The Glass, in: P.Ellis and P. Leach, The Roman Settlement at Fosse Lane, Shepton Mallet: The tesco excavation, 1996-7. pp. W14-17 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY VOL. 155

Birgitta Hoffmann

This report discusses a small globular jug and an unguent bottle found in a feature interpreted as a collection of pyre debris at Dinham Road, Exeter. It dates to the period AD 65-75 and would have been associated with legionary activity there. The fragments of the jug were found placed in a Dr. 18 bowl. The vessels show deliberate smashing prior to deposition with only a selection of the pieces included in the pit. A similar phenomenon can be seen in the pottery

This article represents a summary of recent research into the production and status of Roman glass, with special consideration of decorative Roman glass that has been found in Britain.

This illustrated guide is the first comprehensive classification of vessel glass found in England between 1500 and 1670. The book provides a clear and accessible coverage of vessel forms and individual types - with an indication of their date range, provenance and a description. Each form is illustrated with an excavated example and there is a general summary of the types with reconstruction drawings at the end. Paperback 230 pages (September 2002) Publisher: Council for British Archaeology ISBN: 190277129X

FOSSE LANE 1996: SPECIALIST REPORTS THE ROMAN SETTLEMENT AT FOSSE LANE, SHEPTON MALLET: THE TESCO EXCAVATION, 1996–7 Peter Ellis and Peter Leach with contributions by Umberto Albarella, Lynne Bevan, Brenda Dickinson, Andy Hammon, Kay Hartley, Birgitte Hoffmann, Julie Jones, Joanna Mills, Stephen Minnitt, Donald Mackreth, Stephanie Pinter-Bellows, Fiona Roe, and Roger Tomlin SOMERSET ARCHAEOLOGY AND NATURAL HISTORY VOL. 155 © Somerset Archaeological and Natural History Society and the authors, Somerset Heritage Centre, Brunel Way, Norton Fitzwarren, Taunton TA2 6SF, 2011 1 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 Contents Flint Lynne Bevan ................................................................................................................................. The coins Stephen Minnitt ..................................................................................................................... Roman small finds Lynne Bevan .......................................................................................................... Copper alloy objects ................................................................................................................. Silver object Roger Tomlin ......................................................................................................... Iron objects .................................................................................................................................. Lead objects ................................................................................................................................. Worked bone ................................................................................................................................ Shale .............................................................................................................................................. Wall plaster .................................................................................................................................... The brooches Donald Mackreth .............................................................................................................. The glass Birgitta Hoffmann ................................................................................................................. The Roman pottery Annette Hancocks ............................................................................................... Mortarium Kay Hartley .............................................................................................................. Samian: decorated wares Joanna Mills ....................................................................................... Potters’ stamps Brenda Dickinson ............................................................................................. Graffiti Roger Tomlin ................................................................................................................... The worked stone Fiona Roe ............................................................................................................... The charred plant remains Julie Jones .................................................................................................. The animal bone Umberto Albarella and Andy Hammon ...................................................................... The human skeletons Stephanie Pinter-Bellows .................................................................................... W3 W3 W8 W8 W10 W10 W11 W12 W12 W12 W12 W14 W18 W24 W25 W30 W31 W33 W35 W37 W44 List of figures Fig. 29 Fig. 30 Fig. 31 Fig. 32 Fig. 33 Fig. 34 Fig. 35 Fig. 36 Fig. 37 Fig. 38 Fig. 39 Copper alloy and shale objects; scale as shown ................................................................. Vessel glass; scale as shown ............................................................................................ Romano-British pottery; scale as shown .............................................................................. Decorated samian; scale as shown .................................................................................. Samian stamps; scale 2:1 ................................................................................................ Animal bone: relative importance of the main domesticates ............................................ Animal bone: relative abundance of skeletal elements ...................................................... Animal bone: post-cranial proportions of butchery ............................................................. Animal bone: cattle survivorship curve (from mandibles with 2+ teeth) for earlier Roman periods .................................................................................................................. Animal bone: inter-site comparisons of cattle survivorship curves (from mandibles with 2+ teeth) for earlier Roman periods ......................................................................... Animal bone: inter-site comparison of cattle M3 for earlier Roman periods .................. W9 W15 W25 W27 W31 W41 W42 W42 W43 W43 W44 List of tables Table 4 Table 5 Table 6 Table 7 Table 8 Coin list ................................................................................................................................. Romano-British pottery: fabric, sources and quantities in assemblage ............................. Romano-British pottery: summary of fabrics by period ................................................... Charred plant macrofossils occurrence ............................................................................ Animal bone: numbers of animal bone and teeth ISP including hand and sieve collection) .......................................................................................................................... Animal bone numbers (NISP) and % of the three major domesticates .......................... Animal bone: comparison of cattle and sheep measurements from selected RomanoBritish sites ........................................................................................................................ Human bone: condition and degree of completeness of skeletons ................................... Human bone: demography for Shepton Mallet ............................................................... Human bone: stature, means and ranges for Fosse Lane ................................................. W4 W20 W22 W38 References ........................................................................................................................................... W48 Table 9 Table 10 Table 11 Table 12 Table 13 W2 W40 W40 W44 W45 W45 W46 FOSSE LANE 1996: SPECIALIST REPORTS SPECIALIST REPORTS FLINT Lynne Bevan THE COINS Stephen Minnitt Thirty eight items of humanly worked flint were recovered comprising three cores, two retouched blades, eight scrapers, and 25 flakes; none are illustrated. The raw material was of a generally good quality with the brown, compacted cortex characteristic of flint from secondary deposits, such as river gravels. It was beige and light to medium grey in colour with a high incidence of white recortication. Pebble flint is often of unpredictable quality, and a number of hinge fractures, common when working poor quality flint, were apparent among the flakes and cores. The three cores in the collection had been worked beyond the point of apparent usefulness which is another indication that good-quality flint was at a premium. The core, blade, scraper and flint find spots gave no indication of any activity focus and were evenly spread between areas A and B. While none of the flints were chronologically diagnostic, a generally later prehistoric date during the later Neolithic to early Bronze Age seems most probable, based upon the broad, squat shape of most of the flakes. In contrast to the material found to the south (Bevan 2001a, 105), there was no clearly Mesolithic flintwork. The presence of eight scrapers in the collection is suggestive of habitation foci within the vicinity of the site, but the generally low incidence of flint tools, cores and flakes is not indicative of occupation of any longevity. Instead, this small collection appears to represent a low density and episodic usage of the landscape throughout later prehistory, supporting the conclusions drawn from the 1990 collection (ibid, 106). A total of 298 coins were found of which 296 were Roman (Table 4). Forty-seven derive from excavated contexts, the remainder are unstratified. Coins marked * were identified by Simon Esmonde-Cleary. In view of the detailed analysis of coins found during earlier work at Fosse Lane (Esmonde-Cleary 2001, 211) little comment will be made on the 1996 assemblage. Suffice it to say that the general pattern of coins found in 1996 is closely comparable to the earlier finds and continues the trend of a predominance of coins of 4thcentury date. The occurrence of the Republican denarius should not be considered as particularly significant as such pieces were current into the 3rd century. Unless stated otherwise in the context column the coins are unstratified. The assemblage includes an abnormally high proportion of unidentifiable coins. In spite of careful and individual cleaning of the coins by Mark Davis, Keeper-Conservation, Somerset County Museums Service, surface corrosion (patina), and therefore detail, failed to adhere to the coins. Whilst in part this may be due to soil conditions a significant factor was probably the post-excavation treatment of the coins. Each was individually wrapped in tissue, perhaps while it was still damp, and placed in a polythene bag shortly after excavation. This and a delay of two years prior to cleaning may well have had an adverse effect on the stability of the surface corrosion. W3 3 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 TABLE 4: COIN LIST Reign Reverse Date D. Silanus, denarius Victory in biga 91BC Probably copy of Claudian as, though of unusual style Vespasian, dup. illegible 69-79 Hadrian, sest. Pont Max Tr Pot Cos SC 117 Fort Red Hadrian, sest. Pont Max Tr Pot Cos III SC119-122 Hadrian?, sest. 117-138? Antoninus Pius, dup. Libertas 138-161 Faustina I, sest. Aeternitas 141+ Faustina I, sest. Aeternitas 141+ Marcus Aurelius, sest. illegible 161-180 Lucilla, denarius Iuoni Lucinae 164-169 C1-C2 as or dup. C1-C2 as or dup. Gallienus Gallienus Gallienus Gallienus Gallienus Claudius II Claudius I I Claudius II, divo Tacitus Victorinus Victorinus Victorinus Victorinus Victorinus Tetricus I Tetricus I Tetricus I Tetricus I Tetricus I Tetricus I Tetricus I Tetricus I Tetricus II Tetricus II Tetricus II Radiate Radiate Radiate Radiate Radiate Radiate Radiate Radiate Radiate Claudius II, copy Claudius II, copy Claudius II, copy Claudius II, copy Victorinus copy Victorinus copy Victorinus copy W4 illegible illegible Conservat Pietat Dianae Cons Aug Iovi Propugnatori Aetern Aug Pax Aug illegible illegible Consecratio, altar Salus Aug Aequitas Aug Pietas Aug Pax Aug illegible illegible Fides Militum Hilaritas Augg Laetitia Augg Pietas Augg Salus Aug Spes Aug Salus? illegible Salus Aug Pax Aug illegible Comes Aug Aequitas Consecratio, altar Consecratio, altar Consecratio, altar Consecratio, altar Invictus Pax illegible 260-268 260-268 260-268 260-268 260-268 268-270 268-270 c.270 265-266 268-270 268-270 268-270 268-270 268-270 270-274 270-273 270-274 270-273 270-273 270-274 270-274 270-274 270-274 270-274 270-273 260-280 260-280 260-280 260-280 260-280 260-280 260-280 260-280 260-280 270-290 270-290 270-290 270-290 270-290 270-290 270-290 Context Ref S.F. No. Crawford 337/3 19 512 43 192 F 101 RIC 541a RIC 561a as RIC 950 3010 structure 6 RIC 770 SW room 2062 2041 2006, F224 RIC 171a as RIC 176 RIC 214 RIC 465a RIC 575 as RIC 259 RIC 57 RIC 40 as RIC 58 as RIC 118 as RIC 79 RIC 87/88 RIC 108 as RIC 121 as RIC 130 2098 RIC 266 RIC 247/248 2009 2034 copy as RIC 112 157 182 382 89 231 365 513 366 368 438 325* 179 591 413 102 234 261 311* 169 271 266 226 270 354 87 454 80 273 240* 582 424* 391 351 475 28 185 65 215 235 430 431 448 580 170 284* 285 295* 10 176 376 FOSSE LANE 1996: SPECIALIST REPORTS Reign Reverse Date Context Ref S.F. No. Tetricus I copy Tetricus I copy Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Barbarous radiate Carausius Carausius Maximian, follis but obv legend: MAXIMIANVS NOB C Licinius Constantine I Constantine I Constantine I Constantine I Crispus Crispus Constantine II Constantine I Constantine II Constantine II Constantine II House of Constantine House of Constantine Urbs Roma Urbs Roma Urbs Roma Urbs Roma Urbs Roma copy Urbs Roma copy Constantinopolis Constantinopolis Constantinopolis Constantinopolis Constaninopolis Constantinopolis Constantinopolis copy Constantinopolis copy Constantinopolis copy Constantinopolis copy Constantinopolis copy Constantine II Constantine II Constans Constantius II Constantius II Laetitia Pax Pax Victory Salus Salus Pax Pietas? 270-290 270-273 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 270-290 287-293 287-293 305-307 2020 copy as RIC 86 287* 61 105 243* 254* 370 318* 573 18 37 183 256* 265* 337 472 546 590 228 455 299* Provid Aug illegible Genio Populi Romani Genio Pop Rom 310-312 Victoriae Laetae Princ Perp 319 Victoriae Laetae Princ Perp 319 Beata Tranquillitas 321 Beata Tranquillitas 321 Beata Tranquillitas 321 Beata Tranquillitas 321-323 Beata Tranqlitas 323-324 Providentiae Caes 324-326 Gloria Exercitus 2 stds 330-335 Gloria Exercitus 2 stds 330-337 Gloria Exercitus 2 stds 330-335 Gloria Exercitus 2 stds 330-335 Gloria Exercitus 2 stds 330-335 Wolf and twins 330-335 Wolf and twins 330-335 wolf and twins 330-335 Wolf and twins 330-335 Wolf and twins 330-348 Wolf and twins 330-348 Victory on prow 330-335 Victory on prow 330-335 Victory on prow 330-335 Victory on prow 330-335 Victory on prow 330-335 Victory on prow 330-335 Victory on prow 330-348 Victory on prow 330-348 Victory on prow 330-348 Victory on prow 330-348 Victory on prow 330-348 Gloria Exercitus 1 std 335-337 Gloria Exercitus 1 std 335-337 Gloria Exercitus 1 std 337-341 Gloria Exercitus 1 std 335-337 Gloria Exercitus 1 std 337-341 2006, F224 2005, F223 2068 2004, F213 2009, F224 RIC 356 2029 2020 2060 2058 2063 2005, F223 2020, F223 1036, F122 2074 2060 RIC VI Trier 642a RIC VI London,209c RIC VII London, 209 RIC VII Trier, 213 RIC VII Trier, 303 RIC VII Trier, 305 RIC VII Trier, 308 180 200 46 162 289* 160 76 RIC VII London, 287 199 138 as LRBC I, 49 248 as LRBC I, 49 131 149 167 as LRBC I, 48 327* LRBC I, 51 433 LRBC I, 51 577 as LRBC I, 51 25 as LRBC I, 51 322 copy as LRBC I, 51 314* copy as LRBC I, 51 198 LRBC I, 71 216 LRBC I, 86 320* 142 LRBC I, 185 377 LRBCI, 191 410 LRBC I,201 541 copy as LRBC I, 52 252* copy as LRBC I, 52 286* copy as LRBC I, 52 401* copy as LRBC I, 52 568 copy as LRBC I, 52 165 LRBC I, 93 392 383* LRBC I, 131 321* LRBC I, 94 117 LRBC I, 252 439 W5 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 Reign Theodora Theodora Helena House of Constantine House of Constantine House of Constantine House of Constantine House of Constantine Constans, copy H. of Constantine copy H of Constantine copy H. of Constantine copy H of Constantine copy 515 SW room Constantius II Constantius II Constantius II Constans Constans Constans Constans Constans Constans House of Constantine House of Constantine House of Constantine House of Constantine House of Constantine House of Constantine Constans Reverse Pietas Romana Pietas Romana Pax Publica Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Gloria Exercitus 1 std Date 337-341 337-341 337-341 335-341 335-341 335-341 335-341 335-341 335-348 335-348 335-348 335-348 335-348 Context 3014 Ref as LRBC I, 105 as LRBC I,120 LRBC I, 112 as LRBC I, 87 as LRBC I, 139 As LRBC I, 107 as LRBC I, 100 copy as LRBC I, 87 2001 S.F. No. 272 55 171 511 94 219 594 589 59 437 114 474 3010 structure 6 Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Victoriaeddauggqnn Fel Temp Reparatio, phoenix House of Constantine Fel Temp Reparatio, phoenix Magnentius/Decentius fragment, Felicitas Reipublice Magnentius Gloria Romanorum Magnentius Victoriae dd aug et cae Magnentius? Constantius II Fel Temp Reparatio (FH) Constantius II Fel Temp Reparatio (FH) Falling horseman copy Falling horseman copy Falling horseman copy Falling horseman copy Falling horseman copy Falling horseman copy Falling horseman copy Falling horseman copy Falling horseman copy Falling horseman copy 350 350 351-352 350-353 353-355 350-355 350-360 350-360 350-360 350-360 350-360 350-360 350-360 350-360 350-360 350-360 Falling horseman copy 350-360 3013 str 6 SE room Falling horseman copy Falling horseman copy overstruck on Gloria Exercitus two standards Falling horseman copy overstruck on Gloria Exercitus Valentinian I Gloria Romanorum 350-360 350-360 copy as LRBC II, 25 534 copy as LRBC II, 25 66 350-360 364-378 2098 W6 347-348 347-348 347-348 347-348 347-348 347-348 347-348 347-348 347-348 347-348 347-8 347-348 347-348 347-348 347-348 346-350 LRBC I, 455 LRBC I, 140a as LRBC I, 140 as LRBC I, 148 as LRBC I, 158 LRBC I, 267 as LRBC I, 145 as LRBC I, 158 346-350 447 51 567 257 473 153 196 73 348 52 178 54 545a 372 77 576 214 as Bastien 21 as Bastien 33 RIC 189 2004, F213 2004 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 copy as LRBC II, 25 363 574 50 14 432 205 246* 305* 236* 38 47 109 134 136 173 206 copy as LRBC II, 25 516 copy as LRBC II, 25 423* as LRBC II, 78 350 FOSSE LANE 1996: SPECIALIST REPORTS Reign Valentinian I Reverse Gloria Romanorum Date 364-378 Valentinian I Valentinian I Valentinian I Valentinian I Valentinian I Valentinian I Valens Valens Valens Valens Valens Valens Valens Valens Valens Valens Valens Valens Gratian Gratian Gratian Gratian Gratian Gratian House of Valentinian House of Valentinian House of Valentinian House of Valentinian House of Valentinian House of Valentinian House of Valentinian House of Valentinian House of Valentinian House of Valentinian House of Valentnian House of Valentinian House of Valentinian House of Valentinian? Arcadius Arcadius Arcadius Arcadius House of Theodosius House of Theodosius House of Theodosius House of Theodosius House of Theodosius House of Theodosius House of Theodosius House of Theodosius Gloria Romanorum Gloria Romanorum Gloria Romanorum Gloria Romanorum Securitas Reipublicae Securitas Reipublicae Gloria Romanorum Gloria Romanorum Gloria Romanorum Gloria Romanorum Securitas Reipublicae Securitas Reipublicae Securitas Reipublicae Securitas Reipublicae Securitas Reipublicae Securitas Reipublicae Securitas Reipublicae illegible Gloria Romanorum Gloria Romanorum? Gloria Novi Saeculi Gloria Novi Saeculi Gloria Novi Saeculi Gloria Novi Saeculi Gloria Romanorum Gloria Romanorum Gloria Romanorum Gloria Romanorum Securitas Reipublicae Concordia Auggg Gloria Romanorum Securitas Reipublicae Gloria Romanorum Securitas Reipublicae Securitas Reipublicae Securitas Reipublicae 364-378 364-375 364-378 367-375 364-378 364-378 364-378 364-378 364-378 364-378 364-378 364-378 364-378 364-378 364-367 367-375 367-378 364-378 367-375 364-378 367-375 367-375 367-375 367-375 364-378 364-378 364-378 364-378 364-378 378-383 364-378 364-378 364-378 364-378 364-378 364-378 364-378 364-378? 388-402 388-402 388-402 388-402 388-402 388-402 388-402 388-402 388-402 388-402 388-402 388-402 Victoria Auggg Victoria Auggg illegible illegible Salus Reipublicae Salus Reipublicae Victoria Auggg Victoria Auggg Victoria Auggg Victoria Auggg Victoria Auggg illegible Context 2062 2095 Ref as LRBC II, 94 S.F. No. 49 as LRBC II, 279 as LRBC II, 279 LRBC II, 296 LRBC II, 321 as LRBC II, 96 as LRBC II, 96 as LRBC II, 92 as LRBC II, 282 as LRBC II, 478 64 326* 75 429 145 212 565 417* 222 249 56 152 349 359 127 207 132 545 133 369 223 224 300* 450 113 137 213 237* 147 307* 592 323* 168 361 371* 453 218 188 158 303* 440 510 154 244* 251* 62 63 393 533 575 as LRBC II, 97 as LRBC II, 97 as LRBC II, 97 as LRBC II, 97 LRBC II, 277 LRBC II, 303 LRBC II, 528 LRBC II, 339 2029 2001 2004 2029 2062 2024 2022 as LRBC II, 503 as LTBC II,503 as LRBC II, 503 as LRBC II, 517 as LRBC II, 78 as LRBC II, 78 as LRBC II, 92 as LRBC II, 92 as LRBC II, 96 as LRBC II, 269 as LRBC II, 275 as LRBC II, 276 as LRBC II, 279 as LRBC II, 280 as LRBC II, 280 as LRBC II, 481 as LRBC II, 164 LRBC II, 392 3013 2005, F223 2005, F223 as LRBC II, 796 as LRBC II, 1105 as LRBC II, 389 as LRBC II,389 Twenty one illegible 3rd to 4th-century coins were recorded: SF nos, 78, 217, 352, 353, 355, 362, 364, 367, 428, ?281, 443, 452, 456, 514 (found in layer 3010 structure 6 SW room), 531, 532, 542, 543, 544, 569 and 593 Thirty three illegible 4th-century coins were recorded: SF nos 40, 58, 85, 90, 139, 177, 181, 191, 193, 232, 279, 317* (from layer 2031),426, 427, 441, 444, 445, 446, 449, 470, 517 (layer 3013, Structure 6, SE room), 535, 536, 537, 538, 539, 540, 566, 570, 571, 572, 579 and 581 W7 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 Six illegible 3rd to 4th-century copy coins were recorded: SF nos 82, 97, 209, 269, 414, 418 Three illegible 4th-century copy coins were recorded: SF nos 34, 57 and 135 Twenty one 4th-century copies were recorded: SF nos 116, 123, 156, 161, 184, 186, 190, 204, 208, 210, 220, 221, 230, 247, 316* (from layer 2031), 360, 373, 415* (from F104), 457, 471 and 583 Finally two post-Roman coins were recorded: A French jetton, late C14-early C15; Mitchiner 489; SF 578; and a C19-C20 halfpenny, SF 547 ROMAN SMALL FINDS Lynne Bevan with a contribution by Roger Tomlin Introduction The small finds collection is described below by material rather than by function, given the relatively small number of objects. Analysis of the assemblage as a whole allows some trends and patterns to be identified. Relatively few fixtures and fittings were present, although a fragment of shale inlay from furniture or a tray (Fig. 29.9), may indicate that luxury goods were either present or being made on or in the vicinity of the site. The wall plaster suggests the existence of at least one well-decorated room in a relatively sophisticated building. The small tool assemblage – three chisels, a wedge, two ox goads and a bucket handle – is much as one would expect from an agricultural cum industrial community. The three styli together perhaps suggest the record keeping associated with the corralling and selling of cattle on site, as has been mooted to have occurred at sites such as Barnsley Park, Hambledon, and Rocester, Staffordshire (Ferris and Cooper 1996, 149). The relatively large number of items of jewellery – particularly the ten pins and 16 bracelets – may be of some particular significance as has been suggested for the larger but similarly biased assemblages from Uley, Lydney Park and Great Witcombe (Bevan 1998, 86, 88–9), while the inscribed silver ring strikes a poignant chord even today. The lead ossuary is a relatively rare find on a Romano-British site, and its presence adds another example of an unusual burial rite being practised in Roman Shepton. Catalogue of copper alloy objects Seven items, cat nos 1, 8, 9, 10, 25, 26, 27 are illustrated (Fig. 29) 1 W8 Pin, broken. The globular head was covered with close-set radiating diagonally incised grooves and the neck was delineated by approximately ten horizontal bands. This is similar to a pin head from a post-Roman posthole at Lion Walk, Colchester (Crummy 1983, fig. 31.499, 30–31). SF 166, 110/255, east of Structure 12. Fig. 29.1. 2 Pin, complete, with a small spherical head, a common type of pin which was used throughout the Roman period (Cool 1990, fig. 1.5, 151–2). SF 30, F131. 3–7 Pin fragments: 3, SF 357, 60/305, Structure 6; 4, SF 411, 130/340, Structure 5; 5, SF 115, 75/ 295, Compound 7; 6 SF 91, 80/265, Compound 8; 7 SF 95, 80/280, Compound 8. 8 Bracelet fragment, D-shaped section, broken at both ends, with intricate decoration in the form of a thick raised area of deep grooves and horizontally hatched vertical bands at the front of the armlet which narrows at the shoulder where it is decorated with a vertically hatched horizontal band. SF 381, 125/335, Structure 9. Fig. 29.2. 9 Bracelet fragments, D-shaped section with decoration consisting of diagonal grooves which give way to vertical grooves at one broken end. SF 378. Fig. 29.3. 