% architectural decoration public 56% burial public 0% domestic 50% unknown 5% public 1% ritual 0% public 10% ritual 7% production burial 0% ritual 0% domestic 94% domestic 99% Figure 97: Ratios of the different context categories per commodity type (top: including the material of unknown provenance; bottom: excluding the material of unknown provenance) 182 domestic public hairpins 0% pendants 2% beads 4% hairpins 10% pendants 2% gems 13% beads 7% bracelets 50% gems 9% finger rings 18% bracelets 67% finger rings 18% ritual pendants 4% hairpins 0% pendants 2% beads 8% finger rings 8% gems 3% burial hairpins 0% gems 12% bracelets 33% bracelets 70% beads 56% finger rings 4% Figure 98: Ratios of the different jewellery commodities per context category From the pie charts in Figure 98, it is noticeable that the proportion of the different jewellery commodities fluctuates depending on the context from which the artefact was retrieved. The bracelets constitute the bulk in public and ritual contexts, while the bracelets only represent one third of all jewellery material from burials. The finger rings correspond to about one fifth of both domestic and public contexts, whereas their presence in ritual and burial contexts is minor. The beads are clearly typical burial gifts, as evidenced by the high percentage of beads found in burial sites compared to their low occurrence in domestic, public and ritual contexts. Pendants are not representative because the amount that has been recorded is too little. From the 62 pendants coming from civil settlements, 19 ex. were used as burial gifts, 4 ex. came from domestic contexts, 1 ex. came from public buildings and places, and 1 ex. came from a ritual deposit; most, however, have no detailed provenance information (37 ex.). The gems show a constant share, accounting for about 10% of all jewellery material no matter what the context category, except for the burial context where their presence has only rarely been recorded. However, we assume that it is necessary not only to consider the context character of the gemstones in general, but also to differentiate the gemstones set in finger rings, brooches and helmets [see Chapter 3]. The reason is that the different commodities can be gender-related. The black glass hairpins have not been recorded as having been retrieved from public, ritual or burial contexts. Seeing the value of the pie charts to better understand the link between a type of artefact and its use, we verified the arm rings from a typological viewpoint (Table 81; Figures 99; 100). 183 Table 81: Various types of arm rings in relation to the contexts of provenance civil settlements bracelets domestic public ritual burial unknown Type A 44 10 13 38 189 Type B 12 2 2 13 51 Type C 33 3 1 3 30 Type D 27 12 15 77 207 TOTAL (59 not considered) 116 27 31 131 477 black glass arm rings from civil settlements unknown burial ritual public domestic 0 100 200 Type A 300 Type B 400 Type C 500 600 Type D Figure 99: Stacked-bars showing the quantities of the four arm ring categories per context category Although type B bracelets are not recorded as frequently as type A bracelets, they represent similar ratios with predominant occurrence in domestic and burial contexts. Type C bracelets can be considered characteristic for domestic use, whereas the type D bangles have been recorded most frequently in burials. Type A burial 36% Type B domestic 42% ritual 12% burial 45% public 10% ritual 7% 184 domestic 41% public 7% Type D Type C ritual 3% public 7% burial 8% domestic 21% domestic 82% burial 59% public 9% ritual 11% Figure 100: Comparison of the multiple values of the various bracelet types per context 5.2.1 Domestic context From the 550 recorded artefacts from the domestic context, 364 items have come from glass workshops. The 393 commodity items from domestic contexts displayed in Table 78 and Figure 95 exclude 159 entries of production material or 28,8% of all domestic finds. The vessels represent one fifth of the total with 115 items, whereas the various types of jewellery almost represent one half of the total with 251 items. The architectural decoration and counters represent only a small portion of the totalrespectively, 9 ex. and 16 ex. (Figure 101). domestic context vessels 21% production 29% counters 3% jewellery 45% architectural decoration 2% Figure 101: Functional types of black glass artefacts from domestic civil contexts 5.2.1.1 Vessels The 115 pieces of black glass vessels consist of 16 ex. from Period I, or 13,9%, and 68 ex. from Period II, or 167 59,1%, while none are known from Period III or from Period IV (Figure 102). The Period I vessels from domestic contexts are represented by 10 cast and 5 free-blown shapes which are all tableware. For some the shape remains indefinite, but concerning the cast pieces, the bulk consists either of plates and tables to dish-up food or of cups to drink. The function of the free-blown vessel shapes is less well pronounced due to the high number of indeterminate fragments, but these appear to be of cups. Besides the drinking vessels it includes also flasks to pour drinks or as a toilet bottle. Table 82 demonstrates that most black glass vessels of Period I have been retrieved from urban settlement contexts in the south of France like Fréjus, Marseille, as well as in the west of France like Barzan, but only those from Olbia were from clear domestic 167 31 entries are indefinite because the fragmentation of the items make 27,0% indeterminable. 185 provenance. Also at the Swiss municipium of Avenches and the colonia of Augst/Kaiseraugst vessels in black glass were consumed. Thus far, black glass vessels have been found only rarely at rural settlements such as for st instance the 1 century AD Roman villae of Anthée, Boscoreale, and Porto Potenza Picena. The plate (Form IA.7) from Tongeren (BE) was found in a villa urbana within the municipium. vessels Period IV 0% Period III 0% undefine d 27% Period I 14% vessels Period III 0% Period IV 0% Period I 19% Period II 59% Period II 81% Figure 102: Black glass vessels from domestic civil contexts per period The Period II vessels from domestic contexts are not only more numerous, but the shapes are almost entirely restricted to the carinated beaker, Form IIB.1, and the cup with bulbous body, Forms IIB.4-5. Furthermore, its particular presence in rural dwellings stands in contrast to the Period I vessels which have been found mainly in urban centres. It is clear that for both periods, multiple examples of black glass vessels are available from diverse types of civil settlements, including villae, vici, municipia, and coloniae. Table 82 demonstrates that most Period II vessels come from rural settlements, whether it be rural farmsvillae or villa-like settlementsor dwellings within rural settlements such as vici or diverticuli. Noticeable is also the amount of indeterminate vessel fragments catalogued from within these domestic civil settlements due to its fragmentarity. Nonetheless they most likely are also to be determined as either from the carinated beaker or the bulbous cup. This assertion disapproves a previous statement that the Period II vessels all came from burials (Cosyns, Hanut 2005). In particular, the carchesium beaker, Form IIB.1, appears regularly at Roman villae in Belgium and the Netherlands such as Mortsel (BE) [cat.no.450]; Relegem (BE) [cat.no.496]; Treignes (BE) [cat.nos.582-89]. Besides the rural domestic context of villae, the carchesium has also been reported in rural small towns or vicie.g. a cellar in Braives (BE) [cat.no.223]; the well H5 of the tannery at Liberchies (BE) 168 [cat.no.418]; Tienen (BE) [cat.nos.541-42; 544; 546-47] a well in Velzeke [cat.nos.594-99] and a cellar in VieuxVirton (BE) [cat.no.614]. Further, some examples of black glass vessels have been excavated at larger towns such as the municipia and the colonia, such as Colchester (UK), with at least a rim fragment of a carchesium [cat.no.4027]; Augst-Kaiseraugst (CH) [cat.nos.3754] with the bulbous cup (Form IIB.4-5) that was more popular in the Alpine region. Examples of the two types of bulbous cups (Form IIB.4-5) came from the villae of Anthée (BE); Matagne-la-Petite (BE); Lauw (BE); Mettet (BE); Roly (BE) A rim fragment of a one-handled jug with funnel mouth (Form IIB.11) came from a cellar in Straten (BE). The rod-formed candlestick unguentaria with a triangular solid body, Form IID.1, found in Elkab (EG) during the 1938 excavation campaign of the Royal Museums of Art and History (MRAH-KMKG), Brussels come from the Greek-Roman village within the large mud-brick walls in Elkab, where they were all found together in a dwelling in the southeastern corner inside the city wall close to the 169 SE-gate according to the information from the notebooks on the 14 February 1938. Most pieces from the excavations of Tienen-Grijpenveld, if not all, are regarded as coming from ritual deposits with the exception of those from the flattened tumulus (personal communication by Marleen Martens - 2006). 169 Although the material entered the museum since 1952 the material form the excavations have not been the issue of any context analysis. 168 186 Table 82: Vessels from domestic contexts Period I cat.no. provenance site context shape typology 179 565 1330 2107 2733 2765 2767 2769 2775 3173 3210 3896 Anthée (BE) Tongeren (BE) Olbia (FR) Waldorf (DE) Boscoreale (IT) Pompeii (IT) Pompeii (IT) Pompeii (IT) Porto Potenza Picena (IT) Ribnica (SL) Tarragona (ES) Ephesus (TR) rural urban urban rural rural urban urban urban urban rural urban urban villa d‘Anthée villa at Hondstraat insula VI villa villa domus ‗ara massima‘ domus domus ‗lesbiani‘ villa ‘casa valentini‘ mansio? -, rubbish pit domus (atrium well) plate plate cup skyphos table bowl cup cup cup small dish plate flask Form IA.8 Form IA.8 Form IB.1 Form IB.7 Form IA.7 Form IA.1 Form IA.1 Form IA.1 Form IA.1 Form IA.12 Form IA.8 Form IB.3 174 178 Amay (BE) Anthée (BE) rural, vicus rural beaker cup Form IIB.1 Form IIB.5 224 Brecht (BE) rural beaker Form IIB.1 258 334 430 468-47 449 450 508 519 582-89 594-99 868-71 Elewijt (BE) Lauw (BE) Liège (BE) Matagne-la-Petite (BE) Mettet (BE) Mortsel (BE) Roly (BE) Straten (BE) Treignes (BE) Velzeke (BE) Elkab (EG) rural, vicus rural rural rural rural rural rural rural rural rural, vicus urban domus villa ‗Potstal‘-farm – ‗deep litter house' domus villa villa villa villa villa villa villa villa domus domus beaker cup beaker beaker beaker beaker cup jug beaker beaker unguentarium Form IIB.1 Form IIB.4 Form IIB.1 Form IIB.1 Form IIB.1 Form IIB.1 Form IIB.5 Form IIB.11 Form IIB.1 Form IIB.1 Form IID.1 1367-69 Roussy-le-Village (FR) rural, diverticulum domus cup; beaker Forms IIB.1;4-5 1613-20; 1622-24 Sainte Menehould (FR) rural glass workshop beaker; amphoriskos Forms IIB.1; IIC.1 1946 Mainz-Finthen (DE) rural villa cup 2902-7 Hoogeloon (NL) rural villa cup; beaker 3490;3504 Augst (CH) domus cup 3743;3754; 3762-63; 3773;3778 Kaiseraugst (CH) urban, colonia urban, colonia Form IIB.5 Forms IIB.1;4 Form IIB.17 domus glass workshop cup; aryballos; beaker Forms IIB.45;17;20;21 3776 Kaiseraugst (CH) glass workshop - Regio 17C inkwell Form IIB.18 Period II 5.2.1.2 urban Jewellery Considering the 251 pieces of black glass jewellery from domestic contexts as a whole (Table 79) is too general and need at least to be split up according to the functional specificityarm ring, finger ring, ...  (Table 80; Figure 98). We here verified whether additional information can be obtained from a further breakdown of the specific jewellery commodity type per typological class. 5.2.1.2.1 Arm rings Despite the great proportionate similarity between the bangle types A and B in general (Figure 55) and the ratio established from the 126 entries from domestic contexts (Figure 103), there is a noticeable difference between the general shares of the Types C and Drespectively 8% vs 29% and 48 vs 23%. These proportions show that a distinctive number of Type C bangles have a domestic use whereas the bulk of Type D bangles are retrieved from a different context category [see below]. 187 arm rings in domestic context Type D 23% Type C 29% Type A 38% Type B 10% Figure 103: Pie-chart showing the ratios of the different bangle types from domestic contexts Table 83: Arm rings from domestic contexts cat.no. provenance 176-77 Anthée (BE) 183 Arlon (BE) 194 Basse-Wavre (BE) 212 Braives (BE) 256 Elewijt (BE) 272 Fernelmont (BE) 297-99 Hamois (BE) 331;333 Lauw (BE) 357;367;384 Liberchies (BE) 432 Maillen (BE) 462 Nivelles (BE) 498-507 Roly (BE) 577-81 Treignes (BE) 590 Vechmaal (BE) 592 Velzeke (BE) 616;619 Wancennes (BE) 816-17;842;849-50 Dush Kysis (EG) 1019 Arles (FR) 1029 Carhaix (FR) 1053 Dehlingen (FR) 1413-17;1419-41;1449-50; Sainte Menehould (FR) 1455-57;1463;1469;1472; 1486-1504;1506-13;1550 site villa dwelling villa dwelling dwelling villa villa villa dwelling villa villa villa villa villa dwelling villa dwellings villa workshop context rural urban rural rural rural rural rural rural rural rural rural rural rural rural rural rural rural urban urban rural rural shape A1 (2 ex.) A1 C2 A1 A3 indefinite B2 (2 ex.); C3 A1; C3 A3(2 ex.); B2 indefinite A2 A1; A3; B1; B2; C2; D1 (5 ex.) A2; A3; D1 (2 ex.) A5 C2 A2; C3 D1 (3 ex.); indefinite (2 ex.) A5 B4 C1 A1 (7 ex.); A2; A3 (3 ex.); A4 (2 ex.); A5 (5 ex.); A6 (3 ex.); B2 (3 ex.); C1 (13 ex.); C2 (4 ex.); C3 (9 ex.); D1 (9 ex.);indefinite (4 ex.); waste D1 D1 A2/3 (2 ex.); B1; B5 (2 ex.); D2 (2 ex.) 1761 2051 2423-28;2430 Aldenhoven (DE) Trier (DE) Jalame (IL) dwelling workshop workshop rural urban rural 2750;2752-53 2997-98 3076 3014;3029;3031-54;3056-58;306068;3070-75 Luni (IT) Voerendaal (NL) Braga (PT) Braga (PT) dwelling villa dwelling workshop urban rural urban urban B1; D1; D2 A1 A6; D7 D1 (23 ex.); D2 (19 ex.); D7 3181 3437 3656 3781-83 Cádiz (ES) Augst (CH) Kaiseraugst (CH) Liestal-Munzach (CH) dwelling dwelling dwelling villa urban urban urban rural A1 A3 D1 A2 (2 ex.); A5 Of the 129 arm rings that have been recorded from domestic contexts, a number of entries are related to glass workshops: 64 ex. to that of Les Houis near Sainte Menehould (FR), 7 ex. to that of Jalame (IL), 1 ex. to that of ‗Äussere Reben‘ at Kaiseraugst (CH), and 1 ex. to that of ‗Palais Kesselstatt‘ at Trier (DE). Depending of the 188 region most other material has come from rural settlementsvillaein the north-western provinces or from the dwellings of urban centres in the Mediterranean area (Table 83). Table 84: Arm rings from the black glass manufacturing workshops versus domestic consumers‘ contexts typology Sainte Menehould Les TrierJalame Bragadomestic Houis Kesselstatt Fujacal contexts Type A 21 0 2 0 23 A1 7 8 A2 1 2 5 A3 3 6 A4 2 A5 5 3 A6 3 1 Type B 3 0 3 0 7 B1 1 2 B2 3 4 B3 B4 1 B5 2 Type C 26 0 0 0 7 C1 13 1 C2 4 3 C3 9 3 Type D 9 1 2 44 16 D1 9 1 23 15 D2 2 19 1 D3 D4 D5 D6 D7 1 1 D8 undefined/waste 5 4 TOTAL 64 1 7 44 57 total 46 15 8 9 2 8 4 13 3 7 1 2 33 14 7 12 72 48 22 2 9 173 Considering that the bangles of Types A-C are to be related to Period III and Type D bangles to Period IV, Table 84 demonstrates that the workshop of Les Houis manufactured any type of bracelet material during Periods III and IV. Nevertheless we may say that the workshop mainly produced Types A and C seeing that all variants of both classes are present within the waste material. Concerning the Type B bangles we only inventoried the open B2 variant with 3 examples. 5.2.1.2.2 Finger rings Of the 45 finger rings retrieved from domestic civil contexts, only 5 have come from consumers‘ dwellings: one from a cave fill in a house at the vicus of Braives (BE), one from the villa of Lauw (BE), one from the villa ‗Fosse Levrette‘ at Nivelles (BE), and two from the villa at Great Witcombe (UK). All others have come from glass workshops: 14 ex. from Les Houis near Sainte Menehould (FR), 15 ex. from Medizinische Klinik at Bonn (DE), and 11 ex. from Palais Kesselstatt at Trier (DE). 5.2.1.2.3 Beads Only 18 beads have come from domestic contexts, and of those, 12 have come from workshops in urban centres: Arles (FR) (1 ex.), Trier-Palais Kesselstatt (DE) (10 ex.), and Kaiseraugst (CH) (1 ex.) (Table 85). The black glass beads from domestic contexts are characteristic of the late Roman Periods III-IV, but considering the datable rd th contexts, everything seems to date from the 3 –4 centuries AD [see Chapter 4]. 189 Table 85: Beads from domestic contexts cat.no. provenance site context shape 632-33 835 1014 1373 Dracevica (BS) Dush Kysis (EG) Arles (FR) Saint-Denis (FR) rural rural urban, colonia rural domus domus ‗Sigma House‘ workshop villa ‗Cod‘Esprou‘ spacer-bead barrel-shaped flat ellipsoidal-shaped annular ‗melon bead‘ 2025;2029;2033;2035; 2038;2044;2054;2057;2059;2065 Trier (DE) urban, colonia workshop - ‗Palais Kesselstatt‘ annular,‘ globular,‘ cylindrical,‘ barrel-shaped 3642 4166-67 Kaiseraugst (CH) Nor‘Nour (UK) urban, colonia rural workshop - cylindrical globular-segmented 5.2.1.2.4 Pendants No explicit deduction can be done from the context analysis of pendants from domestic provenance because only 4 pendants have been recorded that were retrieved from a domestic civil settlement. Two came from a workshop, an amphora-shaped pendant from Trier-Palais Kesselstatt (DE) [cat.no.2056] and a jug-shaped pendant from Jalame (IL) [cat.no.2429]. A barrel-shaped pendant was found in the domus ‗casa de Hyppolitus‘ at Alcalá de Henares (ES) [cat.no.3177], and one of indefinite shape was reported from Building 3 at the town of Dush Kysis in the Kharga-oasis (EG) [cat.no.852]. 5.2.1.2.5 Gems In total, 34 gems have been recorded from domestic contexts, of which 5 were still set in a finger ring and 1 in a brooch. The other 28 are loose finds and on the basis of the shapeelliptical with flat surfaces and bevelled edge (Henig type F2 or F4) these are to be considered as gems from finger rings that got lost after being broken off when the finger ring fell on the ground. Most likely, this was the result of getting into fissures of a timber floor, seeing that the rooms within the rural dwellings, villae or domus, were not paved with concrete or mosaics but a wooden floor covering an earthen level (Table 86). Most are blue-on-black gems (28 ex.) imitating the nicolo gemstones, also called nicolo paste gems, and the remaining 6 are monochrome black. Only the gem from Nor‘Nour on the Isles of Scilly (UK) was set in a brooch. The gem is a characteristic plain conical stone in monochrome black glass, and the brooch is an oval gold-plated bronze brooch decorated with miniature 'accolade' stamps [see conclusion in this chapter]. Defining the material of the finger rings in view of the contextual analysis does not provide any additional information due to the variety of setting materials: 3 gems were set in a bronze ring, 1 in a silver ring, and 1 in an iron one. Of the 20 ex., most of the material came from the jewellery workshop excavated at Medizinische Klinik, Bonn (DE) in the canabae close to the north gate of the legionary camp, castra bonnensia [see Chapters 3; 6 and 7]. Table 86: Gemstones from domestic contexts cat.no. provenance 172 221 257 454 593 1799-1803;1805-8;1812;181516;1819;1821;1826-29;1831 2991-92 3964 4052 4061 4165 4202-4 190 site context shape Amay (BE) Braives (BE) Elewijt (BE) Neerharen-Rekem (BE) Velzeke (BE) rural, vicus rural, vicus rural, vicus rural rural, vicus loose finger ring finger ring finger ring loose Bonn (DE) urban, canabae Valkenburg (NL) Carlisle (UK) Fordingbridge (UK) Gorhambury (UK) Nor‘Nour (UK) Wilcote-Shakenoak Farm (UK) urban urban rural rural rural rural domus domus villa villa domus jewellery workshop villa domus villa villa domus villa loose loose loose finger ring finger ring brooch loose 5.2.1.2.6 Hairpins Very few black glass hairpins have been recorded (43 ex.) and even fewer contain context information (26 ex. 170 from civil settlements). Most black glass hairpins have come from domestic contexts ; however, the bulk has come from artisanal contexts. Twenty-three records came from the workshop of Les Houis nearby Sainte Menehould (FR), and possibly the pieces from Fujacal at Braga (PT) and Bon-Villers at Liberchies (BE) have to be linked with a workshop as well. Only the one from the villa of Shakenoak Farm nearby Wilcote (UK) is purely domestic. The Shakenoak piece is a complete hairpin in bronze wire topped with a globular head in black glass rd [cat.no.4201] and was excavated in an unsealed deposit south of Building B, which is dated in the later 3 century AD (Harden 1971, 106, no.152, fig.45:69). 5.2.1.3 Architectural decoration Due to the poor availability of published material and the complexity of investigating the topic individually, a diagnostic contextual analysis for this functional type was unattainable. Furthermore, too few contexts with architectural decoration material in black glass could be recorded within the limited time of the present research project. In total, only 8 entries of architectural decoration material are known to have come from domestic civil settlement contexts, of which 7 entries are from Period I and 1 from Period II (Table 87). The inventoried material came from several Roman villae in Italy and one in France. Black glass tesserae were generally used to decorate wall mosaics within fountains and bath structures of the villa complexes of wealthy citizens and emperors. Also, the laborious application makes these artefacts self-evident markers for public buildings, luxurious domestic spaces, or dumps at building yards from such sites or from glass workshops. Almost solely tesserae have been recorded. It is odd to miss the contemporaneous Period I twisted rods used to border wall and vault mosaics which were used, in particular, in middle class dwellings, but this is probably due to very superficial heuristics. The use of intarsia in the villa of Lucius Verus in Rome is a technique to make mosaics in opus sectile and demonstrates its use from Period II. Table 87: Overview of architectural decoration material in domestic contexts cat.no. place site 1625 Saint-Émilion (Gironde) (FR) villa du Palat 2778 Rome (IT) ‗villa dei Centroni‘, natatio 2779 Rome (IT) ‗domus aurea‘, nympheum 2802 Rome (IT) 'villa of Lucius Verus' 2815 Sperlonga (IT) ‗villa of Tiberius‘, grotto 2816 Tivoli (IT) ‗villa of Hadrian‘, nympheum 2817 Tivoli (IT) ‗villa of Hadrian‘, serapeum 2818 Tivoli (IT) ‗villa of Brutus‘, nympheum 5.2.1.4 material tesserae tesserae tesserae intarsiae tesserae tesserae tesserae tesserae period I I I II I I I I Counters The 16 counters in black glass that have been recorded from domestic contexts do not provide any valuable information within the context analysis. 170 No black glass hairpin has hitherto been recorded in a public or ritual context, while only two examples are known to have come from a burial context. 191 5.2.2 Funerary context A total of 627 artefacts have been recorded from funerary contexts, of which 97 items are vessels, 469 items are jewellery, 1 item is an architectural decoration and 58 items are counters, while no production material has been 171 retrieved from burials or any other funerary context (Figure 104). burial context architectural decoration 0,2% production 0% counters 9,3% vessels 15,5% jewellery 74,8% Figure 104: Functional types of black glass artefacts from Roman burials 5.2.2.1 Vessels The 97 pieces of black glass vessels from known civil funerary contexts consist of 15 pieces of undefined tableware due to their very fragmentary state of preservation. Twenty-one pieces are from Period I burials and 61 ex. or 62,9 % from Period II, while none are from Period III and Period IV burials (Figure 105). vessels Period IV 0% Period III 0% undefine d 15% Period II 63% vessels Period IV 0% Period III 0% Period I 22% Period I 26% Period II 74% Figure 105: Pie charts of the vessels from funerary contexts (left: considering all entries, right: omitting the entries of undefined date) Table 88 clearly demonstrates that the majority of the Period I burials containing vessels in black-appearing glass has come from Mediterranean tombs. There is no uniformity of shapes, as the 23 vessels represent 16 different shapes, and little information is known about the deceased because cremation makes that gender and age has not been determined (Table 90). Too few information is at hand from the burial contexts and due to the cremation practice no information is hitherto available concerning the gender or age of the deceased. It is also not clear if the deceased was rich, but the number of burial gifts and the richness of the ensemble makes it clear that the family of the deceased must have endured social stress (van Lith, Randsborg 1985) and, presumably because of their status within the community, must have felt obliged to give the departed relative a sumptuous burial rite with 171 The remaining two pieces are indefinite shapes: one from Augsburg (DE) [cat.no.1772] and one from Kaiseraugst (CH) [cat.no.3720]. 192 a lot of gifts to impress the populace. It can be tableware such as bowls, plates, cups and jugs, but equally the material is storage ware, for instance re-used as an urn, or even toilet ware, i.e. unguentaria. Out of 21 black glass vessels, 7 are cast, 13 are free-blown and 1 is mould-blown. Only a few vessel shapes reappear several 172 times: 1) the shallow hemispherical bowl, Form IA.1, 1 ex. in the tumulus of Riemst (BE) ; 1ex. in the ‗Tumba de Bahìa Blanca at Cádiz (ES); and 1 ex. in the cemetery ‗En Chaplix‘ at Avenches (CH); 2) the unguentarium with bulbous body, Form IB.3, 1 ex. from tomb 1 at the ―Ex fabbrica Siberia‖ cemetery of Montebelluna (IT); 2 ex. from the Tomba di Salizzole at Verona (IT). 173 Table 88: Tombs with black glass vessels from Period I cat.no. site 1;4570 2585 2586 2758-60 2761 2762 2805 2830-31 497 A4480 A4483 Dürres (AL), tomb V69 Adria-Canal Bianco (IT), tomb 34 Adria-Canal Bianco (IT), tomb 147 Montebelluna (IT), tomb 1 Montebelluna (IT), tomb 3 Montebelluna (IT), tomb 4 S. Elena di Melma (IT), tomba di Silea Verona (IT), tomba di Salizzole Riemst (BE), tumulus Murviel-lès-Montpellier (FR), tomb 1 Saint-Paul-Trois-Châteaux (FR), tomb 7 burial type burial gifts glass objects black glass vessels cremation cremation cremation cremation cremation cremation cremation cremation cremation cremation cremation 5 13 6 28 25 12+ 5 10+ 50 10 14 (38) 4 6 6 13 17 8+ 3 10 12 8 7 (21) 2 1 1 3 1 1 1 2 1 1 1 A wide range of complete pieces comes from the Period II cremation tombs. A preliminary list has not only demonstrated that most of the Period II vessels in black glass from funerary contexts are dated between 170/180 and 230 AD, but also that they are found in rich cremation tombs containing an amount of luxurious grave goods (Cosyns, Hanut 2005, 117, tab.1). An updated list comprising 55 pieces from 36 different contexts does not modify either assertion (Table 89). Table 89: Tombs with black glass vessels from Period II cat.no. site typology n 173 209 210 321 539;543;545;548 1028 1049-51 1847 2882 2979 515 A4618 1052 2898 2012-14 1075 1073 2881 2880 1930 2987 2986 Amay (BE) Bossut-Gottechain (BE) Braffe (BE) Jambes (BE) Tienen (BE) Boulogne-sur-Mer (FR) Cutry (FR) Cologne (DE) Esch (NL) Nijmegen (NL) Schaarbeek (BE) Trier (DE) Cutry (FR) Heerlen (NL) Siesbach (DE) Faulquemont (FR) Faulquemont (FR) Esch (NL) Esch (NL) Krefeld-Gellep (DE) Stein (NL) Stein (NL) 1 1 1 1 1 1 3 1 1 1 1 1 1 1 3 1 1 1 1 1 1 1 Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.1 – carchesium Form IIB.2 Form IIB.2 Form IIB.3 Form IIB.4 Form IIB.4 Form IIB.5 Form IIB.6 Form IIB.6 Form IIB.7 Form IIB.8 Form IIB.9 Form IIB.10 chronology IIIa - IIIc AD (200-260/70) V-VIA AD IId - IIIa AD (170 - 230) IId AD (170-200) IId AD (170/80-200) III AD IId - IIIA AD (180 – 250)174 IIB – IIIA AD IId AD (170-200) IId - IIIA AD (180 - 250) IId - IIIA AD (180-230) IV AD IId - IIIA AD (180 - 250) IIc AD (160-180) IIc AD (173-174) IIB AD IIB AD IIc AD (160-170/80) IIc AD (160-170/80) III AD IIIBAD IIIBAD 172 About the shallow bowl from the tumulus of Riemst (BE) (de Schaetzen 1950), we must bear in mind that the context is dated around the end of the first quarter of the 2nd century AD. Seeing the fragmentary condition and the severe weathering crust covering nearly the entire surface compared to the other well-preserved glass vessels from this rich burial, it is most likely that the much older dated vessel typei.e. Claudian to early Flavianshould be regarded as an heirloom [see Chapter 4: Chronology]. 173 In total 23 items have been reported from burial context, but the full context information is not always available. Hence excluded from the table are the pieces from Bahnasa (EG) [cat.no.812]; Magdalensberg (AT) [cat.no.117]; Avenches (CH) [cat.nos.3614;3593;3597;3617]; Pohlov Gradec (SL) [cat.no.3152]; Zara (HR) [cat.no.772]; Cádiz (ES) [cat.no.3182]. 174 The context of tomb 176 is erroneously dated at the end of the 1st century AD and must have been unwanted switched with that of the 1st century AD tomb 440 containing a Hofheim cup (Isings Form 12) which got dated at the end of the 2nd–early 3rd century AD (Liéger 1997, 66) 193 2879 3154 1869 3622-23 3600 1214 1962 2984 Esch (NL) Ptuj (SL) Elsdorf-Esch (DE) Avenches (CH) Avenches (CH) Lillebonne (FR) Morbach-Wederath (DE) Nijmegen (NL) 1 1 1 2 1 1 1 1 Form IIB.11 Form IIB.11 Form IIB.14 Form IIB.16 Form IIB.16 Form IIB.19 Form IIC.1 Form IIC.1 IId AD (170-200) II AD IIIb-IIIc AD (230-270) IId AD (c.180-200) mid I AD (40-70) IIB – IIIA AD IIB AD IIB AD The burial context and the undamaged vessels (Table 89) are evidence to assume that the black glass vessels were placed into the tomb at the end of the burial rite as secondary burial gifts. Most likely all intact pieces without provenance information originate from burial contexts. Rarely vessel fragments deformed by the heat of the pyre have been excavated. The only recorded examples are the 48 fragments of at least three small bulbous jars (Form IIB.4) from the Siesbach barrow (Abegg 1989, 209, Taf.16, nos.312; 319-326) and a carchesium (Form IIB.1) said to be from a tomb at Auvelais (BE) (unpublished) [cat.no.192]. These few examples demonstrate the fact that black glass vessels were used occasionally as a primary burial gift placed on the pyre along or on top of the deceased. Decorated vessels with applied glass trails have come from rich tombs, as can be observed from various examples: 1) the amphoriskos from Elsdorf-Esch (Gaitzsch 1999, 79) and the jug from Weilerwist-Hausweiler (Follmann-Schulz 1992, 21-22, no.10); 2) with straight upright festoon pattern, e.g. Avenches (Martin Pruvot 1999, 230, pl.149, n°1490), Heerlen (Brouwer 1991, 47, no.16) and Nijmegen (Isings 1964, 176, fig.4); 3) with diagonal upright festoon pattern, e.g. Siesbach (Abegg 1989, 209, fig.17), Cologne (Isings 1964, 176, fig.5). Another type of decoration is the indented body, e.g. Faulquemont (Cabart 2005, 20-21, Cosyns et al. 2006a, 37, fig.6), Mettet in the Musée Archéologique, Namur [cat.no.449] (Mahieu 1919, 163; Isings 1957, 47; Isings 1964, 179, no.18). Various burials from civil settlements in Belgium, the Netherlands, France and Germany yielded intact carchesium beakers (Form IIB.1) in villae, at rural settlements as well as at urban centrese.g. Amay (BE) [cat.no.173], Jambes (BE) [cat.no.321], Lavacherie (BE) [cat.no.335], Tienen (BE) [cat.nos.539; 543; 545; 548], Wancennes (BE) [cat.nos.621-622], Abbeville (FR) [cat.no.998], Boulogne-sur-Mer (FR) [cat.no.1028], Cutry (FR) [cat.nos.1049-1051], Cologne (DE) [cat.no.1847], Esch (NL) [cat.no.2882], Nijmegen (NL) [cat.no.2979]. Bulbous cups (Form II.4-5) came from the barrow at Siesbach (DE) [cat.nos.2012-2014], the cemetery ‗En Chaplix‘ at Avenches (CH) [cat.nos.3615; 3625-27] and Faulquemont (FR) [cat.no.1075]. A rim fragment of a one-handled jug with funnel mouth (Form IIB.11) came from a cellar in Straten (BE) [cat.no.519]. Striking is the total absence, hitherto, of these diverse decorated black glass vessels in Romano-British sites. An intact grape-shaped amphoriskos has been excavated in the cemetery of Nijmegen-Hees, the Netherlands (Figures 41.6; 43) (Isings, van Lith 1992, 17, fig.19). Another (fragmentary) one came from the cremation tomb AG14 of the Roman cemetery at Wederath-Belginum (DE). Interesting to note is that the same tomb yielded the remains of a second very similar vessel in ‗very deep olive green glass‘ (Kaiser 2006, 39, pl.729). At present, all unguentaria with a flattened circular body, Form IIIB.2 (Figure 40.2) in the Rhine region from known contexts have come from burials and, in particular, from inhumation tombs. It also appears to be a typical gift for women and girls when gender and age can be recognized (Table 90). We may therefore assume that the one in black glass was a burial gift for an adult woman, even if the skeletal remains do not allow a determination. 194 Table 90: Tombs with Form IIIB.2 vessels (―-― = undefined) cat.no. site hue burial type gender age reference 1929 Krefeld-Gellep (DE) – tomb 5530 so-called black inhumation - - Pirling 2003, 69, no.5530:4, pls.6:3a-b; 133:2 not catalogued Krefeld-Gellep (DE) – tomb Kr9 pale yellowish-green inhumation female - Pirling 2003, no.Kr9:3, pl.3:3 not catalogued Krefeld-Gellep (DE) – tomb 3198 pale yellowish-green inhumation - - Pirling 1989, no.3198:15, pl.42:9 not catalogued Cologne (DE) – Jakobstrasse tomb 28 decolourized inhumation - - Friedhoff 1991, 219, no.28:2, pl.63 not catalogued Cologne (DE) – Jakobstrasse tomb 227 dark purple inhumation female adult Friedhoff 1991, 282-284, no.227:3, pl.95 not catalogued Cologne (DE) – Jakobstrasse tomb 254 decolourized inhumation female child Friedhoff 1991, 329, no.254:1, pl.102 not catalogued Cologne (DE) – Jakobstrasse tomb 327 dark blue inhumation female? - Friedhoff 1991, 329, no.327:2, pl.115 not catalogued Strasburg (FR) - dark blue inhumation - - Arveiller-Dulong, Arveiller 1985, 138, no.303 not catalogued Trier (DE) – Pallien tomb 17 decolourized inhumation female child Goethert-Polaschek 1977, 137, no.765 not catalogued Trier (DE) – St. Matthias tomb 112 decolourized (2 ex.) green inhumation - - Goethert-Polaschek 1977, 137, nos.767-69 not catalogued Trier (DE) – Biewerer Strasse sarcophagus 1 decolourized inhumation - - Goethert-Polaschek 1977, 137, no.770 None of the catalogued rod-formed balsamaria and unguentaria from Period IV has come from a burial context. It is, however, necessary to execute an exhaustive heuristics and research on location in Syria, Israel and Jordan, as it is evident that there is already such material retrieved from burials. 5.2.2.2 Jewellery In total, 469 pieces of jewellery in black glass are known to have come from civil funerary contexts. Strikingly only a few are from Periods I and II with respectively 5 ex. or 1,1 % and 6 ex. or 1,3 % of all black glass jewellery. Despite the bulk remains of indefinite date, Table 91 clearly demonstrates that the mainstream of the black glass jewellery can be reduced to Periods III and IV with in that order 24 ex. or 5,1 % from Period III and 74 ex. or 15,8 % from Period IV (Figure 106). Table 91: Overview of ratios of the different types of black glass jewellery per period from civil/funerary contexts (n = quantity) Period I Period II Period III Period IV indefinite total % % % % % n % n n n n n arm rings 0 0 0 0 1 0,7 32 21,2 118 78,1 151 32,2 finger rings 0 0 2 11,1 3 16,7 1 5,6 12 66,7 18 3,8 beads 5 1,9 2 0,8 12 4,5 29 10,9 217 81,9 265 56,5 pendants 0 0 0 0 4 19,0 8 38,1 9 42,9 21 4,5 gems 0 0 2 16,7 4 33,3 4 33,3 2 16,7 12 2,6 hairpins 0 0 0 0 0 0 0 0 2 100 2 0,4 TOTAL 5 1,10 6 1,3 24 5,1 74 15,8 360 76,8 469 100 195 hairpins gems Period I Period II pendants Period III beads Period IV indefinite finger rings arm rings 0 50 100 150 200 250 300 hairpins gems Period I Period II pendants Period III beads Period IV indefinite finger rings arm rings 0% 20% 40% 60% 80% 100% Figure 106: Stacked bar charts showing the ratios per period of the different types of black glass jewellery from civil/funerary contexts in amount and in proportion 5.2.2.2.1 Arm rings Table 92 demonstrates clearly the presence of one or maximum two bracelets worn by a girl or adult woman [see Chapter 7]. The young adult woman from inhumation tomb 194 of the late Roman cemetery of Oudenburg (BE) wore a black glass bangle on her left wrist and a bronze one on her right arm (Mertens, Van Impe 1971, 208-209, pl.LX:194:4). A similar ‗dress code‘ can be observed at the inhumation tombs 19a, 1050 and 1239 of the ‗Südostfriedhof‘ cemetery at Intercisa (HU). Bracelets in other materials were also worn on the left wrist, but black glass, jet and bone bangles have never been attested on the right wrist. Apparently, only bronze bangles are recorded on the right wrist. The curiosity of the Orpington case is the peculiarity of the bangle with pressed decoration reminiscent of bronze bracelets (Type B5). It was found in grave 2 of an early Anglo-Saxon cemetery within a late Roman site (Tester 1968, 125-150). It contained the remains of an inhumated woman aged 30. Although the tomb was partly destroyed by grave 3, one fragment of the bangle is thought to have been found in situ beneath the right humerus, suggesting that it was worn just above the elbow. The other fragments were found scattered in the lower grave filling. The occurrence of Roman (fragmented) glass bangles within Anglo-Saxon and Merovingian tombs is attested from early research (Haevernick 1968; De Witte 1977) as well as from recent 196 excavations, such as the material from the cemeteries of Gottechain-Bossut (BE) [cat.nos.202-5;207] 176 (unpublished) and Broechem (BE)175, likewise various other commodity types. Table 92: Black glass bangles from inhumation burials (n = quantity; C = civil settlement; M = military settlement) cat.no. site context type location orientation gender n type of arm ring of tomb 3388 3687 A4607 4120 4170 280-81 278-79 482-83 479 517-18 2210 2194 2204 1058 1059 1004 1005 Augst (CH) tomb 1a Augst (CH) tomb 33 Canterbury (UK) family burial London (UK) tomb B168 Orpington (UK) tomb 40 Furfooz (BE) tomb 22 Furfooz (BE) tomb 2 Oudenburg (BE) tomb 100 Oudenburg (BE) tomb 194 Spontin (BE) tomb C Dunapentele (HU) tomb 19a Dunapentele (HU) tomb 1050 Dunapentele (HU) tomb 1239 Epiais-Rhus (FR) tomb S95 Epiais-Rhus (FR) tomb S159 Amiens (FR) faubourg Beauvais Amiens (FR) rue de Corbie age C D1 1 - - female adult C A1 1 - - female adult C D1 1 left wrist N-S girl child C A1 1 left wrist E-W female? adult C B5 1 left wrist SE-NW female adult M D1 2 left wrist N-S girl child M D1 2 left wrist N-S girl child M D1 2 - E-W - - M B2 1 left wrist N-S female 25-30 C D1 2 wrist E-W female adult C A1 1 left wrist E-W female adult C A1 1 left wrist E-W female adult C A1 1 left wrist E-W girl? child C A1 1 outside coffin NW-SE - - C D1 1 chest N-S female adult C D1 1 - - - - C D1 2 - - girl? child In some cases, the black glass bracelet was not worn by the deceased at the moment of interment, but it was put in a wooden box or eventually in a bag. The boxes are mainly located beside the left shoulder or at the feet and contain intact jewellery. Some burials demonstrate that these boxes were placed beside the left shoulder or at the feet. It appears that all burials yielding boxes containing complete bangles contain a female body and are from clear Roman contexts of Periods III-IV, whereas the tombs containing purses at the height of the waist, always holding fragmentary pieces, are only found in male tombs and erroneously considered from Merovingian burials th th and dated in the late 5 and 6 century AD as some burials from Frénouville (FR) and Kaiseraugst (CH) have to th th be dated late 4 – first half 5 century AD. Table 93: Roman cemeteries containing black glass bracelets site number of tombs tombs with black glass bangles Oudenburg (BE) Tongeren (BE) London (UK) Epiais-Rhus (FR) Frénouville (FR) Kaiseraugst (CH) Krefeld-Gellep (DE) >200 >250 >650 500 >650 >1300 >6500 2 0 2 2 2 1 5 < 1% 0% 0,3% 0,2% 0,3% < 0,1% < 0,1% date range of cemetery reference III-Va AD Id-IV AD Ib-Va AD II-IV AD IIIB-V AD IIIB-V AD IIIB-V AD Mertens, Van Impe 1976 Vanvinckenroye 1994 Barber, Bowsher 2000 Vanpeene 1998 Pilet 1980 Martin 1976 Pirling, Siepen 2006 175 Personal communication by Rica Annaert. Also a base fragment of a Period II carchesium in black glass (Form IIB.1) was found in tomb 297 of the late 5 th–mid 7th century AD cemetery at Gottechain-Bossut [cat.no.209] together with a Period IV Type D1 arm ring [cat.no.205]. 176 197 The brief and partial inventory on black glass bangles from civil cemeteries displayed in Table 93 demonstrates that black glass bracelets were only used on very rare occasions as burial gifts at larger settlements and thus to rd be considered a fairly uncommon consumer good. When taking into account the burials from the 3 century AD onwards –when the glass bangles got into use– we see no substantial change in ratios. For instance the Eastern Cemetery at London (UK) yielded 457 burials from the late Roman period corresponding with 66 % of all burials, resulting in a ratio of 0,43 % for the burials containing a black glass bangle (Barber, Bowsher 2000, 12). Furthermore, Bruno Barber and David Bowsher listed the number of burials with goods on the one handcremation 26 and inhumation 129 resulting in a ratio of 1,29 %and on the other hand the number of female inhumation burials with goods28 burials equals a ratio of 7,14 %. However, it is important to point to a huge number of Roman cemeteries with a total absence of (black) glass bangles, as well as with an occurrence of bangles in jet or shalesuch as the cemetery of Jakobsstrasse at Cologne (DE) (Friedhoff 1991) and the cemeteries at Trier (DE) (Goethert-Polaschek 1977). Ultimately we want to remark that Chapter 7 discusses the possibility of wearing (black) glass bangles in connection with the Christian population. 5.2.2.2.2 Finger rings We catalogued 18 finger rings that were deposited as burial gifts in a tomb. All were late Roman inhumation tombs. The type B2b finger ring from tomb 401/02 of the cemetery at Bregenz (AT) came from a double child burial [cat.no.23]. 5.2.2.2.3 Beads From the 266 recorded beads and spacer-beads found in burial contexts at civil settlements, a good number bears no detailed information on the finding circumstances as for instance all material (81 entries) from the late Roman cemetery at Horbat Qastra nearby Haifa (IL) which remains until now unstudied and unpublished. We only incorporated in Table 94 the burials providing some specific information such as burial type; gender; age; use; chronology. The ratio of rod-formed beads and double-perforated spacer-beads is with 213 ex. vs. 53 ex. a ratio of approximately 4/1. The black glass beads were mainly incorporated within necklaces built up from a variety of bead types and are combined with huge numbers of beads, with an average of 40-50 beads, but easily surpassing 100 beads to provide necklaces with a length of 400-600 mm. Several burials contained a set of four to nine identical or similar beadsi.e. double perforated spacer-beads with one flat sideto form a bracelet. The nine identical theatre mask beads with a severe face came from a double child‘s tomb of the late Roman cemetery at Pécs–Sopianae (HU) and formed a bracelet (Fülep 1984). Table 94 also demonstrates that only on rare occasions black glass beads were yielded from male burials. The black glass beads seems likewise their counterparts in other glass hues to be a plain jewel for girls and women. Most burials yielded one single black glass bead but sets of several pieces of the same type or a combination of various types do occur. The black glass beads were much more frequent in burials of the late Roman/early medieval cemeteries, such as those of Bregenz (AT) (Konrad 1997), Frénouville (FR) (Pilet 1980), Maule (FR) (Arveiller-Dulong 2006), and th Horbat Qastra (IL) (unpublished). We have to take into account that a number of these tombs are dated in the 5 th and 6 century ADi.e. the Merovingian period/early Byzantine period (Table 94). Only one single tomb out of 500 excavated at the cemetery of Epiais-Rhus (FR) yielded black glass beads: a necklace of 42 glass beads included 17 long quadrangular ones in black glass [cat.no.1057]. All burial gifts in the inhumation tomb S441 were placed at the feet of the deceased, and it is most probable that the objects were put in a wooden box. No information is available about the context of the various ‗crumb‘ bead types. This lack of information is similar for most other bead types –annular shaped (13 out of 54); barrel-shaped (2 out of 20); ‗melon‘ beads (7 out of 20)– but the accessible find circumstances show that black glass beads were popular trincket to women and girls. 198 Table 94: Black glass beads used as burial gifts cat.no. provenance type of bead function burial type gender age date reference 1925 Krefeld-Gellep (DE) globular ‗eye-bead‘ necklace - IVB AD Pirling 1979; Swift 2000, 293 Kaiseraugst (CH) annular ‗eye-bead‘ bracelet inhumation inhumation - 3659 female no dated context Martin 1976, 71, Taf.53G; Riha 1990, Taf.68, no.2839 3660 Kaiseraugst (CH) annular ‗eye-bead‘ necklace inhumation female child child no dated context Martin 1976, 60, Taf.43B; Riha 1990, no.2808 3661 Kaiseraugst (CH) annular ‗eye-bead‘ - inhumation female child no dated context Martin 1976, 57, Taf.40E; Riha 1990, no.2800 3662 Kaiseraugst (CH) annular ‗eye-bead‘ - inhumation - - no dated context Martin 1976, 55, Taf.38O; Riha 1990, no.2806 4087 Lankhills (UK) barrel-shaped ‗eye-bead‘ necklace inhumation female - IVc AD (350-370) Clarke 1979, 300, figs.62:G323; 86:436f 3658 Kaiseraugst (CH) barrel-shaped bracelet inhumation - child no dated context Martin 1976, 96, Taf.72A; Riha 1990, 188, no.2888 2745 Lovere (IT) barrel-shaped - plain 1895 Kempten (DE) globular ‗melon‘ bead - inhumation inhumation - child mid I AD Mackensen 1978, 268, Grab 220, pl.90:8 1896 Kempten (DE) globular ‗melon‘ bead - inhumation child mid I AD Mackensen 1978, 268, Grab 238, no.4, pl.100:6 1897 Kempten (DE) globular ‗melon‘ bead - inhumation female adult mid I AD Mackensen 1978, 268, Grab 55, pl.15:7 1898 Kempten (DE) globular ‗melon‘ bead - inhumation female adult mid I AD Mackensen 1978, 268, Grab 55, pl.15:8 1919 Krefeld-Gellep (DE) annular ‗melon‘ bead - inhumation IVB AD Pirling 1966, 16, no.9:4h, pl.8:18h. 4113 London (UK) annular ‗melon‘ bead necklace inhumation male adult IId-III AD Barber, Bowsher 2000, 155, no. 5a, fig.96, 1057 Epiais-Rhus (FR) square tubular plain necklace inhumation - - IId-IIIA AD Vanpeene 1993, 58; 90. 1914 Krefeld-Gellep (DE) square tubular plain - inhumation - - IVB-VII AD Pirling 1966, 57, no.434:1, pl.35:11a-e 1920 Krefeld-Gellep (DE) square tubular plain - VB-VIa AD Pirling 1966, 97, no.781:3, pl.67:36 Frénouville (FR) annular plain inhumation inhumation - 1111 necklace - - 1130 Frénouville (FR) annular plain necklace inhumation - - 2741 Cesena - San Egidio (IT) annular plain bracelet inhumation 2746-47 Lovere (IT) annular plain - inhumation - - 17 Bregenz (AT) annular with zigzag necklace - - 18 Bregenz (AT) annular with zigzag inhumation inhumation 1112 Frénouville (FR) annular with zigzag necklace inhumation - - 3192 Mérida (ES) annular with zigzag 1137 Homblières (FR) cylindrical plain necklace inhumation female adult - Pilloy1886,194-195; Roosens1962,18;Besson1994,17-28 2287 Somogyszil (HU) cylindrical plain - inhumation female child - Burger 1979, 25, Taf.4:2 Augst (CH) cylindrical ‗eye-bead‘ bracelet inhumation female? child 2219 Dunapentele (HU) globular plain - inhumation - - 2736 4022 Cesena - San Egidio (IT) globular plain - inhumation - - Colchester (UK) globular plain - inhumation - - IVb-IVd AD (325-400 AD) Crummy, 1983, 35, fig.37, no.1505 2878 Esch (NL) necklace cremation female? adult? IId AD Van den Hurk 1980, 374-378, fig.6 158 Pöttsching-Zillingthal (AT) 8-shaped spacer bead ‗Trilobitenperle‘ spacer bead - inhumation female adult - Caspart 1935, 13, Taf.IV:6; Haevernick 1983, 272 1888 Göggingen (DE) ‗Trilobitenperle‘ spacer bead - inhumation female adult - Keller 1971, 87, 232 (10) 1889 Göggingen (DE) ‗Trilobitenperle‘ spacer bead - inhumation female adult - Keller 1971, 87, 232 (10), fig.27:5, pl.9:15-18 1890 Göggingen (DE) ‗Trilobitenperle‘ spacer bead - inhumation female? child - Keller 1971, 87, 232 (10) 2202 Dunapentele (HU) ‗Trilobitenperle‘ spacer bead - inhumation - - - Vágó, Bóna 1976, 113-114, no.1326, Taf.27, 1326:3 Fortunati Zuccala 2002, 121, fig.31:3 Pilet 1980, BAR 83, II, 53-54; III, pl.25 Pilet 1980, Pilet 1980, BAR 83, II, 62; III, pl.29 IIIB AD Fadini, Montevecchi 2001, 52, fig.6 IVB – V AD Fortunati Zuccala 2002, 121, fig.31:3 Konrad 1997, 76; 224 Konrad 1997, 76; 230, no1:d Pilet 1980, BAR 83, II, 76; III, pl.34 inhumation Almagro Basch, Marcos Pous 1958, 87 Martin 1976, 96, Taf.72A; Riha 1990, no.2888 IIIB AD Fadini, Montevecchi 2001, 52, fig.6 Vágó, Bóna 1976, 93, no.1179, Taf.21, 1179:1 (top) 199 Keszthely (HU) ‗Trilobitenperle‘ spacer bead - inhumation - child - Haevernick 1983, 271, fig.1:7 Bátaszék-Kövesd (HU) ‗Trilobitenperle‘ spacer bead bracelet inhumation female child IVA AD Gesztelyi 1998, 131, fig.2. 2740 Cesena - San Egidio (IT) ‗Trilobitenperle‘ spacer bead bracelet inhumation female adult - Fadini, Montevecchi 2001, 52, fig.6 2806 Salorno (IT), tomb 98 ‗Trilobitenperle‘ spacer bead - inhumation - - IVA AD Roberti 1952, 8, no.7; Noll 1963, 228; Haevernick 1983, 273 2807 Salorno (IT), tomb 95 ‗Trilobitenperle‘ spacer bead - inhumation - - IVA AD Noll 1963; Haevernick 1983, 273 3891 La Skhira (TN) ‗Trilobitenperle‘ spacer bead - inhumation - - IV AD Fendri 1961; Haevernick 1983, 275 2755 Matarello (IT) ‗theater mask‘ spacer bead - inhumation female adult IVB AD Cambi 1885, 47, pl.IV:10; Haevernick 1983, 273 2265 Pécs – Sopianae (HU) ‗theater mask‘ spacer bead bracelet inhumation - child III-IV AD Fülep 1984, 90; Gesztelyi 1991, 108, fig.1 2739 Cesena - San Egidio (IT) ‗theater mask‘ spacer bead bracelet inhumation female child IIIB AD Fadini, Montevecchi 2001, 51, fig.3; Mandruzzato 2008, 162 1054 Dieulouard-Scarponne (FR) double portrait spacer bead - inhumation - adult IV AD Jacobs, von Schwerzenbach 1910; Swift 2000, 293 33 Leithprodersdorf (AT) single portrait spacer bead - inhumation female adult - Mitscha-Märheim 1957, 42; Haevernick 1983, 272 2252 Keszthely (HU) single female portrait spacer bead - inhumation - child - Sági 1960,1, fig.4; Haevernick 1983, 271; Gesztelyi 1991,108 2253 200 5.2.2.2.4 Pendants Twenty pendants from burial contexts have been recorded so far. These are mainly jug-shaped (8 ex.) or barrelshaped (5 ex.), but poppy-head-shaped pendants as well as discoid stamped pendants also occur. The burials rd th yielding a black glass pendant seem to be inhumations dating from the late 3 –early 5 century AD. The material does not offer good information on the gender and age of the deceased. It is however clear from Table 80 that the pendants were unmistakable personal ornaments the bearer kept wearing when buried (burial context = 79%) or got lost somewhere at home (domestic context = c.17%). The archaeological record cannot verify whether this commodity type had some specific meaning or function to the bearer or the community of the holder, but some assumptions have been proposed pointing to be emblemata of the Christian community (Stern 1977, 112-115) [see Chapter 7]. 5.2.2.2.5 Gems The finger rings with glass gemstones from burials can be assessed from the perspective of gender studies. Unfortunately, this sort of approach could not be worked out. It should be verified whether the use of particular metals is related to gender or social status, as, for instance, is shown by the two finger rings with glass gemstones in pseudo nicolo from the Homblières cemetery nearby Abbeville (FR) and now in the M.A.N. in SaintGermain-en-Laye (Guiraud 1998, 134-136): a silver ring [cat.no.A4575] from a male inhumation tomb (no.59), and a golden ring [cat.no.A4576] from a female inhumation tomb (no.83). The same cemetery yielded also a bronze ring [cat.no.A4574] with nicolo paste gemstone, but it is unclear from what burial it is retrieved. A careful study of the number of burial gifts given along with the deceased, the quality of manufacturing and the expensiveness of the applied materials and techniques can supply indications of the social status of the deceased by means of calculating the social stress the relatives experienced from the the general public (van Lith, Randsborg 1985). 5.2.2.2.6 Hairpins Only two examples are known to have come from a burial context, while pins in other glass hues do occur more regularly as burial gifts, e.g. Colchester (UK) (Crummy 1983, 28, fig.25, nos.462-464), Dunapentele (HU) (Vágó, Bóna 1976). A short, thick polyhedric head of a hairpin was retrieved from the cemetery des Capucins at Beavais (FR) [cat.no.1027] during excavations in 1860-61 (Schuler 1995 RAP 3, 62, fig.19:3). The second piece is a globular head of a hairpin that was found in inhumation tomb 664 of the late Roman/early Medieval cemetery 'Pousse Motte', at Maule (FR) [cat.no.1248] (Arveiller-Dulong 2006, 159, no. 41). All glass, jet, and bone hairpins from burials display a correlation with women and hairdressing and can therefore be considered a good indicator for gender studies; within this research, they indicate the presence of women in army camps. 5.2.2.3 Architectural decoration Despite the statement by Frank Sear (1977) that, particularly during the Julio-Claudian period, the Romans used glass rods to border polychrome mosaic panels on walls and vaults of nymphaea and tombs in Italy, we could not attest much architectural decoration material in black glass from funerary contexts. The only funerary context yielding black glass architectural decoration material is the arcosolium of tomb 88 at Isola Sacra, Ostia (IT). However, clear evidence of such use is known from various sites within the Gulf of Naplese.g. Pompeii, Herculaneumbut also from Rome. In addition to tesserae, twisted rods in black glass were used to embellish burial architecture, such as in a niche of the columbarium of the libertus Pomponius Hylas, where they are used to frame the inscription and to frame the entire mosaic panel together with a frieze of shells (Sear 1977, 65-66, no.25). Frank Sear did not specify the hue of the deeply coloured twisted glass rods, but from the colour picture (Sear 1977, pl.A) we presume that they are black-appearing. The rods framing the entire figurative panel are monochrome, whereas the rods used to frame the inscription are bichrome (‗black‘-white). 5.2.2.4 Counters In total, we recorded 30 burial contexts containing 208 black glass counters (Table 95). A number of these counters formed sets in mainly black and white glass, but also single counters in black as well as opaque white glass appear to have been given as burial gifts [see Chapter 7]. Several finds demonstrate that not only the gaming pieces were given to the deceased but also entire gaming sets, like for instance the 13 counters in monochrome black glass of tomb B435 in the eastern cemetery of London that were found together with 11 opaque white counters and 4 bone dice in a wooden box (Barber, Bowsher 2000, 193, fig.97). The occurrence of st board gaming sets is attested all through the Roman imperial period as emonstrated by the mid 1 century AD th ‗Warrior‘s burial BF64‘ at Stanway near Colchester (UK) and the 4 century AD set from Lullingstone (UK). 201 Table 95: Black glass counters used as burial gifts cat.no. provenance type of counter decoration quantity Berlingen (BE) – tumulus Burdinne (BE) – tomb 1 Clavier-Vervoz (BE) Cortil-Noiremont (BE) – tumulus indefinite large – round medium – round 277 Franchimont (BE) small – round 317 large – round small/large – round medium – round 4094 Herstal (BE) – tumulus Remicourt (BE) tumulus of Hodeige Vorsen (BE) – tumulus Tienen, Grimde (BE) – tumulus 1 Amiens (FR) rue du Faubourg, tomb 4 Frénouville (FR) tomb 315 Poitiers (FR) Des Dunes tomb 59 Poitiers (FR) Des Dunes tomb 94 Strasbourg (FR) Porte Blanche Krefeld-Gellep (DE) tomb 533 Krefeld-Gellep (DE) tomb 1215 Krefeld-Gellep (DE) tomb 6352 Esch (NL) Kollenberg tumulus, grave II Nijmegen (NL) Hunerberg – tomb 49 Merida (ES) tomb 57 Augst (CH) Chichester (UK) Cirencester (UK) Stanway - Colchester (UK) 'The Warrior's burial' BF64 London (UK) - tomb B435 plain plain plain marbled decoration dotted decoration plain 4152 Lullingstone (UK) medium - round 4171 Ospringe (UK) Winchester (UK) Lankhills tomb 51 Winchester (UK) Lankhills tomb 336 Winchester (UK) Grange Road indefinite A4486 225-6 600 228 318-9 615 A4612 1001-2 1110 1341-42 1343-44 1677 1905 1903 1904 2877 2937-41; 2943-44; 2951; 2955-61 3185 3271 3987 3990;3992-94;3997-99 A4582 4079-85 4086 4304 burial type gender age date reference 18 2 12 cremation cremation - - II AD - Roosens, Lux 1973 unpublished unpublished 6 cremation - - IIB AD (163-200) 1 inhumation - - - 12 (of 27) cremation - - IIB AD plain 2 cremation - - - unpublished plain 12 cremation - - - unpublished indefinite plain 5 (of 8) cremation - - IIB AD De Loë 1895, 129 medium - round dotted decoration 17 (of 31) inhumation - - IV AD Dilly, Mahéo 1997, 30 medium - round plain 1 inhumation female adult IV AD Pilet 1980, II, 167; III, pl.74:315-3 small/medium - round plain 2 cremation - - II AD small - round plain 2 inhumation - child II-III AD indefinite dotted decoration 1 inhumation - - IV AD medium - round plain 1(of 3) cremation - - IVA AD Pirling 1966, 70, no.533:22 small - round dotted decoration 10 (of 26) inhumation - - IVA AD Pirling 1966, 141-142, no.1215:21 IVA AD large – round Lefrancq 1989, B.13 (9-8), no.30 unpublished Amand, Mariën 1976, no.33 Simon-Hiernard 2000, 388, nos.364365 Simon-Hiernard 2000, 387-388, nos.360-361 Arveiller-Dulong, Arveiller 1985, 57 small/medium - round plain 9 (of 31) cremation - adult (20-60) small - round plain 8 cremation - - IIb-IIc AD medium - round dotted decoration 15 (of 30) inhumation - - IV AD small - round plain 1 inhumation - child - unpublished large - round small - round small/medium (1) - round plain plain plain 1 1 6 - - - - small - round plain 11 (of 20) cremation male adult mid I AD (43-50) small - round plain dotted decoration plain dotted decoration dotted decoration 13 (of 24) inhumation - - Ib - Id AD (40-80) unpublished unpublished Charlesworth 1982 Crummy, P. et al. 2007, 186-188; 217, fig.89 Barber, Bowsher 2000, 193, fig.97 15 (of 30) inhumation male adult IV AD Meates 1987, 123–5, 139–42 no. 391 Whiting et al. 1931, pl.LVI plain medium - round medium - round indefinite 4 (of 24) cremation - - IId AD 15 (of 24) inhumation male adult IV AD (310-370/90) 1 inhumation female juvenile IVc AD (350-370) 4 (of 18) inhumation female? - Id AD (85-95) Pirling, Siepen 2006, 599-600, no.6352:12, pl.105-106 Van den Hurk 1973, 217, no.II,34, fig.44 Van Enckevort 2008 Clarke 1979, 251-252 Clarke 1979, 251 Biddle 1976, 244-245 202 The obtained information is very limited because anthropological research is seldomly done or doable due to the limited or bad preservation of the bones. Hence, the evaluation of (black) glass counters in burial context does not contribute much to a better understanding of the gender or age. Table 95 demonstrates that the deceased who received (black) glass counters as burial gift could be male as well as female, adult as well as juvenile or even a child. It seems that board gaming sets were burial gifts in rich burials of the upper class and most likely are to be regarded a status symbol. th The occurrence of polychrome counters in only inhumation tombs of the 4 century AD makes it idiosyncratic to the late Roman period, whereas the monochrome counters are attested in inhumation tombs as well as in st cremation tombs all through the Roman imperial period from the mid 1 century AD but mainly concentrate in the nd rd 2 –3 century AD. 5.2.3 Ritual context In connection with the discussion on the use and function of commodities in black-appearing glass [see Chapter 7], we assumed it was important to verify the occurrence of good cases from ritual contexts to help provide material evidence for possible ritual uses, but with 63 entries from the 4475 artefacts only a minor number of black glass artefacts have come from a ritual context (Figure 95; Table 78). Jewellery in black glass perhaps constitutes more than three quarters of all ritual deposits (79,4%), but the very rare occasions on which black glass jewellery has been retrieved from ritual contexts (50 entries) contrasts with the occurrence in burial contexts from civil and military settlements (Figures 96-97; 108). Although jewellery can hardly be associated to ritual deposits, some jewellery types such as for instance bangles have been reported several times. The black glass vessels and counters are even less frequently occurring in ritual contexts, respectively with 7 vessels (11,1%), and 6 counters (9,5%), whereas architectural decoration has not been recorded from ritual deposits at all (Figure 107). ritual production 0% architectural decoration 0% counters 10% vessels 11% jewellery 79% Figure 107: Pie-chart showing the functional types of black glass artefacts from Roman ritual contexts 5.2.3.1 Vessels We can say that black-appearing glass vessels were not an issue in ritual deposits. Hitherto, no more than 7 pieces of black glass vessels have been recorded to have come from civil ritual contexts. Everything seems to point to an undisputed custom, but infrequent practice during Period II within the north-western provincesmore precisely in Germania Inferior and Germania Superior. A fragment of a carinated beaker, Form IIB.1, comes from the diverticulum at Baudecet (BE) [cat.no.195]. Four vessels excavated at the Maasplein in Nijmegen (NL) one is a cup with bulbous body, Form IIB.5 [cat.no.2976]; two are fragments of a carinated beaker [cat.no.2977-78]; one entry remains indefinite [cat.no.2980] are said to have been retrieved from the sanctuary but seeing that this material comes from a cesspit, it is very unlikely that the black glass vessels have received a ritual character. Another indefinite vessel shape comes from Temple 1 in Regio 3C of Augusta Raurica in Augst (CH) [cat.no.3497]. The last piece, a cup with bulbous body, comes from the temple ‗Le Cigonier‘ at Avenches (CH) [cat.no.3613]. 203 5.2.3.2 Jewellery Table 96 shows that the 50 black glass pieces of jewellery from civil ritual contexts are mainly arm rings with 2/3 of the entries considered, while the occurrence of finger rings, beads, gems, and one pendant remains rather marginal within ritual deposits. To our knowledge, no hairpin has been retrieved so far. Seeing the rare occurrence of black glass material in ritual contexts and the general dating of the assemblages in centuries it was rd th impossible to overview the repartition per period. We see however that the main body falls in the 3 4 centuries AD, in particular in Period III and the start of Period IV, whereas there is a total absence in Period I and nearly none in Period II. Most material seems to have been lost during the visit of the sanctuary by the pious visitor Table 96: Overview of the different types of black glass jewellery from civil/ritual contexts type bracelets finger rings beads pendants gems hairpins TOTAL A1-6 (16 ex.); B2 (1 ex.);B5 (1 ex.); C2 (1 ex.); D1 (15 ex.); D7; U2; indefinite A1:3; A2:1; C1:1; indefinite small annular plain; indefinite indefinite flat elliptical nicolo imitations (Henig F2/4); loose or inserted in finger ring and - Total quantity % 37 (39) 68,5 11,1 6 7,4 4 1,8 1 11,1 6 54 (56) 99,9 The 37 arm rings in black glass that have been ascribed to ritual deposits (Tables 96-97) remain quite 177 insignificant compared to the arm rings from other context circumstances (Figure 103). A re-evaluation of the database necessitates questioning whether these artefacts were part of a ritual deposit or just coincidentally lost/broken within a religious public building. Furthermore, the late Roman bangle of Type D1 from the Temple of Hibis at Ain-et-Turba (EG) is probably from a domestic context, seeing that from the late Roman period the Coptic community swallowed the ancient architecture by building mud brick houses in and around the sanctuaries dedicated to a deity of the pharaonic pantheon. It is unclear how to interpret the two bangles from the early Christian basilica complex of Bir Ftouha (TN), but very likely it is not more than the remains of an accident by a female churchgoer. The same should be considered for the 3 arm ring fragments from sector K of the ‗grande tempio‘ in the city centre of Luni (IT). On the other hand, the fragments from Matagne-la-Grande, Tourinnes-SaintLambert, Vervoz-Clavier, Alise-Sainte-Reine, Arras, Estrées-Saint-Denis and Bastendorf came from Gallo-Roman sanctuaries, either from pit deposits, favissae, or within the sanctuary architecture. An intact open bracelet with wide twisted decoration, Type A2, was found in a well at the Cybele sanctuary of Arras (FR) (Trésors archéologiques 1997, no.128). Also an open bracelet with alternating twisted decoration and plain zones, Type A5, came from a well that was re-used as a ritual pit on the vicus-site at Waudrez/ Vodgoriacum (BE) (Hanut, Capers 2003, 5-104). A similar A5-bangle was found with two other bracelets at the late Roman sanctuary of ClavierVervoz (Witvrouw, Witvrouw 1976, 178, fig.40:16-17; 197, fig.40:18). The open A5-bangle was excavated in the southern portico of Building D together with a closed bracelet with fine twists decorated with a twisted double trail in opaque white glass, Type A6. The third bracelet, of undefined type, was found in Temple A. Finally, a large piece of an open bracelet of Type C2 with 6 lengthwise ribs crossed with transverse pinches and a spatula178 shaped ending is presumed to have been found within the favissa at Tourinnes-Saint-Lambert (BE) . Different is the open, plain bracelet with O-shaped section, of Type A1, that came from a shallow pit in the rural settlement of Veldwezelt (BE). The incomplete bracelet was found with an incomplete dish in smoked ware covered with a wrapped-up iron chain with large links and a bronze barrel-shaped lamp or incense burner (unpublished). A related situation has been uncovered in Nijmegen (unpublished) where a nearly complete bangle of Type A4 was found in a pit amid the cemetery, which also contained the skeleton of a dog. The deity to whom the offering might have been dedicated is unknown and is most likely unrelated, because the bangle from ‗La Croix Saint-Charles‘ at Alise-Sainte-Reine (FR) is dedicated to Apollo Moritasgus and the Arrasbangle is dedicated to Cybele. From excavations at the temple of Athena Itonia nearby Philia in Thessaly (GR) were retrieved two intact arm rings type A1 with a quadrangular section at the seam (Kilian-Dirlmeier 2002, 130131, nos.1958-1959, fig.117). The bangle from ‗Le Bois de Noël‘ at Matagne-la-Grande comes from a Constantine sanctuary; nevertheless, it is unclear to whom it may be dedicated (Cattelain, Paridaens 2009). 177 The material from the subaquatic deposits in the cave Trou de Han at Han-sur-Lesse (BE) are also to be considered ritual deposits but are discussed under 5.4.1). 178 Personal communication by Frédéric Hannut. 204 Despite the fact that the temple of Apollo at Alise-Sainte-Reine (FR) had a spring with medicinal effect there is no direct relation to connect the black glass bangle with medical superstition. Concerning the beads, we can be brief as only the small annular bead from the Constantinian sanctuary of ‗Le Bois de Noël‘ at Matagne-la-Grande (BE) is known [cat.no.436] (Cattelain, Paridaens 2009, 111, no.2, fig.50:2). Another bead from the same site might perhaps be a mistake by the author. Table 97: Black glass bangles from ritual context cat.no. site type 368 Liberchies (BE) D1 completeness context sanctuary sanctuary favissa ritual deposit (indigenous dwelling) sanctuary – southern portico sanctuary – southern portico 434 566 591 604 606 Matagne-la-Grande (BE) Tourinnes-Saint-Lambert (BE) Veldwezelt (BE) Clavier-Vervoz (BE) Clavier-Vervoz (BE) A4 C2 U2 A5 A6 fragmentary fragmentary fragmentary fragmentary fragmentary fragmentary 608 Clavier-Vervoz (BE) undefined fragmentary sanctuary – buiding adjacent to the southern portico 625 795 1000 1024 1060-71 1669 1988 1993 1994 1995 2748-49 2751 2855 2856 2857 2968 3887 3889 A4580-81 Waudrez (BE) Ain-et-Turba (EG) Alise-Sainte-Reine (FR) Arras (FR) Estrées-Saint-Denis (FR) Septeuil (FR) Ober-Olm (DE) Pommern a/d Mosel (DE) Pommern a/d Mosel (DE) Pommern a/d Mosel (DE) Luni (IT) Luni (IT) Bastendorf (LU) Bastendorf (LU) Bastendorf (LU) Nijmegen (NL) Carthage (TN) Carthage (TN) Philia (GR) A5 D1 A2 A2 D1 undefined A1 A1 A2 A6 D1 D7 A2 A6 A3 A4 B2 B5 A1 fragmentary fragmentary fragmentary complete fragmentary fragmentary fragmentary fragmentary fragmentary fragmentary fragmentary fragmentary fragmentary fragmentary fragmentary nearly complete fragmentary fragmentary complete ritual deposit in pit temple of Hibis temple of Apollo Moritasgus temple of Cybele sanctuary - fanum Mithraeum unknown sanctuary sanctuary sanctuary temple (Grande tempio) temple (Grande tempio) sanctuary sanctuary sanctuary ritual deposit (cemetery) Christian basilica Christian basilica temple of Athena Itonia 5.2.3.3 Architectural decoration So far, we have not recorded architectural decoration material from (civil) ritual contexts. 5.2.3.4 Counters From the total of 6 black glass counters that are known to have come from civil ritual contexts (Table 98) and the very limited information, we can conclude that counters are rather coincidentally occurring in religious precincts and consequently were never used in rituals during the Roman imperial period. In other words, counters have never had a ritual character in Roman society. Table 98: Black glass counters from ritual contexts cat.no. site type 999 Alba-la-Romaine (FR) small-round 1219-20 Loubers (FR) indefinite 4241 Uley (UK) small-oval 4242 Uley (UK) small-oval 4243 Uley (UK) small-oval decoration undecorated undecorated undecorated undecorated dotted decoration context Northern court undefined Structure IX Structure XIV Structure XIV period Period II no dated context Period (III-)IV Period IV Period IV A small plano-convex counter in monochrome black glass was found in the northern courtyard of the ―Bagnols‖ sanctuary at Alba-la-Romaine (FR). The piece came from a feature related to modification works carried out in the nd second half of the 2 century AD (Period II) at the northern court: an open area behind the large peristylium flanking the eastwest orientated temple (Roussel-Ode 2008, 59, fig.3:10). The piece cannot be connected to a ritual deposit, and consequently it most likely is to be considered as a coincidental loss rather than part of a ritual action. The marked absence of glass counters in ritual contexts makes it obvious that the Romans excluded 205 gaming pieces from any ritual act and gave such material a rather trivial meaning. Furthermore, counters were omnipresent commodities in Roman times and used anywhere for gaming, even up to the threshold of a 179 temple. Two black glass counters were reported from the fanum ‗Camp-Ferrus‘ at Loubers (FR), but without detailed description of the pieces or any further information on the find circumstances (Bessou 1978, Gallia 36, 212). From the West Hill Temple Site at Uley (UK), three black glass counters were retrieved, one of which was in opaque blue glass with 4 dots (Woodward, Leach 1993, 177, fig.135:15-17). The contexts in which the pieces th th th were found are dated 4 -early 5 century AD in general and refined to the second half of the 4 century AD. 5.3 Military contexts The 498 (531) black glass artefacts from military contexts display a preponderance of jewellery and gaming 180 pieces, a small presence of black glass vessels , and a clear absence of architectural decoration material and 181 any indications of manufacturing black glass (Table 99; Figure 108). The absence of architectural decoration material in black glass is in line with the total absence of monumental ornamentation at army camps in general and corroborates the purely functional character of military settlements. The absence of black glass production material at military complexes almost certainly excludes production activities of glass counters and glass jewellery by soldiers. The production of the latter must be considered in a civil workshop at the canabae legionis that grew around an army camp, as is demonstrated by the ‗Medizinklinik‘ workshop close by the camp of Bonn (DE) [see 5.2, ‗Civil contexts‘]. It is, however, not excluded that black glass objects have been manufactured within army camps, seeing the presence of a glass workshop in the castra Bonnensia in Bonn (DE) (Follmann-Schulz 1991). Table 99: Black glass artefacts from military contexts (additional entries between brackets) military context quantity % vessels jewellery architectural decoration counters production total 29 (+2) 240 (+1) 0 227 (+31) 1 497 (531) 5,8 (5,8) 48,3 (45,4) 0,0 45,7 (48,6) 0,2 100,0 military context production 0% 182 military context production 0% vessels 6% vessels 6% counters 46% counters 49% jewellery 45% jewellery 48% architectural decoration 0% architectural decoration 0% Figure 108: Black glass from military contexts (right: with additional entries) The black glass jewellery is dual as it comprises jewellery that can be related solely to men, the gemstones set in finger rings, whereas other jewellery is specifically used by women and girls, e.g. arm rings, hairpins, and 179 It probably was no more than a lost piece from someone gaming in the courtyard, seeing that plenty of carved board games are known from peristylia, agorae, and temples such as Philippi (GR) or Sagalassos (TR). 180 See 5.3.5., ‗Vessels‘, for a more detailed view. 181 The piece of raw glass from Rumst-Molenveld (BE) is taken out of consideration here as it might be intrusive, seeing that its chemical composition does not correspond to the main Roman composition [see Chapter 8, sample PC1f]. 182 The additional recorded material does not change the initial ratios much (see also Figure 98). 206 bracelets or necklaces with spacer beads. The presence of typically female jewellery, such as black glass arm rings and beaded necklaces, points to the presence of women and girls in and/or around the military settlements [see subsequent sections in this chapter and Chapter 7]. The findings presented here are perhaps not adding much new to the current archaeological debate concerning the presence of women and children in and around Roman army camps, but act as independently acquired additional evidence corroborating previous work (Van Driel-Murray 1994; 1997; Allison 2006; 2008). Within the scope of our research, it was impossible to go beyond interpreting the black glass artefact distribution and trying to understand the social structure within Roman army camps in general and the role of women and the family unit of soldiers in particular, because such understandings cannot be deduced from archaeological evidence alone and need interaction with epigraphic sources. The high proportion of black glass counters within army camps corroborates the omnipresence of gaming pieces in various other glass hues as well as in other materials. Board games, on the other hand, definitely played an important role in helping soldiers fight boredom between drill exercises and other camp activities. In view of the abovementioned gender presence in the military community, it is possible that these board games were also used by children. The cataloguing of the black glass artefacts from military contexts has been done in a similar way as for those from civil settlements, i.e. per commodity category within defined contexts of the military unity (domestic, public, ritual, burial, and production) in order to evaluate more in detail who used such commodities and where they were used. The accessibility to detailed context information remained, however, rather limited for most artefacts, making a contextual analysis of the available information not significant. No further context details were at hand 183 for 457 out of 497 items, or 92% (Table 100). Only 15 records came from burial contexts, 3 pieces were found in a ritual context, 10 pieces came from a public context, and 8 pieces came from a domestic context. This meagre result made us decide to consider the material from another point of viewi.e. by considering the material per camp type, as this approach demonstrates the potential to incorporate the artefacts of ‗unknown‘ provenance. The provenance of black glass material from army camps, including those sites with undefined contextual data, illustrates that a substantial part came from late Roman hill forts in Belgium and northern France, th th which were particularly in use during the later 4 and early 5 century AD. Table 100: Overview of the contextual data of black glass undefined context cat.no. site camp type commodity 35-121 Magdalensberg (AT) castellum 235-254 263;266-68 343;345;351 371 455-58 459 465 521-25 637-38 820;822;826-27;829-31; 838 828 818 853-55 1134 1222 1749-51;1753-58 1752 1773-93 1931-32 1933-34 1941;1943;1947 Dourbes (BE) Eprave (BE) Liberchies (BE) hill fort hill fort castellum Nismes (BE) 2 hill forts counter bead vessel arm ring arm ring counter bead arm ring vessel Oudenburg (BE) Taviers (BE) Krivina (BUL) Dush Kysis (EG) castellum burgus castellum castellum 1949 1981 Grandes-Armoises (FR) Lugasson (FR) Vireux (FR) hill fort hill fort hill fort Bad Homburg-Saalburg (DE) Kreimbach-Kaulbach (DE) castellum hill fort Mainz (DE) castellum Miltenberg-Altstadt (DE) Neuwied-Niederbieber (DE) castellum castellum quantity arm ring spacer-bead arm ring 71 1 14 20 4 3 1 4 1 18 5 2 8 bead pendant vessel arm ring finger ring arm ring finger ring gem finger ring arm ring arm ring counter vessel arm ring gem 1 1 3 1 1 9 1 21 2 2 1 1 1 1 1 period I IV IV III-IV IV III II-IV IV III-IV IV IV IV IV II-III III-IV I-III III-IV II-III 183 A consequent and more detailed study on this issue will enable the defining of more context details of a good number of records. 207 2110 2477-78 2908-10;2913-17;2919; 2921-27;2929-36;2942; 2945-50;2953-54 2989 2993-2996 Wiesbaden (DE) Mezad Tamar Nijmegen (NL) castellum castrum castrum gem arm ring counter 1 2 33 I IV I Utrecht (NL) Vechten (NL) castellum castellum Fiães - Vila da Feira (PT) Hru ica (SL) Kaiseraugst (CH) castrum castellum castrum arm ring 1 3 1 1 30 15 I-III I-III 3101 3121-50 3647;3666-68;3672; 3675-78;3680;3685-86 3646 3638;3640 3652-53;3673-74;3682-83 3784-86;3788-91;3793-97 3792 3919 3924 3977-83 4034-4039 4040-43 4069 4070 4180 4185-88 4182;4184 4220-21 4222;4224-26;4229 4223;4228 4227;4230 counter counter gem arm ring finger ring counter bead arm ring bead gem gem gem counter gem gem gem counter gem bead counter gem bead pendant 1 2 7 12 1 1 1 7 6 4 1 1 1 4 2 2 4 2 1 Pfyn (CH) hill fort Birrens (UK) Burnswark (UK) Chesterholm (UK) Corbridge (UK) castrum castrum castellum castrum Hod Hill (UK) Housesteads (UK) Richborough (UK) hill fort castrum castrum South Shields (UK) castrum IV IV IV IV II-III II-III II-III I-III I II-III I-IV I-IV funerary context cat.no. site camp type commodity quantity 117 278-281 479;482;483;488;490 1867 1868 3787 4023 4289 Magdalensberg (AT) Furfooz (BE) Oudenburg (BE) Eining (DE) castellum hill fort castellum castellum vessel arm ring arm ring arm ring spacer-bead bead counter counter 1 4 5 1 1 1 1 1 cat.no. site commodity quantity 1892 1998 4033 Isny-Bettmauer (DE) Regensburg (CH) Corbridge (UK) spacer-bead gem counter 1 1 1 cat.no. site camp type commodity quantity 819;821;824 1873-74 1872 1870-71 3917 3918 Dush Kysis (EG) Epfach (DE) castellum hill fort Birdoswald (UK) castrum arm ring arm ring spacer-bead counter finger ring gem 3 2 1 2 1 1 quantity Pfyn (CH) Corbridge (UK) Usk (UK) castellum castrum castellum ritual context camp type castellum castellum castrum public context period I IV II-IV IV IV I-II I-II period III II I-II period III-IV IV I-II domestic context cat.no. site camp type commodity 217 480 814;825;834;839 836 4181 4183 Braives (BE) Oudenburg (BE) Dush Kysis (EG) burgus castellum castellum Richborough (UK) castrum arm ring gem arm ring bead counter bead 208 1 1 3 2 1 1 period III II-III III-IV I-II 5.3.1 Funerary context The material does not show regularity in commodity type, in period, or type of army camp, besides the very poor representation of black glass material from burials at military settlements: 15 artefacts from 12 contexts from 6 sites. About the vessel fragment of form IA.2 from Magdalensberg (AT), nothing more is known than that it is from tomb 184 3 which is dated 10-40 AD. Since it is a fragment, it could have been intrusive and not a burial gift, unless it had been a primary burial gift and was what remained or got picked up from the pyre. The arm rings have been retrieved from burials of girls or young women, likewise in civil settlements, and are all late types, D1 and D2. Those from tomb 2 at the ‗Unterfeld‘ cemetery of Eining (DE) and tomb 194 of the late Roman cemetery of Oudenburg (BE) were singular pieces, whereas those from tombs 2 and 22 at Furfooz (BE) and tomb 100 at Oudenburg were worn in pairs. The Type D1-bangles from Furfooz were found in inhumation burials, within a bath complex that probably got out of use or was converted into a church. Both deceased were girls wearing glass arm rings at the left wrist; likewise, the girl of Eining and the young woman of tomb 194 at Oudenburg. The use and possible meaning is discussed in Chapter 7. The bones from tomb 100 at Oudenburg were badly preserved, but the two black glass bangles were found together with a bronze arm ring at the area of the hip joint. This position suggests that the hands were placed together on the abdomen. It is, however, impossible to identify on which wrist the glass bangles were worn, even if the position at the right side of the coffin corresponds with the conventional notion. The arm rings all come from inhumation burials which show no uniformity in orientation: either eastwest (Oudenburg T100; ...) or southnorth (Oudenburg 194;...). The bead from Pfyn (CH) is considered to be a pendant of a necklace, as can be seen from burials at civil settlements. This bead as well as the spacer-bead from Eining is from an uncertain funerary context. Both counters from the burials at Corbridge (UK) and Usk (UK) are possible examples of pars pro toto and are also attested at civil settlements as well as with counters in other glass hues [see Chapter 7]. The arm rings and other black glass jewels found in burials connected to a military settlement can be related to female relatives of the military staff, without being able to affirm whether they are restricted to the officers‘ family or also to the wives/girls/sisters of soldiers. A more specific study of all small finds from army camps in relation to gender studies is essential but also necessitates a comparative assessment with imagery (portraits) and written sources, in particular the epigraphic sources. 5.3.2 Ritual context The three known ritual deposits that contained black glass artefacts are the treasures of Vemania fortress nearby Isny (DE) (Garbsch 1971, 137, fig.30), the fortress of Regensburg-Kumpfmühl (DE) (Boos et al. 2000, 23, no.5, Taf.5,1; 6,4), and the hoard of Corbridge (UK) (Allason-Jones, Bishop 1988). The Corbridge hoard is a wooden box filled with military equipment and utensils containing a set of 54 glass counters, of which 23 are in black glass and 31 in white glass. Seeing the use of 12 or 24 counters (6 or 12 in a contrasting colour per player), it is clear that the amount of glass counters is more than a set and should be considered a collective bag wherein all counters of a person or group of persons are put together. The late Roman fortress nearby Isny, dated 260-380 AD, lies on the Roman road Bregenz-Kempten at the Donau-limes in the provincia Raetia. The Vemania treasure, which was found at the northern barracks of the castellum, contained a necklace with 10 so-called ‗Trilobitenperlen‘ combined with cylindrical beads in jet and gold. The stratigraphy within which the treasure was found is dated at the culmination of phases 3 and 4 and th consequently dated around 305 AD (Garbsch, Kos 1988). The early 4 -century treasure was most probably the property of an officer, and consequently the black glass spacer-bead necklace must have belonged to the officer‘s wife, a very plausible assumption seeing that the officer‘s family was allowed within the camp limits (Campbell 2010, 50). On the basis of the ban on marriage for soldiers stated by ancient writers, the widespread model on the organization of army camps has always been that spouses and children of common soldiers were not allowed within the army camp (Phang 2001), and that they must have settled at the limits of the garrison to form the core of a canabae legionis. In particular, the reforms by the emperor Septimius Severus in AD 198 who legitimised the 184 Personal communication by Barbara Czurda-Ruth (2005). 209 marriage of soldiers (Barnes 2008) feed the idea that at least during the 1 2 centuries AD, army camps did not host the women and children of the soldiers. Although the presence of women and children within the military community has been repeatedly confirmed by archaeological evidence (Allason-Jones 1995; Allison 2006; 2008; Van Driel-Murray 1994; 1997), its approval only slowly overcame the biased assumptions by archaeologists and historians due to scepticism on the issue (James 2006). Conversely, epigraphic evidence from Roman military diplomas put forward that specifically auxiliary soldiers took along their wives and children when the unit travelled or remained in a castellum (Roxan 1991; Saller, Shaw 1984; Phang 2001). It is, however, not excluded that the black glass jewellery could have been worn by other women roaming around at army camps, such as freedwomen, concubines, prostitutes, and kitchen maids, focariae. st 5.3.3 nd Public contexts The material at hand does not provide any significant information to postulate a hypothesis. The 12 artefacts from 4 military sites yielded black glass objects from public contexts and consist of 4 counters, 5 arm rings, 1 spacerbead, 1 finger ring and 1 gemstone. The public building of Magdalensberg (AT) is to be dated in Period I; all other material from public contexts that came from Dush Kysis (EG), Epfach (DE) and Birdoswald (UK) is from Period IV. The 2 counters from Magdalensberg were found in cistern NG/20. The 3 arm rings from Dush Kysis came from different areas: 1 Type D1 from the south-western corner of the fortress [cat.no.819]; 1 Type B2 variant from the North Gate [cat.no.821]; and 1 Type D1 from the second courtyard of the Temple [cat.no.824]185. The 5 pieces from the hill fort Lorenzberg at Epfach were 2 counters, 1 spacer-bead and 2 arm rings and mainly came from the horreum (Magazingebäude, Raum A-B) and the so-called Lorenzkapelle (Raum A). Also the finger ring and pseudo-nicolo gemstone from Birdoswald came from a horreum (Building 197) (Wilmott et al. 1997, 283, fig.194, no.85; Henig 1997, 285, fig.195, no.89). The horreum was built during the Severan period under Severus, th Caracalla and Elegabalus (c. 198-215 AD) but underwent a change of use around the mid-4 century AD and th received a sub-floor. The finger ring was found in the floor layer of Phase E of Period 5 (second half of the 4 century AD), and the gemstone was recorded in the backfill on top of the sub-floor. This dump is also considered Phase E of Period 5 (Wilmott et al. 1997). We have to consider these fragmented commodities as the remains of lost objects (in the case of the counters) or of broken objects that have fallen on the ground (arm rings) and small bits and pieces that slipped through the fissures of the wooden floor of, for instance, the horrea or in the sand and vegetation of open courts. On further consideration of the presence of women in the military community, it is very likely that women worked in the horrea or at least were allowed to enter the storerooms. 5.3.4 Domestic contexts Only nine pieces from four military sites yielded black glass artefacts from domestic contexts. The material from the desert fortress of Dush Kysis in the Kharga oasis (EG) possibly gives an idea of the women‘s rooms within the fortress or even the family rooms, considering that also girls wore black glass bangles:   3 arm rings – from room I (Type D1); corridor 16 (indefinite); room 2A (Type D1) 2 beads – room V (small globular-shaped); room 14 (indefinite) Further, there is an arm ring of Type A1 from the small castellum/burgus, ‗Les Sarrasins‘ at Braives (BE) (Vilvorder 1993, 243); a nicolo imitation gemstone in a layered blue-on-black glass decorated with an eagle from a waste pit in the south-western corner of the castellum at Oudenburg (unpublished); a small round counter from waste pit 20 and a small annular bead from a ditch/waste pit from Richborough (UK) (Bushe-Fox 1928, 31, pl.XIV, fig.2; Bushe-Fox 1949, 149, no.236). 5.3.4.1 Vessels Due to the limited context information available, we can remain very brief about the context analysis of the vessels, seeing that the total of not more than 29 (31) pieces of the 651 (681) black glass vessels are known to 185 Personal commuinication by Marie-Dominique Nenna. 210 have come from military settlement contexts. The majority of this small quantity makes it of no use to discuss much more in detail this issue, especially because the recorded material is limited to a few sites from Period I (Table 101). It is interesting that the Period I vessels are all limited to cast, shallow, hemispherical and cylindrical cups and dishes of the forms IA.2 and IA.4 except for the tripod table, form IA.7, which came from the Claudian-Neronian auxiliary fort of Caersws (UK) (Cool, Price 1995, 26; Cosyns, Fontaine 2009, tab. 1, forme 19). The cups and dishes all were retrieved in the Alpine and upper Rhine regionsi.e. Magdalensberg (AT), Mainz (DE), and Windisch/Vindonissa (CH). These sites were Celtic settlements where early legionary camps were established under Augustan rule and obtained a privileged economical status. This occurrence of the black-appearing glass vessels is connected to the presence of other monochrome deeply coloured glass vessels (Berger 1960; CzurdaRuth 1979; 1998; 2001; 2003) but is unlikely to be dated that early. Because it is generally considered that polychrome deeply coloured glass was prevalent in AugustanTiberian times, more importance has to be given to the early material on British sites such as the Caersws‘ tripod, dating from the very early moments of the Roman invasion under the emperor Claudius. Table 101: Overview of ratios of black glass vessels per period from military settlement contexts Period I Period II Period III cat.no. site quantity % quantity % quantity % A4499 Caersws (UK) 1 5,3 0 0 107-121;A4489 Magdalensberg (AT) 16 84,2 0 0 1947 Mainz (DE) 1 5,3 0 0 3806 Windisch (CH) 1 5,3 0 0 492-94 Oudenburg (BE) 0 3 37,5 0 510-14 Rumst (BE) 0 5 62,5 0 459 Nismes (BE) 0 0 1 100 853-855 Dush Kysis (EG) 0 0 0 TOTAL 19 61,3 8 25,8 1 3,2 Period IV quantity % 0 0 0 0 0 0 0 3 100 3 9,7 The vessels of Period II have been retrieved from Belgian sites and are all of carinated beakers, form IIB.1, except for two indefinite fragments from the castellum at Oudenburg (BE). The fragment from the Period III-IV hill fort of Roche Sainte Anne at Nismes is an example of a late use of such carinated beakers because the nd production and distribution of this vessel type was more or less limited to the last quarter of the 2 to the first rd quarter of the 3 century AD [see Chapter 4]. The vessel, therefore, must have come from an early context of the rd initial phase of the fortified settlement, from the third quarter of the 3 century AD, when the north-western provinces were devastated by Frankish incursions. The material from Oudenburg came from the south-eastern corner of the castellum, but mainly from the thick black layer that covers the entire site. The Oudenburg material possibly connected with the last phase of the preceding vicus, i.e. IId AD, and likewise the pieces from the castellum Rumst-Molenveld (Sevenants, Cosyns et al. 2006). The Period IV vessels from military contexts have been so far recorded from Dush Kysis (EG) in the Kharga oasis, rd th where late Roman military installations were built during the 3 4 centuries AD and lasted until the beginning of th the 5 century (Reddé et al. 2004). The vessels are all three rod-formed squat jars, Form IVD.2, decorated with applied trails in opaque blue or yellow glass. The rod-formed vessels such as these jars are characteristic for the region [see Chapter 6] and are to be considered the late variant of the solid body unguentaria which is attested to have been a frequent commodity on military settlements (Brun 2003). 5.3.4.2 Jewellery Table 102: Overview of ratios of the different types of jewellery per c amp type castrum castellum burgus quantity % quantity % quantity % Bracelets Finger rings Beads Pendants Gems Hairpins TOTAL 18 1 13 2 16 0 50 12,2 16,7 44,8 66,7 30,2 0 20,8 69 1 12 1 37 1 121 46,6 16,7 41,4 33,3 69,8 100 50,4 5 0 0 0 0 0 5 3,4 0 0 0 0 0 2,1 hill fort quantity % TOTAL quantity % 56 4 4 0 0 0 64 148 6 29 3 53 1 240 37,8 66,7 13,8 0 0 0 26,7 100 100,1 100 100 100 100 100 211 In total, 239 (240) pieces of jewellery in black glass are known to have come from military settlement contexts, of which 149 ex. or 61,8% are arm rings, 6 ex. or 2,5% finger rings, 29 ex. or 12,0% beads, 3 ex. or 1,2% pendants, 53 ex. or 22,0% gems and 1 ex. or 0,4% hairpins (Table 102). 5.3.4.2.1 Arm rings Table 103: Bracelets from military sites cat.no. site camp type shape n period A2 A1 A3 B1 B2 D1 D1 U1 D1 B2 C1 D1 A1 A2 A4 B1 C3 D1 D2 A3 A4 C2 B2 C3 D1 D2 indefinite D1 A1 A2 D1 indefinite C3 D5 D1 A2 A4 D1 D2 A2 D1 D2 D5 undefined A1 A2 A4 A6 C1 D1 D2 D4 A2 D1 D2 D3 1 1 1 1 1 17 2 1 4 2 1 1 2 2 2 1 4 11 1 2 1 2 1 1 6 2 4 1 1 2 6 1 1 1 2 1 1 2 1 1 9 1 1 15 1 2 1 1 2 6 1 1 1 8 2 1 II-III III III-IV 161 217 235-54 Mautern (AT) Braives (BE) Dourbes (BE) castellum castellum hill fort 263;268 266 278-81 455-58 Eprave (BE) hill fort Furfooz (BE) Nismes (BE) hill fort hill fort 467-70;472-79;481-91 Oudenburg (BE) castellum 521-22 523 524-25 814;819-22;824;826-27; 829-31;834;838-39 Taviers (BE) burgus Dush Kysis (EG) castellum 1134 1749-51;1753 Grandes Armoises (FR) Vireux-Molhain (FR) indefinite hill fort 1868 1873-74 Eining (DE) Epfach (DE) castellum hill fort 1933-34 1943 1949 2477-78 3101 3121-29;3131-44;3146-49 Kreimbach-Kaulbach (DE) Mainz (DE) Miltenberg-Altstadt (DE) Mezad Tamar (IL) Fiães - Vila da Feira (PT) Hru ica (SL) hill fort castellum castellum castrum castrum castellum 3647;3650-51;3654;366668;3672;3675-78;3680;368586 Kaiseraugst (CH) castrum 3784-86;3788-91;3793-97 Pfyn (CH) hill fort III-IV IV III-IV II-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV III-IV The 149 recorded black glass bangles come from 21 military settlements concentrated in the north-western provinces. But the occurrence in Central Europe, the Iberian Peninsula and the south-eastern Mediterranean shows that the ubiquitous black glass bangle is a frequent commodity in army communities all over the Roman Empire. The black glass artefacts mainly crop up at late Roman fortresses, with 69 items from 6 castella 212 (AV=11,5) and 56 items from 8 hill forts (AV=7). There are only 5 items from 1 burgus (AV=5) and 18 items from 3 186 castra (AV=6) (Table 103). The averages only display a general and somewhat distorted representation because a number of sites yielded only few glass bangles due to the minor excavations so far while other sites that have been extensively excavated provided larger numbers of black glass bangles. Twenty pieces are too small to be determined or could not be defined due to the lack of reliable information; 24 pieces are of Type A (5 ex. A1; 10 ex. A2; 3 ex. A3; 5 ex. A4; 1 ex. A6); 6 pieces are of Type B (2 ex. B1; 4 ex. B2); 11 pieces are of Type C (3 ex. C1; 2 ex. C2; 6 ex. C3); and 87 pieces are of Type D (75 ex. D1; 8 ex. D2; 1ex. D3; 1ex. D4; 2 ex. D5). The Type A bangles all come from the north-western army camps of Gallia Belgica (Oudenburg [B]), Germania Inferior (Braives [B], Taviers (BE), Vireux (FR), Germania Superior (Mainz [D]), Kaiseraugst (CH), MiltenbergAltstadt (DE), Pfyn (CH), Noricum (Mautern (AT), and Hru ica (SL). There is not much precise contextual information at hand on these bangles, and they have been retrieved at every camp type (castrum, castellum, small castellum, and burgus). The Type B bangles all came from Belgian sites (Dourbes, Nismes, and Oudenburg) except for a Type B2 variant from the Kharga oasis castellum at Dush Kysis (EG). This bangle type has only been retrieved from hill forts and castella. The Type C bangles also came from various camp types in the north-western provinces except for one C3-variant from the Kharga oasis castellum at Dush Kysis (EG). The Type D bangles were concentrated in army camps of the north-western provinces; nonetheless, a more widespread occurrence appears up to the provinciae Lusitania (FiãesVila da Feira [PT]) and Arabia (Mezad Tamar [IL]). 5.3.4.2.2 Finger rings Only 6 finger rings have come from military settlements. One came from the fortress of Birdoswald (UK) at Hadrian‘s wall and another of Type A4 came from the southern gate in Regio 20Z of the intensively excavated castrum at Kaiseraugst (CH). The four other come from hill forts in France and Germany: Lugasson (indefinite), Vireux-Molhain, and Kreimbach-Kaulbach (1 ex. A1; 1 ex. B6). All point to a use in Period III. 5.3.4.2.3 Beads The 29 black glass beads from military settlements were retrieved from 13 camp sites (4 castra; 6 castella; 3 hill forts). The main types, aside from the folded spacer-beads with double suspension holes (6 ex.), are annular, globular, and cylindrical beads, respectively 7 ex., 4 ex., and 4 ex., but lentil-shaped and polyhedron beads also occur in the military community. The annular, globular, and cylindrical beads display a distribution within the army camps of the north-western provinces, regardless of the 4 beads from the Kharga oasis castellum at Dush Kysis (EG), while the spacer beads were retrieved in Central European camps in the provinciae Raetia, Noricum and Moesia. Monochrome black-appearing beads as well as bichrome and polychrome beads have been recorded. The applied decoration is dotted to create the so-called ‗crumb‘ beads (South Shields (UK)), spiralling (sometimes festooned) or in zigzag (sometimes in double interwoven lines with applied eyes within the created lozenges). Typically, opaque white and opaque blue glass is applied to decorate the beads, but yellow sometimes also has been used. It is unclear whether these beads can be referred to women only or also to the soldiers and their children, seeing the meagre information that can be taken from the archaeological contexts. The 15 small annular beads from Magdalensberg (AT) formed part of a necklace with 221 beads in various other glass hues that came from the st transition level 2a of the cistern N/G and are dated in the first half of the 1 century AD (Czurda-Ruth 1998, no.2009-2230, pl.3). 186 The one from Grand Armoises (FR) is not taken into consideration because the character of the site is unknown to us. 213 5.3.4.2.4 Pendants Only 3 of the recorded pendants came from military settlements. Two small globular shaped pendants came from the castrum at South Shields (UK) [cat.nos.4227; 4230], and a jug-shaped pendant with applied glass trails came from the castellum at Dush Kysis in the Kharga oasis (EG)[cat.no.818]. 5.3.4.2.5 Gems The 52 (53) black glass gemstones recorded in military settlements have come from 16 (17) sites which are 187 situated in the north-western provinces, with a concentration in Britain and the Rhine-region. One piece came from the castellum at Oudenburg (BE), one from the castellum at Liberchies (BE), and one from the castellum at Vechten (NL). All other material has been retrieved in Britain and Germany (Table 104). Most are pseudo-nicolo gemstones in a blue-on-black layered glass (45 ex.), but some are monochrome black (8 ex.). Most are loose finds (35 ex.), but some are still set in a finger ring (15 ex.) and 3 are still set in a discoid brooch (T270-271) [see Chapters 3 and 7]. The 21 pieces from the Saalburg Limes castellum at Bad Homburg (DE) demonstrate that the many army camps showing only one or a very few gemstones give a distorted idea of the gems‘ occurrence and that the number of pieces depends on the intensity of the excavation activities on the site. Table 104: Gemstones from military sites cat.no. site camp type hue shape n period black+blue black black+blue black+blue black+blue black black black+blue black+blue black+blue black+blue black black+blue black black+blue black black+blue black+blue black black+blue black Henig F4 discoid brooch (T270) Henig F2/4 Henig F2/4 Henig F2/4 Henig F2/4 plano-convex Henig F2/4 Henig F2/4 Henig F2/4 Henig F2/4 Böhme 45 1 1 21 1 1 1 1 1 1 1 1 1 6 1 4 1 1 3 1 4 1 III II-III II-III II-III II I I II-III IV II-III II-III II-III 345 480 1773-93 1981 1998 2110 A4617 2996 3918 3919 3924 A4542 Liberchies (BE) Oudenburg (BE) Bad Homburg (DE) Neuwied-Niederbieber (DE) Regensburg (DE) Wiesbaden (DE) Nijmegen (NL) Vechten (NL) Birdoswald (UK) Birrens (UK) Burnswark (UK) Caernarfon (UK) castellum castellum castellum castellum castellum castellum castrum castellum castellum castrum castrum castellum 3977-83 Chesterholm (UK) castellum 4040-43 4069 4070 4186-88 4185 4224-4226;4229 4222 Corbridge (UK) Hod Hill (UK) Housesteads (UK) castrum castellum castrum Richborough (UK) castrum South Shields (UK) castrum 5.3.4.2.6 Henig F2/4 Henig F2/4 Henig F5 Henig F2/4 Henig F2/4 Böhme 45 Henig F2/4 Böhme 45 II-III I-II I II-III II-III IV II-III IV Hairpins Hitherto one globular head of a hairpin in black glass has been recorded from a military context. It is a piece from the late Roman camp at Oudenburg (BE) between Ostend and Bruges which was occupied in interrupted phases between c. 200 and c. 410 AD as a castellum of the Litus Saxonicum (Vanhoute 2009; Vanhoute et al. 2009). Unfortunately, the precise context of this unpublished piece remains unclear to date. Also, two hairpins in pale blue and pale green glass were retrieved from the Roman fort at South Shields (UK) at the east end of Hadrian‘s wall (Allason-Jones, Miket 1984, 275, nos.4.1-4.2). 5.3.4.3 Architectural decoration No architectural decoration material in black glass has been recorded in military settlements so far. 5.3.4.4 Counters The 227 (253) recorded black glass counters that have been retrieved from military settlements are to be considered as a set of examples, as many more should pop up when a more meticulous heuristics is undertaken. Table 105 lists the contexts of 152 ex. or 66,7% from Period I camps, 6 ex. or 2,6% from Period II camps, 2 ex. or 187 This can be explained by the detailed information on this particular material from the general publications by Martin Henig for British material and Antje Krug and Gertrud Paltz-Horster for the German material. 214 1,3% from Period III camps and only 1 ex. or 0,4% from Period IV camps. The remaining 66 pieces or 28,9% came from less well-defined chronological contexts. Table 105: Counters from military sites cat.no. site camp type shape n period small-round small-oval indefinite small-round small-round large-round indefinite small-round small-round indefinite small-round small-oval small-round small-oval small-round medium-round small-oval small-irregular indefinite small-round small-round small-round medium-round small-round small-oval small-irregular large-round small-round small-round indefinite small-round medium-round small-round small-round small-oval medium-oval small-quadrangular medium-round small-round 1 1 94 2 1 1 1 1 2 1 2 1 10 1 26 2 2 2 1 1 3 1 1 24 4 1 3 6 2 6 1 1 2 41 2 1 1 1 4 I 35-105 A4511-35 Magdalensberg (AT) castellum 343;351 465-66 Liberchies (BE) Oudenburg (BE) castellum castellum 1025 1795 1870-71 1941 1954-56 Aulnay (FR) Bonn (DE) Epfach (DE) Mainz (DE) Moers-Asberg (DE) castellum castellum, fabrica hill fort castellum castellum 1970-80 Neuss (DE) castellum 2908-10;2913-17;2919; 2921-27;2929-36;2942; 2945-50;2953-54 Nijmegen (NL) castellum 2989 2993-95 3638;3640 Utrecht (NL) Vechten (NL) Kaiseraugst (CH) castellum castellum castrum 3925-56 Caerleon (UK) castellum A4536-41 4032-39 Caernarfon (UK) Corbridge (UK) castellum castrum 4180-81 Richborough (UK) castrum 4220-4221 4246-90 South Shields (UK) Usk (UK) castrum castellum 4299 4316-19 Water Newton (UK) York (UK) castellum castellum III-IV II-IV I I-IV IV I-IV I I I I-III I-III III-IV I-III III-IV II-IV III-IV I-IV I-IV III-IV I-IV We acknowledge that the catalogued material displayed in Table 105 is only a fraction of all black glass counters from military sites, seeing that most entries have come from the north-western provinces. Despite this discrepancy, the occurrence of small round counters at castella (16) in the Rhine region and Britain outnumber that at castra (4) and hill forts (1). Two thirds of the material has come from Period I contexts, but few lead to a more detailed understanding of the contexts. There is, for instance, no further contextual information available regarding most material from Magdalensberg (AT) except for the two from the fill of cistern NG/20 [cat.nos.52-53] (Czurda-Ruth 1998, VI/10-13). The 32 black glass counters from the fortress of Caerleon (UK) came from the bath complexes (Brewer 1986, 155). Also both pieces from the hill fort of Lorenzberg at Epfach (DE) and the 33 counters from Kops Plateau at Nijmegen (NL) came from a public context, correspondingly a storehouse (Werner st 1969, 180, pl.38:34; 50:9) and a horreum from a fill of the mid-1 century AD (van Lith 2009, 33-34). From the 1965-1976 excavations at the fortress of Usk (UK), 87 small glass counters were retrieved, of which 45 were black, and the remaining were of opaque glass: 38 white, 2 blue-green, 1 green and 1 ultramarine blue (Price 188 1995, 129). All of them except three came from public and domestic features connected with the fortress , e.g. drains, ditches, waste pits and latrines, with a striking concentration (57 ex. = 65,5%) at both sides of the via principalis, ‗particularly in the compound and structures in Area 2 and in Compound 3 to the north of the officer‘s house in Area 5‘ (Price 1995, 129). 188 It is unclear whether the counters are deposits from the occupation, abandonment or destruction. 215 5.4 Other Our data-base also incorporates black glass artefacts from caves and rivers. Because neither can be included within the settlement categories with a civil or military character, we found it appropriate to discuss the black glass artefacts from caves and rivers separately. 5.4.1 Caves Hitherto, only a few black glass artefacts have been reported from cave digs, but we assume more Roman occupancy in caves than only the few attested in the Belgian Ardennes and in the French Alps (Table 106). Because it is self-evident that this peculiar Roman context category is limited to karst regions, it would be interesting to check all caves within these regions for Roman (black glass) artefacts to improve our understanding 189 on the use and function of caves throughout the Roman period. It is interesting that the black glass material catalogued from the known six sitesfive Belgian and one Frenchshow great homogeneity. The black glass material that has been recorded so far is limited to jewellery and more particularly to bracelets, finger rings and 190 hairpins. Table 106: Roman black glass artefacts from caves (n = quantity) cat.no. artefact typology site type 169 304 516 302;305-6;314-15 164-68 163 433 finger ring finger ring finger ring bracelet bracelet bracelet bracelet type B6 var.1 type B6 var.2 type B7 Type A2 Type A3 Type A4 Type A4 308 310 269 301;303;307;312 270 309;311;313 1140-41 bracelet bracelet bracelet bracelet Type A5 Type B2 Type C1 Type C2 hairpin bracelet globular head Type D1 TOTAL arm ring finger ring hairpin Trou des Nutons – Aiseau-Presles (BE) Trou de Han – Han-sur-Lesse (BE) Trou del Leuve – Sinsin (BE) Trou de Han – Han-sur-Lesse (BE) Trou des Nutons – Aiseau-Presles (BE) Trou des Nutons – Aiseau-Presles (BE) ; La Roche aux Corneilles – Marches-lesDames (BE) Trou de Han – Han-sur-Lesse (BE) Trou de Han – Han-sur-Lesse (BE) Trou de l'Ambre – Eprave (BE); Trou de Han – Han-sur-Lesse (BE); Trou de l'Ambre – Eprave (BE) Trou de Han – Han-sur-Lesse (BE) La Grotte du Seuil-des-Chèvres – La Balme, Savoie (FR) period n II-III II-III II-III II-III II-III II-III II-III 1 1 1 5 5 1 1 II-III II-III II-III II-III 1 1 1 5 II-III IV 3 2 22 3 3 To determine the context wherein we have to consider the presence of Roman black glass artefacts in caves, it is necessary to understand why there has been Roman presence in caves as such. Because no burials of Roman imperial times have yet been reported in caves, we have to ascribe caves to be ideal locations to hide from rd th threats in periods of political and military instability and insecurity throughout the 3 5 centuries AD (Warmenbol 1984). All material from other caves show a coherent Period III assemblage apart from ‗La Grotte du Seuil-desChèvres‘ at La Baume (FR), which comprised two Period IV bracelets of Type D1 (Nicod, Sordoillet, Chaix 1998, 49, 71-73, fig.28:1) [see Chapter 4]. Hence, we are tempted to assume cave settlements as clandestine safe harbours with a temporary character during two specific periods on the basis of the chrono-typologies of the retrieved material [see Chapter 3]:  189 th th The Type D1-braceletsdated at the end of the 4 and especially at the start of the 5 century ADare therefore to be seen in connection with the massive movements of the Migration People between 350/360 and 410 AD. Because this subject matter is only a side-issue within the framework of this project, this topic could not be verified properly. Furthermore, it is more relevant to establish first regional idiosyncrasies by studying all Roman archaeological artefacts from caves of a particular region and confront it with the peculiar Roman artefacts from caves in other regions. 190 We have omitted the vessel fragment [cat.no.316] from Trou de Han (BE) because the sampling demonstrated the piece to be of obsidian because of the near-absence of soda and the very high concentration of alumina [see Chapter 8 – sample PC8f]. 216  All other material from the other caves can be connected with the insecurity caused by foreign invaders rd in the third quarter of the 3 century AD but can equally be related to the political instability in the second rd 191 half of the 3 century AD when civil war opposed different factions. The cave of Trou de Han at Han-sur-Lesse (BE) forms an exception. The site has been given a ritual character considering the context situation, the disposition of the deposits and the huge amounts of small finds in various materials (Warmenbol 1996; 1999). Trou de Han takes a special position within the religious contexts because from the late Bronze Age/early Iron Age onwards, a wide variety of materialmainly weapons and jewellerywas thrown at the river curve before the river disappears under the rocks. This sub-aquatic ritual site is not coincidentally situated just at the curve of the river Lesse. After disappearing at one side the river resurfaces at the other side of the hill rock. This place must have received in ancient times a religious connotation that most likely stood in connection with a belief in an afterlife (Warmenbol 1996; 1999). It is not so difficult agreeing to a correlation with the ancient beliefs that the river entered the underworld where daylight got replaced by total obscurity to reappear as reborn at the other side. Part of the gigantic ritual deposit in the river accumulated over centuries consists of Roman artefacts, of which 15 were in black glass. Besides 11 bracelet fragments and 1 intact finger ring, 3 globular pin-heads in black glass of hairpins turned up. No black glass vessel has been retrieved to date apart from the indefinite vessel fragment that is made of obsidian [see Chapter 8]. To sum up, we assume that the few black glass artefacts reported from caves demonstrate a specific type of context in use during a particular period of time. Much more of this material is most likely at hand but it is still ignored, as caves are in the main studied by prehistorians who have no adequate references at hand to date correctly the artefacts in black glass. It is therefore not surprising that the material from ‗La Grotte du Seuil-desChèvres‘ at La Balme (FR) erroneously is linked to the La Tène-period, while it should be linked with a late Roman presence as attested by the pottery and the coins (Nicod, Sordoillet, Chaix 1998, 71-73). Nonetheless, the rd th deposit of black glass jewellery, which perhaps coincides with the turbulent period of the mid-3 to early-5 century AD, has foremost a ritual character. 5.4.2 Rivers The number of catalogued black glass artefacts that were brought up from river contexts is too small to construe a general view from it. Furthermore, the information on the provenance of the catalogued pieces15 objects from 6 sitesis sometimes dubious. Table 107 shows the majority came from one specific area, the river Durme (BE) near Temse and Hamme, where huge amounts of archaeological material were brought up during dredging works th 192 rd th at the start of the 20 century (unpublished). The majority, if not all material, is to be considered 3 4 century AD, but no clear date can be given from the context information. The only understandable river context is that from Trou de Han at Han-sur-Lesse (BE) [see section 5.4.1.]. In contrast with the clear ritual character of the deposits in Trou de Han, all other river sites delivering black glass artefacts are unclear contexts or even from doubtful provenance. Compared to the given specifications of Trou de Han, it is reasonable to believe that the concentration of material from the dredged area in the river Durme is the display of ritual deposits. The black glass artefacts that have been reported in the collection of Van BogaertWauters, Hamme (BE)8 bangles, 3 finger rings and 2 beadsgenerates some questions about the authenticity of its provenance. These 13 black glass artefacts have different provenances: 2 came from Tielrode (BE), 5 came from Hamme (BE) and the remaining part is without provenance, but all are reported to have come from dredging the Durme. In particular, the bangles are rather anomalous. Because several pieces are unusual types for the north-western provinces and very common to the (eastern) Mediterranean region, the black glass bracelets could equally have been (partly) bought on the antique market. Other river finds are a massive silver finger ring with nicolo paste gemstone that was found in the river Witham at Bardney (Lincolnshire, UK); and a golden finger ring with a set-in nicolo paste gemstone that was reportedly 191 In this period, the north-western provinces generated a dissident Gallic Empire (260274 AD) to uncertainty and insecurity for the collaborators of the subsequent oppressed factions. The Gallic Empire resulted from the incapability of Rome to answer the Frankish intrusions in 260 AD and new attacks in 275276 AD emerged after the fall of the Gallic Empire. 192 The objects collected by Alphonse Van Bogaert (18821967) are said to come from the river Durme, not far from Tielrode (BE) where dredging works have been undertaken. He accumulated 5400 objects and deposited them in 1953 at his private museum in Hamme (BE), the Museum Bogaert-Wauters. Because the black glass material is not only atypical for the Roman north-western provinces, and idiosyncratic to the Levant and Egypt [see Chapter 6], we have to assume that Alphonse Van Bogaert was also active on the antique market. We unfortunately were unable to investigate the history of the collected black glass artefacts in the Museum Bogaert-Wauters, but it seems important and necessary to re-evaluate the Van Bogaert collection to distinguish the artefacts from local provenance and those purchased from the antiquarian circuit. 217 dredged up from the river near Mainz and now is in the Landesmuseum at Mainz (DE). Both discoveries yielded no precise details about their provenance or context conditions (Henig 1978², 194, no.62; Krug 1978, 491, no.11, Taf.51:11), but it is very reasonable to believe that both precious jewels were thrown into the river together with other material as an offering when appealing to the gods for help or to fulfil a promised oath after having received what has been implored of the gods. We discuss subsequently the issue concerning the use and function of 193 cheap glass gemstones imitating semi-precious stones set in gold or silver finger rings [see Chapter 7]. Table 107: Roman black glass artefacts from rivers cat.no. artefact type typology 3912 gemstone Henig type F2/4 2089 gemstone Henig type F2/4 2045 arm ring Type A1 292-296 finger ring arm ring spacer-bead arm ring finger ring arm ring finger ring arm ring spacer-bead gemstone type A5 var.2; B6 var.1 Type A2; D1 ‗Trilobitenperle‘ Type B3 indefinite Type B1; B3; D1 (2x) 526 626-631 TOTAL site Period quantity River Witham at Bardney (Lincolnshire - UK) River Rhine near Mainz (DE) (no exact provenance) Channel at Trier (DE) (no exact provenance) River Durme at Hamme (BE) II-III 1 II-III 1 II-III 1 II-III II-IV III II-III II-IV 2 2 1 1 1 5 3 9 1 2 River Durme at Temse (BE) River Durme (BE) (no exact provenance) The black glass jewellery from caves and rivers demonstrate that such material category found acceptance in the traditional ritual practices to be deposited as offerings in the late Roman period. We think that this is due to the absence of black glass jewellery during the early Roman Empire and not because of a new belief with newly introduced ritual practices. Also excluded, to our opinion, is that black glass jewellery was only considered rd suitable as ritual offering material from the 3 century AD onwards. 5.5 Conclusion The rather introductory character of this contextual analysis on black glass artefacts and the provisional disposition of some conclusions put forward marks this as research in the introductory stage. That the investigation incorporates material from a vast area and for a long-lasting period did not make the analysis easier. We faced various heuristic problems, not in the least because much material is still unpublished and unstudied. Hence, it is reasonable to believe that plenty of material might have been overlooked and might appear sooner or later. Bearing in mind the incompleteness of the database, we are fully aware that it was beyond our reach to realize an overall unbiased picture of the contexts wherein the black glass artefact types were consumed 194 throughout the consecutive four periods. Different factors influenced the presented numbers and percentages and consequently these numbers should be considered as guidelines. However, we are convinced that these results, how preliminary they may be, already offer a number of interesting new conclusions on particular issues and provide good evidence to corroborate some assumptions or contest others. The analysis of the available data per context and further per functional type provided interesting results that explained to a certain degree the use of black glass artefact types by taking into consideration the chronological order. Compared to the civil contexts, the frequency of black glass vessels in military milieus is much lower. The black glass vessels of Periods I and II are present in domestic as well as in funerary contexts but absent in ritual contexts. The black glass vessels of funerary contexts have mostly come from very rich tombs that include a large amount of luxurious grave goods. Seeing that most Period IV vessels known today, i.e. the rod-formed balsamaria and unguentaria, are those without provenance in very many museum collections, it is impossible to say something definite about these vessels. All pieces from known provenance are from burial contexts (see below), 193 It is interesting to see that similar gemstones set in gold finger rings have come from ritual depositse.g. the Regensburg-Kumpfmühl treasure. This was partly due to a quantity of material that remained inaccessible or was simply unknown to the author. Secondly, this is caused by the limited information the author could extract from published material or, if unpublished, from card indices in museums and archaeological depots. 194 218 but it would be odd to exclude them from settlements because both vessel shapes are described as toilet ware, a typical daily life commodity [see Chapter 7]. The omnipresence of the black glass bracelets is evident from its frequent occurrence in whatsoever context: production, domestic, burial, ritual or dump on civil and military settlements. But black glass bracelets were in Roman society not just ordinary pieces of adornment. A careful contextual analysis of the available material from burial contexts provided evidence demonstrating the black glass bracelets were only worn by women and juvenile girls on their left wrists. No clear provenance has been registered for a large part of the beads presented here since they were mainly th th retrieved in the 19 or early 20 century before entering the major museum collections, where they never received proper interest. Hence, the data of the pieces with a known provenance remain very limited. For the majority of the pieces, only the country or the site of origin is known, while a precise location or the information on their context has never been recorded. The bulk of the round or elliptical discoid brooches with a plain conical gemstone in black glass set in the centre (Hull Types 270-271) are characteristic at military settlements [see this chapter]. However, they have been regularly reported at temple sites seeing that Grace Simpson and Beatrice Blance report 15 pieces from 11 temple sites in Britain only such as Nornour, Uley, and Cold Kitchen Hill (Simpson, Blance 1998, 277). The placing of brooches in votive deposit finds cannot be considered the result of casual loss but are the effect of ritual evidence like most other brooch types, indicating a pre-Roman origin of the practice that continued throughout Roman times (Simpson, Blance 1998, 278). In brief words we assume that the presence of discoid brooches can help recognize the presence of military men of a certain rank (officers) within the civil settlements after having fulfilled their function. Counters in glass, polychrome and monochrome, are omnipresent at Roman sites, whether they are burial contexts, settlement contexts or military contexts. They appear to be fully absent in ritual contexts. Though for the most part discarded pieces are found singly, these counters can occur in sets in tombs and on odd occasions within a wooden box together with a wooden gaming board [see Chapter 7]. Everything considered we wish to stress that future research could possibly enhance a similar comparative study on the black glass artefacts from civil versus military contexts and from rural versus urban contexts of civil settlements to allow for a better understanding of the consumption patterns of black glass commodities. 219 220 Chapter 6 PRODUCTION, DISTRIBUTION AND TRADE 6.1 Introduction With the study on distribution and trade of black glass artefacts in the Roman Empire, we sought to find out the nature and extent of its consumption during the entire Roman imperial period. The challenge of this chapter is to incorporate the obtained data of this minor group in the Roman material culture within the current view on trade and distribution of Roman glass in general. The obtained data on production and distribution will be discussed in detail in Chapter 10 by considering the different Roman glass workshops generating black glass artefacts, suggesting a macroeconomic influence on regional consumption patterns demonstrating the impact of regulations causing a far-reaching regionalisation in organising the glass production during the Roman imperial period. To determine how to understand black glass artefacts in the context of the production, distribution, and trade of Roman glass we discuss firstly the production of black glass with a reflection on the prevailing concept(s) on the organisation of the Roman glass production. Secondly, we examine the distribution level of black glass artefacts in order to offer useful indications on trade. Considering that our research supplied insufficient evidence concerning the trade of black glass to discuss the issue separately, we decided to integrate the obtained hypothetical schemes and ideas deduced from the study within the section on distribution. The distribution patterning of particular artefact types would help identify local and regional idiosyncrasies indicating a consumption pattern on a local, regional or interregional level, and ultimately to discriminate particular assemblages related to specific glass workshops. One of the goals is to verify whether black glass artefacts are commodities of (1) local, (2) regional or (3) long195 distance trade. 1. 2. 3. A local trade of artefacts involves a production by a wide range of local production centres meant for a local distribution that does not reach further than a regional level. In this situation, the set up of a typology characterises the assemblage of one or more local workshops within a specific region. A regional trade implies the existence of several workshops spread all over the empire for a more regional or interregional distribution. Concerning the regional trade, we should check whether regional differences can be observed and if they evolve differently over time. Long-distance trade implies an empire-wide distribution of the goods produced in one single workshop or eventually in several glass workshops within a specific centre or limited region. In this case, we probably can deduce not much more than a chronological significance for the various types of material. The glass production as well as the consumer‘s behaviour changed significantly during the five centuries of the Roman imperial period. To minimise the exclusion of any additional evidence concerning the impact of black glass production on the various levels of consumption in Roman society, we verify whether the changes in levels of trade and distribution of goods are also regionally bounded. We therefore viewed the material diachronically, using tables and distribution maps of the different functional types per period. For practical reasons, we kept the chronological subdivision of four consecutive periods, as formulated previously. Through this approach, key regions became visible depending on the period or the functional type taken into consideration. Black glass artefacts have been recorded from nearly all present-day countries that were part of the former Roman Empire. In a first part will be discussed the production of black glass vis-à-vis the prevailing models on glass production in general. Subsequently we examine the distribution per functional type and per period. Due to reasons of feasibility and readability of this chapter, we focus on production, the vessels and the arm rings. However, the other commodity categories are integrated in the discussion as well. Finally a suggestion of trade pattern per period will be put forward on the basis of the obtained observations from parts one and two. 195 Aware of a wide variety of publications specialized on this issue, we chose a more general definition of the different levels of consumption. 221 6.2 Production As explained in Chapter 2, the current knowledge of the glass production during the Roman imperial period involves a production process in successive stages. In a first stage, raw materialssand and sodawere melted 196 into raw glass, and in a second stage, chunks of this raw glass got re-melted to produce objects from it. Processed in different furnaces, and most likely also in different workshops, the former stage was executed at socalled primary workshops, whereas the latter stage is characteristic of the so-called secondary workshops. For the entire Roman imperial period, the main production area of raw glass, and accordingly the primary workshops, is considered to be situated on the southern Levantine coast and in Egypt, where also are situated the favourite sources of raw materials (Foy, Nenna 2001; Foy, Nenna 2003; Shortland 2004; Nenna 2008a). From there, large masses of chunks of raw glass were transported empire-wide over sea to supply the secondary workshops, as the numerous shipwrecks reveal (Foy, Nenna 2001; Nenna 2008a; Fontaine, Foy 2007). We may therefore assume that the produced glass artefacts must have a very homogeneous composition within the entire empire that remained relatively uniform and unmodified during the entire Roman imperial period (Aerts et al. 2003; Foy et al. 2003; Picon, Vichy 2003; Silvestri et al. 2006). Yet about twelve variants of silica-soda-lime glass have been recognized hitherto. While all twelve are found in the Levant, only four are present in the west (Picon, Vichy 197 2003). Based on Pliny the Elder (Naturalis Historia, XXXVI, 194) describing the production of raw glass in Italy, st Spain and Gaul from local raw materials during the 1 century AD, we may assume that these western Mediterranean primary glass workshops produced raw glasses with a different composition from that of the Levantine glass that might be absent in the eastern Mediterranean and thus not yet recognised. Despite the 198 analysis of sands from Italy (Turner 1956a-c; Silvestri et al. 2006) and Spain , clear evidence on primary glass production in these regions is still lacking. Concerning the discussion on the black glass production, we verified to which of four proposed models (see below) the available archaeological data is to be linked. To bear out the reliability of the conclusions put forward here, we have rechecked these in Chapter 8 with the results obtained from the chemical analysis. The mass of glass artefacts from excavations contrasts sharply with the limited concrete information known so far regarding the organisation of glass production and the precise location of glass workshops (Baxter et al. 1995). Considering this explicitly for Roman colourless glass, it is most markedly for the here-discussed stronglycoloured glass appearing black. The last decennium is, however, characterised by a continuous growth of knowledge on the Roman glass production from an archaeological point of view as well as from an archaeometric approach. The majority of the excavated workshops cannot be assigned to a specific production; moreover they provide an incomplete view of the there produced assemblages. The focus of the international research, however, concerns at this moment the primary production and the trade of raw glass, not meaning that the secondary workshops are remaining untouched. Some publications compile papers on the organisation of both the primary and secondary glass production on hand for the western part of the empire (Foy, Sennequier 1991) and the eastern part (Nenna 2000). Heidi Amrein dealt with a Claudian-Neronian glass workshop (‗Derrière la Tour‘) in Avenches (CH), providing an exhaustive overview of all known glass workshops at that time (Amrein 2001). Her overview also updates the survey on the Roman glass production by Mara Sternini (1995, 137–200). Another overview has been compiled in the exhibition catalogue ‗Tout feu tout sable‘ describing the Roman glass workshops in France and their productions (Foy, Nenna 2001, 34–60). An extended and improved discussion concerning the primary glass production and its distribution has been worked out recently by Marie-Dominique (2008a). Gladys D. Weinberg published an extensive work on the late Roman secondary workshop of Jalame (IL) (Weinberg (ed.) 1988), whereas a number of contemporaneous workshops have been excavated during open mining coal extraction in the Hambacher Forst in the Rhine region between Jülich-Düren-Cologne (DE) (Seibel 2000; Gaitzsch et al. 2003). Despite all these publications, no information has been made available on the making of black glass or about the production of black glass artefacts. We therefore verified the various components necessary to define a site as a glass workshop (Table 108). 