FELIX ROMULIANA GAMZIGRAD IVANA POPOVI] MAJA @IVI] MIROSLAV LAZI] MIODRAG SLADI] SOFIJA PETKOVI] DRAGOSLAV SREJOVI] MILKA ^ANAK-MEDI] BRANA STOJKOVI]-PAVELKA \OR\E JANKOVI] ALEKSANDAR V. POPOVI] FELIX ROMULIANA – GAMZIGRAD INSTITUTE OF ARCHAEOLOGY, Belgrade Monographs N° 49 INSTITUTE FOR THEOLOGICAL RESEARCH, FACULTY OF ORTHODOX THEOLOGY, University of Belgrade PUBLISHED BY INSTITUTE OF ARCHAEOLOGY, Belgrade INSTITUTE FOR THEOLOGICAL RESEARCH, Faculty of Orthodox Theology, University of Belgrade SERBIAN ORTHODOX CHURCH, Diocese of Timok FOR PUBLISHERS Slavi{a Peri} Bogoljub [ijakovi} EDITORIAL BOARD Nikola Tasi}, Slobodan Du{ani}, Aleksandar Jovanovi} , Milka ^anak-Medi}, Slavi{a Peri}, Maja @ivi}, Ivana Popovi} (editor), Sofija Petkovi} (secretary) REVIEWERS Miloje Vasi}, Nikola Tasi}, Aleksandar Jovanovi} , Miroslav Jeremi} TRANSLATION Mirjana Vukmanovi} PHOTOGRAPHS Neboj{a Bori}, Vladimir Popovi}, Branislav Strugar, Institute of Archaeology, Office for Protection of Cultural Monuments of the Republic of Serbia and National Museum in Zaje~ar COMPUTER DESIGN OF PHOTOGRAPHS Neboj{a Bori} SITE PLANS Documentation of the Office for Protection of Cultural Monuments of the Republic of Serbia Documentation DAI (RGK, Frankfurt a/M) Aleksandar Kapuran DRAWINGS Milovan Risti}, Aleksandar Kapuran GRAPHIC DESIGN Danijela Paracki & D_SIGN, Belgrade PRINTED BY SLU@BENI GLASNIK, Belgrade ISBN 987-86-80093-73-4 (Institute of Archaeology) ISBN 987-86-7405-102-3 (Institute for Theological Research) Printed in 600 copies FELIX ROMULIANA – GAMZIGRAD AUTHORS Ivana Popovi}, Maja @ivi}, Miroslav Lazi}, Miodrag Sladi}, Sofija Petkovi}, Dragoslav Srejovi}, Milka ^anak-Medi}, Brana Stojkovi}-Pavelka, \or|e Jankovi}, Aleksandar V. Popovi} EDITOR Ivana Popovi} BELGRADE 2011 CONTENTS 7 IVANA POPOVI] FOREWORD Transformation of one multi-layered settlement: anonymous settlements at Gamzigrad – Felix Romuliana – Romulianum – Romuliana – anonymous settlements at Gamzigrad 11 MAJA @IVI] Gamzigrad: name, position and economic potential 15 MAJA @IVI] History of archaeological investigations and conservation-restoration works at Gamzigrad 21 MIROSLAV LAZI] Prehistoric settlements and necropoles at Gamzigrad and in its surroundings 29 MIODRAG SLADI] Gamzigrad in the protohistory 33 SOFIJA PETKOVI] Roman Gamzigrad before the imperial palace 43 DRAGOSLAV SREJOVI] Imperial palace 49 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA Architecture and spatial structure of the imperial palace 107 MAJA @IVI] Artistic achievements in the imperial palace 141 IVANA POPOVI] Sacred-funerary complex at Magura 159 DRAGOSLAV SREJOVI] Diva Romula, Divus Galerius 167 SOFIJA PETKOVI] Romuliana in the time after the palace 201 \OR\E JANKOVI] Gamzigrad in the Middle Ages 213 ALEKSANDAR V. POPOVI] Written sources on Galerius 221 SOURCES 222 BIBLIOGRAPHY IVANA POPOVI] FOREWORD TRANSFORMATIONS OF ONE MULTI-LAYERED SITE: ANONYMOUS SETTLEMENTS AT GAMZIGRAD – FELIX ROMULIANA – ROMULIANUM – ROMULIANA – ANONYMOUS SETTLEMENTS AT GAMZIGRAD (Galerius) Ortus Dacia Ripensi, ibique sepultus est; quem locum Romulianum ex vocabulo Romulae matris appelarat. Pseudo-Aurelius Victor, Epitome, 40, 16. The imperial palace Felix Romuliana is one of the best preserved Roman monuments in Serbia and at the same time one of the most apparent examples of the concept and symbolism of the architectural designs at the beginning of the Late Roman period. The remains of the structure are situated near the village Gamzigrad in the Crna Reka region in eastern Serbia. This is mostly the volcanic landscape surrounded by the hills of volcanic masses, most frequently of green-gray andesite. The natural resources including ores, minerals and thermal springs, as well as fact that the region was open to lower Danube Basin, made possible for culture to develop in this area already in the early periods of prehistory. However, this cultural evolution was not straightforward, first of all because of the waves of ethnic migrations. In the period between the 18th and 13th century BC the population of this region, engaged in cattle-breeding, farming and metallurgy, cherished distinct forms of material and religious life recognized as the specific Gamzigrad culture of the Bronze Age period. But, such cultural development was interrupted by the arrival of new ethnic groups, which, during the Iron Age and through the complex processes of the ethnogenesis, would profile in the ancient Balkan tribe of the Triballi. The Triballi had been displaced from these regions in the 4th century BC by the Celts, who ruled unmitigatedly in Pannonia, middle Danube basin and parts of the central Balkans until the arrival of the Romans. After the Roman conquests in the first half of the 1st century AD the Crna Reka region had been included in the Moesia province, i.e. from the year AD 86 in the Upper Moesia (Moesia Superior). Despite being administratively connected with the geographically close middle Danube basin and the Morava Valley, where Roman urban and military centers emerged (Viminacium, Margum, Naissus, Scupi), the eastern Serbia remained outside main communications, so the Romanization process was rather slow. This region experienced cultural and economic renaissance only from the middle of the 3rd century and especially after the Romans 7 IVANA POPOVI] MAP 1. Roman Empire at the beginning of 4th century with locations of imperial palaces and residences walls already in the middle of the 19th century, the character of this structure remained enigmatic for many years. Although systematic archaeological excavations, initiated in 1953, brought to light many archaeological features and objects, the enigma still remained unsolved. The scholars explaining the character of Gamzigrad drifted from the assumption that it was a military camp to the idea that there was the seat of the governor of the nearby mines. But, the distance from the main roads on one hand and discovery of the luxurious mosaic floors and marble sculptures on the other, made such hypotheses unconvincing. When the academician Dragoslav Srejovi} became director of the Gamzigrad investigation project in 1970, the archaeological excavations and conservation works at the site were intensified. Already in 1983 Dragoslav Srejovi} published his opinion that, according to the data from the written sources (PseudoAurelius Victor, Epit. 40, 16), Gamzigrad is in fact Romuliana, the palace which emperor Galerius had built at his birthplace in honor of his mother and where he had been planning to withdraw after his descending from the throne. In that text, whose abandoned Dacia around AD 272, when the frontier of the Empire was once again established along the right Danube bank. The restoration of the old and construction of new fortifications along the Danube resulted in foundation of the new settlements, construction of the roads, opening and exploitation of the mines. After new administrative division of the Empire at the end of the 3rd century the Crna Reka region had been incorporated within the Dacia Ripensis province and consequently this region was once again connected with the lower Danube Basin (map 1). Because of the increasing danger of the invasion of barbarian tribes in the territory of the Empire, the army, which had been amassed along the Danube frontier, became more and more important factor in the public life and the new emperors had been elected from the military ranks. The focus of the Empire shifted to the Danube provinces, where the luxurious villas and palaces had been built. These circumstances had an impact also on the building of the magnificent structure at Gamzigrad. However, even though travelers, archaeologists and geologists noticed the monumentality of the Gamzigrad towers and 8 FOREWORD segments we also reproduce in this book, Srejovi} anticipated the great discovery, which followed just a year later. Namely, the fragmented archivolt made of tuffaceous limestone, with the inscription FELIX ROMULIANA, confirming that it was sacred Romula’s villa or house, has been found in 1984 in the southwestern section of the palace. This finally solved the enigma concerning the character and purpose of the structure at Gamzigrad, i.e. the hypotheses proposed by Srejovi} that it was an imperial palace, the memorial of Galerius, co-ruler of Diocletian and his devoted follower in the realization of new political and ideological system of tetrarchy. And the portrait of this emperor – monumental head made of purple porphyry, discovered in 1993, and one of the best so far known artistic achievements from the period of tetrarchy, was yet another confirmation of the imperial character of the Gamzigrad palace. An inexhaustible intellectual curiosity prompted in 1990 Dragoslav Srejovi} to start besides the investigations of the structures within the palace also archaeological excavations on the nearby hill Magura, where first investigators of Gamzigrad have already recorded the remains of the Roman architecture. During four years of investigation of that hill there have been discovered and explored two mausolea, two tumuli – consecration memorials and the tetrapylon. Combining his analyses of the structures and symbolic meaning of ornamental sculpture in the palace with the new finds and quoted information from the work Epitome ascribed to Aurelius Victor Srejovi} came to conclusion that Magura is distinct sacred mound where burial and apotheosis first of Galerius’ mother Romula and then of Galerius himself took place. So, the palace and sacred complex were, according to this interpretation, unique and complementary complex dedicated to the emperor and his mother, who themselves also became gods by the consecration act at Magura. Suggesting this hypothesis in the work, we convey in this book Srejovi} did not just analyze the ideological hypotheses of the tetrarchical system, but improved to the great extent the existing knowledge about the character of the reign of four co-rulers, two Augusti and two Caesars, promoted by Diocletian in the final decade of the 3rd century. Srejovi} also initiated the start of long-lasting investigations at nearby [arkamen, which turned out to be the palace-memorial complex similar to Romuliana, that was constructed by the emperor Maximin Daia probably for the same reasons as his uncle Galerius built Romuliana. These discoveries revealed that within rather short period of time at the beginning of the 4th century the emperors born in this area had built monumental palaces in eastern Serbia in the area outside main communications, and by the quirk of fate they never lived there. The archaeological and conservation works at Gamzigrad continued after the death of Dragoslav Srejovi} in 1996 and the new discoveries yielded valuable data about the life of Romuliana in the period after Galerius’ death, when the palace was transformed in the important economic center. The change in character of the structure at Gamzigrad is confirmed by its name Romulianum mentioned in the work Epitome, written around AD 380 and suggesting the conclusion that it was some kind of agricultural estate or settlement and not any more the divine Romula’s home, confirmed in the inscription FELIX ROMULIANA from the very beginning of the 4th century. On the other hand, the results of the Serbian–German investigations of the area outside the ramparts, that have commenced in 2004, revealed that there was a substantial settlement there before the construction of the imperial palace at Gamzigrad and that burying in the necropoles located to the south of the ramparts continued from the end of 3rd – beginning of the 4th century until the end of the 4th century – first half of the 5th century. In the mid 6th century the Huns ravaged Romuliana, which experienced its new revival in the Early Byzantine period, when many churches had been built there. The place was once again called Romuliana (Procop., De aedif., IV, 6, 19), but without any epithet indicating its particular importance. Therefore, this settlement, as well as the later Slavic settlements, remained in the shadow of the magnificent palace, which was built by emperor Galerius and whose today visible remains as a testimony about the ruling and ideological policy of this emperor have been proclaimed in 2007 as the monument of World Cultural Heritage protected by UNESCO. 9 MAJA @IVI] GAMZIGRAD: NAME, POSITION AND ECONOMIC POTENTIAL The origin of the name Gamzigrad could not be found either in Greek or Latin, so it seems most plausible that according to the assumption of P. Skok1 the word Gamzigrad is an imperative compound word consisting of the word gam from Proto-Slavic lexis from which originate Serbian words gamziti and gmizati (both meaning crawling). This explanation corresponds with the results of archaeological excavations, as the Slavs found in the Crna Reka valley the ruins of an anonymous town abandoned for centuries with many snakes crawling in these ruins. In the first edition of the Rije~nik by Vuk Karad`i} from 1818 Gamzigrad is not mentioned, but in the edition published in 1852 it says: “Gamzigrad in Crna Rijeka, walls of an ancient town. It is said that many snakes of all kinds are crawling in the ruins”.2 The antique name of Gamzigrad was, as we can see, entirely forgotten in the time of settling of the Slavs. Also, the Slavic, i.e. Serbian toponym Gamzigrad had not been mentioned in the sources written before the 19th century.3 Gamzigrad is located in the Crna Reka valley (map 2), which is a small geographic entity. The valley is surrounded by Vr{ka ^uka in the east, the high plateaus of the Ku~aj Mt. in the west, Crni Vrh, Deli Jovan and Stol in the north, while the ridges of Rtanj and Tupi`nica make the southern border. Crna Reka is characterized by the distinct relief and all other areas within this region are of the same character. It is surrounded by the hills consisting of volcanic masses, mostly of various kinds of 1 — Skok 1971, 548. andesite.4 Their height and distribution 2 — Vuk Karad`i} in his Rije~nik does not mention, however, that the nearare, however, of the kind that in no way diby Walachian village is also called Gamzigrad and that thermal springs in minishes the beauty of green meadows and Timok, present day Gamzigradska Banja, are also named after Gamzigrad. fertile fields at the foothill. It means that the “ruins of the ancient town” had been called Gamzigrad Crna Reka is not strictly confined geobefore the same name was given to the village and thermal springs. graphic entity, but it is just a distinct part of 3 — In contrast to this, some villages in the immediate vicinity of Gamzigrad, e.g. Zvezdan, are mentioned in the 15th–16th century Turkish defters, as well the larger area – eastern Serbia. On the other as in the maps of that time and later. hand, it is widely open only towards the con4 — Already in 1861, German mineralogist A. Breithaupt called the local fluence of the rivers Crni and Beli Timok, variant of green-gray andesite, gray amphibole, of which the Gamzigrad walls i.e. towards the Zaje~ar plain. The approach and most of the structures within it were built gamzigradit (Breithaupt from the south is closed by the ridges of Rtanj 1861, 51–54). 11 MAJA @IVI] 0 MAP 2. 10 km Geomorphological characteristics of eastern Serbia river for intrusion of different ethnic groups, primarily of barbarian descent, that rushed to the Balkans from the lower Danube Basin. Already in the earliest days of prehistory and particularly in the Bronze Age period, close connections had been established via the gateway at the Timok mouth between east Serbia and the west part of the lower Danube Basin, including the areas of present day southwest Romania and northwest Bulgaria. These connections had been interrupted in the time when the regions of east Serbia had been conquered and permanently occupied by the Romans, closing thus the gateway and Tupi`nica, while the saddle ^estobrodica separates Crna Reka from the Morava Valley. The main communication through Crna Reka extends along the geographic parallel: in the east it directly enters the Veliki Timok Valley and in the west the Velika Morava Valley, so in that way it is connected directly to two main communications of meridian direction in the central part of the Balkan peninsula. The eastern Serbia as wider region is known from the ancient times for its natural resources, primarily for its minerals and ores and for its open gateway at the mouth of the Timok 12 GAMZIGRAD: NAME, POSITION AND ECONOMIC POTENTIAL The period of cultural renaissance of the east Serbia took place during the second half of the 3rd century. The Crna Reka region and the Timok Basin experienced sudden cultural advancement: new settlements were established, the old roads were used once again and the new ones were constructed, the mines were opened and fortifications on the Danube banks were built. This great cultural and economic boom started immediately after the withdrawal of the Romans from Dacia – after the year 272, when the frontiers of the Empire were reestablished along the right Danube bank.5 Soon after that the old provinces had been divided into new administrative units, so the regions of east Serbia were included within three newly organized provinces: Crna Reka with the Veliki Timok Valley was included in the province Dacia Ripensis, the regions to the west of the line Iron Gates – ^estobrodica were included in the province Moesia Prima, while the territories to the south of the Rtanj and Tupi`nica were included in Dacia Mediterranea province. These newly established frontiers were more natural than the earlier Moesia Superior frontiers, as cultural and ethnic entities established already in the pre-Roman times had been encompassed within these new boundaries. At the same time the Crna Reka region was once again connected with the western section of the lower Danube Basin. On the other hand, the intensive exploitation of the natural resources of east Serbia also commenced at that time, as the at the Timok mouth for military and administrative reasons. The Romans, however, soon provided the passage to the southwest by constructing the entire system of fortifications along the Timok river, and thus the entire east Serbia was once again closely connected with the Morava Valley. These connections became particularly significant after AD 86, i.e. after establishing of the Upper Moesia (Moesia Superior) province with its centers disposed from the Morava and the Danube confluence to the upper course of the Vardar river (Viminacium, Margum, Naissus, Scupi). The newly established frontiers of the Roman administration (map 3) still did not provide the connection between east Serbia and the lower Danube Basin. The consequences were the slowdown in the development of traditional culture in the Timok Valley and thrusting east Serbia to the cultural periphery, because it was separated from the large urban centers. MAP 3. Balkan provinces of the Empire in Late Roman period 5 — The fortifications on the Iron Gates limes were partially neglected in the time of Roman rule over Dacia and they were restored particularly in the time of tetrarchy – in the end of 3rd and the beginning of 4th century and finally after abandoning because of barbarian campaigns in the beginning of the 5th and during the first half of the 6th century they were thoroughly restored, reinforced and even enlarged, in particular in the time of Byzantine emperor Justinian (527–565). The Iron Gates limes, because it was somewhat isolated by the Homolje Mts, must have had in its hinterland certain joint strategic strongpoint. Ni{ (Naissus) is rather far (although it should not be entirely written off, especially when we consider the broader limes region – from Belgrade as far as Vidin), but Gamzigrad (Romuliana) is much closer to the Iron Gates and the communications are much better. On the basis of investigations conducted at Gamzigrad it could be concluded with certainty that no larger military forces had been stationed there (except if they were not stationed in the neighboring castella), but Gamzigrad could have been the headquarters from where the operations at the limes were coordinated. The results of the recent archaeological investigations reveal that Gamzigrad was in the Late Roman and early Byzantine period an important metallurgical center where the weapons used in military operations at the Danube limes were certainly repaired if not produced. 13 MAJA @IVI] barbarians had certain impact on the intense building activity along the right Danube bank, the increase in army units, production and distribution of military supplies, increased production and storage of provisions. This resulted in the improvement of economic standard of population as it solved the problem of unemployment to the great extent. These economic conditions, as well as the changed historical circumstances in the end of the 3rd century, had the decisive impact on transformation of the Roman settlement at Gamzigrad and also on construction of the palace of emperor Galerius who had been born here. Romans tried to make up for the raw materials they needed after the abandonment of Dacia, first of all the ores (and particularly the gold from Transylvania). The abandoning of Dacia resulted in new immigration of population as the numerous Romanized inhabitants, in particularly those well-off, fled in front of the liberated Dacians and other barbarian tribes. They left their homes and estates on the left Danube bank and came to the northern sparsely populated regions of Moesia on the right Danube bank and settled in Dacia Ripensis. At the same time many of experienced Dacian miners and metallurgists also came to the area south of the Danube. The proximity of the 14 MAJA @IVI] HISTORY OF ARCHAEOLOGICAL INVESTIGATIONS AND CONSERVATION-RESTORATION WORKS AT GAMZIGRAD Over a century before the beginning of systematic archaeological investigations of Gamzigrad a few travelers and interpreters of the ancient past have recorded their encounter with this fascinating monument. First of them was the governor of the Saxon miners, baron S.A.W. von Herder, who in the book Bergmännische Reise in Serbien im Jahre 1835 provided the description and expertise of Gamzigrad. Janko [afarik, who carried out first authentic archaeological site-surveying in Serbia in 1846, suggested the excavations of the archaeological site near the village Gamzigrad. After him, but with no less enthusiasm, German geologist A. Breithaupt also wrote about Gamzigrad. Finally, Austrian archaeologist, historian and FIGURE 1. Gamzigrad in 1864, drawing by Felix Kanitz 15 MAJA @IVI] traveler Felix Kanitz, undoubtedly attracted by the magic of the place, visited the remains of the Galerius’ palace near the present day village Gamzigrad on two occasions (in 1860 and 1864) during his travelling through Serbia. He left us drawings and descriptions of the parts of walls and towers visible at that time, as well as the landscapes with which these strong palace guards are in perfect harmony (Fig. 1). These exceptionally important data are presented in his records about Serbia published in Vienna and Leipzig. There he emphasized in many passages that Gamzigrad is “one of the most splendid monuments of the bygone times” and “one of the largest and the best preserved monuments of the Roman architecture in Europe”. Initial Romantic enthusiasm for Gamzigrad had dissipated by the end of the 19th century. It has been replaced by highly simplified explanation of its strong ruins as the remains of large military camp (castrum) or the seat of an officer in charge of the gold mines in its surrounding (procurator metallorum). Even more so, during the entire first half of the 20th century, in the period of institution of archaeology in Serbia and its flourish, Gamzigrad disappeared in silence. Only in the 1950s, the period which could be characterized as “Neo-romanticism in the Serbian archaeology”, the interest for this unique monument was revived. Already in 1950 Dj. Bo{kovic made new ground plan of the Gamzigrad fortification, recording the position of Beginning of archaeological excavations at Gamzigrad – first excavated shovel, May 15, 1953 FIGURE 2. Vekoslav Popovi}, director of National Museum in Zaje~ar FIGURE 3. Letter of Vekoslav Popovi} to the District Committee of Zaje~ar from 28th January 1953, request for the beginning of excavations at Gamzigrad FIGURE 4. 16 HISTORY OF ARCHAEOLOGICAL INVESTIGATIONS AND CONSERVATION-RESTORATION WORKS AT GAMZIGRAD FIGURE 5. Vekoslav Popovi} and Djordje Mano-Zisi with archaeological team waiting for the train in May 1953 FIGURE 6. FIGURE 7. Djordje Mano-Zisi and Vekoslav Popovi} Dragoslav Srejovi} at Magura at the beginning of excavations in 1991 Ljubinka Vukovi} from 1968 to 1969, directed archaeological excavations at Gamzigrad. The renaissance of Gamzigrad has started in 1970, since when and until his premature death in 1996 the director of archaeological excavations was Professor Dragoslav Srejovi} (Fig. 7). He had much help in Anka Lalovi}, who was his devoted assistant in organizing the logistics and work for often very numerous and diverse professional team, which included the archaeologists Milivoje Veli~kovi}, Emil ^er{kov, Obrenija Vukadin, Smiljka Ka{i}, Ljiljana Bjelajac, Dubravka Nikoli}, Mira Ru`i}, Svetozar Jovanovi}, Miodrag Sladi}, Miroslav Lazi}, Stevan Djuri~i}, Djordje Jankovi}, Slobodan Fidanovski, Aleksandar Ba~kalov, Mila Petra{kovi}, Mirjana Petkovi}, Vesna Biki}, Gordan Janji}, Tonko Rajkova~a, Dragana Antonovi}, Svetozar Stankovi}, Viktor A}imovi}, Marko Vuksan, Moma Cerovi}, Emina Ze~evi}, Olivera Ili}, Anastasios Andonaras, Josip [ari}, Sofija Petkovi}, Ana Premk, Pero Pra{talo, Miroslav Vujovi}, Maja @ivi}, architects – ^edomir Vasi}, Svetlana Lazi}, art historian Vladimir Popovi}, geologist Vidojko Jovi}, conservator of the sculptures Milosav Pavelka and Vladimir Popovi}, art photographer. After Professor Srejovi} (director of the SASA project 1974–1996), director of the SASA project was Professor Milutin Gara{anin (1997–2002) with Dr. Petar Petrovi} as the coordinator of investigations (1997–1998), and after that, since 2002, the most important structures in its interior and at the same time suggested that it is necessary to investigate and protect this important Late Roman monument. Archaeological investigations and conservation-restoration works at Gamzigrad started on the 15th of May 1953 (Fig. 2). Whether by chance or the almighty Jupiter had his hand in it once again, but delving in the secrets of this place, which glorifies the members of his divine family, started just at the time when over 1,700 years earlier one great celebration had been in full swing. Namely, Diocletian, Jupiter’s earthly incarnation, organized the games to celebrate the great victory over the Persian king Narzes, ludi Persici, in the period between 13th and 17th of May in AD 288. The main creator of that triumph was just his adopted son and co-ruler, the founder of Romuliana, invincible Hercules – Galerius, Iovii fillius. We owe our appreciation for rescuing this unparalleled monument from further decay and for its presentation to the world to the stubborn determination of that time director of the Museum in Zaje~ar Vekoslav Popovi} (Figs. 3, 4). His one time professor at Royal Art School Djordje Mano-Zisi, at that time the curator of the Department of Byzantology in the National Museum in Belgrade (Figs. 5, 6), accepted the position of the director of investigations and carried it out until 1960. After him, first Dr. Djordje Stri~evi} from 1960 to 1963, and then 17 MAJA @IVI] FIGURE 8. project director has been Professor Slobodan Du{ani} with Dr. Miloje Vasi} as coordinator (1997–2007). The director of conservation-restoration works on discovered structures was from 1953 to 1986 Professor Milka ^anakMedi} (Fig. 8), with enthusiastic help of V. Popovi} from 1953 to 1967. The director of conservation-restoration works since 1987 is Brana Stojkovi}-Pavelka. The conservation-restoration works on the mosaics had been directed by Milan Duha~ (1954–1957), Rajko Sikimi} (1958–1966), Milorad Medi} (1967–1993) and Vladimir Ra{i} (since 1993), while members of the conservation team were painter Vekoslav Popovi} (organizer of investigation and conservation 1953–1967), painter Milan Tufehd`i} and Vera Toma{evi}, Milivoje Grbi}, Gordana Cvetkovi}-Toma{evi}, Djordje Mitrovi} and Bla`a Jankovi} (Figs. 9, 10) In the period between 2004 and 2008 as a result of the international cooperation with Römisch-germaniche Kommission des Deutchen Archäologischen Instituts from Frankfurt the test-trench excavations outside the Romuliana walls were conducted in order to confirm the results of previously conducted geophysical prospection (geomagnetic and geoelectrical measurements). Archaeological excavations in the area to the south of the rampart were carried out in the years 2005 and 2006 to in- Conservation of imperial palace ramparts at Gamzigrad in 1960s Removing of mosaics from the palace floors in 1960s FIGURE 9. FIGURE 10. Cleaning of removed mosaics in 1960s 18 HISTORY OF ARCHAEOLOGICAL INVESTIGATIONS AND CONSERVATION-RESTORATION WORKS AT GAMZIGRAD presented by the architect Brana Stojkovi}-Pavelka. Project participants on behalf of Germany are Römisch-germaniche Kommission des Deutchen Archäologischen Instituts, director Dr Gerda Sommer von Bülow; Architekturreferat der Zentrale des Deutchen Archäologischen Instituts, Berlin, represented by Dr Ulrike Wulf-Rheidt, architect; Thüringisches Landesamt für Denkmalpflege und Archäologie, Weimar, members of its team being Dr Tim Schüler, geophysicist and Mark Opelt; Lehrstuhl für Vermessungskunde, BTU Cottbus, represented by Rex Haberland and FHTV Berlin, represented by Alexander Pfützner. vestigate the section of the Late Roman necropolis and in 2007 and 2008 an area next to the north rampart was investigated by test-trenching. According to the Contract of international cooperation, participants of the project on behalf of Serbia are Institute of Archaeology in Belgrade, field director Dr. Sofija Petkovi}; Faculty of Philosophy of the Belgrade University, field director Dr. Mihailo Milinkovi}, National Museum in Zaje~ar represented by archaeologist-curator Maja @ivi} MA and Office for Protection of Cultural Monuments of the Republic of Serbia in Belgrade re- 19 MIROSLAV LAZI] PREHISORIC SETTLEMENTS AND NECROPOLES AT GAMZIGRAD AND IN ITS SURROUNDINGS The picturesque valley of the Crna Reka river, surrounded by ancient volcanic mountains which transform into wooded hills, rolling meadows and fertile fields, had always offered to its prehistoric inhabitants everything they needed: guaranteed harvest, rich pastures for their flocks, diverse game, valuable copper, silver and gold ores, as well as thermal mineral springs to achieve the vitality and enhance their pagan spirituality. Because of that it was the homeland of many ancient Balkan peoples who accomplished there magnificent cultural progress – motivated by favorable natural environment of this area and by close connections with their neighbors. The Gamzigrad palace FELIX ROMULIANA is situated in the heart of the Crna Reka region, so it is no wonder that many traces of diverse prehistoric communities have been discovered in its vicinity (Map 1). The first evidence of the Gamzigrad inhabitants from the pre-Roman times have been discovered in the 1960s and 1970s – by chance or in the course of investigations of the Roman palace. These were the fragments of different pottery vessels accompanied by the ground stone tools and rare bronze objects.1 The information that there are important prehistoric sites in the vicinity of Gamzigrad palace prompted academician Dragoslav Srejovi} to start the archaeological excavations of these sites in the final decades of the 20th century. In the period between 1989 and 1997 few sites were investigated, including the Bronze Age necropolis on the Magura hill near the Gamzigrad palace (Map 1/19), multilayered prehistoric settlement at Banjska Stena near Gamzigradska Banja (Map1/19), Eneolithic settlement on the Beligovo hill near Gamzigradska Banja (Map 1/8), rock-shelter dwelling place near Gamzigradska Banja, dating from the Bronze Age (Map 1/10), Bronze and Early Iron Age settlement in the area called Miletov Bunar near Rimski Majdan (Map 1/21) and Bronze Age settlement in the village Zvezdan (Map 1/18). The archaeological material gathered in this phase of investigation indicated entirely new evidence about the distant past of Timo~ka Krajina. The most important was the conclusion that necropolis at Magura and settlement at Banjska Stena had been established in the Bronze Age period by the bearers of distinct (so far unknown) culture, which achieved its complete development in Timo~ka Krajina. 1 — Srejovi} 1983 A, 19–21. It has been identified as Gamzigrad culture 21 MIROSLAV LAZI] MAP 4. Prehistoric sites in Timo~ka Krajina The beginning of this period, during the 4th millennium BC, witnessed the final rise of the Paleo-Balkan farmers, who faced with devastating effect of the development of early copper metallurgy on their traditional economy and also under increasing pressure of the first Indo-Europeans arriving from the south Russian steppes achieved the distinctive cultural unity. In other words, in order to preserve the endangered ancestors’ heritage and to survive, the united farmers from Oltenia in the north to Albania and Pelagonija in the south, established widely distributed complex of related cultures known in archaeology as Bubanj–Sãlkuþa–Krivodol. Thus united, the Paleobalkan peasants created their own world based on and defined as a distinct phenomenon in the prehistory of Serbia.2 The new phase of investigations commenced in 2001 with comprehensive site-surveying of the immediate vicinity of the Late Roman palace, when over 30 sites from different prehistoric periods have been discovered and precisely mapped (Map 1).3 Despite the fact that many prehistoric sites have been investigated and detailed archaeological prospection of the Gamzigrad vicinity has been carried out, the traces of Neolithic cultures have not been confirmed so far in that area. However, considering that Neolithic settlements (Proto-Star~evo, Star~evo and Vin~a cultures) have been discovered at Bor and Knja`evac, it is reasonable to assume that such settlements did exist also in the vicinity of Gamzigrad during the 6th and 5th millennia, as the environmental conditions there were favorable for the development of the early farming cultures.4 The earliest prehistoric settlements documented at Gamzigrad date from the advanced Eneolithic period (Copper Age). 