10 Bracelet of double-stranded twisted wire with a D-shaped section broken at both ends. SF 39, 110/375, east of Structure 2. Fig. 29.4. 11 Bracelet with three ring and dot motifs at one end and an incised line along each edge. SF 290, layer 2026, Structure 8. 12 Bracelet fragment, rectangular-sectioned with decoration in the form of a continuous wave motif enclosed by bands, broken at both ends. SF 159, 100/290, Compound 8. 13 Bracelet fragment, flat with a cut-out crenellated edge, and a perforated terminal at the unbroken end. SF 194, 125/260, east of ditch F236. 14 Bracelet fragment, flat with a cut-out crenellated edge. SF 163, 110/255, Compound 8. 15 ?Wire bracelet. Hooked fragment with squareshaped section and traces of incised diagonal striped decoration on upper face. SF 128, 75/ 300, Structure 17. 16 ?Bracelet fragment. Strip with traces of diagonal and linear decoration. SF 586, unprovenanced. 17 ?Bracelet fragment, curved strip. SF 332, pit F253. FOSSE LANE 1996: SPECIALIST REPORTS Fig. 29 Copper alloy (1-7) and shale (8-10) objects 18 19 20 Bracelet fragment, traces of incised designs on outside face. SF 93, 60/405, west of Structure 1. Finger ring, broken curved fragment widening at shoulder. SF 412, 130/340, Structure 9. Ear-ring pendant. Small segment of wire formed into a rod with chainlink loops at either end, probably a pendant from an ear-ring which once held a glass bead. A pair of similarly sized chainlink ear-ring pendants with dark blue beads in situ was found in a 4th-century context in Colchester (Crummy 1983, fig. 53.1797). SF 3, 70/390, road F101. W9 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 W 10 ?Wire bracelet. Two strands of twisted wire. SF 425, layer 2098, Structure 9. Hooked object, broken, with oval area decorated with circular motif. SF 155, 100/255, east of Structure 12. Bracelet or necklace hook. SF 523, Compound 7. Tweezers with flared blades, one broken. SF 141, 100/270, Structure 11. Horse harness. Long, hexagonal stud with a raised oval boss with deep median groove. Two disc-headed shanks protrude from the reverse. This type of stud, which was probably used as horse harness, has been found on many Roman sites (Allason-Jones and Miket 1984, 3.870, 3.871, 237). SF 104, 55/310, Structure 6. Fig. 29.5. Plate, diamond shaped with raised, split panels at each end and a flat central area with four decorative pierced holes. The reverse has two studs for the attachment of the plate to a strap. SF 110, Structure 6. Fig. 29.6. Mount, shield-shaped, leaded copper alloy, decorated with three deep grooves. SF 529; unprovenanced. Fig. 29.7. Stud, circular rosette shape, crudely decorated with a series of irregular incised lines around the outer circumference, and traces of iron from a central attachment, now corroded. A number of similar studs recovered from 1st-century contexts at Colchester were used as military belt and apron fittings (Crummy 1983, fig. 151:4204, 4205). SF 211, 135/295, east of F236. Small ring, leaded copper alloy, D-shaped section, probably from clothing or horse harness. SF 118, Structure 6. Ring fragment, square section. unprovenanced. Ring fragment, D-shaped section. SF 4, 65/385, south of road F101. Stud, oval, possibly from a small box or item of furniture. SF 16, 60/400, west of Structure 1. Strip, rectangular with uneven ends and three incised lines, rectangular-sectioned. SF 526, unprovenanced. Rod, broken, square-sectioned. SF 336, pit F253. ?Vessel fragment. Plate with grooved slightly raised edge. SF 239, ditch F223. ?Box or furniture fitting. Circular pierced terminal. SF 313, ditch F245. Terminal, rounded end decorated with four lines of incised dots, leaded copper alloy. SF 587, unprovenanced. Fitting. Tapering curved object, leaded copper alloy. SF 521, beneath Structure 7. Ferrule divided into three bead-shaped segments by two grooves. SF 588, unprovenanced. Plate fragment, rectangular, broken, one straight end, serrated edges. SF 298, pit F225. 41–4 Wire fragments. 41 SF 585, unprovenanced; 42 SF 144, 105/280, east of Structure 11; 43 SF 335, pit F253; 44 SF 524, pit F335. 45–51 Rod and strip fragments. 45 SF 201, 115/280, east of Structure 12; 46 SF 262, 140/290, east of F236; 47 and 48 SF 522, F335; 49 SF 26, F133; 50 SF 291, ditch F223; 51 SF 324, unprovenanced. 52–8 Sheet fragments. 52 SF 29, 85/375, Structure 2; 53 SF 125, 95/260, Structure 12; 54 SF 53, 120/ 365, Structure 4; 55 SF 274, 145/290, east of F236; 56 SF 344, layer 1028; 57 SF 305, bank F234; 58 70/390, road F101. Silver ring Roger Tomlin Fragment of a silver ring (not illustrated) , broken at the shoulder with an intact bezel, originally c 22mm in diameter. An inscription which reads ‘ME l MOR’, probably memor, ‘mindful’, is enclosed within the round bezel which measures 8mm in diameter. There is an incised roundel below the legend, which corresponds to the two roundels on each shoulder of the bezel. Memor is a cognomen, and even a cult-title of Minerva at a shrine in North Italy (cf ILS 2603, a votive brought back from Britain), but here it is probably an adjective understood by the donor of the ring, presumably, and by its wearer: either (sis) memor (mei), ‘Remember me’, or (sum) memor (tui), ‘I remember you’. The latter is perhaps to be preferred, in view of the rings inscribed MEMINI TVI (ClL xiii 10024. 71(b) and (c), ‘I remember you’), and MEMINI TVI MEMINI ET AMO (ibid, 72, cf 73, ‘I remember you; I remember and I love (you)’). SF 296; layer 2029, floor Structure 7. Iron objects (none illustrated) 1 2 3 4 5 6 7 8 Part of a buckle. SF 111, Area B. Pin shank. SF 345, grave F208. Fragment from a stylus with a broad-based eraser, broken across shank and very corroded. SF 341, grave F150. Fragment from a stylus with a broad-based eraser, broken across shank and very corroded. SF 108, 80/270, north of Structure 8. Stylus with a rounded eraser, very corroded. Unprovenenaced. Chisel, similar to a possible Iron Age example from Hod Hill, Dorset (Manning 1985, plate 11:B43, 24). 75/265, Structure 8. Chisel or punch, similar in size and shape to a mid 1st-century example from Hod Hill, Dorset (Manning 1985, Plate 5:A23, 10). SF 151, 110/ 270, east of Structure 11. Chisel or punch, similar in size and shape to another mid 1st-century example from Hod Hill, FOSSE LANE 1996: SPECIALIST REPORTS 9 10 11 Dorset (Manning 1985, Plate 5:A25, 10). SF 143, 105/255, east of Structure 12. ?Wedge. 120/365, Structure 4. Bucket handle terminal consisting of a looped circular-sectioned length of iron rod, similar in style to an Iron Age or mid-first century example from Hod Hill in Dorset (Manning 1985, Plate 47:P14, 103). SF 419, 80/375, west of Structure 2. Conical object attached to a circular disc of copper alloy. SF 276, ditch F230. A total of six shoe cleats were found: SF 315, 2025, Structure 7; F133; south of F101; Ditch F134; Structure 2 and unprovenanced. Thirty hobnails were recovered. A total of 15 came from grave F150 with the remainder as surface finds, ten of which came from the areas east of Structure 12, and one each from Structures 1, 2, and 4, road F101 and east of F236 in Area B. Two ox goads were found as surface finds in the area of Structures 4 and 10. A number of miscellaneous items were also found comprising a length of iron rod with a looped end, a hook, two staples, three rings, seven fragments of rod and wire, 21 fragments of plate and over 300 nails. Lead objects (none illustrated) The total weight of the lead objects was just under 4kg. 1 Lead ossuary, circular drum-shaped container, originally equipped with a domed top and central spout which survives as a complete base and a number of small fragments from the sides, top and spout which were severely degraded beyond reconstruction. This small, undecorated ossuary with its integral spout is an example of a ‘pipe burial’, a Roman burial custom known from Italy and the Rhineland and elsewhere (for discussion see Philpott 1991, 28), involving the pouring of libations down a channel protruding above ground level into a coffin or ossuary. The channel was made from various materials, which had usually been re-used or adapted, such as two vertical imbrices leading into a tile cist at St Pancras, Chichester (Down and Rule 1971, grave 323), and the re-used amphora spouts from Osola Sacra, Ostia which would also have been visible as grave markers (Calza and Becatti 1977, 69). At Falerone in Italy, a vertical lead pipe was attached to a small stone sarcophagus containing a cremation-burial and various grave goods (Wheeler 1929, 4). In Britain, a pipe burial involving a lead pipe is known from Creffield Road, Colchester (Wheeler 1929, 4), and a lead pipe which probably fulfilled a similar function was found with a cremation at Mancetter, Warwickshire (Booth 1982, 134–6). The closest parallel is from Caerleon (Wheeler 1929), where a slightly larger drum-shaped ‘canister’ decorated with three reel-patterned bands and with a lead pipe attached to the centre of its lid was found in a stone-lined cist. The pipe had originally protruded beyond the cist and above the former land surface. The burial of the canister, which contained the cremated remains of a mature male, was Hadrianic or later. Philpott has suggested that pipe burials resulted from ‘a strong desire... (that) the dead should have direct contact with the living and that perhaps such libations were seen to nourish the dead in a particularly vivid way’ (Philpott 1991, 28). The act of libation might also have been viewed as a placatory or appeasing gesture towards the spirits of the deceased, perhaps combined with a desire for guidance or the bestowal of good fortune from the other world. The distribution of ossuaries was concentrated in the main military and urban centres, and the occurrence of ossuaries on relatively rural sites such as this example is rare and ‘may indicate an adoption of the ossuarium by high-status romanised natives or the burial of immigrants on their estates’ (Philpott 1991, 28). The large quantities of jewellery recovered from the site would certainly support the presence of high-status individuals, the majority of whom were probably members of the local, Romanised population. Seen in this context, the adoption of a specifically Roman burial practice is less surprising. Base diam 190mm, th 3mm, wt of fragments 1188gm. SF 380, F242 (Fig. 10). 2 Caulking. T-shaped fragment of caulking made from a thick, rounded core of lead fused with folded sheet. 60/285, Structure 14. 3 ?Weight. Roughly conical object. SF 394, unprovenanced. 4 Pipe collar. Collar with furled edge, diam 105mm. SF 33, unprovenanced. 5–6 Circular objects, one with central perforation, both unprovenanced. 7 Rounded fragment with flattened base. unprovenanced. A total of 10 pot cramps were found, five still attached to fragments of pottery, two Black Burnished ware vessels, two sandy reduced coarseware vessels, a Samian vessel and a greyware vessel. The located items came from east of F236 (5), Structure 9, east of Structure 7, east of Structure 11, and east of Structure 12. A total of 13 lead offcuts were found as surface finds, with provenanced items coming from Compound 6 (2) Structures 4, 6, 11 and 17, road F101, and east of F236. Numerous small waste globules from lead working were found. W 11 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 Worked bone (not illustrated) 1 2 Pointed pin fragment. SF 253, F224. Pointed pin fragment. SF 306, F243. Shale There is no evidence from this or the 1990 excavation (Leach 2001) for the on site manufacture of shale items. It is, however, possible that no 1 (Fig. 29.9) may have been manufactured or modified on site. The source of the shale is likely to have been Kimmeridge from which shale was exploited at a number of contemporary sites in Purbeck (Sunter and Woodward 1987). The standard of preservation of the items was generally good apart from some delamination at the edges of no 1, which might have been damaged and subsequently discarded during the manufacturing process. 1 2 3 W 12 Flat fragment of shale. Two motifs are visible on its upper surface; a single spiral and an ‘S’shape, both enclosing incised dots. A fragment has been cut from the upper surface The spiralform decoration is reminiscent of motifs on later Iron Age pottery, in contrast to the more formalised and repetitive decoration seen upon published examples of trays from other Roman sites such as Colchester (Crummy 1983, fig. 75:2022, 2023, 69 and 71), Silchester (Lawson 1975, figs 11.87–9, 12.90), and similar fragments described as ‘plaques’ and ‘tablets’ from Norden near Corfe Castle, Dorset (Thomas 1987, figs 19 and 20.18–22, 30–35). However, similar motifs combined with more regular panels of intricate geometric banding have been identified on an almost complete tray or ‘trencher’ found in a cremation burial from Grange Road, Winchester which had vessels, a spoon and knives, and two pork joints on it (Biddle 1967, fig. 6.233). Although Thomas has suggested that such items might also have been use for other purposes including ‘wall plaques’ and ‘small table tops’ (Thomas 1987, 33), that the Winchester object was used as a tray appears beyond doubt. The dating evidence suggests a general 1st to 2nd-century date for these items (Biddle 1967, 248–59; Thomas 1987, 35), which might reflect a fashion in tableware and serving etiquette. The cut-out section on the decorated face of this fragment might indicate that shaleworking was carried out on the site. SF 6, 85/ 405, Structure 1. Fig. 29.9. Spindlewhorl, elliptical section. Fig. 29.8. Two joining oval-section, lathe-turned fragments from a large bracelet. The item has three grooves on the exterior and a slightly raised bevel on the 4 inside, which might have been designed to hold the bracelet to the upper arm, perhaps over clothing. The armlet is an exceptionally large piece with an internal diameter of 120mm; similarly sized armlets have been recorded at Norden (Thomas 1987, 30) where internal raised bevels were also recorded. SF 27 and SF 31; F133. Fig. 29.10. (not illustrated) Three joining fragments from a child’s bracelet with a D-shaped section. F160. Wall plaster Four fragments of wall plaster were recovered, all of which had retained traces of pigment. A similarly small group was noted in 1990 (Morgan 2001, 230). The pigments ranged in colour from dark red to light pink. As no traces of painted decoration were visible, the fragments appear to have originated from plain painted borders rather than from decorated panels. None are illustrated. 1 2 3 4 Fragment with red/brown paint. SF 518, F224. Fragment with bright pink paint. SF 563, F225. Fragment with traces of pale pink paint. SF 24, 60/310, Structures 6 and 17. Fragment with traces of pink paint. SF 84, 80/ 260, Structure 8. THE BROOCHES Donald Mackreth Where a brooch in the present assemblage is also represented in the 1990 report, the reader is referred to the discussion there, the conclusion only being repeated here (Mackreth 2001). All the brooches, except nos 22 and 23 (of iron), are made from copper alloy; none are illustrated. Colchester Derivatives Brooches 1–3 have or had their springs mounted in the Polden Hill manner: the axis bar through the spring is mounted in a pierced plate at the end of each wing, the chord being held either by a rearward-facing hook or passing through a pierced crest on the head of the bow. 1 Each wing has a relieved beaded ridge separated by a wide flute. The bow has a beaded pseudo hook dying away in the V formed by two more beaded ridges. On each side of the head of the bow is a plain additional moulding. The foot has a slight projection and the catch-plate a small triangular piercing. SF 41, surface find area B. Of the same family as Brooches 7 and 8 in the first report, the conclusion there was that the date-range of the Polden Hill version of the FOSSE LANE 1996: SPECIALIST REPORTS 2 3 general family ran from c AD 65 to about 125, and that those dating to perhaps as late as AD 125 may well have been survivors in use. The head is missing. In the panel on the upper bow, which is outlined by a groove, is the trace of at least one cell, and perhaps of three. Beneath the panel is a slight waist, then two lenticular bosses above a median arris. The bow tapers to a narrow projecting foot. The catch-plate is solid. SF 281, layer 2017. The family to which this brooch belongs has both the Polden Hill spring system, as this example would have had, as well as hinged pins. Discussed under Brooches 12–15 in the first report, the date-range for those with Polden Hill spring systems is mainly later 1st to mid 2nd century. The wings are plain. The bow has a very short additional moulding on each side, and a relieved ridge down the upper part stopping at a marked cross-flute with a raised border top and bottom. The rest of the bow is plain and ends in a small projecting foot. The catch-plate is solid. SF 258, F223. Although not a member of a specific group, the additional mouldings and the ornament in the middle of the bow show affinities with the brooches which have passed in review and a date-range of c AD 75–125/50 is likely. top but rapidly tapers to a thin foot marked by two projections. On the top of the bow is a single short ridge dying away into the V formed by a pair of others meeting in a point below. The catch-plate is solid. SF 330, pit F254 Very like Brooches 25–6 in the first report, this example probably dates from the later 1st century to AD 150/75. Unclassified 7 Trumpet 8 Brooches 4–7 have hinged pins. 4 5 6 In two parts and badly corroded, the surviving wing is plain and only the upper bow survives. This has a flat back, a moulding apparently on each side of the head and a broad one down the middle. SF 130, 105/255, east of Structure 12 There is little to go on and a general date running from the later 1st century to about AD 175 may be suggested. Each wing has a sunken moulding at its end. On the head of the bow is a tab with a circular depression. The bow has a step down each side of the upper part and a projecting foot. Between the lower end of the stepped part are two lozenges each with a central cell for enamel, now missing, and a bordering groove. On each side between the lozenges is a small ‘gablet’. The catch-plate is solid. SF 301, layer 2022 Of the same family as nos 44–8 in the first report (Mackreth 2001, 191), more specifically nos 46–7, the dating arrived at there has not improved since the writing of that report: later 1st to AD 200 by which time any in use were survivors as manufacturing had ceased as near as can be assessed for all Colchester Derivatives by c. AD 175. Each wing has two reels. The bow has a wide There is a ridge running across the head of the bow and the plain wings. The bow is very broad at the top tapering to a wide foot which projects boldly under a step. The profile is distorted, but once had an almost flat area above the main face which has a deep groove on each side and a small central ridge at the very top. SF 81, 55/305, Structure 6 In very broad terms, the overall proportions are like those of Brooches 34–8 of the first report (Mackreth 2001, 189–90), but none there is obviously of the same decorative type. However, the conclusion reached is almost certainly applicable here: there are no grounds for a 1stcentury date, nor any evidence that any should be expected in the 3rd. The pin is hinged, the axis bar being housed in a narrow projection behind the bottom of the trumpet head. There are the remains of a caston loop on a tall pedestal having two grooves. The trumpet head is very broad and shallow at the top, with a deep sweep on each side to a width less than that of the lower bow. The knop is of the usual petalled form with a single ridge top and bottom and all moulded on the front of the bow. The lower bow is broad, has a median arris and tapers slightly to the triple moulded foot. SF 140, 100/275, Structure 11 Although few appear to come from the same moulds, there is a family likeness which, when coupled with the distribution, which lies in the counties from Hampshire to Somerset, points to a distinct variety. However, none is dated and only a general range lying in the 2nd century can be suggested, there is little in the design to suggest that it could have started in the 1st century. Strip 9 The head of the bow is rolled under to house the axis bar of the hinged pin. There is effectively no head-plate, but one is defined roughly by the projecting arcs housing the eyes which are made up of a dot and circle. The lower bow with the W13 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 catch-plate is missing. The bow as it survives has a broad flute down each side and a groove down the middle across which lies a line of horizontal lines. SF 409, F262 An almost exact parallel occurs in the first report, Brooch 82, except that has a rolled over head and was better related to the Alesia – Hod Hill sequence (Mackreth 2001, 197). Here the associations, when married to the patently non-continental way of conveying the impression of there being a bead-row, are entirely with the Strip and, as such, is an example of that type deriving from members of the pre-Aucissa sequence. The determining feature is the presence of eyes. No Aucissa proper has these and it is clear that the general family had abandoned them well before the conquest. This brooch may have survived in use as late as that, but hardly beyond. Fragments 10 11 12 13 14 15 16 17 18 19 20 21 22 23 The lower bow and catch-plate of a brooch of unknown type. Probably before AD 250. SF 530, unprovenanced Complete pin and bilateral spring of five coils probably once mounted between a pair of pierced lugs. Second century. SF 175, 125/255, east of F236 Hinged pin. SF 20, unprovenanced Hinged pin. SF 525, unprovenanced The end of a pin possibly from a brooch. SF 268, 60/280, area of Structure 14 Hinged pin. SF 238, 130/275, east of F236 Hinged pin. SF 245, F224 The lower bow and catch-plate of a brooch of unknown type. Probably before AD 250. SF 1, 60/395, road F101 Spring fragment. SF 242, 130/275, east of F263 Penannular brooch with out-turned terminal end. SF 22, 95/375, east of Structure 2 Spring fragment. SF 332, F253 Hinged pin. SF 416, 130/340, Structure 9 Iron strip brooch with hinged pin. SF 70, 35/ 315, Compound 6 Hinged pin and strip from iron brooch. SF 375, F258 THE GLASS Birgitta Hoffmann A total of 117 Roman glass fragments was found. The fragments can be divided by type as follows: cast: 1; monochrome: 4; colourless: 11l greenish: 28; blue-green: 33; bottle: 19; window glass: 12; and objects: 9. The material derives mainly from the 1st and 2nd centuries and the late Roman period, with diagnostic finds from the later 2nd and early 3rd centuries W 14 (cylindrical beakers etc) being absent. None of the material is of particularly high quality. Pillar-moulded bowls, like no 1 (Fig. 30.1), are an extremely common find on 1st-century settlement sites. They can be found, for example, amongst the finds from the military works depot at Longthorpe (Dannell and Wild 1987, 51) and at Kingsholm (Cool and Price 1985, 45 nos 4–9), both of which were occupied in the (Claudio-) Neronian period, as well as in most southern British towns eg Silchester (Price 1984, 117, nos 1 and 2). At Vindolanda, however, which starts in c AD 85, none were found amongst the nearly 4000 glass fragments from the vicus (Hoffmann forthcoming). Three further examples of deep-blue pillar-moulded bowls are known from the 1990 Shepton Mallett excavations (Price and Cottam 2001, nos 1–3a). Two parallels are known for no 3 (Fig. 30.3), a dark brown fragment from Dorchester, Greyhound Yard (Cool and Price 1993, fig. 84.23) and a blue/green fragment from Usk (Price 1995, fig. 44.68). In both cases the original shape of the vessel can not be reconstructed, but the strong colour might point to a 1st-century date. The high percentage of brown or strong yellow/brown glass in the assemblages from both the 1996/7 (3 fragments) and 1990 sites (5 fragments) is remarkable, and appears to outweigh any other strong colour from the site. Number 4 (Fig. 30.4) comes from a globular flask with wheel-cut lines. This type of bottle is particularly common in the later 3rd and 4th centuries. However, the narrow neck with its very late starting funnel seem to make the vessel more likely to belong to an earlier tradition (Cool and Price 1995, 149, fig. 9.3). Decorated flasks of this type are unusual in Britain, but examples are known from Libya (Price 1985, 78; 99, no 50, fig. 6.4), Kisselbach/Hunsrück (dated to the end of the 1st century (Hopstätter 1942)), KölnSeverinskloster (dated to c. AD 200) and Luxemburger Straße (Fremersdorf 1984, 42, no 107), and from Ladenburg (Hoffmann 1996, 162, kat no L99–100). Number 5 (Fig. 30.5) might be part of one of the 2nd-century drinking vessels with linear cut lines, although not enough survives to be certain. These vessels are known in a number of different shapes and can have either pad bases or pushed-in base-rings, with other base types occurring more occasionally (Cool and Price 1995, 79–82). Handles like nos 6 (Fig. 30.6), 7 and 8 and the brown rim no 2 (Fig. 30.2), are most likely to have come from long-necked conical (or less likely globular) jugs. These occur in a variety of colours, with blue/ green being by far the most common (Cool and Price 1995, 120–3) and range in date from the late Flavian period (Cool and Price 1995, 122) to about AD 150– 170 (Price 1980, 66, fig. 15.6 and 7; Price 1987, 204, fig. 3.20; Price and Cottam 1994, 225, fig. 104.9 and 10). The handles found represent probably three FOSSE LANE 1996: SPECIALIST REPORTS Fig. 30 Vessel glass (1-9), bangle (10) and beads (11-13) different vessels and another four were recovered from the 1990 site. Such a high number of vessels of this type should come as no surprise as they are a common find on sites of the later 1st and 2nd centuries, with examples known from Colchester, Verulamium, Gloucester (Cool and Price 1995, 123), Carlisle, 19 examples (Price 1990, 174; Cool and Price 1991, 166; Price and Cottam 2001, 240) and Vindolanda, at least 17 examples (Hoffmann forthcoming). W 15 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 Indented vessels, cups, bowls and flasks were in use from the 1st to the 4th century, although too little survives of no 9 to allow any statement as to the original shape of the vessel. The same holds true for no 11 (Fig. 30.7), whose inturned rim is most likely to have come from a globular or indented jar. These vessels were fairly common in the 1st and 2nd centuries in Britain (Cool and Price 1995, 112), although the diameter of this example, at c 100 mm, makes it one of the larger examples of the type. Number 10 belongs to a globular bath flask with ‘dolphin handles’, which is a frequent shape on sites from the 1st to the 3rd century (Cool and Price 1995, 156). Bottles, probably four-sided (although other shapes do occur), are represented by nos 15–18. Bottles are usually easily identifiable by their characteristic thickwalled fragments. They form a major percentage of any late 1st and 2nd-century assemblage and the 19 fragments found in 1996/7 are quite in keeping with the 30 found on the 1990 site. These vessels were mainly used for storage purposes and occur in many different sizes (ranging in capacity from c 0.2 to several litres). Number 18 probably comes from one of the smaller bottle types. Number 17 is a small fragment of one of the base marks that are typical of these vessels, in this case circles, for which a wide range of combinations are known. Such bottles date mainly from the mid 1st to the 2nd or early 3rd centuries although they can occasionally be found in 4th-century assemblages (Cool and Price 1995, 184). Fourth-century glass is represented by nos 12 and 13. These curved rims are typical of a number of late Roman conical and hemispherical vessels, although too little survives here to identify the original types. Similar rims can be found in most late Romano-British assemblages as at Silchester (Price 1984, 118, no 3). Thirteen conical beakers with this rim type were found on the 1990 site, and a further 30 specimens have been listed from elsewhere in the region, for example from Bath, Cirencester, Ilchester, Dorchester and the rural sites of Frocester Court and Catsgore (Price and Cottam 2001, nos 28–31; 33; 35–6; 29–41 and 42–45c). The other certainly late Roman type is the greenish vessel with fire-rounded rim, no. 14 (Fig. 30.8) and three other similar such rim fragments were found in 1990 (Price and Cottam 2001, nos 38–40). Similar vessels have been dated elsewhere to the second half of the 4th to the early 5th century, with the most famous examples coming from the glass hoard at Burgh Castle (Harden 1983, 82–3, nos 85–9, fig. 37). The large bangle fragment no 19 (Fig. 30.10) belongs to Kilbride-Jones’ Class I (Kilbride-Jones 1937–1938, 367–72; Stevenson 1956, 208, 218; Stevenson 1976). The applied foil over a usually bluegreen body is characteristic, as is the piece’s rather ‘heavy’ character. Close parallels have been found at Traprain Law, the Roman fort of Camelon (Kilbride- W 16 Jones 1937–8, fig. 1), the native promontory fort of Mains of Ethie, Angus (Wilson 1980, 121) and other sites in Southern Scotland. The southernmost piece known previously comes from the Roman fort at Vindolanda, however, where it was found in a probably residual early 3rd-century context (Hoffmann forthcoming), so its presence in Somerset is somewhat surprising. In view of the finds from Traprain Law, Kilbride-Jones dates this type to the late 1st century (Kilbride-Jones 1937–1938, 367–72). Beads of the same type as no 20 (Fig. 30.11) were dated by Guido (1978, 96) to the 3rd and 4th centuries, but more recent finds from Vindolanda (Hoffmann forthcoming) and Strageath (Price 1989, 197, 202, fig. 102,7) appear to belong to the 2nd century. Although never found in large quantities, examples can be found all over Britain and similar finds from the same region are known from Cirencester, Bradley Hill, Lufton and Mendip (Guido 1978). Bluegreen segmental beads nos 21–23 (Fig. 30.12, 13) appear to be particularly common on later Roman sites, eg Portchester and Lympne (Guido 1978, 201), with a 3rd-century example now known from Vindolanda (Hoffmann forthcoming). Other examples come from Colliton Park/Dorchester, Cadbury Castle, Hengistbury Head, Camerton and Ham Hill (Guido 1978, 200). Bluegreen annular beads like no 24 tend to be found on Roman rather than Iron Age sites but there is currently no more precise dating for the type. Other examples are listed by Guido under her group 6iib (Guido 1978, 66, 143–5). Number 25 is part of Guido’s Group 7v (Guido 1978, 172). This type is quite common in the region, with examples known from Dorchester, Somerleigh, Sydling, Cirencester, Silchester, Amesbury, Meare, Glastonbury and Ham Hill (Guido 1978, 152–5). There is also an earlier find from Shepton Mallet. Numbers 26–27 are plain glass gaming counters, which are a common find on Roman sites of all periods. They are usually associated with board games, although other uses, such as accounting and calculation have occasionally been cited (Price 1995, 129). Black appears to be the most common colour on most sites, usually far outweighing all other colours, including white. The window glass from the site reflects both broad types known from Britain. The bluegreen cast fragments (nos 28 and 29) represent the matt/glossy variety that can be found on Roman sites from the later 1st century, whilst the greenish, thin variety (nos 30–42), which is probably made by blowing a large cylinder and then cutting it open, only came into use in the later Roman period (Harden 1961). The latter material is likely to be associated with the stone buildings of Period 5. In summary, the glass recovered gave a reasonable overview of the glass used in everyday contexts in a FOSSE LANE 1996: SPECIALIST REPORTS settlement of the later 1st and first half of the 2nd century, with, from the evidence of the glass, reoccupation in the 4th century. The present assemblage fits well with the material recovered in 1990 and the only unusual fragment is the Type I glass bangle which, in view of its otherwise mainly southern Scottish distribution, has to be counted as an import. Catalogue of Roman glass 9 10 Cast glass 1 Rim fragment, bluegreen; one rib; no bubbles; no weathering. RD: too small to measure: dims: 21 x 40 Th: 4 EVE: 0.4. SF 174, 110/280, east of Structure 11. Fig. 30.1. 11 Blown glass 2 3 4 5 6 7 8 Brown Rim fragment, brown; rim folded out, up and in; no bubbles; no weathering. D: 30 mm PH: 6 Th: 3 EVE: 0.14. SF 636, F253. Fig. 30.2. Body fragment, dark brown; folded out tube out of body, definitely not base ring; no bubbles; no weathering. D: 120 H: 6 mm Th: 2. SF 505, layer 1019. Fig. 30.3. Also: 3a Dark brown body fragment, from globular vessel; small bubbles; no weathering. Dims: 25 x 47 Th: 2. SF 506, layer 1019. Greenish Nineteen fragments (some joining), greenish; neck, body and base of globular flask with two horizontal linear-cut lines, and two further ones on the lower body; flat base; bubbles; no weathering. PH(neck): 64 W(neck): 16 PH(body): 33 D(base): c.45 EVE: 0.6. SF 11, 75/405, Structure 1. Fig. 30.4. Body fragment with linear cut lines; greenish; ?cylindrical vessel; no bubbles; no weathering. Dims: 21 x 30 Th: 2 EVE: 0.2? SF 255, F224. Fig. 30.5. Bluegreen Handle fragment with central rib; bluegreen; elongated bubbles; no weathering. W(handle): 22–24 PH: 42 Th: 5 EVE: 0.14. SF 629, 105/ 385, east of Structure 2. Fig. 30.6. Lower handle attachment, bluegreen; handle with central rib, drawn out over body and pinched; elongated bubbles; no weathering. W(ext): 43 PH: 36 TH: 7 EVE: 0.14. SF 72, 50/ 275, east of Structure 7. Handle fragment and lower attachment, bluegreen, handle with central rib, drawn out and pinched over the body; elongated bubbles, greenish striations; no weathering. W(handle): 27–34 PH: 67 TH: 5 EVE: 0.17. SF 476, 125/ 255 east of F236. Perhaps from similar vessel: 8a One body fragment, bluegreen. one straight rib; no bubbles; no weathering. Dims: 18 x 24 Th: 1. SF 509, 75/285, north of Structure 8. Body fragment, bluegreen with indent; bluegreen glass with little bubbles; no weathering. Dims: 11 x 18 Th: 1 EVE: 0.14. SF 630, 110/255, east of Structure 12. Handle fragment and shoulder of globular bath flask; handle trails along shoulder, neck and loops from the rim back onto shoulder and back onto the rim; bluegreen; bubbles; no weathering. H(ext): 34 Th: 5 EVE: 0.17. SF 98, 95/255, Structure 12. Two joining rim fragments, bluegreen, rim rolled inwards, many bubbles and greenish striations, no weathering. D: 100 H(ext): 24 Th: 1. SF103, 85/290, Compound 7, and SF 489, 75/290, Compound 7. Fig.30.7. Late Roman glass 12 13 14 Rim fragment, bluegreen/colourless; curved rim, cracked off and left uneven; many small bubbles; no weathering. D: 80 H(ext): 18 Th: 1 EVE: 0.2– 0.4. SF 282, 125/270, east of F236. Rim fragment. Greenish; rim slightly turned out, and cracked off straight; small bubbles; no weathering. D: 80 PH: 12 Th: 2 EVE: 0.2. SF 508, 110/290, east of Structure 11. Rim fragment, greenish; rim fire-rounded; slightly outturned; body ?globular; many bubbles, no weathering. D: 85 H(ext): 21 Th: 1 EVE: 0.4. SF 68, 120/345, Structure 9. Fig. 30.8. Bottles 15 16 17 18 Rim fragment, colourless; small bubbles; matt surface; rim folded out, up and in and flattened on top. D: 45 H(ext): 7 EVE: 0.17? SF 636, F253. Rim fragment, bluegreen; rim folded out, up and in; many bubbles; no weathering. D: 52 H(ext): 21 Th: 5 EVE: 0.28. SF 5, 115/365, Structure 4. Fig. 30.9. Base of straight-sided bottle, bluegreen; small bubbles; dull surface; base design: one circle; not cast. Dims: 13 x 21 Th: 5 EVE: 0.14. SF 599, 105/360, west of Structure 4. Reeded handle fragment, bluegreen; many elongated bubbles, greenish striations; no weathering. W(handle): 35 mm H(to turn): 33 Th: 2 EVE: 0.14. SF 596, pit F256. Glass objects 19 Bangle Bangle fragment (60 degrees); greenish glass W 17 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 20 21 22 23 24 25 26 27 with three uninterrupted and three interrupted opaque yellow trails on the outside. D(inner): 60 H: 18 Th: 13. SF 331, pit F253. Fig. 30.10. 41 42 Beads Fragment, dark blue; rectangular bead with facets at the edges; weathered. PH: 10 Dims (ext): 6x7 D (inner): c. 3. SF 9, 110/360, Structure 4. Fig. 30.11. Bluegreen translucent wound segmental bead, one segment and part of a second surviving; many small bubbles, no weathering. PH: 5 D(max): 3 D(inner): 1. SF 241, 140/295, east of F236. Bluegreen, translucent wound segmental bead, three segments; small bubbles; no weathering. H: 17 D(max): 5 D(inner): 1. SF 328 145/295, east of F236. Fig. 30.12. Bluegreen opaque wound segmental bead, four segments surviving; small bubbles; no weathering. H: 14 D(max): 4 D(inner): 1. SF 202, 125/295, east of F236. Fig. 30.13. Ring bead fragment; bluegreen; opaque; many bubbles, wound, matt surface. D: 22 H: 9 D(inner): 5. SF 92 110/285, east of Structure 11. Bluegreen, opaque globular glass bead; small bubbles; no weathering. H: 4 mm D(max): 4–5 D(inner): 1. SF 283, 145/285, east of F236. Additional glass beads Gaming counters Black, complete; weathered. D: 15 H: 7. SF 112, 60/310, Structure 6. White, splintered; weathered. D: 14 H: 5. SF 250, pit F213. Window glass 28 29 30 31 32 33 34 35 36 37 38 39 40 W 18 Matt/glossy Edge fragment, bluegreen. Dims: 18 x 29 Th: 8. SF 483, 95/290, Structure 11. Fragment, bluegreen. Dims: 12 x 26 Th: 4. SF 636, pit F253. Blown One fragment. SF 598, 100/365, west of Structure 4. One fragment. SF 405, 85/405, east of Structure 1. One fragment. SF 485, 130/275, east of F236. One fragment. SF 487, 60/285, Compound 7. One fragment. SF 488, 60/290, Compound 7. One fragment. SF 494, 30/290, Compound 7. One fragment. SF 520, unprovenanced. One fragment. SF 598, 100/365, area of F134. One fragment. SF 601, 75/410, area of Structure 1. One fragment. SF 603, 100/365, area of F134. One fragment, grozed edge. SF 604, 85/405, east of Structure 1. One fragment. SF 622, 40/305, Compound 7. One fragment. SF 637, 70/310, area of Structure 17. Lynne Bevan Four further glass beads were recorded. 43 Opaque, tubular green bead, a type of bead common in Britain from about the 1st century onwards (Guido 1978, fig. 37.5). SF 44, 90/380, area of Structure 2 44 Well-formed, segmented bead of blue-green glass similar to nos 21–23. F224 45–6 Melon beads made of turquoise faience, one of which was incomplete. Melon beads are usually found in 1st and 2nd-century contexts (Guido 1978, 100). SF 23, Area A and unprovenanced. THE ROMAN POTTERY Annette Hancocks with contributions by Gillian Braithwaite, Brenda Dickinson, Kay Hartley, J.M, Mills and Roger Tomlin Introduction and methods A total of 20,267 sherds of pottery was recovered. The material derived from buildings, pits, ditches and general overall spreads with an appreciable proportion coming from hand clearance following machining. With the exception of the latter material, the pottery was initially scanned for general spot dating; this was undertaken by Jane Evans. The pottery from the initial surface cleaning was not dated but a rapid check for any new forms was made. All the samian from whatever source was recorded. A selection of the pottery from key groups representing good stratigraphic sequences was recorded in detail. This amounted to a total of 4065 sherds, 20% of the stratified material (Table 5). All the pottery examined was generally in the form of small abraded sherds. The key groups were well-stratified, well-dated groups with 3, 5 and 7 closely associated with buildings 1, 7 and 11. The pottery was recorded by context, and fully quantified by count, weight and rim EVE, using the existing Field Archaeology Unit Roman pottery recording system. Fabrics, form types and vessel classes were recorded using the existing type fabric and form series (Evans 2001), and these descriptions will not be repeated here. New forms recognised were integrated into the established form series and catalogued in greater detail, along with information concerning production, such as potters’ stamps and wasters, decorative motifs, post-deposition abrasion, use (sooting and/or residues) and reuse or FOSSE LANE 1996: SPECIALIST REPORTS repair (repair holes, lead rivets and counters). Cross joins and joins were noted where possible. The data are tabulated to show the fabrics present by period (Table 6). The pottery is illustrated by key context group. The key groups Group 1 Ditch F134 (contexts 1059, 1064 and 1066); Period 4 A total of 30 sherds representing a minimum of 3 vessels was recovered. Seven fabrics were identified, with sandy micaceous ware and fine micaceous sandy ware dominant. The material was generally very fragmentary and poorly abraded. A single sherd was recovered from 1066 and three sherds from 1064. No diagnostic forms were recognised with the exception of a Black Burnished ware dish (D5.11) of 3rd-century date from layer 1059, the uppermost fill. Group 2 Spread F133 (context 1002); Period 2 A total of 293 sherds from a minimum of 36 vessels was identified. Eleven fabrics were recognised, the most dominant ones being fine micaceous sandy ware and samian. All of the material was of 2nd-century date. With the exception of two unburnt samian sherds from Les Martres-de-Veyre, the bulk of the samian was very heavily burnt, a factor which made identification and close dating of fabrics difficult. None of the other fabrics identified was burnt. The samian included Hadrianic-early Antonine forms (Dr 37, Dr 36, Dr 18/31R, Dr 33), as well as later ones such as Dr 31 and Curle 23, but none of the latest second century samian forms were present (Fig. 32.6). The pottery was generally very fragmentary. Two new forms were illustrated (Fig. 31.1 and 2). These comprised a Severn Valley ware mortarium stamp (M) and a ‘Belgic’-derived platter (P1.12) in a reduced fine micaceous sandy ware. Forms previously recognised include a single Black Burnished ware bowl (B22.21), a coarse reduced ware bowl and dish (B23.11 and D1.11), sandy reduced coarsewares (B16.21, F6.32, J6.21, JC3.42 and JC4.22), a sandy micaceous coarseware jar (J5.51), a Shepton Mallet Severn Valley mortarium (M1.12) and a Dressel 20 amphora rim. The majority of the material has a utilitarian function associated with food preparation and storage. The only fineware recognised was the samian, some of which was residual. The pottery may have been deposited in rubbish used to make up the ground level. Group 3 Structure 1; Periods 1 and 2 A total of 878 sherds was recorded dating to the 2nd century. Yard F112 (context 1019) - A total of 623 sherds was recovered with 12 different fabrics identified, dominated by Black Burnished ware and sandy reduced coarseware. Forty six vessels were represented consisting of functional wares such as Black Burnished ware and reduced greyware cooking pots, fine sandy greyware beakers, including a face pot fragment (Fig. 31.3), reduced coarse greyware dishes (D1.14; Fig. 31.4), occasional Severn Valley mortaria, and samian cups (Dr 33s) and bowls (Dr 37s). A total of 20 sherds of samian representing 16 vessels was recognised. The material was mainly of Antonine date, with the latest form being a Walters 79. An earlier Les Martres-de-Veyre vessel was present. Shallow pit, F163 (context 1079) - Two fabrics were recognised, a single oxidised Severn Valley ware sherd and three sherds of Black Burnished ware, one of which was a jar type. No samian or other diagnostic material was recovered. Pit F117 (context 1018) - A total of 51 sherds was identified in three fabrics, sandy reduced coarseware, sandy reduced micaceous ware and sandy oxidised ware. These are all early fabrics but no diagnostic dateable forms were observed. Culvert F120 (context 1016) - Eighty eight sherds were recorded in eight different fabrics dated to the later 2nd/early 3rd century. The majority of sherds comprised oxidised Severn Valley ware, Black Burnished ware and locally produced coarse and fine greywares. Only two vessels were represented, a Black Burnished ware cooking pot form (JC3.13) and a reduced coarse greyware bowl (B4.21). Burnt area F143 (context 1022) - A total of 64 sherds was recorded in four fabrics - oxidised Severn valley ware, reduced sandy fine and coarsewares and Black Burnished ware. No diagnostic material was present. Burnt area F162 (context 1025) - Forty eight sherds were recovered. A Dr 35 samian sherd of possible Hadrianic date was present, Six further fabrics were observed including oxidised Severn Valley ware, Black Burnished ware, reduced sandy greywares, sandy micaceous greyware and sandy oxidised wares. At least five vessels were represented of which one was a Black Burnished ware dish (D5.11) and another a new bowl form in a sandy micaceous greyware (Fig. 31.5). Overall very little diagnostic and dateable material survives from Structure 1. A date in the late 2nd/early 3rd century seems likely for its abandonment, with occupation through the 2nd century. The building itself could well have been used in the preparation and storage of foodstuffs, given the presence and dominance of coarsewares over finewares. The majority of the forms recognised have a utilitarian function, with vessels such as mortaria, Severn Valley ware bowls and Black Burnished ware dishes and storage jars associated with food preparation and storage. W19 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 TABLE 5: ROMANO-BRITISH POTTERY: FABRIC, SOURCES AND QUANTITIES IN ASSEMBLAGE Fabric name Qty % Wt (g) % Rim EVE % BBC SANDBRF 5ANDMC SANDRC SANDRF SANDROC SANDRL SANDRM SVOXGR Total reduced 403 356 264 452 357 30 11 165 35 2073 10 9 6 11 9 1 <1 1 1 49 2930 1823 2440 3927 2454 1336 129 2255 300 17594 8 5 7 11 7 4 <1 6 1 49 423 341 141 571 418 27 179 120 2220 4 4 1 6 4 <1 2 1 22 SVOXG SVOXGM SVOXGCC(R) Total Severn Valley 262 42 10 314 6 1 <1 7 1377 384 89 1850 4 1 <1 5 182 27 209 2 <1 2 MISCCR M1SCCW SANDOX SANDOXF SANDOXG SANDOXCCW Totaloxidised 40 2 82 51 4 10 189 1 <1 2 1 <1 <1 5 421 6 566 316 45 37 1391 1 <1 2 1 <1 <1 4 107 20 57 34 218 1 <1 <1 <1 2 SUBTOTAL 2576 61 20835 58 2647 26 BBI SAVNAK Total other wares 1197 84 1281 30 2 32 9090 4465 13555 25 12 37 668 10 678 7 <1 7 OXFW OXFCCW OXFCCR Total Oxford wares 3 4 1 8 <1 <1 <1 <1 130 57 1 188 <1 <1 <1 1 22 22 <1 <1 LNVCC MORTCCW Total traded wares 2 2 1293 <1 <1 32 1 74 13818 <1 <1 39 700 7 SAMIAN DRES2O GIMPORT IMPORT Total imported 158 15 1 2 176 4 <1 <1 <1 4 926 1163 93 10 2192 2 3 <1 <1 6 30 21 15 20 86 <1 <1 <1 <1 1 CGW CREAMI PM PREH1STORlC Total unknown source 1 1 3 15 20 <1 <1 <1 <1 <1 27 7 11 54 99 <1 <1 <1 <1 <1 11 11 <1 <1 4065 100 36944 100 3444 34 GRAND TOTAL W 20 FOSSE LANE 1996: SPECIALIST REPORTS Group 4 Ditches F246 (contexts 2046, 2096, 2097, 2099 and 2100) and F257 (context 2071); Period 1 A total of 288 sherds was recovered from the two ditch fills suggesting a date in the late 1st or early 2nd century. Several previously identified forms were recorded. A total of 53 minimum vessels was observed in 15 different fabrics. Nine sherds of samian representing eight vessels were identified. Amongst the forms recognised was a Dr 37 of Flavian/Trajanic date, with early 2nd-century material possibly no later than the mid 2nd century. The wide range of forms and fabrics recorded was generally of a utilitarian function. F246 (2046) - The upper fill of the boundary ditch contained 84 sherds of 2nd-century date, in 12 different fabrics. Although the sherds are very fragmentary a range of early forms was recognised. At least twelve minimum vessels were identified, with diagnostic rim forms including a Black Burnished ware dish (D5.12), a reduced sandy fineware bowl (B17.11), three samian forms (Dr 37, Dr 18/18/31 and Dr 30). A new flagon form (F9.11) for Fosse Lane, in a reduced sandy micaceous fabric, was recognised (Fig. 31.6). F246 III (2096) - This is the upper fill of the ditch and comprised 64 sherds, representing 10 different fabrics and 15 minimum vessels. No new diagnostic material was recognised, although a variety of previously published forms confirm the residual nature of some of the ceramics within the deposit. Forms identified include Severn Valley ware bowl, beaker and tankard forms (B10.24, BK2.61 and T1.12), Black Burnished ware cooking pot forms (JC3.22, JC3.23 and JC3.31) and lid form (L1.21), sandy reduced fineware bowls (B23.41 and B20.11), sandy reduced coarseware dish (D4.11), sandy reduced micaceous ware wide-mouthed jars (JW3.11 and JW5.21) and samian Dr 37 and Dr 18/18/31 forms. F246 III (2097) - A total of 68 sherds were recovered from the lower fill. The date range for the material was late 1st/early 2nd century. A total of 12 minimum vessels was recognised in at least nine fabrics. The samian was of Flavian/Trajanic date. Diagnostic material consisted of Black Burnished ware cooking pot and dish forms (JC3.11, JC4.11 and D5.11), sandy reduced coarseware lid (L2.11), sandy reduced fineware bowl, dish and wide-mouthed jar forms (B13.11, D6.41, JW2.23 and JW5.11). F246 III (2099) - A single sherd of sandy reduced fine greyware was recovered from the lowest fill of this feature. No diagnostic material was recognised. F246 IV (2100) - A total of four sherds of pottery was recovered from this fill. Three different fabrics were identified, although no dating evidence survived. F257 (2071) - Thirteen fabrics were present in the single fill of this ditch. A total of 67 sherds was represented. All the samian was of Hadrianic/Antonine date. Several diagnostic forms were observed in a range of fabrics, although no new forms were recognised. Dateable material comprised a Severn Valley ware tankard (T1.14), a Black Burnished ware dish (D6.11), a reduced sandy fineware bowl (B23.31), reduced coarseware wide-mouthed bowl (JW2.31), a reduced blue grey, medium to coarse ‘pimply’ micaceous ware beaker (BK3.41), a reduced sandy micaceous jar (J5.51) and a Dr 35. At least 14 vessels were recognised. Group 5 Colluvium layers from Area B (contexts 2030, 2031, 2062); Period 3 A total of 1242 sherds were recovered comprising 177 sherds from 2030, 956 sherds from 2031 and 110 sherds from 2062. An overall total of 43 sherds of samian representing 32 vessels was recorded. Layer 2030 - A total of 177 sherds in 14 different fabrics representing 25 vessels was recognised. A 2ndcentury date was suggested by a Black Burnished ware bowl (B23.31) and lid (L1.21). Also present were a sandy reduced fineware bowl (B21.11), a sandy reduced fineware dish (D6.11), sandy reduced micaceous ware beaker, jar and wide-mouthed jar (BK3.12, J10.15 and JW2.22), miscellaneous colour coat beaker (BK4.22), some Dressel 20 amphorae sherds and five possible late Iron Age body sherds. The samian from this layer was Flavian and early 2ndcentury material, again perhaps no later than the midsecond century. At least three Dr 37s were observed and a Dr 18/18/31. Layer 2031 - A total of 956 sherds was recovered. Nineteen different fabrics were present, representing a minimum of 85 vessels. Severn Valley ware forms identified include bowl (B21.21), beaker (BK3.75) and tankard (T2.12) types. In Black Burnished ware were bowls (B22.21 and B24.12), beaker (BK2.11), dish (D3.11), cooking pot (JC3.23), wide-mouthed jar (JW1.11) and lid types (L1.21). In reduced sandy greyware were a bowl (B21.11), beakers (BK3.41 and BK3.81) and a lid (L6.11); and in reduced black, fine sandy ware were bowls (B20.21 and B22.23), a jar (J14.21) and a wide-mouthed jar (JW2.21). Sandy reduced coarseware forms were a bowl (B22.23), a beaker (BK3.21), a flagon (F5.11), jars (J14.31 and J8.21), a cooking pot (JC3.41) and wide-mouthed jars (JW2.21 and JW3.11). Also present were a sandy reduced micaceous ware bowl (B22.22) and reduced colour coat beakers (BK3.12 and BK3.65). The samian was exclusively Antonine in date with the latest vessels (joins 2094) dated AD 150–190, a Dr 37 bowl (Fig. 32.4). Other samian forms identified included Dr 31 and Dr 18/31 or 31 types. Some Dressel 20 amphorae sherds were present. Two 4th-century coins were found which must be viewed as intrusive. Layer 2062 - A much smaller quantity of pottery W 21 Fabric name Qty P2 Wt % (g) rim Av wt Qty P3 Wt % (g) rim Av wt Qty SANDRF 5ANDBRF 5ANDRM SANDMC SANDRC BBC SVOXGR SANDRL SANDRGC SVOXG SVOXGM SANDDXF SANDOX 5ANDOXCCW SANDOXG SVOXOCC(R) MISCCW MISCCR Total regional 101 618 129 513 14 383 107 1100 157 1305 124 665 22 145 1 9 5 119 116 424 35 345 12 66 3 11 3 6 1 3 830 5712 191 76 46 58 195 163 28 13 7 830 6 4 27 10 8 5 7 9 24 4 10 6 4 2 3 7 1 2 8 41 48 13 8 4 27 4 39 1 196 35 61 15 140 25 454 48 590 36 356 56 61 56 364 42 12 90 8 2227 222 35 30 18 11 12 27 8 14 13 3 2 8 11 BBI SAVNAK OXFW OXFCCR OXFCCW LNVCC MORTCCW Total traded 412 2328 13 444 2 1 2 74 429 2847 165 165 6 34 <1 37 7 69 57 126 563 3542 4105 80 10 90 8 62 33 648 13 1 4 666 PM CGW CREAM PREHISTORIC Total unknown source SAMSG SAMCG DR20 Total imported TOTAL POTTERY P4 Wt (g) 246 1754 197 1139 131 1616 99 731 225 1824 232 1619 1 15 5 47 25 1217 115 568 2 22 33 224 38 459 9 29 4 45 7 83 2 6 39 418 1410 11816 P5 % Av rim wt Qty Wt (g) % rim Av wt 227 210 108 35 301 145 27 102 14 34 50 20 107 1380 7 6 12 7 8 7 15 9 49 5 11 7 12 3 11 12 3 11 8 9 28 12 17 22 34 4 1 4 1 6 2 140 47 110 116 155 208 290 79 17 21 5 26 6 1080 40 39 59 67 10 215 5 4 10 9 9 8 20 17 5 5 4 3 8 5716 473 82 57 6328 366 15 381 9 36 82 14 10 67 1 2 1 71 478 6 48 I 533 57 7 64 7 6 24 1 8 - - - - - - - - 3 1 15 19 11 7 54 72 11 11 4 7 4 4 1 1 27 27 - 27 27 4 87 1 92 14 629 131 774 10 39 21 70 4 7 131 8 1 3 4 3 39 42 - 3 13 10 5 54 14 73 8 229 1032 1269 10 10 2 15 74 17 1 1 2 1 3 4 5 5 1 3 2 1351 9333 1065 7 326 6374 312 20 2168 19485 1782 9 214 1644 284 8 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 W 22 TABLE 6: SUMMARY OF FABRICS BY PERIOD (KEY GROUPS ONLY) FOSSE LANE 1996: SPECIALIST REPORTS was recovered from this layer. Twelve fabrics were identified representing a total of 110 sherds. At least 16 minimum vessels were recognised. A single new diagnostic rim form was identified from the group, a reduced black, fine sandy ware lid (L4.12, Fig. 31.7). The only other dateable piece recovered was a Black Burnished ware cooking pot (JC3.23). The samian comprised four sherds of Dr 31, probably exclusively Antonine in date. As with layer 2031 the pottery evidence was contradicted by the presence of three 4th-century coins. Leaving aside the coin evidence the pottery suggest a 2nd-century date for the three colluvium layers. Group 6 Period 2 Oven F249 (2053, 2091, 2093, 2094); A total of 53 sherds was recovered from the fill (2053) and rakeout (2093) of the oven, with a further 180 sherds recorded from the yard or building floor surface 2094 associated with F249 and from layer 2091 nearby. Seven sherds came from 2053. A single vessel was recognised in a Black Burnished ware bowl form (B22.21). The only other fabric recovered was a Severn Valley ware sherd. Forty six sherds came from layer 2093 and comprised six fabrics, including Black Burnished ware (D5.12), reduced black, fine sandy ware, sandy reduced coarsewares, sandy micaceous wares, sandy oxidised finewares and some prehistoric material. The material was dateable to the 3rd century. At least four vessels were present. The pottery from layer 2094 comprised 8 different fabrics and ten minimum vessels. Severn Valley ware forms included bowl and tankard forms (B9.13 and T2.12). Black Burnished ware was in jar forms (J1.21 and JC3.23), and reduced fine micaceous sandy ware in a bowl form (B20.14). A Dr 37 form dated AD 150–190 was present (Fig. 32.4), and five sherds from three Hadrianic and Antonine vessels including 3 sherds joining with 2031. The pottery from the stone rubble platform to the east of F228 (2091) was very heavily abraded and fragmentary in nature. A total of 10 different fabrics was identified and at least 11 vessels. Very little diagnostic material was recovered, although samian forms Dr 33 and Dr 30 were observed, both of Hadrianic/Antonine date. Group 7 5 Structure 7 (2024, 2025, 2029); Period Three contexts were associated with this structure. A total of 417 sherds was recovered, representing a minimum of 37 vessels. From the floor within Structure 7 (2024), a total of 33 sherds, representing at least 1 vessel, was recovered and dated to the 3rd century. Generally the material was poorly abraded and small in size. Only six fabrics were identified, none providing any diagnostic material, although Black Burnished ware and sandy oxidised wares were present. Layer 2025 represented a rubble deposit in Structure 7. A total of 208 sherds was represented in 15 different fabrics and making 25 minimum vessels, dating to the late 3rd/4th century. Much residual second-century material was recognised, including Savernake ware, Severn Valley ware and samian. No new forms were observed. Diagnostic wares present included Black Burnished ware bowls, dishes and cooking pot forms (B24.11, B24.13, D5.11, JC4.11 and JC3.23); coarse greyware copies of BB1 forms such as dish (D5.11) and cooking pot forms (JC3.23, JC4.11 and JC4.21); sandy reduced coarseware beakers, jars and lids (BK3.81, J9.31 and L3.11); Oxfordshire fine white mortarium (M2.71) and four very small sherds of samian of mixed date. Layer 2029 represented burning in Structure 7. A total of 176 sherds was identified, in 11 different fabrics and 11 minimum vessels. The layer comprised mainly material dating to the early 3rd century. Little diagnostic material was recovered, with the exception of a Black Burnished ware dish (D5.11) and cooking pot (JC3.31), and a coarse reduced sandy greyware dish copy (D5.11). A much later date was suggested by three 4th-century coins. Group 8 Midden (2092); Period 2 The layer contained 159 sherds in ten different fabrics of late 1st/early 2nd century date with a minimum 22 vessels present. Diagnostic forms include Severn Valley ware bowls (B3.11, B7.11 and B9.13), Black Burnished ware dish and cooking pot forms (D1.11 and JC3.23 and JC3.31), reduced coarse greyware BB1 copies of cooking pots (JC3.13, JC3.23 and JC3.31), a reduced fine micaceous sandy ware tankard (T1.11), sandy reduced coarseware beaker and jar (BK3.21 and J6.21), a reduced blue grey medium sandy ware bowl and flagon (B22.22 and F5.11), brown/buff to grey, medium to coarse ‘pimply’ micaceous ware bowls and a wide-mouthed jar (B13.11, B22.22 and JW2.22), and samian comprising two small sherds and two stamped Dr 33 bases (stamps Fig. 33.1 and 33.8) of mixed dates from first through the second centuries. Group 9 Pits F251 (2060) and F256 (2056, 2057, 2070 and 2073), Period 4; and hearth F250 (2054), Period 2 The fill of F251 contained over 264 sherds of pottery from a minimum of 21 vessels in 11 different fabrics. The assemblage dated to the 3rd century and included diagnostic material in the form of Black Burnished ware bowls (B23.11 and B23.31), a dish (D5.11), a W 23 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 jar (J10.41) and cooking pots (JC3.24 and JC3.31). Other forms recognised include sandy reduced coarseware Beaker forms (BK3.21 and BK3.81) and three small sherds of samian including fragments of Dr 37 and Dr 31 forms of Hadrianic-Antonine date. Two 4th-century coins demonstrated that all the pottery was residual. A total of 116 sherds was recovered from F256. The upper fill, 2056, contained 24 sherds representing at least two vessels with eight 8 different fabrics. No new forms were observed, with the only diagnostic material comprising two Black Burnished ware cooking pot forms of 3rd-century date (JC3.23 and JC4.21). In addition, a single sherd of samian (Dr 43/Curle 21) of late 2ndcentury date was present. Layer 2057 contained 31 sherds in 10 fabrics representing at least nine vessels. Black Burnished was the most common fabric represented. Diagnostic material consisted of several forms including a dish (D5.11), bowl (B23.11) and cooking pots (JC3.23 and JC3.31). Additionally a single sherd of prehistoric pottery was identified. Layer 2070 contained 38 sherds in 6 fabrics representing at least five vessels. The material was of late 2nd/early 3rd date. Very little dateable material survived, with the exception of a Black Burnished ware bowl (B23.11) and a sandy reduced fineware jar (J10.32). A good dateable piece of samian was a Dr 37 bowl fragment (Fig. 32.3) dating to AD 150– 80 and a Dr 18/31R or 31R. The lowest fill (2073) provided very little dateable material from 23 sherds. A single minimum vessel was represented, although six different fabrics were observed. The only diagnostic form was a Black Burnished ware dish (D5.11) of 3rdcentury date. A total of 13 sherds was recovered from F250. The pottery consisted of small and very abraded sherds. Very little dateable material was observed, although two fragments of Dr 18/31 and Dr 30/37 samian of Hadrianic-Antonine date formed part of the assemblage. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Illustrated sherds (Fig. 31) 24 1 2 3 4 5 W 24 Mortarium (M); see comments by K Hartley; MSVOXG, 8%. Layer 1002, F133, Period 2, KG2 Platter with Gallo-Belgic derivation; P1.12 SANDRF, 9%. Layer 1002, F133, Period 2, KG2 Decorated sherd, facepot fragment; SANDRF. Layer 1019, Period 2, KG3 Dish with flanged rim; D1.14 SANDRC, 31%. Layer 1019, Period 2, KG3 BB1-type bowl; conical bowl with flat grooved and flanged rims; B23.42 SANDMC, 7%. Layer 1025, F162, Period 2, KG3 25 26 Flagon with bifurcated rim; F9.11 SANDMC, 10%. Layer 2046, F246, Period 2, KG4 Lid with internal groove; L4.12 SANDBRF, 11%. Layer 2062, Period 4, KG5 Bag-shaped beaker; BK3.42 SVOXGR, 16%. Layer 3003, F301, Period 5 Bag-shaped beaker; BK3.44 SVOXGR, 30%. Layer 3003, F301, Period 5 Bag-shaped beaker; BK3.45 SVOXGR, 21%. Layer 3003, F301, Period 5 Bowl with rim overhanging internally flanged bowl (segmental); B2.22 SANDBRF, 25%. Layer 1021, F142, Period 2 Bowl with rim overhanging internally flanged bowl (segmental); B23.43 BBC, 19%. Layer 2037, F211, Period 3 Dish, for graffito see Tomlin below; DSANDBRF, 7%. Layer 2048, F217, Period 4 Dish with plain rim and increasingly splayed rim; D4.12 SVOXG, 16%. Layer 1001, Period 4 Jar with near triangular rim, ?waster; J6.21 SANDRC. Layer 3003, F301, Period 5 Globular jar with inturned rim and external lid seating; J16.12 SANDRM, 12%. Layer 3000, F209, Period 2 Jar with splayed rim of near equal girth, obtuse cross-hatch burnish surmounted by a groove; for graffito see Tomlin below, JC4.11 BB1, 19%. Layer 2019, F225, Period 5 Mortarium; M5.31 GIMPORT, 15%; Antonine (Hartley 1991, fig. 84, C51). Layer 1037, F150, Period 6 Platter, Gallo-Belgic derivation; P3.11 SANDBRF, 10%. Layer 1049, F134, Period 4 Platter, Gallo-Belgic derivation; P4.11 SANDBRF, 10%. Layer 2017, Period 2 Platter, Gallo-Belgic derivation; P5.11 SANDBRF, 6%. Layer 2017, Period 2 Tankard; T4.11 IMPORT. Layer 2020, F223, Period 6 Decorated body sherd, facepot fragment; Gillian Braithwaite comments that this type is not easily paralleled by West Country military examples. It is very late in date. SVOXGR. Layer 3003, F301, Period 5 Decorated body sherd. Layer 2082, F260, Period 4 Decorated body sherd; (T-) IMPORT. Layer 2004, F213, Period 7 Small cup; MS7.11 SANDOX, 7%. Layer 2017, Period 2 Mortarium Kay Hartley Two joining sherds from the bead and upper part of the flange of a mortarium (Fig. 31.1). The bead and upper surface of the flange are blackish grey, almost FOSSE LANE 1996: SPECIALIST REPORTS Fig. 31 Roman pottery certainly the result of overfiring, though the vessel may have been a second rather than a waster. The inside fabric and inner surface are ‘cocoa’ brown (Munsell 10RS/4), and slightly rough to the touch. No obvious slip survives but the brown stains in and near the stamp may be the remains of a discoloured red-brown slip. The fabric is fine-textured and the moderate, tiny and small quartz and slag inclusions are barely visible at less than x20 magnification. Two quartz trituration grits survive. The left-facing stamp is fragmentary, but enough survives to show that it is from an unrecorded die; other stamps from the same die should be identifiable. The large bead, the fabric, the type of stamp and the provenance leaves no doubt that it is a product of the Shepton Mallet pottery workshops, which were active within the period AD 100–140, perhaps early in the century. Samian: decorated wares J.M. Mills Sherds representing some 86 or 87 vessels are described in the catalogue below. There will inevitably have been a few rim and base sherds within the remaining unrecorded material, so that the total number of decorated vessels is probably slightly higher. The decorated wares date from the Neronian period through to the late Antonine. Although there are no apparent gaps in the supply of these vessel types only three of Hadrianic date were recorded from a total of 65 closely dated vessels. Many of the sherds are small and with fragments of decoration surviving. Of the 28 decorated vessels from Southern Gaul, 26 are probably from La Graufesenque. One is a late 1st or early 2nd-century Dr 37 from Montans (Fig. 32.1), the source of the other, a Neronian Dr 29, was not identified. The decorated wares include seven Dr 29s (SG), one Dr 29 or 37 (SG); four Dr 30s (SG 1, W 25 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 CG 3); 67 or 68 Dr 37s (SG, 18; Les Martres 15, CG Lezoux 35 or 36), one Dr 67 (SG) and one Dr 64 (CG). The latter, although not unusual, is a less common form (Fig. 32.6) and is probably the work of Libertus. The discrepancy in the numbers of Central Gaulish Dr 37s results from the possibility that two of the sherds may be from the same vessel. 14 15 16 17 Catalogue 18 The catalogue is presented in approximate date of manufacture order. 19 1 20 2 3 4 5 6 7 8 9 10 11 12 13 W 26 Dr 29, SG. Base with fragment only of lower zone of straight gadroons; AD 55–70. 120/260 Dr 29, SG*. Small body sherd with two small leaves from a scroll design extant. The fabric is almost certainly not from La Graufesenque* although its source is not identified; Neronian. 80/295 Dr 29, SG. Edge of lower zone of decoration fan-shaped plant only remains. Similar plants occur on bowls from La Graufesenque stamped by Germanus (his earlier work, not using his own moulds); Late Neronian and early Flavian type; AD 60–75. 120/260 Dr 29, SG. Carination of bowl with winding scroll in upper zone and zonal decoration below with inhabited medallions. Small spirals used as infills in both upper and lower designs; AD 60– 75. 130/340 Dr 29, SG. Foot ring fragment; Neronian or early Flavian. Layer 3011 Dr 29, SG. Fragment of plain band and bead rows from carination; AD 70–85. Layer 2011 Dr 37, SG. Basal wreath only; AD70–90. 75/285 Dr 37, SG. Body sherd in the style of Germanus with ovolo with corded tongue with rosette terminal; no border below, with only fragments of decoration extant; AD 70–90. Layer 2030 Dr 37, SG. Lower part of decoration with a wreath of S-shaped gadroons with wavy line below and a basal wreath of bifid leaves; AD 70–90. 125/265 Dr 29, SG. Fragment of plain band with large bead row either side and a scrap of ?winding scroll; Early-mid Flavian. 75/280 Dr 37, SG. Body sherd with fragments of two double-bordered medallions or festoons and a long, triangular leaf; AD 75–95. Layer 2048 Dr 37, SG. Ovolo with trident tongue and wavy line below. M Crestio regularly used a version where the trident is blurred as in this example; c AD 75–100. 80/270 Dr 37, SG. Body sherd with fragment of zoned decoration comprising panel infilled with diagonal wavy lines to leave triangular zone inhabited by a lion running to right (similar to 21 22 23 24 25 26 27 28 29 30 31 32 O.1400); Flavian. Layer 3007 Dr 37, SG. Fragment of decoration, too small to identify; Flavian. 50/300 Dr 67, SG. Fragment from top of moulded decoration; Flavian. 100/375 Dr 30, SG. Fragment of decoration, too small to identify; Flavian. 35/320 Dr 37, SG. Fragment of medallion only; Flavian. Layer 2030 Dr 37, SG. Body sherd with fragment of spiral with central rosette; Flavian. 75/280 Dr 37, SG. Fragment of lower edge of decorated zone; Flavian. Layer 3011 Dr 37, SG (overfired). Fragment of decoration with wavy line border; Flavian. 90/400 Dr 37, SG. Small body sherd with fragment of trifid-tongued ovolo; AD 85–110. 105/386 (layer 1001) Dr 37, ?SG. Ovolo with trifid tongue, similar if not identical to ovolo of Mecator. Plain, single bordered medallion below contains a trilobed motif which appears on an unpublished stamped bowl from Leicester (Brenda Dickinson pers. comm.); AD 85–110. Layer 3007 Dr 37, SG. Trident-tongued ovolo with bead row below. Fragment of animal body below; FlavianTrajanic. Layer 2096 Dr 37, SG. Trident-tongued ovolo; FlavianTrajanic. Layer 2017 Dr 37, SG. Blurred ovolo only; Flavian-Trajanic. 55/275 Dr 37, SG. Fragment of moulded decoration; Flavian or Trajanic. u/s Dr 29 or 37, SG. Body sherd with horizontal bead row fragment; 1st century. Layer 1040 Dr 37, SG (Montans) (Fig. 32.1). Body sherd with dog (with collar) running left, similar to motifs used in Central Gaul eg O.1980, O.1989A. Bowls made at Montans sometimes have a similar dog running to the right. The fabric is very pale with occasional large (<1mm) red/ brown ?iron inclusions and is not a fabric used by the later Montans potters and so is unlikely to be later than the very early second century (Brenda Dickinson pers. comm.). Another sherd of similar fabric was recovered from 80/395 and may derive from the same vessel; Trajanic. 75/395 Dr 37, CG (Les Martres-de-Veyre). Fragment of ovolo; Trajanic. Layer 1019 Dr 37, CG (Les Martres-de-Veyre). Rim sherd with ovolo Rogers B44 with wavy line (Rogers A24) below. Potter of the Rosette; AD 100–120. Layer 2030 Dr 37, CG (Les Martres-de-Veyre). Body sherd with ovolo Rogers B28. Probably Potter X-2 or X-3; AD 100–120. 105/360 Dr 37, CG (Les Martres-de-Veyre). Body sherd with fragment of ovolo Rogers B37 with wavy FOSSE LANE 1996: SPECIALIST REPORTS Fig. 32 Decorated samian 33 34 35 36 line below. Fragment of rosette Rogers C280 and fragment of ?Perseus (O.234); Potter X-4 (Igocatus); AD 100–120. 105/360 Dr 37, CG (Les Martres-de-Veyre). Rim with ovolo Rogers B29; Potter X-4 (Igocatus); AD 100–120. 115/275 Dr 37, CG (Les Martres-de-Veyre). Body sherd, ovolo double impressed and not identifiable, wavy line below with tail of seahorse only surviving. The seahorse is probably that used by Potter X-12 (Stanfield and Simpson 1990, pl 40, 462); AD 100–120. 125/260 Dr 37, CG (Les Martres-de-Veyre). Fragment with rosette Rogers C292 and wavy line above (or below); Potter X-12; AD 100–120. 75/275 Dr 37, CG (Les Martres-de-Veyre). Fragment with horizontal bead row and festoon Rogers F8. 37 38 39 40 41 Probably Potter X-13 or X-14. AD 100–120. 90/ 270 Dr 37, CG (Les Martres-de-Veyre). Small body sherd, extant decoration comprises the junction of Rogers motif F74 with another, possibly J169, rather than a festoon or medallion with leafy border. The base of a ?trifid leaf sits at the junction; AD 100–120. Layer 2076 Dr 37, CG (Les Martres-de-Veyre) (Fig. 32.2). Fragment of ovolo with bead row below; ovolo overlaps; AD 100–120. Layer 2017 Dr 37, CG (Les Martres-de-Veyre). Small sherd with blurred ovolo; AD 100–120. Layer 2043 Dr 37, CG (Les Martres-de-Veyre). Chip with fragment of blurred ovolo; AD 100–120. Layer 2012 Dr 37, CG (?Les Martres-de-Veyre). Scrap of W 27 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 42 43 44 45 46 W 28 moulded decoration; ?AD 100–120. 110/285 Dr 37, CG (Les Martres-de-Veyre) (Fig. 32.3). Seven sherds from the same vessel, not all joining. The ovolo is that of Libertus (Rogers B214) with a poorly executed bead row below. The bowl is sloppily finished as is often the case at Les Martres-de-Veyre. The bowl is decorated with Ulysses in his ship, a variant of O.982A; sherds with fragments of three of these were recovered. The decoration is infilled with a variety of nautical motifs including turtles (O.2156 variant), fish (O.2417), and tritons or sea horses (not in Oswald). A bead row towards the lower edge of the decoration represents the sea, below which, on the sherd where this element of the design survives, swim fish. This same combination of motifs can be seen on an unstamped bowl from London, although seahorses replace fish in the lower zone on that bowl (Stanfield and Simpson 1990, pl 53, 626); AD 115–130. Layer 2017 and 60/275 Dr 37, CG (Les Martres-de-Veyre) (Fig. 33.3). See also stamps below. Body sherd with seahorse facing right similar to that on an unstamped bowl from London (Stanfield and Simpson 1990, pl 53, 626). Fragment of a possible stamp ]O or O[ at edge of sherd (stamp cat. no. 3). The slip is damaged and the sherd is of incomplete thickness and worn, however, the curvature of the sherd, the depth and detail of the moulding as well as the fabric suggest that this may be a fragment of the Libertus bowl (see 2017 etc) above; ?AD 115–30. 75/285 Dr 37, CG. Lowest part of panelled decoration of large double-bordered medallions with vertical bead rows between, these having terminal rings and flanked by astragalus Rogers R12. The two figures within the medallions are Abundance O.802 and Bacchus O.571. To the right is caryatid 0.1199. The work of Divixtus; all these elements can be seen on stamped bowls of his (Stanfield and Simpson 1990, pl 115, 3 and pl 116, 8) from Silchester and Coventry, and Carlisle and Leicester respectively; AD 140–175. u/s 4 Dr 37, CG (KG9; Fig. 32.4). Body sherd with ovolo Rogers B135 and corded border A36 below. The extant decoration comprises a plain festoon with ‘seahorse’ O.52A. In the style of Ivstus; AD 150–180. Layer 2070 Dr 37, CG. Base and lower part of decoration only comprising vertical borders Rogers A10, stands Rogers Q84, and the legs of figure O.569 but without the cup. Possibly the work of Illixo, who used the figure (with the cup) and the borders. The bowl retains a lead rivet from a repair to the base. The lowest part of the interior is heavily worn with no slip remaining; AD 150– 47 48 49 50 51 52 53 54 55 180. 55/285 Dr 37, CG. Fragment with ovolo Rogers B145 used by Cinnamus, Carantinus and Illixio; AD 150–180. u/s 3 Dr 37, CG. Body sherd, heavily burnt. Ovolo fragment probably Rogers B143, with bead row below. Figure within plain double bordered medallion is dancer O.348 with broken hand; Cinnamus; AD 150–180. 85/285 Dr 37, CG. Body sherd with fragment of panelled decoration with vertical bead row and figure of Venus at an altar O.322 to the right. Probably Cinnamus; AD 150–180. Layer 1001 Dr 37, CG. Body sherd from bowl in the style of Albucius. With dog running to left (possibly a small version of O.1984) and leaf tips in the ground, Rogers J145, but the edge only; AD 150–180. 125/265 Dr 37, CG (KG 5 and 6; Fig. 32.5). Two groups of joining sherds from the same vessel with ovolo Rogers B105 with blurred bead row below. The bowl is decorated with a freestyle design of wild animals, dogs and horses, with and without riders. The figures used include a variant of horseman O.246, horse O.1911, and bear O.1589. Above the bear are two very poorly impressed leaf-tip fillers. This could be the work of either Paternaus or Albacus, however, the leaftips may be those of Albacus (compare with bowls from Corbridge and London – Stanfield and Simpson 1990, pl 123.33, 35, 38, 41 and 42); AD 150–190. Layers 2031 and 2094 Dr 37, CG. Body sherd with ovolo Rogers B20; Secundus i; AD 125–140. 110/270 Dr 64, CG (Fig. 32.6). Rim fragment with two parallel grooves at top of decoration. Of the three figures two are complete and identifiable, cupid O.422 and kneeling stag O.1752; both were used by Butrio. The discus (Stanfield and Simpson 1990, fig. 12.4) used by Libertus appears below and to each side of the cherub. The stag and discus can be seen together on an unstamped form 64 beaker from London (Stanfield and Simpson 1990, fig. 52, 613) which is attributed to Libertus. There is no evidence for Butrio using the discus, suggesting that this may be the work of Libertus; Hadrianic, 80/395 Dr 37, CG (Fig. 33.2). Stamp in plain area below decoration (stamp no 2); AD 125–150. 60/265 Dr 37, CG. A bowl with the distinctive buds of the Cerialis ii – Cinnamus ii group (Rogers J178), here in very clear impressions and so probably indicating an early product of the potter. Other figures include Venus (O.286), winged centaur (O.735A); bear (O.1609) with head squashed (?by a thumb). The striated divider and row of circles below both used by this group; AD 135–150. 115/380 FOSSE LANE 1996: SPECIALIST REPORTS 56 57 58 59 60 61 62 63 64 65 66 Dr 37, CG. Rim sherd with ovolo Rogers B108 with wavy line A23 below. Maccius and Maccarirra used the ovolo; AD 130–160. Layer 2043 Dr 37, CG. Body sherd with ovolo Rogers B47 (or a variant of it) and a small rosette Rogers C120. The lack of a border below the ovolo is paralleled at Lincoln on a bowl in the style of Tetturo (Brenda Dickinson pers. comm.); AD 130–160. 