196 For further reading on primary and secondary glass production in Roman imperial times, see Sternini 1995; Nenna ed. 2000; Amrein 2001; Foy, Nenna (eds.) 2001; Foy, Nenna (eds.) 2003; Nenna 2008. 197 This finding forms the start of our query in Chapter 8 to set up a consequent strategy in choosing samples of a variety of commodity types from different periods and different areas. 198 This is an issue within the ongoing ERC-StG research project, ‗Archaeometry and Archaeology of Ancient Glass: Production as a Source for Ancient Technology and Trade of Raw Materials‘ (ARCHGLASS) under direction of Patrick Degryse of the Centre for Archaeological Sciences at KULeuven. 222 Table 108: List of glass workshop sites with possible black glass production glass site period date range furnace lumps Avenches (CH) I Ib-Ic AD X X ‗Derrière la Tour‘ Lyon (FR) I Ib-IIa AD X X ‗La Montée de la Butte‘ Kaiseraugst (CH) II IId – IIIA AD X X Äussere Reben Lavoye (FR) Sainte Menehould (FR) ‗Les Houis‘ Augst/ Kaiseraugst (CH) Liberchies (BE) Ore ac (HR) Braga (PT) Horbat Qastra (IL) Trier (DE) ‗Palais Kesselstatt‘ crucibles waste - X - - X X products vessels utensils vesels utensils vessels II+III III – IVA AD - X - X II+III III – IVA AD - X - X III III III III+IV III+IV III – IVA AD III – IVA AD III – IVA AD IV AD IV – V AD - X - X - X X X X - jewellery vessels? jewellery vessels? jewellery jewellery jewellery jewellery jewellery IV IVb AD - - X - jewellery The secondary glass workshop of Aüserrere Reben at Kaiseraugst (CH) is most interesting for the here-presented research. Two buildings at the junction of Regio 17B and Regio 17C yielded on either side a glass workshop respectively active during Period II and Period III (Rütti 1991, 150–153). Black-appearing glass materialfragmented vessels as well as production wasteare in particular coming from the building of Regio nd 17B dating mid-2 century to about 220 AD (Fischer 2009). Although the workshop at Kaiseraugst-Aüserrere Reben only produced vessels, it is clear from indirect material that somewhere in Augusta Raurica a workshop produced black glass jewellery (in particular, arm rings) in Period II and/or III. For most sites, only indirect material is available, and therefore the potential workshops cannot be determined due to the lack of clear evidence. Either there is a multitude of fragmented material and production waste or there is none to connect a production of black glass with the available structures such as furnaces. Both Period I workshops at Avenches (‗Derrière la Tour‘) and Lyon (‗La Montée de la Butte‘) contained furnace structures, but there is no evidence indicating what vessel types were fashioned in black glass nor in what quantity, although black-appearing artefacts have been produced in both workshops. In a recent contribution, we compared the late Roman jewellery in black glass from the secondary workshops of ‗Les Houis‘ near Sainte Menehould (FR) with that from ‗Palais Kesselstatt‘ in Trier (DE), correspondingly of Period II/III and Period IV, to describe the differentiation of rural and urban workshops (Cosyns 2009). Another contribution in the same publication deals with the late Roman glass workshop of Braga (PT), where black glass jewellery was produced (da Cruz 2009). The most indirect assignment to a glass workshop is that from Horbat Qastra (IL) based on the huge amount of similar material from the late Roman–early Byzantine cemetery of Period IV (Gorin-Rosen, Katsnelson 2008, 74) [see below and Chapters 8 and 10]. In Table 108 is assembled the evidence arguing for the presence of a secondary workshop responsible for the manufacturing of black glass artefacts, whereas Table 109 lists the sort of production per workshop. Mainly from urban settlements, Trier (DE), Braga (PT), Augst (CH), Avenches (CH), and Lyon (FR), some workshops have a rural character. The workshops from ‗Les Houis‘ near Sainte Menehould (FR) and those at Lavoye (FR), both in the Argonne forest, appear to be rural workshops that were engaged in large-scale production of black glass (jewellery) objects. Other rural workshops considered to have produced black glass artefacts are located in Liberchies (BE) and Ore ac (HR). All four examples of rural workshops are located on a main road-axis connecting them easily to major urban settlements. 223 Table 109: List of black glass products from sites with glass workshops finger site production vessels bracelets rings Avenches (CH) X Lyon (FR) X Kaiseraugst-Äussere Reben (CH) [Augst/ X X [X] Kaiseraugst] LavoyeBerthancourt/Champ X X X Damloup (FR) Sainte Menehould-Les X X X X Houis (FR) Liberchies (BE) ? X X Ore ac (HR) ? X Braga (PT) X X X Horbat Qastra (IL) ? Trier (DE) X X X 6.2.1 beads pendants gems hairpins counters - - - - - - - [?] - - - - - - - - - - X - X X - - X X X X X X X - Organisation of the production of coloured glass Here above and in Chapter 2 is explained the organisation of the production process of glass in various levels (i.e., primary and secondary workshops). But while a consensus exists on the production of raw glass in primary workshops and finished goods in the secondary workshops, it is less well-defined where the colouring of the glass must have taken place during the Roman imperial period. The colouring of raw glass is commonly considered as a production process carried out in the primary workshops where the raw glass got produced or in specialised intermediate workshops from which coloured glass chunks or cakes were exported. This practice is already attested for the late Bronze Age considering the conical cakes in blue glass from Tell-el-Amarna (EG) on the Ulu Burun shipwreck (TR) (Nicholson 1995; Nicholson et al. 1997; Jackson, Nicholson 2010) and the specialised workshops from Qantir (EG), where the production of strongly coloured glass cakes by colouring imported raw glass took place (Rehren 1997). In the Hellenistic period, a comparable organisation in supplying secondary rd workshops still prevailed, as can be demonstrated from the deep ultramarine blue glass chunks on the 3 century BC shipwrecks Sanguinaires A (FR) (Alfonsi, Gandolfo 1997) and Lequin 2 (Pomey, Long 1992), or for the late La st Tène period with the black-appearing dark purple glass cakes on the 1 century BC sites of Odijk (NL) (Schuuring 2007, 126–128, fig.10.3), Basel-Gasfabrik (CH) (Berger 1974, 63; Berger, Furger-Gunti 1980, 63), Manching (DE) (Krämer 1960, 197, pl. XXIVb; Gebhard 1989b, 148, pl. 37, 484), Hengistbury road (UK) (Henderson 1987a, 160– 163; 1987b, 180–186) that all must have a Mediterranean provenance seeing the similarity in composition with the early Roman counters in deep purple glass (Cagno et al. forthcoming). Whether the colouring got executed in specialised intermediate workshops or in the primary workshops themselves remains unanswered, but it is certain that the glass that arrived in the secondary workshops was coloured. This model is still widespread in the Roman imperial period, as demonstrated by recent archaeological research on the location of the primary workshops and on the trade routes for providing the secondary workshops (Sternini 1995, 127–135; Foy, Nenna 2001, 100–112; Nenna 2007, 125–131). This production and distribution model, however, seems to experience important changes during the early Roman imperial period (i.e., Period I) in relation to the strongly coloured glass. Various shipwrecks from the Roman imperial period provide hard evidence for the long-distance distribution of large chunks of raw glass on a large scale to supply secondary workshops all over the empire from decolourised and ‗naturally coloured‘ glass (Foy et al. 2000; Nenna 2008a, 130–131). None of the retrieved chunks or finished products was coloured black, but solely ‗naturally‘ coloured or decolourised. This fact pleads for 1) a regional colouring process in a centre specialised in the production of semi-finished products such as glass cakes, implying an intermediate production stage to supply the local market of cakes of coloured glass; or for 2) a local colouring process within the production needs of each concerned secondary workshop. The colouring process of black-appearing glass in the north-western provinces from Period II has recently been demonstrated (Van der Linden et al. 2009), but in Chapter 8 we present a well-argued case in favour of the above-mentioned hypothesis 2. Various possible models can be proposed to explain the organisation of the glass production from the raw materials to the finished products. Some consistent schemes have recently been set up concerning the possible 224 supply of raw glass from the primary workshops to the secondary workshops (Freestone et al. 2002, 258-259, 199 figs.1–2; Rehren 1997; Komp 2009, 205–208, fig.98; Freestone et al. 2009, fig.3) : o Scheme 1 (Figure 109): The production of raw glass from raw materials, the colouring as well as the remelting of raw glass chunks to manufacture artefacts is done in one production centre or in various glass workshops in close proximity. This scheme emphasises a glass production through one or very few large self-supplying production centres to generate finished products for an empire-wide distribution. PRODUCTION CENTRE RAW GLASS – COLOURING –ARTEFACTS MARKET Figure 109: scheme 1 o Scheme 2 (Figure 110): This scheme underlines a glass production by one or few primary workshops specialised in manufacturing coloured glass and supplying secondary workshops all over the Roman Empire with raw glass ‗naturally coloured‘; decolourised; as well as deeply coloured, including blackappearing glass. PRIMARY PRODUCTION RAW GLASS - COLOURING SECONDARY WORKSHOP ARTEFACTS MARKET Figure 110: scheme 2 o Scheme 3 (Figure 111): This model defines the activity of intermediate specialised workshops responsible for manufacturing coloured glass, including black glass from incoming raw glass to supply the secondary workshops. The two proposed variants are given because it is very likely that these specialised intermediate workshops were focussed on the making of one specific glass hue to produce black-appearing glass, depending on the level of technical know-how at hand. Consequently, each colour becomes idiosyncratic for various secondary workshops in one region, and each colour provides evidence of a trade route and distribution pattern with one origin. Perhaps because it seems easier to transport colouring agents to the existing glass workshops than to implant a glass workshop where colouring oxides are extracted, the idea of intermediate centres specialised in colouring glass near a local exploitation of a colouring agent seems plausible. Not wanting to exclude this model, it is important to stress that such a chaîne opératoire is difficult to trace. 199 The model in Ian Freestone et al. (2009) will be discussed in Chapter 8, seeing that it verifies how to understand the possible batch variations and the implications in the composition of the glass matrix. 225 PRIMARY PRODUCTION RAW GLASS INTERMEDIATE CENTRE COLOURING (GREEN; PURPLE; BROWN; BLUE) SECONDARY WORKSHOP SECONDARY WORKSHOP SECONDARY WORKSHOP SECONDARY WORKSHOP ARTEFACTS ARTEFACTS ARTEFACTS ARTEFACTS PRIMARY PRODUCTION RAW GLASS INTERMEDIATE CENTRE INTERMEDIATE CENTRE INTERMEDIATE CENTRE INTERMEDIATE CENTRE COLOURING COLOURING COLOURING COLOURING SECONDARY WORKSHOP ARTEFACTS GREEN SECONDARY WORKSHOP SECONDARY WORKSHOP ARTEFACTS ARTEFACTS PURPLE SECONDARY WORKSHOP SECONDARY WORKSHOP ARTEFACTS ARTEFACTS BROWN SECONDARY WORKSHOP SECONDARY WORKSHOP ARTEFACTS ARTEFACTS BLUE SECONDARY WORKSHOP ARTEFACTS Figure 111: scheme 3 (top: variant a – bottom: variant b) o Scheme 4 (Figure 112): The colouring is done at the secondary workshop, resulting in a workshopspecific black glass. When using a proper recipe every secondary workshop is featuring an idiosyncratic black glass but when most secondary workshops within a regione.g. a civitas or a provinciaprocessed black glass according to a strict recipe and were supplied with the same raw glass and coloring agents no distinction will be noticeable concerning the chemical composition. PRIMARY PRODUCTION RAW GLASS SECONDARY WORKSHOP SECONDARY WORKSHOP SECONDARY WORKSHOP SECONDARY WORKSHOP COLOURING COLOURING COLOURING COLOURING + ARTEFACTS + ARTEFACTS + ARTEFACTS + ARTEFACTS Figure 112: scheme 4 226 Apart from the first model, the proposed models assume the supply of raw glass at the secondary workshops for the local production of glass artefacts. Model 1 and 2 involve a limited number of people acquainted with the glass technology and/or investing in the glass production. Both models imply a rather restricted production capacity and thus a limited availability of goods in glass. Model 1, with self-supplying and all-controlling production centres, fits well in the view of regional production of raw glass and consumer goods, as can be deduced from Pliny the 200 Elder‘s descriptions on the production of raw glass in Italy, Spain and Gaul to supply the local market. Hitherto, valuable evidence is lacking with the exception of the primary production in Syro-Palestine and Egypt. Model 2 entails the presence of secondary workshops nearby the primary workshops and in the large centres. Model 3 implies a much more sophisticated and levelled organisation of the glass production process, but still with the technological capacities monopolised by a restricted number of centres. Model 3 may be regarded as a much more developed version of model 2 and thus more befitting a classical organisation of glass production. Both models 2 and 3 are workable for the Hellenistic and early imperial period (i.e., Period I), as shown from the above201 mentioned example on the late La Tène glass cakes. But as far as we know, there is no clear evidence available revealing how the colouring process was organised. Finally model 4 implies an organisation that experienced an enormous change to remain economically cost-effective, where the primary workshops had to focus on the production on ‗naturally coloured‘ and decolourised glass. The secondary workshops themselves produced the colours required for the local and regional market. Model 4 is applicable to the north-western secondary workshops of Period II concerning the production of black-appearing glass (Van der Linden et al. 2009). However, it goes without saying that the Roman glass production in general was much too complex to fit within one single model, and thus more than one model must have been prevailing at the same time. The Roman market was not solely supplied from the local secondary workshops. The distribution patterns of particular functional types and shapes, discussed further in this chapter, demonstrate the existence of production centres with a regional or even inter-regional character of the trade radius. For instance, the Period II shipwreck from Embiez (FR) with a shipload containing piled-up, finished products in glass, such as vessels and window panes, besides the mass of raw glass chunks (Foy et al. 2005; Fontaine, Foy 2007) demonstrates that simple commodities in glass got distributed and traded overseas over long distances. None of the shipwrecks with glass as shiploads contained black-appearing chunks or consumer goods. The challenge is now to distinguish the peculiarities of the workshops producing black glass artefacts with the available archaeological data. This information and that in relation to the eventual intermediate workshops responsible for colouring the glass (black) remains, as expected, rather very limited at this point. Within the archaeometric approach, we verified the results from chemical and optical analysis on a large number of artefacts to provide additional information on the production and distribution of black glass and to exclude any biased statements. An answer is formulated on the questions whether secondary workshops have been supplied from one or various intermediate or primary workshop(s) or conversely that each secondary workshop produced its own black glass [see Chapter 8 and 9]. The proposed theoretical models do not take into account contaminations due to the addition of cullet to the batch, because there is hitherto no archaeological evidence proving the recycling of black glass cullet. From ancient writers, only the well-known quote by Pliny the Elder (NH XXXVI, 194) referring to the re-melting of glass to produce black glass is familiar to us. Also, the archaeological evidence is rare. Believed to be indirectly connected with recycling of black glass are the huge amounts of glass tesserae at the late Roman secondary glass workshop of Les Houis near Sainte Menehould (FR). These tesserae are considered a sort of fine chopped readymade raw material that can be added to the batch as colorant, or to be used as an easily measurable system to add specific amounts of coloured glass to the crucible in order to manufacture a certain amount of artefacts (Foy 2007, 43). In Chapter 8, on chemical analyses, and Chapter 10, on the historical approach, the issue is discussed furthermore in detail. 6.2.2 Analysis of secondary workshops A number of sites have provided direct or indirect evidence for the presence of secondary glass workshops that produced black glass artefacts during the Roman imperial period. Only a few of these secondary workshops have revealed clear evidence of the black glass assemblage they produced (Table 110). 200 Model 1 is, for instance, valid if not characteristic of the late Bronze Age and the medieval period. The study of the Qantir cakes in the Egyptian Nile delta has demonstrated that model 3 was already valid from the late Bronze Age (Rehren 1997). 201 227 Table 110: List of secondary glass workshops processing black glass (between [ ] and in italic are uncertain production sites) production centres period evidence Lyon (FR) [Avenches (CH)] I: Early Roman period (Ib–Id/IIa AD) I: Early Roman period (Ib–Ic AD) Kaiseraugst (CH) II: Mid Roman period (IId–IIIA AD) Sainte Menehould Les Houis (FR) (II)-III: Mid - Late Roman period (III–IV AD) Lavoye (FR) (II)-III: Mid - Late Roman period (III–IV AD) [Liberchies (BE)] [Orešac (HR)] Braga (PT) (II)-III: Mid - Late Roman period (III–IVA AD) III: Mid - Late Roman period (III–IVA AD) III-IV: Late Roman period (III–IV AD) Aquileia (IT) III-IV: Late Roman period (III–IV AD) Trier (DE) IV: Late Roman period (IVB AD) [Horbat Qastra (IL)] IV: Late Roman period (IV–V AD) 6 Period 1 Period 3 Period 2 Period 4 1. Lyon (F); 2. Avenches (CH); 3. Kaiseraugst (CH); 4. Liberchies (B); 5. Orešac (HR); 6. Lavoye (F); 7. Ste Menehould (F); 8. Braga (PT); 9. Aquileia (IT); 10. Trier (D); 11. Horbat Qastra (IL) 4 7 chunks production waste; chunks furnace; production waste; misshapen artefacts; raw glass production waste; chunks; misshapen artefacts production waste; misshapen artefacts; raw glass misshapen artefacts misshapen artefacts production waste; chunks; crucible misshapen artefacts production waste; chunks crucibles; misshapen artefacts misshapen artefacts 10 3 2 1 8 9 5 11 Figure 113: Distribution map of the known or presumed glass workshops that produced black glass artefacts 228 Figure 113 demonstrates that the present knowledge on secondary glass workshops producing black glass artefacts is very incomplete, with only a few workshops per period. Most of these are situated in the north-western provinces, with Braga (PT) and Ore ac (HR) in its periphery. Only Horbat Qastra (IL) lay in the Mediterranean. This does not imply that black glass was essentially produced in north-west Europe, but only demonstrates the actual lack of information on this issue for the Mediterranean region. Especially the (near) absence of a black glass-producing workshop in the Levant and the Balkan is striking and stands in contrast with the high frequency of late Roman jewellery [see Chapter 6.3.2. Jewellery]. A more thorough heuristic study of black glass material from these areas would probably impose a modification of this observation. The interpretation of the results described in Table 111 and the column charts in Figure 114 is twofold: 1) the high number of black glass artefacts from the sites with a (presumed) glass workshop show that black glass consumables can be a frequent find category depending on the intensity of archaeological activities and the excavated areas/levels; 2) the consumption centres with black glass production display 1286 pieces from only 11 sites, which is equal to 28,7% of the 4475 recorded artefacts from 651 sites. The production centres thus show an average of 117 per site, 202 whereas the other 640 sites represent an average of only 5 pieces. Table 111: Overview of various concentrations of black glass commodities from sites with a clear or supposedly secondary glass workshop where black glass has been manufactured architectural sites vessels jewellery counters utensils decoration production TOTAL Aquileia (IT) Augst (CH) Braga (PT) Kaiseraugst (CH) Lavoye (FR) Liberchies (BE) Ore ac (HR) Sainte Menehould (FR) Trier (DE) Horbat Qastra (IL) TOTAL 4 128 0 39 30 22 0 15 0 0 238 129 129 74 46 15 56 38 105 47 82 721 7 94 2 5 0 16 0 0 1 0 125 1 1 0 0 0 0 0 4 0 0 6 1 0 0 0 0 0 0 1 2 0 4 0 10 9 53 13 0 0 100 7 0 192 142 362 85 143 58 94 38 225 57 82 1286 600 500 400 production 300 200 architectural decoration utensils counters 100 0 jewellery vessels 202 When taking into account the number of sites with 10 to 25 piecese.g., Xanten, Nijmegen, Avenches, Jerusalem, or Sens, the main average is rather 2 pieces per site. 229 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% production architectural decoration utensils counters jewellery vessels Figure 114: Column charts demonstrating the concentrations of various black glass commodities from ites with a clear or supposedly secondary glass workshop where black glass has been manufactured  PERIOD I No clear evidence is available that reveals how the production of black glass artefacts was organised or that identifies the location of any particular workshops in Period I. The two chunks of raw glass in black-appearing dark green glass are perhaps the sole evidence from the secondary glass workshop of ‗La Montée de la Butte‘ at Lyon (FR) [cat.nos.1224–1225]; nevertheless, these make clear that the workshop also made use of blackappearing raw glass besides the ‗naturally coloured‘ and decolourised glass. Because no artefacts in black glass have been retrieved, it is impossible to determine what the Lyon workshop produced in black glass. It is, however, possible to form an idea of what the Lyon workshop might have produced in black glass by looking at the assemblage in other glass hues that the workshop manufactured. Furthermore, another contemporaneous Lyon workshop, ‗La Manutention no. 2‘, produced quite similar material (Foy, Nenna 2001, 76–79, nos. 61–71). Thus, even if there is no relationship with what has been produced, we may assume in all probability that the Lyon workshop ‗La Montée de la Butte‘ produced vessels and twisted rods in black glass. The Claudian-Neronian workshop from Avenches (CH) delivered a huge amount of strongly coloured glass production waste (Amrein 2001), but none of the retrieved material is really black-appearing unless we consider some of the pieces in very deep purple glass as ‗black‘a not too exaggerated idea seeing the various blackappearing artefacts (e.g., vessels, rods and counters) that have been produced in such deep purple glass. The st bulk of the 1 century AD black-appearing artefacts is produced in a deep brown to deep green glass and in a purple glass. A glass workshop producing free-blown vessels in deep purple black-appearing glass was almost certainly located in north-east Italy. This hypothesis is based on the amount of particular vessel shapes having idiosyncratic features in other hues as well. Some examples from the Canal Bianco necropolis at Adria (IT) include the cup with a large vertically outward-folded rim [cat.no.2585] or malformed pieces at the limit of being a bad production [cat.no.2586] that would not have been imported from far distances. The examples all come from burial gifts from the Aquileia area (Adria, Aquileia, Este and Montebelluna). Few are those in a cobalt rich deep blue glass, despite, for instance, the amount of eye-catching cameo glass vessels. Presuming that the Avenches and Lyon workshops also produced black glass, some features can be put forward that seem characteristic of Period I workshops in general, seeing the correspondence with the above-mentioned workshop in the Aquileia area:    Both workshops are situated in the centre of a large town; The production shows specialised craftsmanship but also demonstrates very diverse production; The organisation of the deeply coloured glass in Period I workshops is considered comparable to schemes 2–3. Furthermore, we have to stress that both more or less contemporaneous sites are located in the north-western part of the Roman Empire. The very low number of production centres identified and the limited information on 230 available objects does not allow us to establish a marked difference in the production and distribution of the artefact types produced.  PERIOD II nd During the 1970s, two glass workshops were excavated in Kaiseraugst that processed black glass in the 2 to rd early 3 century AD (Rütti 1991, 150–153; Fischer, 2009). Interesting to notice is the presence of four different types of crucibles (Fischer 2009, 91-106), providing an early date for the use of crucibles which are generally rd regarded as introduced during the 3 century AD. At the same time, there were found numerous fragments of quartz-rich clay covered with gullies of deep blue-green glass appearing black (Figure 115). These have to be regarded as part of the interior of the furnace construction and can thus have been parts of a batch basin, the st nd technique assumed to be used during the 1 and 2 century AD. Probably these fragments are the remains from the opening where the glassblower spilled molten glass when gathering a hot gob of glass. Because a glassblower is never clumsy the thickness of the glass suggests that the furnace was active for a long time. Figure 115: Fragments from the Regio 17B glass workshop of Kaiseraugst (CH) (picture by author, by courtesy of Römermuseum, Augst) Seven different glass hues, with a predominance of the green hues, have been identified in the crucibles (Table 112). Table 112: List of attested glass hues from the workshop at Regio 17C (left) and Regio 17B (right) (taken from Fischer 2009, 95, fig.108) hue excavation 1974 excavation 1978 yellow green-white pale green green dark green dark green/black white 2 9 10 2 3 - 3 5 13 64 45 5 1? We distinguished translucent deeply coloured blue-green glass full of round pinprick air bubbles from the blackappearing glass flows on the basin fragments from the workshop ‗Äuserre Reben‘ in Kaiseraugst. This is also the glass we find in many objects, vessels and jewellery that have been excavated all over the town Augusta Raurica (Augst/Kaiseraugst). Various shapes of free-blown as well as some very few examples of mould-blown vessels were retrieved from the workshop in Building 1 of Regio 17B. Previously, it has been stated that glass workshops producing vessels would not generate jewellery and vice versa (Stern 1999). Although the glass workshops at the facing corner buildings of Regio 17B and 17C provided no signs indicating a combined production of vessels and bracelets, we noticed that the very same blue-green glass was utilised for the bracelets found in Augusta Raurica [see Chapter 8]. The organisation of the workshop of Äussere Reben at Kaiseraugst matches up with scheme 4, seeing the various indications that reconstruct, to a certain extent, the production process of black glass. For instance, it is 231 striking that a bottom piece of a black glass batch included an unmelted tessera-shaped cubic inclusion of 7x5, 5x5, 5 mm in a transparent pale blue-green glass [cat.no.3703] (Figure 116 top). Does this piece suggest that the glass workers in the Kaiseraugst workshop added glass tesserae systematically to the batch to work on a continuous basis? Another piece of production waste from the same workshop appears to be a thick curving gutter/trail and contains four transparent, pale bluish-green glass tesserae in a row [cat.no.3695] (Figure 116 bottom). Figure 116: top) batch fragment from Äussere Reben workshop at Kaiseraugst (CH) with unmelted tessera in ‗naturally coloured‘ glass; bottom) production waste from the same workshop showing four unmelted cubic-shaped inclusions in transparent pale blue-green glass (pictures by author, by courtesy of Römermuseum Augusta Raurica) But when looking to the theoretical visualisation of the possible mixtures that can occur when glass artefacts got recycled as cullet, one can easily observe the complexity of recognising glass workshops. This intricacy of glass mixtures is already discussed on the basis of chemical analysis on colourless glass (Freestone et al. 2009, 133– 134, fig.3) and will be treated in a following chapter [see Chapter 8]. The Argonne workshops at Les Houis near Sainte Menehould (FR) and at Lavoye (FR) are mainly known for the production of black glass jewellery and tesserae, but the small set of tiny vessel fragments that has been retrieved from both sites possibly also indicates the production of glass vessels (Cosyns 2009b, 93). These included 36 vessel fragments in black glass from which a good number could be attributed to the carinated beakers or carchesia and the mould-blown, grape-shaped amphoriskoicharacteristic vessel shapes for the late Antonine 203 and Severan period. Nonetheless some presume that the high fragmentation degree needs to be interpreted as evidence for recycling glass, i.e. cullet (Chew 1989). This assumption is based on the hypothesis proposed by Marianne Stern (1999) stating that in Roman times the glassblowing technique was a very specialized procedure that needed skilled craftsmen defining the organization of a glass workshop. Accordingly it is generally assumed 203 The interpretation of this material and its implications for the dating of the site have been discussed earlier [see Chapter 4]. 232 that a glass workshop either manufactured artefacts through glassblowing thus solely producing containers or using any other technique but glassblowing and concentrating on jewellery production. Glassworkers active in vessel production would thus never produce beads, bangles or any other type of jewellery. The Argonne workshops yielded no furnace structures but the presence of one or probably several glass workshops can be assumed from the retrieved crucibles, fragments of blowing pipes and other utensils related to the manufacturing of vessels alongside the production of jewellery. The supposedly workshop(s) in Liberchies (BE) responsible for the manufacturing of black glass artefacts can be deduced from the high number of material retrieved during excavations; and bad productions of carchesia and a so-called conical counter that is to our opinion a misshapen pseudo-nicolo gem [cat.no.345]. Both commodity rd types supports a dating in the first half of the 3 century AD.  PERIOD III The lack of evidence indicating the continuity of the production of black glass vessels during Period III, and elsewhere in the Roman Empire than the north-western provinces, is striking [see Chapter 3]. The available information on late Roman glass workshops manufacturing black glass artefacts demonstrates that the production rd th 204 of jewellery was characteristic of the mid 3 to 5 century AD (Periods III and IV). Since the workshops of Les Houis and Lavoye provided glass material in a wide variety of colours, it is clear that the glass workshop(s) not only diversified in shapes but also in hues. Thus, it is realistic to assume that secondary workshops in general when producing black glass artefacts never limited their production to this one single glass hue. Conversely, it may well be that a range of known secondary workshops, where no evidence of black glass has been found, also produced black glass artefacts. No clear evidence of other contemporaneous glass workshops producing black glass artefacts is available. The th glass workshop attested in Braga (Bracara Augusta)dated roughly 4 century ADfalls within this and the rd th following period since earlier activities in the late 3 century AD or later activities in the early 5 century AD cannot be excluded [see Chapter 4]. Based on what has been retrieved from the Fujacal area, the workshop solely produced jewelleryi.e., bracelets, finger rings, beads, pendants, gems and hairpins (da Cruz 2009, 100rd th 102, fig.3). The presumed (black) glass workshop in Ore ac (HR) seems to have been active in the 3 –4 century AD according to the type of yielded material.  PERIOD IV The glass workshop Trier-Palais Kesselstatt was found during salvage excavations in 1922 (Loeschke 1925), but th no structures were registered or reported. The fill of the zone is dated at the end of the 4 century AD on the basis of the pottery, the glass and the coins, including coins struck under Valentinian (364-375 AD) and th Theodosius I (379-395 AD). The glass workshop is believed to have started somewhere in the first third of the 4 century AD (Goethert 2007, 395). It is clear that the high barrel-shaped crucibles are different from the mainly open types in Period II and III from Kaiseraugst and Les Houis. The seven black beadsannular, globular and barrel-shapedwere all decorated with applied glass trails in opaque white and/or blue. The site produced one black vessel-shaped pendant with applied glass trails in opaque white, one plain arm ring (type D1) in black glass and four black finger rings out of 11 of type C 1 and 2 with a portrait pressed in (Loeschke 1925, 337-340; th Goethert-Polaschek 1984; Cosyns 2009) (Table 113). All material was rather typical of the second half of the 4 century AD onwards. We may conclude that the retrieved objects represent discarded workshop material that got th dumped at the end of the 4 century AD. Table 113: List of glass finds from the late 4th century glass workshop at Trier, Palais Kesselstatt functional type black blue yellowish green amber yellow arm ring finger ring bead pendant crucible 1 4 7 1 3 1 1 - 2 1 - 3 1 - colourless 1 - Horbat Qastra lacks any remains of furnaces or glass working waste, but is supposed to have had a glass workshop based on the statement by Yael Gorin-Rosen that a site showing a large amount of similar-looking 204 Continuity of black glass jewellery production is observable in the East up to the mid 7th century AD, whereas in the West only beads remain in circulation, mainly until the mid 6th century AD but also continuing up to the mid 7th century AD [see Chapter 4]. 233 material is sufficient to determine the presence of a glass workshop within Horbat Qastra or in its vicinity (GorinRosen, Katsnelson 2008, 74). The workshop, producing black glass beads and pendants, must have been active 205 during Period IV. The material is related to that from Trier-Kesselstatt, with annular, globular and barrel-shaped beads with applied trails in opaque white and blue glass as well as vessel-shaped and poppy-head pendants. Some material such as the discoid pressed beads and the small melon beadsannular and globularwere more characteristic of Period III, whereas the so-called ‗crumb‘ beads were typical from Period IV onwards and were popular all through the early Byzantine period. The black glass jewellery from the cemetery of Horbat Qastra thus th th gives a date range from the early 4 up to the 6 century AD. Before discussing the issue of distribution, we first want to point to the sometimes limited availability of direct evidence for locating glass workshops responsible for the manufacturing of black glass artefacts. But when are we allowed to determine the existence of a glass workshop? The prevailing opinion is the presence of direct evidence such as furnaces, tools and production waste, but seldom are all found together (Table 109). We previously listed a number of sites yielding direct and/or indirect evidence of workshops where black glass was made and/or used to manufacture artefacts (Table 110). Unfortunately, the sites are characterised by a lack of concrete archaeological evidence useful in the interpretation, and when available the information is not always easy to read. An ultimate solution in discerning a potential local workshop without direct evidences has been articulated by Yael Gorin-Rosen: … a site may be identified as a production center if large quantities of similar vessel types are evident alongside production debris, although the original location of the workshop is unknown, … … large quantities of vessels of the same limited number of types sharing similar decorations, fabric and workmanship, are sufficient to determine the presence of a glass production center, despite the lack of furnace residue or glassworking waste. Occasionally, it is even possible to identify the fingerprints of a specific craftsman in particular features, such as the fashioning and application of handles or trails onto the vessels. (Gorin-Rosen, Katsnelson 2008, 74) We have tested this statement‘s validity by using the black glass material from Augusta Raurica, Augst/Kaiseraugst (CH). Gorin-Rosen‘s assertion has been re-evaluated again in Chapter 8 by means of the chemical composition of samples taken from one particular type of vessels (Form IIB.4) as well as some bracelets, all of them found at various places in the colonia. 6.3 Distribution and trade Next to the overview of glass workshops processing black glass, it is important to verify the distribution of the different commodity categories in black-appearing glass to obtain a better view on the trade patterns. For reasons of feasibility and to increase the readability of this chapter, we describe the distribution of the black glass artefacts per function, focussing on vessels and arm rings, without excluding the other commodity categories from the reviewing discussion. Aiming to find out where particular forms have been produced, we not only take into account the typological shape, but also consider the adopted glass hue. The importance of the latter is because variations in glass hue for the same form type can point to an idiosyncratic production of a particular workshop. 6.3.1 Vessels Based on the results from previous chapters, it is clear that the production of vessels in black-appearing glass had a discontinuous success throughout the Roman imperial period [see Chapter 3 and 4]. The here-assembled material shows that the black glass vessels from the successive periods were consumer goods within distinct regions. Monochrome strongly coloured glass of Period I was fashionable from the reign of Tiberius to that of st Nero, but it disappeared fully by the end of the 1 century AD in favour of decolourized glass (Cosyns, Fontaine 2009). The catalogued cast and free-blown vessels from Period I show an empire-wide distribution with a concentration in the western part of the Roman Empire. During Period II, a fashion for black-appearing glass vessels re-emerged in the north-western provinces (Cosyns, Hanut 2005). This regional production of free-blown rd and mould-blown black glass vessels ceased around the middle of the 3 century AD. A circulation of black glass th vessels reappeared in the 4 century AD, and the vessels stayed in use all through period IV. This renewed 205 Because the contexts have not been studied yet and the material is still under study, it might well be that a part falls under Period III. 234 production, which consisted entirely of rod-formed, solid-bodied toilet ware, shows once again a regional production/consumption, this time in the south-eastern Mediterraneanthe Levant and Egypt.  PERIOD I We were able to list only 161 pieces of Period I vessels, but we recognize that the list is a work in progress and 206 that many more examples will emerge. From 42 sites were retrieved 83 pieces of cast vessels and 42 pieces of free-blown vessels from 24 sites. Interesting is that either cast vessels or free-blown vessels have been retrieved from a settlement. Only few sites yielded both cast and free-blown vessels: Liberchies, Augst, Kaiseraugst, Avenches, Arles, Olbia, Adria, Pompeii and Heerlen. Few pieces with no detailed provenance are excluded from the presented computation. All were found in Italy: one cast piece, a shallow hemispherical cup of Form IA.3 of the former Bocchi collection and now in the National Archaeological Museum at Adria [cat.no.2825]; and two freeblown jugs, respectively Form IB.8 and 9, in dark purple glass in the Musée du Louvre, Paris [cat.nos.2826; 2828]. st Remaining a rare artefact from the 1 century AD, the black glass vessels have been found in the greater centres all over the Roman Empire, but with a greater frequency in the west. This distribution corresponds to the Roman glassware in general within the 1st century AD in the west and that of monochrome strongly coloured glassware in particular (Isings 1957; Cool, Price 1995; Price, Cottam 1998; Foy, Nenna 2003; Hanut 2003; Price 2006). Various areas of concentration have been noticed in this western part of the Roman Empire (Cosyns, Fontaine 2009): 1) the Gulf of Naples; 2) the north of Italy; 3) Gallia Narbonensis, and in particular the Rhone/Saone valley up to Switzerland and 4) the greater limes-zone in the north-western provinces, e.g. Austria, Germany, Belgium, the Netherlands and Great Britain. It is obvious that the areas where concentrations of black glass material have been observed are influenced by various aspects not related to the true distribution and trade of goods, but to the availability of published material and the accessibility of unpublished material. The concentration of Period I vessels in black-appearing glass was also affected by historical circumstances, such as the eruption of Mt. Vesuvius in 79 AD that put the entire society in the Gulf of Naples under seal. This outnumbered presence in Pompeii (IT) is of course due to its intensive excavations, as is the case for sites like Magdalensberg (AT), Fréjus (FR), Olbia (FR) or Augst/Kaiseraugst (CH). 1.Durres (AL) 2.Magdalensberg (AT) 3.Anthée (BE) 4.Han-sur-Lesse (BE) 5.Liberchies (BE) 6.Riemst (BE) 7.Tongeren (BE) 8.Zara (HR) 9.Augst (CH) Form IB.10 Forms IA.2;4 Form IA.8 Form IA.2 Form IA.14 Form IB.7 Form IA.1 Form IA.8 Form IC.1 Form IA.8 Forms IA.8;13;14 Forms IB.1;19 10.Avenches (CH) 11.Kaiseraugst (CH) 12.Locarno (CH) 13.Windisch (CH) 14.Moers-Asberg (DE) 15.Mainz (DE) 16.Waldorf (DE) 17.Cádiz (ES) 18.Taragona (ES) 19.Arles (FR) 20.Barzan (FR) 21.Cruzy (FR) 22.Forêt d'Argonne (FR) 23.Fréjus (FR) Forms IA.1;10 Forms IB.5 Form IC.1 Form IA.3/4;14 Form IB.19 Form IB.2 Form IA.4 indefinite Form IA Form IA.2 Form IB.7 Form IA.1 Form IA.8 indefinite Form IA Form IB.17 indefinite Form IA Form IA.7 Form IA.7/8 Form IA. 1 1 16 1 1 2 1 1 2 2 2 1 1 1 1 1 2 3 1 6 2 1 1 1 1 1 2 6 2 2 1 1 1 1 1 1 1 1 1 1 1 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 mouldblown freeblown cast blue 7 technology brown 9 purple hue green 1 107-121;A4489 179 316 417;425 407;413 497 565 562 772 3469;3471;3570; 3474;3506 3481;3487;3536;3545; 3548;3554 3597;3611 3614 3617 3751;3758 3767;3776 A4497 3806 1957 1947 2107 3182 3120 1022-23 A4484 1026 A4494 1092 A4495 form black Table 114: List of sites with black glass vessels of Period I cat.no. site 1 1 1 1 1 206 It is noteworthy to pinpoint the absence of Period I vessels in a lot of countries that were part of the Roman Empire. The recorded material from 50 sites is spread over 14 countries of the 32 present-day countries that correspond with the vastness of the Roman Empire. 235 A4476 1138 1154 1227 A4480 A4477-79 1321-22; 1333;1335 1340 1365 A4491 A4498 A4483 24.Gof de Fos (FR) 25.Jublains (FR) 26.Lattes (FR) 27.Marseille (FR) 28.Murviel-lès-Montpellier (FR) 29.Nîmes (FR) 30.Olbia (FR) A4481-82 2584-87 31.Orange (FR) 32.Reims (FR) 33.Richebourg (FR) 34.Saintes (FR) 35. Saint-Paul-Trois-Châteaux (FR) 36.Vaison-la-Romaine (FR) 37.Adria (IT) 2726 2829 2733 2734 2743-44 2758-62 2765-74 38.Aquileia (IT) 39.Bologna (IT) 40.Boscoreale (IT) 41.Calvatone (IT) 42.Este (IT) 43.Montebelluna (IT) 44.Pompeii (IT) 2275 2789 2805 2813 2830-31 2894;2896-97 2982-83 3105 3152 3173 A4499 A4500 A4501-7 4051 4145-47 794 810;812 856 874 3896 TOTAL 45.Porto Potenza Picena (IT) 46.Rome (IT) 47.S. Elena di Melma (IT) 48.Santelpido (IT) 49.Verona (IT) 50.Heerlen (NL) 51.Nijmegen (NL) 52.Torre de Ares (PT) 53.Polhov Gradec (SL) 54.Ribnica (SL) 55.Caersws (UK) 56.Carlisle (UK) 57.Colchester (UK) 58.Fishbourne (UK) 59.London (UK) 60.Abydos (EG) 61.Bahnasa (EG) 62.El-Ashmunein (EG) 63.Ghurob (EG) 64.Ephesus (TR) Form IA.4 indefinite Form IA Form IA.8 Form IA.2 Form IB.10 Forms IB.10;11 Form IA.2 Form IB.1 Form IA.7 Form IA.7/8 Form IA.3 Form IA.2 Form IB.8 Form IB.5;16 Forms IA.3;8 Forms IB.12;13 Form IB.10-11 Form IB.5 Form IA.7 Form IA.3 Forms IB.3;4 Forms IB.13;14;15 Forms IA.1;4;5;6 Forms IB.1;5;8;9 Form IA.1 Form IB.11 Form IA.9 Form IB.11 Form IB.10 Form IA.4 Form IB.1 Form IA.8 Form IA.3 Form IB.1 Form IA.12 Form IA.7 Form IA.7 Form IA.7 Form IA.7 Form IA.11 Form IB.16 Form IA.10 Form IB.5 Form IB.4 Form IB.3 1 1 1 1 1 1 1 1 1 2 1 1 2 1 3 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 2 2 1 1 1 1 1 1 1 1 1 1 4 1 4 1 2 4 1 6 3 1 1 1 1 2 4 1 1 1 1 2 2 2 1 2 1 1 1 2 1 1 1 1 1 7 1 3 1 1 1 2 1 2 5 1 1 1 1 1 1 2 1 45 29 1 36 14 8 87 1 1 1 45 2 The material listed in Table 114 illustrates an empire-wide distribution of the black glass vessels in Period I, although most recorded material is distributed in the west of the Roman Empire with only few sites in Egypt and Turkey for the east. Only one cast vessel came from the eastern Mediterranean area: the large bowl with horizontal rim from Bahnasa (EG). Another cast cup with shallow cylindrical body (Form IA.4) now in the Haaretz museum, Tel Aviv, reportedly came from Israel (unpublished) [cat.no.A4496]. Besides the free-blown unguentaria with globular body (Form IB.4) from Ephesus (TR) [cat.no.3896], and Ghurob (EG) [cat.no.874], others remained unprovenanced although reportedly coming from the eastern Mediterranean such as the one in the Kestner Museum in Hannover [cat.no.A4509]. Other free-blown vessels of eastern origin include the globular juglet with a vertical handle in opaque blue glass (Form IB.2) supposedly from Syria (Slitine 2005, 71) [cat.no.A4508] and the large cantharus (Form IB.5) from El-Ashmunein (EG) that is now in the Louvre [cat.no.856]. This rare occurrence in the eastern part of the Roman Empire is most probably a reflection of the narrowness of the material at hand. However, a more intensive investigation of the black glass vessels from Period I in the eastern provinces could not be carried out during this research project. On the other hand, the rareness of these vessels is also due to the lack of extensive excavations of early Roman sites in the eastern Mediterranean. In this area, the early Roman levels mostly remain sealed by structures from later periods. Valuable and major architectural finds generally stop excavations at the early Byzantine or late Roman levels to preserve the uncovered structures. Nevertheless, it may be advanced that the much more limited occurrence of black glass vessels in the eastern Mediterranean corroborates the prevailing view proposing a western production of non-blown vessels during the early imperial period in Italy, most likely in Rome as well as other urban centres (Grose 1991; Price 2004, 15). Because the majority of the black-appearing vessels made of purple glass were found in Italy, it is assumed that in particular the free-blown vessels were produced in an Italian workshop and more precisely in the north of Italy, with Aquileia as the most frequently cited location (Beretta et al. 2006, 207). The concentration of black glass 236 vessels for the Veneto region is, however, influenced by the very intensive and long-lasting excavations in that region. Despite the fact that we could not accumulate as much material from the eastern part as from the western part of the Roman Empire, the available examples provide enough evidence to demonstrate that the black glass vessels are omnipresent and appear to be part of the then-popular deeply coloured monochrome glass vessels from late Tiberian period up to the Neronian period, even if the low frequency is ranking this commodity as a very occasional article of trade. Nevertheless, it is necessary to verify more carefully the differences and similarities between the eastern and western empire in the distribution and consumption of strongly coloured vessels. PERIOD II 173-74 178 185-90 192 193 195 199 209 210 222 224 255 258 259 291 320 321 324-28 Amay (BE) Anthée (BE) Arlon (BE) Auvelais (BE) Bassenge (BE) Baudecet (BE) Bois-et-Borsu (BE) Bossut-Gottechain (BE) Braffe (BE) Braives (BE) Brecht (BE) Eben-Emael (BE) Elewijt (BE) Ellignies-Ste-Anne (BE) Grobbendonk (BE) Izier (BE) Jambes (BE) Kontich (BE) 329-30 334 335 408-12;414-16; 418-24;426-28 429-30 438-47 449 450 459 492-94 Kruishoutem (BE) Lauw (BE) Lavacherie (BE) Liberchies (BE) 496 508 510-14 515 519 539;541-48 540 563-64 582-89 594-99 610 614 621-23 998 1028 1049-52 1073-75 Relegem (BE) Roly (BE) Rumst (BE) Schaarbeek (BE) Straten Tienen (BE) 1183-88;1189-94;1196;11981203;1205-1207;1209-12 1214 1295-96 1367-69 1613-19;1623 1620;1622;1624 A4488 Lavoye (FR) Liège (BE) Matagne-la-Petite (BE) Mettet (BE) Mortsel (BE) Nismes (BE) Oudenburg (BE) Tongeren (BE) Treignes (BE) Velzeke (BE) Vervoz-Clavier (BE) Virton-St. Mard (BE) Wancennes (BE) Abbeville(FR) Boulogne-sur-Mer (FR) Cutry (FR) Faulquemont (FR) Lillebonne (FR) Morains (FR) Roussy-le-Village (FR) Sainte-Menehould (FR) Altbachtal-Trier (DE) Form IIB.1 Form IIB.4 Forms IIB.1;4 Form IIB.1 Form IIB.15 Form IIB.1 Forms IIB.4 Form IIB.1 Form IIB.1 Form IIB.1 Form IIB.1 Forms IIB.4 Form IIB.1 Form IIB.1 Form IIB.1 Form IIB.1 Form IIB.1 Form IIB.13 +indefinite forms IIB Form IIB.1 Form IIB.4 Form IIB.1 Form IIB.1 Forms IIB.1;4 Forms IIB.1;4;5 Form IIB.4 Form IIB.1 Form IIB.1 Form IIB.1 +indefinite forms IIB Form IIB.1 Form IIB.5 Form IIB.1 Form IIB.2 Form IIB.11 Form IIB.1 Form IIC.2 Form IIB.1 Form IIB.1 Form IIB.1 indefinite form IIB Form IIB.1 Form IIB.1 Form IIB.1 Form IIB.1 Forms IIB.1;3 Forms IIB.5;6 +indefinite form IIB Forms IIB.1;20 +indefinite forms IIB Form IIB.19 Form IIB.1 Forms IIB.1;4;5 Form IIB.1 Form IIC.1 Form IIB.4 2 1 1 2 1 1 15 2 1 1 18 1 10 2 10 1 1 1 3 5 1 1 1 1 1 1 1 3 1 1 1 1 3 1 1 1 1 5 1 1 9 5 1 1 9 1 1 1 1 1 1 7 5 1 8 6 1 1 2 1 1 4 3 1 2 1 1 4 1 2 20 1 2 3 8 3 1 1 1 3 25 1 2 3 8 3 1 237 rod-formed mould-blown free-blown 1 1 1 5 2 1 6 1 1 1 1 1 1 2 1 1 1 1 1 1 1 5 1 1 1 1 1 blue brown technology purple hue blue-green form black Table 115: List of sites with black glass vessels of Period II cat.no. site green  1840;1842-43;50;1857 1845;1848;1858 Cologne (DE) 1869 1876 1929-30 1946 1962 1999 2012-14 A4618 2108 2178 2788 2879-82 2883 2895;2898 2902-7 2975-79;2981 2984 2986-87 3154 3463;3465;3472-73;3475; 3477;3488;3490-91;3495; 3498;3504;3512;3516-17; 3519-20;3523;3525-30; 3535;3537;3539-40;3542; 3544;3546;3550;3553;3556; 3559-60;3563;3565-66; 3576;3581-82 3462;3493-94;3521;3532; 3551 3594;3598;3600;36023;3607;3609;361213;3615;3618;3621-23;3625-28 3634-36 3743;3746;3748;3750;375357;3760;3762-63;3772-74; 3778-79 3777 4026-27 868-71 788 2851-52 2866 Total Elsdorf-Esch (DE) Eshweiler-Hastenrath (DE) Krefeld-Gellep (DE) Mainz (DE) Morbach-Wederath (DE) Rheinbach-Flerzheim (DE) Siesbach (DE) Trier (DE) Weilerwist-Hausweiler (DE) Zûlpich (DE) Rome (IT) Esch (NL) Heel (NL) Heerlen (NL) Hoogeloon (NL) Nijmegen (NL) Stein (NL) Ptuj (SL) Augst (CH) Forms IIB.1;4;11;19;21 Form IIC.1 Form IIB.14 Form IIB.13 Form IIB.8 Form IIB.5 Form IIC.1 Form IIB.5 Form IIB.4 Form IIB.2 Form IIB.8 Form IIB.21 Form IIC.1 Forms IIB.1;6;7;11 Form IIB.1 Forms IIB.1;4 Forms IIB.1;4 Forms IIB.1;4;11;13 Form IIC.1 Forms IIB.4; Forms IIB.11 Forms IIB.1;3;4;5;6; 12;14;17;20 3 4 3 3 1 1 1 1 1 1 1 1 1 1 3 1 1 1 3 1 1 1 1 1 2 1 1 3 1 2 6 4 4 1 2 6 6 2 1 28 11 2 1 42 1 4 1 1 1 Avenches (CH) Forms IIB.1;4;5;16; 17 1 8 Courroux (CH) Kaiseraugst (CH) indefinite Form(s) IIB Forms IIB.1;4;5;17; 20;21 3 1 8 Form IIC.2 Forms IIB.1;16 Form IID.1 Form IIC.1 Forms IIB.11;18 Form IIB.10 8 1 Form IIC.2 Colchester (UK) Elkab (EG) Cyprus Lebanon Luxemburg 1 3 6 9 18 3 17 8 1 2 1 2 4 4 1 1 52 200 1 25 1 1 7 19 0 2 1 281 18 The black glass vessels of Period II show a very distinct distribution pattern with a clear regional character that contrasts with the supraregional occurrence during Period I. All vessel shapes corroborate the model of productiondistributionconsumption of black glass vessels within the provinces Gallia Belgica, Germania Inferior and Germania Superior, with a few exceptions found in neighbouring provinces such as the one-handled jug (form IIB.11) from Grave 4 of the cemetery at Ptuj (SL) (Tu ek 1993, 411-412, pl. 6.1; Lazar 2003, 130-131, fig.39) falling in the Roman province Pannonia Superior and the dolphin-shaped vessel from Lillebonne (FR) in Normandy [cat.no.1214] within a transitional area between the Roman provinces Gallia Lugdunensis and Gallia Belgica, just north of the River Seine. A good number of the pieces show the material to be rare if not unique, with the exception of two well-spread vessel shapes: the carinated beaker with out-turned fire-rounded rims, Form IIB.1, and the bulbous jar, Forms IIB.4-5. For the very thin-walled vessel shape Form IB.1 two things can be said:  207 This vessel type is the most frequent type of all black glass vessel shapes exclusively distributed in Gallia Belgica and Germania Inferior. The 139 records catalogued up to now originate from 46 sites dispersed over Belgium (86 ex.), France (30 ex.), the Netherlands (10 ex.), Germany (3 ex.), Switzerland (8 ex.) and England (1 ex.), ranking the carinated beaker as a rather common commodity. The Colchester fragment of the carinated beaker forms no inconsistency because the material culture in southeast Britannia matches the distinctive eastwest axial distribution with Germania inferior and (the northern part of) Gallia Belgica, a region now known as the north of France, Belgium, the south of the 207 Netherlands and the northern Rhine region in Germany (Table 115). Small quantities of black glass Very useful and conducive to a better understanding of the distribution pattern of black glass vessels are observations on glass on the whole as well as on other materials, such as pottery. 238 4  material produced in Gallia Belgica and Germania Inferior reached the Romano-British market and, more specifically, south-east Britain. This delimited concentration area in south-east Britain coincides with the rd th distribution of the mould-blown cylindrical glass bottles of the late 3 and 4 century AD (Isings 1957, Form 128; Price 1978, 76-77; 1985, 213) plus a number of pottery imports in late Roman Britain (Millett 1990, 157-180). On the other hand, the south-western part of Britain, where black glass artefacts are nearly absent, seems a significant consumers‘ market for imports from Gallia Aquitania via Normandy and Brittany (Millett 1990), where black glass pieces are similarly more or less absent. One might think of the impact of customs barriers, which are suggested to have existed between the different provinces, and their impact on trade and distribution. These can help to set a value on its position in the nd th interprovincial trade patterns with the north-western provinces on the continent in the late 2 to 4 century AD. The research data for Britain of the material described above illustrate good similarities with the distribution of black glass material. Odder is the material‘s near absence in the Rhine-region while occurrences have been noted in Switzerland. The black glass carchesium can therefore be considered idiosyncratic for Gallia Belgica and Germania Inferior, even though some pieces have been found in Germania Superior as well; This vessel type is in some way ubiquitous outside Gallia Belgica and Germania Inferior but in decolourized and ‗naturally coloured‘ glass, whereas in the area of the two north-western provinces, Gallia Belgica and Germania Inferior, the shape is nearly exclusively produced in black glass. Only a few examples are known in other hues, such as the carchesium in colourless glass from Couvin (BE) (Musée Archéologique, Namur) or the piece of unknown provenance in an almost colourless greenish glass (Römisch-Germanischen Museum, Cologne, Inv. no. 63, 60 – La Baume 1973, D8, pl.16:4). Striking is that the same shape is well represented in neighbouring provinces such as Gallia Lugdunensis, but the carinated beakers from those provinces are of blue-green and green glass, while those in black glass are completely lacking, e.g. Mazières-en-Mauges with 25 ex. (Simon 2000, 149; 152-153, fig.87:2-12). The Mediterranean carchesia are not to be compared with those from the northwestern provinces, as they show technological differences. Produced in a pale coloured or decolourized glass, these carinated beakers have a cut rim with ground edge, and the foot is folded and not applied. The Cypriot carchesium at the National Museum of Scotland (Lightfoot 2007, 77, no. 162), and numerous other similar beakers in large museum collections, show that the Romans were familiar with this specific shape. The total absence of black glass bulbous cups (Forms IIB.4-5) in Britannia is not surprising, as its distribution demonstrates a northsouth axis along the Rhine region, excluding Gallia Belgica from the distribution of this form type, with the exception of some rare occurrences on settlements in the most eastern part of this province within the civitas Trevirorum, such as Trier (DE); Arlon (BE); Faulquemont (FR); Roussy-le-Village (FR). The bulbous cups are characteristic of Germania Superior and Germania Inferior, with a distribution as far west as Anthée (BE); Matagne-la-Petite (BE); Kontich (BE); Hoogeloon (NL), and with Nijmegen (NL) and Avenches (CH) as the most northern and most southern extremities of its distribution range. Seeing that 46 of the 73 entries or 2/3 come from either Augst/Kaiseraugst (CH), with 36 ex., and Avenches (CH), with 10 ex., we may assume with an acceptable certainty that both vessel types were produced in the area, and most likely in Augst/Kaiserugst and probably in the workshop ‗Äussere Reben‘ [see Chapters 8 and 9]. A third vessel typethe shallow plate or bowl with fire-rounded rim (Form IIB.17)is thus far only known from the nd mid-2 century AD in Avenches, Colchester and the vicus of Castleford (Cool, Price 1998, 156, fig.53, no. 62) st nd and the late 1 to early 2 century AD in Deansgate (Cool, Price 1995, 100). Hitherto, the known examples are unfortunately too restricted for us to be able to outline any distribution pattern, but might be of a wider range than the two other vessel shapes. It appears that black glass vessels were consumer goods characteristic of the provinces of Gallia Belgica, Germania Inferior, Germania Superior, but black glass artefacts always remained a minor side-product on the Roman (glass) market in general and a relatively occasional and singular glass product in particular.  PERIOD III The recorded material of Period III is insufficient to play a role in the discussion on distribution patterns. Hitherto, the material is represented by a few examples of a limited number of vessel shapes [see Chapter 3], all of which was recorded in the north-western provinces: Tongeren (BE) [cat.no.563] (Form IIIB.1), Krefeld-Gellep (DE) [cat.no.1929] (Form IIIB.2) and Amiens (FR) [cat.no.1006] (Form IIIB.3). Considering the amount of globular unguentaria with flattened body and aryballoi with annular body in deeply coloured glass, we assume that the black appearance of these few pieces was not deliberately obtained. These shapes are mainly known in ‗naturally‘ coloured and decolourized glass and show a similar distribution within the north-western provinces, with a 239 concentration in the northern Rhine region (Welker 1974, 138, no. 30; Goethert-Polaschek 1977, form 139; Friedhoff 1991, 154, pl. 84:7; Foy, Nenna 2001, no. 205).  PERIOD IV The nearly total absence of provenance information for the recorded 48 pieces of rod-formed vessels makes a distribution map of the Period IV vessels irrelevant. However, the few pieces with provenance information corroborate the assumption that the rod-formed, solid-bodied squat jars and elongated slender unguentaria are idiosyncratic to the Levant and Egypt (Table 116) (Stern 2001, 186-189)] and corresponds to the distribution of the blown and rod-formed, solid-bodied unguentaria in deeply green and blue-green glass in the preceding periods (Harden 1936). Table 116: List of sites with black glass vessels of Period IV cat.no. Provenance Type 853-55 868-71 930 3807 3808 3821 Dush Kysis (EG) Elkab (EG) Shurafa (EG) Aleppo (SY) Hama (SY) Homs (SY) Form IVD.2 Form IID.1 Form IVD.2 variant 4 Form IVD.2 Indefinite form Form IVD.2 Quantity 3 4 1 1 1 1 The black glass vessels in Period IV have only been produced by using the rod-formed technique. Two remarks can be put forward: 1) the great uniformity makes it quite reasonable to assume that the rod-formed, tall, slender balsamaria and wide, shallow jars are to be regarded as a product of one single workshop or a very few number of workshops within a limited region where the same fashion as well as technological know-how and traditions were available during a well-defined period of time; 2) these rod-formed vessels were most likely produced in the same glass workshop(s) where the assortment of beads, pendants and other trinkets were made of black glass and decorated with applied trails in opaque yellow, red or turquoise glass. The statement by Maud Spaer contradicts Marianne Stern‘s assumption that glass vessels and glass jewellery were produced by different workshops (Stern 1999). These vessels, decorated like the beads and pendants of the period, were not manufactured by the common contemporary vessel-manufacturing techniques and are likely to have been produced in beadmaking workshops. … These unique vessels were in all likelihood made in the workshops specializing in beads and other varieties of glass ornaments, in ways closely related to beadforming technologies. (Maud Spaer 2001, 31-32, fig.5) Most likely the rod-formed vessels have to be seen as an exception to the rule. But there is as far as we know no evidence concerning the causes why a bead-making workshop has produced vessels, but probably the production of rod-formed vessels by bead-making workshops was caused by coincidental circumstances. Black glass was commonly used by glass workshops producing jewellery, given the high frequency of bracelets, beads and pendants in black glass during Period IV. Conversely, black glass was fully absent at glass workshops where glass vessels were manufactured. We can think of two possible explanations: 1) the rod-formed vessels were produced by glassworkers unacquainted with the technique of blowing glass, or 2) regulations interdicted the bead makers to blow glass vessels. Both situations refer indirectly and directly to possible socio-economic implications, but we leave open the debate about what regulations may have governed glass workshops in the late Roman period. A striking observation is that a limited assortment of toilet ware was solely produced by the rod-formed technique [see Chapter 3]. Such non-blown vessels very well fit to the assemblage of a shop specialized in the production of jewellery for female customers. Jewellery shops have been rarely excavated so rd far. The 3 century AD shop of Medizinische Klinik in Bonn (DE) close to the gate of the legionary fort castra Bonnensis is an important exception in view of the fabrication of pseudo-nicolo gems [see Chapter 5 and in this chapter, section 6.3.2.5 on gems]. The recently excavated and so far unpublished shop at Kalat Seman (SY) is th the only Levantine jewellery shop with black glass jewellery, but it seems to date to the mid-7 century AD and 208 thus it would be categorized as post-Period IV. A suitable deduction concerning the distribution network of the workshop(s) manufacturing rod-formed toilet ware is that it (they) supplied a range of shops in towns within a vast area. But is it correct to suppose that a jeweller‘s shop was provided by one single glass workshop? It is equally possible that the shopkeeper was a glassworker 208 Personal communication by Danièle Foy (2009). 240 himself or a member of a family of glassworkers and that the distribution only gives an idea of the travelling area of attended fairs. This assumption can confirm whether only one single or some workshops produced such rodformed vessels but could not be verified in more detail within the scope of this study. In brief, we may conclude that the black glass vessels in Period I demonstrate an empire-wide distribution, whereas those from Period II and IV show a much more limited pattern confined to a region of a few provinces. The Period II vessels show a wide variety of shapes and techniques producing luxurious tableware, whereas the Period IV vessels are restricted to very crudely made, rod-formed toilet ware in only two particular shapes, even though a wide range of variants is noticeable. 6.3.2 Jewellery With 2975 pieces out of a total of 4475, the black glass jewellery forms the majority (66,5%) of the recorded material. This commodity category has been registered in every country corresponding to the former territory of the Roman Empire. On the other hand, the total absence of black glass arm rings, finger rings and hair pins outside the Roman Empire makes these artefact types purely Roman commodities. The other jewellery types, such as beads, pendants and gems, should be considered purely Roman as well, although they occasionally occur beyond the Roman Empire. It is obvious that this high frequency makes black glass jewellery an ideal commodity category to provide additional information for a better understanding of an empire-wide or interregional long-distance trade as to the impact factor of socio-cultural and economical dynamics. The evaluation of the provenance of the black glass jewellery verifies whether 1) these objects can help trace back idiosyncratic assemblages from one or a few centres from a specific region (socio-cultural impetus) or whether 2) these objects demonstrate through the uniformity of standardized material a sort of ‗globalized‘ design in most glass workshops all over the Roman Empire, each producing the same material for their own regional or local market (economical impetus). Recognizing a sort of regional differences in taste preferences vis-à-vis the sameness in the Roman material 209 culture, as is also the case for other materials , would help in detecting regional and chronological differences in production and in consumer behaviour. 6.3.2.1 Arm rings 210 At present, 1453 black glass bangles have been registered within the entire Roman Empire. However, the current database remains an incomplete list, seeing that unpublished material in archaeological depots and museums was catalogued only in some regions, while for most regions we were limited to what has been published and made accessible to the author. Documentation of material in some countries that are situated within the territory of the former Roman Empire is missing, due to the absence of publications and the difficulty in 211 making contacts with the responsible persons. The provenance of the inventoried material, described in Table 117, shows that the glass bracelets in general and those in black glass particularly were a common commodity all over the Roman Empire. When reducing the present-day countries to large areas comprising interrelated Roman provinces, the compiled data show a marked discrepancy between Area I (the north-western provinces) and the rest of the former Roman Empire. As we discussed previously, the main reason is that a number of production centres within Area I provided the bulk of the artefacts [see section 6.2. on production]. But to evaluate the distribution of (black) glass bracelets within the entire Roman Empire, it is necessary to assess this commodity type on different levels of the consumption marketthat is, on the regional level (per province) and on the interregional level (per several neighbouring provinces) to facilitate the discerning of patterns per type. 209 For instance, the so-called Coptic textiles show that in the late Roman period (i.e., from the 3rd century AD) there existed all over the Empire a cultural mainstream. 210 With the exception of one piece that is said to be from Horodnica in Ukraine (Haevernick 1960, Taf.14a) and three pieces from the Czech Republic (oppidum Hradi tě/Stradonice and Libceves), no provenance information is available for 34 ex. relative to the country and for 202 ex. at the level of the place of provenance. 211 No material has been recorded from Morocco, Libya, Rumania, Macedonia, Albania, Montenegro or Bosnia-Hercegovina. 241 Table 117: Overview of the number of black glass bangles per country and by Roman area Area I. North-western Europe II. III. Italian Peninsula + islands Central Europe IV. Eastern Europe V. Aegean VI. Levant VII. VIII. IX. Egypt Northern Africa Iberian Peninsula Country Quantity Belgium France Germany Grand-Duchy of Luxemburg Netherlands Switzerland United Kingdom Italy Austria Croatia Slovenia Bulgaria Hungary Serbia Cyprus Greece Turkey Israel Jordan Lebanon Syria Egypt Algeria Tunisia Portugal Spain 246 352 75 10 19 125 22 42 4 77 36 13 35 2 7 1 3 113 12 2 54 84 3 3 54 21 A more selective assortment of the black glass bangles is nonetheless necessary to demonstrate a particular production of black glass bangle types idiosyncratic to a specific province or a vaster area comprising several provinces such as Area Iencompassing specific interregional idiosyncrasies not attested outside the provinces of this areabut also comprising specific regional peculiarities only attested within one of provinces of the area such as those discernable between Britannia. The glass bangles from the British Isles are a clear outsider compared to the rest of the Roman Empire showing a very idiosyncratic regional production in various ways. It has been assumed previously that the Romano-British glass bangles are locally manufactured but using Roman raw glass they are considered Roman commodities rather than being of indigenous British origin (Jones 1996, 122-123). The Romano-British glass bangles however demonstrates peculiar shapes and decoration patterns using glass colours and designs which are atypical to the continental productions [see Chapter 3]. Additionally the Romano-British material shows to be representative to the Early Roman times and consequently not contemporaneous with the purely (late) Roman continental bangle production. The Romano-British glass bangles are therefore to be considered a local production that most likely evolved from the late Iron Age tradition throughout the early imperial period [see Chapter 4]. The idiosyncratic Romano-British production of black glass is also attested through the chemical composition of the applied glass identifying a completely different raw glass when compared to the standard available Roman glass in general and to the continental black glass in particular [see Chapter 8]. The results thus provide good evidence for a local 212 production of black glass to manufacture the idiosyncratic Romano-British bangles. Howard E. Kilbride-Jones, who established a first typo-chronology on the Romano-British glass bangles (KilbrideJones 1938), believed in a so-called local Scottish production seeing the high concentration in south Scotland between Hadrian‘s wall and the Antonine‘s wall with Traprain Law (UK) as a possible production centre based on the high concentration of this material (Kilbride-Jones 1938, 394). Later work revised this theory (Stevenson 1956; 1976) by emphasizing that this type of artefact occurs on a rather considerable number of Roman military sites in northern Britain, a finding that only got confirmed by Jennifer Price (Price 1985, 214; 1988) who‘s most recent inquiries on the matter has shown that these bangles are common down to Wales and South England (Price 2004, 233). These characteristic colourful glass bangles may therefore be considered as true Romano-British pieces. 212 This was made possible by sampling six pieces from London preserved at the LAARC [cat.nos.4118-4119; 4122; 4126-4127; 4130] [see Chapter 8]. 242 Because the Romano-British glass bangles show a distribution that is confined to the British Isles with a nearly 213 complete nonappearance on the continent they are an indicator for a regional production within the periphery of the Empire (Tables 116-119). In contrast continental black glass bangles are scanty in Britain and essentially limited to the south-east of England. The large late Roman cemetery at Winchester-Lankhills (UK) is illustrative, with its total absence of bracelets in black glass while there were uncovered a vast amount of bracelets within a wide variety of materialsi.e. bronze, iron, jet, shale, ivory, bone (Clarke 1979, 301-314). The poor occurrence in Britain of black glass jewellery is entirely owing to the popularity of the similarly appearing local jet and shale (Allason-Jones 1996). By querying the accumulated database we can generate an overview of the distribution of the four main types (A; B; C; D) in general and look at the distribution of the various subtypes as well to that of the characteristic elliptical open bangles. First we verify the distribution of the main types per country and according to the outlined areas. But for practical reasons we reduced the geographical subdivision into four major areas when looking at the subtypes roughly coinciding the late Roman Western and Eastern Empire subdivided into a northern and southern sectioni.e. 1) the north-west European provinces including Belgium; Netherlands; Luxemburg; France (except Corsica); Germany (Rhine-region); Switzerland; UK; 2) the north-east European provinces including northern Italy; Austria; Hungary; Slovenia; Croatia; Serbia; Bulgaria; Rumania; Germany (Danube region); 3) the west Mediterranean provinces including Spain; Portugal; Italy (excluding the north) ; Morroco; Algeria; Tunesia; Corsica; 4) the east Mediterranean provinces including Egypt; Libya; Israel; Jordan; Lebanon; Cyprus; Syria; Turkey; Greece. Table 118: Overview of the Type A black glass bangles per country and per area Area Country Quantity Open Area I Belgium France Germany Grand-Duchy of Luxemburg Netherlands Switzerland United Kingdom 117 124 30 4 11 83 3 372 2 1 1 1 5 3 3 1 1 1 3 24 2 26 7 2 9 1 2 3 421 22 18 8 1 9 2 60 0 0 1 1 0 0 0 61 subtotal Area II Area III subtotal Area IV subtotal Area V subtotal Area VI subtotal Area VII Area VIII subtotal Area IX subtotal TOTAL Italy Austria Croatia Slovenia Bulgaria Hungary Serbia Cyprus Greece Turkey Israel Jordan Lebanon Syria Egypt Algeria Tunisia Portugal Spain 213 Very few pieces have been reported on the Continent, such as the fragment from Valkenburg Z.H. (NL) (van Lith 1977, 131, figs. 1-2) and the one from Dorestad (Isings 2009). 243 From Table 118, it is clear that the 424 pieces of Type A on hand are mainly from Area I (north-western provinces) with 88,4% and only in a minor range attested in Area V (Levant). All other areas show an insignificant occurrence. Additionally, two pieces have no provenance, and two pieces came from outside the Roman Empire. The incidence of a Roman black glass bangle at the oppidum Hradi tě, Stradonice (CZ) is probably due to the military presence at the hill fort that most likely functioned as a (temporary) outpost in late Roman times to create a buffer zone immediately beyond the limes frontier. The second piece is reported to have come from Horodnica (UKR) (Haevernick 1960, pl. 14a). When making the query more specific to verify the distribution of the open bangles, only 63 of the 424 Type A bracelets remain, and all were found in Area I. Three open pieces are from outside Area I. One is now in the Ringling Museum of Art, Sarasota (USA) and comes from the di Cesnola collectionthe reason why presumed to be from Cyprus [cat.no.780], and the two others are the above-mentioned pieces from outside the Roman Empire: the oppidum Hradi tě, Stradonice (CZ) [cat.no.791] (Karasová, Schönfelder 2004, 222, fig.2:86 top left) and Horodnica (UKR) [cat.no.3903]. Table 119 demonstrates that the plain and twisted arm rings (Types A1-3) are the most frequent subtypes within the entire Roman Empire, encompassing 82,3% of all Type A bangles with known provenance. Significant is that most were recorded in the north-western European provinces (71,9%). All three other variants Types A4-6, with alternating plain and twisted decoration, fine-wide twisted decoration, and applied twisted glass trail(s) in contrasting opaque colour(s), respectively, are less common and essentially occurred in the north-western European provinces, with the exception of few examples in the Mediterranean: 1 ex. from Tipasa (AL) [cat.no.4]; 1 ex. from Corinth (GR) [cat.no.2179]; 2 ex. from Jerusalem (IL) [cat.nos.2436; 2448] and 1 ex. from Tell Keisan (IL) [cat.no.2495]. The pie-charts in Figure 117 emphasize the predominant presence of all Type A variants in the north-western provinces. Some are even exclusively attested in this area, whereas the plain (A1) and twisted with applied trail (A6) are also retrieved from eastern Mediterranean sites. Table 119: Detailed list of Type A bangle subtypes per vast geographical section North-western European provinces Belgium; Netherlands; Luxemburg; France; Germany (Rhine region); Switzerland; UK A1 A2 A3 A2/3 A4 A5 A6 A 112 79 84 25 30 18 21 369 26,86 18,94 20,14 6,00 7,19 4,32 5,04 88,49 North-eastern European provinces Austria; Hungary; Slovenia; Croatia; Serbia; Bulgaria; Romania; Germany (Danube region) A1 A2 A3 A2/3 A4 A5 A6 A 3 3 0 0 0 0 0 6 0,72 0,72 0,00 0,00 0,00 0,00 0,00 1,44 Western Mediterranean provinces Spain; Portugal; Italy; Morocco; Algeria; Tunisia A1 A2 A3 A2/3 A4 A5 A6 A 2 1 0 2 0 0 1 6 0,48 0,24 0,00 0,48 0,00 0,00 0,24 1,44 A6 A Eastern Mediterranean provinces Egypt; Libya; Israel; Jordan; Lebanon; Cyprus; Syria; Turkey; Greece A1 A2 A3 A2/3 A4 A5 14 2 1 15 0 0 4 36 3,36 0,48 0,24 3,60 0,00 0,00 0,96 8,63 244 A1 4% 2% 2% 1% 2% A2 northwest European provinces northwest European provinces 11% northeast European provinces northeast European provinces west Mediterranean provinces west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 85% 93% A3 0% A4 1% 0% 0% 0% 0% northwest European provinces northwest European provinces northeast European provinces northeast European provinces west Mediterranean provinces west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 99% 100% A6 A5 0% 0% 0% northwest European provinces northeast European provinces 100% 4% northwest European provinces 15% northeast European provinces 0% west Mediterranean provinces west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 81% Figure 117: Pie-charts visualizing the distribution range of black glass bracelets of Type A Both tables are evidence of an intensive production of bangles with an O-shaped section within the north-western provinces and more particularly in the three provinces of Gallia Belgica, Germania Inferior and Germania Superior in accordance with the region of the black glass vessel production and consumption. A swift plot on a departmental map of France illustrates clearly that most material circulated in Gallia Belgica and Germania Superior, while a much lesser intensity is visible for Gallia Lugdunensis and a near absence for Gallia Aquitania and Gallia Narbonnensis (Figure 118). When looking to the area of Gallia Belgica and the two Germaniae, it can be noticed that some areas show concentrations of the material while others show a rather trivial occurrence. It is worth checking whether these areas are corresponding to particular civitates. 245 Figure 118: Distribution map of departments in France with black glass bracelets of type A Table 120: Overview of the Type B black glass bangles per country and per area Area Country Quantity Open Area I Belgium France Germany Grand Duchy of Luxemburg Netherlands Switzerland United Kingdom 19 30 2 3 2 4 60 7 6 13 1 1 0 25 9 34 8 3 11 2 3 5 124 3 10 1 14 0 0 0 0 0 0 14 subtotal Area II Area III subtotal Area IV subtotal Area V subtotal Area VI subtotal Area VII Area VIII subtotal Area IX subtotal TOTAL 246 Italy Austria Croatia Slovenia Bulgaria Hungary Serbia Cyprus Greece Turkey Israel Jordan Lebanon Syria Egypt Algeria Tunisia Portugal Spain Table 121: Detailed list of Type B bangle subtypes per vast geographical section North-western European provinces Belgium; Netherlands; Luxemburg; France; Germany (Rhine region); Switzerland; UK B1 B2 B3 B4 B5 B 22 31 11 1 2 67 17,46 24,60 8,73 0,79 1,59 53,17 North-eastern European provinces Austria; Hungary; Slovenia; Croatia; Serbia; Bulgaria; Romania; Germany (Danube region) B1 B2 B3 B4 B5 B 5 1 0 0 0 6 3,97 0,79 0,00 0,00 0,00 4,76 Western Mediterranean provinces Spain; Portugal; Italy; Morocco; Algeria; Tunisia B1 B2 B3 B4 B5 B 4 3 0 3 3 13 3,17 2,38 0,00 2,38 2,38 10,32 Eastern Mediterranean provinces Egypt; Libya; Israel; Jordan; Lebanon; Cyprus; Syria; Turkey; Greece B1 B2 B3 B4 B5 B 31 5 1 1 2 40 24,60 3,97 0,79 0,79 1,59 31,74 With 127 ex. are the Type B arm rings, which occurred less frequently than Type A. Three of the Type B arm rings are from unknown provenance but might be old finds from Nijmegen (NL). From Tables 120 and 121 it is clear that Type B bangles have been recorded from all over the Roman Empire. However, the pie-charts in Figure 119 accentuate the absence of some subtypes in certain areas and showing a preponderance of solely the subtypes B2 and B3 in the north-western provinces. B4 and B5 are predominant in the western Mediterranean are whereas subtype B1 is most popular in the eastern Mediterranean. B1 B2 northwest European provinces 36% 50% west Mediterranean provinces northwest European provinces 13% northeast European provinces 2% northeast European provinces 8% west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 6% 8% 77% 247 B3 B4 0% 0% 0% northwest European provinces 8% northwest European provinces northeast European provinces 20% northeast European provinces 20% west Mediterranean provinces west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 92% 60% B5 northwest European provinces 29% 28% northeast European provinces west Mediterranean provinces 0% east Mediterranean provinces 43% Figure 119: Pie-charts visualizing the distribution range of black glass bracelets of Type B The Type C arm rings are, despite some exceptionsi.e. two fragments from Egyptian provenance [cat.nos.820;823] and a piece from Cyprus [cat.no.A4673]all originating from the north-western provinces (Table 122). Figure 120 demonstrates that all three eastern Mediterranean pieces are of Type C3. But how do we have to interpret their presence in Cyprus and in the Egyptian western desert oasis? We believe that a possible production somewhere in the eastern Mediterranean is to be excluded because the open bangles are undeniably idiosyncratic to the north-western provinces (Tables 117; 119; 121) and because this variant was only very rarely attested in the eastern Mediterranean (Table 123). Atypical for the region, they rather have to be regarded as imports from somewhere in the north-western provinces. We believe that these bangles were no direct imports through trade but arrived indirectly as personal possessions of women [see Chapter 7] coming from the northwestern provinces following their husband abroad or marrying a trader from Cyprus and Egypt they met in northern Gaul. Although epigraphy proves the presence of European citizens in the eastern Mediterranean there is unfortunately no evidence to prove the Gaulish ethnicity of the owners of these bangles as it is equally possible that the women were originally coming from Egypt and Cyprus returning to their native country with objects they acquired in Gaul. C4 east Mediterranean provinces C3 west Mediterranean provinces northeast European provinces C2 northwest European provinces C1 0 10 20 30 40 50 Figure 120: Bar-charts visualizing the distribution range of black glass bracelets of Type C 248 Table 122: Overview of the Type C black glass bangles per country and per area Area Country Quantity Open Area I Belgium France Germany Grand Duchy of Luxemburg Netherlands Switzerland United Kingdom 31 55 5 4 1 8 104 0 0 1 1 0 2 2 0 107 31 55? 5 4 1 8 104? 0 0 1 1 0 2? 2? 0 107? subtotal Area II Area III subtotal Area IV subtotal Area V subtotal Area VI subtotal Area VII Area VIII subtotal Area IX Italy Austria Croatia Slovenia Bulgaria Hungary Serbia Cyprus Greece Turkey Israel Jordan Lebanon Syria Egypt Algeria Tunisia Portugal Spain subtotal TOTAL Table 123: Detailed list of Type C bangle subtypes per vast geographical section North-western European provinces Belgium; Netherlands; Luxemburg; France; Germany (Rhine region); Switzerland; UK C1 C2 C3 C4 C 37 28 36 2 103 34,91 26,41 33,96 1,89 97,17 North-eastern European provinces Austria; Hungary; Slovenia; Croatia; Serbia; Bulgaria; Romania; Germany (Danube region) C1 C2 C3 C4 C 0 0 0 0 0 0,00 0,00 0,00 0,00 0,00 Western Mediterranean provinces Spain; Portugal; Italy; Morocco; Algeria; Tunisia C1 C2 C3 C4 C 0 0 0 0 0 0,00 0,00 0,00 0,00 0,00 Eastern Mediterranean provinces Egypt; Libya; Israel; Jordan; Lebanon; Cyprus; Syria; Turkey; Greece C1 C2 C3 C4 C 0 0 3 0 3 0,00 0,00 2,83 0,00 2,83 249 Table 124: Overview of the Type D black glass bangles per country and per area Area Country Quantity Open Area I Belgium France Germany Grand Duchy of Luxemburg Netherlands Switzerland United Kingdom 68 115 31 1 4 28 247 34 1 68 20 123 12 23 2 37 6 6 58 1 1 60 57 57 51 9 60 590 0 0 0 0 0 0 0 0 subtotal Area II Area III subtotal Area IV subtotal Area V subtotal Area VI subtotal Area VII Area VIII subtotal Area IX Italy Austria Croatia Slovenia Bulgaria Hungary Serbia Cyprus Greece Turkey Israel Jordan Lebanon Syria Egypt Algeria Tunisia Portugal Spain subtotal TOTAL Table 125: Detailed list of Type D bangle subtypes per vast geographical section North-western European provinces Belgium; Netherlands; Luxemburg; France (except Corsica); Germany (Rhine region); Switzerland; UK D1 D2 D3 D4 D5 D6 D7 D8 D 221 4 2 1 3 0 1 0 231 54,17 0,98 0,49 0,25 0,74 0,00 0,25 0,00 56,62 D8 D North-eastern European provinces Austria; Hungary; Slovenia; Croatia; Serbia; Bulgaria; Romania; Germany (Danube region) D1 D2 D3 D4 D5 D6 D7 104 13 7 0 8 1 0 0 133 25,49 3,19 1,72 0,00 1,96 0,25 0,00 0,00 32,60 D7 D8 D Western Mediterranean provinces Spain; Portugal; Italy; Morocco; Algeria; Tunisia; Corsica D1 D2 D3 D4 D5 D6 39 33 1 0 3 3 5 3 79 9,56 8,09 0,25 0,00 0,74 0,74 1,23 0,74 19,36 D7 D8 D Eastern Mediterranean provinces Egypt; Libya; Israel; Jordan; Lebanon; Cyprus; Syria; Turkey; Greece D1 D2 D3 D4 D5 D6 44 49 0 15 6 5 0 0 119 10,78 12,01 0,00 3,68 1,47 1,23 0,00 0,00 29,17 250 Table 124 demonstrates that Type D arm rings are omnipresent, in large quantities. 214 Considering the absence of bangles with an open elliptical shape, only closed circular bangles were produced from Period IV. The pie-charts in Figure 121 visualize clearly the regional difference of the variants listed in Table 125 showing some variants idiosyncratic to a large geographical zone or totally absent in one of the areas. The Type D4 bangles most likely were solely manufactured in the eastern Mediterranean and on occasion imported in the north-western provinces such as the piece with honey comb impressed decoration from Kaiseraugst (CH) [cat.no.3677], while the Type D7 and 8 variants seem characteristic to the western Mediterranean. D1 D2 4% northwest European provinces northwest European provinces 11% northeast European provinces 10% west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 33% D4 D3 0% 6% 0% 10% west Mediterranean provinces 50% 54% 25% northeast European provinces 13% 20% 0% northwest European provinces northwest European provinces northeast European provinces northeast European provinces west Mediterranean provinces west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 70% 94% D5 D6 northwest European provinces 15% 30% 0% 11% northeast European provinces west Mediterranean provinces west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces D7 0% 17% 33% 56% 40% 15% 83% northwest European provinces northeast European provinces 0% 0% D8 0% 0% northwest European provinces northwest European provinces northeast European provinces northeast European provinces west Mediterranean provinces west Mediterranean provinces east Mediterranean provinces east Mediterranean provinces 100% Figure 121: Pie-charts visualizing the distribution range of black glass bracelets of Type D 214 A more thorough investigation would sweep away the actual absence of information on some countries. 251 6.3.2.2 Finger rings A look at the distribution of finger rings in black glass demonstrates that from the 204 pieces recorded thus far, a large part came from only a few sites: Augst/Kaiseraugst (17 ex.), Braga (9 ex.), Bonn (15 ex.), Sainte Menehould (13 ex.) and Trier (16 ex.) (Table 126). Those are not by coincidence the sites where workshops were located, whether secondary glass workshops or jewellery shops like the one excavated at Bonn-Medizinische Klinik. Fourteen pieces have no detailed provenance information, of which 11 ex. come from the north-western provinces (France, 5 ex.; Germany, 5 ex.; Belgium, 1 ex.) and 3 pieces from the south-eastern Mediterranean (Syria, 2 ex.; Egypt, 1 ex.). The pieces from Menen (BE) [cat.no.448] and the Durme area (BE) [cat.no.626], could not be defined and were consequently omitted from the list in Table 126. The catalogued black glass finger rings show a distribution that is comparable to that of the bracelets, with 54 of the 65 sites concentrated in Area I. The black glass finger rings are consequently an additional example to illustrate the (inter)regional distribution of consumer goods in Roman imperial times, implying the existence of various workshops producing this trinket. Table 126: List of sites with black glass finger rings cat.no. provenance 23 143 169 184 213-14 260 295-96 304 332 370;380;386-88 437 463 516 551 2859-60 1762 1796-98;1804;1809-11; 1813-14;1817;1820; 1822-24;1830 1875 1931-32 1944 1992 2000 2001 2017 2022-23;2027;2030;2032; 2034;2036;204041;2043; 2047;2049-50;2052-53; 2055 2162 2083;2086;2093;2099;2101 3320;3326;3329-30; 3354;3372;3376;3392; 3396;3398;3401-2; 3404;3410;3414;3427 3588-89 3646;3648 1008 1010;1015 1036 1037-40 1079 1149-50 1222 252 quantity Bregenz (AT) Petronell (AT) Aiseau-Presles (BE) Arlon (BE) Braives (BE) Eprave (BE) Hamme (BE) Han-sur-Lesse (BE) Lauw (BE) Liberchies (BE) Matagne-la-Grande (BE) Nivelles (BE) Sinsin (BE) Tongeren (BE) Dalheim (LU) Albachtal-Trier (DE) Bonn (DE) 1 1 1 1 2 1 2 1 1 5 1 1 1 1 2 1 15 Type A 1 2 1 1 5 1 1 2 15 Type B 1 1 1 1 1 - Type C 1 1 - undefined Hastenrath-Eschweiler (DE) Kreimbach (DE) Mainz (DE) Pommern am Mosel (DE) Rheingönheim (DE) Rheinzabern (DE) Tholey (DE) Trier (DE) 1 2 1 1 1 1 1 16 1 1 1 4 1 1 1 1 1 - 12 - Xanten (DE) Germany Augst (CH) 1 5 16 14 3 2 2 - 1 - Avenches (CH) Kaiseraugst (CH) Arcis-sur-Aube (FR) Arles (FR) Chevincourt (FR) Cierges (FR) Forêt d‘Argonne (FR) Laneuvelotte (FR) Lugasson (FR) 2 2 1 2 1 4 1 2 1 1 1 1 1 - 1 1 1 - 1 4 - 2 1 1 1 1 - 1265 1268 1271;1273 1303;1308-9 1354 1377-78;1393 1627;1629;1635;1637 1451-52;1458-62; 1464-68;1476;1505 1647-48;1654;1659; 1661;1667 1673 1748 1752 1707-10 3917 4005 4048 4053 4138 4155-56 4174 4237-38 4297 4064-65 2598;2602;2605;2610; 2634;2644-46;2715 3015;3017;3023; 3025-28;3069 3089;3098-99 768-70;773 2209 3809-19 3820 3862;3877 3900 898 980 6.3.2.3 Mazières-en-Mauges (FR) Mesnil de Poses (FR) Metz (FR) Niederbronn-les-Bains (FR) Reims (FR) Saint-Dié-des-Vosges (FR) Saint-Etienne-Roilaye (FR) Sainte-Menehould (FR) 1 1 2 3 1 3 4 14 2 1 4 13 1 2 1 1 1 - 1 1 1 - Sens (FR) 6 3 - 3 - Sézanne (FR) Vertault (FR) Vireux-Molhain (FR) France Birdoswald (UK) Cirencester (UK) Exeter (UK) Fosse Lane (UK) London (UK) Malton (UK) Poundbury (UK) Thisleton (UK) Wanborough (UK) Witcombe (UK) Aquileia (IT) 1 1 1 4 1 1 1 1 1 1 1 2 2 2 9 3 1 1 3 1 1 1 1 1 2 4 1 2 1 1 1 2 2 - Braga (PT) 8 7 - 1 - 3 4 1 11 1 2 1 1 1 197 3 1 8 1 105 2 35 1 1 1 1 34 1 2 1 1 1 23 Conimbriga (PT) Zara (HR) Dunapentele (HU) Hauran (SY) Homs (SY) Syria Sardis (TR) Karanis (EG) Egypt TOTAL Beads The diversity of bead types and the huge number of rod-formed and pressed beads makes it a complex category to assess. Because the purpose of this chapter is to demonstrate local, regional and supraregional distribution, we only will focus on the spacer beads, which fall within the category of pressed beads. Not a negligible issue concerns the distribution of beads beyond the borders of the Roman Empire, as for instance in northern Europe beyond the Rhine- and Danube-limes, since black glass beads are involved. This socalled Barbaricum roughly corresponds to Germany, Poland and the southern part of the Scandinavian countries. A certain focus by some Scandinavian and German scholars on the glass beads (Eggers 1940; TempelmannMaczyńska 1985; Erdich, Voss 1997) and on glass vessels (Eggers 1951; Lund Hansen 1987) created a vivid debate on how to comprehend and value these artefacts within the chaine opératoire. It is self-evident that the glass vessels such as the well-known ‗painted‘ cylindrical cups with double base-ring (Isings Form 85b) are to be accepted as Roman products. But what about the beads? Were they all produced within the Roman Empire, or was there also a ‗German‘ production? No clear view on the production of glass beads has been possible hitherto because bead-making workshops have rarely been attested within the Roman Empirefor example, Trier (DE) (Loeschke 1925; Schulze 1978; Goethert-Polaschek 1984), Tibiscum (RO) (Benea 1983)and nearly fail when looking beyond its border (Erdich, Voss 1997, 86). Furthermore, the various bead types recorded in northern Central Europe cannot be linked to one of the recognized bead-making workshops. In addition, we have to bear in mind that north-central European bead types are equally present in the Roman Empire unless we have to see these beads as Roman which circulated into the adjacent area just beyond the limes. Some scholars propose/are convinced of local production of beads in northern Central Europe (Guido 1978; 1999). We probably have to see 253 the presence of such material as the effect of trade from the Roman Empire on Barbaricum, just as it is acknowledged for the glass vessels, sigillata ware, coins and bronze material. The complication concerning the th th black glass beads is that comparisons have been made with 5 to 7 century AD burial contexts in the Roman frontier zone bordering Barbaricum, i.e. from Great Britain over the Rhine and Danube regions to the Black Sea. It is true that, for instance, black annular beads with zigzags in opaque white, red, yellow or blue glass trails present rd th in 3 to 5 century AD dated cremation tombs in Barbaricum are equally attested in, for instance, Anglo-Saxon, th th Merovingian and Aleman tombs of the 5 to 7 century AD. It is tempting to presume a German tradition and consequently a production by the ‗Völkerverwanderung‘ people (Guido 1999). But this is a too simplistic analysis and to some extent a biased assumption by some scholars. What do we do with similar beads in, for instance, Cyprus, Israel or even Egypt? Do we have to see them as examples of long-distance trade from Barbaricum? What other material from northern Central Europe entered the Roman Empire that far to confirm long-distance trade? Furthermore, how do we interpret the much wider variety of bead types in the Roman Empire [see Chapter 3] and especially in the late Roman-early Byzantine Eastern Mediterranean compared to the very limited number of types in Barbaricum, where only annular and short, cylindrical black glass beads have been recorded? When looking to the frequency of these black glass beads with opaque coloured trails, the types found in north Germany (Lower Saxony and Slesvig-Holstein) and following Magdalena Tempelmann-Maczyńska‘s typology: 2 ex. T263; 17 ex. T293; 3 ex. T300; 1 ex. T301; 3 ex. T302; 1 ex. T303 (Erdich, Voss 1997, 93). The publications on glass beads by Magdalena Tempelmann-Maczyńska and Margaret Guido focus on a specific regionthe first specifically on the material from outside the Roman Empire in Barbaric Europe and the latter on the British glass beads. It is therefore important to take notice of the regional character of the information on the productiondistributionconsumption on Roman glass beads. The result is that some shapes are present or absent in Guido‘s list, and others are incorporated within a list of non-Roman beads. Their publications unfortunately can be the root for biased assumptions. We think, for instance, of the so-called ‗Trilobitenperlen‘ 215 discussed for the first time in detail by Thea Haevernick (1983). It is clear that a massive task waits to be developed, starting with a general synthesis on the Roman glass beads from the entire empire. This is, however, a project for an entire team or even a group of teams inventorying the material from delimited regions. Only then will data be obtained that make it possible to understand the variety of regional idiosyncrasies and almost certainly also the chronological variability. Despite the incorporation of rod-formed, barrel-shaped beads in a typology on Barbaric Central Europe (Tempelmann-Maczyńska 1985, 185) and ascribed as imports by the Migration People (Guido 1999), it is important to stress that they appear rather singularly outside the Roman Empire. Conversely, these late Roman productions [see Chapter 4] are omnipresent within the Roman Empire. Table 127 demonstrates that all rod-formed beads in black appearing glass are characteristic to the European continent whether it be cylindrical beads or both variants of so-called ‗melon‘ beads. Table 127: List of sites with rod-formed beads in black appearing glass (n = quantity) cat.no. provenance decoration n size annular ‗crumb‘ bead annular ‗crumb‘ bead annular ‗crumb‘ bead annular ‗eye-bead‘ large ‗melon‘ bead - plain ‗melon‘ bead - plain small small ‗melon‘ bead - plain ‗melon‘ bead - plain ‗melon‘ bead - plain ‗melon‘ bead - plain ‗melon‘ bead - plain ‗melon‘ bead - plain ‗melon‘ bead - plain ‗melon‘ bead - plain small small small small small small small small 2376;2410 2492-93 2503 2 7-13;15-19;22;24-25 26-30 31 106 122 135;140;154 1011 3325;3331; 3333;3360; 3363;3374; 3416 A4509 229 1232 3186;3191 649 4166-67 1373 215 Horbat Qastra (IL) Tell Abu Shusha (IL) unknown provenance (IL) Tipasa (DZ) Bregenz (AT) Eisenstadt (AT) Hallstatt a.d. Lahn (AT) Magdalensberg (AT) Müllendorf (AT) Petronell (AT) Arles (FR) Augst (CH) 2 2 1 1 15 5 1 1 1 3 Cutry (FR) De Panne (BE) Dunapentele (HU) Maule (FR) Mérida (ES) Nicopolis (BG) Nor‘Nour (UK) Saint-Denis (FR) 1 1 7 1 2 1 2 1 All known double-perforated beads such as the so-called ‗Trilobitenperlen‘ in Britain are in jet. 254 534 3174 1113 A4597 3352;3338;3342 3336 3424 198 A4598 1895-98 A4599 536 208 796 809 836-37 932;936-39; 943; 945-50; 952-53; 956; 958; 974; 976; 978 1239;1255 1742;1744 1764-65 1835 1989 2035;2038; 2059;2065 2189;2289; 2291;2298-99;2301;2304 2198;2220; 2223-26;2228; 2230;2237 2248 2251 2294 2309 2311 2341-42;2350;2357; 2365;2385-86;2393-95 2623 2962 3002;3005-6;3009; 3012-13;3016;3018-21; 3024;3030 3663;3673-74;3679 Tienen (BE) Trebnje (SL) Frénouville (FR) Anreppen (DE) Augst (CH) Augst (CH) Augst (CH) Blicquy (BE) Haltern (DE) Kempten (DE) Oberwinterthur (CH) Tienen (BE) Bossut-Gottechain (BE) Akhmim (EG) Bahnasa (EG) Dush Kysis (EG) unknown provenance (EG) 1 1 1 1 3 1 1 1 1 4 1 1 1 1 1 2 19 ‗melon‘ bead - plain ‗melon‘ bead - plain ‗melon‘ bead - spiralling trail (yellow) ‗melon‘ bead - spiralling trail (white) ‗melon‘ bead - spiralling trail (white) ‗melon‘ bead - spiralling trail (red-brown) ‗melon‘ bead - spiralling trail (yellow) ‗melon‘ bead - spiralling trail (white) ‗melon‘ bead - spiralling trail (white) ‗melon‘ bead - spiralling trail (white) ‗melon‘ bead - spiralling trail (white) ‗melon‘ bead - spiralling trail (white) unspecified type unspecified type unspecified type unspecified type unspecified type small small small large large large large large large large large large - Maule (FR) Vermand (FR) Altbachtal (DE Burgheim (DE) Pähl (DE) Trier (DE) 2 2 2 1 1 4 unspecified type unspecified type unspecified type unspecified type unspecified type unspecified type - unknown provenance (HU) 7 unspecified type - Dunapentele (HU) 9 unspecified type - Gyór (HU) Kapsovár (HU) Szóny (HU) Vasasszonyfa (HU) Veszprém (HU) Horbat Qastra (IL) 1 1 1 1 1 10 unspecified type unspecified type unspecified type unspecified type unspecified type unspecified type - Aquileia (IT) Nijmegen (NL) Braga (PT) 1 1 13 unspecified type unspecified type unspecified type - Kaiseraugst (CH) Bónaduz (CH) Tamins (CH) Krefeld-Gellep (DE) Mainz (DE) Saint-Martain-de-Fontenay (FR) Lisieux (FR) Villa d‘Ancy (FR) Villers-Erquery (FR) Dunapentele (HU) Lankhills (UK) Lydney (UK) Augst (CH) Burgheim (DE) Dunapentele (HU) Sagvar (HU) Trier (DE) Virton (BE) 4 1 1 1 1 4 1 1 1 3 1 1 1 1 3 3 1 1 unspecified type Guido 4 Guido 4 Guido 4 Guido 4 Guido 4 Guido 4 Guido 4 Guido 4 Guido 4 Guido 4 Guido 4 crossing zigzag + eyes spiralling trailed cylindrical bead spiralling trailed cylindrical bead spiralling trailed cylindrical bead spiralling trailed cylindrical bead spiralling trailed cylindrical bead - Pressed beads Pressed beads are present all over the Roman Empire, whether or not in the so-called opaque black glass. For instance, the ‗Trilobitenperlen‘ have been recorded from Spain, France, Belgium, Germany, Switzerland, Austria, Italy, Slovenia, Croatia, Hungary, Bulgaria, Romania, Ukraine, Greece, Turkey, Cyprus, Egypt and Tunisia (Table 128). The ‗Trilobitenperlen‘ were ubiquitous in the entire Roman Empire. The wide variety of types in and large sets of double-perforated, pressed beads known from Raetia, Noricum, Pannoniae, Dalmatia as well as the SyroPalestinian area make these types idiosyncratic to the eastern Alpinewestern Balkan area as well as the Levant. The spacer beads are therefore an example of regionally bounded productions for a more ‗globalized‘ (read empire-wide) consumption pattern. An interesting exception is Britannia, where such beads also do occur, but in jet and not in black glass (Allason-Jones, Jones 2001). As we could see from the bracelets [see Chapter 6.3.2.1.], the British Isles produce in jet all kinds of comparable beads with double lateral perforation. Most probably these examples were made from Whitby jet in the neighbouring area of York, but we have to bear in mind that various sources of jet were known in Roman times as well as that other material looks similar to jet, such as cannel coal, 255 shale or durain. In contrast to the flat and irregularly shaped beads in glass, those in jet are very regularly cut and polished and are almost semi-spherical. The ‗Trilobitenperlen‘ in jet that most resemble the ones in black glass are those found in and around Cologne (Hagen 1937). Ubiquitous, double-perforated pressed beads were not present in every settlement and generally occured individually or in small quantities, as can be seen from the examples from burial gifts [see Chapter 5]. indefinite discoid shaped 8-shaped lion’s head theatre mask Medusa head provenance portrait cat.no. ‘Trilobitenperlen’ Table 128: List of sites with double-perforated pressed beads in black appearing glass north-western provinces 292 529 2061 1054 1339 1705 2878 3222 3366 3644;3665 3780 3800-2 Hamme (BE) Tielrode (BE) Trier (DE) Dieulouard-Scarponne (FR) Orange (FR) unknown provenance (FR) Esch (NL) Aarau (CH) Augst (CH) Kaiseraugst (CH) Kloten (CH) unknown provenance (CH) 1 1 1 1 1 1 1 (14) 1 1 2 1 3 3 north-eastern provinces 1766 1867 1872 1888-90 1891 1892 1963 1967-68 1990 1991 1996 6 26-30 31 32 33-34 122 123-34;136-39;141;14445;147-53;155-57 158-59 160 162 632-33 634 635 636 637-38 652 658-59 660 256 Altenerding-Klettham (DE) Eining (DE) Epfach (DE) Göggingen (DE) Hassleben am Lichwege (DE) Isny-Bettmauer (DE) Munich (DE) Murnau (DE) Pähl-Aidenried (DE) Pliening (DE) Regensburg (DE) Baden (AT) Eisenstadt (AT) Hallstatt (AT) Krungl bei Aussee (AT) Leithprodersdorf (AT) Müllendorf (AT) 1 1 1 3 1 1 (10) 1 2 1 1 2 1 4 1 1 2 1 Petronell (AT) 26 3 Pöttsching-Zillingthal (AT) Stinkenbrunn (AT) Wattens (AT) Dracevica (BS) Proboj (BS) Rudine (BS) Stolac (BS) Krvina (BG) Nicopolis (BG) Odessos (BG) Plovdiv (BG) 1 1 1 1 2 1 1 1 1 1 1 2 1 661 664 717-18;720-32 738-41;744-46;748;75052;754 774-775 789 2180 2181 2190 2191-92 2193 2195-97;2201-2;220-7; 2212-18;2221-22;2227; 2229;2231-36;2238-42 2243 2244-47 2249 2250 2252-56 2258 2280-81 2265 2282-83 2293;2295-97 2308 2310 2188;2290;2300;2302;23056 3115 3120 3150 3153 3168;3172 3175 3106 3107 3901-2 3904 3905 3906 Sofia (BG) Betegenica (HR) Sisak (HR) 1 1 11 4 Split (HR) 7 5 unknown provenance (HR) Csúny (CZ) Werria (GR) Alattyán (HU) Bezenye (HU) Cikó (HU) Csákvár (HU) 2 1 1 1 1 2 1 Dunapentele (HU) 21 Felsodobsza (HU) Fenék (HU) Gyór (HU) Halimba (HU) Keszthely (HU) Mosonszentjános (HU) Pilismarót-Basaharc (HU) Pécs (HU) Sitke (HU) Szóny (HU) Várpatola (HU) Vasasszonyfa (HU) 1 4 unknown provenance (HU) 2 Celje (SL) Drnovo (SL) Hru ica (SL) Ptuj (SL) Ribnica (SL) unknown provenance (SL) Istros (RO) Cetatea Bisericuta (RO) Chersones (UA) Kerch (UA) Kuban (UA) unknown provenance (UA) 1 1 1 1 2 1 1 1 1 1 1 9 1 1 2 1 2 2 1 1 1 3 1 1 1 1 3 1 1 western Mediterranean provinces 2596-97;2603-4;26069;2611-17;2619-20;2622; 2630-33;2635-40;264243;2660;2666;2677-78; 2680-85;2695;2703;27056;2708-14; 2730 2716;2718-19;2722 2735 2739-40 2742 2755 2763 2776 2783 2806-7 2814 2819-23 3088 3180 3183 3184 Aquileia (IT) Ascoli Piceno (IT) Castel Trosino (IT) Cesena – San Egidio (IT) Chiusi (IT) Mattarello (IT) Naunia/Nonsberg (IT) Porto Torres (IT) Rome (IT) Salorno (IT) Siracuse unknown provenance (IT) Conimbriga (PT) Belo (ES) Elche (ES) Italicà (ES) 24 1 24 1 19 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 2 3 257 3190;3193 3200 3212 3213 3890 3891 3894 Mérida (ES) Moreruela de Tábara (ES) Torredonjimeno (ES) unknown provenance (ES) Carthage (TN) La Skhira (TN) Tunis (TN) 2 1 1 1 1 1 1 eastern Mediterranean provinces 785 797 994-95 2501;2517-18;252324;2527-28;2537 3895 6.3.2.4 unknown provenance (CY) Akhmim (EG) unknown provenance (EG) 1 1 1 1 unknown provenance (IL) 3 1 Ephesus (TR) 1 1 1 2 5 Pendants A correct overall picture on the distribution of this type of commodity cannot be offered because of the exclusion of the pieces in other glass hues (mainly deep blue and yellowish amber) and the restricted information 216 concerning the provenance of the 199 inventoried pendants in black appearing glass. Table 129 gives an idea about the distribution of the catalogued black glass pendants showing a concentration in the south-eastern Mediterranean (55,7%) with a relative distribution in the rest of the Mediterranean (26,4%), and remaining rather rare at the European continent (respectively 12,1% for the north-western provinces and 5,8% for the northeastern provinces). The listed sites corroborate the report on jug-shaped pendants in various glass hues by Marianne Stern (1977, 112-115) who demonstrated a frequent occurrence in the south-eastern Mediterranean compared to the rather scattered finds on the European continent. Presumed to have been amulets for the early Christian community (Loeschke 1925, 359-360; Eisen 1927, 520-521; Stern 1977, 112-113) led to the hypothesis on the origin of this type of consumer‘s good suggesting that these characteristic pendants were, likewise a good number of beads and bangles, pilgrim‘s souvenirs from the Holy Land (Foy 2010, 306-310) [see Chapter7]. Striking however is the fairly high occurrence with 21 records (15,8%) in large centres in the West like Aquileia (IT) of the late Roman barrel-shaped pendants with two or three suspension loops. These pendants with applied spiralling glass trails in an opaque contrasting colour nevertheless show a distribution all over the Roman Empire, as confirmed with examples from Horvath Qastra (IL) [cat.no.2390] and elsewhere in Israel (Spaer 2001, 102-103, fig.47; 113, nos. 173-174), Braga (PT) (da Cruz 2009, 99, fig. 3:421-422), and Colchester (UK) (Crummy 1983, 35, fig.37:1504). This concentration can be explained by the presumed presence of at least one workshop in Aquileia that processed black glass in the late Roman period and possibly produced such pendants [see this Chapter 6.2 Production]. Also the concentration of jug-shaped pendants in Rome can be considered the result of a local workshop (Stern 1977, 110-113). Because most other are limited to the south-east Mediterranean it is not excluded that the (black) glass pendants were early Christian pilgrim‘s souvenirs from the Holy Land. But based on the idea of a similar consumption at possible places of pilgrimage in Western Europe it is equally possible to deduce from the distribution pattern that black glass pendants were produced in western workshops like various other jewellery objects to supply the early Christian communities in western centres such as Rome, Aquileia and Trier with fashionable symbolic trinket. From this point of view it is clear that a thorough study on the (black) glass pendants is compulsory to verify whether there are indications to detect regional differencessuch as technical and morphological or decorative featuresor whether we are dealing with an individual or bulk import from the south-eastern Mediterranean. Most interesting is the distribution of these black glass artefacts as possible evidence for the presence of early Christian communities of Jewish, Syrian, Greek or Egyptian origin at centres or areas in the western part of the Roman Empire. If the various types and subtypes can be connected to particular pilgrimage places a detailed study of their distribution would reveal the importance of holy places by showing the intensity of pilgrimage and power of attraction of a sanctified place. 216 From 25 records no provenance information is available and from another 50 records only the country of provenance is known. 258 Table 129: List of various types of black glass pendants per region North-western European provinces Belgium; Netherlands; Luxemburg; France (except Corsica); Germany (Rhine region); Switzerland; UK poppy-head other globular barrel-shaped jug-shaped discoid-shaped shaped TOTAL 3 1 2 13 0 2 21 1,7% 0,6% 1,1% 7,5% 0% 1,1% 12,1% North-eastern European provinces Austria; Hungary; Slovenia; Croatia; Serbia; Bulgaria; Romania; Germany (Danube region) poppy-head globular barrel-shaped jug-shaped discoid-shaped other shaped TOTAL 0 1 0 9 0 0 10 0% 0,6% 0% 5,2% 0% 0% 5,8% other TOTAL Western Mediterranean provinces Spain; Portugal; Italy; Morocco; Algeria; Tunisia; Corsica poppy-head globular barrel-shaped jug-shaped shaped discoid-shaped 4 15 5 17 2 3 46 2,3% 8,6% 2,9% 9,8% 1,1% 1,7% 26,4% other TOTAL Eastern Mediterranean provinces Egypt; Libya; Israel; Jordan; Lebanon; Cyprus; Syria; Turkey; Greece poppy-head globular barrel-shaped jug-shaped discoid-shaped shaped 0 2 10 58 7 20 97 0% 1,1% 5,8% 33,3% 4,0% 11,6% 55,7% 6.3.2.5 Gems We are well aware of the very limited amount of black glass gems we have inventoried (336 ex.), in particular because the very ubiquitous nicolo paste gems have not been catalogued systematically. The discussion on its distribution is therefore to be taken as an opening appraisal indicating the necessity of a more enhanced assessment on this specific segment of the entire glass gemstones‘ production and consumption. 6.3.2.5.1 Monochrome gems o Plain conical gems A first functional type is the conical gem set in the centre of discoid brooches [see Chapter 7]. The plain conical glass settings in the centre of round and oval gilded bronze disc-shaped brooches, respectively Hull Type 270 and Hull Type 271 [see Chapter 3], are generally considered of RomanoBritish manufacture (Mackreth 1992, 60). This assumption is strengthened by their total absence in Scotland and the disproportionate concentrations in Britain vis-à-vis the continent (Simpson, Blance 1998, 277; Bayley, Butcher 2004, 178-179). Regardless of the multiple finds of this specific brooch from Britannia, we believe that this type is equally present on the continent, notwithstanding its relatively rareness (Table 130). The low occurrence on the continent is partly due to the fact that large numbers of these brooches have lost the central setting and when preserved it is regularly considered black stone (Vrielynck, 2007, 54, fig.6). Seeing the examples in present-day Belgium, Germany and Switzerland, these brooches are unmistakably present in the Rhine region and thus at least common in the provinces Germania Inferior and Germania Superior. Punzverzierte ovale Scheibenfibel mit Glaseinlage (Hull Type 271) were also attested beyond the Roman Empire into Barbaricum (Riha 1979, type 3.17; Böhme 1972, Type 45), and 217 despite the lack of evidences such fibulae were possibly used in the Danube region too. Is it possible that the rd discoid brooches are part of the military dress code and therefore to be considered as militaria of at least the 3 century AD in the north-western or northern frontier zone? 217 We have not been able to verify whether this type occurs all the way down the Danube limes. 259 Table 130: List of plain conical-shaped gems in monochrome black appearing glass set in disc brooches (n.c. = not catalogued; Hull Type T270 = round; Hull type T 271 = elliptical) cat.no. provenance brooch type reference 206 Bossu-Gottechain (BE) T270 Vrielynck, 2007, 54, fig.6 A4698 Tongeren (BE) T270 Vanvinckenroye 1984, 96-97, pl.91 3586 Avenches (CH) T270 Guisan 1975, 26, no.3.15, pl.11:15 3684 Kaiseraugst (CH) T271 Riha 1979, 88, no.309, Taf.13 A4699 Hofheim (DE) T270 Ritterling 1913, 253-254, pl.X A4700 Miltenberg (DE) T270 ORL 38, pl.IV, no.14 A4701 Regensburg (DE) T270 Lamprecht 1904, pl.IV, no.7 A4702 Regensburg (DE) T271 Lamprecht 1904, LVIII, pl.VIII, no.7 A4703 Saalburg (DE) T270 Böhme 1972, pl.29, no.1134 A4704 Zugmantel (DE) T271 ORL 8, pl.X, no.26 A4705 New Grange (IRE) T271 Carson, O‘Kelly 1977, 52, pl.VIIa, no.E.56.976 A4706 New Grange (IRE) T270 Carson, O‘Kelly 1977, 53, pl.VIIa, no.E.56.1711 2996 Vechten (NL) loose gem unpublished A4707 Benwell (UK) T271 Bayley, Butcher 2004, 178-179 A4708 Bramble Hill (UK) T271 Yorkshire Arch. Journal XXXIX, 1958, 54 A4709 Brough (UK) T271 Bayley, Butcher 2004, 178-179 A4542 Caernarfon (UK) T271 Casey et al. 1993, 166, no.13 A4710 Caister-on-Sea (UK) T270 Darling, Gurney 1993, fig.42, no.13 A4711 Canterbury (UK) T270 Journal Brit. Arch. Ass. XVI, 1860, 274; 324, fig. 4, pl.XXIII A4712 Chesters (UK) T270 Budge 1903, 381, no.1105 A4713 Chesters (UK) T271 Budge 1903, 381, no.1102 A4714 Chesters (UK) T271 Budge 1903, 381, no.1112 A4715 Cirencester (UK) T270 Bayley, Butcher 2004, 178-179 A4716 Corbridge (UK) T270 Bayley, Butcher 2004, 178-179 A4717 Corbridge (UK) T270 Bayley, Butcher 2004, 178-179 A4718 Dover (UK) T270 Philip 1981, fig.32, no.72 A4719 Emscote (UK) T270 Ant. Journal V, 1925, 269, pl.XXIX, no.2 A4720 Faversham (UK) T270 Guide Ant. RB.BM 1922, 61, fig.76 A4721 Gatcombe (UK) T271 Branigan 1977, fig.26, no.505 A4722 Icklingham (UK) T270 Bayley, Butcher 2004, 178-179 A4723 Ham Hill (UK) T270 St George Gray 1910, fig.4 A4724 Harpham (UK) T270 Yorkshire Arch. Journal XXXIX, 1958, 54 A4725 Henley Wood (UK) T270 Watts, Leach 1996, 82, fig.88, no.21 A4726 Hinksey (UK) T270 Bayley, Butcher 2004, 178-179 A4727 Hockwold (UK) T270 Gurney 1986, fig.40, no.8 A4728 Housesteads (UK) T270 Bayley, Butcher 2004, 178-179 A4729 Housesteads (UK) T271 Arch. Aeliana XI, 1934, pl.XXIX, no.4 A4730 Kettering (UK) T270 Proc. Soc. Antiq. London XXII, 1907-9, 50 A4731 Kirkby Thore (UK) T271 Royal Arch. Inst. York Meeting 1848, 15 A4732 London (UK) T270 Bayley, Butcher 2004, 178-179 A4733 London (UK) T271 Guide Ant. RB.BM 1922, 61, fig.76 A4734 Lowbury Hill (UK) T271 Atkinson 1916, fig.13, pl.9, no.35 A4735 Lowbury Hill (UK) T271 Atkinson 1916, fig.13, pl.9, no.36 4165 Nor‘Nour (UK) T271 Arch. Journal CXXIV, 1967, fig.24, no.237 A4736 Richborough (UK) loose gem Bayley, Butcher 2004, 134, no.388 4185 Richborough (UK) T270 Bushe-Fox 1949, Richborough IV, 121, no.63; Bayley, Butcher 2004, 135, no.390 A4737 Silchester (UK) T270 unpublished A4738 Silchester (UK) T271 Boon 1974², fig.19, no.1 4233 Stonea Grange (UK) T270 Jackson, Potter 1996, 320, no.79, fig.99 A4739 Stony Stratford (UK) T270 unpublished A4740 South Shields (UK) T270 Allason-Jones, Miket 1984, 116, no.138 n.c. South Shields (UK) T270 Allason-Jones, Miket 1984, 116, no.139 n.c. South Shields (UK) T270 Allason-Jones, Miket 1984, 116, no.140 n.c. South Shields (UK) T271 Allason-Jones, Miket 1984, 115, no.136 4222 South Shields (UK) T271 Allason-Jones, Miket 1984, 116, no.137 A4741 South Wiltshire (UK) T270 unpublished A4742 Swaffham (UK) T270 Victoria County Hist. Norfolk I, 321 A4743 Uley (UK) T270 Woodward, Leach 1993, fig.125, no.5 260 A4744 Uley (UK) T270 Woodward, Leach 1993, fig.125, no.6 A4745 Uley (UK) T270 Woodward, Leach 1993, fig.125, no.7 A4746 Upper Hopton (UK) T270 Brown 1986, 39, fig.25, no.184 A4747 Vindolanda (UK) T270 Henig 1975, pl.IV 4296 Wanborough (UK) T270 Butcher 2001, 64, fig.26, no.137 4298 Wanborough (UK) T270 Butcher 2001, 64-65, fig.26, no.138 A4748 Water Newton (UK) T270 unpublished A4749 Wroxeter (UK) T270 BusheFox 1912/3, 23, fig.9, no.1 To have this assumption corroborated, we checked some publications on brooches in adjacent regions. The extensive study by Jacques Philippe (2000) on the brooches from the Seine-et-Marne region which corresponds to the civitates of the Parisii, Meldi and Senones in the provincia Gallia Lugdunensis (prima during the three first th centuries and quatuor in the 4 century AD) and neighbouring Gallia Belgica at the south show a striking absence of such discoid brooches in this area. This nonappearance might be an indicator that discoid brooches with plain conical gemstones (in black glass) are limited to the frontier zone in the broad sense, including the provinces Britannia, Gallia Belgica, Germania inferior and Germania superior. What emerges as well from Jacques Philippe‘s catalogue is that the representation of a thick-cheeked human head in a beaded circle at the centre of brooches from Seine-et-Marne region is comparable to the one from the discoid brooch of Wancennes (BE) [cat.no.617]. Whereas the Wancennes piece is made of a black-appearing green glass, the medallions of the Seine-et-Marne brooches are made in gilded bronze sheet by using the repoussé-technique (Philippe 2000, fig.33:171; fig.36:X1b,c,f,i; fig.63:433). 6.3.2.5.2 Bichrome gems o Nicolo paste gems A second group of gems made with black glass are the ubiquitous bi-chrome glass gemstone imitating the 218 semiprecious nicolo gemstones and generally called ‗nicolo paste‘. rd It became in particular successful during the 3 century AD, as can be deduced from the long catalogue lists incorporating this type of jewellery. The booming of this glass substitute for nicolo is said to be due to thwarted rd trade caused by the arising conflict with the Sassanid Dynasty by the mid 3 century AD. These Persian rulers destabilized the socio-economical situation in the eastern Roman provinces consistently and as a result the longnd distance trade with India, Afghanistan and China shrank. Its popularity from the 2 century AD onwards is in correlation with the culminating conflict in the east regarded as responsible for the success of the imitations in glass. Furthermore, it is clear that by using moulds to press the figure into the glass, highly skilled engravers were not needed to embellish a glass gem, and time did not need to be allotted for the engraving work. The here-discussed gems in very deeply coloured glasses appearing black were produced in the West and more particularly in the north-western provinces Britannia, Gallia Belgica and Germania Inferior and Superior. Their nd rd th production is to be dated from the later 2 century AD up to the very end of the 3 -early 4 century AD. More indepth research will provide more precise information concerning the workshops where gems have been produced and regarding the chronology of production. As stated in the introduction, interesting information is provided by a detailed study of the moulded glass gemstones to identify the distribution range of one single set of serial produced gemstones. Martin Henig (1978²) discussed the frequent depiction of the so-called ‗Bonus Eventus‘ figure on nicolo paste gemstones in Britannia but also the repeatedly occurrence in various places in the United Kingdom of gemstones with the representation of a specific ‗Bonus Eventus‘-type from the same mould (Henig 1997, 285): Birdoswald (Henig 1997, 285, no.89, fig.195) [cat.no.3918], Lowbury Hill (Henig 1978², no.192) [cat.no.4149], Poundbury (Henig 1978², no.App.42) [cat.no.4175]. Very similar but apparently from a different mould is Carlisle (Henig 1997, 285) [cat.no.3964], and St. Albans (Henig 1978², no.App.4) [cat.no.A4609]. Martin Henig proposes that although identical, they ultimately have been made from the same prototype but from various worn moulds (Henig 1997, 285). When several examples of one single type are found in the same place, it becomes even more intriguing and interesting. A good example is the set of 21 gemstones found at ‗Medizinische Klinik‘ near the gateway of the military camp of Bonn 218 Nicolo is a layered chalcedony quartz of the agate/onyx family characterised by a blackish basal layer and an ultramarine blue top layer extracted in the Near East up to India and Afghanistan. 261 together with a large number of finger rings in glass, jet and bronze (Platz-Horster 1984, 11-16; 38-48, pls 4-7, nos.11-31) [cat.nos.1799-1803;1805-8;1812;1815-16;1819;1821;1825-29;1831]. This context has been interpreted as a jewellery (work)shop and contained all kinds of finished and semi-finished products, showing the intermediate stage ready to be employed for completing a jewel. Three gemstones have the same imprint with a tree in the centre, a shepherd at the left and several goats all around. Gertrud Platz-Horster (1984, 14; 44-45) mentions eight identical equivalents come from Cologne (DE) [cat.no.A4753], Xanten (DE) (still set in a massive silver finger ring) [cat.no.2169], Vechten (NL) [cat.no.A4752] and Cruseilles (FR) [cat.no.A4751]. A similar example but in yellow glass is reported to have been found in Aquileia (IT) (Sena Chiesa 1966, 204, pl.24:463). Two other gemstones from Medizinische Klinik, Bonn (Platz-Horster 1984, 14; 38-39; nos.11-12) [cat.nos.18151816], showing a sitting Jupiter, are similar to one from Cologne [cat.no.A4754] except for the fact that the right arm leaning on the back of a chair is not visible on the former intaglios. Opposite to the explanation Martin Henig gave on the Birdoswald Bonus Eventus piece and its copies, the pieces with various quality of impression from Medizinische Klinik, Bonn makes clear that all identical pieces must have been made in the same workshop and consequently from the same mould. It could be that the mould was impressed in the glass with different force or the glass of the different gems did not have the same viscosity. This issue is worked out further in Chapter 7. Only few large round nicolo paste gems have been recorded, but this is partly due to the failing of an exhaustive heuristic research on nicolo paste gems in general and partly because much material is not recognized 219 correctly. These large round nicolo paste gems are mainly loose finds from Roman civil settlements, such as France (Sens), Switzerland (Augst; Kaiseraugst), and Portugal (Conimbriga), or still set in a Roman artefact, like the eight pieces in the military helmet from Berkasovo (SB) (Figure 132). 6.3.2.6 Hairpins Hairpins in glass are not frequently attested, but their occurrence is more widespread than is generally assumed. Pins in a translucent yellowish-brown, bluish-green or pale green glass are known from England (Crummy 1983; Allason-Jones, Miket 1984, 275) to the Danube region (Topál 2003). Those in black glass more or less remained ignored until now (Cosyns 2009). Despite the low rate of distribution of the hairpins in black glass, we can conclude from the recorded pieces that at least one glass workshop responsible for the production of these pins was located in the north-western provinces. The glass workshop of Les Houis near Sainte Menehould (FR), bordering the Argonne forest, yielded thus far 23 pieces [cat.nos.1418;1442-48;1453;1470-71;1473-75;1477-85] 14 pin-heads and 9 shaft fragments in black 220 glass. But another 9 piecessaid to have been retrieved from the Argonne Forestare possibly from Les Houis as well [cat.nos.1076-78;1082-87]. Other hairpin fragments are the polyhedric head from a late Roman burial at the cemetery of Beavais (FR) (Schuler 1995, 62, fig.19:3) [cat.no.1027], and a globular pinhead from tomb 644 of the cemetery ‗Pousse Motte‘ at Maule (FR) (Arveiller-Dulong 2006, 159, n°41) [cat.no.1248]. Two shaft fragments are known from the vicus Geminacum, the settlement in Liberchies (BE) on the road BoulogneBavay-Tongeren-Cologne [cat.nos.372-373]. Three globular pin-heads came from the ‗cave-sanctuary‘ Trou de Han in Han-sur-Lesse (BE) [cat.nos.309;311;313] and one came from the military camp in Oudenburg (BE) [cat.no.471]. Seeing that Liberchies, Trou de Han and Oudenburg are not that distant from Sainte Menehould, it is most likely that they were produced in the glass workshop of Les Houis although it is not excluded that other glass workshops produced similar hairpins in black glass such as the supposed workshop at Liberchies. Although located fairly apart, the pieces from Augst (CH) [cat.no.A4592] and Wanborough (UK) [cat.no.A4620] perhaps corroborate one single workshop responsible for producing glass hairpins. Furthermore the twisted hairpin tapering towards one end, now at the Petrie Museum of Egyptian Archaeology [cat.no.988] is of a totally different category in shape, decoration and technique. Besides, the piece retrieved from Braga (PT) [cat.no.3003] is the only piece from outside the north-western provinces. Despite the lack of explicit evidence we believe that more workshops manufactured hairpins entirely in (black) glass or by combining a glass head with a metallic pin. Besides the hairpins entirely made in glass, there are several examples known of a pin in iron or bronze with a black glass head. An intact piece comes from the villa of Shakenoak (UK) [cat.no.4201]. Because such pieces have been catalogued from the Sainte Menehould material, it is once again very likely that the workshop of Les Houis was responsible for the production of this variant of hairpin. 219 In Chapter 4 we give details on the re-use of large amounts of Roman nicolo paste gems in medieval times, the so-called ‗Alsengemmen‘. Possibly many more have been retrieved from the site, but we only considered the pieces at the M.A.N. in Saint-Germain-en-Laye and only looked at the black glass artefacts excluding all hair pins produced in another glass hue. 220 262 Notwithstanding the few pins recorded in black glass, we assume that glass hairpins must have been a much more ubiquitous commodity in Roman daily life. This is to a large extent due to the difficulty to recognize the commodity type with only the pinhead remaining. 6.3.3 Architectural decoration Where the glass cake production has taken place still remains an open question. Were these glass cakes produced by a few larger ‗intermediate‘ workshops retaining the specialized knowledge to colour raw glass as described in schemes 2 and 3? Were such discs traded over long distances, or was their production much more widespread and manufactured by most local secondary workshops? Les Houis produced not only a mass of glass tesserae in various colours, but also some fragments of glass cakes. Although more than 450 tesserae in black glass are within the collections of the M.A.N., Saint-Germain-en-Laye, none of the glass cakes is made of black glass. Fragments of eight glass cakes being about 10-15 mm thick have been reported at Augusta Raurica (CH) (Schmid 1993, 172-173). Besides one ultramarine blue piece and four greyish blue fragments, three of them were nd rd made of a dark green glass with red trails. When from dated contexts, they are dated from the late 2 to early 3 century AD. Debora Schmid also reports 43 similar glass cake fragments from the villa of Leudersdorf (CH) as well as some from Trier (DE) which are related to the making of tesserae (Schmid 1993, 173). The opaque and barely translucent dark green glass cakes from Augst/Kaiseraugst (CH) could very well have been locally produced since a glass workshop was found there in Regio 17C in the lower town of Augusta Raurica. The nd workshop processed black glass during the later 2 century AD and the entire Severan period to produce blackappearing glass vessels (Fischer 2009). In Chapter 8, it is described that the vessels and the glass lumps show a homogeneous and very distinct chemical composition. Because the bracelets also demonstrate the same distinct composition, we may assume that the deeply coloured glass for the vessels and the bracelets were made in one and the same glass workshop [see Chapter 8]. By checking the composition of the three dark green glass cake fragments, we could compare them and verify whether they were made from the same (colouring) recipe. But if the Augusta Raurica-workshop produced the glass cakes, do we then have to regard it as a larger, more specialized workshop, or is it an example of a local category? When taking into account the weight of a tessera (c.1,5 grams) and the estimated weight of a glass cake is above 2 kilogram (Marii, Rehren 2009, 296), a glass cake can provide more than 1000 tesserae, even considering the loss by generating waste of flakes. When taking into account the size of a tessera c.4-5 mm side for the smallsized tesserae or 8-10 mm for the medium-sized tesserae, one average glass cake produces enough tesserae to cover about 0,3-0,5 m². The small-sized pieces are used for various details, whereas the medium-sized pieces are used to fill in the main parts. 