2 — Srejovi}, Lazi} 1997; Lazi} 1998. 3 — Unpublished. Documentation in the Center for Archaeological Investigations of the Faculty of Philosophy in Belgrade and the Office for Protection of Cultural Monuments of the republic of Serbia in Belgrade. 4 — Jovanovi} B. 2004, 33–55; Sladi}, Jovanovi} 1997, 167–175. 22 PREHISORIC SETTLEMENTS AND NECROPOLES AT GAMZIGRAD AND IN ITS SURROUNDINGS FIGURE 11. Site Banjska Stena of animal bones have also been found.6 Similar situation was encountered also at Banjska Stena, where the remains of modest settlement from the same period have been discovered in the lowest layers. The inaccessibility of these settlements, the character of the impoverished dwelling places and the accompanying archaeological finds reveal that in the vicinity of Gamzigrad in the end of the 4th millennium BC lived the last offsprings of the bearers of this cultural complex, who disappeared completely in the complex and dynamic ethnocultural mixtures of the ensuing period. In the beginning of the 3rd millennium BC the Coþofeni culture was established in Oltenia, northwest Bulgaria and east Serbia. The pottery of that population is recognizable for the ornaments executed by incision, band impression and the spiritual experiences, inherited from their ancestors – skillful farmers and master craftsmen. In the fertile river valleys and the Balkans hilly terrains they produced around the middle of the 4th millennium the supreme artistic and artisan objects – lavishly decorated pottery, often with gilded surfaces, magnificently made anthropomorphic figurines, variously modeled ritual vases and the like.5 The settlements of the population of the Bubanj–Sãlkuþa– Krivodol cultural complex have been discovered in the vicinity of Gamzigrad at the sites Beligovo and Banjska Stena in the immediate vicinity of Gamzigradska Banja (Map 1/8, 9). These were rather small settlements established at inaccessible elevations along the right bank of Crna Reka. As the investigations at Beligovo revealed, the houses were of rectangular plan, with stone foundations and walls made of logs and wattle plastered with daub. The pottery was made of refined clay with addition of fine sand. The prevailing shapes are various bowls and pyriform amphoras with roughened outer surface. The tools made 5 — Tasi} 1979, 87–114. 6 — Nikoli}, \uri~i} 1997, 79–88 23 MIROSLAV LAZI] FIGURE 12. Group of funerary structures at Magura FIGURE 13. Funerary structure at Magura The first more detailed information about the Bronze Age in Timo~ka Krajina was reached in the 1980s after the excavations at site Trnjane near Brestova~ka Banja, where around twenty circular structures made of stone and with pottery urns in the center have been investigated. The settlement has also been discovered in the immediate vicinity of this necropolis.10 Few years later, funerary structures made of broken stone with urns in the center (Figs. 12, 13) have also been investigated on the hill Magura – one kilometer to the east of FELIX ROMULIANA palace and under the layers from the antique period (Map 1/19). The results of excavations at Trnjane and at Magura made possible attribution of the Bronze Age sites in Timo~ka Krajina to the Gamzigrad culture, dated to the period from the 18th to the 13th centuries BC. The bearers of this culture were recognized as stock-breeders, farmers and metallurgists who established close contacts with their contemporaries in the Morava Valley, Banat and Oltenia.11 Many settlements of the Gamzigrad culture have been encountered in the immediate vicinity of the Gamzigrad palace and test excavations have been carried out at some of them. series of the button-like appliqués. They were accompanied in east Serbia by communities of the Kostolac culture, originating from the Danube Basin and the Morava Valley. These two different ethnocultural groups achieved in eastern Serbia complete unity and established joint settlements discovered in the Iron Gates, in the vicinity of Majdanpek and near Bor. In their generally inaccessible settlements was found the pottery decorated in the Coþofeni culture style, but with techniques and motifs characteristic of the pottery decoration of the Kostolac culture.7 In the vicinity of Gamzigrad the joint settlements of the members of Coþofeni and Kostolac culture has been discovered at the Banjska Stena hill near Gamzigradska Banja (Fig. 11). The investigations of this settlement, however, did not yield enough informations about its size and characteristics. There was found just pottery, which, according to the method of production and decoration, does not differ from the contemporary pottery from other regions of the east Serbia. Somewhat later pottery from Banjska Stena, decorated with cord impressions, confirm that still another wave of immigrants from the east followed in this region of the Danube Basin by the end of 3rd millennium. The archaeological finds from east Serbia and the neighboring regions do not, however, offer enough data about the end of that dark and unstable period.8 Nevertheless, we know that after many tumultuous centuries the complete and unhindered development of the prehistoric communities in that region started only in the Bronze Age.9 7 — Tasi} 1997, 81–82; Tasi} 2004, 91–99. 8 — Roman, Dodd-Opriþescu, János 1992, 57. 9 — Srejovi}, Lazi} 1997, 235–236. 10 — Jovanovi} B., Jankovi} 1996, 185–201. 11 — Srejovi}, Lazi} 1997, 225–247; Lazi} 1998, 147–158. 24 PREHISORIC SETTLEMENTS AND NECROPOLES AT GAMZIGRAD AND IN ITS SURROUNDINGS covered the traces of the funerary rituals. Many scattered pottery fragments bear witness to the intentional (ritual) breaking of the vessels, and flat stones with incised spirals and netlike motifs found by the urns indicate the religious and mythological notions concerning the afterlife of the dead.12 The members of the Gamzigrad culture established close contacts with the inhabitants of the surrounding areas and that made possible the precise dating of this culture. The imported pottery originating from the Danube Basin and Oltenia, as well as the pottery imitating the products of the Vatin culture, Verbicioara III and Para}in culture, are dated in the first half of the 2nd millennium BC. Gamzigrad culture experienced fascinating progress during that period. In the rich and densely populated Timo~ka Krajina emerged distinctive forms of material and spiritual culture, evident on the pottery and recognizable in the funerary practice. Certain vessels (amphoras with distinctive handles) are encountered also in the northwest Bulgaria allowing for the assumption that Gamzigrad culture was present also in that area where it made the symbiosis with the Verbicioara III culture, whose home territories were in the south and southwest Romania.13 The forcible intrusion of the channeled pottery peoples from the north interrupted in the beginning of the 13th century BC (Br D/Ha A1) the development of the traditional Bronze Age cultures in the Danube Valley and the central Balkans. In one of the latest graves at Magura has been found a spearhead dating from that period, indicating the period of armed conflicts with the intruders. The outcome of these dramatic events is well-known. The settlements of the Gamzigrad culture perished in conflagrations including even the fortification at Banjska Stena, whose strong palisades did not withstand the attacks of the conquerors. The newcomers founded their own settlements on the burnt ruins of the previous settlements. It is confirmed by the characteristic channeled pottery of brown color and battle axe (celt) cast of bronze from the site Miletov Bunar and also by the similar pottery of the newcomers collected in the course of site surveying.14 After this tumultuous period, next few centuries witnessed the progress of the Early Iron Age cultures in east Serbia. As this period is not sufficiently studied in Timo~ka Krajina, general The spacious settlements were distributed over the gentle slopes along the banks of smaller waterways, near the springs and along the banks of the Crni Timok river (Map 1). The single fortified settlement with walls and wooden palisades had been established at Banjska Stena – an important strategic point controlling the approach to the large and fertile valley in the lower course of the Crna Reka (Map 1/9). The aboveground houses of the Gamzigrad culture, of rectangular plan and built of wattle and daub, have been discovered in the area of village Zvezdan (Map 1/18, 21). The pottery was made of clay with admixture of fine sand and fired to the brown and gray nuances. Massive pots for storing provisions and preparing food, decorated with the applied molded bands, predominate by quantity and size among the various pottery shapes. There were also large polished bowls with two antithetically modeled handles, as well as the beakers with one or two arched handles. The pottery of higher quality had been imported from the Danube Basin, as it is confirmed by the vessels from Banjska Stena originating from the territory of the Vatin culture in Banat and the Verbicioara III culture in Oltenia and northwest Bulgaria. In addition to the pottery there were also found stone tools (chipped stone blades, ground stone axes, whetstones, etc.), as well as the tools made of horns and animal bones (mallets, awls, perforators and the like). Despite the fact that there were also discovered the traces of processing the sulphide copper ore (characteristic vessels, slag, copper granules), it is conspicuous that there were no objects made of bronze. In contrast to the settlements dispersed in the lower zones, the necropoles of the Gamzigrad group are located on the dominant hills. The necropolis next to the sacred complex from the antique times was investigated at the Magura hill (Map 1/19, plan 1), and another necropolis (completely destroyed) was situated couple hundred meters to the north – near the tetrapylon (Map 1/20). These are large cemeteries of cremated individuals with pottery urns containing remains of the deceased and grave offerings buried in the center of funerary structures of circular shape consisting of carefully arranged broken stones. Few tens of such structures, 1.50 to 3.50 meters in diameter and with more than 80 urns, have been discovered at Magura (plan 1). The grave offerings in the burials are infrequent. Most often two-handled beakers and clay spindle whorls have been found. The objects made of bronze – one spearhead and one damaged arrowhead have been found in just two burials. Within the funerary structures were also dis- 12 — Srejovi}, Lazi} 1997, 232–233; Lazi} 1998. 13 — Berciu 1961, 123–161. 14 — Sladi}, Ru`i} 2001, 159–160. 25 MIROSLAV LAZI] 0 5 PLAN I 10 m Prehistoric graves inside and around consecration memorials at Magura 26 PREHISORIC SETTLEMENTS AND NECROPOLES AT GAMZIGRAD AND IN ITS SURROUNDINGS one handle and polished beakers decorated by channeling or with series of stamped motifs of horizontal letter S. From the same period also dates a fragment of double-looped fibula with spherical nodules (knots) on the bow.17 Although detailed analysis of the pottery from the sites dating from somewhat later phases of the Early Iron Age has not been performed, it is almost certain that the surroundings of the Gamzigrad palace has been inhabited from the 6th to the end of the 4th century BC by the bearers of the postBasarabi style, which is recognizable for its distinctive pottery decoration (tremolo). These were the Triballi – strong PaleoBalkan tribe mentioned in the writings of antique historians and geographers.18 The famous campaign of Alexander the Great in the Danube Valley in the year 335 BC and his encounter with the Celtic envoys indicate rapid decline of the Triballi, because only few decades later the Celts will absolutely dominate the expanses of Pannonia, middle Danube Basin and central Balkans, thus marking the end of Early Iron Age in this areas. conclusions could be drawn only on the basis of investigations in the Iron Gates and in the vicinity of Bor, as well as on the basis of scarce archaeological finds from the prehistoric layers of the Gamzigrad fortification or those found by chance in its vicinity.15 Investigations in the Iron Gates and Klju~ region revealed that more frequent use of iron resulted in dynamic ethno-cultural integrations, so on the pottery from those regions, dating from the beginning of the Early Iron Age, could be recognized characteristics originating from Pannonia and the Morava Valley (Kalaka~a cultural horizon) as well as from Romania (Ostrov group). The complete cultural unity in the territory of east Serbia was achieved some time later, in the 8th and 7th centuries BC, when the Basarabi style in pottery decoration prevailed over vast area from the south Russia in the east to the Alpine regions in the west.16 The pottery decorated in that way has been found within the Gamzigrad fortification and at few more sites in its vicinity (Map 1). The prevailing shapes are shallow bowls with inverted and facetted rims, conical cups with 15 — Vasi} R. 1997. 16 — Vasi} R. 1997, 93–94. 17 — Srejovi} 1983 A, 20, sl. 12 a–g. 18 — Papazoglu 1969, 11–68; Srejovi} 1983 A, 20. 27 MIODRAG SLADI] GAMZIGRAD IN THE PROTOHISTORY The territory of the village Gamzigrad encompasses the banks of the Crni Timok River. It is characterized by gentle and vivid landscape with many valleys and slopes, which abruptly descends from the surrounding hills towards the village. This is rolling and picturesque landscape pleasing to the eye. The generous nature offered abundance of water, much different game and the soil suitable for cultivation, providing for long and peaceful life in this area. Even today, if walking these picturesque and green areas, we could understand those people who wanted to settle here many thousand years ago. The same feeling certainly harbored also the inhabitants of one of the oldest settlements in the Gamzigrad region, the hillfort at Banjska Stena. The site is situated on the hilly promontory above Gamzigradska Banja, on the road towards Metovnica, and it was no doubt specially selected for the settlement because it has an exceptional strategic position, which made possible safe and long existence for the hillfort inhabitants. Even later, long time after the cessation of life in this fortified hillfort settlement, the configuration of the terrain still offered safe sojourn also to the Celtic visitors. The long-lasting excavations had been carried out at this important site and in one of the trenches investigated in 1996, the archaeologists discovered finds, which from cultural and chronological point of view could be dated in the Early La Tène period. There were found sparse fragments of the typical Celtic gray pottery made on wheel and also one zoomorphic fibula with the protome shaped as mythical hybrid animal on circular backward turned foot (Fig. 14). Despite being made of bronze and not of some precious metal, it is still fascinating. The amazing imagination of the artisan, or better to say artist, was realized to perfection considering also the highest precision of manufacture. The design of the backward turned foot and the protome on it is articulated in four basic elements: the mane on the bow, horns, eyes and trunk raised in such a way to touch the back of the head. The fibula bow is also originally stylized with entwined ropes resembling the net, or, more probably, the headstall.1 It seems when watching the fibula sideways that artisan wanted to represent the fish coming to the surface. The first-class design of the fibula is particularly emphasized by the small surface enclosing 1 — Sladi} 2003, 38. all these elements. 29 MIODRAG SLADI] FIGURE 14. Fibula from Banjska Stena FIGURE 15. Fibula from Vi{icina Ba{ta in the second half of the 4th century BC.14 Their distribution from that territory could be followed within a wide area, from Liter in Hungary as the westernmost point to Veliko Trnovo in Bulgaria, where is the easternmost point of their occurrence. We must conclude that fibula from Banjska Stena exceeds all to date discovered specimens in the imaginativeness of the designed motifs and persistence of the artist on the precise execution of details. Discovery of this fibula revealed that Gamzigrad territory was certainly on the route of penetration of the Celtic influences. However, although stay of the Celts at this location was short-lived, it left the indelible trace in the material culture. The fibulae of this type are rare if we take into consideration the extent of the territory where they have been found (around twenty specimens were discovered), but grouping of the fibulae indicate their home region. They have been discovered so far in Hungary at Sopron–Bécsidomb,2 Györ–Ujszàllàs,3 Liter,4 Pilismarot–Basaharc,5 Szentendre and Pűspőkhatvan,6 in the Sava Valley at Donja Dolina (three specimens),7 in the Serbian Danube Valley in Viminacium (two specimens)8 and one specimen each at Banjska Stena9 in Timo~ka Krajina, at Pecica in Romania and in Veliko Trnovo in Bulgaria.10 The importance of these decorative objects in the north Bosnia along the Sava river, where at Donja Dolina have been found few specimens of these fibulae, helped Z. Mari} to determine the time of first intrusions of the Celtic culture into the Bosnian territory, as he dated them in the period between the years 350 and 300 BC.11 On the basis of these conclusions B. ^ovi} also ascribed to these fibulae, among other things, the key role in the periodization of the Early Iron Age phase Donja Dolina – Sanski Most (3b), when the conditions were set for the acceptance of the La Tène culture in this area.12 N. Majnari}Pand`i} who, also, recognized in these finds the very first Celtic cultural intrusions into the Sava Valley also supported these conclusions.13 It could be easily said that all fibulae of this type share common stylistic trait, linking them to the same genetic core, which, judging by the number of discovered specimens, originated in the area of northern Pannonia and Romanian Banat 2 — Hunyady 1942–1944, T. XVIII,5; Szabo 1971, 19, fig. 3. 3 — Mir~eva 2004, 166. 4 — Marton 1934, T. V, 2; Szabo 1971 A, 138. 5 — Mir~eva 2004, Obr. 3. 4. 6 — Szabo 1971, pl. 3; Szabo 1975, 71–86. 7 — Truhelka 1904, 3–27. 8 — Popovi} P. 1996, 109, fig. 4/13. 9 — Sladi} 2003, 39, sl. 1. 10 — Mir~eva 2004, Obr. 1. 11 — Mari} 1963, 67. 12 — ^ovi} 1987, 260–266. 13 — Majnari}-Pand`i} 1996, 37. 14 — Szabo 2001, 59. 30 GAMZIGRAD IN THE PROTOHISTORY made bowls with “S” profiled rim and the coarse hand-made pottery, first of all the so-called situla-type pots. It is wellknown that Roman culture at that time still accepted local tradition until the 2nd century, since when it aggressively imposed its cultural achievements. Therefore, this period when traditional pottery shapes still continue, is rather often also called the Latenized phase. This is, in other words, the time of downfall of the culture of the Scordisci and their gradual assimilation into the Roman civilization, which, step by step united the entire area where the Roman administration was established and it is known that the territory of present-day Timo~ka Krajina was included in the Upper Moesia province. There are still few more questions for the archaeologists. First of all, where the discovered objects come from, from the nearby vicus or from some other urban center and who were their producers? In searching for answer to the first question we have the problem of insufficiently investigated area. The dilemma concerning the second question is whether these people were the Timacii from the Timok Valley or Picenses from the valley of the Pek River. Judging by antique sources, which provide relatively precise demographic picture and by archaeological material discovered at the site, the Timacii were not in the time we are speaking about such an important tribe, which could leave important trace in the cultural history of this area, because they are not mentioned in the Ptolemy’s list of tribes living in the Upper Moesia in the beginning of the 2nd century AD.19 Therefore, our choice is the Picenses, who were the mixture of the Scordisci and the resettled Dacians,20 as it is also confirmed by the archaeological finds. The Picenses, who were good miners and metallurgists, reached in their quest for the resources perhaps as far as this part of Timo~ka Krajina. The finds from Vi{icina Ba{ta are considerable and significant new information suggesting that there were settlements in this region in the beginning of the Roman occupation, i.e. much before the period of tetrarchy, when the palace of Galerius as the greatest cultural symbol of this area had been built. *** After the first wave of immigration, which was some kind of a vanguard, the Celts did not return to the Gamzigrad area, as it could be concluded considering absence of their traces in the later periods. Many centuries had passed before the elements of their culture recorded at the location known as Vi{icina Ba{ta appeared again in this region. The site is situated in the northeast periphery of the village Gamzigrad at the end of the street leading uphill towards the area called Varzari, on the estate of Vi{ica Jovanovi}. The material has been found by pure chance in the course of digging a pit in the yard and thanks to the scrupulous finder it has been culturally evaluated and published.15 The find includes fragments of pottery vessels and few metal objects among which is particularly interesting one in many ways unique wire-made fibula (Fig. 15). Even though archaeological context of the find is unknown, as material was most likely washed away from the higher ground, it has been concluded that material dates from the period which has not been confirmed in this region before. The stylistic and typological analysis of the pottery material, making most of the found objects, provided general direction for cultural and chronological determination of the complete find. The pottery shapes and technology of manufacture indicate that the material dates from the end of 1st and the beginning of the 2nd century AD. The identical pottery vessels have been recorded throughout the areas where the Roman culture penetrated, first of all in Srem, Ba~ka16 and the Morava Valley,17 where many sites with almost identical pottery have been investigated. According to O. Brukner such vessels are characteristic of the Pannonian production in the time of the Flavians, with conspicuous elements of the La Tène culture taken over from the Scordiscan and to a certain extent from the Dacian tradition.18 It is mainly suggested by the vessels, which greatly resemble the pottery produced in the workshops of the Scordisci. We actually think of large repertoire of the wheel- 15 — Sladi} 2005, 211–222. 16 — Brukner 1987, 25–44; Brukner 1995, 91–136. 17 — Madas 1970, 133–134. 18 — Brukner 1987, 36. 19 — Papazoglu 1969, 328. 20 — Mirkovi} 1968, 38. 31 SOFIJA PETKOVI] ROMAN GAMZIGRAD BEFORE THE IMPERIAL PALACE The archaeological investigations at Gamzigrad confirmed the existence of Roman settlement before construction of the palace of emperor Galerius. The sporadic data, obtained by earlier excavations, have been interpreted as the remains of farming estate (villa rustica) founded in the middle of the 3rd century.1 However, the analysis of structures partially discovered in the 1980s in the sector of the Jupiter temple, in front of the entrance to the south tower of the west gate of the later Romuliana fortification (tower 19) and in the sector of east gate, in front of the south tower of the earlier fortification (tower I), as well as the more recent archaeological investigations in the sector of thermae and outside the palace walls shed light on the Roman Gamzigrad in quite a new way (plan II). After almost three decades of investigations of the Roman Gamzigrad, portion of large building, 11.5 x 10.5 m in size and oriented in northwest-southeast direction, was discovered in 1981 to the south of the Jupiter’s temple in the south section of the fortification. There were encountered two longitudinal rooms, around 4 meters wide each, and in the east longitudinal “corridor” 1.5 m wide, that was most probably the connection with the central atrium. The rooms were entered from the “corridor” through the doors, 0.90 wide, and with brickbuilt doorposts. The walls of the structure, 0.55 m thick, were built of brick and stone in the opus mixtum technique, while the foundations were of rubble stone embedded in lime mortar.2 The building was demolished in the beginning of the 4th century, during the construction of imperial palace, i.e. of the large structure and south portico of the temenos of Jupiter’s temple. Bronze coins minted in the time of emperors Aurelian and Probus were found in the layer of soot, above the mortar floor of the “corridor”, and they determine the terminus post quem non for construction of this structure, i.e. it was abandoned during the final quarter of the 3rd century. On the basis of finds from the rooms, including the fragments of pottery, lamps and bronze fibulae (Fig. 16) dating from the 2nd and the beginning of the 3rd century, it could be assumed that this villa rustica, as identified by Dragoslav Srejovi}, had been built in 1 — Srejovi} 1983, 21–23; Srejovi}, Vasi} 1994, 56–57. the first half of the 3rd century3 (plan III,1). 2 — Srejovi} 1983, 21–22, sl. 14. In front of the entrance to the south 3 — Srejovi} 1983, 23, sl. 13. tower of west gate of later fortification 33 SOFIJA PETKOVI] PLAN II Structures from the phase before Galerius’ building activity within fortified palace north was enclosed a small rectangular room, 1.90 m x 0.90 m on the outside and 0.80 m x 0.80 m on the inside, with remains of the vault (channel) extending to the south, and this was most probably the praefurnium, the mouth of furnace of the hypocaust heating system. The structure extends further to the east and south and considering that its east part overlies the remains of rampart of earlier fortification, it most probably dates from the period before construction of the later fortification of imperial palace and after destruction of the earlier one. Whether the assumed thermae were the segment of some larger structure (villa), or they were just independent public baths, is impossible to determine on the basis of investigated section of the structure. It is also impossible to determine the over-all plan of the complete building (plan III, 2). The objects, including fragments of ceramic and glass vessels, iron tools and damaged zoomorphic terracotta, dating from the 3rd century, have been found in the investigated rooms of the building. (tower 19) in the area of portico along the west rampart the structure with the remains of the hypocaust system was partially investigated in 1986. This structure extends in the northsouth direction and one of the pillars of the portico of later fortification damaged its west wall preserved only in the foundation zone. The north, east and south walls of the structure, built of stone and tegulae in the opus mixtum technique and 0.55 m thick, are better preserved, at some spots up to 0.90 m above the floor level. The discovered northwest section of the building ground plan, covering around 36 square meters, resembles the Roman baths (thermae). In the northwest corner is one room, 2.20 x 1.90 m, and extending along the north wall is another one, 1.70 x 1.20 m in size, and both with the hypocaust heating system. An apse, 2.40 m in diameter, was added on the outside of the north wall. The remains of floor made of hydraulic mortar were recorded between the east wall of the structure and small central room with hypocaust. In the south wall towards the 34 ROMAN GAMZIGRAD BEFORE THE IMPERIAL PALACE 0 5 10 m 1 0 10 20 m 0 5 10 m 2 0 10 20 m 3 4 1) Structure in the area of Jupiter’s temple, villa, built in 3rd century; 2) structure in the area of west gate, thermae or villa, built in the period between construction of earlier and later palace fortification; 3) structure in the area of east gate built in the 3rd century, before construction of earlier fortification; 4) structure in the area of thermae built in the 3rd century, before construction of Galerius’ thermae PLAN III because of the weight of tower staircase vault. The northeast corner of the structure with two rooms separated by the partition wall running in the east-west direction has been discovered. The walls are built of half-dressed and rubble stone lime mortar. The interior wall face is coated with layer of lime mortar with The building dating from the time before the construction of this fortification has been partially investigated in front of the south tower of east gate of earlier fortification (tower I). The south pilaster of the tower I gateway is leaning on the north wall of mentioned structure, which had sunk into the ground 35 SOFIJA PETKOVI] FIGURE 16. Bronze fibulae (2nd–3rd c.) found in the structure in the area of Jupiter’s temple under the floor of stone slabs has been recorded the mortar floor decorated with mosaics, then the leveling horizon of yellowbrown clay that contained fragments of prehistoric and Roman pottery vessels and, finally, the mortar floor of the earlier structure the and layer of greenish-brown clay, identified as the virgin soil. The earlier level of mortar floor most probably dates from the earlier phase of the Roman building activity at Gamzigrad. Already in 2005 the investigations carried out in the area between the caldarium (room with hot baths) and tepidarium (room with tepid baths) of the thermae and next to its east façade brought to light the walls of earlier Roman structure. The objective of the excavations was to define the partition wall between praefurnium (mouth of furnace) and sudatorium (sweating room) of Galerius’ thermae, and on that occasion a wall around 5 m long and running in the north-south direction has been investigated. The wall, 0.90 m thick, leveled when the chuff. There is a door opening in the partition wall with doorposts built of brick and threshold of ashlars bonded by mortar. The walls, around 0.80 m thick, are partially preserved up to the 0.90 m from the socle, i.e. up to the level of the mortar floor. The foundation zone of the walls made of stone rubble laid in mortar is even wider than the above ground wall segment (plan III, 3). The foundations of the structure are embedded into the layer of greenish-brown clayey soil that contained fragments of Roman and prehistoric pottery. In the building interior were found four bronze coins greatly damaged by patina and most probably minted in the mid 3rd century, two pottery lamps, one footed beaker and two three-handled vessels (Fig. 17). The foundation zone of the wall of earlier structure, built of broken stone laid in lime mortar, was discovered between the north pillar of the earlier fortification portico that flanks the entrance to the north tower of east gate (tower II) and southeast corner of the “building with corridor” of palace II. Unfortunately, it is not possible to determine the size and purpose of this structure on the basis of the discovered wall (plan III, 3). During the more recent archaeological excavations carried out between 2002 and 2005 in the sector of thermae, in the course of investigations of the baths from the phase of construction of emperor Galerius’ palace, the parts of an earlier building were encountered under this structure. The walls of earlier structure were discovered in the east section of Galerius’ thermae. They were partially used as the foundation for the imperial thermae and they were built of broken and halfdressed stone laid in lime mortar. They are 0.60–0.90 m thick and preserved in the foundation zone or up to 0.50 m above the socle, i.e. the floor level. In the course of test-trench excavations carried out in 2002 in the apodyterium (dressing room) of the Galerius’ thermae, Pottery vessel from 3rd century found in the structure in the east gate area FIGURE 17. 36 ROMAN GAMZIGRAD BEFORE THE IMPERIAL PALACE Foundation of podium of square plan in the area of thermae, from the northeast FIGURE 18. Water channel from the 3rd century negated by pillar of south portico of earlier fortification and south façade of Galerius’ thermae, from the west FIGURE 19. South portico of earlier fortification with water channel from 3rd century, from the west thermae had been built, is preserved above the ground as one course of ashlars bonded by lime mortar. Another wall, 0.60 m thick and running in the east–west direction represented by two courses of ashlars bonded by lime mortar is added to its south end, and this wall is negated by east façade and apse of the caldarium of Galerius’ thermae. The analysis of the foundation walls of thermae on the east façade has confirmed that section of an earlier building existed in the zone of caldarium and frigidarium (room with cold pool) and the above described walls belonged to that struc- ture. The rectangular annex of the structure, 12 x 8 m in size, is divided by partition walls into three rooms and it extends in the east into a larger room (atrium?). The building in the zone of apodyterium extended also to the west as it is confirmed by the mentioned discovery of mortar floor (plan III, 4). Next to the southeast corner of the earlier Roman building, outside the Galerius’ thermae, the remains of the square foundation, 3.8 x 3.8 m in size, most probably the podium for smaller cult structure, altar or statue have been investigated. This structure is negated by the later Galerius’ buildings, including FIGURE 20. 37 SOFIJA PETKOVI] Southeast corner of the portico of earlier fortification and junction of water and drainage channels into main sewage system (cloaca), from the northwest FIGURE 21. Southeast corner of Galerius’ thermae, pillar of south portico of earlier fortification, water channel from the 3rd century and older lead pipe negated by its construction (2nd–3rd c?) FIGURE 22. mae. This has been established considering the relations between the channels and pillars of the portico of earlier fortification (Fig. 19). The water supplying channel running in the east–west direction was negated by the construction of the pillar of earlier fortification south portico and at the west end, where it turns towards northwest, it is interrupted by the south façade thermae, and it is covered with mortar substructure of the floor of that phase (Fig. 18).4 The objects, which could be ascribed to the earlier Roman structure, are very scarce due to leveling of the terrain when the palace was built and include the fragments of the 3rd century pottery and glass vessels. The southeast corner of the portico of earlier fortification, as well as the system of water pipes and drainage channels predating the construction of the palace, were investigated in 2004–2005 and in 2007 to the south and east of Galerius’ ther- 4 — Petkovi} 2008 B, 66, sl. 2. 