85/270 Dr 37, CG. Body sherd with ovolo B144 with bead row below. Cerialis ii/Cinnamus ii; AD 135–170. Layer 2068 Dr 37 CG. Joining body sherds with ovolo Rogers B144 with bead row below. Moulding poor. Extant decoration comprises figure in arcade between vertical bead rows; Cinnamus/ Cerialis group; AD 135–170. 60/285 Dr 37, CG. Joining body sherds with ovolo Rogers B144, with small bead row below. Part only of panelled design extant comprising vertical bead row, open ring and fragment of motif O.234 (Perseus). Cerialis ii/Cinnamus ii group; AD 135–170. 105/365 Dr 37, CG. Body sherd with upper part of decoration comprising ovolo Rogers B107 with bead row below. Decoration is panelled with only the head of Pan O.709 extant. Paternus; AD 150–190. 50/405 Dr 30, CG. The decoration is panelled with bead rows dividing. The surviving elements include shell Rogers U76 in a plain double-bordered medallion and a column including vase? Rogers T16, Q59 and astralagus R18. Almost certainly the work of Doeccus; AD 160–190. Layer 2050 Dr 37, CG. Lower part of design with single line delimiting the decorated zone. The extant decoration comprises the lower legs and feet, probably of Venus O.278 and, to the left, fragments of a winding scroll of two plain lines terminating with acorn Rogers U87. Both the Venus and the acorn appear on sherds from Corbridge (Stanfield and Simpson 1990, pls 125.8, 126.16 respectively). The Venus has a damaged left foot as, from the illustration, the Corbridge example appears to have. The scroll is similar to the Corbridge example and may be seen, although in reverse, on a sherd from Dorchester (Pengelly 1982, fig. 33, 38); Iullinus style; c AD 160–190. Layer 2066 Dr 37, CG. Body sherd with ovolo (Stanfield and Simpson 1990, fig. 44.2), rosette Rogers C274 and beaded ring E58; Style of Doeccus; AD 160– 190. 75/280 Dr 37, CG. Scrap of decoration, rosette Rogers C144. Probably Paternus; AD 160–195 (Brenda Dickinson pers. comm.). 55/275 Dr 37, CG. Chip with fragment of rosette- 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 tongued ovolo and wavy line below; Hadrianicearly Antonine. 75/275 Dr 37, CG (burnt). Incomplete ovolo with corded tongue; Hadrianic or early Antonine. Layer 2021 Dr 37, CG (heavily burnt) KG 2 (Fig. 32.7). A substantial portion of a panelled bowl with ovolo Rogers B17 with large bead row (?A2) below, and panels divided with bead rows. The panels alternate between large human figures and a festoon inhabited by a cockerel with another motif (missing) below. Figures include Bacchus O.566, Venus O.278, Cock O.2348 (or variation) and a small figure similar to O.658B. Attribution is difficult, the ovolo is probably that of Paternus iv (Brenda Dickinson pers. comm.), however his repertoire of figures is not yet known; probably Hadrianic or early Antonine. Layer 1002 Dr 37, CG. Two sherds, not joining but likely to be from the same vessel. Fragment of panelled decoration with vertical bead row and small fragment of a human figure (unidentified); Hadrianic or Antonine. Layer 1019 Dr 37, CG. Body sherd with fragment of inhabited scroll, motif too fragmentary to identify; Hadrianic-Antonine. Layer 2003 Dr 30 or 37 CG. Scrap with decoration; Hadrianic or Antonine. 50/300 Dr 37, CG. Body sherd with small fragment of ovolo; Hadrianic or Antonine. 70/410 Dr 30 or 37, CG. Decorated fragment. Hadrianic or Antonine. 75/405 Dr 30 or 37, CG. Decorated fragment (ext surface only); Hadrianic or Antonine. 75/405 Dr 37, CG. Fragment of lower edge of decorated zone; Antonine. Layer 2031 Dr 37, CG. Fragment of decoration; Antonine. Layer 2031 Dr 30, CG. Body sherd with fragment of stag running to left with the antler tines pointing upwards. No exact parallel in Oswald; Antonine. Layer 2066 Dr 30 or 37 CG. Body sherd with vertical bead row and ?foot of human figure to left; Antonine. Layer 1040 Dr 37, CG. Body sherd with ovolo ?B17; Antonine. 55/405 Dr 37, CG. Body sherd with most of external surface missing. No identifiable decoration survives; Antonine. 55/405 Dr 37, CG. Fragment only, with single bordered medallion; Antonine. 115/260 ?Dr 30, CG. Two chips, probably Dr 30; Antonine. 50/315 Dr 37, CG. Lower part of decoration with vertical bead row with ring at terminal; standing Bacchus to the left (O.566); Antonine. 70/400 Dr 37, CG. Fragment of panelled decoration with vertical bead row and part of figure, possibly W 29 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 85 86 87 Venus (similar to O.279); Antonine. 75/400 Dr 37, CG. Fragment of panelled decoration with vertical bead row and leg of standing figure, probably Perseus (O.234) to the right; Antonine. 90/385 Dr 37, CG. Chip with fragment of decoration; Antonine. 90/385 Dr 37, CG. Body sherd with upper part of decoration. There is no ovolo on this bowl. The only figure is a cupid O.440; Mid–late Antonine. 75/405 4 5 Potters’ stamps Brenda Dickinson Each entry gives: excavation number, potter (i, ii, where homonyms are involved), die, form, reading, published example (if any), pottery of origin, date. Superscript (a), (b) and (c) indicate: (a) A stamp attested at the pottery in question. (b) Not attested at the pottery in question, but other stamps of the potter known from there. (c) Assigned to the pottery on the evidence of fabric, distribution, etc. Ligatured letters are underlined 6 7 Catalogue (Fig. 33) 1 2 3 W 30 Dagomarus 3b 33a [DAGOMA]RVSFV (final letter uncertain) Les Martres-de-Veyre a. . Dagomarus worked at Les Martres under Trajan and at Lezoux in the Hadrianic period. There are many vessels with this stamp in Les Martres fabric in the London Second Fire groups. c AD 100–120. Layer 2092, KG8 Gratus ii la 37 (stamped upside down below the decoration, after moulding) [GR]ATI.M Lezouxb. This appears on a bowl from York with a mould-stamp of Quintilianus i (Stanfield and Simpson. 1990, pl. 68, 7). It occurs also in the Rhineland, where the import of Central Gaulish ware seems to have ceased around the middle of the second century, and at South Shields. c. AD 125–150. 60/265 Libertus ii 6a 37 O[FLIBERTI] (Dechelette 1904, 282, no 110, 39). The presence of Libertus wasters at Lezoux is evidence that he worked there, but this bowl is in one of the fabrics in the range produced at Les Martres-de-Veyre, and so will have been from a mould traded there, or brought from Lezoux. Dating evidence for this potter is slight, but his limited distribution and his use of non-standard fabrics suggest that he was rather earlier than the main body of Hadrianic potters working at Lezoux. The similarities between his style of decoration and that of Butrio suggest Trajanic–Hadrianic date. 8 9 10 Layer 2017 Mettius 1a’ 18/31R METTI.MA Lezouxc. This is from a broken die which originally had ends to the frame. It occurs at Mumrills and Newstead, while a stamp from a different die is known from the Saalburg Erdkastell (before AD 139). c AD135–160. Layer 2072 Osbimanus-Cadgatis 2a 33 OSBIMLCL Lezouxa. This stamp records a partnership, or joint venture, of two potters, whose names are extrapolated from a stamp on which they are more complete. It was used only on cups, mainly of form 33, but with one example on form 46. Both potters worked independently at some stage in their careers. Their separate ranges of forms, which include 18/31R, 31R, 79 and Ludowici Tx, suggest activity c AD 150–180. 50/390 Pateratus 5a PATIIRATV Lezouxb. This stamp is known from Birdoswald, Chesterholm and Corbridge, and has been noted once on form 18/ 31R. Pateratus’s use of other dies on forms 18/ 31, 27 and 81, combined with the site evidence, suggests a range c AD 135–165; see also graffito. 85/300 Paterclinus 4a 18/31R or 31R [PLTER]CLINI (Juhasz 1935, 207). A stamp noted on Hadrian’s Wall and at hinterland forts recommissioned c AD 160; it has been noted on form 79 or Ludowici Tg. This evidence suggests mid to lateAntonine activity, but his use of other dies on form 27 may mean that he was active by the middle of the second century. c AD 150–180. Layer 2067 Severus vi 3d 33 SIIVIIRIM. Lezouxb. There is no internal dating for this stamp, but the potter’s output includes plain forms made at Lezoux in the later second century, such as 31R, 79 and 80, and decorated ware which belongs stylistically to the same period. Vessels stamped with some of his other dies occur at northern forts recommissioned c AD 160–c AD 160–190. Layer 2092, KG8 Suobnedo 2b 31 SVO[BNE]DOF (Dickinson 1990, fig. 183. 41) Lezouxb. This is only the second example noted on a dish of a stamp more normally used on cups of form 33. Dating evidence of Suobnedo is sparse, but he is known to have made form 79/80, and so Antonine activity is certain. c AD 140–170. Layer 1055 Tituro 5A 31 [TITVR]ONIS (Dannell 1971, 315, 92) Lezouxa. One of the less-common stamps of a potter working in the mid- to late-Antonine period. Stamps from his other ides occur in the Wroxeter Gutter find of the early 160s, on Hadrian’s Wall and on some of the later 2ndcentury forms, such as 31R, 79 and 80. c AD 160–190. Layer 2012 FOSSE LANE 1996: SPECIALIST REPORTS 11 12 13 ]CI (?) on form 33, burnt, Central Gaulish. Antonine. 95/265 S [or SX [on form 31, Central Gaulish. Mid- to late-Antonine. 80/305 (not illustrated) Frame only form 18/31. Central Gaulish. Hadrianic/early Antonine. ?Layer 1053 Graffiti Roger Tomlin Graffiti was recognised on four sherds (see below), two of which are samian. This includes a deeply scored ‘X’ on the underside of stamped vessel cat. no. 6. 1 2 3 4 Rim sherd of a black burnished ware jar with splayed rim, found with material of the late 3rd/ 4th century. Coarsely incised after firing: two vertical strokes close together, now incomplete, and a third, also incomplete, overlying more or less horizontal scoring. This could be read as [...N]I +, the genitive case-ending of a personal name followed by a mark of identification, but not enough survives for certainty. Layer 2019 (Fig. 31.17) Base sherd of a buff jar dated to the 3rd century. Scored underneath after firing: two parallel lines intersecting with a broad line at right angles. Not a letter, but a mark of identification. Layer 2048 (Fig. 31.13) Base sherd of a Dr. 33 cup stamped PATIIRATV (Lezoux), scratched after firing within the footring: ‘+’. A mark of identification. 85/300 Rim sherd, probably of a Dr. 31 bowl (ex catalogue), scratched after firing with four incomplete letters just below the carination. Only the tops of them survive, so the reading is not certain, but possibly: [...]MPRO[...]. This could be part of the name Sempronius, but it must be emphasised that with these irregular capitals being incomplete, the reading is in doubt. 85/ 390 Discussion Annette Hancocks Dating Dating of the pottery as a whole was inhibited by its fragmentary nature, whilst using the dating evidence was rendered less sure by the residual character of the majority of the assemblage. Nevertheless good quality dating evidence was found deriving principally from the samian and amphorae but also from Black Burnished ware forms. The samian gave the best evidence for 1st and 2nd-century activity. Samian reached the site from the mid-1st century Fig. 33 Samian stamps; scale 2:1 throughout the exporting period and into the 3rd century. The usual peaks and troughs within the assemblage are apparent with the bulk of the W 31 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 material being of mid to late-Antonine date. The quantity of Neronian and early Flavian material is not great and may indicate that activity on the site did not really begin until the 70s. For the 1990 collection Brenda Dickinson observed that within the Southern Gaulish decorated wares the ratio of Dr 29:Dr 37 is about 1:3 and concluded that activity on the site may have started as late as the mid-80s (Dickinson 2001, 144). The same ratio occurs in this assemblage. The range of forms present is fairly standard (15/ 17, 15/31, 18, 18/31, 18/31R, 27, 29, 30, 31, 31R, 32, 33, 35, 36, 37, 38, 45, 67/72, 64, 79, 79R, 80, Curle 15 and Curle ?21). There appear to be more Dr 33 cups than Dr 27 and more Dr 31/31R than Dr 18/ 31–18/31R bowls, but this was not quantified. The Dr 64 beaker and several globular jars (including plain examples, one with moulded and one with cut-glass decoration) are the less common forms, but are not unexpected in an ‘urban’ context. A small quantity of pre-Flavian material was identified, although the majority of the Southern Gaulish material is Flavian or later. The latest vessel from Southern Gaul is from Montans and probably no later than the early 2nd century. Other early 2ndcentury products are from Les Martres-de-Veyre and include some fine decorated bowls and plain forms including at least one form 15/17. The bulk of the assemblage is from Lezoux, with the greater proportion of this material probably dating to the mid and later Antonine period. Some of the latest vessels to reach the site are from Eastern Gaul and include forms 45, 32 and 36, although there are probably only a handful of these vessels. No decorated or stamped vessels from Eastern Gaul were identified. At least one vessel of 3rd-century date was noted. In general terms the trends observed here agree with those observed for the material excavated in 1990 to the south (Dickinson 2001), although later forms (45, 79R, 80) which were absent in that earlier assemblage were noted here. Trade Some 35% of the stratified pottery was supplied by non-local producers (Tables 5–6). Of this imported material, BB1 was the only traded ware reaching the site in any quantity (30% by count) with all the other sources represented by much smaller quantities. At the 1990 site the proportion of BB1 in the assemblage as a whole was 37% (Evans 2001, table 3) and at Cannards Grave 41% (Laidlaw 1997), and the figures from both sites contrast markedly with the representation of BB1 at Ilchester of 50–60% (Leach 1982, 142–3) and at Catsgore of 70% (Leech 1982, 159; Leach 1984, 25). This fall off in the proportion of BB1 is thought to reflect Durotrigan ethnicity and to be explained by Shepton Mallet’s location at a civitas boundary (Evans 2001, 158). By mapping these W 32 distributions, Allen and Fulford (1996) have also suggested that the main route taken by the BB1 supply trains from Ilchester was down river to the Severn rather than north up the Fosse Way, and it may be that the distribution of BB1 owes as much to the choice of communications routes as to the preference of the inhabitants of a civitas. The analysis of the 1990 assemblage showed that Severn Valley wares formed a surprisingly low proportion of the collection (Evans 2001, 126). Not only should the fall off in BB1 be offset by an increase in Severn Valley ware to fit the theory that the pottery is an indication of civitas groups, but Shepton Mallet as a producer of Severn Valley ware should have had high local use. However, the 1996 pottery gave a similar figure, 9%, for the proportion of Severn Valley ware to the 4% recorded in 1990. There was a sharp difference in the dominance of reduced over oxidised wares at both sites, in comparison with sites to the north in the Severn Valley ware heartland. In the 1996 collection, greywares (including BB1 and Savernake ware) represented 82% as opposed to oxidised wares which represent only 5% by count. Evans has suggested that in ceramic terms Shepton Mallet can be best paralleled by sites north of the Mendips from the Chew Valley to Bath. An interesting link with sites across the Bristol Channel in Wales was also noted (Evans 2001, 159, 160). Study of the form/fabric occurrence in the 1996 collection and the proportion of different vessel classes in different fabrics showed that BB1 jars and bowls occurred in roughly equal proportions, while tankards occurred in Shepton Mallet Severn Valley ware along with the occasional bowl form. There was an obvious lack of locally produced forms in the assemblage. From this it can be inferred that the locally produced products were traded further afield, whilst there was a heavy reliance on the marketing of BB1/local BBC for utilitarian forms. The assemblage from Cannards Grave had a higher proportion of BB1 that at the 1990 site, but this was made up of a limited range of utilitarian forms (Laidlaw 1997). The pottery as a whole suggested that the local inhabitants’ access to pottery from non-local sources was rather restricted. The dominant supply of BB1 appeared to have been unaffected by changes in the later centuries. This is reflected in the continued reliance on BB1 pottery as opposed to the New Forest and Oxfordshire products of the 3rd and 4th centuries. Even within the earlier phases of occupation early local products such as Savernake ware from Wiltshire did not prevail in any noticeable quantity. There were no New Forest products and only a small quantity of Oxfordshire products in the key group assemblages. As was usually the case in the Province, imported wares were dominated by Central Gaulish samian and a small quantity of amphorae. FOSSE LANE 1996: SPECIALIST REPORTS Status/function Overall several forms were present through the life of the settlement including bowls (B20s), jars, jar/ cook pots, beakers and the occasional dish and tankard form. The extent of residuality makes it difficult to assess the chronological significance in the forms present. With the exception of the Central Gaulish samian, very little fineware material was recovered from the key groups. This may be an indication that the area excavated in 1996 was of a lower level of status to that excavated in 1990 or may simply reflect the type of contexts chosen. In general, looking at the 1990 and 1996 material together, a general lower status can be suggested than in comparison to the civitas capital at Ilchester (Leach 1982), but both collections were of markedly higher status than at Cannards Grave (Laidlaw 1997). Within the material looked at in detail in 1996, few New Forest and Oxfordshire region fine wares were found. The only late finewares were two sherds of possibly residual Gaulish glazed wares. This contrasts with the slight increase of fine ware proportions seen in 1990 in the later periods (Evans 2001, 157). The range of forms at the Tesco site also seems smaller than at the 1990 site. Analysis of vessel class demonstrated the high proportion of jar forms compared to other forms such as bowls and dishes, and this was particularly the case from Period 4 onwards. This high proportion is likely to reflect a limited range of forms. Generally the samian trends observed in 1996 agreed with those presented by Brenda Dickinson for material excavated in 1990 (Dickinson 2001) and was comparable with assemblages from Ilchester (Leach 1982 and 1994b). All the samian from Area A was of Central Gaulish origin (predominantly Lezoux or Les Martresde-Veyre), whilst within Area B, particularly F247, a small amount of South Gaulish samian was observed. The majority of samian recovered from this area was of Central Gaulish origin. However, nearly all the decorated samian was from Southern Gaul (La Graufesenque). recovered from the midden F247 is significantly heavier (31g) and may represent a primary rubbish deposit. The Period 4 material had weights of 8g from a floor, 13 and 8g from hearths and ovens and 12g from a pit, all suggesting secondary deposition. A few pieces of coarseware and samian, particularly bowls, demonstrated evidence of prolonged use, with lead rivets in situ or drilled holes for rivets apparent. The same was true in the 1990 collection with the proportion of material so treated higher than at Ilchester (Evans 2001, 162). While the pottery in general was in a fragmentary and abraded condition, much of the samian was relatively unabraded although some was heavily burnt. The condition of the pottery in general is likely to be an indication of the degree of disturbance on site through time with material being re-excavated to be redeposited in the range of below ground intrusions observed. The condition of the samian presumably reflects its durability and its loss and burial in the earlier centuries of occupation. THE WORKED STONE Fiona Roe Objects The objects reported on here amount to several pieces from a small millstone, six quern fragments, two whetstones, a rubber and a small dish or mortar; none is illustrated. Only three of these finds are from stratified contexts. However the materials utilised, and particularly the Mendip Old Red Sandstone, all had long periods of use so that exact phasing for them may not be of any great significance. Although no stone artefacts were recorded in contexts earlier than the late 2nd century, the likelihood is that some could have been current earlier than this. A delay may often have occurred between the period of use of a stone object and its final burial as discarded rubbish, since pieces of stone left lying on the ground surface would often survive well when abandoned to the elements. Taphonomy Millstone Variations in sherd size in different groups allows consideration of the level of residuality and of different kinds of artefact deposition around the site. In Period 2, assemblages with larger mean sherd weights were found in ditch deposits (10g and 13g) than in material from Structure 1 (5g and 4g). These latter are likely to have been the remnants of secondary rubbish which had then been incorporated in floor surfaces. The average weight of sherds from F133 of 6g indicates that the material had been redeposited in the spread. In Period 3 the mean average sherd weight of pottery from ditches is slightly higher at 13g than in earlier phases but must nevertheless reflect its arrival in the ditch after deposition elsewhere. Material Fragments of a small millstone, no 1, made from Upper Greensand, were reused as cobbling in the open yard F112. Roman millstones appear to come in two size ranges. Many were up to a metre or more in diameter, while a smaller group varies from between 540 and 630mm in diameter (King 1988). This, and a second millstone from the 1990 excavation, fit within the smaller size range, with diameters of c. 580mm for the 1996 object and 540mm for a complete lower stone from the 1990 excavations (Roe 2001a, 235). The pieces of Upper Greensand are now weathered, but it can be seen that the grinding surface was coarsely grooved. Three further unstratified greensand fragments, no 2, may belong to the same millstone. W 33 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 The greensand probably came from the Pen Pits, a quarrying area some eleven miles (17.7km) from the site (Pitt Rivers 1884, 6). Greensand from the Pen Pits was used quite frequently for querns on Roman sites in the area (Roe 2001a), despite the nearby availability of Mendip Old Red Sandstone. At the time of writing further examples of Roman millstones utilising Pen Pits greensand were not known but have since been recorded. Greensand, though unspecified, was used for Roman millstones at Halstock, Dorset (Lucas 1993, 96, nos 4–6), and at Figheldean, Wilts (Mepham 1993, 36). 