6.3.4 Counters The counters in black glass appear, like all other glass counters, to be a ubiquitous commodity throughout the th whole Roman Empire from the Augustan period to the 5 century AD. However, they are assumed more numerous in the western provinces and especially in modern-day France, Germany and Britain (Catalogue Roanne 1987, 103). Focusing on solely the black glass counters of this area, a difference between the continent and Britannia is noticeable, even though an exhaustive study on the subject is necessary to provide a more accurate picture. A first survey revealed that the medium- and large-sized counters were mainly consumed on the continent, whereas the small pieces are idiosyncratic to the British material. The medium-sized counters (diameter between 20 and 27 mm) seem absent on the British Isles, and the occurrence of large pieces is negligible in contrast to the small ones (Cosyns forthcoming). In Chapter 4, it is demonstrated that the survey on datable black glass counters did not add much new information to refine the discussion on the chronological differentiation based on size variations. Conversely, a geographical differentiation seems more likely. Small counters occur on a wide variety of sites in Britain throughout the Roman st nd period, and those in opaque white and black glass appear especially on military sites of the 1 and 2 century AD (Price 1985, 214). The counters recorded from the fortress of Usk (UK) are a good example (Price 1995, 129). The 87 pieces are all small-sized, with a diameter ranging from 12,5 to 20,1 mm. Conversely, small counters are also present on the continent but only sporadically, such as examples from Tienen (BE) (Cosyns et al. 2006b, 33, fig.1b), Penteville (BE) (Plumier 1986, 82-83, fig.56-31) and Franchimont (BE) (unpublished). The large discoid counters common on the continent are then again unusual in Britannia. This survey demonstrated the 263 corroboration with the overall archaeological record indicating a regional difference between the continent and Great Britain. With regard to the distribution of monochrome glass counters, we focussed on the provenance but did not ignore the material from the consecutive periods (Table 131). From the 96 catalogued pieces from Magdalensberg, we 221 were able to find out the shape of only two counters. We did not verify whether the distribution of the different sizessmall, medium, large and shaperound, oval, irregular includes regional and or chronological differences because we only collected information on black glass counters. Only a general but detailed synthesis including all glass counters in any colourmonochrome, bichrome and polychromecan demonstrate the intrinsic value of this fairly neglected commodity type. Table 131: List of monochrome black glass counters (n = quantity) cat.no. site n northwestern provinces 175 A4486 211 225;226 228 282 290 317 318-19 336-351; 2109 461 465-66 530-31 576 600 615 999 1025 1110 1219-20 1229 1313-16 1341-44 1370-71 1639 1642 1679-81 1795 1851-56 1859-65 1870-71 1893-94 1904-5 1935-37 1941 1950-56 1970-80 2002-2010 2021 2112-27 2877 2884-90 2908-61;4466 2989 2990 2993-95 3223-3316 3637-41 3925-56 3965 3987 3989-99 4007-10 4032-39 221 Anthée (BE) Berlingen (BE) Braives (BE) Burdinne (BE) Cortil-Noirmont (BE) Gembloux-Penteville (BE) Grobbendonk (BE) Herstal (BE) Hodeige (BE) Liberchies (BE) Nivelles (BE) Oudenburg (BE) Tienen (BE) Treignes (BE) Vervoz (BE) Vorsen (BE) Alba-la-Romaine (FR) Aulnay (FR) Frénouville (FR) Loubers (FR) Mathay (FR) Olbia (FR) Poitiers (FR) Ruscino (FR) Saint-Pathus (FR) Sanxay (FR) unprovenanced (FR) Bonn (DE) unprovenanced (DE) Duisburg-Rheinhausen (DE) Epfach (DE) Kempten (DE) Krefeld-Gellep (DE) Ladenburg (DE) Mainz (DE) Moers-Asberg (DE) Neuss (DE) Rottweill (DE) Trier (DE) Xanten (DE) Esch (NL) Heerlen (NL) Nijmegen (NL) Utrecht (NL) Valkenburg (NL) Vechten (NL) Augst (CH) Kaiseraugst (CH) Caerleon (UK) Carmarthen (UK) Chichester (UK) Cirencester (UK) Colchester (UK) Corbridge (UK) 1 18 1 2 1 1 1 1 2 14 1 2 2 1 1 1 1 1 1 2 1 4 4 2 1 1 3 1 6 7 2 2 2 3 1 7 11 9 1 16 1x8 7 39 1 1 3 94 5 32 1 1 10 4 8 These two could be studied more in detail during the ROCT 2005 workshop in Leuven (my gratitude goes to Eleni Kaudelka-Schindler and Barbara Czurda-Ruth). All other material published by Barbara Czurda-Ruth in consecutive contributions on the glass from Magdalensberg received only marginal interest because she focused on vessels and only mentioned the glass hue and the quantity of counters (Czurda-Ruth 1979, nos.2167-2228; Czurda-Ruth 1998, nos.2286-2292). This is the same for more recent material we forgot to insert in the initial database and is included in the addendum catalogue (Czurda-Ruth 2004, 287, nos. 378-402). 264 4059-60 4062;4079-86 4092-4112 4171 4180-81 4196-200 4207 4220-21 4234 4239 4241-43 4246-90 4292 4316-20 4244 Gorhambury (UK) Winchester (UK) London (UK) Ospringe (UK) Richborough (UK) Wilcote (UK) Silchester (UK) South Shields (UK) Templeborough (UK) Tisinghurst (UK) Uley (UK) Usk (UK) Wanborough (UK) York (UK) unprovenanced (UK) 2 2 21 1 2 5 1 2 1 1 3 45 1 5 1 northeastern provinces 35-105 3111-12 3151 3155-64 Magdalensberg (AT) Celje (SL) Izola (SL) Ribnica (SL) 96 2 1 10 west Mediterranean provinces 3185 2589-95 3000-1 3087 Mérida (ES) Aquileia (IT) Braga (PT) Conimbriga (PT) 1 4 2 1 east Meditteraneanprovinces 876-90 Karanis (EG) 15 Glass counters showing a dotted decoration on the top surface with spots in opaque red, opaque white, opaque th ultramarine blue, translucent turquoise blue or combinations of these are idiosyncratic to the 4 century AD and thus overlapping parts of Period III and Period IV [see Chapter 4]. Up to now, we were able to record black glass counters with dotted decoration from 17 sites within Belgium, Netherlands, France, Germany, England and the north of Italy (Table 132). Sets of 10 to 15 black glass dotted counters have been retrieved from burials in Amiens, Krefeld-Gellep, Nijmegen, Winchester-Lankhills and Lullingstone. All other places yielded only 1 or 2 pieces. Table 132: List of bichrome and polychrome dotted black glass counters cat.no. site quantity shape 277 343; 345; 351 1001;1002 1317 1677 1903 2590-92 2937-41; 2943-44; 2951; 2955-61 3995 4152 4180 4220 4243 4199 4299 4079;4081-86 4308 4244 hue of dotted decoration Franchimont (BE) Liberchies (BE) Amiens (FR) Olbia (FR) Strasbourg (FR) Krefeld-Gellep (DE) Aquileia (IT) Nijmegen (NL) 1 3 15 1 1 10 3 15 small - round small - round medium - round ? ? small - round small - round medium - round white white; blue/red red; blue/red white red blue-green/red blue; white; blue/red blue/red Cirencester (UK) Lullingstone (UK) Richborough (UK) South Shields (UK) Uley (UK) Wilcote-Shakenoak (UK) Water Newton (UK) Winchester-Lankhills (UK) Woodcutts (UK) unprovenanced (UK) 1 15 1 1 1 1 1 1; 14 1 1 medium - round medium - round medium - round small - round small - oval medium - oval medium - round medium - round ? small - round white/red white/blue ? blue/red blue blue/red white/red blue; blue/red ? white/red 265 Figure 122: Distribution map of late Roman dotted counters A distribution map of these typical late Roman glass counters demonstrates a distribution mainly in the northwestern provinces, i.e. Britannia, Gallia Belgica and Germania Inferior (Figure 122). This restricted area of consumption is evidence for the production of dotted counters within these provinces. A more thorough investigation of the material is necessary to point out whether all material can be reduced to one single workshop or reflects the production of various workshops. Insofar that the here-presented inventory is representative, it is reasonable to assume one single workshop responsible for the production of the decorated glass counters and that this workshop was located somewhere in south-east Britannia, in the lower Rhine region or somewhere in the north of Gallia Belgica. No such counters have been retrieved at the two most suitable workshops, Les Houis near Sainte Menehould and Trier Palais Kesselstatt. Only Liberchies, where a workshop is assumed to have produced black glass artefacts, yielded such counters, but no indications can argue for such production somewhere at the vicus Geminiacum or at the castellum of Liberchies. The pieces of Olbia and Aquileia most likely are the products of another workshop that was situated in the Mediterranean. Considering the basic production technique of plano-convex glass counters [see Chapter 2], it is in a sense contradictory to produce misshapen counters. This causes problems in defining a workshop where glass counters were fabricated, not in the least because the less well-manufactured irregular counters also got into circulation. Consequently, it is self-evident that no or nearly no such material was discarded as waste at a glass workshop producing glass counters, which diminishes considerably the availability of evidence to prove the making of counters. 6.4 Conclusion One of the aims within this chapter was to define the distribution pattern of various deeply coloured glasses appearing black. The evaluation of the glass workshops where black glass was processed and the distribution patterns of the different functional types in black glass enabled us to get an idea on how its production was organized and to what level the different black glass artefact types were consumed. The study of the black glass artefacts, although being a minor commodity within the Roman glass production and consumption, has provided evidence to speak of a production of artefacts with regional character and a distribution on an interregional level (Table 133). Conversely, no archaeological evidence was found to locate workshops where glass was made black, whereas the number of secondary workshops where black glass was processed also remains very little. Within the archaeometric approach various recipes used to produce black glass in the Roman Empire during the imperial period are discussed [see Chapter 8] as well as their physical properties [see Chapter 9] indicative of the generally accepted model that the secondary workshops were supplied by large primary workshops in the south266 eastern Mediterranean (Nenna 2008a), either with black coloured raw glass or with ‗naturally coloured‘ and decolorized raw glass that got coloured by an intermediate workshop or a secondary workshop. Table 133: Overview of the distribution of black glass vessels during the consecutive periods period distribution Period I (1st–mid 2nd century AD) Period II (mid 2nd – mid 3rd century AD) Period III (mid 3rd–mid 4th century AD) Period IV (mid 4th–5th century AD) entire empire north-western provinces (largely restricted to the provinces Gallia Belgica, Germania inferior and Germania superior) south-eastern provinces (mainly Levant and Egypt)  PERIOD I During Period I the general distribution of black glass vessels covers the entire Roman Empire, though seeing the examples at hand only in marginal quantities and foremost in the west. It is not clear whether these black and other strongly coloured glass vessels were the result of a production in one or a few workshops in the south-east Mediterranean where also the primary workshops were located and traded all through the Roman empire. The glass workshops of, for instance, Lyon-La Manutention and Avenches-Derrière la Tour do not provide clear evidence of black glass vessels, but the incidence of a vessel production in strongly coloured glass makes such assemblages very probable in both workshops. It remains thus far unclear whether every assemblage was idiosyncratic or unvaryingly imitating each other. Probably a difference is to be made between the cast and the blown vessels. The concentration in the north of Italy of free-blown vessels in deep purple glass appearing black and the evidence of blown vessels in strongly coloured glass in the workshop of Avenches-Derrière la Tour might indicate that blown vessels in strongly coloured glass were locally produced in the west, whereas cast vessels were imported from the east. The widespread distribution and consumption of standardized vessel shapes in black glass, like in any other strongly coloured glass, monochrome and polychrome, must however be considered as a sort of veneer of st Romanisation featuring the material culture during the earlier 1 century AD up to the Neronian times. The extremely standardized-appearing material culture in the entire Roman Empire starts to change from the Flavian period onwards, influenced by a changing supply and demand due to the harmonisation of the material culture in the entire empire. The success of the romanitas policy under Julio-Claudian rule made long-distance supply fail and unprofitable. Perhaps delocalisation of production centres all over the empire was necessary and more profitable, creating opportunities for glass production and distribution on interregional, regional and finally local levels. While an interregional level of glass production in very large centres was clearly introduced in the first half of Period I, a regional level got established in the second half of Period I, what resulted in regional idiosyncrasies.  PERIOD II It is thus not surprising that the glass assemblages in circulation during Period II seem to have a regional character. Limited to Gallia Belgica, Germania Inferior and Germania Superior, some examples do occur in the neighbouring areas, such as the southeast of Britannia, Gallia Lugdunensis and even further in Pannonia, probably by means of the Rhine and Danube. It has been attested that the glass workshop ‗Äussere Reben‘ in Kaiseraugst (CH) (Germania Superior) produced black glass vessels in Period II. But seeing that mainly globular cups (Form IIB.4-5) are retrieved in Augst/Kaiseraugst and the black glass carchesium (Form IIB.1) is barely present in the entire province Germania Superior while very popular in Gallia Belgica and Germania Inferior, other workshops are likely to have produced contemporaneously black glass vessels in Gallia Belgica and/or Germania Inferior. The workshop at Les Houis near Sainte Menehould (Gallia Belgica) is almost certainly another producer of black glass vessels. nd rd The consumption of black glass vessels during the late 2 and early 3 century AD is widespread within the borders of Gallia Belgica, Germania Inferior and Germania Superior, with some minor occurrence in the neighbouring areas (e.g., the southeast of Britannia and Gallia Lugdunensis) or even further away (e.g., Pannonia). The limited area of distribution of black glass vessels within the boundaries of the north-western provinces stands in sharp contrast with the nearly total absence outside these provinces. This regional consumer behaviour stands in great contrast with the general distribution and consumption of black glass vessels in Period I all over the entire Roman Empire, though in marginal quantities. This consumption pattern can be brought in relation to a regional and interregional echelon of Romanisation, or in other words to a regionalism or interregionalism within the Roman society involving cultural, social, economical and political aspects. Other recent research such as that of Wim De Clercq (2009) on house patterns and pottery consumption within the northern part of the civitas menapiorum in Gallia Belgica provides evidence for a local level of romanitas or ‗Roman-ness‘. 267 We are convinced of a Roman Empire built up in various levels. No one doubts the fact that the Empire does not emanate a homogeneous (material) culture but is built up out of a melting pot of numerous cultures, entities and ethnicities. The assimilation degree of genuine ‗Roman-ness‘ and the speed of its absorption is determined by local, regional and inter-regional elements within the societies ruled by Rome. Consequently, we see a civitas as the local level, a provincia, comprising several civitates, with the regional level and neighbouring provinciae as the inter-regional level. Features on the local level (i.e., micro level) were built up from below and were affected by tribal traditions involving religion, habits, social organization in families and within small communities. Features on the inter-regional level (i.e., macro level) conversely were imposed from above and came from Rome. These enforced political ramifications caused repercussions on the social and economic organisation of the local people within one or several neighbouring provinces. The regional level (i.e., meso level) was influenced by the interaction between the local socio-cultural traditions and the centrally steered politico-economic Roman administration. If we can see artefactual differences on the micro level within one region (i.e., province), this suggests a difference in the degree of Romanisation within one region. For instance, the civitas menapiorum in present-day western Belgium, south-west Netherlands, and north-west France, largely the area west from the River Scheldt, forms together with the civitates nerviorum, morini and remagii the province Gallia Belgica. Although in the same region with similar people and normally with a similar tax regime, there is a strong difference of Romanisation visible between, for instance, the Menapii and the Remii. The black glass counters are markedly present on Roman sites, and it has become plainly clear that jewellery like bracelets, finger rings and even beads were also main products on the Roman market. But whereas the former is rd rather featured in the early imperial period up to the early 3 century AD, the latter was more typical for the late rd th Roman period in the later 3 to 5 century AD. Identical bracelets have been attested in, for instance, Belgium, Cyprus, Egypt, France, Germany, Israel, Italy, the Netherlands, Portugal, Switzerland, Turkey.  PERIOD III Within this period, the distribution of black glass vessels disappears. Most material in this glass hue is jewellery, rd even if architectural decoration material and counters remain in production as well. During the 3 century AD, there was a markedly standardized execution of the various commodity categories in black glass. The black glass material is therefore a striking example to demonstrate the globalising Roman material culture in the late Roman period.  PERIOD IV th th The present knowledge on workshops producing black glass artefacts during the second half of the 4 and 5 century AD remains rather limited. Up to now, only the sites of Trier-Palais Kesselstatt (DE), Braga-Fujacal (PT) and Les Houis near Sainte Menehould (FR) have demonstrated the presence of secondary glass workshops to have produced jewellery in black glass during this last Roman period. It is obvious that more workshops in the western half of the Roman Empire must have been responsible for the production of artefacts in black glass, and that many more are still to be recognized in the eastern half, such as Ore ac (SL) and Horvat Qastra (IL). The existence of secondary glass workshops producing black glass artefacts in the eastern Mediterranean is apparent from the distribution of the black glass vessels. No glass workshop has been excavated yet where such material has been produced, but it is very probable that a workshop has been active in or near Haraun (SY). The production of black glass vessels in Period IV is limited to rod-formed shapes and shows a consumption pattern on a regional level within the boundaries of the Roman provinces Aegyptus, Arabia, Iudea and Syria. The regional character of its distribution is evidence to locate at least one workshop in Egypt and/or the Levant producing such type of vessels, although it is more probable that many more workshops produced these idiosyncratic glass vessels. Considering the adopted rod-formed technique, which is specific for the production of all kinds of jewellery, we believe that the jewellery-producing workshops were also responsible for making the rodformed vessels [see Chapter 7]. Possibly the production of rod-formed vessels in black glass was trend-related to a regional consumer market, but it can equally be that the mass production/mass consumption of glass artefacts th within the south-eastern Mediterranean, where primary glass production originated, was at the end of the 4 th century AD and in particular during the 5 century AD, when there was a market demanding for such type of commodity. Future research is compulsory to verify whether the specific shapes and the different glass hues used for the trailed decoration are related to a regional or a chronological production. The distribution of jewellery in general is less well definable due to its widespread occurrence. When examining the distribution maps of the different jewellery commodities separately, a regional differentiation can be noticede.g., arm rings. We demonstrated previously that the arm rings characteristic for Period IV are of Type D [see Chapter 4]. Where in the north-western provinces almost solely plain narrow D-shaped bangles (Type D1) circulated, a larger variety of bangle types occur in the Mediterranean until Switzerland, southern France and 268 northern Italy. The Type D-bangles are even more frequent in the north-eastern European (Balkan) provinces and in particular in the (eastern) Mediterranean. Seeing the different techniques used (gob-gathered and cone-rolled in the north-western provinces and rod-made in the east Mediterranean provinces), it is clear that the pieces from the north-western provinces were locally produced and not imported as finished products from the eastern Mediterranean. The workshop of Trier-Palais Kesselstatt known so far, that produced beads and pendants but possibly also bangles or other jewellery, makes a comparative evaluation impossible. 269 270 Chapter 7 USE AND FUNCTION In 1867 Karl Marx was surprised to learn that five hundred different kinds of hammers were produced in Birmingham, England, each one adapted to a specific function in industry or the crafts. A hammer is thus not merely a hammer. Depending on who will use it and for what purpose, the shape can differ. ... why are there so many different kinds of things? (Basalla 1988, 2) Colour operates in every culture as a symbol, a metaphor, and as part of an aesthetic system (Jones, McGregor 2002, cover page) 7.1 Introduction Within this chapter on use and function, we attempt to improve our understanding of how black glass artefacts were used in Roman times by checking the correlation between the diversity of artefact types and their possible purposes. However, defining the connotation (meaning) an artefact received in the past is much more complex compared to identifying its denotation (function). Hence, the use and function of the various commodity types in black glass can only be given to a certain degree, if already a purpose can be ascribed to the artefacts from the available archaeological data. We are aware that this query is bringing us far beyond the basic discussions in archaeological glass studies, i.e., describing the finds, linking them with the excavated contexts, and dating the artefacts. Material culture studies, however, should not remain restricted to solely such general outcome but should try to include an assessment of the use and function (and if possible even the meaning) of the concerned artefacts to help better reconstruct and understand the regionalism of the consumption pattern within the ‗globalized‘ Roman material culture as demonstrated in the previous chapters. In other words, it is important to identify the regional and local cultural identities within the ‗globalized‘ multicultural Roman society which are influenced by individual traditions and drives. Nonetheless, we have to be aware that the consumer is an individual that got unwittingly influenced by cultural and societal issues, which are of great importance in the study of consumer behaviour (Miller 1991). The here-presented assessment wants to surpass the limited use of colour to help classify the wide variety of artefacts by including information from Chapter 5 with the aim to gain greater insight of the lifestyle and material culture during the Roman imperial period. Besides the discussed causes and the summarized socio-cultural and socio-economical motives which can influence the fabrication of consumer goods in black-appearing glass, the reason to make a specific shape in a specific material in a particular hue are indicative for an approach based on skeuomorphism. Due to practical reasons, this approach remained minor. Nonetheless, this approach provides new facts to help better understand certain issues of the Roman material culture. When tackling the issue of the use and function the Romans attributed to artefacts made in black-appearing glass, the question that immediately arises is what drove the Romans to produce glass artefacts in such a hue. In Chapter 1, we described that the coloration purpose in Antiquity is much more complex than presently accredited and goes beyond solely aesthetic motivations. This logical assumption corroborates the artefactual data discussed in previous chapters, showing an interest for commodities in black-appearing, deeply coloured glass st that are very much products of their times. Regarding the 1 century AD, for instance, it is clear that the blackappearing vessels are to be understood within the fashion alteration from polychrome to monochrome deeply coloured glass vessels. On the other hand, it is interesting to take into account that most black glass jewellery is currently appreciated as being very stylish and tasteful. But do we have to exclude beforehand all aesthetic motivation by the Romans and go for a purely functional meaning? With regard to the vessels, it is obvious that the opacity caused by the black appearance of the glass increases 222 the significant advantage of preserving sensitive products from decay caused by UV-radiation from sunlight. As described in Chapter 1, colours in general can also bear meaningful and sometimes powerful connotations, particularly in the worlds of religion, tradition and superstition. The real challenge thus is attempting to unravel the thought behind the consumers‘ behaviour and motivation in the acquisition of black glass artefacts or specific goods in general. What influenced the Romans to acquire black- 222 Recent biotechnological research in beer production has demonstrated that a cheaper but less UV-resistant glass for the production of beer bottles would give the opportunity to sunlight to produce toxic bacteria and make the beer unfit for drinking. 271 appearing glass vessels or jewellery? Is it related to gender, age, and ethnicity? Or could it have had a symbolic meaning (e.g., a religious symbol or a social status symbol)? This consumer research on black glass artefacts provides additional information concerning the use and function of black glass material. Of course it cannot be the purpose to verify what it meant to own black glass artefacts because then one could easily lapse into mistakes in the sense of Aristotelan physiognomica (‗you are how you look like‘). However, it is a widely observable fact that there always are (have been) ostentatiously materialistic people within every society level having the underlying idea of ‗you are what you have‘ transformed into ‗you are what you wear‘, leaving aside whether they do (did) it deliberately or unconsciously. For types of jewellery, it does not necessarily have to do with merely beautification. Jewellery can be given a religious connotation or a symbolic meaning believed to hold superstitious power. For practical reasons, this chapter does not discuss all categories of artefact types in black glass. We discarded material such as the mould-blown vessels because these categories do not add any value to the here-described discussion. 7.2 Reflections on the use and meaning of colours in Roman society We discussed in Chapter 1 that Beat Rütti discerned 76 colour variations to describe the Roman glass vessels from Augst/Kaiseraugst (CH) and that he grouped them into seven colour sets: blue, blue-green, green, yellow (+ 223 orange), brown, purple (+ pink). and white (Rütti 1991, 110). But did it all matter to the Romans? By contrast, in Antiquity great importance was attached to the simplex color to designate the basic colours used by the 224 tetrachromists, ancient painters working only with white, black, yellow and red (Brécoulaki 2006). When taking into consideration Roman pottery, black and red are the two dominant colours, but pale yellowish to whitish pottery is also popular. Besides pottery, many objects in the non-ceramic material such as glass and metal demonstrate a similar set of colour variations easily to classify into these four major colour groups: silver-white and black, gold-yellow, bronze-black and green, and brass-yellow and red. st Within this perspective, it is easy to reduce the true hue of each of the burial gifts from the 1 century AD tomb at Antran in the south of France (Pautreau (ed.) 1999) and of those from the contemporaneous tomb 1 at Montebelluna in the north of Italy (Casagrande, Ceselin 2003, pl. X) into the four major colour tonalities so imperative in Roman times (Figure 123). These examples support the assumption that in Antiquity, more meaning was given to the brightness of the hue and to its saturation or its intensity and much lesser interest to the hue or chroma. Within the modern colour perception, we can conclude that the Romans rather differentiated colours as vivid colours, pale colours (white) and dull colours (black). It is not excluded that each of the four groups received connotations associating a colour with a deity, one of the four elements,..., and not solely to self-evident functional- or material-related reasons. This means, for instance, that the success of monochrome deeply st coloured glass vessels during the middle of the 1 century AD and its decline during the Flavian period in favour of colourless glass was not merely a change in fashion but was most likely the result of socio-cultural evolution instigated by the policy of the emperor. Within that view, it is believable that under Flavian rule the political condemnation of everything what was directly and indirectly linked to the rule of Nero influenced the success of colourless glass and the decline of the vessels in deeply coloured glass. 223 The only relevance in applying such detailed subdivisions at present is because the hue of the glass is related to the chemical composition [see Chapter 8]. 224 Pliny the Elder mentions lime from Melos (white), ocres from Attica (yellow), earth from Sinope (red) and atramentum (black) (Naturalis Historia XXXV, 50). 272 Figure 123: The burial gifts from 1st century AD tombs at Antran (FR) (left) and Montebelluna (IT) (right) (pictures taken from Pautreau (ed.) 1999, fig.28 (left); Casagrande, Ceselin 2003, pl. X (right)) 7.3 The vessels The small quantities of vessels produced in black-appearing glass in the first place demonstrate a rather marginal production. But in relation to the large set of forms and shapes in black and other hues, its production responded to a consumer market in need of a wide variety of goods. Conversely, these productions fit the contemporary fashionable glass assemblages. Seeing the number of shapes produced by means of the free-blowing technique, of which only one piece is known in (black) glass, we may wonder if these objects are not representing the 225 specialized design of (a) glass artist(s) creating individual artefacts. This rather individual production contrasts sharply with the common bulk productions and establishes idiosyncrasies of a region and period. In Chapters 3 and 4 we demonstrated that the black glass vessels were produced within all four periods established. When tackling the issue of use and function of the vessel shapes, it is necessary to explain what we understand as tableware, toilet ware and storage ware. We identify tableware with all vessel shapes used for 226 drinking (cups, beakers, bowls), pouring/serving drinks (flasks, jugs, bottles ) and eating or serving food (plates, bowls, dishes). The toilet ware is nearly restricted to the unguentaria or balsamaria but includes also aryballoi. The storage ware includes all sorts of jars but also the mould-blown square bottles and the so-called ink-wells are here considered as vessels used for storage. Further discussion is considered in general per technique and additionally some specific shapes have been examined in detail to verify to what degree we can attribute the functionality and/or meaning of an artefact from archaeological data. The mould-blown vessels are omitted from discussion, as the black glass alternative of the square bottles and grape-shaped amphoriskoi does not provide any additional information on the use and function of these vessel shapes. A general labelling of functionse.g., tableware, toilet ware, storage wareis given on the basis of shape following the conventional model in glass studies. But concerning, for instance, the tableware, there is not always a distinct differentiation possible. It is very obvious to consider jugs, flasks and bottles as pouring vessels to serve drinks, but it is less clear what should be attributed to some cups, bowls and jars (e.g., drinking vessels, eating vessels or vessels for serving food). The here-used functional attributions are based on the prevailing interpretations, but we are aware of the multifunctionality one single shape can imply or the possible erroneous connotations assumed. It was, however, not the purpose here to list the functions given to a number of specific vessel shapes or groups of vessels that have received through history a functional name in Latin or Greek such as the small bilobated cup (Form IA.2) which is generally described as acetabulum. Some remained very simple (e.g., olla), but the name is mainly based on the 225 226 This hypothesis is lent better support from the number of individual creations of the snake-thread-decorated vessels. Bottles can also be regarded as storage vessels because their impermeability allows them to stock liquids easily. 273 correspondence between the external properties of a vessel and the descriptions in literary sources, such as carchesium, modiolus, or skyphos (Isings 1957, 50; 55), or its resemblance to similar shapes in other materials and of which the functionality was already attributed, such as amphoriskos or aryballos. Some designations include, however, a certain functionality (e.g., acetabulum or unguentarium). 7.3.1 Cast vessels Nearly all cast vessels in black-appearing glass are tableware. When referring to the 13 acknowledged shapes, it becomes clear that the casting technique was adopted to manufacture open vessel shapes such as cups, bowls and dishes (Table 134). The cups can be considered as drinking vessels, whereas the bowls can be attributed a multiple use for eating and/or drinking. The remaining shapes (e.g., dishes, plates and tables) concern solely food consumption purposes. Table 134: List of cast shapes per function (numbers coincide with Form IA shapes discussed in Chapter 3) cups beakers or goblets bowls jars flasks jugs bottles dishes plates tripod tables DRINKING POURING DRINKS EATING SERVING FOOD 2; 6 1A; 3; 9 - 9* - 1A-B; 10; 13 4A-B; 11; 12 - 1A-B; 5;10; 13 4A-B; 11; 12 7 8 An exception is the boat-shaped vessel (Form 9), which in all probability has been used as storage ware. It has been reported that a Pompeian piece contained jewellery at the time of its discovery, which makes the vessel a ‗decorative container on women‘s dressing-tables‘ (Glass of the Caesars no. 24). This idea corroborates the presence of a boat-shaped vessel in female tombs such as the black example from the ‗tomba di Silea‘ at S. Elena di Melma (IT) [cat.no.2805] or the deep blue one from the rich tomb at Sankt Aldegund (DE). 7.3.2 Free-blown vessels The free-blown vessels produced in black glass are essentially tableware. For the benefit of readability, we choose to focus on one distinct vessel shape of Period II in order to demonstrate the potential outcome of a more in-depth discussion on use and function. We choose to consider the carinated beaker or carchesium (Form IIB.1) because this vessel shape enjoyed a certain success in relation to the entire glass assemblage of the north-western provinces. The relatively popularity of carinated beakers in black glass in only the provinces of Gallia Belgica and Germania Inferior during Period II as described in the previous chapter most likely was a vessel shape in relation to a particular consumer‘s behaviour. Still it remains unclear whether it was designed purely for the consumption of a specific drink, or whether it was used on special if not ritual occasions within a regional socio-cultural tradition. To our knowledge, there is no evidence available to explain the use of Form IIB.1 beakers, but the continuity of this characteristic shape, from Period I (ceramics) over Period II (glass) up to Period III (silver), implies a continuity of a particular drinking custom. Hence, we assume that the carchesium was designed for a specific use that was characteristic in the provinces Gallia Belgica and Germania Inferior [see Chapter 6]. The creation of a vessel with pronounced carination at the lower part of the body just above the bottom is not coincidental or simply inspired by aesthetical design but rather applied on purpose. This means that the carinated beakers were designed to retain something not meant to be swallowed when drinking, such as a residue that is inevitably connected to the preparation of a type of drink, whether it be celebratory, ceremonial or medicinal. The retention of a residue is most likely the function that should be attributed to the similarly carinated small jug from Schaarbeek (BE) [cat.no.515]. Cups with bulbous bodies (Forms IIB.4-6) provide the consumer the possibility to empty the cup completely until the last drop. These cups, however, might have been used for the same drinking custom as the carinated beakers when the drink was first prepared in a carinated jug; likewise, the Schaarbeek jug, seeing that the cup within the tomb is a metallescent cup with bulbous body of the type Niederbieber 31 (Hanut 1999, 8, fig.3:2). Interesting in 274 this issue is the intact burial content of tomb 176 at Cutry (FR) (A. Liéger 1997, 148 pl. 28) yielding three carchesia in black glass [cat.no.1049-1051] together with two small jugs with globular body (Form IIB.3), one of which is also in black glass [cat.no.1052] (Figure 124). Seeing that the entire content can easily flow out of hemispherical vessels, the Form IIB.4-6 vessels are designed to hold a liquid that is to be drunk in its totality. The liquid needed therefore to be filtered or decanted, a function that can be assigned to the carinated jug from the Schaarbeek tomb. Conversely, the carchesia were most likely conceived to receive an unfiltered drink. It is not excluded that the same drink was consumed in the carchesium and the bulbous cup. If so, the drink was either prepared in a carinated jug for the bulbous cup or prepared in a bulbous jug for the carinated beaker or in the carinated beaker itself. Figure 124: Overview of burial gifts of tomb 176 from the cemetery at Cutry (illustration taken from Liéger 1997, pl. 28) Only a thorough study of globular and carinated vesselsserving/pouring versus drinkingand their possibly combined occurrence can provide additional information on this issue. Yet, some pieces may contain indications giving an idea about its use, like for instance the base fragment of a carchesium from Liberchies-Bon Villers, Belgium [cat.no.428] or the base fragment from Matagne-la-Petite (BE) [cat.no.438]. The scratches limited to a particular area of the inner bottom of the carinated vessel cannot be caused from the impact of its stay in the soil, in particular seeing that the carinated part of the bottom show no scratches at all. The explanation must therefore be that the beaker was employed to prepare something by using a sharp and hard utensil to stir and beat something, thus causing the scratches (Figure 125). Conversely, we noticed plenty of these carinated beakers where these specific marks are missing. It is, of course, evident that these carinated beakers could have had a versatile use depending on the purpose given by its consumer. However, scratches can be avoided when stirring and beating ingredients with similar utensils in a softer material. Secondly, the scratches may be caused during the manufacturing process by a poorly skilled person, an apprentice for instance, when shaping the vessel from the inside with a jack. 275 Figure 125: (left) Base fragment from Liberchies-Bon Villers (BE); (right) from Matagne-la-Petite (BE) showing marks of use (photographs by author, by courtesy of the Musée Archéologique, Nivelles and VIOE) Additional information on the use and function of carinated vessels can be obtained from the carinated jug with high narrow neck and high vertical handle (Form IB.14). These so-called lagynoi have counterparts in pottery and metal ware. The lagynos is a vessel shape that was designed as metal ware and was copied in pottery and glass (Isings 1957, 17); accordingly, the lagynoi in both materials imitate this specific shape in metal ware (Lierke 1999, 61-63, figs. 150-151). The lagynoi in glass imitate very closely the examples in more traditional materials, showing features without any sense or use but in metal ware. The imitation of a production process atypical to a material and only executed to improve the inclination of the consumer to purchase a vessel in another (novel) material for traditional purposes and occasions is an example of skeuomorphism, demonstrating traditionalism in shaping, and the continuity of a specific shape into different materials for a specific use characterizes cultural conservatism. That is the reason why some features with a genuine link to the metal ware are reduced to a corsage on copies in pottery and glass in order to increase the recognisability with the original vessels. From this viewpoint, the origin of the carchesium (Form IIB.1) is to be considered in metal ware with counterparts in pottery and glass. The black shiny aspect of the black glass and black burnished ware probably stands in relation to silverware in angular folded silver sheet. The original production of the lagynosincluding a local variant in black glazed waregoes back to the early Hellenistic period but are very common throughout the entire Hellenistic period (Rotroff 1997, 127; 226-229), st whereas those produced in monochrome deeply coloured glass are dated around the middle of the 1 century AD. In Flavian-early Antonine times, the production of glass lagynoi knew a last revival by means of the free-blown globular and conical bottles with tall necks and single vertical handles, corresponding to Isings Forms 52 and 55. We believe that the reason to deliberately copy such very particular ‗antique‘ shapes in pottery and metal in glass workshops during the Julio-Claudian era is related to the function of this specific vessel shape: lagynoi are considered as a typical vessel at symposia to decant wine: In ancient sources the word appears from the 4th century onward to designate a wine jug distinguished by its long, thin neck and single handle and by the gurgling sound it made when the wine was poured. It was used at symposia, where the drinker often brought it along as his contribution to the feast. (Susan Rotroff 1997, 226) It is thus imaginable that the Form IB.14 jugs were produced in free-blown strongly coloured glass during Claudian-Neronian times to supply vessels for symposia. This type of social activity was primarily to celebrate the introduction of young men into aristocratic societyi.e., a sort of social initiation. To come back to the carchesia, we are inclined to assume from the aforementioned information on lagynoi that carinated beakers were used to drink a wine still containing residues and that the carinated jug of Schaarbeek functioned as a decanter so that globular cups were usable to drink the same wine. But why it had to be black in Gallia Belgica, Germania Inferior and Germania Superior remains unsolved. Seeing the frequent appearance in funeral contexts, it is plausible that there is a connection to a ritual/religious meaning. When taking into consideration a deity connected with wine and death/resurrection, there is a diversity of possibilities: a chtonic deity; the Egyptian god Osiris; Dionysus; Dis Pater; Liber; etcetera. 276 7.3.3 Rod-formed vessels This category of vessels has been systematically catalogued as toilet ware. In the first place the Period II candlestick unguentaria with a triangular solid body (Form IIID.1) which correspond in shape with the thin-walled candlestick unguentaria (Isings Form 82) and also look similar to the blown toilet bottles with a solid base [see Chapter 3]. Jean-Pierre Brun (2003, 385) rejects Carol Meyers‘ idea (1992, 30) that these solid bodied vessels with very small capacity were perfume containers to mislead the purchaser. He demonstrated through finds from Roman desert fortresses in the Egyptian desert that the latter type was used to hold kohl (Brun 2003, 383-385, figs.6:8; 7). The pieces from Elkab (EG), which form hitherto the sole basis to incorporate this shape into our classification, are somehow unlike their blown counterparts. The shape of the four pieces from Elkab allows us to identify them as vessels with a hollow tubular neck and a horizontally folded rim, but they can hardly be used as containers. The very small capacity of the content is disproportionate to the mass of glass used. They therefore were possibly not used to hold kohl, and likewise the blown toilet bottles with a solid base. On the other hand, the Elkab pieces are very similar to some stemmed pieces from Augst (Rütti 1991, 165-166, figs. 105-106), even though that these stems are solid. Unfortunately, there is no clear assertion of the correct interpretation for the function of these objects either. One hypothesis says they should be seen as slick stones or linen smoothers (Steiger et al. 1977, 330), although others tend to speak about ingots used in glass production (Toma ević-Buck 1980, 249-251). We have no knowledge of any chemical analysis executed on the Augst pieces, but those done on at least the medieval plano-convex ‗linen smoothers‘ provided evidence to refute the idea that they have been used as cakes of raw glass to be processed in the batch. Not only a too-high concentration of impurities has been demonstrated, but also several attested elements in the linen smoothers are not present in manufactured glass artefacts (Gratuze et al. 2003b, 106; Foy 2004, 32). A plausible alternative remaining is that the Augst and Elkab pieces might be considered as Roman forerunners of the medieval plano-convex linen smoothers with rounded edge, with the distinctive difference that the Roman pieces have a stem. However, some finds provide evidence to grant an apotropaic meaning to these glass cakes, as they were built into the façades of houses or in the foundations of walls or thresholds (Foy 2004, 32; Munier 2009, 68). A second group are the Period IV vessels, including two main groups of rod-formed vessels, the slender tubular vessels (Form IVD.1) and the squat jars (Form IVD.2). Although some of these crudely shaped weighty objects have been described as pendants (Whitehouse 2001, no.969) we have to consider them as toilet or medicinal vessels to hold an unguent or make-up. Birgit Schlick-Nolte mentions that both vessel shapes probably formed a set (Schlick-Nolte, 2002, 105, no.V-69). She noticed that many museum collections acquired both vessel shapes together and that they match from various viewpoints. Both shapes are crudely made in a similar technique with comparable decoration. Although this is far from a decisive factor, she proposed different cosmetics content for 227 each vessel shape, i.e. the tall and slender tubular vessels to hold the black eye-paint, kohl , while the broadmouthed squat jars an unguent (Schlick-Nolte 2002, 104-105). Besides the generally assumed use of squat jars as ointment containers, these vessels are presumed also to have contained medicinal preparations (Israeli 2003, 234). An alternative explanation could be that the tubular vessel held one or more applicators for the eye makeup, while the squat jar was used to contain the kohl. An indication to consider these black glass thick-walled vessels as kohl-tubes is the resemblance in shape and decoration with a number of the sometimes very kitschy manufactured free-blown kohl-tubes, in particular the modelling of the handles and the way of decorating the body (Figure 126). These for their part are determined as kohl-tubes on the basis of burial finds in Palestine containing traces of kohl or thin applicators in bronze, bone or glass with one end spatula-shaped and the other rounded (Israeli 2003, 227). Furthermore, the connection with the blown double and quadruple kohl tubes with complex handles reminding basketry (Israeli 2003, 230-233, nos.287-294) demonstrates these vessel types form a very idiosyncratic commodity category for specific uses during a relatively short period of time and within a confined region. Despite the explicit regional character of the consumption of rod-formed vessels in black glass within the Levant and Egypt [see Chapter 6], it is important to notice a much more widespread tradition in the Near-East and in North Africa to paint the eyelashes and eyebrows with kohl. Alternative receptacles in other materialse.g. clay; metal; woodmust have been more fashionable in other regions. 227 Kohl was in Antiquity always made from galena, a lead sulphate (PbS), to use (1) for cosmetic purposes as eyeliner to accentuate the eyebrows and eyelashes, 2) for medical purposes with ophthalmological curative effect, and 3) for divine attributions (Janot, Vezie 1999). 277 Figure 126: (middle) rod-formed kohl-tube; (others) free-blown kohl-tubes (pictures taken from Israeli 2003, 228-229, nos. 283286 except middle Schlick-Nolte 2002, 104, no.V68) The difference in manufacturing technique and utilized glass‗naturally coloured‘ and decolorized glass for freeblown vessels and opaque dark blue and black appearing glass for rod-formed vesselssuggests that two different types of workshops were responsible for the manufacturing of these kohl tubes and jars, one specialized in blowing glass vessels and one where glass-blowing was not executede.g. a workshop specialized in the production of (black) glass jewellery. The ubiquitous occurrence in the Levant and Egypt of both rod-formed and free-blown kohl tubes assumes a widespread manufacturing of kohl tubes and small unguent jars, not so much because they were fashionable consumables but rather because the use of kohl and unguents must have been very fashionable and an affordable cosmetic/medicinal product. A more detailed holistic research on this specific commodity type will improve our view on regional consumption patterns in the late Roman period. 7.4 Jewellery More than in the study of Roman jewellery, the assessment of the use and function of jewellery receives a vivid debate in prehistoric studies where it has been approached as ‗a polysemantic component‘ (Vanhaeren 2005) embracing a multitude of symbolic values (Table 135). Hence, personal ornaments are not solely to be regarded as decorative objects but equally seen as proxies for symbolic awareness and emblematic thinking (McBrearty, Brooks 2000). Sally McBrearty and Alison Brooks describe symbolic behaviour as ‗the ability to represent objects, people and abstract concepts with arbitrary symbols, vocal or visual, and to reify such symbols in cultural practice‘ (McBrearty, Brooks 2000, 457); and consequently to objectify these symbols into artefacts. Although the different functions defined by Marian Vanhaeren (2005, 527-533) are with reference to small-scale, hunter-gatherer societies, most of the proposed connotations are universal: Table 135: List of proposed functional connotations and practical use of jewellery (based on Vanhaeren 2005) Function (= cause) Use (= motivation) Aesthetical expression and self assertion Courtship Ethnic marker Social marker Individual marker Ritual objects Offerings Amulets and talismans Prophylactics Exchange media Inalienable possessions Communication systems Counting devices 278 to beautify the human body to attract members of the opposite sex as an indicator of ethnic identity reflecting the affiliation of an individual to one or more social groups to characterise individuals that have acquired or inherited a unique social status, e.g. a shaman or a king to identify the ceremony leaders and attendants or involving the ‗rites de passage‘ such as birth, initiation, and marriage to attract the goodwill of the worshipped entity or to acknowledge it for a received favour to protect the wearer from particular misfortunes, e.g., illness, death, miscarriage, loss of beloved as a healing item embracing peculiar curative properties as gift within a gifts exchange network to reinforce social ties as a link to the family ancestors or to the group identity to store and transmit information carrying mnemotechnical devices where colours and motifs conceal a general indication of the content for instance to recite the correct number of prayers required We briefly verified the connectivity of the proposed symbolic signalling with the various artefact types within the jewellery in black glass (Table 136). We perhaps consider this category of adornments, at present, as rather unappealing or even tasteless; it is, however, self-evident that the wide variety of personal ornaments made in black glass during the Roman imperial period are to be considered primarily as aesthetical expressions to adorn the body, the hair and the clothing. We assume that in late Roman society, when sobrietas was a nourished virtue, the black glass jewellery was part of a tendency towards a sobriety in fashion that considered the wearing of these clumsy but plain and pure ornaments as acceptable. On the other hand, we have to bear in mind that these types of jewellery most likely also received a visual meaning according to an encoded concept. Hence, we cannot exclude courtship as we remain unable to provide evidence. The use of black glass jewellery is not excluded from being used as an ethnic marker but this also cannot be verified. Otherwise, the wearing of specific bangles, finger rings, etc. must be regarded as only a minor ethnic marker within a local or regional dress code. On the basis of the wearing of glass bangles by a limited number of girls and young women, it is not excluded to categorize at 228 least the arm rings as a possible individual marker. Yet, from various contexts we could see that bracelets equally have been used as ritual objects by girls and young women who wore black glass bracelets at the left wrist or cherished as amulets and talismans by men in Period IV who kept fragments of glass artefacts in their purses [see Chapter 5]. This implies that black glass jewellery does not have to be ruled out as having a prophylactic function, despite the absence of hard evidence. The presence of bracelets and finger rings with or without gemstones has been attested in ritual deposits such as the one at the cave Trou de Han (BE) and the socalled treasure of Regensburg-Kumpfmühl (DE). It is not excluded that the other jewellery types were used in offerings. There is no evidence or indication to demonstrate the remaining values, but they do not have to be ruled out. Future detailed research including all glass jewellery will increase our understanding on this issue. This short overview on the allotment of values makes clear that most black glass jewellery could be used for multiple reasons, except for the hairpins which seem to have been solely used for aesthetic reasons. Table 136: List of different values per functional type in black glass (*= positive; -= negative; ?= uncertain) value bracelets finger rings beads pendants hairpin Aesthetics Courtship Ethnic marker Social marker Individual marker Ritual object Offering Amulet/talisman Prophylactics Exchange medium Inalienable possession Communication system Counting device 7.4.1 *** ? ? ? *** *** *** *** ? ? ? ? ? *** ? ? ? ? *** *** ? ? ? ? ? ? *** ? ? ? ? *** ? ? ? ? ? ? ? *** ? ? ? ? *** ? ? ? ? ? ? ? *** ? ? ? ? *** ? ? ? ? - gems *** ? *** *** ? *** ? ? ? ? ? - Arm rings The idea to attribute a particular meaning to the black glass bangles originated with a paragraph in Jan Kock and Torben Sode‘s archaeo-ethnographical campaign in the heart of northern India, describing the use and meaning of glass bangles in present-day India (Kock, Sode 1995, 14). Glass armrings now form an essential part of the traditional dress of most Indian women, and a great deal of symbolism and superstition is attached to their use. Bangles are for women only, and if a man decided to walk around wearing one it would be taken as a sure sign of helplessness and loss of manly strength. Green armrings are restricted to young women who are to marry or are very newly married. When a married man dies, the widow, as a symbol of the dead, must break her bangles and no longer wear any. If a woman is unlucky enough to break one of her armrings it means bad luck, or is a warning that something bad will happen. For this reason people will even avoid stating directly that an armring has been broken. Many girls prefer to hide the event by saying that the number of armrings has increased. Bangles are primarily used for adornment, however. The selection of colours and patterns is practically infinite and as many are often worn together they are something that can be both seen and heard. (Kock, Sode 1995, 14) 228 This is purely based on the hypothesis that the black glass bangles are imitations of jet bangles and the link with the belief in Roman times that jet could prove the girl‘s/woman‘s virginity (Pliny the Elder, NH XXXVI, 142). 279 Chapters 4 and 5 on chronology and context analysis, respectively, have demonstrated that black glass bracelets are a commodity that does not occur in the early Roman imperial period. The complete absence of glass bangles in Roman Italy and the Mediterranean during the Hellenistic and early imperial period proves that the Romans initially had no interest in producing or using glass bangles. Conversely, this type of glass jewellery is 229 characteristic to the Celtic La Tène (LT) culture on the continent and many fragments of LT glass bracelets st nd have been excavated in 1 and early 2 century AD contexts, in particular at early Roman rural settlements in the north-western provinces of the Empire. The presence of Celtic Iron Age glass bangles in early Roman contexts amplifies the difficulty of determining the period when the LT bangles were produced (Gebhard 1989b, 102; st Roymans et al. 1993). The hitherto total absence of archaeological evidences of a 1 century AD glass workshop producing Celtic bangles makes that any proposed rational reasoning remain hypothetical. It is, however, not the question whether this La Tène material was still produced after the Roman conquest which is important here, but verifying whether information can be deduced from the contextual findings of such material. It is significant that all recovered pieces are only small fragments of glass bangles rarely longer than 20 mm. This is to us an indication that the late Iron Age glass bracelets were not produced in the early Roman period. There is, however, continuity in usage that can be granted different interpretations: 1) the bits and pieces kept the significance the bangles originally were given and are as such to be regarded as a ‗pars pro toto‘; or 2) the fragments received a new connotation. It is noteworthy that the material, besides its fragmentary condition, is sometimes preserved in a much worn state, an example being the colourless piece with yellow opaque glass from the early Claudian quadrangular enclosure in Tienen (BE) (Cosyns et al. 2006b). Hence, such bangle fragments point to the survival of a Celtic tradition/belief/ritual of the indigenous ‗Romanised‘ peoples in temperate Europe. Similar fragmentary bangle 230 pieces have been noted previously in regard to the early Romano-British glass bangles. A study by Thea Haevernick proposes to see these bits and pieces as amulets with apotropaic meaning (Haevernick 1968). This would make the small Tienen fragments part of a ritual deposit and would mean that the single-event deposition in the ditch was preceded by a ritual offering (Cosyns et al. 2006b). nd The occurrence of fragmented LT-glass bangles in early Roman contexts disappears in the 2 century AD, and th th this practice re-emerges in the late 4 or 5 century AD in the West when the production of glass bangles stops again. Where late Roman tombs only show re-used pieces of Roman black glass bangles, the early Merovingian tombs rather contain fragments of LT glass bracelets (Haevernick 1968; De Witte 1977). The pieces are mainly kept in a purse at the waist together with a wide range of objects, which appear mainly in a broken state. Examples have been attested at mixed late Roman-early medieval cemeteries such as for instance the Type A1 th th fragment in a late 4 -5 century AD tomb at Maule (FR) (Arveiller 2006, 160, no.59) [cat.no.1260]. Most field archaeologists are in search of well-dated artefact types to date the excavated layers and features, and are less interested in considering the possible causes that influenced the changing material culture. The evolution in material culture is unmistakably affected not only by trends but also by the introduction of new technologies and rd by economical situations. The gradual fading of the Romano-British glass bangle from the 3 century AD coincides with the booming of the jet consumption on the British Isles for the production of bangles and other jewellery (Allason-Jones 1999) as well as with the black glass production on the Continent. Without having proof of correct validity for the here-proposed causes of these changes, we see a clear discrepancy between the very idiosyncratic Romano-British glass bangles and the Romano-British jet bangles on the one hand and great uniformity between the latter and the monochrome black glass bangles from the continent. It is easy to speak of a shifting trend during the Severan period involving a gradual change from traditional indigenous glass bangles by new Roman(ised) bangles in jet. But being aware of the intricacy in valuing the various factors involved and the gradualness of the alteration, any explanation will seem too simple and above all incomplete. However, the success of the black jewellery (in jet and glass) could only be possible when it was known to the Roman population beforehand. Although hazardous, it is important to try to understand what social/economical/cultural changes might have generated this radical shift of materials and styles within jewellery in the Severan period. The majority of the Roman glass bracelets come from settlement contexts and is residual material [see Chapter 5], and consequently the material is unsuitable to verify its use. Only a marginal part of all Roman glass bracelets from known contexts is suitable to study its use and function. The most useful information appears to be available from burial archaeology. Its intrinsic value is of importance to come to a certain degree of understanding but at 229 The peculiar types of Romano-British glass bangles, produced during the first centuries AD and almost solely limited to the British Isles, are a local product and reminiscent of a non-Roman production. 230 A large part of the retrieved Romano-British material is small and seems to have been ground or cut as for the purpose of applying it in something or mounting it to join several fragments together (Kilbride-Jones 1938, 370371). 280 least to a consideration of the use and meaning of black glass bracelets in the Roman period. Chapter 5 listed those black glass bracelets from burial contexts with a number of acknowledged parameters. It is unfortunately no longer clear how to interpret this trend since a recent comparative study on the use of bracelets from transalpine cemeteries (Bolla 2005, 62; Bolla 2011) shows that about 75% of all bracelets are worn at the left under arm and only about 25% is worn at the right under arm. However, it is important to stress that on the right side always one single bracelet is recorded, while on the left arm it can vary from one to six. Unfortunately it is unclear if attention is given to the difference in materials of the bracelets. The publications on Roman cemeteries in the North-western provincesVermand (Eck 1891); Krefeld-Gellep (Pirling 1966; 1974; 1979; 1989; 1997; Pirling, Siepen 2000; 2003; 2006); Trier (Goethert-Polaschek 1977); Tongeren (Vanvinckenroye 1984); Oudenburg (Mertens, Van Impe 1971); London (Barber, Bowsher 2000)illustrate that (black) glass bracelets are only worn at the left side like the bangles in jet and bone, whereas the bracelets in silver, gold and bronze can be worn on either side. Presumably, this must have had a reason. Despite the comprehensive overview recently provided by Henning Wirth (2010) on the perception and attitude of the Romans towards the left-hand and left-handedness, a lack of ancient written sources results only in assumptions on the possible different use and meaning of wearing bracelets in the various materials. From ancient writers we learn, however, that the Romans gave a magical connotation to the left hand seeing that various ritual actions had to be executed with for instance the left thumb and left ring finger, also called digitus medicinalis (Wirth 2010) [see this chapter 7.4.2. Finger rings]. Table 137: List of tombs with (black) glass bracelets (n = quantity) cat.no. site context sex age side n type 2016 Tawern (DE) female - - 1 A1 3687 Kaiseraugst (CH) female adult - 1 A1 4120 London (UK) inhumation inhumation tomb 33 inhumation tomb B168 inhumation tomb 194 inhumation tomb inhumation tomb 721 inhumation tomb 2 inhumation tomb 1749 inhumation tomb rue de Corbie inhumation tomb 108 inhumation tomb 2 inhumation tomb 22 inhumation tomb 100 inhumation tomb C inhumation tomb 1 (HA382) inhumation tomb 2 (HA382) inhumation tomb 165 (HA132) inhumation family burial female juvenile 13-18yrs left 1 A1 female adult 25-30yrs adult c. 30yrs left 1 B2 unclear 1 B5 inhumation undefined 479 Oudenburg (BE) 4170 Orpington (UK) 14 Bregenz (AT) 1868 Eining (DE) 1910 Krefeld-Gellep (DE) 1003; 1005 1741; 1745-46 Amiens (FR) Vermand (FR) 278-79 Furfooz (BE) 280-81 Furfooz (BE) 482-83 Oudenburg (BE) 517-518 Spontin (BE) 1983-86 Niederzier (DE) 1982 Niederzier (DE) 1987 Niederzier (DE) A4607 Canterbury (UK) 2756 Milan (IT) female female adult - 1 D1 female child left 1 ? female child left 1 D1 female child left 1 D1 female adult left 3 ? female juvenile left 2 D1 female juvenile left 2 D1 undefined unknown unclear 2 D1 female adult - 2 D1 female adult left 4 D1 female adult left 1 D1 female adult left 1 D1 left 1 D1 left 1 D2 female child c. 11yrs child c. ½-1½yrs The examples presented in Table 137 bear out that above all girls and female juveniles wore black glass bracelets, e.g. tombs from Tawern (s.n. 1950, 76-77, fig.40), Krefeld-Gellep (Pirling 1974, tomb 1749, n°1), Amiens (Dilly, Mahéo 1997, 123, X.) and Furfooz (Nenquin 1953, 74). The Amiens tomb, found in 1820, contained a lead coffin enclosing the inhumation of a child wearing two octagonal jet bracelets and two plain black glass bracelets with a D-shaped section (Type D1; Riha type 3.34; Spaer type A.2) [cat.no.1004]erroneously described as jet bracelets. Because the lead coffin is decorated with a chrismon monogram (Mahéo, Dilly 1997, 281 120) it is most likely the deceased was a Christian. Also the two burials of female juveniles at Furfooz (BE) can be considered of early Christians because the burial place was formerly a bath-house of the late Roman fortification th settlement at Furfooz that possibly was reused as a funerary chapel or church from about the middle of the 4 century AD (Nenquin 1953, 74, nos. F6-9, pls. X,7-8; 88; 98). Despite the absence of evidence to support this assumption it is tempting to connect these plain Type D1 bangles in black appearing glass with early Christian practice and belief. The interest of Beer Cart Lane within the Roman city centre of Canterbury (UK) is its contextual particularity (Bennett 1980, 406-410). A large pit containing a multiple burial was excavated in 1980, just below the Angloth Saxon layers of the 5 century AD and very probably dating from the very end of the Roman period. This is a remarkable grave, seeing that the interment is in the first place situated in the middle of the town and not in a cemetery outside the settlement, as normally done in Roman society, and secondly because the grave comprised a whole family of two adults and two children as well as one or two dogs, instead of individual burials. As regards the burial gifts of each skeleton, the total absence of objects in relation to the adult male is striking, especially in contrast to the other three individuals. The adult female had eight bronze bracelets on the left wrist and another th th th one on the right wrist, which are dated 4 century AD, four bronze keys, a small knife and razor (mid-4 to mid-5 century AD), three amber beads and ten dark brown or black glass beads. One juvenile is a girl of about 11 years old and has a black glass bangle (type D1; Riha type 3.34; Spaer type A.2), three bronze bracelets and fragments of a fourth one, as well as an ivory bracelet, a small bronze key and 15 small glass beadsmostly blue and green, th th characteristic of the 4 5 century AD. The second juvenile, of uncertain gender, was about eight years old and only wore two bronze bracelets. The assessment of the burial contexts demonstrated that only girls and (young) women wore these ornaments and only at the left wrist, which corroborates with the two distinct size categories: 1) the small bangles with a diameter below 60 mm (= for girls) and 2) the large bangles with a diameter above 60 mm up to 95 mm (= for women). This observation is evidence to assume that this type of jewellery is to be associated with the female gender. However, occasionally male burials have been recorded that have included a black glass bracelet fragment, such as tomb 753 in Kaiseraugst (CH) (Martin 1976, 65-66, Taf. 48A:18) [cat.no.A4608]. Being in a fragmentary 231 condition (length of 18 mm), it was kept in a purse together with several finger rings in silver and bronze and a set of coins covering three centuries ranging from Trajan up to Valens or Gratian. The bangle fragment could not be worn at the wrist but must have received a new symbolic meaning that caused the object to be cherished. Hence, there is reason to believe the fragmented object had certain significance for the person who carried it with him, making it an artefact of itself. The custom to carry (in a purse) a small fragment of a glass bracelet (mainly below 20 mm but never exceeding 25 mm) appears especially in late Roman and early Merovingian tombs, but principally it consists of small pieces of La Tène bangles in ultramarine blue glass. The specific circumstances in which these pieces were found have caused them to be interpreted as amulets having an apotropaic connotation (Haevernick 1968). It is tempting to follow the same course by attributing a similar meaning to the Roman black glass fragment in the Kaiseraugst purse. The inhumation tomb 575 of the Drouly cemetery at Frénouville (FR) yielded a leather purse containing a large fragment of a twisted bracelet type A6, with applied trails in opaque white glass [cat.no.1118], two glass vessel fragmentsone rim fragment in pale blue glass and one body rd fragment in yellowish-green glassand two late 3 century AD coins, one of which is a pierced antoninianus th th rd th (Pilet 1980, II, 285; III, pl.150:5). The 5 and 6 centuries AD dated tomb makes the 3 and early 4 century glass bangle fragment, like various other items in the purse, part of an assortment of antique trinkets. Another such example is available from tomb 598 of the same cemetery at Frénouville (FR). A fragment of a late Roman bracelet type D1 [cat.no.1121] was found at the belt height on the left side together with a hook, two nails and an iron pin (Pilet 1980, II, 298; III, pl. 155 top). Besides the incompleteness of the bangle (only one-third of its circumference), its position together with other small items suggests that the bangle fragment must have been th th kept in a purse. But how do we have to interpret the presence of late 4 and early 5 century fragments in purses th th of people from the late 5 and 6 centuries AD? It is evident that, once broken glass bangles lose their initial function. However, the initial function of a piece can remain as such, but when it loses its utility it becomes waste and becomes unwanted. Given that people cared for these small bits and pieces, these fragments were given a new meaning. It is significant that the Merovingians integrated into their material culture a great number of recycled artefacts from the late Bronze Age onwards that were fragmented for the large part (Haevernick 1968). The Merovingians not only re-used the Roman glass bangle fragments but also fragmented La Tène glass th th bracelets. Also, other Roman black glass artefact types do occur in late 5 and 6 century AD tombse.g., the 231 The bags must be considered as purses as they are always found at the waist where a belt is located and because the content consists of a wide variety of small fragmented items. 282 Period II carchesium base fragment from tomb 297 of the Merovingian cemetery in Grez-Doiceau (BE) [cat.no.209]. Dating from the Severan period, there is a discrepancy of three centuries, which would be comparable to walking around nowadays with a purse full of small artefact fragments from Napoleonic times. It is characteristic but most likely that these bits and pieces have been invested with an apotropaic meaning or, as Jacqueline Lemière suggests, a talismanic connotation (Lemière 1980, 339). It is of course impossible to retrieve what meaning the owner gave to the objects, but they must have been of emotional and/or spiritual value, indicating a form of superstition. It is typical of all times and cultures to attribute certain powers or meanings to specific colours and materials, but it could as well have been the only tangible bond remaining with his deceased beloved wife or daughter. Another issue is whether the black glass bangles were solely used as bracelets. The problem is that the burial context gives only a partial insight of the use of this type of jewellery in Roman times. It goes without saying that th the Romans could have given other uses to these glass bangles. For instance, the late 4 century AD tomb 404 of the Drouly cemetery at Frénouville (FR), near Caen, yielded two small bangles and a hairpin in a transparent glass (Pilet 1980, (I) 96; (II) 206, (III) pl. 107, fig.2). One bangle is made in a clear yellow glass, and the other, as well as the hairpin, in a clear green glass. With a diameter of 47 mm and 49 mm, correspondingly, they clearly reside in the small-sized bangle category. Christian Pilet describes them as being used as ‗parure de coiffure‘ (hairdressing accessories) together with the hairpin. He shows in a reconstruction drawing of the hair style the combined use of the two small bangles to have the hair ‗en queue de cheval‘ (ponytail) passed through both rings by folding the hair double and fastening it by piercing the ponytail with the hairpin above the rings (Figure 127). Even if the artefacts were not found at the head, this hypothesis sounds very plausible. Not willing to exclude this possibility for the Roman glass bangles, it is conspicuous that such use has not been recorded so far for those in black glass. In connection with the bangles, we have few hairpins in black glass and none of them come from a context that could give more information on their (combined) use (with bangles). Figure 127: Hair-style reconstruction proposed by Pilet 1980, (I) 95, fig.2; pl.107:404. Some information, however, can be compiled from the connection with jet bracelets. Not only is there a striking resemblance with the bracelets in black glass (Allason-Jones 1996), but they also appear to have had a similar use. Based on the burial gifts, the bracelets in both materials are characteristic for girls and young women, supporting the assumption of a religious or magical significance (Allason-Jones 2005, 184). Lindsay AllasonJones also mentions some passages by ancient writers, for instance Galen, De simpl.med.facult. IX, 203 who describes the use of jet for medicinal use by burning and swallowing (Allason-Jones 1996, 15). Referring to the curing effect of jet, she stresses that for jet no apotropaic meaning as such is known from ancient written sources: Gagates is a stone, so called from Gages, the name of a town and river in Lycia. It is asserted, too, that at Leucolla the sea throws it up, and that it is found over a space twelve stadia in extent. It is black, smooth, light, and porous, differs but little from wood in appearance, is of a brittle texture, and emits a disagreeable odour when rubbed. Marks made upon pottery with this stone cannot be effaced. When burnt, it gives out a sulphureous smell; and it is a singular fact, that the application of water ignites it, while that of oil quenches it. The fumes of it, burnt, keep serpents at a distance, and dispel hysterical affections: 283 they detect a tendency also to epilepsy, and act as a test of virginity. A decoction of this stone in wine is curative of tooth-ache; and, in combination with wax, it is good for scrofula. The magicians, it is said, make use of gagates in the practice of what they call axinomancy; and they assure us that it will be sure not to burn, if the thing is about to happen as the party desires. (Pliny the Elder, Naturalis Historia XXXVI, 142 [transl. by Bostock et al. 1855 - online @Perseus, Plin. Nat.36, Probably this medicinal quality, combined with its electrostatic property, provides an explanation why jet was regarded as an appropriate material to make amulets. Seeing that items in jet got copied in other black-appearing, non-electrostatic organic materials as well as black glass, may we then assume these materials were also considered to have the same talismanic connotations? Written sources remain silent on the magical and apotropaic function of objects in jet or any other black material. Only archaeological sources provide evidence of apotropaic representations on jet and black glass artefacts such as the head of Medusa. But this emblematic representation par excellence can occur in any kind of material, e.g. metal, stone, pottery or in paintings and mosaics. Considering all these material types as apotropaic would thus be very inconsistent. Hence, not the substance jet or any other black material retains a magical meaning, but only the representation itself. That some particular finds relate the use of jet with Eastern mystery religionssuch as the burial at Bainesse (UK) said to be of a gallus, a believer of the goddess Cybele, or the jet dagger from the Walbrook Mithraeum in London (AllasonJones 2005, 184-185)only demonstrates the status of the material. However, it is interesting to connect this with the twisted bracelet in black glass that got retrieved from the well at the Cybele sanctuary of Arras (FR) [cat.no.1024] (Jacques 1997, no. 128). Black glass finger rings and bracelets can be decorated with Jewish, Biblical, Christian, Mithraic or Classical pagan symbols, so that there is no specific connotation to be given to it. This means that the black glass bangles worn by the girls and women demonstrate the religious belief and/or ideological conviction of the wearer. 7.4.2 Finger rings Unlike the contrast in the particular use of the gender related black glass arm rings, the context analysis did not contribute to the enhancement of our knowledge on the use and function of black glass finger rings. On the whole, the use of black glass finger rings has to be considered in a larger framework, incorporating a general discussion on the use and function of all finger ringsincluding those in jet, iron, bronze, silver and gold, as there is no possibility to distinguish the material of the finger rings represented in figurative art or descriptions brought up by ancient writers. Furthermore a search for written sources discussing finger rings and their representation in paintings, e.g. in the so-called Faiyum-portraits, or in sculpture, e.g. the funerary portraits and reliefs, would lead us too far from the scope of our research. Despite the limited discussion on the use and function of black glass finger rings, we would like to refer to Henning Wirth‘s dissertation on the left hand and left-handedness (Wirth 2010). Henning Wirth mentions that the generally known ring finger or fourth finger next to the little finger of the left hand was called digitus medicinalis. th th Macrobius, a Roman grammarian and Neo-Platonist of African origin who lived at the end of the 4 –early 5 century AD, wrote in his Saturnalia on the use of finger rings and why wearing these especially at the fourth finger of the left hand. He presented two different beliefs. At first explains the belief of the Egyptian of a hallowed finger: Comme Disaire eut achevé de parler, Aviénus ramassa sur la table son anneau, qui venait de tomber du petit doigt de sa main droite; et les assistants lui ayant demandé pourquoi il le mettait à une autre main et à un autre doigt qu'à celui qui est consacré à le porter, il leur montra sa main gauche enflée par suite d'une blessure. … dis-moi pourquoi l'on s'est généralement accordé à porter les anneaux principalement à la main gauche, et au doigt qui est à côté du plus petit, et qu'on appelle médicinal ? …un nerf parti du coeur se prolonge jusqu'au doigt de la main gauche qui est à côté du plus petit, et qu'il s'y termine en s'enlacant dans les autres nerfs du même doigt. Voilà pourquoi les anciens voulurent que ce doigt fût entouré d'un anneau, comme d'une couronne. Ce que tu dis de l'opinion des Égyptiens, Disaire, est si vrai, qu'ayant vu dans leurs temples leurs prêtres, qu'ils appellent prophètes, parcourir les simulacres de leurs dieux pour oindre ce seul doigt d'essences odoriférantes, et leur en ayant demandé le motif, j'appris de leur premier pontife que c'était à cause du nerf dont tu viens de parler, et de plus, à cause du nombre qui est signifié par ce doigt; car étant plié, il désigne le nombre six, nombre entièrement plein, parfait et divin. Le pontife me démontra par plusieurs arguments les causes qui constituent la perfection de ce nombre. Je les passe sous silence, comme étant peu appropriés à notre conversation actuelle …voilà ce que j'ai appris dans cette Égypte, dépositaire de toutes les connaissances sacrées, sur le motif qui a fait affecte l'anneau à un doigt plutôt qu'à un autre. (Macrobius, Saturnalia 7,13 [http://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/7*.html]) The second belief was that of Etruscan tradition viewing socio-cultural based functional reasons granting only free men the permission to wear a finger ring on the left hand: 284 … je me souviens d'avoir lu sur ce même sujet dans Atéius Capito, l'un des hommes les plusinstruits du droit pontifical. Capito, après avoir établi que la religion défend de sculpter les statues des dieux avec des anneaux aux doigts, passe à l'explication du motif pour lequel on porte l'anneau à ce doigt et à cette main. "Les anciens, dit-il, portaient l'anneau autour de leur doigt,comme sceau et non comme ornement; c'est pourquoi il n'était permis d'en porter qu'un seul; et encore ce droit n'appartenait qu'aux hommes libres, à qui seuls pouvait être accordée cette confiance qu'on attache à un sceau. Ainsi, les esclaves ne jouissaient point du droit de porter l'anneau. Soit qu'il fût de fer, soit qu'il fût d'or, l'anneau était orné de ciselures, et chacun le portait à son gré, à quelque main ou à quelque doigt que ce fût. Dans la suite, ajoute-t-il, un siècle de luxe amena l'usage d'inciser les sceaux sur des pierres précieuses. Cet usage devint bientôt universel; en sorte qu'il s'établit une émulation de vanité, pour élever de plus en plus le prix des pierres destinées à être ciselées. De là, il arriva que la main droite, qui agit beaucoup, fut affranchie de l'usage de porter des anneaux, usage qui fut transporté à la main gauche, laquelle reste plus oisive; et ceci pour éviter que la fréquence de l'usage et du mouvement de la main droite n'exposât les pierres précieuses à être brisées. … on choisit parmi les doigts de la main gauche celui qui est à côté du petit, parce qu'il fut trouvé plus apte que les autres à recevoir la garde précieuse de l'anneau. En effet, le pouce … ne reste pas oisif, même à la main gauche. Il est toujours en activité de service, …. Le doigt qui est placé à côté du pouce fut trouvé trop nu, puisqu'il n'est point défendu par la juxtaposition d'un autre doigt; car le pouce est placé tellement au-dessous, que c'est tout au plus s'il dépasse sa racine. Le doigt du milieu … et le plus petit furent négligés, comme peu convenables, l'un, à cause de sa longueur, l'autre, à cause de sa courte taille, et l'on choisit celui qui est enclavé entre ces deux, et qui fait peu de service, comme étant, à cause de cela, le plus convenablement disposé pour la garde de l'anneau". (Macrobius, Saturnalia 7,13 [http://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/7*.html]) Macrobius leaves the reader free to decide which of these two opinions to accept: Telle est la version du droit pontifical; que chacun suive à son grél'opinion des Étrusques, ou celle des Égyptiens. (Macrobius, Saturnalia 7,13 [http://penelope.uchicago.edu/Thayer/L/Roman/Texts/Macrobius/Saturnalia/7*.html]) It is clear that already in Roman times ritual meaning faded into incomprehensive rooted traditions ending up in a world of superstition and symbolism that had to give way for objectivity and functionality. 7.4.3 Beads From contextual finds as well as from images such as the so-called Faiyum portraits, it is clear that glass beads were mainly used as body ornaments in necklaces, earrings and bracelets, but glass beads were also used as dress ornamentation sewn into clothes. It is self-evident that black glass beads should be regarded as ornament, but we have seen earlier that many other functions can give a reason to wear beads (Tables 135-136). We verified whether the Romans gave particular meaning to beads of specific shape, colour, decoration, as well as for the combination and amount of beads put on. Aware of the fact that tastes changes, it is questioned whether the use and function of black glass beads changed through Antiquity. Concerning the use of black glass beads, we distinguished two major groups on the basis of the perforation in the middle of the object: single perforation versus double perforation [see Chapter 3]. The perforation was meant to put a string or wire through to create a beaded chain. The use of strings in perishable organic material such as from cut leather or from twisted hair or flax can be assumed from the many sets of loose beads that clearly formed a bracelet or necklace. The applied metal wires to hold the beads together show the use of gold wire, 232 silver wire, bronze wire, and even iron wire. Examples of beads made of black glass strung on metal wire are at present rather scarce, such as the tear-drop shaped beads from Wancennes (BE) integrated in earrings of bronze wire (Mignot 1984, 224) [cat.no.618] or the discoid pressed beads of the necklace in gold wire in the Archaeological Museum of Ancona (IT) (unpublished) [cat.no.A4606]. To detect any possible purposeful distinction between single and double perforated beads, we compared the characteristics of each group. One of the most striking features of single perforated beads is that these are made in such a way that they can turn around the string or wire. Furthermore, the shape and decoration of single perforated beads are made in such a way that it does not matter if the bead moves all the time. In contrast, the so-called spacer beads have a decorated frontal side and an undecorated rear side, necessitating the creation of beads with double perforation to avoid the bead turning upside-down. The double perforation avoided having all 233 the time the single-sided decoration back-to-front and kept the top surface perceptible. It is clear that the Roman glassworker had a functional motive in mind to make these beads have double perforation that was dictated by the technological aspect of the production of the so-called ‗spacer beads‘. These were made by folding and pressing a decorative design, while the single perforated beads were all rod-formed [see Chapter 2]. Hence we can conclude that according to the way of wearing beads, the Romans adapted technical features of 232 233 This is almost solely attested on beads in blue and green glass. An additional help for this is the flat reverse side and the rather broad size of the spacer beads. 285 bead shapes. Because the various black glass spacer-bead types resemble those in jet, it is possible that the rd th Roman glassworkers of the 3 4 century AD on the continent imitated the shape of the popular but rare jet spacer beads. Concerning the function of the most idiosyncratic category of late Roman glass beads, the double perforated pressed spacer bead has received multiple names. First, scholars named the type after their appearance, like 234 Rudolf Noll who used ‗Rippenglasperlen‘ (Noll 1963, 68) and Thea Haevernick who named the largest category ‗Trilobitenperlen‘ because of its resemblance with the trilobite fossils (Haevernick 1983). Other scholars gave a name in association with its presumed function: Tamás Gesztelyi spoke of ‗Glaskameen‘ and ‗Glasbullen‘ (Gesztelyi 1991; Gesztelyi 1998, 129), while Maud Spaer and Bernard Gratuze correspondingly spoke of ‗spacers‘ (Spaer 2001, 76) and ‗écarteurs‘ (Gratuze 2001b). Such hypothetical function is based on the logical reasoning that the two perforations were used to keep two chains parallel to create a necklace or bracelet with a double row. This explanation only makes sense when the necklace and bracelet combine small single perforated beads with one or two double-perforated beads. Seeing that necklaces and bracelets could equally consist of spacer beads alone, it is very likely that (an)other reason(s) was(were) responsible for producing beads with two perforationsrespectively the ‗Trilobitenperlen‘ necklace from the fortress of Vemania near Isny (DE) (Garbsch 1971, 137, fig.30) [cat.no.1892] and the theatre-mask beaded bracelet from tomb 12 of the San Egidio cemetery at Cesena (IT) (Fadini, Montevecchi 2001, 51, fig.3; Mandruzzato 2008, 162) [cat.no.2739]. It is interesting to see that within the multitude of possible names, people choose a metaphoric description, an interpretative functional one or one based on plain visual features. These beads occur in a wide variety, but the more amorphous ribbed pieces are most common. A second large group are the portrait beads. Further, there is the representation of the Medusa, theatre mask, lion‘s head, and plenty of other images linked with pagan, Jewish, Christian and Mithraic symbolism [see Chapter 3]. All in all it has been accepted from their first observations that these artefacts were jewellery and particularly beads to be used in necklaces and bracelets (Hampel 1905, 68; Noll 1963, 68; Haevernick 1983). A deviant proposal was presented by Tamás Gesztelyi, who questioned the validity of this communis opinio (Gesztelyi 1991; Borbély-Kiss et al. 1994, 836; Gesztelyi 1998, 129). His argumentsthe poor execution, the double perforation and the mainly individual occurrencewere put forward against the idea these beads were originally produced as jewellery. He proposed to see these beads as glass sealings analogous to those in lead. Where the lead sealings are regarded as common, he suggested the double perforated pressed glass beads were employed to seal containers of precious medicine to guarantee the genuineness and unadulterated nature of the product. Gesztelyi deduced the link between these beads and medicine on the basis of the impressed representations decorating the top surface. Gesztelyi‘s remark on the so-called poor execution of these beads says more about his personal taste than his objective appreciation. He probably is more fond of the very regularly cut and smoothly polished equivalents in jet (Figure 128). But to what extent did the Roman consumer of black glass jewellery care about the biased remarks th of a late 20 century scholar on 1500-year-old trends? The shape and the quality of double-perforated pressed beads are not only the result of the degree of workmanship but are very much influenced by the nature of glass metal and the production method, i.e. folding and pressing. But it is perhaps interesting not to limit the classification of the spacer beads to plain and very elaborate types but also include the fact that some pieces are well-finished while others appear more clumsily executed. Figure 128: (left) ‗Trilobitenperle‘ in black glass from Augst; (right) ‗Trilobitenperle‘ in jet from Augst (both drawings taken from Riha 1990, Taf. 39:1345; 1347) 234 Thea Haevernick only named the double perforated beads ‗Trilobitenperlen‘ which were decorated with transverse ribbings parallel with the perforations (discriminating plain and crosswise decorated transverse ribbings). All other beads were named theatre mask beads, lion head beads, portrait beads, etc., which are confusingly called ‗Trilobitenperlen‘ as well. 286 The second and third remark, respectively on the double perforation and the individual occurrence, seem even more odd to us. He stated that because other beads found together with the so-called ‗Trilobitenperlen‘ in burial contexts only have one hole, the ‗Trilobitenperlen‘ initially must have had another use and purpose (Gesztelyi 1998, 129). But there is no reason why double-perforated beads could not have been combined with singleperforated beads. The necklace from Isny (DE) is a perfect example. Furthermore, necklaces and bracelets with one single pendant or bead are appropriate at all times, while a necklace with large ‗Trilobitenperlen‘ all around the neck would be very unpractical. Most probably, therefore, the Isny-necklace has a good number of small cylindrical beads on either sides of the ‗Trilobitenperlen‘. Gesztelyi proposed that the double-perforated pressed beads were used for sealing (‗Glasbulle‘ as he calls them), like clay, lead or wax seals. It is out of the question that the ‗Trilobitenperlen‘ would substitute for those in clay, lead or wax. The technological features of glass are much too elaborate and not a very practical alternative. In addition, the hot glass only becomes workable from a temperature of at least 600°C, which means that all organic material would catch fire. However, he plainly states that the double-perforated beads were made in advance and a string or wire got put through to seal the object. However, the hundreds of such beads reported from all over the Roman Empire show no remains of a metal wire, nor one together with a vessel. We wish to stress that Gesztelyi‘s proposal lacks the real function of a seal and that is to seal up something. The presence of the entire gamut of ‗Trilobitenperlen‘ in burial contexts of women and children only or in treasures together with other female ornaments are arguments urging this type of bead is a characteristic ornament for women and girls [see Chapter 6]. On the other hand, it appears to have been improper for the honourable men (and boys) in Roman society to wear jewellery, with the exception of some explicit items, such as a wedding ring or a bulla. Questioning the interpretation of a find in Pécs, Sopianae (HU) (Fülep 1984, 90, fig.35; 99:10/8, pl.LII7), Gesztelyi only acknowledged he was mistaken after a similar find from a tomb at Bátaszék-Kövesd (HU) was reported: Aufgrund dieser Umstände ist die Verwendung auch der Glaskameen mit figuralen Darstellungen als Glieder eines Armbandes oder einer Halskette offensichtlich. (Gesztelyi 1998, 131) Gesztelyi still pleads for a correlation with a medicinal meaning (Gesztelyi 1998, 131). He nevertheless recants his prior statement that the word VΓIE on two double-perforated glass beads is to be linked to the goddess Hygieia. He now is more convinced that the inscription should be regarded as a motto wishing for good health and that the bead must thus have been considered by the wearer as a magic amulet. Other inscriptions are also seen as amulets, such as those bearing ZOHN, like the double-perforated discoid bead in the Israel Museum, Jerusalem (IL) (Spaer 2001, 76, no. 58). The Greek word stands for ‗good fortune‘ and has the same connotation as VIVAS in Latin. The word ZOHN is split in two at either side of a frog which is depicted in view from above. A similar bead, albeit in a yellowish orange glass, comes from the cemetery of Keszthely-Dobogó (Gesztelyi 1998, 133, fig.4). Gesztelyi sees the beads and pendants in yellowish orange glass as imitations of cornelian and gold. A bracelet made of ‗Trilobitenperlen‘ in black glass and enveloped in gold foil from the cemetery of Dunapentele (HU) (Alföldi 1957, 442) [cat.no.2221] or the ebony wooden bracelet with gold leaf decoration preserved on two areas from a tomb in the cemetery of Tyre (LB) (Chéhab 1986, 153-155, pl.XVa:g3) illustrates most certainly the wish/need to make the object seem as if it is made from gold. Striking is the resemblance of the piece from Tyre with the gold bracelets found in the treasure of Eauze in south-west France (Schaad 1992, 36-39, pl.2). It is possible that the Eauze example was a comparable piece made in thicker gold sheet enveloping a wooden core that has not been preserved. A part of the double-perforated pressed beads in black glass are adorned with a portrait, whether it is single or double, whether only a woman or a man is represented [see Chapter 3]. This decorative element is not limited to double-perforated pressed beads in black glass. It is also attested on a type of finger rings as well as on pendants. Furthermore, this kind of jewellery is equally present in a yellowish orange glass. The use of small, impressed portraits is known for lead seals as well as for silver and gold jewellery. This material makes clear that portraits on the pendants and beads in glass are not to be regarded solely as impersonal images of men and women or as emblematic representations of husband and wife a propos the double portraits. The portraits seem to be interpreted as the depiction of the emperor and eventually in combination with the empress or together with his brother/son(s). These objects can therefore be correlated with the imperial military authorities. The lunula-shaped golden pendant from the treasure of Regensburg-Kumpfmühl, Germany (Boos et al. 2000, 2426, figs. 10-11, pl. 5:2) is decorated with a double portrait: at the left side a woman looking to the right and at the 287 right side a bearded man looking to the woman. The treasure comprises 638 coins25 aurei, 610 denarii, 2 asses and 1 quadransgiving a closing date of c. 166 AD (Boos et al. 2000, 64-65). The lead seals from the Roman fort of South Shields, England mainly consist of three male heads, one of which is bearded, and have been taken to represent Septimius Severus together with his two sons Geta and Caracalla. The main arguments for this reading are that (1) the South Shields‘ lead seals with triple portraits come from early rd 3 century AD dated contexts, suggesting a correlation with the military activities of the Severi at Hadrian‘s wall and the Scottish campaigns in 209-211 AD (Allason-Jones, Miket 1984, 326-330, nos. 8.7-21); (2) the bearded figure, regarded as Septimius Severus, is the central figure and is the only one to wear a wreath. The assumption matches the historical events: Caracalla received the title of Caesar in 195 AD and Geta in 198 AD when Caracalla became co-emperor and he was given the title Augustus. Geta obtained the title of Augustus only in 209 AD. The seals representing Septimius Severus and his two sons hence have to be dated between 198 and 211 AD. Opposite to Richmond, who proposed that the arrival of provisions in bulk must have arrived not later than 209 AD (Richmond 1936, 99), there are elements that prove the supplies must have arrived between 209 and 211 AD since some seals show the imperial trio with the letters AVG inscribed above or over their heads (Allason-Jones, Miket 1984, 327, nos. 8.9; 8.17). 7.4.4 Pendants Like the beads, pendants almost certainly received an aesthetic appreciation by the Romans, but a symbolic meaning given by the wearer of a specific pendant is not excluded, given the disk-pendants with pressed images of Classic, Christian and Jewish symbols or specific emblematic animals referring to mythological or biblical stories and/or figures. The pendants with talismanic connotations are to be interpreted as amulets or apotropaic emblemata, bringing good fortune to the wearer or to ward off the evil-eye. Within the scope of this project it was, however, impossible to go beyond the rather one-dimensional significance and to examine the possible relations between the configured images and the written sources. The fact that similar pendants received Christian, Jewish and Pagan symbols may point to the triviality of religion to the glassworker who made whatever the customers desired according to their religious faith. The open-mindedness in manufacturing religious trinket for people with another conviction allows us to assume that the late Roman society was tolerant vis-à-vis the people‘s religious conviction even though it is hard to provide evidence. It at least stands in contrast with the polemical publications of contemporaneous pagan and Christian writers. Furthermore the uniformity of the glass jewellery demonstrates a homogeneous and standardized material culture during the late Roman period that is not influenced by the conviction of the customer. Danièle Foy (2010, 310) mentions work by Sophie Kauffmann (2007) to propose that certain pendants possibly were considered Christian symbols by the Church whereas others might have been banned as heretic material. Due to the lack of good evidence it is impossible to prove that these little knick-knacks were either customary to the entire Christian community or whether they received particular emblematic meaning from a particular faction within the early Christian church segregation such as the Arian or Monophysite. It should be checked from the written sources th th resulting from the various synods of the 4 -5 century AD whether these little knick-knacks were forbidden, tolerated by the Catholic Church. In Chapter 3 we described the range of pendant shapes made of black glass. Besides the discoid medallion, other pendant forms correspond to the shape of a vessel, an animal (e.g., a dolphin), a human head or bust, or a plant 235 (e.g., a poppy-head ). For practical reasons, we only discuss the well-known vessel-pendants shaped as a jug or jar, a barrel or a vase. In particular, we checked whether the crudely shaped vessel-pendants match particular vessel-shapes and whether less distinct pendant types also could be linked to vessel shapes. The form of the jugshaped and jar-shaped pendants explains that these pendant types with ovoid or a globular body and vertical 236 handle represent real glass vessels shaped in a coarse style, corresponding with Isings Forms 120-124 (Figure 129). Marianne Stern considered these jug-shaped pendants as Christian amulets (Stern 1977, 112-113) but rejects Gustav Eisen‘s theory that they are to be connected to the two jugs Joseph of Arimathea had with him in his sarcophagus which contained the blood and the sweat of Jesus Christ (Eisen 1927, 520-521). These jugshaped pendants could equally have received a talismanic character to Jewish or Pagan people as is already established by the armrings with stamped decoration showing pagan, Christian or Jewish motifs. It is possible that 235 In Chapter 3, we explained that unlike others (e.g., Spaer 2001) we do not consider the bulbous pendant with discoid base as a vessel-shaped pendant but assume it represents a poppy-head. 236 Something we have left out of consideration is to what vessels in other materials the pendants refer. 288 it is not an import into the West from pilgrims who went to the Holy Land (Foy 2010). Still today some are confident that these vessel-shaped pendants have to be regarded as Joseph of Arimathea jug amulets (see online catalogue of Yale University, New Haven) [cat.no.A4697]. Figure 129: Comparison of jug-shaped pendants and their real counterparts in vessel glass (drawings taken from Spaer 2001 and Harden et al. 1989) The vase-shaped glass pendants most likely have to be interpreted as a cantharus considering the resemblance with the cantharus-headed pins in jet and bone (Allason-Jones 1996, 39-40). Lindsay Allason-Jones (1996, 16) remarks that the cantharus is a drinking vessel related to the god Dionysus/Bacchus but also characteristic to the Christian Eucharist. Wearing a trinket ornated with a cantharus, thus makes it impossible to determine whether there is any relation to Bacchic cult or Christianity. A wide variety of pendants are vessel-shaped, but some are less clearly related to an existing vessel shape, such 237 as the barrel-shaped pendants which appear to match the elongated barrel-shaped beads. We thus want to demonstrate that these more amorphous and geometric-shaped pendants are in fact stylized representations of particular vessel types. Furthermore we discuss that these vessel-shaped pendants are the counterpart of real vessels in bronze, (glazed) pottery and glass related to ritual act. Both barrel-shaped beads and pendants show identical form, proportions and decorative designs with applied trails spiralling at both ends, and eventually with zigzags at the centre. The barrel-shaped pendants, both large and small, differ from the beads by the presence of one or more suspension loops. The presence of one or more suspension loops is also attested on real barrel-shaped vessels. There is a sort of resemblance with for instance the barrel-shaped bottles (Morin-Jean Type 132; Isings Forms 89;128), but more striking is the similarity with some blown-glass vessels, such as the barrel-shaped beakers from the Rhine region (Fremersdorf 1959, 67-68, pls. 94-95) having one vertical looped handle at the centre of the body with trailed decoration on either end of the vessel (Figure 130-left) and the so-called perfume-containers, small horizontal barrel-shaped bottles with four small foot knobs and a short neck with dolphin-shaped handles (Morin-Jean type 135) (Figure 130-middle) (MorinJean 1913, 177, fig.233). Various pieces from the Netherlands are reported like a small colourless piece with opaque yellow glass trails from a tomb at Stein (NL) (Isings 1971, 16, no.38, fig.3:38), or that from Brunssum (Pirling 1993, 213, fig.2:6) and Maastricht (van Lith 1987, 54-55, figs. 9-10). It is significant that nearly all are confined in the wider Rhine region from Krefeld-Gellep (DE) as far upstream as Kaiseraugst (CH) (Fünfschilling 238 2000). An incomplete piece in olive green-tinged colourless glass with twelve spiralling opaque yellow glass trails at either ends has been found in tomb 3920 of the late Roman-early Frankish cemetery at Krefeld-Gellep 237 Maud Spaer takes these trailed decorated beads and pendants together and describes the pendants as trailed bead-pendants (Spaer 2001, 113, nos. 173174). 238 The only known exception thus far is the Vatican piece mentioned by Fritz Fremersdorf (1975, 73, no. 708, pl. 30:708) which is assumed to have come from a catacomb tomb seeing that the piece contains remains of mortar from being walled up in a loculus, similar to the gold-glass vessels from the Roman catacombs. 289 (Pirling 1997, pl. 37:4a-b; Pirling, Siepen 2006, 294, Gellep 714, pl. 46:714). The top of the central part, which must have contained the handles and mouth, is missing, but the basal part shows four massive knobs bulging out which serve as feet for the vessel. Pirling‘s reconstruction is presumably based on the intact piece from Bonn (DE) in colourless glass with opaque pale blue glass trails spiralling at either ends and decorating the rim (Follmann-Schulz 1988, 127-128, no. 495, pl. 55). The vessel has a short cylindrical neck tapering towards the horizontally folded rim and two vertical dolphin-shaped handles. Plenty of other analogous pieces come from Cologne (DE) (Fremersdorf 1959, 67-69, pls. 91; 96-99; Doppelfeld 1966). Seeing that most examples have been found in Cologne and made in the same way as the colourless snake-thread vessels, it is tempting to consider nd rd these vessels as having been produced in Cologne at the end of the 2 and first half of the 3 century AD. Those th rd th in greenish glass are dated 4 century AD as is discussed by Sylvia Fünfschilling (2000, 246). This 3 –4 century AD date fits perfectly the production of the trail-decorated, elongated barrel-shaped beads and pendants in black glass. Figure 130: Comparison of barrel-shaped pendants (left: Trier - DE) and their counterparts as real vessels in glass (Cologne DE) and bronze (Veldwezelt - BE) (drawing left by author; drawing glass vessel taken from Harden et al. 1989, no.54; picture right taken by VIOE, by courtesy of VIOE, Tongeren) Parallel pieces in other materials are known as well: for example, the bronze unguentarium from Vertault (FR) with a foot and two suspension loops fitting a handle (Feugère 1994, 157, no. 46, fig.17). The piece is 73 mm long and has a maximum diameter of 28 mm. Or the bronze barrel-shaped vessel on central stemmed foot from a ritual deposit at Veldwezelt (BE) which is determined as a lamp or incense burner (unpublished) (Figure 130 right). Other comparisons can be made to the horizontal barrel-shaped flasks in glazed ceramics with a twohandled neck in the middle of the body but without feet (Pirling, Siepen 2007, 294, Gellep 714; 285, Gellep 214). Tomb 249 of the late Roman cemetery at the Jakobsstrasse in Cologne (DE) yielded a barrel-shaped ‗beaker‘ in terra nigra, type Gellep 96, measuring 57-58 mm in height and 35 mm in diameter (Friedhoff 1991, 298, pl. 101:1/13). The remains of white paint on the concentric ribbings at either end give a contrasting effect to the vessel referring to a wide range of barrel-shaped beads and pendants. But perhaps we should speak of a beaker, 239 particularly seeing the very small dimensions which approximating those of Turkish tea cups. Furthermore, the recipient contained a follis minted between 313 and 317 AD representing Constantine I. Interesting to add is the resemblance of Renate Pirling‘s Gellep 214 barrel-shaped glass jug with one small handle with spiralling applied glass trails at either end (Pirling, Siepen 2006, 285-286). A comparable vessel, but without handle, was found in Maximsstrasse, Trier (DE) (Goethert-Polaschek 1977, 149-150, Form 92a, no. 900). 7.4.5 Gems It is well-known that gems were considered precious and that they were very much appreciated for the embellishment of various types of jewellery. It is thus not surprising that the gems in monochrome or bichrome black glass, like any other, appear to have had multiple purposes during the Roman imperial period. Previously we established various types of glass gemstones of which a good number are still set in the original piece of jewellery [see Chapter 3]. The Roman gemstone, fully and partly manufactured in black glass, received a wide range of uses. Perhaps for the most part set into the bezel of finger rings, some glass gems have been set in brooches or even helmets. In this section we verify 1) the relationship of specific types of black glass gemstones with particular commodity categories and 2) the differences between the material of the piece of jewellery wherein the gemstone is set. To assess the use and function of gems, the observation is twofold, as the gemstones have 239 Ulrich Friedhoff in his note 12 of page 109 refers incorrectly to Trier Form 46 which is a globular beaker with applied spiralling glass threads and tapering towards the rim. 290 to be considered in terms of the function of the commodities wherein they are set. We therefore looked to the gems and the ‗carrier‘ one by one, taking into account shape, colour (combinations) and decoration. 240 241 A range of variations have been observed from the 340 recorded gemstones : 1) the glass intaglios can be manufactured in monochrome black glass (40 ex. or 11,8%) or in combination with another glass hue such as the blue-on-black nicolo imitations (300 ex. or 88,2%); 2) the gems are mainly elliptical but can be circular; 3) the decoration can be engraved or pressed. A specific use per type of gem emerged when taking into account those gemstones still in their original situation. This observation allowed us to ascribe the masses of gemstones found loosely to an explicit commodity such as brooches, finger rings or necklaces. In a second stage, we verified what meaning can be ascribed to the various types of gems. A closer look at the gemstones demonstrates that in most cases pseudo-nicolo intaglios embellish the bezel of a finger ring (102 records out of 300 = 34%) while only few have been used to adorn helmets (5 records or 1,7%). All other 193 records are loose gemstones. It is significant that there is a marked difference in shape between the pseudo-nicolo gems for finger rings and those set in helmets. The finger ring gemstones are all small, elliptical in shape, and have a flat surface (Henig type F) and bevelled edge, either narrow with a steep angle (Henig variant 2) or wide with a gentle angle (Henig variant 4), whereas the helmet gems are large and circular with bevelled edges [see Chapter 3]. But the most striking difference is that the helmet gems remained plain while the finger 242 ring gems were never left undecorated. The brooch gems are solely occurring on circular and elliptical discoid brooches (Riha Typ 3.17; Böhme Typ 45; Bayley T270-271) [see Chapter 3]. This is a specific type of brooch that is connected with military dress, seeing the frequent and sole occurrence in military contexts [see Chapter 5]. The gems set in those brooches are mainly cast in monochrome black glass, but on rare occasions they are bichrome blue-on-black layered. The monochrome black glass gems are either plain conical or flat plano-convex with a moulded or pressed decoration 243 showing the head of Medusa. The black glass gem set in the gilded bronze brooch of Oudenburg [cat.no.480] represents an eagle pointed to the left with the head looking to the right and with a wreath in its beak. A close parallel can be seen in the golden gem in the golden finger ring of the Schützenstrasse in Trier (DE) (Krug 1995, 55, no. 17, Taf. 39A; 46:17; 57:17). Krug also mentions some other gems with a comparable motif. The bichrome gems are shaped as are the pseudo-nicolo gems with flat top and basal surface and bevelled edge 244 and decorated with impressed figurative motifs. The discoid brooches are always cast in bronze and completed by gilding, giving the object a certain status. Nonetheless, these gilded bronze brooches with glass settings are seemingly direct imitations of those in gold with a precious gemstone. Not interested in putting much emphasis on the contrast with their cheaper substitutes in glass, we assume (without much evidence) that the different renderings of discoid brooches are related to military rankings. It is, however, clear that a sort of uniformity is noticeable within the recorded assortment. Those with a simple plain conical glass stone are smaller and less sophisticatedly rendered compared to those with the moulded Medusa head inserted, such as the very elaborate piece from the Roman cemetery of Wancennes (BE) [cat.no.617] (Figure 131). This piece stands in great contrast with the rather basic and abstracted design of the cast discoid brooches with conical gems. The outer side of the gilded bronze plate is decorated in repoussé-technique with a row of eyes flanked with a row of pearls and an inscription ―PERSEUS CONCIDERA[T] ° CAPUD GORG[ON]IS‖ (trans.: ‗Perseus knocked off the head of the Gorgon‘). It is known that the Romans considered the head of Medusa to have apotropaic connotations (Henig 1984, 185; Pailler 1992, 79; 82) as fending off evil from the individual, a function adopted later in Christian iconography in the person of Saint George slaying evil personified as a dragon/monster. It is thus very tempting to see these plain conical glass settings as an abstract version of these Medusa gems and assign them equally an apotropaic meaning. 240 As mentioned earlier the catalogued material is very partial as it remained essentially limited to Germany (Krug; Platz-Horster) and the United Kingdom (Henig). A more detailed study of this specific commodity type comprising the material from a much vaster area is necessary to verify the findings of this investigation. 241 The word intaglio is here used as a synonym for gemstone, although it is for the glass gems somewhat contradictory because the word – derived from the Italian word ‗tagliare‘ meaning ‗to cut‘ – refers to the cutting, carving or engraving into a flat surface to create an image. The glass gemstones on the other hand are nearly all decorated by using a mould with an image in relief [see Chapters 2 and 3]. 242 Within the Merovingian period the pseudo-nicolo intaglios are regularly re-used in brooches as is illustrated by material from Trier. 243 It is not excluded that the so-called plano-convex counters were occasionally used as gemstones in the discoid brooches. 244 Because these bichrome gems remain rare we omitted them from the present discussion. 291 Figure 131: The very elaborate discoid brooch with inscription from the Roman cemetery of Wancennes (BE) (picture taken by author, courtesy by Musée Archéologique, Namur) As far as we know, the plain conical and ‗Medusa‘-head gemstones were only used on brooches of the same typei.e., the circular and elliptical discoid brooch, which is related to the military dress. Based on these observations, the loose finds can be assigned to such militaria. This not only implies that the loose ‗Medusa‘-head gemstone from Wanzoul (BE) [cat.no.624] was originally set in a gilded discoid brooch, but more important is that it most likely was of a retired military officer who got buried at a civilian cemetery. The alleged officer Wancennes (BE) who was buried at the cemetery ‗de Chaurnai‘ probably can be considered the owner of the villa 'la Couturelle' at Wancennes. When looking at the finger rings, we observe that the gems for finger rings are mainly blue-on-black glass paste gems, imitating the semiprecious nicolo stones (102 records out of 111 = 91,9%, but also monochrome black 245 gems imitating black onyx (9 pieces = 8,1%) do occur. From Table 138 it is clear that monochrome black glass or pseudo-nicolo gemstones adorned finger rings made in various materials. From the 111 gemstones still set in finger rings, we recorded 20 iron finger rings (4 monochrome black and 16 pseudo-nicolo gems); 75 bronze rings, of which 4 are silvered [cat.nos.3980; 4066; 4076; 4224] and 7 are gilded [cat.nos.3960; 3967; 3975; 4072; 4074-4075; 4213]; 8 silver rings (all pseudo-nicolo gems); and 7 golden finger rings (all pseudo-nicolo gems). Table 138: List of glass gemstones set in finger rings (n = quantity) Cat.no. Site n 1998 2089 3963 4068 4168 4235 4240 1881 3912 3913 3986 4021 4061 4226 4307 3960 3967 3975 4072 4074 4075 Regensburg (DE) unknown (DE) Carlisle (UK) Havering atte Bower (UK) Odiham (UK) Terling (UK) Tuddenham (UK) Frankfurt am Main (DE) Bardney (UK) Bartlow (UK) Chew Valley Lake Site (UK) Colchester (UK) Gorhambury (UK) South Shields (UK) Wittering (UK) Caerleon (UK) Carrawburgh (UK) Chessalls, Kingscote (UK) Housesteads (UK) Ipswich (UK) Kenchester (UK) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 245 246 Material finger ring gold silver gilded bronze Colour Type of gemstone bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome Henig type F2/4 Henig type F2/4 Henig type F4 Henig type F2/4 Henig type F2 Henig type F2/4 Henig type F2 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2 ? Henig type F2 Henig type F4 Henig type F2 All recorded white-on-black glass paste gems as an imitation of white-banded onyx are loose finds. It was impossible within the view of our research to take into account the typology of the finger rings as well, but we are aware that this would provide even more interesting information. 246 292 4213 3980 4066 4076 4224 221 454 567 1833-34 1878-80;1883;1885 1887 1945 1981 2067 2077;2079;208182;2085;2087-88;2100 2110 3907-8 3909 3915 3916 3961 3966 4002 4013;4016;4018-20 Silchester (UK) Chesterholm (UK) Harlow (UK) Kettering (UK) South Shields (UK) Braives (BE) Neerharen-Rekem (BE) Tourinnes-St-Lambert (BE) Burgheim (DE) Frankfurt am Main (DE) Germersheim (DE) Mainz (DE) Neuwied-Niederbieber (DE) Trier (DE) unknown (DE) 1 1 1 1 1 1 1 1 2 5 1 1 1 1 8 Wiesbaden (DE) Alchester (UK) Ashel Barn, Kingscote (UK) Beauchamps, Wickford(UK) Bedward Row (UK) Camerton (UK) Carrawburgh (UK) Cirencester (UK) Colchester (UK) 1 2 1 1 1 1 1 1 5 4031 4043 4044 4052 4055;4058 4067 4071 4073 4089 4132 4160 4161 4175 4177 4179 4186;4188 4192-93 4211;4213 Combe Hay (UK) Corbridge (UK) Cow Roast Site (UK) Fordingbridge (UK) Godmanchester (UK) Hassocks (UK) Housesteads (UK) Icklingham (UK) Lockleys, Welwyn (UK) London (UK) Nettleton (UK) Newgate (UK) Poundbury (UK) Reculver (UK) Ribchester (UK) Richborough (UK) Saint-Albans (UK) Silchester (UK) 4225;4229 4231-32 4302 4309-10 257 1773;1778;1789-90 1877;1884 1902 2090;2103 3911 4011-12 4077 4124;4131 South Shields (UK) Springhead (UK) Whitton, Ipswich (UK) Woodeaton (UK) Elewijt (BE) Bad Homburg (DE) Frankfurt am Main (DE) Kobern-Gondorf (DE) unknown (DE) Auchendavie (UK) Colchester (UK) Kirkby Thore (UK) London (UK) 4169 4189 4210 4315 Old Kilpatrick (UK) Rodmarton (UK) Silchester (UK) Wroxeter (UK) 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 2 2 1 1 2 2 1 2 1 4 2 1 2 1 2 1 1 1 1 1 1 1 silvered bronze bronze iron monochrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome ? Henig type F2 Henig type F2 Henig type F2 Henig type F4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 bichrome bichrome bichrome bichrome Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome Henig type F2/4 Henig type F2 Henig type F2 Henig type F2/4 Henig type F Henig type F2/4 Henig type F2 Henig type F2 Henig type F2 (2) Henig type F4 (2) Henig type F2/4 (1) ? Henig type F2 ? Henig type F2 Henig type F4 + F2/4 Henig type F2 Henig type F2 ? Henig type F2 Henig type F2 Henig type F2 Henig type F4 Henig type F2 Henig type F2 Henig type F4 Henig type F2 Henig type F2 Henig type F2 ? Henig type F2 Henig type F2 Henig type F2 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 Henig type F2/4 monochrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome bichrome monochrome bichrome bichrome bichrome bichrome bichrome bichrome monochrome bichrome bichrome bichrome bichrome monochrome bichrome monochrome bichrome bichrome bichrome bichrome It is noticeable from Table 138 that besides the expected finger rings in bronzeplain, silvered and gildedand iron, glass gemstones in imitation of semiprecious stones are also used to embellish finger rings in luxurious materials such as gold and silver. Apparently, the Roman craftsmen used any material to manufacture finger rings in combination with glass gemstones. Conversely, semiprecious stones are equally occurring on cheap iron finger rings like the one from Caersws (UK) holding a brown chalcedony or agate intaglio (Zienkiewicz 1989, 73, no. 37, fig.37:37). These observations seem at first to corroborate the statement by Kathy Sas (1993, 110): People in Antiquity attached very little importance to the difference in quality and value between semiprecious stones, organic materials and glass pastes in manufacturing gems and other jewellery. The appraisal of the gem depended on the quality of the engraving, although specific preferences existed for materials and their colours in certain periods. (Kathy Sas 1993, 110) 293 However, when taking into account a passage from Pliny the Elder (NH XXXVII, 197-200) on real and fake semiprecious stones, we are more inclined to think that the Romans did mind the careful imitations of semiprecious stones and thus rather saw them as a fraud or forge. ..., we will make it our business to point out the methods of detecting these false stones, seeing that it is only proper to put luxury even on its guard against fraud. ...The modes of testing stones are numerous: first, by their weight, the genuine stone being the heavier of the two; next, by their comparative coolness, the genuine stone being cooler than the other to the mouth; and, next to that, by their substance; there being blisters perceptible in the body of the fictitious stone, as well as a certain roughness on the surface; filaments, too, an unequal brilliancy, and a brightness that falls short before it reaches the eye. The best mode of testing is to strike off a fragment with an iron saw; but this is a thing not allowed by the dealers, who equally refuse to let their gems be tested by the file. Dust of Obsian stone will not leave a mark upon the surface of a genuine stone: but where the gem is artificial, every mark that is made will leave a white scratch upon it. (Pliny the Elder Naturalis Historia XXXVII, 197-200) [translation by Bostock et al. 1855 - online @Perseus, Plin. Nat. 37.76]) The consumer thus either was not able to recognize genuine from false, or was well aware of the replicas but was not willing to pay the excessive prices for authentic stones. The gold finger ring with nicolo-imitating gemstone from the Regensburg-Kumpfmühl treasure in blue-on-black glass illustrates that a clear answer is beyond reach: 1) as it could well be considered a fraud committed by a Roman goldsmith and that has not been noticed by the compiler of the treasure seeing the three other golden finger rings have a genuine nicolo gemstone set in; 2) because it is equally possible that the fraud was ordered by the customer purely guided by snobbishness and greediness. Consequently, it is very plausible that another reason played a role of importance to set pseudonicolo glass gemstones into golden and silver finger rings than only forgeries of which the consumer was not aware. Since we are unaware of any study on the composition of the golden and silver finger rings, an analysis of the chemical composition of the finger rings would be useful to distinguish a difference in quality of the precious metal between those with nicolo gems and those with pseudo-nicolo gems. Dave Zienkiewicz describes that the rendering of the Caersws-gemstone is rather rudimentary and considers it a cheap gem for trinket jewellery any auxiliary soldier could afford (Zienkiewicz 1989, 73-74). This hypothesis is challenged by similar renderings on gemstones set in golden finger rings such as the one from the RegensburgKumpfmühl treasure (DE) showing an incised stylized bird-like figure which is assumed to represent a peacock (Boos 2000, 23, no. 5, pl. 6:4) [cat.no.1998]. The style of the rendering is in all probability influenced by sequential fashions as explained by Kathy Sas (1993) [see Chapter 4, Table 68] rather than a calculation of costs. It is clear that the idea of inexpensive semiprecious gemstones being inserted in cheap materials such as iron finger rings or ‗economically cut with a minimum of wheel-cut lines‘ is a biased assumption based on modern ideas and perhaps personal appreciations, and it does not recognize the use and meaning the Romans gave to materials, colours and symbols. It is, however, generally known that during the early Roman imperial period only people of senatorial rank and the equites were authorized to wear a finger ring in gold (Pliny, NH XXXIII,29). It is only rd around the transition to the 3 century AD that golden finger rings in the army become more widespread, not excluding that the retrieved golden finger rings were property of one or more members of the equestrian or the senatorial order active in the Roman army (Boos 2000). Thus it might well be that the auxiliary soldiers were only allowed to wear an iron finger ring but without restrictions in the sort of gemstone. It is also generally known that in early Roman times, precious and semiprecious stones had great value, making these expensive stones easier to keep safe compares to a bag full of coins. ‗Helmet‘ gems made of black glass are the bichrome pseudo-nicolo gems with a characteristic shape. These large, round pseudo-nicolo gems are nearly all loose finds from Roman civil settlementsi.e., France (Sens), Switzerland (Augst; Kaiseraugst), and Portugal (Conimbriga). The gilded silver sheet military helmet from Berkasovo (SB) [cat.no.3108] is therefore imperative for the interpretation of these loose finds. The Berkasovo helmet is decorated with a large number of stones, of which eight pieces are large circular pseudo-nicolo gems (Mackensen 2009, pl. 25) (Figure 132). We are, however, aware that it is somewhat premature to assume from the Berkasovo helmet that all large round pseudo-nicolo gems were exclusively produced to be used to adorn armor, although we believe it is most likely that these gemstones were attached to the ornamental garments of official dress, if not solely military dress. Furthermore the Berkasovo helmet demonstrates that we have to consider the so-called medieval ‗Alsengemmen‘ as re-used Roman artefacts [see Chapter 4]. Whether it be from Roman military dress or not, it is clear that the use and function of the large, round pseudo-nicolo gems received a totally different use and meaning through time. 294 Figure 132: large pseudo-nicolo gemstones set in a gilded silver sheet military helmet from Berkasovo (SB) (picture taken from Mackensen 2009, pl. 25). 7.4.6 Hairpins Asserting from the beginning that the pins in glass are hairpins might seem presumptive. Nevertheless, none of the recorded black glass hairpins can provide any direct evidence on their use. The large amount from Les Houis near Sainte Menehould (FR) have been found in a workshop context and the few others known so far, from Liberchies (BE), Trou de Han (BE) and Oudenburg (BE), come from a settlement, a ritual deposit and a military camp, respectively. Besides the assumed use in headdress, it has been described on several occasions that this type of artefact is related to cosmetics (Harden 1936, 285; Allason-Jones, Miket 1984, 275). … kohl-sticks, i.e. tapering rods for putting on eye-paint and other toilet uses… The kohl-stick of glass is much rarer than its parallels in other materials, e.g., bone or bronze. (Harden 1936, 285-286) … their shortness and the bulge of the stem would make them inefficient at keeping long hair in position, and impractical in fastening clothing. It is possible that they were used as ‗droppers‘ for unguent bottles as the spherical heads rest comfortably in the necks of Isings Forms 27 and 82, and would efficiently hinder evaporation of perfume or medicine. (Allason-Jones, Miket 1984, 275) More information about their use in headdresses had to come from elsewhere. On the one hand, additional information is provided by the glass pins in other hues which have been found in burial contexts, such as the four pins in pale green glass originating from the inhumation tomb G537 at Butt Road, Colchester (UK) (Crummy 1983, 28, nos. 461-464). All four were found together close to the woman‘s head. Most likely, all four were used together to keep the hair in a certain style, and it is very likely that the pin with annular head was used jointly with one of the three pins with globular head. The pin with annular head, with an opening of c. 6 mm, left enough space for another pin to pass through, given that the maximum diameter of the shaft is c. 5 mm. That a headdress might have been sophisticated is not only lavishly explained by Janet Stephens (2008) but also demonstrated by the seven pins in blue and green glass beneath the head from the inhumation tomb of a woman at Dorchester (Roach Smith 1848-1880, III, 34, pl. XI.3). More burial contexts yielded similar-appearing pins in other materials, e.g. jet (Allason-Jones 1996) and bone (Crummy 1979). Confirming the use of such short pins in headdress, it turns out to be that these pins may be identified as hairpins even if it is never excluded that they had one or more additional applications. 