38 ROMAN GAMZIGRAD BEFORE THE IMPERIAL PALACE FIGURE 23. FIGURE 24. Terracotta from the area of thermae, head of Dionysus, 3rd century Group of bronze fibulae found in southeast section of fortified palace (thermae area ?), end of 1st – first half of 3rd century early imperial period, from the end of 1st to the middle of the 3rd centuries were discovered in the leveling layers under the floors of Galerius’ buildings. Although not abundant, they bear witness to the life at Gamzigrad in the 2nd and first half of the 3rd centuries. Most important among the finds from that period is the group of fibulae found in the southeast section of the fortified palace (Fig. 24).7 The later fortification of Romuliana had, besides main east gate and west gate that marked the line of main palace communication (cardo), two more smaller gates: one in the east rampart, between tower 3 and southeast corner tower (tower 5), and one in the north rampart, between tower 8 and tower 10 (plan II). In the light of most recent discoveries the purpose of north “small gate” could be explained as a way of communication between the remaining section of the earlier Roman settlement at Gamzigrad and the imperial palace. Besides the traces of the Roman settlement within the fortification dating prior to the palace construction, the archaeo- of Galerius’ thermae and its cavity was carefully closed with the fragments of tegulae and mortar (Fig. 20).5 The parallel water and draining channels running in the northwest–southeast direction unite with the mentioned water supply line into wider channel (cloaca), which diverted excessive water from the water system and the waste water towards the east, into the present day Draganov potok (Fig. 21). The lead water pipe running in the north–south direction was discovered in the sector of thermae. At its north end the pipe was disassembled in the course of construction of already described water channel of the east–west direction. This pipe was the component of the system of siphons, which provided water for the Roman settlement before the construction of earlier fortification (Fig. 22).6 The layer of yellow-brown and greenish-brown clay into which had been buried the mentioned channels and the lead water pipe does not abound in portable finds, of which particularly important are the coins from the middle and second half of the 3rd century (Gordian III, Valerian, Trajan Decius, Florianus, Aurelian, Probus), fragments of the 3rd century pottery and glass vessels, lead mirror and fragment of ceramic appliqué representing the head of Dionysus (Fig. 23). In the course of archaeological excavations of the fortified imperial palace Felix Romuliana, the objects dating from the 5 — Petkovi} 2008 A, 61–62, sl. 5–6. 6 — Petkovi} 2008 B, 66, sl. 5. 7 — Petkovi}, @ivi} 2006, 439–458. 39 SOFIJA PETKOVI] explored by test trenching in 2007. The central building is a podium, around 4 m in diameter, with a “crypt” in its interior, most probably for smaller cult structure, triumphal column or monumental statue. The podium of broken stones laid in lime mortar had ashlar facing, which is nowadays missing. On the podium are the remains of cylindrical building, 3.50 m in diameter, made of bricks bonded by lime mortar (Fig. 25). There were also investigated two of 16 column bases, 2 x 2 m in size, built of ashlars bonded by lime mortar. These bases supported masonry pillars or monumental columns around 8 meters tall (Fig. 26). Partially investigated circular building of imperial character was demolished and leveled most probably in the time of construction of Galerius’ palace. In the leveling horizon of brown clay were encountered the fragments of prehistoric pottery, from the Neolithic and the Bronze Age, atypical fragments of Roman pottery and glass vessels and iron objects. Among the scarce objects found in the debris of demolished structure worth mentioning are two tegulae with the stamp of 4th Flavian legion (legio IV Flavia), that could be most probably dated in the 3rd century. Because of the scarcity of finds it is difficult to determine when this indubitably important structure had been built. It could have been erected to honor significant imperial military campaigns and victories in Upper Moesia or in the neighboring provinces. Therefore, two chronological periods are possible: time of Domitian’s and Trajan’s Dacian wars at the turn of the 1st to the 2nd century, and the victory over the Goths achieved by Claudius Gothicus and Aurelian in 268–269.9 These events emphasized the importance of territory on the right Danube bank in Upper and Lower Moesia and emperor Aurelian established new province known as Dacia Nova or Dacia Ripensis. In addition to the military-strategic importance, the abundance in ores, intersections of the main roads of the Roman empire and migration of affluent population from the Roman Dacia made possible establishing of large farming estates (villae rusticae) and restoration of old and foundation of new towns, mining-metallurgical and business centers (vici, civitates).10 logical prospection of the area extra muros by geophysical methods conducted in 2006–2007 revealed the ground plans of structures next to the north Romuliana rampart (plan IV).8 Among the recorded structures of residential and economic character particularly distinguished in the settlement to the north of the palace was the structure of circular plan, around 35 meters in diameter, with circular peristyle of 16 monumental columns and circular building in the center. This structure was PLAN IV Section of Roman settlement north of fortified palace recorded by geophysical investigations 0 100 m 8 — Bülow, Schüler 2009, 231–249. 9 — Zosim., Hist., I,43; Aur. Vict., Caes., 34,5. 1 0 — Du{ani} 1995, 223–224, Note 39. 40 ROMAN GAMZIGRAD BEFORE THE IMPERIAL PALACE Base of column or pillar of peristyle of circular building in the settlement north of fortified palace, from the south FIGURE 25. a FIGURE 26. b a) Podium with crypt in the center of circular structure in the settlement north of fortified palace, from the south; b) podium in the center of circular structure in the settlement north of fortified palace, from the north opinion is based on the finds of bronze coins minted in the 2nd century for the ore-bearing regions of Upper Moesia.11 As an element of imperial propaganda on the occasion of establishing the new province and foundation of new and restoration of old mining-metallurgical and business centers could have also been erected the monumental structure of circular plan in the Roman settlement at Gamzigrad. A. Móscy assumed the existence of municipium Aurelianum, town in the Timok Valley, the administrative center of the orebearing region metalli Aureliani established in the time of the Antonines dynasty in the northeast part of Upper Moesia, that sometime later became part of the province Dacia Ripensis. This 11 — Móscy 1974, 131–134. Among the coins of the Roman mining regions in the territory of Serbia there is a specimen with representation of Diana on the obverse and inscription METAL AVRELIANIS on the reverse. According to analogies with coins from mining regions metalli Ulpiani – municipium Ulpianum and metalli Dardanici – municipium Dardanorum (MVNICIPIVM DD), for the mining region metalli Aureliani, the center was in hypothetical municipium Aurelianum. 41 SOFIJA PETKOVI] town (municipium) and was reduced to the fortified settlement Aureliana.14 Late Roman Romuliana was just that type of settlement. It is not impossible that toponyms Aureliana and Romuliana refer to the same settlement, named in the beginning of the 4th century Romulianum after the deified empress and Galerius’ mother.15 For the time being archaeological evidence indicates the existences of urban settlement dating from the period before Galerius building activities and in that territory the imperial palace was erected. The earlier settlement was larger than the fortified palace and extended further to the north and southeast (plan II, plan IV). The ore-bearing region metalli Aureliani, constituted in the time of the Antonines, most probably in the reign if Marcus Aurelius (161–180), gained in importance after abandoning the Dacia province and its rich gold and silver mines in AD 272. It could be assumed, on the basis of archaeometallurgical investigations that gold, silver, lead and iron were exploited in this mining region in the Roman period.12 Almost three centuries later Procopius mentions fortifications Aureliana and Romuliana in the area of Aquae, modern Prahovo, that were restored by emperor Justinian I.13 It could mean that the assumed municipium Aurelianum lost its status of 1 2 — Simi} 1969, 24–30; Du{ani} 1980, 32–35. 13 — Procop., De aed., IV,4. 14 — Jovanovi} 2004, 175–179. 15 — Aur. Vict., Caes., 40,16; Srejovi} 1995, 299–300. 42 DRAGOSLAV SREJOVI] IMPERIAL PALACE OWNER AND BUILDERS Gamzigrad is the distinctive architectural monument, because it entirely depends on specific political program. The monuments of this type are usually small in number as the political programs are almost always short-lived. Particularly short-lived was the political program which motivated the construction and appearance of Gamzigrad, as well as few more Late Roman edifices of the kind. At the beginning we should determine the category of monuments to which Gamzigrad should be ascribed. It has become clear long time ago that it is neither the military camp with praetorium nor the administrative or religious center, but there still remained surmises about the fortified country estate, luxurious health resort, imperial summer house or palace and even a settlement.1 Today, when it has been confirmed that both fortifications were built within a short period of time and that both temples erected to glorify the emperor are associated with them, it is quite certain that Gamzigrad should be included into official, even more so in the court architecture. The resemblance between Gamzigrad and imperial residences, particularly the Diocletian’s palace in Split, has been noticed rather early,2 and more recently Gamzigrad was compared also with other monuments of the Roman and Byzantine court architecture.3 It has not been noticed, however, that in general development of the Roman court architecture there is a distinct individualized category of monuments related exclusively to the political system of tetrarchy. As this system functioned only around thirty years (from AD 295 to approximately AD 324), all the monuments in this category had been built within this short period, so it is futile to compare them with court architecture from earlier and later periods. The only architectural monument, which is formally, essentially as well as chronologically close to Gamzigrad, is the palace of Diocletian in Split. The political program of Diocletian is well-known and the motifs that made this great ruler to undertake the construction of one such monumental architectural 1 — ^anak-Medi} 1978, 158–176. structure in the east coast of the Adriatic so 2 — Mano-Zisi 1956. much far from his capital, have been estab3 — ^anak-Medi} 1978, 155–158. lished. 43 DRAGOSLAV SREJOVI] proclaimed them Caesars. After that event the emperors left their capitals. Diocletian left Nicomedia and went to the east coast of the Adriatic to spent rest of his years in peace and glory in the palace he built in his homeland near Salona, at the site of present day Split. Maximian with the same idea withdrew from Milan to his estates somewhere on the border of Campania and Lucania. Diocletian’s palace in Split is the first monument in the history of antique court architecture constructed with an idea to be safe residence during life and after death of one great ruler who voluntarily renounced the throne and secular power, but not the authority and acquired rights. In order to satisfy entirely this basic requirement, Diocletian’s private and political biography, his intimate wishes and emotions, and at the same time about his ideological notions articulated during his twenty-year rule, must have been taken into account. For the first time the architects were entrusted with the task to create the safe place for permanent, well-deserved rest of the soldier emperor, but it should have not been either fortification or summer house. They were supposed to built everything appropriate for the Roman emperor, but different from the official palace in the capital and to add the temple where emperor would be venerated during his life and after death, and which at a certain moment would receive his earthly remains and would be transformed into the temple-mausoleum and enable the complete architectural undertaking to become the permanent monument in honor of the founder of new system of ruling the world – the tetrarchy. This memorial character of Diocletian’s palace in Split has not been so far clearly understood nor sufficiently emphasized. The palace at the site of modern Split was built exclusively for Diocletian and only to honor his person and his deeds. Its purpose was twofold: to provide dignified and peaceful old age for the great ruler and after his death – the eternal memorial, to be for a short time just luxurious palace and then to become the place of permanent pilgrimage, some kind of political and ideological testament to all nations of the newly established world empire. Only few rulers in Roman history followed Diocletian’s political program, hence in the history of Roman court architecture there are perhaps only two or three monuments identical in essence to the palace in Split. The palace in Gamzigrad is certainly one of such monuments, and his owner was without doubt one of the tetrarchs who consistently followed Diocletian’s reform of the central The program of Diocletian had three basic objectives: to raise the ruler to the level of deity, to reform the central government and to enlarge the army. The architecture of Roman Split as well as of Roman Gamzigrad depends greatly on the realization of these objectives, particularly on Diocletian’s reform of the state administration. That reform, realized in AD 293, introduced the tetrarchy as the distinctive system according to which the imperial authority was divided between four rulers – two emperors having the title of Augustus and their two adopted sons having the title of Caesar. Every Augustus and every Caesar was an independent ruler in his part of the state, as he had his capital, treasury, army and the executive power. In order to secure the longevity of this system, Diocletian equalized the divine and dynastic filiation: he proclaimed himself Jupiter and his co-ruler Hercules. The right to the Roman throne could have been acquired only by the act of adoption, i.e. the inclusion into the Jupiter’s family. Thus each of four rulers became a deity, carefully guarded and approached according to the distinctive ceremony. Special units for protection of the emperor known as palatini, scholae and candidati, dressed in the lavish uniforms, were organized at the court, and around thirty courtiers or confidential counselors (silentiarius) were always present in front of the imperial chambers. The ceremony of the access to the emperor was in charge of the highest courtier (magister officiorum) and audience officers (invitatores admissionales). Thus, even the court itself became sacred and moving around it and access to the emperor was transformed into an authentic ritual. The first objective of Diocletian’s political program – raising the ruler to the level of deity and the institution of corresponding ceremony – was not only achieved, but it was maintained permanently until the late Middle Ages. His reform of the central government, otherwise essential for understanding Roman Split as well as Roman Gamzigrad, was not, however, longlasting: it outlived its creator for only couple of years. Diocletian’s program of reforms of the imperial government had one important regulation: when the Augusti celebrate twenty years of their rule (vicennalia) they will voluntarily renounce the throne and give up their positions and titles to the Caesars. On the 1st of May 305 this regulation had been followed for the first and last time: Diocletian and Maximian transferred the title of Augustus to their adopted sons Galerius and Constantius Chlorus, while they adopted experienced military commanders Maximinus Daia and Flavius Severus and 44 IMPERIAL PALACE government. As their names, actions and destinies are wellknown in history, it is possible to establish whose wish was to build the palace at Gamzigrad, i.e. who of the Diocletian’s heirs could have wished to retire in Dacia Ripensis after twenty-year rule and to chose one nameless place there for his final residence, to transform into in the palace mausoleum, into the magnificent monument in honor of tetrarchy and into the symbol of indestructibility of the new system of ruling the world. It is absolutely certain that the structure resembling the palace in Split had been built by Diocletian’s co-ruler Maximian Herculius. It is known from the written sources that in AD 305 Maximian acted in all details according to the wishes of Diocletian, i.e. he voluntarily renounced the throne, left the capital and retired to peaceful life in Lucania or Campania, where he already had prepared residence.4 Only Constantius Chlorus and Galerius, who became the rulers of the Roman Empire in AD 305 and confirmed in all their actions the devotion to Diocletian, could have planed and built the structures where they would retire after celebrating the twenty-year rule. Already their legitimate heirs – Severus, Maximin, Constantine and Licinius – could have not do that as the events following after AD 306 resulted in final disintegration of the Diocletian’s system of governing the state. After taking the title of Augustus, Constantius Chlorus probably made plans for construction of the palace where he would retire after celebrating his vicennalia. However, as he died in Britain already in the following year (AD 306), it is not very probable that building of that palace had ever started. The completely different situation is with Galerius who could have built himself a palace mausoleum as he had reasons to believe that he would, like Diocletian, celebrate the twentieth anniversary of his rule and spend his last years in safety and peace in the pleasant countryside of his homeland, far from his capital. On the basis of available historical and archaeological data it could be concluded that Gamzigrad is that very palace mausoleum of Galerius. All historical evidence concerning events in the period from the death of Constantius Chlorus (306) to the rise of his son Constantine to the position of the sole ruler of the Roman Empire (324), points out unambiguously that in this long and painful agony of the system of tetrarchy only Galerius was ready to obey Diocletian’s regulations consistently, to protect them by all available means and substantiate them by his personal example.5 Even if entirely biased, malicious testimonies about Galerius’ personality and reign are carefully considered, for instance the text by Lactantius De Mortibus Persecutorum (On the Death of Persecutors), it is possible to get impression about Galerius’ comprehensive devotion in maintaining the system of tetrarchy and his strong intention to renounce the throne voluntarily as soon as he celebrates vicennalia and to spend his old age in peace and glory. Lactantius even speaks few times about Galerius’ intentions and actions in relation to the vicennalia celebration. When describing the situation in the Empire in AD 306 he explicitly says that Galerius already then made a decision that from AD 312–313, after he celebrates vicennalia, the Empire should be ruled by four persons he had chosen (Licinius, Severus, Maximin, Candidianus), while he himself “would spent safe and peaceful old age in the shelter of unconquerable walls”.6 That Galerius never gave up that idea is witnessed also by Lactantius, who directly relates difficult economic conditions in the Empire in AD 310 with the beginning of preparations for the vicennalia celebration. After describing all mistreatments following collection of taxes for that celebration, this bitter enemy of Galerius shouts: “Who is the one not deprived of all goods in order to squander all sources still existing in the Galerius’ empire because of the celebration, which should have never been celebrated”.7 This celebration actually never took place, because during that very year (AD 310) Galerius was taken ill and already in the beginning of May 311 died in terrible pains in Serdica (Sofia), not too far from his homeland. Lactantius informs us: “This event became known in Nicomedia in the mid May, at the time when celebration of vicennalia was expected on the 1st of March of the following year”.8 The intention of Galerius to celebrate the twentieth anniversary of his rule on the 1st of March 312 and to spend “safe and peaceful old age in the shelter of unconquerable walls” certainly means the construction of structure resembling Diocletian’s 4 — Eutr., IX, 27, 2; Zosim., II, 10, 2; Zonar., XII, 32; Lactant., De mort. pers., XXVI, 7. 5 — Ensslin 1930, 2516–2528. 6 — “… ita cum imperii summam tenerent Maximianus et Severus et secundum Caesarum nomen Maximianus et Candidianus, inexpugnabili muro circumsaeptus securam et tranquillam degeret senectutem.” (Lactant., De mort. pers., XX, 4). 7 — Lactant., De mort. pers., XXXIV, 4. 8 — Lactant., De mort. pers., XXXV, 4. 45 DRAGOSLAV SREJOVI] friends from his native country. The fact that Galerius emphasized that he has divine father did not result from his wish to hide his humble origin as is suggested by the malicious Christian authors,14 but from the ideology of tetrarchy. When he was adopted by Diocletian (in AD 293), Galerius became the member of Jupiter’s family, son of god. Therefore it is quite understandable that from that time he felt close to those two deities – Heracles and Dionysus, whose mother was a mortal woman and father the supreme god. On the other hand, Heracles and Dionysus as savior deities, who only after victorious campaigns in the world withdraw in glory to the Olympus, could have taken the prominent place in the ideology of tetrarchic system. Close connection between Galerius and Heracles and Dionysus is clearly emphasized in the decoration of the monuments built in his capital Thessalonica: in the scene of offering sacrifice on the Galerius’ triumphal arch15 and on the so-called small arch decorating the entrance to the large octagon of imperial palace.16 The fact that Heracles and Dionysus have the most prominent place in the decoration of the palace at Gamzigrad is still another proof that this structure had been built by Galerius’ orders. It has not yet been established with certainty whether Galerius had his official cult. It is only certain that Diocletian proclaimed him “second Romulus and Alexander”17 and it is also known for certain that annual festival in honor of the tetrarchs (Iovii et Herculii) has been celebrated probably on the day when the Augusti took the name of Jupiter and the Caesars of Hercules, on the day understood as their joint and actual birthday, geminus natalis.18 It means that Galerius had two pairs of parents and two birthdays. His first birthday and devotion to his mother obliged him to pay special respect to the chthonic “mountain deities”, and his second birthday and palace in Split, i.e. the palace mausoleum. It is absolutely certain that building of that structure must have started much before AD 312, but the question is at what location. The most important information about location of Galerius’ palace mausoleum is certainly the fact that Galerius had not been buried in his capital, Thessalonica, but in his homeland in Dacia Ripensis at the place where he had been born and which had been named Romulianum after his mother’s name Romula (“Ortus Dacia Ripensi, ibique sepultus est: quem locum Romulianum ex vocabulo Romulae matris appellarat”).9 This information shows that Galerius acted in all details as Diocletian and that he built the palace where he would safely spend his old age and after death find eternal peace in his homeland, far from his capital. In Dacia Ripensis, Galerius’ homeland, Gamzigrad is the only place where has been recorded magnificent edifice appropriate in all aspects to provide for a great ruler during his life “safe and peaceful old age” and after death – eternal memory of his person and his deeds. The position of Gamzigrad, as well as everything created there or brought from other places – both fortifications, temples, palaces, public buildings, architectural decoration, mosaics and sculptures – ideally corresponds, on one hand, with the ideology of tetrarchy and on the other, with Galerius’ private and political biography and even with his character. Galerius was born in a place “not too far from Serdica” in Dacia Ripensis;10 his father was simple peasant and mother was a barbarian woman who fled from the left Danube bank to Dacia Ripenses because of the attack of the Carpians.11 As a young man Galerius was tending cattle and because of that got a nickname Armentarius (Herdsman).12 He was handsome, strong man, lover of good food and exceptionally brave soldier, for all his life deeply devoted to his foster father Diocletian and strongly attached to his homeland, his fellow countrymen and relatives, particularly to his mother Romula, who was highly superstitious women, ardent admirer of the “mountain deities” and fervent enemy of the Christians.13 His actions at the time when he was Caesar and Augustus clearly show that he was not ashamed of his place of birth in the middle of nowhere, his poor descent, his relatives and friends from his youth. By naming his birthplace Romulianum he raised and immortalized his homeland and his mother, and by choosing the son of his sister, Maximin, and his fellow countryman Licinius for his co-rulers he demonstrated his devotion to the family and old 9 — Ps.-Aur. Vict., Epit., 40, 16. 10 — Eutr., IX, 22, 1. 11 — Lactant., De mort. pers., IX, 2; Ps.-Aur. Vict., Epit., XL, 10. 12 — Aur. Vict., Caes., XXXIX, 24. 13 — Lactant., De mort. pers., XI, 1–3. 14 — Lactant., De mort. pers., IX, 9. 15 — Laubscher 1975, T. 40. 16 — Hoddinott 1963, R1. 9 b–c. 17 — Ps.-Aur. Vict., Epit., XL, 17. 18 — Seston 1950. 46 IMPERIAL PALACE native man and one escapee from the left Danube bank got a son around AD 250 and that son became from the herdsman first the excellent soldier in the legions of Aurelian and Probus, then the Diocletian’s co-ruler, and finally the first man of the Roman Empire. The original fortification at Gamzigrad with contemporary buildings and the temple of Cybele corresponds entirely with Galerius’ decision to raise his birthplace in Dacia Ripensis in order to show devotion to his mother who ardently venerated “mountain deities”, probably the goddess from the Ida Mountain. The construction of later fortification, palace I and large temple in the center of Gamzigrad, had certainly started already at the moment when Galerius decided to transform his birthplace into the unconquerable fortress where he would retire after celebration of his vicennalia. As the central place within this new structure was given to the temple built as the place of veneration of emperor himself and possibly as his future mausoleum, the position of this temple caused the change of orientation of the decumanus axis. In the course of construction of this palace mausoleum the care was taken of all the structures built earlier, i.e. the section of Romulianum built until AD 306 was incorporated within the framework of the new imperial residence. devotion to the newly acquired father bounded him closely to Jupiter and Hercules. These facts, as well as already mentioned data that Galerius named his birthplace Romulianum after his mother Romula, that he was buried there and that he intended to spend his old age “in the shelter of unconquerable walls” make possible the reconstruction of his architectural undertakings in his homeland. Galerius, probably already in the time when he was Caesar, undertook considerable building activities at his birthplace in order to make it worthy of his mother’s name. Later, when he became the first person of the Empire in AD 306 and decided to retire to his homeland after twenty years of rule, like Diocletian did, he took steps to magnify even more that very place where he was to spend his old age and find eternal peace and to make it the symbol of eternal duration of the tetrarchic system. He could have started to build Romulianum only after he finished the wars with Sarmatians, Quadi and Bastarnae on the Danube, probably in AD 297. On the other hand, Galerius certainly started construction of the “unconquerable walls”, which would secure his safe and peaceful old age only after the death of Constantius Chlorus in AD 306, i.e. when he had chosen his heirs and became the most powerful person in the Empire. It seems that on the basis of these data from Galerius’ private and political biography could be understood the entire process of construction of Gamzigrad. The country estate whose remains have been discovered in the south section of Gamzigrad is probably the very site in Dacia Ripensis where one (Translation of the part of the following text: Carski dvorac, Gamzigrad – kasnoanti~ki carski dvorac, Galerija SANU 45, Beograd, 1983, 24–66) 47 MILKA ^ANAK-MEDI] BRANA STOJKOVI]-PAVELKA ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE The settlement, which reached its highest architectural and artistic rise when it became the imperial palace Romuliana, had been built in the center of trapezoid plateau slightly sloping down towards the southeast. The settlement had been founded on the plateau according to distinct plan, and it was transformed in few segments in the course of time. It is surrounded by fortifications built on two occasions, that also follow the natural slope of the terrain. This is most apparent in the difference in altitude, which is 11.84 meters between the thresholds of fortification gate on the east and west side. The buildings constructed in the southeast quarter of Romuliana are at even lower level. The mounds indicating remaining still unearthed FIGURE 27. Aerial view of Romuliana and its environs 49 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA the rituals of inauguration, limitation and orientation (inauguratio, limitatio, orientatio), performed by augurium using the groma. This ritual had been used for outlining main communication route, decumanus and most of the architectural entities constructed inside Romuliana. By designating the east–west main road also the position of main fortification gates had been established. The ramparts were outlined from the gates, following the natural morphology of the terrain. Because of that, only the south rampart was parallel with decumanus, while three other sides create the trapezoidal ground plan of the settlement. This plan was determined by the waterway in the east and by the elevated plateau in the north, where numerous earlier villas and architectural entities had already existed at that time. At the foothill was a deep trench, perhaps regulated in antique period and used as vallum. The vallum was, however, certainly created along the south rampart, as it is confirmed by geomagnetic prospection,3 and it existed on the west side, as we can conclude on the basis of the records of one 19th century traveler.4 First to be outlined and created on the selected plateau was the defensive zone, established according to the recommendations in ancient technical manuals that had been continuously used. These recommendations were stated by Vitruvius in the sixth chapter of his first book, emphasizing that in the process of founding certain settlement, the internal arrangement of structures is determined only after the surrounding fortification had been built: When thus the city walls are erected all around, says Vitruvius, after that follows in the interior the division of building site, planning of roads and streets according to the points of the compass.5 Following the established sequence, first the design of the ground plan of the first Romuliana fortification had been determined, and after that it was built. Its west rampart is creating sections of the antique structures are visible here and there on this slightly sloping terrain (Fig. 27). Nevertheless, completely or partially investigated architectural entities in the Romuliana interior offer clear picture about its spatial composition and irregularities in its ground plan design. We shall try to explain their causes and course of creation of the agglomeration by investigating architectural forms, function and date of origin of individual buildings and architectural entities. It is well-known that Roman planned settlements are based on the orthogonal design employed to determine main communications, as well as further division in insulae, conforming to the system resulting from the modular scheme originating from the Hellenistic tradition. Such division was applied already in founding Hellenistic towns according to the principle of Hippodamus, that he used first in establishing the urban structure of his hometown Miletos, rebuilt after Persian destruction in 494 BC, and also later, when planning the spatial structures of Piraeus.1 There was applied the Pythagorean principle of harmony, represented in numbers and respecting their meaning either for the physical structure of the settlement or for the character of the community within it. From that time onwards identical orthogonal schemes and distinct numbers, which had particular significance and meaning, were applied without interruption in Greek and Roman architecture. The same principle in settlement planning was not ignored even in the Late Roman period. This system was particularly observed in founding imperial residences, as is confirmed by the Diocletian’s palace in Split, but also when planning agglomerations of the closed structure, surrounded by walls with towers like the military camps, which influenced to a considerable degree the spatial structure of Late Roman imperial palaces.2 SPATIAL STRUCTURE Thanks to the discovered archaeological remains we are able to reconstruct the course and manner of establishing the area which Romuliana was to occupy. The central stone (omphalos) had been laid in the center of selected plain. This is a rather large stone slab placed on masonry foundation and with two engraved lines on the surface. One is denoting the east direction and the other the west direction. This stone was used for ceremonial definition of the sacred areas (area sacra) and the lines of main communications, i.e. for designating magic – cosmological cross within the ground plan design. It included 1 — Lavedan 1926, 189; Böethius 1948, 3–33; Castagnoli 1956; Martin 1956, 15 sqq; Mamford 1988, 173, 174; Mili} 1990, 107, 108. 