2000, 263). The fragment came from yard F110. Whetstones Both of the whetstones, nos 11 and 12, apparently result from the casual use of broken slabs of building stone. They are made from Old Red Sandstone, but a dark red, micaceous variety, different from any used for querns, and with a separate source, probably Prior’s Hill above Wells (Green and Welch 1965, 13), where there are old quarries some five miles (8km) from the site. Another six Old Red Sandstone whetstones came from the 1990 excavations (Roe 2001b). Querns Other stone artefacts Old Red Sandstone from the Mendips was the most frequently used corn-grinding material at Fosse Lane, as might be expected with a good source for stone at Beacon Hill only 2.25 miles (3.8 km) from the site (Green and Welch 1965, 13; Leach 1993, 139). There are four querns of probable Beacon Hill sandstone from the 1996 excavations, and another six of Old Red Sandstone from the 1990 excavations, these latter examples dating from Period 2 (Roe 2001a). The 1996 finds of querns do not appear in the archaeological record until the third century (Period 4). One, no 8, which is unstratified, is a complete lower stone of the Roman disc type, 430mm in diameter and weighing 20.8kg. This weight, for a comparatively slim disc quern, suggests that anything larger, such as the millstone with a diameter of 580mm, no 1, would have needed mechanical means for turning the upper stone. One quern fragment, no 5, is made unexpectedly of Upper Old Red Sandstone from the Forest of Dean/ Wye Valley area, about 38 miles (61km) from the site. The fragment was part of a quern of Roman disc type from an undated context, the rubble from Room 2 in Structure 7, which suggests that it may not have been in use before the late fourth century. Imported Old Red Sandstone could not exactly have been needed at Fosse Lane, so its presence is something of an enigma. The sixth quern fragment, no 6, is made from an altogether different type of stone, a silicified Jurassic sandstone or roach from the Harptree Beds. This is a sandstone with casts of fossil shells, and the gaps where they have eroded out would have provided a good grinding surface. The nearest source for this type of stone is at Oakhill (Green and Welch 1965, 109), just off the Fosse Way, a short distance from Beacon Hill, and so only 2.5 miles (4km) from the site. The quern itself is a somewhat thick beehive type with affinities to Iron Age varieties of quern, and this typology is reflected in the other finds made from the same sandstone. Beehive querns utilising the Harptree Beds roach occurred at both the Meare and Glastonbury Lake villages (Roe 1995, 166) and there were also a number of similar finds from Iron Age (Middle Cadbury) contexts at South Cadbury (Roe W 34 Mendip Old Red Sandstone was also used for a small rubber/hammerstone, no 9. A small dish or mortar, no 10, only some 110–120mm in diameter, was made from another material that was available close to the site, shelly limestone from Doulting. More recent quarries in the Doulting stone are sited some 2.1 miles (3.4km) from Fosse Lane. The same stone was also used for details on Roman buildings at Fosse Lane (Leach 2001), and also for a child’s coffin, nos 15 and 16 (Figs 22 and 23). The small vessel probably represents the contingent use of a spare piece of building stone. Building stone It was possible to use stone extensively for building at Fosse Lane, as good materials were available either on or within a few miles of the site, and the local Lias in particular was utilised in various capacities. None of the catalogued items, as it happens, is dated earlier than the 4th century. By then (Period 5), there is some evidence, no 13, for the use of ashlar masonry, made from the local Jurassic limestone which was brought in from around 2–3 miles (3.2–4.8km) south east of the site. The remaining pieces are all from the undated late or post-Roman period. Bath stone, which was conveniently available up the Fosse Way, from about 16 miles (25.7km) away, was used for a column, no 14, which, with a diameter of about 170mm, was fairly slender. Bath stone was also used for an altar from the 1990 excavations, and seems to have been employed generally in the area for any stonework needing detailed shaping. At King’s Weston, for instance, the villa was entered through a portico with six columns of Bath stone (Branigan 1969, 19), and the same stone was also used for architectural features at Camerton (Wedlake, 1958, 49) and at Gatcombe (Horwell 1977, 103). The final item to be discussed is a stone coffin, used for the burial of a small child, nos 15 and 16. A second coffin, made from local Downside stone, was found in the 1990 excavations (Leach 2001, 253). The lid FOSSE LANE 1996: SPECIALIST REPORTS of the child’s coffin is made from Doulting stone, which in places outcrops adjacent to the Downside stone. In general these Roman coffins tend to occur in areas of suitable stone (Woodward 1993, 227), and particularly where Jurassic limestones were available, so that finds from Fosse Lane fit well into the known pattern. Discussion Sites in and around the Mendips were well provided, within a dozen miles or so, both with good local grinding stone and with building stone, and this assemblage from Fosse Lane shows how the local rocks were put to good use. However, the inhabitants of the small town were not deterred from using imported materials as and when the opportunity arose. The evidence from the 1996 excavations is in some ways misleading, since although the 1996 whetstones and the small mortar are made of local stone, the 1990 excavations produced six well-travelled whetstones and a mortar of Kentish Rag, and also another mortar of Purbeck marble (Roe 2001c, 177). The Bath freestone was probably used on many Somerset sites for shaped or carved architectural details, so its use at Shepton Mallet is of no surprise, but the presence of a quern of Upper Old Red Sandstone from the Forest of Dean is less easy to explain. One possibility is the link with the exportation of Severn Valley Wares, as part of the widespread trading opportunities that were available during the Roman period. Catalogue 1 2 3 4 5 6 Part of small millstone, diam c. 580mm, th c. 66–73.5mm, now weathered; grinding surface was coarsely grooved, other surface pecked to shape; three large fitting pieces. Upper Greensand, probably from the Pen Pits. SF 386, F112 Three fragments probably of the same millstone. Upper Greensand, probably from the Pen Pits. SF 406 and SF 407, Area B, SF 408, Area A Fragment upper stone rotary quern, Roman disc type, diam c. 410mm, th at rim 48.5mm, pecked neatly to shape, traces of grooving on grinding surface. Mendip Old Red Sandstone conglomerate, probably from Beacon Hill. SF 404, ditch F230 Half lower stone of rotary quern, diam c. 355mm th at edge 36mm, in centre 77.5mm; pecked grinding surface. Mendip Old Red Sandstone pebbly sandstone, probably from Beacon Hill, SF 385, F113, Structure 2 Fragment of rotary quern, Roman disc type, probably upper stone; top surface and edge are pecked to shape, th at edge 43mm. Upper Old Red Sandstone from Forest of Dean. SF 293, layer 2025, Structure 7 destruction Half lower stone of rotary quern, diam c. 370mm, max th at edge 74mm, in centre 110mm, grinding surface worn smooth, trace of socket for spindle. Jurassic sandstone or ‘roach’ from the Harptree Beds. SF 402, F110 7 Half upper stone of small rotary quern, diam c. 365mm, max th 60.5mm; fairly smooth grinding surface, now weathered, trace of slot in side for handle. Mendip Old Red Sandstone, pebbly sandstone probably from Beacon Hill. SF 403, Area B 8 Complete lower stone large rotary quern, Roman disc type, diam c. 430x445mm, th at rim 48mm; grinding surface mostly worn smooth, underside more roughly shaped, central hole min diam 43mm. Mendip Old Red Sandstone, probably from Beacon Hill. SF 405, surface find Area B 9 Small rubber/hammerstone, utilising pebble, 78x54x25mm. Mendip Old Red Sandstone. SF 119, surface find Building 6 10 Fragment from small dish or mortar; diam c. 110–112mm, depth 45mm. Doulting stone. SF 121, surface find Building 6 11 Whetstone, utilising rectangular slab of sandstone, wear on one flat surface, iron stained; 100x66x20mm. Dark red micaceous Mendip Old Red Sandstone, could be from Prior’s Hill. SF 302, layer 2030, under F228 12 Whetstone, utilising broken triangular slab, worn smooth on four surfaces; 126x96x31mm. Dark red micaceous Mendip Old Red Sandstone, could be from Prior’s Hill. SF 304, 2095, rakeout from F264 13 Three large worked fragments, burnt; two fitting, with flat surface, the third with a flat surface and a carved area; re-used building stone? Oolitic limestone, coarser grained than no 14 and with some large pieces of shell; local Jurassic. SF 292, layer 2029, in Structure 7 14 Two fragments from column, diam c. 170mm. Oolitic limestone with a few shell fragments; Bath stone. SF 120, Structure 6 demolition 15 Coffin lid fragment, th 98mm. Shelly limestone, Doulting. SF342, grave F208 (Fig. 22) 16 Stone box coffin. SF 343, grave F208 (Fig. 23). In addition a further fragment of rotary quern, SF 122, was found in Building 7 destruction. This was not seen by the author. THE CHARRED PLANT REMAINS Julie Jones Introduction and methods Extensive environmental sampling was carried out as part of the excavation with samples recovered from ditches, pits and other features associated with activities in the settlement. Most of the samples were 30 litres and these were flotation sieved to a 250 micron float and 500 micron residue and allowed to W 35 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 dry. Thirty-two flots which contained charred plant macrofossils and charcoal were assessed by Vanessa Straker (Leach and Ellis 1996). In view of the low concentration of material in most of the samples it was recommended that only ten were analysed further. These were examined under a low-powered microscope. Preservation was variable with much of the cereal grain being in poor condition, in contrast to the chaff and weed seeds which were generally well preserved. The results are shown in Table 7. Plant nomenclature and habitat information follows Stace (1991) and cereal grain and chaff determinations are based on Jacomet (1987). Results Area A From spread F133, context 1002 (sample ES14) Substantial quantities of barley grain were recovered, including tail grain (smaller than 5mm) with some examples well enough preserved to confirm the presence of hulled barley. Only 8 wheat grains were present with a few silicified wheat/barley awns and a single charred oat awn. Weed seeds were also limited to a few grasses, sedges and docks. From ditch F134, context 1061 (sample ES29) Cereal remains included 53 wheat and 11 barley grains. with accompanying wheat glume bases and spikelet forks. Glume bases of Triticum spelta confirmed the presence of spelt wheat but most of the glumes and spikelet forks were too abraded to identify further. There was no accompanying chaff with the oat grains to confirm whether these would have been the domesticated variety or whether the oat was present as a crop weed. Charred seeds include typical arable weeds including orache, blackbindweed and scentless mayweed. The Brassica/ Sinapis seed may be a crop weed but could also represent the cultivation of plants of the cabbage family such as mustard, rape or cole. Hazelnut fragments from this deposit may represent food debris or could have been collected with firewood. From burnt area F143, context 1022, within Structure 1 (sample ES4) This sample only contained 8 wheat grains and a single barley grain. No chaff was present and the weed assemblage includes a few arable weeds, with some more typical grassland taxa including clover/medick and ribwort plantain. Area B From spread of pottery and burnt material, context 2017 (sample ES7) This sample produced a few wheat grains with three W 36 glume bases and a low concentration of weed seeds. From ditch F236, context 2038 (sample ES23) This contained four wheat grains, wheat glume bases and spikelet forks, three oat grains and nine weed seeds, mostly grasses. From pit F243, context 2650 (sample ES10) Numbers of grains were low but chaff included spelt and wheat glume bases with two cereal sprouts. Seeds were limited to single examples of arable weeds with one hazel nut fragment. From pit F243, context 2041 (sample ES9) Limited numbers of both wheat and barley grains with a few oats were present. Amongst the wheat chaff some well-preserved glume bases again confirm this as spelt. A range of weed seeds included some typical arable weeds such as orache, black bindweed and brome. From ditches F245, context 2047, and F246, context 2046 (samples ES6 and ES11) Both ditch fills produced a few examples of wheat, barley, a single oat grain and some wheat chaff. Slightly higher numbers of seeds included dock, orache, bartsia/eyebright, cleavers, stinking chamomile as well as species typical of grassy places, including clover/medick, vetches and grasses. From oven F249, context 2053 (sample ES2) This sample contained limited cereal and weed remains. Discussion The range of cultivated plants and their associated weed assemblages recovered throughout the different phases of activity at Fosse Lane are very similar, with both wheat and barley grains present throughout, wheat on the whole being predominant. The single exception is from Phase 2 where 243 barley grains and only eight wheat grains were found amongst the spread of burnt clay associated with F133. No cereal chaff was found with the grains in this sample apart from a few silicified wheat/barley awns and a single oat awn. There were also very few weed seeds. This is therefore likely to represent the remains of a cleaned crop which had become burnt. None of the grains showed signs of sprouting, although a few detached cereal sprouts were recovered from other samples. On the whole the cereal grain was poorly preserved with much of the surface layers lost through burning and many grains were encrusted with sediment. Although all the wheat was identified as Triticum sp, many of the wheat grains were of the straight parallel sided form, typical of spelt wheat and the presence of FOSSE LANE 1996: SPECIALIST REPORTS spelt glume bases confirms the presence of this crop. Some grains had a more rounded dorsal surface with steeply angled embryo characteristic of bread wheat (Triticum aestivum), although no free-threshing wheat chaff was recovered. Triticum spelta is the wheat most commonly found in the Romano-British period. The chaff recovered represents the debris from the threshing of spelt, which does not have free-threshing grains and would have required parching to render the glumes brittle to free the grain. The soils around Shepton Mallet are shallow, welldrained, brashy, calcareous clayey soils over limestone with some deeper calcareous clayey soils. Much of the land use today is cereals and short term grassland with stock rearing and dairying. The environs of the Romano-British settlement would therefore have been suitable for local cultivation. Spelt, which is the most commonly occurring crop recovered here, is a hardy cereal, ideal for winter sowing, which thrives on heavy soils. Weeds of autumn-sown grain crops (the Secaletea) such as wheat, germinate in the autumn and grow rapidly in the spring along with the crop and are then harvested with the cereals. This group of weeds includes examples recovered in many of the samples, including stinking chamomile, cleavers and some species of vetch. Stinking chamomile also suggests tillage of heavier soils. Other weeds found include black bindweed and scentless mayweed. The oats could also have occurred as weeds, as it was not possible to tell whether these were the domesticated or wild variety, as no oat chaff, necessary to confirm this, was found. The brome is also likely to have occurred as an arable weed but may have been collected and utilised with the crop. It has been described as a famine food by some authors. Jones (1981) remarks that even in this century in Denmark, seeds of brome were collected and ground to flour as a famine food when the rye crop failed. The fact that fine chaff and weed seeds have been recovered from the Roman town does suggest that the crops were being brought into the town partly processed, the spelt wheat perhaps in spikelet form, with the final cleaning and removal of grain from chaff taking place in the settlement. However, apart from the find of cleaned barley from F133, charred crop and weed remains are scarce suggesting that if any large-scale processing was taking place in the settlement, this was in a different part of the town. There are a number of other Romano-British sites where plant macrofossil analysis has been carried out in the local region. Six samples were recovered following the excavations at Fosse Lane in 1990 (Straker 2001). These contained a similar range of cultivated and wild plants to those recovered in 1996. In the different phases examined in 1990 both wheat and barley grains were present with spelt glume bases and spikelet forks confirming the presence of spelt. Only a limited weed assemblage was however recovered. A burnt layer beneath a floor in a late 2nd/ 3rd century building and a hearth inside a 3rd-century building were the most productive deposits. The hearth was composed of 70% cereal chaff, mostly spelt spikelet forks and glume bases, with 21% grain, largely wheat and only 9% weeds and was interpreted as cereal chaff being used as tinder. Deposits of chaff are frequently interpreted as fuel and have been recorded locally at Roman sites at Catsgore (Hillman 1982) and Kenn Moor (Jones 2000) from corn drying ovens. The layer under the floor (6044) was composed of 69% grain, most probably spelt, with 29% largely spelt glumes and only 2% weeds. It is not known how the layer of burnt grain became incorporated into the floor deposit, although this building is interpreted as a barn, but there was no evidence that it had burnt down. Other Romano-British sites in the locality include Catsgore (Hillman 1982), three sites at Ilchester (Murphy 1982) as well as more recent excavations at Ilchester (Stevens 1999). From the latter, a rich assemblage of charred plant remains dominated by grains and chaff of spelt, with only sparse barley and free-threshing wheat was recovered. A range of arable weeds from a number of habitats indicate the cultivation of a range of soil types. Stevens also suggests the cereals were stored at different times either fully or partly processed for use by individual households in the Roman town. THE ANIMAL BONE Umberto Albarella and Andy Hammon Introduction and methods The excavated material would seem to represent a typical Roman urban assemblage. Cattle dominate, followed by sheep and then pig. Horse, dog and domestic chicken were present, but appeared in low numbers. Fragmentation and butchery suggest that most bones derive from food refuse. Due to its small size it was considered inconvenient to divide the assemblage into seven chronological groups. Consequently, Phases 1, 2, 3 and 4, and Phases 5 and 6 have been grouped together. These will be referred to as ‘Earlier Roman’ and ‘Later Roman/post-Roman’ respectively. Phase 7 has not been considered because it was likely to be affected by residuality and contamination problems. The majority of animal bone was recovered by hand collection during the excavation. No bulk sieving specifically for the recovery of animal bone was undertaken. A very small amount of animal bone came from samples taken for flotation. The quantity of animal bone recovered from W 37 PHASE 2 1002 2041 ES14 ES9 42 70 ContextNo Sample No Flotsize(ml) CHARRED PLANT REMAINS Cereals Grain Triticum sp cf Triticum sp Triticum sp (tail grain) Hordeum sp Hordeum sp (hulled) Hordeum sp (tail grain) Hordeum sp (hulled/tail grain) cf. Hordeum sp Avenasp cf. Avena sp Cereal indet Total Chaff Triticum spelta (glume base) Triticum sp (glume base) Triticum sp (spikelet fork) Triticum sp (rachis internode base) Triticum/Hordeum sp (awns silicified) Avena sp (awn) Cereal sprout (coleoptile) Total Weeds RANUNCULACEAE Ranunculus spp (inside of) Ranunculusfiammula L. POLYGONACEAE Fallopia convolvulus (L.)A.Love Rumexsp Wheat 8 Barley 180 45 12 6 Oat 13 Wheat Wheat/Barley few 1 2047 ES11 25 PHASE 3 1022 1061 ES4 ES29 150 550 PHASE 4 2017 ES7 45 PHASE 5 2038 2053 ES23 E52 40 23 11 7 1 6 2 4 5 11 7 13 2 2 2 1 1 4 3 3 28 292 9 43 19 49 5 9 2 2 1 6 20 6 20 2 11 6 4 15 36 57 10 1 28 15 10 6 1 5 HABITAT 2650 ES10 4 2 1 4 13 3 33 116 5 18 6 13 3 7 16 2 3 2 6 3 9 13 25 5 2 1 1 Buttercup Lesser Spearwort Black-bindweed Dock 2046 ES6 28 1 76 11 13 0 104 3 26 5 1 3 1 2 1 4 2 45 7 DG M 1 19 4 2 12 4 4 1 3 1 CD DG SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 W 38 TABLE 7: CHARRED PLANT MACROFOSSILS: OCCURRENCE ContextNo Sample No BRASSICACEAE Brassica/Sinapis spp FABACEAE Lathyrus/Vicia spp Lathyrus/Vicia/Pisum spp Thifohum/Medicago sPp Vicia c.f. tetrasperma (L.)Schreber APICACEAE Apium graveolens L. PLANTAGINACEAE Plantago lanceolata L. SCROPHULARIACEAE Odontites/Euphrasia spp RUBIACEAE Galium aparine L. ASTERACEAE Anthemis cotula L. Poaceaeindet Indet Total OTHER REMAINS Buds Charcoal (fragments >2mm) 2046 ES6 2047 ES11 PHASE 3 1022 1061 ES4 ES29 Mustard/Rape/Cole Vetch Vetch/Pea Clover/Medick Smooth Tare PHASE 5 2038 2053 ES23 ES2 HABITAT 2650 ES10 1 1 1 Wild Celery 1 Ribwort Plantain 2 CD# 1 1 2 4 3 1 1 4 1 3 1 1 Cleavers 2 Stinking Chamomile Knapweed Oxeye Daisy Scentless mayweed 2 1 2 3 1 1 2 CHSo 1 CDd 1 DG G-rich soils CD 1 Rush 1 4 2 1 GMRw 1 6 1 1 1 2 3 2 16 1 8 8 16 44 <20 100+ 50+ G CD 2 1 1 DG DG# DG G ws Bartsia/Eyebright Smooth/soft/rye Brome Crested Dog’s-tail Heath-grass Meadow-grass/ Cat’ s-tail Grass PHASE 4 2017 ES7 2 DG/CD 1 G Ew sandy.peaty G 5 1 G 3 18 1 2 26 2 1 54 13 9 31 8 3 <40 1 <200 300+ 50+ <50 <10 50+ 4 6 W 39 Habitats: C: Cultivated/Arable. D: Disturbed. E: Heath/Moor. G: Grassland. H:Hedgerow. M: Marsh. R: Rivers/streams. S: Scrub. W: Woodland. d: dry soils. n: nitrogen rich soils. o: open habitats. w: wet soils. #: cultivated plant/of economic importance FOSSE LANE 1996: SPECIALIST REPORTS Centaurea sp Leucanthemum vulgare Lam Tripleurospermum inodorum (L.)Schultz. Bip JUNCACEAE Juncus sp POACEAE Bromus racemosus hordaceus/secalinus Cynosurus cristatus L. Danthonia decumbens (L.)DC Poa/Phleum spp PHASE 2 1002 2041 ES14 ES9 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 the samples was very small and prior to the assessment it had been incorporated into the hand-collected material. Consequently, the hand and sieve collected material had to be treated as a single assemblage. The samples taken for fine sieving and flotation varied between 20–30 litres, unless the context from which they derived was smaller. In such cases the whole context was sampled. The coarse residue was washed over an approximately 2mm mesh. The presence of residual material cannot be estimated from the animal bone alone; a figure of about 40% was suggested by the pottery analysis for the later group. The later Roman/post-Roman material has, therefore, to be treated with caution. Dogs had gnawed 53% of the ‘countable’ post-cranial skeletal elements. This would suggest that a fairly high proportion of the animal bones recovered from Fosse Lane may not result from primary deposition, but secondary deposition caused by scavenging dogs. Consequently spatial analyses need to be treated with caution, although some groups, such as the concentration of cattle mandibles in ditch F236, context 2039, seem unaffected. In general bone preservation (cortical integrity) was fairly good with little exfoliation. However, some contexts were not so well preserved, with poor cortical integrity and extensive abrasion. There was little variation in colour and it followed no discernable pattern. The animal bones were heavily fragmented and this is indicative of material deriving from butchery and kitchen waste. The mammal bones were recorded following a modified version of the method described in Davis (1992) and Albarella and Davis (1994). This system considers a selected suite of anatomical elements as ‘countable’; it does not include every bone fragment that is identifiable. Mandibular fragments were considered to be ageable when there were two teeth present with recognizable wear. Von den Driesch (1976) defines the majority of measurements used. All pig measurements follow the definitions of Payne and Bull (1988). The differentiation of sheep and goat was attempted on the following elements: deciduous lower premolars (dP3 and dP4); humerus; metacarpal; tibia; astragalus; calcaneum; and metatarsal. The criteria defined by Boessneck (1969) were used for all elements except the teeth (Payne 1985) and the tibia (Kratochvil 1969). The Chicken/Guinea Fowl/Pheasant (Gallus/Numida/ Phasianus) distinction was attempted on the following elements: articular end of the scapula, shaft of the carpometacarpus, proximal end of the femur and shaft of the tarsometatarsus. Mandibular teeth, both in situ and loose, were aged using wear patterns. The system recommended by Grant (1982) was used for cattle and pig, whereas the wear of sheep/goat teeth was recorded according to Payne (1973; 1987). W 40 Occurrence and frequency of species The following mammal and bird species were identified: Cattle (Bos taurus), Sheep (Ovis aries), Sheep/goat (Ovis/Capra), Pig (Sus scrofa), Equid (Equus sp.), Dog (Canis familiaris), Red Deer (Cervus elaphus), Domestic Chicken (Gallus gallus), Chicken/ Guinea Fowl/Pheasant (Gallus/Numida/Phasianus) and Rook/Crow (Corvus frugilegus/corone). Two bird bones have been provisionally identified as the distal radius of a large corvid, most likely a Raven (Corvus corax) (context 2009) and the proximal radius of a Woodcock (Scolopax rusticola) or a related species (context 2050). No small mammals or fish bones were recovered, but their absence may be due to a recovery bias. As is well known the bones of these small animals can easily be overlooked during excavation. In 1990 no fish bone and only one specimen of a small mammal – this was a Water Vole (Arvicola terrestris) – were found (Pinter Bellows 2001). Most of the caprine remains could not be identified at species level and had to be recorded as sheep/goat (Ovis/Capra). However, it may be inferred that the majority was in fact sheep (Ovis aries), as 29 specimens belonged to this species and none could be identified as goat (Capra hircus) – including horncores (Tables 8 and 9). However, goat must have been present on site as nine goat specimens – compared to 56 sheep and 218 sheep/goat – were recorded from TABLE 8: NUMBERS OF ANIMAL BONE AND TEETH NISP (INCLUDING HAND AND SIEVE COLLECTION) Species Earlier RB Cattle (Bos taurus) 319 Sheep/Goat (Ovis/Capra) 173 Sheep (Ovis aries) 22 Pig (Sus scrofa) 36 Dog (Canis familiaris) 7 Equid (Equus sp.) 10 Red Deer (Cervus elaphus) 1 Chicken (Gallus gallus) 1 Chicken/Guinea Fowl/ Pheasant (Gallus/Numida/ Phasianus) 3 Rook/Crow (Corvus frugilegus/corone) Total 572 Later/ post-RB 147 55 7 25 6 11 0 1 Total 466 228 29 61 13 21 1 2 1 4 1 254 1 826 TABLE 9: ANIMAL BONE NUMBERS (NISP) AND PERCENTAGES OF THE THREE MAJOR DOMESTICATES Earlier RB Species N % Cattle (Bos taurus) 319 60 Sheep (Ovis aries)* 195 36 Pig (Sus scrofa) 22 4 Total 536 Later/ post-RB N % 147 63 62 27 25 10 234 Total 466 257 47 770 * combines both Sheep/Goat (Ovis/Capra) and Sheep (Ovis aries) FOSSE LANE 1996: SPECIALIST REPORTS the 1990 excavations (Pinter-Bellows 2001). Goat, although at its commonest in the Romano-British period, is never well represented in British assemblages (Albarella and Davis 1996; Dobney et al nd). Four different methods were employed to quantify the relative importance of species present (Fig. 34) They were as follows: 1 2 3 4 NISP: Number of identified specimens. A simple count of ‘counted’ bone fragments by skeletal element and the total number of fragments per species. MNI: Minimum number of individuals. The most frequently occurring