295 7.5 Architectural decoration 7.5.1 Tesserae The small cubic blocks chopped from discoid glass cakes or raw glass chunks are undeniably used as inlay stones for mosaics, but some scholars assume a supplementary meaning for the tesserae as being used as raw material in the bead production (Lundström 1976, 5; Foy 2007), something we also described by means of material from the workshop ‗Äussere Reben‘ at Kaiseraugst (CH) in Chapter 6. When looking at the mosaics, there seems to have been a preference to use glass tesserae in walls and vaults. In Chapter 5, we described the use of glass tesserae in Julio-Claudian times to adorn nymphaea and fountains but also bath complexes and burial architecture such as arcosolia. In late Roman times, glass tesserae were very much adopted to embellish the walls and vaults of basilicas. Floor mosaics, mainly polychrome ones, sometimes incorporate glass tesserae in various colours, even though rd the black tesserae used in floor mosaics appear to have been made of stone and not of glass. The 3 century AD floor mosaic of Mariamin near Hama (SY), representing young female musicians on a podium, is mainly made of stone tesserae but contains also glass tesserae to accentuate details in the dress of the musicians as well as in the border (Zaqzuq, Duchesne-Guillemin 1970, 94). Interesting here is the variation in the number of tesserae per dm² depending on the zones where the count was done, i.e. closely linked with the design of the image. For the body of the ‗Spring‘ epitome, 491 tesserae per square decimetre were counted, while 455/ dm² were used for the face of one of the figures in the border, and only 323 pieces per dm² for the main surface of the central panel, representing a podium built from wooden planks. Depending on the rendering of the depiction, smaller or larger tesserae were used. The dimensions of the mosaic stones range between 4-5 mm in length and 3-5 mm in width, which coincides with those in black glass from the glass workshop at Les Houis near Sainte Menehould (FR) [cat.nos.1411-12]. If we go back to for instance the floor mosaic of Mariamin measuring 5,37m x 4,25m, the 247 surface of c. 22,82 m² or 2282,25 dm² must count c. 1 million tesserae. Others calculated an average of 15 to 20 kg of tesserae to cover 1m² (Brun et al. 1991, 48-49), meaning that the Mariamin floor mosaic necessitated between 342 and 456 kg of tesserae. These huge numbers put in perspective the relatively high number of 455 black glass tesserae retrieved from Les Houis and now in M.A.N., Saint-Germain-en-Laye (FR), which is merely sufficient to cover about 1 dm². Conversely, the inevitability of such enormous amounts of tesserae by mosaicists when paving floors, walls and vaults illustrates a very time-consuming groundwork to chip the glass chunks into millions of mosaic stones. We therefore assume that the accumulation of tesserae from the workshop at Les Houis could equally have been prefabricated material so that mosaicists could pick up ready-made bits and pieces in all shapes and sizes and in a wide variety of colours. This hypothesis does not exclude the aforementioned assumption by Agneta Lundström and Danièle Foy the likely use of tesserae as raw material in the bead making. Another striking element of the polychrome floor mosaic of Mariamin is the very limited but apparently deliberate occasions on which glass tesserae were implemented to accentuate elements, yet none of the glass tesserae in the discussed mosaic are made of black glass (Table 139). Table 139: Areas where glass tesserae have been used on the mosaic of Mariamin nearby Hama (SY) representation hue reference Belt around the body of a bull Band around middle of a putto Small sharp knife in right hand of a putto Shoulder strap around a hunting putto with bow emerald green pale blue green emerald green Zaqzuq, Duchesne-Guillemin 1970, 99, fig.9 Zaqzuq, Duchesne-Guillemin 1970, 100, fig.8 Zaqzuq, Duchesne-Guillemin 1970, 100 Zaqzuq, Duchesne-Guillemin 1970, 100, fig.6 The intarsia are also used as mosaic inlays by means of the opus sectile technique and will not be discussed separately. The only difference with the tesserae is the production technique and the resulting characteristics in use. Where the tesserae are very small and roughly shaped to compose pointillistic images, the intarsia were cast 247 To estimate the total amount of used tesserae, we have counted half of the surface with the smallest mosaic stones and the other half with the largest examples what resulted in 560.231 small pieces (1141 dm² x 491 pieces/dm²) and 368.543 large pieces (1141 dm² x 323 pieces/dm²) or a total of 928.774 tesserae for the entire surface. Taking into account the minuscule size of 1-2 mm for the glass tesserae from the Nero‘s Domus Aurea in Rome (Sear 1977) the total amount of chips and flakes to cover the same surface increases exponentially. 296 or cut into a specific ‗jigsaw‘ shape forming a comparatively large part of the image. The very brief review here is because little is to be said at present on the basis of only black glass intarsia. 7.5.2 Rods The use and function of black glass rods, twisted or plain, can only be considered in view of the entirety of glass rods, regardless from the hue and not just by limiting it to the accumulation of rods in black glass only. The brief overview of prior investigations discussing the use of glass rods includes different readings of this type of commodity. Two main groups of rods can be distinguished: 1) those for architectural decoration; and 2) those used as utensils. Various features can help differentiate the architectural twisted glass rods from the twisted rods used as utensils when incomplete. The rods used as borders of polychrome mosaic panels can easily be distinguished from the so-called stirring rods, as the ‗architectural‘ rods 1) are not produced in clear transparent glasses but in a 248 translucent glass; 2) are much more worn ; and 3) usually have remains of concrete all over the surface (if not fully removed at the conservation and restoration lab of the museum). Besides, in architecture only twisted rods in monochrome or bichrome glass seem to have been used, whereas the glass rods used as utensils can be plain as well. The latter group has characteristic ends, commonly with a looped end or a bird at one side and a pointed or discoid end at the other side. Fulfilling all these features, the bits and pieces from the KMKG-MRAH, Brussels are clearly remains of an architectural decorative wall or vault panel. The twisted rods were inserted at regular intervals in the mosaic panel parallel to a row of shells (Sear 1977). But multiple twisted rods placed side by side also occur to form a wide frame for a polychrome figurative mosaic panel or a black-and-white inscription (Goldstein 1979, 263-264, no. 791; Grose 1989, 358, fig.170). That twisted glass rods were used to border the polychrome wall mosaic panels is clearly described by Frank Sear in his ‗Roman Wall and Vault Mosaics‘ (1977). The report on a wall mosaic from the ‗Villa on Posilipo‘ near Naples (IT) has been considered the first to ascribe the twisted glass rods as ―forming part of the decoration of a mural mosaic‖ (Gunther, Manley 1912, 104). However, Emile de Meester de Ravestein already mentioned about these twisted rods that they were used to ―… cacher les jointures des plaques de verre dont on ornait les murs des appartements‖ (de Meester de Ravestein 1884², 463). Most twisted rods, as far as attested, have been inserted in wall mosaics of the fountain niches of Nymphaea in private houses. Other places of incidence are bath complexes such as the bath complex of Aquileia (Sear 1977) and the natatio of the ‗Villa dei Centroni‘ in Rome [cat.no.2278] (Figure 133a-b). Twisted rods have also been employed to decorate funerary architecture, as is seen in the columbarium of Pomponius Hylas (Sear 1977) [see Chapter 5]. This type of architectural decoration element was particularly popular within a relatively short time span during the st st very beginnings of the Roman imperial period, i.e. from the late 1 century BC to the mid 1 century AD, whereas nd the stirring rods are considered also to have circulated during the Flavian period and the first half of the 2 century AD [see Chapter 4]. Numerous functions have been attributed to the utensils. Clasina Isings (1957, Form 79) described such objects rather vaguely as stirring rods, whereas Sidney Goldstein (1979, 263-264, no.791) and David Grose (1989, 358) speak about ―cosmetic applicators‖ aside from ―stirrers or stirring rods‖. Dominique Simon-Hiernard (2000, 373378) explains more in detail the use and speaks about the need to spread out, mash and blend cosmetic products. Marianne Stern (2001, 365, no. 228) demonstrated that other functions are attributable to these twisted rods: 1) to point at a text when reading a papyrus (Knauer 1993, 25) and 2) to spin yarn (Barber 1991). On very rare occasions, twisted rods were used to make handles, such as the cup at the Newark Museum (Auth 1976, 97, no. 105) or the two beautiful kantharoi of the former Constable-Maxwell collection (Constable-Maxwell 1979, 50, nos. 66-67). The latter can be difficult to distinguish from fragmented arm rings of Type A2-3. The bichrome plaques are not discussed separately here as little information on their use and function is available, and this material category did not receive the full focus in this study. 248 This degradation of the glass surface is probably due to the impact of the humidity conditions and the mortar wherein the rods are set. 297 a. b. c. d. e. f. g. Figure 133: Examples of wall and vault mosaics with twisted glass rods inserted: a.-b.: natatio, villa dei Centroni, Rome; c.-d.: bathcomplex, Aquileia; e.-g.: nymphaeum, Casa del Torello, Pompeii (illustrations taken from Sear 1977, pl. 15:2-3 [a-b]; pl. 44:3 [c]; pl. 43:4 [d]; pl. 9:1;3 [e-f]; fig.13 [g]) 298 7.6 Counters It is clear that in use throughout five centuries and all over the entire Roman Empire, glass counters unmistakably must have had multiple purposes. The study of this rather ubiquitous commodity is however quite recent and the discussion on its use still remains underdeveloped. Gustavus Eisen remarks in his paper on button beads (Eisen 1916, 299) that they never have been the object of serious study, albeit their frequent occurrence in publications. Gustavus Eisen, himself saw the plano-convex glass counters as button-beads, which were used in pairs cemented together back to back and strung as necklaces (Eisen 1916, 299). This idea got countered by Donald Harden (1936, 292), who claimed that the mentioned example is a forgery made from genuine material meant for the antique market. According to Kate Elderkin (1928) the counters are to be described as clothing-fasteners like the pierced examples, and she explained the pieces lacking a hole to have been mounted in a holder. Donald Harden (1936, 292) is not negative for the proposal but questions why there are so few pierced compared to the high quantities of solid plano-convex counters. Donald Harden (1936, 291-292) on his turn suggested that the irregular and oval shaped counters were employed in metal holders as gemstones in jewellery. On the basis of information by ancient writers and some remarkable archaeological finds, the plano-convex glass artefacts are generally described as gaming pieces used in various Roman board games, such as ludus 249 duodecim scriptorum and ludus latrunculorum (Austin 1934; 1935). Gaming sets including 24 to 30 counters and a board such as those found in late Roman tombs of Lullingstone (UK), Krefeld-Gellep (DE), Amiens (FR), and Nijmegen (NL) are plain examples. Furthermore, written sources support this idea; Pliny the Elder (NH XXXVI, st 199) explains that (in the 1 century AD) the making of calculi in all kinds of colours was done by the recycling of broken glass and names these pieces oculi for gaming boards [see Chapter 2]. The gaming sets mainly encompass counters in two contrasting colours, which means that the games involved only two players. On rare occasions, sets show four contrasting colours, two dark coloured (black and blue) and two bright coloured (white and yellow, blue or green). One can wonder whether this was to enable games with four players or whether all dark and bright stones were used as a set for only two players. An alternative utility put forward was that these counters could have been used in administration (Price 1995, 129). For instance, it has been proposed to call the small pieces with a diameter below 20 mm as counters and the large ones, with a diameter ranging between 27 mm and 32 mm, as gaming pieces (Price 1985, 214). In the same line is the explanation describing the use of glass counters for arithmetic intent as counting pieces, calculi, with a transportable counting board, abacus or quinarius, which principally included 24 counters (Schärlig 2003, 518519). The high number of glass counters concentrated in one place in the Pompeian house with the Indian statue, Insula 8.5 (Fontaine 2009), could in all probability be connected with such sets of abaci used during arithmetic lessons, rather than collected glass waste for recycling purposes. This plausible but hard-to-prove interpretation of the remains then assumes the room possibly functioned as an improvised schoolroom just before the calamity by the fatal eruption of Mount Vesuvius. But domestic classes with a private teacher and personal teaching material were not unusual in ancient times (Casson 1998, 17). Other functions can be expected too, such as an inlay in jewelleryfor instance, a gem set in a finger ring, brooch or braceletexplaining the single discoveries such as a small black glass counter from the auxilliary fort of Segontium at Caernarfon (UK) [cat.no.A4536] (Casey et al. 1993, 210, no. 514). The potential decorative function of these plano-convex counters is, as expected, not limited to jewellery. For instance, a late Roman mirror frame in plaster at the Musée du Verre, Charleroi (BE) (Musée du Verre 1999, 51, no.11) [cat.no.A4611], shows five black glass counters in the corners, one large at the top and four small at the sides (Figure 134). An example of implement as architectural decoration can be seen in the stucco ceilings of the Domus Transitoria in Rome, where large numbers of counters have been inserted within the design of the frescoes (Bacchelli, Pasqualucci, Mastrodonato 2000, 86). 249 The former game included 24 gaming stones, 12 per player as the name of the game says, whereas for the latter one is supposed to have consisted of 30 gaming pieces, 15 per contrasting colours, of which black and white formed the most characteristic set. 299 Figure 134: plaster mirror from Egypt decorated with plano-convex counters in black glass (picture taken from Musée du Verre 1999, 51, no.11, courtesy by Musée du Verre, Marcinelle) This brief enumeration of multiple functionality of the plano-convex glass counter makes the alleged difference between the smaller ‗counter‘ and the larger ‗gaming piece‘ introduced by Jennifer Price (1985, 206-214) [see Chapter 3] is of no use. Moreover, what name would then be given to the medium-sized pieces? Notwithstanding the multifunctional applications of glass counters, they in the first place must have been used as gaming stones, as can be deduced from a set of eye-catching burial gifts dating from the earliest times to the end 250 st of the Roman imperial period. The early 1 century AD ‗Warrior‘s burial‘ BF64 at Stanway (UK) not only contained a maplewood board but also 20 gaming pieces, 11 blue and 9 white (Crummy et al. 2007, 186-187, fig.89). From the ‗Doctor‘s burial‘ CF47, also in Stanway, comes another gaming board in the similar wood with 26 gaming pieces, 13 blue and 13 white, which appeared to have been placed in position on the board as if in st readiness to start a game (Crummy et al. 2007, 216-218, fig.89). A mid-1 century AD set of twenty counters in different colours, dimensions and materials (10 glass; 10 bone) comes from a tomb at Moers-Schafheim (DE), th near the military fortress of Asciburgium (DE) (Kraus 2005, 389). From a 4 century AD tomb at Lullingstone (UK) was retrieved a set of small counters (15 white and 15 so-called brown) all found on top of a lead coffin together with the remains of a wooden game board (Liversidge 1973, 350, fig.133l). Tomb B435 at the East Cemetery of London yielded a set of 24 glass counters (11 white and 13 black) jointly with four bone dice all within a wooden box (Barber, Bowsher 2000, 135, fig.97; 193-195). Other sets have been found in Amiens (Dilly, Maheo 1997), Krefeld-Gellep (Pirling 1966, 125-126, pls. B; 101:3) and recently in Nijmegen (van Enckevort 2008). The practice of placing a gaming set in a tomb is to give the deceased the equipment for the afterlife. Such sets remain rare as a burial gift, whereas single plano-convex counters have been retrieved more regularly. Individual counters seem to occur exclusively in inhumation tombs, which make that custom characteristic for the late Roman period. Not denying the earlier discussed possibility of a jewellery inset, this individual occurrence could equally be understood as a pars pro toto, functioning as a symbolic burial gift. The reasons why will remain silent since personal motive cannot be revealed without written sources. Occasionally, some texts survive and supply potential explanations and motivations. Large quantities of single finds have been retrieved during the intensive excavations of Roman forts on British soil, such as Corbridge (Bishop, Dore 1988, 204), Caerleon (Allen 1986, 155-156), or Usk (Price 1995), showing that board games were very popular with soldiers to kill time and that pieces got regularly lost. 250 A gaming board with calculi in yellow, white, blue and green glass is known from the late 1st century BC dated elite burial at Welwyn Garden City (UK) (Harden 1967, 14-16; Price 1995, 129; Cool, Philo 1998, 190, tabs. 34-35). The presence of an early Roman board game outside the empire demonstrates that such boards and joined glass counters must have been already fashionable during the Roman Republic. 300 7.7 Conclusion A general but superficial consideration of the use and function of the Roman black glass artefacts would have ended up in the simple conclusion that these artefacts are not more than a part of their functional categorye.g., a vessel is used as a container, a bracelet is used as a jewel and a counter is used for gaming. However, we have to bear in mind that a part of these commodities were not solely made in black glass just to supply the consumers market with objects in an additional hue. Particular reasons or specific purposes could have been part of the cause in giving a functional object an additional meaning. In particular, within a society like the Roman Empire featuring a material culture with a huge diversity of forms and a wide variety of materials, the utilized material and their related colour can be influential in giving the reached shape a particular function or significance. For instance, if the Romans used only one type of beaker and dish, these would be the sole objects to drink and eat from. But when a wide range of shapes are contemporaneously available, the diversity can be required for various functional or traditional purposes different shapes in particular hues depending on: 1) the drinks or dishes; 2) the circumstances (private or public) and types of events (social, religious, political); 3) the social class, gender or age. We, however, think that besides socio-cultural motives also socio-economic elements may have played a significant role. Socio-economical motives can make or break the popularity of a certain commodity in a specific material. The question is whether prevailing shapes are imitated in a type of material that becomes fashionable, such as glass at the start of the Roman imperial period, or whether this newly introduced material only slowly won importance on the consumer market, catching attention of the consumer by producing the usual stuff but in fancycoloured polychrome and monochrome glass. Different reasons can influence the rendering of an artefact. Firstly, the shape of an artefact is subject to changing fashions. The popularity of the utilized materials and shapes is susceptible to these trends. Through materializing the succession of trends, we can come to a detectable chronology [see Chapter 3]. Secondly, the shape of an artefact is dependent on the technological know-how of the glassworkers and the local/regional prevailing style. These geographical differences can infer distribution patterns [see Chapter 6]. Thirdly, the use of specific artefacts manufactured in a particular material, in a particular shape and with a particular design, can be inflicted from socio-cultural criteria. Where the inductive method of reasoning can provide a wide range of information on the first two, only little information is obtained about the latter through this method and imposes a deductive method of reasoning. By means of the bracelets, we pointed out that these Roman black glass artefacts were more than only supplying the consumer with a greater choice in the vast assortment of jewellery. Conversely, this category of Roman jewellery might nowadays be regarded as unrefined or even unappealing objects, making us wonder what must have been the reason of success, seeing the popularity of black glass jewellery in the late Roman period. A part of the answer has been given from a socio-cultural perspective. Merely made for aesthetical reasons, the black glass jewellery must have served as amulets or tokens. Besides socio-cultural motives, other reasons played a role. Seeing that wearing black glass bangles were clearly not for adornment purely, they had a certain meaning received in Roman times and eventually were used for specific purpose(s), as they were worn by only girls and young women. A finger ring is usually regarded as a type of jewellery that in Roman society has been worn by adults of both sexes, and due to the lack of clear evidence, the black glass finger rings do not contradict this assumption, as do the arm rings. On the other hand, it is generally accepted that the discoid and oval brooches have to be interpreted as types of dress accessories with military connotations and thus solely worn by male adults, without regard to the presence of an inlayed black glass gem. The use of black glass bangles cannot be related to a particular religion because they can contain Christian, Jewish, Pagan or Mithraic motifs, even if religion or superstition is never far away. Hence, it is also excluded that the black coloured bracelets symbolized sobriety, a highly valued virtue under for instance the sophists and the Christians. Conversely, the left-carrying side of all noticed black glass bangles from burial contexts is to be linked with an apotropaic meaning, likewise those in jet. A confirmation to this idea has been discussed previously by Lindsay Allason-Jones (1996) à propos the use and meaning of jet with references to ancient writers such as Pliny the Elder, Galen and Solinus, who wrote on the medicinal and apotropaic power of this fossilised wood. Such meaning and functional use can also be given to most late Roman pendants and for a number of beads as well. Even if this belief is situated within a twilight zone of superstition and symbolism, it would be wrong to neglect the significance of these religious trinkets because they act as metaphors of the late Antique cultural entity and picture well the customary traditions and beliefs, being evidence for a certain conservatism and traditionalism in late Roman times, something we find once more in the orthodoxy of shapes and colours. 301 Because it would be too far-reaching to investigate the black glass consumption more thoroughly, we thought it apt to continue the discussion on the Roman consumer behaviour vis-à-vis the consumption of black glass artefacts in Chapter 10. This essay on different levels is meant, insofar possible, to result in a more penetrating inquiry confronting the opposite approaches of micro-, meso- and macroeconomic level in consumer behaviour studie 302 PART II THE ARCHAEOMETRIC APPROACH 303 304 Chapter 8 CHEMICAL ANALYSIS 8.1 Introduction The archaeological approach in the preceding chapters has provided clear evidence of production of black glass artefacts all through the Roman imperial period and all over the Roman Empire. The production of black glass thus cannot have been considered an error. Otherwise, such material would then have been discarded instantly, excluding a systematic production and consumption of specific artefacts in black glass. Moreover, we can assume from, for instance, material of very rich burials and treasures that black glass artefacts were not considered misfits, but rather appreciated and cherished for their black appearance. Hence, it is clear that black glass was not produced unintentionally nor obtained by mistake. It is, however, not excluded that during the production of deeply coloured glass, some batches accidentally could have become so intensely coloured that they acquired a 251 black appearance. Jennifer Price and Sally Cottam state in their book on Romano-British glass vessels that the black appearance of glass is due to the intensity of the coloration of the glass metal, whether in yellow/brown, yellow/green, green, blue or purple glass (Price, Cottam 1998, 15). Their assertion was based on previous observations on the Roman vessel glass from Colchester, wherein ‗black‘ glass objects were systematically described as appearing black while the true colour of the strongly coloured glass matrix was observed (Cool, Price 1995). Such nuance is also attested in the publications on glass vessels from Avenches (Bonnet-Borel 1997; Martin Pruvot 1999) and Augst (Rütti 1991). Concerning the colorants used in antiquity to obtain a black glass, Justine Bayley attributes it to the addition of iron, but stresses the relativity inherent in expressing the observations made with the naked eye: A glass containing large amounts of iron, with or without manganese, appears opaque black. In fact, the glass is neither opaque nor black. If a thin enough piece is examined, it is usually seen to be transparent dark olive, the apparent opacity being due to the depth of colour rather than the presence of opacifying agents. (Bayley 1999, 90) The awareness of using various glass colourse.g., green, brown, purple, bluein Roman times to produce glass artefacts with a black appearance is definite for the archaeometric approach. The Roman glassworkers deliberately coloured glass with various specific strong colorizing oxides to obtain a glass matrix with a very deep hue. Hence, the Roman production of black-appearing objects was accomplished by using different chemical recipes. From an archaeological point of view, the question arises whether chronological and/or geographical idiosyncrasies are identifiable in the different compositions produced all through the five centuries of the imperial period and all over the empire. The results from chemical analyses supply additional information to answer one of the main concerns that preoccupies a material specialist in the study of archaeological artefacts: achieving a better understanding of Roman (economical) society through an optimized typo-chronology and a more nuanced understanding of the distribution patterns. We therefore assumed a diachronic study of a well-defined assemblage, such as the Roman black glass material having the potential to supply additional information about the organization of Roman glass production and consumption in general. Classical subdivisions have been obtained by looking at external properties of the artefactse.g., size, colour, shape, and decoration and production techniqueand, when possible, they are completed with information from the context from which the artefact has been extricated. The subsequent step was to verify whether functional types, chronological periods and/or geographical delimitations are idiosyncratic for a specific chemical 252 composition in (black) glass production and consumption. The chemical approach to the analysed black glass pieces is not exclusively confined to the understanding of the production, the distribution and the use of Roman black glass. The assessment also verified to what extent the data of this present study can be compared to Roman glass production and consumption during the imperial period. Accordingly, the obtained results have been evaluated by means of the current knowledge on Roman glass in general and the available results from prior analysis of black glass in particular. 251 The 12th century monk Theophilus described, for instance, the possibility of obtaining from the same batch, and consequently with the same chemical composition, a glass hue ranging from pale pink to nearly black or purple-brown (see Forbes 1966; Jackson, Smedley 2008). 252 Some significant results have been obtained for the ‗naturally‘ coloured and decolourized glass (Brill 1988; Nenna et al. 1997; Foy et al. 2000; Freestone et al. 2002; Freestone et al. 2009; Bingham, Jackson 2008; Foster, Jackson 2009; 2010). 305 A description and discussion on the entire evolution of chemical analysis on Roman glass is not within the realm 253 of possibility here because that would lead us as far back as Klaproth (1801). In fact, we can go back to Pliny the Elder, who also wrote on the composition of Roman glass, describing the complexity of producing glass and implying that it involved more than only mixing and heating sand and soda. Soon since man‘s skill is ingenious, he was not content to mix just soda, but magnetite, since it is believed to attract to itself the melted glass just as it does to iron. In the same way, shining stones came to be added to the melt in many places, then shells and pit sand. (Pliny the Elder, Naturalis Historia XXXVI, 192 [transl. Eichholz, D.E., 1962, Pliny X, 150]) The continuously growing number of chemical analysis methods and constantly improving applied techniques support the research on archaeological glass, resulting in more detailed data sets to answer more complex questions. However, they also make this part of archaeological glass research more intricate. Particularly during the past 15 years, amazing progress has been made through improved measuring techniques refining the analysis results. The applied measuring systems are SEM-EDX to characterize the major elements, and LA-ICPMS to characterize the minor and trace elements [see this chapter section 8.2. Methodology]. Important aspects in the study of Roman black glass include the network former (silica), the network modifier (soda) and the colouring agent. Although other silica sources do exist, sand is the main constituent in Roman times. Consisting, for the largest part, of silica (Si), sand also incorporates other elements like a number of metallic impuritiese.g., iron (Fe), titanium (Ti) and alumina (Al) (Henderson 1985; Brill 1988; Freestone 2006). Also lime (Ca), acting as a stabilizing agent making the glass resistant to water, seems to be present in the sand (Newton, Davison 1989). The use of different sands resulted in the production of raw glass with different ratios among the various oxides. The high melting temperature made it necessary to add a flux to the batch. Therefore, various alkalis were suitable. During the Roman imperial period, trona (Na 2CO3.NaHCO3.2H2O) and natron (Na2CO3.10H2O) were used as a flux to lower considerably the melting temperature of silica. These naturally occurring salt evaporites were the primary sources of soda for glass production during the entire Roman imperial period. They were extracted from the dry lake beds in the western Nile delta between Alexandria and the Wadi El 254 Natrun (Nenna 2000; Shortland et al. 2006; Nenna 2008a; Foster, Jackson 2009). Hence, it is not surprising that primary glass workshops have been recently pinpointed at Wadi Natrun, Taposiris Magna and Marea (Nenna 2005, 24). An interesting passage in Pliny‘s Naturalis Historia explains that soda is not too different from salt, being a multipurpose mineral collected for manufacturing soap, for medicine, for mummification, for dyeing and for cooking (Ignatiadou et al. 2005, 64-65), and that the Egyptian soda was considered of lower quality possibly because it was produced by artificial methods from Nile water and rain (Pliny the Elder, Nat. Hist. XXXI, 106-110). That the lower quality soda was good for dyeing and in particular for glass production makes it likely that the Romans only used Egyptian soda in the production of glass because of its low cost and not because of its quality. The use of mineral soda in Roman times gives a very characteristic composition with a high average of 15%-20% of soda (NaO2), a low concentration of magnesium oxide (MgO) below 1,5% and generally even below 1%, and a low potash (K2O) concentration below 1,5%. When vegetable soda is added to the batch, the composition of the glass shows a much higher MgO and K2O contentboth above 1,5%. Also, the more elevated ratios of manganese (MnO), phosphor (P2O5) and chlorine (Cl) are to be correlated with the addition of a vegetable alkali flux. Manganese could also be added deliberately to the batch to decolourize the glass (c.1%) or to colour it pink (c.1,5%), purple (c.2%), dark purple (c.2,5%) or black-appearing (above 2,5%) (Brill 1999). Also, antimony (Sb2O5) was used during the Roman imperial period to make colourless glass. Besides manganese oxide (MnO), the Romans generally used iron oxide (FeO and Fe 2O3) as a colouring agent to produce black glass, but cobalt (CoO) was also used. All these elements make possible a comparative study through SEM-EDX analysis to characterize and categorize groups on the basis of major elements, substantial for the archaeological research. The use of LA-ICP-MS refines these queries by making available a large number of minor and trace elements related to the used sand, flux or colouring agent. 253 The major publications from the 20th century that give an overview on chemical analysis of archaeological glass are Neumann 1925; 1927; 1928; Turner 1956ab; 1962; Sayre, Smith 1961; 1967; Bezborodov 1975; Biek, Bayley 1979; Bimson, Freestone (eds.) 1987; Henderson 1985; 1989; Hartmann et al. 1997; Brill 1999. 254 Various locations are known to have produced salt evaporites useful for glass production (Ignatiadou et al. 2005, 6467). 306 Table 140: List of sites with prior analysed black glass artefacts (* = LA-ICP-MS; ** = SEM-EDX; *** = AAS) late Bronze Age – Bronze Age late Iron Age Roman early Iron Age Lac du Bourget (FR)* Bouche-Rolland (FR) Neufchâteau-le Sart (BE)* Aquileia (IT)* Mons, Saint Georges (FR)* Pokravska (RU)** S. Vito Tagliamento (IT)* Mez-Notariou (FR)* Aveyron (FR)* Quintigny, Champ-de-Mont (FR)* Liberchies (BE)*** Roche Noire, Montagnieu (FR)* Taviers (BE)*** Tumulus de Courtesoult (FR)* Tongeren (BE)*** Hasanlu (IN)* Vervoz (BE)*** Pella (JO) Chotin (UKR) Prior research on the chemical composition of black glass has not been restricted to only Roman material (Table 140). All analysed pre-Roman black glass samples are limited to beads made from green glass. Bronze Age beads from various sites in France, Central Europe (Chotin) and Iran (Hasanlu) were analysed by means of LAICP-MS (Gratuze 2000; 2001c; 2002; 2004). The beads from Pella, Jordan, dated 1050-900/850 BC, have th recently been studied by using SEM-EDX (Reade et al. 2009); likewise, the Sarmatian glass beads from the 4 to nd 2 century BC dated kurgans from Pokrovka, Russia (Hall, Yablonsky 1997). The late Iron Age beads from the rd early 3 century BC dated chariot grave of Neufchâteau-le Sart (Gratuze, Cosyns 2007) have been analysed with LA-ICP-MS at the CNRS centre Babelon-IRAMAT, Orléans. rd th A number of 3 to early 5 century AD Roman bracelets and counters from Italy and southern France have been analysed at the same centre (Gratuze 2001a; 2001b; Moretti, Gratuze 2002). Another set of Roman black glass samples has been analysed much earlier by using atomic absorption spectrometry (AAS) (De Witte 1977, I, 6373). These artefacts are bracelets of green-black glass from various Belgian sites: Liberchies, Taviers, Tongeren and Vervoz. Due to the absence of comparative studies at that time, the conclusion remained restricted to the remark that all five late Roman bracelets from Belgian sites were produced using a mineral soda-rich glass with high iron content (De Witte 1977, I, 65-66). The iron was deliberately added in high concentrations as a colouring agent. However, the currently available analysis results on Bronze Age and Iron Age material demonstrate similar high iron content (Table 141). Various French sites show an iron content ranging between 14% and 17% that was used during the late Bronze Age and early Iron Age to obtain black-appearing glass beads. The Central European black glass th beads from Chotin have a similar high iron content, while the 9 century BC dated beads from Hasanlu have only a slightly lower iron content of 11,2%. The only exception comes from the four black glass beads from tomb 89 at Pella (JO), showing a much lower iron content. 307 Table 141: Chemical composition of black appearing era site/sample Na2O MgO Al2O3 SiO2 P2O5 Cl K2O CaO TiO2 MnO Fe2O3 CoO CuO NiO ZnO PbO SrO ZrO2 Bronze Age Lac du Bourget (FR)* - bead 13,2 0,5 1,5 61,1 1,1 - 1,3 3,9 0,1 0,03 15,2 0,002 0,03 0,002 0,03 - - - Mez Notariou (FR) - bead 13,9 1,38 2,4 68,5 1,8 0,6 1,33 1,9 0,2 0,07 7,1 0,001 0,18 - - 0,188 - - Enveitg, Brangoli (FR)◘ - bead 16,7 3,83 2,08 58 0,48 0,9 1,89 7,65 0,096 0,61 7,36 0,002 0,024 - - 0,32 - - Courtesoult (FR)* - bead 15 0,33 0,67 61,3 0,24 1 0,27 6,5 0,09 0,18 14,1 0,001 0,14 0,003 0,002 - - - Mons (FR)* - bead 11,8 0,33 1,5 66,1 0,18 0,5 0,67 4 0,17 0,67 14,3 0,006 0,6 0,005 0,009 - - - Quintigny (FR)* - bead 12,7 0,3 1,4 63 0,24 0,5 0,81 3,2 0,06 0,09 17,1 0,002 0,064 0,004 0,009 - - - Chotin (UKR)* - bead 7,7 0,4 1,6 67,4 0,73 - 0,9 5,5 0,1 0,5 15 - 0,001 0,01 0,009 - - - Hasanlu (IN)* - bead 17 0,9 0,8 64 - - 0,8 1,6 0 0,01 11,2 - 0,005 - 0,004 - - - Pella (JO) ** - Temple - bead 17,04 1,48 1,62 66,2 0,17 - 1,34 2,26 0,18 0 9,72 - - - - - - - Pella (JO) ** - Tomb 89 - bead 17,45 4,39 1,34 66,1 0,13 - 3,3 6,83 0,01 0 0,43 - - - - - - - Bouche-Rolland (FR)◊ - bead 14,6 0,93 3,09 59,2 0,17 0,4 0,7 7,23 0 1,8 8,43 0,002 0,034 0,003 - 0,58 - 8E - bead 14,1 0,69 2,75 63,6 0,09 0,9 0,87 7,1 0,054 0,11 9,56 0,007 0,014 0,002 0,003 0,003 0,047 0,005 8G - bead late Bronze Age – early Iron Age late Iron Age late Roman 308 Neufchâteau-le Sart (BE)*** 14,3 0,61 2,99 62,6 0,15 0,8 0,7 6,8 0,056 0,11 10,8 0,002 0,025 0,002 0,002 0,002 0,047 0,006 Pokrovka (RU) **** - bead 11,44 0,34 1,56 53,3 - - 0,41 2,83 0,542 0,03 11,68 0,001 0,261 0,005 0,017 17,25 0,017 0,018 Aveyron (FR)*** - bead 15,7 0,78 2,77 63,2 0,15 1 0,85 8,33 0,13 0,94 6,05 0,04 - - - - - - Aquileia (IT)*** - counter 15,2 0,47 2,56 63,2 0,14 1,2 0,72 7,52 0,07 0,51 7,87 0,26 - - - - - - San Vito (IT)*** 1 - bracelet 20 13,8 0,56 3,04 62,4 0,22 1 0,42 8,99 0,09 1,06 8,26 0,01 - - - - - - San Vito (IT)*** 2 - bracelet 21 15,3 0,55 2,88 60,8 0,19 0,9 0,49 9,31 0,08 0,96 8,3 0,03 - - - - - - San Vito (IT)*** 3 - bracelet 22 16,1 0,9 2,56 62,7 0,15 1 0,51 7,24 0,11 0,7 7,71 0,03 - - - - - - San Vito (IT)*** 4 - bracelet 23 15,3 0,57 2,65 64 0,2 1 0,52 7,35 0,08 0,76 7,28 0,06 - - - - - - San Vito (IT)*** 5 - bracelet 40 14,8 0,7 2,68 63,4 0,19 0,8 0,64 7,28 0,09 0,92 8,07 0,11 - - - - - - Liberchies (BE)***** 1 - bracelet 14,67 0,55 2,35 65,6 - - 0,66 6,08 0,11 0,4 6,14 - - - - - - - Liberchies (BE)***** 2 - bracelet 13,97 0,73 3,46 67,9 - - 1,6 6,47 0,16 0,35 2,37 - - - - - - - Taviers (BE)***** 1 - bracelet 14,84 0,47 2,37 63,6 - - 0,71 6,42 0,09 0,51 6,17 - - - - - - - Taviers (BE)***** 2 - bracelet 15,04 0,52 2,38 65,9 - - 0,98 7,06 0,11 0,46 5,18 - - - - - - - Tongeren (BE)***** - bracelet 17,08 0,86 2,3 63,8 - - 0,67 6,11 0,27 1,43 5,6 - - - - - - - Vervoz (BE)***** - bracelet 15,18 0,54 2,14 63,4 - - 0,78 6,34 0,08 0,39 6,66 - - - - - - - These beads are characterized by high magnesium and high potash content (HMHK) and are different from the contemporaneous temple area context, dated 1050-900/850 BC (Reade et al. 2009, 47-48, tab.1), demonstrating a much lower magnesium and potash content (LMLK). The former group is in fact comparable to the Bronze Age Nuzi glass (Reade et al. 2009, 48, tab.2). Hence, we may conclude that the knowledge about colouring the glass melt black by means of iron ore, ferrous salts or iron slags is not a Roman invention. Conversely the identification of a piece as Roman on the basis of its 255 high iron concentration is clearly too simplistic. The Sarmatian black glass bead was made of leaded soda-silica-lime glass but has been regarded as a Chinese production from the Han period on the basis of its chemical composition, in particular because of its low barium 256 content (Hall, Yablonsky 1997, 1243, tabs 1-2). The very high concentrations of titanium and lead oxide and a lower silica, magnesium and aluminium content distinguish the chemical composition of the Sarmatian black glass bead from other late Iron Age beads but relate it to very late Roman material (Table 141). A re-assessment of the pieces analysed by Hubert De Witte in 1977 is at present possible thanks to the current knowledge on ancient (black) glass. The Tongeren sample (no. 5), a bracelet of type D1, was made from a deviating raw glass demonstrating a much higher titanium and manganese content, corresponding to Ian Freestone‘s late Roman HIMT glass (Freestone 1994; Freestone et al. 2002) and Maurice Picon‘s ‗Groupe I‘ (Picon, Vichy 2003; Foy et al. 2003, 45-55; 80-83, Ann.1), even if the initial iron content, at impurity level, cannot be read by the addition of a high concentration of iron oxide. It can, however, be assumed, since the iron content in HIMT glass normally ranges between 1,3% and 2,3%. HIMT glass is supposed to have been introduced th somewhere at the start of the 4 century AD, but only became successful and used empire-wide by the middle of that same century. We will demonstrate that it is no coincidence that none of the four other Belgian samples were produced with HIMT glass [see below in this chapter]. Interestingly, Hubert De Witte already stressed that the type D bracelets have been manufactured by means of a distinct technology vis-à-vis the other types [see th Chapter 3]. The bracelet types A, B and C were produced up to the first half of the 4 century AD, whereas the th type D bracelets only came into use after the mid 4 century AD [see Chapter 4]. When comparing these considerations to the analysis results from the contemporaneous material from southern France and northern Italy, one immediately notices that there is no HIMT used here due to the low titanium content. The high manganese content of 4 out of the 7 samples is correlated to the decolourizing of the glass used to produce black glass, whereas the 3 other bracelets must have been made from ‗naturally coloured‘ bluish-green glass. When opposing the silica and soda content (Figure 135), the composition of all black glass samples appears to be a homogeneous silica-soda glass, except for the Sarmatian bead from Pokrovka (RU) and the Central European one from Chotin. The silica content ranges between 60%-70% and the soda content averages between 12%18%. The two East European beads differ somewhat from this homogeneous group: the Chotin bead has a significant lower soda content of 7,7%, while the Sarmatian bead differs on the basis of a lower silica content of 53,3%. It is striking to see that the Bronze Age beads have a similar SiNa ratio as do the Roman samples. 255 But it does not devalue the picture presented recently on the technical evolution the Roman glassworkers underwent during the two first centuries of our era (Van der Linden et al. 2009). 256 Mark Hall and Leonid Yablonsky propose a low barium content because they could not detect the barium level when determining the sample by EPMA. 309 Si vs Na 20 HIMT 18 16 14 12 Sarmatian bead late Roman 10 Na2O 8 late Iron Age late Bronze Age-early Iron Age Bronze Age 6 4 2 0 0 10 20 30 40 50 SiO2 60 70 80 Figure 135: Biplot opposing silica to soda oxide content (wt %) Ca vs Si 80 70 60 Sarmatian bead 50 late Roman late Iron Age 40 SiO2 late Bronze Age-early Iron Age Bronze Age 30 20 10 0 0 2 4 6 8 10 CaO Figure 136: Biplot opposing lime to silica oxide content (wt %) The biplot opposing the lime and silica content demonstrates two distinct groups, a low lime group and a high lime group (Figure 136). The Roman material appears very homogeneously within the high lime group with a content ranging between 6-9,5%. Seeing that the two late La Tène beads and fall within this same group we can assume th that this composition is bonded with the classic silica-soda-lime glass from about the 8 century BC (Gratuze et al. 2003a). That the two late Iron Age beads from Neufchâteau-le Sart (BE) are also similar to the Roman glass is th th not so surprising, since silica-soda-lime glass was common from about the 9 8 century BC onwards. It is possible to determine whether vegetable soda or mineral soda was used as flux by plotting the potash content against the magnesium content (Figure 137), because the use of plant ash as alternative flux show both oxides are proportionally correlated. Consequently, the glass of the Pella bead, with its high potash and high magnesium content, is unmistakeably made with plant ash as the fluxing agent. Mineral soda glass normally shows a potash and magnesium content below 1% and never much above, so we have to consider it to be contamination from the mineral soda. 310 K vs Mg 4,5 VEGETABLE SODA GLASS 4 3,5 3 PELLA 2,5 late Roman late Iron Age MgO late Bronze Age-early Iron Age 2 Bronze Age 1,5 MINERAL SODA GLASS HIMT 1 Liberchies 2 0,5 0 0 0,5 1 1,5 2 2,5 3 3,5 K2O Figure 137: Biplot opposing potash to magnesium oxide content (wt %) When plotting the silica content against the alumina content, it is clear that the Bronze Age and early Iron Age black glass artefacts have a much lower alumina content (< 2%) compared to the late Iron Age and Roman pieces (> 2%) (Figure 138). The alumina content is correlated to the sand; hence, we may conclude that 1) the alumina content is not related to the concentration of silica oxides in the glass matrix, but to the applied sand sources; 2) during the Bronze Age and early Iron Age, different sands (with low alumina content) were used than those used from the late Iron Age on. Even though the Sarmatian bead is different from the other late Iron Age beads and matches the Bronze Ageearly Iron Age ratio, we may conclude that there is a chronological dimension to the general change of origin of sand extraction to produce (black) glass; 3) the alumina content of the Liberchies 2 piece is with 3,5% much higher than the 2,0-3,0% generally attested on Roman glass; 4) two late Bronze Age/early Iron Age beadsi.e. from Mez Notariou (FR) and Enveitg (FR) coincide the late Roman cluster. It can be assumed that these two coincidentally matches due to an equivalent (or the same) sand source, unless both beads have been mistakenly determined and are in reality late Roman pieces. The plot in Figure 138 demonstrates that the Roman cluster consist of two separate smaller clusters, i.e. with an alumina content of 22,5% and 2,5-3%. In contrast with the Bronze Age and early Iron Age material which have a similar alumina content as the HIMT sample, the late Iron Age material (excluding the Sarmatian bead) falls within the sub-cluster with more elevated alumina content. The Liberchies 2 piece shows a more discrepant composition. 311 Si vs Al c.1,5% <1% 80 2-2,5% 2,5-3% 3,5% HIMT 70 Liberchies 2 late Roman 60 late Iron Age 50 late Bronze Age-early Iron Age Sarmatian bead 40 SiO2 30 Bronze Age LATE IRON AGE & ROMAN BRONZE AGE & EARLY IRON AGE 20 10 0 0 0,5 1 1,5 2 Al2O3 2,5 3 3,5 4 Figure 138: Biplot opposing alumina to silica oxide content (wt %) Besides a chronological difference, also a regional difference can be identified within the Bronze Ageearly Iron Age black glass beads when opposing the alumina and the magnesium content (Figure 139). The samples from the Near East have a magnesium content of about 1% and above, whereas those from the European Continent have a Mg concentration below 0,5%. Probably the glassworkers in the Near East added vegetable soda to the batch. Actually, this conclusion is already attestable in Figure 137 but less apparent. The difference in alumina content makes it possible to discriminate the late Iron Age and Roman samples from the Bronze Age–early Iron Age black glass. Mg vs Al 5 4,5 4 3,5 late Roman late Iron Age late Bronze Age-early Iron Age Bronze Age 3 2,5 MgO 2 NEAR EAST 1,5 1 0,5 0 Liberchies 2 EUROPEAN CONTINENT 0 0,5 1 1,5 2 2,5 3 3,5 4 Al2O3 Figure 139: Biplot opposing alumina to magnesium oxide content (wt %) This bipolarity is also demonstrated in Figure 140 when plotting the alumina content against the iron oxide concentration. The Bronze Ageearly Iron Age beads have a much higher Fe concentration than did the late Iron AgeRoman artefacts, correspondingly 10%-18% and 5%-11%. Noticeable is that the iron content of the Roman 312 material does not exceed the 8% by much. The late Iron Age beads are fully responsible for the higher iron content in this group. One sample in each group shows a deviant ratio: the beads from tomb 89 in Pella (JO) have almost no iron content, whereas that of the Roman bangle from Liberchies (no. 2) does not go much beyond 2%. This explains its dark green hue not appearing black. Furthermore alumina clusters, most likely due to different sand sources as it has nothing to do with the concentration of silica in the glass matrix (Figure 138) Fe vs Al 18 c.1,5% 16 BRONZE AGE & EARLY IRON AGE 14 12 LATE IRON AGE & ROMAN Sarmatian bead <1% late Roman late Iron Age late Bronze Age-early Iron Age Bronze Age 10 Fe2O3 8 6 2,5-3% 2-2,5% 4 3,5% 2 0 0 0,5 1 1,5 2 2,5 3 3,5 4 Al2O3 Figure 140: Biplot opposing alumina to iron oxide content (wt %) The most important plots clustering the various colouring agents in achieving black glass are those that oppose iron oxide with the titanium content on the one hand to the manganese content on the other hand. In the first place, we can check whether the iron oxide was introduced as a contaminant in the sand like the titanium oxide, or whether it was added deliberately. When plotting the iron oxide against the titanium oxide content, there is a marked cluster for the Roman material around 0,1% Ti-oxide, clearly differing from the late Iron Age beads (Figure 141). Only the Pella tomb beads and perhaps the Liberchies bangle have been made with impure or not purified sand. All others have deliberately added iron oxide to the batch, making it impossible to verify whether impure sand was used too. When opposing iron oxide to the manganese concentration (Figure 142), it is even more clear that the Roman samples are different from all pre-Roman examples, having a much higher Mn content equal or above 0,4% and a lower iron oxide content below 8%. By plotting the Mn content against the Ti content in Figure 143, one Roman bangle catches the eye. The type D1 th th bangle from Tongeren, characteristic for the second half of the 4 5 century AD, has a higher titanium and manganese content featuring the HIMT composition (Freestone 1994) or Maurice Picon‘s ‗Groupe 3‘ (série 3.1) th th dated late 4 -early 5 century AD (Foy et al. 2003, 61-62, tab.12). 313 Fe vs Ti 18 16 14 Sarmatian bead 12 late Roman 10 Fe2O3 8 late Iron Age late Bronze Age-early Iron Age Bronze Age 6 Tongeren HIMT 4 Liberchies 2 2 0 0 0,1 0,2 0,3 TiO2 0,4 0,5 0,6 Figure 141: Biplot opposing titanium to iron oxide content (wt %) Fe vs Mn 18 16 14 BRONZE AGE LATE IRON AGE 12 10 Fe2O3 8 late Roman late Iron Age late Bronze Age-early Iron Age Bronze Age LATE ROMAN 6 Tongeren HIMT 4 Liberchies 2 2 'naturally coloured' 0 0 0,2 0,4 mix of both 0,6 0,8 decolourized 1 MnO Figure 142: Biplot opposing manganese to iron oxide content (wt %) 314 1,2 1,4 1,6 Mn vs Ti 1,6 Tongeren HIMT 1,4 1,2 MnO 1 late Roman 0,8 late Iron Age 0,6 late Bronze Age-early Iron Age Bronze Age 0,4 Liberchies 2 0,2 Sarmatian bead 0 0 0,1 0,2 0,3 0,4 0,5 0,6 TiO2 Figure 143: Biplot opposing titanium to manganese oxide content (wt %) K vs Na 20 NEAR EASTERN LBA-EIA VEGETABLE SODA GLASS 18 Tongeren HIMT 16 Liberchies 2 14 12 late Roman 10 Na2O late Iron Age EUROPEAN CONTINENTAL LBA-EIA 8 late Bronze Age-early Iron Age Bronze Age 6 MINERAL SODA GLASS 4 2 0 0 0,5 1 1,5 2 2,5 3 3,5 K2O Figure 144: Biplot opposing potash to soda oxide content (wt %) From the acquired knowledge on the chemical composition of black glass artefacts from the different periods, it 257 was possible to establish the differences between Roman and pre-Roman black glass. These plots give a general view confined by the limited number of samples and reduced variety, giving rise to various questions concerning a possible production variation within the Roman imperial period itself or a continuation of the various raw glass recipes within the vast research area that the Roman Empire encompassed. 257 It was also our intention to integrate some data from post-Roman times, but the PIXE-PIGE analysis data on Merovingian black glass beads from the vast cemetery of Bossut-Gottechain/Grez—Doiceau, Belgium were provisionally not accessible. 315 A. Concerning the technological issue:  Are different sands recognized in Roman black glass production?  Does the used flux show a great homogeneity or were different fluxes used?  Did Romans use ‗naturally coloured‘ raw glass to make black glass or was decolourised glass also used? o o  Is a chronological and/or geographical pattern observable in the use of naturally coloured‘ and/or decolourized raw glass? Did the Romans use manganese or antimony to decolourize the raw glass ? Did the Romans colour glass by using one particular recipe to obtain a black appearance? o o o Was the applied colorant a product of its time? Which furnace condition was applied? What temperatures were necessary or used? B. Concerning the archaeological issue:  Can the different sands, fluxes, decolourizers, and colorants recognized be linked with o o o specific functional types particular form types within one functional type specific periods, regions or even workshops  Was black glass produced locally in secondary workshops, or did the Romans import the at-hand raw glass coloured black in primary workshops or specialized intermediate workshops?  To what extent can the chemical composition provide a clearer understanding of the typo-chronological idiosyncrasies?  Are chronological and regional differences observable in the production of black glass during the Roman imperial period? C. Concerning the historical issue:  When various specific recipes were in use during the Roman imperial period: o o o Can this diversity in black glass production have its origins in a chronological development? Can a contemporary diversity of recipes be easily reduced to regional variability due to specific distribution and trade restrictions or regulations? Can this diversity be attributed to the function of the finished product? In other words, is there a difference in chemical composition between the glass used for the production of vessels, jewellery, tesserae and/or counters?  Was black glass mass-produced in one single, or several, centres and then further distributed all over the Empire, or was it produced locally? The latter fits in with the traditional system of supply and demand for welldefined regional markets causing regional and local idiosyncrasies, whereas the former rather claims the existence of a proto-industrialized organisation of glass workshops, implying a much greater sameness of forms and types. But is one single model appropriate for the entire Roman imperial period or did various models emerge through the consecutive centuries?  To what extent is it possible to identify secondary glass workshops that produced artefacts in (black) glass on the basis of the different recipes used? For instance, can workshops be identified on the basis of specific elements pointing to a different source of raw glass or to a different colouring technique? In other words, is it likely to reduce each recipe to one single secondary glass workshop that produced artefacts in black glass in a well-defined period? Is it equally possible that black glass (artefacts) from one and the same recipe were produced in various Roman glass workshops within a specific region or even at random? 258 To achieve this goal, it was advisable to sample black glass artefacts from a long period. This diachronic sampling already had provided us with information on the technology used to colour the glass black that apparently developed through the successive centuries (Cosyns et al. 2006a; Van der Linden et al. 2009). To enable further exploration, we tried to assemble a comprehensive set of sampled black glass artefacts including: 258 We have compared in this introduction the prior analysis results of Roman black glass with, as much as possible, pre- and post-Roman black glass reflecting the present-day research on this topic. 316     diverse functions, e.g., vessels, jewellery, counters, architectural decoration material and production material all five centuries of the Roman imperial period different areas within the Roman Empire various glass hues, e.g., green, purple, brown, blue, blue-green This unavoidably necessitated a large number of samples. We were able to integrate in this study the analysis 259 results of Roman black glass artefacts for a total amount of 366 samples , nearly all embedded in an acrylic 260 resin [see Appendix 1]. Although we tried to make a cross-section of all commodity categories, all areas, all colours and all periods we realize we mostly sampled vessels and bracelets in green glass from north-western Europe (Table 142). Table 142: Overview of samples per assorted topics [see Appendix 1] (n= number of samples) function region or country colour period n n n (see map) vessels bracelets finger rings beads hairpins pendants gems counters rods tesserae raw glass crucible glass obsidian indefinite TOTAL 184 149 7 19 2 1 11 11 1 40 3 5 5 438 UK Benelux France Germany Alps (Switzerland – Austria) Italy Balkan Aegean Levant Egypt Maghreb Iberia unprovenanced 20 229 33 10 70 8 25 16 7 8 2 10 438 green purple brown blue blue-green black 268 54 19 17 23 46 other 11 438 n Period I Period II Period III Period IV Period I/II Period II/III Period III/IV Period I/III Period I/IV modern indefinite 47 137 15 33 32 27 58 5 6 7 29 pre-Roman post-Roman 35 7 438 The significant contribution of Prof. Dr. Koen Janssens and his team at the Micro and Trace Analysis Centre (MiTAC) of the University of Antwerp made it possible to conduct the chemical analysis on a huge number of black glass samples by means of using the scanning electron microscope with energy dispersive X-ray spectroscopy (SEM-EDX) and by means of laser ablation with inductively coupled plasma mass spectrometry (LA-ICP-MS). The cooperation with MiTAC not only added greatly to the acquisition of an outline on how to define in detail the artefacts to sample, but it also gave the opportunity to learn SEM-EDX analysis as it is done in the laboratory at Wilrijk. I was apprenticed to prepare samples in a resin and to measure the samples. The detection, de-convolution and subsequent quantification of the spectra were done by the MiTAC team. However, Dr. Olivier Schalm, dra Veerle Van der Linden and drs Simone Cagno successively received my help with measuring the samples. By taking samples myself of all artefacts to be analysed at the museums and archaeological depots in Belgium and abroad, and by preparing most resins, we tried to minimize the nuisance to Prof. Dr. Koen Janssens‘ team. All useful information has been acquired with their helpfulness, their accuracy and above all their commitment to contribute to a better insight into the black glass matter in particular, and into Roman glass in general. A concerted action between MiTAC of the University of Antwerp (UA) with the Department of Chemistry at the University of Technology of Warsaw has been the opportunity to have a first selection of resins analysed via LAICP-MS by Ewa Bulska‘s team in collaboration with Anna Nowak and Barbara Wagner. Subsequent selections of resins for LA-ICP-MS analysis were executed in the lab of Prof. Dr. Vanhaecke at the Department of Chemistry of the University of Ghent, Belgium (UGent). Prior to these analyses, we already had the occasion to use LA-ICP-MS to compare the SEM-EDXmeasurements of some resins by Bernard Gratuze at the CNRS centre Babelon IRAMAT in Orléans in the 261 prospect of the ULB-CReA publication edited by Prof. Dr. Eugène Warmenbol on the glass material of the subaquatic site Trou de Han at Han-sur-Lesse (BE). 259 An additional 66 samples have been incorporated in the study (57 non black; 4 obsidian; 5 modern). Three samples from Tienen (BE) [see Appendix 1, PC4eg] could not be sampled and were analysed by SEM-XRF and the semi-quantitative results have been converted to % weight. 261 Université Libre de Bruxelles – Centre de Recherche Archéologique (CReA). 260 317 8.2 Methodology 8.2.1 SEM-EDX The analysis, executed at the MiTAC of the University of Antwerp, has been done by SEM-EDX under the supervision of Prof. Dr Koen Janssens. Thanks to coaching by his assistants, Dr Olivier Schalm, Dra Veerle Van der Linden and Drs Simone Cagno, I was able to help in a substantial way through selecting and sampling all black glass artefacts, through the preparation of the majority of the resins and through helping when measurements were taken. The samples had to be prepared to enable the measuring. Therefore, the selected small glass fragments (mainly averaging between 1 and 2 mm per side but never exceeding 5 mm) were embedded into an acrylic resin. The resin blocks were mechanically ground with silicon carbide paper and polished with diamond paste up to 0,25 μm to obtain a smooth surface of unaltered glass. This surface treatment is required because SEM-EDX is a superficial measurement system bombarding a sample with X-rays in a predetermined angle to be received by the SEM to allow an efficient detection of the elemental spectra. The preparation of the resins contained different stages:  sticking a selection of samples on double-sided tape that was applied on the lid of a plastic matrix  photographing and drawing a first sketch of the samples (seen from the inner/back side) related with an identification list of the samples (Figure 145)  obtaining a cold-curing resin for metallographic testing by mixing the Technovit 4004 (Kulzer) powder with the Technovit 4004 (Kulzer) liquid (Figure 146)  pouring the resin into the matrix to encase the samples  polishing of the resin to get a clean, smooth and scratchless surface to enable the analysis measurements (Figure 147) o use of polishing paper (320; 500; 800; 1200; 2400) with running water o use of polishing platform with diamond powder (3 µm; 1 µm; .25 µm) and a lubricant  photographing and drawing of a definitive sketch (= mirror image of first drawing) that can be scanned to result in a digital drawing [see Appendix 1]; a b Figure 145: a) preparing the samples in a plastic matrix before filling the plastic matrix with resin; b) making of a sketch of the samples 318 Figure 146: The material used to make a resin for encasing the samples Figure 147: (left) The final stage of grinding and polishing the resin surface by using diamond paste of 0,25 µm.; (right) Resin with clean, smooth and scratchless surface ready for analysis measurements Figure 148: Vacuum carbon evaporation device (BALZERS UNION) used to apply a thin carbon coating 319 After the polishing process, a thin carbon coating was applied on the top surface of the resin by using a vacuum carbon evaporation device to give the resin a conductive surface to disperse the electronic charge from the electron beam (Figure 148). A JEOL 6300 scanning electron microscope equipped with an energy dispersive Xray detector was used. The acquisition time was 100 seconds, the accelerating voltage 20kV, the electron beam current 2nA, and the magnification 500x. An electron beam generated from a heated tungsten filament is accelerated through a vacuum column and bombards the sample. As a result of the excitation as well as de-excitation of the sample‘s electrons, characteristic X-rays are emitted, whose energy is related to specific elements. After being received by an Energy-Dispersive X-ray detector positioned at an angle of 30°, the recorded X-ray intensities result in a plot of the sample‘s chemical spectrum. Subsequently, the measured spectra are de-convoluted by a computer 262 programme (least squares fitting) and quantified using a standard-less ZAF-programme (Schalm, Janssens 2003). The quantification of the elements is tested on the efficiency of the measurements vis-à-vis the used standards. The analysis results used in the tables are the calculated average from four to five measurements taken on different zones of the sample. 8.2.2 LA-ICP-MS The analysis technique by LA-ICP-MS has been used in addition to SEM-EDX to enable the insertion of concentration data of minor and trace elements far below 0,1%. The latter gives a fast and easy outline on all oxides with a concentration above 0,1%, whereas LA-ICP-MS provides very sharp results on the minor and trace elements. Another interest in the application of this technique resides in the fact that it can be considered as a nondestructive analysis since the laser only burns microscopic ablation craters with diameters smaller than 100 µm, and thus they are barely visible with the naked eye. Laser ablation was executed on a selection of the black glass samples in three different laboratories: 1. 2. 3. In two consecutive sessions 16 resins were measured by Olivier Schalm and Simone Cagno at the Department of Chemistry of the University of Warsaw, Poland with the support of Anna Nowak, Barbara Wagner, and Ewa Bulska. Five resins (PC1-3; PC6; PC7) were analysed in March 2007, and in December 2007 another eleven resins (PC10-16; PC18; PC21-23) were subjected to LA-ICP-MS 263 testing. During a visit at the Centre Ernest BabelonCNRS, Institut de Recherche sur les Archéomatériaux (IRAMAT), Orléans for the analysis of the intact late Bronze Age glass beads from Trou de Han (BE) rd (Figure 149) and the beads from the early 3 century BC chariot tomb excavated at Neufchâteau-le-Sart (BE) (Table 141), we seized the opportunity to check a set of Roman black glass samples (resins PC3; PC6; PC7). This verification by Bernard Gratuze (UMR 5060) was only meant to reveal the deviation between his measurement results and those obtained using SEM-EDX at MiTAC, University of Antwerp and through LA-ICP-MS at the University of Warsaw [see Appendix 1]. A selection of the most recently acquired samples has been analysed by Simone Cagno at the Department of Chemistry of the University of Ghent (UGent) with the support of Isolde De Coninck (resins PC 24-28) and Andrei Izmer (resins PC 35; 36; 40; 41) under the supervision of Prof. Dr Frank Vanhaecke. With a laser, it is possible to take microscopic samples from a specific zone on the surface of the embedded samples. Using inductive coupling, the evaporated mass is taken along with a carrying axle towards the plasma beam. The plasma beam is able to dissociate, atomise and ionise the injected mass. By means of supersonic propulsions the ions thus formed are then launched from the plasma through an intermediate vacuum. Afterwards, a system of electrostatic lenses extracts and focalises the positively charged ions, diverting them to a quadrupole mass filter. This filter makes it possible to select the ions of which the ratio mass/charge (m/Z) was preselected. A detector then records the ions transmitted through the mass filter. Each isotope of each element corresponds to a unique value corresponding with a specific ratio mass/charge (m/Z). This provides an easy identification of the elements present in the sample. Furthermore, the amount of detected ions directly depends 262 263 The ZAF programme takes into account the atomic number (Z), absorption (A), and fluorescence (F). The first results on the chemical composition obtained from these resins resulted in a publication (Vander Linden et al. 2009). 320 on the concentration from the source of each element within the glass sample, thus making it possible to calculate the percentage of each element. In Warsaw, an inductively coupled plasma mass spectrometer ELAN 9000 (Perkin Elmer, SCIEX, Canada) + + equipped with the laser ablation system LSX-200 (CETAC, USA) was used. The LSX-200 combined a stable, environmentally sealed 266 nm UV laser (Nd-YAG, solid state, Q-switched) with a high sampling efficiency, variable 1–20 Hz pulse repetition rate and maximum energy up to 6 mJ/pulse. The ablation was done horizontally on two different areas of the sample, making a scratch of 100 µm wide each. In Orléans, an inductively coupled plasma mass spectrometer VG Plasma Quad II type PQXS, in combination with a UV-Microprobe laser ablation system, was used. The ablation was executed by vertically pointing on 3 to 4 different areas of the sample, provoking perforations of 80 µm in diameter each. To ensure non-interference of impurities and defects, the first seconds of the produced gas was not taken into account for quantification. In Ghent, an inductively coupled plasma (ICP) mass spectrometer (Thermo PlasmaLab type XSeries 2) was used to quantify the different isotopes present in the sample. The glass sample embedded in the resin was ablated by a Geolas 200M laser-Lambda Physik Compex and the evaporated glass mass, carried by the argon gas, was conveyed to the ICP mass spectrometer. The ablation was done horizontally on two different areas of the sample, making perforations of 60 µm diameter each with a length of 600 µm. A pre-cleaning was completed each time under a 1 Hz pulse repetition with a maximum energy up to 1,6 mJ/pulse at a rate of 50 µm/s on a surface of 120 µm width to ensure a clean surface for the quantitative analysis. Afterwards, the ablation was done under a 10 Hz pulse repetition rate with a maximum energy up to 0,4 mJ/pulse at a rate of 10 µm/s,. To analyse with LA-ICP-MS, the resins with embedded glass samples were placed inside a quartz cell filled with Argon (Ar) that acts as a flux (Figure 149). Figure 149: LA-ICP-MS installation at CNRS-IRAMAT, Orléans, France (photographs by author, courtesy of Bernard Gratuze) In Orléans, two detection methods were used. The first was analogue and provided the measurement of the major elements. The second method was impulsivemeaning that the computer counted the pulses for every oxide indicated beforehand, supplying a quantification of the minor and trace elements. Several measurements were executed depending on the number of elements to be checked. The first was done to accumulate the data on the major elements of the glass matrix (elements with a concentration above 1% weight such as silica, soda, lime, alumina, and potash). The others were needed to collect the data on the selected minor and trace elements chosen in correlation to the queries appropriate to the investigation pursued. Per compilation of data accumulation, three ablations were carried out. A conversion programme was set up in Visual Basic within Excel, giving the opportunity to process the acquired data set. The applied computation system was based on the principle of an internal standard. The calibration of the machine was obtained by using NIST glass standard SRM 610 and Corning Museum of Glass (CMG) standards B, C and D. In Warsaw and in Ghent, the selection of both methods was done automatically. 321 Furthermore, the nature of the studied matrices (lime glass, potash glass, lead glass, glass with antimony or with tin and lead) of the archaeological glass was analysed a priori by neutron activation by fast cyclotron. The correctness of the method was regularly checked by measuring the international glass standards (or by using previously analysed archaeological glass). The limits of detection ranged between 0,1%-0,01% for the major elements and between 20-500 ppb (parts per billion) for the minor and trace elements. The precision of the results is of an order of magnitude of 5%-15%, depending on the elements and their measured percentages. To obtain the LA-ICP-MS data, the resin blocks were inserted separately in the laser ablation sample cell on the XYZ translation stage. The exact positioning of the sample was obtained by means of a CCD camera as a + viewing system under PC control. The LSX-200 was centrally controlled by the Cetac Windows software which allows a selective ablation of the chosen areas of the investigated samples. Each measurement cycle consisted of pre-cleaning the sample surface with a laser beam (no signal was registered during this step) followed by the laser ablation accompanied by registering of signals. Before ablation, a blank was always registered within the 11 23 26 27 first 18 s of argon flow only. Signal intensities were recorded for the following isotopes: B, Na, Mg, Al, 28 31 32 39 42 43 49 55 57 65 75 85 86 87 88 89 90 91 121 133 137 Si, P, S, K, Ca, Ca, Ti, Mn, Fe, Cu, As, Rb, Sr, Sr, Sr, Y, Zr, Zr, Sb, Cs, Ba, 151 178 206 207 208 264 43 Eu, Hf, Pb, Pb, and Pb. Ca was used as an internal standard. Before quantification, the blank signal was subtracted for each element individually. During the LA-ICP-MS measurements in Warsaw, the CMG standard B was employed as an external standard, and its composition was validated with the use of CMG standard D and NIST SRMs 2709, 2710 and 2711. More details are provided in (Wagner, Bulska 2004; Šmit et al. 2005; Wagner et al. 2007). In Orléans, the NIST standard 610 and CMG standards B, C and D were used, while those utilized in Ghent were limited to NIST 610 and 612. The contents of chromium (Cr), nickel (Ni), and zinc (Zn) were in almost all cases below the limit of quantification for the VG Plasma Quad II type PQXS at the CNRS lab of Bernard Gratuze in Orléans, while this was not the case for the very recent and more efficient Thermo PlasmaLab type XSeries 2 installation at the Department of Chemistry in Ghent. Brome (BE) and arsenic (As) were the only elements for which semi-quantitative values could be given, because in the CMG standard B used here, these elements are present only in an insignificant degree above the limit of quantification. On the other hand, it is a problem quantifying arsenic (As) because its mass is 75 and interference can occur with the chlorine (Cl) that is measured together with the argon plasma gas. 8.3 Studied material Within the available time of the project, we were able to sample 368 black glass artefacts of the 4475 (4754) inventoried black glass items embedded in 49 resins, but various non-black glass artefacts, or non-Roman 265 material, were integrated in these 49 resins [see Appendix 1]. This not inconsiderable amount of sampled artefacts (roughly 1 in 12) included vessels, bracelets, finger rings, beads, hairpins, pendants, gaming counters, tesserae and raw glass. These sampled objects come from different areas in the north-western Roman provinces Gallia Belgica, Germania Inferior and Germania Superior, although a not negligible quantity of the sampled material covers the Mediterranean area (Italy, Croatia, Slovenia, Cyprus, Israel, Egypt, Tunisia) (Table 143). This extensive set of samples provides important additional information to the prior obtained results on French, Italian (Moretti, Gratuze 2002), and some late Roman material from Braga (PT) (pers. comm., Bernard Gratuze). The sampled material applicable in the research presented here covers the entire Roman imperial period over five st nd rd centuries. The samples, however, are to a large extent material of the 1 century AD and the late 2 3 century AD (Table 144). When paying attention to the different glass colours appearing blacke.g., purple, blue, brown, greenone can see that the majority of the sampled artefacts are made of green glass (Table 145). 264 With LA-ICP-MS, it is not possible to quantify separately the various isotopes of a particular element. For various practical reasons, specific isotope analysis has not been carried out within the research project on Roman black glass. 265 The set of 49 resins contain in total 439 samples but the 71 omitted samples are either non-Roman or of non-black glass. This is the result of a gradually accumulated list of samples over the past years incorporating occasional opportunities cropping up on the sideline. Although these samples have no direct interest to the black glass research, they appeared to be important as comparative material. Thirty-eight samples are from La Tène bracelets, 3 samples are from decolourized or ‗naturally coloured‘ glass, 5 samples are from obsidian, 4 samples are modern, and 4 samples are indistinct vitreous material. The compositions of these samples have not been incorporated in Appendix 1 but are available on demand. Besides, some samples have been polished away or seemed to have been so affected that a good and relevant measurement was not possible. 322 Table 143: List of sampled Roman black glass material according to provenance (between square brackets, from previous research by others (Morreti, Gratuze 2002)) country sites Belgium Netherlands Germany France England Switzerland Austria Slovenia Croatia Italy Portugal Cyprus Tunisia Egypt Israel Amay; Anthée; Auvelais; Bassenge; Dilsen; Dourbes; Elewijt; Florennes; Grobbendonk; Izier; Kontich; Kruishoutem; Lauw; Liberchies; Liège; Marche-les-Dames; Matagne-la-Petite; Menen; Nismes; Nivelles; Oudenburg; Pry; Roch-Ste-Anne; Roly; Rumst; Tienen; Tongeren; Treignes; Velzeke Heerlen; Hoogeloon; Nijmegen; Trier Faulquemont; Grandes Armoises; Marseille; Olbia; Reims; Ste Menehould; [Gap] London; Colchester Augst; Avenches Magdalensberg Celje; Ribnica; Trebnje Oresac Rome; [Aquileia]; [San Taglimento] Braga Carthage; Sidi Djididi Elkab; Ghurob Horbat Qastra Table 144: List of sampled Roman black glass material according to type and date range per generalized period prepostI II I/II II/III IV III/IV I/IV indefinite Roman Roman vessels TOTAL 35 109 26 - - - - - 1 13 184 arm rings - - 6 41 33 32 - 33 - 3 148 finger rings - - - 2 - 4 - - - - 6 beads - - - 2 - 17 - - - - 19 pendants - - - - 1 2 - - - - 3 hairpins - - - 2 - - - - - - 2 counters - - - - - - 11 - - - 11 rods and tesserae 10 - - - - - - - 1 - 11 raw glass + crucibles 2 16 - - 3 3 - - 8 6 38 other - 3 - - - - - - 5 8 16 47 128 32 47 37 58 11 33 15 30 438 TOTAL Table 145: List of sampled Roman black glass material according to type and colour of the glass matrix black green purple brown blue blue-green other TOTAL vessels 7 142 11 12 8 2 2 184 arm rings 8 91 34 1 3 8 3 148 finger rings 1 5 - - - - - 6 beads 14 4 - - 1 - - 19 pendants 2 1 - - - - - 3 hairpins - 2 - - - - - 2 counters - 5 5 1 - - - 11 - 1 2 3 3 2 - 11 raw glass + crucibles 10 12 1 2 1 10 2 38 other 4 5 1 - 1 1 4 16 TOTAL 46 268 54 19 17 23 11 438 rods and tesserae First, the material has been subdivided by means of a chronological component (e.g., period I and period II) and a geographical component (e.g., zone a and b) (Van der Linden et al. 2009) (Table 146). 323 Table 146: Chronological subdivision of the sampled material as proposed in Van der Linden et al. 2009 Period I Period II (ante 150 AD) (post 150 AD) East Mediterranean (Egypt; Israel; Cyprus) Ia Ib Northwest Europe (UK; Benelux; France; Germany; Switzerland; Austria) IIa IIb st 266 267 Group Ia: No samples could be taken from 1 century AD black glass artefacts from the East Mediterranean. Group IIa: This section incorporates 113 early Roman samples from Northwest Europe. Five sampled cast st vessels come from the early 1 century AD site in Magdalensberg (AT) (Resin PC1g-k). Group Ib: Concerning the samples of the later period in the Mediterranean, we have four rod-formed unguentaria with a solid body from Elkab, Egypt (PC2a-b; d-e). Further are sixteen samples, mainly beads, from the late Roman fort at Horbat Qastra, Israel. Also a pendant and two bracelet fragments from the same site could be sampled (PC24a-h; PC25a-h). Of the two samples from Cyprus, a pendant and a chunk of raw glass (PC28a-b), only the pendant could be dated on a typological basis. Group IIb: All samples from Matagne-la-Petite, Tienen, and Rumst, Belgium have been taken from carinated beakers Isings Form 36b, with the exception of one piece of production waste. The beakers are all dated between nd rd the last quarter of the 2 century and the second quarter of the 3 century AD (Cosyns, Hanut 2005). Further, in the set of resins, two Faulquemont samples were incorporated from two different beakers (Isings Form 94variants), one with indented body and base-ring and the other without foot. The non-Roman material consists of two categories: 1) Resins PC3i-n, PC5k, PC9-10, PC29a-b, PC38a-b, PC39e, and PC42f are late La Tène glass bangles st st dating from 1 century BC1 century AD, mainly in deep purple glass appearing black 2) Modern material, for instance sample PC1f from Rumst (BE), which shows a composition equalling the th 268 th th black glass bottles of the 17 century in northern France ; the late 18 –early 19 century bottle from Wenduine th th (PC5l); and the late 19 –early 20 century medallion and vessel fragment from Han-sur-Lesse (PC8d and f). These samples were deliberately inserted to see how they contrast with the Roman black glass to understand better the Roman black glass idiosyncrasies. The entire set included some non-black-appearing Roman material, for instance, the three colourless bracelets from the Belgian sites of Tongeren (PC3i-j) and Emblem (PC3m) and two colourless body fragments of the beakers with Greek inscriptions from the Kielenstraat in Tongeren (BE) (Cosyns et al. 2005, 182-183). Most interesting is the material from the recently published glass workshop from Kaiseraugst (CH) that also produced black glass artefacts (PC46-47) (Fischer 2009) [see this chapter below]. Several obsidian fragments are integrated within the set of black glass samples, sometimes intentionally. One piece comes from the sub-aquatic site at Han-sur-Lesse (BE) (PC8e). Another sample is taken from a chunk from Monte-Arci, Sardinia (PC44h) and used for the Velzeke experiment in the scope of the obsidian vessel fragment R1610 in RMAH, Brussels (Cagno et al. 2010). Several pieces come from east Turkey (PC29g; PC37g-h). Other non-black glass is the material from Vremde (BE) (PC3a-d). Finally, the samples polished away or fully altered are PC24b-c, a late Roman bead, and a pendant from Horbat Qastra (IL), respectively. 266 We verified whether samples from other areas – e.g., Central Europe (Austria; Slovenia; Croatia) and the West Mediterranean (Portugal; Italy; S-France; Tunisia) – can result in a further elaboration of geographical subdivision – respectively as Group IIIab and Group IVab. 267 A demand for sampling Egyptian material including 1st century AD vessels in the Petrie Museum of Egyptian Archaeology, London was accepted in January 2009 and was executed on 24t February 2009. Unfortunately, not all granted pieces have been sampled in particular the bracelets as well as some particular vessels. 268 Personal communication from Bernard Gratuze. 324 8.4 Results 8.4.1 SEM-EDX Roman glass production is characterised by a soda-lime-silica glass with ratios in the order of 16:10:70. This production relied on the melting of two primary ingredients: sand and mineral soda. Apart from the silica content, the sand also contained a considerable concentration of lime (8%-10%), which could immediately be used as stabilizing agent, and some impurities such as iron and alumina. The latter appeared in a concentration of about 3% (Brill 1988). Other trace elements present in ancient glass are Fe 2O3, TiO2, BaO, SrO and B2O3 (Brill 1988). Figure 150: Ternary plot opposing Al2O3-Fe2O3-MnO (weight %) (taken from Vander Linden et al. 2009, fig.4) 269 The material studied shows a very homogeneous silica-soda-lime group with a concentration of 54%-72% for silica (SiO2), 12%-21% for soda (Na2O) and 3,5%-11% for calcium (CaO) (Figure 150). This homogeneous composition is characteristic throughout Antiquity, roughly from the 8th century BC to the 8th century AD (Gratuze 2004). Analysis of Roman glass demonstrates that soda was the main flux in glass production. During this period, the primary source of soda was natron and trona, naturally occurring salts found in dry lake beds (Shortland et al. 2006), with its main exploitation in the larger area around Wadi El Natrun, located in Egypt (Brill 1988; Nenna 2003). Figure 151 demonstrates that during the Roman imperial period the bulk of the black glass artefacts were made by using mineral soda with the typical low magnesiumlow potash contentboth below 1,5%. A number of samples, however, show clear increased levels of magnesium and potash above 1,5%. Varying directly, the magnesiumpotash content is clearly correlated. We can thus conclude that besides the general use of natron/trona, Roman glassworkers also produced black glass by using plant ash alkali. The addition of vegetable soda via a soda-rich plant ash to the batch is only applied in production of the low iron black glass artefacts. Besides the increased levels of magnesium (MgO) and potash (K 2O), the use of plant ash alkali results in higher concentrations of phosphor (P2O5) and strontium (SrO). 269 A number of samples are deviant from these dividing lines. Because they appear interesting for the entire research on Roman black glass, they cannot be neglected, yet they will only be discussed in the margin. 325 Resins PC1-48 7,0 6,0 5,0 4,0 vegetable soda K2O 3,0 Resins PC1-48 2,0 1,0 0,0 0,0 1,0 2,0 3,0 4,0 5,0 6,0 7,0 MgO mineral soda Figure 151: Type of flux: magnesia opposing potassium (weight %) Some preliminary results have been presented during the previous years (Cosyns et al. 2006a; Van der Linden et al. 2009). The first results were obtained from a set of 39 samples [see Appendix 1, resins PC1-3] covering vessels, counters, jewellery and production waste from Periods I and II (Cosyns et al. 2006a, 39-40, tab.2). Two distinct clusters are discernable when plotting the ratios of titanium oxide (TiO2) and alumina (Al2O5) (Figure 152) and that between iron oxide (Fe2O3) and alumina (Figure 153). 0,45 0,40 0,35 Period I +Egypt Period II 0,30 Rumst Magdalensberg Elkab Tienen Faulquemont Dilsen 0,25 TiO2 0,20 0,15 0,10 Europe Period II 0,05 0,00 0,00 0,50 1,00 1,50 2,00 2,50 3,00 Al2O3 Figure 152: Type of sand: Al2O3 versus TiO2 (weight %) for the resins PC1-3 326 3,50 4,00 Opposing the alumina and titanium contenttwo oxides which are present in sandwe can detect two clusters nd rd corroborating the use of two different sands. The late 2 and 3 century AD material from Rumst, Tienen, and Dilsen (Belgium) and Faulquemont (France) [Cluster 1] are characterised with a low Ti content, below 0,1 (5)%, st and low Al content, ranging between 1,25% and 2%, while the 1 century AD material from Magdalensberg rd (Austria) and the early 3 century AD pieces from Elkab (Egypt) [Cluster 2] have a higher Ti and Al content: above 0,15%, up to 0,40%, and between 2% and 3,5%, respectively. We may conclude here that the artefacts in Cluster 1 are produced with a raw glass from pure sand, whereas the sand of Cluster 2 shows higher ratios of impurities. In Figure 153, Cluster 1 features a high iron content above 6% up to 12%, whereas Cluster 2 is characterised by a low iron content between 1% and 2%. The correlation of the iron and alumina content in Cluster 2 makes it clear that the presence of iron oxide is at the impurity level. The iron content of Cluster 1 artefacts exceeds by far the impurity level and is consequently deliberately added. In opposition to Cluster 2, we see this time an anticorrelated ratio pointing to the addition of quite pure iron as a colouring agent. Deviant for Cluster 2 is the Form IIB.6 piece from Faulquemont (FR) [see Appendix 1 Resin PC3e], which demonstrates a low iron content. 14,00 12,00 Europe Period II 10,00 Rumst Magdalensberg Elkab Tienen Faulquemont Dilsen 8,00 Fe2O3 6,00 4,00 Period I +Egypt Period II 2,00 0,00 0,00 0,50 1,00 1,50 2,00 2,50 3,00 3,50 4,00 Al2O3 Figure 153: Biplot visualizing alumina versus iron content (weight %) for the resins PC1-3 Figure 153 shows that the concentration of manganese increases in proportion to the iron content. This correlated ratio, clearly visible for Cluster 2, is blurred by the addition of iron oxide as a colorant. Both clusters have a very low manganese concentration, indicating that the presence of manganese oxide, below 0,4%, is at the impurity level from the added flux. One of the Dilsen bracelets [see Appendix 1 Resin PC3h] has a manganese content of nearly 1% that points to a deliberate addition as a decolourizer. One of the Faulquemont pieces seems either a rd Period I artefact instead of Period II or a Period II import from Egypt seeing its similarity to the early 3 century AD Elkab pieces. 327 14,00 Europe Period II 12,00 10,00 Rumst Magdalensberg Elkab Tienen Faulquemont Dilsen 8,00 Fe2O3 6,00 4,00 Period I +Egypt Period II 2,00 0,00 0,00 0,20 0,40 0,60 0,80 1,00 MnO Figure 154: Biplot visualizing manganese versus iron content (weight %) for the resins PC1-3 The abovementioned Figures 153-154 demonstrate that there was a group of the Roman imperial period blackappearing glass with a high iron content [Cluster 1] and a group with a low iron content [Cluster 2] (Cosyns et al. 2006a, 39-40). Making regional (a = eastern Mediterranean and b = northwestern provinces) and chronological (I = Period I and II = Period II) subdivisions (Van der Linden et al. 2009), Cluster 1 is only present in the northwestern provinces during Period II (IIb), whereas Cluster 2 is illustrated by samples from northwestern artefacts of Period I (Ib) and East Mediterranean artefacts of Period II (IIa). No artefacts of group Ia were sampled at that time. When plotting most material, the two clusters become more vague, but still persist. The number of dots on the 0 line for titanium is due to the fact that SEM-EDX cannot give accurate calculations of concentrations below 0,1% weight. The plot shows we have to enlarge the range of both clusters. Cluster 1 has a titanium level up to 0,25% and an alumina content ranging from 1,25%-2,5%, yet there is a distinct concentration within the cluster between 0,1%-0,2% of TiO2 and between 1,5%-2,0% of Al2O3 (red circle in Figure 155). Cluster 2 differs from Cluster 1 with a titanium level between 0,2% and 0,65% and an alumina content between 1,8% and 4%. The deviant group with an alumina level above 4% is most likely to be connected with obsidian and will be discussed later in this chapter, in the section concerning the archaeological relevance [part 9.5.]. The distinct high Al – high Ti content of Cluster 2 is to be related to Period I (IIa) and the East Mediterranean group (Ib), while both northwestern groups (IIa-IIb) are characterised by a low Al  low Ti content. Noticeable is a certain portion of Group IIb that overlaps with the East Mediterranean group with an alumina-titanium ratio between 1,5%-0,25% and 2,5%-0,15%. 328 0,70 0,60 0,50 0,40 TiO2 0,30 0,20 0,10 0,00 0,00 1,00 2,00 3,00 4,00 5,00 6,00 7,00 8,00 9,00 Al2O3 Figure 155: Biplot visualizing alumina versus titanium content (weight %) for the resins PC1-48 Figure 156: Al2O3 versus Fe2O3 (weight %) for the resins PC1-48 In Figure 156, opposing Al2O3 with Fe2O3, the two clusters detected previously corroborate with the use of two different techniques in colouring the glass black by adding a high portion of iron oxide, or through a low concentration of iron. Seeing that both oxides of the artefacts in Cluster 2 are directly proportional, they can be considered as impurities in the sand. Because of the addition of iron to the batch for colouring the artefacts in Cluster 1, this proportional occurrence is not more clearly visible. The material of Period II is confined by an alumina content below 2%. This variation in the glass is also noticeable in Figure 157, where we clearly can 329 demonstrate it to be linked with the chronological evolution in the glass production. The high iron samples with increased alumina and titanium levels correspond with the characteristic late Roman HIMT glass. This distinct th group with high iron, high manganese and high titanium content appeared during the 4 century AD and seems to have been a dominant player on the market during Period IV. The Period II material is characterized with a low titanium content below 0,2%. The reason for this low concentration is probably due to the use of pure iron for colouring the glass. The higher titanium ratio of the Period IV HIMT glass is due to impurities in the sand and not by the addition of iron, seeing that both elements are not correlated. On the other hand, this correlation is clearly visible in the low iron cluster. In Figure 158 the two main colouring agents, iron and manganese, are opposed. The samples with high iron and high manganese content all have a green hue. The samples with an opposite ratio, with low iron and high manganese content, are purple. The low iron–low manganese group encompass brown, blue, and green glasses. The high iron group is characterised with a low manganese content below 0,6% (left rectangle), but a number of samples show a more elevated manganese content with two subcategories. One has a high manganese concentration above 1% and up to 1,8%. The other with a MnO level ranging between 0,6% and 1% is to us a mix of low and high manganese raw glass, and pleads for recycled glass. 18,0 Resins PC1-48 LEVANTINE 16,0 14,0 12,0 10,0 Fe2O3 8,0 Resins PC1-48 HIMT 6,0 4,0 2,0 0,0 0,0 0,5 1,0 TiO2 Figure 157: TiO2 versus Fe2O3 (weight %) for the resins PC1-48 330 1,5 2,0 Resins PC1-48 7,0 6,0 GREEN 5,0 4,0 Fe2O3 3,0 BROWN 2,0 PURPLE 1,0 0,0 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 MnO Figure 158: MnO versus Fe2O3 (weight %) for the resins PC1-48 Figure 159 shows that the green samples with low iron content have an elevated magnesium content up to 6%, compared to those in blue and purple glass with an overlap from 1,5% and 2%. Black-appearing blue glass artefacts have a low magnesium content well below 1% and with 1% to 3% more elevated iron content compared with the black-appearing, purple-coloured glass. Two clusters appear for the brown-coloured black artefacts: either the ratio iron-magnesium is below 0,5% or both oxides range around 2% each. This refers to two productions of black glass using the same colouring technique, but by two melting methods with or without the addition of (sodic) plant ashes on top of the soda. The dots within the low iron range remaining dark blue are black samples of which the true hue could not be determined. 331 Figure 159: MgO versus Fe2O3 (weight %) for the resins PC1-48 Summary The entire set of 438 samples supplied an increase of information for a better understanding of Roman black glass technology. A great part of the samples show the typically low K 2O and MgO values (below 1.5%) of the Levantine glass made with evaporites (e.g., natron). Some samples show moderately increased, to very distinct, higher amounts between 1,5% and 4%, indicating that plant ash alkali were used as a fluxing agent rather than mineral soda (Jackson et al. 2008). Black-appearing glass was made from dark green, brown, purple and blue glass with iron and manganese as the main colorants. Two groups can be recognized on the basis of their high contents: Fe 2O3 > 3,5% and MnO >2%. Consequently, an iron ore or a manganese ore was added to the batch in order to obtain correspondingly a dark green and dark purple colour. The remaining group (green and brown samples) gets its hue from the raw materials themselves and possibly by furnace conditions [see Chapter 9]. A few dark blue samples get the colour from cobalt and/or copper. The low iron green and some brown glass were made from impure sand with the addition of natron/trona and plant ash as flux. This (colouring) technique is characteristic for Period I. The high iron green glass was made from pure sand with natron as the flux, and iron ore as the colouring agent. All samples from the northwestern provinces of Period II form a homogeneous group, characterised by high iron content and low titaniummanganesemagnesium content. We thus may assume that this group of black glass artefacts was produced from a ‗naturally coloured‘ raw glass, made from very clean sand. We will check the LA-ICP-MS results to see whether the Period II black glass might be also decolourised by the addition of antimony (Sb 2O5). Some of the high iron green glass shows a more elevated contamination of impurities, e.g., alumina, manganese and titanium. This is due to the use of HIMT glass. This subgroup of black glass can thus chronologically be defined in Period IV. The technique adopted to make dark-coloured green glass that appeared black was practiced in Period I using impure sand with high iron contamination, but the sand also contained other oxides besides iron. From Period II, a colouring technique was introduced in the northwestern provinces to make black glass by adding pure iron ore to the batch. The hypothesis that emerges is that black glass was produced locally in one or more (secondary) glass workshops in the northwestern provinces from imported ‗naturally coloured‘ or decolourized raw glass. From the 332 low impurity level of various metal oxides (alumina, titanium, manganese), we may assume that clean sand and pure mineral soda has been used [see this chapter section 8.5.]. 8.4.2 LA-ICP-MS Although LA-ICP-MS measurements can also generate data on the major elements, the technique is elaborate and expensive. We therefore limited the evaluation of this technique to a well-defined selection of samples in order to verify their minor and trace elements. Among the minor and trace elements are characteristic impurities in the sand, or in the flux. These cannot, however, be measured with SEM-EDX. Contrasting them between themselves or with specific major elements can provide additional information for a better understanding of Roman glass production. Figure 160: Plot opposing chromium with vanadium oxide (weight %) Figure 160 demonstrates a homogeneous cluster of low chromiumlow vanadium oxide ratio (15-25 ppm vs 1030 ppm) of the Period II glass that should be understood as glass from clean sand. This contrasts with the rather heterogeneous cluster of high chromium–high vanadium oxide ratio (35-80 ppm vs 25-75 ppm). An interesting detail is that the late Iron Age bangles show a low chromium–high vanadium oxide ratio (40-60 ppm vs 10-20 ppm). In Figure 161 the opposed strontiumtitanium oxide ratio helps to verify the used sand and to refine the search. The result is striking, showing in general two clusters independently varying directly: one with a ratio of 6,42 and the other with a ratio of 1,93. This corroborates Figure 160 on the use of two different sands: correspondingly high titanium sand and low titanium sand. Various samples point to an Egyptian origin of the high titanium glass, whereas the low titanium glass seems to be of Levantine origin. This corresponds integrally with the two clusters in the previous plot (Figure 160). On the other hand, we see three specific clusters in the low titanium group on the basis of their different strontium content (Table 147). 333 Table 147: Different glass types on the basis of the strontium-titanium ratios SrO TiO2 Cluster 1 Cluster 2 Cluster 3 c.0,05% c.0,07% 0,1-0,15% <0,1% 0,1-0,2% c.0,2% 1,0000 0,9000 0,8000 resins PC11-23 0,7000 CASTRA OUDENBURG AUGST TiO2 0,6000 AVENCHES LONDON BRACELETS 0,5000 ELKAB MAGDALENSBERG 0,4000 OLBIA MATAGNE 0,3000 TIENEN RUMST 0,2000 ORESAC STE MENEHOULD 0,1000 0,0000 0,0000 0,0500 0,1000 SrO Figure 161: Plot opposing strontium with titanium oxide (weight %) Figure 162: Plot opposing zirconium with titanium oxide (weight %) 334 0,1500 0,2000 However, when plotting titanium against zirconium, another impurity element linked with sand, only Clusters 1 and 2 are still observable, while Cluster 3 has been integrated in the large heterogeneous high titanium group (Figure 162). We consequently have to consider the high strontium content of Cluster 3 as a contamination from added plant ash and not as solely a sand impurity. Cluster 3 features the Avenches (CH) samples, which have a high magnesium and potash content (Figure 151) [see Appendix 1, resins PC22-23; 35]. This strontium addition from the plant ash alkali is much more clearly visible in Figure 167, where some material from Olbia (FR) [see Appendix 1, resin PC11] corroborates the Avenches material. Particularly Figure 163 shows that the influence of the flux impurities can be countered by verifying the zirconium content. Figure 163: Plot opposing chromium with vanadium oxide (weight %) From the different plots between titaniumstrontium and zirconium, it has been demonstrated that the black glass vessels of Period II were produced from more than one type of raw glass. It is interesting to link this with the location of the different raw glass types to discuss the regionalism of the glass distribution and trade [see Chapter 10; case study 1]. From the SEM-EDX analyses, we were able to verify the use of manganese, but antimony is another element that was frequently used as a decolourizer (Braun 1983). We checked which decolourizer had been used (Figure 164). The Period I material from Magdalensberg (AT) has only some presence of manganese that has to be considered as an impurity in the sand. It is clear that the pieces were made from ‗naturally coloured‘ glass. Period II material is characterised by the presence of antimony on the whole ranging between 0,2% and 0,5%. Antimony decolourized raw glass was applied in Period II to produce black glass artefacts. 335 0,9000 Elkab 0,8000 Magdalensberg 0,7000 Faulquemont Period II/III Rumst 0,6000 Sb2O5 Tienen 0,5000 Augst Trou de Han 0,4000 Dilsen 0,3000 La Tène 0,2000 Trier-Kesselstatt Castra 0,1000 0,0000 0,00 HIMT+decolourized Oudenburg La Tène (purple) Olbia 0,50 1,00 1,50 2,00 2,50 3,00 3,50 4,00 MnO Figure 164: Plot opposing manganese with antimony oxide (weight %) 8.5 Archaeological Relevance 8.5.1 Augst vs Avenches One of the key questions we wanted to see answered through archaeometry was whether the distribution of the nd rd late 2 and early 3 century AD black glass vessels originated from one or more glass workshops, almost exclusively found in the provinces Germania Inferior, Germania Superior and Gallia Belgica [see Chapter 6]. We therefore examined samples from vessel fragments coming from Belgium, France, the Netherlands and Switzerland: Oudenburg (BE), Kruishoutem (BE), Velzeke (BE), Grobbendonk (BE), Tienen (BE), Rumst (BE), Matagne-la-Petite (BE), Tongeren (BE), Nijmegen (NL), Hoogeloon (NL), Augst (CH), Avenches (CH). From LA-ICP-MS results it has been demonstrated that the Period II material is very homogeneous and made from one single area, showing two or three subtypes. When considering solely the two Swiss sites, Augst and Avenches distinct clusters are detectable. While the Augst samples make one homogeneous group with a very distinct chemical composition, we see that the Avenches samples are split up into different clusters (Figures 165-167). 336 Na vs K 3,00 2,50 Avenches Rumst - Belgium Faulquemont - France Tienen - Belgium Augst - Switzerland Hoogeloon - Netherlands Auvelais-Belgium Matagne-la-Petite - Belgium Avenches-Switzerland Nijmegen -Netherlands Oudenburg-Belgium Grobbendonk-Belgium Kruishoutem-Belgium Velzeke-Belgium K 2,00 1,50 1,00 Augst 0,50 0,00 0,00 5,00 10,00 15,00 20,00 25,00 Na Figure 165: Plot opposing soda with potash oxide (weight %) Al vs Ti 5,00 4,50 4,00 Rumst - Belgium Faulquemont - France Tienen - Belgium Augst - Switzerland Hoogeloon - Netherlands Auvelais-Belgium Matagne-la-Petite - Belgium Avenches-Switzerland Nijmegen -Netherlands Oudenburg-Belgium Grobbendonk-Belgium Kruishoutem-Belgium Velzeke-Belgium 3,50 Al 3,00 2,50 Augst 2,00 1,50 Avenches 1,00 0,50 0,00 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 Ti Figure 166: Plot opposing alumina with titanium oxide (weight %) 337 Mn vs Fe 18,00 16,00 green 14,00 Rumst - Belgium Faulquemont - France Tienen - Belgium Augst - Switzerland Hoogeloon - Netherlands Auvelais-Belgium Matagne-la-Petite - Belgium Avenches-Switzerland Nijmegen -Netherlands Oudenburg-Belgium Grobbendonk-Belgium Kruishoutem-Belgium Velzeke-Belgium 12,00 Fe 10,00 HIMT 8,00 olive green 6,00 Augst 4,00 brown 2,00 Avenches purple 0,00 0,00 0,50 1,00 1,50 2,00 2,50 Mn Figure 167: Plot opposing manganese with iron oxide (weight %) The uniformity of the Augst-group may have been caused by taking samples from vessel fragments of a similar shape, i.e. globular beakers (Forms IIB.4/5). Since all pieces are from different objects excavated at different locations and at a certain point in time, the chemical homogeneity of the sampled material rather tends to suggest a single batch or a precise production technique in which one specific recipe was used successively and repeatedly applied in one single glass workshop. When selecting the carinated beakers (Form IIB.1) the most characteristic black glass vessel shape in the North-western provinces of the Roman Empire during Period IItwo distinct clusters appear in the plots when comparing different oxides significant for the sand and the colorants used. It is most probable that there were two contemporaneous glass workshops producing their own black glass by using two different recipes. The other option, encompassing two consecutive periods of activity within one single workshop, can be ruled out by the archaeological study, which shows that one cluster is specific for the sites in Germania Inferior and that the other cluster is material from sites in Gallia Belgica. It seems very unlikely that, by coincidence, the sampled material from Germania Inferior and Gallia Belgica, presented here, was made during successive generations of glass workers from the same glass workshop. We therefore believe that there must have been at least two contemporary glass workshops in the North-western provinces where black glass (vessels) were produced. One of these has to be located in the province of Gallia Belgica, while the other must have been active in the province of Germania Inferior. Because the black glass artefacts from the sites in Gallia Belgica are unanimously related and dissimilar to the material from sites in Germania Inferior we have a strong argument to state that both glass workshops brought the same vessel shapes into production for a regional market. This can have important consequences in the interpretation of the glass production and glass distribution in general during Period II [see Chapter 10]. 8.5.2 Deviant material Concerning the production of black glass we had on several occasions the possibility of sampling lumps of (raw) glass from various sitese.g. Rumst (BE), Vremde (BE), Treignes (BE), Sainte-Menehould (FR), Lyon (FR), Nijmegen (NL) and Cyprus as well as glass from crucibles from Trier (Table 148). However, the compositions at first offered little valuable information on the production or consumption. The heterogeneous composition displays a wide variety of non-standard compositions for Roman glass, with the exception of the material from the crucibles of Palais Kesselstatt, Trier (DE), showing a normal soda and lime content. The Trier crucibles demonstrate a composition equal to the produced black glass artefacts from the Period IV, namely HIMT glass. The very anomalous composition of the other samples show an uncharacteristic composition of soda, which is 338 nearly absent; and potash remains low for most samples. Striking is the high alumina and magnesium content. This finding has recently been discussed by Bernard Gratuze (2009, fig.1) who sees this as characteristic of ‗famille 1‘ of Picon‘s ‗groupe 3‘. However, his group of blue glass contains alumina between 4% and 8% while the Al-content of some black glass samples are much higher, even up to 15%. Also striking is the extremely high lime content, averaging between 20-25%, which constitutes together with the silica about 70-80% of the melt. Most of these rough chunks of Al-Ca glass show a combined high content of iron and manganese whereas the two Nijmegen pieces from resin PC23 and the Cypriot sample have just a high iron content. The two samples from Sainte Menehould demonstrate low concentrations for both oxides. Conspicuous is the high Mn and high Fe content in the opaque turquoise lump of glass from Treignes. Would it be that manganese decolourized glass was th used in the production of black glass in the later 4 century. It could perfectly be that HIMT glass was used in Trier. The Mg-content of sample PC13a exceeds by far the proposed 1.5% limit of HIMT glass (Mirti et al. 1993; Freestone 1994; Foy et al. 2003) and can only be explained by the addition of plant ashes since the sample also exceeds the proposed 1.5% limit of potash. Table 148: Major elements of the Roman glass lumps and crucible glass obtained from SEM-EDX (n.d. = not detected; italic = not appearing black). sample provenance Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO 6,71 21,45 0,21 0,41 2,15 n.d. 0,1 n.d. 1,22 21,46 0,74 2,45 10,96 <0,1 n.d. n.d. 1,17 23,66 0,76 2,58 9,92 <0,1 n.d. n.d. 1,20 24,44 0,83 2,68 8,01 n.d. n.d. n.d. 1,25 21,81 0,82 2,52 8,85 n.d. 0,4 2,50 8,37 0,61 1,54 6,85 62,49 0,2 n.d. n.d. 0,6 0,97 5,80 0,45 1,60 7,30 2,04 61,19 n.d. 6,90 0,19 1,63 7,25 10,71 57,68 n.d. 0,9 n.d. 0,63 2,98 0,3 n.d. 4,06 23,09 0,40 0,35 0,64 <0,1 1,42 15,22 41,76 n.d. n.d. 1,51 38,58 0,17 0,44 0,30 Nijmegen(NL) 0,8 3,06 8,63 36,77 2,9 0,6 n.d. n.d. 1,10 27,52 7,86 0,28 10,98 PC23j Nijmegen(NL) 1,0 3,06 8,97 36,98 3,6 n.d. n.d. 1,15 27,20 5,82 0,37 11,81 PC28b Cyprus 0,8 7,91 14,75 45,32 0,2 <0,1 <0,1 3,28 15,76 1,04 0,16 10,66 PC42i Treignes (BE) 0,1 1,31 6,15 57,63 0,1 0,1 <0,1 2,48 20,54 0,43 1,52 9,39 PC31d Nijmegen(NL) 0,5 1,7 6,2 54,4 0,2 0,3 0,0 1,0 21,0 1,8 2,5 10,4 PC31g Nijmegen(NL) 0,2 1,9 7,3 56,8 0,1 0,1 0,0 1,4 16,6 1,0 1,5 13,0 PC48i Treignes (BE) PC50a Lyon (FR) 0,3 4,8 2,1 0,3 9,0 11,6 55,0 74,3 <0.1 n.d. 0,1 n.d. <0.1 0,2 2,3 4,7 21,5 0,8 0,6 0,3 1,5 n.d. 7,5 3,1 PC50b Lyon (FR) 4,8 0,3 11,6 74,4 n.d. n.d. 0,2 4,6 0,7 0,3 n.d. 3,0 n.d. 52,90 1,8 <0,1 8,54 51,53 8,95 52,34 1,77 9,03 53,62 15,8 2,46 2,96 58,34 Trier (DE) 17,2 1,01 2,59 PC13c Trier (DE) 18,0 1,05 PC14h Ste Menehould(FR) 0,1 PC14j Ste Menehould(FR) PC23i PC1f Rumst (BE) 1,6 3,8 2,95 58,86 PC3a Vremde (BE) 0,1 1,6 8,55 PC3b Vremde (BE) 0,1 1,7 PC3c Vremde (BE) <0,1 1,51 PC3d Vremde (BE) 0,4 PC13a Trier (DE) PC13b Fe2O3 Inaccurate measurements have to be excluded, since the given values are averages from five measurements on different places and calibrated with various standards. The first samples, Rumst and Vremde, were so atypical we thought them to be the result of intrusive modern production from non-ferric industry. On the other hand most other samples, mainly consisting of chunks of raw glasse.g. Sainte Menehould (FR), Nijmegen (NL) (PC23), Treignes (BE) (PC42), Cyprusdemonstrate the same anomalies. This was thought to be due to the type of material, but now two vessel fragments from Nijmegen (resin PC31) and a counter from Treignes (resin PC48) have increased the confusion. Especially these last three examples indicate such results must be taken seriously being part of the Roman glass production. But this premise poses a dilemma since we then take for granted that a glass workshop must have been active somewhere on the sites under consideration here or in the near vicinity. In view of production material we wonder whether this nearly zero soda material was not an in between production to be added as colorant and opacifying agent in the glass batch instead of seeing it as production waste. A different composition is observable for the early Byzantine material from Tunisia. A tessera in black glass from th th the early Byzantine church of Sidi Jididi was the start of some unexpected glass consumption in the 5 -6 century 270 AD. Only after analysis of the seven sampled lumps of black glass from the contemporaneous church of Bir 271 Messaoudi in Carthage (TN) we understood its importance. A significant detail is that the lumps from Bir Messaoudi were found in an adjacent room from the church together with a wide range of material pointing to a 272 mosaic workshop (unpublished). The tessera and the lumps have a similar composition deviating from the 270 We wish to thank Danièle Foy (CNRS, Aix-en-Provence) to have given at our disposal the sample. We wish to thank Prof. dr Roald Docter (UGent, Ghent) for the kind permission to study and sample the material for integration in the black glass research. The analysis on one sample (PC28c) failed in providing its chemical composition. 272 Personal communication Roald Docter (UGent). We are grateful to Roals Docter for putting the material at our disposal. 271 339 norm. All six samples from Bir Messaoudi and the Sidi Jididi-tessera show a very homogeneous composition of mixed-alkali glass with high alumina and iron content while lime is lacking (Table 149). In contrast with the preceding group we see an almost complete absence of magnesium as well as a very low concentration of titanium and manganese. When comparing the Tunisian glass cakes and tesserae with those from Petra, Jordan (Marii, Rehren 2009, fig.2) we notice a marked difference in the phosphor and lime-content. The Petra material has high concentrations of calcium, between 5% and 15%, and of phosphor, between 0.10% and 1%. Since no further information on the composition of this material was available we were unable to verify them with the glass lumps in Table 148. Table 149: Major elements of early Byzantine mosaic material (n.d. = not detected) sample provenance PC11n PC28d PC28e PC28f PC28g PC28h PC28i Sidi Jididi (TN) Carthage (TN) Carthage (TN) Carthage (TN) Carthage (TN) Carthage (TN) Carthage (TN) PC50a PC50b Lyon (FR) Lyon (FR) Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 7,5 7,9 7,9 8,0 8,0 5,3 8,1 4.8 4.8 0,0 0,1 0,1 0,1 0,1 0,1 0,0 0.3 0.3 6,4 6,2 5,9 6,2 6,0 10,1 6,3 11.6 11.6 71,7 71,4 71,6 71,4 71,3 75,2 71,3 74.3 74.4 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 0,3 0,4 0,5 0,5 0,5 0,3 0,4 0.2 0.2 4,4 4,3 4,4 4,4 4,4 4,9 4,4 4.7 4.6 0,3 0,3 0,3 0,3 0,3 0,8 0,3 0.8 0.7 0,2 0,3 0,3 0,3 0,3 0,2 0,3 0.3 0.3 0,3 0,3 0,3 0,3 0,3 0,1 0,3 n.d. n.d. 8,8 8,6 8,7 8,6 8,8 2,9 8,6 3.1 3.0 0,1 n.d. n.d. n.d. n.d. n.d. n.d. th We know that integrating mid 6 century AD glass in the present research is colouring beyond the lines, but necessity knows no law. Combining the chemical composition of this specific Tunisian material with its context makes us understand that in the early Byzantine period lumps of black glass were used to make tesserae for rd th mosaics in large public buildings such as churches and that the composition is clearly different from the 3 -4 century glass lumps analysed here. We assume that early Byzantine mosaic workers recycled large lumps of glass to produce tesserae. From this point of view we may presume the possibility of tesserae production in large glass workshops like the one at Sainte Menehould where tesserae in various opaque colours were produced, including those with a black appearance. Unfortunately we were not able to sample the Sainte Menehould material at the M.A.N. in Saint-Germain-en-Laye (FR). On various occasions scholars have pinpointed on the use of tesserae in glass workshops for bead making in particular for the early medieval period (Lundström 1976, 5). 8.6 Conclusion Although the analysis of black glass may concern a limited segment within the Roman glass production, the results contribute to a better understanding of Roman glass workshops of different periods and from different areas. The obtained results on major elements (SEM-EDX) as well as on minor and trace elements (LA-ICP-MS) provide knowledge about recipes using specific sands, fluxes and colorants, and consequently have the potential to yield clusters linked to chronological and geographical differences in the production technology. Information on the chemical composition can thus help in tracing the specific production technologies developed through history and in identifying possible production centres and trade routes. The data on the chemical composition of 438 samples of Roman black glass acquired via various measuring methods have shown it is relevant to examine extensively the chemical composition of a thin section of archaeological glass material. It has given new clues about specific distribution patterns and accordingly can help clarify trade routes. It allows us to understand better the production techniques used by the Roman glassworker and the levels within this production process. In the introduction, we proposed various queries to verify the Roman black glass. From the obtained results, a number of questions concerning the technological issue could be answered. For instance, we were able to define the use of different sands to produce black glass in the Roman imperial period. Two major groups were recognized which appear to be related to the two main centres of glass production. First, there is an ‗Egyptian‘ raw glass with high concentrations of alumina, manganese and iron as major impurities and high concentrations of titanium, vanadium and chromium. The heterogeneity of its composition can be explained from the long period of five centuries. Second, there is a ‗Levantine‘ raw glass with a very homogeneous composition made from pure sand with a low concentration of impurities (Table 150). Table 150: Values of magnesium, potash and strontium in %wt Al2O3 TiO2 Egyptian Levantine 340 1,5-4,5 1,3-4,0 >0,2 0-0,2 ZrO2 Cr2O3 V2O5 0,006-0,032 0,004-0,022 0,0035-0,0080 0,0015-0,0025 0,0025-0,0075 0,0010-0,0030 Different fluxes were used to melt the raw glass during the production process of black glass. Mineral soda, natron or trona, was the main source to lower the temperature of the batch, but vegetable soda was used as well, though on a lesser scale. The plant ash alkali used resulted in a high concentration of magnesium, potash and strontium, clearly distinguishable from the mineral soda glass. It appears that the vegetable soda glass can be correlated to the ‗Egyptian‘ group, based on its absence in the glass of the ‗Levantine‘ group (Table 151). Table 151: Values of used flux soda, magnesium, potash and strontium in %wt Na20 MgO Egyptian Levantine 15,0-20,0 13,0-17,0 0-3,5 < 1,5 (mainly 0,4-0,8) K20 SrO 1,0-2,5 0,5-1,5 0,01-0,07 0,04-0,1 The Romans used ‗naturally coloured‘ as well as decolourized raw glass to make black glass. Depending on the period and the origin, various categories can be discerned. The ‗Levantine‘ group is characterized by antimony presence, whereas antimony is totally absent in the ‗Egyptian‘ group. Depending on the period, this group demonstrates a low manganese or high manganese content, respectively Period I and Period IV (Table 152). Table 152: Values of used decolourizers manganese and antimony in %wt Sb2O5 MnO Egyptian Levantine < 0,1 0,1-0,6 0,3-2,2 0-0,8 The Romans not only used different types of raw glass in the production process of black glass. They also coloured the glass by using particular colouring agents. Besides iron (green and brown) and manganese (purple) they in a lesser degree used cobalt and/or copper (blue) to colour the glass matrix so that it appeared black. At this stage of the investigation of black glass samples we can only draw out rough lines for the chronology of the various colours. Where brown-, blue- and purple-coloured black glass are mainly limited to Period I, the green glass can be linked to Periods I-IV. Refining this huge group recognizes a low iron group for Period I and II, and a high iron group for Periods II-IV where Period IV black glass is characterised by an elevated manganese and titanium content. The applied furnace conditions could not be verified through the chemical analysis, nor were the necessary or 273 used temperatures to melt the black glass calculated. Concerning the archaeological issue, the use of different sands, fluxes, decolourisers, and colourants cannot be linked with specific functional types and consequently not to particular form types within one functional type. The only obtainable subdivision per functional type or form type is relative to the raw glass used when it is confined to a specific and limited period. The perfect example is the Period IV bracelet of type D1 which is made from the late Roman HIMT glass. It is also true that the obtained results can provide information on the quality of the glass used to produce specific finished products. For instance, the Period II vessels in the north-western provinces are made from very pure decolourised raw glass; in other words, a high quality glass required high standards to guarantee finished products of excellent quality for the consumer. The apparently random occurrence of the different glasses and the long continuity of their use make it difficult, if not impossible, to discern specific workshops, production periods or consumption areas. Consequently the results corroborate model 4 in the organization of Roman glass production [see Chapter 6] that points to a locally organised production of black glass artefacts within secondary workshops from imported raw glass, whether or not already coloured black and distributed regionally. This would limit the interpretation possibilities of the obtained data only to the rough determination of the date and distribution patterns of general groups of artefacts within a specific period. A more detailed examination of the obtained results indicates that the chemical composition can provide a clearer understanding of the typo-chronological idiosyncrasies and the organization of the (black) glass production. In Chapter 10, we verify the validity of the abovementioned hypothesis and present a more multi-levelled model implying an ‗and-and‘ situation instead of an ‗or-or‘ situation. The obtained data demonstrated that the various specific recipes give information on chronological development in the black glass production related to the economic organization of glass production in particular and the Roman economy in general [see Chapter 10]. Question was whether each recipe could be reduced to one single 273 The calculation of the theoretical melting temperatures is obtainable through the plotting of the reduced base compositions of glass in an adapted silica-soda-lime ternary plot (Rehren 2000; 2001). 341 secondary glass workshop that produced artefacts in black glass in a well-defined period, because it is equally possible that black glass (artefacts) from one and the same recipe were produced in various Roman glass workshops within a specific region or even at random. To improve our understanding of the chemical composition of archaeological glass, it is sometimes necessary to check and verify the analysis results from different angles. We demonstrated that the samples from Augst (CH) and Avenches (CH) are characterised by a high homogeneity. Ensuing from this is the possibility of identifying secondary glass workshops in both towns that have contemporaneously produced artefacts in (black) glass on the basis of the different recipes [see in this chapter 8.4]. We were not able to link the contemporary diversity of recipes to a regional variability, due to an apparently free distribution and trade without restrictions or regulations. This interference is even more amplified due to the possibilities of recycling broken glass and the mixing of various groups. From the chemical data, we can demonstrate that the raw glass used for the production of black glass was massproduced in several centres in Egypt and the Levant. Consequently, two different models of black glass supply are possible in the West: 1) solely imported as finished products, and 2) produced locally from imported raw glass. Generally speaking, it is not unlikely that during Period I black glass vessels were not only imported but also produced locally in Western Europe. In Chapter 6 we discussed a chunk of black raw glass which was retrieved from the glass workshop of La Montée de la Butte at Lyon (FR). We, however, lack all proof for black glass vessel production at that particular workshop. However, since the workshop specialized in the production of vessels, it is very likely that the Lyon workshop also produced vessels in black glass. From archaeological data, we know that chunks of raw glass were exported empire-wide by ship and then further distributed all over the Empire (Nenna 2008a) [see Chapter 6]. The local production of black glass vessels in Period II is demonstrated by the glass workshop of Kaiseraugst (CH) on the one hand, and by the homogeneous composition of the Kaiseraugst/Augst samples. The high iron colouring technique, newly introduced in Period II (Van der Linden et al. 2009), is clearly attested from the Kaiseraugst/Augst samples, excluding the import of chunks of black glass. Moreover, the homogeneity of the Kaiseraugst/Augst samples provides evidence of local colouring from imported ‗Levantine‘ antimony decolourized raw glass. But how should we understand this emergence? Was this simply a logical consequence of the constant development of the glass-working knowledge? Was it because of economic stress? Or was it culturally determined and just a trend? The ―new‖ colouring technique introduced in Period II to produce black glass was at least established to supply the regional workshops in the area where black glass was consumed. Furthermore, the appearance of an idiosyncratic assemblage of black glass vessels let us hypothesize that there could have been an intermediate glass workshop in Western Europe, where large quantities of imported ‗naturally coloured‘ and decolourized raw glass was coloured in a variety of vivid colours, including black. The former model fits in the traditional system of supply and demand for well-defined regional markets, thus causing regional and local idiosyncrasies, whereas the latter claims the existence of a proto-industrialized organisation of glass workshops, thus implying a much greater sameness of forms and types and a compositional homogeneity. However, the question is whether one single model is appropriate to explain the glass production throughout the entire Roman imperial period, or not various models emerged all through the consecutive centuries. In Chapter 9, we discuss how to interpret this innovation and assess how to value the low iron and the high iron vessels in black glass. Everything considered we described on the basis of a large set of samples that specific elements point to a different source of raw glass or to a different colouring technique to produce black glass during the Roman imperial period. This research demonstrates that a diachronic selection of very characteristic glass samples from a wide area offers added value to the glass research when the sample selection involves the entire range of black glass colours, chronological and geographical subdivisions, functional types, and the large site contexts. 342 8.7 Addendum This section is added to discuss the use of non-destructive analysis methods that have been applied on a selection of museum objects in the KMKG-MRAH, Brussels and Grand Curtius, Liège which were impossible to samplerespectively the portable XRF (see 8.7.1.) and µ-XRF (see 8.7.2.). On the other hand we describe less commonly used measuring methods within the archaeometry of glass, i.e. PIXE-PIGE (see 8.7.3.) and Raman spectroscopy (see 8.7.4.). We verified whether these alternative techniques could provide similar results or possibly additional information. Some have the advantage of portability, so that they can be carried out on location in museums (p-XRF; µ-XRF), whereas others need specialized labs (PIXE-PIGE; Raman) and involving time-consuming and sometimes complicated permission procedures. Some of these analyses and verifications could only be carried out during the last year and months of the present research programme. Consequently, we integrated them as appendices in this research; however, considering their potential contribution to the glass research, these non-destructive methods deserve more attention. The obtained results have a promising effect, even though we were unable to work out the use of these measuring methods extensively within this manuscript. Non-destructive methods can supply data on the chemical composition to objects of interest to the research questions when sampling is impossible. In order to answer particular questions in relation to various aspects of the production process on Roman black glass Prof. Koen Janssens of the Centre for Micro and Trace Analysis (MiTAC) at the University of Antwerp (UA) was willing to provide a small portable XRF and a larger, more efficient, µ-XRF supplying corresponding qualitative and semi-quantitative information on location and to allow one of his experienced assistants, Simone Cagno, to take measurements. Finally, Prof. Herman Terryn enabled us to perform Raman spectroscopy on a large set of samples at the ESEMC of the Vrije Universiteit Brussel (VUB) within the Horizontal Orchestrated Action (HOA) project 15. We are also grateful to Prof. Dr. Koen Janssens who granted us the possibility to have a set of samples tested in the synchrotron at the Hamburger Synchrotronstrahlungslabor HASYLAB at Deutsches Elektronen-Synchrotron DESY, a Research Centre of the Helmholtz Association in Hamburg. In 2009 Simone Cagno analysed three thin sections of the resins PC2, PC6, PC32 containing purposely selected black glass samples, in order to verify whether crystalline structures were present in the glass opacifying the matrix, and thus responsible for the obtainment of a black appearance. The immediate cause leading to the verification of this aspect was the result of the concentrations of Pb-oxide together with S-oxide within the glass matrix. 