2— There are many texts about military camps. They were described already by antique writers: Polybius and Philon of Byzantium and later also Vitruvius and Hyginus (in the time of Traian) and then Julius Africanus, Vegetius and anonymous Byzantine writer from the time of Justinian. Their works were discussed in many texts and commentaries quoted in ^anak-Medi} 1978, 150 and note 392. 3— Bülow von 2007, Abb. 3. 4— Ma~aj 1882, 95. 5— Vitruvius, I, 6, 1. 50 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE FIGURE 28. Porta decumana of earlier fortification known as porta decumana (Fig. 28), is facing the sunset and human demise and it is the residence of spirits of the underworld, and because of that, this gate usually led to the necropolis of the settlement. Because of the morphological characteristics of the terrain, there was no north entrance, which had been most important in Roman settlements and imperial palaces, as it was leading to the central open area – in the cities to the main square – where the most important shrine of the agglomeration had been located on the south side, which was considered the best.7 There was only the east–west communication, passing through the settlement center, while transversal cardo was missing, but it still existed in the mind of architects and resulted in central position of the temple in north section of the interior and directly opposite large south temple. Both these streets in Diocletian’s palace in Split were lined on both sides with porticos supported by masonry pillars.8 It seems that in Romuliana an acute angle of approximately 78°, with the south side of the fortification, as its deviation was caused by configuration of the terrain in the north. The north rampart extends along the mentioned trench from the corner tower, where it meets the west rampart at an angle of 110°. This rampart had two deviations at the first and second third of its length. The north part of east rampart is outlined at the right angle to the east part of north rampart and it extends in the same direction until reaching the east fortification gate, whence it is outlined at right angle to the south rampart. The towers had been erected as integral elements of the rampart and porticos were constructed on the inside, along all ramparts and in front of the towers. Main entrance to the palace was through the east fortification gate, as it was established with certainty, because the approaching road, marked by tetrapylon on the saddle of east mountain range, has been identified (plan XLIV, 1). There were found enough remains of that entrance, so its one time appearance could be inferred (Fig. 108).6 The east gate in military camps is called porta praetoria (the gate of the commander), as it provides favorable outcome of the battle for the soldiers leaving the camp via that gate, and according to an even earlier tradition, originating from the east Mediterranean, it is the residence of gods. The fortification gate on the opposite, west, side, 6— Srejovi}, Vasi} 1994, 108–119. 7— ^anak-Medi} 1978, 78, with earlier relevant literature, cf. note 156. 8— Buli}, Karaman 1927, restitution of original palace plan; Marasovi} 1968, prilog 34. 51 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 29. Ramparts and towers of later fortification from the south Roman law by most severe punishments, it is quite certain that such radical change of the Romuliana defensive system could have been carried out only by edict issued according to the divine right of Emperor Galerius himself.14 This undertaking substantially changed the appearance of the palace and its visual aspect. So, it became regarded as identical to sacrum palatium of the Late Roman times, according to its new, lavishly decorated fortification. The real reasons for modification of original Romuliana fortification will probably remain incomprehensible. It is well known that ramparts and towers were considered sacred (res sancta) and that they denoted sacred boundary, protected by wooden posts supported the portico along south palace façade in the northwest section of the interior.9 Similar porticos existed in many European regions, for example at Caerwent in Britain. This street in Romuliana was together with porticos 29 meters, i.e. 100 feet wide. Both main streets in Diocletian’s palace in Split were around 90 feet wide each, while in the military camps, which were models for the imperial palaces, these streets were, according to recommendation of Hyginus, 60 feet wide.10 Besides main communications there were also peripheral ones, which included also porticos constructed in an interrupted series along the ramparts. Neither these streets nor the main east–west communication had never been entirely completed.11 The ramparts of the first fortification and their porticos had been in use for a rather short time. They were constructed together with towers in the final decade of the 3rd century12 and there is also assumption that their building started not before AD 303.13 Soon after their construction, the concept of imperial palace was changed, so the ramparts and porticos were pulled down, but the existing towers were preserved in full height and incorporated into the new strong and monumental fortification system surrounding earlier fortification (Fig. 29). When it is known that violation of the city walls and even jumping over them had been sanctioned according to the 9— In one transversal trench explored near southwest corner of the palace in northwest quarter of palace D1, was discovered a compact layer of roof tiles and a layer of soot underneath. Considering this, it was asumend that these are the remains of wooden portico along the south wall of that palace, ^anak-Medi} 1978, 160 and note 431. 10— Lavedan 1926, 185. 11— Vasi} ^. 1997, 57. 12— ^anak-Medi} 1978, 48–50. 13— Vasi} ^. 1997, 54. 14— Vasi} ^. 1997, 46, 47 with arguments and sources. 52 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE 0 PLAN V 10 20 30 40 50 100 m Site plan of Romuliana with partially or completely uncovered structures in the interior boundaries. The year of Romula’s death had not been recorded, so it is assumed that she encouraged the persecution of Christians in 303/304, consequently, that she was alive at that time.17 The analysis of coins found at the base of Romula’s tumulus suggested the end of 3rd century as the date of her funeral.18 But, this rules out the preceding assumption. So it remains most probable that mentioned tectonic disturbance the divine authority. At Romuliana later ramparts were 10.95 m far from and parallel to earlier walls, so it was not the case of shifting pomerium, that had been allowed for the towns if coinciding with successful conquests or expansion of the Empire.15 On the contrary, the construction of stronger and more luxurious fortification at Romuliana is the consequence of decision of the emperor to transform it in his residence after renouncing the throne, planned for the year AD 312. Besides this crucial reason, it is not impossible that first fortification was destroyed in tectonic disturbance, recognized in conspicuous cracks and shifting of walls of some of its towers.16 The possibility was also considered that perhaps the death of Romula, after whom Romuliana was named, was, among other things, reason for abandoning the first established sacred 15— Homo 1951, 91–97. 16— ^anak-Medi} 1978, 169. 17— Srejovi} 1995 B, 301; Vasi} ^. 1997, 58. 18— It was assumed that apotheosis of Romula took place either during AD 294 or in the beginning of AD 295 at the latest, Vasi} M. 2007, 50. 53 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA Palace in northwestern section of the interior (D1), aerial view FIGURE 31. Centrally placed large temple – peripteros, aerial view FIGURE 30. The palace D1,2 was located in the most prominent place, at the highest point of the inside fortification. Its main entrance was facing east fortification gate and it was approached from that direction (Fig. 30). The walls of its rooms were outlined in such a way that initial measuring segment was laid parallel to the decumanus. The palace extends within the northwest quarter toward north, as far as the portico of earlier fortification, whose pillars were already demolished at that time and incorporated into the palace north walls. This palace has rhythmically arranged pilasters only along the south and west façade, and these pilasters determine the rhythm of the pillars of the portico. It was the most luxuriously conceived and decorated palace, so it is reasonably considered as the residence of the emperor. Although it was of residential character, the areas for the official activities of the emperor, taking place when he was staying in the palace, were also planned. instigated abandoning of the first fortification, even more so as it could have caused the fires recorded archaeologically in the explored section of west portico and portico in southeast corner and in the layer, which precedes the one from the time of later fortification.19 Therefore, there are grounds for the assumption that when decision was made to built new fortification the earlier one had not been completely finished.20 It has been assumed that the later fortification had been built from the 1st of May 305 to the end of building season in AD 306, and that this undertaking started when Diocletian withdrew from office and left the ruling of the east part of the Empire to Galerius.21 It is possible that the construction of buildings in the interior started before the new fortification was completed, but certainly not before the work on the north rampart and the towers was finished, because their construction would have been almost impossible after the construction of the palace in the north half of Romuliana interior.22 As Vitruvius recommends, the locations for temples and squares had been designated first, and after that for other public buildings.23 They were outlined in the northwest and south part in the orthogonal disposition, according to direction of decumanus, and in the northeast quarter of the settlement they follow the direction of surrounding ramparts (plan V). They could be distinguished as private residences with temple on the north, left side, and the public buildings situated on the south, right half of interior space. 19— Petkovi} 2008 A, 61–63; Petkovi} 2008 V, 64–67; Petkovi}, @ivi}, Kapuran 2009, in print. About layer of soot under the roof tiles in west portico, ^anak-Medi} 1978, 47. 20— It is well-known that Diocletian’s palace in Split was not finished not only until Diocletian’s abdication in 305, but not even later, i.e. until his death (313). 21— Vasi} ^. 1997, 56. 22— Vasi} ^. 1997, 43. 23— Vitruvius, I,7,1–2; Vasi} ^. 1997, 44 and note. 98 where the original text by Vitruvius, concerning the temple locations, is quoted. 54 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE The investigations carried out so far, provided the ground for reliable establishing of chronological relations between the architectural entities. Building of both palaces started simultaneously on the north side, but first was finished the palace whose complex and elaborate plan covered northeast quarter of the interior (D3). East segment of that building negates internal portico of the earlier fortification, confirming without doubt that it was built after the portico had already been demolished.24 Building of the south section of palace D1 started simultaneously with palace D3. This section has the pilasters along the south façade identical to the pilasters along the interior of the earlier fortification gate and they were built of identical material. This indubitably confirms that they were executed at both places by same masons and within rather short time span. The remaining segment of this palace was built only after the ramparts and porticos of the first fortification had been demolished and new fortification erected. Only after completion of this large building enterprise the palace D1 was completed and its annex D2 was also finished. Building and decoration of the palace took rather long time to be completed.25 The northeast part of the palace, considering the selection of architectural features and direction of the east wall, was observing the position of the temple situated in the same area as palace D3. When the temple was finished and its temenos was under construction, west segment of the neighboring palace D3 was reshaped and adjusted to it.26 This confirms without doubt that east palace was completed before the temple was erected, but there is also possibility that its earlier phase with section of the podium had been finished earlier. The sequence of building structures in the south section of the interior has also been established. First central large temple (I) was built, whose building perhaps commenced before the later fortification was entirely finished, and after that the neighboring buildings were constructed. Next structure constructed in that section of interior are the thermae. It has been concluded that it was built after the parts of earlier fortification had been demolished. After thermae, the building with porch (J) was built, and slightly later also the building with General view of thermae (X) in southeast section of the interior FIGURE 32. Second palace of the private character is located in the northeast quarter of the settlement (D3). It was outlined parallel to the east section of the north rampart and was extending in the east–west direction. Between the two palaces was the mentioned small temple (C), functionally connected with the palace in the northeast half of Romuliana. The architectural entity in this half of the interior was completed with long building with corridor D4, outlined to the south of the palace D3 and with atrium D5, which linked this building with the palace in the northwest section (D1,2). Division on the south side, which was intended for the buildings of public character, started with centrally placed temple (Fig. 31), surrounded in the west, south and east by buildings of diverse purpose. They are all in the orthogonal arrangement, corresponding with decumanus, but they do not constitute coherent entity, but are freely arranged within the area. South and west building are at approximately the same distance from the temple. South building has a portico (G), and the west (F) is the largest single roofed structure. It was located to the opposite of palace D1 on the south side of decumanus, but was deviating from its axis more than the palace. Along its west side and rather far from decumanus is one enclosed structural entity (E), complying also to the basic orthogonal grid, whose starting point is at the decumanus axis. Building H, which has been identified (Fig. 32) on the basis of its ground plan and internal installations as the public thermae, is situated to the east of large temple (I). 24— Vasi} ^. 1997, 40, 41. 25— In the attempt to establish the absolute chronology in construction of this palace, it was assumed that the building lasted from AD 306 to AD 311, Vasi} ^. 1997, 57. 26— ^anak-Medi} 1995, 54. 55 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 PLAN VI 50 m Plan of earlier defensive system, uncovered segments are marked black portico (G). Although spacious building F has not been systematically investigated, it could be assumed that it was built after the building with portico (G), but before the westernmost architectural entity E.27 On the basis of presented data concerning the contents and chronology of the fortification and buildings in the interior of Romuliana, we got a clear picture of its composition. It has been revealed that division of internal space was not based on any strict and coherent scheme and that it seems that disposition of buildings was not completely predicted in advance, according to the premeditated plan. In favor of this conclusion speaks the evidence that some parts of the palace in the northwest quarter were altered in the building process, while some other data suggest that palace was subsequently enhanced with luxurious stone porticos in its peristyles, but we would discuss that later. Also, there is evidence that the palace D3 was subsequently connected with the temenos of small temple (D). Nevertheless, it is obvious that from the beginning the binary principle of interior division had been accepted. So, the whole structure was divided in two conceptually and functionally different halves, two successive fortifications, two fortification gates, two temples and two residential palaces.28 This is certainly the consequence of the dual purpose of Romuliana. It was at first intended to be the residence of Galerius’ mother Romula, to whom he was very attached and thus built her a palace at his birthplace, but then emperor decided to spend his life after renouncing the throne in this palace, close to her mausoleum. Such division is also related to the Roman comprehension of meaning of distinct segments of the settlement located on its north and south side. The south side was considered the right side (part dextrata), while the north side was considered the left side (pars sinistra), and each of them had its distinct meaning.29 Romuliana, which was partially built at the end of 3rd century and intensely built during the two decades of the next century and its function being confirmed with certainty, contributes to the comprehension of court architecture and its progress in the Late Roman period, as we will try to show in more details by studying individual structures and architectural entities in its interior. 27— Vasi} ^. 1997, 45. 28— Srejovi} 1983 S, 45. 29— Sui} 1976, 90. 56 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE 0 PLAN VII 5 10 m Plan of uncovered segment of defensive walls and towers of earlier fortification in the west (III), as well as both corner towers on the east side (V, VI), were also investigated. Two towers of the north rampart (VIII, IX) are partially explored, while recent archaeological excavations at Romuliana were aiming at discovery of the south portico in front of the corner tower V. Even though all towers and porticos of the earlier fortification have not been archaeologically investigated, it was possible to establish in general the entire fortification plan (plan VI). The number and disposition of towers could have been assumed with considerable certainty even before the recent investigations on the basis of the Kanitz plan from 1864, when the towers of the earlier fortification had been visible at many locations,30 and later, according to the significant mounds at their locations.31 West and east fortification gate is of identical plan, only the later fortification gate in the west is shifted southward in relation to the earlier gate, so south octagonal tower of the earlier fortification became propugnaculum of the new gate. The shifting is insignificant on the east side, so both octagonal towers maintained their former function. While the west fortification gate – porta decumana, have been completely unearthed, on the opposite, east side, only fortification gate and areas in front ARCHITECTURAL COMPOSITIONS The positions and purpose of certain buildings and architectural entities in the Romuliana interior presented in the preceding chapter represent their contents and spatial structure in general outlines. The data about them will be completed with analysis of their architecture, including also the elements for their dating, based on the study of stylistic traits, and they will be presented according to the types of architectural entities. EARLIER FORTIFICATION The ground plan and architectural design of the earlier fortification are known to us thanks to systematic and test trench excavations carried out so far. Thus it is known that east and west fortification gate was flanked with towers built on octagonal plan, while other towers were of square shape. There is one tower each between the gates and corner towers on the west and east side, while on the north and south side, where there were no big fortification gates, but possibly only poternae, there are two towers at equal distance on both sides. The remains of the first fortification at Romuliana that are completely explored include part of the west rampart with fortification gate (porta decumana – towers XV, XVI) and the neighboring tower of square plan (XIV), with the remains of the pillars of inner portico. The pillars of portico and octagonal towers of east fortification gate (I, II) are only partially explored. The neighboring square tower on the south side of east rampart 30— Kanitz 1868, T. IV, 4; V, 1,2,4. 31— ^anak-Medi} 1978, sl. 127. 57 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 33. counterforts with strong pillars (1.9 x 1.9 m) opposite them had been built on the inside of the gates and octagonal towers. These pillars are preserved to the greater height than the others along the rampart, indicating that they and structure they supported existed also after the construction of new fortification. The ramparts with the octagonal towers on the outside are structurally connected with the mentioned counterforts and they extend northward and southward from that junction (plan VII). The counterforts built next to the entrance create with opposite pillars the space which has the form of propugnaculum near other contemporary fortification gate. Thanks to the west fortification gate (Fig. 34), where both lateral towers have been completely unearthed (plan VIII, IX), we know the character of the ground and partially of the first floor of north tower (XVI), and also the upper structure of portico in front of theoctagonal towers could be surmised. It is possible, however, to suppose with considerable certainty that square bays created by counterforts and pillars had cross vaults, while counterforts and pillars were interconnected by longitudinal and transversal arches (plan X).32 There is sufficient data for establishing the elevations of octagonal towers. The entrance to both west towers was in the east and to the east towers on the west side. It was vaulted with the barrel vault of which segments have been preserved. The towers had three windows at the ground floor level. Judging by the windows in north octagonal tower of west fortification gate, they were 15 cm wide on the outside and 97 cm on the inside. The Octagonal towers of west fortification gate of earlier fortification of the entrances to octagonal towers (I, II) have been discovered. Also, only on the west side the fortification gate was discovered as far as the threshold, and on the inside of the gate to the brick paved floor. There is difference in altitude, considering the thresholds of the fortification gates on both sides, and that difference is over 1 meter. It has been concluded that it is the consequence of filling up the area next to the earlier fortification. There are jambs on both sides of the gates of earlier fortification, so we know that the gates were 4.25 m wide. Massive 0 1 2m 0 Cross-section of north octagonal tower of west fortification gate of earlier fortification 1 2m PLAN IX West façade of north octagonal tower of west fortification gate of earlier fortification PLAN VIII 58 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE 0 5 10 m PLAN XI Façade of west fortification gate of earlier fortification, ideal reconstruction 0 5 brick string course was at 2.65 m from the threshold, we know what the gate actually looked like. Because that string course denotes the point from which the arched structure above the entrance had started. In the course of archaeological excavations no stones from the vault have been encountered, so it was reasonably assumed that vault was made of bricks, as all other apertures on the towers. Semicircular niches in the façades of south and north tower facing the entrance (Fig. 34) also were the elements of the external composition of the gate. Such niches next to the large portal appear at few gates of the Late Roman fortifications. They are, however, usually on the gate façade to the left and right of the main entrance.35 There is no direct evidence for the total height of octagonal towers, but it could be established according to the remains of portico on the inside of the wall. The lean-to portico roof was resting on tower and rampart, but the towers must have been 10 m North octagonal tower of west fortification gate of earlier fortification, ideal reconstruction PLAN X identical slanting windows exist at the ground floor level of Diocletian’s palace in Split.33 In the same tower at Romuliana are visible the recesses for beams supporting the upper storey, as well as the beams supporting the staircase running along its walls, starting from the east wall. The staircase led to the first floor of which there are parts preserved in the same tower. There were three windows at that level, wider than those at the ground floor and starting from the floor.34 On the basis of the data about octagonal towers of the west fortification gate it is possible to surmise the complete outside appearance not only of towers, but also of fortification gate, although its jambs are not preserved to the full height. Thanks to the fact that internal sides of the towers façades are preserved to a slightly greater height and that the segment of the 32— ^anak-Medi} 1978, 33. 33— Marasovi}, Mcnally 1972, 21–23, pl. 7–10. 34— ^anak-Medi} 1978, 36, 37. 35— These niches, for which Frigerio thinks that they originate from Asia Minor, were particularly popular in the 3rd century and were employed on the fortifications in Bosra and Nicaea in that time. 59 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 5 10 m Modular grid and method of designing west fortification gate of earlier fortification (M = 8 feet, foot = 29.6 cm) PLAN XII South façade of north octagonal tower of west fortification gate of earlier fortification FIGURE 34. The way of outlining the fortification gate and the design procedure could have been also established with considerable certainty. On the basis of individual measures it was concluded that the entire fortification was designed using the foot 29.2 to 29.7 cm in size. Thus the thickness of ramparts is 5.5 and 6.5 feet, which is slightly less than suggested by the ancient technical manuals,39 while the length of space in front of the tower was 24 feet, the internal span in the east–west direction was 19 feet and in the north–south direction 20 feet, while the external span of octagonal towers was 32 feet. The total width of city gate is 16 feet, and outer sides of the octagons vary from 12 to 13 feet. The result is that half of the gate opening (8 feet) was used as the project module in creation of the complete fortification gate.40 On that module is based the square grid, into which the achieved spatial composition fits rather well (plan XII). The fortification gates flanked by octagonal towers had been built at the entrances to towns and military camps in a one storey higher, in order to give access to the walkway on the defensive wall. Such walkway with parapets exists on many Late Roman fortifications.36 The trimming of the tower façades could also be surmised on the basis of available data. They were rather simple-looking on the outside, and careful execution of the joints suggests that they were not plastered and that picturesque impression was attained by method the building material had been employed. Thus the lower section was built only of bricks (opus testaceum) and the top section was constructed of alternating courses of brick and stone (opus mixtum). The façades were also decorated with cornices of ceramoplastic elements, discovered in great quantity during excavations. These elements include the bricks with denticulated profiles or with cyma on one side, and many consoles. So, we know that crowning courses on towers consisted of few courses of molded bricks resting on the consoles of baked clay.37 On the top of them was the wooden roof structure covered with ceramic roof tiles. The towers were surmounting the defensive walls, which were, according to the recommendations of the antique writers, supposed to be 20 ells high, if built in the lowland area.38 According to the hypothetical reconstruction of Romuliana walls, they were 35 feet, i.e. 23 ells, high. The façade of the fortification gate was in line with the ramparts, and there were probably wide windows above the entrance (plan XI), similar to those above the silver gate of Diocletian’s palace in Split. 36— Richmond 1955, fig. 3, 5. 37— ^anak-Medi} 1978, 40, 41. 38— Information is provided by anonymous Byzantine from the time of Justinian, who took over many recommendations by Philon of Byzantium and just supplemented them, Berchem 1954, 266. 39— There is suggested that defensive walls should be 5 ells or 7.5 feet thick. 40— ^anak-Medi} 1978, 41, 42. 60 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE FIGURE 35. Square tower (XIV) of earlier fortification FIGURE 36. Windows in the west façade of tower XIV of earlier fortification to the height of 5.6 m (Fig. 35). It has spacious arched entrance in the east wall and two windows in the opposite wall that are of identical size as windows on the ground floor of the octagonal tower. However, their window sill is sloping to the inside to a greater degree than it is the case with the windows in the neighboring tower. There were not found the recesses for beams of the storey structure in the square tower, but in spite of that it could be assumed that upper storey structure was made of wood. There is no data about total height of the discussed square tower, but as all towers of well-preserved antique fortifications are usually of the same height, it could be assumed that this tower had ground floor and two stories, and that it was of the same height as the octagonal towers, i.e. 14.9 m. The most characteristic on the façades of this tower are the windows. They are arched on the outside with two concentric semicircular arches, which have strings of two rows of brick on their extradoses (Fig. 36). There are various molded ornaments above window extradoses at many antique buildings, but two rather long period of time, and the towers of identical shape are flanking the gates at Diocletian’s palace in Split. The scholars were of the opinion that this type of Split fortification gate was made after Porta Caesarea in Salona from the republican times. There are, however, chronologically closer examples like Porta Praetoria in Como and fortification gate in Vindonissa, from the 3rd century.41 Of the square towers of the earlier fortification, the most comprehensively explored one is that next to the north rampart (XIV), located between the fortification gate and the north corner tower (XII). The towers differ according to the position of their entrance and to the way of their connection with the adjoining ramparts. The ramparts join the corner towers (V, VI, XI, XII) in the middle of two adjoining walls, in one of which was the entrance, while the ramparts join other square towers at the third of the length of lateral façades, so these towers are 3.41 m inside the fortification. The entrances to these towers are facing the interior of the settlement. The distance between the square and octagonal towers is from 29.93 to 31.38 m.42 The antique writers suggest that distance between the towers should not exceed the arrow range, and so it was concluded that distance between the towers should not exceed 30 m.43 The architecture of the square towers could be comprehended using as an example the tower next to the west rampart (XIV), that has been completely unearthed and preserved up 41— Frigerio 1934–35, 113 fig. 86; 260 fig. 203 A, B, C; 261, fig. 204. 42— It is 30.29 m between east fortification gate and tower III, between that tower and corner tower B is somewhat greater: 31.38 m, and on the west side, between north octagonal tower and square tower XIV, rampart is 29.93 m long, Vasi}, ^. 1997, 23. 43— Grenier 1934, sv. castrum. 