element denotes the MNI for that species. MNI was calculated as follows: incisors and phalanges divided by 8; deciduous and permanent premolars by 6; M1/ 2 by 4; all other elements by 2, except metapodials. The metacarpal and the metatarsal were calculated in the following way: metacarpal (MC1 + MC2/2 + MC1/2 + MP2/4)/2; metatarsal (MT1 + MT2/2 + MT1/2 + MP2/4)/21 Harcourt: Live weight ratios based on Harcourt (1979), calculated on the meat contribution of sheep and goats being equal to 1, pigs being equal to 1.5 and cattle being equal to 10. Manching: Live weight ratios based on the midpoint values taken from the Manching data set by Dobney et al (nd). Cattle live weight was considered to be 275 kg, sheep live weight 37.5 kg and pig live weight 85 kg. Therefore, 7.3 sheep equals a single cow and 2.3 sheep equal 1 pig. Fig. 34 Relative importance of the main domesticates. Upper: Earlier RB; Lower Later/post RB All four methods were applied to cattle, sheep and pig, whereas horse, all other mammal and bird species were quantified using NISP only (due to their lower frequency). All four methods clearly demonstrate that cattle were the most abundant species and economically the most important animal, followed by sheep and then pig. The frequency of the main taxa in the Earlier Roman and Later Roman/post-Roman periods is similar, although sheep seems slightly less common in the later period. The pattern described above occurs using all the methods with the exception of MNI for Later Roman/post-Roman. This shows cattle and sheep to be present in equal numbers, which probably results from the small size of the assemblage. Unsurprisingly, the live weight estimations show cattle to have occurred in far greater proportions for both periods. According to this system the contribution of beef to the meat supply would range between 84% and 94% per cent, that of mutton between 4% and 12% and of pork between 1% and 4% (Fig. 34). The predominance of cattle is quite common for a Romano-British assemblage and falls within the range of 59% cattle to 32% sheep expected on non-military Romanised sites (King 1984). The results discussed above are similar to those from the earlier excavation at Fosse Lane, although in that assemblage sheep were the most numerous species using NISP (PinterBellows 2001). Skeletal representation and butchery Skeletal representation has only been considered in detail for cattle and sheep, due to the small number of bones of all other species. In Figure 35 each cattle and sheep skeletal element has been shown as a percentage of the most commonly occurring element (according to MNI). Proportions of butchered bone and the type of butchery have been summarized in Figure 36. For both species and periods a similar pattern appears. This pattern generally reflects the taphonomic processes affecting the assemblage at Fosse Lane. Certain bones are more durable than others, therefore dense bones, such as the calcaneum and astragalus, survive better than less compact and late fusing bones, such as the two ends of the femur and the proximal humerus. For a similar reason teeth also survive well. Brain (1981) summarizes the postdepositional destruction of skeletal elements and the Fosse Lane cattle and sheep generally adhere to this pattern. In addition, the recovery bias favouring the larger bones will also lead to some elements being under represented. The skeletal representation of cattle and sheep at Fosse Lane does not indicate any specialized activities, such as, for example, bone/horn working on a commercial scale, and can entirely be explained on the basis of recovery and taphonomic effects. W 41 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 Fig. 35 Relative abundance of skeletal elements: above cattle %MNI, below sheep and sheep/goat %MNI Maltby (1981) has noted that the type of butchery in cattle for the Roman period varies depending on the nature of the site. Rural sites have prevalence for cut bone, whereas urban sites have a prevalence for chopped bone. At Fosse Lane (for both Earlier and Later Roman/post-Roman) the butchery on postcranial elements was divided as 69% chopped and 31% cut. As noted above, the high level of fragmentation is typical of kitchen and butchery waste. The fair number of butchery marks found on the bones also confirms this (Fig. 36). The deposit in ditch F236 (context 2039), dated to the Earlier Roman, contained twelve cattle mandibles and relatively large quantities of loose teeth. Five of the twelve mandibles had been chopped. These may derive from primary butchery waste, therefore suggesting that the primary dismemberment of freshly slaughtered animals may have been carried out on site. Three horse bones showed butchery evidence with three cut and one chop marks present. All occur on the lower limb bones (two first phalanges and one metatarsal) from context 2010. Skinning probably caused the cut marks. The first phalanx (context 2010) displaying both cut and chop marks is more difficult to interpret. The chop could have been made for a variety of reasons, including a rather crude way to separate the hide and the consumption of horse marrow either for dogs or people. Despite the W 42 assemblage being rather small, the skeletal representation (Fig. 35) and butchery (Fig. 36) is informative. Most body parts from the main domesticates were represented; this would suggest that animals were being butchered on site. Fig. 36 Post-cranial proportions of butchery for Earlier and Later periods combined (Sheep = sheep/ goat) Age at death The small numbers of ageable mandibles has meant that it has only been possible to construct a survivorship/mortality curve for the Earlier Roman cattle (Fig. 37). The fusion data generally support the FOSSE LANE 1996: SPECIALIST REPORTS Fig. 37 Earlier RB: cattle survivorship curve based on mandibles where two or more teeth are present Age stages (O’Connor 1988): Juvenile = M1 not in wear; Immature = M1 in wear; M2 not in wear; Sub-adult = M2 in wear, M3 not in wear; Adult - M3 in wear, not heavily worn; Elderly = M3 heavily work (j or beyond sensu Grant 1982 pattern suggested by the ageable mandibles. The Earlier Roman period saw the majority of cattle being kept till they were adult. This would suggest that they were primarily used for secondary products, such as dairy or traction. Animals would only being slaughtered once they had stopped producing milk or they would have become too weak to pull the plough. The fact that Roman written sources, such as Columella and Varro, hardly mention cow milk but at the same time emphasize the importance of sheep and goat dairy products (White 1970) probably suggests that cattle were mainly used as draught animals. The lack of very young individuals – normally associated with milk exploitation (Payne 1973; Legge 1981) – seems to support this suggestion. Because of the small size of this assemblage only very general information may be obtained from the sheep ageing data. Sheep, for both periods, were kept to at least one year and up to eight years (Payne’s stages D–H) before being slaughtered. Consequently, it is not known whether they were primarily utilized for milk, meat, wool or a combination. Only three very young ‘countable’ post-cranial elements were recorded, although they were not young enough to indicate on-site breeding. Only very few pig mandibles were present but fusion data suggest that two thirds of the post-cranial ‘countable’ pig bone was either unfused or fusing (seven of eleven elements). This would suggest that the majority of individuals were being killed while either immature or sub-adult. This is the optimum age for slaughter – the pig would have almost reached its maximum weight by this age. The number of cattle mandibles from the Earlier Roman period has allowed a comparison with other Roman sites (Fig. 38). The Fosse Lane survivorship curve is comparable with other sites at Lincoln (1st– 3rd centuries), Exeter (1st–3rd centuries), and Tanner Row, York (1st–3rd centuries) up to the sub-adult age Fig. 38 Earlier RB: inter-site comparison of cattle survivorship curves for Lincoln (Dobney et al 1996); Exeter (Maltby 1979) Phases R1-R7; Tanner Row (O’Connor 1988) Periods 3-7; for age stages see Fig. 37 caption stage. At which point it decreases slightly, falling just below the other curves. This may tentatively suggest a higher reliance on beef than at the other sites. Morphology and size Little metric data was collected from the Fosse Lane animal bones. This is due to the small size of the assemblage and the high degree of fragmentation. As a result it has not been possible to chart the changes in size and morphology over time and a greater emphasis has been placed on an inter-site comparison. Maltby (1981) states that the greatest lateral length (GLI) of the astragalus is the most commonly taken measurement in Roman cattle assemblages. Unfortunately, the Fosse Lane assemblage only produced three measurements (from both periods). Although, it has not been possible to calculate a mean based on so few measurements, it is possible to demonstrate that all are within the range suggested by the other sites (Table 10). Pinter-Bellows (2001) uses the distal breadth (Bd) of the sheep tibia to make an inter-site comparison. Again, it has not been possible to calculate a mean for this Fosse Lane assemblage (only five measurable specimens from both periods), but all are within the range suggested by the other sites (Table 10). Thus, all that can be said of the Fosse Lane cattle and sheep is that they were of a similar size to animals from other Roman sites. Mandibular third molar’s (M ) were the only 3 element from the second Fosse Lane assemblage to occur frequently enough to allow a reliable comparison (Fig. 39). Length and width were compared to specimens from sites on Ermine St in Cambridgeshire (Albarella 1998) and from Dodder Hill (Davis 1988). Albarella (1998) concluded that, taking into account regional and chronological variation, the greater size of the Ermine St specimens may be attributed to larger imported continental cattle making a greater contribution to the local genotype. The Fosse Lane specimens show greater variation, but are generally similar to the Dodder Hill cattle. This, therefore, supports Pinter-Bellows’ (2001, 298) W 43 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 TABLE 10 COMPARISON OF CATTLE AND SHEEP MEASUREMENTS FROM SELECTED ROMANO-BRITISH SITES Site No Mean Range (1/10 mm) (1/10 mm) Cattle astragalus (GLI) Carlisle, 73/74–100/1051 Corstopitum, Roman Fosse Lane, Roman2 Exeter, 300–400 West Stow, Roman Baylham House, 100-200 Alcester, Late Roman Shakenoak, Late Roman Winnall Down, Early Roman 16 9 18 18 4 10 30 454 16 Sheep and sheep/goat tibia (Bd) 31 Carlisle, 73/74–100/1051 Balksbury 1973, Roman 7 Frocester Court, 100-300 12 Exeter, 55-100 21 11 Fosse Lane, Roman2 Exeter 100-300 30 Alcester, 100-200 9 Exeter, 300-400 15 Winnall Down 8 Frocester Court, Late Roman 13 Shakenoak, Late Roman 26 Baylham House, 100-200 22 Alcester, Late Roman 59 West Stow, Roman 9 571 580 582 583 608 613 614 616 616 528-641 530-630 500-625 543-620 597-630 560-658 539-679 530-720 561-684 226 228 228 231 232 233 236 239 239 240 245 245 255 256 206-248 210-273 200-250 213-292 210-249 214-259 211-260 223-270 219-256 230-270 220-280 216-288 212-294 221-274 1 Stallibrass 1991 Pinter-Bellows 2001 All others Maltby 1981 2 Fig. 39 Inter-site comparison of Earlier RB cattle M3. A1(M) (Albarella 1997), Dodder Hill (Davis 1988) the second Fosse Lane assemblage seem to suggest that this may have not necessarily been the case. Throughout prehistoric and Roman periods it was more normal for cattle to be raised on low ground and sheep to be raised on higher ground (Grant 1984). One reason for this is that sheep in low-lying areas were prone to infestations of liver fluke (Dobney et al nd). THE HUMAN SKELETONS Stephanie PinterBellows Summary hypothesis that the Fosse Lane cattle were of a relatively small size. Summary and conclusions Because of the small nature of this assemblage it has only provided general information regarding the economy and activities of the Fosse Lane inhabitants. In general the Fosse Lane assemblage would seem to be very similar to that from the 1990 site (PinterBellows 2001). However, due to its small size it does not provide much additional evidence to support the earlier conclusions. Preservation varied, but the majority of bones were not badly abraded or exfoliated. This would suggest that bones were deposited reasonably quickly. However, as suggested by the high level of scavenger gnawing, most bones may derive from secondary deposition. Cattle dominate this assemblage, regardless of which method was used to quantify species abundance (MNI, NISP, Harcourt’s live weight ratios and the Manching live weight ratios). This differs slightly from the relative species abundance at the 1990 site (Pinter Bellows 2001), which has a greater number of sheep. It was concluded that this may suggest that Fosse Lane was not a heavily Romanised site (ibid). The findings of W 44 A total of ten individuals were recovered. The majority of the skeletons were found in two groupings. Stature and the relatively low rate of pathologies show this to be a healthy population sample. Material and methods A total of eight inhumations, together with a very small amount of human bone from other contexts, was recovered from across the site. Two of the contexts with human bone contained articulating foetal bones which could be shown by age and skeletal elements not to be part of the recognised inhumations and were counted in the total number of individuals. The skeletons were found in two groupings and a few isolated burials. These burials are only a small part of the community which existed at the time, a larger percentage were excavated in 1990. It must be kept in mind throughout this report that the relatively small number of skeletons means that the description of the individual skeletons does not necessarily accurately reflect the mortality conditions which prevailed generally during this period. An inventory was made of the bones and teeth from each skeleton. Standard metric and non-metric data FOSSE LANE 1996: SPECIALIST REPORTS was recorded; information on sexing and ageing characteristics was collected; and bone and dental pathologies was noted. The demographic characteristics of each skeleton were established following the criteria and procedures presented in Bass (1971), Brothwell (1981), Phenice (1969), and Stewart (1979). Priority for gender determination was given to innominate morphology. Cranium morphology was also used, and, whenever possible, supplemented by univariate measurements of the femur and humerus head, the glenoid fossa of the scapula, the maximum length of the talus, and other robusticity indicators. Morphological traits of the pelvis and cranium from skeletal series of known sex are reported generally to be around 95% accurate (Krogman 1962), univariate measurements range from 80 to 90% accuracy (Buikstra and Mielke 1985; Dittrick and Suchey 1986; Steele 1976). Sexing was only attempted for adult skeletons (a term used here to indicate those above the age of approximately 20 years). Sub-adult age was determined through dental development (Logan and Kronfeld as presented in Downer 1975) diaphysis lengths (Scheuer et al. 1980; Workshop of European Anthropologists 1980) and epiphyseal union (Krogman 1962; Brothwell 1981). Adult age was evaluated using the recommendations of the Workshop of European Anthropologists (1980) for pelvis and cranium, Suchey et al (unpub) also for pubic symphysis metamorphosis, auricular surface metamorphosis by Lovejoy et al. (1985) and dental attrition (Miles 1962; Brothwell 1981). A range of possible ages was first established, using all indicators applicable and then a final best estimate of age was determined by the smallest range of agreement among the indicators. The adults here have been separated into three groups, Young adults (20–29.9), Middleaged adults (30–49.9), and Old adults (50+). As children’s maturation is considered to be a more regular process than the degeneration of the adults’, the sub-adults have been divided into more age categories: Foetal – Birth, Birth – 0.9 years, 1–4.9, 5–9.9, 10–14.9, 15–19.9. One other category is also used, Adults age unknown. Measurements were taken following descriptions in Bass (1971) and Brothwell (1981). Stature was calculated following Trotter (1970). Cranial nonmetric variants were recorded as described by Buikstra (1976) and the post-cranial non-metric variants following Finnegan (1978). Pathological conditions were evaluated through gross anatomical observation and radiographic examination. Criteria for probable diagnosis stemmed from Steinbock (1976), Ortner and Putschar (1981) and Rogers et al. (1987). Coding for dental pathologies followed Hillson (1979). Condition and taphonomy The preservation of the majority of the skeletal TABLE 11: CONDITION AND DEGREE OF COMPLETENESS OF SKELETONS <20% 20-40% 40-60% 60-80% 80%+ Total Good 2 2 2 6 Condition Fair 1 1 2 4 Total Poor 0 3 0 1 2 4 10 material was good (Table 11). Preservation was scored as good, fair or poor on the basis of a visual inspection of the remains. Sixty percent were in good condition, 40% fair and none in poor condition. The degree of completeness varied. Forty percent were over 80% complete, but 30% has less than a fifth of the skeleton present. The less complete skeletons are more often found in the shallow graves and those disturbed by ploughing. Both preservation and completeness are very similar to that observed for the skeletons from the 1990 excavations. Sex and age Of the six adult inhumations, five had characteristics allowing the sex to be determined. Metric standards developed during analysis of the 1990 skeletal material (Pinter-Bellows 2001) were used for univariate measurements of the femur and humerus head, the glenoid fossa of the scapula, and the maximum length of the talus. Two skeletons were diagnosed as female and three as male (Table 12). This fairly even gender ratio was also seen in the 1990 excavation. It is generally assumed that cemeteries with an even sex distribution are likely to be those where a representative selection of the whole adult population was buried, probably in family groups. This certainly appears to be the case archaeologically and osteologically at Fosse Lane. TABLE 12: DEMOGRAPHY Age Foetal -birth Birth -.9 1-4.9 5-9.9 10-14.9 15-19.9 Young adults (20-29.9) Middle-aged adults (30-49.9) Old adults (50+) Adults age unknown ? 3 1 Total 5 Sex M 3 F Total 3 1 1 1 1 4 1 1 3 2 10 W 45 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 In turning to the age profile of this small group (Table 12), 40% of the individuals are sub-adults (those individuals approximately under the age of 20 years). This figure is around that which would be expected in a representative population sample. Three of the four sub-adults recovered are late foetal or neonate. It was not possible from the skeletal evidence to determine whether they were stillborn or died in the immediate post-natal period. The other sub-adult died within the first year of life. In a larger grouping of skeletons where statistical chance figures less heavily into the age of death, the skeletons of several older children would have been expected. The adult age distribution shows that the greatest percentage of aged adults were in the middle-aged category (four out of the five adults which could be aged). This is a quite common distribution. Stature The four adults whose stature could be calculated fit comfortably within the distribution from the 1990 excavation (Table 13). Developmental stress (which can affect stature) does not seem to have been a problem for Fosse Lane with the mean statures for both males and females equal to the tallest means of the other population samples (Pinter-Bellows 2001). This similarity of stature shows that the sub-adults experienced no more overall stress during development than the sub-adults from other sites examined. Skeletal and dental malformations and anomalies One minor skeletal anomaly was noted. The dens epiphysis of the 2nd cervical vertebra never fused to the body of the dens in the middle-aged male (HB 8) leading to a pseudarthosis (a supernumerary articulation). This appears to have caused little trouble for the individual, the articulation on the 1st cervical vertebra is just slightly larger and on two planes, and while slightly rough does not show much osteophyte development. Pathology Dental disease For the six adults, four maxillae and four mandibles survive, with all tooth positions available for study. Only one of these four individuals, a middle-aged male (HB 8), shows evidence of ante-mortem tooth loss. Two individuals have caries, a middle-aged female (HB 1) with four and a middle-aged male (HB 8) with five. Caries are formed when sugars in the presence of harmless bacteria ferment and the resultant plaque has a pH low enough to demineralise enamel, cement, and dentine. In the male one of these caries has resulted in a point abscess, a local circumscribed area of infection (osteomyelitis). Two individuals, a middle-aged female (HB 1) and a middle-aged male (HB 7) have dental calculus. Dental calculus may be considered as a mineralized dental plaque. It takes the form of a concretion on the teeth consisting mainly of calcium salts and, in life, organic material in which flourishes numerous bacteria; it is associated with poor oral hygiene. Periodontal disease is perhaps the best skeletal indicator of general oral hygiene (Manchester 1984, 34). It involves inflammatory pitting of the alveolar margins and the progressive alveolar resorption resulting in exposure of tooth root. Two individuals show evidence of this, a young adult female (HB 2) and a middle-aged male (HB 7). Enamel hypoplasia is an area with a deficiency of enamel of a developmental origin in a tooth. The area affected is usually a band or a line of pitting. It occurs during acute severe stress and has a non-specific aetiology. It is associated with general systemic disorder and nutritional deficiency. There are also some forms of enamel hypoplasia which are hereditary or occur through trauma to the tooth; in order to identify and discount these types, in this study enamel hypoplasia was only counted if at least two teeth from different classes were affected. It was noted on permanent teeth of both of the females (HB 1 and HB 2). TABLE 13: STATURE, MEANS AND RANGES Sex N Measurements Tesco (1996) Female Male 2 2 157 cm (5' 1") & 159 cm (5' 2") 174 cm (5' 7") & 180 cm (5' 8") Sex N Mean Range 162 cm (5' 4") 171 cm (5' 7") 157 cm (5' 2") - 168 cm (5' 6") 162 cm (5' 4") - 178 cm (5' 9") Showerings (1990) Female 10 Male 12 *n is the number of individuals for which the elements were observable W 46 FOSSE LANE 1996: SPECIALIST REPORTS Metabolic and nutritional disease One individual, an infant whose tooth eruption/ calcification was of a nine month old ± three months (HB 5), has porotic hyperostosis. It is presented as a response to chronic anaemia (Moseley 1974). This anaemia can be caused by a number of factors, such as physiological susceptibility, poor nutrition or parasitic infestation. In this individual, the associated anaemia would probably be iron deficiency in childhood probably caused by malnutrition and infections; an iron inadequacy can also be exacerbated by high lead intake. The anaemic state stimulates a proliferative reaction of the marrow in an attempt to make good its deficiency. The marrow space enlarges at the expense of the outer layer of bone. This becomes thin and porous and the marrow cells extend through it. The subsequent deposition of new bone upon these cells has given rise to the classic ‘hair-on-end’ appearance of the outer table in this individual of the frontal, parietals and occipital. The outer surface of the long bones is also raised and very porotic. In this individual the chronic anaemia has lead to a ‘failure to thrive’ (obviously followed by death) as while the dental age of this individual is estimated at approximately nine months, measurements of the long bones gives a developmental age of a 40 week ± two week old foetus. An aetiologically related condition, cribra orbitalia, has been observed of the trabecular type in a middleaged male (HB 7). Cribra orbitalia is an increase of the diploic bone in the orbital roofs resulting in an increase in the thickness of the orbital plate and sievelike lesions or pits appearing in the thin cortical bone layer of the orbital roofs. It is usually considered to be due to anaemia in young childhood, and not to develop in response to anaemia in adulthood (StuartMacadam 1985, 398); so this individual survived the early anaemia, malnutrition and infections which killed the infant. Infection and inflammation The majority of infections affect the soft tissues of the body, and many, from influenza and measles to meningitis and pneumonia, run their course too rapidly for the infection process to spread to the bones. Bone lesions are, therefore, going to represent the chronic infections, which are more likely to involve a bacteria than a virus, viral infections being resolved more rapidly. The Fosse Lane individuals show only evidence of non-specific infections, inflammatory in nature, for which the pathogenic agent is unknown. Lesions which are superficial and appear to involve only the fibrous covering of the bone (the periosteum) are termed periostitis. Periostitis is recognised as a deposition of irregular new bone upon the outer surface of a bone. It