8.7.1 p-XRF and µ-XRF From a museological point of view, it was impossible to sample the so-called obsidian vessel fragment R1610 and a set of black glass objects from the Royal Museums of Art and History, Brussels (Cagno et al. 2010). This excluded the SEM-EDX and LA-ICP-MS techniques and thus the possibility of acquiring quantitative results on those pieces. A first session was done with the p-XRF technique and the following session was executed by using a micro X-ray fluorescence (µ-XRF) installation. Both instruments give qualitative information on the main part of the chemical elements that are present in the glass since the light elements are not detectable. The p-XRF and the µ-XRF are sensitive analysis methods; however, due to open-air measurements, the air attenuates the X-rays at low energy so that the first 12 elements of the periodical table are not gaugeable. This means that important elements within the glass such as soda (Na), magnesium (Mg) and alumina (Al) are missing. The required installation to obtain a full and consequently a quantitative measurement can be obtained by integrating a vacuum box into which the artefact is placed. This very recently developed technique (Tantrakarn et al. 2009) was not at our disposal within the current black glass research project. 8.7.2 Portable X-Ray Fluorescence (p-XRF) X-ray fluorescence (XRF) is a very welcome measuring technique in obtaining fast information on the composition of museum objects that cannot be sampled. In concreto, this technique implies the excitation of the material by bombarding the artefact with X-rays and subsequently detecting the radiation emitted by the object. Because this 343 reliable and non-destructive measuring technique can be used safely with a compact and portable instrument, easily movable to any museum, this method is extremely appropriate for heritage-related applications. An X-ray beam, exciting the atoms of the sample placed on the XRF pistol, causes the expulsion of electrons from the inner orbitals. This gap hence creates instability, resulting in electrons from outer orbitals gravitating towards this empty position. During this process, the characteristic X-ray radiation is detected. Consequently, a qualitative measuring of the composition is possible, since the emittance of X-ray radiation varies depending on the chemical element. The portable energy dispersive X-ray fluorescence machine (p-XRF) employed is the TRACeR III-V. This portable, wide range elemental analyzer is equipped with a rhodium X-ray tube and a high resolution Peltier 274 cooled Silicon PIN (SiPIN) diode detector. The operating conditions of the X-ray tube are 15 kV, 15 mA, while 100 seconds has been adopted as measuring time. The p-XRF is connected to a laptop run by p-XRF 32 software (Figure 168). Figure 168: The installation of the p-XRF Interesting preliminary results could easily obtain by comparing the X-ray spectrum of the R1610 piece with, on the one hand, the spectra of two pieces in unmistakably raw obsidian and, on the other hand, those of two clearly black glass samples, one modern and one Roman. When looking to the general plot, we see no very clear differences between the five analysed pieces, except for the peaks in three specific zones (Figure 169). The first zone, which has peaks between 1.4-1.9 KeV corresponding to the K-lines of alumina (Al) and silica (Si), stand in correlation to the sand. The second zone, which lies between 3.3-4.1 KeV and matches the K-lines of potash (K) and calcium (Ca), refers to the flux/stabilizer within the glass melt. Finally, the Fe- and K-lines between 6.3-7.1 KeV refer to the colorant. The red line of the R1610 piece floats between the high black glass curves and the low obsidian curves. It is worthwhile to look at these zones in more detail by selecting them separately and by blowing up the specific peaks. When looking at the AlSi peaks (Figures 169 left; 170), we see that the vessel fragment R1610 demonstrates an X-ray spectrum that clearly aligns with that of the obsidian pieces. The black glass pieces do not contain aluminium and show a lower intensity of the Si-peak compared to the obsidian artefacts. Within the KCa zone (Figures 169 middle; 171), we can observe that the X-ray spectrum of vessel fragment R1610 aligns rather well with that of obsidian piece 2, whereas obsidian piece 1 has a slightly more diverging spectrum with a lower potash and a higher calcium content. Nevertheless, the difference is not as strong as with the two black glass artefacts which show a much lower intensity for the potassium peak (K) and a much more intense peak for calcium (Ca). 274 Personal communication by Veerle Van der Linden 2008. 344 Concerning the iron X-ray spectra (Figures 169 right; 172) the converging curves of the two black glass pieces (very high peak) and the two obsidian pieces (very low peak) show an intermediate situation for R1610 (red line). To summarize, the use of a portable-XRF can easily verify whether a black glass artefact is made of volcanic glass or from sand. But when the artefact is determined as obsidian it does not confirm whether it was made from a raw block of obsidian by drilling or by remelting and casting. Al-Si K-Ca Fe Figure 169: Energy spectrum of R1610 compared with the spectra of black glass and obsidian pieces Figure 170: Detailed view of the X-ray spectra of the aluminium (Al) and silicium (Si) peaks (taken from Cagno, Cosyns 2009a, unpublished internal report) 345 Figure 171: Detailed view on the X-ray spectra of the potash (K) and calcium (Ca) peaks (taken from Cagno, Cosyns 2009a, unpublished internal report) Figure 172: Detailed view of the X-ray spectra of the iron (Fe) peaks (taken from Cagno, Cosyns 2009a, unpublished internal report) 8.7.3 Micro X-Ray Fluorescence (µ-XRF) A different transportable XRF system used is the more efficient micro energy dispersive X-ray fluorescence machine (µ-XRF). This movable, wide-range elemental analyzer is equipped with a rhodium X-ray tube and a high 275 resolution Peltier cooled silicon PIN (SiPIN) diode detector. The operating conditions of the X-ray tube were 35 kV, 0,15 mA, while 300 seconds has been adopted as measuring time. The µ-XRF was connected to a laptop likewise the procedure when using a p-XRF. The measurement done by µ-XRF provided additional information concerning the vessel fragment R1610 (no. 6), but we took the opportunity to have several other items of the Royal Museums of Art and History analysed that were impossible to sample for SEM-EDX or LA-ICP-MS. As shown in Tables 153--154, the 11 objects concerned were four counters, two arm rings, two drinking vessels, two pseudo-nicolo gems, both set in their original ring and the obsidian vessel R1610. 275 Personal communication by Veerle Van der Linden 2008. 346 Table 153: List of additional artefacts analysed with µ-XRF cat.no. sample no. site type of object 317 317 228 228 194 2790 552 335 199 not catalogued 221 counter1 counter2 counter1 counter2 arm ring obsidian vessel arm ring drinking vessel (carchesium) drinking vessel (bulbous cup) pseudo-nicolo gem in silver ring pseudo-nicolo gem in bronze ring 1 2 3 4 5 6 7 8 9 10 11 Herstal (BE) Herstal (BE) Cortil-Noirmont (BE) Cortil-Noirmont (BE) Basse-Wavre (BE) Rome? (IT) Tongeren-Koninkshem (BE) Lavacherie (BE) Bois-et-Borsu (BE) Liberchies (BE) Braives (BE) Concerning the counters, two pieces come from the 12 black glass counters of a set of 27 found in the rich tomb of Herstal (BE) (Amand, Mariën 1976, no.33). Counter no. 1 shows high iron content and traces of lead and copper. The presence of the latter two oxides are features of recycled glass that entered the glassmaking process. The counter is therefore most presumably made from recycled glass. The high iron content, together with the general glass composition (e.g., low potassium content) is proof that the counter was made with technology that came into use from the second half of the 2nd century AD onwards only (Cosyns et al. 2006a; Van der Linden et al. 2009). The other counter (no. 2), even though its iron oxide appears lower, can still be regarded as a piece issued from the same technology. We therefore presume quite securely that the two nd analysed counters must have been produced within the second half of the 2 century AD and are consequently from the Herstal tomb. This date fits perfectly with some well-dated glass pieces from the tomb, such as the nd rd footed beaker with snake-thread decoration which is typical for the later 2 to early 3 century AD. Concerning the set of counters from the rich tomb of Cortil-Noirmont, two different colouring techniques can be observed in the two analysed counters. The composition of counter no. 4 is similar to the aforementioned no. 2, whereas counter no. 3 is characterised by a black glass with a dark purple hue, obtained by the addition of the manganese (Mn) ore to an originally ‗naturally coloured‘ or aqua-coloured glass. The visibly purple-on-white eyedecoration shows a slightly different composition, with less iron and manganese and more antimony; a further investigation seems to be needed to explain the nature of this decoration. Table 154: Qualitative data (net intensities) of samples analysed by µ-XRF (data taken from Cagno, Cosyns 2009b, unpublished internal report) ID 1F 1B 2F 2B 3F 3F 3F 3B 4F 4B 5F 5B 6 7 color black black black black black purple white black black black black black black black Si 3479 3196 3525 2902 3354 3152 2856 2956 4758 4549 1546 3066 4816 3812 Cl 1677 1546 2122 1590 1655 1008 853 1701 2030 1946 226 867 716 1498 K 8494 5808 15007 12306 11396 9285 9159 10177 8349 7618 1881 8012 40943 7996 Ca 65151 56023 84178 71232 78085 59576 55650 78775 67065 65125 34011 70174 10543 78719 Qualitative data (net intensities) Ti Mn Fe 1931 7199 203059 1751 6527 195113 2673 12056 55711 2183 10368 48169 2245 98872 41904 569 22523 22393 562 6661 15710 2188 91929 39764 2864 5591 39273 2651 5400 38275 978 6673 106137 1970 16103 245371 4381 4921 118603 1621 19309 55315 Cu 2956 261 743 626 1007 1137 968 1102 744 821 437 1107 340 20495 Rb 368 - Sr 482 4668 1390 1172 1318 1049 669 1209 889 948 288 521 96 846 Zr 8 15 197 143 183 121 118 156 203 153 35 108 497 73 Sb 1081 701 335 5305 5512 387 810 2085 Pb 15145 10542 389 357 453 406 918 901 172 370 10299 From Table 155 it is clear that the µ-XRF measurements on the vessel fragment R1610 with incised floral motifs (no. 6) corroborates those resulting from the portable XRF (see above). The vessel fragment can definitely be distinguished from Roman black glass and is thus to be interpreted as made from obsidian (Figures 173-174). The difference with other obsidian pieces and the strong opacity of the vessel fragment makes it likely that the matrix must have been manipulated by adding colouring and/or opacifying agents to the batch. Thus, instead of being carved from one single block of obsidian, it is more likely that the vessel was made from crushed obsidian and iron. The possibility of an addition of natron (sodium source) to lower the melting temperature can also be considered; however, it was not possible to detect sodium with the µ-XRF instrument. Quantitative analysis only can give an explicit answer on this issue. 347 Table 155: Semi-quantitative data (elemental ratios) of the samples measured by µ-XRF (data taken from Cagno, Cosyns 2009b, unpublished internal report) sample no. site K/Ca Mn/Fe hue glass type 1 2 3 4 5 6 7 8 9 10 11 Herstal (BE) Herstal (BE) Cortil-Noirmont (BE) Cortil-Noirmont (BE) Basse-Wavre (BE) Rome? (IT) Tongeren-Koninkshem (BE) Lavacherie (BE) Bois-et-Borsu (BE) Liberchies (BE) Braives (BE) 0.04 0.23 green soda glass 0.32 0.23 green soda glass 2.48 0.26 purple soda glass 0.22 0.15 green soda glass 0.07 0.20 green soda glass 6.93 0.04 black obsidian 0.18 0.37 green soda glass No results were obtained from these artefacts. The intact vessels cause problems to enable a correct measuring and the metal of the finger rings caused interference as it is too close to the measuring spot of the gem. The bracelet from Basse-Wavre (no. 5) is an open type with wide, flat D-shaped section and decorated with rd th lengthwise ribbings (type C2). Typologically dated 3 century AD up to the middle of the 4 century AD (Cosyns 2004), it corresponds to the high iron content characteristic for the material of the post-150 AD Roman period (Van der Linden et al. 2009). On the other hand, we observed a higher manganese content (Mn) in this piece compared to the other pieces analysed (Figure 174). This offers a more precise identification, since higher th th manganese content is typical for late Roman HIMT glass, produced from the 4 century AD onwards to the 6 century AD and most likely in one or more Egyptian glassmaking sites (Freestone et al. 2002). Hence, we may assume with a certain degree of confidence that the bracelet from Basse-Wavre should be dated in the first half of th rd the 4 century AD rather than in the 3 century AD. On the other hand, the results point to the existence of longdistance trade because at least the raw glass material was imported from Egypt or in any case the south-eastern Mediterranean. To this distinctive raw glass was visibly added a high quantity of iron as colorant. Alternatively, we had the opportunity to investigate this hypothesis by means of an experiment (Cagno et al. 2010). We received from Alain Riols about 1.5 kilograms of black obsidian from the Monte Arci on Sardinia, Italy st (Riols 2008). We added industrially produced sodium carbonate (Na 2CO3) and Roman iron from a 1 century AD context of the villa at Treignes (BE), kindly made available by Prof. Eugène Warmenbol of ULB-CReA. Only preliminary results of this work in progress can be given thus far. This first experiment was conducted at the Provincial Archaeological Museum (PAM) at Velzeke during the last week of May 2009 with the kind permission of Marc Rogge and under the supervision of Mark Taylor and David Hill and their team of enthusiastic and skilful nd volunteers. The furnace used is a life-size reconstruction of the 2 century furnace from Cesson-Sévigny built by the skilled hands of François van den Dries, Robert van Zijll de Jong and Peter Van der Plaetse (PAM). To prepare the obsidian for an easy melting, all but two obsidian blocks were crushed and ground at the Department of Geology of the Vrije Universiteit Brussel (VUB) in collaboration with Drs. Rémy Mas and with kind permission of Prof. Dr Philippe Claeys. To facilitate the crushing and grinding of the raw material, we first sawed the blocks into small pieces by using a diamond blade sawing machine with water conduction. Afterwards, the small blocks were put into a mechanical crusher. The resulting small flakes were then put into an agate grinder to obtain obsidian powder. The final outcome was sieved, resulting in a total mass of 1.373 kilograms dividable into 276 grams with a granular size below 50 µm, 675 grams with a size ranging between 50 µm and 2 mm and 323 grams above 2 mm size. Because the latter size was thought to be too large to melt easily and quickly, this amount was not used in the PAM experiment. The total amount to serve the purpose was 938 grams of obsidian powder. To have two different compositions tested, the amount was divided between two crucibles with a similar mixture apart from the iron content. The ratio of obsidian to sodium carbonate was set at c. 469 g/140 g, which corresponds to about 70%-30%. Considering that a part of the 609 g would evaporate during heating, a loss of about 30% of the total glass melt was estimated, in order to calculate the amount of colouring to be added to the batch. To colour the obsidian melt black, we added one crucible of 5% of powdered iron from an unworked iron lens of a Flavian iron production activity on the villa estate at Treignes. Two complete Roman iron nails from the same site were put in the other crucible. After three hours of heating, the two batches started to melt but were still very sticky, like solidifying caramel. The glass melt was of a deep blue-green and full of air bubbles. Two days later, the glass metal could still be inflated but remained a clear, translucent, very dark greenish colour without having a black appearance. Some conclusions can be drawn from the experiment. We saw that it is possible to melt and inflate a simple mixture of obsidian with sodium carbonate, meaning that the addition of arsenic and borax as done by Alain Riols and Allain Guillot is not necessary. Further, we became aware of the necessity of an increased quantity of iron as well as of the addition of an opacifying agent. 348 The bracelet from Tongeren-Koninkshem (no. 7) is of the same type as no. 5, except for its lower iron content. The traces of lead and copper indicate that recycled glass (possibly mosaic tesserae) was used during the glassmaking process. Some general conclusions can be advanced based on the analysis results by means of µ-XRF discussed above , even though the set of analysed material remained limited:  all analysed objects are of a high iron deep green glass appearing black, except for one piece in deep purple glass (no. 3) and the sherd in obsidian (no. 6) (Figure 173); all analysed deeply green-coloured artefacts can be classified in the post-150 AD high iron group; the two bracelets, one from Basse-Wavre (no. 5) and one from Tongeren-Koninkshem (no. 7) are made th 276 of HIMT glass and accordingly are to be dated in the first half of the 4 century AD (Figure 174); one counter from Herstal (no. 1) and one bangle (no. 7) are made from recycled glass (Figure 175).    3.00 2.50 1 2 3 4 5 6 7 Black/purple Mn/Fe 2.00 1.50 1.00 0.50 Black/green Obsidian 0.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 K/Ca Figure 173: Manganeseiron ratios (Mn/Fe) versus potashcalcium ratios (K/Ca) in the analysed samples (taken from Cagno, Cosyns 2009b, unpublished internal report) 276 Based on the typo-chronology, these bracelets went out of circulation around the mid 4th century AD (350360 AD). 349 1.4 1 2 3 4 5 6 7 Purple glass 1.2 1 Mn* 0.8 0.6 Obsidian 0.4 HIMT glass 0.2 Aqua / Levantine glass 0 0 2 4 6 8 10 12 Fe* Figure 174: Manganese intensities (Mn) versus iron intensities (Fe) in the analysed samples (normalised values) (taken from Cagno, Cosyns 2009b, unpublished internal report) 0.25 1 2 3 4 5 6 7 0.2 Traces of Cu-Pb = recycling Pb* 0.15 0.1 0.05 0 0 0.05 0.1 0.15 0.2 0.25 Cu* Figure 175: Lead intensities (Pb) versus copper intensities (Cu) in the analysed samples (normalised values) (taken from Cagno, Cosyns 2009b, unpublished internal report) 350 0.3 8.7.4 Raman spectroscopy It was never a goal as such to verify black glass samples with Raman spectroscopy, but it turned out to be an 277 opportunity that emerged from HOA15-project NOCEM that started in 2007at the Vrije Universiteit Brussel (VUB). This technique has been barely used on archaeological glass because it generally has been estimated not adequate for glass analysis in comparison to other methods providing quantitative measurements such as SEMEDX and LA-ICP-MS [see Chapter 8]. Nevertheless work on glass and glazes by means of portable micro-Raman spectrometers demonstrated recently the possibilities for archaeological research on-site and without the necessity of destroying the artefacts by taking samples (Colomban 2008; Prinsloo, Colomban 2008; Simsek et al. 2010). It was therefore worth checking whether the Raman spectra from the Roman black glass samples offered good readings seeing the interest of the HOA15 project to verify the various measurement systems available in a portable version in order to estimate its value for the creation of a fast measuring portable instrument to detect the chemical composition of the glass artefact or at least aspects of its composition. Current research by means of the standard chemical analysis techniques SEM-EDX and LA-ICP-MS demonstrated a clear compositional distinction between the Roman black glass fragments from Period I, with a low iron content (Fe2O3 = <3%), and Period II, with a clearly higher iron concentration (Fe 2O3 = >6%) (Van der Linden et al. 2009). Since the Roman black glass was coloured with high amounts of iron in various concentrations, the first task that emerged was to check if Raman spectra correlated with the iron oxide concentrations in the Roman black glass artefacts. Thus, the experiments are rather to be considered as an investigation in view of the HOA15-project to test the validity of Raman spectroscopy for measuring glasses with low iron content, as in decolourised or ‗naturally‘ coloured glass, and for verifying other metal oxides in the glass, as, for instance, copper or cobalt.  Methodology A DILOR XY spectrometer of the type HORIBA (Jobin Yvon Inc.) was used to measure the Raman spectra of 49 Roman black glass artefacts. The spectrometer was equipped with an Olympus BH2 microscope (50x magnification, 8 mm focal length), a single monochromator, a notch filter, and a liquid nitrogen-cooled charge-1 coupled device detector CCD3000 with a resolution of ~2 cm . The excitation source providing a radiation of 514 nm was a mixed-gas laser of the type Coherent Innova 70C Argon/Krypton with an output power of 50mW. All samples were measured at room temperature (Tr) twice with acquisition times lasting 60 and 300 seconds. Caution regarding the possible influence of weathering was here not applicable since all black glass samples used were previously embedded in an acrylic resin, ground and diamond polished for SEM-EDX and LA-ICP-MS analysis. The measured Raman spectra were normalized by treating with a 5-segment linear baseline subtraction and a minor smoothing using the LabSpec (Dilor) software (Baert et al. 2011). All measurements were performed by Kitty Baert at the Department of Electrochemical and Surface Engineering, Materials and Chemistry (ESEMC) of the Vrije Universiteit Brussel (VUB) and were integrated in a paper on the use of Raman spectroscopy as non-destructive method for glass analysis (Baert et al. 2011). The SiO4 tetrahedral is generally used as vibration unit to describe the Raman spectra (Robinet et al. 2006; Colomban et al. 2006; Colomban 2008; Mulevanov et al. 2009). When the SiO4 tetrahedral is bonded to 4 oxygen atoms, the pure silica -1 -1 (Si) is fully polymerised (Q4= ~1150-1250 cm ), whereas the isolated tetrahedron (Q0= ~800-850 cm ) has no covalent bond. In between lie the SiO4 tetrahedral structural units bonded to 1, 2 and 3 oxygen atoms, -1 -1 -1 correspondingly Q1= ~900-950 cm ; Q2= 950-1000 cm ; Q3= 1050-1100 cm (Figure 176). The degree of -1 polymerisation of the glass can be calculated from the ratio of the bending modes (~500 cm ) and the stretching -1 modes (~1000 cm ), using the polymerisation index which is defined as Ip = A500/A1000 (Robinet et al. 2006; Colomban et al. 2006; Colomban 2008). The Raman spectra thus supply information about the covalently bonded structural units (Qn) of the glass samples, where n indicates the number of bridging oxygens. (Qn= stretching components, and Qn‘= bending components) (Baert et al. 2011). 277 NOCEM stands for ‗Non-destructive Optical Analysis of Cultural Heritage Materials‘. 351 Q3 Intensity (a.u.) 1200 Q2 Q1 Q4 Qn' 600 Q0 0 400 600 800 1000 1200 Wavenumber (cm-1) Figure 176: Example of a Raman spectrum of a sodium-silicate glass (taken from Baert et al. 2011, fig. 1) We did not verify the polymerisation index (Ip) of the different types of black glass, despite the fact that recent research demonstrated the calculation of the polymerisation index (Ip) from the area ratio of Si-O bending (A500) and stretching (A1000) envelops indirectly indicate the processing temperatureIp of 0,3 = ≤ 600°C; Ip of 1 = 278 1000 °C; Ip of 7 = 1400 °C (Colomban et al. 2006; Prinsloo, Colomban 2008). Inherently linked with the processing temperatures the Raman index of polymerisation (Ip) can be used to verify the flux ratiothe higher the index the lower the flux contentand perhaps also the type of flux added to the glass metal batch or to some additional elements such as PbO.  Results The correlations between these different spectral components supply information on the chemical composition of the analysed glass samples. All 49 investigated black glass artefacts are in reality deeply coloured, which only becomes visible when a strong light source is used. Previous research has demonstrated that the Roman black glass artefacts can be divided into three sub-groups based on their ironmanganese concentration (Van der Linden et al. 2009). A first group (Group A) represents the artefacts in deep green glass with high iron and low manganese content, respectively 613% and <1%. The second group (Group B) comprises the artefacts in deep purple glass with low iron and high manganese content, respectively 0.5%-3% and 2%-3%. The third group (Group C) consists of the artefacts in 279 deep brown glass with a low to moderate iron and manganese content, respectively 1%-3% and <1% [see Appendix 3]. The obtained Raman spectra show that the iron/manganese ratio in glass artefacts can be evaluated. Clear Raman spectra can be observed from glass with an elevated concentration of iron (Figure 177), but the presence of a high manganese concentration causes a strong fluorescence, resulting in a scattered spectrum by measuring a high background signal. From both groups with moderate manganese concentration, useful Raman spectra can be measured (Figure 178). However, to analyse the spectra, a background correction was performed on the signals from the samples of Group A and Group C in order to neutralise the presence of the minimal background signal. This appeared useless for the high manganese artefacts in deep purple glass (Group B), as unreliable results ensued from it. 278 279 The polymerisation index of black glass beads from India is calculated at 1,7 and that of obsidian at 3,5 (Colomban et al. 2006). Although Mn and Fe can have different oxidation states, they have been fixed at 2+ in this study. 352 In a preliminary test, the Raman spectra of four black glass samples from Rumst (BE) were recorded [see Appendix 3]. The (not normalised) spectra of these four samples with rather equal chemical composition are apparently quite similar (Baert internal report 18.03.2008) (Figure 177). Figure 177: The not normalised Raman spectra from 4 black glass samples with high iron content intensity (a.u.) 8000 GroupA GroupB GroupC 6000 4000 2000 0 200 400 600 800 1000 1200 1400 wavenumber (cm-1) Figure 178: Raman spectra of Grobbendonk (Group A = high Fe–low Mn); Olbia_3 (Group B = low Fe–high Mn); Avenches_4 (Group C = low Fe–low Mn) (taken from Baert et al. 2011, fig.4) [a.u. = arbitrary units] -1 It is striking that in addition to the classic broad stretching band around ~1000 cm and weak bending band at -1 ~500 cm , the Raman spectra of all black glass samples within Group A and Group C show a pronounced peak -1 at 415 cm . Based on the fact that ferric-sulphide chromophore is resonance-enhanced (Prinsloo, Colomban 3+ 22008) Kitty Baert considers the peak to be linked with the presence of ferric-sulphide Fe -S -chromophore in the black glass artefacts (Baert et al. 2009). The fact that the sulphur was not detected with SEM-EDX is evidence for 353 280 a very low content and consequently makes the Raman spectra very sensitive for sulphur. Instead of sulphur, -1 iron can be responsible for this pronounced peak at 415 cm . The different peaks related to iron can possibly be 2+ 3+ reduced to the presence of Fe and Fe oxides, respectively as FeO and Fe2O3. At least the broad stretching -1 band at ~1000 cm visible in the Raman spectra is due to the iron concentration in the black glass (Olivier et al. 2005). A next step was normalising the recorded Raman spectra of the selected black-appearing glass samples to come to a semi-quantitative evaluation (Figure 179). The Raman spectra were standardised starting from the intensity at Q3. The broad stretching bands all fall within the Q2 and Q3 order of the SiO4-tetrahedral stretching signal, -1 respectively around 980 and 1080 cm and showing two distinct peaks. The higher the iron content of the samples, the more the stretching band moves to Q2, meaning a decrease of the glass polymerisation. There is thus a clear link between the iron content in glass and the Q2/Q3-ratio. On the other hand, the SiO bending -1 bands at ~500 cm where the polymerisation index becomes very low (Ip = <0,10) show up as very weak -1 compared to the stretching signal. These peaks at 415 cm are considered resonance-enhanced in the bluegreen part of the spectrum and related to the presence of Fe-S chromophore and (Prinsloo, Colomban 2008). Fe2O3 11% 1.0 10% intensity (norm.) 8% 0.8 6% 2% 0.6 0.4 0.2 0.0 400 600 800 1000 1200 wavenumber (cm-1) Figure 179: Raman spectra of Roman black glass artefacts (Appendix 3) with a various concentrations of iron oxide (taken from Baert et al. 2011, fig. 5) [norm. = normalised] In brief the investigation with Raman spectroscopy on 49 Roman black glass artefacts has demonstrated that this non-destructive technique is a useful alternative for the study of archaeological glass artefacts. The amount of iron oxides added as colouring agents in the glass melt to produce black- appearing glass is visible in the Raman spectrum because iron oxide systematically influences the silicate glass structure. Early Roman black glass features a low iron content and the later Roman black glass, after 150 AD, a high iron content; the obtained Raman spectra are thus valuable for dating the artefact, even though the dating remains merely approximate. The fact that the iron content is clearly observable on the Raman spectra means that this nondestructive technique can be helpful in discriminating the provenance of ‗naturally‘ coloured and decolourised glass (such as the late Roman-early Byzantine HIMT glass) from that of other originse.g., Levantine I, II and Egyptian I, II (Freestone et al. 2002; Picon, Vichy 2003; Foy et al. 2003). 280 It is unknown to the author in what degree sulphur evaporated from the batch by bonding with oxygen. 354 Raman spectroscopy can therefore provide fundamental information about materials and techniques that were applied in a certain period or region. Applicable for dating glass artefacts, this knowledge should be checked by analysing more objects of one specific chronological group in order to test the limits of Raman spectroscopy on archaeological and historical glass. A geographical dimension can be added to the future analysis by checking similar artefacts from various regions. 355 356 Chapter 9 OPTICAL ANALYSIS 9.1 Introduction 281 The initial impetus for executing optical analysis on black-appearing glass is the lack of a consensus on when 282 to classify strongly-coloured glass as black-appearing. This has consequences in the quantification of the studied material and thus for its interpretation. The very specific and highly debatable issue of designating and describing the hue of archaeological glass artefacts thus necessitated an objective optical verification what can be realized by means of absorption spectroscopy. This characterization technique is especially appropriate because the blackness of glass artefacts depends on particular properties of the material, e.g., the colouring and opacifying agents and the thickness of the object (Bamford 1977; Bayley 1999, 90; 92). Optical analysis can act as an unprejudiced measuring method eliminating the biased colour vision of the human observer who is influenced by his visual sensitivity to light signals but also by the luminosity of the different light sources and the interaction of objects with light (Hunter, Harold 1987). To obtain high-quality results standardized conditions by means of artificial daylight illuminants are adopted for the visual evaluation of colours. On the other hand, the previous chapter has demonstrated the benefit of measuring the chemical composition of black glass artefacts for dating, tracing the type of applied production technique and identifying certain distribution patterns. Hence, the aim of this chapter is to verify whether additional information on the production and consumption of Roman black-appearing glass artefacts is obtainable by using physical methods to confirm or to refine the conclusions from chemical analysis described in the previous chapter. In addition, we have verified to what extent UV-Vis-NIR spectroscopy provides answers to some specific archaeological questions. An additional outcome of the UV-Vis-NIR spectroscopy research is that it is a simple technique which does not require an excessive handling of the artefact, is non-destructive, and is not expensive (Meulebroeck et al. 2011). Therefore this technique holds considerable advantages in the study of archaeological glass artefacts compared to the expensive equipment necessary for the standard chemical analysis. The techniques we used to determine the chemical composition of Roman black glass artefacts were SEM-EDX and LA-ICP-MS [see Chapter 8]. These 283 techniques however include several clear disadvantages , making that these chemical analysis methods are always executed on a limited number of well-selected artefacts.     the inevitability of complex techniques and expensive instrumentation, meaning that measurements can only be performed ex-situ in a major institute requiring the skills of trained people the strict export restrictions of archaeological artefacts the limited selection of samples offered because of destructive sampling methods the time-consuming analysis per sample Optical analysis by means of absorption spectroscopy, specifically UV-Vis-NIR spectroscopy, has already been 284 applied to archaeological glass research in the past. In section 9.2 (Absorption Spectroscopy on Roman Glass Artefacts, current publications on the absorption spectroscopy of archaeological glass correlating the spectral data of Roman glass colours with the chemical composition are discussed. The use of optical spectroscopy on st nd archaeological glass artefacts has already been applied in studies on 1 and 2 century Roman glass from Coppergate, York (Bingham, Jackson 2008), on Roman glass from Colchester (Green, Hart 1987), on Late Roman Palestine glass from Galilee (Schreurs, Brill 1984) and Jalame (Brill 1988), on Parthian and Sasanian th dated glass artefacts coming from excavation sites of modern Iraq (Mirti et al. 2008), on 7 century glass th th fragments of the Crypta Balbi in Rome (Mirti, Lepora 2000), on 15 and 16 century German forest glass (Sellner et al. 1979), and on post-Medieval glass from Surrey (Green, Hart 1987). Despite the many advantages, little interest has been shown in exploiting UV-Vis-NIR spectroscopy fully within the archaeometric approach. 281 This study was a pilot project of an ongoing research project at the Vrije Universiteit Brussel – the HOA15-project NOCEM. Its goal was to test the possibility of using UV-VIS-NIR spectroscopy (spectroscopic analysis within the range of UltraViolet – VISible – Near InfraRed wavelengths) to characterize archaeological and historical glass artefacts and verify whether similar information can be generated from chemical analysis. Consequently, this chapter is the result of a close collaboration with Dr. Wendy Meulebroeck from the Department of Applied Physics (TONA). 282 In Chapter 1 [Defining ‗Black‘ and the Perception of Colour] we explained that black is more than a limited perception of the naked eye for recognizing colours. Additionally, each culture describes and nominates colours differently, and attributes specific meanings and connotations to particular hues. 283 Even though there is also the existence of cheaper, portable and more user-friendly instrumentation such as for instance the portable Raman (Colomban 2008; Colomban, Tournié 2009; Ricciardi 2009, 129-149) and portable XRF providing quantitative measurements when used under vacuum (Nakai et al. 2005; Nakai et al. 2009; Tantrakarn et al. 2009). 284 In recent years, simple portable optical spectrum analysers became commercially available, opening perspectives towards in-situ measurements. 357 In section 9.3 (Methodology), we discuss the different ways of using absorption spectroscopy (through light in transmission with or without the use of an integrated sphere, and through light in reflection). Both physical and 285 chemical properties affecting the transmission of light were calculated with the Lambert-Beer formula. We attempt to determine whether optical analysis on Roman black glass artefacts can answer the following archaeological questions and thus supply more information for a better interpretation of archaeological glass: 1. 2. 3. 4. 5. Is a distinction between the black and non-black appearance as opposed to the strongly-coloured glass measurable? Does the colorimetry of glass correspond to its chemical composition? Is it possible to calculate the variety of shades per specific colour type resulting in a black-appearing glass matrix? Can specific glass colours be matched to specific periods or typological groups? Can UV-Vis-NIR spectroscopy provide answers to technological aspects of black glass production? In section 9.4 (Analysis Results), these questions are examined. The first two questions concern a purely optical 286 analysis and have not been the subject of prior research. An objective measuring system distinguishing black glass from strongly-coloured glass validates the attribution of material as black, seeing the arbitrary perceptiveness of the human eye. We applied UV-Vis-NIR spectroscopy to establish without bias the different hues and colour intensities of black-appearing glass (question 1). Through UV-Vis-NIR spectroscopy, the spectrum of an analysed glass artefact is illustrated with characteristic absorption bands and transmission peaks. The respective absorption bands hold very valuable indications about possible spectral regions of interest. Prior research has been done on the identification of specific absorption bands related to the applied colouring agents (Schreurs, Brill 1984; Green, Hart 1987; Brill 1988; Mirti, Lepora 2000; Bingham, Jackson 2008). Various glass compositions were established on the basis of different oxides to obtain black glass. This resulted in the characterization of the absorption bands responsible for colouring the glass batch black. Based on previous research, which established the calculation of the colour coordinates of the samples and drew conclusions on applied colouring agents and production facilities (Brill 1988; Mirti, Lepora 2000; Mirti et al. 2008), UV-Vis-NIR spectroscopy is used in the study of strongly-coloured Roman glass that has a black appearance in order to establish the correlation between the transmission spectra of the artefact, and the added colouring agent(s) in the glass batch that give this black appearance (question 2). The identification of the various metal oxides used for making the glass matrix appear black will be compared with the available information on the colouring capacities of the chemical elements (Brill 1988) (question 3). Can optical analysis provide answers to archaeological questions on the geographically and chronologically related technological differences of the black glass production (questions 4 and 5)? The UV-Vis-NIR spectroscopy results on black glass artefacts were compared to previously obtained knowledge on the reduction-oxidation furnace conditions of glass production in general (Schreurs, Brill 1984; Green, Hart 1987; Brill 1988; Bingham, Jackson 2008). Can the obtained optical spectra reveal information about the applied production methods for black-appearing glass, e.g., furnace conditions (question 5)? In section 9.5 (Use of Photonics), we make a first attempt to fully apply the analysis results of section 9.4 to the archaeological questions asked concerning black glass. 9.2 Absorption spectroscopy on Roman glass artefacts To keep a logical structure, we thought it helpful to describe briefly and in rough outline the discussion on the chemical composition of Roman glass. The basic glass hue during the Roman imperial period is the so-called ‗naturally coloured‘ glass. The hue of this large group encompasses a wide variety of tinges including pale blue, pale bluish-green (also called ‗aqua‘), pale 285 The Lambert-Beer formula is I = I0.e.(-a.c.x), resulting from the Lambert law on thickness and the Beer law on concentration (Ingle, Crouch 1988). 286 The three other issues can only be discussed in relation to the results from the chemical analysis. 358 olive-green, and yellowish-green due to impurities in the sand (Brill 1988, 269). The most typical ‗naturally‘ coloured glass has a bluish-green effect due to the natural presence of FeO/Fe 2O3, commonly ranging between 0,3-1,5% (Schreurs, Brill 1984). Hence, to obtain a decolourized or pale-coloured (nearly colourless) glass the Romans neutralized the blue-green colouring effect of the iron oxide impurities by adding antimony (Sb) and/or manganese (Mn) to the batch. The Roman glassworker also deliberately added iron, manganese, copper and cobalt to obtain a wide range of colours, but furnace conditions (oxidising or reducing) must have played a significant role as well. Because the Roman strongly-coloured glass artefacts appearing black were variously manufactured from intentionally coloured glass, it is most relevant to focus on these transition metal oxides added to the batch to colour the glass matrix. Different methods were used to obtain a black-appearing glass:     by adding large concentrations of the standard colouring agents (e.g., iron, manganese) (Newton, Davison 1989) by adding non-soluble particles to the batch (the opacity is maximum when the size of the particles is in the order of magnitude of the visible wavelengths) (Scholze 1991) by adding opacifying agents to the batch (these special types of elements are characterised by their very high absorption for visible wavelengths. The most important opacifying agent used in Roman times was antimony by means of calcium antimonate and lead antimonate) (Mirti, Lepora 2000; Lahlil et al. 2009; 2010) by using different furnace conditions (Schreurs, Brill 1984; Green, Hart 1987; Brill 1988; Bingham, Jackson 2008) It is common knowledge that the basic principle of producing glass consists of the melting of three main constituents: a network former and two types of network modifiers [see Chapter 8]. In ancient glass, sand generally produced the network former, silica (SiO2). The three-dimensional tetrahedral network has a silicon atom at the centre and an oxygen atom at each corner, two of which are shared by an oxygen atom of the adjacent tetrahedral. This strongly bonded structure is responsible for the strength and durability of glass products, but it also means that the melting temperature of the pure network former (c.1410°C, based on the table of Mendeleyev) is too high to have been reached in ancient furnaces. A network modifier (a so-called flux) was added in order to loosen the strong network bonding and to make the network more open. Very suitable for this operation are the alkali elements such as soda (Na) and potash (K), due to the presence of free electrons. The 287 melting temperature of silica can be lowered 200°C to 900°C by fluxing the batch depending on the ratio added. The disadvantage of this process is that the glass is made less durable and will lose its resistance to water, resulting in a very porous material. To avoid this, the Romans used stabilizing agents, e.g., calcium and 288 magnesium, probably by adding lime (calcium oxide) or magnesia (magnesium oxide) to the melt. Pliny the Elder (Nat. Hist. XXXVI, 192) not only mentioned the addition of soda to the batch, but also magnetite, glittering stones, shells and even sand from quarries. It is thus very likely that the shells or calcium-rich and/or magnesia289 rich sands were used, consolidating to some extent the strength of the network. Depending on the quality of the sand that was used, various oxides got into the batch as impurities. The most common trace elements present in ancient glass are, for instance, Al 2O3, Fe2O3, TiO2, BaO, SrO and B2O3 (Brill 1988, 262-263). Alumina generally appeared in a concentration of between 2-3%, while iron impurity levels in the sand fluctuate between 0,3-1,5% (Schreurs, Brill 1984). Besides the natural presence of iron (as FeO) in the sand as an impurity, much higher iron concentrations can also be caused by the addition of the ferrous oxide Fe 2O3 haematite, a red iron oxide or by means of the ferrous ferric oxide Fe3O4 (= Fe2O3*FeO) magnetite, a black iron oxide (Heck, Hoffmann 2000). In Roman times, ‗naturally coloured‘ glass, as well as decolorized glass, was used in the production of black glass [see Chapter 8]. To obtain a wide range of colours, the Romans glassworkers added small portions of oxides of transition materials such as cobalt (Co), copper (Cu), iron (Fe), nickel (Ni) or manganese (Mn) (Schreurs, Brill 1984; Brill 1988). However, the colouring agents used in Roman times to obtain a black-appearing glass was mainly iron oxide (from brown to green), or otherwise manganese (purple) sometimes in combination with cobalt and or copper (violet). Characterizing the absorption band of lead oxide is important in distinguishing the Roman 287 Pliny the Elder mentions the proportions of sand and natron at 3:1 in weight (Nat. Hist. XXXVI 194). Seeing that natron is much lighter than sand results in a rough proportion of 1:1 in volume. This means that the melting temperature of the glass matrix could be lowered to about 1150°C producing an already good workable viscose glass at about 1000°C. Robert Brill (Brill 1988, 268) mentions that the glass from the late Roman secondary workshop at Jalame (IL) has a calculated ratio of Belus River sand and Wadi al-Natrun natron equivalent to 4,92:2 (~ 5:2 ratio) or when calcium is also taken into account for this silica-soda-lime glass a 5:2:1 ratio. This ratio results in a glass matrix that melts at a temperature much below 1150°C so that in (late) Roman times workshops could economize on fuel. 288 A number of 1st century AD black glass samples with low iron content show an increased magnesia content. Related to a higher potash and phosphor content, this points to use of a vegetable soda rather than a mineral one. 289 See Wedepohl, Baumann 2000; Freestone et al. 2003 and Freestone 2005; 2006. 359 material from that of the very late Roman and early medieval glass emerging in Period IV. Equally important for this period is the introduction of HIMT glass, a LMG glass with a high level of sand impurities such as titanium (Ti) and Manganese (Mn). The elevated Ti-content generally exceeds 0,3% and that of Mn usually goes above 1,0% , whereas the Roman glass of the prior periods using much purer sand has a titanium content under 0,25% but for a large part ranging below 0,1%, and a manganese content less than 0,5% [see Appendix 3], even though a range between 0,02-0,15% is considered the impurity level of pure Levantine sand (Brill 1988, 259). Particular types of transition metals were added to the batch in various small concentrations of about 1%, depending on the glass colour defined in advance that had to be produced (Newton, Davison 1989). Since any impurity influences the colouration properties of the raw material, it is clear that absorption spectroscopy can help determine the composition of the glass. Glass was decolorized in Roman times by using manganese and/or antimonite oxide (Braun 1983, 146-147; Vichy et al. 2007; Lauwers 2008; Foster, Jackson 2010). The most common hue of Roman glass is the ‗naturally coloured‘i.e., blue-green, pale bluish-green (described by Brill as ‗aqua‘), and pale bluewhereas the intentionally coloured glasses range between deep blue, green, red-purple, yellow-green, olive-green and yellowbrown (Schreurs, Brill 1984; Green, Hart 1987). The yellow-green colour of the glass, for example, is due to the appearance of iron oxide (FeO), the most important colouring impurity in sand. The colour of an artefact depends, however, not only on its chemical composition but also on the applied technology. The variation in colour is the result of the addition of small fractionsnot more than a few percentof transition metal ions in particular oxidation states. Most of the colouring elements applied in Roman timesiron (Fe), manganese (Mn), copper (Cu) and cobalt (Co)could thus result in a broad range of colours not only by using colouring oxides separately or by combining various colouring elements to the batch, but also by playing on the furnace atmosphere (oxidizing or reducing conditions). Iron can imply a continuous range of colours, starting from pale bluish-green going towards blue-green, green, yellow-green and olive-green (Schreurs, Brill 1984; Newton, Davison 1989; Scholze 1991). These colours are a 2+ 3+ result of the relative proportions of ferrous iron (Fe ) and ferric iron (Fe ) ions. In general, it is stated that the 2+ 3+ saturation of the colour depends on the total iron concentration, whereas the hue is fixed by the ratio Fe /Fe 2+ 3+ (Mirti et al. 2008). This Fe /Fe ratio is affected by the furnace conditions during the melting and production process, i.e., the oxidizing or reducing conditions. The result is that two glasses with the same iron content could 2+ have a different colour due to the difference in Fe concentration, which indicates a production under different oxidation states (Brill 1988, 9). 2+ 3+ The conversion from Fe into Fe could be obtained in different ways. A first possibility is to work under oxidizing conditions. This could be reached by the addition of extra oxidizing agents (see below) or by changing the furnace conditions with short, bright flames. This is possible if an air flow is sent through the melt for a certain time period, transforming ferrous iron into ferric iron (Newton, Davison 1989; Meulebroeck et al. 2010). The reverse reaction can also take place if the glassmaker works under more reducing (smoky) conditions (Newton, Davison 1989). Production facilities which were able to control this reaction could make the glass more blue or more green. The 2+ reduced form of iron (Fe ) will impose a blue colour, whereas under more oxidizing conditions iron will oxidize to 3+ 3+ 2+ Fe and will result in a green hue caused by the yellow colour of Fe together with the blue colour of Fe . The colour intensity after reduction is stronger than after oxidation (Scholze 1991). An often applied technique to make colourless glass is therefore to work under oxidizing conditions (Scholze 1991). Various references cite the possible absorption bands of the two types of iron ions (Schreurs, Brill 1984; Green, 2+ Hart 1987; Brill 1988; Scholze 1991). The most pronounced is the absorption band of Fe between 900-1200 nm with a shoulder around 450 nm (Brill 1988). This tail in the visible part of the electromagnetic spectrum is apparently responsible for the pale blue hue. The weaker the absorption band around 1100 nm, the more oxidized the furnace conditions (Meulebroeck et al. 2010). The presence of other elements in the batch could influence the final colour considerably. One important element is sulphate, which can be present as a contaminant through the added soda. Natron, mainly used by the Roman glassworkers as flux, is an evaporite that occurs in combination with sodium sulphate (Schreurs, Brill 1984). Under reducing furnace conditions and in the presence of sulphate ions, the following reactions take place (Schreurs, Brill 1984; Van der Linden et al. 2009): 360 Fe SO4 (2-) + 8H (+) (3+) (-) (2+) + e  Fe (-) (2-) + 8e  S (1) + 4H2O (2) 2+ A large fraction of the iron is reduced to Fe , while sulphate ions can be reduced to sulphide ions. The sulphide 3+ (3+) (2-) can combine with the remaining Fe to form the ferro-sulphide Fe -S . This chromophore modifies the blue 2+ colour of the glass to green, olive-green or amber-brown, depending on its relative concentration to that of Fe with the olive-green and the amber colours obtained under the maximum reduction state (Schreurs, Brill 1984). This is the result of the characteristic intense absorption band around 405 nm of the ferro-sulphide (giving by itself 2+ a red tint) in combination with the absorption bands of Fe , although other sources speak about 420 nm (Schibille 3+ et al. 2007). Other possible absorption bands due to iron are all due to the presence of Fe , such as a series of weak, sharp bands mainly in the region 375-450 nmi.e., 380 nm, 415 nm, and 440 nm (Schibille et al. 2007) and a very strong band at 258 nm. These bands give a very pale lemon yellow colour but are normally both 2+ present together with the bands of Fe , in which case they impart a greenish-blue tint, or together with the absorption band of the ferro-sulphide when their effect is negligible (Green, Hart 1987). Another important colouring element in Roman times is manganese. Ancient glass contains, in most cases, small 2+ concentrations of manganese as a constituent of the alkali flux. This is under the form of the colourless ion Mn . When the concentration exceeds 0,4%, it should be considered as intentionally added (Brill 1988). Known absorption peaks are those from MnF6 at 461 nm, and Mn(H2O6) at 476 nm (Green, Hart 1987). Under more 3+ oxidizing conditions, the red colouring ion Mn with an absorption maximum between 470 and 520 nm (Green, Hart 1987; Baxter 1995; Schibille et al. 2007) is formed via the reaction: Mn 3+ +e (-)  Mn 2+ (3) In almost all ancient glasses, both iron and manganese were simultaneously present: iron as an impurity in the sand and manganese as a contaminant in the flux. The combined effects of reactions (1) and (3) show that there is an equilibrium between the two states of oxidation of the iron and the manganese (Newton, Davison 1989): Fe 2+ + Mn 3+  (yellow+purple) 2+ Fe 3+ + Mn 2+ (4) (blue+colourless) 2+ 3+ 2+ Because the Fe and the Mn ions are the more stable states, the equilibrium tends to move to Fe + Mn (4 2+ right). When the conditions during melting of the glass are fully reducing, the equilibrium has been forced to Fe 3+ 2+ + Mn (4 left), and the iron contributes to a bright blue colour due to the Fe ions while the manganese is in the colourless form, resulting in (pale) blue glass. When the conditions are fully oxidizing, the equilibrium has been moved to the right, and the iron contributes a brownish yellow colour and the manganese a purple colour, resulting in a brownish violet glass (Newton, Davison 1989). Indications about the relation between the ratio of the manganese and the iron content, and the resulting colour hue and its implications within the archaeological discourse are thus far poorly discussed. 2+ 3+ From reaction (4) it is clear that Fe and Mn could not be present in the same glass matrix. As a consequence, 3+ the absorption bands of Mn between 470 and 520 nm do not appear together with the absorption band around 2+ 1100 nm of Fe (Green, Hart 1987). When the conditions of reaction (4) are intermediate, a variety of colours is obtainable such as green, yellow and pink (Newton, Davison 1989). Also a colourless glass is obtainable through the manipulation of furnace conditions. This physical process of decolourizing occurs when the purple from the st manganese just balances the yellow from the iron (Scholze 1991). From the 1 century AD onwards, the Roman glassworkers performed chemical decolourizing as an alternative method (Newton, Davison 1989) by adding the oxidizing agent pyrolusite (MnO2) (Schreurs, Brill 1984; Green, Hart 1987; Newton, Davison 1989; Scholze 1991): Mn 4+ + 2Fe 2+  Mn 2+ + 2Fe 3+ (5) 2+ In this reaction, the decolourizing takes place in three different ways: transformation of the strong colouring Fe 3+ 2+ into the less colouring Fe , eliminating the blue colour, and the formation of the colourless Mn and the oxidation 2of the S of the ferro-sulphide chromophore to SO2, eliminating the brown-yellow colour (Green, Hart 1987). It is clear that depending on the furnace conditions, manganese (Mn) could give different colours, with violet, green and colourless as the three most important ones (Brill 1988). All the described reactions required a precise control. With respect to reaction (4), if too much manganese was added, it would act as a decolourizer for the glass, which would otherwise be greenish in colour (Newton, Davison 1989). In reaction (5), on the other hand, a 3+ too-high concentration of MnO2 could result in oxidation to the red colouring ion Mn via reaction (3) (Schreurs, Brill 1984; Green, Hart 1987), or retention of the brown colour of the ferro-sulphide (Green, Hart 1987). 361 It is significant that the addition of manganese is linked with the presence of small portions of iron, copper, vanadium, nickel and barium (Brill 1988). Further investigation should lead to the recognition of the peaks in the absorption spectra connected with vanadium, nickel and barium. Within the previous chapter discussing chemical analysis, we provided enough evidence of the deliberate addition of iron and manganese oxide to obtain black-appearing glass. Two other colouring agentscopper and cobaltare much less frequently used by the Romans to produce black-appearing glass, or are only used in combination with iron and manganese oxide. Although the deep blue coloured glass artefacts integrated within our study are very few, we also discuss the properties of the glass artefacts coloured through copper and cobalt, as they can turn out black as well. Copper results in a light blue colour, and cobalt in a deep blue, violet, or indigo hue (Green, Hart 1987; Brill 1988; Newton, Davison 1989; Scholze 1991; Shortland et al. 2006). Both metal oxides are thus responsible for a typical colour. However, copper can also produce a similar ultramarine blue glass, as with cobalt under certain furnace conditions, if the concentration in the glass matrix is high enough. The concentration of copper generally ranges between 2-5%, while higher percentages would give a black colour (Newton, Davison 1989). The absorption band 2+ around 835 nm is connected with Cu (Mirti 1995). Different ores of copper and cobalt could be used to bring this element into the batch, mainly under the form of oxides and carbonates (Newton, Davison 1989). These ores contain other elements such as alumina, manganese, nickel, zinc and magnesium (Shortland et al. 2006). In many samples, tin and lead are found together with copperwith a Cu:Sn:Pb ratio of 90:3:7pointing to the addition of bronze fillings (Brill 1988; Newton, Davison 1989; Shortland et al. 2006). The latter elements are characteristic of a group of late Roman glass within Period IV, such as the late material from Trier-Kesselstatt (DE) [see Chapter 8]. 2+ Under oxidizing conditions, CuO (Cu ) colours the glass blue or green, depending on the other constituents of the batch. In combination with PbO, the copper metal oxide results in a green colour, while the glass would turn . . turquoise blue when combined with natron (Na2CO3.10H2O) or trona (Na2CO3 NaHCO3 2H2O) (Newton, Davison 1989). Under reducing conditions, CuO results in a red colour (Newton, Davison 1989; Heck, Hoffmann 2000): 2FeO + 2CuO  Cu2O + Fe2O3 (6) Cobalt is a much stronger colouring element than copper is. Only a very small concentration is needed within ppm 2+ range, with a concentration up to 50 ppm that is spectroscopically measurable. Cobalt, present as Co seeing as 3+ Co is unstable (Green, Hart 1987), has three successive absorption bands around 600 nm (Scholze 1991) and several less strong absorption bands in the near-infrared region. Depending on the added form of cobalt and the presence of other elements, these bands could shift considerably (Green, Hart 1987). In Roman samples, the first absorption band is located at 535 nm, whereas in medieval glass samples from Wealden, this band is shifted towards 520 nm (Green, Hart 1987). + + 2+ This difference is the result of a difference in Na , K and Ca concentrations (Green, Hart 1987). The other absorption bands for Roman glass containing cobalt are situated at 595 nm and 650 nm (Mirti 1995). In many cases, these elements were not used as the only colouring element. A mix of copper, cobalt, and manganese was usual to obtain blue glass (Newton, Davison 1989), whereas blue and green colours can be 2+ 2+ 3+ 2produced in a combination of Co with Fe or Fe /S (Green, Hart 1987). Table 156 summarizes the results found in literature. This first evaluation gives an indication of the possible colorants used, the relation with the furnace conditions, and the colour resulting from it. We also mention the wavelength at which the transmission is maximal (Tmax). This value depends on the obtained colour. 362 Table 156: Table summarizing the possible colorants used on the analysed Roman black glass artefacts, and the relationship between the applied furnace conditions and the resulting colour (taken from Brill 1988, 272-273, Tables 9-10) colorant content furnace colour Tmax comments conditions (nm) Fe2+ (FeO) Fe3+ (Fe2O3) 2+ high concentration reducing aqua (bluish) 490 low concentration high concentration reducing oxidising aqua (greenish) 520 480 moderate concentration oxidising 3+ Fe + Fe (Fe3+ predominates) reducing oxidising slightly reducing ferro-sulphide complex Mn3+ 480 + 510 green (blue+green+yellow) 555 olive (green+yellow+red) amber (yellow+red) 585 S: 0,2 – 1,4% Fe: 0,3% around 3% strongly reducing strongly reducing oxidising purple (violet + red) 400 2 - 13% oxidising (deep) blue and green ferro-sulphide complex + Fe2+ Cu2+ Co2+ 0,10% a pale blue-green colour is the common natural colour of raw glass deep blue 670 700 shift the natural aqua colour can be intensified with the addition of copper. By adding lead, the green colour produced by copper could be darkened intense colouring 9.3 Methodology Two methods are applicable to measure colours   290 : Light measurement in transmission Light measurement in reflection The transmission spectra were measured with a high-performance optical spectrum analyser SA320/Instrument Systems by using white light, i.e., a beam equalling a spectral resolution of 1,5 nm and emitting a white light incorporating all hues in the visible spectrum between 380 and 780 nm. The operating source of light is a stabilized halogen source with a constant emission pattern after a 30-minute warm-up. The white light sent in transmission through the glass samples has been focused through a microscope objective which focuses the beam to a spot size of approximately 1,5 mm diameter at a distance of about 2 mm from the artefact. An optical fibre bundle positioned very close (1 mm distance) to the other side of the artefact collects the transmitted light within an angle of 43° and guides it to the entrance slit of the high performance optical spectrum analyser, SA320/Instrument Systems. The optical fibre bundle is put in connection with a computer calculating and plotting 291 the spectrum of the transmitted light between 300 and 1600 nm. The source and the receiving fibre were placed in one line with the sample in between (Figure 180). 290 We only carried out light measurement in transmission, although we tested the light measurement in reflection on a small set of samples. Due to the particular properties of black glass, the reflection of light in a straight on-line set-up produced no information because no absorption could be measured, as the white light source became mirrored, reflecting all light back. More complex constructions possibly would give a solution, but no alternative was tested within the reach of this study. 291 Quantification up to the level of a nanometre (nm) of the caught light in transmission is done by using a calculation programme written in 1997 in IS-SPECWIN-Ascii for Microsoft Windows 95, and makes conversion into a graph with a precision of up to 0,12 nm possible. 363 T transparency white light blue light sample optical fibre bundle  wavelength Figure 180: schematic set-up of the optical installation for UV-Vis-NIR spectroscopy To admit the possibility of a comparative study of the measured values, it was essential that the distance between the source of light and the artefact on the one hand, and that between the artefact and the receiving optical fibre on the other hand, remained invariably similar. Therefore, it was repeatedly necessary to move both the sending and the receiving part of the set-up, as the various artefacts measured had different external shapes. The colour of an object not only depends on the pigments inside the object, but also under which lighting conditions it is perceived, and at what angle it is viewed. The light source under which we wanted to represent the colour had to be defined in terms of the angle at which the artefact has been irradiated (Figure 181). eye white light white light reflection light source sample transmission light source Figure 181: schematic view of all possible influences on the colour of an artefact before reaching the naked eye. The fragments of carinated beakers (Form IIB.1) from the villa of Matagne-la-Petite (BE) and the military camp at Rumst (BE) can be held up as examples, showing unanimously in Figure 182 their black appearance when put on a white paper and using only reflective light. When these fragments are put on a milky white glass with an 18 watts neon light in transmission, some fragments exhibit the true hue of the glass (a: in red dotted ellipses; b1; c1), demonstrating a (bottle) green colour. Conversely, when the white neon light is replaced by a halogen lamp of 500 watts, the same fragments turned greenish to orange brown, frequently called olive-green (b2; c2). 364 Matagne Rumst a b2 b1 c2 c1 Figure 182: (a) A selection of black-appearing glass vessels with only reflective light; (b) the Rumst piece with neon (whitish) light (18 Watt) in transmission (b1) versus a strong halogen (yellowish) light (500 Watt) in transmission (b2); (c) a Matagne piece with neon (whitish) light (18 Watt) in transmission (c1) versus a strong halogen (yellowish) light (500 Watt) in transmission (c2) (photographs by the author, courtesy of VIOE) To be able to refer to the colour information mentioned in Table 156, a white light source with a constant intensity in the visible region was chosen. In most cases, a standardized source proposed by the Commission International de l‘Eclairage (CIE) is used, such as the ‗Illuminant A‘ source, which simulates natural daylight. If a comparison of the colour coordinates is done between this light source and a theoretical light source, the colour will appear warmer for the ‗Illuminant A‘ source. This is because for this source the red colours dominate. Accordingly, we decided to represent the colour under the illumination of a (theoretical) light source with a constant intensity in the spectral region of our interest. Starting from the measured transmission spectra, we calculated the x and y colour coordinates which are represented on the CIE1931 horseshoe graph (Figure 186, left). This colour system has been chosen because it represents all colours that the human eye can physically perceive. The point of interest was to come to a classification based on the applied colorants. The artefact was placed between the light source and the receiving optical fibre to enable the measurement of the spectral emittance E (W/m².nm) with the spectrum analyser. The transmission spectrum of each artefact TA(λi) has been calculated into the defined spectral region by dividing the spectral emittance of the used light source, ES(λi), with the spectral emittance of an artefact, EA(λi), for wavelengths λi between 300 and 1600 nm: TA (i )  ES (i ) .100% E A (i ) Initially, the main interest of this research was limited to the analysis of the spectral shape of the black-appearing glass artefacts; accordingly, the use of an integrating sphere was at that time trivial. This, however, resulted in the loss of a fraction of the transmitted light not being received by the optical fibre, consequently making absolute quantifications impossible (Figure 183, top). But when trying to determine the darkness/blackness of the glass artefact and the viscosity of the glass matrix, it has been necessary to measure the absolute spectra, implicating the use of an integrating sphere (Figure 183, bottom). 365 transmitted light white light optical fibre bundle It sample integrating sphere white light optical fibre bundle It T coloured glass sample  Figure 183: Schematic views of the spectrum measurement using light in transmission without (top) or with (bottom) the use of an integrating sphere When measuring the absolute spectra, the calculation of the full transmission (T) of a glass artefact is necessary. This implies the use of an integrated sphere catching all the light (IT), because without the optical fibre bundle will not collect all the transmitted light conveyed from the artefact (Figure 184). A calculation starts from the known value of the transmitting light source (It). Taking into account the interface reflections (R), the true transmitting light (Io) demonstrates a refraction angle affected by the pad length (approximately equivalent to the thickness) of the glass artefact (x), as well as by the concentration (c) and the absorption coefficient (a). The higher the optical density (OD), the lower the luminosity of an object or in other words the lower the transparency and brightness of the glass matrix and accordingly the darker the glass hue. From a certain point the optical density (OD) makes the glass appears black to the naked eye at a certain point of absorption (Weyl 1959; Bamford 1977). log10T = c.x.a.log10e – 2.log10.(1-R) log10T = OD x R It I0 T a, c I integrating sphere Figure 184: Schematic view of the measuring of transmission (T) of glass artefacts Different aspects were investigated after calculation of the analysed transmission spectra:     366 establishing the objectively measured limit between strongly coloured and black appearing glass testing the degree of correlation between the colorimetry of the glass artefact and its chemical composition by comparing the spectral shapes to the various subgroups defined from the chemical composition, particularly on the basis of the iron content vis-à-vis the manganese content (Figure 194) identifying the characteristic absorption bands related to the colorants used in the glass matrix drawing some conclusions about the applied production process, for it has been indicated that furnace conditions used during the production process can influence the colour of the glass mass The colour of the glass artefacts can be calculated on the basis of various parameters: x X X Y  Z y X   TA ( )S ( ) x( )d Y X Y  Z z Z with z  1  x  y X Y  Z Y   TA ( )S ( ) y( )d   the calculated transmission spectrum TA(λ) the spectrum of the source S(λ)  the colour matching functions x , y and z Z   TA ( )S ( ) z( )d (Figure 186, top right) The used values of the colour matching functions x , y and z for calculations were also taken from Wendy Meulebroeck‘s doctoral thesis (Meulebroeck 2004, 63). The calculations on the black glass artefacts were done within the spectral region of visible wavelengths between 380-780 nm (Figure 185, Table 157). Figure 185: Solar spectrum within the visible (VIS) [between 380-780 nm], from ultraviolet (UV) [below 380 nm] to nearly infrared (NIR) [above 780 nm], including the Fraunhofer lines featuring the absorption lines matching the wavelength of specific chemical elements (see http://en.wikipedia.org/wiki/Fraunhofer_lines) Table 157: overview of wavelength intervals for the different colours within the visible light spectrum The colours of the visible light spectrum Colour ultraviolet Near ultraviolet Violet Blue Green Yellow Orange Red Near Infrared Infrared Wavelength interval < 300 nm 300-400 nm 400-450 nm 450-490 nm 490-560 nm 560-590 nm 590-630 nm 630-700 nm 700-1000 nm > 1000 nm To make plotting possible on the CIE1931 horseshoe graph, the colour coordinates were calculated for every transmission spectrum (Figure 186, top left) (DeCustatis 1997). This has been done by means of the two chromatic values ‗x‘ and ‗y‘. The system, however, is based on three primary colour stimuli ‗X‘, ‗Y‘, and ‗Z‘, which correspond approximately to the colours red ( x ), green ( y ), and blue( z ), respectively. These stimuli were historically chosen so that each visual colour could be represented as a mixture of these stimuli. For every wavelength, the so-called colour matching functions x , y and z (Figure 186, below) reflect how much of every stimulus ‗X‘, ‗Y‘, and ‗Z‘, one has to mix so that the resulting substance has the same colour as monochromatic light at the considered wavelength (λ) (Meulebroeck 2004, 61-63). 367 Tristimulus values 4 2 1.6 _ z _ y 1.2 _ x 0.8 0.4 0 380 480 580 680 780 Wavelength (nm) Figure 186: (top left) CIE1931 horseshoe curve; (top right) Colour matching functions showing normalized typical human cone cell responses (Z, Y and X) to monochromatic spectral stimuli (taken from Fortner, Meyer 1997); (bottom) the spectra of the three primary colour stimuli (taken from DeCustatis 1998) The entire assignment has also been done in the reverse order to ascertain the calculated results. First, the chemical composition of the artefact and the applied production process were prefigured from the measured spectra. Afterwards, the calculated results were verified with the data acquired from the chemical analysis explained in Chapter 8. By linking the results on the optical characterization of the black-appearing glass with the data from the chemical composition via the two minor-destructive analysis methods, i.e. SEM-EDX and LA-ICPMS, useful information for archaeological research was obtained. The measuring of the transmission spectrum of each artefact generates three general remarks with respect to this set-up. The variety of questions and interests imposed the employment of various methods to analyse the spectral shape of the black glass artefacts. For certain purposes, such as the determining of the blackness, it was necessary to measure absolute spectra, and, as a consequence, the use of an integrating sphere was required. Principally, the integrating sphere has not been used because a useful quantification is already obtainable when special care is given to the applied measurement geometry, i.e. the light source, the angle, and the distance between the glass artefact and the light source, and the detector. In particular, the properties of glass do specify the blackness of a glass artefact, although we have to take notice of various external circumstances influencing visual perception. We have to consider (1) the spectral properties of the lightinge.g. daylight vs. artificial light, halogen light vs. fluorescent light; (2) the measuring rangee.g. direct light vs. diffuse light, reflection (R) vs. transmission (T); and (3) the powers of perception on the part of the observer. We therefore have applied standardised methods, using a standardised light source excluding all interference of external sources of light. The calculated colour is related to the chosen measuring range as well as the chosen light source. Hence, the calculation of the numerical values in the various colour spaces of the measured glass artefact is done by multiplying the three tristimulus values (Table 158):   368 the spectral power distribution of the used light sourcei.e. Illuminant A; the spectral reflectance of the measured artefact  the colour matching functions. Table 158: The calculated tristimulus values in the three colour spaces sample hue L* Velzeke Ab black appearing greyish green 0,084 Rumst b deep dull greenish yellow 32,69 Grobbendonk very pale greyish green 92,14 a* -0,0074 -6,18 -1,54 b* 0,018 28,84 12,07 a b c d Figure 187: The spherical colour spaces with vertically the lightness of the colour (L*) and horizontally the hue in two axes, green-red (a*) and yellow-blue (b*) (a-c: provided by Wendy Meulebroeck; d: taken from http://en.wikipedia.org/wiki/Munsell_Color_System) Within the spherical colour diagram, we have to consider three axes: 1) the vertical Z-axis L* that stands for lightness between white and black representing the value of a colour; 2) the horizontal X-axis a* that stands for the colour transition from green to red; and 3) the horizontal Y-axis b* that stands for the colour transition from blue to yellow. Figure 187 demonstrates the horizontal X- and Y-axes give the saturation of the chroma or hue from grey in the centre over dull to vivid hues at the outer peel. The lower the L*-value on the vertical Z-axis, the darker the glass hue, with very dark glass from L*10 to black at around 0. The higher the L*-value, the paler the 292 293 glass is, with very pale glass from L*90 to colourless at around 100. Both horizontal axes a* and b* show vivid colours when the value is high, between +/- 60 to +/- 40; dull, up to +/- 20; and greyish to grey, below +/-20 292 293 For glass, we speak of it as almost colourless if it is transparent and lightly tinged white when opaque. In this case, we speak of it as colourless glass when it is transparent and white when opaque. 369 (Figure 187a-c). We can conclude that the value L* can be used as a potential parameter to quantify the blackness of glass artefacts. 9.4 UV-Vis-NIR Spectroscopy results on Roman artefacts in stronglycoloured glass appearing black This section highlights the UV-Vis-NIR spectroscopy results on a selection of Roman black glass objects within the framework of the five questions formulated in section 9.1 [see Appendix 3]. We therefore selected black artefacts in various glass colours whose chemical composition was analysed [see Appendix 1]. A number of 294 samples remained indeterminate because the true colour appeared beyond measure. The largest part of these so-called opaque black glass samples is due to the presence of an opacifying agent, such as antimony, that causes a high absorption in the visible part of the electromagnetic spectrum. However, some artefacts appeared too thick to allow transmission of light. Not visible for the naked eye, these black-appearing pieces disclose their true colour rather by accident through a thin fractured surface. Making the transmission of the used halogen white light source not properly measurable, these samples have also been excluded from this research. Hence, all samples investigated by way of UV-Vis-NIR spectroscopy are characterised by the transmittance of the visible light that could be received accurately and accordingly to reveal the colour of the transparent and translucent glass matrix. We were able to transfer to the TONA lab at the Vrije Universiteit Brussel a total of 96 samples for UV-Vis-NIR spectroscopy [see Appendix 3], very kindly donated by various institutions from different countries. The accumulated set of samples comprises for the most part Belgian sites (Amay, Bassenge, Bonsin, Florennes, Grobbendonk, Izier, Kontich, Kruishoutem, Maffe, Matagne-la-Petite, Nismes, Oudenburg, Rumst, Tienen, Tongeren, Treignes, Trou de Han, Velzeke). A number of the analysed material, however, comes from various other places of the West European Continenti.e., France (Olbia), Italy (Rome), the Netherlands (Nijmegen, Utrecht), Portugal (Braga) and Switzerland (Avenches). The majority of the measurements are taken from vessels (68 ex.), but a certain number are jewellerye.g. 16 bracelets, 1 finger ring, and 1 hairpin. The remaining samples include counters (4 ex.), architectural decoration 295 material (4 ex.), and linen smoothers (2 ex.). The majority of the analysed material is from Period II (48 ex.), and the remaining number of samples are from Period I (22 ex.), and III (22 ex.). None of the selected material seemed to belong to Period IV except for the counter from Braga (PT) due to its HIMT composition. The counters (2 ex.) are, however, in general not well dated, as they appear in all four periods, and remain no issue for intensive assessment. The linen smoothers (2 296 ex.) are most probable post-Roman and one piece of raw glass from Rumst (BE) remained indeterminate. The analysed artefacts are almost entirely made from various green glasses (76 ex.), except in only a few cases where the glass matrix is purple (3 ex.), brown (1 ex.), blue-green (1 ex.) or blue (4 ex.). A remaining 11 pieces the true hue of the black appearing smaples remained indeterminate for the naked eye. The obtained results from UV-Vis-NIR spectroscopy are discussed in connection with the previously formulated questions. o Is a distinction between the black and non-black appearance as opposed to the strongly-coloured glass measurable? The initial experiment was executed without using an integrated sphere in the set-up, and resulted in measuring the transmission spectra characteristic to the glass hue of an artefact (Figure 188), and verifying whether a scale of blackness could be determined. 294 The non-black appearing pieces integrated in the discussion here below on the analysis results have been omitted from the appendix. Although the linen smoothers were excavated at the late Roman military camp of Oudenburg (BE), we are fully aware that this type of artefact is characteristic for the medieval period. From this point of view, we wished to verify what differences or similarities can be detected, particularly because no chemical analysis has yet been carried out. 296 See Introduction, on the delimitation of the studied material. 295 370 Spectroscopic measurements Roman glass: session 2 Rumst1 Rumst2 Rumst4 Rumst 5 Floren1 Floren2 Floren3 Floren4 120 UV IR Transmission (%) 100 80 60 40 20 0 350 400 450 500 550 600 650 700 750 800 Wavelength (nm) Figure 188: transmission spectra of Roman strongly-coloured glass appearing black from Rumst and Florennes between 350 297 and 800 nm Through the calculation of the x and y coordinates, the measured transmission spectra can be plotted on the CIE1931 horseshoe graph (Table 159; Figure 189). It shows that the blue pieces are in fact to be described as purplish blue, while the green samples are mainly plotted at the fringe of yellow green and greenish yellow (Rumst1; 2; 5), but can be yellowish green (Floren4), or even yellow (Rumst4; Floren3). The visual perception thus corresponds only roughly to the objectively measured colour coordinates, demonstrating the observation limits of the naked eye and the individual‘s biased perception in describing colours. On the other hand, it does not give any precise information about the degree of blackness. Table 159: x and y-coordinates for plot on CIE1931 horse-shoe graph (* hue perceived by the naked-eye) appendix no. provenance sample hue* x y 64 Rumst (BE) Rumst1 green 0,41 0,49 65 Rumst (BE) Rumst2 green 0,45 0,53 67 Rumst (BE) Rumst4 green 0,50 0,49 68 Rumst (BE) Rumst5 green 0,42 0,49 16 Florennes (BE) Floren1 blue 0,15 0,04 17 Florennes (BE) Floren2 blue 0,15 0,06 18 Florennes (BE) Floren3 green 0,49 0,50 19 Florennes (BE) Floren4 green 0,34 0,56 297 This material will be discussed within this chapter in section 9.5, verifying the discrepancy that can occur within one single object, and whether it is possible to recognize batches. 371 Figure 189: Plot of the Rumst and Florennes samples on the CIE1931 horseshoe graph To achieve further improvement in establishing a way to determine the blackness of glass artefacts, the absolute spectra of the transmitted light were calculated in a new set by using an integrated sphere. We therefore analysed two pieces of deep green glass appearing blackfrom Velzeke (BE) [cat.no.595] and Rumst (BE) [cat.no.510] and compared it with three non-black pieces from Grobbendonk (BE), respectively in purple (72.GRO.520), blue (70.GRO.349), and pale green (no inv.no.). Figure 190 demonstrates a clear difference between the transmission spectra of the two black fragments and that of the three coloured glass fragments from Grobbendonk. The latter have marked transmission peaks within the visible spectrum compared to the blackappearing fragments. The Velzeke piece has a very high absorption degree within the visible wavelength between 400 nm and 700 nm, whereas the Rumst piece shows a faint transmission peak within the yellow range so that it is easily perceived as olive green glass when held to a light source. Conspicuous is the similarity of both spectra below 400 nm in the ultraviolet spectrum as well as from about 800 nm in the infrared spectrum. These results lead to the Rumst piece being regarded as (poorly) translucent glass, whereas the Velzeke fragment is considered to be (nearly) opaque black. 372 Velzeke Ab Rumst b 72.GRO.520 70.GRO.349 light green 100 Transmission (%) 10 1 300 500 700 900 1100 1300 1500 0,1 0,01 0,001 Wavelength (nm) Figure 190: The different transmissions of light according to the particular hue of glass By converting the measured spectra to their optical density (OD), we see that the transparent pieces from Grobbendonk result in an OD far below 1, whereas the two black pieces have an OD far above 1 (Figure 191). Remarkable is that the nearly opaque black piece from Velzeke has an OD of 4, between 380 and 700 nm within the visible spectrum. The translucent Rumst piece fluctuates between 1 and 3 within the visible spectrum. When calculating the integrated optical density (OD) between 300 nm and 1600 nm, the transparent coloured pieces are even below 0,5, while the two black pieces are 2,5 (Rumst) and c.3,5 (Velzeke) (Figure 192). Velzeke Ab Rumst b 72.GRO.520 70.GRO.349 light green 4,5 4 3,5 3 O.D. black glass 2,5 2 1,5 1 3 coloured, non-black appearing glass samples 0,5 0 350 450 550 650 750 850 950 1050 Wavelength (nm) Figure 191: The transmissions of light converted into optical density (OD) 373 4,00 3,50 increasing blackness Integrated O.D. (300-1600 nm) 3,00 2,50 2,00 1,50 1,00 0,50 0,00 Sample Velzeke Ab Rumst b black appearing glass 72.GRO.520 70.GRO.349 light green non-black appearing coloured glass Figure 192: integrated optical density of black-appearing glass versus non-black appearing coloured glass To confirm these results, we measured a supplementary set of strongly-coloured black- appearing glass, comprising sometimes more than one measurement per artefact (Table 160). At first glance, a distinct difference is to be seen between the Period I material from Olbia (FR) and the Period II material from the various Belgian sites (Grobbendonk, Kruishoutem, Matagne-la-Petite, Rumst, Velzeke) the first 20 measurements. The Period I material shows an integrated optical density between 300 nm and 1600 nm below 2,5, whereas that of the Period II material is above 2,5 (Figure 193). One piece from Olbia (C014a) has a density even below 1. Possibly this is due to the very low iron content of translucent deep blue glass. 298 Table 160: List of integrated optical density for 31 measured black appearing glass samples from 21 objects appendix integrated OD integrated OD provenance sample period colour no. 300-1600 nm 400-700 nm 29 Kruishoutem (BE) II green-yellow 2,93 2,35 19 Grobbendonk (BE) II green-yellow 3,10 2,72 89 Velzeke (BE) Aa II green-yellow 3,21 3,32 89 Velzeke (BE) Ab II green-yellow 3,40 4,07 90 Velzeke (BE) B II green-yellow 3,33 3,85 91 Velzeke (BE) Da II green-yellow 3,39 3,85 91 Velzeke (BE) Db II green-yellow 3,13 3,32 91 Velzeke (BE) Dc II green-yellow 2,68 2,66 92 Velzeke (BE) F II green-yellow 3,02 2,93 93 Velzeke (BE) G II green-yellow 2,45 1,42 31 Matagne-la-Petite (BE) A II green-yellow 3,14 3,19 32 Matagne-la-Petite (BE) B II green-yellow 2,93 2,69 33 Matagne-la-Petite (BE) C II green-yellow 3,07 3,22 34 Matagne-la-Petite (BE) D II green-yellow 3,18 3,24 35 Matagne-la-Petite (BE) E II green-yellow 3,18 3,41 36 Matagne-la-Petite (BE) G II green-yellow 3,12 3,29 62 Rumst (BE) A=1 II green-yellow 2,61 1,54 63 Rumst (BE) B=2 II green-yellow 2,49 1,50 65 Rumst (BE) C=4 II green-yellow 2,88 2,59 298 Some pieces have been measured several times to see whether homogeneity can be observed and to what extent a margin must be taken in consideration. In 10.4.4 the possibilities of using OD measurement for archaeological queries are discussed. 374 66 44 41 41 41 41 41 41 41 43 45 45 Rumst (BE) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) Olbia (FR) D=5 037-20 037-30a 037-30b 037-30c 037-30d 037-30e 037-30f 037-30g 061-3 C014a C014b II I I I I I I I I I I I green-yellow pink-purple yellow-orange yellow-orange yellow-orange yellow-orange yellow-orange yellow-orange yellow-orange pink-purple blue blue 3,20 1,36 2,98 2,35 2,48 1,77 1,81 1,53 1,99 2,08 0,63 2,43 3,49 3,71 3,44 2,87 3,09 2,70 2,71 2,32 2,61 1,89 1,77 3,83 4,00 3,50 Integrated O.D. (300-1600 nm) 3,00 2,50 2,00 1,50 1,00 C014 b 061_3 C014 a 037_30 f 037_30 g 037_30 e 037_30 c 037_30 d 037_30 b 037_20 037_30 a Rumst c Rumst d Rumst a Rumst b MatagneE MatagneG MatagneC MatagneD MatagneB Velzeke G MatagneA Velzeke F Velzeke Dc Velzeke Db Velzeke B Velzeke Da Velzeke Ab Velzeke Aa Kruishoutem Grobbendonk 0,50 Sample Figure 193: Overview of the integrated optical density of the 22 black-appearing artefacts within the UV-VIS-IR (300-1600 nm) In conclusion, from both tests we can say that the degree of lightness of the glass hue can be determined from the measured OD level. The OD is calculable on only those objects revealing their true hue when light in 299 transmission is used. Two limits have importance: all material with an OD of 2,5 or more have a green(ish) hue and contain a high level of iron. Those with an OD between 1,5 and 2,5 have a much lower iron content and appear more orange-brown. The only piece far below the 1,5 limit, Olbia C014a, is made of blue glass. The link between the OD level and the glass hue demonstrates why the human eye more easily accepts black-appearing green glass as black, while black-appearing blue glass was only defined as dark blue. Hence, the measurement of light transmission by using an integrated sphere, and the conversion of the normalized spectral curve to obtain the integrated OD, does not make the limit between strongly-coloured glass and black calculable. It only explains why human perception is deficient through less easily ascribing Period I material as black-appearing, compared to the Period II material with a higher OD; defines the opacitytranslucency-transparency of the glass artefacts; and shows why the high iron green glass became the material used most successfully in Roman times to produce black-appearing artefacts. 299 UV-VIS-NIR spectroscopy of black glass artefacts with an opacity making the transmission of light impossible could not be done, so their OD remains unknown. 375 o Does the colorimetry of glass correspond to its chemical composition? Is the chemical composition responsible for the colour of a glass artefact, or to put it differently, is it possible to determine the chemical composition of glass from the colour measurement? In a preceding section (10.2), we assembled a useful outline of the various identified chemical elements on the measured transmission spectra of Roman glass artefacts. The characteristic absorption bands can estimate what metal oxides were used for colouring the glass. Table 161: Overview of the examined artefacts, their colour coordinates and concentration in iron oxide, manganese oxide, copper oxide and cobalt oxide appendix sample/site hue calculated concentration of no. colour metal oxides (CIE 1931) SEM-EDX LA-ICP-MS x y Fe2O3 MnO CuO CoO 2 Avenches 1 green-yellow 0,40 0,52 9,17 0,29 0,0053 0,0060 3 Avenches 2 green-yellow 0,48 0,50 8,99 0,18 0,0082 0,0055 4 Avenches 3 green-yellow 0,48 0,50 9,04 0,19 0,0084 0,0057 5 Avenches 4 yellow-orange 0,60 0,39 1,65 0,48 0,0039 0,0007 6 Avenches 5 yellow-orange 0,60 0,37 1,53 0,36 0,0074 0,0005 7 Avenches 6 yellow-orange 0,49 0,47 1,65 0,41 0,0047 0,0007 8 Avenches 7 yellow-orange 0,54 0,46 1,69 0,38 0,0048 0,0007 9 Avenches 8 yellow-orange 0,55 0,44 2,32 0,41 0,0669 0,0007 10 Avenches 9 yellow-orange 0,55 0,44 2,23 0,37 0,0048 0,0006 11 Avenches 10 yellow-orange 0,52 0,46 2,13 0,40 0,0100 0,0006 15 Braga green-yellow 0,58 0,29 10,77 0,96 [0,18] 19 Grobbendonk green-yellow 0,51 0,48 11,09 0,45 28 Kruishoutem 1 green-yellow 0,31 0,34 10,78 0,25 29 Kruishoutem 2 green-yellow 0,49 0,49 11,13 0,23 31 Matagne-la-Petite 1 green-yellow 0,51 0,48 9,77 0,22 0,0200 0,0103 32 Matagne-la-Petite 2 green-yellow 0,51 0,48 8,88 0,22 0,0120 0,0056 33 Matagne-la-Petite 3 green-yellow 0,52 0,47 10,38 0,22 0,0069 0,0076 34 Matagne-la-Petite 4 green-yellow 0,52 0,47 8,86 0,22 0,0113 0,0053 35 Matagne-la-Petite 5 green-yellow 0,53 0,47 9,06 0,21 0,0119 0,0054 36 Matagne-la-Petite 6 green-yellow 0,46 0,48 12,71 0,22 0,0040 0,0068 37 Matagne-la-Petite 7 green-yellow 0,46 0,49 10,46 0,25 0,0094 0,0074 38 Matagne-la-Petite 8 green-yellow 0,45 0,47 11,92 0,28 0,0102 0,0079 40 Nismes green-yellow 0,44 0,47 8,83 0,22 41 Olbia 1 yellow-orange 0,59 0,41 1,45 0,63 0,026 0,0035 42 Olbia 2 yellow-orange 0,56 0,44 1,60 0,37 0,036 0,0035 43 Olbia 3 pink-purple 0,44 0,18 0,59 2,60 0,068 0,0048 44 Olbia 4 pink-purple 0,48 0,26 0,46 2,14 0,0067 0,0065 0,22 45 Olbia 5 blue 0,15 0,11 2,35 2,14 0,15 [0,28] 48 Oudenburg 1 green-yellow 0,43 0,49 10,52 0,45 0,0104 0,0032 49 Oudenburg 2 green-yellow 0,45 0,51 9,29 0,49 0,0096 0,0039 50 Oudenburg 3 green-yellow 0,42 0,50 10,54 0,50 0,0103 0,0035 58 Rome1 blue-green 0,28 0,43 1,1 0,2 0,0471 0,0519 59 Rome2 blue-green 0,22 0,35 1,4 0,3 0,0901 0,0596 60 Rome3 brown 0,64 0,36 0,4 0,4 0,0046 0,0004 61 Rome4 purple 0,58 0,26 0,5 2,0 0,0020 0,0015 67 Tienen 1 green-yellow 0,51 0,48 7,66 0,43 0,0060 0,0037 68 Tienen 2 green-yellow 0,51 0,46 8,77 0,21 0,0061 0,0041 69 Tienen 3 green-yellow 0,42 0,48 8,28 0,16 0,0056 0,0058 70 Tienen 4 green-yellow 0,41 0,46 7,72 0,14 0,0075 0,0030 71 Tienen 5 green-yellow 0,45 0,52 8,48 0,17 0,0060 0,0038 72 Tienen 6 green-yellow 0,44 0,50 8,63 0,15 0,0039 0,0090 73 Tienen 7 green-yellow 0,45 0,48 11,44 0,16 0,0063 0,0038 74 Tienen 8 green-yellow 0,47 0,49 12,09 0,08 0,0056 0,0066 89 Velzeke 1 green-yellow 0,46 0,52 12,23 0,24 90 Velzeke 2 green-yellow 0,46 0,52 12,31 0,26 91 Velzeke 3 green-yellow 0,51 0,49 11,41 0,30 92 Velzeke 4 green-yellow 0,52 0,47 12,40 0,72 93 Velzeke 5 green-yellow 0,40 0,46 8,31 0,59 - 376 From Chapter 8, it is clear that iron oxide and manganese oxide are the two main substances responsible for the black appearance of glass artefacts. We have distinguished four different groups of black glass artefacts by using the standard chemical analysis techniques SEM-EDX and LA-ICP-MS on the basis of the colouring, and opposing the iron-manganese concentration chart (Figure 194; Table 162). 3 purple 2,5 blue 2 HIMT MnO 1,5 Braga 1 c.350 AD orange-yellow 0,5 brown ? blue-green 0 0 2 c.150 AD 4 6 8 10 12 14 Fe2O3 green-yellow yellow-orange pink-purple blue blue-green brown Figure 194: Iron content versus manganese content (%wt) of the first set of 48 artefacts, showing the clear compositional difference between artefacts produced at different periods in time. Table 162: list of the four groups of black glass based on the colouring by means of iron and manganese content class iron oxide % manganese oxide % glass type UV-Vis-NIR hue Period I II III IV high low moderate high >3 0,5-3 1-3 >3 low high low high <0,6 2-3 <0,6 >0,6 HILM LIHM MILM HIHM=HIMT green-yellow blue to pink-purple yellow-orange yellowish green II-III I I IV All artefacts within Class I are characterised by a high iron content above 3% and a low manganese oxide concentration below 0,6% (HILM) that made the glass turn greenish-yellow. The samples measured show a very high iron content between 8-12% and a manganese content below 0,3%, even though a cluster is observable having a manganese content ranging around 0,5% and 0,7%. This cluster is possibly the result of using manganese decolourized raw glass or colourless cullet added to the batch. The artefacts of Class II combine a low to moderate iron content below 3% with a high manganese oxide concentration of 2-3% (LIHM) resulting in pink-purple coloured glass. Those with cobalt and copper oxide have an ultramarine blue hue. The samples within Class III have a low manganese content similar to that of Class I, but the iron content is much lower, below 3%, comparable to Class II. The samples of this Class show a yellow-orange colour (LILM). The only sample measured within Class IV is the piece from Braga (PT) and features a high iron content, above 3%, and a high manganese content (HIHM). This entirely corroborates the late Roman black glass material made th from the well-known HIMT-glass that came into circulation within the 4 century AD (Mirti et al. 1993; Freestone 1994; Foy et al. 2003; Freestone et al. 2005). It is not excluded that the higher manganese content of artefacts 300 within the Class I cluster is caused by a mix of Class IV material with recycled Class I cullet. 300 Seeing the low titanium content of the concerned pieces from Velzeke, Grobbendonk, Tienen and Oudenburg, as well as their early dating in Period II, we assume that these black glass objects were produced with decolorized raw glass. 377 nd The full range of samples from Class I are characteristic for Periods II and III, i.e., from the second half of the 2 th century up to the early 4 century AD. Classes II and III are to be dated in Period I, and more specifically within st th the 1 century AD. Class IV is characteristic for Period IV from about the mid 4 century AD, but possibly this type rd th of raw glass slowly got introduced from the late 3 or early 4 century AD. The detection of a sequential production of black-appearing glass artefacts obtained from UV-Vis-NIR spectroscopy corroborates the conclusions which were drawn in earlier research (Van der Linden et al. 2009). Testing strongly-coloured glass means that we focussed on the various colorants added to the batch to obtain black-appearing glass. To verify its value for archaeological research, it is necessary to test UV-Vis-NIR spectroscopy on large sets of ‗naturally coloured‘ and decolourized glass artefacts. Linked with the chemical analysis, more fine tuning reduces the black glass groups into four large classes, taking notice of some important sub-categories. The measured samples have thus been classified according to four major classes corroborating the observed chemical groups of Roman black glass artefacts. Class I Class II Class III Class IV high iron/low manganese group low iron/high manganese group low iron/low manganese group high iron/high manganese group HILM LIHM LILM HIHM (=HIMT) Class I Three subgroups of green coloured glasses can be recognized within Class I, featuring a cluster with high iron and low manganese content. The samples here enclosed have an iron content between 8% and 13%, but lower percentages down to 4% or even 3% have been analysed [see Chapter 9]. The chemical composition of the first subgroup (a) is characterised by a very low amount of manganese (below 0,4%) corresponding to the impurity level typical for ‗naturally‘ coloured glass. The second subgroup (b) of green glasses has a slightly more elevated manganese content (0,4-0,6%) but is also present here as an impurity. The third subgroup has a manganese content above 0,6% and up to 1%, which means that the glass is much paler and borders the zone of Class IV that encloses the artefacts made from HIMT glass and having a manganese content above 1%. We assume, therefore, that it is more appropriate to bunch these different high iron levels into one single group, Class I, but still keep them in separate sub-categories and denominate them as Class Ia; Ib; Ic. Class II The chemical composition of this group is characterised by a high manganese concentration between 2% and 3%, and thus intentionally added. The iron oxide content below 0,4% corresponds to the impurity level in the sand, but it can go up to around 2%, and be regarded as a second subgroup. The elemental composition means that the Class II material is characterised by a purple hue; however, Olbia 5 has a deep blue hue. This is most probably due to high copper and cobalt content, excluding any influence from the concentration of iron oxide or manganese oxide. The three samples within subgroup IIa ( Rome 4; Olbia 1 and 2), and likewise the one sample within st subgroup IIb (Olbia 5), are all dated in the 1 century AD. Even though we suggest one major group, Class II, featuring a low iron/high manganese content, we have to bear in mind that each specific hue might represent a sub-category. It is unclear at this stage whether this class is st featuring particular colours within a specific period, namely the 1 century AD, or all through time. Class III The artefacts within this class have a low iron and low manganese content and are all to be dated in Period I, st specifically in the 1 century AD. Two sub-categories can be distinguished. Subgroup IIIa has an iron level above the impurity level, though barely elevated with a concentration below 3%. The manganese concentration, on the other hand, is negligible. This subgroup encloses eleven samples from various sites: Olbia 3 and 4, Rome 1 and 2, and Avenches 4-10. Subgroup IIIb is characterised by the iron and the manganese concentration far below 0,4%, and therefore they are to be considered as impurities. Hitherto only one single pieceRome 3has been recognized from all selected samples. Class IV Within the high iron black glass material occurs with a high manganese content that is above 1%, featuring also a more elevated titanium content above 0,3%-0,4%, in contrast with the < 0,2% normally seen. This type of glass, 378 which was observed only about 15 years ago, is nominated HIMT (Mirti et al. 1993; Freestone 1994); it would then form our Class IV. Only the Braga sample [see Appendix 3] falls within this class. It is, however, worthwhile to focus on this material since HIMT glass is typical for the late Roman II/Byzantine I period and in use for the th production of black glass artefacts from about the mid 4 century AD onwards, such as the sampled material from Trier (DE), Castra (IL), and partly from Oudenburg (BE) [see Chapter 8]. o Is it possible to calculate the variety of shades per specific colour type resulting in a black-appearing glass matrix? Spectral Group 1 120 Tmax = 495 nm Transmission (%) 100 Velzeke(sample 5 Velzeke E) 80 60 40 1100 nm; Fe2+ 380 nm & 420 nm & 440 nm; Fe3+ Kruishoutem Kruishoutem 1 (KK/PROSP78KB) 20 UV absorption; Fe3+ 0 350 450 550 650 750 850 950 1050 1150 1250 Wavelength (nm) Figure 195: Transmission spectra of Spectral Group 1 In Figure 195 the spectra of the 2 artefacts belonging to this group are shownVelzeke 5; Kruishoutem 1. The peak transmission of both artefacts is situated around 495 nm. The low transmission of ultraviolet absorption ranging between 350 and 380 nm, and the series of weak absorption bands in the region between 375 and 450 3+ nm (at 380, 420 and 440 nm), features the presence of Fe ions (Meulebroeck et al. 2011). The spectral shape is considered to be characteristic for strongly reduced glasses. This assumption is based on the presence of a broad intense absorption band at 1100 nm and a shoulder around 450 nm, which are apparently generated by 2+ Fe ions (Meulebroeck et al. 2011). Table 158 consolidates the discovery about the reducing furnace conditions. 2+ 3+ It is clear that the main colouring agent is iron (with a higher concentration of Fe than Fe following Table 158). No absorption bands of other elements are present. For this spectral group, we expect a blue-green ‗aqua‘ colour. Hence, it is not surprising that the samples of this group correspond with the samples of Class I defined earlier, based on the chemical analysis. It has been previously stated that iron could be present as an impurity in the sand (as FeO) or could be added as an extra, e.g., as magnetite (Fe2O3.FeO). Various references cite potential absorption bands of both types of iron 2+ 3+ 2+ ions (Fe = ferrous iron and Fe = ferric iron). The most pronounced absorption band is the one of Fe between 900-1200 nm with a shoulder around 450 nm. Other possible absorption bands attributable to iron are a series of feeble pointed bands mainly in the region 375-450 nm (380 nm, 415 nm and 440 nm), and a very strong band at 3+ 3+ 2258 nm because of the presence of Fe . In the presence of sulphate ions, the ferro-sulphide Fe - S chromophore could be formed with an intense absorption band around 405 nm (Meulebroeck et al. 2011). 379 Spectral Group 2 120 Tmax between 564 & 588 nm Transmission (%) 100 80 Matagne (sample E) 60 Velzeke (sample G) 40 Avenches (88/6501-9) 20 0 350 400 450 500 550 600 650 700 750 Wavelength (nm) Figure 196: Transmission spectra of Spectral Group 2 samples from Avenches, Matagne-la-Petite and Velzeke 120 Oudenburg 100 Transmission (%) Braga 80 60 40 Tienen 20 0 350 450 550 650 750 850 Wavelength (nm) Figure 197: Transmission spectra of Spectral Group 2 of samples from Braga, Oudenburg and Tienen This spectral group, with 30 samples from Period II or III, is by far the largest sampled group, containing artefacts from Avenches 1-3, Braga 1, Grobbendonk, Kruishoutem 2, Matagne-la-Petite 1-8, Nismes, Oudenburg 1-3; Tienen 1-8 and Velzeke 1-2; 4; 6-7 [see Appendix 3]. The calculated colour coordinates in Figure 194 demonstrated one matching spectral shape for each set of samples from Oudenburg and Tienen. For matters of clearance, we have decided to put only one spectrum per site in the graph. Both graphs show that the spectral shape of this group is featuring the presence of iron. All samples are characterised by a low transmission in the ultraviolet and the blue spectral region (between 350 and approximately 450 nm), and a high transmission of green and red wavelengths (between 560 and 590 nm). This results in a brownish-green or olive-green glass hue. 380 With the colour coordinates corroborating the iron and manganese concentrations of the samples (Table 158), this means that this spectral group corresponds with Class I in the iron-manganese chart. Important here is that a comparison of these results with the data obtained from chemical analysis shows that these calculated colour coordinates and spectral shapes correspond with a high iron concentration situated between 8% and 13% and a low manganese concentration below 1%, but for the most part averaging between 0,3% and 0,6%. Further evaluations and comparisons are needed to verify these thus- far relative estimations. The calculated colour coordinates could give us a first indication about the iron and manganese concentration. Analysis of the spectral shape could give us information about the applied fabrication process. In the next part we will use this information to briefly highlight how this tool can be applied in answering different types of archaeological questions. Even though we analysed a large amount of samples, we only show in Figure 197 the spectra of three artefacts. 3+ All others with respect to the spectral shape and the absorption bands of Fe have disappeared. This is an 2+ indication that a large fraction of the iron is reduced to Fe (see chemical reaction [2]). The formed sulphide is 3+ combined with the remaining Fe ions to form the ferro-sulphide (see chemical reaction [3]). The transmission maxima are positioned between 564 nm and 588 nm (Figure 196). Following Table 158, this indeed proves that 2+ Fe ions are present together with the ferro-sulphide chromophore. The high absorption around 405 nm is a result of the presence of ferro-sulphide. In other words, a large absorption in the ultraviolet and blue part of the electromagnetic spectrum indicates the presence of iron. Moreover, we observe for all the samples a transmission 2+ maximum in the green part of the spectrum (between 564 nm and 588 nm). This proves that Fe ions are present together with the ferro-sulphide chromophore (Brill 1988). Based on the presence of the absorption of iron (and only iron), we can conclude that the main colouring agent is iron. It is not a surprise that this group corresponds with the earlier-defined group II. Spectral Group 3 120 100 Olbia (CO14) Transmission (%) Olbia (037-20) 80 60 Tmax: 395 nm Rome (R1609 - no.77) 40 450-550nm; Mn3+ 20 0 350 450 550 650 750 850 950 1050 1150 1250 Wavelength (nm) Figure 198: Transmission spectra of Spectral Group 3 The spectral shape of this group of artefacts with colour coordinates falling in the pink and red-purple region of the CIE1931 horseshoe curve indicates the presence of manganese. The three artefacts represented in Figure 198Olbia1, Olbia5, Rome4are characterised by a transmission maximum (Tmax) around 395 nm and a very strong absorption band between 450 and 550 nm, although a poor transmission percentage goes on to 650 nm. There is again a very high transmission within the infrared spectrum, with a faint peak at c.950 nm and a stronger 3+ one at c.1250 nm. The strong absorption band of 450-550 nm corresponds with the presence of Mn ions and indicates that the glass batch was heated under oxidising conditions (Table 158) (Meulebroeck et al. 2009). Roman glass contains, in most cases, small concentrations of manganese by nature as an impurity added 2+ through the alkali flux, but from a Mn -concentration exceeding 0,4%, it is accepted to be intentionally added 2+ 2+ (Brill 1988). This is initially under the form of the decolourizing ion Mn , but under more oxidising conditions Mn 381 3+ is transformed into the red colouring ion Mn with an absorption maximum between 470 and 520 nm (Weyl 1959; Bamford 1977; Green, Hart 1987; Baxter 1995), and a strong transmission in the infrared spectrum. A marked contrast with the previous group(s) is what concerns the spectra, the near-absence or poor presence of iron. Hence, this third spectral group corresponds with the earlier defined chemical low iron/high manganese group (Class III). The data obtained via chemical analysis, showing a high manganese content for these purple coloured glass samples, only strengthens this theory. Spectral Group 4 120 100 Transmission (%) 4b Olbia (061-3) Rome (R1609 - no.59) 80 Rome (R1609 - no.56) 60 40 absorption between 580 & 660 nm 4a 20 0 350 400 450 500 550 600 650 700 750 800 Wavelength (nm) Figure 199: Transmission spectra of Spectral Group 4 Spectral Group 4 embodies blue coloured glass showing three transmission peaks in the visible part of the electromagnetic spectrumrespectively 470-500 nm, 560-570 nm, 710-720 nmand a strong absorption band between 580 and 660 nm. We define two subgroups (Figure 199): Subgroup 4a contains two samples (Rome R1609 – nos. 56;59 = Rome 1-2) with a high absorption in the ultraviolet spectrum and the region between 400 and 450 nm. This is explained by the presence of both iron 3+ 2+ oxides, but with a predominance of Fe with regard to Fe (Brill 1988) and implies the use of switching furnace conditions, reducing, and oxidising. Both artefacts fall in the low manganese and low iron group, but the iron content is slightly above the impurity level, meaning that iron possibly got into a batch by the purposeful addition 2+ of a colorant, but perhaps not consciously added. Depending on the Fe concentration, the colour of these glass artefacts should range between bluish green (aqua) and greenish. Subgroup 4b (Olbia 061-3 = Olbia 3) has a similar spectrum but with a higher transmission in the blue region of the spectrum and a peak at about 420 nm. This indicates the presence of a second colouring agent. Looking to the chemical composition of the Olbia sample, manganese might be a candidate. This will shift the colour towards a more blue hue. Group 4b corresponds with chemical Group IV, whereas the samples of group 4a match with those of chemical Group V. From the iron-manganese chart (Figure 194), it is apparent that the Group 4b sample contains a high manganese, and a moderate iron concentration. The deep blue colour hue of the sample (colour coordinates x and y are respectively equal to 0.15 and 0.11), however, is obviously due to the presence of additional metal oxides in the batch. We discussed in a previous section that copper oxide and cobalt oxide are two colouring agents utilized by the Romans to produce a characteristic blue glass, from pale blue to a very deep blue, violet, or indigo hue (Green, Hart 1987; Brill 1988; Newton, Davison 1989; Scholze 1991; Shortland et al. 2006). Yet it is known that in many cases copper and cobalt were not used singularly as colouring agents. A mix of copper, cobalt, and manganese was usually utilized to obtain blue glass, even though an intentional addition of iron within this mix seems to occur too. The results of the chemical analysis indicate that the blue colour of the Olbia3 fragment was obtained via a mixture of iron, manganese, copper, and cobalt. The presence of these colouring elements is reflected in its spectrum, with a clear absorption band around 420 nm indicating the 3+ presence of the Fe iron oxide. The transmission maxima at 463 nm and 718 nm of the Olbia3 artefact are most 3+ probably to be linked to the presence of the manganese oxide Mn that causes a transmission maximum around 400 nm, and sometimes also at 700 nm (see above). However, the most distinct absorption bands are those of 382 cobalt. Previously, we explained that the transmission spectra are depending on the added form of cobalt and the presence of other elements, causing three successive absorption bands around 600 nm that can shift + + 2+ considerably. This difference is the result of a variation in ratios of Na , K , and Ca . The Olbia3 sample has absorption bands at 535 nm, 594 nm, and 645 nm, which is distinctive for Roman cobalt glass. This deviation in spectral position is explained by the additional presence of copper oxide, but this premise could not be confirmed within the investigation presented here, and needs additional research. Spectral group 1 Spectral group 2 Spectral group 3 Spectral group 4 yellow-green  greenish-yellow  yellow  yellowish-orange  orange  reddish-orange 0.6 CIE1931 yellowgreen greenishyellow yellow y (CIE1931) 0.5 0.4 0.3 yellowishorange pink orange 0.2 red-purple blue 0.1 reddishorange 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 x (CIE1931) Figure 200: Calculated colour coordinates of the first four spectral groups In Figure 200 we represent the calculated colour coordinates on the CIE1931 horseshoe curve. Spectral Group 5 120 Tmax = 715 nm 100 Transmission (%) Rome (R1609 - no. 69) 80 60 40 1100 nm; Fe2+ 20 0 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 Wavelength (nm) Figure 201: Transmission spectra of Spectral Group 5 Only one artefact (Rome R1609 – no. 69 = Rome 3) of the samples examined here falls within Spectral Group 5. Its spectral shape is quite similar to that of Spectral Group 2, with a strong absorption all the way in the ultraviolet and the blue between 350 nm and 540 nm. The transmission peak at 405 nmpresent in the spectral shape of 3+ Spectral Group 2, and due to the presence of Fe indicative of the addition of ferro-sulphide chromophoreis 2+ totally absent here. There is an absorption band at 1100 nm (indicating the presence of Fe ). The difference with Spectral Group 2 is the position of the transmission maximum (at 715 nm), which is shifted to a longer wavelength. The colour will therefore swing towards a more red(dish) glass. From these observations we can infer that the samples of Spectral Group 5 have been produced under more reducing conditions than those of Spectral Group 2. From the chemical analysis, we know that both the iron and manganese concentrations are present as impurities. In the remaining two spectral groups, we were not able to identify characteristic absorption bands apart from the ferro-sulphide absorption between 400 and 450 nm. Both groups have a different shape in the rest of the spectral 383 region, indicating different furnace conditions during the production process. Depending on this, the glass will be browner or redder coloured. Spectral Group 6 120 Transmission (%) 100 80 60 Avenches (89/7851-9) Avenches (89/7179-03) Avenches (89/7851-13) Olbia (294.2) 40 20 0 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 Wavelength (nm) Figure 202: Transmission spectra of Spectral Group 6 This spectral group consists of seven pieces from Avenches (Avenches 89/7179-03; 89/7851-9; 89/7851-13 = Avenches 8-10) and one from Olbia (Olbia 294.2 = Olbia 2). It features a high transmission of the green and red wavelengths rendered by two transmission maxima present in the visible part of the spectrum at about 600 nm and about 700 nm (Figure 199). All these artefacts have a high absorption between 400 and 450 nm, indicating 3+ the presence of Fe in the form of the ferro-sulphide chromophore complex, corroborating the data from chemical analysis. The calculated colour coordinates are situated in the yellowish-orange, orange, and reddish-orange region. type I type II 0.49 y (CIE1931) 0.47 0.45 0.43 0.41 0.39 0.37 0.35 0.40 0.45 0.50 0.55 0.60 0.65 x (CIE1931) Figure 203: plot of the calculated colour coordinates of the two types of artefacts belonging to Spectral Group 6 A striking detail is that all early artefacts from Avenches fall within the yellowish-orange Class III, whereas the contemporaneous material from Olbia integrated here shows a heterogeneity of differently coloured glasses appearing black. A plausible explanation is that the Avenches pieces come from one single production centre, while the Olbia material might have been imported from various production centres, one of which can be assumed in Avenches. Further down it will be explained that absorption spectroscopy can provide additional information to interpret the chemical cluster more precisely [see Spectral Group 4; Table 158]. 384 Spectral Group 7 120 Transmission (%) 100 Avenches (89/07170-46) Avenches (88/06918-02) Avenches (89/07170-43) 7a 80 Olbia (037-30) 60 40 Avenches (88/06975-05) 7b 20 0 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 Wavelength (nm) Figure 204: Transmission spectra of Spectral Group 7 This spectral group clusters artefacts with an increasing transmission in the visible part of the spectrum, and all through the infrared spectrum. A first transmission maximum is situated between 585 and 610 nm. We distinguish two subgroups: group 7a (Avenches 89/07170-43; 89/07170-46 = Avenches 4-5), and group 7b (Avenches 88/06918-02; 88/06975-05 = Avenches 6-7; Olbia 037-30 = Olbia 1), where the increase in transmission is respectively prolonged and stopped around 800 nm. In Figure 205, we resume the transmission spectra of the first four spectral shapes we defined above. Each group corresponds with a distinctive transmission spectrum and is defined in the iron-manganese graph (Figure 194). Spectral Group 1 represents the high iron/low manganese artefacts with colour coordinates situated in the yellowgreen, greenish-yellow, yellow and yellowish-orange regions. Spectral Groups 2 and 3 correspond with low iron/high manganese artefacts. The colour coordinates of Spectral Group 2 belong to the red-purple and pink region, while those of Spectral Group 3 fall in the blue area. Spectral Group 4 embodies low iron/low manganese artefacts with colour coordinates situated in the yellowish-orange, orange, and reddish-orange region. Spectroscopic analysis, however, enabled division of this spectral group into two subgroups related to the furnace conditions during the production process, i.e., the reducing and oxidising techniques. Consequently the same conclusions can be drawn from the spectroscopic measurement results as from chemical analysis on the chronological specificity in black-appearing glass production. In addition, the measured spectral shapes with their characteristic absorption bands and corresponding colour coordinates provide evidence on the used furnace conditions, something that cannot be obtained from chemical analysis. Table 163: concordance list of the different chemical classes and spectral groups class spectral group Period glass type I 1; 2 II + III HILM II I LIHM III 5; 6; 7 I LILM IV IV HIHM = HIMT 385 Transmission (%) Transmission (%) Spectral group 2 120 100 80 60 40 20 0 350 550 750 950 120 Spectral group 3 100 80 60 40 20 1150 0 350 550 Wavelength (nm) 750 950 1150 Wavelength (nm) 3 2.5 Spectral group 4 1.5 1 0.5 Spectral group 1 120 100 Transmission (%) Transmission (%) MnO (wt%) 2 type I 80 60 40 type II 20 0 350 550 750 950 1150 120 100 80 60 40 20 0 350 Wavelength (nm) 0 0 2 4 550 750 950 1150 Wavelength (nm) 6 8 10 12 14 Fe2O3 (wt%) Figure 205: Similar spectral shapes corresponding with different chemical compositions. The separate groups which are defined in the iron-manganese chart all correspond with a typical transmission spectrum o Can specific glass colours be matched to specific periods or typological groups? From Figures 192-193 and Table 160, discussed in the introduction of this chapter section, it is clear that the measuring of the integrated OD is not providing useful data concerning the dating of the glass, even though the OD for the Period II material is quite systematically higher compared to the Period I artefacts. Velzeke fragments Aa and Ab, as well as Da and Db, are showing close resemblance whereas Velzeke (no. 91 Dc) displays a deviation of about 0,5 compared to the two other samples, Velzeke (no. 91, Da and Db). On the other hand, we see that the seven Olbia samples (no. 41) show even more discrepancies. The first three have an elevated integrated optical density of about 2,5 or 3 (037-30a-c), whereas the other four balance between 1,5 and 2. 386 0.6 group2 group7a 0.5 group4a group7b y (CIE1931) 0.4 group6 0.3 group5 group1 group3 0.2 0.1 group4b 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 x (CIE1931) G2 G4a G7a G1 G7b G6 G5 G3 G4b Figure 206: Plot of the various spectral groups on the CIE1931 horseshoe graph In the first stage six separate groups were defined, featuring a specific iron and manganese content based on the specific glass hues perceived by the naked eye. Through UV-Vis-NIR spectroscopy, the transmission spectra of 96 black glass artefacts have been verified [see Appendix 3]. On the basis of the spectral shape, seven separate groups were defined (Figure 206). Each group consists of artefacts with a similar, specific spectral shape. Therefore we verified whether there is a possibility to obtain the same clustering via optical spectroscopy as from chemical analysis. Starting from the measured optical transmission spectra in the region between 300 and 1600 nm, we investigated different aspects:    the spectral shape and the link of specific spectral shapes towards the different defined subgroups the characteristic absorption bands in the measured spectra in order to estimate what colorants were used. Relying on the chemical analysis, only the two main colouring substances of the selected artefacts, i.e., iron and manganese, were taken into consideration. The absorption bands of these two components are described in literature (Brill 1988, 272-273) (Table 156). the tracing of the production conditions used to make the batch of glass appear black by calculating the colour coordinates and the spectral shapes. Certain furnace conditions applied during the production process result in certain colours due to a characteristic transmission. 387 These findings can be useful to archaeologists in drawing conclusions about chronological divisions in the production process of (black) glass during the Roman imperial period, or in existing knowledge of trade routes. Figure 206 demonstrates clearly that the range of colours in which Roman black glass samples could appearblue, green, brown and purpleis dependent on the addition of specific metal oxides. A first test was to check whether the results from absorption spectroscopy through the spectra within UV-Vis-NIR range corroborate the chemical categorization of the black glass. In view of a comprehensive approach on the matter, we have selected samples from these four different glass hues. These samples not only had to represent a broad chronological and geographical areaspread over four centuries and coming from sites in the northern and southern part of the western European Continent (Table 161) but also their chemical composition had to be known. Therefore, we selected some material we had analysed earlier in collaboration with Prof. Koen Janssens at MiTAC (UA) [see Chapter 8; Appendix 1]. Some of these samples have been the subject of recent research showing a clear compositional distinction between fragments dated before and after 150 AD (Van der Linden et al. 2009). Table 164: Overview of the examined samples region Period number of samples North-western European Continent Grobbendonk (BE); Kruishoutem (BE); Matagne-la-Petite (BE); Nismes (BE); Velzeke (BE) Oudenburg (BE); Tienen (BE) II II II III/IV II II/IV II 1 2 8 1 6 3 8 29 South-western European Continent Avenches 4-10 (CH) ; Avenches 1-3 (CH) ; Olbia (FR) ; Rome (IT) Braga (PT) I II I I IV 7 3 5 4 1 20 49 Figure 206 and Table 161 give an overview of the first set of 49 samples investigated. One set of 29 samples 301 comes from 7 Belgian sites and is more or less entirely from identical contemporaneous materiali.e., tableware Forms IIB.1 and IIB.4except for the sites of Nismes (a bracelet) and Oudenburg (three vessel fragments and a bracelet, but dated not earlier than 200 AD). A second part comes from various places in the wider southern European Continent. This set is more st th heterogeneous in timefrom the early 1 to the 4 century ADand in functionarchitectural decoration, counter, tableware and jewellery [see Appendix 3]. 301 Grobbendonk, Matagne-la-Petite, Nismes and Tienen are situated in the civitas tungrorum within the province Germania inferior, whereas Kruishoutem, Oudenburg and Velzeke are situated in the civitates menapiorum and nerviorum within the province Gallia Belgica. 388 Figure 207: Overview of the examined artefacts from a wide geographical provenance and produced over a long-lasting period o Can UV-Vis-NIR spectroscopy provide answers to technological aspects of black glass production? We have verified here how to identify the different furnace conditions in the production process of Roman black glass and the possibilities for the archaeological research on Roman glass. Important for this concern is to detect 2+ 3+ the presence of Fe and Fe . In traditional chemical analysis, it is customary to quote the entire amount of iron 302 oxide solely as Fe2O3, but not all the iron is in this state (Newton, Davison 1989). Seeing that UV-Vis-NIR spectroscopy can detect both iron oxides, it is worth checking the furnace conditions used in the production of black glass. Important information about the applied production process is provided through the colour of the artefact and the spectral shape in the visible part of the measured spectrum. The analysis of the spectral position of the transmission maxima not only shows the presence of iron [see part 10.2.]. It can also reveal information about the corresponding iron ion contents which are related to the used furnace conditions. 2+ 3+ Under oxidizing conditions, Fe can be transformed into Fe (Davison, Newton 2003). A high concentration of 3+ Fe results in a transmission peak around 480 nm. The reverse reaction takes place under reducing conditions. 2+ 2+ A high concentration of Fe gives rise to a transmission maximum around 490 nm. For lower Fe levels, this 2+ 3+ 3+ maximum is shifted towards 520 nm. If Fe and Fe ions are present together, Fe predominates when the transmission maximum is situated around 510 nm (Brill 1988) (Table 156). Under reducing furnace conditions and 2+ in the presence of sulphate ions, a large fraction of the iron is reduced to Fe , while sulphate ions can be 3+ 3+ 2reduced to sulphide ions. The sulphide can combine with the remaining Fe to form the ferro-sulphide Fe -S 2+ (Hruschka, 1987). The presence of this chromophore together with Fe ions results in transmission maxima (Tmax) between 555 nm (slightly reducing) and 585 nm (strongly reducing). Under even more reducing conditions, the ferro-sulphide complex remains and the maximum is positioned around 670 nm (Brill 1988). The higher the wavelength of the spectral position of the transmission maximum, the stronger the reduction conditions within the furnace, resulting in a colour shift from green over olive to amber (Table 158). The presence of the ferro-sulphide 2+ chromophore thus modifies the blue colour of Fe in the glass to green, olive-green or amber-brown, depending 2+ on its ratio in proportion to that of Fe (Schreurs, Brill 1984; Brill 1988). The glass turns olive-green to amber under maximum reduction conditions. Hence, the production conditions are partly translated into the hue of the 2+ 3+ glass artefact. Depending on the relative concentrations of Fe and Fe , an aqua bluish or greenish tint is obtained. In Figure 208, the transmission maximum of the 30 artefacts belonging to Spectral Group 1 are correlated to the calculated x (CIE1931) colour coordinates. The higher the spectral position of the transmission maximum, the more reduced the furnace conditions and the more amber-like the artefact would appear. So, optical analysis provides by means of absorption spectroscopy an easy method to identify the furnace conditions 302 This is mainly because the determination of a chemical composition in oxide form is not adequate enough. This is also the case for manganese, which is solely represented as MnO. 389 in which each artefact has been produced. This result gives the archaeologists an extra tool in tracing regional differences and chronological evolutions concerning the production process of black-appearing glass. 3 x (CIE1931) 0.55 2 0.50 1 0.45 0.40 e mor green 0.35 560 565 570  575 tions ondi c g cin redu olive 580  amber 585 590 Transmission maximum (nm) 590 green --> olive 585 580 575 2 3 more reducing conditions 1 570 565 560 e ek nch e e (s s ( am 63 Ve Ve ple /24 1 lz l ek zek G - 1) e e VD (s (s Av am am C) en ple ple c Av he B - A) en s (8 no . ch 8 es /65 7) (8 72 8 N is /6 9) Ve M me 501 lz a ek ta s ( -9) g A2 e (s ne am (s 824 Kr ui pl am ) sh e pl e D ou - M I) te m 3K 7 (K 2. G 1 M K90 RO ) at ag sa .78 M ne mpl at e ( ag sam A ) pl M ne e at ( ag sam G) n M Ve at e (s ple lz agn am A) ek e e (s ple (s a am J) M mp pl at e ag le F C ) M ne -V at ag (sa 73 ) M ne mpl at e ag (sa B ne mp ) (s le D am ) pl e E) 555 Ve lz Av Position transmission maximum (nm) Figure 208: Transmission maxima (Tmax) of Spectral Group 1 as a function of the x colour coordinate. The higher both values, the more reduced the furnace conditions Figure 209: Position of the transmission maxima in the samples of Group 2 Although all artefacts appear similarly coloured, two subtypes can be discerned by looking to the entire spectrum of Spectral Group 6 (Figures 202-203). Subgroup 1 has a much higher transmission peak between 600 and 750 nm, as can be seen from the spectral shape of the samples Avenches 8, Avenches 9, Avenches 10, and Olbia 2. These artefacts have a high transmission for yellow and red wavelengths (Figure 202). The presence of a transmission maximum in the region between 610 and 670 nm indicates that the production process of these artefacts is similar to the one applied to the artefacts of the first described spectral group. Here also, the ferrosulphide complex is formed. The difference lies in the furnace conditions. The artefacts belonging to spectral group 6a are formed under even stronger reducing furnace conditions, resulting in brown tinges. 