61 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 PLAN XIII 1 their elevation on the basis of the remaining parts and of the evidence from octagonal towers. The fortification rampart is 1.75 m thick and row of pillars somewhat smaller than the pillars in front of the octagonal towers (156 x 156 cm) were built on the inside, at the distance of 5.30 m. They are also more modest, as they do not have extended base finished with slanting bricks laid in few rows, as it was the case with the pillars in front of the octagonal towers. The pilasters were built (opposite the pillars) along the outer rampart in the section of portico between the north octagonal tower of the west fortification gate and the rectangular tower (XIV). These pilasters change their position in the section where the wall supporting staircase was built parallel to the rampart. The staircase gave access to the walkway on top of the rampart (plan VII). According to discovered foundations and above ground parts of pillars along the south rampart, next to the corner tower V, it is concluded that the corner pillar at the meeting of portico lines from different directions was of more elaborate plan (shaped as latter L) and that it has slanting base, like pillars in front of the octagonal towers.45 Some segments of portico were vaulted, but there is evidence that in some segments there was also wooden upper structure.46 Their façades consisted of arcades opening to the interior and as their original height could have been established, thus the entire height of portico has been established as well. It is reasonable to assume that the arcades started from the stringcourses at 2.65 m, and were equal to stringcourse on the octagonal tower. The height of their apexes could have been determined on the basis of the span between the pillars, and it could be assumed that the total height of outer walls to the stringcourse under the roof was 5.92 m, i.e. 20 feet. The smallest height of the ridge of the portico roof could have been 10.37 m or 36 feet (plan X), as thus was achieved the slope compatible to the type of cover used in antique architecture. It was not possible to establish the complete longitudinal appearance of the porticos. As the terrain was sloping towards the southeast, the porticos could not have been continuous, 2m West façade of square tower XIV of earlier fortification rows of bricks at the same spot and in the wall plane are rather exceptional. They were encountered on the rectangular towers of the fortification at Thessalonica.44 The same ornamental pattern like on the octagonal towers was employed on the other façades. The picturesque impression was achieved by alternating courses of stone and four or five rows of brick. Besides, the bricks were also used around the windows and at the corners in such a way that few rows of bricks are interrupted vertically at the same spot, thus creating the denticulated two-color motif, which imitates the building with stone blocks (plan XIII). This tower was also covered with the wooden structure having hipped roof covered with roof tiles like the octagonal towers. The measuring of this and other identical square towers was not a complex problem. As they partially project inside the settlement, the measuring started from the outer wall face of the rampart and the external and internal walls were marked at equal distance of 17.5 m. This resulted in total width and length of the towers being 35 feet, and as the walls are 6 feet thick, internal span is 23 feet. Defensive rampart and portico are preserved to a considerably smaller extent than the towers, as they had been demolished, as we mentioned earlier, in the process of building the later fortification. At that time the area they covered was raised for 1.10 m above the floor of the entrance gate of earlier fortification and their remaining parts were covered with earth. It is, however, possible to establish the outline of porticos and 44— Towers with identical ornaments within fortification at Thessalonica were built either before Thessalonica became the capital of emperor Galerius, or in the rein of Galerius at the latest, i.e. in the first years of the 4th century. 45— Petkovi}, @ivi}, Kapuran 2009, in print. 46— ^anak-Medi} 1978, 47. 62 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE but they had to follow the descending of the terrain in a sort of steps. They certainly made the striking frame for the structures constructed within. The described earlier fortification is related, according to its features, to the Late Roman fortification architecture. The basic composition of the west and east side of the fortification, with octagonal towers next to the gates, square corner towers and a tower, each of the same shape at half distance between the corner and entrance towers, repeats the shape and composition of corresponding sides of the fortification of Diocletian’s palace in Split. Even the length of these sides is almost identical at both locations. The difference is in the position of central square towers in relation to the defensive walls. While at Diocletian’s palace all towers were built on the outside of the ramparts, central towers at Romuliana are partially projecting toward the fortification interior. There are also differences in the interior, as in Split there are cryptopoticos besides the porticos, while such structures are missing in the earlier fortification at Romuliana. The system employed in Split, with towers leaning to the ramparts on the outside, is considered to be later and used from the 4th century onwards, but, according to the examples from the northwest parts of the Empire, it is confirmed that other type of towers had been built also in the time of Diocletian.47 The employed structural composition in Romuliana also complies with that epoch, as the vaults and arches had been built in the end of the 3rd and the beginning of the 4th century in the same way as they were built at the earlier fortification at Romuliana. The dating of the earlier fortification is influenced by the results of recent archaeological excavations in the southeast corner of the Romuliana interior, where a layer with archaeological finds from the 3rd century was discovered on top of the floor of earlier fortification portico.48 The more precise time span for the building of earlier fortification is established by the bricks with stamps of Legio V Macedonica, encountered in the fortification walls, so we know that this legion participated in the building activities. Therefore, the earlier fortification at Romuliana could have been built only after AD 270, when the mentioned legion returned to the nearby Oescus.49 The possibility for even more precise dating is offered by the building technique employed in the construction of towers and defensive walls, as there are many data about the building technique in the given periods, and about the building material used in certain parts of the Empire.50 According to these studies, it could have been established that traits identical to those of the earlier fortification at Romuliana have the walls constructed not before the end of the 3rd century. On the earlier fortifications, dating before the reign of Aurelian, the bricks are thinner, and particularly important, the layer of mortar between the courses is thinner than the bricks. Only in the second half of the 3rd century the thickness of mortar reaches the thickness of bricks, as is the case at Romuliana, and sometimes even exceeds it, and that would be the characteristic of architecture from the time of Constantine. The rows of bricks within the courses are also important for precise dating of certain structures, as it was concluded that the number of brick rows increased with time. According to the analysis of many Late Roman fortifications, including the analysis of employed fortification system, as well as the building technique, it could be concluded that the time of construction of the first fortification at Romuliana could be established, according to the employed fortification system, around the seventh decade of the 3rd century, but the employed building technique – with lacing-courses consisting of three, four and five rows of bricks and with broad joints – does not allow for much earlier date in the 3rd century, but indicates that fortification had been built in the closing years of that century. This conclusion is supported by the fact that three monuments: palace in Split, the earliest section of Thessalonica fortification with Galerius’ palace and earlier fortification at Romuliana are similar, according to the manner of vaulting, general characteristics of the masonry bond and aspiration to picturesqueness by selection of the building material. Because of that the first fortification at Romuliana was in the former studies dated in the time of Diocletian, i.e. in the final years of the 3rd century.51 Thus, its construction could have been related 47— Petrikovits 1971, 178–218. 48— Two layers of leveling between earlier and later fortification have been recorded. In both layers were found objects from the second half of the 3rd and the first half of the 4th century, and objects found immediately above the floor of the earlier fortification portico dated from the second half of the 3rd century, Petkovi} 2008 A, 61–63; Petkovi} 2008 V, 64–67; Petkovi}, @ivi}, Kapuran 2009, in print. 49— ^anak-Medi} 1978, 90 and note. 202, with information about whereabouts of that legion, Vasi}, M. 1997, 152, 154. 50— Lugli 1957, sv. opus testaceum, 542–629; Bob~ev 1961, 153–201; WardPerkins 1958, 52–104; Gall 1958, 181–202. 51— ^anak-Medi} 1978, 48–50. 63 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 37. Ramparts and towers of later fortification on the east side was still alive. These building activities were perhaps slowed down in AD 297, when Legio V Macedonica was transferred to the east, because of the wars against Persians. But the legion returned to Oescus in AD 299 and works at Romuliana were probably continued in full swing, even more so as after that victory Galerius stayed in the Danube Valley from AD 299 to AD 302 and part of 303, and then probably supervised the building of his palace.54 to Diocletian’s Caesar Galerius and his decision to build a palace in his homeland, dedicated to his mother Romula. According to the latest assumption, Galerius started to build earlier fortification only after the end of wars against Persians and after the celebration of his decennalia in AD 303.52 The mentioned 3rd century finds encountered in the layer above the floor of earlier fortification portico suggest, however, that earlier assumption was more correct, as they also suggest the end of the 3rd century as the period of construction of this fortification. It seems, however, more probable that Galerius finished his undertaking at Romuliana before transferring his capital to Thessalonica in AD 300 or AD 305 at the latest, as since that time he was very intensely building there. In Thessalonica Galerius built a magnificent palace, monumental rotunda, large octagon and many other substantial buildings, and there were also much works on strengthening the defensive system, if its earliest segment had not even been built at that time. Many masons and stone carvers, as well as huge financial resources, were required for the Thessalonica undertakings.53 Therefore, it seems more probable that Galerius erected earlier fortification at Romuliana before the beginning of works in his capital and that he started works at Romuliana while his mother Romula LATER FORTIFICATION Later fortification was built parallel to the earlier one at a distance of about 10 meters. Therefore, it replicates the irregular plan of previous fortification, and only on the north side direction of rampart was partially straightened so only one deviation remained at the second third of its length (plan V). The east side also remained of irregular direction. The east and west fortification gate of the new fortification were generally con- 52— Vasi}, ^.1997, 54. 53— ^anak-Medi} 1978, 173. 54— Vasi}, M. 2007, 51, 52. 64 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE ceived in the same way as those of earlier fortification, and only the shapes of towers are different. Between the towers flanking the gates and the corner towers there is one tower each on every side, similar to the earlier fortification, but number of towers along the south and north rampart is doubled, so on these sides there are four towers between the corner towers. The towers are of different size, but of identical composition. They are polygonal on the outside and circular on the inside, where there are three massive pillars in the center. The towers flanking the gates are twelve-sided, corner towers are sixteensided on the outside, and others are ten-sided, except two middle ones on the north and south side, that are twelve-sided, like the towers next to the fortification gates. Thus, all sides of later fortification were designed in the same manner, considering the number and the form of the towers, while difference between east and west and other two sides is in the fact that towers along the south and north rampart were at equal distance, while on two other sides towers next to the fortification gates are closer to each other. All towers were built on the outside of the rampart. In addition to two main fortification gates there are also two poternae. One is between southeast corner tower and neighboring tower (3), and the other is on the north side, between towers 8 and 6 (plan V). Later fortification is more comprehensively investigated than the earlier one. Systematic investigations included new west fortification gate together with flanking towers, and gate and area between the towers surrounding the gate, as well as the entrance into them and their outer side, were discovered on the east side. The remaining towers of the south section of east rampart and all towers and ramparts on the south side (Fig. 37) were cleaned of rubble and deposits, and besides the southwest corner tower (15) also the neighboring tower on the west side has been exposed. The test trenches were excavated inside some of the towers, and thus the supporting elements were discovered in their center. The pillars of the porticos were interpolated between the towers of the earlier fortification along the inside of all ramparts. Fortification gates of the later fortification have been discovered on both sides and were completely explored on the west side. It is known, on the basis of discovered segments, that the gates on both sides were identical, so we may discuss the east gate on the basis of the completely explored west one. The span of the entrance is 4.42 m, and its outer side is of segmental 0 1 2m Plan and cross-section of north polygonal tower of west fortification gate of later fortification, ideal reconstruction PLAN XIV plan, and the octagonal tower of the earlier fortification is on the west side of the inside. So, this tower was used as propugnaculum, because the new entrance was shifted 8 meters to the south. New gate on the east side is shifted only 2.7 m, so the towers of the earlier fortification remained in use. The complete gateway includes also the polygonal towers, 22.48 meters on the outside and 15.20 meters on the inside, flanking the gate (plan VII). Three strong pillars in the center of the towers have convex outer side, concave inner side, and radial lateral sides. They supported the upper structure. The towers have entrances facing inside the fortification and the rectangular vaulted area is in front of the entrance. The stairs leading to 65 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA the platform on the rampart start on one side of this vaulted area, and on the other side is the staircase leading to the upper stories of the tower and to the gallery above the entrance. The complete elevation of the towers could have been inferred thanks to many remnants of the upper structure of the north tower of west fortification gate. One of the collapsed pillars with finishing stringcourse has been found inside. There were enough fragments to determine its original height, which was 9.45 m, i.e. 30 feet. There were conical vaults between the pillars and in the peripheral ring-like section were barrel vaults, of which the remains have also been found. The central section of the tower was, by all appearances, higher and supported upper wooden structure (plan XIV). There was a platform above the barrel vault in the peripheral section. Its height could have been established on the basis of the data about the interior structure and about the staircases leading to the stories, and these data were also coordinated with the elements for reconstruction of the gate façade. Thus, it was established that platform was at the height of 45 feet, i.e. 14.10 m. It had outside facing arcades carved of white limestone and of which many fragments have been encountered.55 Both platforms on the north and south tower of fortification gate were interconnected by the bridge, of which there are impressions which are better preserved on the east fortification gate.56 The way of designing and later marking of the west fortification gate, as well as of the polygonal towers, has been established. By checking the measures of some of their parts, it was concluded that they consisted of whole numbers of antique measuring units, and that they were from 31.5 to 31.8 cm.57 For designing had been used the modular grid, whose basic element corresponds, as in the earlier fortification, to the half width of the fortification gate, and it is 5 ells. The interior wall face was used as the starting segment from which the measures were taken. There are five modules from the wall to the center of the circular tower, and from the middle of the gate to the same tower center, along the abscissa, there are seven modules. The dimensions of areas in front of the entrances to the towers fit into the same modular system. It was also possible to establish the scheme for measuring and designating the twelve-sided towers, using the circle inscribed in the square, the sides of which are ten measuring units long (plan XV).58 Between the towers whose elevation and interior structure could have been comprehended there was the fortification gate. Its façade is preserved in the lower section up to 6.78 m. It was 0 5 10 m Modular grid and method of designing west fortification gate of later fortification (M = 5 ells, ell = 47.2–47.7 cm) PLAN XV built up to 2.23 m in the opus quadratum technique, of accurately dresses blocks of tuffaceous sandstone with some white limestone blocks next to the jambs, and from that height, above the stringcourse, in the opus listatum technique, with alternating courses of stone blocks and two and three rows of brick. On the outside there are simple door posts in the lower section, ending in stringcourses supporting the semicircular arch. The arch was built of voussoirs of tuffaceous sandstone. There is one semicircular niche with semi-dome vault to the left and right of the entrance respectively. The rectangular groove was carved on the lateral sides of the doorposts from the threshold up the preserved wall, and it was used for lowering and raising the cataracta. Such device existed in all known Roman fortification gates, and at Romuliana it helped to determine the height of the first gallery above the entrance. 55— Detailed description of all archaeological remains from that tower and about bases for their theoretical reconstruction: ^anak-Medi} 1978, 54–58. 56— Vasi}, ^. 1997, 32. 57— The feet 31.5–31.8 cm in size were used in eastern parts of the Empire, where they are known from the 6th century onward, Underwood 1948, 64–74; Schlibach 1970; Lugli 1957, 189. 58— More detailed description of the designing method in ^anak-Medi} 1978, 58–60. 66 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE 0 10 20 30 40 50 cm PLAN XVI Fragments of archivolt and inscription Felix Romuliana from the first gallery of west fortification gate of later fortification ornament of the west fortification gate of Romuliana revealed that there were two different galleries. One had stone decoration made of gray tuffaceous sandstone and the other one had the decoration made of white limestone. This is the same material used for ornaments on the platform under the roof of polygonal towers, and as they are molded in the same way, it was assumed that these two galleries were at the same altitude. Among the pieces carved of tuffaceous sandstone there are six consoles of identical dimensions and identical basic division of the frontal section, but each pair is decorated with different motif. According to the carved bearings it was concluded that they were projecting up to 70 cm. They helped in distinguishing the elements they supported, their lower segment and stringcourse adjoining them. Important evidence was provided by the bases resting on consoles and supporting free-standing columns. They did not only determine the dimensions of the given columns, but they also revealed an essential fact that they had grooves for the parapets on two opposite sides. Thanks to one completely preserved column shaft the height of the lower part of the decorative ensemble of the first gallery has been established. There were also found the corresponding capital fragments and one almost complete capital. They had On the basis of the preserved part it could be concluded that the courses of stone and brick were not at the same level across the entire façade and that they were leveled only at 6.78 m from the ground, and that led to the conclusion that from the apex of gate arch to the height where courses are leveled there was some architectural part. It has been assumed that the tablet with the important inscription about the founder of Romuliana had been placed there. We could draw conclusions about the upper destroyed part of the façade on the basis of the remains of architectural decoration found in front of the entrance and next to it on the inside. There have been discovered over one hundred fragments of architectural ornamental elements, including bases, column shafts, consoles, capitals, abaci, pilasters, archivolts, imposts, voussoirs and many types of cornices and stringcourses. After studying and systematization of the fragments, two distinct groups, which were decorating galleries of the gate façades, have been distinguished. It has been concluded that they were the elements of the galleries, according to the renowned general concepts and division of the façades of the fortification gates in the cities and imperial palaces. There were one or two galleries above the settlement entrance, with series of apertures facing outside. The identified 67 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA Carved decoration of the first gallery of west fortification gate of later fortification Carved decoration from first gallery of west fortification gate of later fortification FIGURE 38. PLAN XVII panum was the two-layered finishing course (Fig. 38). In addition to the mentioned ornaments this entity also contained the pilasters, which created the stepped aperture of the gallery. We concluded, considering the ornamental motifs, that like in the case of consoles there were pairs with distinct decoration. Thus, four pairs have been identified. The carved decoration of the first gallery also included the parapet panels, and the fragments of three panels are preserved (plan XVII).59 Almost identical repertoire of the architectural ornaments has been found in front of the east fortification gate. There distinct repertoire of motifs and the execution of acanthus leaves was also specific, and they helped considerably in the dating of carved decoration on the first gallery of the west fortification gate. One abacus, which thanks to its elaborate execution helped in reconstruction of the composition of decorative ensemble on the first gallery, also belongs to this group. It has been established with certainty that above the abacus was the archivolt, of which enough fragments have been found to determine its original size. The archivolt has semicircular intrados and the extrados ending at an angle of approximately 52°. At the top of the tympanum created by sloping sides was the piece with the inscription FELIX ROMULIANA (plan XVI), as it has been concluded on the basis of its thickness, which is identical to other fragments of this archivolt. Above the tym- 59— ^anak-Medi} 1978, 62–65, where discovered fragments are described in detail and argumentation for their identification is presented. 68 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE FIGURE 39. Carved decoration of the second gallery of west fortification gate of later fortification Carved decoration from second gallery of west fortification gate of later fortification PLAN XVIII were carved figural representations whose iconography determines more exact date of construction of later fortification.61 The arrangement of apertures on the first gallery has been established so far only for west fortification gate, according to the number of key elements of its carved decoration and division of the lower section of the façade. Thus, it was concluded that, according to pilasters, there were at least four openings, and on the basis of other elements their height and span was established. Very important was the evidence on the column bases, indicating that parapet panels were inserted continually between them. The axial distance between the free-standing columns is 236.5 cm or 7.5 feet, and almost the same was the total height of carved decoration of the gallery, while its inner have also been discovered many fragments of decoration of the gallery above the entrance. There were voussoirs, many types of cornices, column shafts, consoles, parapet panels and pilasters carved of tuffaceous sandstone and white limestone. There was also found one almost complete archivolt of the same shape as one found in front of the west fortification gate.60 There were encountered the motifs known from the decoration of gallery at the west gate, but there are some new ones, including the eagle with wreath, head of Gorgon, laurel wreath flanked with two peacocks, while some cornices have on the frontal side the motifs not encountered on the west side. Also some motifs had been carved in a different way, like ivy leaves for instance. Particularly important among the new motifs is the representation of military standard (signum) on two pilasters. It has been concluded, according to these finds, that east fortification gate was more lavishly decorated than the west one. Particularly important is that on some decorative elements 60— Srejovi} 1986 a, 92; idem 1993, 206, 207. 61— Srejovi} 1986 a, 93; Vasi}, M. 2007, 37, 38. 69 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 PLAN XIX 5 10 m Longitudinal cross-section of west fortification gate of later fortification, ideal reconstruction with laurel wreath and two peacocks, as the arch span is the same as that of the archivolts on the first gallery of the west gate. The carved borders of the second gallery of the west fortification gate differ from the first gallery openings not only in the fact that they were carved of different stone, but also in general composition of decoration. It has been established with certainty that these were arched openings flanked with freestanding columns with capitals and arcades made of individual voussoirs. Among the discovered elements of the second gallery there were identified two columns carved together with the bases. The columns had Ionic capitals with imposts, of which we discovered just one, and there were many voussoirs (Fig. 39). Not all voussoirs had the same height of the frontal side and span was identical with the span of semicircular niches on the ground floor level (plan XX). This fact suggested that one aperture on the gallery was directly above the ground floor niche and that they both had identical carved decoration. This assumption is corroborated by the fact that in that case the distance between two furthest apertures is divisible into three intercolumnations of the same size, which is in accordance with renowned designs of upper zones of the façades of Roman city gates. The information about three apertures with identical carved decoration suggested the conclusion that third one was in the middle, and that the mentioned inscription was in the tympanum of its archivolt.62 There was certain information that there were two more apertures, but its decorative ensemble was unknown. It has been assumed that other two apertures had arches, because the favorite ornament at Late Roman façades was the series of arches with alternating triangular and semicircular or segmental tympanums.63 The hypothetical reconstruction of the east fortification gate is not completed, but it could be concluded on the basis of discovered fragments of architectural decoration that there were also two galleries, and first of them had similar decoration as the same gallery on west fortification gate. We came to this conclusion on the basis of tectonic of pilasters and archivolts 62— Location of that inscription was first assumed to be above the entrance to the cruciform structure E (Srejovi} 1985, 57, 58), and later it was assumed that together with archivolt it belongs to the earlier fortification gate (Vasi}, M. 2007, 51), but it is not in accordance with archaeological data about the finding place of this archivolt and the archivolt of same form, discovered in front of theeast gate of later fortification. 63— It was encountered on earlier monuments, like on the library in Ephesus, then on the Bucoleon palace in Istanbul, and from later times on the stucco decoration in the baptistery of the Greek orthodox in Ravenna: Mango 1965, 315, 317, sq, fig. 7, 8; Krautheimer 1965, pl. 57. 70 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE the galleries, and the vertical groove in the door posts. Thus it could be assumed that the floor of the first gallery was higher than the top of that door, as much as is the height of the door to the top, in order to contain the raised cataracta. As the fortification gate was 15 feet high to the top, it was assumed that the first gallery was at double that height (plan XIX). These conclusions have been verified by exploring the manner of design of the fortification gate. This analysis suggested the conclusion that square grid,64 divided into modules of 7.5 feet or 5 ells, had been used for their design, and that position of all the segments of second range ornaments was established on the basis of thus achieved modular grid by farther decomposition of the squares (plan XX).65 The redesigned octagonal tower of the earlier fortification got new appearance and function when it had been transformed into propugnaculum of the new fortification gate. All tower walls are preserved up to considerable height, except west and east one, where the passages were made. In order to make it suitable for the new purpose and in agreement with new fortification gate, the passage on the west side is laid symmetrically with the new gate, and on the east it had to be shifted 1 meter to the north, because of existing counterforts. Stone architraves decorated with diverse floral and figural motifs and found inside the octagonal tower originate from the new door of the octagonal tower. It is obvious, according to their dimensions, that they correspond to the additionally opened door, and it is also corroborated by the motifs carved on them, some of which are encountered on the fortification gate, where they are used as protection from the earthquakes and all other disasters66 The façades of the type identical to the west fortification gate at Romuliana have been in use for rather long time. They appear already in the Hellenistic time,67 and they had been continuously built in Roman architecture, following the identical concept since the 2nd century BC, and they are preserved in many Roman towns. They all had one or two galleries above the entrance. Their decoration was mostly reduced to the simple that helped in distinguishing the voussoirs from the gallery above the entrance from somewhat higher voussoirs from platforms on the polygonal towers. It should be mentioned that there are grooves for the parapet panels on two opposite sides of the column bases, so it confirms that there was an uninterrupted row of apertures. One of these parapet panels has been discovered in fragments. The total height of the carved decoration of the second gallery is established on the basis of complete columns and other members of the ensemble, and the span is established on the basis of difference in width between the voussoirs of extrados and intrados. It has been concluded that the axial distance between columns was 2.30 m, and the height of opening was 2.7 m. Thanks to these finds the decorative entity of the second gallery above the fortification gate could have been conceived (plan XVIII). The height of reconstructed openings of the first and second gallery of the west fortification gate of later fortification was established considering the data about the staircase leading to Façade of west fortification gate of later fortification with modular grid used in designing PLAN XX 64— Identical geometric figure for modeling Gallo-Roman monuments was identified by Formigé 1943; idem 1949. 65— Detail description of use of square grid in ^anak-Medi}, 1978, 74. 66— Picard 1962, passim. 67— Martin 1968, 171–184; Swoboda 1961, 79 sqq. 71 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA PLAN XXI West fortification gate of later fortification, ideal reconstruction are only the towers at the gates in Verona and Spello, as well as the towers at Eburacum in England.71 The towers flanking entrances at all mentioned locations are more distant from each other than at Romuliana, and Romuliana gate is distinguished also by its concave plan. There are gates with façades designed as exedras, like in the south parts of Gaul from the 1st century AD and in Tipaza and Vindonissa from the 2nd and 3rd centuries, and also some fortifications on the German limes have external exedras.72 It could not be concluded on the basis of the known examples that the entrance of semicircular or concave plan was characteristic of the distinct type of fortified settlement, but it is conspicuous that Nero’s Golden House, Hadrian’s villa in Tivoli, one castle in Bavaria (Harlach) and luxurious villa at Piazza Armerina also had external exedras. architectural sculpture. Among the most beautiful examples from the Late Roman period are the façades of the fortification gates in Diocletian’s palace in Split and fortification in Nicaea, from the end of 3rd century. The evolution of the fortification gate façades could be estimated besides the preserved monuments also on the basis of the Thracian coins from the time of Hadrian and from even much earlier times from other east regions.68 The possible carved decoration on the gates could be inferred from the sarcophagi of Sidamara type, for which it is assumed that they repeat in their decoration one type of town gates.69 The gate, which could help to the greatest extent in explanation of emergence and concept of the Romuliana fortification gate, despite being of later date, is the entrance gate to the imperial palace in Constantinople, dating from the third decade of the 4th century, as there is a series of richly decorated arcades with niches above the entrance.70 The fortification gates at Romuliana have certain distinct characteristics, including the external shape and the size of the flanking towers. There are large towers at many Roman fortifications, including Porta Nigra in Trier or gates at Autun and Nimes. There are towers of considerable size at few entrances of Aurelian’s fortification in Rome, but almost all of them are of somewhat smaller size than the towers at Romuliana. They differ also in shape, as they are twelve-sided, and of such shape 68— Smith 1956, 38–50. 