is usually thin and localized in area, but can be thicker and cover a bone more extensively. Three individuals exhibit periostitis. One young adult female (HB 2) has unhealed periostitis in a line 21mm long by 9mm wide perpendicular to the spine on both sides of the body of the left scapula. The sternal end of ten ribs, both left and right have porous new bone on the external surface. Two middle-aged males have healed periostitis on the lower leg; one (HB 7) on the left tibia and fibula and one (HB 8) on the left fibula. Periostitis, in the absence of general pathology, has been noted in association with repeated and minor trauma to the lower legs (Manchester 1984), chronic venous insufficiency (Resnick and Niwayama 1988) and tropical ulcers with soft tissue infection (Molleson 1993). Trauma Trauma may affect soft tissue, bone, or both. For trauma to the bone – fractures – there is direct evidence (as discussed below). However, the majority of injuries are sustained by the soft tissues. Evidence for these injuries is indirect and depends on the severity of the damage to the soft tissues adjacent to the bone. There are four individuals who have injuries to their entheses or syndesmoses. The osseous sites of tendons, which attach the muscles to the bone, and ligaments, which attach near particular joints and limit movement, entheses and syndesmoses respectively, can have alterations occur in the fibres. Disruption of the fibre bundles that are anchored to the underlying bone can cause hyperostotic (bone growth) or osteolytic changes (bone loss) to that area of the bone. They are often degenerative in nature and are ‘common in older individuals’ (Resnick and Niwayama 1981, 1297). If the individual in which the lesion occurs is young and robust and disease is not suggested as the causative agent, then direct and work-related trauma or loading stresses may be explored. The middle-aged female (HB 1) has 1mm high enthesophytes (extra bone) on the proximal hand phalanges where the flexor digitorum superficialis attach. This is seen on the proximal and medial phalanges of the right but not the left hand on a middle-aged male (HB 1). The young adult female (HB 2) has enthesophytes on the left and right humeri on the distal, anterior surface where the brachialis attaches and on the proximal posterior surface of the right fibula where the tibialis posterior attaches. She also has a small roughened circumscribed enthesolithic lesion (area of bone loss) on the left femur on the distal posterior where the medial head of the gastrocnemius attaches. The adult of unknown sex and age (HB 4) has an enthesophyte on the right ulna on the proximal end where the anconeus attaches; a middle-aged male (HB 8) has an enthesophyte for the small muscle of the left ulna. W 47 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 Another pathology thought to sometimes originate from stress to soft tissue, in this case the intervertebral discs, is Schmorl’s node. While its aetiology is not completely understood, it is believed that if the disc located between the vertebrae is subject to too much strain it may rupture. The bubble of escaped material then presses against the body of the adjacent vertebra, which gradually yields to the pressure, allowing a small cavity to be formed in its body. Schmorl’s nodes provide evidence of torsional and compressional injuries that have occurred as a result of the sudden application of loading stresses if they are not degenerative or related to various diseases (Resnick and Niwayama 1981, 1404). One middle-aged male (HB 3) exhibits these nodes on five mid to lower thoracic vertebrae. The one bone fracture found was probably accidental (using criteria from Manchester 1983, 58). A middle-aged male (HB 3) has a healed, simple, closed oblique fracture to the tip of the medial malleolus of the left tibia with less than 5mm of proximal-anterior displacement; and a simple, closed oblique fracture of the distal articular end of the left fibula with slight displacement. Neoplastic disease Neoplasma or new growth may be classified as malignant or benign. No cases of malignant neoplasms have been found in this sample. There is one case of a benign neoplasm, an osteoma in bone cells in circumscribed areas, normally in the periosteum, which grow more than the surrounding tissue. The cells eventually mature and the osteoma is not progressive. They are small hemispherical hard projections which are usually about the size of a pea. They are usually symptomless and are not considered to be of clinical significance. A middle-aged male (HB 8) has one on the proximal right thumb phalange, just distal of midway on the medial side of the shaft. Conclusions The sample gives a look at what appears to be two family groups and a few others from part of a small community. These individuals were relatively healthy. The adult skeletons are of medium height with mean statures for the males close to the means for other Roman populations in Britain and the females slightly taller than the means of many. There are not a high number of osseous pathologies. REFERENCES Albarella, U, 1998, ‘The animal bones’, in P Ellis et al., Excavations Alongside Roman Ermine Street, Cambridgeshire, 1996, BAR 276, 99–104 W 48 Albarella, U., and Davis, S., 1994, The Saxon and Medieval Bones Excavated 1985–1989 from West Cotton, Northamptonshire, English Heritage, Ancient Monuments Lab Report 17/97 _____ , 1996, ‘Mammals and birds from Launceston Castle, Cornwall: decline in status and the rise in agriculture’, Circaea, 12 (1), 1–156 Allason-Jones, L., 1989, Earrings in Roman Britain, BAR 201, Oxford. Allason-Jones, L., and Miket, R., 1984, The Catalogue of Small Finds from South Shields Roman Fort, Soc. Antiq. Newcastle Upon Tyne Monograph Series No. 2. Allen, J.R.L., and Fulford, M.G., 1996, ‘The distribution of south-east Dorset Black Burnished category I pottery in south west Britain, Britannia, 27, 223–81. Bagwell, T., and Webster, C., 2005. ‘Somerset Archaeology, 2004’, SANH 148, 114–15. Barrett, J.C., Freeman, P.W.M., and Woodward, A., 2000, Cadbury Castle, Somerset: The Later Prehistoric and Early Historic Archaeology, English Heritage Archaeol Rep 20. Bass, W., 1971, Human Osteology: A Laboratory and Field Manual of the Human Skeleton, Columbia. Bevan, L., 1998, ‘Bracelets; pins’, in P. Leach, Great Witcombe Roman Villa, Gloucestershire. A Report on Excavations by Ernest Greenfield 1960–1973, BAR 266, 86–90. _____ , 2001a, ‘The flint’, in Leach 2001, 100–7. _____ , 2001b, ‘Fired clay, brick and tile’, in Leach 2001, 225. Biddle, M., 1967, ‘Two Flavian burials from Winchester’, Antiq Journ, 47, 224–50. Birbeck, V., 1997, Excavations on Iron Age and Romano-British Settlements at Cannards Grave, Shepton Mallet, Wessex Archaeol Rep. 38985. _____ , 2002. ‘Excavations on Iron Age and RomanoBritish settlements at Cannards Grave, Shepton Mallet’, SANH 144, 41–116. Boessneck, J., 1969, ‘Osteological differences between sheep (Ovis aries Linné) and goat (Capra hircus Linné), in D. Brothwell and E. Higgs (eds), Science in Archaeology, London, 331–58. Booth, P., 1982, ‘A Romano-British burial from Mancetter’, Trans Birmingham Warks Archaeol Soc, 92, 134–6. Brain, C.K., 1981, The Hunters or the Hunted? Introduction to Taphonomy, London Branigan, K., 1969, The Romans in the Bristol Area, Bristol. _____ , 1977, Gatcombe: The Excavation and Study of a Romano-British Villa Estate, 1967–76, BAR 44, Oxford. Broomhead, R.A., 1999. ‘Ilchester Great Yard archaeological excavations 1995’, SANH 142, 139–91. Brothwell, D., 1981, Digging Up Bones, 2nd edn, FOSSE LANE 1996: SPECIALIST REPORTS London. Buikstra, J., 1976, Hopewell in the Lower Illinois Valley – a Regional Study of Human Biological Variability and Prehistoric Mortuary Behavior, Evanston, Illinois. Buikstra, J., and Mielke, J., 1985, ‘Demography, diet, and health’, in The Analysis of Prehistoric Diets (R Gilbert Jr and J Mielke eds), Orlando, Florida, 359–422. Burnham, B.C., and Wacher, J.S., 1990, The Small Towns of Roman Britain, Berkeley. Calza, G., and Becatti, G., 1977, Ostia, Rome. Coles, J., and Minnitt, S., 1995, ‘Industrious and Fairly Civilised’: The Glastonbury Lake Village, Taunton. Cool, H.E.M., 1990, ‘Roman metal hair pins from southern Britain, Archaeol Journ, 147, 148–82. Cool, H.E.M., and Price, J., 1985, ‘Glass’, in H. Hurst, Kingsholm, Gloucester, 41–54. _____ , 1991, ‘The Roman vessel and window glass’, in T. Padley, The Metalwork, Glass and Stone Objects from Castle St, Carlisle: Excavations 1981–2, 165–76. _____ , 1993, ‘Roman glass’, in P. Woodward et al., Excavations at the Old Methodist Chapel and Greyhound Yard, Dorchester, 1981–1984, Dorchester, 150–67. _____ , 1995, Roman Vessel Glass from Excavations in Colchester, 1971–85, Colchester Archaeol Rep, 8. Corney, M., 1997, ‘The origins and development of the ‘small town’ of Cunetio, Mildenhall, Wiltshire’, Britannia, 28, 337–50. Costen, M., 1992, The Origins of Somerset, Manchester. Crabtree, P.J., 1989, West Stow Early Anglo-Saxon Animal Husbandry, East Anglian Archaeology, 47. Crummy, N., 1983, The Roman Small Finds from Excavations in Colchester 1971–9, Colchester. Dannell, G.B., 1971, ‘The samian pottery’, in B Cunliffe, Excavations at Fishbourne 1961–1969, Leeds, 260–318. Dannell, G.B., and Wild, J.P., 1987, Longthorpe II. The Military Works-Depot: An Episode in Landscape History, Britannia Monographs 8. Davey, J.E., 2005. The Roman to Medieval Transition in the Region of South Cadbury Castle, Somerset, BAR Brit Ser 399, Oxford. Davis, S., 1988, Animal Bones from Dodder Hill, a Roman Fort Near Droitwich (Hereford and Worcester) Excavated in 1977, English Heritage Ancient Monuments Lab Rep 140/88. _____ , 1992, A rapid method for recording information about animal bones from archaeological sites, English Heritage Ancient Monuments Lab Rep19/92. Déchelette, J., 1904, Les Vases Céramiques Ornés de la Gaule Romaine, Paris. Dickinson, B., 1990, ‘The samian ware’, in McCarthy 1990, 213–36. _____ , 2001, ‘Samian ware’, in Leach with Evans 2001, 144–9. Dittrick, J., and Suchey, J., 1986, ‘Sex determination of prehistoric central California skeletal remains using discriminant analysis of the femur and humerus’, American Journ of Physical Anthropology, 70, 3–9. Dobney, K., Jaques, D., and Irving, B., undated, Of Butchers and Breeds. Report on Vertebrate Remains from Various Sites in the City of Lincoln, Lincoln Archaeological Studies 5, Lincoln. Down, A., and Rule, M., 1971, Chichester Excavations 1, Chichester. Downer, G.C., 1975, Dental Morphology, Bristol. Driesch, A. Von den, 1995, A Guide to the Measurement of Animal Bone from Archaeological Sites, Peabody Museum Bulletin 1, Cambridge Mass., Harvard University. Ellis, P., 1984. Catsgore 1979, Further Excavation of the Romano-British Village, Bristol. Ellis, P. and Leach, P., 2000 The Roman small town at Shepton Mallet, Somerset. The Tesco Excavation 1996, Birmingham University Field Archaeology Unit No. 449. Esmonde Cleary, S., 2001. ‘Roman coins’ in Leach with Evans 2001, 211–24. Evans, C.J., 2001. ‘The Roman pottery’, in Leach with Evans 2001, 107–69. Farwell, D.E., and Molleson, T.I., 1993, Poundbury Volume 2, The Cemeteries, Dorchester. Ferris, I.M., and Cooper, J., 1996, ‘Copper alloy objects’, in I. Ferris and A. Esmonde-Cleary, Excavations at New Cemetery, Rocester, Staffordshire, 1985–1987, Trans Staffs Arch Hist Soc, 35. Finnegan, M., 1978, Non-metric variation of the infracranial skeleton, Journ of Anatomy, 125, 23– 37. Fowler, P., and Rahtz, P., 1972. ‘Somerset AD 400– 700’, in P. Fowler (ed), Archaeology and the Landscape, Bradford on Avon, 187–221. Fremersdorf, F., and Polonyi-Fremersdorf, G., 1984, Die farblosen Gläser der Frühzeit in Köln 2 und 3 Jahrhundert, Denkmäler des römischen Köln 9. Gillam J.P., 1976, ‘Coarse fumed ware in northern Britain’, Glasgow Archaeol Jnl, 4, 57–80. Grant, A., 1982, ‘The use of tooth wear as a guide to the age of domestic ungulates’, in B. Wilson et al. (eds), Ageing and Sexing Animal Bones from Archaeological Sites, BAR 109, Oxford, 91–108. _____ , 1984, ‘Animal husbandry in Wessex and the Thames valley’, in B. Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain, Oxford, 102–19. Graham, A., and Newman, C., 1993, ‘Recent excavations of Iron Age and Romano-British W 49 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 enclosures in the Avon Valley, Wiltshire’, Wiltshire Archaeol Magazine, 86, 8–57. Green, M.A., and Welch, F.B.A., 1965, Geology of the Country around Wells and Cheddar, Geological Survey of Great Britain, Memoir for Sheet 280, London. GSA 1990a. Report on Geophysical Survey: Mendip Business Park, Shepton Mallet, Report 90/51, Geophysical Surveys Bradford. GSA 1990b. Report on Geophysical Survey: Fosse Way, Shepton Mallet, Report 90/52, Geophysical Surveys Bradford, 1990b. Guido, M., 1978, The Glass Beads of the Prehistoric and Roman Periods in Britain and Ireland, Res Rep Soc Antiq 35, London. Harcourt, R.A., 1979, ‘The animal bones’, in G Wainwright (ed), Gussage All Saints: An Iron Age Settlement in Dorset, London, 150–60. Harden, D.B., 1961, ‘Domestic window glass: Roman, Saxon and Medieval’, in E. Jope, Studies in Building History, London, 49–52. _____ , 1983, ‘The glass hoard’, in S. Johnson, Burgh Castle. Excavations by Charles Green 1958–61, East Anglia Archaeology Report 20, 78–89. Hartley, K.F., 1991, ‘The mortaria’, in Holbrook and Bidwell 1991, 189–215 _____ , 2001, ‘Shepton Mallet mortaria’, in Leach 2001, 130–2. Hillman, G., 1982, ‘Evidence for spelt malting at Catsgore’, in Leech 1982, 137–41. Hillson, S.W., 1979, ‘Diet and dental disease’, World Archaeol, 11, 147–62. Hoffmann, B., 1996, Römisches Glas in BadenWürttemberg von 70 bis 260, unpub Phd thesis, Freiburg University. _____ , forthcoming, The Roman Glass from Vindolanda, Vindolanda Research Reports. Holbrook, N., and Bidwell, P.T., 1991, Roman finds from Exeter. Exeter Archaeol Report 4. Hooke, D., 1989. ‘Early medieval estate and settlement patterns’, in M Aston et al., The Rural Settlements of Medieval England, London. Hopstätter, H., 1942, ‘Ein römisches Brandgrab bei Kisselbach (Hunsrück)’, Germania, 26, 211–13. Horwell, D., 1977, Stone objects, in Branigan 1977, 99–102. Jacomet, S., 1987, Prahistorische Getreidefunde. Botanisches Institut der Universitat Abteilung Pflanzensystemematik und Geobotanik, Basel. Jones, J., 2000. ‘Plant macrofossils’, in S. Rippon, ‘The Romano-British exploitation of coastal wetlands: survey and excavation of the North Somerset Levels, 1993–7’, Britannia, 31, 122–38. Jones, M., 1981, ‘The development of crop husbandry’, in M. Jones and G. Dimbleby, The Environment of Man: the Iron Age to the AngloSaxon Period. BAR British Series 87. Juhasz, G., 1935, Die Sigillaten von Brigetio, W 50 Dissertationes Pannonicae, Ser 2, no 3, Budapest. Kilbride-Jones, H.E., 1938, ‘Glass-armlets in Britain’, Proc Soc Antiq Scotland, 72, 366–72. King, A.C., 1984, ‘Animal bones and the dietary identity of military and civilian groups in Roman Britain, Germany and Gaul’, in T. Blagg and A. King (eds), Military and Civilian in Roman Britain: Cultural Relationships in a Frontier Province, Oxford, BAR 136, 187–218. King, D., 1998, Mechanised corn milling in Roman Britain, Quern Study Group. Kratochvil, Z., 1969, ‘Species criteria on the distal section of the tibia in Ovis ammon F. Aries L. and Capra aegrus F. Hircus L.’, Acta Veterinaria, 38, 483–90. Krogman, W., 1962, The Human Skeleton in Forensic Medicine, Springfield, Ill. Laidlaw, M., 1997, ‘Romano-British pottery’, in Birbeck 1997, 24–32. Lawson, A.J., 1975, ‘Shale and jet objects from Silchester’, Archaeologia, 105, 241–75. Legge, A.J., 1981, ‘The agricultural economy’, in R. Mercer, Grimes Graves, Norfolk: Excavations 1971–72, DoE Res Rep 11, London, 79–103. Leach, P.J., 1982. Ilchester Volume 1: Excavations 1974–5, Bristol. _____ , 1984, ‘The pottery, in Ellis 1984, 23–8. _____ , 1990. An Archaeological Assessment of the Mendip Business Park, Fosse Lane , Shepton Mallet 1990, BUFAU. _____ , 1991. An Archaeological Evaluation at Bullimore Farm, Shepton Mallet, Somerset 1991, BUFAU. _____ , 1992. An Archaeological Evaluation of Development Land at Fosse Lane, Shepton Mallet, Somerset 1992, BUFAU. _____ , 1994a. Mendip Business Park, Shepton Mallet: Site A 4/94, Archaeological excavations in advance of development, BUFAU Report 298. _____ , 1994b. Ilchester Volume 2: Archaeology, Excavations and Fieldwork to 1984, Sheffield. _____ , 2001. Roman Somerset, Dorchester. _____ , 2009. ‘Prehistoric ritual landscapes and other remains at Field Farm, Shepton Mallet’, SANH 152, 11–68. Leach, P.J., and Ellis, P., 1996. Fosse Lane, Shepton Mallet, the Tesco excavations, BUFAU interim report. Leach, P.J, with Evans, C.J., 2001. Excavation of a Romano-British Roadside Settlement in Somerset: Fosse Lane, Shepton Mallet, 1990, Britannia Monograph 18, London. Leech, R.H., 1982. Excavations at Catsgore 1970– 1973, Bristol. Lovejoy, C.O., Meindl, R., Mensforth, R., and Barton, T., 1985, ‘Chronological metamorphosis of the auricular surface of the ilium: a new method for FOSSE LANE 1996: SPECIALIST REPORTS the determination of adult skeletal age at death’, American Journl of Physical Anthropology, 68, 15–28. Lucas, R.N., 1993, The Romano-British villa at Halstock, Dorset: Excavations 1967–1985, Dorset Nat Hist and Archaeol Soc Mono Series 13. Mackreth, D., 2001, ‘Brooches’, in Leach 2001, 179– 200. Maltby, M., 1979, The Animal Bones from Exeter 1971–1975, Sheffield: Exeter Archaeological Reports 2. _____ , 1981, Iron Age, Romano-British and AngloSaxon animal husbandry: a review of the faunal evidence, in M Jones and G Dimbleby (eds) The Environment of Man: The Iron Age to the AngloSaxon Period, BAR 87, Oxford, 254–61. Manchester, K., 1983, The Archaeology of Disease, Bradford. _____ , 1984, Eccles skeletal report, unpub. Manning, W.H., 1985, Catalogue of the RomanoBritish Iron Tools, Fittings and Weapons in the British Museum, London. McCarthy, M., 1990, A Roman, Anglian and Medieval site at Blackfriars Street, Carlisle, Cumb West Archaeol Soc Res Series 4. Mepham, L., 1993, ‘The stone’, in Graham and Newman 1993, 36. Miles, A.E.W., 1962, ‘Assessment of the ages of a population of Anglo-Saxons from their dentitions’, Procs Royal Society of Medicine, 55, 881–6. Molleson, T.I., 1993, ‘The human remains’, in Farwell and Molleson 1993, 142–213. Morgan, G., 2001, ‘Mortar and plaster analysis’, in Leach 2001, 226–30. Moseley, J.E., 1974, ‘Skeletal changes in the anemias’, Seminars in Roentgenology, 9(3), 169– 84. Murphy, P., 1982, ‘Plant remains from Roman deposits at Ilchester’, in Leach 1982, 286–90. Neale, F., 1976. ‘Saxon and medieval landscapes’, in R Atthill (ed) Mendip, A New Study, Newton Abbot, 75–101. O’Connor, T., 1988, ‘Bones from the General Accident Site, Tanner Row’, Archaeology of York fascicule 15/2, The Animal Bones, CBA Res Rep, London. Ortner, D., and Putschar, W., 1981, Identification of Pathological Conditions in Human Skeletal Remains, Smithsonian Contributions to Anthropology 28, Washington, DC. Oswald, F., 1936–7, Index of Figure-Types on TerraSigillata, Liverpool. Payne, S., 1973, ‘Kill-off patterns in sheep and goats’, Anatolian Studies: Journal of the British Institute of Archaeology at Ankara, 23, 281–303. _____ , 1985, ‘Morphological distinctions between the mandibular teeth of young sheep, Ovis, and goats, Capra’, Journal of Archaeological Science, 12, 139–47. _____ , 1987, ‘Reference codes for the wear states in the mandibular cheek teeth of sheep and goats’, Journal of Archaeological Science, 14, 609–14. Payne, S., and Bull, G., 1988, ‘Components of variation in measurements of pig bones and teeth, and the use of measurements to distinguish wild from domestic pig remains’, Archaeozoologia, 2, 27–65. Pengelly, H., 1982, ‘The samian;, in J. Draper and C. Chaplin, Dorchester Excavations Volume 1: Excavations at Wadham House 1968; Dorchester Prison 1970, 1975, and 1978; and Glyde Path Road 1966, Dorset Natur Hist Archaeol Soc Mono 2, 78–9. Phenice, T., 1969, ‘A newly developed visual method of sexing the os pubis’, American Journ of Physical Anthropology, 30, 297–301. Philpott, R., 1991, Burial Practices in Roman Britain, a Survey of Grave Treatment and Furnishing AD 43–410, BAR British Series 219, Oxford. Pinter-Bellows, S., 2001, ‘Animal bone’, in Leach 2001, 289–303. Pitt Rivers, A., 1884, Report on Excavations in the Pen Pits near Penselwood, Somerset. Price, J., 1980, ‘The Roman glass’, in G. Lambrick, ‘Excavations in Park Street, Towcester’, Northamptonshire Archaeology, 15, 35–118. _____ , 1984, ‘The objects of glass’, in M. Fulford, The Silchester Defences 1974–80, Britannia Monograph 5,116–18. _____ , 1985, ‘Early Roman vessel glass from burials in Tripolitania: a study of finds from Forte della Vite and other sites now in the collections of the National Museum of Antiquities in Tripoli’, in D. Buck and D. Mattingly, Town and Country in Roman Tripolitania. Papers in Honour of Olwen Hackett, BAR Int Ser, 274, 67–105. _____ , 1987, ‘Glass from Felmongers-Harlow in Essex. A dated deposit of vessel glass found in an Antonine pit’, Congrés AIHV, 10, 185–206. _____ , 1989, ‘The Roman glass’, in S. Frere and J. Wilkes, Strageath: Excavations within the Roman Fort, 1973–1986, 192–203. _____ , 1990, ‘Roman vessel and window glass’, in McCarthy 1990, 163–79. _____ , 1995, ‘The glass vessels’, in W. Manning, J. Price and J. Webster, The Roman Small Finds. Report on the excavations at Usk 1965–1976, Cardiff, 139–91. Price, J, and Cottam, S, 1994, ‘Glass’, in S. Cracknell and C. Mahany, Roman Alcester: Southern Extramural Area. 1964–1966 Excavations, CBA Res Rep 97, 224–9. _____ , 2001, ‘The Roman glass’, in Leach 2001, 170–6. Rahtz, P.A., 1977. ‘Late Roman cemeteries and beyond’ in R. Reece (ed) Burial in the Roman W 51 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY, 2011 World, CBA Res Rep 22, 53–64. Resnick, D., and Niwayama, G., 1988, Diagnosis of Bone and Joint Disorders, 2nd edn, London. Roe, F.E.S, 1995, ‘Stone’, in Coles and Minnitt 1995, 161–6. _____ , 2000, ‘Worked stone’, in Barrett et al. 2000, 148. _____ , 2001a, ‘Querns and millstones’, in Leach 2001, 235. _____ , 2001b, ‘Whetstones’, in Leach 2001, 235–7. _____ , 2001c, ‘Stone mortars’ in Leach 2001, 177– 8. Rogers, G.B., 1974, Poteries Sigillées de la Gaule Centrale, 1 Les motifs non figures, Sup Gallia 28. Rogers, J., Waldron, T., Dieppe, P., and Watt, I., 1987, ‘Arthropathies in palaeopathology: the basis of classification according to most probable cause’, Journ of Archaeol Science, 14, 179–83. Scheuer, J.L., Musgrave, J.H., and Evans, S.P., 1980, ‘The estimation of late fetal and perinatal age from limb bone length by linear and logarithmic regression’, Annuals of Human Biology, 7, 257– 65. Stace, C., 1991, New flora of the British Isles, Cambridge. Stallibrass, S., 1991, A Comparison of the Measurements of Romano-British Animal Bones from Periods 3 and 5, Recovered from Excavations at Annetwell Street, Carlisle, Ancient Monuments Laboratory Report 133/91. Stanfield, J.S., and Simpson, G., 1990, Les Potiers de la Gaule Centrale, Gonfaron. Steele, D.G., 1976, ‘The estimation of sex on the basis of the talus and calcaneus’, American Journ of Physical Anthropology, 45, 581–8. Steinbock, R.T., 1976, Paleopathological Diagnosis and Interpretation: Bone Disease in Ancient Human Populations, Springfield, Ill. Stevens, C., 1999. ‘The plant remains’, in Broomhead 1999, 156–65. Stevenson, R.B.K., 1956, ‘Native bangles and Roman W 52 glass’, Proc Soc Antiq Scotland, 88, 208–21. _____ , 1976, ‘Romano-British glass bangles’, Glasgow Archaeological Journal, 4, 45–54. Stewart, T.D., 1979, Essentials of Forensic Anthropology, Springfield, Ill. Straker, V., ‘Charred plant macrofossils’, in Leach with Evans 2001, 303–7. Stuart-Macadam, P., 1985, ‘Porotic hyperostosis: representative of a childhood condition’, American Journ of Physical Anthropology, 66, 391–98. Suchey, J.M., Brooks, S.T., and Katz, D., nd, Instructions for use of the Suchey-Brooks system for age determination of the female os pubis, unpub report. Sunter, N., and Woodward, P.J., 1987, Romano-British Industries in Purbeck, Dorchester. Thomas, C., 1987, ‘The worked stone’, in Sunter and Woodward 1987, 30–6 Todd, M., 2007. Roman Mining in Somerset, Exeter. Trotter, M., 1970, ‘Estimation of stature from intact long bones’, in T. Stewart (ed) Personal Identification in Mass Disasters, National Museum of Natural History, Washington, DC, 71– 83. Webster, C.J., and Croft, R.A., 1994. ‘Somerset archaeology, 1993’, SANH 137, 127–56. Wedlake, W.J., 1958, Excavations at Camerton, Somerset, Bath. Wheeler, R.E.M, 1929, ‘A Roman pipe burial from Caerleon, Monmouthshire’, Antiquaries Journ, 9, 1–7. White, K.D., 1970, Roman Farming, London. White, R.W., 2007. Britannia Prima: The Romans in the West of Britain, London. Wilson, E.M., 1980, ‘Excavations at West Mains of Ethie’, Proc Soc Antiq Scotland, 110, 114–21. Woodward, A.B., 1993, Coffins and grave linings, in Farwell and Molleson 1993, 227. Workshop of European Anthropologists, 1980, ‘Recommendation for age and sex determination’, Journ of Human Evolution, 9, 517–49.

RELATED PAPERS

RELATED TOPICS

Log In

The Glass, in: P.Ellis and P. Leach, The Roman Settlement at Fosse Lane, Shepton Mallet: The tesco excavation, 1996-7. pp. W14-17 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY VOL. 155

The Glass, in: P.Ellis and P. Leach, The Roman Settlement at Fosse Lane, Shepton Mallet: The tesco excavation, 1996-7. pp. W14-17 SOMERSET ARCHAEOLOGY AND NATURAL HISTORY VOL. 155

Related Papers

RELATED PAPERS

RELATED TOPICS