390 reducing vs oxydising technique 1,0 0,9 0,8 YELLOW-ORANGE 0,7 strong reducing technique 0,6 Rome 1 MnO 0,5 Rome 2 BROWN Rome 3 0,4 oxydising technique Avenches 0,3 Olbia BLUE-GREEN + BLUE 0,2 0,1 0,0 0,0 1,0 2,0 3,0 Fe2O3 Figure 210: Measured transmission spectra of the artefacts belonging to Spectral Group 6 [= Class III]; left: subtype I; right: subtype II. The artefacts of subgroup b are most probably produced via an alternative production process under totally different furnace conditions (oxidising instead of reducing). From the iron-manganese chart, we know that for all artefacts both metals are present in moderate concentrations. Depending on the furnace conditions, the following reactions can take place: Fe 2+ + Mn 3+   (reducing furnace condition) Fe 3+ + Mn 2+ (4) (oxidizing furnace condition) 3+ 3+ From this reaction it can be seen that under fully oxidising conditions the ions Fe and Mn are formed. The iron contributes a brownish yellow colour and the manganese a purple colour, resulting in a brownish violet glass (Newton, Davison 1989). It is clear that artefacts with a so-called brown colour could be obtained via two totally different production processes. In Table 165, we give an overview of such artefacts together with the corresponding furnace conditions and colour coordinates. Figure 210 demonstrates that the number of obtainable colour hues under strong reducing furnace conditions is smaller than those obtained under oxidising furnace conditions. The colour coordinates indicate that the oxidising technique can lead to a wider range of hues than via the reducing technique. The colour of the artefacts of subgroup 6 and 7 is restricted to the yellowish-orange region, while the second production technique can result in hues ranging from yellowish-orange to reddish-orange hues (Figure 210). This is possibly due to the difficulty of regulating the furnace conditions resulting in the corresponding reaction, but further research on a larger group of artefacts is needed to verify this hypothesis. Table 165: Furnace conditions and colour coordinates of the brown coloured artefacts spectral sample furnace condition colour group x y CLASS 1 6 6 6 6b 7a 7a 7b 7b 7 Avenches8 Avenches9 Avenches10 Olbia2 Avenches4 Avenches5 Avenches6 Avenches7 Olbia1 strong reducing strong reducing strong reducing strong reducing oxidizing oxidizing oxidizing oxidizing oxidizing 0,55 0,55 0,52 0,56 0,60 0,60 0,49 0,54 0,59 0,44 0,44 0,46 0,44 0,39 0,37 0,47 0,46 0,41 391 The results presented here demonstrate that UV-Vis-NIR spectroscopy is an objective method to distinguish dark coloured glass from black-appearing glass. The calculation of the optical density clarifies why glass in a green hue is more easily considered as black (appearing) in comparison with glass in a blue, a brown or a purple hue [see in this Chapter section 9.4.1]. It has been demonstrated that useful information can be obtained about the chemical composition of the glass artefact by using UV-Vis-NIR spectroscopy, even if this remains limited to the iron and manganese oxides when considering the strongly-coloured samples [see in this Chapter section 9.4.2]. In addition, UV-Vis-NIR spectroscopy offers information on the applied production process and more particularly on the used furnace conditions during the glass processing to obtain black-appearing glass [see in this Chapter section 9.4.5]. 9.5 Use of photonics and its efficiency in relation to archaeological questions The obtained measurement results from UV-Vis-NIR spectroscopy provided information on the coloration and production process of Roman strongly-coloured glass, but to what extent can UV-Vis-NIR spectroscopy contribute to the research on archaeological glass in general? In a first case study, we will demonstrate how the use of UV-Vis-NIR spectroscopy can help the archaeologist in understanding the various excavated features containing glass artefacts. In a second case study, we will demonstrate how the use of UV-Vis-NIR spectroscopy can help improve the analysis process. The inexpensive method UV-Vis-NIR spectroscopy can facilitate a more convenient and purposive sample selection by means of a preliminary inquiry, selecting the most interesting pieces out of a large group of artefacts suitable for chemical analysis. o Case study 1: contribution to archaeological analysis The idea of a potential contribution to the archaeological analysis of UV-Vis-NIR spectroscopy arose when a set of measurements at various places on one single piece demonstrated to give identical measurements with a minimal deviation. As an example, we tested the intact mould-blown cylindrical vessel from Tongeren (BE) [cat.no.562] (Figure 211, 303 left). In Figure 211 (right) the averaged transmission spectrum obtained from measurements taken at ten different places at the surface of the object is shown. Analysis of the spectral shape and calculation of the colour coordinates (x=0.58 and y=0.41) classifies this object in Spectral Group 7a. Therefore, the expected chemical composition is that of Class III with a low manganese and low iron oxide concentration, below 1% and below 3%, respectively. This result causes confusion seeing that cylindrical vessels without base-rings, such as the one from Tongeren, are dated 150-170/80 AD (Cosyns et al. 2006a). So do we have to consider the beaker as a Class III product of Period I and consequently adapt the date range of this vessel type? Or do we have to make a connection with the low iron/low manganese content of the rod-formed solid base unguentaria from Elkab rd 304 [cat.no.868-871] which are dated in the first quarter of the 3 century AD? This would mean that the Tongeren beaker was either a finished product imported from the South-eastern Mediterranean or made in a European glass workshop using imported black appearing raw glass. A third option, although to us not very convincing 303 Seeing as the vessel is intact, UV-VIS-NIR spectroscopy can be valuable for estimating the iron and manganese concentrations and providing immediate information on the applied production process without the necessity of sampling. 304 It is very plausible that the technological knowledge to produce high iron black glass was not yet introduced in the north-western provinces around the mid 2nd century AD, but only after 170/80 AD. On the other hand, it is equally possible that this type of beaker has been produced in a Mediterranean workshop, or that when locally produced, the low manganese/low iron raw glass was imported into the north-western provinces. This is not an isolated case, as the Form IIB.6 vessel from Faulquemont (FR) (Cabart [cat.no.1073] shows a low manganese/low iron content as well. On the base of an intact piece from grave III of Esch (NL), which is dated 160-180 AD (van den Hurk 1973, 224-225, no.III.36) [cat.no.2881], this shape is dated to Period II. Conversely, an example of this vessel shape has been retrieved in Augst coming from a context dated between 70 to 150 AD (Rütti 1991, II, no.1186) [cat.no.3540]. It could thus very well be that the bulbous beaker from the rich barrow of Esch is to be regarded as an heirloom, and consequently the Tongeren cup and Faulquemont piece should then be dated in Period I. 392 120 Transmission (%) 100 80 60 40 20 0 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 Wavelength (nm) Figure 211: Roman glass vessel from Tongeren/Belgium (photograph by the author, by courtesy of PGRM, Tongeren) (left). Transmission spectrum of the object (right) When excavating various fragments of low frequency idiosyncratic shapes such as the black glass vessels in different layers or in different features, the spectral shape and the calculation of the colour coordinates make possible the characterization of the artefacts belonging to one single object or produced from the same batch. But are we able to extend these differences to distinct glass workshops? In Figure 212 we show the calculated colour st coordinates of 26 fragments from nine contexts in Avenches (CH), which are roughly dated from the 1 to the rd early 3 century AD (Periods I+II). Depending on the circumstances of discovery, we may assume nine different artefacts, but may we simply say that the nine contexts stand for nine objects in black glass? Based on the colorimetry, only four separate colour groups were determined, putting forward the use of four different recipes as an argument (Table 166). 63/2411 (1) 89/07170-46 (1) 88/6501-9 (3) 89/07179/3 (8) 88/06918-02 (3) 89/7851-09 (2) 88/06975-05 (1) 89/7851-13 (4) 89/07170-45 (3) 0.6 1. 2. 3. 0.5 y (CIE1931) 0.5 0.4 0.48 0.46 0.3 4. 0.44 0.2 0.42 0.1 0.4 0.49 0.5 0.51 0.52 0.53 0.54 0.55 0.56 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 x (CIE1931) Figure 212: Colour coordinates of different samples excavated at Avenches (top) Table 166: List of the four different colour groups from Avenches hue n inventory no. yellow-green amber brown 1 ex. 3 ex. 19 ex. pink 3 ex. 63/2411 88/6501-9 88/06918-02; 89/7851-13; 89/07170-45; 89/07170-46; 89/07179/3; 89/785109; 89/7851-13 88/06918-02; 88/06975-05; 393 The four groups are definitely made by using different production processes and are therefore at least made from four different batches. It is not excluded that one single glass workshop was responsible for all four assemblages. Can the use of UV-Vis-NIR spectroscopy provide a better fine-tuning of the analysis results and determine a more correct number of objects than from archaeological contexts or external features? Or can the colour coordinates and spectral shapes help by bringing together fragments from different contexts or by separating fragments within one context? The yellowish-green fragment (63/2411) forms a first object while the three amber fragments from 305 context 88/6501-9 show such similarity that we may consider them as being from one single object. The same can be said for the three fragments falling within a fourth zone [4], but interesting here is that the piece from context 88/06975-05 corresponds with two out of three fragments from context 88/06918-02. The third piece falls within the third colour group, including the bulk of the black glass material from Avenches (Figure 212). Consequently, we can assume that the three corresponding pieces from both contexts come from the same piece or a similar piece from the same batch, besides the fact that both contexts are related. Hence, the artefacts within colour groups 3 and 4 are to be regarded as contemporaneous. Figure 213 clearly illustrates that a more detailed analysis of the samples of the third colour group can be made by plotting the spectral shape of each piece‘s different production processes. Two distinct groups can clearly be identified: a set falls within Spectral Group 6 and a second group falls within Spectral Group 7a. Both fall within the Chemical Class III with a low iron and low manganese content. 120 Group 6: 89/07179/3 (8), 89/7851-09 (2), 89/7851-13 (4) 100 Transmission (%) Group 7a: 88/06918-02 (1), 89/07170-45 (3), 89/07170-46 (1) 80 60 40 20 0 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 Wavelength (nm) Figure 213: Spectra of the samples belonging to Spectral Group 6 and 7a [Class III] The identical spectral shape of the two fragments from contexts 88/06918-02 and 89/07170-46 in Spectral Group 306 7a are to be considered from one single artefact , making both contexts match. Consequently, all other artefact groups need to be considered together. The three fragments from context 89/07170-45 are somewhat different compared to the first two, although also belonging to Spectral Group 7a. Their spectral shapes clearly show to be equal, and are therefore to be considered as being from one single object. Looking back to the calculated colour coordinates, we see a difference of 0,02 on the x-coordinate and 0,04 on the y-coordinate between these three samples (Figure 214). Striking is the matching spectral shape of the Tongeren cup with those of the first two fragments within Spectral Group 7a of Avenches. Hence, the Tongeren cup and the two Avenches fragments most likely come from one single glass workshop. It is quite speculative to attempt to locate the glass workshop in Avenches, even though some arguments may plead for this very tempting idea [see Chapter 9]. 305 If external features would have been recognised, the three fragments should then be considered to be from different objects produced with glass from the same batch. 306 Seeing the rarity of the black glass vessel fragments, we may assume that both fragments from the same batch are consequently from one single object. This reasoning cannot be done for the more ubiquitous ‗naturally coloured‘ and decolourized vessels, and similar results only correspond to the level of the batch. 394 The spectral shapes of the twelve fragments from contexts 89/07179/3 (8 ex.) and 89/7851-13 (4 ex.) within Spectral Group 6 are identical and can therefore also be considered to be from one single object or from more artefacts from one single batch. This implies that other artefact groups can be evaluated as one. Both fragments from context 89/7851-09 fit, but differ from both previous contexts within Spectral Group 6. We assume that these two fragments come from an object made from a different batch but manufactured in the same workshop using the same recipe. From Figure 214, we may conclude that the 19 fragments from 6 different contexts refer to an MNI of 4 artefacts from 2 batches. We can conclude that UV-Vis-NIR spectroscopy can link material from a particular site to a specific production process, and subsequently these artefacts which were produced under similar production conditions can be connected to one and the same workshop, even if the workshop is not yet located. Conversely, UV-Vis-NIR spectroscopy on archaeological glass artefacts definitely can help in the set-up of the Harris-matrix of an excavation by identifying the fragments belonging to one single artefact. The results from the Avenches material can be confirmed with the measurements made on the ten samples excavated at Tienen-Grijpenveld (BE) (Figure 214-215). From an archaeological point of view, we may speak of an MNI of 2 when taking into consideration only the rim and base fragments. However, on the basis of the contexts, this might raise the MNI to a minimum of 3 and maximum of 5 (Figure 216). On the basis of the chemical composition, these samples have to be categorized within the high iron/low manganese Class I, and based on the transmission spectra, they belong to Spectral Group 1. In Figure 215, we show the calculated colour coordinates of these samples. From the measurements repeatedly taken on different areas, we may conclude that the artefacts were rather uniformly coloured. The maximum colour difference measured at different positions on the same artefact equalled 0.06. Based on this information, we could identify the artefacts which certainly do not belong to a same object, as their colour differences are too high, such as the pieces from context TI/02/TR/20 in relation to those from TI/98/TR/03 (Figure 215). The difference in colour coordinates of the samples belonging to TI/02/TR/20 and TI/01/TR/79 groups, on the contrary, are within the 0.06 range. Consequently, from an optical point of view, the fragments from these two contexts could potentially belong to the same object or to two or more different vessels made from one single glass batch. Seeing that the rim fragment of the tumulus carchesium (TI/02/TR/20) is intact, the rim fragment from context TI/01/TR/79 is to be considered from the same batch. We also know that the samples of Spectral Group 2 have a transmission maximum between 564 nm and 588 nm. The position of this maximum depends on the furnace conditions as described in part 10.4.5. Accordingly, this maximum is a potential selection parameter for artefacts produced under the same furnace conditions. TI/98/TR/03/022 TI/02/TR/20/262 TI/98/TR/03/026 TI/01/TR/79/059 TI/02/TR/20/078 TI/01/TR/79/142 TI/02/TR/20/079 TI/02/TR/20/80 0.56 0.54 0.52 y (CIE1931) Tmax = 570 nm 0.50 Tmax = 580 nm 0.48 Tmax = 578 nm 0.46 Tmax = 588 nm 0.44 0.42 0.40 0.36 0.38 0.40 0.42 0.44 0.46 0.48 0.50 0.52 0.54 0.56 x (CIE1931) Figure 214: Grouping of the Tienen artefacts based on their colour coordinates and position of the transmission 395 TI/98/TR/03/022 TI/98/TR/03/026 TI/01/TR/79/059 TI/01/TR/79/142 TI/02/TR/20/080 TI/02/TR/20/079 TI/02/TR/20/078 TI/02/TR/20/262 Figure 215: Maximum (left) and on external examination (right) Based on the different transmission peaks, we identified four clusters for the eight measured samples, subgroups with a maximum transmission respectively at 570 nm (samples group TI/02/TR/20), 578 nm (samples group TI/01/TR/79), 580 nm (sample TI/98/TR/03/22), and 588 nm (sample TI/98/TR/03/26). The eight artefacts belong to a minimum of four different objects made from two different batches. This approach adds extra information to the conclusions which were drawn after excavation, when the artefacts were submitted to a first examination based on their external properties. It was concluded that the artefacts of group TI/02/TR/20 belong to the same object (Figure 214). Another set comprises ten vessel fragments of Period II from the excavations at Steenakker/Kapelleveld in Kontich (BE). The set can be subdivided into five subcategories on the basis of their transmission maxima, correspondingly 574.5 nm; 577.5 nm; 579 nm; 583.5 nm; 585 nm (Figure 216). From the analysis results, it was established that the deviation on the transmission maximum of one artefact is negligible. Therefore, we can assign these samples to at least five different glass vessels. 396 KFL-1429 (sample1) K87-542 (sample5) K87-542 (sample9) KFL-1429 (sample2) K87-542 (sample6) KFL-6603 KFL-1429 (sample3) K87-542 (sample7) sample 4 K87-542 (sample8) 0.52 y (CIE1931) Tmax = 574.5 nm 0.50 Tmax = 577.5 nm Tmax = 585 nm 0.48 Tmax = 579 nm Tmax = 583.5 nm 0.46 0.42 0.44 0.46 0.48 0.50 0.52 0.54 x (CIE1931) Figure 216: Grouping of the Kontich artefacts based on their colour coordinates and position of the transmission maximum The use of UV-Vis-NIR spectroscopy makes it possible to recognize the characteristic spectral shapes dependent on the chemical composition of the sample, and thus makes it possible to determine the fragments made from the same recipe. Additionally, the calculation of the colour coordinates x and y makes it possible to detect differences on the level of the batch. We are, however, well aware that we only took into consideration the very low frequency idiosyncratic Roman black glass vessels of Period II. It is therefore necessary to verify the analysis results and the conclusions ensuing from it with the more ubiquitous ‗naturally coloured‘ and decolourized material, developing an approach for a much more reliable number of materials from a specific excavation. o Case study 2: contribution to archaeometric analysis The results of UV-Vis-NIR spectroscopic analysis on strongly-coloured glass artefacts make it possible to detect a number of colouring elements. The chemical concentrations can thus be estimated without damaging the object. Perhaps the use of UV-Vis-NIR spectroscopy on glass barely produces a surplus value to chemical analysis methods thus far, but it can counter the restrictions of these rather complex, expensive, and not easily granted procedures. The consequence is that investigations on the chemical composition of archaeological glass artefacts remain limited, even if during the last 15 years chemical analysis has increased in quality and in quantity. The analysed sets, however, always are limited and have thus to be selected carefully in view of research queries. This means that the selected amount of samples remains rather undersized vis-à-vis the quantity of excavated glass artefacts. When a lot of glass artefacts are excavated at the same site, the objects are first grouped on the basis of their external properties. Most often the different groups contain several samples which are similar in composition. Hence, an individual selection results too easily in a number of unnecessary samples or opportunities that go by unnoticed. Clear groups can be recognized on the basis of the spectral shapes which are due to the specific chemical compositions of the artefacts. Using UV-Vis-NIR spectroscopy, all (diagnostic) material can be checked quickly, making an overall but well-defined selection of the available study material possible. Eight vessel fragments from the villa of Matagne-la-Petite (BE) and another eight from the south-western area of the vicus Tienen (BE) are considered in this case-study. The external factors of the fragments allow for the assumption of at least four vessels from Matagne and two from Tienen. When looking to the chemical analysis results from all sampled black glass vessels of the villa at Matagne-la-Petite (BE) and the south-western area of the vicus Tienen (BE) [see Chapter 9], the homogeneity of the chemical composition allows the hypothesis of one single batch. When verifying the composition of the samples by means of chemical analysis, generally only one artefact would have been taken when combined with the easy and non-destructive absorption spectroscopy. This optical analysis method on all 16 pieces from both sites has demonstrated that all samples have a similar spectral shape and so also possess a similar colour. We can conclude that the analysed artefacts from Matagne-la-Petite and Tienen must have a similar composition. 397 One example is the clustering of artefacts in the MnO ratio versus the Fe 2O3 ratio diagram. The artefacts we considered were strongly-coloured glass artefacts appearing black from the Roman period. Depending on the fabrication period and place, these samples contain a different MnO and Fe 2O3 percentage. The ―low iron‖ group nd is traditionally produced in the Mediterranean region. At the beginning of the 2 century AD, a characteristic production technique was developed in the northwestern region of the Roman Empire where uncoloured glass was imported from the south and locally coloured by adding iron (Cosyns et al. 2006a, 40-41; Van der Linden et al. 2009). We concluded that the same information could be extracted from the x, y colour diagram, which shows that for this type of research, absorption spectroscopy could be used as a low-cost alternative for the expensive X-raybased techniques. The idea consists of drawing a parallel between the specific regions in the MnO – Fe2O3 diagram and the different regions of the CIE1931 curve. In Figure 217 we illustrate this on eight samples excavated at Matagne-la-Petite. Three of them are located in the ―yellow‖ region, and five others in the ―yellowish-orange‖ region. For reasons of clarity, only a limited number of samples are shown here. 0,90 Avenches Matagne CIE1931 curve Olbia Tienen Velzeke Oudenburg Grobbendonk Kruishoutem Nismes 0,80 0,70 0,60 y 0,50 0,40 0,30 0,20 0,10 0,00 0,00 0,10 0,20 0,30 0,40 x 0,50 0,60 0,70 Figure 217: Mapping of a number of samples on the CIE1931 colour diagram It is clear that glass production in Roman times took place under very precise and controlled conditions. A determination of the chemical composition in oxide form for this reason provides only limited information relevant to the applied production process. A traditional chemical analysis of a glass sample, for instance, only quotes the amount of iron oxide as Fe2O3, while not all the iron within the glass matrix is in this state (Newton, Davison 1989). The use of UV-Vis-NIR spectroscopy can therefore prove interesting to explore the possibilities and the benefits of using this technique. 398 0.6 group2 group7a 0.5 group4a group7b y (CIE1931) 0.4 group6 0.3 group5 group1 group3 0.2 0.1 group4b 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 x (CIE1931) Figure 218: Calculated colour coordinates of the different spectral groups recognized on Roman black glass G2 G4a G7a G1 G7b G6 G5 G3 G4b Figure 219: Plot on the CIE1931 horseshoe graph of the calculated colour coordinates of the different spectral groups recognized on Roman black glass Seeing the correlation between the chemical composition and optical properties with the archaeologically dated material, it is possible to put forward that each class and spectral group implies chronological associations (Figures 218-219). 9.6 Conclusion The optical spectroscopy of the Roman black-appearing monochrome strongly-coloured glass artefacts demonstrates that UV-Vis-NIR spectroscopy is a very straightforward, easy, non-destructive and cheap technique which is complementary to the chemical analysis methods and, to a certain extent, workable as an alternative to the accepted standard analysis methods (Meulebroeck et al. 2008; Meulebroeck et al. 2011). Analysis of the 399 spectral shape tells if the glass was processed under reducing or under oxidizing furnace conditions. The identification of characteristic absorption bands gives an idea about the used colouring agents, whereas the calculated colour coordinates indicate the concentration of the colouring agents. Thus, besides information on the used colouring agents, a specificity that is generally obtained from chemical analysis, UV-Vis-NIR spectroscopy shows that the measured transmission spectra also provide information on the applied furnace conditions. The objective measuring through optical analysis, providing a distinction between dark coloured versus so-called black glass, clarifies why glass in a green hue is regarded more easily as black (appearing) in comparison with glass in a blue, a brown, or a purple hue. The application of UV-Vis-NIR spectroscopy on Roman strongly-coloured glass artefacts appearing black illustrates well the benefit of photonics to characterize various technological aspects employed in the production of the glass artefacts. Optical spectroscopy has the potential to act as an alternative method to the ion- and X-raybased chemical analysis techniques or at least as an additional method in order to provide additional information to approach the archaeological question. Earlier research has shown that a clear compositional distinction could be observed between fragments dated to the period before and after 150 AD. Different groups with a well-defined iron and manganese concentration could be identified. These conclusions were based on traditional chemical analysis results. Our research proved that these separate groups which are defined in the iron-manganese chart all correspond with a typical transmission spectrum. This means that the same conclusions concerning chronological differences in black-appearing glass production can be drawn from the spectroscopic measurement results. The identification of characteristic absorption bands of the artefact‘s transmission spectrum allowed us to define the applied colouring agents. The calculated colour coordinates indicate the concentration of the colouring agents. The high-iron group corresponds with a yellow-green hue. The characteristic absorption bands of iron are visible in the spectrum. For the high-manganese group, we distinguished two groups based on the spectral information. A first group is clearly coloured by manganese (pink-purple colour), the second group contains a mixture of iron, manganese, copper and cobalt as colouring agents. The remaining group with the moderate iron and manganese concentration results in brown coloured artefacts. Concerning the absolute concentration of an artefact, we have to remark that although optical spectroscopy is a semi-quantitative method, not able to give an exact value, it can be used in a first step to classify a large group of artefacts in separate classes with a similar spectral behaviour. In this way, only a restricted number of samples need to be analysed through these complex, high-cost, and destructive ion- and X-ray-based techniques. The recent introduction of the low-cost portable optical spectrum analysers makes this measurement method even more feasible. Furthermore, the analysis of the spectral shape of the transmission spectrum reveals information about an artefact‘s production conditions. This information could not be derived from standard chemical analysis results. We learned that the green glass was produced under reducing furnace conditions. The transmission spectrum is characterised by a transmission maximum in the green part of the electromagnetic spectrum. However, the more reduced the furnace conditions are, the higher the spectral position of the transmission maximum. Furthermore, the hue of the glass matrix will turn orange. The artefacts coloured with manganese required an oxidizing furnace environment. This was also the case for some (type II) of the brown coloured samples. The second type of brown artefacts, on the other hand, was produced under reducing conditions. The calculated colour coordinates of the different groups were compared and two main conclusions could be drawn. For all the groups, the calculated colours correspond with the colours we expected. If we link this to the six groups we defined in the iron-manganese diagram, we conclude that every group has a specific colour: o o o o o o Group I [= group 1] = very pale aqua blue-green colours Group II [= group 2] = green, olive and amber colours Group III [= group 3] = purple colours Group IV [= group 4b] = blue colours (group 4a: aqua bluish or greenish colours) Group V [= groups 6; 7] = brown colours Group VI [= group 5] = brown-red colours The paler the colour perceived by human eyes, the smaller the concentrations of the colorants. Brown colours need much lower concentrations than, for example, pale blue glass. All this lets us conclude that starting from the 400 calculated colour coordinates, it is possible to make a first estimation of the iron and manganese concentrations. Secondly, we learned that starting from the colour or from the chemical analysis data, it is not possible to decide about the applied production process. As we explained earlier, brown colours, for example, can be obtained via several different ways, where every single production condition is linked to a specific spectral shape of the spectrum. Analysis of the spectral shape is therefore necessary if we want to get knowledge of the applied production method. This information will be useful to archaeologists in different types of applications including the ones discussed in part B and in identifying workshops in a certain region or in identifying possible knowledge transfer between workshops in different regions. The conclusions drawn in this study are valid for this specific type of Roman glass, but can be generalized for other types of glass with different chronological and geographical backgrounds. A first general conclusion of our research is related to the colour of the artefacts. Up till now archaeologists have used colour information during their research, but these colour adjudications have been carried out in a very subjective manner by just physically looking at the object. We concluded that a quantitative determination of the colour of the artefact can be beneficial in solving several types of archaeological problems. A next step should be the calculation of the second derivative spectra (Hruschka 1987) to allow fine-tuning the results and to identify the characteristic absorption bands for different chemical elements. Another technique that could not be verified within the present study is the application of the principal component analysis (PCA) to the spectra (Martens, Naes 1989), which will allow the classification of the artefacts based on a difference in composition. Our research proved that PCA can act as a tool to identify samples belonging to the same production batch. 401 402 PART III THE HISTORICAL APPROACH 403 Chapter 10 HISTORICAL ANALYSIS In the 1950s it was customary to assert that archaeologists could be properly concerned with the how, when and where of the past but that the question ‗why‘ generally lay beyond the limits of archaeological inference. However, a subject relegated to a mere descriptive role will rapidly become sterile, practised by those to whom the catalogue or corpus is an end in itself rather than a means of further enlightenment. (Peacock 1982, xi) Personally it feels hubristic to attempt to discuss the whole. The risk of giving offense, whether by omission or injudicious comment, is enormous. (Woolf 2004, 417) 10.1 Introduction The challenge is to demonstrate that an integrated view of a minor segment within the Roman material culture can provide pattern(s) to help further develop the understanding of the Roman imperial economy in general and that of the glass production and consumption in particular. In the preceding chapters, we discussed one by one a series of different aspects of the production, distribution and consumption of Roman (black) glass built up from the analysis of the artefacts. Apart from the great value of the prior conclusions obtained from each approach separately, it is definitely most important to survey the archaeological and archaeometric data jointly in view of a historical approach. The aim of this chapter is thus not to recapitulate the results from all previous chapters, but rather to combine the established observations with historical information in order to come to a holistic approach. Hence, we confront in this chapter the results from the archaeological and archaeometric approaches with the available information from historical sources and models. Due to the lack of ancient written sources on (black) glass production and consumption, it is compulsory to compare the acquired archaeological and archaeometric data with hypothetical models current within Roman historical research. Conversely, we are well aware of the difficulty of such undertaking and the limitations of the possible outcome, seeing the high degree of specialized research within the different approaches of Roman archaeology that led to a discipline that became so highly fragmented that no individual can master it anymore (Woolf 2004, 417). The specificity of the heuristic survey to acquire knowledge on the classic as well as the current models was experienced as a first obstacle within this attempt. In the face of the overwhelming multiplicity of specialities within the Roman historical research discoursepolitical history, military history, social history (with the booming gender studies, not in the least in view of the military theme), cultural history (including ritual, customs and belief), economic history, technology historywe decided to focus on economic history. We believe this choice relevant because the scope of our research is to characterize the production and consumption of idiosyncratic commodities during specific periods throughout the Roman imperial period. Although the glass production and consumption is an economic issue, we are well aware of the complexity of interwoven involvement of various factors on different levelse.g. the impact of the socio-cultural development through the centuries and within the different areas of the empire; the impact of regional consumers‘ behaviour and traditions on the (d)/(r)evolutions on the microeconomic level; the technological changes and innovations instigated by mesoeconomic pressure; and the macroeconomic changes influenced by or resulting from the turbulent political and military events during the Roman imperial period. In an attempt to incorporate at best the production and consumption of the Roman (black) glass in the Roman society, we confront the here-compiled data and the generated rules with the generally proposed models on Roman economy. Nevertheless, some socio-cultural aspects have already been paid attention to, discussing the use and function of the commodities in black-appearing glass [see Chapter 7]; some technological aspects have been observed in view of the manufacturing of the artefacts through the consecutive periods [see Chapters 2 and 3] as well as in view of the production of glass [see Chapters 8 and 9]. Consequently, we confront in this chapter the data from both approaches within an historical perspective by tackling particular issues to improve our understanding of the Roman glass production, distribution and consumption within a proposed model. The archaeological and archaeometric data related to the historical approach illustrate the presented Roman black glass from different angles within the Roman socio-economic framework. But to what extent can the obtained results on Roman black glass fit the conventional models on Roman economy or that on Classical Antiquity in general? 404 To demonstrate the value of a holistic approach and its representativeness, we verified the production and consumption of Roman black glass from different angles by approaching the compiled data on three successive levels in economics.    The macroeconomic level integrates the impact of external factorslike, for instance, political decisions concerning the regulation of taxation and tax collection; military actions such as border wars, civil wars, or army restructuring measuresthat influenced indirectly the glass production by resulting in economic crisis or prosperity. Hence, from this level we compare the black glass production with the entire Roman material culture and draw parallels with the prevailing models of the Roman economy. The mesoeconomic level involves the organization and evolutions the production of (black) glass underwent and the unavoidable technological innovations necessary for the glass workshops to survive workshops producing similar objects in other materials and compete with other, less innovative glass workshops. The mesoeconomic level thus evaluates the Roman economy on a (inter)regional level because competition and the inherent innovations are a purely regional matter. The microeconomic level incorporates the data on the local level that influenced or involved the glass workshop directly or a very specific aspect within the chaîne opératoire of black glass artefacts. This three-levelled approach is particularly common in economics, where they speak of micro-, meso- and macroeconomic factors, but it is also well embedded within psychological and sociological research, where all three echelons involve explicit questions to examine the assertions put forward. Such a multi-levelled querying is barely used in archaeological research, in particular when considering aspects within the study of the material culture such as Roman glass. Through this approach, the compiled data can be viewed from different angles, thus providing a much wider perspective of the material studies such as Roman (black) glass and surpassing the diagnosis of bits and pieces from one period or from one site or a region. For the benefit of readability, we discuss the black glass in relation to the proposed query per level in a separate case study. We confront the accumulated data on Roman black glass consumption (micro-level) with the prevailing general views on the Roman glass production and consumption (meso-level). Both categories are subsequently considered vis-à-vis the Roman economy (macro-level) (Figure 220). Historical approach Macro-level Meso-level Roman economy Roman (black) glass production Micro-level Roman black glass consumption Figure 220: Schematic visualization of the three-levelled economic approaches Because no fine breakdown by subject can be formulated for the three interacting and interrelated levels, we merge in the conclusion the deductions from each level described in the different case-studies. This broader, multi-levelled perspective assesses Roman black glass consumption by analysing the relationship of the production and distribution of black glass artefacts to the overall glass production and distribution and to the Roman economy. Despite the complexity of the issue, we do no more than only presenting a brief summary on the different recent reliable models. Besides, we are aware of the difficulty and risk that is involved when attempting to implement the archaeological data into historical models in particular because of our limited knowledge on the matter as 307 archaeologist. 307 Illustrative is the metaphor by Michel Polfer (2001, 7): ―Vouloir étudier le rôle que l‘artisanat a pu jouer dans la vie économique de l‘Empire romain revient à effectuer une ballade en terrain mine.‖ 405 10.2 Case study I: Macro-level In this case study we demonstrate the role of political decisions and economic regulations affecting the production, distribution and consumption of the entire glass assemblage within a well-defined period. Therefore, we use the (inter)regional consumption pattern of black glass vessels in Period II because it emerges as a perfect indicator for macroeconomic interference by Rome indirectly influencing the regionalism of the Roman material culture nd rd during the later 2 and early 3 century AD. In prior chapters, we demonstrated that during the period 150-250 AD the produced vessels were essentially freeblown shapes which for the most part involved tableware used for drinking or serving drinks [see Chapter 3]. The majority of these shapes are unique or seldom attested in black glass, but the high carinated beakers (Form IIB.1) and the globular cups and jars (Forms IIB.4/5) were produced in relatively elevated numbers with a distribution restricted to the north-western provinces Gallia Belgica, Germania Inferior and Germania Superior [see Chapter 6]. Furthermore, the distribution pattern of the carinated beakers demonstrates an eastwest circulation in Gallia Belgica and Germania Inferior and appears absent in the upper Rhine region, which corresponds with the province Germania Superior. The globular beaker, on the other hand, shows a southnorth distribution pattern along the Rhine from Augst (CH) to Nijmegen (NL), while it is absent in the west, corresponding to the province Gallia Belgica [or at least the western part coinciding with the later Belgica secunda] (Figure 221). Figure 221: Distribution map of vessel Forms IIB.1 and 4/5;glass workshops where black glass was processed This well-defined distribution is evidence for the black glass vessel shapes as regional consumer goods produced somewhere in these provinces. Hence, the vessel shape Form IIB.1 must have been produced in one or more workshops in Gallia Belgica and/or Germania Inferior, while the workshop(s) responsible for the production of the vessel shapes Forms IIB.4/5 is/are to be situated in Germania Inferior and/or Germania Superior. From the information at hand, we locate three (possible) glass workshops where black glass was processed to manufacture vesselsForm IIB.1: Liberchies (BE)?; Sainte Menehould – Les Houis (FR)?; Forms IIB.4/5: Kaiseraugst (CH) [see Chapter 6] (Figure 221). Careful examination of the technological aspects of all Form IIB.1 fragments shows various features that are evidence for the various technical movements adopted by the glassworker when shaping the vessel and might be helpful in determining the fingerprint of a glass workshop: 1) rounded rim or an inward-folded rim; 2) a massive pontil scar (d.:15-20 mm) or a thin annular-shaped mark (d.:25-30 mm); 3) large annular foot or a small annular foot (jagged in an early or later stage of the blowing process) (d.:20-50 mm); 4) a cylindrical to conical-shaped body. At first glance, these features seem not systematically present on specific vessels, but this is probably due to the fact that we were not able to execute a detailed analysis of this approach. Consequently, no definite 406 distribution pattern could be discerned from all these elements. Because we believe that such technological approach is very helpful in understanding the ability of a glassworker and for recognizing the fingerprint of the various glassworkers producing the same vessel shapes, we consider it valuable to explore this issue more in depth. The chemical composition of the black glass vessels reveals that both vessel shapes fall within the high iron group characteristic from Period II onwards. Two distinct subclasses can be observed in this group: one with a titanium content below 0,10 % and another with a titanium content ranging between 0,10-0,17 % (Figure 222). 308 This difference in titanium ratios is, like that of alumina oxide , related to the quality of the sand used. The two ratio clusters of impurities indicate that two different types of raw glass were processed to produce black glass vessels. Logical reasoning results in the assumption of two glass workshops, separately importing raw glass from a different primary glass workshop. 18 16 14 12 10 Fe2O3 Roman black glass vessels 8 6 4 Lo Fe 2 0 0 0,1 0,2 0,3 0,4 0,5 TiO2 Figure 222: Bi-plot discriminating two subclasses according to the titanium content (wt %) Figure 223 shows that the two different raw glasses applied are not linked with one of the two vessel shapes; nonetheless, almost all sampled globular cups and jars (Isings Form 94) fall within the low titanium subclass. Because this shape is limited to the provinces Germania Inferior and Germania Superior, our next step was to verify whether both subclasses are correlated to the production area, meaning that the carchesia samples with low titanium content also originate from one or more glass workshops in the provinces Germania Inferior and Germania Superior. 308 The alumina content, which is also related to the used sand, shows an alumina content above 3,50 % for the subclass with more elevated titanium content, while low titanium glass has a low alumina content below 2,50 %. 407 0,70 0,60 0,20-0,25% 0,10-0,16% 0,50 0,40 MnO Isings 36b 0-0,07% Isings 94 0,30 0,20 0,10 0,00 0,00 0,05 0,10 0,15 0,20 0,25 0,30 TiO2 Figure 223: Bi-plot of the titanium – manganese ratio (wt %) of the vessel shapes Form IIB.1 aand Forms IIB.4/5 Figure 224 provides evidence for what was assumed from Figure 223. The titanium content encompasses a regionally bonded differentiation showing that low titanium glass is characteristic for the black glass vessel production in the provinces Germania Inferior and Germania Superiori.e. Augst (CH); Avenches (CH); Tienen (BE); Matagne-la-Petite (BE); Rumst (BE)whereas the more elevated titanium content is clearly restricted to 309 sites within the province Gallia Belgicai.e. Velzeke (BE); Liberchies (BE); Kontich (BE). The two pieces from Colchester (UK) that were analysed proved to be imports from Gallia Belgica and not from Germania Inferior. 14,00 AUGST 12,00 AVENCHES 10,00 COLCHESTER TIENEN 8,00 Fe2O3 MATAGNE 6,00 VELZEKE 4,00 RUMST LIBERCHIES 2,00 0,00 0,00 KONTICH 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40 TiO2 Figure 224: Bi-plot of the titanium – manganese ratio (wt %) of the vessel shapes Form IIB.1 aand Forms IIB.4/5 Also interesting to notice is that both titanium subclasses are present within the low iron group as well, alongside a third subclass with even more elevated titanium content ranging between 0,20-0,25 % (Figure 224dashed ellipses). The ensuing stage of the reasoning is to explain that workshops in each area used a different raw glass to produce black glass vessels independently. The production in both regions of an assortment of shapes in black 309 All three Belgian sites (Tienen, Matagne-la-Petite, and Rumst) were part of the civitas tungrorum, which is usually supposed to have been part of Germania Inferior, whereas Velzeke, Liberchies, and Kontich were settlements in the civitas nerviorum, which is generally assumed to have been part of Gallia Belgica. 408 glass and in particular the carinated beakers (Form IIB.1) points to the use of similar consumption goods exceeding the borders of a province, indicating common drinking traditions in a wider area. The homogeneity of the material culture in the three north-western provinces contrasts with that of the neighbouring provinces. This correspondence of material culture in different provinces is apparently the result of culturally determined consumer behaviour due to traditions, ritual and regional ethnicity, and is in no way considered connected with the political organization of the Roman Empire (Woolf 2004). The surprise was big when a distribution map of all material from both titanium subclasses did not visualize the expected random picture but a neat distribution pattern corresponding with either the province Gallia Belgica or the provinces Germania Inferior and Superior (Figure 225). Figure 225: Distribution map of the low titanium subclass (green) and high titanium subclass (red) The relationship between the archaeological data and the chemical analysis not only showed that the two recipes used to produce black glass vessels have regional variations, but also that those variations are connected with (at least two) different workshops, which used their personal recipes to produce similar vessel shapes. Besides this geographically-bounded production per province, the abovementioned analysis on the regional idiosyncratic consumption of black glass vessels during Period II identified also a province-bounded distribution. Widening the focus to the whole range of Roman glass vessels of these three provinces is needed to verify this conclusive outcome and would provide many new opportunities in the appraisal of regional glass production and the complex consumption patterns in order to better understand the (d/r)evolutions of the Roman imperial material culture. The final step should be to describe the cause(s) leading the regional economy to a strictly provincial production, distribution and consumption market, but without being aware of all current discussions and issues in the field, it would be too inconsiderate to enter into a penetrating debate. The complexity of such discussions means that we only can offer modest remarks on the matter. We believe that the observation put forward results from a consumption tax imposed by Rome on consumer goods when crossing provincial borders. The important work by Siegfried De Laet (1949) on the portorium, indirect taxation on the circulation of goods, demonstrates that the Roman Empire built up a complex administration to collect custom charges on trade: 1) quadragesima, a fortieth part or 2,5% levied on goods entering a large region (but possibly also internally seeing the occurrence of internal stations); 2) toll paid to be authorized to use the main road, a mountain pass, or cross a watercourse; and 3) octroi, a duty levied on various goods entering a town or city. The corroboration of the distribution pattern of both chemical groups with the provincial limits between, for instance, Gallia Belgica and Germania Inferior allows us to assume that the tax on trade turned out to a significant price increase of the possibly already costly black glass vessels, so that it was any longer profitable to attend markets beyond the border. But one can wonder whether 409 the keen distribution within provincial borders resulted coincidentally from the formulated tax regulations or whether Rome wanted to stimulate local production and trade all over the empire. The presence of at least one glass workshop in either province during Period II where the same vessel shapes were manufactured in black glass speaks in favour of the hypothesis to promote local consumption. Probably the smuggling of such vessels across the province border got discouraged above all by the loss of benefit as a result of 1) increasing control for the remuneration of the taxes on (long-distance) road-transport and 2) the severe retribution for those aiming to avoid the declaration of trading goods scared off to commit fraud. Siegfried De Laet (1949, 427) speaks of commissum or the confiscation of fraudulent goods as sentence for fraud. Despite our limited knowledge on the political economic history of imperial Rome, we believe that the economic restructuring measures by Rome under Marcus Aurelius and Commodus are to be considered curative interventions to solve an existing crisis rather than preventive interventions in anticipation of an upcoming crisis. Most likely, the reorganisation by Marcus Aurelius was intended to maintain the tax revenue from the provinces so as to preserve the cash flow to finance the rising army expenses in a period with numerous incursions at the frontier, in particular at the Rhine and Danube Limes; a dramatic population implosion due to the Black Death; 310 increasing inflation; and incessant devaluation. According to Siegfried De Laet (1949, 404), the reorganisation of tax collection from the private administered socii by conductors into a state-controlled administration under a procurator got introduced gradually and was only completed under Commodus. It is clear that we need to verify the degree of influence from centralised decisions by the emperor and the Senate in Rome by checking the corroboration with the so-called black glass shapes in ‗naturally coloured‘ and decolourised glass from neighbouring provinces, as well as with the contemporaneous artefacts in ‗naturally coloured‘ and decolourised glass idiosyncratic to the north-western provinces, like, for instance, the snake-thread decorated vessels, the grape-shaped amphoriskoi, or the Mercury bottles. This case study produced an exponential increase of knowledge and understanding in Roman glass research on production and consumption by demonstrating a strong impact of the political organisation of the empire during Period II on the Roman economy, culminating in a heterogeneous and complex material culture. This case study corroborates Greg Woolf‘s idea of a Roman imperial economy consisting of distinct, loosely regional economies but contradicts the ‗loosely and sluggishly integrated‘ Roman economy alongside the economic politics (Woolf 1992). 10.3 Case Study II: Meso-level In this case study, we verify whether an approach on the mesoeconomic level of Roman black glass can contribute to the understanding of the character of the Roman economy. The underlying idea of this case study is to bear out whether the Roman black glass production and consumption corroborates a pre-industrial Roman economy or one with a proto-industrial character. We therefore check to what degree the technical innovations and/or trend changes recognised in black glass assemblages affected the production of Roman glass workshops in a wider region or, in other words, the Roman glass industry in general. A related issue is assessing the correlation or variation between the existing manufacturing techniques and the prevailing styles of commodity shapes in glass within the different regions of the Roman Empire. No real consensus exists hitherto on how to evaluate the Roman economy, but different models have been proposed. For about a century, the field was dominated by two major factions (Pleket 1998; Polfer 2001): the ‗primitivists‘ or ‗substantivists‘ with Moses Finley (1965; 1973) as the major representativepropagating ‗immobilism‘ and an autarkic state with a pre-industrialised imposed economic system (Figure 226 left); and ‗modernists‘ or ‗positivists‘ with Michael Rostovtzeff (1957)propagating an industrialised mass production, capitalism built on exploitation of the plebs. The endless discussions between both factions are no longer an issue of debate since both models are played out due to the enhancement of ‗factual knowledge and interpretation of ancient technology and economics‘ (Greene 2000). Firstly, there was a building up of a more moderate primitivism involving a more dynamic and more complex pre-industrialised economy describing the Roman political economy as an integrated economic system based on unified local markets (Carandini 1986; 1989) (Figure 226 middle) or based on an interaction between loosely integrated regional economies and the economic policy by Rome (Woolf 1992) (Figure 226 right), though the perception of a production and exchange that is 310 According to some sources, the cost of living tripled during the later 2nd century AD under Marcus Aurelius and Commodus (De Laet 1949, 404). 410 based on farming and local consumption with undersized production scale and rudimentary trade activities was maintained. Finley 1965; 1973 Carandini 1986; 1989 Woolf 1992 • political economy • political economy • political economy • influencing • based on • intermediate • series of essentially LOCAL economic systems • unified LOCAL markets, credit and agriculture produce • series of REGIONAL economies loosely integrated into a single system Figure 226: Key models of the Roman economy A contrasting model to the agricultural character of the primitive pre-industrialised Roman economy proposed by Moses Finley (1965; 1973) is the model of a proto-industrialised Roman economy put forward by Ken Dark (2001):       an economic system in which long-standing crafts-working practices are coordinated to undertake mass production aimed at long-distance trade a money-based market exchange system clusters of rural craft-based production for regional or geographically larger markets products mediated through urban centres to these regional or wider markets use of traditional technologies evidence of coordination to produce standardised products The Roman glass industry matches the criteria to be fulfilled attributing a proto-industrialised character to the Roman economy. Two specific archaeological contexts demonstrate well the assertion: 1) the furnace tank st nd capacity of c. 25 tons at the 1 2 century AD primary workshop of Beni Salama in the western Nile Delta (EG) nd rd (Nenna 2008, 127-128); 2) the late 2 early 3 century AD shipwreck of Ouest-Embiez I, which held a major shipload of raw glass estimated at 17 tons (Fontaine, Foy 2007; Andreau 2010, 114). To generate a conceivable implication of the impact of this gigantic production capacity on the consumer market, 311 we present here some numbers. When taking an average weight of 100 grams per drinking vessel , the Egyptian primary workshop at Beni Salama processed in one single production activity enough raw glass to make, roughly speaking, 250.000 vessels, corresponding to the procurement by 250.000 households of one vessel. 312 Taking the ‗low count‘ as the generally estimated population within the Roman Empire, meaning c. 60 million and a rough estimation of 10 persons per household, we come up with c. 6 million households. This approximate calculation means that the Beni Salama workshop needed to produce 40 times its capacity annually. Seeing the st 313 very low price of glass artefacts from the early 1 century AD onwards , we can take for granted that the glass was a commonly consumed commodity during the entire imperial period. This estimation demonstrates that the Roman society had the market potential during the imperial period to consume easily 60 million vessels at the same time. This means that 6000 tons of raw glass was necessary to supply the 6 million households every year with 10 vessels with an average of c. 100 grams. But to enable the manufacturing of one piece per person per year, the Beni Salama workshop should produce annually 240 times its capacity of 25 tons. Consequently, the Beni Salama workshop would have to produce 25 tons of raw glass every 1,5 days. This approximation is, however, not realistic, as the production process of such a magnitude needed in all probability about 30 days or 311 The information is based on the weight of the Roman goblets and cups from the Ernesto Wolf Collection (Stern 2001), ranging between 50160 grams (Nenna 2008, 128). 312 Walter Scheidel makes a distinction between the generally accepted ‗low count‘ and his more elevated ‗high count‘ estimating the population of the Roman Empire above 100 million (Scheidel 2004; 2006, 9). 313 See Strabo, Geographica XVI, 2.25 [transl. Jones, H.L., 1961, Strabo VII, 271-273]. 411 more, and the primary workshops most likely were intermittently manufacturing and not on a continuous basis. Besides, we have to take into account that the provided numbers concerning one single vessel shape are to be considered as a strict minimum. This plain estimation thus illustrates a high potential that more primary workshops with the Beni Salama capacity must have existed contemporaneously in Egypt and the Levant to supply the entire Roman market of raw glass. It is plain that the shiploads with finished products and chunks of raw glass are clear evidence of supplying the 314 oversea markets with very standardised products on the one hand and the supply of the secondary workshops in the different regions and provinces (as we saw in case study I) to produce local commodities with a regional distribution. Obviously, the long-distance sea-trade must have been intense to supply the local workshops empirewide with raw glass to manufacture their proper material for the local or regional consumption market. Bearing in mind that the estimated 17-ton shipload of the Embiez-Ouest I shipwreck is equivalent to the transport of one large furnace tank or several smaller ones, the abovementioned empire-wide consumption calculation approximates an annual long-distance sea trade in raw glass easily reaching 250 shiploads. Besides the inevitable economic law of supply-and-demand, there was another advantage to favour a shipload with large broken chunks of raw glass rather than shipping brittle finished products for long-distance (sea) trade. We are not aware of a possible difference in taxation degree between raw materials and finished products, but mindful of Diocletian‘s Price List (Whitehouse 2004; Barag 2005)displaying differences in price according to the origin of 315 the glass (Egyptian vs Levantine) ; according to the state (raw glass vs finished product); and according to the type of artefact (vessel vs window pane)an equal tax rate was less favourable for trading finished products as these were more expensive and took more space. The load with finished products would need storage room for 1.000 large wooden boxes of 100 piled-up fragile vessels, whereas the load of broken chunks could be easily dumped in the cargo-space as ballast, without any risk to endure damage. Conversely, the shortening of the distribution range for finished products makes it important to counteract more easily the fluctuations in supply and demand within a society that is characterised by a slow itinerant distribution system, as in Roman imperial times, but that also decreases transport costs. The starting point for the here-stated need of benefit is not to be understood from the modern capitalistic viewpoint but the necessity to keep the production cost-effective so that there always remains a positive balance in what went out to pay every component within the production and distribution costs. Different methods are valuable to increase the cost effectiveness of the (black) glass production without raising the price of glass objects for consumers:   lowering the production costs  replacing the more expensive, purer ingredients by cheaper and less pure substitutes (by using ingredients with more impurities but without automatically involving the lowering of the physical quality of the glass metal)  changing the ratio main components by lowering the ratio of the more expensive ingredients in favour of the cheaper constituents  using more economic technical production methods like fuel-saving methods lowering the distribution costs  replacing the expensive provider by a cheaper provider  replacing the distant provenance of basic ingredients (sand, soda, colourants) by using ingredients from less far-away sources st From the early 1 century AD onwards, there is clear evidence for an emerging large-scale Romanised material culture all over the Roman Empire. This is yet about a century after the invention of glass blowing, somewhere st around the mid-1 century BC. The innovation was revolutionary to the glass industry, in particular because it helped (re-)map the glass industry in relation to competing craft industries such as pottery and metalware. Glassblowing meant that glass vessels also could become a standard commodity, since this technique enables a fast, cheap and easy production of recipients in all kind of colours and shapes and not negligible by using less material than casting or rod-forming. It is not unlikely that taxes and transport costs caused market economic pressures and problems in supplynot in the least due to the fragility of the productinducing the Romans to transfer knowledge as well alongside the raw materials towards distant places in the west. Our study on black- 314 We believe that imported finished products can only be traded over long distances when it concerns a very standardised product, as otherwise the imported commodities would hold little appeal for the consumer market. 315 We are aware of the unsolved issue whether the ―Alexandrian‖ and ―Judean‖ glass are pointing to the provenance of primary production (Whitehouse 2004) or to types of glass (Barag 2005). 412 appearing glass has clearly demonstrated that the workshops in the north-western provinces accumulated their nd knowledge during the second half of the 2 century AD to colour autonomously imported decolorised and ‗naturally‘ coloured glass (Van der Linden et al. 2009) [see Chapter 8]. The technological improvement of the production process to downsize the fuel consumption is not necessarily limited to the use of better performing fuel. Besides the evidence of using charcoal and coal by the Romans in the metallurgical craft industry (Webster 1955; Dearne, Branigan 1995), there are no clear indications of use in the 316 glass production to our knowledge, or at least it is not reported. Another technical modification that clearly reduced fuel consumption could be obtained by introducing crucibles. They replaced the built-in basins, but no clear consensus is available when this exactly happened and in what time envelope this technological innovation became common in the glass workshops of the entire Empire. It is generally accepted that the use of crucibles nd probably was introduced somewhere at the end of the 2 century AD, but this innovation became definitely rd nd rd frequent from the 3 century AD onwards. Striking it is to see that the 2 early 3 century AD glass workshop rd th where black glass was melted within basins made from quartz-rich tempered clay (Fischer 2009), but all 3 4 century dated workshops where black glass was meltedlike, for instance Trier, Les Houis/Sainte Menehould and Bragayielded only crucibles. Additional motives most likely made the use of crucibles more desirable than the built-in basins, but the issue is beyond the scope of this case study. Another technique to reduce fuel consumption is by recycling old glass. Adding cullet to the batch is not only beneficial to the reducing of the melting point of raw glass, it also influences the rapidity with which glass becomes ready for production. Hence, the alteration of the composition is useful to economise on fuel and a higher malleability includes the option to produce containers with a thinner wall, thus providing the opportunity to diminish the glass mass required per object and as a consequence produce more pieces with the same amount of glass mass. The arguments put forward prove that most glass workshops must have assimilated the nd rd abovementioned practices from the late 2 early 3 century AD to withstand the continuous economic pressure rd caused by inflation and devaluation that made them resistant to the long-lasting crisis of the 3 century AD. From the information assembled on black glass, it appears that the glass workshops were able to maintain production th by using the classic Levantine glass up to the mid-4 century AD by continuously improving the technical production methods and improving the production capacity by diminishing the quantity of glass for the manufacturing of one piece. Around that time, the sole technological innovation remaining to maintain the postulated profit in the production of black glass was to convert to a cheaper but less qualitative raw glass made from less pure ingredients, what resulted in the heyday of HIMT glass. This is however much later compared to rd the presumed introduction of the ‗naturally coloured‘ and decolourised HIMT glass during the 3 century AD to produce vessels and window panes (Mirti et al. 1993). It is possible that there was not equal pressure on the glass workshops responsible for the production of jewellery in black glass. To support the widespread assumption that the glass workshops have introduced innovative practices, the black rd th glass makes available evidence for Periods II-IV and in particular for the 3 4 centuries AD. To withstand the economic crises, the glass workshops first improved the production process and only afterwards transformed the production process. The initial innovations probably had a purely in-house practical charactere.g. better performing fuel; more efficient furnace constructions; introduction of cruciblesin order to continue the routine processing of the imported raw glass. As a result, we consider the economic stress that engendered an extensive practical knowledge of furnace technology to the Romans in order to create alternatives and innovate to economise on the production costs. Another primary cost that could be cut down is to reduce the use of the high quality raw glass with low impurity level, i.e. the Levantine glass. Such change must have been a/the stimulus to th introduce and promote the HIMT glass so idiosyncratic to the 4 century AD its emergence most likely took place rd in the 3 century AD. We believe, however, that the Roman glass workshops must have either introduced in substitution of the slowly expanding process of HIMT glass consumption, and probably even beforehand, the use of cheaper raw glass from the same provider/provenance (i.e. the Levant), or improved their technological knowhow by fabricating independently a more cost-effective mixture. The use of local sand has been reported already by Pliny the Elder [see Chapter 2], and thus the knowledge of primary glass production was seemingly already st introduced from the 1 century AD onwards. Characterising the composition of the raw glass produced in each primary workshop can help us better understand its distribution, seeing that the primary workshops must have supplied multiple secondary glassworking sites all over the Roman Empire (Freestone 2006; Van der Linden et al. 2009). The identification of the different sands is therefore giving new impulses in the understanding of the 316 We have not been able to check the scientific literature on the types of fuel used during Roman times to tend the fire in the various types of furnaces, but a general work would be very suitable and in particular concerning the glass furnaces. 413 different sand sources used in ancient times, although the hitherto-obtained information remains somewhat unsatisfactory to discriminate clearly the different sand sources since a far too small amount of material has been analysed so far (Silvestri et al. 2006, Degryse, Schneider 2008; Degryse, Shortland 2009). The benefit of introducing absorption spectroscopy is that the method provides an easy obtainable outline on the composition of a vast number of artefacts, firmly contributing to the advance of the research on raw glass materials [see Chapter 9]. Despite the information on the primary glass production in the West, it is generally accepted that during Roman times the raw glass production was centred on a relatively small number of primary workshops nearby the raw material sources where it was produced on a large scale (Freestone, Gorin-Rosen 1999; Freestone et al. 2000; Foy et al. 2003; Freestone 2003; Foy et al. 2004; Nenna 2008). But as has been stated before, only a limited amount of information is known regarding the production, the organisation or the precise location of Roman glass workshops (Baxter et al. 1995). Primary workshops, for instance, have up to now only been archaeologically investigated in the western Nile delta (Nenna 2008) and in Israel (Gorin-Rosen 2000), whereas most are late Roman to early Byzantine/early Islamic. Broken into chunks, the raw glass was distributed to local as well as to far-distant secondary workshops where glass objects were manufactured by melting and shaping nd glass. From the late 2 century AD onwards, these secondary workshops performed also the colouring of the imported ‗naturally coloured‘ and decolourised raw glass (Van der Linden et al. 2009). Hitherto, no clear evidence is known concerning the colouring of glass for the early imperial period (Jackson et al. 2009), but most probably it rd st was done in the same way as in the Hellenistic period, seeing that the various shipwrecks from the 3 and 1 317 century BC contain chunks of ‗naturally coloured‘ and coloured raw glass (Foy et al. 2000, 52). One single primary workshop is thus considered able to supply multiple secondary glassworking sites, supporting the idea of vast areas with glass assemblages showing a homogeneous composition. In reality, it is more complex as the imported chunks of raw glass were usually remelted by adding cullet to lower the melting point of the batch and coloured when necessary when the production of glass objects in a specific colour was desired. On the other hand, the discussion is still open whether one or a few workshops were specialised in the colouring of raw glass, functioning as intermediate centres dispatching specifically coloured glass cakes to the secondary workshops, or whether the secondary workshops themselves were responsible for colouring the imported ‗naturally coloured‘ and decolourised raw glass for proper production of consumer articles (Van der Linden et al. 2009; Cosyns et al. in preparation). The latter assumption takes for granted that the imported raw glass was solely uncontaminated nd decolourised or ‗naturally coloured‘ glass. The use of different recipes to produce black glass in the later 2 and rd first half of the 3 century AD within a relatively restrained regioni.e. the north-western provincesis a clear indication that raw glass production and glass colouring must have been separate executions since the colouring in the province Gallia Belgica as well as in the provinces Germania inferior and superior was then identical, whereas the provenance of the imported raw glass was different [see Chapter 8 and case study I in this chapter]. Most probably, the colouring was carried out in larger secondary workshops supplying the bulk of the smaller secondary workshops. This hypothesis is based on the fact that the composition of the black glass material within nd rd the provinces Germania inferior and Germania superior of the late 2 early 3 century AD show a strong homogeneity, whereas there must have been ample glass workshops active within that specific period in this vast territory. But when the imported raw glass was already coloured black before arriving at the secondary workshops, it is quite obvious that during the long-lasting Roman imperial period and within the different parts of the enormous Roman Empire, different modes must have been in use to make the glass appear black as there have been various recipes utilised. The chemical analyses of the black glass material show a mainstream in using genuine raw glass for the production of black-appearing glass artefacts but also confirm the introduction of recycling glass in the Roman Empire and the progressively growing impact of cullet in the Roman glass production. From technological point of viewe.g. used sand; used flux; used colourants and their concentrationsa good correspondence is observable with the four recognised periods [see Chapter 4], even though Period II and Period 318 III show great compatibility (see Table 167):  317 The black glass samples of Period I artefacts are characterised by various colourants added, whether it is manganese oxide to obtain purple glass, cobalt oxide to obtain a deep blue glass or iron oxide to st obtain a brownish to greenish brown glass. For that reason the 1 century material is produced in the Such kind of trade involving the distribution of coloured raw glass from primary workshops towards secondary workshops was already common as early as the Bronze Age, such as the blue-coloured raw glass in the Uluburun shipwreck (Jackson, Nicholson 2010). 318 The difference between the production of black glass from Period II and Period III is not the used composition but the applied production technique and consequently the idiosyncratic artefact types [see Chapter 3]. 414   then-customary techniques featuring the so-called Hellenistic traditions. No evidence from the black glass samples can demonstrate the use of or addition of recycled glass. nd By the end of the 2 century AD, a new technique got spread to produce black glass (read modified recipe for colouring without great changes in the use of sand or flux). The black glass samples of Period II and III artefacts are characterised by almost only greenish glass with high iron concentrations. The majority of the used glass is of a bluish-green hue, but sometimes elevated concentrations of antimony oxide have been observed what is due to the use of decolourised glass [see Chapter 8]. Although it remains undecided whether the decolourised glass was raw glass or cullet added to the batch of ‗naturally‘ coloured raw glass, the use of decolourised glass seems most likely as none of the samples show a mix of the decolourisers antimony and manganese oxides [see Chapter 8.4.2]. th In the second half of the 4 century AD, again a change in technology occurred, but this time it concerns the used glass matrix and consequently not restricted to the production of black glass. The black glass samples of Period IV artefacts are characterised by the introduction of HIMT glass and in particular from th the last decennium of the 4 century AD. But others show to be the result of recycling glass, as is demonstrated by the presence of lead oxide. Table 167: Timetable of the various glasses used to produce black glass artefacts period time raw glass colouring remarks Period I 1st–mid 2nd century AD Levantine addition of vegetable soda - Period II Period III Period IV mid 2nd–mid 3rd century AD mid 3rd–mid 4th century AD mid 4th–5th century AD Levantine Levantine HIMT Low Iron High manganese High cobalt High Iron High Iron High Iron high impurity level of manganese and titanium It is clear that the detection of recycled glass in the Roman Empire and above all during the late Roman imperial period generates new issues of debate regarding how we have to value the growing impact of recycling glass within the glass production as such and within the Roman economy in general. The recycling of glass in the Roman Empire has so far constantly been interpreted as a negative economical push-factor. It is therefore difficult to accept that the glass production in a growing and flourishing economy is partly running on cullet with a growing proportion of recycled glass in the manufacturing of glass artefacts. Consequently, it is customary to consider the th high ratio of recycled glass during the 4 century AD as an evidence of an economy in decline. Pull-factors can equally have influenced innovations and changes within glass production in the Roman Empire. We are for instance convinced that the introduction of HIMT glass, a glass of inferior quality compared to the then-prevailing Levantine glass, is inherently related to the growing proportion of recycled glass in the secondary glass th workshops in the 4 century AD. The most cost-effective arrangement by the secondary glass workshop must have been the adding of cullet, i.e. broken and refuse glass, to the batch. The recycling of glass caused an increasing independence of the secondary workshops towards the supply of raw glass from the primary workshops in the south-eastern Mediterranean. The discussion about the causes and effects of recycling is an important but mainly neglected issue in the debate of identifying glass production and the distribution patterns of its centres. The recent paper by Harriet Foster and Caroline Jackson (2010) on 128 decolourised late Roman glass samples from 19 British sites th demonstrates that the circulating glass in 4 century AD Britain was decolourised in three different ways: 1) by manganese (13 ex.); 2) by antimony (69 ex.); and 3) by manganese/antimony (69 ex.). Although this ascertainment has been notified by various prior assessments (Braun 1983; Vichy et al. 2007; Lauwers 2008) [see Chapter 8], it was never given any valuable explanation how to understand the presence of the mixed group. Within the scope of recycling, the study of decolourised glass has shown to be imperative in being aware of contaminations of the proposed theoretical model in Figure 227 that the three distinctive groups encompass two separate recipes (1-2) and one mixture of both recipes (1+2). Foster and Jackson explain the presence of the mixed manganese-antimony group as clear evidence of recycling by amalgamating manganese decolourised cullet with antimony decolourised cullet (Foster, Jackson 2010). It has not been considered, however, whether the late Roman secondary glass workshops in Britain (and in the rest of the [north-] western provinces) could perfectly have been provisioned with both types of raw glass, meaning that a mix was deliberately produced from raw glass and not from recycling cullet. It is, however, not excluded that cullet from both recipes and from the mixture was added to the batch as well. The meaning given to recycling is, to our opinion, too much assuming a Roman ecological solicitude, while it appears more correct to think of the outcome of a pure economic practice. It is essential to take into account that both types of decolourised glass occurred in concurrency, and secondly that 415 they were produced in different regions: 1) antimony decolourised glass is considered Levantine; 2) manganese decolourised glass is considered Egyptian. Primary workshop 1 Recipe 1 Mixture 1-2 Mixture 1-4 Mixture 1-2-4 Primary workshop 2 Recipe 2 Mixture 1-2-3 Mixture 1-2-3-4 Mixture 1-3-4 Recipe 4 Primary workshop 4 Mixture 2-3-4 Mixture 2-3 Mixture 3-4 Recipe 3 Primary workshop 3 Figure 227: Theoretical visualisation of the possible mixtures that can occur when cullet is added to the batch th Recent research by Foster and Jackson established that during the 4 century AD huge numbers of glass objects from British sites were produced with recycled glass and discusses the causes responsible for the success of th recycling in Britannia in the 4 century AD (Harriet, Jackson 2010). We in particular wish to confer to the proposed causes supposedly at the base of the large proportion of recycled glass used in late Roman Britain because we are persuaded that the obtained analysis results from the black glass research can help in refining and perceiving the projected interpretation. We are, however, not convinced by the offered explanation of th macroeconomic factors influencing the change because at least the first half of the 4 century AD demonstrates a political stability and a military revival. We believe that the organisation of glass production in the Roman Empire is bonded to complex interaction on the meso-level between the glass craft industry and the Roman economy. th Not believing in a decline of long-distance trade within the Roman Empire not until the late 4 century AD, the mixed manganese-antimony group thus should be regarded as an example for an intensive import into Britain from Egypt and the Levant. Yet some remarks can be put forward concerning the statement that the mixed manganese-antimony decolourised glass was the result of recycling cullet and in particular that this recycling was due to the diminishing availability of raw glass in late Roman Britain. Foster and Jackson attribute the intensity of recycled (decolourised) glass to the extreme downsizing of the longdistance trade with Egypt and the Levant due to the general insecurity on the trade routes and instability in the entire empire (Harriet, Jackson 2010). The large set of samples taken from a definite group void of vessels and collected from various Romano-British sites allows the obtained results to be considered representative of late 416 Roman glass production. Without doubt, we have to conclude that recycling glass was imperatively present in th Roman Britain in the 4 century AD, but the proposed effects seem to be rather shallow. To assume that the intensive use of recycled glass was the inevitable effect from a defective supply of raw glass caused by the stagnating long-distance trade due to the political incapacity to guarantee security within the Roman Empire in the th 4 century AD is perhaps a logical reasoning. However, it is, from our viewpoint, a biased conclusion (and perhaps even a misleading one) that is based on a one-dimensional view of the Roman economy that does not take notice of economical rules and forces as well as variables active within a complex economic system like that of the late Roman Empire. It is not very persuasive to blame the political-military incapability to explain the raise of th th recycling glass in the 4 century AD, especially because the first half of the 4 century AD is considered the ‗Constantinian renaissance period‘ resulting from Constantine‘s reforms of coinage as a reaction to the draconian 319 regulations set in the Price Edict by Diocletian in 301 AD (Corcoran 2000). The Roman economy flourished th more than half a century to fall abruptly into a new crisis only from the last quarter of the 4 century AD. What might have had an impact on the declining import of raw glass from the south-eastern Mediterranean are the possible taxes with regard to the definitive division of the Empire in an eastern and a western half from 364 AD under Valentinian and Valens. This divisionintroduced by Diocletian some 80 years beforehand – must have influenced differently the evolving of the consumer habits within both parts of the Empire. Roughly speaking we th have to separate the first half of 4 century AD (‗renaissance period‘) from the second half of that century (‗collapse period‘) when considering the economic situation and organization in the west and more particularly the th turn is noticeable at the start of the last quarter of the 4 century AD. It is however with the withdrawal of the th Roman legions and the conflicts with the Migration People from the early 5 century AD that the economy and consumption patterns of the western Empire contrast sharply with that of the eastern Empire. Long-distance trade th in the west survived in the 5 century but on a much smaller scale than before due to the growing inability to invest in largeand accordingly cheapshiploads. So we have to agree that the political and military involvement in the Roman economy in the west was dramatic for trade, but not because of incapacity to guarantee security within the Roman Empire. We are more convinced that the withdrawal of the Roman legions and the establishment of the new rulers created an economic vacuum where money was in short supply, considering that only the emperor was acquainted to strike golden coins, essential to bulk trade. It is within these more complex situations of long-lasting economical decline that we have to consider the necessity to increase the impact of recycled glass as raw material in the production of glass vessels. As already stated the introduction of HIMT glass must be linked to the growing proportion of recycling glass in the th secondary glass workshops in the 4 century AD but indirectly. The raison d'être of HIMT glass can be multiple and remains to this point no more than tentative:    by a big player on the market of raw glass o wishing to increase its market segment o wishing to consolidate its market by reacting to the increasing use of cullet o wishing to consolidate its benefit due to the increasing prices for transport and the continuous devaluation of the currency by a small player on the market of raw glass o wishing to increase its market segment by a new player on the market of raw glass o wishing to create a market segment with continuous growth margin It is important to distinguish the raison d‘être of HIMT glass from its success because the former is due to the creativity, initiative and originality of individual players on micro-level creating opportunities in a specific market segment of the Roman economy, whereas the success is the result from the situation and dynamics of the Roman economy on a meso-level and macro-level which is interrelated to the political and social evolution of a society. th When we see that up to the 4 century AD Levantine glass of very pure quality was the main raw glass distributed and consumed within the entire Roman Empire and that this market position gradually crumbles in favour of HIMT glass, the exercise is to value the cause(s) responsible for the shrinking market of Levantine glass and why HIMT 319 Study of the Price Edict established that Diocletian‘s attempt to reorganize the Roman economy stranded as it generated an inverse effect on the economy. Producers stopped the production of specific goods and merchants could circumvent the severe price regulations through barter what amplified the bartering system spectacularly and inherently contributed to the failing of tax revenues. The application of the Price Edict therefore only lasted a few years up to 305 AD and caused a new vacuum until Constantine‘s economical reforms (Corcoran 2000). 417 becomes a rising provider of raw glass. Not able to provide an answer on this issue within the present work, we will confer to some research questions:    had the introduction of HIMT glass imperative effect on the reducing market of Levantine glass? was the gradually crumbling market segment of Levantine glass an opportunity for a new player like HIMT glass? did Levantine glass remain an key market player that had to allow an upcoming giant provider of raw glass? When looking at the Price Edict by Diocletian from 301 AD stating Alexandrian glass as higher-priced compared to the Judean glass (Whitehouse 2004; Barag 2005), it is tempting to believe either that HIMT was not yet a market player or that the success of HIMT glass only came later. The dominance of Levantine glass all through the first three centuries AD is thus not solely due to the higher quality of the raw glass, but moreover because of a more competitive price compared to the Alexandrian glass. th On the other hand the quantity and diversity of glass artefacts from the 4 century AD archaeological contexts displays the increasing success of glass commodities to the prejudice of for instance vessels in bronze and ceramics. th The approach and interpretation of the high degree of recycled glass for the production of vessels in the 4 century AD consequently must have had a variety of imperative economic basic rules of supply and demand at the basis needs to be taken into account. Generally speaking it can be stressed that an evolution of the logistical organization of secondary workshops is apparent throughout the Roman imperial period:     Phase 1: glass production on the basis of imported raw glass (Hellenistic period + Period I(A)) Phase 2: glass production on the basis of imported raw glass + recycled glass on low level (Period IB-III) Phase 3: glass production on the basis of imported raw glass + recycled glass on intensive level (Period IIIB?-IV or later) Phase 4: glass production on the basis of recycled glass (Period IV and later) The production process of glass in Roman times was perhaps based on a continuously improving know-how by trial and error through mixing the different necessarily raw materials as scientific knowledge was unknown until th the 19 century. Well-thought-out techniques improved however the production capacity by diminishing the quantity of glass for the manufacturing of one piece but the use of cheaper ingredients or even the complete modification of the recipe was needed in order to maintain the wanted/required profit. In particular as political and/or economic reasons compelled the glass workshops to modify their supply of raw materials and accordingly confronted with variable quality. 10.4 Case-study III: Micro level In this case-study on micro-economy we verify whether (d/r)evolutions in the consumption pattern of particular black glass assemblages provides evidence on the changes meso- and macro-level of the Roman economy and to what extent consumers‘ behaviour produced an effect on the successive transformations in the Roman material culture and the Roman economy. It is true that the (dis)continuity of individual techniques cannot be viewed within one particular period but on the contrary needs to be examined in view of the preceding and consecutive periods (Mannoni 2007, lviii-lix). A diachronic analysis is needed in order to understand the causes of innovations and the abandonment of a specific manufacturing technique. This would enable us to view in what perspective stands the production of black glass vis-à-vis the glass production in general and the Roman economy. Because there is to our knowledge little direct evidence available from written sources on the prices and price evolutions of (black) glass and its evolving consumption we have to remain for the most part hypothetical. Hence, our assumptions are likewise in both preceding case-studies based on the evidence from the archaeological and archaeometric approaches in view of the conventional historical models. The framework of this case-study is supported by two recent papers of importance showing an interest intensification on the issue/ the first paper deals about the continuity and rupture of craft techniques in relation to the principles of Material Culture (Mannoni 2007), while the other treats on consumption and consumerism in the Roman Empire (Greene 2008). A general approach of the black glass material let presume of an uninterrupted continuity in production and consumption of black glass, whereas a more careful look [see Chapters 4; 6] demonstrates a discontinuity in the production of the different functional types and their various artefact types. The black glass production points to a 418 specificity of assemblages during well-defined and relatively short periods of time and regulary coupled to a regional distribution. When considering for instance the black glass vessels three distinct periods have been noticed320 [see Chapter 3]:  Period I: empire wide distribution; from Tiberian to Flavian times (c. 20-70 AD)  Tiberius-Claudius = cast vessels/ Claudius-Nero = blown vessels; duration of about half a century or 2 generations  Period II: limited to the north-western provinces; late Antonine-Severan period (c. 170/80-230 AD); solely blown vessels; duration of about half a century or 2 generations  Period IV: limited to the Levant and Egypt; during the 4 -5 defining of the length of its production and consumption th th centuries AD; rod-formed vessels; no clear It is tempting to assume for Period IV a similar duration of the black glass vessel production and consumption as for Period I and II but at present it is impossible to prove or invalidate the assumption. In view of Kevin Greene‘s paper (2008) on the necessity to differentiate between consumption - and consumerism it appears that the production of black glass vessels lasted for about two generations. The decrease in value arising from changes in style or techniques and thus the process to become outmoded, i.e. the obsolescence of the black glass vessels was very slow. This observation does not speak in favour of an intense degree of consumerism during the Roman imperial period. The change in production technique from cast to blown vessels in Period I demonstrates that the consumers‘ behaviour did not change with the technological transformation in the glass workshops. On the contrary, the glass workshops kept similar shapes in the production line by imitating their cast predecessors alongside the introduction of new shapesi.e. closed shaped (micro-level). Thus we see here an example of the glass workshop adapting the organization of its production through diversification of the assortment to expand its market share (meso-level). The introduction of such innovations also helped improve the production effectiveness in order to remain competitive (meso-level), but equally it was probably essential to implement ameliorating costeffective factors to carry on being profitable in a period of macroeconomic restructuring measures (macro-level). Tiziano Mannoni (2007) recently defined a long list of potential factors to investigate the possible dynamics that might have caused continuity, changes or rupture in the crafts industry in Roman imperial period (Table 168). We verified what factors can be linked to the production and consumption of black glass artefacts. The underlying idea of this case-study was in the first place to demonstrate to what degree the assemblages produced in black glass follows the prevailing fashion of shapes and functional types produced in any other glass hue. Or does a production differentiation exist depending on the glass hue? In other words, did the Romans use black glass to produce idiosyncratic material not available in other glass hues? Table 168: List of potential factors influencing technological change as proposed by Tiziano Mannoni 2007 scenarios of technical change factors continuity 1 product no longer useful 2 product replaced by another product X 3 new uses for old products modifying functionality X 4 demand for specific product disappears causing production to cease or evolve 5 new experiments undertaken X 6 irrational ideas inspire change X 7 technical ability retained by re-ordering production modifying the production capacity X 8 organization of production is modified X 9 adoption of techniques from other manufacturing trades or other cultures X 10 changes in artistic techniques related to taste X 11 subtle changes in overall demand X 12 generalized economic collapse 13 limited availability of raw materials X 14 limited availability of specialized labour X 15 limited availability of unspecialized labour X 16 changes in communication needs X discontinuity X X X X X Tiziano Mannoni (2007) distinguishes sixteen different factors that can influence technical change featuring either continuity or discontinuity know-how and technology (Table 168). Most factors of technical change seem to implicate the continuity of production but some factors can generate both, continuity and discontinuity. It is 320 Certain shapes are dated in Period III but appear to be considered as a offshoot of Period II assemblages. 419 obviously true that the continuity or discontinuity of a functional type or an artefact category produced in black glass resulted from a combination of different factors as well as that various factors have had no or little impact on the black glass assemblages. Furthermore Table 168 is not limited to microeconomic factors seeing that for instance factor 12 is generated from a macroeconomic level. We believe that not all factors implying discontinuity had an effect on the black glass production. For instance factor 1 is not applicable on the black glass artefacts as most shapes or functional types remained in use after the black glass was anymore fashionable. The black glass vessels from Period I are only produced from the late Augustan times to the early Neronian period and correspond to the assemblages produced in monochrome strongly-coloured glass with a change in production under Claudius when the cast technique got abandoned in favour of the free-blowing technique (factor 2). By the late Augustanearly Tiberian period the strongly-coloured polychrome vessels got out of fashion gradually and become replaced by strongly-coloured monochrome vessels (factor 11). Hence, the production of black glass vessels in Period I followed the then current fashion. Period II features the breakthrough of black glass production with the introduction of high iron black glass (factors 5; 8). The short lifespan of the vessels in comparison with the jewellery and counter production is partly due to the disappearance of the demand of this specific commodity type (factor 4), but also as a result of the incapacity to remain profitable caused by macroeconomic factors such as inflation and devaluation (factor 12) resulting from the political and military instability after the fall of the Severan Dynasty. Period III features the continuity and stability of the glass workshops processing black glass in analogy to the rd th production organization and artistic techniques established during Period II. The artefacts from the mid 3 –mid 4 century AD demonstrate a conventional production featuring an assortment of minor decorative variations per commodity type as demonstrated from the diversity of bangle and finger ring types. The immovability of these microeconomic elements must be regarded as a reflection of an overall stability that must have reigned in the rd craft industry even though the political and military crisis that reigned during a large part of the 3 century AD. On the basis of the changes in styles and techniques the production of black glass artefacts underwent around th the mid 4 century AD we are tempted to consider this as evidence of a total conversion of the glass industry (factors 5;8;9;11) during the first decades of Period IV. The technical side demonstrates a modified use of raw glass and the intensity of recycling. Changes in taste or fashion (factor 10) are recognizable in the introduction of glass counters decorated with a dotted design and the various bichrome and polychrome beads but characterize the continuity of a production. The change in production technique seen in the manufacturing of the bangles and finger rings must have been the result of adoption of techniques (factor 9) what can be explained by the Migration th People who entered by force and settled in the European frontier zone firstly but from the early 5 century AD by overwhelming the entire Western Roman Empire. 10.5 Conclusion The deductions within the abovementioned case-studies support the usefulness of a holistic approach on a minor segment within the wide range of archaeological small finds by providing additional information to a better understanding of the multi-levelled socio-economic organization in the Roman society. Mainly caused by macroeconomic changes and reforms the long-distance trade of finished (black) glass vessels, so illustrative for the late Republican and Julio-Claudian period, got reduced and slowly taken over by regional productions from the Claudian-Neronian period in the large centres such as Lyon (FR) and Avenches (CH). This evolution further evolves in a more widespread occurrence of glass workshops from the Flavian period and not exclusively in major centres. Hence the import was not put to a halt but it got reduced to a fraction of the entire consumption of glass vessels. The discussed segment of black glass vessels taken as a key indicator to nd rd regionalism of the Roman material culture of the late 2 –early 3 century AD in case-study I provides evidence for a regionalized glass production and consumption well within the province borders. Furthermore it confirms an imperial economy with binding state interference. The optimization of the taxing machinery necessary to increase considerably the tax income was probably imperative to finance the continuously growing army expenses of the war apparatus at the extremely large limes under the Severi. The black glass vessels provide clear evidence the Roman material culture show a strong ‗regional‘ homogeneity within the Roman provinces Gallia Belgica, Germania inferior and Germania superior and that these provinces have to be regarded as one large ―bonded‖ region. The Roman ‗globalized‘ market was thus not more than a 420 patchwork of regional and local production and consumption. In view of this observation and because glass workshops produced artefacts in all sorts of glass hues we see the acute necessity to reassess the entire range of Roman glass vessels from these three provinces to verify whether these corroborate the findings of this research. Furthermore it will generate an updated general work very suitable for the archaeologist confronted with Roman glass. Assessing all Roman vessels within this well-delimited but vast region would give the opportunity to detect much better the similarities and differences within the successive levels, i.e. local, regional, inter-regional, and long-distance. Further potential research on the Roman black glass production and consumption should verify what can be deduced from a multi-levelled approach in view of specific glass commodity categories irrespective of the glass hue, but in particular in view of those made in various other materials. On a regularly base we have demonstrated that an approach of the black glass material in relation to skeuomorphism can supply additional information on the chronology of black glass artefacts, on the use and function, on the distribution. In other words the relation between the production and consumption of particular vessel types and jewellery types in black glass (micro) with their equivalents in all other glass hues (meso) and in comparison with similar shapes in other materials (macro) will enable to evaluate how a multi-levelled approach on specific material can give valuable information on chronology, trade and use. 421 GENERAL CONCLUSION The theme of my thesis, ‗The production, distribution and consumption of black glass in the Roman Empire during st th the 1 5 centuries AD. An archaeological, archaeometric and historical approach,‘ involves a holistic approach within Roman material culture studies. The study of black glass artefacts covering a wide range of functional types spanning five centuries from a vast region may seem not enough focussed on a specific research issue or query. But the main objective of this research project on a minor segment of the Roman material culture was to provide reliable, solid data on diachronic consumption patterns to add a supplementary opinion on the character of the production and distribution of glass within the entire Roman Empire. Hence, this manuscript provided some answers on the different operating processes influencing Roman glass production in general and consumption of Roman black glass within the Roman economy. By describing different issues and aspects of the Roman black glass from an archaeological, archaeometric and historical approach, the empire-wide character of glass production (globalization) and the local or (inter)regional character of glass consumption (regionalism) was evaluated. The study of the consumption of goods to establish geographical and chronological idiosyncrasies in the consumer‘s behaviour was one issue, while determining the evolution and changes of the consumer‘s behaviour was another. The different approaches to analysing Roman black glass developed a more complete picture of its historical context by deriving social, economic and cultural/religious implications from the obtained data, thus providing a better reading of the use of these commodities within ancient society. The current study on Roman black glass resulted in a comprehensive overview based on the wide range of publications at hand and on research on location of unpublished collections from different places in Europe and the Mediterranean to provide a cross section of the local and regional material within different areas of the former Roman Empire. We provided the reader with a multilevel format discussing typology, chronology, context and distribution on the basis of a compiled database. The dataset comprised material from (old) museum collections as well as from (recent) archaeological field work, resulting in an overview of 4,475 entries from 651 sites. The consecutive chapters each provided specific value-added information contributing to a visualization of the consumer‘s behaviour in the Roman imperial economy. This was obtained by supplying data on the various aspects of this explicit and narrow segment of artefacts within the Roman material culture. The first seven chapters considered the research material from various archaeological viewpoints, providing information about the different commodities produced and consumed in the Roman Empire. The archaeometric approach in chapters 8 and 9 demonstrated that the knowledge on archaeological glass benefits greatly from the implementation of chemical and optical analyses. The results from both approaches were essential to generating a better interpretation of the chronological evolution and regional differences of Roman (black) glass production. Finally, within chapter 10 the obtained insights from the archaeological and archaeometric approaches were compared to the current models on Roman economy and society. The challenge of this research was to integrate and sublimate the results from the archaeological, archaeometric and historical approaches into a holistic view on Roman black glass and confront it with the main glass production and consumption which continuously evolved during the imperial period and experienced different drives within the Roman Empire. This multilevel approach and the interdisciplinary character of the research, with an interaction between the archaeological, the archaeometric and the historical disciplines, provided added value to archaeological glass research in general by supplying additional information on the interaction of regional and global consumption of Roman glass. Some results corroborated the prevailing assumptions, but others provided new information on Roman (glass) consumption in the broad sense. In particular, the archaeometric results contributed substantially to defining the chronological and/or geographical idiosyncrasies by discerning the different material sources and techniques applied to produce (black) glass. The first three chapters are to be considered introductory; they provided a clear outline of the approach we found necessary to tackle the Roman black glass material. Our approach allowed us to obtain well-defined artefact groups showing relevant chronological and geographical prospects for the analysis of the consumption patterns of the material. The first chapter described from different angles the identification of black and the complexity of its perception in order to define what we consider black glass. Subsequently, we explained the colour description and perception in antiquity, showing that in antiquity the involvedness in colour perception was already recognized but that neither rules nor parameters were established to categorize the colours with basic colour terms. This was supported by a linguistic and psychological approach to colour perception, making us conscious of the biased assumption on how to look at Roman glass hues due to the current determination of colours influenced by the 422 scientific colour spectrum and digitization. Further on, we showed the lack of consensus in describing the glass hues in archaeological material studies, seeing the wide range of systems and circumstances influencing the observation into too-general classes or in over-subdivided groups. Finally, we described our choice to consider all dark-coloured glass as black-appearing when it is perceived as black by the naked eye as well as when it appears black when placed on a white paper and only light in reflection is used. A preliminary colour survey of the blackappearing artefacts showed that those in dark green glass formed the bulk material. The Romans must have been confronted with the optical issue that green glass was more easily considered black (appearing) in comparison with glass in a blue, a brown or a purple hue. Besides this optical issue, they gave preference to iron oxide, the colouring agent with the strongest colouring effect, which was a very cheap and easily affordable pigment. The second chapter was not merely a glossary on technological terms, but it looked at the applied techniques in the production of Roman black glass commodities to justify the importance of recognizing and understanding technological features to build up an effective typo-chronology. The first part discussed the different aspects of glass production, including raw glass, finished products, and tools, whereas the second part focused on the techniques of production and decoration per commodity category. This outline demonstrated some novel insights on the production processes of some particular commodity types: we demonstrated the use of an open mould or wide metal ring to produce a part of the carinated beakers, the use of prefabricated glass rods to produce arm rings and the necessity to distinguish the poured/cast counters from the fused ones. Finally, a survey of the various glass hues applied per commodity was presented in relation to the applied technique, defining the application of specific techniques and glass hues to manufacture specific commodity categories. The tables and plots showed that specific techniques were applied to produce specific artefact types and that the adopted techniques changed depending on the period and/or region. In the third chapter, we provided typological mainframes on various commodity categories to enable quantifications for a detailed evaluation of the inventoried material. Some commodity categories, such as vessels and all sorts of jewellery, were very susceptible to classification systems. The vessels were presented according to the different production techniques but were considered within the margins of the existing general classifications. Besides a wide range of quite unique pieces, some shapes were very popular. A new typology of the arm rings and the finger rings was deemed necessary, as the existing ones are not useful to field archaeologists in need of datable material. Most other commodity categories remained at the level of a side note, but the gems and the counters were shown to be a ubiquitous commodity, with very promising research potential that necessitates a more detailed study. The next three chapters provided archaeological information on the chronology of the various types of commodities produced in black glass by querying and computing the database. This resulted in the removing of a number of biased assumptions concerning the date and provenance of certain (shapes of) artefact categories. Chapters 4 and 5 gave a more detailed chronology and context analysis, respectively, of the various artefact categories. The here-presented overview remains, however, limited because the information at hand is rather unsatisfactory, and a more elaborate research needs a meticulous preliminary study on the contextual chronology. We described the continuity in the use of black glass and the regular change of success of the different commodity categories during the consecutive periods. Vessels were mainly produced during the Julio-Claudian st dynasty throughout the second third of the 1 century AD (cast and free-blown); again under ‗Severan rule‘ in the th th north-western provinces (free-blown and mould-blown) and in the 4 5 century in North Africa and the Levant st (rod-formed). Except for some particular bead and gem types that fall into the 1 century AD, all jewellery in black nd rd glass is characteristic for the late Roman period from the late 2 early 3 century AD. Architectural decoration material and counters in black glass occur all through the five centuries. Our preliminary results demonstrated the interesting potential of a more thorough study of the counters. Through context analysis, the provenanced material indicated that the consumption of black glass not only was socio-economically driven but also gender-based. From the burial gifts it was seen that rich burials yielded the black glass vessels and that all jewellery was worn by girls and (young) women. The gemsgenerally inserted in finger ringsformed an exception as they were equally worn by men or form part of purely male dress, like the gems inserted on militaria such as the discoid brooches and ornamental helmets. The presence of black glass jewellery at military sites was perhaps minor but was evidence for the presence of entire families living close to or together with the soldiers in the late Roman period. The sixth chapter on production and distribution described the ubiquitous occurrence of most commodity categories. The knowledge of the glass workshops responsible for the wide range of black glass artefacts remains 423 scanty. Yet, a more detailed look at the distribution demonstrated that the production and consumption of the empire-wide presence of specific commodity types such as bracelets had a regional character. This regionalism points to a globalized Roman material culture with regional production centres, producing artefacts according to the needs of the consumer within the supply-and-demand model, but not in the least due to ruling socio-cultural traditions and rites. Chapter 7 can be considered an extension of Chapter 5, as it discussed the use and function of black glass artefacts on the basis of the contextual results. Not only did we see that the black glass jewellery was genderspecific to girls and ‗young‘ women on the basis of the burial gifts, but we also demonstrated that black glass vessels were only common burial gifts in rich tombs. Furthermore, the Romans did not estimate glass very highly but must have had difficulty in discriminating the black glass jewellery from the pieces using black semiprecious stones. The next two chapters were the result of an intensive collaboration with Prof. Koen Janssens of the University of Antwerp (Chapter 8) and with Prof. Hugo Thienpont and dr.ir. Wendy Meulebroeck of the Vrije Universiteit Brussel (VUB) (Chapter 9). Both approaches were necessary for adding valuable information to the obtained results from archaeological analysis to verify the issues production, trade and distribution. Chapter 8 described the results from chemical analysis by using SEM-EDX and LA-ICP-MS to define the chemical composition of the sampled material. This research demonstrated that a large diachronic selection of 366 very characteristic glass samples from a wide area enhanced our understanding of the colour of the glass matrix, the chronological and geographical subdivisions, the functional types, and the large site contexts. The results demonstrated that a chronological subdivision can be made on the basis of the different sands (Levantine or HIMT), fluxes (mineral soda or vegetable soda), and colorants (iron, manganese, or cobalt/copper) detected. In particular, the irongreen glasses showed a chronological difference, with low iron during Period I and high iron from Period II. The chemical analysis also demonstrated the use of decolourized glass to produce black glass, although ‗naturally coloured‘ glass was most frequent (manganese or antimony). The chemical analysis also helped in defining the provenance of the glass, either the raw glass (Levantine or Egyptian) or the locally coloured glass in the secondary workshops (Gallia Belgica or Germania Inferior/Superior). The colouring technique to obtain black glass in Period II is related to a regional technological specificity of one or more workshops within the north-western provinces, i.e. Gallia Belgica, Germania Superior and/or Germania Inferior and fits production scheme 4 as discussed in Chapter 6, while the ‗naturally coloured‘ and decolourized raw glass from Egypt and the Syro-Palestinian coast is an indicator of long-distance trade supplying the secondary workshops empire-wide. Such results helped to specify the distribution patterns of specific assemblages within particular periods and gave new clues to the existing trade routes and the organization of the Roman economy in imperial times. Additionally, a number of artefacts were determined to be obsidian, showing that this material was used by glassworkers and in glass workshops. Other artefacts were shown to be the result of purposeful recycling (as revealed by the presence of lead). In Chapter 9 on optical spectroscopy, different issues were described. First, we analysed how to eliminate the subjectivity of the hue description by making it quantifiable through the plotting of the measurements using UVVis-NIR spectroscopy on a CIE1934 horseshoe graph. The use of an integrated sphere while measuring the transmission of light made it possible to define the optical density (OD) of the glass matrix and quantify the intensity of strongly coloured glass; thus, it provided an objectively measured limit between strongly coloured and black-appearing glass. Further research demonstrated that the colorimetry of the glass matrix corresponded to its chemical composition, and consequently the colours matched specific periods and typological groups. Additionally, the UV-Vis-NIR spectroscopy provided information regarding the technological aspects of black glass such as furnace conditions (reducing and oxidising), which helped in interpreting the bi-plots from chemical analysis (e.g., the greenish-brown low iron group). We found that optical spectroscopy offers opportunities in archaeometry, as results can be obtained more quickly and easily. The use of absorption spectroscopy allows in situ analysis of large quantities of material to acquire a first insight on the assemblage at hand, thus facilitating the selection of samples for detailed chemical analysis. Optical spectroscopy also provides opportunities for archaeological research, as the transmission spectra and calculated colour coordinates on the CIE1934 horseshoe graph have been shown to correlate with the specific chemical composition of the glass matrix and the furnace conditions. The use of UV-Vis-NIR spectroscopy therefore helps in defining the Harris matrix of an excavation. The historical approach using insights gained from the archaeological and archaeometric approaches described in the final chapter has not been fully exploited. Three case studies show however that this holistic approach 424 unfolds information on the micro-, meso-, and macroeconomic levelimpossible to discern from a unilateral surveyallowing a clearer picture of the interactions within and the organization of the Roman economy. The macroeconomic analysis demonstrates that we have to make a distinction between the apparently at random cultural regional consumption pattern and the politico-economic regulated production and distribution within province boundaries of (black) glass commodities. Nevertheless this observation needs to be confirmed by the examining the assemblages in other glass hues, but above all by verifying the validity to define all crafts being subject to the centralized macroeconomic policy in Rome. The mesoeconomic analysis demonstrates the difficulty to understand the issue on recycled glass and its impact and implications on the Roman glass production. Besides, the (black) glass production and the crafts in general demonstrate that the Roman economy should be regarded in imperial times as a proto-industrial economy notwithstanding its primitiveness and limitations. Finally the microeconomic analysis confirms the involvedness of a variety of criteria influencing the success of a commodity type and the declining productions. The present study clearly demonstrated that a holistic approach provides the tools to acquire new insights for a better understanding of the consumption patterns in ancient society generally and of glass in Roman times more specifically. 425 426 PART IV THE CATALOGUE 427 428 THE CATALOGUE This catalogue does not pretend to be exhaustive or complete. It only intends to give an adequate data list of the recorded material to enable an as accurate as possible analysis with the aim to enable quantifications and observations on the production and consumption of Roman black glass artefacts from the entire Roman Empire st th from the 1 century AD until the 5 century AD. However lots of blanks remain in the descriptive fields and in particular concerning the illustrations. This is partly due to the unevenly availability of data from the published material recorded, and partly because the accessible information on the unpublished artefacts has not been 321 comprehensive either. Even though the majority of the data-base consists of published material, a significant number of records in the catalogue concerns unpublished artefacts. We are very much indebted to all museum conservators and archaeologists to allow us the opportunity to integrate unstudied material from museum 322 collections and from old excavations as well as from recently excavated pieces. In the attempt to give a full picture of the black glass artefact the recorded data incorporate various material groups (production material; vessels; jewellery; architectural decoration; counters). The large quantities and high diversity of black glass material made a deliberate selection of basic data fields unavoidable in order to facilitate relevant queries on the material to accomplish the proposed research aims. To retain the readability of the extensive and very diverse data-list we separated the data per material type into five separate files. The sequence is that established in Part I: The archaeological approach: 1) production material; 2) vessels; 3) jewellery (arm rings; finger rings; beads; pendants; gems; hairpins); 4) architectural decoration; 5) counters. The material of a number of publications and online catalogues has been overlooked by the author but is added subsequently in the present catalogue. Hence a good number of objects have not been included in the pie-charts 323 and figures, though we tried to implement as much as possible material in the related discussions and tables. The publication by Maurice H. Chéhab (1985-1986) on the material from the necropolis of Tyre (LB) is known to the author but we could not manage to incorporate the high amount of black glass artefacts (mainly arm rings, beads and pendants). 321 It is also important to take account of numerous indeterminate items among the unpublished material. Some of the unpublished material at the time of the research has been recently published, such as the Croatian material from Ore ac (Jelincek 2010) and the Portuguese material from Braga (da Cruz 2009). Other material is still in process like the Egyptian material from Douch Kyssis (Nenna forthcoming). 323 We are however well aware of many more publications dealing with Roman black glass artefacts. 322 429 Figure 228: Access 2007 catalogue Form 430 The database is available on cd-rom in Access 2007 and set in a traditional form (Figure 228), including the following fields: Id: catalogue number; is a unique number that facilitates referring in the text, presented in the extensive file 1 (see below) Depot: name or abbreviation of the museum or archaeological depot and its location Inventory number: reference number of artefact (either that of the museum or that of the excavation) Picture: the pictures are either by the author or scanned from publications mentioned in the field ‗reference‘; in few occasions we received pictures from museums or archaeological depots Drawing: the drawings are either by the author or scanned from publications mentioned in the field ‗reference‘; in few occasions we received drawings from museums or archaeological depots Country: refers to the country of provenance Place: refers to the present-day town where the excavations have taken place (where possible we cite the ancient name) Site: refers to the excavation area where the piece was retrieved Site character: distinguishes ‗civil settlements‘ and ‗military settlements‘ Context: distinguishes ‗domestic‘, ‗public‘, ‗burial‘ and ‗ritual‘ contexts; the material from ‗glass workshops‘ is taken separately but their material is considered with the material from domestic contexts. Commodity: distinguishes ‗production‘, ‗jewellery‘, ‗vessel‘, ‗utensil‘, ‗gaming piece‘, ‗architectural decoration‘ and ‗sculpture‘ Function: this field is left blank for the commodities ‗production‘, ‗gaming piece‘ and ‗sculpture‘; for the commodity ‗jewellery‘ we distinguish ‗ornament‘ in general (also used for ‗architectural decoration‘) and ‗body ornament‘, ‗dress ornament‘ and ‗hair ornament‘ specifically; for the commodity ‗vessel‘ are distinguished ‗tableware‘, ‗toilet ware‘ and ‗storage ware‘ Form type: this field characterizes the type of artefact (the type of production material; the type of vessel; the type of jewellery ;...) Typology: this field provides either a drawn up typology or one or more established typologies Colour: when possible the true hue of the glass is added ‗/‘ (e.g. black/blue) and when the artefact is bi-chrome or polychrome the applied colours are added ‗+‘ (e.g. black/blue + white/red) Technique: distinguishes ‗cast‘, ‗free-blown‘ and ‗mould-blown‘ for the vessels and ‗rod-made‘, ‗pressed‘, ‗folded‘ and ‗gathered + cone-rolled‘ for the jewellery; ‗rod-formed‘ is used for vessels, jewellery, utensils and architectural decoration; the production technique of the counters is left open Description: this field attempt to give a generalized but individual description of the shape and decoration Completeness: distinguishes the different options of preservationi.e. ‗complete‘, ‗almost complete‘, ‗1/2‘, ‗1/3‘, ‗1/4‘, ‗1/8‘, ‗fragmentary‘ Dimensions: provides measurements of the diameter (‗d.‘), length (‗l.‘), width (‗w.‘), wall thickness (‗th.‘), height (‗h.‘) in mm; Origin: defines details about the find circumstances Date: distinguishes the four periods definedi.e. I (0-150 AD), II (150-250 AD), III (250-350 AD), IV (350-500 AD); the different periods can be used in combination Reference: gives bibliographical references where the concerned item is described unless it concerns an unpublished piece Remarks: mentions additional information such as the sample numbers for the chemical analyses 431 Museums Abbreviations of museums occur in the text but mainly are used in the catalogue for practical reasons CM CRAN KMKG-MRAH M MAE MAN PAM PGRM County Museum/City Museum Centre de Recherche Archéologique Nationale at the Université Catholique de Louvain-laNeuve (UCL), Louvain-la-Neuve (BE) Royal Museums for Art and History, Brussels (BE) Museum Museum of Archaeology and Ethnology, Cambridge (UK) Musée des Antiquités Nationales, Saint-Germain-en-Laye (FR) Provinciaal Archeologisch Museum, Velzeke (BE) Provinciaal Gallo-Romeins Museum, Tongeren (BE) Besides a general file we present separate files concerning a particular group of black glass artefacts: File 1: All inventoried material (comprising the entries in alphabetical order respectively per ‗country‘; per ‗place‘; per ‗commodity‘) File 2: Production File 3: Vessels     cast vessels mould-blown vessels free-blown vessels rod-formed vessels The material labelled as ‗jewellery‘ within the field ‗commodity‘ is too disparate and too vast to use this wideranging commodity all together within one file. Therefore we present the different ‗form types‘ separately. File 4: Arm rings File 5: Finger rings File 6: Beads File 7: Pendants File 8: Gems File 9: Hairpins File 10: Architectural decoration    tesserae rods plaques File 11: Counters 432 The cd-rom also includes a number of papers in pdf on Roman black glass published by the author or in coauthorship: Baert, K., Meulebroeck, W., Wouters, H., Cosyns, P., Nys, K., Thienpont, H. and Terryn, H., 2011. Using Raman spectroscopy as a tool for the detection of iron in glass. Journal of Raman Spectroscopy, 42: n/a. doi: 10.1002/jrs.2935 Meulebroeck, W., Cosyns, P., Baert, K., Wouters, H., Cagno, S., Janssens, K., Terryn, H., Nys, K., Thienpont H., 2011. Optical spectroscopy as a rapid and low-cost tool for the first-line analysis of glass artefacts: a step-by-step plan for Roman green glass. Journal of Archaeological Science, 38, 2387-2398. Meulebroeck, W., Baert, K., Wouters, H., Cosyns, P., Ceglia, A., Cagno, S., Janssens, K., Nys, K., Terryn, H., Thienpont H., 2010. ‗The identification of chromophores in ancient glass by the use of UV-Vis-NIR spectroscopy‘ in Berghmans, F., Mignani, A.G., van Hoof, C.A. (eds.) Optical Sensing and Detection, Proceedings of SPIE 2010 Vol. 7726, 77260D1-12. Cagno, S., Cosyns, P., Janssens, K., 2010. La production de vaisselle en vitrum obsianum au Ier siècle après J.