69— More about the evolution of façades of the type of Romuliana fortification gates in ^anak-Medi} 1978, 83–86. 70— Mango 1959, 99–107. 71— Bechert 1971, Abb. 37. There are certain similarities between the towers of the late Romuliana fortification and towers of the fortifications at ^ezava and Hajdu~ka vodenica, Kondi} 1984, 135, 13. 72— Schulze 1909, 284–304; Frigerio 1934–35, 112–116. 72 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE medallions and two wreaths on the other. There are figural compositions in three medallions on the first pilaster, and on the second the figural compositions are depicted in two medallions. Two male busts are carved in each medallion and it is assumed that because they are represented in pairs, these are the portraits of the tetrarchs. It was attempted in further studies to identify the portrayed persons.79 It has been assumed that the figures in the top medallion on the first pilaster represent Augusti who abdicated, and in the other two one Augustus and one Caesar still reigning. In favor of such conclusion has been mentioned the difference in their clothes. The persons in the top medallion are dressed in togas and cloaks, and others have paludamenta fixed with fibulae. This assumption determined the beginning of the construction of the later Romuliana fortification into the period after the 1st of May 305, when Diocletian and Maximian abdicated.80 More recent interpretations of these images are somewhat different. According to them, on the pilasters were depicted the tetrarchs from the third tetrarchy, established at the meeting at Carnuntum in AD 308, when Galerius and Licinius were elected as Augusti, and Maximinus Daia and Constantine I were proclaimed Caesars.81 Even if it is the case, the building of the later fortification must have certainly started earlier, probably immediately after AD 305, because the construction of such large and strong fortification certainly took many years, and we know that the final works on the west fortification gate had been carried out between AD 308 and AD 311.82 Therefore, considerable section of the new Romuliana fortification had already existed In addition to the explained characteristics, the fortification gate at Romuliana are distinguished also for its multi colorness, achieved by using diverse building material and enhanced by selection of two-colored stone for architectural decoration that was particularly popular in the Late Roman architecture.73 Also, the opus listatum building technique was once again popular in the Late Roman times and employed on the buildings of Maxentius and Constantine.74 The fortification gates at Romuliana are important not only because of their architecture, but also because of stylistic traits of sculpture on the first and second gallery. Besides the Corinthian and Ionic capitals there were also numerous other stone ornaments with various motifs: vine foliage with leaves and grapes, laurel three-leaf bands and laurels in bundles, palmettes, broad individual acanthus leaves and four-leaf and fiveleaf rosettes. Their repertoire and manner of modeling correspond to the Late Roman period.75 Certain inconsistency in manufacture that could be noticed, indicates that they were made by the masters with different experience and that local stone masons were also employed. The capitals have been used for general dating of the west fortification gate at Romuliana, because main phases in their evolution have been established.76 First it was proven that the four-leaf capitals of free-standing columns from the first gallery are of individual shape, which appears in the same compact form in the end of the 3rd and in the 4th century.77 Their shape and modeling of acanthus leaves on the pilasters helped even more precise dating. The characteristic of these acanthus leaves is curving and breaking of few tips and joining with the tips of the next leaf. According to this and the grooves between the segments and the channels starting from the tips, the capitals from the first gallery of the west fortification gate could be compared only to the capitals dated to the 4th century. The more distinctive indication is the fact that the Corinthian capitals from Diocletian’s palace in Split are very close to our capitals in the way of modeling.78 The time span established according to the analysis of the stylistic traits of architectural decoration of the west gate gallery is narrowed down even more, thanks to the representations on two pilasters from the gallery of east fortification gate. They have the vine with leaves and grapes on their lateral sides, and on the front side there is a representation of the military standard with four circular medallions and a laurel wreath in the field framed with ivy leaves on one pilaster, and there are three 73— Azevedo 1970, 228. 74— Lugli 1957, II, tbl. CXCIV, 2, 3, 4; CCV, 2; Krautheimer 1965, 26, 34. 75— First to determine that the architectural decoration from the first gallery of the west fortification gate at Gamzigrad is of Late Roman character, dating it in the 4th century, Schlunk 1970, 161 sq. 164, Taf. 46. 76— Kautzche 1936, passim; Nikolajevi}-Stojkovi} 1957, 9 sqq; Bob~ev 1970, 122. 77— ^anak-Medi} 1975, 247–253. On four-leaf capitals from Sirmium, Singidunum and Viminacium, I. Nikolajevi} 1965, 653–660. 78— ^anak-Medi} 1978, 92. 79— Srejovi} 1986 A, 93; Srejovi} 1994, 145, 146, Fig. 1, 2; Vasi}, M. 2007, 37, 38. 80— Srejovi} 1994, 143–152; Vasi}, ^. 1997, 55, 56. 81— Vasi}, M. 2007, 52. 82— ^anak-Medi} 1978, 96. 73 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA entrance is of distinct shape and it got the upper structure, because the wall is very thick at that point. The junction of two ramparts was leveled, and one exedra each was made on the outer and inner side of the tower. They were vaulted with the semi domes on each side (Fig. 40). The entrance to the northeast corner tower (6) was designed in a different way because the corner tower (VI) of earlier fortification was very close to it. So, the earlier tower had to be altered and adapted to a passage leading to the new corner tower. Not a single corner tower has been completely unearthed, but only their upper sections, where window openings widening toward the interior have been encountered. Their dimensions are close to the dimensions of the windows on the twelve-sided towers. It has not been excavated deep enough to reach the pillars in the tower centers, so their number has not been established with certainty, but it is assumed that there were three of them, like in the twelve-sided towers. However, it is not impossible that there were four pillars. The twelve-sided towers located in the middle of the north and south wall have the identical shape and structure as the towers within the fortification gate. From their vestibule lead the staircases in two directions to the walkway along the rampart. Only in the tower 12 is just one staircase and it is facing east. The ten-sided towers were located between the corner and middle ones. Some of them were excavated on the outside, where circular stone foundation having an outer diameter of 22.55 m was discovered. All ten-sided towers are circular on the inside, as it was concluded after the test trench excavations. Their outer diameter of above the ground section is from 22.28 to 22.78 m and the inner diameter is from 15.10 to 15.20 m. They all have entrances from the fortification interior reached via the vaulted entranceways. From these towers usually lead the staircases on both sides to the walkway on the rampart, but there are examples with the staircase at just one side. There are windows vaulted with the conical vaults in the walls of these towers. The tower 10, next to the north wall, has the staircase leading to the walkway just on the east side, and for the staircase had been used the tower VIII of the earlier fortification. Defensive ramparts and porticos have not been explored to the same extent. The defensive walls were unearthed from the outside at half length of the east and west side, and the outer side of the south rampart was explored completely. The walls are preserved up to the height of around 10 meters. The larger until then, and certainly the fortification gate with the described pilaster. Other polygonal towers of the later fortification have the same basic structure as the towers next to the fortification gate, but the only identical ones are those in the middle of the north and south rampart. The others differ in size and number of sides on the outside. The corner towers are the largest, and they are sixteen-sided on the outside and circular on the inside. The diameter of their circular foundation is 26.90 m, while the external diameter of, the upper section is 26.23 m, and the internal is 19.05–19.10 m. The center of the circular tower plan is on symmetry axis of the angle made by two ramparts of different direction, and their meeting point is the entrance. The FIGURE 40. Entrance to the corner tower 15 of later fortification 74 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE segment of portico was excavated along the west rampart, and also the entire south half on the east side. There were uncovered the pillars of portico next to corner tower VI of the earlier fortification and one pillar next to the tower VIII. On the basis of the unearthed remains of porticos it was confirmed that they exist on all sides and that they were built between the towers of the earlier fortification. Sometimes they directly lean to the earlier towers, like on the west side, where one pillar was built against the octagonal tower, and the same situation is on the north side next to tower VIII. The pillars were built against the octagonal tower of the fortification gate on the east side and against the neighboring rectangular tower III. This had happened because of the unequal distance between the towers of the early fortification, while the distance between the pillars of new portico had to be uniform. The outer defensive wall, 3.60 or 3.62 m thick, was erected when the earlier one had already been demolished. This is confirmed by the building material used for new structure. The wall was built in the opus mixtum technique with lacing courses consisting of three rows of brick at a distance of 1.8–2.5 m. In the first lacing course were mostly used old, damaged and broken bricks, which are also thinner than bricks used for the fortification gates of the later fortification and the upper section of the external side of entire south rampart and segment of east rampart. The original height of the ramparts could have been established on the basis of rather reliably determined height of the platform in twelve-sided towers of the west fortification gate leading to the walkway on top of the rampart. The platform was 45 feet, i.e. 30 ells high. Hence, we know that the fortification wall at Romuliana was considerably higher than the defensive walls of the fortifications built in the lowland, that were 20 ells high, according to the recommendations of ancient writers.83 The pillars of the new portico were built at 4.65 m distance on the average from the defensive walls. The number of pillars between the towers is not always identical, as the distances between towers were not identical. Thus, there were five pillars between the tower I and III and four between the tower III and tower V, and also four between the north octagonal tower of the west fortification gate (XVI) and the neighboring square tower (XIV). The pillars of the portico are of cruciform plan everywhere except where they lean to the towers of the earlier fortification and where one arm of the cross is missing. The average distance between the pillars is 4.5 m. The pillars were 0 1 2 3 4 5m Segment of the portico façade of later fortification, ideal reconstruction based on established project module (M = 5 ells, ell = 47.7 cm) PLAN XXII preserved up to the height of 2 to 2.5 meters and only exceptionally up to 4.5 m. The pillars have bases finished with the slanting blocks around its entire perimeter except on the middle section of the inner side. It is assumed that they also terminated in slanting cornices under the arches, like the pillars in the interior of polygonal towers. The shape of the pillars reveals that they were supporting the upper structure consisting of longitudinal and transversal semicircular arches. These pillars created almost symmetrical square bays, which were by all appearances vaulted with the cross vaults. There was an attempt to establish the height of the pillars and of the upper structure of the portico on the basis of previously established height of the ramparts onto which the portico was leaning and on the basis of the staircases leading to the platforms, but most decisive for comprehension of the entirety of portico was the analysis of proportional relations. The recommendations by Vitruvius about good proportions between the height and width of the portico have not been neglected either.84 83— See before, note 38. 84— Vitruvius, V,9,3. 75 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 PLAN XXIII 1 2 3 4 5m Plan of tetrastyle prostylos (temple C), ideal reconstruction façade above the entrance. This fact, as well as three Corinthian capitals carved of the same limestone and found at 10 to 15 meters from the later ramparts in two different sections of the settlement, suggest the possibility that gallery existed also above the portico and that those shafts and capitals were part of it. However, it could not be excluded that identical architectural elements and similar carved decoration have been used on some buildings in the Romuliana interior. The time of construction of the later fortification at Romuliana is determined on the basis of the stylistic traits of carved decoration and figural compositions on the described pilaster of the east fortification gate. Thus, it has been assumed that building started after the 1st of May 305.85 It was also assumed that works continued for two building seasons and that the fortification was completed in the autumn of the following year.86 Although it was assumed that the huge working force took part in this undertaking, including also Legio V Macedonica producing necessary bricks, this time period seems too short for building such an exceptionally strong and high defensive It could be concluded upon the analysis of the ground plan that portico was designed in the same way as the façade of the city gate, i.e. that the module used for measuring ground plan and elevation was 7.5 feet, i.e. 5 ells, as was also the thickness of the pillars. The distance between them is somewhat less than two modules and is in some bays 14 and in others 14.5 feet, resulting from the length of the portico. From the established modular grid results that the height of pillars was 2 modules, as it was the distance between them and the distance to the surrounding ramparts (plan XXII). The coordination of the height of portico with obtained position of platform on the twelve-sided towers reveals that there were also 2 modules from the foot of the arches to the top of its façade, so the entire façade was 30 feet or 9.42 meters high. Thus, the ratio between the height of pillars of the portico of the later fortification at Romuliana and its depth was like it is suggested by Vitruvius. With certain reservation we suggest the assumption that the upper section of the portico had a series of openings facing outside. This is indicated by certain elements of the carved decoration found in the vicinity of the west rampart. There was discovered large number of limestone column shafts of the type used on the second gallery above the fortification gate. There are more shafts than could have been used on the 85— Vasi}, ^. 1997, 56. 86— Vasi}, ^. 1997, 56. 76 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE size. So the temple did not have antae, but only corner pilasters (plan XXIII). There was spacious staircase on the frontal side, that could have been reconstructed with certainty as the first row of steps was discovered in situ and there was trace of next two steps on the wall. Thus it was established that there were 13 steps, 21.13 x 33 cm in size, all of them identical, except the first one, which was wider. The number of steps is odd, as Vitruvius suggested for the approaching steps of the temples, but their dimensions do not correspond to the staircases recommended by Vitruvius.89 The sanctuary (cella) of the temple is almost of square plan, 7.07 x 7.24 m or 22 x 22.5 feet. It is very similar in length to Diocletian’s temple in Split, but it is much narrower (7.29 x 5.86 m). There are strong corner reinforcements in the cella of tetrastyle prostylos at Romuliana. They were added after the walls reached the height of 1.8 m, as it was concluded on the basis of masonry junction with longitudinal and transversal walls discernible from that height. The reinforcements had not been built in the same manner as the perimetral temple walls, but, judging by the vertical line scratched in fresh mortar on the wall of the cella to mark their position, it could be concluded that they had been planned from the beginning. The supporting walls parallel to the west wall and at a distance of 90 and 150 cm had been built to create a crypt of cruciform plan at the same time when the corner reinforcements had been added. Next to the crypt and at the lowest level of the west wall there is a small rectangular vaulted niche with the bottom under the crypt floor. In the niche was a receptacle made of marble slabs. Traces of the stairs for descending into the crypt have not been discovered. They could not have been adjacent to the walls because of corner reinforcements,90 so it seems that the crypt was partially covered with slabs and that it was entered on special occasions by wooden stairs. The crypt and the supporting walls creating it were connected with the upper structure, where there were two niches in the west wall. The supporting walls could be explained, on the basis of preserved system. It could perhaps be extended to the autumn of AD 307, when Galerius with the legions started the campaign against Maxentius. But even until then the complete fortification could not have been completed and works on some of its segments were probably running parallel with the construction of the buildings in Romuliana interior. In any case, some interior works, like flooring of towers and porticos, had not been finished. This is confirmed by the coin found in the floor substructure in the polygonal tower of the west fortification gate with the obverse reading GAL MAXIMIANVS PF AVG, and the reverse GENIO A – VGUSTI. The coin originates from Thessalonica mint and it is dated in the period from December 308 to May 311.87 SACRED STRUCTURES The most important positions in the Romuliana interior had been taken by two temples located in the imagined axis of cardo, in the center of north and south side. The north temple was of the tetrastyle prostylos type and the south one was of the peripteros type. Apart from these buildings, the structural complex, which is also assumed to have been of sacred character, was discovered in the southwest section. Tetrastyle prostylos or small temple (C), as it is often called, was situated between the palace D1 and palace D3 in such a way that it fits into the grid of the walls of palace D3. Its longer side is in the east–west direction and an altar was erected at 8.25 m from it to the east. The temple entrance is facing east, so, considering its direction, it does not comply with the instructions of Vitruvius. According to Vitruvius, the temple should be oriented in such a way that when people approach the altar for sacrifice or other religious rites they should be facing east celestial sphere.88 The preserved Roman temples bear witness that this suggestion of Vitruvius had not been always observed. The podium of the small temple, around two meters high, is preserved. It consists of the perimetral and two parallel transversal walls that together with the staircase area create the rectangular ground plan 10.45 x 16.57 m in size. It was concluded, according to the size of the temple and its plan, that it was the tetrastyle prostylos. Its east wall was the stylobate of frontal columns, and second transversal wall separated pronaos from the cella. The stylobate length from the south wall axis to the axis of north wall is 8.95 m or 28.5 feet, if one foot is 31.4 cm. It means that the axial distance between the columns was 9.5 m, and the depth of pronaos to the wall of cella was of the same 87— After reading and interpretation of Professor Vladislav Popovi}, ^anak-Medi} 1978, 96 and note 228. 88— Vitruvius, IV,5,1. 89— Vitruvius, III,4,4. 90— Reinforcement in the northwest corner collapsed in the stepped shape, so it was incorrectly assumed to be the staircase, Srejovi}, Lalovi}, Jankovi} 1981, 87, 88. 77 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 41. the priest, as it was connected with adyton. It has been assumed that the cult of Cybele in the tetrastyle prostylos at Romuliana was associated with the cult of emperor.91 When the tetrastyle prostylos is concerned, there is most comprehensive material evidence about the stereobate (stereoba,thj), and insignificant information about its upper section. Thanks to the fact that stereobates of the Roman temples were identical,92 it was possible to reconstruct its original form according to the bases of the lateral façades and traces of upper courses within the walls, as well as according to the pieces of the final segment of the stereobate (Fig. 41). On top of the base finished in slanting blocks there was (like in all known temples) one rather high vertical segment – orthostat – concluded to be 88 cm high. The finishing segment of stereobate had two layers, of which the lower one was molded and documented by discovered fragments, and the upper one consisted of the flat rectangular slabs. The back, west side of the stereobate was flat (plan XXIV). Further investigations of the original structure of the tetrastyle prostylos were carried out in two directions. First, the height of frontal columns has been established on the basis of their bottom diameters calculated according to the fragments of bases corresponding to the width of their stylobate, and following the scales recommended by Vitruvius for the Corinthian columns and the size of the column shafts in the neighboring atrium of palace D1. The conclusion drawn from this comparative analysis is that the diameter of base, i.e. the Stereobate of tetrastyle prostylos examples, as the bases of columns, which together with niches created tripartite altar (adyton). The structure in the crypt underneath the adyton had been related to the rituals connected with the shrines of Magna Mater, i.e. Cybele, where there are crypts entered by the believers to sprinkle the eyes, tongue and forehead with blood of the sacrificed animal. So the crypt in Romuliana prostylos was fossa sanguinis, the type of the underground baptistery, but it differs from the other known structures of the kind, because it was entered only by 0 PLAN XXIV 1 2 3 4 5m Stereobate of tetrastyle prostylos, ideal reconstruction according to which it was restored 78 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE Tetrastyle prostylos, ideal reconstruction (3D) Ideal reconstruction of tetrastyle prostylos FIGURE 42. PLAN XXV apex of the arches was at 5.45 meters, i.e. 18 feet, what was also the span between the corner reinforcements, and that would be in accordance with the grid according to which it was designed. According to the possible height of the arches it was possible to calculate that the cornice at the base of the barrel vault was at 20 feet or 9.6 m and the height of the vault was 33 feet, while the roof ridge was at 37 feet or 18.50 modules, what is corresponding to the results of investigations of the original temple façade (plan XXV, fig. 42). The most lavish among the fragments of architectural decoration from the tetrastyle prostylos, considering motifs and module, was 2 feet, and total height of the shaft 9 modules or 18 feet. Considering these measures it results that the column with capital, plinth and base was 21 feet high. It could have been calculated that epistylium together with architrave, frieze and the cornice had 3 additional modules and that cornice on the lateral sides was in line with the cornice within epistylium at the height of 27 feet. The tip of tympanum on the front façade was established considering the slope of roof used in the antique architecture. Thus it was calculated that the total height of the temple without stereobate and to the roof ridge was 18.50 modules or 37 feet (plan XXV). Second course of investigations of the original structure of the tetrastyle prostylos is based on the analysis of its interior structure. Even though the remains of upper structure are not preserved, it could be reasonably assumed that along the longitudinal walls of the cella there were arches, and above them was the barrel vault (plan XXIIII). It could be assumed that the 91— Srejovi}, Lalovi}, Jankovi} 1981, 68–70, with parallels and earlier literature on fossa sanguinis. 92— Stereobate composition of Roman temples quoted Vitruvius in his book III, in the chapter On temples foundation, Vitruvius, III,4,1. 79 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 43. The south section of the temenos has not been encountered, but the architectural elements, which perhaps belonged to the colonnade of the north or west portico, have been discovered. Thus, many column shafts carved together with bases and identical to the columns from the second gallery of west fortification gate have been found within the temenos. If they did not come from the north rampart, they certainly belonged to the colonnade of the temenos portico. The palace D3 was reconstructed when the temenos was built in order to connect the temple temenos and the palace. At that time was demolished the semicircular apse of its central hall and large staircase towards the temenos was built instead, while rather wide opening of the apse was diminished and the door was created. For the pilaster within the portico interior with adjoining colonnade the beginning of the curve of the palace D3 apse had been used. The portico was completed when temenos was extended westward. It has been concluded, considering the building technique and material used, that walling up of the apse and building of temenos and upper section of temple, together with the corner reinforcements and the outer wall of the crypt, had been carried out at the same time. It could be concluded, considering the junction of temenos and northwest corner of the palace D1, that finishing works on the temple were carried out only after the palace had already been built. Peripteros had been selected as the form for the main temple of Romuliana, identified during the excavations as large temple (I). It is situated in the center of Romuliana, on the best, south side. Many segments of its upper section have been found besides the remains of stereobate. This temple was also oriented in the east-west direction with the approaching staircase on the east side and with massive altar at 9 meters from the entrance. The temple foundation is slightly sloping to the east and its size is 32.4 x 23.8 m. On top of the foundation are the remains of stereobate, of which the core and few fragments of the outer facing are preserved in situ. The impressions of facing stone blocks are visible in the core, and according to this, we know that the orthostat was built in Portal border of tetrastyle prostylos, detail craftsmanship, are the fragments of the portal, carved of gray tuffaceous sandstone (Fig. 43). As in the crypt under the pronaos were discovered the threshold fragments collapsed right next to its original place, the complete portal appearance could have been devised according to the data in Vitruvius about the proportions of the portals.93 The fragments of cornice and frieze of white limestone, assumed to be the elements of interior stringcourse under the base of the barrel vault, were also found in the course of archaeological excavations.94 A whole with the tetrastyle prostylos created the temenos, whose remains were encountered north and west of the temple. It consists of the enclosing wall with portico along the north side and the part of the west side. The remains of that portico include the column bases with the trace of plinth of one of the columns. The north wall of the temenos, that is leaning to the side rooms of the palace D3 in the east, was completely uncovered. In the west it was also not incorporated into the northeast corner of the palace D1, but is just leaning onto it. The temenos wall is not parallel with the north temple wall, because the idea was to provide approach to the temenos from the palace D3. Besides the north section of temenos its west section was also encountered, and its external wall was at unequal distance from the east wall of palace D1. Its east section consisted of the wall built next to northwest corner of the temple, and it was the stylobate of the west colonnade of the temenos. 93— Vitruvius, IV,4. 94— Similar stringcourse in the interior at the base of barrel vault is encountered in Jupiter’s temple in Diocletian’s palace in Split, Buli}, Karaman 1927, sl. 55. 80 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE 0 5 10 m PRESERVED PARTS OF THE CORE MADE OF MORTARED STONE PLAN XXVI Plan of peripteros at the crypt level with modular grid (M = 9 ells, ell = 42.2–42.7 cm) analysis to establish the geometric-structural scheme which could have been employed in designing and measuring the ground plan of stereobate. It is based on the project module of 9 x 9 ells, derived from the east side of stereobate. The staircase consists of three units, and the lateral segments of one unit each. The ratio between length and width expressed in modules is 7:5. There were also data about 13 meters wide approaching stairway. The lateral sides enclosing the stairs are preserved, as well as their masonry base, so their slope could have been established. The results were coordinated with the data about the position of the preserved section of the cella floor, according to which the height of the pronaos floor was also known. the opus quadratum technique. The 4.5 meters high stereobate created wide podium along the lateral sides of the temple. It includes the crypt consisting of two linked rooms reached from the cella by the stairs running along the south part of east wall. Both rooms have recently reconstructed longitudinal barrel vaults. In each room are two deep windows splaying to the inside and vaulted with the conical vaults (plan XXVI).95 Few detached limestone blocks, some of them with beveled sides, and fragments of stringcourses from the outer side of the stereobate have been found, while the method of construction of the vertical part of stereobate and its height were inferred from the impressions of the removed stone blocks (Fig. 44).96 The analysis of the available data confirmed that the stereobate was 30.20 m long and 21.76 m wide and the individual measures suggest that it was measured by ell of 47.2–47.7 cm, so it was 63 ells long and 45 ells wide. It was possible by further 95— Srejovi} 1983 S, 43–46. 96— Stojkovi}-Pavelka 1988/89, 136, 137. 81 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 44. Remains of large temple – peripteros These data were compared with the recommendations of Vitruvius about the size of steps,97 and it was concluded after all calculations that there were 19 steps 23.15 cm high.98 The perimetral walls of the cella, 2.6 meters high, are preserved of the upper section of the temple. They extend on the east side into antae built of blocks of white limestone in contrast to the walls of cella built in the opus mixtum technique. In the east corner of the north wall is an entrance and landing of a staircase leading upwards and incorporated into the wall. Of the one time adyton on the west side just the part of the outer wall 10 cm high and with shallow niches separated by four wide pilasters is preserved. It has been assumed, according to the preserved temples with the three-sided porches,99 that the west wall extended to the north and south, closing the porches on these sides (plan XXVIII). Many architectural elements from the outside of the upper temple section have been found. The most important for reconstruction of the temple entirety are two types of the column shafts, bases and capitals. The smaller shafts, 47 cm in diameter and 3.87 m high, were carved of the white marble from Proconnesus, while the larger ones were of green-colored serpentine breccia, 65 cm in diameter and 5.52 m high (calculated according to diameter). There were also found three Ionic capitals, which correspond, according to their size, to the marble shafts, and one Corinthian capital.100 Two variants of the temple elevation have been envisaged on the basis of these Ideal reconstruction of peripteros front façade after hypothesis of Dragoslav Srejovi} PLAN XXVII finds. According to the first variant, the temple of the octastyle peripteros type had two stories (plan XXVII). In the lower section were supposed the taller columns with the Corinthian capitals, and on the first floor were the shorter columns with the Ionic capitals.101 However, the possibility that taller columns emphasized the frontal façade and shorter decorated side façades102 has not been considered impossible. In the course of further investigations of the temple elevation the priority 97— Vitruvius, III,4,4. 98— More details about designing method and staircase reconstruction, Stojkovi}-Pavelka 1988/89, 138. 99— Ward-Perkins 1974, Fig. 6, 45, 65, 83. 100— Srejovi} 1983 S, sl. 43, 44. 101— It has been concluded that this variant does not correspond to any canonic model, Duval 1987 A, 80. 102— Srejovi} 1983 S, 45. 82 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE 0 5 10 m PRESERVED PARTS OF THE WALLS OF CELLA, PAVEMENT AND ADYTON PLAN XXVIII 0 PLAN XXIX 5 Plan of peripteros, ideal reconstruction 10 m 0 Possible elevation of peripteros façade with two-level roof PLAN XXX 83 5 10 m Possible elevation of peripteros façade with simple gable roof MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA was given to the second variant, and it was examined in detail.103 In that analysis also two types of columns have been taken into consideration. As those made of white marble are too short for the large and certainly high frontal façade, it has been assumed that they were arranged along the north and south side of the cella, and that the green, more massive columns were used for the front façade. Their color would also satisfy the tendency for polychromy, characteristic of the Late Roman architecture. In that case the columns of different height were used within a single entity. Such combination, if there was continuous three-sided porch, required the establishing of an interlink to connect the columns of diverse height in one structural unit. Three possibilities, which satisfy structural conditions, have been investigated. Two variants are with two-leveled roof, where lateral sides rest on the marble columns with Ionic capitals and architrave (plan XXX). The designs suggested by these variants are unknown on the temples of peripteros type. In principle, such design existed on the mausoleum of emperor Diocletian in Split, but it has an octagonal ground plan. It had lower porches around the temple, except on the façade, where the higher columns were used, thus creating the deep tetrastyle pronaos and it, primarily, indicates the diversity of forms created in the time of tetrarchy. Such phenomenon in the Late Roman period is confirmed by the temple with porch of identical height on all four sides with the cella surmounting it, and it was, according to ideal reconstruction the form of the Altbachtal temple in Trier.104 Therefore it seems that more probable for the peripteros at Romuliana is the third variant. According to it, single roof covering not only cella, but also side porches, was assumed (plan XXX). The difference in column heights had been overcome by constructing the arches on the side porches. The arches would also make possible greater spans, so just eleven columns would be necessary105 The Ionic capitals intended for the side colonnades could not support the arches independently, so in this variant there must have been imposts of the type used also for the second gallery of the west fortification gate. The corner column in this variant should have been replaced with the stronger pilaster, as the marble column would not be able to sustain the thrust of the arches from two directions, meeting at an angle of 90°. On the façade and opposite to antae must have been masonry pillars, while there should have been also four Corinthian columns, supporting epistylium central arch (plan XXX). In both suggested variants the arched structure PRESERVED PARTS OF THE TEMPLE 0 PLAN XXXI 5 10 m Transversal section of peripteros, ideal reconstruction was envisaged on the central, high columns, and this design had been employed rather early for the temples with the combination of columns and pillars on the façade, like on Hadrian’s temple in Ephesus.106 In favor of the last suggested design speaks also the internal structure of the Romuliana peripteros. The considerably stronger north and south wall of the cella do not leave any doubt that it was vaulted with the longitudinal barrel vault. Investigations of the possible height of the vault and position of the stringcourse resulted in conclusion that interior space of the cella was inscribed in the square, so the vault apex is at the height corresponding to the width of the cella, and stringcourse is at half that height (plan XXXI). The temple was covered with marble tegulae, whose remains have been found. They were certainly inserted in the masonry substructure above the vault that was descending also over the side porches. Such design 103— Stojkovi}-Pavelka 1988/89, 135–145. 104— Ward-Perkins 1974, 234, fig. 272. 105— About the number of columns along the lateral sides in analyzed variants see more details in Stojkovi}-Pavelka 1988/89, 141. 106— Ward-Perkins 1974, 280, fig 349. 84 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE FIGURE 45. It has been assumed that the large peripteros at Romuliana had dual purpose and was dedicated to two deities. It had been supposed that the crypt was dedicated to some chthonic deity, while the upper part with cella was dedicated to some Olympian deity. The dedication of the upper part of the temple is assumed on the basis of discovered fragments of sculptures and their iconographic traits. Thus, it has been concluded that the colossal statue of emperor and two figures of Hercules were in the cella of the large peripteros at Romuliana, while the discovered figures of youths and girls with torches in their hands were in the sanctuary in the crypt.109 The more precise identification of discovered deities confirmed that one figure was of Hercules and other of Jupiter, so the peripteros was dedicated to Jupiter and other deities of the tetrarchy.110 It should be, however, mentioned that use of the crypt for certain rituals must have been impeded by the fact that the staircase for entering the crypt is next to the east cella wall and partially under the main portal, so it must have been covered with slabs and opened only on special occasions. The cruciform building (E) was built in the center of the large courtyard in the southwest section of Romuliana. The courtyard of square plan was enclosed within a wall having 31.5 m long sides. It created closed space around the cruciform building. There are enough remains of 60 cm thick surrounding wall to establish the size of the courtyard. It seems that it had porticos with pillars and columns on the inside.111 The building constructed in the center of the courtyard consists of few rooms clustered around the large central hall, some kind of the inner yard (plan XXXII). The entrance to this building complex is in the middle of the north surrounding wall and just opposite of it is the door leading to the vestibule of the cruciform building (II). There are two pilasters creating shallow niches in the north wall, next to the entrance. This rather small room (6.45 x 4.40 m) was lavishly decorated. Of that decoration there were found the remains of floor mosaics. Next to the sides of vestibule is one rectangular room each (I, III), and their significance and Peripteros, ideal reconstruction (3D) also included the construction of supporting vaults above the porches. We know that there was certainly such space in the north and it served some purpose. The assumed internal structure of peripteros complies mostly with the last suggested variant of the temple exterior (Fig. 45), primarily because its higher porches along the sides contributed to the static stability of the large cella vault and because there was a room above the north porch, suggested by the beginning of staircase leading to it. In addition, this variant seems more plausible because its assumed structural composition has the analogy with the architectural elements in Diocletian’s palace in Split. These elements are the junctions of low arcades with higher columns of the peristyle, shape of its prothyron, three-quarter engaged columns and arched structures in the colonnades. So, new structural elements had already been employed in Diocletian’s palace in Split and it is characteristic of the Roman architecture in the time of tetrarchy, when new spatial forms had often been introduced.107 The design of Romuliana peripteros was probably also influenced by some Syrian temples with three-sided porches, like the large temple in Hosa Sfira, the temple in Kalat Fahra with six columns on the façade and the temple in Hibaria with the vaulted crypt.108 107— Ward-Perkins 1968, 284, 312. 108— Krenker–Zschietzschmann 1938, Taf. 14, 20, 89, 90. 109— Srejovi}, Lalovi}, Jankovi} 1981, 72. 110— Vasi} M. 2007, 46. 111— Srejovi} 1985, 61. 85 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 5 10 m Plan of cruciform structure (E) in southwest part of the interior Assumed stereometric view of cruciform structure (E) PLAN XXXII PLAN XXXIII distinct purpose is confirmed by remains of lavish mosaic floor. The central room of square plan (IV), with sides 6.45 m long, was also entered from the vestibule. This room differs from the others by the type of floor, which was made of rectangular marble slabs. The rooms at the sides of central hall have not been explored, but it could be assumed that they also had mosaic floors like the room VII, which projects southward from the main building mass.112 Its position indicates that it was the most important room in this building complex. The remains of frescoes have also been found in all investigated rooms. There was no information about the upper structure of the cruciform building. Considering the thickness of the walls (90 cm) and outside reinforcing pilasters, here and there it was covered with vaults. In the central room, judging by the corner reinforcements, was the cross vault, and in the lateral rooms V, VI, VII and vestibule II were barrel vaults, perpendicular to the central room. From the outside this building had cruciform articulation (plan XXXIII). Its distinctive spatial composition together with luxurious floors and fresco painted walls suggests that it served a special purpose. This is confirmed by the quality of mosaics, which are equal to the mosaics in the Galerius’ residential palace (D1), so it seems reasonable to assume that they were executed by the same artisans.113 It has been assumed that this building was a shrine for Galerius’ mother Romula, who, according to the antique writers, exceptionally venerated the mountain deities and offered them sacrifices every day, followed by ritual feasts. On the basis of this information it was assumed that there was at Romuliana the special shrine of the mountain deities, with altar and distinct area for sacred feasts of the believers, so the cruciform building in the southwest section of Romuliana was intended for that purpose, i.e. that it was monumental triclinium.114 This hypothesis was challenged on the assumption that it was a spatial entity used as a reception area.115 This assumption is contradicted by exceptionally lavish trimming of this building, its central plan and cruciform exterior, as well as the enclosing wall surrounding the yard resembling temenos. Nevertheless, the assumption that it was the shrine for Galerius’ mother Romula could be considered out of question, because, according to our former information, it had been built after Romula’s 112— Srejovi} 1985, 53, 54; Vasi} ^. 1997, 19. 113— Srejovi} 1985, 50, 54. 114— Srejovi} 1985, 62, 63 115— Duval 1987 A, 78. 86 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE FIGURE 46. Palace in northwest insula (D1, 2), aerial view and probably triangular tympanum (Fig. 47). The south hall at the entrance (1), 7.5 x 42.7 m in size (plan XXXIV), was connected to the east cryptoporticus (16), the open courtyard next to its east side (2) and the large west hall (3), placed at the right angle to the entrance hall via two doors in the north wall and one large door on the west, frontal side. Both lateral walls of the entrance hall are on the outside divided by pilasters (Fig. 48), but of uneven size and at different interdistance. The outside pilasters have slanting bases made of bricks, and on the sides facing the courtyard the pilasters start from the flat platform, 25 cm high. The identical bases have the pilasters in one interior courtyard of Galerius’ palace in Thessalonica. There the pilasters are preserved to a greater height, so we know that they extend into the blind arches and that their foot was covered with variegated stone slabs. The interior decoration of the entrance hall consisted first of all of the mosaics covering the entire floor area. The floor consists of three mosaic carpets (plan XXXV). The emblem depicting labyrinth, that was of prophylactic character, is in one panel near the entrance. In addition to the lavish floor, the death. The comparative studies of the purpose of buildings with similar spatial composition at Palatine, in Diocletian’s palace in Split and at Piazza Armerina it was confirmed that this cruciform building at Romuliana was a triclinium,116 but the exceptional luxury of its rooms indicates that it was intended for very formal feasts. PRIVATE PALACES AND PUBLIC BUILDINGS Palace D1, 2 covers the entire northwest insula of the Romuliana interior (Fig. 46). It consists of few large halls, one peristyle, two atriums and two open courtyards, and it also includes few rooms with heating installations and a group of economic and auxiliary rooms. All the rooms are arranged according to the orthogonal grid and oriented according to the cardinal points, and fitting into the same grid is also the easternmost room – cubiculum, as well as the group of rooms surrounding second atrium (D2). Three palace entrances have been discovered, and the main entrance was in the east. In front of it was the large area paved with rectangular marble slabs. The entrance is 4.12 m wide and flanked by two strong pilasters, supporting the free-standing columns. Fragment of one massive shaft of green-colored serpentine breccia was found just in front of the entrance. They supported upper structure, consisting of capitals, epistylium 116— Wulf-Rheidt 2007, 71, 72, Abb. 13. 87 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA a b Entrance to the residential complex – palace D1: a) from the east; b) from the southeast, ideal reconstruction (3D) FIGURE 47. is the semicircular niche, projecting outside (Fig. 49). The channels for supplying and draining water were leading to the niche, so it could be assumed that it was used for the ritual of purification. There were on the outside of the west wall of this hall rhythmically arranged pilasters, which terminated, as in the previous room, with blind arches interior decoration consisted of stone slab facing of the walls, stringcourses and magnificent wooden ceiling. The west hall (3) of almost same width (7.59 m), but of considerably smaller length (30.50 m), was approached from the entrance hall via three steps. In the center of its west wall is an auxiliary entrance to the palace, and nearby, in the same wall, Pilasters on south façade of palace D1 Semicircular niche in the west wall of hall 3, palace D1 FIGURE 48. FIGURE 49. 88 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE 0 PLAN XXXIV 5 10 m 0 5 10 m Plan of palace D1, 2, graphic design with denoted rooms and atrium PLAN XXXV Plan of palace D1, 2, ideal reconstruction such niches, could have been excluded, because of the steps in front of it, ending next to the wall at the foot of supposed throne. The walls of this large hall are much stronger than the walls of the entrance and of the west hall, indicating that it was vaulted with the longitudinal barrel vault. Judging by the discovered portable finds and remains of the mosaics, it was the most luxuriously decorated room. It had three mosaic carpets on the floor, and the middle one, which was of the same width as semicircular niche, was decorated with figural compositions. It was concluded on the basis of preserved segments that a hunting scene was represented. The more complete idea about the luxurious decoration of this hall, and particularly about its upper destroyed section, could be acquired according to the numerous fragments of stucco decoration, frescoes and variegated stone facing with special fragments shaped as lesene. The date of the finishing works on decoration of this hall could be established on the basis of one coin found in the mosaic substructure. Its obverse reads as VAL LICINNIANUS LICINNIUS PF AVG, and the reverse GENIO A–VGVSTI. This The interior decoration of the west hall (3) was somewhat less luxurious, judging by the remains of the mosaic floor executed with the larger cubes (tesserae) than the others, and with only one carpet with rather simple geometric motifs. This hall also had wooden ceiling. From the west hall via wide door and down three steps one descends into the large hall, 10.95 x 36.06 m in size (4), with the semicircular niche on the east side (Fig. 50). Particularly important for establishing the purpose of this hall is that niche, facing the entrance. It was flanked with the free-standing columns, of which the bases could be identified. They supported the columns with epistylium and triangular tympanum, emphasizing the importance of the niche and probably its cult character. The floor of the niche is raised 92 cm above the mosaic floor and there are two steps, so it is 53 cm above the last step. There were discovered the traces of a square field on the floor of the niche, but lower than the surrounding wall, for which it has been assumed to be the podium for the stone throne. The possibility that the niche housed the statue of some distinguished person, what was usually the purpose of 89 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 50. East segment of hall 4 of palace D1 coin was minted in Cyzicus and is dated between AD 309 and around AD 311.117 The room (5) of octagonal plan, located south of the hall 4, is entered up the three steps (Fig. 51). Under its floor are the heating installations, and in the south wall is the arched mouth of the furnace (praefurnium). The channels for conducting hot air have not been encountered, although the floor was made of hydraulic mortar – opus signinum. This room had luxurious stone facing, of which many slabs of Greek green marble with white veins have been found. The purpose of this room could be determined not on the material evidence, but according to the similar structures and written sources. The closest analogy exists in the Galerius’ palace in Thessalonica, where there is also a small octagonal room, next to the main hall with throne. On the other hand, the most important among the written sources is one book of much later date, the book by Constantine VII Porphyrogenitus, called On Ceremonies (De ceremoniis), where he collected and described the treasures of the ancient heritage.118 According to Preserved section of octagonal room (5), palace D1 FIGURE 52. Peristyle (6), palace D1 FIGURE 51. Constantine VII Porphyrogenitus, the room where vestments and insignia had been kept was of that shape. He also adds that it had an important role in the imperial rituals.119 Dressing of the emperor had symbolic and sacred meaning and was sometimes accompanied by ritual washing, like in Thessalonica for instance, where washing facilities existed in the octagonal room. 117— After reading and interpretation of Professor Vladislav Popovi}, ^anak-Medi} 1978, note 239. 118— Constantin VII Porphirogénète 1935, 1 sq. 119— Constantin VII Porphirogénète 1935, 26 sq. 90 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE PLAN XXXVI 0 1 2m 0 1 2m Longitudinal section of peristyle (6), ideal reconstruction PLAN XXXVII Transversal section of peristyle (6), ideal reconstruction Two doors were leading from the hall with throne (4) to the peristyle (Fig. 52) with porches along all four sides (quadriporticus). The peristyle was also approached from the long Some installation of similar purpose probably existed also in palace D1 at Romuliana, as the floor of the octagonal room was made of hydraulic mortar. Thus it was possible, on the basis of analogies and later sources, to determine the purpose of this room and also to understand the function of the most luxurious large hall 4 and type of possible imperial ceremonies taking place there.120 120— That purpose of the octagonal room was accepted later (Vasi} M. 2006, note 19). It was questioned by Dintchev (Dintchev 2007, 24), assuming that it was the resting room. 91 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 54. FIGURE 53. Northeast atrium (8), palace D1 North section of peristyle with entrance to the north hall (7), ideal reconstruction (3D) concluded that the Corinthian capitals were on higher and stronger columns, creating with the opposite pilasters the pronounced bay in front of the north hall entrance (7). These pilasters were subsequently added to the walls, which had already been plastered. Hence, we know that the entire peristyle colonnade was constructed later, but certainly soon after the construction of the surrounding walls. This could be concluded on the basis of traits of the Corinthian capital from the higher column, that has very close analogies with the capitals of Diocletian’s palace in Split. The shorter columns joined the taller ones at 88.5 cm under their terminal plane by the beams grooved into the taller shafts. They were, including plinth, capital and architrave, 3.58 m high, as was also the depth of the porches. The conclusion about the complex of the peristyle porches was supported by the find of the complete column shaft in the neighboring atrium 8. There was employed the same design with two higher columns in front of the entrance to the space next to the east atrium wall (Fig. 53), and along the other sections of the three-sided porch were shorter columns of the same size and composition as the shorter columns in the peristyle. It was reasonable to assume that epistylium with triangular tympanum was on top of high columns at both places, thus attaching special importance to the entrances in front of which they stood (Fig. 54). corridor – cryptoporticus (16) and there was also door connecting the peristyle with the west section of the palace, with economic and auxiliary rooms and atrium 8. The most representative entrance from the peristyle was leading to the large north hall (7). The peristyle is of rectangular shape, 15.30 x 23.50 m, the ratio of width and length being 2:3, so it could be concluded that one of designs recommended by Vitruvius for atrium construction had been employed.121 There have been found enough elements of original decoration of the peristyle to reconstruct its original appearance. There was almost completely preserved stylobate of the porch colonnades. There are column plinths, 10 to 15 cm high, arranged on the stylobate, except in front of the large north hall (7), where they were higher. So, because of that, there is different intercolumniation on the north side and the plinths there supported wider bases and stronger columns than those of the rest of the colonnade. In the course of excavation of the peristyle area, there were found few tens of fragments of two types of bases and many column shafts, many fragments of Ionic capitals and one damaged Corinthian capital. The study of these fragments resulted in establishing the original appearance of the peristyle porch (plan XXXVI, XXXVII). As two types of bases have been encountered, and the plinths supporting them were not of the same size, it has been concluded that the porches were not of the same height. The Ionic capitals correspond to the smaller columns, whose height could have been established with certainty, so it was 121— Vitruvius, VI,3,3: On the Courtyards and other Rooms in the House. 92 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE North section of hall 2, palace D1 FIGURE 56. View of north hall FIGURE 55. heating installations under the floor, that are connected to the praefurnium found next to its west wall. There is, on the inside, another wall of considerable thickness, parallel with the wall of the semicircular apse. According to this and to the bases of the free-standing columns next to the pilasters at the beginning of the apse, it has been assumed that there were the free-standing columns at both locations. This is indicated by the examples with series of columns along the inside of the apses, while in the columns at the beginning had been recognized some kind of triumphal arch like in two large halls in Piazza Armerina, and such feature seems to have existed in the large hall in Galerius’ palace in Thessalonica and in Aula Palatina in Trier.122 The assumption about such purpose of the wall along the inside of the apse has been disputed and evidence was presented that this 90 cm thick wall was used for sigma couches of the stibadium used for formal dining in the Late Roman times.123 The apse was vaulted with the semi-dome and rest of the hall was covered with horizontal wooden structure with decorated ceiling and not with the vault, although there are strong reinforcements on the outside of longitudinal walls that certainly The decoration of the peristyle center made a whole with the monumental entrance to the north hall (7). The peristyle was paved with large rectangular marble slabs all over, except the small section (2.1 x 3.7 m) on the east side, as there was no substructure for the slabs. The main feature of the peristyle center was a fountain, of which basin and fragment of pedestal have been found. According to the supplying and draining channels, it was concluded that it was standing just opposite the north hall (7), in the axis of its entrance door. It considerably increased the beauty and ambiental quality of the peristyle, and mosaics on the floors of its porches also contributed to its appearance. The hall whose importance is denoted by the pronounced bay, 18.5 x 11.2 m in size, in front of the entrance had on the north side a large semicircular apse, separated from the remaining space by strong pilasters (Fig. 55). In front of these pilasters were the free-standing columns, of which just the bases were discovered. This hall was connected with other open area – atrium 8, via the subsidiary door in the east wall. The hall has 122— ^anak-Medi} 1978, 107. Some other additional architectural elements on interior wall of the apse used to divide it in remarkable segments, were assumed also by Srejovi} 1983 S, 40. 123— Vasi} M. 2007, 72, with earlier literature about other examples of the same features in notes 9 and 10. 93 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA FIGURE 57. The floor execution was in accordance with the purpose of this hall. Thus the floor in the semicircular apse was higher for one or possibly two steps than the floor in the remaining section of the hall. The floor in the remaining section of the hall was decorated on the lateral and south side with the mosaics, including the emblem near the entrance with the representation of Dionysus with leopard. Rather large mosaic field of uniform composition, concluding on the basis of the mosaic borders, was in front of the apse (plan XXXV). There was probably depicted certain composition from the cycle of Dionysus. In the middle of the room was discovered the floor paved with bricks carelessly arranged in star-like geometric motifs, but it was just the substructure of much more lavish floor, executed in the opus sectile technique. Along the edge of the central field was the impression in mortar of the stone border, which surrounded the higher floor in the center. The difference in height between the central and side sections of the floor, and their decoration, suggest that special installation for dining existed in the center of this hall. There is no doubt considering the mosaic theme from the cycle of Dionysus’ triumph and the central raised zone that this hall was used as the dining room (triclinium).126 It was certainly the triclinium of complex character, intended for the dining of emperor and his retinue, but also of many other guests.127 The rooms built along the east wall of northeast atrium (8) represent the distinct assemblage within the palace D1. This assemblage includes circular vestibule, one room of the tetraconchal plan and one of the triconchal plan (Fig. 26). Only the room of tetraconchal plan had heating installations, of which the channels were discovered under the floor. Next to this room to the north were two smaller rooms, where, by all appearances, were the praefurnia. Despite the fact that the tetraconchal room, considering its ground plan, corresponds to thermae, there have not been found swimming pools nor supplying and draining water channels and such installations are also missing in other two rooms (9,11). Nevertheless, these General view of rooms 9, 10, 11, palace D1 supported the arches, because the identical ceiling and outside decoration was encountered at Aula Palatina in Trier (Fig. 56).124 The north hall (7) was lavishly decorated likewise the hall with the throne. On the wall was the veneer of marble slabs with molded borders also of marble. Above the borders were variously decorated smaller panels in the inlaying technique, as it is concluded on the basis of discovered tiles. There were various molded bands, tiles shaped as sectors of the circle, made of various kinds of stone and of diverse colors, and some borders were decorated on the front side with the motif of continuous waves. The slabs with carved lesenes were also found in this hall, besides various decorative borders.125 Identical repertoire of decorative tiles is known from the Galerius’ palace in Thessalonica. 124— On large halls from the time of tetrarchy more in Tóth 1978–79, 189–195. 125— Documentation with technical drawings in ^anak-Medi} 1978, 214, 215. 126— As is first assumed by N. Duval, Duval 1971, 120. 127— Dintchev 2007, 9. 94 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE ning. Somewhat inferior execution and more modest decoration of this entity do not leave any doubt that this atrium was the economic yard with the surrounding rooms of similar purpose. One of them was certainly the kitchen, one bakery, and others were used as storerooms and rooms for food preparation. The spatial organization of the palace D1,2 was partially influenced by the topographic situation. Also some former architectural designs had impact on the selection of the ground plan design and its concept. In contrast to the Hellenistic–Roman type of palaces and houses with axial arrangement of most important rooms, that had been built in the Roman Empire for rather long period of time,130 in the Pannonian regions was used the type of house with the orthogonal disposition of rooms, known from Carnuntum and Aquincum.131 The same was the spatial composition of the palace D1,2 at Romuliana. The palace the outside civil settlement at Carnuntum, that was of residential character, was of the same type, and there perhaps had taken place the meeting of four rulers, when peace in the Empire was in jeopardy in AD 308. There are closer analogies for the north section of palace D1 at Romuliana among the palaces believed to be the temporary residences of emperors in the central European regions, but also in the south parts of the Empire. The most similar to Romuliana, considering the spatial composition, is the 4th century palace in Laufenbach. There was a spacious peristyle in front of the large hall and next to its right and longer side was another courtyard with threesided porch and farther along this atrium was a triclinium with eight-sided vestibule surrounded by rooms (plan XXXVIII).132 Considering the functional structure of the residential palace at Romuliana, it has greater similarity with some other imperial palaces. Thus, the identical functional designs were previously employed on Palatine in Rome, in Diocletian’s palace in Split and certain characteristics of the same structure would have been repeated also in Aula Palatina in Trier.133 Regarding certain spatial arrangements palace D1 also resembles Galerius’ rooms were previously explained as parts of some kind of baths.128 More recent studies of the Late Roman custom of dining on the sigma couches got closer to more precise identification of their purpose. After analyzing many buildings with the conchs, their purpose and installations, the conclusion was drawn that such rooms in the Late Roman period were used as stibadia, so such purpose was also ascribed to the rooms 9, 10 and 11 at Romuliana.129 This group of rooms at Romuliana (9, 10, 11) was lavishly decorated. Thus the vestibule floor was decorated with mosaic and floors in other rooms were executed in the opus sectile technique, the walls had marble socle and were decorated with frescoes and geometric ornaments of multi-colored stone tiles. Another internal open courtyard (14) was reached via small rooms 12 and 13 and from that courtyard was entered the small east room with semicircular niche on the east side (15), 6.20 x 6.20 m in size without niche. It was also one of the representative rooms and was probably used at least in the beginning as the sleeping chamber. It had the heating installations under the floor, which was decorated with mosaics, segment of which is preserved in the semicircular niche. Small atrium with series of rooms (D5) was added to this room later, and this entire complex was used as the sleeping quarter. The series of rooms built around another atrium west of the peristyle and north hall (7) was also an integral part of the described section of the palace D1. There have been conducted just the test-trench excavations, so only the arrangement of rooms is known. The northwest atrium (1, palace D2) was approached by the door in the west wall of peristyle. This atrium, 24.80 m long and 15.70 m wide, was larger than the peristyle, and it had porches along all for sides, supported by strong masonry pillars. The porches were paved with various irregular stone slabs in the opus segmentatum technique. Along the south side of atrium and partially along its north and west side is arranged number of rooms, some of which are very large (6 and 7). The north wall of this section of palace D2 incorporated one of the the pillars of earlier fortification portico, and this confirms that it was built only after the portico had been demolished. Not a single wall of this spatial entity was structurally connected with the walls of the halls they are leaning to (4 and 7), but as the door facing westward were already opened in the west wall of peristyle (6), it is obvious that the construction of the northwest group of rooms had been planned from the begin- 128— ^anak-Medi} 1978, 108, 110. 129— Vasi} M. 2006, 69–72, 75; Vasi} M., 2007, 42, 43; Dintchev 2006, 24. 130— See comprehensive study about examples in Dacia and Thrace by V. Dintchev (Dintchev 2006, 9–24). 131— Swoboda 1958, 136, 147. 132— Tóth 1973, fig. 5/1; Tóth 1978/79, Abb. I. 133— Wulf-Rheidt 2007, 70, 71, Abb. 9, 10. 