-C. Un état de la question à partir de l‘étude du fragment R1610 (MRAH-Bruxelles). Bulletin de l‘Association Française pour l‘Archéologie du Verre (AFAV) 2010, 9-14. Van der Linden, V., Cosyns, P., Schalm, O., Cagno, S., Nys, K., Janssens, K., Nowak, A., Wagner, B., Bulska, E., 2009. Deeply coloured and black glass in the nortern provinces of the Roman Empire: differences and similarities in chemical composition before and after 150 AD. Archaeometry 51/5, 822-844. Cosyns, P., 2009. ‗Sainte-Menehould (FR) and Trier (DE): Two Roman workshops of black glass jewellery in the northwest provinces reconsidered‘ in Annales du 17e Congrès de l‘AIHV (Antwerpen 2006), Brussels, University Press Antwerp (UPA), 8895. Cosyns, P., Fontaine, S.D., 2009. ‗La vaisselle en verre d‘apparence noire dans les provinces occidentales au Ier siècle ap. J.C.‘ in Annales du 17e Congrès de l‘AIHV (Antwerpen 2006), Brussels, University Press Antwerp (UPA), 80-87. Cosyns, P., Cagno, S., 2009. Internal report: μ-XRF Analyses of Roman black glass objects and a vitrum obsianum piece Brussels, KMKG Jubelparkmuseum, 16/03/2009, Brussels/Antwerp, VUB-SKAR & UA-MiTAC. Cagno, S., Cosyns, P., 2009. Internal report: Compositional Analysis on the Obsidian Vessel Fragment R1610 in the Royal Museums of Art and History (KMKG-MRAH) by using portable-Xrf, Brussels/Antwerp, VUB-SKAR & UA-MiTAC. Meulebroeck, W., Cosyns, P., Wouters, H., Baert, K., Nys, K., Terryn, H., Thienpont H., 2008. Optical Characterization and Analysis of Archaeological Glass Artefacts‘ in Proceedings of MOC 2008 Cosyns, P., Cagno, S., 2008. Internal report: The SEM-EDS and LA-ICP-MS-analyses of late Roman black glass from Trier, Brussels/Antwerp, VUB-SKAR & UA-MiTAC. Cagno, S., Cosyns, P., 2008. Internal report: The EPMA and LA-ICP-MS-analyses of late La Tène glass bracelets from Odijk, Tiel and Meteren (ACVU), Brussels/Antwerp, VUB-SKAR & UA-MiTAC. Meulebroeck, W., Cosyns, P., 2008. Internal report: Archeometrisch onderzoek d.m.v. Fotonica-analyse op Romeins zwart glas afkomstig van de opgravingen te Tienen-Grijpenveld (Erfgoedcel, Tienen), Brussels, VUB-TONA/SKAR. Gratuze, B., Cosyns, P., 2007. La composition chimique des perles en verre de la tombe à char de la nécropole laténienne de Neufchâteau-Le Sart par LA-ICP-MS. Arduina 63, 1-7. Cosyns, P., Hurt, V., 2007. Les perles en verre de Neufchâteau Le-Sart. Arduina 62, 1-5. Cosyns, P., Janssens, K., Vander Linden, V., Schalm, O., 2006. ‗Black glass in the Roman Empire: a work in progress‘ in Creemers G., Demarsin B., Cosyns, P., (eds), Roman Glass in Germania Inferior. Interregional Comparisons and Recent Results, International colloquium Tongeren, 13/05/2005, Atuatuca 1, Hasselt, PGRM Tongeren, 30-41. Cosyns, P., Hanut, F., 2005. Black glass of 2nd to 3rd century date in Northern Gaul : a preliminary survey, in Annales du 16e congrès de l‘Association Internationale pour l‘Histoire du Verre London 2003, 113-118. Cosyns, P., 2004. Les bracelets romain en verre ‗noir‘: une étude typologique et contextuelle, in Bulletin de l‘A.F.A.V. 2003-04, 15-18. Cosyns, P., 2003. Romeinse glazen armbanden in België, Romeinendag 08/02/2003, Leuven, 12-16. 433 434 APPENDICES APPENDIX 1: LIST OF BLACK GLASS SAMPLES ANALYSED BY SEM-EDX RESIN PC1 SEM-EDX at MiTAC, UA, Antwerp a-f Vlaams Instituut voor het Onroerend Erfgoed (VIOE), Zellik (BE) g-k not located, examined by courtesy of Eleni Schindler-Kaudelka (AT) PC1 cat.no. place inventory no. a 514 Rumst – BE 88 RU 51.40 b 510 Rumst – BE 88 RU 86.88 c 512 Rumst – BE 89 RU 120.2 d 511 Rumst – BE 89 RU 107.204 e 513 Rumst – BE 89 RU 107.205 f 509 Rumst – BE RU (?) _ (?) 18 g 115 Magdalensberg – AT M70 6.7 h 117 Magdalensberg – AT M49 GRAB III 50A i 114 Magdalensberg – AT TGF IV 4.591 NR2 RIMFRAG j 113 Magdalensberg – AT TAF IV 4.591 NR2 BODYFRAG k 119 Magdalensberg – AT GLAS A VARIA 2614 BASEFRAG period II II II II II modern? I I I I I hue green green green green green green green green green green green function vessel – form IIB.1 vessel – form IIB.1 vessel – form IIB.1 vessel – form IIB.1 vessel – form IIB.1 production waste / raw glass vessel – form IA.2 vessel – form IA.2 vessel – form IA.2 vessel – form IA.2 vessel – form IA.4 PC1 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a b c d e f g h i j k 16,5 18,4 16,5 17,0 16,7 1,6 16,2 19,6 20,1 20,1 20,0 0,4 0,5 0,6 0,5 0,5 3,7 3,3 1,9 1,9 1,9 1,5 1,6 1,3 1,8 1,6 1,7 3,0 2,5 1,9 3,1 3,1 2,7 64,4 63,6 66,4 63,9 65,9 58,9 64,0 66,1 66,0 65,7 65,5 n.d. n.d. n.d. n.d. n.d. 1,8 0,5 n.d. n.d. n.d. n.d. n.d. 0,1 n.d. n.d. n.d. n.d. n.d. 0,1 0,3 0,3 0,2 0,9 1,0 0,7 0,9 0,7 0,1 0,9 0,8 1,0 1,0 1,0 0,6 0,4 0,7 0,6 0,7 6,7 2,2 1,1 1,0 1,0 1,0 5,6 5,3 6,2 5,5 6,2 21,5 7,8 6,9 4,7 4,8 6,4 n.d. n.d. n.d. n.d. n.d. 0,2 0,2 0,2 0,4 0,4 0,2 0,2 n.d. 0,3 0,2 0,3 0,4 0,3 0,2 0,1 0,1 0,1 9,7 9,4 6,7 9,9 7,2 2,1 2,0 1,1 1,6 1,6 1,3 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 435 RESIN PC2 SEM-EDX at MiTAC, UA, Antwerp synchrotron in Hamburg 07.2009 a-b, d-e Koninklijke Musea voor Kunst en Geschiedenis (KMKG), Brussel (BE) c not located, examined by courtesy of Hubert Cabart (FR) f-n Erfgoedcel, Tienen (BE) PC2 cat.no. place inventory no. a 868 Elkab – EG KMKG E 7776 A b 869 Elkab – EG KMKG E 7776 B c 1075 Faulquemont – FR FAULQUEMONT ST24 d 870 Elkab – EG KMKG E 7776 C e 871 Elkab – EG KMKG E 7776 D f 543 Tienen – BE TI 02 TR/20-78 (tienen 3) g 548 Tienen – BE TI 02 TR/20-80 (tienen 5) h 542 Tienen – BE TI 01 TR-79/059 (tienen 7) i 544 Tienen – BE TI 98 TR/03.26/5 (tienen 2) j 545 Tienen – BE TI 02 TR/20-262-3 (tienen 6) k 541 Tienen – BE TI 01 TR-79/142 (tienen 8) l 539 Tienen – BE TI 02 TR/20-79 (tienen 4) m 546 Tienen – BE TI 98 TR-03/22 (tienen 1) n 547 Tienen – BE TI 98 TR-02/00.3 period II II II II II II II II II II II II II II hue brown brown green brown brown green green green green green green green green green function vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form IID.1 IID.1 IIB.5 IID.1 IID.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 PC2 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a b c d e f g h i j k l m n 15,7 16,7 16,6 14,5 17,0 16,6 15,9 16,9 16,3 17,0 17,4 16,6 15,8 15,7 2,0 2,0 1,8 1,1 1,1 1,1 0,5 0,7 1,0 0,5 1,0 0,4 0,5 2,0 3,7 3,1 3,3 2,2 1,6 1,6 1,7 1,8 1,6 1,4 1,6 1,7 1,6 3,7 65,8 66,0 66,7 72,5 64,5 64,6 63,2 64,2 64,5 62,6 65,1 65,7 64,1 65,8 0,2 0,2 0,1 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 0,2 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 0,1 n.d. 0,9 0,9 0,8 0,9 0,7 0,8 0,7 0,7 0,8 0,8 0,7 0,8 0,7 0,9 2,5 2,4 1,6 1,0 0,9 0,9 0,6 0,7 0,9 0,5 0,9 0,6 0,7 2,5 6,3 6,6 6,7 6,2 5,7 5,9 5,8 6,0 5,9 5,1 5,4 6,0 6,1 6,3 0,4 0,3 0,3 0,2 0,1 0,1 n.d. n.d. 0,1 n.d. 0,1 n.d. 0,1 0,4 0,1 0,2 0,2 0,1 0,2 0,2 0,2 0,2 0,2 0,1 0,1 0,4 0,2 0,1 2,4 1,8 1,9 1,4 8,3 8,5 11,4 8,8 8,6 12,1 7,7 7,7 10,1 2,4 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 436 RESIN PC3 SEM-EDX at MiTAC, UA, Antwerp a-d, f-h, l-m private collection Goossens, Borgerhout (BE) e not located, examined by courtesy of Hubert Cabart (FR) i-k PGRM, Tongeren (BE) n Archeologische Dienst Provincie Oost-Vlaanderen, Gent (BE) PC3 cat.no. place inventory no. period hue a n.c. Vremde – BE II a26 modern? black b n.c. Vremde – BE III b3 modern? black c n.c. Vremde – BE III a modern? black d n.c. Vremde – BE II ? modern? black e 1073 Faulquemont – FR Faulquemont ST14 II green f 233 Dilsen – BE Dilsen 1 II-III green g 234 Dilsen – BE Dilsen 2 IV green h 232 Dilsen – BE Dilsen 3 IV green i n.c. Tongeren – BE Tongeren 20_346 I colourless j n.c. Tongeren – BE Tongeren 17_849 I colourless k n.c. Broechem – BE Broechem H7d 0 purple l n.c. Eindhoven – NL Eindhoven H3a 0 purple m n.c. Emblem – BE Emblem XII 19 0 purple n n.c. Aalter – BE Aalter Langevoorde 0 colourless function production waste ? production waste ? production waste ? production waste ? vessel – form IIB.6 bracelet – A2 bracelet – D1 bracelet – D1 vessel – mould-blown cup with Greek inscription vessel – mould-blown cup with Greek inscription bracelet – late La Tène bracelet – late La Tène bracelet – late La Tène bracelet – late La Tène PC3 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a b c d e f g h i j k l m n 0,1 0,1 n.d. 0,4 16,4 16,7 16,6 17,2 14,7 15,1 15,3 16,4 15,8 16,7 1,6 1,7 1,5 1,8 2,9 0,6 0,7 0,6 0,4 0,3 0,7 0,9 1,8 0,3 8,6 8,5 9,0 9,0 1,8 1,6 1,7 1,7 2,5 1,6 1,9 1,6 1,9 1,7 52,9 51,5 52,3 53,6 63,4 64,5 61,5 63,2 72,6 70,8 69,7 65,1 62,5 69,1 n.d. n.d. n.d. n.d. 0,1 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 0,1 n.d. 0,1 n.d. n.d. 0,1 n.d. n.d. 0,1 n.d. n.d. n.d. n.d. 1,0 1,0 0,9 1,0 0,8 0,9 0,9 0,9 0,6 0,8 1,2 1,2 1,2 1,2 1,6 0,5 0,5 0,6 1,0 0,7 1,0 1,0 1,9 0,9 21,5 23,7 24,4 21,8 10,7 6,8 6,8 6,4 6,7 7,8 7,8 10,8 10,6 8,7 0,7 0,8 0,8 0,8 0,1 n.d. 0,1 0,1 0,1 0,1 n.d. 0,1 0,1 0,1 2,5 2,6 2,7 2,5 0,4 0,1 0,4 0,9 0,7 2,2 2,4 2,7 4,0 1,2 11,0 9,9 8,0 8,8 1,6 8,2 10,8 8,3 0,5 0,5 0,4 0,4 0,9 0,4 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 437 RESIN PC4 SEM-EDX at MiTAC, UA, Antwerp a-g Erfgoedcel, Tienen (BE) PC4 cat.no. place a 531 Tienen – BE b n.c. Tienen – BE c 540 Tienen – BE d n.c. Tienen – BE XRF at MiTAC, UA, Antwerp (not sampled) PC4 cat.no. place e 530 Tienen – BE f 537 Tienen – BE g 535 Tienen – BE inventory no. Ti 98 TR/06/164 Ti 02 TR/079/468 Ti 97 TR/02/018 Ti 98 TR/02/128 period I-III I II I hue purple green purple brown function counter – large vessel – form IA.14 vessel – form IIC.2 vessel – form IA.14 inventory no. Ti 02TR/79/440 Ti 99 TR 018/262 Ti 00 TR/016/252 period I-III II-III II-III hue brown black black function counter – small bead – globular with suspension loop bead – micro annular PC4 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,1 1,5 1,6 64,6 n.d. n.d. 0,8 1,5 9,0 0,1 3,4 1,3 n.d. n.d. n.d. b 16,2 2,5 1,6 65,3 n.d. n.d. 0,8 1,7 9,3 0,1 0,4 1,9 n.d. n.d. n.d. c 15,2 0,5 1,8 69,8 n.d. n.d. 0,8 0,8 8,4 n.d. 2,0 0,6 n.d. n.d. n.d. d 17,6 0,4 4,4 70,5 n.d. n.d. 1,1 1,5 4,0 0,1 n.d. 0,4 n.d. n.d. n.d. e 14,2 2,6 1,7 62,6 0,2 0,1 0,9 2,9 12,0 0,3 0,5 2,1 n.d. n.d. n.d. f 13,2 0,6 1,7 64,1 n.d. n.d. 0,9 1,0 8,9 0,1 0,4 9,1 n.d. n.d. n.d. g 18,2 0,7 1,4 63,5 n.d. 0,5 1,0 0,8 5,7 0,2 6,9 1,1 n.d. n.d. n.d. 438 RESIN PC5 SEM-EDX at MiTAC, UA, Antwerp a, d-k PGRM, Tongeren (BE) b-c not located, examined by courtesy of Hubert Cabart (FR) l private collection, Leuven (BE) PC5 cat.no. place inventory no. a n.c. Broekom – BE 84.B.320 b 1359 Reims – FR DES 03/ US 148 c 1360 Reims – FR DES 03/ 442 d 497 Riemst – BE 2052 e 565 Tongeren – BE hondstraat f 553 Tongeren – BE 1562 g 554 Tongeren – BE 69.TON.21b h 559 Tongeren – BE 79.F.243 i 560 Tongeren – BE 71.G.50 j 590 Vechmaal – BE 2.D4.44 k n.c. Vechmaal – BE 2.D4.11 l n.c. Wenduine – BE 2005-1 period 0 II-III II-III I I II-III II-III II-III II-III II-III 0 modern hue purple green green green purple green green green green green purple brown function bracelet – late La Tène bracelet - A3 bracelet - A3 vessel - form IA.1 vessel - form IA.11 bracelet - C1 bracelet - A5 bracelet - A1 bracelet - A3 bracelet - A4 bracelet – late La Tène vessel; cilindrical bottle PC5 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 19,8 0,8 2,3 66,9 n.d. n.d. 0,6 0,6 6,2 n.d. 1,4 0,3 n.d. n.d. n.d. b 21,7 0,6 2,0 64,0 n.d. 0,1 0,6 0,6 4,5 0,1 0,1 5,9 n.d. n.d. n.d. c 19,9 0,7 2,0 66,9 n.d. n.d. 0,6 0,6 5,4 n.d. 0,3 3,7 n.d. n.d. n.d. d 24,9 3,0 2,2 60,2 0,4 0,2 0,8 1,9 5,1 0,1 0,2 1,0 n.d. n.d. n.d. e 20,1 0,8 1,8 67,1 n.d. n.d. 0,6 0,6 7,0 n.d. 1,8 0,3 n.d. n.d. n.d. f 17,1 0,7 2,6 62,3 0,1 n.d. 0,5 0,9 5,5 0,1 0,3 4,1 n.d. 5,8 n.d. g 19,3 0,6 2,1 62,4 n.d. 0,1 0,5 0,6 5,2 0,1 0,3 8,8 n.d. n.d. n.d. h 18,7 0,8 2,4 64,1 n.d. 1,7 0,4 0,6 5,8 0,1 0,5 5,1 n.d. n.d. n.d. i 18,8 0,8 2,0 65,0 n.d. n.d. 0,6 0,6 5,9 n.d. 0,5 5,6 n.d. n.d. n.d. j 16,9 0,7 2,7 61,1 0,1 n.d. 0,4 0,7 5,4 0,1 0,4 4,7 n.d. 6,9 n.d. k 21,8 1,1 1,9 64,3 n.d. 0,1 0,7 1,3 5,4 0,1 2,6 0,7 n.d. n.d. n.d. l 3,0 3,6 2,9 64,3 0,7 0,1 0,1 1,9 21,8 0,1 0,1 1,4 n.d. n.d. n.d. 439 RESIN PC6 SEM-EDX at MiTAC, UA, Antwerp synchrotron in Hamburg 07.2009 a-k Augusta Raurica, Augst (CH) PC6 cat.no. place inventory no. a 3504 Augst – CH 1939-2120 b 3490 Augst – CH 1948-2409 c 3535 Augst – CH 1968-12534 d 3756 Augst – CH 1976-3747 e 3523 Augst – CH 1976-9745ab f 3757 Augst – CH 1975-8206ab g 3760 Augst – CH 1973-13799 h 3530 Augst – CH 1981-7046 i 3437 Augst – CH 1937-6527 j 3371 Augst – CH 1938-3093 k 3445 Augst – CH 1937-6525 period II II II II II II II II II-III II-III II-III hue green green green green green green green green green green green function vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form bracelet - A2 bracelet - A2 bracelet - A1 IIB.5 IIB.5 IIB.5 IIB.5 IIB.5 IIB.5 IIB.5 IIB.5 PC6 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 20,5 0,8 1,9 65,6 n.d. n.d. 0,6 0,5 4,3 n.d. 0,2 5,6 n.d. n.d. n.d. b 20,8 0,8 2,0 65,9 n.d. n.d. 0,6 0,5 4,3 n.d. 0,3 4,8 n.d. n.d. n.d. c 20,8 0,7 2,0 66,1 n.d. n.d. 0,6 0,5 4,3 n.d. 0,2 4,8 n.d. n.d. n.d. d 20,0 0,9 1,9 65,3 n.d. n.d. 0,7 0,6 5,1 n.d. 0,2 5,3 n.d. n.d. n.d. e 20,0 0,8 1,9 64,7 n.d. n.d. 0,7 0,6 5,2 n.d. 0,3 5,8 n.d. n.d. n.d. f 20,7 0,7 2,0 65,2 n.d. n.d. 0,7 0,5 4,9 n.d. 0,2 5,0 n.d. n.d. n.d. g 21,4 0,6 1,8 63,8 n.d. 0,1 0,7 0,5 4,5 n.d. 0,2 6,4 n.d. n.d. n.d. h 21,4 0,5 1,9 64,2 n.d. n.d. 0,7 0,6 4,8 n.d. 0,2 5,8 n.d. n.d. n.d. i 19,1 0,7 1,9 65,0 n.d. n.d. 0,7 0,6 5,9 n.d. 0,4 5,6 n.d. n.d. n.d. j 19,7 0,7 1,9 64,1 n.d. n.d. 0,6 0,6 5,9 n.d. 0,4 6,1 n.d. n.d. n.d. k 19,3 0,5 1,8 65,2 n.d. n.d. 0,7 0,6 5,9 n.d. 0,4 5,5 n.d. n.d. n.d. 440 RESIN PC7 SEM-EDX at MiTAC, UA, Antwerp a-f Musée du Monde Souterrain, Han-sur-Lesse (BE) PC7 cat.no. place inventory no. a 314 Trou de Han – BE A 66-59 b 312 Trou de Han – BE A 66-60 c 307 Trou de Han – BE A 66-61 d 310 Trou de Han – BE A 71-8 e 301 Trou de Han – BE A 71-9 f 303 Trou de Han – BE A 71-10 period II-III II-III II-III II-III II-III II-III PC7 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO a 16,4 0,6 1,9 65,3 n.d. n.d. 0,7 0,7 b 16,4 0,6 1,9 63,4 n.d. n.d. 0,7 0,9 c 16,1 0,6 1,8 63,8 n.d. n.d. 0,6 d 16,7 0,6 1,8 65,4 n.d. n.d. e 16,6 0,7 1,8 63,9 n.d. f 16,5 0,7 1,9 65,9 n.d. hue green green green green green green function bracelet - A2 bracelet - C2 bracelet - C2 bracelet - B2 bracelet - C2 bracelet - C2 TiO2 MnO Fe2O3 CuO PbO ZnO 6,7 0,1 0,5 7,2 n.d. n.d. n.d. 6,3 n.d. 0,2 9,5 n.d. n.d. n.d. 0,9 6,4 0,1 0,2 9,3 n.d. n.d. n.d. 0,7 0,7 6,7 n.d. 0,4 7,0 n.d. n.d. n.d. n.d. 0,7 0,9 6,5 0,1 0,2 8,7 n.d. n.d. n.d. n.d. 0,7 0,7 6,7 0,1 0,5 6,4 n.d. n.d. n.d. 441 RESIN PC8 SEM-EDX at MiTAC, UA, Antwerp a-f Musée du Monde Souterrain, Han-sur-Lesse (BE) PC8 cat.no. place inventory no. a 305 Trou de Han – BE A 73-3 b 308 Trou de Han – BE E 76-79 c 306 Trou de Han – BE AX-72 d n.c. Trou de Han – BE E 77-6 e n.c. Trou de Han – BE no inv. No. f 316 Trou de Han – BE W1 period II-III II-III II-III modern? 0 i hue green green green green blue purple function bracelet - A2 bracelet - A5 bracelet - A2 cameo/medallion chunk of obsidian vessel – form IA.2 PC8 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 15,3 0,6 2,2 63,1 n.d. n.d. 0,6 1,0 6,5 0,1 0,3 10,4 n.d. n.d. n.d. b 16,1 0,6 1,9 65,4 n.d. n.d. 0,7 0,8 6,6 0,1 0,4 7,3 n.d. n.d. n.d. c 18,3 0,1 1,0 68,3 n.d. n.d. 0,1 0,7 0,6 0,2 10,4 0,5 n.d. n.d. n.d. d 16,1 1,6 3,6 62,8 n.d. n.d. n.d. 1,7 9,1 0,3 3,8 1,1 n.d. n.d. n.d. e 0,2 4,1 19,4 38,0 n.d. 0,3 n.d. 1,6 36,1 0,2 n.d. n.d. n.d. n.d. n.d. f 0,8 2,1 24,1 52,3 n.d. n.d. n.d. 4,4 6,0 1,0 n.d. 9,3 n.d. n.d. n.d. RESIN PC9 & PC10 Both resins are excluded from the appendix because they were not incorporated in the catalogue. Both resins contained solely late La Tène dark purple glass bracelets from the Dutch sites of Meteren, Odijk and Tiel in the east of the Netherlands which are date c.120-50 BC. We are very grateful to the Archeologisch Centrum Vrije Universiteit Amsterdam (ACVU) for having kindly given to our disposal these particular sample set and for having granted permission to integrate the analysis results in the general output of the chemical analyses on deeply coloured glass. 442 RESIN PC11 SEM-EDX at MiTAC, UA, Antwerp a-n CNRS – Var (FR) PC11 cat.no. place a 1321 Olbia – FR b 1324 Olbia – FR c 1325 Olbia – FR d 1315 Olbia – FR e 1323 Olbia – FR f 1335 Olbia – FR g 1326 Olbia – FR h 1318 Olbia – FR i 1322 Olbia – FR j 1331 Olbia – FR k 1330 Olbia – FR l 1329 Olbia – FR m 1227 Marseille – FR n 3892 Sidi Djididi – TN inventory no. 061-3 (olbia 5) 281-FNS C014-2 FNS (olbia 4) 072-5 M6648 037-20 (olbia 3) C014-1 FNS HS 13.08.1965 1002-1 294-2 (olbia 2) 037-30 (olbia 1) 056-8 JV 29.92 1035 period I I I I I I I IV I I I I I (post?) IV hue blue green purple purple brown purple green green green green green green green green function vessel vessel vessel vessel vessel vessel vessel bracelet - D1 vessel vessel vessel vessel vessel arch. decoration - tessera PC11 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 18,8 1,0 2,4 64,6 n.d. 0,1 0,6 1,2 6,4 0,1 2,1 2,4 0,3 n.d. n.d. b 19,1 2,2 5,3 62,9 n.d. n.d. 0,8 1,7 5,8 0,3 0,1 1,9 n.d. n.d. n.d. c 18,0 0,7 2,1 67,5 n.d. n.d. 0,8 0,8 7,5 n.d. 2,1 0,5 n.d. n.d. n.d. d 18,2 0,7 2,0 66,0 n.d. n.d. 0,6 0,9 7,8 n.d. 3,3 0,5 n.d. n.d. n.d. e 16,9 0,1 0,7 74,2 n.d. n.d. n.d. 0,5 7,2 n.d. 0,2 0,2 n.d. n.d. n.d. f 17,1 0,7 2,3 67,2 n.d. n.d. 0,7 0,9 7,9 n.d. 2,6 0,6 n.d. n.d. n.d. g 16,5 2,9 2,1 63,4 0,5 n.d. 0,5 3,3 8,8 0,2 0,3 1,5 n.d. n.d. n.d. h 17,5 1,1 2,2 61,9 n.d. n.d. 0,7 0,7 6,7 0,3 2,4 6,5 n.d. n.d. n.d. i 18,2 2,2 5,3 63,7 n.d. n.d. 0,5 1,7 6,1 0,3 n.d. 1,9 n.d. n.d. n.d. j 19,1 1,8 2,7 65,6 n.d. n.d. 0,7 1,4 6,5 0,3 0,4 1,6 n.d. n.d. n.d. k 18,5 1,8 2,6 65,3 n.d. n.d. 0,9 1,5 6,9 0,3 0,6 1,4 0,1 n.d. n.d. l 18,4 2,8 3,1 62,7 n.d. n.d. 0,7 2,1 8,5 0,2 0,3 1,5 n.d. n.d. n.d. m 20,2 2,3 3,0 61,6 n.d. n.d. 1,0 1,4 7,2 0,3 0,6 2,5 n.d. n.d. n.d. n 7,5 n.d. 6,4 71,7 n.d. n.d. 0,3 4,4 0,3 0,2 0,3 8,8 n.d. n.d. n.d. 443 RESIN PC12 SEM-EDX at MiTAC, UA, Antwerp a-g Koninklijke Musea voor Kunst en Geschiedenis (KMKG), Brussel (BE) h-m Vrije Universiteit (VU), Amsterdam (NL) PC12 cat.no. place inventory no. period a 2793 Rome – IT R1609/49 I b 2794 Rome – IT R1609/56 I c 2795 Rome – IT R1609/57 I d 2800 Rome – IT R1609/59 I e 2796 Rome – IT R1609/69 I f 2791 Rome – IT R1609/77 I g 2801 Rome – IT R1609/106 I h 2905 Hoogeloon – NL no number II i 2907 Hoogeloon – NL KA 80-1 II j 2904 Hoogeloon – NL KA 87 II k 2903 Hoogeloon – NL KA 80-2 II l 2902 Hoogeloon – NL KA 85-1 II m 2906 Hoogeloon – NL KA 85-2 II hue blue blue-green blue-green blue brown purple purple green green green green green green function arch. decoration - rod arch. decoration - rod arch. decoration - rod arch. decoration - rod arch. decoration - rod arch. decoration - rod arch. decoration - rod vessel - form IIB.1 vessel - form IIB.4 vessel - form IIB.1 vessel - form IIB.4 vessel - form IIB.1 vessel - form IIB.1 PC12 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 17,4 0,5 2,5 68,9 n.d. n.d. 0,6 0,8 7,6 n.d. 0,2 1,3 0,1 n.d. n.d. b 18,0 0,6 2,4 68,7 n.d. n.d. 0,6 0,9 7,4 n.d. 0,2 1,1 n.d. n.d. n.d. c 18,4 0,5 2,6 68,3 n.d. n.d. 0,6 0,9 7,3 n.d. 0,2 1,1 n.d. n.d. n.d. d 17,5 0,6 2,2 68,4 n.d. n.d. 0,9 0,7 8,1 n.d. 0,3 1,4 n.d. n.d. n.d. e 15,8 0,5 2,1 70,6 n.d. n.d. 0,8 1,1 8,3 n.d. 0,4 0,4 n.d. n.d. n.d. f 16,8 0,6 2,4 68,8 n.d. n.d. 0,8 0,9 7,3 n.d. 2,0 0,5 n.d. n.d. n.d. g 15,9 1,0 2,2 67,2 n.d. n.d. 0,7 0,9 7,8 n.d. 3,4 1,0 n.d. n.d. n.d. h 17,2 0,6 1,9 64,6 n.d. n.d. 0,7 0,7 6,1 n.d. 0,2 8,1 n.d. n.d. n.d. i 19,1 3,1 1,6 63,4 n.d. n.d. 0,9 2,0 7,5 0,2 0,4 1,8 n.d. n.d. n.d. j 17,5 0,5 1,8 63,4 n.d. n.d. 0,6 0,7 5,4 n.d. 0,1 9,9 n.d. n.d. n.d. k 19,3 3,1 1,6 63,3 n.d. n.d. 0,9 2,0 7,5 0,2 0,4 1,8 n.d. n.d. n.d. l 16,4 0,6 1,9 62,1 n.d. n.d. 0,6 0,7 6,0 n.d. 0,3 11,3 n.d. n.d. n.d. m 16,4 0,5 1,7 61,6 n.d. n.d. 0,7 0,6 5,6 n.d. 0,2 12,6 n.d. n.d. n.d. 444 RESIN PC13 SEM-EDX at MiTAC, UA, Antwerp a-k Rheinishes Landesmuseum, Trier (DE) PC13 cat.no. place a 2073 Trier-Palais Kesselstatt – DE b 2073 Trier-Palais Kesselstatt – DE c 2073 Trier-Palais Kesselstatt – DE d 2056 Trier-Palais Kesselstatt – DE e 2051 Trier-Palais Kesselstatt – DE f 2031 Trier-Saarstrasse – DE g 2867 Wasserbillig – LU h 1958 Möhn – DE i 2048 Trier-Barbarathermen – DE j 2041 Trier – DE k 2022 Trier-Fausenburg – DE PC13 Na2O MgO Al2O3 SiO2 inventory no. EV 22-343 182a-2 EV 22-343 182a-1 EV 22-343 182a-3 EV 22-343 382.561.193 16600 9137 9233 2140 EV 22-343 (38.2558) 8894 period IV IV IV IV IV III-IV III-IV III-IV IV III-IV III-IV hue green green green green green green green green green green green function production glass from crucible production glass from crucible production glass from crucible pendant - vessel-shaped bracelet - D1 bracelet - C1var.3 bracelet - C3 var.3 bracelet - A1 bracelet - D1 finger ring finger ring – A1 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 15,8 2,5 3,0 58,3 0,2 n.d. 0,4 2,5 8,4 0,6 1,5 6,9 n.d. n.d. n.d. b 17,2 1,0 2,6 62,5 n.d. n.d. 0,6 1,0 5,8 0,5 1,6 7,3 n.d. n.d. n.d. c 18,0 1,0 2,0 61,2 n.d. 0,2 0,9 0,6 6,9 0,2 1,6 7,2 n.d. n.d. n.d. d 14,7 1,2 2,8 57,9 n.d. n.d. 0,4 1,5 6,7 0,4 1,4 8,8 0,6 3,5 3,5 e 18,7 1,1 2,1 63,7 n.d. n.d. 0,8 0,7 6,0 0,4 1,8 4,7 n.d. n.d. n.d. f 16,3 0,5 2,1 62,3 n.d. n.d. 0,6 0,8 6,5 n.d. 0,3 10,5 n.d. n.d. n.d. g 16,7 0,6 2,2 67,0 n.d. n.d. 0,6 0,8 7,2 n.d. 0,5 4,4 n.d. n.d. n.d. h 16,6 0,6 2,0 63,9 n.d. n.d. 0,6 0,7 6,3 n.d. 0,4 8,8 n.d. n.d. n.d. i 17,8 0,8 1,9 61,1 n.d. n.d. 0,8 0,7 6,1 0,2 1,3 7,3 0,2 1,8 1,8 j 17,3 1,0 2,2 60,3 n.d. n.d. 0,7 0,7 6,1 0,4 1,7 7,8 0,1 1,7 1,7 k 17,0 1,1 2,2 60,6 n.d. n.d. 0,8 0,8 6,1 0,3 1,6 8,2 0,2 1,1 1,1 445 RESIN PC14 SEM-EDX at MiTAC, UA, Antwerp a-m Musée des Antiquités Nationales (M.A.N.), Saint-Germain-en-Laye (FR) PC14 cat.no. place inventory no. period a Ste Menehould – FR 83401 - B1 III-IV b Ste Menehould – FR 83401 - B2 III-IV c Ste Menehould – FR 83401 - B3 III-IV d Ste Menehould – FR 83401 - B4 III-IV e Ste Menehould – FR 83401 - B5 III-IV f Ste Menehould – FR 83401 - B6 III-IV g 1604 Ste Menehould – FR 83401 - P1 III-IV h 1600 Ste Menehould – FR 83401 - P2 III-IV i 1602 Ste Menehould – FR 83401 - P3 III-IV j 1595 Ste Menehould – FR 83401 - P4 III-IV k 1570 Ste Menehould – FR 83401 - P5 III-IV l 1508 Ste Menehould – FR 83401 - Q88 III-IV m 1576 Ste Menehould – FR 83401 - Q117 III-IV hue green green green blue-green blue-grey blue-grey purple greyish green green green brown red function bracelet bracelet bracelet bracelet bracelet bracelet bracelet raw glass bracelet raw glass bracelet bracelet bracelet PC14 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 17,8 0,6 2,0 66,7 n.d. n.d. 0,8 0,7 6,6 n.d. 0,3 4,4 0,3 n.d. n.d. b 18,6 0,5 2,1 66,5 n.d. n.d. 0,8 0,8 6,3 n.d. 0,3 3,9 n.d. n.d. n.d. c 18,3 0,6 2,0 66,5 n.d. n.d. 0,8 0,7 6,5 n.d. 0,3 4,1 0,2 n.d. n.d. d 17,9 0,6 2,4 65,8 n.d. n.d. 0,7 0,8 6,3 n.d. 0,3 3,7 0,3 1,2 1,2 e 16,7 0,5 4,4 65,5 n.d. n.d. 0,8 0,7 6,5 n.d. 0,3 1,8 0,2 2,6 2,6 f 18,3 0,5 2,3 66,7 n.d. n.d. 0,8 0,7 6,6 n.d. 0,3 2,2 0,2 1,3 1,3 g 16,3 0,6 2,4 66,8 n.d. n.d. 0,5 0,7 8,4 n.d. 3,6 0,7 n.d. n.d. n.d. h 0,1 3,0 10,7 57,7 n.d. n.d. n.d. 4,1 23,1 0,4 0,3 0,6 n.d. n.d. n.d. i 11,5 1,6 3,6 68,1 n.d. n.d. n.d. 4,5 6,5 0,3 0,4 2,2 0,2 1,0 1,0 j n.d. 1,4 15,2 41,8 n.d. 0,6 n.d. 1,5 38,6 0,2 0,4 0,3 n.d. n.d. n.d. k 16,4 0,6 1,9 64,5 n.d. n.d. 0,6 0,8 6,8 n.d. 0,4 8,1 n.d. n.d. n.d. l 16,7 0,7 2,3 70,1 n.d. n.d. 0,7 1,3 6,7 0,1 0,3 1,0 n.d. n.d. n.d. m 16,7 0,6 2,0 62,6 n.d. n.d. 0,6 0,9 6,5 n.d. 0,4 7,5 0,4 2,0 2,0 446 RESIN PC15 SEM-EDX at MiTAC, UA, Antwerp a-g London Archaeological Archive Centre (LAARC), London (UK) PC15 cat.no. place inventory no. period a 4142 London – UK 23121 I-II b 4148 London – UK BBB 05 I-II c London – UK 2406a I-II d* 4143 London – UK 18.022 I-II e 4145 London – UK A22557 I-II f 4091 London – UK 2685 I g 4090 London – UK 14499 I hue olive green greyish-green ultramarine blue black olive green amber brown blue function vessel vessel vessel vessel vessel arch. decoration - rod rod *polished away and therefore resampled in resin PC17 PC15 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 18,4 2,8 2,1 63,6 n.d. n.d. 0,8 2,1 7,6 0,3 0,4 1,7 n.d. n.d. n.d. b 15,6 0,6 2,2 64,2 n.d. n.d. 0,5 1,1 6,7 n.d. 0,4 8,7 n.d. n.d. n.d. c 15,9 0,4 1,8 72,3 n.d. n.d. 0,8 0,6 5,9 n.d. 0,9 1,1 0,3 n.d. n.d. d 2,3 1,9 17,0 54,0 2,9 3,9 n.d. 1,8 8,6 1,6 0,6 4,4 1,0 n.d. n.d. e 17,5 3,8 1,6 63,7 0,4 n.d. 0,8 2,7 7,7 0,2 0,3 1,3 n.d. n.d. n.d. f 18,4 0,4 4,8 69,7 n.d. n.d. 1,0 1,5 3,7 n.d. n.d. 0,4 n.d. n.d. n.d. g 17,8 0,5 2,5 68,7 n.d. n.d. 0,6 0,9 7,7 n.d. 0,2 1,1 n.d. n.d. n.d. 447 RESIN PC16 SEM-EDX at MiTAC, UA, Antwerp a-f London Archaeological Archive Centre (LAARC), London (UK) PC16 cat.no. place inventory no. period hue a 4130 London – UK 98.94/6 I-II brownish green b 4126 London – UK 98.94/7 I-II brownish green c 4122 London – UK 98.94/8 I-II bottle green d 4127 London – UK 98.94/26 I-II olive green e 4118 London – UK 98.94/30 I-II brownish green + white/red/blue/yellow f 4119 London – UK 98.94/34 I-II brownish green function bracelet – Romano-British bracelet – Romano-British bracelet – Romano-British bracelet – Romano-British bracelet – Romano-British bracelet – Romano-British PC16 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 20,3 5,9 1,3 64,7 n.d. n.d. 0,8 2,3 3,5 0,2 n.d. 0,9 n.d. n.d. n.d. b 20,3 4,5 1,6 65,5 n.d. n.d. 1,1 2,4 3,3 0,3 n.d. 1,0 n.d. n.d. n.d. c 19,1 5,7 1,4 66,0 n.d. n.d. 1,0 2,2 3,4 0,2 n.d. 0,9 n.d. n.d. n.d. d 21,3 2,0 6,6 55,2 n.d. 0,6 1,0 3,3 5,7 1,0 n.d. 3,3 n.d. n.d. n.d. e 20,9 5,2 2,3 63,4 n.d. n.d. 0,8 1,9 3,7 0,3 n.d. 1,4 n.d. n.d. n.d. f 20,4 4,4 4,2 60,6 n.d. n.d. 0,7 3,3 3,9 0,4 n.d. 1,9 n.d. n.d. n.d. 448 RESIN PC17 SEM-EDX at MiTAC, UA, Antwerp a London Archaeological Archive Centre (LAARC), London (UK) PC17 cat.no. place inventory no. period a 4148 London – UK BBB 05 I hue greyish-green function vessel PC17 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,4 0,6 2,5 64,4 n.d. n.d. 0,5 1,0 6,1 n.d. 0,4 7,9 n.d. n.d. n.d. 449 RESIN PC18 SEM-EDX at MiTAC, UA, Antwerp a-k Muze Prokrajinska, Celje (SLO) PC18 cat.no. place a 716 Oresac – HR b 705 Oresac – HR c 693 Oresac – HR d 694 Oresac – HR e 701 Oresac – HR f 704 Oresac – HR g 713 Oresac – HR h 683 Oresac – HR i 684 Oresac – HR j 686 Oresac – HR k 687 Oresac – HR inventory no. no.2 no.3 no.7 no.8 no.12 no.15 no.27 no.31 no.32 no.34 no.35 period IV IV IV IV IV IV IV IV IV IV III-IV hue green green green green green green black green green green green function bracelet - D1 bracelet - D1 bracelet - D5 bracelet - D2a bracelet - D1 bracelet - D1 bracelet - D1 bracelet - D5 bracelet - D2a bracelet - D3 bracelet - A2 PC18 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,8 0,6 2,2 62,7 n.d. n.d. 0,8 0,6 7,8 n.d. 0,9 7,5 n.d. n.d. n.d. b 16,7 0,9 2,2 60,7 n.d. n.d. 0,7 0,9 7,6 n.d. 0,3 9,9 n.d. n.d. n.d. c 17,2 0,8 1,7 59,1 n.d. n.d. 0,7 0,5 5,9 0,1 0,8 13,1 n.d. n.d. n.d. d 15,6 0,7 2,1 60,3 n.d. n.d. 0,6 0,8 8,6 n.d. 0,9 10,3 n.d. n.d. n.d. e 18,1 0,6 1,7 62,1 n.d. n.d. 0,9 0,5 5,7 n.d. 0,8 9,5 n.d. n.d. n.d. f 17,4 0,7 1,8 59,3 n.d. n.d. 0,8 0,7 5,9 0,1 0,8 12,5 n.d. n.d. n.d. g 17,8 0,7 1,8 57,1 n.d. n.d. 0,8 0,6 6,3 n.d. 0,3 14,6 n.d. n.d. n.d. h 17,7 0,7 1,8 60,1 n.d. n.d. 0,8 0,7 6,0 n.d. 0,8 11,2 n.d. n.d. n.d. i 17,8 0,8 1,9 63,6 n.d. n.d. 0,9 0,6 5,7 0,1 1,2 7,3 0,1 n.d. n.d. j 18,3 0,6 1,8 60,9 n.d. n.d. 0,9 0,6 5,7 0,1 0,9 10,2 n.d. n.d. n.d. k 17,4 0,6 2,0 65,7 n.d. n.d. 0,8 0,6 7,3 n.d. 0,5 5,1 n.d. n.d. n.d. 450 RESIN PC19 SEM-EDX at MiTAC, UA, Antwerp a-l Muze Prokrajinska, Celje (SLO) PC19 cat.no. place a 3113 Celje – SL b 3115 Celje – SL c 3116 Celje – SL d 3174 Trebnje - SL e 3170 Ribnica - SL f 3163 Ribnica - SL g 3173 Ribnica - SL h 3155 Ribnica - SL i 3165 Ribnica - SL j 3172 Ribnica - SL k 3168 Ribnica - SL l 3167 Ribnica - SL inventory no. R24440/S 296 R24633/PN3166 R24628/PN4003 R10/06 S159/PN1087 S186 S298 S304 S490 S492 S502 S544/PN3220 period IV III-IV III-IV III-IV I-III I I-III III-IV III-IV III-IV IV hue black/olive green black/olive green black/olive green black/green black/green black/olive green black/olive green black/purple black black/olive green black/green black PC19 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO function bracelet - D1 bead - Trilobitenperle raw glass bead - melon bead bracelet - Spaer B2b counter - small vessel counter - large bracelet - B1 bead - Trilobitenperle bead - Trilobitenperle bracelet - D1 Fe2O3 CuO PbO ZnO a 16,7 0,7 2,0 64,7 n.d. n.d. 0,8 0,8 7,2 n.d. 0,4 6,6 n.d. n.d. n.d. b 16,2 0,7 2,2 60,5 n.d. n.d. 0,7 0,7 7,9 n.d. 0,7 10,4 n.d. n.d. n.d. c n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. d 15,8 0,5 1,6 59,2 n.d. n.d. 0,8 0,5 5,8 n.d. 0,2 15,6 n.d. n.d. n.d. e 17,7 0,9 2,0 63,9 n.d. n.d. 0,8 0,7 6,4 0,1 0,6 6,9 n.d. n.d. n.d. f 15,1 0,4 1,9 64,2 n.d. n.d. 0,7 0,8 6,6 n.d. 0,4 9,8 n.d. n.d. n.d. g 20,6 2,1 2,9 63,9 n.d. 0,1 0,9 1,5 5,9 0,3 0,2 1,6 n.d. n.d. n.d. h 16,3 1,7 2,3 65,1 n.d. n.d. 0,9 0,8 7,8 0,2 3,9 1,2 n.d. n.d. n.d. i 17,1 0,9 1,9 62,3 n.d. n.d. 0,8 0,6 6,3 0,1 0,5 9,2 0,2 n.d. n.d. j 17,4 0,9 2,0 59,9 n.d. n.d. 0,8 0,6 4,7 0,2 0,8 12,7 n.d. n.d. n.d. k 17,3 0,8 1,8 61,6 n.d. n.d. 0,7 0,7 6,9 0,1 0,7 9,3 n.d. n.d. n.d. l 18,1 0,6 1,7 62,6 n.d. n.d. 0,8 0,6 5,9 0,1 0,9 8,5 n.d. n.d. n.d. 451 RESIN PC20 SEM-EDX at MiTAC, UA, Antwerp a-b Muze Prokrajinska, Celje (SLO) c not located, examined by courtesy of Mario Da Cruz (PT) d-j Musée Archéologique, Namur (BE) PC20 cat.no. place inventory no. a 3166 Ribnica – SL S551/PN3883 b 3157 Ribnica – SL S740 3001 c Braga – PT BRA26 d e f g h i j 177 179 192 433 495 457 458 Anthée – BE Anthée – BE Auvelais – BE Marche-les-Dames – BE Pry – BE Nismes – BE Nismes – BE period IV I-III AMN A5775/V236 AMN A5775/V238 AMN A4604c AMN A2569 AMN A1872 AMN A2823 AMN A2824 hue black/green black/olive green III-IV black/green III-IV II II III-IV III-IV III-IV III-IV black/green black/purple black/green black/green black/green black/green black/green PC20 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO function bracelet - D2 counter - small raw glass/ glass cake from crucible bracelet - A1 vessel - form IIB.4/5 vessel - form IIB.1 bracelet - A4 bracelet - A3 bracelet - C1 bracelet - B2 TiO2 MnO Fe2O3 CuO PbO ZnO a 18,4 0,9 1,9 62,3 n.d. n.d. 0,8 0,6 7,3 0,1 0,9 6,7 n.d. n.d. n.d. b 15,5 0,5 2,0 64,6 n.d. n.d. 0,8 0,8 6,6 n.d. 0,4 8,9 n.d. n.d. n.d. c 17,9 0,9 1,9 60,4 n.d. n.d. 0,9 0,5 5,4 0,1 1,0 10,8 0,2 n.d. n.d. d 16,2 0,6 1,8 62,4 n.d. n.d. 0,6 0,7 6,1 n.d. 0,2 11,3 n.d. n.d. n.d. e 19,6 1,4 2,5 62,9 n.d. n.d. 0,9 0,8 8,4 0,2 2,0 1,3 n.d. n.d. n.d. f 16,8 0,4 1,6 63,2 n.d. n.d. 0,8 0,6 5,8 n.d. 0,2 10,5 n.d. n.d. n.d. g 16,6 0,5 2,0 65,9 n.d. n.d. 0,6 0,8 6,6 n.d. 0,4 6,3 n.d. n.d. n.d. h 14,9 0,6 2,7 65,3 n.d. n.d. 0,4 1,3 6,5 0,1 0,3 7,7 n.d. n.d. n.d. i 16,4 0,6 1,9 64,4 n.d. n.d. 0,6 0,8 6,7 n.d. 0,4 8,0 n.d. n.d. n.d. j 15,7 0,6 1,9 64,3 n.d. n.d. 0,6 0,8 6,8 n.d. 0,4 8,8 n.d. n.d. n.d. 452 RESIN PC21 SEM-EDX at MiTAC, UA, Antwerp synchrotron in Hamburg 07.2009 a-j Vlaams Instituut voor het Onroerend Erfgoed (VIOE), Zellik (BE) PC21 cat.no. place inventory no. a 438 Matagne-la-Petite – BE 79.MP.48.A (Matagne 1) b 439 Matagne-la-Petite – BE 79.MP.48.B (Matagne 2) c 440 Matagne-la-Petite – BE 79.MP.48.C (Matagne 3) d 441 Matagne-la-Petite – BE 79.MP.48.D (Matagne 4) e 442 Matagne-la-Petite – BE 79.MP.48.E (Matagne 5) f 446 Matagne-la-Petite – BE 79.MP.48.F g 443 Matagne-la-Petite – BE 79.MP.48.G (Matagne 6) h 444 Matagne-la-Petite – BE 79.MP.48.H i 445 Matagne-la-Petite – BE 79.MP.48.I (Matagne 7) j 447 Matagne-la-Petite – BE 79.MP.48.J (Matagne 8) period II II II II II II II II II II hue black/green black/green black/green black/green black/green black/green black/green black/green black/green black/green function vessel - form vessel - form vessel - form vessel - form vessel vessel vessel vessel vessel - form vessel - form IIB.1 IIB.1 IIB.1 IIB.1 IIB.5 IIB.5 PC21 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,6 0,5 1,8 63,6 n.d. n.d. 0,7 0,7 6,0 n.d. 0,2 9,8 n.d. n.d. n.d. b 16,8 0,5 1,7 64,1 n.d. n.d. 0,7 0,7 6,2 n.d. 0,2 8,9 0,1 n.d. n.d. c 16,3 0,5 1,9 63,0 n.d. n.d. 0,7 0,7 6,2 n.d. 0,2 10,4 n.d. n.d. n.d. d 16,7 0,5 1,9 64,0 n.d. n.d. 0,7 0,7 6,2 n.d. 0,2 8,9 n.d. n.d. n.d. e 16,5 0,5 1,8 64,2 n.d. n.d. 0,7 0,7 6,2 n.d. 0,2 9,1 n.d. n.d. n.d. f 16,7 0,5 1,8 64,2 n.d. n.d. 0,7 0,7 6,1 n.d. 0,2 8,9 n.d. n.d. n.d. g 15,8 0,5 1,7 62,1 n.d. n.d. 0,6 0,7 5,6 n.d. 0,3 12,7 n.d. n.d. n.d. h 16,3 0,5 1,8 62,8 n.d. n.d. 0,7 0,7 6,0 n.d. 0,2 11,0 n.d. n.d. n.d. i 16,0 0,5 1,9 63,0 n.d. n.d. 0,6 0,7 6,4 n.d. 0,3 10,5 n.d. n.d. n.d. j 15,9 0,5 1,7 62,2 n.d. n.d. 0,7 0,7 6,2 n.d. 0,2 11,9 n.d. n.d. n.d. 453 RESIN PC22 picture not available SEM-EDX at MiTAC, UA, Antwerp a-l Musée Romain, Avenches (CH) PC22 cat.no. place a 3612 Avenches – CH b 3628 Avenches – CH c 3627 Avenches – CH d 3626 Avenches – CH e 3625 Avenches – CH f 3594 Avenches – CH g 3603 Avenches – CH h 3617 Avenches – CH i 3624 Avenches – CH j 3630 Avenches – CH k 3623 Avenches – CH l 3599 Avenches – CH inventory no. MRA 63/2411 (Avenches 1) MRA 88/6486-06 MRA 88/6501-9 (Avenches 2) MRA 88/6572-9 (Avenches 3) MRA 88/6809-5 MRA 88/6918-02B (Avenches 4) MRA 88/6975-05 (Avenches 5) MRA 89/7170-43 (Avenches 6) MRA 89/7170-45 MRA 89/7170-46 (Avenches 7) MRA 89/7179-03 (Avenches 8) MRA 89/7851-9 (Avenches 9) period II II II II II I-II I-II II I-II I-II I-II II hue green green green green green green green green green green green green function vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel PC22 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,7 0,5 1,9 63,7 n.d. n.d. 0,8 0,7 6,2 n.d. 0,3 9,2 n.d. n.d. n.d. b 16,6 0,5 1,8 64,1 n.d. n.d. 0,7 0,7 6,2 n.d. 0,2 9,1 n.d. n.d. n.d. c 16,6 0,5 1,7 64,2 n.d. n.d. 0,7 0,7 6,2 n.d. 0,2 9,0 n.d. n.d. n.d. d 16,5 0,5 1,7 64,2 n.d. n.d. 0,8 0,7 6,2 n.d. 0,2 9,0 0,1 n.d. n.d. e 19,1 2,5 2,1 63,6 0,2 n.d. 1,0 1,3 6,9 0,2 1,3 1,6 n.d. n.d. n.d. f 18,0 2,6 1,6 63,9 0,1 n.d. 0,9 1,9 8,6 0,1 0,5 1,6 n.d. n.d. n.d. g 17,2 2,9 1,8 63,0 n.d. n.d. 0,9 1,6 10,4 0,1 0,4 1,5 n.d. n.d. n.d. h 18,6 3,2 1,6 62,2 0,2 n.d. 0,9 2,1 8,9 0,2 0,4 1,7 n.d. n.d. n.d. i 18,5 3,4 1,7 62,1 0,2 n.d. 0,9 2,1 8,8 0,2 0,4 1,6 n.d. n.d. n.d. j 18,2 3,3 1,6 62,6 0,2 n.d. 0,8 2,1 9,0 0,2 0,4 1,7 n.d. n.d. n.d. k 14,3 3,3 1,5 63,4 0,2 n.d. 0,7 2,4 11,2 0,2 0,4 2,3 0,1 n.d. n.d. l 15,3 3,3 1,6 63,4 0,1 n.d. 0,8 2,2 10,6 n.d. 0,4 2,2 n.d. n.d. n.d. 454 RESIN PC23 picture not available SEM-EDX at MiTAC, UA, Antwerp a-e Musée Romain, Avenches (CH) f VIOE-buitendienst, Zarren (BE) g-j GAD, Nijmegen (NL) PC23 cat.no. place a 3599 Avenches – CH b 3595 Avenches – CH c 3597 Avenches – CH d 3596 Avenches – CH e 3600 Avenches – CH f 448 Menen – BE g 2978 Nijmegen – NL h 2977 Nijmegen – NL i 2973 Nijmegen – NL j 2974 Nijmegen – NL inventory no. MRA 89/7851-11 MRA 89/7851-13 (Avenches 10) MRA 90/8248-348 MRA 90/8248-349 MRA 90/8248-351 MEK-07-473 MP1-3/299 MP1-3/542 BO5-132-10 BO5-126-122 PC23 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 14,8 3,3 1,5 63,2 0,1 n.d. 0,7 2,4 11,0 0,2 0,4 2,3 n.d. n.d. n.d. b 15,1 3,3 1,6 63,7 n.d. n.d. 0,7 2,4 10,3 0,1 0,4 2,1 n.d. n.d. n.d. c 19,2 2,1 2,9 62,6 n.d. n.d. 1,1 1,2 8,5 0,3 0,3 1,6 0,1 n.d. n.d. d 20,1 0,4 2,2 66,0 n.d. 0,1 1,1 1,1 8,5 n.d. n.d. 0,4 n.d. n.d. n.d. e 19,5 1,3 2,6 65,4 n.d. n.d. 0,9 1,4 6,9 0,3 0,4 1,4 n.d. n.d. n.d. f 16,8 0,4 1,7 64,8 n.d. n.d. 0,8 0,7 6,1 n.d. 0,2 8,3 n.d. n.d. n.d. g 16,9 0,8 1,8 64,3 n.d. n.d. 0,8 0,9 6,7 n.d. 0,3 7,4 n.d. n.d. n.d. h 16,7 0,7 2,4 65,2 n.d. n.d. 0,7 0,9 6,3 0,1 0,3 6,7 n.d. n.d. n.d. i 0,8 3,1 8,6 36,8 3,0 n.d. n.d. 1,1 27,5 7,9 0,3 11,0 0,1 n.d. n.d. j 1,0 3,1 9,0 37,0 3,6 n.d. n.d. 1,2 27,2 5,8 0,4 11,8 n.d. n.d. n.d. period I-II I-II I I-II I ? II II - hue green green green green green black/green black/green black/green black/brown black/brown function vessel vessel vessel vessel vessel remelted finger ring? vessel - form IIB.1 vessel - form IIB.1 raw glass/glass cake raw glass/glass cake 455 RESIN PC24 SEM-EDX at MiTAC, UA, Antwerp a-h Israel Antiquities Authority (IAA), Jerusalem (I) PC24 cat.no. place inventory no. a 2402 Horbat Qastra – IL A2207/L6127/B6437 b 2403 Horbat Qastra – IL A2207/L5246/B5721 c 2401 Horbat Qastra – IL A2207/L6128/B6609 d 2359 Horbat Qastra – IL A2207/L6129/B6428 e 2406 Horbat Qastra – IL A2482/L6703/B62780/1; 32I f 2406 Horbat Qastra – IL A2482/L6703/B62780/1; 32K g 2405 Horbat Qastra – IL A2482/L6703/B62780/1; 32M h 2405 Horbat Qastra – IL A2482/L6703/B62780/1; 32N PC24 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl period III/IV III/IV III/IV III/IV III/IV III/IV III/IV III/IV K2O CaO hue function bead jewellery/bead jewellery/pendant bead bead bead bead bead black black black black black black black black TiO2 MnO Fe2O3 CuO PbO ZnO a 18,9 1,2 1,3 63,5 n.d. n.d. 1,0 0,5 5,9 0,2 0,9 6,1 n.d. 0,5 n.d. b n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. c n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. d 16,5 1,3 1,9 61,9 n.d. 0,2 0,7 0,6 6,3 0,5 1,3 8,7 n.d. n.d. n.d. e 16,5 1,4 1,9 61,6 n.d. 0,3 0,8 0,7 6,0 0,5 1,3 9,0 n.d. n.d. n.d. f 15,1 0,7 1,8 66,5 n.d. 0,2 0,7 0,8 8,6 0,1 0,7 4,7 n.d. n.d. n.d. g 16,6 1,4 1,8 59,5 n.d. 0,4 0,7 0,6 6,2 0,4 1,2 11,2 n.d. n.d. n.d. h 13,4 1,8 1,9 71,7 n.d. 0,4 1,0 0,7 5,7 0,2 0,6 2,6 n.d. n.d. n.d. 456 RESIN PC25 SEM-EDX at MiTAC, UA, Antwerp a-h Israel Antiquities Authority (IAA), Jerusalem (I) PC25 cat.no. place inventory no. a 2412 Horbat Qastra – IL A2482/L6704/B62731-1; 11/18E b 2336 Horbat Qastra – IL A2207/L5257/B5750E c 2343 Horbat Qastra – IL A2207/L5257/B5750F d 2346 Horbat Qastra – IL A2482/6689/62623/5C e Horbat Qastra – IL A2482/L6723/B62620-12 f 2372 Horbat Qastra – IL A2482/L6725/B62633/9A g 2341 Horbat Qastra – IL A2482/L6725/B62635-3C h Horbat Qastra – IL A2482/L6725/B62635-3D period III/IV III/IV III/IV III/IV III/IV III/IV III/IV III/IV PC25 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO hue black black black/grey black/blue black/red black black black/blue function bead bead bead bead bracelet bead bead bracelet a 20,3 2,1 1,5 64,4 n.d. 0,4 1,2 0,6 5,2 0,2 0,8 3,3 n.d. n.d. n.d. b 13,1 1,1 1,7 71,1 0,3 0,4 0,7 0,6 7,6 0,2 2,0 1,2 n.d. n.d. n.d. c 17,3 0,9 1,4 61,3 0,1 0,3 0,8 0,6 6,6 0,2 0,9 9,6 n.d. n.d. n.d. d 14,3 0,8 1,6 60,5 n.d. 0,5 0,7 0,8 8,1 0,1 0,7 11,1 0,6 n.d. n.d. e 18,3 1,0 2,0 67,2 n.d. 0,4 0,8 0,7 7,2 0,2 0,8 1,3 n.d. n.d. n.d. f 16,2 0,9 1,8 62,3 n.d. 0,3 0,7 0,6 9,2 0,2 0,9 7,0 n.d. n.d. n.d. g 13,8 0,7 1,7 62,8 n.d. 0,2 0,7 1,3 8,3 0,1 0,5 9,8 n.d. n.d. n.d. h 16,8 0,7 1,3 70,0 n.d. 0,5 1,1 0,5 6,9 0,2 0,8 1,3 n.d. n.d. n.d. 457 RESIN PC26 SEM-EDX at MiTAC, UA, Antwerp a-g VIOE-buitendienst, Zarren (BE) PC26 cat.no. place a 494 Oudenburg – BE b 493 Oudenburg – BE c 492 Oudenburg – BE d 467 Oudenburg – BE e Oudenburg – BE f Oudenburg – BE g Oudenburg – BE inventory no. OU.SPEG.7919 76.OU.43 76.OU.66 OU.SPEG.1786 OU.SPEG.4911 OU.SPEG.22429 OU.SPEG.4000F period II II II III III III III hue black/yellowish green black/yellowish green black/yellowish green black/olive green black/olive green black black function vessel – form IIB.1 vessel – form IIB vessel – form IIB bracelet – C3 bracelet bracelet bracelet PC26 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,2 0,7 1,8 64,5 n.d. 0,2 0,6 1,0 6,6 0,2 0,4 7,9 n.d. n.d. n.d. b 16,5 0,7 1,7 64,4 n.d. 0,2 0,7 1,0 6,4 0,1 0,3 7,9 n.d. n.d. n.d. c 16,6 0,8 1,7 65,2 n.d. 0,2 0,7 1,0 6,5 0,2 0,4 6,9 n.d. n.d. n.d. d 15,9 0,7 1,7 61,9 n.d. 0,2 0,6 1,0 6,0 0,1 0,3 11,5 n.d. n.d. n.d. e 16,7 0,7 1,8 67,4 n.d. 0,2 0,7 0,7 7,4 0,2 0,4 3,8 n.d. n.d. n.d. f 14,5 0,6 1,7 58,2 n.d. n.d. 0,5 0,6 6,2 0,1 0,4 8,0 n.d. 9,2 n.d. g 15,7 0,7 1,9 65,3 n.d. 0,3 0,7 0,7 7,1 0,1 0,5 7,1 n.d. n.d. n.d. 458 RESIN PC27 SEM-EDX at MiTAC, UA, Antwerp a-i VIOE-buitendienst, Zarren (BE) PC27 cat.no. place a 474 Oudenburg – BE b 475 Oudenburg – BE c 477 Oudenburg – BE d 478 Oudenburg – BE e 468 Oudenburg – BE f 485 Oudenburg – BE g 470 Oudenburg – BE h 469 Oudenburg – BE i 472 Oudenburg – BE inventory no. OU.SPEG.2100 OU.SPEG.8905 OU.SPEG.4998G OU.SPEG.22917 OU.SPEG.4960 OU.SPEG.4000N OU.SPEG.8902 OU.SPEG.24904 OU.SPEG.4000J period III III III III IV IV IV IV IV hue black/green black/bottle green black/blue-green black/olive green black black/green black black/green black function bracelet – B1 bracelet – C3 bracelet – A1 bracelet – A2 bracelet – D1 bracelet – D1 bracelet – D1 bracelet – D1 bracelet – D1 PC27 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 19,0 1,1 1,7 63,0 n.d. 0,2 0,7 0,7 6,3 0,3 1,4 5,5 n.d. n.d. n.d. b 17,0 0,8 2,1 65,5 n.d. 0,3 0,6 1,0 5,8 0,1 0,4 6,4 n.d. n.d. n.d. c 16,2 0,7 1,7 65,6 n.d. 0,7 0,6 0,6 7,9 0,1 0,6 5,1 n.d. n.d. n.d. d 16,7 0,8 1,9 65,2 n.d. 0,2 0,7 0,7 7,1 0,1 0,6 6,1 n.d. n.d. n.d. e 19,5 1,0 2,2 65,0 n.d. 0,2 0,6 0,6 5,8 0,5 1,5 3,1 n.d. n.d. n.d. f 17,3 0,8 1,7 63,1 n.d. 0,3 0,6 1,0 6,4 0,2 0,5 8,1 n.d. n.d. n.d. g 17,4 0,8 1,7 62,5 n.d. 0,2 0,6 1,0 6,4 0,1 0,5 8,8 n.d. n.d. n.d. h 19,0 1,1 1,7 61,4 n.d. 0,2 0,7 0,6 6,2 0,3 1,6 7,0 n.d. n.d. n.d. i 18,2 1,1 1,9 60,6 n.d. 0,2 0,6 0,6 5,7 0,4 1,5 9,1 n.d. n.d. n.d. j 19,0 1,1 1,7 63,0 n.d. 0,2 0,7 0,7 6,3 0,3 1,4 5,5 n.d. n.d. n.d. 459 RESIN PC28 SEM-EDX at MiTAC, UA, Antwerp a-b Department of Antiquities Cyprus, Nicosia (CYP) c-i Universiteit Gent (UGent), Gent (BE) PC28 cat.no. place inventory no. a unknown – CY PFM Larnaka RG 427 b unknown – CY PFM Larnaka RG 437 c 3886 Carthage – TN BM00/33010 d 3885 Carthage – TN BM00/33012 e 3880 Carthage – TN BM01/33008 f 3881 Carthage – TN BM01/33011 g 3882 Carthage – TN BM01/33013 h 3883 Carthage – TN BM01/33017 i 3884 Carthage – TN BM01/33009 period III/IV post IV post IV post IV post IV post IV post IV post IV hue black black black black black black black black black function pendant raw glass chunk raw glass chunk raw glass chunk raw glass chunk raw glass chunk raw glass chunk raw glass chunk raw glass PC28 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 20,2 1,0 1,7 63,0 n.d. 0,4 1,2 0,7 6,3 0,2 0,6 4,8 n.d. n.d. n.d. b 0,4 9,6 11,9 48,8 n.d. n.d. n.d. 2,4 17,6 0,8 0,1 8,3 n.d. n.d. n.d. c n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. d 7,9 0,1 6,2 71,4 n.d. n.d. 0,4 4,3 0,3 0,3 0,3 8,6 n.d. n.d. n.d. e 7,9 0,1 5,9 71,6 n.d. 0,1 0,5 4,4 0,3 0,3 0,3 8,7 n.d. n.d. n.d. f 8,0 0,1 6,2 71,4 n.d. n.d. 0,5 4,4 0,3 0,3 0,3 8,6 n.d. n.d. n.d. g 8,0 0,1 6,0 71,3 n.d. n.d. 0,5 4,4 0,3 0,3 0,3 8,8 n.d. n.d. n.d. h 5,3 0,1 10,1 75,2 n.d. n.d. 0,3 4,9 0,8 0,2 0,1 2,9 n.d. n.d. n.d. i 8,1 n.d. 6,3 71,3 n.d. n.d. 0,4 4,4 0,3 0,3 0,3 8,6 n.d. n.d. n.d. 460 RESIN PC29 SEM-EDX at MiTAC, UA, Antwerp a-f Vrije Universiteit Brussel (VUB-SKAR), Brussel (BE) g Vlaams Instituut voor het Onroerend Erfgoed (VIOE), Zellik (BE) PC29 cat.no. place inventory no. period a n.c. Florennes – BE FLO1 0 b n.c. Florennes – BE FLO2 0 c 273 Florennes – BE FLO3 III-IV d 274 Florennes – BE FLO4 III-IV e 275 Florennes – BE FLO5 III-IV f n.c. Turkey (east) no number g 291 Grobbendonk – BE 72.GRO.78 II hue very dark ultramarine blue very dark ultramarine blue black/olive green black/bottle green black/blue-green greyish black/olive green function bracelet – late La Tène bracelet – late La Tène bracelet – A1 bracelet – A2 bracelet – C1 obsidian chunk vessel – form IIB.1 PC29 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 19,2 0,6 1,7 67,8 n.d. 0,3 1,0 0,7 5,9 0,1 1,3 1,0 0,3 n.d. n.d. b 15,9 0,9 1,9 68,7 n.d. 0,1 0,8 0,7 8,5 0,1 0,7 1,2 0,3 n.d. n.d. c 17,0 0,6 1,9 66,3 n.d. 0,2 0,7 0,8 6,7 0,1 0,6 5,0 n.d. n.d. n.d. d 17,2 0,6 1,8 66,9 n.d. 0,2 0,7 0,7 6,8 0,2 0,6 4,3 n.d. n.d. n.d. e 17,3 0,7 1,7 64,7 n.d. 0,3 0,7 0,9 6,3 0,2 0,3 6,7 0,3 n.d. n.d. f 4,6 0,2 10,2 78,3 n.d. 0,1 n.d. 5,0 0,4 0,2 0,1 0,9 n.d. n.d. n.d. g 17,7 0,5 1,5 64,3 n.d. 0,2 0,9 0,5 5,7 0,1 0,3 8,1 n.d. n.d. n.d. 461 RESIN PC30 SEM-EDX at MiTAC, UA, Antwerp a-h Provinciaal Archeologisch Museum (PAM), Velzeke (B) PC30 cat.no. place inventory no. a 330 Kruishoutem – BE KK/PROSP KB 78 b 329 Kruishoutem – BE KK/90/PROSP JVW c 596 Velzeke – BE V73/C790/A1/K4 [F] d 599 Velzeke – BE VDC 95/IV/332a2 [G] e 595 Velzeke – BE V74-C907/WP3/K14Q/1 [A] f 594 Velzeke – BE V74/C907/WP3/K14 [B] (nr.7) g 598 Velzeke – BE V74/C907/WP3/K14a [C] (nr.8) h 597 Velzeke – BE V74/C907/WP3/K14a [D] period II II II II II II II II hue black/yellowish green black/bottle green black black black black black black function vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form vessel - form IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 IIB.1 PC30 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 17,2 0,8 1,6 61,9 n.d. 0,2 0,8 0,6 5,7 0,1 0,2 10,8 n.d. n.d. n.d. b 18,0 0,7 1,3 64,5 n.d. 0,2 0,7 0,5 5,5 0,1 0,2 8,2 n.d. n.d. n.d. c 16,9 0,6 1,4 63,6 n.d. 0,2 0,8 0,6 6,0 0,1 0,5 9,2 n.d. n.d. n.d. d 17,0 0,7 1,6 65,8 n.d. 0,2 0,7 0,9 6,5 0,1 0,4 6,1 n.d. n.d. n.d. e 18,1 0,7 1,3 63,5 n.d. 0,2 0,8 0,6 5,5 0,1 0,2 9,0 n.d. n.d. n.d. f 17,6 0,8 1,4 63,8 n.d. 0,2 0,8 0,6 5,5 0,1 0,2 9,1 n.d. n.d. n.d. g 16,2 0,7 1,4 62,1 n.d. 0,2 0,8 0,6 6,0 0,1 0,3 11,5 n.d. n.d. n.d. h 17,2 0,7 1,4 64,6 n.d. 0,2 0,7 0,7 5,7 0,1 0,2 8,4 n.d. n.d. n.d. 462 RESIN PC31 SEM-EDX at MiTAC, UA, Antwerp a-g GAD, Nijmegen (NL) PC31 cat.no. place a Nijmegen – NL b Nijmegen – NL c Nijmegen – NL d Nijmegen – NL e Nijmegen – NL f Nijmegen – NL g Nijmegen – NL inventory no. GN1.7.39 GN1.7.054 WW1.15.586 GN1.7.198A WW1.3.079 GN1.7.198B MP.6.76 period I-II I-II I-II I-II I-II I-II II hue blue purple green green green purple greyish green function vessel vessel vessel vessel vessel vessel raw glass PC31 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 18,2 0,7 1,9 66,9 n.d. 0,2 0,9 0,6 8,0 0,1 2,0 0,6 n.d. n.d. n.d. b 20,1 2,9 2,3 63,2 0,5 0,4 0,9 1,1 6,2 0,4 0,3 1,8 n.d. n.d. n.d. c 0,5 1,7 6,2 54,4 0,2 0,3 n.d. 1,0 21,0 1,8 2,5 10,4 n.d. n.d. n.d. d 16,6 1,7 1,9 64,9 0,1 0,2 0,7 0,8 6,4 0,6 4,5 1,5 n.d. n.d. n.d. e 16,1 1,3 1,9 62,4 n.d. 0,7 0,7 6,7 0,2 0,2 6,9 3,0 n.d. n.d. n.d. f 0,2 1,9 7,3 56,8 0,1 0,1 n.d. 1,4 16,6 1,0 1,5 13,0 n.d. n.d. n.d. g 17,5 0,5 2,1 68,5 n.d. 0,1 0,9 0,5 7,0 0,1 0,5 2,0 0,3 n.d. n.d. 463 RESIN PC32 SEM-EDX at MiTAC, UA, Antwerp a-d PGRM, Tongeren (B) PC32 cat.no. place a 334 Lauw – BE b 331 Lauw – BE c 332 Lauw – BE d 333 Lauw – BE inventory no. PGRM10538 PGRM10547 PGRM10871 PGRM10872 period II II-III II-III II-III hue olive green olive green olive green olive green function vessel – form IIB.4 bracelet – A1 finger ring – A1 bracelet – C3 PC32 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 17,4 0,6 1,5 62,1 0,1 0,3 0,8 0,5 5,7 0,1 0,1 10,8 n.d. n.d. n.d. b 16,4 0,7 2,0 64,7 0,1 0,2 0,6 0,8 7,0 0,2 0,4 6,9 n.d. n.d. n.d. c 18,7 3,2 1,6 63,4 0,5 0,3 1,1 1,8 7,1 0,2 0,4 1,7 n.d. n.d. n.d. d 12,7 0,6 3,0 58,2 n.d. 0,7 0,4 1,1 7,6 0,2 0,4 6,8 n.d. 7,9 0,4 464 RESIN PC33 SEM-EDX at MiTAC, UA, Antwerp a-f Grand Curtius, Liège (B) PC33 cat.no. place a 430 Liège – BE b 193 Bassenge – BE c 320 Izier – BE d 174 Amay – BE e 174 Amay – BE f 174 Amay – BE inventory no. Curtius L738 Curtius 2001AR1520 Curtius IO1205 Curtius 4A30a Curtius 4A30b Curtius 4A30c PC33 Na2O MgO Al2O3 SiO2 period II II II II II II hue olive green olive green olive green olive green olive green olive green function vessel – form vessel – form vessel – form vessel – form vessel – form vessel – form IIB.1 IIB.15 IIB.1 IIB.1 IIB.1 IIB.1 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,8 0,6 1,8 62,7 0,1 0,3 0,7 0,6 6,2 0,1 0,4 9,7 n.d. n.d. n.d. b 17,3 0,6 1,9 64,0 n.d. 0,2 0,8 0,6 5,9 0,2 0,3 8,3 n.d. n.d. n.d. c 17,0 0,6 1,9 64,5 0,1 0,3 0,8 0,6 6,1 0,1 0,3 7,8 n.d. n.d. n.d. d 17,6 0,5 1,7 63,6 0,1 0,2 0,8 0,6 5,8 0,1 0,3 8,9 n.d. n.d. n.d. e 17,5 0,5 1,7 63,7 0,1 0,3 0,8 0,6 5,7 0,1 0,2 8,8 n.d. n.d. n.d. f 17,6 0,6 1,7 63,6 n.d. 0,2 0,8 0,6 5,7 0,1 0,3 8,8 n.d. n.d. n.d. 465 RESIN PC34 SEM-EDX at MiTAC, UA, Antwerp LA-ICP-MS at UGent, Ghent 04.06.2009 a-l Musée Communal, Nivelles (B) PC34 cat.no. place a Liberchies – BE b Liberchies – BE c Liberchies – BE d 421 Liberchies – BE e 339 Liberchies – BE f 347 Liberchies – BE g 346 Liberchies – BE h 350 Liberchies – BE i 428 Liberchies – BE j Liberchies – BE k 464 Nivelles – BE l 463 Nivelles – BE inventory no. BV12018.165a BV12018.165b BV12044.16 BV12062.46 BV12113.30 V29R V38R V40R V56R V58R no number B no number A period II II II II I-IV I-IV I-IV I-IV II II-III II-III II hue olive green green green green purple purple olive green green green green black green function vessel vessel vessel vessel – form IIB.1 counter – medium counter – large counter – medium counter - small vessel – form IIB.1 bracelet – A2 (production waste?) finger ring – A1 vessel – form IIB.1 PC34 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,5 0,2 1,9 61,9 0,1 0,3 0,7 0,7 5,8 0,2 0,3 11,3 n.d. n.d. n.d. b 16,5 0,2 1,8 62,0 0,2 0,3 0,7 0,7 5,8 0,2 0,3 11,3 n.d. n.d. n.d. c 17,0 0,4 2,0 67,0 0,1 0,3 0,7 0,7 6,7 0,2 0,4 4,6 n.d. n.d. n.d. d 17,6 2,4 1,8 64,8 0,5 0,3 0,9 1,7 8,6 0,2 0,3 1,1 n.d. n.d. n.d. e 17,5 1,4 2,1 63,8 0,2 0,3 0,9 1,2 8,3 0,2 3,0 1,2 n.d. n.d. n.d. f 19,6 0,7 1,7 65,7 0,1 0,4 0,9 0,8 5,6 0,2 3,4 0,8 n.d. n.d. n.d. g 17,2 0,2 1,7 62,8 0,1 0,3 0,8 0,7 5,7 0,1 0,3 10,1 n.d. n.d. n.d. h 16,6 0,2 1,9 64,4 0,1 0,2 0,7 0,7 6,1 0,2 0,3 8,6 n.d. n.d. n.d. i 16,8 0,3 1,7 63,1 0,1 0,3 0,7 0,6 5,9 0,1 0,2 10,1 n.d. n.d. n.d. j 17,6 2,4 1,3 65,9 0,6 0,3 0,9 2,5 6,4 0,2 0,3 1,6 n.d. n.d. n.d. k 16,6 0,3 1,9 63,1 0,1 0,3 0,7 0,7 5,9 0,1 0,2 10,1 n.d. n.d. n.d. l 15,5 0,2 2,1 63,9 0,1 0,2 0,8 0,7 6,6 0,1 0,5 9,2 n.d. n.d. n.d. 466 RESIN PC35 SEM-EDX at MiTAC, UA, Antwerp LA-ICP-MS in Ghent 04.06.2009 synchrotron in Hamburg 07.2009 a-f Musée Romain, Avenches (CH) PC35 cat.no. place inventory no. a 3625 Avenches – CH MRA88/06809.05 b 3594 Avenches – CH MRA88/06918.02A c 3624 Avenches – CH MRA89/07170.45A d 3624 Avenches – CH MRA89/07170.45B e 3624 Avenches – CH MRA89/07170.45C f 3624 Avenches – CH MRA89/07170.45D period II ? II II II II hue green green green green green green function vessel - Form IIB.5 vessel - Form IIB.17 vessel – indefinite vessel – indefinite vessel – indefinite vessel – indefinite PC35 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 18,6 2,4 1,6 63,2 0,6 0,4 1,0 1,7 8,2 0,2 0,5 1,6 n.d. n.d. n.d. b 19,8 1,4 1,5 68,0 0,1 0,4 1,1 0,6 5,8 0,2 0,1 0,9 n.d. n.d. n.d. c 18,1 2,5 1,5 61,7 0,7 0,9 1,2 2,3 8,7 0,2 0,5 1,8 n.d. n.d. n.d. d 19,1 3,0 1,5 61,3 0,7 0,4 0,9 1,9 8,7 0,2 0,4 1,7 n.d. n.d. n.d. e 19,2 2,9 1,5 61,6 0,7 0,4 0,9 2,0 8,6 0,2 0,4 1,7 n.d. n.d. n.d. f 19,3 3,1 1,5 61,3 0,7 0,4 0,9 1,9 8,7 0,2 0,4 1,6 n.d. n.d. n.d. 467 RESIN PC36 SEM-EDX at MiTAC, UA, Antwerp LA-ICP-MS in Ghent 04.06.2009 a-j Thermenmuseum, Heerlen (NL) PC36 cat.no. place inventory no. a 2893 Heerlen – NL 25917 b 2899 Heerlen – NL 26003 c 2901 Heerlen – NL 26004 d 2895 Heerlen – NL 3236 e Heerlen – NL ? f 2896 Heerlen – NL 1856 g 2900 Heerlen – NL 26039 h 2897 Heerlen – NL 23771 i 2892 Heerlen – NL 5919 j Heerlen – NL 1812 period I-II I-II I-II II I-II I I-II I I I-II hue green olive green brown green purple olive green green black/green brown purple function vessel vessel vessel vessel - form IIB.1 vessel vessel - form IA.4 vessel vessel - form IA.4 rod vessel PC36 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 17,1 0,9 2,3 60,8 0,1 0,3 0,7 0,7 5,8 0,6 1,8 9,0 n.d. n.d. n.d. b 19,7 2,4 2,5 61,7 0,8 0,4 0,9 2,3 6,7 0,3 0,5 1,8 n.d. n.d. n.d. c 19,1 2,1 1,8 64,3 0,5 0,3 1,2 1,6 7,0 0,3 0,4 1,4 n.d. n.d. n.d. d 16,0 0,3 1,8 62,3 0,1 0,3 0,8 0,7 6,1 0,1 0,3 11,3 n.d. n.d. n.d. e 16,2 0,5 2,4 63,8 0,1 0,5 0,7 0,8 13,0 0,2 1,2 0,5 n.d. n.d. n.d. f 17,0 3,4 2,1 62,4 1,0 0,2 0,7 2,2 8,4 0,3 0,5 1,8 n.d. n.d. n.d. g 16,5 0,3 1,7 63,6 0,1 0,2 0,7 0,7 6,1 0,1 0,3 9,6 n.d. n.d. n.d. h 18,1 3,1 2,5 63,0 0,5 0,6 1,1 1,6 7,1 0,3 0,3 1,7 n.d. n.d. n.d. i 18,4 n.d. 4,8 69,9 0,1 0,2 1,1 1,4 3,6 0,2 0,1 0,4 n.d. n.d. n.d. j 17,2 0,4 2,1 67,7 0,1 0,2 0,9 0,7 7,8 0,1 2,3 0,5 n.d. n.d. n.d. 468 RESIN PC37 SEM-EDX at MiTAC, UA, Antwerp a-e AVRA, Kontich (B) f-h Vrije Universiteit Brussel (VUB), Brussel (BE) PC37 cat.no. place inventory no. a 325 Kontich – BE KFL no number b 328 Kontich – BE KFL-6603(209) c 324 Kontich – BE K87-542 d 324 Kontich – BE K87-542 e 327 Kontich – BE KFL1429 f n.c. Ham – BE NO INV. NO. g n.c. Turkey (east) NO INV. NO. h n.c. Turkey (east) NO INV. NO. period II II II II II - hue green olive green green green green UM blue greyish greyish function vessel- form IIB vessel – form IIB.13 vessel – form IIB.1 vessel – form IIB.1 vessel – form IIB vessel obsidian chunk obsidian chunk PC37 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,8 0,2 1,7 62,9 0,1 0,2 0,7 0,7 6,2 0,1 0,3 10,0 n.d. n.d. n.d. b 16,8 0,3 1,9 64,6 0,1 0,3 0,8 0,7 6,2 0,2 0,4 7,8 n.d. n.d. n.d. c 17,1 0,2 2,0 64,9 0,1 0,3 0,8 0,7 5,7 0,2 0,2 8,0 n.d. n.d. n.d. d 17,1 0,3 1,9 64,6 0,1 0,3 0,8 0,7 6,1 0,1 0,2 7,8 n.d. n.d. n.d. e 17,2 0,3 2,0 64,4 0,1 0,2 0,8 0,7 6,1 0,1 0,3 7,8 n.d. n.d. n.d. f 16,6 0,3 2,1 65,0 0,1 0,3 0,8 0,6 6,3 0,2 0,7 7,1 n.d. n.d. n.d. g 4,3 n.d. 10,7 78,4 0,1 0,1 0,1 4,8 0,5 0,1 0,1 0,7 n.d. n.d. n.d. h 4,3 n.d. 10,7 78,6 n.d. 0,1 0,1 4,7 0,5 0,1 0,1 0,8 n.d. n.d. n.d. 469 RESIN PC38 SEM-EDX at MiTAC, UA, Antwerp a-k Centre de Recherche Archéologique National (CRAN), Louvain-la-Neuve (BE) PC38 cat.no. place inventory no. period hue a 366 Liberchies – BE BV.88.T55Aa III-IV green b 369 Liberchies – BE BV.88.T55Ab III-IV green c 413 Liberchies – BE BV.92.594M I green d 407 Liberchies – BE BV.91.T692.5 I olive green e 368 Liberchies – BE BV.12224.132 III-IV green f 370 Liberchies – BE BV.12247.101 III-IV green g 425 Liberchies – BE BV.04.T6.R3 I-II brown h 372 Liberchies – BE BV.67.225 III green i 380 Liberchies – BE BV.95.T155.E7 III-IV green j 411 Liberchies – BE BV.96.T105.F4 II green k 414 Liberchies – BE BV.06.T91.A1 II olive green function bracelet bracelet vessel vessel bracelet finger ring vessel hairpin finger ring vessel vessel PC38 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 18,5 0,6 1,1 59,9 n.d. 0,4 1,0 0,4 6,2 0,2 0,1 11,8 n.d. n.d. n.d. b 15,1 1,2 2,4 65,6 0,1 0,5 1,5 5,3 2,4 0,3 1,4 4,2 n.d. n.d. n.d. c 16,0 1,0 2,1 66,3 0,1 0,2 0,6 1,4 6,7 0,2 0,4 5,1 n.d. n.d. n.d. d 16,5 0,8 2,0 66,8 n.d. 0,2 0,7 1,0 6,6 0,2 0,4 4,9 n.d. n.d. n.d. e 18,5 1,1 2,3 64,2 n.d. 0,2 0,6 0,9 5,5 0,3 1,3 5,0 n.d. n.d. n.d. f 16,7 0,7 1,8 64,2 n.d. 0,2 0,6 0,7 6,3 0,1 0,4 8,3 n.d. n.d. n.d. g 19,4 0,7 1,6 68,1 0,1 0,2 1,1 0,7 7,5 0,1 0,1 0,4 n.d. n.d. n.d. h 16,9 1,0 1,8 66,7 0,1 0,2 0,7 1,1 7,1 0,1 0,4 4,0 n.d. n.d. n.d. i 16,4 0,7 1,5 64,2 0,1 0,2 0,7 0,7 6,6 0,1 0,4 8,5 n.d. n.d. n.d. j 16,0 0,8 1,9 64,1 0,1 0,3 0,7 0,7 6,7 0,2 0,6 8,1 n.d. n.d. n.d. k 17,4 0,6 1,5 64,5 n.d. 0,2 0,8 0,7 5,9 0,1 0,2 8,1 n.d. n.d. n.d. 470 RESIN PC39 SEM-EDX at MiTAC, UA, Antwerp a-f Musée Communal, Nivelles (BE) g-i private collection, Leuven (BE) PC39 cat.no. place inventory no. a 410 Liberchies – BE BV.97.T162.B24 b 381 Liberchies – BE BV.00.T62.R4 c 417 Liberchies – BE BV.05.T40.A4 d 383 Liberchies – BE BV.04.T24.R2 e 373 Liberchies – BE BV.06.T73.B2 f 378 Liberchies – BE BV.88.T42.A1 g n.c. unknown – DE NO INV. NO. h n.c. Leuven – BE NO INV. NO. i n.c. Leuven – BE NO INV. NO. period II 0 I II-III II-III - hue olive green purple brown green green purple purple green brown function vessel – form IIB.1 bracelet – late La Tène vessel – form IA.14 bracelet – A4 hairpin bracelet – U3 raw glass vessel vessel PC39 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 18,6 2,7 1,6 62,9 0,7 0,3 0,8 2,0 8,1 0,2 0,4 1,7 n.d. n.d. n.d. b 17,6 0,3 2,0 68,3 0,2 0,3 0,8 0,8 7,3 0,1 1,8 0,5 n.d. n.d. n.d. c 18,6 0,3 2,1 68,3 0,2 0,2 1,1 0,8 7,7 0,1 0,1 0,4 n.d. n.d. n.d. d 16,6 0,5 2,0 67,4 0,1 0,2 0,8 0,8 6,8 0,1 0,4 4,3 n.d. n.d. n.d. e 16,1 0,4 2,0 64,6 0,1 0,2 0,7 0,7 6,6 0,2 0,5 7,8 n.d. n.d. n.d. f 17,5 0,4 2,0 68,8 0,1 0,4 0,8 0,7 6,5 0,1 2,3 0,4 n.d. n.d. n.d. g 5,1 4,4 2,7 61,4 0,2 0,6 n.d. 1,4 16,2 0,7 5,3 1,9 n.d. n.d. n.d. h 4,2 0,4 1,6 48,2 1,1 n.d. 0,4 5,3 n.d. 1,4 0,4 3,8 n.d. n.d. n.d. i 6,0 2,7 8,1 59,0 0,2 0,3 0,1 3,3 15,4 0,7 1,3 3,0 n.d. 31,5 0,3 471 RESIN PC40 SEM-EDX at MiTAC, UA, Antwerp LA-ICP-MS in Ghent 04.06.2009 a-g Petrie Museum of Egyptian Archaeology, London (UK) PC40 cat.no. place inventory no. a 993 unknown – EG UC 22180 b 917 Lahun – EG UC 25402 ccxxxii c A4507 unknown – EG UC 25792A d A4577 unknown – EG UC 25792B e A4578 unknown – EG UC 25792C f 812 Bahnasa – EG UC 22628b g 794 Abydos – EG UC 43242 period IV IV hue olive green olive green brown olive green olive green pale blue ultramarine blue PC40 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO function vessel – form IVD.4 bracelet – type D1 vessel vessel vessel vessel vessel Fe2O3 CuO PbO ZnO a 20,6 0,5 1,8 61,7 0,1 0,3 1,1 0,5 5,7 0,2 1,0 6,5 n.d. n.d. n.d. b 17,7 0,5 1,6 55,8 0,1 0,4 0,8 0,6 5,7 0,2 0,4 16,2 n.d. n.d. n.d. c 19,3 4,0 1,6 66,1 0,4 0,3 0,6 2,3 3,8 0,3 0,1 1,2 n.d. n.d. n.d. d 21,7 5,4 2,4 60,5 0,3 0,5 0,7 2,4 4,1 0,3 0,1 1,5 n.d. n.d. n.d. e 20,5 5,4 2,1 62,3 0,3 0,3 0,6 2,1 4,2 0,4 0,1 1,6 n.d. n.d. n.d. f 20,7 2,7 1,5 61,6 0,7 0,4 1,0 1,6 7,6 0,2 0,4 1,6 n.d. n.d. n.d. g 20,2 0,5 2,2 66,8 0,1 0,5 1,0 0,4 5,3 0,3 0,2 2,5 n.d. n.d. n.d. 472 RESIN PC41 SEM-EDX at MiTAC, UA, Antwerp LA-ICP-MS in Ghent 04.06.2009 a-f Colchester Archaeological Museum, Colchester (UK) PC41 cat.no. place inventory no. a 4027 Colchester – UK 615 b 4025 Colchester – UK 713 c 4028 Colchester – UK 206 d 4029 Colchester – UK 201 e 4030 Colchester – UK 196 f 4026 Colchester – UK 694 period II II? I I I II hue green green green green green green function vessel – form IIB.1 vessel – form B vessel – form IA vessel – form IA vessel – form IA vessel – form IIB.16 PC41 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,6 0,3 1,8 62,0 0,1 0,3 0,8 0,7 6,0 0,2 0,2 11,1 n.d. n.d. n.d. b 17,3 0,2 1,7 63,1 0,1 0,3 0,7 0,7 6,0 0,2 0,4 9,3 n.d. n.d. n.d. c 18,4 2,6 1,4 64,4 0,5 0,3 0,9 1,8 7,7 0,2 0,3 1,5 n.d. n.d. n.d. d 18,5 2,5 1,4 64,2 0,6 0,3 0,9 1,8 7,7 0,2 0,3 1,5 n.d. n.d. n.d. e 21,5 1,3 2,7 64,4 0,3 0,4 1,2 0,8 4,7 0,7 0,2 1,9 n.d. n.d. n.d. f 21,4 2,2 1,6 63,5 0,4 0,4 1,2 1,3 6,3 0,2 0,3 1,2 n.d. n.d. n.d. 473 RESIN PC42 SEM-EDX at MiTAC, UA, Antwerp a-i Musée du Malgré-Tout, Treignes (BE) PC42 cat.no. place inventory no. a 586 Treignes – BE TRII.94GC.86 b 587 Treignes – BE TRII.80.V.545 c 589 Treignes – BE TRA.N188 d 588 Treignes – BE TRII.190 e 579 Treignes – BE TR.78.MG.0312 f 581 Treignes – BE TR.30648 g n.c. Treignes – BE TRII.97.10CII/21 h n.c. Treignes – BE TRII.99.1B1C.II128 i n.c. Treignes – BE TRII.95.CALD.22 period II II II II IV IV - hue green green green green blue-green purple UM blue UM blue turquoise function vessel - form IIB.1 vessel - form IIB.1 vessel - form IIB.1 vessel - form IIB.1 bracelet - D1 bracelet - D1 vessel vessel raw glass PC42 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,8 0,8 1,8 63,4 0,1 0,3 0,8 0,6 6,0 0,1 0,3 9,0 n.d. n.d. n.d. b 17,2 0,6 1,6 64,0 n.d. 0,2 0,8 0,7 6,0 0,2 0,2 8,2 n.d. n.d. n.d. c 16,4 0,7 1,7 63,6 0,1 0,2 0,7 0,7 6,5 0,2 0,4 8,8 n.d. n.d. n.d. d 17,6 0,7 1,4 62,1 n.d. 0,3 0,8 0,5 5,6 0,1 0,3 10,2 n.d. n.d. n.d. e 17,2 0,9 1,8 65,1 0,1 0,2 0,6 0,9 6,7 0,2 0,5 5,6 n.d. n.d. n.d. f 16,4 0,9 1,8 69,5 n.d. 0,3 0,7 0,6 7,0 0,1 2,1 0,5 n.d. n.d. n.d. g 17,2 0,7 1,7 68,8 0,1 0,3 0,9 0,6 7,4 0,2 0,6 1,1 0,3 n.d. n.d. h 16,2 0,6 1,5 71,3 0,1 0,2 0,8 0,7 6,1 0,1 0,9 1,1 0,4 n.d. n.d. i 0,1 1,3 6,2 57,6 0,1 0,1 n.d. 2,5 20,5 0,4 1,5 9,4 n.d. n.d. n.d. 474 RESIN PC43 SEM-EDX at MiTAC, UA, Antwerp LA-ICP-MS in Ghent 04.06.2009 a-g Musée du Malgré-Tout, Treignes (BE) PC43 cat.no. place inventory no. a 499 Roly – BE RO1a b 506 Roly – BE RO1b c 504 Roly – BE RO1c d 505 Roly – BE RO1d e 501 Roly – BE RO1e f 503 Roly – BE RO2 g 507 Roly – BE RO3 period IV IV IV IV IV II-III II-III hue green green green green green green green function bracelet - D1 bracelet - D1 bracelet - D1 bracelet - D1 bracelet - D1 bracelet - B2 bracelet - B1 PC43 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 19,8 0,7 2,0 63,2 0,1 0,3 0,7 0,7 6,1 0,4 1,6 4,4 n.d. n.d. n.d. b 19,1 0,6 1,9 61,2 0,1 0,3 0,7 0,7 6,0 0,4 1,6 7,5 n.d. n.d. n.d. c 18,8 0,6 1,8 60,2 0,1 0,3 0,8 0,7 5,8 0,4 1,4 9,2 n.d. n.d. n.d. d 17,9 0,8 2,1 59,7 0,1 0,3 0,6 1,0 5,9 0,5 1,6 9,6 n.d. n.d. n.d. e 18,9 0,6 2,0 61,3 0,1 0,3 0,7 0,7 6,0 0,4 1,6 7,5 n.d. n.d. n.d. f 16,7 0,3 1,8 64,5 0,1 0,3 0,7 0,8 6,8 0,1 0,4 7,4 n.d. n.d. n.d. g 16,9 0,2 1,9 65,0 0,1 0,3 0,7 0,7 6,9 0,1 0,5 6,7 n.d. n.d. n.d. 475 RESIN PC44 SEM-EDX at MiTAC, UA, Antwerp a-h Römermuseum-Augusta Raurica, Augst (CH) j Vrije Universiteit Brussel (VUB), Brussel (BE) PC44 cat.no. place inventory no. a n.c. Augst – CH 1949_3835 b 3353 Augst – CH 1959_9368 c 3347 Augst – CH 1959_9370 d 3343 Augst – CH 1959_9371 e 3367 Augst – CH 1959_9373 f 3359 Augst – CH 1959_9374 g 3516 Augst – CH 1964_9670 h 3772 Augst – CH 1967_8910 i n.c. Monte- Arci, Sardinia – IT no number PC44 Na2O MgO Al2O3 SiO2 period II II II II II II II II - hue green green green green green green green green grey (layered) P2O5 SO3 Cl K2O CaO function indefinite bracelet – A2/A3 bracelet – A1 bracelet – A2/A3 bracelet – A2/A3 bracelet – A2/A3 vessel – form IIB.3 vessel – form IIB.1 chunk of obsidian TiO2 MnO Fe2O3 CuO PbO ZnO a 19,1 2,7 1,4 64,4 0,3 0,4 1,1 1,4 7,4 0,2 0,3 1,4 n.d. n.d. n.d. b 17,6 0,5 1,7 65,4 <0,1 0,3 0,8 1,1 6,1 <0,1 0,2 6,3 n.d. n.d. n.d. c 17,1 0,4 1,9 66,9 <0,1 0,3 0,8 0,5 5,7 0,2 0,2 6,0 n.d. n.d. n.d. d 17,4 0,5 1,6 65,9 <0,1 0,3 0,9 0,5 6,2 0,2 0,2 6,2 n.d. n.d. n.d. e 18,0 0,5 1,6 65,8 <0,1 0,3 0,8 0,5 6,1 0,1 0,2 5,9 n.d. n.d. n.d. f 18,3 0,4 1,7 68,3 <0,1 0,3 0,9 0,6 6,3 0,1 0,2 2,8 n.d. n.d. n.d. g 18,6 1,4 1,9 66,7 <0,1 0,3 1,0 1,0 6,6 0,2 0,3 1,9 n.d. n.d. n.d. h 16,3 0,5 1,8 61,8 <0,1 0,2 0,7 0,6 5,7 0,2 0,2 12,0 n.d. n.d. n.d. i 3,6 0,2 11,9 74,9 <0,1 n.d. <0,1 6,1 0,9 0,4 0,1 1,8 n.d. n.d. n.d. 476 RESIN PC45 SEM-EDX at MiTAC, UA, Antwerp a-i PC45 a b c d e f g h i Römermuseum-Augusta Raurica, Augst cat.no. place inventory no. 3777 Kaiseraugst – CH 1980_36023 3714 Kaiseraugst – CH 1981_3680 n.c. Augst – CH 1983_33067 Kaiseraugst – CH 1990_001 Kaiseraugst – CH 1991_002 3657 Augst – CH 2008_003_F05756.1 3667 Augst – CH 2008_003_F05756.4 3677 Augst – CH 2008_003_F05917.1 n.c. Augst – CH 1994_02_D3166_49 period II II hue green green blue-green green purple green green green yellowish grey II II IV IV IV I-IV function vessel – form IIC.2 production waste production waste production waste bracelet – Type D5 bracelet – Type D1 bracelet – Type D4 obsidian chunk PC45 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 16,8 0,5 1,9 66,2 <0,1 0,3 0,9 0,7 6,4 0,1 0,3 5,9 n.d. n.d. n.d. b 16,5 0,5 2,0 63,4 <0,1 0,2 0,8 0,7 6,1 0,1 0,3 9,3 n.d. n.d. n.d. c 17,4 0,5 2,1 69,3 <0,1 0,2 0,8 0,7 6,7 0,2 0,5 1,6 n.d. n.d. n.d. d 17,2 0,7 1,8 67,3 <0,1 0,2 0,9 1,2 6,3 0,1 0,3 4,0 n.d. n.d. n.d. e 15,7 0,5 2,0 69,3 <0,1 0,2 0,9 0,6 7,7 0,1 2,4 0,5 n.d. n.d. n.d. f 17,9 0,7 1,8 64,1 <0,1 0,3 1,0 0,5 6,3 0,2 0,8 6,4 n.d. n.d. n.d. g 17,6 0,7 1,7 62,7 <0,1 0,3 0,9 0,5 6,3 0,2 0,8 8,3 n.d. n.d. n.d. h 17,4 0,9 1,8 60,4 <0,1 0,3 1,0 0,5 5,8 0,2 1,0 10,8 n.d. n.d. n.d. i 0,3 0,3 6,8 75,3 2,6 <0,1 <0,1 5,3 4,3 0,7 1,3 3,0 n.d. n.d. n.d. 477 RESIN PC46 SEM-EDX at MiTAC, UA, Antwerp a-g PC46 a b c d e f g Römermuseum-Augusta Raurica, Augst cat.no. place inventory no. 3725 Kaiseraugst – CH 1978_5433 Kaiseraugst – CH 1978_5660 3739 Kaiseraugst – CH 1978_5661 3732 Kaiseraugst – CH 1978_5668 3730 Kaiseraugst – CH 1978_5670 3729 Kaiseraugst – CH 1978_5671 3705 Kaiseraugst – CH 1978_5672 period II II II II II II II hue blue-green blue-green blue-green blue-green blue-green blue-green yellowish green function raw glass - glass raw glass - glass raw glass - glass raw glass - glass raw glass - glass raw glass - glass raw glass - glass workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b PC46 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 17,1 0,5 2,1 68,1 <0,1 0,3 0,9 0,8 6,9 0,2 0,5 2,7 n.d. n.d. n.d. b 17,2 0,4 1,9 70,3 <0,1 0,2 1,0 0,7 7,0 0,1 0,3 0,8 n.d. n.d. n.d. c 17,6 0,5 1,8 67,5 <0,1 0,3 0,9 0,7 6,4 0,1 0,3 4,0 n.d. n.d. n.d. d 18,0 0,4 1,7 68,8 <0,1 0,3 0,9 0,6 6,4 0,1 0,3 2,4 n.d. n.d. n.d. e 16,6 0,5 1,9 68,2 <0,1 0,2 0,8 1,0 6,5 0,1 0,3 3,7 n.d. n.d. n.d. f 17,0 0,5 1,8 66,6 <0,1 0,3 0,9 0,7 6,7 0,1 0,3 5,1 n.d. n.d. n.d. g 17,7 0,5 1,8 67,7 <0,1 0,2 0,9 0,8 6,3 0,2 0,3 3,6 n.d. n.d. n.d. 478 RESIN PC47 SEM-EDX at MiTAC, UA, Antwerp a-g PC47 a b c d e f g Römermuseum-Augusta Raurica, Augst cat.no. place inventory no. 3703 Augst – CH 1978_5728 3723 Augst – CH 1978_5733 3696 Augst – CH 1978_5746 3695 Augst – CH 1978_5748 Augst – CH 1978_5761 3721 Augst – CH 1978_24075 3722 Augst – CH 1978_24079 period II II II II II II II hue yellowish green blue-green blue-green blue-green grey green green function raw glass - glass raw glass - glass raw glass - glass raw glass - glass raw glass - glass raw glass - glass raw glass - glass workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b workshop regio 17b PC47 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 17,6 0,6 1,8 67,7 <0,1 0,2 0,9 0,8 6,4 0,1 0,2 3,7 n.d. n.d. n.d. b 17,5 0,6 1,8 67,0 <0,1 0,2 0,9 1,0 6,2 0,1 0,3 4,4 n.d. n.d. n.d. c 16,6 0,5 1,8 66,6 <0,1 0,2 0,9 0,7 6,5 0,1 0,3 5,6 n.d. n.d. n.d. d 17,0 0,5 1,8 67,6 <0,1 0,2 0,9 0,7 6,5 0,1 0,3 4,3 n.d. n.d. n.d. e 16,7 0,5 1,9 68,7 <0,1 0,2 0,9 0,9 6,9 0,1 0,3 2,9 n.d. n.d. n.d. f 16,4 0,5 1,7 62,1 <0,1 0,2 0,8 0,6 6,0 0,1 0,2 11,3 n.d. n.d. n.d. g 17,0 0,5 1,7 66,3 <0,1 0,2 0,9 0,7 6,2 0,1 0,2 6,1 n.d. n.d. n.d. 479 RESIN PC48 SEM-EDX at MiTAC, UA, Antwerp a b c-e f-j PC48 a b c d e f g h i j not located, examined by courtesy of Mario Da Cruz (PT) Vlaams Instituut voor het Onroerend Erfgoed (VIOE), Zellik (BE) not located, examined by courtesy of Jean-Marc Doyen (BE) Musée du Malgré-Tout, Treignes (BE) cat.no. place inventory no. 3000 Braga – PT Q18_15_1639 258 Elewijt – BE NDO 0006.06 459 Roche-Ste-Anne – BE no number 252 Dourbes – BE no number 1134 Grandes-Armoises – FR no number 578 Treignes – BE TRII 80.VII 577 Treignes – BE 28542 583 Treignes – BE 28935 584 Treignes – BE TR6 576 Treignes – BE 11103 period I-IV II II IV II-III II-III II-III II II I-IV hue purple green green blue-green blue-green olive green green olive green green green function counter – medium vessel – form IIB.1 vessel – form IIB.1 bracelet – D1 bracelet – A1 bracelet – A2 bracelet – A3 vessel – form IIB.1 vessel – form IIB.1 counter – medium PC48 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 19,0 0,9 1,6 66,9 <0,1 0,3 0,9 0,8 6,4 0,2 2,3 0,8 n.d. n.d. n.d. b 17,2 0,5 1,7 63,6 <0,1 0,2 0,8 0,6 5,8 0,1 0,2 9,2 n.d. n.d. n.d. c 16,5 0,6 1,8 63,7 <0,1 0,2 0,6 0,8 6,4 0,1 0,4 9,0 n.d. n.d. n.d. d 18,4 0,9 2,0 62,1 <0,1 0,2 0,7 0,6 5,9 0,4 1,5 7,3 n.d. n.d. n.d. e 16,0 0,7 1,8 62,7 <0,1 0,4 0,6 0,6 7,3 0,1 0,5 9,2 n.d. n.d. n.d. f 16,7 0,5 1,9 64,9 <0,1 0,2 0,7 0,7 7,1 0,1 0,5 6,6 n.d. n.d. n.d. g 17,0 0,6 1,9 64,9 <0,1 0,2 0,7 0,7 6,5 0,2 0,4 7,0 n.d. n.d. n.d. h 16,8 0,6 1,9 63,8 <0,1 0,2 0,7 0,7 6,2 0,1 0,3 8,6 n.d. n.d. n.d. i 17,1 0,5 1,7 63,9 <0,1 0,2 0,7 0,6 5,8 0,1 0,2 9,0 n.d. n.d. n.d. j 0,3 2,1 9,0 55,0 <0,1 0,1 <0,1 2,3 21,5 0,6 1,5 7,5 n.d. n.d. n.d. 480 RESIN PC49 This resin included the samples from the experiment executed at PAM – Velzeke in May 2009 melting grounded obsidian with industrial soda and by colouring with iron – one with grounded iron oxide and the other with the addition of a iron nail. The result of this experiment is described in a paper addressed at the Bulletin of the AFAV and attached to the CD-ROM included to this volume. Because the samples are from newly made glass batches we did not catalogued them. We are very grateful to the people from the PAM – Velzeke for having granted permission to execute the experiment, but above all we wish to show our gratitude to Mark Taylor and David Hill as well as to François Van den Dries without whom the experiment would never been brought to a succesful end. RESIN PC50 no drawing available no picture available SEM-EDX at MiTAC of the University ofAntwerp (UA), Antwerp (BE) LA-ICP-MS at University of Ghent (UG), Ghent (BE) a-b PC50 a b Lyon (FR) cat.no. place 1224 Lyon – FR 1225 Lyon – FR inventory no. 290 291 period I I hue so-called black so-called black function raw glass (obsidian) raw glass (obsidian) PC49 Na2O MgO Al2O3 SiO2 P2O5 SO3 Cl K2O CaO TiO2 MnO Fe2O3 CuO PbO ZnO a 4,8 0,3 11,6 74,3 n.d. n.d. 0,2 4,7 0,8 0,3 n.d. 3,1 n.d. n.d. n.d. b 4,8 0,3 11,6 74,4 n.d. n.d. 0,2 4,6 0,7 0,3 n.d. 3,0 n.d. n.d. n.d. 481 482 APPENDIX 2: LIST OF BLACK GLASS SAMPLES ANALYSED BY LA-ICP-MS Due to the extensive list of minor and trace elements measured through inductively coupled plasma mass spectrometry by means of laser ablation we were not able to implement the a readable table in a paper version. We therefore refer to an excel-file on the attached cd-rom. The 20 resins analysed by means of LA-ICP-MS are PC1; 2; 3; 6; 7; 11; 12; 13; 14;15;16;18; 21; 22; 23; 24; 25; 26; 27; 28. The set encompasses a large part of Period II material but we selected those resins to compare with different periods in connection to the distinct black glass types (low iron; high iron; Levantine – Egyptian; HIMT) and various regions (Belgium; Croatia; France; Germany; Israel; Italy; Netherlands; Switzerland; Tunisia; UK) 483 APPENDIX 3: LIST OF BLACK GLASS SAMPLES ANALYSED WITH RAMAN SPECTROSCOPY Overview of 49 black appearing glass artefacts (metal oxides in % weight) (data taken from Baert et al. forthcoming, Table 2) All measured samples are soda-rich glass (~60-65% SiO2 / ~15-20% Na2O) concentration of metal oxides (wt%) inventory no. site period Fe2O3 MnO 9.2 9.0 9.0 1.7 1.5 1.7 1.7 2.3 2.2 2.1 0.3 0.2 0.2 0.5 0.4 0.4 0.4 0.4 0.4 0.4 x LA-ICP-MS sample SEM-EDX cat.no. chemical analysis method 3594-95; 3603; 3612; 3615; 3617; 3623; 3626-27; 3630 Avenches_1 Avenches_2 Avenches_3 Avenches_4 Avenches_5 Avenches_6 Avenches_7 Avenches_8 Avenches_9 Avenches_10 63/2411 88/6501-9 88/6572-9 88/06918-02 88/06975-05 89/07170-43 89/07170-46 89/7179-03 89/7851-9 89/7851-13 Avenches (CH) 291 Grobbendonk 72.GRO.78 Grobbendonk (BE) Post 150 AD 8.1 0.3 x - 329-330 Kruishoutem_1 Kruishoutem_2 KK/PROS_KB78 KK90/PROSPJVW Kruishoutem (BE) Post 150 AD 11.2 8.2 0.3 0.2 x - 438-447 Matagne_1 Matagne_2 Matagne_3 Matagne_4 Matagne_5 Matagne_6 Matagne_7 Matagne_8 Matagne_9 Matagne_10 79.MP.48.A 79.MP.48.B 79.MP.48.C 79.MP.48.D 79.MP.48.E 79.MP.48.F 79.MP.48.G 79.MP.48.H 79.MP.48.I 79.MP.48.J Matagne-la-Petite (BE) Post 150 AD 9.8 8.9 10.4 8.9 9.1 8.9 12.7 11.0 10.5 11.9 0.2 0.2 0.2 0.2 0.2 0.2 0.3 0.2 0.3 0.2 x x 1227 Marseille JV_29.92 Marseille (FR) Pre 150 AD 2.5 0.6 x x Post 150 AD Pre 150 AD visually perceived colour Raman spectrum green-yellow x figure 178 yellow-orange green-yellow green-yellow figures 178179 figure 179 green-yellow green-yellow 484 1315; 1318; 1321-25; 1329-31; 1335 510-514; 547 542-43 594-599 Olbia_1 Olbia_2 Olbia_3 Olbia_4 Olbia_5 Olbia_6 Olbia_7 Olbia_8 Olbia_9 Olbia_10 Olbia_11 Rumst_1 Rumst_2 Rumst_3 Rumst_4 Rumst_5 Tienen_1 Tienen_2 Tienen_3 037-30 294.2 037-20 C014 061-3 281 072-5 M6648 HS 13.08.1965 1002-1 056-8 88RU_51.40 88RU_86.88 88RU_120.2 88RU_107.204 88RU_107.205 TI/98/TR/02/00.3 TI/02/TR/20/078 TI/02/TR/79/059 Velzeke_1 Velzeke_2 Velzeke_3 Velzeke_4 Velzeke_5 Velzeke_6 V74-C907/WP3/K14Q/A V74/C907/WP3/K14/B V74/C907/WP3/K14/C V74/C907/WP3/K14a/D V73/C790/A1/k4/F VDC95/IV/332a2/G Pre 150 AD Olbia (FR) Post 150 AD Pre 150 AD Rumst (BE) Post 150 AD Tienen (BE) Post 150 AD Velzeke (BE) Post 150 AD 1.5 1.6 0.6 0.5 2.4 1.9 0.5 0.2 6.5 1.9 1.5 9.7 9.4 6.7 9.9 7.2 10.1 8.3 11.4 0.6 0.4 2.6 2.1 2.1 0.1 3.4 0.2 2.4 0.1 0.3 0.2 0.0 0.3 0.3 0.3 0.2 0.2 0.2 9.0 9.1 11.2 8.4 9.2 6.1 0.2 0.2 0.3 0.2 0.5 0.4 x x yellow-orange yellow-orange pink-purple pink-purple blue green purple brown green green green x x green-yellow x x green-yellow x - green-yellow figures 178179 figure 179 figure 179 485 Amay (BE) Avenches (CH) Avenches (CH) Avenches (CH) Avenches (CH) Avenches (CH) Avenches (CH) Avenches (CH) Avenches (CH) Avenches (CH) Avenches (CH) Bassenge (BE) Bonsin (BE) Bonsin (BE) Braga (PT) Florennes (BE) Florennes (BE) Florennes (BE) Florennes (BE) Florennes (BE) Grobbendonk (BE) Izier (BE) Kontich (BE) Kontich (BE) Kontich (BE) Kontich (BE) Kontich (BE) Kontich (BE) Kontich (BE) Kruishoutem (BE) Kruishoutem (BE) Maffe (BE) Matagne-la-Petite (BE) Matagne-la-Petite (BE) Matagne-la-Petite (BE) Matagne-la-Petite (BE) Matagne-la-Petite (BE) Matagne-la-Petite (BE) Matagne-la-Petite (BE) Matagne-la-Petite (BE) Nijmegen (NL) Nismes (BE) Olbia (FR) Olbia (FR) Olbia (FR) 1 1 2 3 4 5 6 7 8 9 10 1 1 2 1 1 2 3 4 5 1 1 1 2 3 4 5 6 7 1 2 1 1 2 3 4 5 6 7 8 1 1 1 2 3 W24 63/2411 88/6501-9 88/6572-9 89/07170-43 89/07170-46 88/06918-02 88/06975-05 89/7179-03 89/7851-9 89/7851-13 SN2001-AR.1520 SN2001-AR.1525A SN2001-AR.1525B BRA26 FLO1 FLO2 FLO3 FLO4 FLO5 72.GRO.78 I.O.1205 KFL no number KFL.6603 K87.542a K87.542b KFL.1429 K87.542c K87.542d 90/KK/PROSP78KB 90/KK/PROSPJVW I/7613 78.MP.A 78.MP.B 78.MP.C 78.MP.D 78.MP.E 78.MP.G 78.MP.I 78.MP.J 56184X A2824 037-30 294.2 C061-3 vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel bracelet bracelet bracelet bracelet bracelet bracelet bracelet bracelet vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel bracelet bracelet vessel vessel vessel II II II II I I I I I I I II III III II/IV I I III III III II II II II II II II II II II II II II II II II II II II II III III I I I green green green green green green green green green green green green green green olive green ultramarine blue ultramarine blue green green green bottle green green green green green green green green green olive green olive green green green green green green green green green green bottle green bottle green green green blue green-yellow green-yellow green-yellow yellow-orange yellow-orange yellow-orange yellow-orange yellow-orange yellow-orange yellow-orange green-yellow yellow-orange green-yellow green-yellow yellow-orange yellow-orange yellow-orange yellow-orange yellow-orange green-yellow green-yellow green-yellow green-yellow yellow-orange yellow-orange blue 2 2 2 7a 7a 7b 7b 6 6 6 2 2 2 2 2 2 2 2 2 2 2 7b 6 4b 0,40 0,48 0,48 0,60 0,60 0,49 0,54 0,55 0,55 0,52 0,15 0,15 0,49 0,34 0,51 0,49 0,51 0,51 0,52 0,52 0,53 0,46 0,46 0,45 0,44 0,59 0,56 0,15 y 0,52 0,50 0,50 0,39 0,37 0,47 0,46 0,44 0,44 0,46 0,04 0,06 0,50 0,56 0,48 0,49 0,48 0,48 0,47 0,47 0,47 0,48 0,49 0,47 0,47 0,41 0,44 0,11 x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x resin x chemical analysis method LA-ICP-MS calculated colour values (CIE 1931) SEM-EDX spectral group measured colour perceived colour Period commodity type inventory no. 174 3612 3627 3626 3617 3630 3594 3603 3623 3615 3595 193 200 201 3001 not catalogued not catalogued 273 274 275 291 320 325 328 324 324 327 324 324 330 329 431 438 439 440 441 446 443 445 447 not catalogued 458 1330 1331 1321 sample no. cat.no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 site appendix no. APPENDIX 4: CONCORDANCE LIST OF THE ROMAN BLACK GLASS FRAGMENTS ANALYSED WITH UV-VIS-NIR SPECTROSCOPY PC22a PC22c PC22d PC22h PC22j PC22f PC22g PC22k PC22l PC23b PC33b PC20c PC29 PC29 PC29 PC29 PC29 PC29f PC33c PC37 PC37 PC37 PC37 PC37 PC37 PC37 PC30a PC30b PC21a PC21b PC21c PC21d PC21e PC21g PC21i PC21j PC20j PC11k PC11j PC11a 486 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 1335 1325 494 492 493 484 471 not catalogued not catalogued not catalogued 465 466 2794 2800 2796 2791 512 510 509 514 513 546 544 543 539 548 545 542 541 562 578 576 577 583 not catalogued not catalogued 302 315 306 308 304 316 2989 595 594 597 596 599 598 Olbia (FR) Olbia (FR) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Oudenburg (BE) Rome (IT) Rome (IT) Rome (IT) Rome (IT) Rumst (BE) Rumst (BE) Rumst (BE) Rumst (BE) Rumst (BE) Tienen (BE) Tienen (BE) Tienen (BE) Tienen (BE) Tienen (BE) Tienen (BE) Tienen (BE) Tienen (BE) Tongeren (BE) Treignes (BE) Treignes (BE) Treignes (BE) Treignes (BE) Treignes (BE) Treignes (BE) Trou de Han (BE) Trou de Han (BE) Trou de Han (BE) Trou de Han (BE) Trou de Han (BE) Trou de Han (BE) Utrecht (NL) Velzeke (BE) Velzeke (BE) Velzeke (BE) Velzeke (BE) Velzeke (BE) Velzeke (BE) 4 5 1 2 3 4 5 6 7 8 9 10 1 2 3 4 1 2 3 4 5 1 2 3 4 5 6 7 8 1 1 2 3 4 5 6 1 2 3 4 5 6 1 1 2 3 4 5 6 037-20 CO14 OS7919 76OU66 76OU43 60OU10 OS7938 OS4911 OS3022 OS1000A OS22935 OS1169 R1609 – 56 R1609 – 59 R1609 – 69 R1609 – 77 89RU.120 88RU86.88 89RU… 88RU51.40 89RU107.205 TI/98/TR/03/22 TI/98/TR/03/26-5 TI/02/TR/20-78 TI-02-TR/20-79 TI-02-TR/20-80 TI-02-TR/20-262-3 TI-01-TR-79/059 TI-01-TR-79/142 SC128 TRII.80.VII 11103 28.542 28.935 TR6.1984-85A TR6.1984-85B A66-58 A65-7 AX-72 E76-79 A70-1 W1 9892 V74-907/WP3/K14Q/1 V74-C907/WP3/K14/2 V74-C907/WP3/K14/4 V73-C790/A1/K4 VDC95/IV/332a2 V74-C907/WP3/K14/3 vessel vessel vessel vessel vessel bracelet hairpin bracelet linen smoother linen smoother counter counter architectural decoration architectural decoration architectural decoration architectural decoration vessel vessel raw glass vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel vessel bracelet A3 counter bracelet A2 vessel vessel vessel bracelet bracelet bracelet bracelet finger ring vessel counter vessel vessel vessel vessel vessel vessel I I III III III III III III (post) IV (post) IV III III I I I I II II ? II II II II II II II II II II I III ? III II II II III III III III III I ? II II II II II II purple purple bottle green bottle green bottle green bottle green black bottle green black black black black blue-green blue brown purple green green black green green green green green green green green green green black green black green green green green green green green green black black black olive green olive green olive green olive green olive green olive green pink-purple pink-purple green-yellow green-yellow green-yellow green-yellow green-yellow indeterminate green-yellow green-yellow yellow-orange yellow-orange green-yellow green-yellow green-yellow green-yellow green-yellow green-yellow pink green-yellow green-yellow yellow-orange yellow-orange green-yellow - 3 3 4a 4a 5 3 2 2 2 2 2 2 2 2 - 0,44 0,48 0,42 0,45 0,43 0,41 0,45 0,50 0,42 0,51 0,51 0,42 0,41 0,45 0,44 0,45 0,47 - 0,18 0,26 0,50 0,51 0,49 0,49 0,53 0,49 0,49 0,48 0,46 0,48 0,46 0,52 0,50 0,48 0,49 - x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x - x x x x x x x x x x x x x x x x x x x x x x x x - 2 2 2 1 2 0,46 0,46 0,51 0,52 0,46 - 0,52 0,52 0,49 0,47 0,52 - x x x x x x - PC11f PC11c PC11 PC11 PC26a PC26c PC26b PC12b PC12d PC12e PC12f PC1c PC1b PC1f PC1a PC1e PC2m PC2i PC2f PC2l PC2g PC2j PC2h PC2k PC48f PC48c PC48g PC48h PC48i PC48i PC8b PC8c PC8f PC30e PC30f PC30h PC30c PC30d PC30g 487 488 LIST OF REFERENCES The Portland Vase, Journal of Glass Studies 32, 1990. Abegg, A. (1989) Der Römische Grabhügel von Siesbach, Kreis Birkenfeld, Trierer Zeitschrift 52, 171-278. Aerts, A., Velde, B., Janssens, K., Dijkman, W. (2003) Change in silica sources in Roman and post-Roman glass, Spectrochimica Acta Part B 58, 659-667. Alarcao, J. (1976) Verres, in: Alarcao, J., Etienne, R. (eds) Fouilles de Conimbriga. VI Céramiques diverses et verres, Paris, 154-233. Albizatti, C. (1923) Quattro vasi romani nel tesoro di S. Marco, in: Atti della Pontificia Accademia Romana di Archeologia, Memorie 1.1, 37-62. Alföldi, M.R., Barkoczi, L., Fitz, J., Poczy, K.-S., Radnóti, A., Salamon, A., Sagi, K., Szilagyi, J., Vágó, E.-B. (1957) Intercisa II. (Dunapentele) Geschichte der Stadt in der Römerzeit. (Archaeologia Hungarica, Dissertationes Archaeologicae Musei Nationalis Hungarici, Series nova 36), Budapest. Alfonsi H., Gandolfo P. (1997) L‘épave Sanguinaires A (Cahiers d‘Archéologie Subaquatique 13), 35-74. Allason-Jones L. (1995) ‗Sexing‘ small finds, in: Rush, P. (ed.) Theoretical Roman archaeology: second conference proceedings, Avebury, 22-32. Allason-Jones L. (1996) The Roman Jet in the Yorkshire Museum, York. Allason-Jones, L. (2005) Coals from Newcastle, in: Crummy, N. (ed.) Image, Craft and the Classical World. Essays in Honour of Donald Bayley and Catherine Jones, Montagnac, 181-186. Allason-Jones, L., Miket, R. (1984) The Catalogue of Small Finds from South Shields Roman Fort (Society of Antiquaries of Newcastle upon Tyne. Monography Series 2) Newcastle Upon Tyne. Allason-Jones, L., Jones, J.M. (2001) Identification of ‗jet‘ artefacts by reflected light microscopy, European Journal of Archaeology 4:2, 233-251. Allen, D. (1986) The Glass Vessels, in: Zienkiewicz, J.D., The Legionary Fortress Baths at Caerleon. II. The Finds, Cardiff, 98156. Allison, P.M. (2006) Mapping for gender. Interpreting artefact distribution inside 1st- and 2nd-century A.D. forts in Roman Germany, Archaeological Dialogues 13 (1), 1-20. Allison, P.M. (2008) The women and children inside 1st and 2nd-century forts: comparing the archaeological evidence, in: Brandl, U. (ed.) Frauen und Römisches Militär.Beiträge eines Runden Tisches in Xanten vom 7. bis 9. Juli 2005, (British Archaeological Reports International Series 1759), Oxford, 120-139. Amand, M., Mariën, M.-E. (1976) La tombe de Herstal, in: Mariën, M.E. (ed.) Inventaria Archaeologica. Corpus des Ensembles Archéologiques. Belgique 2-B11, Brussels. Amrein, H. (2001) L‘atelier de verrier d‘Avenches. L‘artisanat du verre au milieu du Ier siècle après J.-C. (Cahiers d‘Archéologie Romande 87. Aventicum XI), Lausanne. Andreau, J. (2010) L‘économie du monde romain (Le monde: une histoire). Paris. Arveiller-Dulong, V. (2006) Le mobilier en verre de la nécropole de Maule (Yvelines), Bulletin archéologique du Vexin français et du Val-d‘Oise 38, 143-163. Arveiller-Dulong, V., Arveiller, J. (1985) Le Verre d‘époque romaine au Musée archéologique de Strasbourg (Notes et Documents des Musées de France 10), Paris. Arveiller-Dulong, V., Guillot, D., Roumegoux, Y. (1994) Les fouilles de la cour de l‘archevêché de Sens : la verrerie, Revue archéologie de l‘Est 45 :1, 169-191. Arveiller-Dulong, V., Vanpeene, N., Sennequier, G. (2003) Verreries du Nord-Ouest de la Gaule: Productions et importations, in: Foy, D., Nenna, M.D. (eds) Echanges et commerce du verre dans le monde antique, Actes du colloque de l‘Association Française pour l‘Archéologie du Verre (Aix et Marseille 2001), Montagnac, 147-160. Arveiller-Dulong, V., Nenna, M.-D. (2005) Les verres antiques du musée du Louvre, Paris. Atkinson, D. (1916) The Romano-British Site on Lowbury Hill in Berkshire (University College Reading Studies in History and Archaeology), Reading. Auth, S.H. (1976) Ancient Glass at the Newark Museum. From the Eugene Schaefer Collection of Antiquities, Newark (N.J.). Bacchelli, B., Pasqualucci, R., Mastrodonato, V. (2000) Glass in interior decoration and furniture in the Roman Imperial period, in: Annales du 14e Congrès de l‘Association Internationale pour l‘Histoire du Verre (Venice-Milan 1998), Lochem, 86-88. Baert, K., Meulebroeck, W., Wouters, H., Cosyns, P., Nys, K., Thienpont, H., Terryn H. (2011) Using Raman spectroscopy as a tool for the detection of iron in glass, Journal of Raman Spectroscopy. (available online since 1 April 2011, DOI: 10.1002/jrs.2935) Bagatti, B., Malik, J.T. (1958) Gli Scavi del ‗Dominus Flevit‘ (Monte Oliveto — Gerusalemme). p.1: La necropoli del periodo romano (Pubblicazioni dello studium biblicum franciscanum. collectio maior 13), Jerusalem. Bamford, C.R. (1977) Colour Generation and Control in Glass (Glass Science and Technology 2), Amsterdam. Baradez, J. (1961) Nouvelles fouilles à Tipasa, Libyca 9, 7-200. 489 Barag, D. (2001) Stamped Pendants, in: Spaer, M. Ancient Glass in the Israel Museum. Beads and Other Small Objects, Jerusalem, 173-185. Barag, D. (2002) Late Antique and Byzantine Glass, in: Bianchi, R.S. (ed.) Reflections on Ancient Glass from the Borowski Collection. Bible Lands Museum Jerusalem, Mainz, 305-328. Barag, D. (2005) Alexandrian and Judean Glass in the Price Edict of Diocletian, Journal of Glass Studies 47, 184-186. Baratte, F. (1989) La Tombe de Lillebonne, in: Baratte, F., Painter, K. (eds) Trésors d‘orfèvrerie gallo-romains (exhibition catalogue), Paris, 106-110. Barber, E.J.W. (1991) Prehistoric Textiles: The Development of Cloth in Europe and the Near East with Special Reference to the Aegean, Princeton. Barber, B., Bowsher, D. (2000) The Eastern Cemetery of Roman London. Excavations 1983-1990 (MoLAS Monograph 4), Lavenham (Suffolk). Barovier, R., Mollo, R., Framarin, P., Sciaccaluga, M., Geotti, A. (2003) Glass Throughout Time. History and Technique of Glassmaking from the Ancient World to the Present, Milan. Barthélemy, A. de (1904) Une verrerie romaine près de Sainte-Menehould, Bulletin archéologique 1904, 82-85. Basalla, G. (1988) The Evolution of Technology, Cambridge. Bass, G.F., Brill, R.H., Lledó, B., Matthews, S.D. (2009) Serçe Limanı, Vol. II: the glass of an eleventh-century shipwreck, College Station (Texas). Baxter, M.J., Cool, H.E.M., Heyworth, M.P., Jackson, C. (1995) Compositional variability in colourless Roman vessel glass, Archaeometry 37:1, 129-141. Bayard, D., Massy, J.-L. (1983) Amiens romain (Revue archéologique de Picardie. Numéro spécial 2). Bayley, J. (1992) Metalworking ceramics, Medieval Ceramics 16, 3-10. Bayley, J. (1999) Appendix notes on the composition of coloured glasses, in: Guido, M., Welch, G.M. (eds) The Glass Beads of Anglo-Saxon England c. AD 400-700. A Preliminary Visual Classification of the more Definitive and Diagnostic Types, London, 89-93. Bayley, J. (2009) Early Medieval Lead-Rich Glass in the Bristish Isles-A Survey of the Evidence, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 255-260. Bayley, J., Butcher, S. (2004) Roman brooches in Britain: A Technological and Typological Study based on the Richborough Collection (Reports of the Research Committee of the Society of Antiquaries of London 68), London. Beck, H.C. (1927) Classification and Nomenclature of Beads and Pendants, Archaeologia 77, 1-76. Beckmann, B. (2004) Neuere Untersuchungen zum Römischen Limeskastell Miltenberg-Altstadt (Bayerisches Landesamt für Denkmalpflege 85), Kallmünz. Benea, D. (1983) Die Glasperlenwerkstatt von Tibiscum und die Handelsbeziehungen mit dem Barbaricum, in: von Freeden, U., Wieczorek (eds) Perlen. Archäologie, Techniken, Analysen. Kolloquium Vor- und Frühgeschichte 1, Bonn, 279-292. Bennett, J. (1980) Researches and discoveries, Archaeologia Cantiana 96, 406-410. Bequet, A. (1883) Nos fouilles en 1883-1884. Wancennes. Sa villa et ses cimetières antiques, Annales de la Société Archéologique de Namur 16, 363-387. Beretta, M., Di Pasquale, G. (eds) (2006) Arts et sciences. Le verre dans l‘empire romain, Florence-Milan. Berger, L. (1960) Römische Gläser aus Vindonissa (Veröffentlichungen der Gesellschaft Pro Vindonissa 8), Basel. Bernheimer, G.M. (2002) Catalogue of Classical Glass Objects, in: Bianchi, R.S. (ed.) Reflections on Ancient Glass from the Borowski Collection. Bible Lands Museum Jerusalem, Mainz, 305-328. Berlin, B., Kay, P. (1999) Basic Color Terms. Their Universality and Evolution (The David Hume Series on Philosophy and Cognitive Science Reissues), Stanford. Bertrang, A. (1954) Le Musée Luxembourgeois. Archéologie – Histoire – Folklore, Annales de l‘Institut Archéologique de Luxembourg 85, 3-155. Bezborodov, M.A. (1975) Chemie und Technologie der Antiken und Mittelalterlichen Gläser, Mainz. Bianchi, R.S. (2002) Reflections on Ancient Glass from the Borowski Collection, Bible Lands Museum Jerusalem, Mainz. Biek, L., Bayley, J. (1979) Glass and other vitreous materials, World Archaeology 2:1, 1-25. Bimson, M., Freestone, I.C. (eds) (1987) Early Vitreous Materials (British Museum. Occasional Paper 56), London. Bingham, P.A., Jackson, C.M. (2008) Roman blue-green bottle glass: chemical-optical analysis and high temperature viscosity modelling, Journal of Archaeological Science 35, 302-309. Böhme, A. (1972) Die Fibeln der Kastelle Saalburg und Zugmantel (Saalburg Jahrbuch 29), Berlin. Bolla, M. (2005) L‘inumazione a Verona, Aquileia Nost 76, 189-262. Bolla, M. (2011) Tipologie tombali e riti funerari, in: Invernizzi, R. (ed.) … et in memoriam eorum. La necropoli romana dell‘area Pleba di Casteggio, Cremona, 105-129. 490 Bonnet-Borel, F. (1997) Le verre d‘époque romaine à Avenches – Aventicum. Typologie générale (Documents du Musée Romain d‘Avenches 3), Avenches. Bonomi, S. (1996) Vetri antichi del Museo Archeologico Nazionale di Adria (Corpus delle collezioni archeologiche del vetro nel Veneto 2), Venice-Murano. Boon, G.C. (1974²) Silchester: The Roman Town of Calleva, London. Boos, A. (2000) Zusammensetzung und Interpretation des Schatzes, in: Boos, A., Dallmeier, L.-M., Overbeck, B. (eds) Der Römische Schatz von Regensburg-Kumpfmühl, Regensburg, 19-60. Boos, A., Dallmeier, L.-M., Overbeck, B. (2000) Der römische Schatz von Regensburg-Kumpfmühl, Regensburg. Borbély-Kiss, I., Fülöp, Z., Gesztelyi, T., Kiss, Á.Z., Koltay, E., Szabó, G. (1994) The PIXE-PIGE method for the classification of late Roman glass, Nuclear Instruments and Methods in Physics Research B 85, 836-839. Bornstein, M.H. (2006) Hue categorization and color naming: physics to sensation to perception, in: Pitchford, N.J., Biggam, C.P. (eds) Progress in colour studies. 2 : Psychological aspects. Amsterdam, 35-68. Boulanger, K., Jude, R., Cabart, H. (2009) Le mobilier en verre de Cutry « La Hache » (Meurthe-et-Moselle), Bulletin de l‘Association Française pour l‘Archéologie du Verre 2009, 27- 30. Boulogne, S. (2007) Reflet d‘un art populaire: les bracelets de verre coloré de la période médiévale (1171-1517) et ottomane (1517-1864) du Bilād al-Šām, unpublished doctoral dissertation at Paris-IV Sorbonne, Paris. Boulogne, S. (2008) Les bracelets de verre coloré polychromes des sites de Damas, Mașyaf, Tell abū Sarbūt et Hirbat Fāris au Bilād al-Šām mamelouk et ottoman: essai de synthese, in Bulletin d‘Etudes Orientales 57, 127-154. Branigan, K.(1977) Gatcombe Roman Villa – the Excavation and Study of a Romano-British Villa Estate, 1967-76 (British Archaeological Reports British Series 44), Oxford. Braun, C. (1983) Analysen von Gläsern aus den Hallstattzeitmit einem Exkurz über römische Fenstergläser, in: Haevernick, T.H., Frey, O.H., Glasperlen der vorrömischen Eisenzeit 1 (Marburger Studien zur Vor- und Frühgeschichte 5), Mainz, 129-178. Brécoulaki, C. (2006) Considérations sur les peintres tétrachromistes et les colores austere et floridi: l‘économie des moyens picturaux contre l‘emploi de matériaux onéreux dans la peinture ancienne, in: Rouveret, A., Dubel, S., Naas, V., Couleurs et matières dans l‘antiquité. Textes, techniques et pratiques (Études de littérature ancienne 17), Paris, 29-42. Brewer, R.J. (1986) The beads and glass counters, in: Zienkiewicz, J.D., The Legionary Fortress Baths at Caerleon. II. The Finds, Cardiff, 146-156. Brill, R.H. (1988) Scientific investigations of the Jalame glass and related finds, in: Weinberg, G.D. (ed.) Excavations Conducted by a Joint Expedition of the University of Missouri and the Corning Museum of Glass, Columbia, 257-271. Brill, R.H. (1999) Chemical analyses of early glasses, Corning (New York). Brosh, N. (2003) Glass in the Islamic Period, in: Israeli, Y., Ancient Glass in the Israel Museum. The Eliahu Dobkin Collection and Other Gifts, Jerusalem, 319-383. Brouwer, M. (1991) Glas uit de Oudheid (Rijksmuseum voor Oudheden 29), Amsterdam. Brown, R.A. (1986) The Iron Age and Romano-British Settlement at Woodcock Hall, Britannia 17, 1-58. Brown, K.R., (1991) Glazen situla met Dionysisch tafereel, in: Vrieze, J. (ed.) De Schatkamer van San Marco Venetië, ‘s Gravenhage, 85--87. Brugmann, B. (2004) Glass Beads from Anglo-Saxon Graves: A Study on the Provenance and Chronology of Glass Beads from Anglo-Saxon Graves, Based on Visual Examination, Oxford. Brulet, R. (1971) La nécropole gallo-romaine du gué d‘Hameau à Monceau-sur-Sambre, Documents et Rapports de la Société Royale Archéologique et Paléontologique de Charleroi 55, 25-85. Brulet, R. (1974) La Roche à Lomme à Dourbes. Fortification de Bas-Empire romain et refuge médiéval (Archaeologia Belgica 160), Brussels. Brun, J.-P. (2003) Le verre dans le Désert Oriental d‘Egypte: contextes datés, in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 377-387. Brun, N., Pernot, M., Velde, B. (1991) Ateliers de verriers et tesselles de mosaïque, in: Ateliers de verriers de l‘Antiquité à la période Pré-industrielle, 4e Rencontres de l‘Association Française pour l‘Archéologie du Verre (Rouen 1989), Rouen, 47-53. Bruno, V.J. (1977) Form and Colour in Greek Painting. London. Budge, E.A. (1903) An Account of the Roman Antiquities preserved in the Museum at Chesters Northumberland, London. Burger, A. (1966) The Late Roman Cemetery at Sagvar, Acta Archaeologica Academiae Scientiarum Hungaricae 18, 99-234. BusheFox, J.P. (1913) Excavations on the Site of the Roman Town at Wroxeter, Shropshire, in 1912 (Report of the Research Committee of the Society of Antiquaries of London 1), Oxford. Butcher, S. (2001) The Brooches, in: Anderson, A.S., Wacher, J.S., Fitzpatrick, A.P. (eds), The Romano-British ‗Small Town‘ at Wanborough, Wiltshire. Excavations 1966-1976 (Britannia Monograph Series 19), London, 41-74. Cabart, H. (2005) Les verres gallo-romains de Pontpierre (Moselle), Bulletin de l‘Association Française pour l‘Archéologie du Verre 2005, 18-22. 491 Cagno, S., Cosyns, P. (2009a) Compositional Analysis on the Obsidian Vessel Fragment R1610 in the Royal Museums of Art and History (KMKG-MRAH) by Using Portable-XRF, Internal report, UA-MiTAC & VUB-SKAR, Antwerp-Brussels. Cagno, S., Cosyns, P. (2009b) µ-XRF Analyses of Roman Black Glass Objects and a Vitrum Obsianum Piece in the Royal Museums of Art and History (KMKG-MRAH) by Using Portable-XRF, Internal report, UA-MiTAC & VUB-SKAR, AntwerpBrussels. Cagno, S., Cosyns, P., Janssens, K. (2010) La production de vaisselle en vitrum obsianum au Ier siècle après J.-C.. Un état de la question à partir de l‘étude du fragment R1610 (MRAH-Bruxelles), Bulletin de l‘Association Française pour l‘Archéologie du Verre 2010, 9-14. Cagno, S., Cosyns, P., Schalm, O., Van der Linden, V., Nowak, A., Wagner, B., Bulska, E., Nys, K., Janssens, K. (in preparation) The EPMA and LA-ICP-MS-Analyses of Late La Tène Glass Bracelets from Odijk, Tiel and Meteren (the Netherlands). Campbell, D.B. (2010) Women in Roman forts: Residents, visitors or barred from entry?, Ancient Warfare IV-6, 48-53. Capart, J. (1940-1954) Fouilles d‘Elkab. Documents, Livraisons I-III, Brussels. Carson, R.A.G., O‘Kelly, M.J. (1977) A Catalogue of the Roman Coins from Newgrange, Co. Meath and Notes on the Coins and Related Finds, Proceedings of the Royal Irish Academy. Section C: Archaeology, Celtic Studies, History, Linguistics, Literature 77, 35-55. Casagrande, C., Ceselin, F. (2003) Vetri antichi delle Province di Belluno, Treviso e Vicenza (Corpus delle Collezioni Archeologiche del Vetro nel Veneto 7), Venice. Casey, P.J., Davies, J.L., Evans, J. (1993) Excavations at Segontium (Caernarfon) Roman Fort, 1975-1979 (CBA research reports 90), London. Casson, L. (1998) Everyday Life in Ancient Rome, Baltimore-London. Cattelain, P., Paridaens, N. (2009) Le sanctuaire tardo-romain du ―Bois des Noël‖ à Matagne-la-Grande. Nouvelles recherches (1994-2008) et réinterprétation du site (Études d‘Archéologie 2, Artefacts 12), Brussels. Chéhab, M. H. (1985) Fouilles de Tyr: la nécropole (vol.III) (Bulletin du Musée de Beyrouth 35), Paris. Chéhab, M. H. (1986) Fouilles de Tyr: la nécropole (vol.IV) (Bulletin du Musée de Beyrouth 36), Paris. Chenet, G. (1925) Anciennes verreries d‘Argonne, Bulletin de la Société Archéologique champenoise 3, 253-286. Chenet, G. (1941) La céramique gallo-romaine d'Argonne du IVe siècle et la terre sigillée décorée à la molette, Mâcon. Chew, H. (1989) Sainte-Menehould, in: À travers le verre. Du Moyen-Âge à la Renaissance (Catalogue d‘exposition), Musée départemental des Antiquités, Rouen, 57. Claes, P. (1954) Découverte de l‘époque romaine à Taviers (Namur). Inscription lapidaire – Figurine de cheval – Céramique – Monnaies, etc., Annales de la Société Archéologique de Namur 47, 1953-1954, 225-256. Claes, P., Millau, E. (1962) Fouilles aux Bons-Villers (Liberchies). Années 1959 à 1961, Documents et Rapports de la Société Royale Archéologique et Paléontologique de Charleroi 50, 1-8. Clairmont, C.W. (1977) Catalogue of Ancient and Islamic Glass: Benaki Museum, Athens. Clarke, G. (1979) The Roman Cemetery at Lankhills (Winchester Studies 3 Part II), Oxford. Cleland, L., Stears, K., (eds) (2004) Colour in the Ancient Mediterranean World (British Archaeological Reports International Series 1267), Oxford. Colomban, P. (2008) On-site Raman identification and dating of ancient glasses: A review of procedures and tools, Journal of Cultural Heritage 9, 55-60. Colomban, P., Tournié, A., Bellot-Gurlet, L. (2006) Raman identification of glassy silicates used in ceramics, glass and jewellery: a tentative differentiation guide, Journal of Raman Spectroscopy 37, 841-852. Colomban, P., Tournié, A. (2007) On-site Raman identification and dating of ancient/modern stained glasses at the SainteChapelle, Paris, Journal of Cultural Heritage 8, 242-256. Colson, P. (1903) Découverte d‘une verrerie gallo-romaine à Sainte-Menehould, Revue archéologique 1903, 277. Constable-Maxwell (1979) Ancient Glass. Catalogue of the Constable-Maxwell Collection of Ancient Glass (sale catalogue Sotheby 4th-5th June 1979), London. Cool, H.E.M., Price, J. (1995) Roman Vessel Glass from Excavations in Colchester, 1971-1985 (Colchester Archaeological Report 8), Colchester. Cool, H.E.M., Price, J. (1998) The vessels and objects of glass, in: Cool, H.E.M., Philo, C. (eds) Roman Castleford Vol. I. The small finds (Yorkshire Archaeology 4), Wakefield, 141-198. Cool, H.E.M., Philo, C. (ed.), 1998. Roman Castleford Excavations 1974-85.Volume I: the Small Finds (Yorkshire Archaeology 4), Wakefield. Cooney, J.D. (1976) Catalogue of Egyptian Antiquities in the British Museum. IV. Glass, London. Corcoran, S. (2000) The Empire of the Tetrarchs, Imperial Pronouncements and Government AD 284-324. Oxford. Corti, C. (2007) Perle in pasta vitrea di epoca altomedievale nell‘area del delta padano: alcuni dati a confronto, in: Atti delle IX Giornate Nazionali di Studio, Il Vetro nell‘Alto Adriatico, Ferrera 13-14 dicembre 2003, Imola, 71-77. 492 Cosyns, P. (2003) Romeinse glazen armbanden, gevonden in België, in: Journée d'archéologie romaine-Romeinendag 2003, Leuven, 15-19. Cosyns, P. (2004) Les bracelets romains en verre noir: une étude typologique et contextuelle, Bulletin de l'A.F.A.V. 2004, 15-18. Cosyns, P. (2009) Sainte-Menehould (F) and Trier (G). Two Roman workshops of black glass jewellery in the North-western provinces reconsidered, in: Annales du 17e Congrès de l‘Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Brussels-Antwerp, 88-95. Cosyns, P., (in press), Beyond the Channel! That‘s Quite a Different Matter. A Comparison of Roman Black Glass from Britannia, Gallia Belgica and Germania Inferior, in: Bayley, J., Freestone, I. (eds), Glass of the Roman Empire. A celebration of the contribution of Professor Jennifer Price to the study of archaeological glass, The Wallace Collection, London 14th-15th March 2006. Cosyns, P., Hanut, F. (2005) Black glass of second to third-century date in northern Gaul: a preliminary survey, in: Annales du 16e Congrès de l'Association Internationale pour l'Histoire du Verre (London 2003), Nottingham, 113-118. Cosyns, P., Vanderhoeven, A., Vynckier, G., Janssens, K., Schalm, O., Van der Linden, V. (2005) Two fragments of mold-blown glass beakers with Greek inscriptions from Tongeren (Belgium), Journal of Glass Studies 47, 179-183. Cosyns, P., Janssens, K., Schalm, O., Van der Linden, V. (2006a) Black glass vessels and jewellery in the Roman Empire: a work in progress, in: Creemers, G., Demarsin, B., Cosyns, P. (eds) Roman Glass in Germania Inferior. Interregional Comparisons and Recent Results, International colloquium Tongeren, 13/05/2005 (Atuatuca 1), Hasselt, 30-41. Cosyns, P., Martens, M., Debruyne, T. (2006b) Contextual analysis of glass in the Roman vicus Tienen. Preliminary results, in: Creemers, G., Demarsin, B., Cosyns, P. (eds) Roman Glass in Germania Inferior. Interregional Comparisons and Recent Results, International Colloquium Tongeren, 13/05/2005 (Atuatuca 1), Hasselt, 98-105. Cosyns, P., Fontaine, S.D. (2009) Le verre noir romain du 1er siècle: un aperçu, in: Annales du 17e Congrès de l‘Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Brussels-Antwerp, 80-87. Cottam, S., Price, J. (2009) The early Roman vessel glass, in Goudineau, Ch., Brentchaloff, D., Le Camp de la Flotte d‘Agrippa a Fréjus, Paris, 185-275. Crawford, O.G.S. (1921) Man and his past, London. Crummy, N. (1979) A chronology of Romano-British bone-pins, Britannia 10, 157-163. Crummy, N. (1983) The Roman Small Finds from Excavations in Colchester 1971-9 (Colchester Archaeological Report 2), Colchester. Crummy P., Benfield, S., Crummy, N., Rigby, V., Shimmin, D. (2007) Stanway: An Élite Burial Site at Camulodunum (Britannia Monograph Series 24), London. Cunliffe, B W (1971) Excavations at Fishbourne 1961-1969; Vol 2; the Finds (Society of Antiquaries of London Research Report 27), London. Czurda-Ruth, B. (1979) Die Römischen Gläser vom Magdalensberg (Kärntner Museumsschriften 65. Archäologische Forschungen zu den Grabungen auf dem Magdalensberg 6), Klagenfurt. Czurda-Ruth, B. (1998) Die Römischen Gläser vom Magdalensberg 2, in: Piccottinni, G. (ed.) Die Ausgrabungen auf dem Magdalensberg, 1980 bis 1986 (Magdalensberg Grabungsbericht 16), Klagenfurt, 459-507. Czurda-Ruth, B. (2001) Die Rolle des Magdalensberges in der römischen Glasforschung, in: Leitner, F. (ed.) Carinthia Romana und Die Römische Welt. Festschrift für G. Piccottini zum 60. Geburtstag, Klagenfurt, 425-428. Czurda-Ruth, B. (2004) Die Römischen Gläser vom Magdalensberg 3, in: Piccottini, G. (ed.) Die Ausgrabungen auf dem Magdalensberg, 1986 bis 1990 (Magdalensberg Grabungsbericht 17), Klagenfurt, 261-300. Czurda-Ruth, B. (2007) Hanghaus 1 in Ephesos. Die Gläser (Forschungen in Ephesos VIII/7), Vienna. Da Cruz, M. (2009) Black glass jewellery from Bracara Augusta, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), AntwerpBrussels, 96-103. Dark, K. (2001) Proto-industrialization and the Economy of the Roman Empire, Montignac. Darling, M.J., Gurney, D. (1993) Caister-on-Sea, excavations by Charles Green 1951- 55 (East Anglian Archaeology 60), Gressenhall. Dasnoy, A. (1966) Quelques ensembles archéologiques du bas-empire provenant de la région namuroise (Spontin, Flavion, Tongrinne, Jamiolle, Jambes, Treigne), Annales de la Société Archéologique de Namur 53, 169-231. Dasnoy, A. (1970) La nécropole de Furfooz. Révision des notes et documents anciens, Annales de la Société Archéologique de Namur 55, 1969-1970, 121-194. Dearne, M.J., Branigan, K. (1995) The use of coal in Roman Britain, Antiquaries Journal 75, 71– 105. De Clercq, W. (2009) Lokale gemeenschappen in het Imperium Romanum. Transformaties in rurale bewoningsstructuur en materiële cultuur in de landschappen van het noordelijk deel van de civitas Menapiorum (provincie Gallia-Belgica, ca. 100 v. Chr. – 400 n. Chr.), unpublished doctoral thesis, Universiteit Gent, Ghent. Décobert (2003) Le natron égyptien au Moyen Age; II. Productions et importations, in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 125-130. DeCustatis, C. (ed.) (1997) Handbook of Applied Photometry. Woodbury–New York-Washington DC. 493 Degryse, P., Schneider, J. (2008) Pliny the Elder and Sr-Nd isotopes: tracing the provenance of raw materials for Roman glass production, Journal of Archaeological Science 35:7, 1993-2000. Degryse, P., Shortland, A. (2009) Trace elements in provenancing raw materials for Roman glass production, Geologica Belgica 12:3-4, 135-143. Degryse, P., Schneider, J. (2009) Neodymium and strontium isotopes in the provenance determination of primary natron glass production, in: Isotopes in Vitreous Materials (2009) Degryse, P., Schneider, J., Lauwers, V., Waelkens, M., Muchez, P. (2009) Radiogenic isotopes in the provenance determination of raw materials: a case of lead and glass recycling at Sagalassos (SW Turkey), Journal of Nordic Archaeological Science 16, 15-23. Del Marmol, E. (1851) Découvertes d‘antiquités dans les tumulus de Champion et dans quelques localités voisines, Annales de la Société Archéologique de Namur 2, 57-82. Demakopoulou, K.(1988) 203. Beads; 204. Mould, in: Demakopoulou, K. (ed.) The Mycenaean World. Five Centuries of Early Greek Culture, 1600-1100 BC, Athens, 219. De Meester de Ravestein, E. (18842) Musée de Ravestein. Notice, Brussels. Desbat, A. (2003) Les verres des fouilles du pseudo sanctuaire de Cybèle à Lyon, in : Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 397-403. de Schaetzen, P., (1950) La tombe belgo-romaine de Riempst, Bulletin de l‘Institut Archéologique Liégeois 67, 37-60. Destexhe, G. (1989) Le cimetière Gallo-Romain de Wanzoul, Archéologie hesbignonne 8, 1-236. Destexhe, G. (2003) Les cimetières gallo-romains et merovingiens du Jointy à Verlaine. Rapport des fouilles, Archéologie hesbignonne 17, 1-265. De Witte, H. (1977) Glazen armbanden uit de voorromeinse en romeinse periode, gevonden in België, unpublished Ma-thesis Universiteit Gent, Ghent. Diani, M.G. (1998) Contributo alla carta di distribuzione di alcune forme vitree di età romana colate a stampo e soffiate a stampo. Note su alcuni recenti ritrovamenti in Lombardia, in: Il vetro dall‘antichità all‘età contemporanea: aspetti tecnologici, funzionali e commerciali. Atti 2e giornate nazionali di studio Association Internationale pour l‘Histoire du Verre -comitato nazionale Italiano (14-15 dicembre 1996 Milano), Milan, 31-40; 295-296. Dilly, G., Mahéo, N. (1997) Verreries antiques du Musée de Picardie, Amiens-Paris. Doppelfeld, O. (1966) Römisches und fränkisches Glas in Köln (Schriftenreihe der Archäologischen Gesellschaft Köln 13), Cologne. Doyen, J.-M. (1980) La fortification romaine de la "Roche Sainte-Anne" à Nismes (Viroinval, province de Namur), in: Congrès de Comines 28-31/VIII/1980, Actes II, 269-280. Dudley, D. (1967) Excavations on Nor'Nour in the Isles of Scilly, 1962-6, Archaeological Journal 124, 28-64. Eck, T. (1891) Les deux cimetières gallo-romains de Vermand et de Saint-Quentin. Récit complet des fouilles faites durant les années 1885, 1886 et 1887, Paris. Eggers, H.J. (1940) Das römische Einfuhrgut in Pommern, Baltische Studien N. F. Band 42, 1-36. Eggers, H.J. (1951) Der römische Import im freien Germanien (Atlas der Urgeschichte 1), Hamburg. Eisen, G. (1916) Button-Beads, American Journal of Archaeology 20:3, 299-307. Eisen, G., Kouchakji, F. (1927) Glass. Its Origin, History, Chronology, Technic and Classification to the Sixteenth Century, New York. Eisen, G. (1930) Lotus- and Melon-Beads, American Journal of Archaeology 34:1, 20-43. Elderkin, K.McK. (1928) Buttons and Greek Garments, American Journal of Archaeology 32:3, 340-341. Erdich, M., Voss, H.-U. (1997) Die Perlen der Germanen des 1.-5. Jahrhunderts in Mecklenburg-Vorpommern, SchleswigHolstein und Niedersachsen, in: von Freeden, U., Wieczorek, A. (eds.) Perlen. Archäologie, Techniken, Analysen, Akten des Internationalen Perlensymposiums in Mannheim vom 141. Bis 14.November 1994, Bonn 77-93. Erdmann, E. (1977) Die Glasfunde von Mezad Tamar (Kasr Gehainije) in Israel,Saalburg Jahrbuch 34, 1-46. Fadić, I. (1998) Trasparenze Imperiali: vetri romani dalla Croazia, Milan. Fadini, N., Montevecchi, G. (2001) Un bracciale con «Trilobitenperlen» da una necropoli cesenate, in: Atti della V Giornata Nazionale di Studio Vetri di ogni tempo, Massa Martana (Perugia) 30 ottobre 1999, Milan, 51-55. Feugère, M. (1987) Une forme rare de verrerie romaine: Le socle quadrangulaire de Cruzy (Hérault) (Archéologie en Languedoc 1987), 62-63. Feugère, M. (1994) La vaisselle gallo-romaine en bronze de Vertault (Côte-d'Or). Revue Archéologique de l‘Est 45:1, 137-168. Feugère, M. (2008) Techniques, productions, consommations: le sens des objets, Facta 1, 2007, 21-30. Filarska, B. (1962) Szkła Starozytne II, Starozytne Ozdoby i elementy dekoracji w szkłe, Warsaw. Finley, M.I. (1973) The Ancient Economy, New York-London. Fischer, A. (2009) Vorsicht Glas! Die römischen Glasmanufakturen von Kaiseraugst (Forschungen in Augst 37), Augst. 494 Fiori, C., Roncuzzi-Fiorentini, I., Vandini, M. (2001) Colours and stories of the Italian mosaics, in: Claussen, N. (ed.) The Art of Ceramics: The Blend of Art and Science in Technology. Discussions of the Academy of Ceramics Forum, 2000, Ravello/Sorrento, Italy (Monographs in materials and society 5), Faenza, 149–72. Follmann-Schulz, A.-B. (1988) Die römischen Gläser aus Bonn (Beihefte der Bonner Jahrbücher 46), Cologne. Follmann-Schulz, A.-B. (1991) Fours de verriers romains dans la province de Germanie Inférieure, in: Ateliers de verriers de l'Antiquité à la période pré-industrielle. Actes des 4ième Rencontres de l‘association française pour l‘archéologie du verre, Rouen 1989, 35 – 40. Follmann-Schulz, A.-B. (1992) Die Römischen Gläser im Rheinischen Landesmuseum Bonn (Führer des Rheinischen Landesmuseums Bonn und des Rheinischen Amtes für Bodendenkmalplege 138), Cologne. Fontaine-Hodiamont, C. (1994) Une technique particulière pour la fabrication des coupes en verre mosaïque de Hollogne-auxPierres (Belgique). Le témoignage de Pline l‘Ancien (Histoire Naturelle 36, 199), Bulletin van de Koninklijke Musea voor Kunst en Geschiedenis 65, 27-63. Fontaine-Hodiamont, C. (ed.) (2010) D‘Ennion au Val Saint-Lambert. Le verre soufflé-moulé. Actes des 23e Rencontres de l‘Association Française pour l‘Archéologie du Verre, Bruxelles-Namur 2008 (Scientia Artis 5), Brussels. Fontaine, S.D. (2002) Le mobilier en verre d‘Olbia de Provence. Fouilles de Jacques Coupry, le puits public et la zone centrale (1961-1967), unpublished Ma-thesis, Université de Provence. Fontaine, S.D. (2006) Le mobilier en verre, in: Bats, M. (ed.) Fouilles à Olbia de Provence. L'époque romaine (Etudes Massaliètes 9), Aix-en-Provence, 54-61; 307-382 and pl. XIV-XVI. Fontaine, S.D. (2009) Le mobilier en verre de la Maison de la Statuette Indienne (I.8,5): contribution à l‘étude socio-économique d‘une insula de Pompéi, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 115-120. Fontaine, S., Foy, D. (2007) L‘épave Ouest Embiez 1, Var. Le commerce maritime du verre brut et manufacturé en Méditerranée occidentale dans l‘Antiquité, Revue de la Narbonnaise 40, 235-268. Forbes, R.J. (1966) Glass, in: Forbes, R.J., Studies in Ancient Technology. Volume V, Leiden, 112-216. Fortner, B., Meyer, T.E. (1997) Number by Colors: A Guide to Using Color to Understand Technical Data, New York. Fortunati Zuccala, M. (1998) Una inumazione plurima dalla necropolis romana di Lovere, in: Masseroli, S. (ed.) Vetro e vetri. Preziose iridiscenze, Milan, 115-122. Fossing, P. (1940) Glass vessels before Glass-blowing, Copenhagen. Foster, H., Jackson, C.M. (2009) The composition of ‗naturally coloured‘ late Roman vessel glass from Britain and the implications for models of glass production and supply, Journal of Archaeological Science 36, 189–204. Foster, H., Jackson, CM. (2010) The composition of late Romano-British colourless vessel glass: glass production and consumption, Journal of Archaeological Science 37 3068-3080. Foy, D. (2004) Les verres, in: Ben Abed-Ben Khader, A., Fixot, M., Bonifay, M., Roucole, S. (eds) Sidi Jdidi, I. La basilique sud (Collection de l‘Ecole Française de Rome 339), Rome, 317-329. Foy, D. (2007) Produits semi-finis colorés pour mosaïstes, orfèvres et verriers. Antiquité tardive et Moyen Age, Bulletin de l‘Association Française pour l‘Archéologie du Verre 2007, 43-52. Foy, D. (2010) Souvenirs de pèlerinages dans l'Antiquité tardive: vaisselle, ampoules et breloques de verre découvertes en Narbonnaise, in: Delestre, X., Archéologie des rivages méditerranéens (Arles 2009), Paris, 303-311. Foy, D., Sennequier, G. (1991) Ateliers de verriers de l‘Antiquité à la période pré-industrielle (Actes des 4èmes Rencontres de l‘association française pour l‘archéologie du verre, Rouen 1989), Rouen. Foy, D., Vichy, M., Picon, M. (2000) Lingots de verre en Méditerranée (IIIe s. av. J.-C.- VIIe s. ap. J.-C.), approvisionnement et mise en œuvre; les données archéologiques et les données de laboratoire (Annales du 13e Congrès de l'Association Internationale pour l'Histoire du Verre, Venise-Milan 1998), Lochem, 51-57. Foy, D., Nenna, M.-D. (2001) Tout feu, tout sable. Mille ans de verre antique dans le Midi de la France, Aix-en- Provence. Foy, D., Nenna, M.-D. (2003) Productions et importations de verre antique dans la Vallée du Rhône et le Midi méditerranéen de la France (Ier-IIIe siècles), in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 227-296. Foy, D., Picon, M., Vichy, M., Thirion-Merle, V. (2003) Caractérisation des verres de la fin de l‘Antiquité en Méditerranée occidentale: l‘émergence de nouveaux courants commerciaux, in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 41-86. Foy, D., Jezegou, M.-P., Fontaine, S.D. (2005) La circulation du verre en Méditerranée au début du IIIe siècle: le témoignage de l‘épave Ouest Embiez 1 dans le sud de la France (fouilles 2001-2003), (16e Congrès de l‘Association Internationale pour l‘Histoire du Verre, London 2003), Nottingham, 122-126. Foy, D., Nenna, M.-D. (2006a) Corpus des signatures et marques sur verres antiques, vol. 1 : La France, Aix-en-Provence – Lyon. Foy, D., Nenna, M.D. (2006b) Corpus des signatures et marques sur verres antiques, vol. 2 : Pays-Bas, Belgique, Luxembourg, Allemagne, Autriche, Suisse, Slovénie, Hongrie, Croatie, Espagne, Portugal, Maghreb, Grèce, Chypre, Turquie, mer Noire, Proche-Orient, Égypte, Soudan, Cyrénaïque, France (Addenda), Aix-en-Provence – Lyon. 495 Freestone, I.C. (1994) Appendix: chemical analysis of "raw" glass fragments , in: Hurst, H.R. (ed.) Excavations at Carthage:the British mission. Vol II, 1 The Circular Harbour, North Side. The Site ands Finds other than Pottery, Oxford, 290. Freestone, I.C. (2005) The Provenance of Ancient Glass through Compositional Analysis, Materials Research Society Symposium Proceedings 852, 008.1., 1-14. Freestone, I.C. (2006) Glass production in Late Antiquity and the Early Islamic period: a geochemical perspective, in: Maggetti, M., Messiga, B. (eds) Geomaterials in Cultural Heritage, London, 201-216; 257. Freestone, I. C., Gorin-Rosen Y. (1999) The Great Glass Slab at Bet She'arim, Israel: An early Islamic Glassmaking Experiment, Journal of Glass Studies 41, 105-116. Freestone, I.C., Gorin-Rosen, Y., Hughes, M.J. (2000) Composition of Primary Glass from Israel and the Production of Glass in Late Antiquity and the Early Islamic Period, in: Nenna, M.-D. (ed.) (2000) La route du verre. Ateliers primaires et secondaires du second millénaire av. J.-C. au Moyen Âge (Travaux de la maison de l‘orient méditerranéen 33), Lyon, 65-84. Freestone, I.C., Ponting, M., Hughes, M.J. (2002) The origins of Byzantine glass from Maroni Petrera, Cyprus, Archaeometry 44, 2, 257-272. Freestone, I.C., Leslie, K.A., Thirlwall, M., Gorin-Rosen, Y. (2003) Strontium Isotopes in the Investigation of early Glass Production: Byzantine and early Islamic Glass from the Near East, Archaeometry 45, 1, 19-32. Freestone, I.C., Price, J., Cartwright, C. (2009) The batch: Its recognition and significance, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 130-135. Fremersdorf, F. (1951) Figürlich geschliffene Gläser. Eine Kölner Werkstatt des 3. Jahrhunderts (Römisch-Germanische Forschungen 19), Berlin. Fremersdorf, F. (1958) Römisches Buntglas in Köln (Die Denkmäler des Römischen Köln 3), Cologne. Fremersdorf, F. (1959) Römische Gläser mit Fadenauflage in Köln (Die Denkmäler des Römischen Köln 5), Cologne. Fremersdorf, F. (1975) Antikes, islamisches und mittelalterliches Glas, sowie kleinere Arbeiten aus Stein, Gagat und verwandten Stoffen in den vatikanischen Sammlungen Roms (Museo Sacro, Museo Profano, Museo Egizio, Antiquarium Romanum) (Catalogo del Museo Sacro Vaticano 5), Vatican City. Friedhoff, U. (1991) Der römische Friedhof an der Jakobstrasse zu Köln (Kölner Forschungen 3), Mainz. Fülep, F. (1977) Roman Cemeteries on the Territory of Pécs (Sopianae) (Fontes Archaeologici Hungariae), Budapest. Fülep F. (1984) Sopianae, The History of Pecs During the Roman Era, and the Problem of the Continuity of the Late Roman Population, Budapest. Fünfschilling, S. (2000) Glas in Form eines Fässchens aus Kaiseraugst, Castrum Rauracense, Jahresberichte aus Augst und Kaiseraugst 21, 245-248. Fünfschilling, S. (2005) Verres romains à Augusta Raurica/Castrum Rauracense: quelques conclusions tirées du matériel, Bulletin de l‘Association Française pour l‘Archéologie du Verre 2005, 2-5. Fünfschilling, S., Hedinger, B. (2008) Glas, in: Ad Fines. Das spätrömische Kastell Pfyn. Befunde und Funde (Archäologie im Thurgau 8.1), Frauenfeld, 154-165. Gaitzsch, W. (1999) Reicher Bernsteinschmuck aus einem Römischen Brandgrab, Archäologie im Rheinland 1998, 77-79. Gaitzsch, W., Wedepohl, K.H., Hartmann, G., Tegtmeier, U. Follmann-Schulz, A.-B. (2003) Spätrömische Glashütten im Hambacher Forst - Produktionsort der ECVA-Fasskrüge. Archäologische und naturwissenschaftliche Untersuchungen, Bonner Jahrbücher 200, 2000, 83 – 241. Gandert, O.-F. (1956) Die Alsengemmen, Bericht der Römisch-Germanischen Kommission 36, 156-222. Garbsch, J., Kos, P. (1988) Das Spätrömische Kastell Vemania bei Isny I. Zwei Schatzfunde des Frühen 4. Jahrhunderts (Münchner Beiträge zur Vor- und Frühgeschichte 44), Munich. Gazenbeek, M., Van der Leeuw, S. (2003) L‘Argonne dans l‘Antiquité. Etude d‘une région productrice de céramique et de verre, Gallia 60, 269-317. Gebhard, R. (1989a) Pour une nouvelle typologie des bracelets celtiques en verre, in: Feugère, M. (ed.), Le verre pré-romain en Europe occidentale, Montagnac, 73-83. Gebhard, R. (1989b) Le verre de Manching: données chronologiques et rapport des analyses, in: Feugère, M. (ed.), Le verre pré-romain en Europe occidentale, Montagnac, 99-106. Gebhard, R. (1989c) Der Glasschmuck aus dem Oppidum von Manching (Die Ausgrabungen in Manching 11), Stuttgart. Gérardin, C., Sireix, C., Dubernet, S., Pernot, M., Raux, S. (2010) Les bracelets celtiques en verre du site de Lacoste (Moulietset-Villemartin, Gironde), Bulletin de l‘Association Française pour l‘Archéologie du Verre 2010, 5-9. Gesztelyi, T. (1991) Spätrömische Glasbullen, Acta Classica Universitatis Scientiarum Debreceniensis 27, 107-114. Gesztelyi, T. (1998) Spätrömische Glaskameen mit zwei Kanälen, Acta Antiqua Academiae Scientiarum Hungaricae 38, 129136. Gesztelyi, T. (2001) Spätrömische Gemmenfunde in Pannonien, in: Akten des 6. Internationalen Kolloquiums über Probleme des Provincialrömischen Kunstschaffens, Historisches Museum der Stadt Budapest, 11-16 Mai 1999, Aquincum (Budapest Régiségei 35), Budapest, 109-116. 496 Goethert-Polaschek, K. (1977) Katalog der Römischen Gläser des Rheinischen Landesmuseums Trier (Trierer Grabungen und Forschungen 1), Mainz. Goethert-Polaschek, K. (1984) Trier, Kaiserresidenz und Bischofssitz. Die Stadt in spätantiker und frühchristlicher Zeit, Mainz. Goethert, K. (2007) Keramikproduktion in Trier, in: Demandt, A., Engemann, J. (eds) Imperator Caesar Flavius Constantinus. Konstantin der Grosse. Ausstellungskatalog, 396-403. Goldstein, S.M., Rakow, L.S., Rakow, J.K. (1982) Cameo Glass. Masterpieces from 2000 Years of Glassmaking, Corning. Gorin-Rosen, Y. (2000) The Ancient Glass Industry in Israel: Summary of the Finds and New Discoveries, in: Nenna, M.-D. (ed.) La route du verre. Ateliers primaires et secondaires du second millénaire av. J.-C. au Moyen Âge (Travaux de la maison de l‘orient méditerranéen 33), Lyon, 49-63. Gorin-Rosen, Y., Katsnelson, N. (2008) Glass Production in Light of the Finds from Khirbat-el-Ni‘ana, Atiqot 57, 73-154. Graff Y. 1962, Découverte d'un fortin romain aux Bons-Villers (Liberchies), Romana Contact 3, 7-36. Gratuze, B. (2000) Etude de perles en verre provenant des niveaux de l'Age du Bronze et de l‘Age du Fer du site de Roche Noire (Montagnieu, 01), internal report CNRS Institut de Recherche sur les Archéomatériaux, Orléans. Gratuze, B. (2001a) Etude de la composition de bracelets protohistoriques en verre du Museo Civico ″Federico de Rocco" (S. Vito al Tagliamento, Italie), internal report CNRS Institut de Recherche sur les Archéomatériaux, Orléans. Gratuze, B. (2001b) Etude de la composition d‘un écarteur de collier en verre de l‘Aveyron (Collection B. Tournier, Musée de Gap), internal report CNRS Institut de Recherche sur les Archéomatériaux, Orléans. Gratuze, B., (2001c) Etude des perles protohistoriques en verre des tumulus de Mons (Saint-Georges, 15), internal report CNRS Institut de Recherche sur les Archéomatériaux, Orléans. Gratuze, B. (2002) Etude d'éléments de parure provenant du site protohistorique de Mez-Notariou (Ouessant, 29), internal report CNRS Institut de Recherche sur les Archéomatériaux, Orléans. Gratuze, B. (2004) Etude des perles protohistoriques en verre du tumulus de Courtesoult (Haute-Saône), internal report CNRS Institut de Recherche sur les Archéomatériaux, Orléans. Gratuze, B. (2009) Les premiers verres au natron retrouvés en Europe occidentale: composition chimique et chrono-typologie, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 8-14. Gratuze, B., Blet-Lemarquand, M., Barrandon, J.-N. (2001) Mass spectrometry with laser sampling: A new tool to characterize archaeological material, Journal of Radioanalytical and Nuclear Chemistry 247, 3, 645-656. Gratuze, B., Billaud, Y. (2003a) La circulation des perles en verre dans le Bassin méditerranéen de l'Age du Bronze moyen jusqu'au Hallstatt, in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 11-15. Gratuze, B., Foy, D., Lancelot, J., Tereygeol, F. (2003b) Les "lissoirs" carolingiens en verre au plomb: mise en évidence de la valorisation des scories issues du traitement des galènes argentifères de Melle (Deux-Sèvres), in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 101-107. Gratuze, B., Cosyns, P. (2007) La composition chimique des perles en verre de la tombe à char de la nécropole de Neufchâteau-Le Sart par LA-ICP-MS, Arduinna 63, 1-7. Green, L.R., Hart, F.A. (1987) Colour and chemical composition in ancient glass: an examination of some Roman and Wealden glass by means of ultraviolet-visible-infra-red spectrometry and electron microprobe analysis, Journal of Archaeological Science 14, 271-282. Greene, K. (2000) Technological innovation and economic progress in the ancient world: M.I. Finley re-considered, Economic History Review 53:1, 29-59. Greene, K. (2008) Learning to consume: consumption and consumerism in the Roman Empire, Journal of Roman Archaeology 21, 64-82. Grose, D.F. (1989) The Toledo Museum of Art. Early Ancient Glass: Core-formed, Rod-formed, and Cast Vessels and Objects from the Late Bronze Age to the Early Roman Empire, 1600 B.C. to A.D. 50, New York. Grose, D.F. (1991) Early Imperial Roman Cast Glass: The Translucent Coloured and Colourless Fine Wares, in: Newby, M., Painter, K., (eds) Roman Glass: Two Centuries of Art and Invention (Occasional papers 13), London, 1-18. Gunther, R.T., Manley, J.J. (1912) A Mural Glass Mosaic from the Imperial Roman Villa near Naples. With a note on the analysis of the green and blue glass, Archaeologia 63, 99-105. Guido, M. (1978) The Glass Beads of the Prehistoric and Roman Periods in Britain and Ireland (Reports of the Research Committee of the Society of Antiquaries of London 35), London. Guido, M. (1999) The Glass Beads of Anglo-Saxon England c. AD 400-700. A Preliminary Visual Classification of the more Definitive and Diagnostic Types, London. Guiraud, H. (1988) Intailles et camées de l‘époque romaine en Gaule. Territoire français vol. 1 (Gallia Supplément 48), Paris. Guiraud, H. (1989) Bagues et anneaux à l‘époque romaine en Gaule, Gallia 46, 173-211. Guiraud, H. (1996) Intailles de Lons-le-Saunier, Gallia 52, 359-406. 497 Guiraud, H. (1998) Intailles du Musée des Antiquités Nationales, Antiquités Nationales 30, 131-142. Guiraud, H. (2008) Intailles et camées de l‘époque romaine en Gaule. Territoire français vol. 2 (Gallia Supplément 48), Paris. Gurney, D. (1986) Settlement, religion and industry on the Fen-edge. Three Romano-British sites in Norfolk (East Anglian Archaeology Report 31), Dereham, 61-67. Gustin, M. (1981) La verrerie, in: Brulet, R. (ed.) Braives Gallo-Romain I. La zone centrale (Publications d‘Histoire de l‘Art et d‘Archéologie de l‘Université Catholique de Louvain 26), Louvain-la-Neuve, 113-121. Gustin, M. (1983) La verrerie, in: Brulet, R. (ed.) Braives Gallo-Romain II. Le quartier des potiers (Publications d‘Histoire de l‘Art et d‘Archéologie de l‘Université Catholique de Louvain 36), Louvain-la-Neuve, 87-92. Gustin, M. (1985) La verrerie, in: Brulet, R. (ed.) Braives Gallo-Romain III. La zone périphérique occidentale (Publications d‘Histoire de l‘Art et d‘Archéologie de l‘Université Catholique de Louvain 46), Louvain-la-Neuve, 50-54. Gustin, M. (1987) La verrerie, in: Brulet, R. (ed.) Liberchies I. Vicus Gallo-Romain. Bâtiment méridional et la Fontaine des Turcs (Fouilles de Pierre Claes (1956-1964) (Publications d‘Histoire de l‘Art et d‘Archéologie de l‘Université Catholique de Louvain 54), Louvain-la-Neuve, 83-85. Gustin, M. (1990) La verrerie, in: Brulet, R. (ed.) Braives Gallo-Romain IV. La zone centre-ouest (Publications d‘Histoire de l‘Art et d‘Archéologie de l‘Université Catholique de Louvain 77), Louvain-la-Neuve, 80-88. Haberey, W. (1942) Spätantike Gläser aus Gräbern van Mayen, Bonner Jahrbücher 142, 249-284. Haevernick, T.E. (1952) Antike Glasarmringe und ihre Herstellung, Glastechnische Berichte 25, 212-215. Haevernick T.E. (1960) Die Glasarmringe und Ringperlen der Mittel- und Spätlatènezeit auf dem Europäischen Festland. Bonn. Haevernick, T.E. (1963) Beiträge zur Geschichte des Antiken Glases. XI. Obsidianarbeiten (Jahrbuch des RömischGermanischen Zentralmuseums Mainz 10), Mainz, 122-130. Haevernick, T.E. (1968) Perlen und Glasbruchstücke als Amulette (Jahrbuch des Römisch-Germanischen Zentralmuseums Mainz 15), Mainz, 120-133. Haevernick, T.E. (1983) Trilobitenperlen (Folia archaeologica 25, 1975, 105-129), in: Haevernick, T.E. (ed.) Beiträge zur Glasforschung. Die wichtigsten Aufsätze von 1938 bis 1981, Mainz, 265-276. Haevernick, Th.E., Haberey, W. (1963) Glattsteine aus Glas, Jahrbuch des römisch-germanischen Zentralmuseums Mainz 10, 130-138. Hagen, W. (1937) Kaiserzeitliche gagatarbeiten aus dem rheinischen Germanien, Bonner Jahrbücher 142, 77-144. Hahnloser, H.R. (ed.) (1971) Il Tesoro di San Marco, Florence. Halbertsma, R. (2006) Schitterend glas. Uit Egypte, Griekenland en Rome, Zwolle. Hall, M., Yablonsky, L. (1997) Chemical analyses of glass beads found in two Sarmatian burials, Archaeometry 39:2, 369-377. Hampel, J. (1905) Alterthümer des frühen Mittelalters in Ungarn, Braunschweig. Hanut, F. (1999) Les sépultures gallo-romaines de la chaussée de Haecht (Schaerbeek), Vie archéologique 52, 5-23. Hanut, F. (2003) Périodisation de la verrerie au Ier siècle ap. J.-C. dans le nord de la Gaule, Bulletin de l‘Association Française pour l‘Archéologie du Verre 2002-2003, 27-34. Hanut, F., Capers, P. (2003) Un puits cultuel du vicus de Waudrez (Hainaut, Belgique): étude du matériel archéologique et interprétation chronologique, Vie Archéologique 60, 5-103. Harden, D.B. (1936) Glass from Karanis, Found by the University of Michigan Archaeological Expedition in Egypt 1924-29 (University of Michigan Studies. Humanistic Series 41), Ann Arbor. Harden, D.B. (1967) Gaming pieces, in: Stead, I.M., A La Tène III burial at Welwyn Garden City, Archaeologia 101, 14-16. Harden, D.B. (1971) Glass, in: Brodribb, A.C.C., Hands, A.R., Walker, D.R., Excavations at Shakenoak Farm, near Wilcote, Oxfordshire. Part 2, s.l., 98-108. Harden, D.B. (1979) Glass vessels, in: Clarke, G., The Roman Cemetery at Lankhills (Pre-Roman and Roman Winchester, Winchester Studies 3), Oxford, 209-220. Harden, D.B., Toynbee, J.M.C. (1959) The Rothschild Lycurgus Cup, Archaeologia 97, 179-212. Harden, D.B., Hellenkemper, H., Painter, K., Whitehouse, D. (1987) Glass of the Caesars, Milan. Harding, G.L. (1950) A Roman Family Vault on Jebel Yofeh, Amman, The Quarterly of the Department of Antiquities in Palestine 14, 81-94. Hartmann, G., Kappel, I., Grote, K., Arndt, B. (1997) Chemistry and technology of prehistoric glass from Lower Saxony and Hesse, Journal of Archaeological Science 24, 547-559. Heck, M., Hoffmann, P. (2000) Coloured opaque glass beads of the Merovingians, Archaeometry, 42, 2, 341-357. Henderson, J. (1985) The raw materials of early glass production, Oxford Journal of Archaeology, 4, 3, 267-291. Henderson, J. (1987a) Glass, in: Cunliffe, B. (ed.) Hengistbury Head, Dorset. Volume 1: The Prehistoric and Roman Settlement, 3500 BC-AD 500 (Monograph 13), Oxford, 160-163. Henderson, J. (1987b) Glass working, in: Cunliffe, B. (ed.) Hengistbury Head, Dorset. Volume 1: The Prehistoric and Roman Settlement, 3500 BC-AD 500 (Monograph 13), Oxford, 180-186. 498 Henderson, J. (1989) Scientific Analysis in Archaeology and Ist Interpretation (Oxford University Committee for Archaeology. Monograph 19.), Oxford. Henig, M. (1978²) A Corpus of Roman Engraved Gemstones from British Sites (British Archaeological Reports British Series 8), Oxford. Henig, M. (1975) Vindolanda jewellery (Northern history booklet 62), Newcastle upon Tyne. Henig, M. (1984) Religion in Roman Britain, London, Batsford. Henig, M. (1997) Intaglios, in: Wilmott, T., Hird, L., Izard, K., Summerfield, J., Birdoswald: excavations of a Roman fort on Hadrian‘s Wall and its successor settlements: 1987 – 92 (English Heritage Archaeological Report 14), London, 283-285. Henkel, F. (1913) Die Römischen Fingerringe der Rheinlande und der Benachbarten Gebiete, Berlin. Hentrich, H., von Saldern, A. (1974) Glassammlung Hentrich: Antike und Islam (Kataloge des Kunstmuseums Düsseldorf I: Glas; 3), Düsseldorf. Hill, J.N., Evans, R.K. (1972) A model for classification and typology, in: Clarke, D. (ed.), Models in Archaeology, London, 231273. Hochuli-Gysel, A. (1993) Römisches Glas aus dem Südwesten von Frankreich, in: Annales du 12e Congrès de l‘Association Internationale pour l‘Histoire du Verre (Vienna 1991), Lochem, 79-88. Hodder, I. (2002) Archaeological theory, in Cunliffe, B., Davies, W., Renfrew, C. (eds) Archaeology. The Widening Debate, Oxford-New York, 77-90. Honroth, M., et al. (2007) Vom Luxusobjekt zum Gebrauchsgefäss. Vorrömische und römische Gläser, Stuttgart. Hruschka, W.R. (1987) Data analysis: wavelength selection methods, in: Williams, P., Norris, K. (eds) Near-Infrared Technology in the Agricultural and Food Industries, Saint-Paul, 35-55. Humphrey, J.W., Oleson, J.P., Sherwood, A.N. (1998) Greek and Roman Technology: A Sourcebook, London – New York. Hunter, R.S., Harold, R.W. (1987) The Measurement of Appearance, New York. Husseini, S.A.S. (1938) A rock-cut tomb-chamber at Ain Yabrūd, The Quarterly of the Department of Antiquities in Palestine 6, 54-55. Ignatiadou, D., Dotsika, E., Kouras, A., Maniatis, Y. (2005) Nitrum Chalestricum: the natron of Macedonia, in: Annales du 16e Congrès de l'Association Internationale pour l'Histoire du Verre (London 2003), Nottingham, 64-67. Iliffe, J.H. (1933) Vaulted tomb at Askalon, The Quarterly of the Department of Antiquities in Palestine 2, 182-183. Iliffe, J.H. (1934) Rock-cut tomb at Tarshīhā, The Quarterly of the Department of Antiquities in Palestine 3, 9-16. Ingle, J.D.J., Crouch, S.R. (1988) Spectrochemical Analysis, New Jersey. Isings, C. (1957) Roman Glass from Dated Finds (Archaeologica Traiectina 2), Groningen-Djakarta. Isings, C. (1964) Glass from Ulpia Noviomagus, Bulletin van de Vereeniging tot bevordering der kennis van de antieke beschaving te ‘s-Gravenhage 39, 174-179. Isings, C. (1971) Roman Glass in Limburg (Archaeologica Traiectina 9), Groningen. Isings, C. (1980) Glass from the Canabae Legionis at Nijmegen, Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 30, 281-346. Isings, C. (2009) Glass, in: van Es, W.A., Verwers, W.J.H. (eds) Excavations at Dorestad. 3, Hoogstraat 0 II-IV (Nederlandse Oudheden 16), Amersfoort, 259-280. Isings, C., Van Lith, S.M.E. (1992) Romeins glas (Museumstukken 3), Nijmegen. Israeli, Y. (2003) Ancient Glass in the Israel Museum. The Eliahu Dobkin Collection and Other Gifts, Jerusalem. Israeli, Y. (2005) What did Jerusalem‘s First Century BCE Glass Workshop produce?, in: Annales du 16e Congrès de l'Association Internationale pour l'Histoire du Verre (London 2003), Nottingham, 54-57. Jacobs, J., von Schwerzenbach, C. (1910) Die römische Begräbnisstätte von Brigantium, Jahrbuch für Altertumskunde 4, 5-66. Jackson C.M., Smedley, J.W. (2008) Theophilus and the use of beech ash as a glassmaking alkali, in: Martinón-Torres, M., Rehren, T. (eds) Archaeology, History and Science. Integrating approaches to ancient materials, Walnut Creek (California), 117130. Jackson, C.M., Nicholson, P.T. (2010) The Provenance Of Some Glass Ingots From The Uluburun Shipwreck. Journal of Archaeological Science 37, 295-301. Jackson, C.M., Price, J. (forthcoming), Analysis of a late Roman assemblage from South Shields in: Annales of the 18th Congress of the Association Internationale pour l‘Histoire du Verre (Thessaloniki 2008). Jacques, A. (1997) 128. Bracelet, in Trésors archéologiques du Nord de la France. Gallo-Romains et Mérovingiens, Valenciennes, 99. Janot, F., Vezie, P. (1999) Les charmes de la galène, Bulletin de l‘Institut Français d‘Archéologie Orientale 99, 217-232. Janssens, D. (1977) Een Gallo-Romeins grafveld te Maaseik (Archaeologia Belgica 198), Brussels. 499 Jelinčić , K. (2007) Kasnoantičke narukvice od staklene paste s lokaliteta Virovitica-Ki korija jug/Glass Paste Bracelets from late Antiquity from the Site of Virovitica-Ki korija South, Prilozi Instituta za arheologiju u Zagrebu 24:1, 213-220. Jelinčić , K. (2009) Rimske staklene narukvice kao prilog poznavanju antičkoga Ore ca/Roman glass bracelets as a means to understanding Ore ac in Antiquity, Opuscula Archaeologica 33, 105-124. Jones, C. (1996) The Jewellery of Roman Britain. Celtic and Classical Traditions, London. Jones, A. (2002) Archaeological Theory and Scientific Practice, Cambridge. Jones, A., MacGregor, G. (2002) Colouring the past: the significance of colour in archaeological research, Oxford. Kaiser, M.-S. (2006) Das keltisch-r mische Gräberfeld von Wederath-Belginum. Die Aschengruben und Aschenflächen Ausgegraben 1954-1985 (Trierer Grabungen und Forschungen 6), Mainz. Karageorghis, V. (2000) Ancient Art from Cyprus. The Cesnola Collection in the Metropolitan Museum of Art, New York. Karasová, Z., Schönfelder, M. (2005) Die Funde aus dem Oppidum »Hradiste« bei Stradonice (Okr. Beroun, CZ) im RömischGermanischen Zentralmuseum, Mainz (mit einem Beitrag von Jirí Militky), Jahrbuch des Römisch-Germanischen Zentralmuseums 51, 219-242. Karwowski, M. (2004) Latènezeitlicher Glasringschmuck aus Ostösterreich (Mitteilungen der Prähistorischen Kommission 55), Vienna. Kauffmann, S. (2007) Obtenir son salut : quel objet choisir ? La réponse du christianisme, in: Delattre, C. (ed.) Objets sacrés, objets magiques : de l'Antiquité au Moyen Age (THEMAM, Textes, histoire et monuments de l'Antiquité au Moyen âge), Paris, 117-140. Keller, E. (1971) Die Spätrömischen Grabfunde in Südbayern (Münchner Beiträge zur Vor- und Frühgeschichte 8), Munich. Kilbride-Jones, H.E. (1938) Glass armlets in Britain, Proceedings of the Society of Antiquaries of Scotland 72, 366-395. Kilian-Dirlmeier, I. (2002) Kleinfunde aus dem Athena-Itonia-Heiligtum bei Philia (Thessalien) (Forschungsinstitut für Vor- und Frühgeschichte 48), Mainz. Klaproth, M.H. (1801) Sur quelques vitrifications antiques, Histoires et mémoires de l‘Académie Royale des Sciences, Inscriptions et Belles Lettres Berlin 49, 3-16. Klejn, L.S. (1982) Archaeological Typology (British Archaeological Series International Series 153), Oxford. Knauer, E.R. (1993) Roman Wall Paintings from Boscoreale: Three Studies in the Relantionship between Writing and Painting, Metropolitan Museum Journal 28, 13-46. Kock, J., Sode, T. (1995) Glass Beads and Glassmakers in Northern India, s.l. Komp, J. (2009) Römisches Fensterglas. Archäologische und archäometrische Untersuchungen zur Glasherstellung im Rheingebiet, Aachen. Konrad, M. (1997) Das Römische Gräberfeld von Bregenz-Brigantum. 1 : Die Körpergräber des 3. bis 5. Jahrhunderts (Münchner Beiträge zur Vor- und Frühgeschichte 51), Munich. Korfmann, M. (1966) Zur Herstellung nahtloser Glasringe, Bonner Jahrbücher 166, 48-61. Kraus, K. (2005) Ein Römisches Gräberfeld bei Moers-Schwafheim am Niederrhein, in: Von Anfang an. Archäologie in Nordrhein-Westfalen (Schriften zur Bodendenkmalpflege in Nordrhein-Westfalen 8), Mainz, 387-390. Krug, A., (1975) Römische Gemmen und Fingerringe im Museum Frankfurt, Germania 53, 113-125. Krug, A., (1977) Römische Fundgemmen. 2. Wiesbaden und Berlin, Germania 55, 77-84. Krug, A., (1978) Römische Fundgemmen. 3. Speyer, Worms, Bad Kreuznach, Mainz und Saalburg, Germania 56, 476-503. Krug, A., (1980) Römische Fundgemmen. 4. Neuwied, Frieburg, Florstadt, Darmstadt, Hanau, Aschaffenburg und Koblenz, Germania 58, 117-135. Krug, A. (1995) Römische Gemmen im Rheinischen Landesmuseum Trier (Schriftenreihe des Rheinischen Landesmuseums Trier 10), Trier. Küçükerman, O. (1988) Glass beads. Anatolian Glass Bead Making. The Final Traces of Three Millennia of Glass Making in the Mediterranean Region, Istanbul. Kunina, N. (1997) Ancient Glass of the Hermitage Collection, St. Petersburg. Kunkel, O. (1961) Zur Frage Keltischer Glasindustrie – Nach einer Manchinger Fundgruppe, Germania 39, 322-329. Kurzrock, R. (ed.) (1983) Farbe. Material-Zeichen-Symbol (Forschung und Information 33 Schriftenreihe der RIASFunkuniversität), Berlin. La Baume, P. (1973) Glas der Antiken Welt I, Cologne. Lahlil, S., Biron, I., Galoisy, L., Morin, G. (2009) Technological processes to produce antimonate opacified glass throughout history, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 571-578. Lahlil, S., Biron, I., Cotte, M., Susini, J., Menguy, N. (2010) Synthesis of calcium antimonate nano-crystals by the 18th dynasty Egyptian glassmakers, Applied Physics A - Materials Science & Processing 98, 1-8. Lamprecht, H. (1906) Der grosse römische Friedhof in Regensburg (Verhandelungen des historischen Vereins für Oberpflatz und Regensburg 58,1), Nördlingen. 500 Lauwers, V. (2008) The glass of Sagalassos. Towards a geochemical and typo-chronological interpretation, unpublished doctoral dissertation at Katholieke Universiteit Leuven, Leuven. Lauwers, V., Degryse, P., Poblome, J., Waelkens, M. (2005) Le Verre de Sagalassos: De nouvelles preuves d‘une production locale de verre, in: Cabart, H. (ed.) Eclats de verre, Toulouse, 26-29. Lauwers, V., Degryse, P., Waelkens, M. (2007) Evidence for Anatolian glassworking in Antiquity: the case of Sagalassos (Southwestern Turkey), Journal of Glass Studies 49, 39-46. Lavan, L. (2007) Explaining Technological Change: Innovation, Stagnation, Recession and Replacement, in: Lavan, L., Zanini, E, Sarantis, A. (eds) Technology in Transition A.D. 300-650 (Late Antique Archaeology 4), Leiden, 15-40. Lazar, I. (2003) The Roman glass of Slovenia (Opera Instituti Archaeologici Sloveniae 7), Ljubljana. Leech, R. (1986) The excavation of a Romano-Celtic Temple on Lamyatt Beacon, Britannia 17, 259-275. Lefrancq, J. (1988) Le contexte archéologique belgo-romain de la «Grotte des Nutons» à Presles (Documents et rapports de la Société Royale d‘Archéologie et de Paléontologie de Charleroi 60), Charleroi. Lefrancq, J. (1989) Les tombes de Noirmont. Tumuli belgo-romains (Inventaria Archaeologica. Belgique 4, B13-14), Brussels. Lehance, H., Willems, J. (1987) Le cimetière belgo-romain d‘Amay. Les tombes mises au jour en 1986, Vie archéologique 7, 4664. Lémant, J.P. (1985) Le cimetière et la fortification du Bas-Empire de Vireux-Molhain, dép. Ardennes (Römisch-germanisches Zentralmuseum zu Mainz. Forschungsinstitut für Vor- und Frühgeschichte. Monographien 7), Bonn. Lemière, J. (1980) Catalogue des monnaies in : Pilet, C., La Nécropole de Frénouville. Etude d‘une population de la fin du IIIe à la fin du VIIe siècle (British Archaeological Reports International Series 83), Oxford, vol.II:327–39. Liéger, A. (1997) La nécropole gallo-romaine de Cutry (Meurthe-et-Moselle) (Etudes Lorraines d‘Archéologie Nationale 3), Nancy. Lightfoot, C.S. (2003) From East to West: the early Roman glass industry, in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 341-347. Lightfoot, C.S. (2007) Ancient Glass in National Museums Scotland, Edinburgh. Liversidge, J. (1973) Britain in the Roman Empire, London. Loeschke, S. (1925) Frühchristliche Werkstätte für Glassschmuck in Trier, in: Trierer Heimatbuch, Trier, 337-360. Ložar, R. (1938) Rimska najdba iz Polhovega Gradca, Glasnik Muzejskega društva za Slovenijo 19, 87-108. Lund Hansen, U. (1987) Römischer Import im Norden: Warenaustausch zwischen dem Römischen Reich und dem freien Germanien während der Kaiserzeit unter besonderer Berücksichtigung Nordeuropas (Nordiske fortidsminder Serie B, 10), Copenhagen. Lundström, A. (1976) Bead Making in Scandinavia in the Early Middle Ages (Antikvariskt arkiv 61. Early Medieval Studies 9), Stockholm, 3-19. Maaskant-Kleibrink, M. (1975) Classification of ancient engraved gems, Leiden. Maaskant-Kleibrink, M. (1978) Catalogue of the engraved gems in the Royal Coin Cabinet, The Hague: The Greek, Etruscan and Roman collections, The Hague. Macalister, R.A.S. (1912) A history of civilization in Palestine, Cambridge. Mackreth, D. (1992) Roman brooches from Gastard, Corsham, Wiltshire, The Wiltshire Archaeological and Natural History Magazine 85, 51-62. Macquet, C. (1990) Les lissoirs de verre, approche technique et bibliographique (note à l‘occasion de l‘étude des exemplaires découverts à Saint-Denis), Archéologie Médiévale 20, 319-334. Magyar, Z. (2009) Trilobitenperlen from Dunaszekcső (Hungary) Janus Pannonius Múzeum Évkönyve 50-52 (2005-2007), 100104. Mahieu, A. (1919) La villa belgo-romaine de Bauselenne à Mettet, Annales de la Société Archéologique de Namur 33, 49-208. Mainman, A.J., Rogers, N.S.H. (2000) Craft, Industry and Everyday Life: Finds from Anglo-Scandinavian York, in: Addyman, P.V. (ed.) The Small Finds (The Archaeology of York 17), Bradford. Maiuri, B.T. (1971) Museo Nazionale Napoli, Novara. Mannoni, T. (2007) The Transmission of Craft Techniques according to the principles of Material Culture: Continuity and Rupture, in: Lavan, L., Zanini, E, Sarantis, A. (eds) Technology in Transition A.D. 300-650 (Late Antique Archaeology 4), Leiden, 41-60. Mariën, M.E. (1970) Le Trou de l‘Ambre au Bois de Wérimont à Eprave (Monographie d‘Archéologie Nationale 4), Brussels. Mariën, M.E. (1983) Tombe belgo-romaine à Arsimont (Archaeologia Belgica 255), Brussels, 95-126. Marii, F., Rehren, T. (2009) Archaeological coloured glass cakes and tesserae from the Petra Church, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 295-300. Martens, H., Naes, T. (1989) Multivariate calibration, Chichester. 501 Martin-Kilcher, S. (1976) Das Römische Gräberfeld von Courroux im Berner Jura (Basler Beiträge zur Ur- und Frühgeschichte 2), Derendingen. Martin, M. (1976) Das spätrömisch-frühmittelalterliche Gräberfeld von Kaiseraugst, Kt. Aargau (Basler Beiträge zur Ur- und Frühgeschichte 5B), Derendingen-Solothur. Martin-Pruvot, C. (1999) Le verre, in: Castella, D. et al. (eds) La nécropole gallo-romaine d‘Avenches ―En Chaplix‖. Fouilles 19871992, Volume 2: Etude du mobilier (Cahiers d‘Archéologie Romande 78. Aventicum 10), Lausanne, 167-295. Matheson, S.B. (1980) Ancient Glass in the Yale University Art Gallery, Yale. Mauget, L. (1903) Découverte d‘une verrerie d‘art gallo-romaine aux Houis, écart de Sainte-Ménehould, Mémoires de la Société d‘Agriculture, Commerce, Sciences et Arts du département de la Marne, année 1902-1903 1, 3-15. Mauget, L. (1904) La verrerie des Houis, Revue Archéologique 1904, 305-306. Mauget, L. (1907) Découverte d‘une verrerie d‘art gallo-romaine, aux Houis écart de Sainte-Menehould, Association Française pour l'Avancement des Sciences (AFAS) 36e session, Reims, 897-908. McBrearty, S., Brooks, A.S. (2000) The revolution that wasn‘t: a new interpretation of the origin of modern behaviour, Journal of Human Evolution 39, 453-563. Meconcelli Notarianni, G. (1979) Vetri antichi nelle collezioni del Museo Civico Archeologico di Bologna, cataloghi nuova serie 1, Bologna. Mertens, J., Van Impe, L. (1971) Het Laat-Romeins grafveld van Oudenburg, Archaeologia Belgica 135 (2 vols), Brussels. Mertens, J., Rémy, H. (1973) Un refuge du Bas-Empire à Eprave (Archaeologia Belgica 144), Brussels. Mertens, J., Brulet, R. (1974) Le castellum du bas-empire romain de Brunehaut-Liberchies, (Archaeologica Belgica 163), Brussels. Mertens, J., Vanvinckenroye, W. (1975) Een Romeins gebouwencomplex extra muros te Tongeren (Archaeologia Belgica 180), Brussels. Meulebroeck, M. (2004) Optische detectietechnieken voor lasersorteermachines: productidentificatie en kwaliteitscontrole, unpublished doctoral dissertation at Vrije Universiteit Brussel, Brussels. Meulebroeck, M., Cosyns, P., Wouters, H., Baert, K., Nys, K., Terryn, H., Thienpont, H. (2008) Optical characterization and analysis of archaeological glass artefacts, in: Proceedings of the 14th Microoptics Conference 2008, Brussels, 197-198. Meulebroeck, W., Baert, K., Wouters, H., Cosyns, P., Ceglia, A., Cagno, S., Janssens, K., Nys, K., Terryn, H., Thienpont, H. (2010a) The identification of chromophores in ancient glass by the use of UV-Vis-NIR spectroscopy, in: Berghmans, F., Mignani, A.G., Van Hoof, C.A. (eds) Conference proceedings Photonics Europe - Optical Sensing and Detection, Brussels, 77260D. Meulebroeck, W., Wouters, H., Baert, K., Ceglia, A., Terryn, H., Nys, K., Thienpont, H. (2010b) Optical spectroscopy applied to the analysis of medieval and post-medieval plain flat glass fragments excavated in Belgium, in: Berghmans, F., Mignani, A.G., Van Hoof, C.A. (eds) Conference proceedings Photonics Europe - Optical Sensing and Detection, Brussels, 77261E. Meulebroeck, W., Cosyns, P., Baert, K., Wouters, H., Cagno, S., Janssens, K., Terryn, H., Nys, K., Thienpont, H. (2011) Optical spectroscopy as a rapid and low-cost tool for the first-line analysis of glass artefacts: a ste-by-stepplan for Roman green glass, Journal of Archaeological Science 38, 9, 2387-2398. Meyer, C. (1992) Glass from Quseir al-Qadim and the Indian Ocean trade, Chicago. Mignot, P. (1984) Les cimetières gallo-romains du Haut-Empire en Famenne, Annales de la Société Archéologique de Namur 63, 2, 149-248. Miller, D. (1991). Material culture and mass consumption. Oxford. Millet, M. (1990) The Romanization of Britain. An essay in archaeological interpretation, Cambridge. Mirti, P., Casoli, A., Appolonia, L. (1993) Scientific analysis of Roman glass from Augusta-Praetoria, Archaeometry 35, 225-240. Mirti, P., Appolonia, L., Casoli, A., Ferrari, R.P., Laurenti, E., Canesi, A.A., Chiari, G. (1995) Spectrochemical and structural studies on a Roman sample of Egyptian blue, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 51, 437446. Mirti, P., Lepora, A., Sagui, L. (2000) Scientific analysis of seventh-century glass fragments from the Crypta Balbi in Rome, Archaeometry, 42, 2, 359-374. Mirti, P., Pace, M., Negro Ponzi, M.M., Aceto, M. (2008) ICP-MS analysis of glass fragments of Parthian and Sasanian epoch from Seleucia and Veh Ardasir (Central Iraq), Archaeometry 50, 3, 429-450. Monk, L. (2001) Glass, in: Anderson, A.S., Wacher, J.S., Fitzpatrick, A.P. (eds), The Romano-British ‗Small Town‘ at Wanborough, Wiltshire (Britannia. Monograph Series 19), London, 162-172. Moretti, C., Gratuze, G. (2002) Vetri romani di Aquileia et di altri siti europei: analisi chimiche e studio comparativo, Rivista della Stazione Sperimentale del Vetro 2, 19-28. Morin-Jean (1913) La Verrerie en Gaule sous l‘Empire Romain, Paris. Motte, S., Martin, S. (2003) L‘atelier de verrier antique de la Montée de la Butte à Lyon et ses productions, in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 303-319. Moulin, J. (1993) La verrerie, in: Brulet, R., Demanet, J.-C. (eds) Liberchies II. Vicus Gallo-Romain. Sondages de Pierre Claes et Edmond Milliau (1957, 1961 et 1963). Zone d‘Habitat au sud de la voie antique. Fouilles de Pro Geminiaco (1969-1986) 502 (Publications d‘Histoire de l‘Art et d‘Archéologie de l‘Université Catholique de Louvain LXXXII), Louvain-la-Neuve, 30 ; 167-171 fig. 7; 83. Mulevanov, S., Minko, N., Kemenov, S., Osipov, A., Bykov, N. (2009) Vibrational spectroscopy investigation of the structure of multicomponent phosphorus-containing silicate glasses, Glass and Ceramics 66, 117-119. Munier, C. (2009) Nevers – 12 rue Saint-Genest: etude du verre médiéval (VIIe, IXe-XIIe siècles), Bulletin de l‘Association Française pour l‘Archéologie du Verre 2009, 59-70. Naas, V (2006) Omnia ergo meliora fuere, cum minor copia (Pline l‘Ancien, NH, XXXV, 50): matières et couleurs au service d‘un discours moral dans la minéralogie de Pline l‘Ancien, in: Rouveret, A., Dubel, S., Naas, V., Couleurs et matières dans l‘antiquité. Textes, techniques et pratiques (Études de littérature ancienne 17), Paris, 201-211. Nakai, I., Yamada, S., Terada, Y., Shindo, Y., Utaka, T. (2005) Development of a portable X-ray fluorescence spectrometer equipped with two monochromatic X-ray sources and silicon drift detector and field analysis of Islamic glasses at an excavation site in Egypt, X-Ray Spectrometry 34, 46–51. Nakai, I., Tantrakarn, K., Kato, N., Kawai, N., Nishisaka, A., Yoshimura, S. (2009) XRF analysis of 16th century BC transparent glass beads excavated from a hillside in northwest Saqqara, Egypt, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), AntwerpBrussels, 27-31. Nenna, M.-D. (2000) Ateliers de production et sites de consommation en Egypte. Ve siècle av. J.-C. – VIIe siècle ap. J.-C.. Premier bilan, in: Annales du 14e Congrès de l‘Association Internationale pour l‘Histoire du Verre (Venice-Milan 1998), Lochem, 20-24. Nenna, M.-D. (ed.) (2000) La route du verre. Ateliers primaires et secondaires du second millénaire av. J.-C. au Moyen Âge (Travaux de la maison de l‘orient méditerranéen 33), Lyon. Nenna, M.D. (2003) Productions et importations de verre antique dans la vallée du Rhône et le Midi méditerranéen de la France (Ier-IIIe siècles), in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 227-296. Nenna, M.-D. (2005) Ateliers de production et sites de consommation en Egypte. Ve siècle av. J.-C. – VIIe siècle ap. J.-C. Premier bilan, in: Annales du 16e Congrès de l'Association Internationale pour l'Histoire du Verre (London 2003), Nottingham, 20-24. Nenna, M.-D. (2008a) Production et commerce du verre à l‘époque impériale: nouvelles découvertes et problématiques, Facta 1, 2007, 125-147. Nenna, M.-D. (2008b) Nouveaux acquis sur la production et le commerce du verre antique entre Orient et Occident, Zeitschrift für schweizerische Archäologie und Kunstgeschichte [Handwerk und Gesellschaft in den römischen Provinzen. Internationaler Kongress CRAFTS 2007, (Zürich 01/03/2007)] 65, 61-66. Nenna, M.-D., Vichy, M., Picon, M. (1997) L'atelier de verrier de Lyon, du Ier s. ap. J.-C., et l'origine des verres "romains", Revue d'Archéométrie 21, 81-87. Nenquin, J.A.E. (1953) La nécropole de Furfooz (Dissertationes Archaeologicae Gandeses I), Bruges. Neumann, B. (1925) Antike Gläser, ihre Zusammensetzung und Färbung, Zeitschrift für Angewandte Chemie 38, 776-780. Neumann, B. (1927) Antike Gläser, ihre Zusammensetzung und Färbung, Zeitschrift für Angewandte Chemie 40, 963-967. Neumann, B. (1928) Antike Gläser, ihre Zusammensetzung und Färbung, Zeitschrift für Angewandte Chemie 41, 103-204. Newby, M. (2008) Byzantine Mould-Blown Glass from the Holy Land with Jewish and Christian Symbols. The Shlomo Moussaieff Collection, s.l.. Newton, R., Davison, S. (1989) Conservation of Glass, London. Nicholson, P.T. (1995) Glassmaking and glassworking at Amarna: some new work, Journal of Glass Studies 37, 11-19. Nicholson, P.T. (2000) The Glass, in: Sidebotham, S.E., Wendrich, W.Z. (eds.) Berenike 1998. Report of the 1998 Excavations at Berenike and the Survey of the Egyptian Eastern Desert, including Excavations in Wadi Kalalat, Leiden, 203-209. Nicholson, P.T., Jackson, C.M. (2000) Tell el-Amarna and the Glassmakers‘ Workshop of the Second Millennium B.C., in: Nenna, M-D. (ed.) La Route du Verre: Ateliers primaires et secondaires du second millénaire av. J.-C. au Moyen Âge (Travaux de la Maison de l‘Orient Mediterranéen 33), Lyon, 11-21. Nicholson, P.T., Jackson, C.M., Trott, K.M. (1997) The Ulu Burun glass ingots, cylindrical vessels and Egyptian glass, Journal of Egyptian Archaeology 83, 143-153. Nicod, P.-Y., Sordoillet, D., Chaix, L. (1998) De l‘épipaléothique à l‘époque moderne sur le site du Seuil-des-Chèvres (La Balme, Savoie), Revue archéologique de l‘Est 49, 31-85. Noll, R. (1963) Das römerzeitliche Gräberfeld von Salurn (Archäologische Forschungen in Tirol 2), Innsbruck. Oelmann, F. (1976²) Materialen zur Römisch-Germanischen Keramik. Heft 1: Die Keramik des Kastells Niederbieber, Bonn. Ogden, J. (1982) Jewellery of the ancient world, London. Olivier, F., Boizot, B., Ghaleb, D., Petite, G. (2005) Raman and EPR studies of β-irradiated oxide glasses, Journal of NonCrystalline Solids 351, 1061-1066. Overbeck, B. (2000) Die Münzen des Schatzes, in: Boos, A., Dallmeier, L.-M., Overbeck, B., Der römische Schatz von Regensburg-Kumpfmühl, Regensburg, 61-81. 503 Pailler, J.-M. (1992) Eikonis, Lakonis, Gorgonis, Eutychousin: Hypothèses de lecture. 1re Hypothèse, in: Schaad, D. (ed.) Le trésor d‘Eauze, Toulouse, 77-83. Palumbo, A.D. (1999) Ricerche sulla necropolis dell‘Università cattolica di Milano. Elementi vitrei fra gli oggettti d‘uso e di ornamento personale delle tombe a inumazione, in: Ferrari, D., Meconcelli, G. (eds.) Il vetro fra antico e moderno, Atti della III Giornata Nazionale di Studio (Milan 1997), 29-32. Peddemors, A. (1975) Latèneglasarmringe in den Niederländen, Analecta Praehistorica Leidensia 8, 93-145. Phang, S.E. (2001) The Marriage of Roman soldiers (13 BC - AD 235). Law and family in the imperial army, Leiden, Boston, Cologne. Philip, B. (1981) The Excavation of the Roman Forts of the Classis Britannica at Dover 1970-1977, Dover. Philippe, J. (2000) Les fibules de Seine-et-Marne du 1er siècle av. J.-C. au 5e siècle ap. J.-C. (Mémoires archéologiques de Seine-et-Marne 1), Châlons-en-Champagne. Picon, M., Vichy, M. (2003) D‘Orient en occident: l‘origine du verre à l‘époque romaine et durant le haut moyen âge, in: Foy, D., Nenna, M.-D. (eds) Echanges et commerce du verre dans le monde antique (actes du colloque international de l‘association française pour l‘archéologie du verre, Aix-en-Provence et Marseille, juin 2001) (Monographies Instrumentum 24), Montagnac, 17–31. Pilet, C. (1980) La Nécropole de Frénouville. Etude d‘une population de la fin du IIIe à la fin du VIIe siècle (British Archaeological Reports International Series 83), Oxford. Pilet, C. (1994) La Nécropole de Saint-Martin-de-Fontenay. Recherches sur le peuplement de la plaine de Caen du Ve siècle avant J. -C. au VIIe après J. -C. (Gallia Supplement 54), Paris. Pini, I. (1981) Spätbronzezeitliche ägäische Glassiegel, Jahrbuch des Römisch-Germanischen Zentralmuseums Mainz 28(1981) 63-69. Pirling, R. (1966) Das Römisch-Fränkische Gräberfeld von Krefeld-Gellep (Germanische Denkmäler der Völkerwanderungszeit serie B. Die Fränkischen Altertümer des Rheinlandes 2), Berlin. Pirling, R. (1974) Das Römisch-Fränkische Gräberfeld von Krefeld-Gellep 1963-66 (Germanische Denkmäler der Völkerwanderungszeit Serie B. Die Fränkischen Altertümer des Rheinlandes 8), Berlin. Pirling, R. (1979) Das römisch-fränkische Gräberfeld von Krefeld-Gellep 1964-1965 ( Germanische Denkmäler der Völkerwanderungszeit. Serie B. Die fränkischen Altertümer des Rheinlandes 10), Berlin. Pirling, R. (1989) Das römisch-fränkische Gräberfeld von Krefeld-Gellep 1966-1974 (Germanische Denkmäler der Völkerwanderungszeit. Serie B. Die fränkischen Altertümer des Rheinlandes 13), Wiesbaden. Pirling, R. (1993) Ein römischer Grabfund in Brunssum in Limburg, Archäologisches Korrespondenzblatt 23, 2, 213-221. Pirling, R. (1997) Das Römisch-Fränkische Gräberfeld von Krefeld-Gellep 1975-1982 (Germanische Denkmäler der Völkerwanderungszeit Serie B. Die Fränkischen Altertümer des Rheinlandes 17), Stuttgart. Pirling, R., Siepen, M. (2000) Das römisch-fränkische Gräberfeld von Krefeld-Gellep 1983-1988 (Germanische Denkmäler der Völkerwanderungszeit serie B, Die Fränkischen Altertümer des Rheinlandes 18), Stuttgart. Pirling, R., Siepen, M. (2003) Das Römisch-Fränkische Gräberfeld von Krefeld-Gellep 1989-2000 (Germanische Denkmäler der Völkerwanderungszeit Serie B. Die Fränkischen Altertümer des Rheinlandes 19), Stuttgart. Pirling, R., Siepen, M. (2006) Die Funde aus den römischen Gräbern von Krefeld-Gellep : Katalog der Gräber 6348-6361 (Germanische Denkmäler der Völkerwanderungszeit. Serie B. Die fränkischen Altertümer des Rheinlandes 20), Stuttgart. Pitchford, N.J., Biggam, C.P. (eds) (2006a) Progress in Colour Studies, vol. 1, Amsterdam. Pitchford, N.J., Biggam, C.P. (eds) (2006b) Progress in Colour Studies, vol. 2, Amsterdam. Platz-Horster, G. (1984) Die antiken Gemmen im Rheinischen Landesmuseum Bonn (Kunst und Altertum am Rhein 113), Cologne. Platz-Horster, G. (1987) Die antiken Gemmen aus Xanten im Besitz des Niederrheinischen Altertumsvereins, des Rheinischen Landesmuseums Bonn, der Katholischen Kirchengemeinde St. Viktor und des Regionalmuseums Xanten (Kunst und Altertum am Rhein 126), Bonn. Platz-Horster, G. (1994) Die antiken Gemmen aus Xanten II im Besitz des Archäologischen Parks-Regionalmuseums Xanten, der Katholischen Kirchengemeinde St. Mariae Himmelfahrt Marienbaum sowie in Privatbesitz (Führer und Schriften des Regionalmuseums Xanten 35), Cologne. Pleket, H.W. (1998) Tussen Rostovtzeff en Finley. De Economie van het Romeinse Keizerrijk in vergelijkend Perspectief, Lampas 31:4, 276-289. Plumier, J. (1986) Tumuli belgo-romains de la Hesbaye occidentale: Séron, Hanret, Bois de Buis, Penteville (Documents inédits relatifs à l‘archéologie de la région namuroise 2), Namur. Polfer, M. (2001) L‘archéologie de l‘artisanat et le débat sur la nature de l‘économie romaine : quelques réflexions critiques, in : Polfer, M. (ed.) L‘artisanat romain : évolutions, continuités et ruptures (Italie et provinces occidentales). Actes du 2 e colloque d‘Erpeldange (26-28 octobre 2001), (Monographies instrumentum 20), Montagnac, 7-17. Pomey, P., Long, L. (1992) Les premiers échanges maritimes du midi de la Gaule du VIe au IIIe s. av. J.-C. à travers les épaves, in: Bats, M., Bertucchi, G., Conges, G., Treziny, H. (eds.) Marseille grecque et la Gaule, Actes du Colloque international d‘Histoire et d‘Archéologie et du Ve Congrès archéologique de Gaule méridionale, Marseille 18-23 novembre 1990 (Etudes Massaliètes 3), 189-198 504 Popović, I. (1992) Les bagues. Les bijoux romains de Musée National de Belgrade, Belgrad. Price, J. (1978) Trade in Glass, in: Plat Taylor, J. du, Cleere, H. (eds) Roman Shipping and Trade: Britain and the Rhine Provinces (Council for British Archaeology Research Report 24), London, 70-78. Price, J. (1985) The glass, in: Bidwell, P.T. (ed.) The Roman Fort of Vindolanda at Chesterholm, Northumberland (Archaeological Report 1), London, 206-214. Price, J. (1988) Romano-British glass bangles from eastern Yorkshire, in: Price, J., Wilson P.R. (eds), Recent Research in Roman Yorkshire; studies in honour of Mary Kitson Clark (Mrs Derwas Chitty) (British Archaeological Reports British Series 193), Oxford, 339-366. Price, J. (1995) The glass vessels, in: Manning, W.H., Price, J., Webster, J., The Roman small finds (Report on the Excavatons at Usk 1965-1976), Cardiff, 139-191. Price, J. (2002) Broken bottles and quartz-sand: Glass production in Yorkshire and the North in the Roman period, in: Wilson, P., Price, J. (eds) Aspects of Industry in Roman Yorkshire and the North, Oxford, 81-93. Price, J. (2004) Roman Glass in Spain: The western provincial context, in: Fuentes Domninguez, A. (ed.) Jornadas sobre El vidrio en la España Romana, La Granja 1-2 November 2001, Cuenca, 13-31. Price, J. (2006) Glass in settlements and burials in Roman Britain, in: Creemers, G., Demarsin, B., Cosyns, P. (eds) Roman Glass in Germania Inferior. Interregional Comparisons and Recent Results, International colloquium Tongeren, 13/05/2005 (Atuatuca 1), Hasselt, 42-48. Price, J., Cottam, S. (1996) The Roman Glass from Fishbourne 1983 and 1985-6, in: Cunliffe, B.W., Down, A.G., Rudkin, D.J. (eds), Chichester Excavations IX. Excavations at Fishbourne 1969-1988, Chichester, 161-188. Price, J., Cottam, S., (1998) Romano-British glass vessels: A handbook, Practical Handbook in Archaeology 14, York, Council for British Archaeology. Prinsloo, L., Colomban, P. (2008) A Raman spectroscopic study of the Mapungubwe oblates: glass trade beads excavated at an Iron Age archaeological site at South Africa, Journal of Raman Spectroscopy 39, 79-90. Rahmani, L.Y. (1967) Jason‘s Tomb, Israel Exploration Journal 17:2, 61-100. Ravagnan, G.L. (1994) Vetri antichi del Museo Vetrario di Murano, Collezioni dello Stato (Corpus delle Collezioni Archeologiche del Vetro nel Veneto 1), Venice. Reade, W., Freestone, I.C., Bourke, S. (2009) Innovation and continuity in Bronze and Iron Age glass from Pella in Jordan, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour ‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 47-54. Reddé, M., Ballet, P., Lemaire, A., Bonnet C. (2004) Kysis: Fouilles de l'Ifao à Douch, oasis de Kharga (1985-1990) (Documents de fouilles de l'Ifao 42), Cairo. Rehren, T. (1997) Ramesside glass colouring crucibles, Archaeometry 39, 355-368. Rehren, T. (2000a) Rationales in Old World base glass compositions, Journal of Archaeological Science 27 (12), 1225-1234. Rehren, T. (2000b) New aspects of ancient Egyptian glassmaking, Journal of Glass Studies 42, 13-24. Renard, C. (ed.) (1999) Musée du Verre, Charleroi. Ricciardi, P. (2009) Raman spectroscopy for the non destructive characterization of ancient pottery, porcelain and mosaic glass, on-line doctoral dissertation at University of Firenze, Florence. Richmond, I.A. (1936) Roman Lead Sealings from Brough-under-Stainmore, Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society 36 NS, 104-125. Riemenschneider, U. (1989) Antike Gläser in Museum Altenessen, Boreas 12, 159-198. Riha, E. (1979) Die römischen Fibeln aus Augst und Kaiseraugst (Forschungen in Augst 3), Augst. Riha, E. (1990) Der Römische Schmuck aus Augst und Kaiseraugst (Forschungen in Augst 10), Augst. Riols, A. (2008) La bouteille en «verre noir façon d‘Angleterre» en basalte. Les recherches de Jean-Antoine Chaptal (1783) et les expérimentations d‘Allain Guillot (2005) Bulletin de l‘Association Française pour l‘Archéologie du Verre 2008, 114-115. Ritterling, E. (1913) Das Frührömische Lager bei Hofheim im Taunus (Annalen des Vereins für Nassauisches Altertumkunde und Geschichtforschung 40), Wiesbaden. Rober, A. (1983) Le sanctuaire gallo-romain de Matagne-la-Grande (Archaeologia Belgica 252), Brussels. Robin, L. (2008) L‘atelier de verrier de la Montée de la Butte à Lyon (milieu Ier s. ap. J.C. – début IIme s. ap. J.C.), aspects techniques et typologiques, Bulletin de l‘Association Française pour l‘Archéologie du Verre 2008, 42-46. Robinet, L., Coupry, C., Eremin, K., Hall, C. (2006) The use of Raman spectrometry to predict the stability of historic glass, Journal of Raman Spectroscopy 37, 789-797. Rodet-Belarbi, I., Van Ossel, P. (2003) Les épingles à tête anthropomorphe stylisée: un accessoire de la coiffure féminine de l'Antiquité tardive, Gallia 60, 319-368. Roosens, H., Lux, G.V. (1973) Grafveld met Gallo-Romeinse tumulus te Berlingen (Archaeologia Belgica 147), Brussels. Rotroff, S. (1997) Hellenistic pottery: Athenian and imported wheelmade table ware and related material (The Athenian Agora 29), Princeton. 505 Roussel-Ode, J. (2008) Le verre de la cour nord du sanctuaire des Bagnols à Alba-la-Romaine (Ardèche), Bulletin de l‘Association Française pour l‘Archéologie du Verre 2008, 58-61. Rouveret, A., Dubel, S., Naas, V. (eds) (2006) Couleurs et matières dans l‘antiquité. Textes, techniques et pratiques (Études de littérature ancienne 17), Paris. Roymans, N., van Rooijen, T. (1993) De voorromeinse glazen armbandproduktie in het Nederrijnse gebied en haar culturele betekenis, Vormen uit vuur 150, 2-10. Roztovtzeff, M.I. (1957²) Social and economic history of the Roman Empire, Oxford. Rütti, B. (1988) Die Gläser (Beiträge zum römischen Oberwinterthur - Vitudurum 4; Berichten der Züricher Denkmalpflege, Monographie 5), Zürich. Rütti, B. (1991) Die Römischen Gläser aus Augst und Kaiseraugst (Forschungen in Augst 13, 1-2), Augst. Sági, K. (1981) Das römische Gräberfeld von Keszthely-Dobogó (Fontes Archaeologici Hungariae 13) Budapest. Sánchez-Gijón Martínez, Á. (1966) Tumba de Bahía Blanca, Cádiz, Archivo Español de Arqueología 39:113-114, 183-193 Sas, K. (1993) A study of roman engraved gemstomes - intaglios and cameos - in Belgium, Helinium 33/1, 108-137. Sas, K., Thoen, H. (eds) (2002) Schone Schijn/Brillance et Prestige. Romeinse juweelkunst in West-Europa/La joaillerie romaine en Europe occidentale (exhibition catalogue), Leuven. Saunders, B. (2000) Revisiting basic color terms, Journal of the Royal Anthropological Institute 6, 81-99. Saunders, B., van Brakel, J. (1997) Are there nontrivial constraints on colour categorization?, Behavioural and Brain Sciences 20,167-228. Sayre, E.V., Smith, R.W. (1961) Compositional categories of ancient glass, Science 133, 1824-1826. Sayre, E.V., Smith, R.W. (1967) Some Materials of Glass Manufacture in Antiquity, in: Levey, M. (ed.) Archaeological Chemistry. A Symposium (3rd Symposium on Archaeological Chemistry, Atlantic City/New Jersey 1967), Atlantic City/New Jersey, 279-321. Schaad, D. (ed.) (1992) Le trésor d‘Eauze, Toulouse. Schädler, U. (2007) The Doctor‘s Game – New Light on the History of Board Games, in: Crummy P., Benfield, S., Crummy, N., Rigby, V., Shimmin, D., Stanway: An Élite Burial Site at Camulodunum (Britannia Monograph Series 24), London, 359-375. Schalm, O., Janssens, K. (2003) A flexible and accurate quantification algorithm for electron probe x-ray microanalysis based on thin-film element yields, Spectrochimica Acta B 58, 669-680. Schärlig, A. (2003) Rechnen ganz einfach. Der Römische Abacus – sein Griechischere Vorfahre und seine Nachkommenschaft bis nach Japan, Antike Welt 34, 517-520. Scheidel, W. (2004) Human Mobility in Roman Italy, I: The Free Population, Journal of Roman Studies 94, 1-26. Scheidel, W. (2005) Marriage, Families and Survival in the Roman Imperial Army: Demographic Aspects (Princeton/Stanford Working Papers in Classics), Princeton-Stanford. Scheidel, W. (2006) Population and demography (Princeton/Stanford Working Papers in Classics), Princeton-Stanford. Schibille, N., Marii, F., Rehren, T. (2008) Characterization and provenance of late antique window glass from the Petra church in jordan, Archaeometry 50:4, 627–642. Schlick-Nolte, B. (2002) Ancient Glass Vessels, in: Bianchi, R.S. (ed.) Reflections on Ancient Glass from the Borowski Collection. Bible Lands Museum Jerusalem, Mainz, 43-109. Schmid, D. (1993) Die römischen Mosaiken aus Augst und Kaiseraugst (Forschungen in Augst 17), Augst. Schneider-Schnekenburger, G. (1980) Churrätien im Frühmittelalter auf Grund der archäologischen Funde (Veröffentlichung der Kommission zur Archäologischen Erforschung des Spätrömischen Raetien der Bayerischen Akademie der Wissenschaften, Münchner Beiträge zur Vor- und Frühgeschichte 26), Munich. Scholze, H. (1991) Glass Nature, Structure and Properties, New York. Schreurs, J.W.H., Brill, R.H. (1984) Iron and sulphur related colours in ancient glasses, Archaeometry 26, 2, 199-209. Schuermans, H. (1864) Exploration de quelques tumulus de la Hesbaye, Bulletin des Commissions Royales d‘Art et d‘Archéologie 4, 3-559. Schuler, R. (1995) Nécropoles et sépultures d'époque romaine à Beauvais: état des connaissances, Revue archéologique de Picardie 3-4, 49-140. Schulze, M. (1978) Zur Interpretation spätkaiserzeitlicher Glasperlen, Archäologische Korrespondenzblatt 8, 51-68. Schulze-Dörrlamm, M. (1990) Bemerkungen zu Alter und Funktion der Alsengemmen, Archäologisches Korrezpondentsblatt 20, 215-226. Schuuring, I. (2007) 10. Glas, in: Schurmans, M., Verhelst, E., Oudheden uit Odijk. Bewoningssporen uit de Late IJzertijd, Romeinse tijd en Merovingische tijd aan de Singel West/Schoudermantel (Zuidnederlandse Archeologische Rapporten 30), Amsterdam, 125-131. Sear, F.B. (1977) Roman wall and vault mosaics (Mitteilungen des Deutschen Archaeologischen Instituts Rom 23), Heidelberg. Seibel, F. (2000) Technologie römischer Glashütten am Beispiel der Ausgrabungen im Hambacher Forst, in: Annales du 14e Congrès de l‘Association Internationale pour l‘Histoire du Verre (Venice-Milan 1998), Lochem, 156-159. 506 Sellner, C., Oel, H.J., Camara, B. (1979) Untersuchung alter Gläser (Waldglas) auf Zusammenhang von Zusammensetzung, Farbe und Schmeltzatmosphäre mit der Elektronenspektroskopie und der Elektronenspinresonanze (ESR), Glastechnische Berichte 52, 255-264. Sena Chiesa, G. (1966) Gemme del Museo Nazionale di Aquileia, Aquileia. Sena Chiesa, G. (1978) Gemme di Luni (Archaeologica 4), Rome. Sennequier, G. (1985) Verrerie d‘époque romaine (Collections des Musées départementaux de la Seine-Maritime 2), Rouen. Shindo, Y. (1996) Islamic glass bracelets found in the Red Sea region, in: Annales du 13e Congrès de l‘Association Internationale pour l‘Histoire du Verre (Netherlands 1995), Lochem, 269-276. Shindo, Y. (2001) The Classification and Chronology of The Islamic Glass Bracelets al-Tur, Sinai, Cultural Change in the Arab World (Senri Ethnological Studies 55), Osaka, 73-100. Shortland, A.J. (2004) Evaporites of the Wadi Natrun: seasonal and annual variation and its implication for ancient exploitation, Archaeometry 46, 497-516. Shortland, A., Schachner, L., Freestone, I., Tite, M. (2006) Natron as a flux in the early vitreous materials industry: sources, beginnings and reasons for decline, Journal of Archaeological Science 33, 521-530. Shugar, A., Rehren, T. (2002) Formation and composition of glass as a function of firing temperature, Glass Technology 43C, 145-150. Silvestri, A., Molin, G., Salviulo, G. (2006) Sand for Roman glass production: An experimental and philological study on source of supply, Archaeometry 48, 415-432. Silvestri, A., Molin, G., Salviulo, G. (2008) The colourless glass of Iulia Felix, Journal of Archaeological Science 35, 331-341. Simon, L. (2000) Verrerie, parures en verres, intaille, in: Berthaud, G. (ed.) Mazières-en-Mauges gallo-romain (Maine-et-Loire). Un quartier à vocation artisanale et domestique, Angers, 147-158. Simon, L. (2010) Des ―tablettes‖ en verre moulé d‘époque romaine découvertes à Bordeaux (Gironde), Aquitania, 26, 2010, 147-157. Simonett, C. (1941) Tessiner Gräberfelder (Monographien zur Ur- und Frühgeschichte der Schweiz 3), Basel. Simon-Hiernard, D. (2000) Verres d‘époque romaine. Collection des Musées de Poitiers, Poitiers. Simpson, G., Blance, B. (1998) Do brooches have ritual associations?, in: Bird, J. (ed.) Form and Fabric. Studies in Rome‘s material past in honour of B.R. Hartley (Oxbow Monograph 80), Oxford, 267-279. Simsek, G., Colomban, P., Milande, V. (2010) Tentative differentiation between Iznik tiles and copies with Raman spectroscopy using both laboratory and portable instruments, Journal of Raman Spectroscopy 41:5, 529–536. Slitine, F. (2005) Histoire du verre. L‘antiquité, Paris. Šmit, Ž., Janssens, K., Bulska, E., Wagner, B., Kos, M., Lazar, I. (2005) Trace element fingerprinting of façon-de-Venise glass, Nuclear Instruments and Methods in Physics Research B 239, 94-99. Smith, R.W. (1949) The Significance of Roman Glass, The Metropolitan Museum of Art Bulletin 8:2, 49-60. Smith, R.W. (1957) Glass from the Ancient World. The R.W. Smith Collection, Corning. Spaer, M. (1988) The Pre-Islamic glass bracelets of Palestine, Journal of Glass Studies 30, 51-61. Spaer, M. (1992) The Islamic glass bracelets of Palestine: preliminary findings, Journal of Glass Studies 34, 44-62. Spaer, M. (2001) Ancient Glass in the Israel Museum. Beads and Other Small Objects, Jerusalem. Steiger, R., Schwarz, G.T., Strobel, R., Doppler, H. (1977) Augst, Insula 31. Ausgrabungen und Funde 1960/61 (Forschungen in Augst 1), Augst. Stephens, J. (2008) Ancient Roman hairdressing: on (hair)pins and needles, Journal of Roman Archaeology 21, 110-132. Stern, E.M. (1977) Ancient glass at the Fondation Custodia (Collection Frits Lugt) Paris (Archaologica Traiectina Edita ab Academiae Rheno-Traiectinae Instituto Archaeologico 12), Groningen. Stern, E.M. (1995) Roman mold-blown glass: the First through Sixth centuries, The Toledo Museum of Art, Rome. Stern, E.M. (1999) Roman glassblowing in a cultural context (Commerce, society, imperial Rome), American Journal of Archaeology 103 (3), 441-484. Stern, E.M (2001) Roman, Byzantine, and Early Medieval Glass: 10 BCE – 700 CE. The Ernesto Wolf Collection, Ostfildern. Stern, E.M., Schlick-Nolte, B. (1994) Early Glass of the Ancient World: 1600 BC – 50 AD. The Ernesto Wolf Collection, Ostfildern. Sternini, M. (1995) La Fenice di Sabbia. Storia e tecnologia del vetro antico, Bari. Stevenson, R.B.K. (1956) Native bangles and Roman glass, Proceedings of the Society of Antiquaries of Scotland 88, 208-221. Stevenson, R.B.K. (1976) Romano-British glass bangles, Glasgow Archaeological Journal 4, 45-54. Strohm, G., (1920) Eine spätrömische Glashütte in den Argonnen, Korrespondenzblatt der Römischen Germanischen Kommision 2, 30-34. Swift, E. (2000) Regionality in Dress Accessories in the late Roman West (Monographies Instrumentum 11), Montignac. 507 Swift, E. (2003) Late-Roman bead necklaces and bracelets, Journal of Roman Archaeology 16, 336-349. Symonds, R.P. (1992) Rhenish Wares. Fine Dark Coloured Pottery from Gaul and Germany (Oxford University Committee for Archaeology Monograph 23), Oxford. Tait, H. (1995) Five Thousand Years of Glass, London. Tantrakarn, K., Kato, N., Hokura, A., Nakai, I., Fujii, Y., Glu čević, S. (2009) Archaeological analysis of Roman glass excavated from Zadar, Croatia, by a newly developed portable XRF spectrometer for glass, X-Ray Spectrometry 38, 121-127. Tartari, F. (2004) La nécropole du Ier-IVe sc. De notre ère à Durrachium, Durrës. Tempelmann-Maczyńska, M. (1985) Die Perlen der römischen Kaiserzeit und der frühen Phase der Volkenwanderungszeit im mitteleuropäischen Barbaricum (Römischen-Germanischen Forschungen 43), Mainz. Tester, P.J. (1968) An Anglo-Saxon cemetery at Orpington. First Interim Report, Archaeologia Cantiana 83, 125-150. Toma ević-Buck, T. (1980) Ein Depotfund in Aug. Raurica, Insula 42 (Forschungen in Augst 4), Augst. Topál, J. (2003) Roman cemeteries of Aquincum, Pannonia. The Western Cemetery (Bécsi Road) II, Budapest. Torrecilla Aznar, A. (2004) Los vidrios romanos de la Villa de El Saucedo (Talavera la Nueva, Toledo), in: Domínguez, A.F. (ed.) Jornadas sobre ‗El vidrio en la Espaňa Romana‘ (La Granja 2001), 323-349. Tremmel, B. (2006) V. Glasgefässe und Glasperlen aus Haltern, Anreppen und Oberaden, in: Roth-Rubi, K., Rudnick, B., Schneider, G., Ellinghaus, C., Tremmel, B., Müller, M., Varia Castrensia. Haltern, Oberaden, Anreppen (Bodenaltertümer Westfalens 42), Mainz am Rhein, 235-285. Trowbridge, M.L. (1930) Philological Studies in Ancient Glass (University of Illinois Studies in Language and Literature 13, 3-4), Urbana. Turner, W.E.S. (1956a) Studies in ancient glasses and glassmaking processes. Part III The chronology of glassmaking constituent, Journal of the Society of Glass Technology 40, 39-52. Turner, W.E.S. (1956b) Studies in ancient glasses and glassmaking processes. Part IV The chemical composition of ancient glasses, Journal of the Society of Glass Technology 40, 162-186. Turner, W.E.S. (1956c) Studies in ancient glasses and glassmaking processes. Part V Raw material and melting processes, Journal of the Society of Glass Technology 40, 277-300. Tu ek, I. (1993) Rimsko grobi če ob novi obvoznici ob Potrčevi cesti v Ptuju. - V (with English summary: The Roman cemetery on the By-pass at Potrčeva Cesta in Ptuj, 411-412), in: Ptujski arheološki zbornik, Ptuj, 385-448. Tytgat, C. (1992) Verre, in: Karageorghis, V., Picard, O., Tytgat, C. (eds) La nécropole d‘Amathonte: tombes 113-367. 6: Bijoux, armes, verre, astragales et coquillages, squelettes (Etudes chypriotes 14), Nicosia, 85-86. Vágó, E.B., Bóna, I. (1976) Der spätrömische Südostfriedhof. Die Gräberfelder von Intercisa 1, Budapest. Van den Hurk, L.J.A.M. (1973) The tumuli from the Roman period of Esch, Province of North Brabant I, Berichten van de Rijksdienst voor Oudheidkundig Bodemonderzoek 23, 189-236. Van den Hurk, L.J.A.M. (1975) The tumuli from the Roman period of Esch, Province of North Brabant II, Berichten van de Rijksdienst voor Oudheidkundig Bodemonderzoek 25, 69-92. Van den Hurk, L.J.A.M. (1977) The tumuli from the Roman period of Esch, Province of North Brabant III, Berichten van de Rijksdienst voor Oudheidkundig Bodemonderzoek 27, 91-138. Van den Hurk, L.J.A.M. (1980) The tumuli from the Roman period of Esch, Province of North Brabant IV, Berichten van de Rijksdienst voor Oudheidkundig Bodemonderzoek 30, 367-392. Vanderhoeven, M. (1958) Verres tardifs et mérovingiens du Musée Curtius, Liège. Vanderhoeven, M. (1962) De Romeinse glasverzameling in het Provinciaal Gallo-Romeins Museum (Publicaties van het Provinciaal Gallo-Romeins Museum Tongeren 2), Tongeren. Vanderhoeven, M. (1976) De Gallo-Romeinse tumulus van Riemst, Limburg 55, 3-29. Van der Linden, V., Cosyns, P., Schalm, O., Cagno, S., Nys, K., Janssens, K., Nowak, A., Wagner, B., Bulska, E. (2009) Deeply coloured and black glass in the northern provinces of the Roman Empire: differences and similarities in chemical composition before and after 150 AD, Archaeometry 51, 5, 822-844. Van Driel-Murray, C. (1994) A question of gender in a military context, Helenium 34/2, 342-362. van Enckevort, H. (2008) Begraven sieraden en verbrand geld op de Nijmeegse Hunerberg, Ulpia Noviomagus 13, Nijmegen. Vanhaeren, M. (2005) Speaking with beads: The evolutionary significance of personal ornaments, in: d‘Errico, F., Backwell, L. (eds) From Tools to Symbols, Witwatersrand, 525-553. Vanhoute, S. (2009) The Saxon Shore fort at Oudenburg (Belgium): new excavations results, in: Morillo, A., Hanel, N., Martin, E. (eds) Limes XX. XX Congress of Roman frontier studies, Madrid, 1383-1392. Vanhoute, S, Bastiaens, J., De Clercq, W., Deforce, K., Ervynck, A., Fret, M., Haneca, K., Lentacker, A., Stieperaere, H., Van Neer, W., Cosyns, P., Degryse, P., Dhaeze, W., Dijkman, W., Lyne, M., Rogers, P., Van Driel-Murray, C., van Heesch, J., Wild, J.P. (2009) De dubbele waterput uit het laat-Romeinse castellum van Oudenburg (prov. West-Vlaanderen): tafonomie, chronologie en interpretatie, Relicta 5, 9-141. van Lith, S.M.E. (1977) A Romano-British bangle from Valkenburg Z.H., Ex Horreo IPP 1951-1976. Cingula IV, 130-134. 508 van Lith, S.M.E. (1987) Glas aus Asciburgium (Funde aus Asciburgium 10), Duisburg. van Lith, S.M.E. (1994) Die römischen Gläser von Neuss. Gesamtkatalog der Ausgrabungen 1955-1978, Bonner Jahrbücher 194, 205-340. van Lith, S.M.E. (2009) Römisches Glas aus Nijmegen (Rijksdienst Oudheidkundig Bodemonderzoek Nederlandse archeologische rapporten 38), Amersfoort. van Lith, S.M.E., Randsborg, K. (1985) Roman Glass in the West: A Social Study, Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 35, 413-532. Van Ossel, P., Defgnée, A., (eds) (2001) Champion, Hamois. Une villa romaine chez les Condruses. Archéologie, environnement et économie d‘une exploitation agricole antique de la Moyenne Belgique (Etudes et Documents Archéologie 7), Namur. Vanpeene, N. (1993) Verrerie de la nécropole d‘Epiais-Rhus (Val-d‘Oise) (Cahier Archéologique 8), Guiry-en-Vexin. Vanpeene, N. (2007) Découverte fortuite de verreries en Haute-Normandie à Lyons-la-Forêt (Eure), Bulletin de l‘Association Française pour l‘Archéologie du Verre 2007, 25-27. Vanvinckenroye, W. (1984) De Romeinse zuidwest begraafplaats van Tongeren (Opgravingen 1972-1981) (Publicaties van het Provinciaal Gallo-Romeins Museum te Tongeren 29), Tongeren. Vanvinckenroye, W. (1988) De Romeinse villa op de Sassenbroekberg te Broekem (Publicaties van het Provinciaal GalloRomeins Museum te Tongeren 38), Hasselt. Venclová, N. (1990) Prehistoric Glass in Bohemia, Prague. Vessberg, O. (1952) Roman Glass in Cyprus, Opuscula Archaeologica 7, 109-165. Vichy, M., Thirion-Merle, V., Picon, M. (2007) Note sur le recyclage du verre dans l‘Antiquité et sur les groups de recyclage, Bulletin de l‘Association Française pour l‘Archéologie du Verre 2007, 55-57. Vilvorder F. (2001) La verrerie, in: Brulet, R., Demanet, J.-C., Vilvorder, F. (eds) Liberchies IV, vicus gallo-romain. Travail de rivière: fouilles du Musée de Nivelles (1986/87 et 1991-97) (Publications d‘histoire de l‘art et d‘archéologie de l‘Université catholique de Louvain CI), Louvain-la-Neuve, 104-110. Vilvorder F. (1997) La verrerie et l‘artisanat du verre, in: Brulet, R., Demanet, J.-C. (eds) Liberchies III, vicus gallo-romain. Les thermes: fouilles du CRAN (1973 et 1989-90) et zone d‘habitat au nord de la voie antique: fouilles de Pro Geminiaco (1979-80 et 1987-94) (Publications d‘histoire de l‘art et d‘archéologie de l‘Université catholique de Louvain XCIV), Louvain-la-Neuve, 146151. Vilvorder F. (1993) La verrerie, in: Brulet, R. (ed.) Braives Gallo-Romain V. La fortification du Bas-Empire (Publications d‘Histoire de l‘Art et d‘Archéologie de l‘Université Catholique de Louvain LXXXIII), Louvain-la-Neuve, 111-115. Vitrum 1990: Vitrum. Le verre en Bourgogne, Autun. Vollenweider, M.-L. (1979) Catalogue raisonné des sceaux, cylindres, intailles et camées II: Les portraits, les masques de théatre, les symboles politiques, Mainz. von Saldern, A. (1980) Ancient and Byzantine Glass from Sardis (Archaeological Exploration of Sardis 6), Cambridge (Massachusetts). von Schnurbein, S. (1982) Die unverzierte Terra Sigillata aus Haltern (Bodenaltertümer Westfalens 19/1-2), Münster. Vrielynck, O. (2007) La nécropole mérovingienne de Bossut-Gottechain, commune de Grez-Doiceau, Brabant wallon (Catalogue exposition. Espace archéologique Saint-Pierre, Namur, 24.06.2006 - 13.05.2007), s.l.. Wagner, H. (2006) Glasschmuck der Mittel- und Spätlatènezeit am Oberrhein und den angrenzenden Gebieten (Ausgrabungen und Forschungen 1), Remshalden. Wagner, B., Bulska, E. (2004) On the use of laser ablation inductively coupled plasma mass spectrometry for the investigation of the written heritage, Journal of Analytical Atomic Spectrometry 19, 1325-1329. Wagner, B., Nowak, A., Bulska, E., Kunicki-Goldfinger, J., Schalm, O., Janssens, K. (2008) Complementary analysis of historical glass by scanning electron microscopy with energy dispersive X-ray spectroscopy and laser ablation inductively coupled plasma mass spectrometry, Microchim Acta 162, 415–424. Warmenbol, E. (1984) Essai d‘interprétation des vestiges d‘époque romaine (Ier-IIIème s.) trouvés au Trou del Leuve de Sinsin (Namur, Belgique), Amphora 37, 1-23. Warmenbol, E. (1996) L‘or, la mort et les Hyperboréens. La bouche des Enfers ou le Trou de Han à Han-sur-Lesse, in: Archäologische Forschungen zum Kultgeschehen in der jüngeren Bronzezeit und frühen Eisenzeit Alteuropas. Ergebnisse eines Kolloquiums in Regensburg 4-7 Oktober 1993 (Regensburger Beiträge zur Prähistorischen Archäologie 2), Bonn, 203-234. Warmenbol, E. (1999) Le soleil des morts. Les ors protohistoriques de Han-sur-Lesse (Namur, Belgique), Germania 77, 39-69. Watts, L., Leach, P. (1996) Henley Wood, temples and cemetery: excavations 1962-69 by the late Ernest Greenfield and others (Council for British Archaeology Research Reports 99), London. Webster, G. (1955) A note on the use of coal in Roman Britain, Antiquaries Journal 35, 199-217. Wedepohl, K.H. (2009) The source of lead as a major constituent in ancient glass, in: Janssens, K., Degryse, P., Cosyns, P., Caen, J., Van‘t dack, L. (eds) Annales of the 17th Congress of the Association Internationale pour l‘Histoire du Verre (Antwerp 2006), Antwerp-Brussels, 616-624. 509 Wedepohl, K.H., Baumann, A. (2000) The use of marine molluskan shells for Roman glass and local raw glass production in the Eifel area (Western Germany). Naturwissenschaften 87, 129–132. Weinberg, G.D. (ed.) (1988) Excavations at Jalame: site of a glass factory in late roman Palestine, Columbia (Missouri). Wesemael, E., Hensen, G., Doucet, H. (2004) Een merkwaardig vondstencomplex aan de Romeinse weg te Straten (Bevingen), bij Sint Truiden, Tongeren. Weyl, A. (1959) Coloured Glasses, London. Wheeler, R.E.M., Wheeler, T.V. (1932) Report on the Excavation of the Prehistoric, Roman, and Post-Roman Site in Lydney Park, Gloucester (Report of the Research Committee of the Society of Antiquaries of London 9), London. Whitehouse, D. (1999) The Date of the Glass from Karanis, Journal of Glass Studies 41, 168-170. Whitehouse, D. (2003) Roman Glass in The Corning Museum of Glass. Vol. 3, Corning (New York). Whitehouse, D. (2004) Glass in the Price Edict of Diocletian, Journal of Glass Studies 46, 189-191. Wick, S. (2008) Keltische Glasarmringe, Archäologie der Schweiz 31,1, 30-33. Willems, J., Amand, M. (1960) A propos du cimetière belgo-romain d‘Amay (Liège), Latomus 45, 759-767. Wilmott, T., Hird, L., Izard, K., Summerfield, J. (1997) Birdoswald: excavations of a Roman fort on Hadrian‘s Wall and its successor settlements: 1987 – 92 (English Heritage Archaeological Report 14), London. Wirth, H. (2010) Die linke Hand: Wahrnehmung und Bewertung in der griechischen und römischen Antike (Heidelberger Althistorische Beiträge und Epigraphische Studien 4), Stuttgart. Witvrouw, J., Witvrouw, D. (1976) Le Sanctuaire Belgo-Romain de Clavier-Vervoz, Bulletin du Cercle Archéologie HesbayeCondroz 14, 147-216. Woimant, G.-P. (1995) L'Oise (Carte archéologique de la Gaule 60), Paris. Woolf, G. (2004) The Present State and Future Scope of Roman Archaeology: A Comment, American Journal of Archaeology 108, 417-428. Woodward, A. & Leach, P. (1993) The Uley shrines, excavation of a ritual complex on West Hill, Uley, Gloucestershire: 1977-9 (English Heritage Archaeological Report 17), London. Zahn, R (1929) Sammlung Baurat Schiller, Sales Catalogue Lepke, Berlin. Zampieri, G. (1998) Vetri antichi del Museo Civico Archeologico di Padova (Corpus delle Collezioni Archeologiche del Vetro nel Veneto 3), Venice. Zaqzuq, A., Duchesne-Guillemin, M. (1970) La mosaïque de Mariamin (conservée au musée de Hama), Annales Archéologiques Arabes Syriennes. Revue d‘Archéologie et d‘Histoire 20, 93-104. Zazoff, P. (1983) Handbuch der Archäologie. Die antiken Gemmen, Munich. Zepezauer, M.A. (1989) Perles à décor occulé spiralé de La Tène moyenne et finale, in: Feugère, M. (ed.) Le verre préromain en europe occidentale, Montagnac, 107-120. Zepezauer, M.A. (1992) Glasperlen der vorrömischen Eisenzeit III. Mittel- und spätlatènezeitliche Perlen (Marburger Studien zur Vor- und Frühgeschichte 15), Marburg. Zouhdi, B. (1974) Bracelets et bagues de verre du Musée National de Damas, in: Annales du 6e Congrès International d‘Étude Historique du Verre (Cologne 1-7 juillet 1973), Liège, 85-99. Zwierlein-Diehl, E. (1973) Die antiken Gemmen des Kunsthistorischen Museums in Wien I: Die Gemmen von der minoischen Zeit bis zur frühen römischen Kaiserzeit, Munich. Zwierlein-Diehl, E. (1979) Die antiken Gemmen des Kunsthistorischen Museums in Wien II: Die Glasgemmen, die Glaskameen und die Gemmen der späteren römischen Kaiserzeit, Munich. 510 LIST OF FIGURES Figure 1: Schematic view how to reach a typo-chronology ......................................................................................................... xv Figure 2: a) Carchesium of Heel (NL); b+c) Bulbous cup of Heerlen (NL) (photographs by author, by courtesy of Limburgmuseum, Venlo); d+e) Bulbous jar (picture d) from Foy, Nenna 2001, 85, fig.93; picture e) from Harden et al. 1987, 111, fig.44) .......................................................................................................................................................................................... 5 Figure 3: a) Counter from Cortil-Noirmont (BE) (left part reconstructed); b) Globular cup from Heerlen (NL) (photographs by author, a) by courtesy of KMKG-MRAH, Brussels; b) by courtesy of RMO, Leiden) ..................................................................... 6 Figure 4: a) Black glass counter from Nijmegen (NL) with light in reflection (left) and light in transmission (right); b) Black glass counter from Oudenburg (BE) exposed to white light in transmission (photographs by author, a) by courtesy of GAD, Nijmegen; b) by courtesy of VIOE, Brussels) ................................................................................................................................................ 6 Figure 5: Pie-chart with different hues of Roman black appearing glass artefacts ........................................................................ 7 Figure 6: Pie-chart with different hues of Roman black appearing glass artefacts (excluding the unidentified group) ................... 8 Figure 7a-c: Pie-chart with the ratios between undecorated vs. decorated black appearing glass artefacts (excluding the unidentified group) ....................................................................................................................................................................... 9 Figure 8a-d: Pie-charts defining the ratios of glass hues utilized for the core of the artefact (a + b) and for the applied decoration (c + d) ........................................................................................................................................................................................ 10 Figure 9: a) Irregular circular mark from a massive punty gob; b) Thick ring-shaped mark from a hollow punty gob or ring pontil; c) Thin irregular annular scar from the moil of the blowing pipe (photographs by author) ........................................................... 17 Figure 10a-d: Various uses of jacks and pincers a) to model a base-ring [left: cat.no.595; right: cat.no.414]; b) to apply glass trails [Kaiseraugst - cat.no.3721]; c) to model handles [Liberchies – top: cat.no.413 ; bottom: cat.no.407]; d) reconstruction of how pincers were utilized to apply glass trails on an artefact (drawings and pictures by author, by courtesy of PAM, Velzeke; Musée Archéologique, Nivelles; Augusta Raurica, Augst; CRAN, Louvain-la-Neuve) ................................................................. 18 Figure 11: Wheel-cut fragment in the Petrie Museum for Egyptian Archaeology, London (photograph by author, by courtesy of Petrie Museum) ......................................................................................................................................................................... 23 Figure 12: Vessel with rolled-in glass particles in a contrasting colour to be marvered from the Canal Bianco necropolis, Adria (IT) (photograph taken from Barovier et al. 2003, 231, no.3; cat.no.2586) .................................................................................. 24 Figure 13: Reconstruction of possible production process (drawing by author) .......................................................................... 25 Figure 14: Pitted surface due to a thermal shock on bottom of the Heel carchesium (NL) [cat.no.2883] (photograph by author, by courtesy of Limburgmuseum, Venlo) .......................................................................................................................................... 25 Figure 15: (left) Thin elongated pointed elliptical scar of a punty on apex of the top surface of an open elliptical bracelet from Poitiers-Vienne (FR) [cat.no.1345] (drawing by author, by courtesy of MAN, Saint-Germain-en-Laye); (right) Closed, circular bracelet with clear seam from Kaiseraugst (CH) [cat.no.3687] (drawing taken from Riha 1990, pl.74:2930) ............................... 27 Figure 16: Possible reconstruction of production process (drawing by author) ........................................................................... 29 Figure 17: Basal side of the so-called ‗Trilobitenperle‘ from Tielrode (BE) [cat.no.529] showing concentric chill marks (photograph by Marc De Meirelaer, by courtesy of Museum Van Bogaert-Wauters, Hamme) .................................................... 29 Figure 18: Tesserae in various shapes and sizes from Les Houis, near Sainte Menehould (FR) (drawings by author, by courtesy of MAN, Saint-Germain-en-Laye) ............................................................................................................................................... 32 Figure 19: Two counters from the castellum Oudenburg (BE) a) with a flat and smooth lower surface [cat.no.465]; b) with a slightly irregular, pitted surface [cat.no.466]; c) counter from a re-fused glass sherd from Utrecht (NL) [cat.no.2989] (photographs by author, a-b) courtesy by VIOE, Brussels; c) courtesy by ACVU, Amsterdam) .................................................. 33 Figure 20: Ratios of the various production techniques for black glass vessels .......................................................................... 35 Figure 21: Ratios of the various production techniques for black glass vessels in relation to the adopted glass hue .................. 36 Figure 22: Ratios of the cast black glass vessels in relation to the adopted glass hue ................................................................ 36 Figure 23: Ratios of the free-blown black glass vessels in relation to the adopted glass hue ...................................................... 37 Figure 24: Ratios of the mould-blown black glass vessels in relation to the adopted glass hue .................................................. 38 Figure 25: Ratios of the rod-formed black glass vessels in relation to the adopted glass hue ..................................................... 38 Figure 26:a-b) Stacked bar charts of the different commodity types per technique (in terms of quantity and percentage); c-d) Stacked bar charts of the different techniques applied per commodity type (in terms of quantity and percentage) ..................... 40 Figure 27: Pie-charts of the colour ratios of the bracelets; (left) with undefined hue incorporated;(right) with defined hues only . 41 Figure 28: Pie-charts of the colour ratios of the finger rings; (left) with undefined hue incorporated (right) with defined hues only41 Figure 29: Pie-charts of the colour ratios of the beads; (left) with undefined hue incorporated; (right) with defined hues only ..... 42 Figure 30: Pie-charts of the colour ratios of the pendants; (left) with undefined hue incorporated (right) with defined hues only . 42 Figure 31: Pie-charts of the colour ratios of the gems; (left) with undefined hue incorporated; (right) with defined hues only ...... 43 Figure 32: Pie-charts of the colour ratios of the hairpins; (left) with undefined hue incorporated; (right) with defined hues only .. 44 511 Figure 33: Period I cast vessel shapes in strongly coloured glass appearing black (Forms IA) .................................................. 48 Figure 34: First century AD free-blown vessel shapes in strongly coloured glass appearing black (Forms IB) ........................... 52 Figure 35: Typology of free-blown shapes (drawings nos. 1-14 from Cosyns, Hanut 2005, fig.2; nos. 15-20 drawn by author; no.19 from Isings 1957) ............................................................................................................................................................. 56 Figure 36: (top) Base fragment of a small bulbous cup (Form II.B.4) from the Roman villa at Lauw, Belgium; (bottom) Mouldblown globular beaker from Bois-et-Borsu (photograph by author, by courtesy of PGRM Tongeren (top) by courtesy of KMKG (bottom)) ................................................................................................................................................................................... 58 Figure 37: Small bulbous cup in Württembergisches Landesmuseum Stuttgart, formerly in the Ernesto Wolf Collection (illustration from Stern 2001, 193, no. 84) .................................................................................................................................. 59 Figure 38: (left) Faulquemont, France (photograph by Hubert Cabart); (right) MET 17.194.166, Metropolitan Museum of Art, New York (photograph by Chris Lightfoot, courtesy by Metropolitan Museum of Art, New York) ........................................................ 59 Figure 39: a two-handled jug from unknown provenance (LU) (photograph by Jeannine Geyssant by courtesy of MNAH, Luxemburg) ............................................................................................................................................................................... 61 Figure 40: Typology of Period III free-blown shapes (drawings taken from Pirling, Siepen 2006 (1-2); Dilly, Mahéo 1997 (3)) ... 65 Figure 41: Typology of mould-blown vessels (1-5 = Forms IC; 6-7 = Forms IIC; 8 = Form IVC) (all drawings by author; photograph (3) taken from Whitehouse 2001, 65, no.536) ......................................................................................................... 66 Figure 42: (left) The hexagonal bottle (Form IB.4) and (right) the double head-shaped bottle (Form IB.5) from the Toledo Museum of Art (photographs taken from Stern 1995, colour plates 7; 23). ................................................................................. 68 Figure 43: The grape-shaped amphoriskos from Nijmegen-Hees (Photograph by the author, courtesy of the Museum Het Valkhof, Nijmegen) .................................................................................................................................................................... 69 Figure 44: The grape-shaped amphoriskos from Bologna, Italy(illustration taken from Meconcelli Notarianni 1979, no. 213) .... 69 Figure 45: Two fragments of mould-blown prismatic bottles from the ‗Aussere Reben‘ workshop at Kaiseraugst (CH) [cat.no.3777 (left); cat.no.3462 (right)] (photographs by the author, by courtesy of the Römermuseum, Augst) ......................... 70 Figure 46: Rod-formed unguentaria with triangular shaped solid body and tall neck from Elkab (EG) (drawing by F. Roloux, by courtesy of the MRAH-KMKG, Brussels; photograph by author, by courtesy of the MRAH-KMKG, Brussels) ............................ 71 Figure 47: squad candlestick unguentarium with solid body (photograph by courtesy of the Yale University Art Gallery, New Haven) ...................................................................................................................................................................................... 72 Figure 48: Rod-formed alabastra with twisted body without foot (left) or with foot (right) (drawings by author, by courtesy of Grand Curtius, Liège) ................................................................................................................................................................ 73 Figure 49: The various subtypes of squat jars with their various decorations (drawings by author, by courtesy of Grand Curtius, Liège; photographs taken from Arveiller, Nenna 2005, no. 1156 and Israeli 2003, no.394) ........................................................ 75 Figure 50: Squat globular jar with pressed decoration (Form IVD.2:3) (photograph and drawings by author, by courtesy of Grand Curtius, Liège) ........................................................................................................................................................................... 76 Figure 51: (top) Truncated conical (Form IVD.3); (bottom) biconical jar (Form IVD.4) ................................................................ 77 Figure 52: Typology of Roman glass bracelets set up by Hubert De Witte (taken from De Witte 1977, 21) ................................ 83 Figure 53: Typology of the black glass bracelets (drawings by the author)................................................................................. 86 Figure 54: Open bracelet of type A4 from Nijmegen showing with an unusual crested spatula-shaped end (drawing by the author, by courtesy of the Gemeentelijke Archeologische Dienst, Nijmegen) ......................................................................................... 88 Figure 55: Pie-charts demonstrating a distribution between the different types and between the different variants per type ...... 92 Figure 56: Typology proposed by Helène Guiraud based on the material from France (taken from Guiraud 1989, figs. 9, 11; 21; 26; 32; 37; 41; 45; 47) ............................................................................................................................................................... 94 Figure 57: Proposed typology of Roman black glass finger rings (drawings by the author) ........................................................ 95 Figure 58: Column-chart and pie-chart correspondingly showing the ratios in terms of quantity (top) and percentage (bottom) of the different types of black glass finger rings ............................................................................................................................. 96 Figure 59: Pie-chart displaying the ratios of the various type A finger rings................................................................................ 97 Figure 60: Pie-chart displaying the ratios of the various subtypes of the simple annular finger ring............................................ 97 Figure 61: Pie-chart displaying the ratios of the various type B finger rings................................................................................ 99 Figure 62: (left + middle) Finger ring from Trou de Han (drawing by the author, courtesy of Musée du Monde Souterrain, Hansur-Lesse); (right) finger ring fragment from Wanborough (drawing taken from Monk 2001, fig. 67)......................................... 102 Figure 63: Plain hexagonal finger ring in dark green glass appearing black from Trou del Leuve at Sinsin (BE) (drawing taken from Warmenbol 1984, 8, fig.6) ............................................................................................................................................... 102 Figure 64: Theoretical classification of Type C finger rings ...................................................................................................... 103 Figure 65: Pie-chart displaying the ratios of the various type C finger rings ............................................................................. 104 Figure 66: Classification of the Roman glass beads from the British Isles (after Margaret Guido 1978, fig.37); Encircled with full red line the types known in black glass; encircled with dotted red line the types in black glass recorded on the continent ....... 105 512 Figure 67: Chart based on the different types of black glass beads (by author) ........................................................................ 107 Figure 68: Pie-chart showing the ratios of the different bead classes in black glass ................................................................. 108 Figure 69: Roman necklace combining a golden chain with black glass discoid beads (photograph by author, by courtesy of the Archaeological Museum of Ancona) ......................................................................................................................................... 112 Figure 70: overview of Roman cast and pressed beads produced in black glass (3: drawings by author;1: taken from Van den Hurk 1980, fig.6:m; 2; 5-8: taken from Spaer 2001, 76) ............................................................................................................ 113 Figure 71: Organization chart of portrait beads ........................................................................................................................ 115 Figure 72: Schematic view of the various Roman pendant types distinguished in black glass (drawings by author or taken from Spaer 2001) ............................................................................................................................................................................. 116 Figure 73: Diversity of the poppy-head-shaped pendants: a. plain; b. ribbed; c. wavy/zigzag line; d. mottled (drawing by author)118 Figure 74: Diversity of the barrel-shaped pendants: top: single looped – left: elongated (Trier); middle+right: short (Cyprus; Colchester); bottom: multiple looped - Aquileia (top: drawings by author; Colchester taken from Crummy 1983, no.1504; photograph taken from Mandruzatto 2008) .............................................................................................................................. 118 Figure 75: Diversity of the jug-shaped pendants (all drawings by autho, except left taken from Tempelmann-Maczyńska 1985; other) ....................................................................................................................................................................................... 119 Figure 76: Overview of black glass gemstones: 1) plain conical; 2) plano-convex; 3) small elliptical; 4) large round ................. 121 Figure 77: Overview of gem shapes (illustration taken from Henig 1978², fig.1; after E. Zwierlein-Diel & J. Boardman) ............ 123 Figure 78: Glass hairpins with a) spherical head; b) conical head; c) discoid head; d) elliptical head ....................................... 125 Figure 79: Bronze hairpin with black glass head from Shakenoak (UK) (taken from Harden 1971, fig.45:69 – scale 1:1) ......... 125 Figure 80: Selection of black glass tesserae shapes from the site at Les Houis, nearby Sainte Menehould (FR) in Musée des Antiquités Nationales (M.A.N.), Saint-Germain-en-Laye (photograph by the author, courtesy of M.A.N.) ................................. 126 Figure 81: Set of tesserae from the site at Les Houis, nearby Sainte Menehould (FR) in Grand Curtius Museum, Liège (BE) since the late 19th century (photograph by author, courtesy of Grand Curtius Liège) ................................................................ 127 Figure 82: Set of 455 black tesserae (Inv.no. 83401-B) from the site at Les Houis, nearby Sainte Menehould (France) in Musée des Antiquités Nationales (M.A.N.), Saint-Germain-en-Laye (France) (photograph by the author, courtesy of M.A.N.)............. 127 Figure 83: An overview of the R1609 material from the KMKG-MRAH, Brussels (photograph by the author, courtesy by KMKGMRAH) ..................................................................................................................................................................................... 128 Figure 84: Chart of the rod varieties based on technology ........................................................................................................ 129 Figure 85: Stirring rod from tomb 272 at Amathous (CY) (drawing by the author, by courtesy of the District Museum of Limassol)130 Figure 86: An overview of the R1609 rectangular bichrome plaques from the KMKG-MRAH, Brussels (photograph by the author, courtesy by KMKG-MRAH, Brussels) ....................................................................................................................................... 131 Figure 87: Schematic view on the various features used to classify Roman glass counters...................................................... 132 Figure 88: Plot of the dimensions of Roman black glass counters (in mm) ............................................................................... 133 Figure 89: Plot opposing the dimensions of the plain against the decorated counters in black glass (in mm) ........................... 134 Figure 90: Variety of blue glass used to apply a dotted decoration on counters ........................................................................ 136 Figure 91: Ratios of black glass counters per period from dated contexts ................................................................................ 167 Figure 92: Plot of the black glass counters according to the dimensions .................................................................................. 169 Figure 93: Scheme of the subdivisions of this chapter .............................................................................................................. 176 Figure 94: Ratios of sites and entries in relation to the settlement character (incorporating or excluding the items from unknown provenance) ............................................................................................................................................................................. 178 Figure 95: Pie-charts showing the ratio of black glass artefacts from the various Roman civil contexts .................................... 180 Figure 96: Column chart showing the quantity of the various commodities per context category .............................................. 181 Figure 97: Ratios of the different context categories per commodity type (top: including the material of unknown provenance; bottom: excluding the material of unknown provenance) .......................................................................................................... 182 Figure 98: Ratios of the different jewellery commodities per context category .......................................................................... 183 Figure 99: Stacked-bars showing the quantities of the four arm ring categories per context category ...................................... 184 Figure 100: Comparison of the multiple values of the various bracelet types per context ......................................................... 185 Figure 101: Functional types of black glass artefacts from domestic civil contexts ................................................................... 185 Figure 102: Black glass vessels from domestic civil contexts per period .................................................................................. 186 Figure 103: Pie-chart showing the ratios of the different bangle types from domestic contexts ................................................. 188 Figure 104: Functional types of black glass artefacts from Roman burials ................................................................................ 192 513 Figure 105: Pie charts of the vessels from funerary contexts (left: considering all entries, right: omitting the entries of undefined date) ........................................................................................................................................................................................ 192 Figure 106: Stacked bar charts showing the ratios per period of the different types of black glass jewellery from civil/funerary contexts in amount and in proportion ....................................................................................................................................... 196 Figure 107: Pie-chart showing the functional types of black glass artefacts from Roman ritual contexts .................................. 203 Figure 108: Black glass from military contexts (right: with additional entries) ........................................................................... 206 Figure 109: scheme 1 .............................................................................................................................................................. 225 Figure 110: scheme 2 .............................................................................................................................................................. 225 Figure 111: scheme 3 (top: variant a – bottom: variant b) ........................................................................................................ 226 Figure 112: scheme 4 .............................................................................................................................................................. 226 Figure 113: Distribution map of the known or presumed glass workshops that produced black glass artefacts ........................ 228 Figure 114: Column charts demonstrating the concentrations of various black glass commodities from ites with a clear or supposedly secondary glass workshop where black glass has been manufactured ................................................................. 230 Figure 115: Fragments from the Regio 17B glass workshop of Kaiseraugst (CH) (picture by author, by courtesy of Römermuseum, Augst)............................................................................................................................................................ 231 Figure 116: top) batch fragment from Äussere Reben workshop at Kaiseraugst (CH) with unmelted tessera in ‗naturally coloured‘ glass; bottom) production waste from the same workshop showing four unmelted cubic-shaped inclusions in transparent pale blue-green glass (pictures by author, by courtesy of Römermuseum Augusta Raurica) ........................................................... 232 Figure 117: Pie-charts visualizing the distribution range of black glass bracelets of Type A ..................................................... 245 Figure 118: Distribution map of departments in France with black glass bracelets of type A .................................................... 246 Figure 119: Pie-charts visualizing the distribution range of black glass bracelets of Type B ..................................................... 248 Figure 120: Bar-charts visualizing the distribution range of black glass bracelets of Type C .................................................... 248 Figure 121: Pie-charts visualizing the distribution range of black glass bracelets of Type D ..................................................... 251 Figure 122: Distribution map of late Roman dotted counters.................................................................................................... 266 Figure 123: The burial gifts from 1st century AD tombs at Antran (FR) (left) and Montebelluna (IT) (right) (pictures taken from Pautreau (ed.) 1999, fig.28 (left); Casagrande, Ceselin 2003, pl. X (right)) .............................................................................. 273 Figure 124: Overview of burial gifts of tomb 176 from the cemetery at Cutry (illustration taken from Liéger 1997, pl. 28) ......... 275 Figure 125: (left) Base fragment from Liberchies-Bon Villers (BE); (right) from Matagne-la-Petite (BE) showing marks of use (photographs by author, by courtesy of the Musée Archéologique, Nivelles and VIOE) ........................................................... 276 Figure 126: (middle) rod-formed kohl-tube; (others) free-blown kohl-tubes (pictures taken from Israeli 2003, 228-229, nos. 283286 except middle Schlick-Nolte 2002, 104, no.V68) ............................................................................................................... 278 Figure 127: Hair-style reconstruction proposed by Pilet 1980, (I) 95, fig.2; pl.107:404. ............................................................ 283 Figure 128: (left) ‗Trilobitenperle‘ in black glass from Augst; (right) ‗Trilobitenperle‘ in jet from Augst (both drawings taken from Riha 1990, Taf. 39:1345; 1347) ............................................................................................................................................... 286 Figure 129: Comparison of jug-shaped pendants and their real counterparts in vessel glass (drawings taken from Spaer 2001 and Harden et al. 1989) ........................................................................................................................................................... 289 Figure 131: The very elaborate discoid brooch with inscription from the Roman cemetery of Wancennes (BE) (picture taken by author, courtesy by Musée Archéologique, Namur) ................................................................................................................. 292 Figure 132: large pseudo-nicolo gemstones set in a gilded silver sheet military helmet from Berkasovo (SB) (picture taken from Mackensen 2009, pl. 25). ........................................................................................................................................................ 295 Figure 133: Examples of wall and vault mosaics with twisted glass rods inserted: a.-b.: natatio, villa dei Centroni, Rome; c.-d.: bathcomplex, Aquileia; e.-g.: nymphaeum, Casa del Torello, Pompeii (illustrations taken from Sear 1977, pl. 15:2-3 [a-b]; pl. 44:3 [c]; pl. 43:4 [d]; pl. 9:1;3 [e-f]; fig.13 [g]) ............................................................................................................................ 298 Figure 134: plaster mirror from Egypt decorated with plano-convex counters in black glass (picture taken from Musée du Verre 1999, 51, no.11, courtesy by Musée du Verre, Marcinelle) ...................................................................................................... 300 Figure 135: Biplot opposing silica to soda oxide content (wt %) ............................................................................................... 310 Figure 136: Biplot opposing lime to silica oxide content (wt %) ................................................................................................ 310 Figure 137: Biplot opposing potash to magnesium oxide content (wt %) .................................................................................. 311 Figure 138: Biplot opposing alumina to silica oxide content (wt %) .......................................................................................... 312 Figure 139: Biplot opposing alumina to magnesium oxide content (wt %) ................................................................................ 312 Figure 140: Biplot opposing alumina to iron oxide content (wt %) ............................................................................................ 313 Figure 141: Biplot opposing titanium to iron oxide content (wt %) ............................................................................................ 314 Figure 142: Biplot opposing manganese to iron oxide content (wt %) ...................................................................................... 314 514 Figure 143: Biplot opposing titanium to manganese oxide content (wt %) ................................................................................ 315 Figure 144: Biplot opposing potash to soda oxide content (wt %) ............................................................................................. 315 Figure 145: a) preparing the samples in a plastic matrix before filling the plastic matrix with resin; b) making of a sketch of the samples ................................................................................................................................................................................... 318 Figure 146: The material used to make a resin for encasing the samples ................................................................................ 319 Figure 147: (left) The final stage of grinding and polishing the resin surface by using diamond paste of 0,25 µm.; (right) Resin with clean, smooth and scratchless surface ready for analysis measurements ......................................................................... 319 Figure 148: Vacuum carbon evaporation device (BALZERS UNION) used to apply a thin carbon coating ............................... 319 Figure 149: LA-ICP-MS installation at CNRS-IRAMAT, Orléans, France (photographs by author, courtesy of Bernard Gratuze)321 Figure 150: Ternary plot opposing Al2O3-Fe2O3-MnO (weight %) (taken from Vander Linden et al. 2009, fig.4) ........................ 325 Figure 151: Type of flux: magnesia opposing potassium (weight %) ........................................................................................ 326 Figure 152: Type of sand: Al2O3 versus TiO2 (weight %) for the resins PC1-3 .......................................................................... 326 Figure 153: Biplot visualizing alumina versus iron content (weight %) for the resins PC1-3 ...................................................... 327 Figure 154: Biplot visualizing manganese versus iron content (weight %) for the resins PC1-3 ................................................ 328 Figure 155: Biplot visualizing alumina versus titanium content (weight %) for the resins PC1-48.............................................. 329 Figure 156: Al2O3 versus Fe2O3 (weight %) for the resins PC1-48 ............................................................................................ 329 Figure 157: TiO2 versus Fe2O3 (weight %) for the resins PC1-48 ............................................................................................. 330 Figure 158: MnO versus Fe2O3 (weight %) for the resins PC1-48............................................................................................. 331 Figure 159: MgO versus Fe2O3 (weight %) for the resins PC1-48............................................................................................. 332 Figure 160: Plot opposing chromium with vanadium oxide (weight %)...................................................................................... 333 Figure 161: Plot opposing strontium with titanium oxide (weight %) ......................................................................................... 334 Figure 162: Plot opposing zirconium with titanium oxide (weight %) ......................................................................................... 334 Figure 163: Plot opposing chromium with vanadium oxide (weight %)...................................................................................... 335 Figure 164: Plot opposing manganese with antimony oxide (weight %) .................................................................................... 336 Figure 165: Plot opposing soda with potash oxide (weight %) .................................................................................................. 337 Figure 166: Plot opposing alumina with titanium oxide (weight %)............................................................................................ 337 Figure 167: Plot opposing manganese with iron oxide (weight %) ............................................................................................ 338 Figure 168: The installation of the p-XRF ................................................................................................................................. 344 Figure 169: Energy spectrum of R1610 compared with the spectra of black glass and obsidian pieces ................................... 345 Figure 170: Detailed view of the X-ray spectra of the aluminium (Al) and silicium (Si) peaks (taken from Cagno, Cosyns 2009a, unpublished internal report) ..................................................................................................................................................... 345 Figure 171: Detailed view on the X-ray spectra of the potash (K) and calcium (Ca) peaks (taken from Cagno, Cosyns 2009a, unpublished internal report) ..................................................................................................................................................... 346 Figure 172: Detailed view of the X-ray spectra of the iron (Fe) peaks (taken from Cagno, Cosyns 2009a, unpublished internal report) ...................................................................................................................................................................................... 346 Figure 173: Manganeseiron ratios (Mn/Fe) versus potashcalcium ratios (K/Ca) in the analysed samples (taken from Cagno, Cosyns 2009b, unpublished internal report) ............................................................................................................................. 349 Figure 174: Manganese intensities (Mn) versus iron intensities (Fe) in the analysed samples (normalised values) (taken from Cagno, Cosyns 2009b, unpublished internal report) ................................................................................................................. 350 Figure 175: Lead intensities (Pb) versus copper intensities (Cu) in the analysed samples (normalised values) (taken from Cagno, Cosyns 2009b, unpublished internal report) ................................................................................................................. 350 Figure 176: Example of a Raman spectrum of a sodium-silicate glass (taken from Baert et al. 2011, fig. 1) ............................. 352 Figure 177: The not normalised Raman spectra from 4 black glass samples with high iron content ......................................... 353 Figure 178: Raman spectra of Grobbendonk (Group A = high Fe–low Mn); Olbia_3 (Group B = low Fe–high Mn); Avenches_4 (Group C = low Fe–low Mn) (taken from Baert et al. 2011, fig.4) [a.u. = arbitrary units] ............................................................ 353 Figure 179: Raman spectra of Roman black glass artefacts (Appendix 3) with a various concentrations of iron oxide (taken from Baert et al. 2011, fig. 5) [norm. = normalised] ........................................................................................................................... 354 Figure 180: schematic set-up of the optical installation for UV-Vis-NIR spectroscopy ............................................................... 364 Figure 181: schematic view of all possible influences on the colour of an artefact before reaching the naked eye.................... 364 Figure 182: (a) A selection of black-appearing glass vessels with only reflective light; (b) the Rumst piece with neon (whitish) light (18 Watt) in transmission (b1) versus a strong halogen (yellowish) light (500 Watt) in transmission (b2); (c) a Matagne piece 515 with neon (whitish) light (18 Watt) in transmission (c1) versus a strong halogen (yellowish) light (500 Watt) in transmission (c2) (photographs by the author, courtesy of VIOE) ........................................................................................................................ 365 Figure 183: Schematic views of the spectrum measurement using light in transmission without (top) or with (bottom) the use of an integrating sphere ............................................................................................................................................................... 366 Figure 184: Schematic view of the measuring of transmission (T) of glass artefacts ................................................................ 366 Figure 185: Solar spectrum within the visible (VIS) [between 380-780 nm], from ultraviolet (UV) [below 380 nm] to nearly infrared (NIR) [above 780 nm], including the Fraunhofer lines featuring the absorption lines matching the wavelength of specific chemical elements (see http://en.wikipedia.org/wiki/Fraunhofer_lines) ................................................................................................... 367 Figure 186: (top left) CIE1931 horseshoe curve; (top right) Colour matching functions showing normalized typical human cone cell responses (Z, Y and X) to monochromatic spectral stimuli (taken from Fortner, Meyer 1997); (bottom) the spectra of the three primary colour stimuli (taken from DeCustatis 1998) ....................................................................................................... 368 Figure 187: The spherical colour spaces with vertically the lightness of the colour (L*) and horizontally the hue in two axes, green-red (a*) and yellow-blue (b*) (a-c: provided by Wendy Meulebroeck; d: taken from http://en.wikipedia.org/wiki/Munsell_Color_System) ................................................................................................................ 369 Figure 188: transmission spectra of Roman strongly-coloured glass appearing black from Rumst and Florennes between 350 and 800 nm ............................................................................................................................................................................. 371 Figure 189: Plot of the Rumst and Florennes samples on the CIE1931 horseshoe graph ........................................................ 372 Figure 190: The different transmissions of light according to the particular hue of glass .......................................................... 373 Figure 191: The transmissions of light converted into optical density (OD) .............................................................................. 373 Figure 192: integrated optical density of black-appearing glass versus non-black appearing coloured glass ........................... 374 Figure 193: Overview of the integrated optical density of the 22 black-appearing artefacts within the UV-VIS-IR (300-1600 nm)375 Figure 194: Iron content versus manganese content (%wt) of the first set of 48 artefacts, showing the clear compositional difference between artefacts produced at different periods in time. .......................................................................................... 377 Figure 195: Transmission spectra of Spectral Group 1 ............................................................................................................ 379 Figure 196: Transmission spectra of Spectral Group 2 samples from Avenches, Matagne-la-Petite and Velzeke .................... 380 Figure 197: Transmission spectra of Spectral Group 2 of samples from Braga, Oudenburg and Tienen .................................. 380 Figure 198: Transmission spectra of Spectral Group 3 ............................................................................................................ 381 Figure 199: Transmission spectra of Spectral Group 4 ............................................................................................................ 382 Figure 200: Calculated colour coordinates of the first four spectral groups .............................................................................. 383 Figure 201: Transmission spectra of Spectral Group 5 ............................................................................................................ 383 Figure 202: Transmission spectra of Spectral Group 6 ............................................................................................................ 384 Figure 203: plot of the calculated colour coordinates of the two types of artefacts belonging to Spectral Group 6 .................... 384 Figure 204: Transmission spectra of Spectral Group 7 ............................................................................................................ 385 Figure 205: Similar spectral shapes corresponding with different chemical compositions. The separate groups which are defined in the iron-manganese chart all correspond with a typical transmission spectrum .................................................................... 386 Figure 206: Plot of the various spectral groups on the CIE1931 horseshoe graph ................................................................... 387 Figure 207: Overview of the examined artefacts from a wide geographical provenance and produced over a long-lasting period389 Figure 208: Transmission maxima (Tmax) of Spectral Group 1 as a function of the x colour coordinate. The higher both values, the more reduced the furnace conditions ................................................................................................................................. 390 Figure 209: Position of the transmission maxima in the samples of Group 2 ............................................................................ 390 Figure 210: Measured transmission spectra of the artefacts belonging to Spectral Group 6 [= Class III]; left: subtype I; right: subtype II. ............................................................................................................................................................................... 391 Figure 211: Roman glass vessel from Tongeren/Belgium (photograph by the author, by courtesy of PGRM, Tongeren) (left). Transmission spectrum of the object (right) ............................................................................................................................. 393 Figure 212: Colour coordinates of different samples excavated at Avenches (top) .................................................................. 393 Figure 213: Spectra of the samples belonging to Spectral Group 6 and 7a [Class III] .............................................................. 394 Figure 214: Grouping of the Tienen artefacts based on their colour coordinates and position of the transmission ................... 395 Figure 215: Maximum (left) and on external examination (right)............................................................................................... 396 Figure 216: Grouping of the Kontich artefacts based on their colour coordinates and position of the transmission maximum .. 397 Figure 217: Mapping of a number of samples on the CIE1931 colour diagram ........................................................................ 398 Figure 218: Calculated colour coordinates of the different spectral groups recognized on Roman black glass ......................... 399 Figure 219: Plot on the CIE1931 horseshoe graph of the calculated colour coordinates of the different spectral groups recognized on Roman black glass ........................................................................................................................................... 399 516 Figure 220: Schematic visualization of the three-levelled economic approaches ...................................................................... 405 Figure 221: Distribution map of vessel Forms IIB.1 and 4/5;glass workshops where black glass was processed ..................... 406 Figure 222: Bi-plot discriminating two subclasses according to the titanium content (wt %) ...................................................... 407 Figure 223: Bi-plot of the titanium – manganese ratio (wt %) of the vessel shapes Form IIB.1 aand Forms IIB.4/5 .................. 408 Figure 224: Bi-plot of the titanium – manganese ratio (wt %) of the vessel shapes Form IIB.1 aand Forms IIB.4/5 .................. 408 Figure 225: Distribution map of the low titanium subclass (green) and high titanium subclass (red) ......................................... 409 Figure 226: Key models of the Roman economy ...................................................................................................................... 411 Figure 227: Theoretical visualisation of the possible mixtures that can occur when cullet is added to the batch ....................... 416 Figure 228: Access 2007 catalogue Form ................................................................................................................................ 430 517 518 LIST OF TABLES Table 1: Country subdivision codes ............................................................................................................................................xxi Table 2: Overview of the different colours of Roman black glass .................................................................................................. 7 Table 3: Categorization of decorated artefacts (n = quantity) ....................................................................................................... 8 Table 4: List of colours and colour combinations (n = quantity) .................................................................................................. 10 Table 5: Overview of colour combinations between core and decoration of bichrome artefacts .................................................. 11 Table 6: Overview of colour combinations between core and decoration of polychrome artefacts .............................................. 11 Table 7: The ratio of different techniques utilized in the production of black glass vessels ......................................................... 20 Table 8: Differences between the so-called ‗Trilobitenperlen‘ in jet and black glass ................................................................... 30 Table 9: Overview of various types of black glass production material ....................................................................................... 35 Table 10: Overview of various adopted techniques in black glass vessel production .................................................................. 35 Table 11: Overview of various adopted glass hues in the production of black appearing vessels ............................................... 36 Table 12: Overview of various adopted glass hues in the production of black appearing cast vessels ........................................ 36 Table 13: Overview of various adopted glass hues in the production of black appearing free-blown vessels .............................. 37 Table 14: Overview of various adopted glass hues in the production of black appearing mould-blown vessels .......................... 37 Table 15: Overview of various adopted glass hues in the production of black appearing rod-formed vessels ............................. 38 Table 16: Quantitative overview of the various jewellery types for each technique ..................................................................... 39 Table 17: Quantitative overview of the various glass hues applied for arm rings ........................................................................ 40 Table 18: Quantitative overview of the various glass hues applied for finger rings ...................................................................... 41 Table 19: Quantitative overview of the various glass hues applied for beads ............................................................................. 42 Table 20: Quantitative overview of the various glass hues applied for pendants ........................................................................ 42 Table 21: Quantitative overview of the various glass hues applied for gems .............................................................................. 43 Table 22: Quantitative overview of the various glass hues applied for hairpins ........................................................................... 44 Table 23: List of ovoid beakers, Form IIB.13 .............................................................................................................................. 62 Table 24: Overview of features used to categorize the rod-formed vessels ................................................................................ 71 Table 25: Types of decoration of the rod-formed vessel shapes known in black glass................................................................ 71 Table 26: List of rod-formed squat jars (Form IVD.2) .................................................................................................................. 74 Table 27: List of rod-formed conical and biconical jars (Forms IVD.3-4) ..................................................................................... 77 Table 28: chronological overview of the different techniques of black glass vessel production ................................................... 78 Table 29: Frequency list of the different shapes in black glass recorded per period and per technique (n = quantity) ................. 79 Table 30: Concordance list of the proposed typologies on Roman glass bracelets. [* our taxonomy is a final elaboration of our previously proposed classification system (Cosyns 2003; 2004) which is based on the typology proposed by Hubert De Witte (1977)] ....................................................................................................................................................................................... 86 Table 31: List with amounts of Roman glass bracelets per type and colour (n = quantity) .......................................................... 91 Table 32: List of different classes of glass finger rings per colour with % in black glass (based on the catalogue of Henkel 1913, 154-158) .................................................................................................................................................................................... 93 Table 33: Concordance table of the established typologies ........................................................................................................ 94 Table 34: List of finger rings in black glass (n = quantity) ........................................................................................................... 96 Table 35: List of type A finger rings in black glass ...................................................................................................................... 96 Table 36: List of type B finger rings in black glass ...................................................................................................................... 99 Table 37: List of type C finger rings in black glass decorated with protuberances (dimensions in mm) ..................................... 103 Table 38: Overview of the rod-formed beads in black glass ..................................................................................................... 108 Table 39: Overview of annular beads in black glass (n = quantity) ........................................................................................... 109 Table 40: Overview of globular beads in black glass (n = quantity) .......................................................................................... 110 Table 41: Overview of cylindrical beads in black glass (n = quantity)........................................................................................ 110 Table 42: Overview of melon beads in black glass (n = quantity) ............................................................................................. 111 Table 43: Overview of barrel-shaped beads in black glass (n = quantity) ................................................................................. 111 519 Table 44: Overview of the miscellaneous shapes of beads rarely occurring in black glass (n = quantity) ................................. 112 Table 45: Dimensions (in mm) of pendants with bulbous body and gooseneck suspension loop ............................................. 117 Table 46: Ratio of various coloured pressed discoid pendants in large collections (YG = yellowish green; YB = yellowish brown; P = purple; G = green; B-G = blue-green; B = blue) ................................................................................................................. 119 Table 47: Number of gemstones from Belgium related to finger rings (taken from Sas 1993, 126) .......................................... 123 Table 48: List of various glass colours of twisted rods from R1609 at the KMKG-MRAH, Brussels .......................................... 128 Table 49: List of so-called stirring rods (dimensions in mm) ..................................................................................................... 130 Table 50: List of black glass counters per size and shape (* see note 81) ............................................................................... 135 Table 51: List of monochrome black glass counters plain vs dotted decoration (* see note 81)................................................ 135 Table 52: The dimensions of two sets of decorated gaming counters ...................................................................................... 136 Table 53: detailed timetable of the Roman imperial period....................................................................................................... 137 Table 54: Number of shapes per functional types per period ................................................................................................... 140 Table 55: Number of shapes per functional type per period and per technique (A= cast; B= free-blown; C= mould-blown; D= rodformed).................................................................................................................................................................................... 140 Table 56: List of entries per applied technique per period ........................................................................................................ 140 Table 57: Chronological overview of all vessel types produced in black-appearing glass [see Chapter 3] ................................ 141 Table 58: List of type A bangles from dated finds .................................................................................................................... 148 Table 59: List of type B bracelets from dated contexts ............................................................................................................. 149 Table 60: List of type C-bracelets from dated contexts ............................................................................................................ 150 Table 61: List of type D bracelets from dated contexts............................................................................................................. 152 Table 62: Chronological overview of the black glass arm rings ................................................................................................ 154 Table 63: Overview of the ratios per finger ring shape and subtype ......................................................................................... 155 Table 64: List of finger rings from dated finds .......................................................................................................................... 156 Table 65: List of sites with black glass pendants from dated finds ........................................................................................... 159 Table 66: Chronological bar chart of the types of discoid pressed pendants (based on Barag 2001) ....................................... 160 Table 67: Overview of the decorative categories per type of pressed discoid pendant (Barag 2001) ....................................... 160 Table 68: General typo-chronology of the engraving styles applied on intaglios proposed by Kathy Sas (Sas 1993, 119-122) 161 Table 69: List of sites with black glass tesserae....................................................................................................................... 163 Table 70: Timetable of various architectural decoration materials in Period I (according to Sear 1977) ................................... 163 Table 71: List of sites with black glass twisted rods ................................................................................................................. 165 Table 72: Quantification of the counters per size for each period ............................................................................................. 168 Table 73: Quantification of the counters per size for each period ............................................................................................. 168 Table 74: Quantification of the counters per used technique for each period ........................................................................... 168 Table 75: Quantification of the monochrome and polychrome counters per period .................................................................. 169 Table 76: Quantification of the counters per glass hue for each period .................................................................................... 169 Table 77: Overview of number of sites and number of items per context (* omitting the artefacts from unknown provenance and the additionally recorded material – see introduction) .............................................................................................................. 177 Table 78: Overview of black glass artefacts per context (* incorporates the items related to secondary glass workshops) ....... 179 Table 79: Overview of the various commodity classes per context type ................................................................................... 180 Table 80: Overview of ratios of the different types of jewellery per period from civil settlement contexts (n = quantity) ........... 181 Table 81: Various types of arm rings in relation to the contexts of provenance ........................................................................ 184 Table 82: Vessels from domestic contexts ............................................................................................................................... 187 Table 83: Arm rings from domestic contexts ............................................................................................................................ 188 Table 84: Arm rings from the black glass manufacturing workshops versus domestic consumers‘ contexts ............................. 189 Table 85: Beads from domestic contexts ................................................................................................................................. 190 Table 86: Gemstones from domestic contexts ......................................................................................................................... 190 Table 87: Overview of architectural decoration material in domestic contexts .......................................................................... 191 Table 88: Tombs with black glass vessels from Period I ......................................................................................................... 193 520 Table 89: Tombs with black glass vessels from Period II ......................................................................................................... 193 Table 90: Tombs with Form IIIB.2 vessels (―-― = undefined) ..................................................................................................... 195 Table 91: Overview of ratios of the different types of black glass jewellery per period from civil/funerary contexts (n = quantity)195 Table 92: Black glass bangles from inhumation burials (n = quantity; C = civil settlement; M = military settlement) .................. 197 Table 93: Roman cemeteries containing black glass bracelets ................................................................................................. 197 Table 94: Black glass beads used as burial gifts ...................................................................................................................... 199 Table 95: Black glass counters used as burial gifts .................................................................................................................. 202 Table 96: Overview of the different types of black glass jewellery from civil/ritual contexts ....................................................... 204 Table 97: Black glass bangles from ritual context ..................................................................................................................... 205 Table 98: Black glass counters from ritual contexts .................................................................................................................. 205 Table 99: Black glass artefacts from military contexts (additional entries between brackets) .................................................... 206 Table 100: Overview of the contextual data of black glass ....................................................................................................... 207 Table 101: Overview of ratios of black glass vessels per period from military settlement contexts ............................................ 211 Table 102: Overview of ratios of the different types of jewellery per c amp type ................................................................... 211 Table 103: Bracelets from military sites .................................................................................................................................... 212 Table 104: Gemstones from military sites ................................................................................................................................ 214 Table 105: Counters from military sites .................................................................................................................................... 215 Table 106: Roman black glass artefacts from caves (n = quantity) ........................................................................................... 216 Table 107: Roman black glass artefacts from rivers ................................................................................................................. 218 Table 108: List of glass workshop sites with possible black glass production ........................................................................... 223 Table 109: List of black glass products from sites with glass workshops .................................................................................. 224 Table 110: List of secondary glass workshops processing black glass (between [ ] and in italic are uncertain production sites)228 Table 111: Overview of various concentrations of black glass commodities from sites with a clear or supposedly secondary glass workshop where black glass has been manufactured ..................................................................................................... 229 Table 112: List of attested glass hues from the workshop at Regio 17C (left) and Regio 17B (right) (taken from Fischer 2009, 95, fig.108)..................................................................................................................................................................................... 231 Table 113: List of glass finds from the late 4th century glass workshop at Trier, Palais Kesselstatt .......................................... 233 Table 114: List of sites with black glass vessels of Period I ...................................................................................................... 235 Table 115: List of sites with black glass vessels of Period II ..................................................................................................... 237 Table 116: List of sites with black glass vessels of Period IV.................................................................................................... 240 Table 117: Overview of the number of black glass bangles per country and by Roman area .................................................... 242 Table 118: Overview of the Type A black glass bangles per country and per area ................................................................... 243 Table 119: Detailed list of Type A bangle subtypes per vast geographical section ................................................................... 244 Table 120: Overview of the Type B black glass bangles per country and per area ................................................................... 246 Table 121: Detailed list of Type B bangle subtypes per vast geographical section ................................................................... 247 Table 122: Overview of the Type C black glass bangles per country and per area ................................................................... 249 Table 123: Detailed list of Type C bangle subtypes per vast geographical section ................................................................... 249 Table 124: Overview of the Type D black glass bangles per country and per area ................................................................... 250 Table 125: Detailed list of Type D bangle subtypes per vast geographical section ................................................................... 250 Table 126: List of sites with black glass finger rings ................................................................................................................. 252 Table 127: List of sites with rod-formed beads in black appearing glass (n = quantity) ............................................................. 254 Table 128: List of sites with double-perforated pressed beads in black appearing glass........................................................... 256 Table 129: List of various types of black glass pendants per region ......................................................................................... 259 Table 130: List of plain conical-shaped gems in monochrome black appearing glass set in disc brooches ............................... 260 Table 131: List of monochrome black glass counters (n = quantity).......................................................................................... 264 Table 132: List of bichrome and polychrome dotted black glass counters ................................................................................ 265 Table 133: Overview of the distribution of black glass vessels during the consecutive periods ................................................. 267 Table 134: List of cast shapes per function (numbers coincide with Form IA shapes discussed in Chapter 3) .......................... 274 521 Table 135: List of proposed functional connotations and practical use of jewellery (based on Vanhaeren 2005) ...................... 278 Table 136: List of different values per functional type in black glass (*= positive; -= negative; ?= uncertain) ............................ 279 Table 137: List of tombs with (black) glass bracelets (n = quantity).......................................................................................... 281 Table 138: List of glass gemstones set in finger rings (n = quantity) ........................................................................................ 292 Table 139: Areas where glass tesserae have been used on the mosaic of Mariamin nearby Hama (SY) ................................. 296 Table 140: List of sites with prior analysed black glass artefacts (* = LA-ICP-MS; ** = SEM-EDX; *** = AAS) ......................... 307 Table 141: Chemical composition of black appearing .............................................................................................................. 308 Table 142: Overview of samples per assorted topics [see Appendix 1] (n= number of samples) .............................................. 317 Table 143: List of sampled Roman black glass material according to provenance (between square brackets, from previous research by others (Morreti, Gratuze 2002)) ............................................................................................................................ 323 Table 144: List of sampled Roman black glass material according to type and date range per generalized period .................. 323 Table 145: List of sampled Roman black glass material according to type and colour of the glass matrix ................................ 323 Table 146: Chronological subdivision of the sampled material as proposed in Van der Linden et al. 2009 .............................. 324 Table 147: Different glass types on the basis of the strontium-titanium ratios .......................................................................... 334 Table 148: Major elements of the Roman glass lumps and crucible glass obtained from SEM-EDX (n.d. = not detected; italic = not appearing black). ............................................................................................................................................................... 339 Table 149: Major elements of early Byzantine mosaic material (n.d. = not detected) ............................................................... 340 Table 150: Values of magnesium, potash and strontium in %wt .............................................................................................. 340 Table 151: Values of used flux soda, magnesium, potash and strontium in %wt ...................................................................... 341 Table 152: Values of used decolourizers manganese and antimony in %wt ............................................................................ 341 Table 153: List of additional artefacts analysed with µ-XRF ..................................................................................................... 347 Table 154: Qualitative data (net intensities) of samples analysed by µ-XRF (data taken from Cagno, Cosyns 2009b, unpublished internal report) ......................................................................................................................................................................... 347 Table 155: Semi-quantitative data (elemental ratios) of the samples measured by µ-XRF (data taken from Cagno, Cosyns 2009b, unpublished internal report) ......................................................................................................................................... 348 Table 156: Table summarizing the possible colorants used on the analysed Roman black glass artefacts, and the relationship between the applied furnace conditions and the resulting colour (taken from Brill 1988, 272-273, Tables 9-10)....................... 363 Table 157: overview of wavelength intervals for the different colours within the visible light spectrum ..................................... 367 Table 158: The calculated tristimulus values in the three colour spaces .................................................................................. 369 Table 159: x and y-coordinates for plot on CIE1931 horse-shoe graph (* hue perceived by the naked-eye) ............................ 371 Table 160: List of integrated optical density for 31 measured black appearing glass samples from 21 objects......................... 374 Table 161: Overview of the examined artefacts, their colour coordinates and concentration in iron oxide, manganese oxide, copper oxide and cobalt oxide ................................................................................................................................................. 376 Table 162: list of the four groups of black glass based on the colouring by means of iron and manganese content ................. 377 Table 163: concordance list of the different chemical classes and spectral groups .................................................................. 385 Table 164: Overview of the examined samples........................................................................................................................ 388 Table 165: Furnace conditions and colour coordinates of the brown coloured artefacts ........................................................... 391 Table 166: List of the four different colour groups from Avenches ............................................................................................ 393 Table 167: Timetable of the various glasses used to produce black glass artefacts ................................................................. 415 Table 168: List of potential factors influencing technological change as proposed by Tiziano Mannoni 2007 ........................... 419 522

Log In

Roman black glass

Roman black glass

Related Papers

RELATED PAPERS