95 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 PLAN XXXVIII 10 20 30 40 was done by some architects who knew the spatial structure of the temporary imperial residences. Palace D3 in the northeast quarter has its longer side oriented in the southeast–northwest direction and all its rooms are within the orthogonal arrangement regarding the axial axis. The large central hall (3) has been completely explored and other segments of this palace were investigated by test trenches, so we know the disposition of rooms, but not their purpose. The palace was entered from the east via spacious vestibule (1), in front of which was, as it seems, some kind of portico. The peristyle, 18 x 32 m in size, with porches on all four sides, of which just the stylobate remained, was entered from the vestibule. Somewhat higher plinths, on which the columns were set, were encountered on the stylobate. The plinths bear witness that there were four columns along the shorter sides and eight columns along the longer ones. The plinths were carved of the same limestone as the plinths in the peristyle (6) of palace D1. The peristyle of palace D3 is surrounded by rooms. They are somewhat larger on the south side (20,18, 16) and series of eight almost identical smaller rooms was built in the north tract. There is a corridor and passage between rooms 12 and 11, leading to the north fortification gate – poterna. All the rooms on the east and north side and the corridor along the north side of the large hall 3 were approached from the peristyle. From the peristyle were entered the rooms 11 and 12, one of which had a masonry couch along its east wall. The rooms on the south side of peristyle were functionally separated and they had a special entrance via vestibule 17 from another open courtyard reached by the corridor along the west side of room 15. This corridor was connected via door on the south side with the building with corridor (D4). Opposite the entrance in the longitudinal axis of peristyle is the large hall 3, 11 x 22 m in size, reached via wide door in the east wall opening to the peristyle. It had a semicircular apse on the west side. This hall had the very strong side and front walls (2.5 m), indicating that it had a longitudinal barrel vault, certainly raising high and surmounting the porches of the peristyle if there were no rooms on the upper floors.134 This hall was decorated with frescoes and rather thick layer of polished mortar was discovered on the floor. Perhaps the luxurious mosaic floor had been also planned in this room and the mortar 50 m Graphic design of palace in Laufenbach, after E. Tóth palace in Thessalonica. In addition to the spatial composition there is resemblance in the decoration of certain rooms and in many details of craftsmanship that relates Romuliana palace with that imperial residence. This is particularly conspicuous in execution and decoration of the interior. Thus, the wall facing in some unearthed rooms of Thessalonica palace consists of the monochrome marble socle ending in the molded border identical with the border in the north hall (7) or in the tetraconchal and triconchal rooms of the stibadium at Romuliana. Also, in Thessalonica were discovered the fragments of the wall facing of variegated stone originating from the same quarries as the stone for the wall facing at Romuliana. There were encountered the identical slabs with carved lesenes like in palace D1 at Romuliana, so these details indicate that same artisans and masons were working on both imperial residences. As the coin discovered in the mosaic substructure in the hall 4 of palace D1 dates the works at Romuliana in the period between AD 309 and 311, it is obvious that the masons from Thessalonica were brought to Romuliana only after they completed most of the works in the imperial capital. Before that was already established the spatial outline of the palace as a whole, and that 134— Srejovi} 1983 S, 45, 46; Vasi} ^. 1997, 14, 15. 96 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE was just temporary solution. The same situation was encountered in the palace D1, where in almost all large halls and in the north atrium (8) has been found the polished mortar floor at the lower level, that was used for the short period of time.135 In order to connect this palace with temenos of the tetrastyle prostylos, the apse in the hall 3 was demolished and the staircase leading to the ground around the temple was built instead. The original wide opening of the apse was diminished at that time and this rebuilding was executed in the opus listatum technique (like the façade of the fortification gate of the earlier fortification), while the remaining walls of the hall 3 were built in the opus mixtum technique.136 This suggests two stages in its evolution. According to the spatial relationship between its east range and the tower II and portico of the earlier fortification, building started when the portico had already been demolished, while the alterations on the main hall apse (3) were carried out when the temenos of the tetrastyle prostylos had been built. The spatial arrangement of the palace D3 with its pronounced longitudinal axis, the peristyle surrounded by rooms and the centrally placed most important room results from the type of Hellenistic–Roman houses.137 According to its composition it is appropriate for a residence and its location within Romuliana suggests that it was used by some person close to the emperor. Because of that, it was assumed earlier that it was intended for Galerius’ mother Romula and her retinue. Such purpose of the palace D3 would be in accordance with its connection to the temenos of the tetrastyle prostylos, which was of private character, as it is indicated by the fact that it was separated from other sections of Romuliana. The suggested attribution of the palace D3 was brought into question after dating Romula’s death before the end of the 3rd century.138 Therefore, only future and more comprehensive investigations of this palace could provide the data about its purpose. Building with corridor D4 was parallel to the palace D3, built along its south side and making with it an open courtyard along its south tract. It was connected with the palace D3 by the door and was probably creating a functional entity with it. This building has been investigated only by test trenching, so we know only its position, direction of north and south wall and the beginning of partition walls. It is a rectangular structure, 18 meters wide and 80 meters long, oriented in the southeast–northwest direction. It is longitudinally divided in two segments. There is a 4 m wide corridor along the south wall. 0 PLAN XXXIX 10 20 30 40 50 m Plan of palace D3, 4 and atrium D5, after ^. Vasi} From that corridor were entered the rooms arranged along the north wall (plan XXXIX). This structure has four entrances on the south side and on the north side are two openings connecting it with another interior courtyard, which was perhaps the south section of temenos of the tetrastyle prostylos (plan XL). Despite the fact that the building D4 has not been completely explored, it was possible to conclude on the basis of partition walls that there was one large hall, 10 x 30 m in size.139 It is not known what the façade of this building looked like, and on its sides facing the main communications, on the east and south façade, there were no pilasters indicating the series of blind arches. It seems most probable that in front of these façades of the building with the corridor were wooden porches, like in front of the palace D1, creating thus harmonious visual whole, despite not being of the same height. 135— ^anak-Medi} 1978, 113, nap. 280. 136— ^anak-Medi} 1995, 54. 137— More on houses and palaces in Roman architecture: Swoboda 1924, passim; Paribeni di 1940, 131–148; McKay 1984, 26–68. 138— Vasi} M. 2007, 50, 51. 139— Srejovi} 1983 S, 46, 47, sl. 41; Vasi} ^. 1995, 316, Fig. 3; Vasi} ^. 1997, 14, 15. 97 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 PLAN XL 10 20 30 40 50 m Spatial relations of palace, building with corridor and tetrastyle prostylos in north section of Romuliana, after ^. Vasi} porticus of the palace D1 (16), where the door to enter the atrium was subsequently opened. The atrium and surrounding rooms were built after the building with corridor (D4), because their walls lean to the west wall of the building D4. But already during the construction of the building with corridor the west atrium had been planned, because the door was opened to the atrium from the room 8. This atrium makes the functional entity with the east room in the palace D1 (15 – plan XXXIV) and was built soon after the palace D1 was completed. Their close date of construction is confirmed by the stylobate and the column plinth identically shaped and of identical material as the corresponding segments in peristyle of palace D1. Of the interior decoration of atrium there are just remains of the floor in the west porch, executed in the opus segmentatum technique by the small asymmetrical stone pieces of different shape. The building with corridor had lavish interior decoration. The walls were covered with frescoes, of which fragments were discovered, and the floors in some rooms were made of asymmetrical small stone slabs in the opus segmentatum technique. It was built after the palace D3 and it certainly existed when the temenos of the tetrastyle prostylos was created, as there is a door between them. Atrium D5 and surrounding rooms are located between the palace D1 and the building with corridor (D4). Atrium had porches along three sides, of which only a segment of the disturbed stylobate and one column plinth were discovered, while the masonry podiums for the other plinths were also found. On the basis of this evidence, the stylobate has been almost completely reconstructed. It has been established that atrium had three columns on each side and smaller rooms (cubicula) were built on the east and south side. The atrium walls on the west side were just leaning to the east wall of crypto- 98 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE Building with portico (G) is situated to the south of the peripteros (large temple I) and it has been only partially investigated. It is of rectangular plan, 19.40 x 51.20 m, and its longer sides are oriented in the east–west direction. The 4 meters wide portico, of which the strong masonry pillars (1.5 x 1.5 m) have been discovered, is along its north side. The interior of the building with portico consisted, according to our present knowledge, of the narrow entrance hall (4 x 18 m) and the large hall (18 x 44.60 m), covering its entire interior (plan XLI). The building has two entrances. The main one, 2.8 m wide and leading into the large hall, is in the middle of the north wall, while the subsidiary entrance is on the east side and was leading to the entrance hall. It is not possible to draw conclusions about the purpose and about the upper structure of the building with portico, as it has not been completely explored. The great width of 18 meters could not have been spanned with the wooden architrave structure without series of free standing supporters in the center of the hall, while the use of the longitudinal barrel vault is out of question, because the side walls are not strong enough. The building with portico was built of blocks of strong tuffaceous sandstone and few rows of brick in the opus mixtum technique. The pillars and doorposts were built entirely of brick. It is known that the building with portico was plastered on the outside and that the pillars of the portico and north façade were covered with frescoes. It has also been established that it had the heating installations under the floor.140 Five-aisled building F was built to the west of the peripteros (large temple I) and its longer sides were oriented in the north–south direction. It is of rectangular plan, 51.20 meters long and 19.40 meters wide, and consists of vestibule (4 x 18 m) in the north and five-aisled hall, 44.6 meters long and 18 meters wide. The building F has two entrances. The subsidiary entrance is in the west wall of the vestibule and the main entrance is in the middle of the north façade (plan XLIIII). In the interior are four rows of six strong pillars (1.20 x 1.20 m) and opposite the pillars are along the north and south wall the pilasters with the slanting bases, identical to the façade pilasters of the palace D1. The aisles are not of identical width. The side aisles are 4.2 m and the central is 5.05 m wide. Despite the fact the that five-aisled building has not been completely explored, its upper structure could be envisaged. It consisted of longitudinal and transversal arches, creating the square bays covered with the cross vaults. Such upper struc- 0 PLAN XLI 10 20 m Plan of building with portico (G), after Srejovi} 0 PLAN XLII 10 20 m Plan of five-aisled building (F), after Srejovi} ture is not suggested only by detached supporters, but also by the method of reinforcing the perimetral walls of large hall. They were divided on the outside by pilasters, which certainly terminated in blind arches. Considering that there are pilasters also on the outside of the north wall of the large hall and that the wall of the vestibule was leaning to its northeast corner, it could be assumed that the vestibule was added later. 140— Srejovi} 1983 S, 51; Vasi} ^. 1997, 18. 99 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA 0 1 2 3 4 5m PLAN XLIII There is very little information about the interior decoration of the five-aisled building. Thus, the segment of the wall decoration was preserved on the mortar with some multi-colored geometric motifs.141 Judging by the lively colors and discovered column shafts in the interior, it has been assumed that this building had architectural decoration, which did not fall behind the decoration of palace D1.142 Nevertheless, such decoration would not be appropriate if the five-aisled building was horreum, as it has been assumed.143 Building with porch J was built between the building with portico (G) and thermae (Fig. 58). Only the north porch, 4.71 x 10.90 m in size, has been completely explored. On its frontal side are pillars, of which the corner ones are shaped as letter L, and two central ones are rectangular. The south wall of the porch has pilasters at the corners and in the middle is the door leading to the south rooms.144 Thermae H are located in the furthermost southeast section of Romuliana interior, at the lowest level of the terrain in comparison with other unearthed structures. The structure was lo- Plan of thermae (X) cated next to the columns of later fortification portico. The building complex was identified as baths, according to the arrangement and purpose of the rooms (Fig. 59). The baths were entered from the porch raised on two steps from the surrounding terrain and was attached to the portico extending along the building south of peripteros (J). The structure consisting of few parallel rooms is entered by the high threshold. First to enter is the spacious vestibule (1), and around it are arranged few rooms of different purpose. The archaeological investigations revealed the floor paved with slabs of marble and tuffaceous sandstone in different pattern. In the central area square slabs were arranged in the checkerboard pattern, while the rest of the room was covered with slabs 141— Srejovi} 1983 S, 51, 52; Vasi} ^. 1997, 18, 19. 142— Srejovi} 1983 S, 52. 143— Duval 1987 A, 80. 144— Vasi} ^. 1997, 18. 100 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE FIGURE 58. View of excavated segments of building with porch (J) FIGURE 59. Thermae (H) from the north The floor of this room was decorated with the mosaic with geometric motifs and the remains of it have been found in the northwest corner. On the east side of the vestibule and opposite the entrance to semicircular room (2) was the shallow recessed area, approached from the north via two steps of tuffaceous sandstone. The floor was paved with bricks. The appearance the of recessed area indicates that this was cold water bath, but there were no traces of hydraulic mortar on the floor or of the installations for water drainage. The use of tuffaceous sandstone for the staircase approaching the semicircular room (2), the construction of the rectangular pool and floor of marble and tuffaceous sandstone slabs could be ascribed to the same phase of later interventions on the thermae architecture. Next to this rectangular pool was constructed a horseshoeshaped structure (5), surrounding the closed pool area, which is separated from the vestibule by the high partition. On the walls and floor of the pool were found the traces of hydraulic mortar, and on the east side was discovered a lead pipe for emptying the pool. This was undoubtedly the pool with the cold water – frigidarium, where the visitors bathed before entering the warm sections of the baths. In the middle of the vestibule south wall was the door opening leading to the heated area of the baths. This area was divided into many rooms and all of them had the underfloor heating. Hot air was conducted from the furnace situated in of different dimensions. The remains of mosaic floor, sloping toward the center of the room, were found next to the south wall of the vestibule. The spatial and chronological relationship between these two floors is so far impossible to establish with certainty. On the west side of the vestibule is a semicircular extension (2) – five-sided on the outside and separated from the central area by a solid wall. It was approached via massive steps. Certain evidence suggests that this apse was added later. Two plastered and painted semicircular niches existed originally in the west wall of the vestibule. They were probably used for the sculptures of Asclepius and Hygieia, gods associated with hygiene and baths. These two niches were walled up at one time and the south one was transected when the opening for five-sided apse had been made. Also, there are conspicuous vertical joints, which separate the apse wall and the vestibule wall. Although the apse has been archaeologically investigated and mortar floors were discovered, it is not possible to draw conclusions about its purpose. In the beginning of excavations many wall mosaic cubes, some even with gold paste, have been found in the debris. The one time lavish decoration and raised position of the apse indicate that it had been built for some prominent person, perhaps for Galerius himself. The rectangular room (3), entered by two passageways separated by the masonry pillar, is situated to the north of main hall. This room with masonry bank along the walls is identified as apodyterium – area, where visitors changed their cloths. 101 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA the southernmost section of the structure. The room entered first (6) was the furthest from the furnace, the air was not so hot and tepid water was used – tepidarium. This area has a semicircular niche on the west side and there the bathtubs were probably located, as there were no traces of the hydraulic mortar indicating the existence of the pool. The partition preserved only in the foundation zone separated tepidarium from the hottest section of the thermae – caldarium (7). This area was closest to the furnace and there was the niche on the west side, identical to the one in tepidarium, and the semicircular pool placed in the east. The pool was on top of the hypocaust covered with hydraulic mortar. The mortar lined channel, which defines the lowest level of the pool, was used for emptying it. Next to the mouth of the furnace (8) – praefurnium, where the air is hottest, there was the thick layer of hydraulic mortar on top of hypocaust structure, while the bricks – tubuli for vertical heating of the walls were discovered along the wall. This area is assumed to have been used as sweating room – sudatorium. Another caldarium room (9), with the hypocaust installations on the small posts of ceramic pipes, was additionally built next to the thermae south wall. The furnace for heating this area was most probably located on the east side. The channels for supplying and draining water, that are considered earlier than the room 9, have been discovered on the outside of the south side of this room (9). Two channels were used to supply water and one to drain the waste water. It is assumed that one channel of rather large diameter (0.80 cm) was conducting the waste water through the east rampart and outside the fortification, and that it was the main channel of the earlier fortification. These channels transected one even earlier lead water pipe running in the north–south direction.145 All this bears witness that some of the discovered channels date from the first phase of the thermae at the latest, and that the large main drain and the lead pipe are even of earlier date. The appearance of the upper sections of building H that are missing could be supposed, because the baths and particularly their heated rooms were vaulted; the semicircular niches the with semi-domes and the long rooms with the barrel vaults. In the scholarly literature concerned with the architecture of Romuliana the building we discussed was identified as thermae, what is a common name for the Roman baths.146 In the more recent studies the term thermae is generally suggested for the large baths for public use, built by the emperor or by the state, while the structures of balnea type were usually of smaller size and intended for the small number of visitors.147 On the other hand, the term balnea was used in the traditional terminology for the spas, so this terminological question should be the subject of the future investigations. But, as the structure H at Romuliana was certainly visited not only by the emperor, but also by his retinue, it was of public character, so the term thermae is quite appropriate. The established phases in its development and later annexes suggest its long duration and its installation shows the adaptability of its architecture to the importance of the visitors. *** It could be noticed, considering the presented data about the buildings and building complexes at Romuliana, that they have not been equally investigated, and some are not sufficiently known, but by a lucky combination of circumstances the most characteristic buildings and building complexes have been completely explored. Thus we are informed about the architecture of the earlier and later fortification, of the palace in the northwest quarter of the interior, that was identified as the imperial residence (D1,2), of the sacred structures and of one of rather important public structures – the thermae. The knowledge about their architecture is completed by the investigations of their spatial organization and purpose, of their sculptural decoration and equipment of their interior. The later fortification and its main gate are distinguished for their architecture and distinct characteristics. Their architecture of the fortification gate is unique in its spatial composition, certain structural designs and particularly in its stone-carved decoration. They are considered among the most monumental fortification structures of the Late Roman period. There were employed the complex structural designs in vaulting the interiors of the polygonal towers with the conical vaults erected on the circular plan. The similar complex structural design was 145— Petkovi} 2008 A, 61–63; Petkovi} 2008 V, 64–67; Petkovi}, @ivi}, Kapuran 2009, in print. 146— Vasi} ^. 1997, 17, 18; Lalovi}, Ru`i}, Jovanovi} 1997, 199; Lalovi} 1998, 126; Lalovi} 2000, 125. 147— Yegül 1992, 43. 102 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE FIGURE 60. West fortification gate of late Romuliana fortification, ideal reconstruction (3D) FIGURE 61. North fortification gate of Diocletian’s palace in Split employed also in the crypts of peripteros (temple I), where conical windows penetrate the barrel vault at right angle. The structural whole of the fortification gate would not be distinguished from other known capital fortification gates except for its height and span of the polygonal towers (Fig. 60), if they did not have the lavish carved decoration unknown on the structural complexes of the same purpose. One of the most impressive is considered to be the Golden Gate of Diocletian’s palace in Split, embellished by the sculptures in the niches and on the pedestals near the top, but the execution of its secondary details with the series of arcades on the free-standing columns was reduced to classic architectural forms (Fig. 61). Considerably later, Golden Gate in Istanbul, from the time of emperor Theodosius I (378–395), could be compared in size and monumentality with the Romuliana fortification gate. The gate in Istanbul, according to certain unconfirmed assumptions, had lavish carved decoration on the secondary sections above the ground floor level.148 The character of the sculptural decoration and composition of its ornamental ensembles contribute, besides the structural entity of Romuliana fortification gates, to their distinct characteristics. Thus, we can see on their capitals that they have the reduced repertoire of motifs characteristic of the sculpture from the beginning of the 4th century. The modeling of motifs also corresponds to that time. Other elements with sculptural decoration: consoles, stringcourses, pilasters, with diverse motifs covering all free surfaces, correspond, according to their characteristics, to the Late Roman period as accumulating of sculptural decoration on secondary segments of the buildings is characteristic of that period. When modeling is considered, sculptural decoration is not uniform. Some pieces are works of local, unskilled stonecutters, but the sculptural decoration of Romuliana fortification gates corresponds in general, according to the motifs and modeling, to the time of tetrarchy. The inconsistency of execution suggested the conclusion that this was provincial production,149 but this is correct only to some extent. Certain signs of decline in the building technique could be perceived in the structure of decorative ensembles on the first and second gallery of the fortification gates. The elements of these ensembles do not have any more classic forms and it is 148— Grabar 1963, 58. 149— Duval 1987 A, 69. 103 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA particularly important that their composition and entirety are considerably different from the composition of the classic secondary parts. Because where there are on the earlier Roman structures the arched structures flanked with the engaged or free-standing columns, the columns are usually extended above the arches and connected with the architrave structure, but it was not the case either in Diocletian’s palace in Split nor at Romuliana. Also, the elements of decoration in the earlier architecture had been made of larger blocks, which were parts of the walls, and not as in Romuliana, where some archivolts were made on slabs just leaning to the wall. In the same way the connection of these archivolts and the stringcourse supporting them is an improvisation not based on the classic structural principles. The palace in northwest section of Romuliana, considering the size of rooms and the lavish interior decoration, fits into the Late Roman court architecture, whose most essential characteristic is monumentality and flamboyance. Considering the spatial structural entities, the residential complex at Romuliana has no parallels in formal approach, but only in adopted architectural program. There could be noticed only formal similarities between the design of its north section and the spatial disposition of most important rooms in the Flavian palace at Palatine. There is also the triclinium placed opposite aulae regiae and surrounded on both lateral sides by the atrium with the three-sided porch. Identical or similar disposition was employed in many Late Roman villas and imperial residences, including already mentioned Laufenbach palace. Considering the functional structures of the imperial palace, there is conspicuous resemblance between Palatine, Diocletian’s palace in Split and Romuliana.150 The permanent and temporary imperial residences have many common features also in decoration of the interiors of their ceremonial halls. The most essential characteristic is the abundance of flashy ornaments, lavish wall paintings, incrustation, stucco and mosaic decoration realized in the spirit of the so-called second Pompeian style. There are interior wall surfaces divided vertically and horizontally with stone or painted architectural features. In the lower section usualy the stone veneer imitated the orthostat, while the upper section was usually divided with stone and painted lesenes into the series of panels. These panels had the coffers executed in the incrustation technique of thin ornamental bands in the strict geometric pattern or these panels were painted in vivid colors with prevailing bright red, blue and intense yellow.151 Such was also the interior wall decoration in Galerius’ residential palace at Romuliana. The architecture of the sacred structures was dual in its character. The tetrastyle prostylos has the distinct ground plan with the crypt and corner reinforcements in the cella, but its elevation, which could have been rather reliably established, was based on the classic architectural forms. Also, its assumed proportions follow the forms recommended by Vitruvius for that type of temples. Second temple, which is of peripteros type (temple I), is distinctive and belongs to the small group of temples with crypts and his distinct elevation based on two types of free-standing supporters, and there were, by all appearances, also arches: one in the center of the façade and others on the side porches. This temple, according to its envisaged exterior, represents the transitional form, which is characteristic of the time of tetrarchy. In that period were created new architectural entities of specific structure and new structural designs were employed, that differed from the classic structural compositions known in the architecture of Diocletian’s palace in Split.152 Its composition was also influenced by the fact that its columns – brought from Proconnesus or from Pentelikon and columns of the serpentine breccia from Greece – were probably of standard size.153 Its use in one structural entity imposed new architectural designs and composition of the façades of Romuliana peripteros. But this temple fits into the courses of late antique architecture not only by its form, but also by the polychromy employed in its decoration. The most unusual architectural composition was employed on the cruciform sacred structure (E). Its ground plan and cruciform upper section relates it to tricliniums, so it could be assumed that because of its decoration and luxury it was used for some rituals. This structure together with other buildings of Romuliana represents the important transition period in the art of structural design, when new forms appear, new models and new visual forms are created. The evaluation of the imperial palace is possible on the basis of the analysis of its spatial structure and the comparison 150— Wulf-Rheidt 2007, 70, 71, Abb. 9, 10. 151— Gerke 1973, 9–49; Lavin 1967, 99–113; Reusch 1966, 187–235; Azevedo, de 1959, 3 sqq; Azevedo, de 1970, 223–259; Ward-Perkins 1974, 295–325; 152— Ward-Perkins 1974, 284, 312; Marasovi} 1982, 100–102. 153— Stojkovi}-Pavelka 1988/89, 144, 145; Ward-Perkins 1974, 284. 104 ARCHITECTURE AND SPATIAL STRUCTURE OF THE IMPERIAL PALACE with the renowned imperial residences from the Late Roman period. They include palaces in Trier, Milan, Thessalonica, Antioch and other centers, built in the final decade of the 3rd and the first decade of the 4th century. We have attempted to determine their spatial organization on the basis of the archaeological remains and written sources and then to establish their meaning and relation to the structure of the city they had been built in.154 It was not possible to draw precise conclusions about them, because they are insufficiently explored. It was possible to establish, on the basis of available data, only the basic characteristics of their spatial structure, including the ceremonial axial arrangement of most important spatial units from the entrance via peristyle to the hall, which had the important ritual purpose. This spatial entity was in urban imperial palaces connected to the hippodrome, and there was also the mausoleum within the fortress or in the palace vicinity.155 The best known imperial palace with such urban perspective and remarkable background is Diocletian’s palace in Split, although it had not been erected for official purposes. Because of that, it was the subject of many studies and basis for establishing the idea about the distinct type of the tetrarchic imperial palaces.156 This thesis has been disputed by Duval, who has analyzed in detail also the other palaces of the Roman emperors, including the Theodorich’s palace in Ravenna and Large palace in Constantinople and some other palaces of wealthy owners, and also the temporary imperial residences (Palace of the Dux in Apollonia, palace – villa in Laufenbach, Piazza Armerina, palace on Mljet). According to his analysis, Duval concludes that despite the fact that the imperial residences had been built in many centers, there is neither the standard plan for the imperial residences nor the emperors built the secondary residences, including the palace for tetrarchs, to live in them after abdication, modeled after the official palace from the time of their reign.157 Duval is right when he thinks that there was not the established type of the spatial structure of tetrarchic imperial palace. This is confirmed to the greatest extent by the spatial structure of Romuliana, that is not coherent and consists of many independent structural entities. The basis of entire interior composition is the main communication route and the location for two temples was determined in relation to that route and the palace. The larger temple was located on the best side, following the principles of antique urbanism. Around the temple are freely arranged the independent buildings, which have only one thing in common, and that is to fit into the orthogonal grid, determined by the axis of the main communication. Such spatial structure of Romuliana, divided into few independent entities, has no points of contact with the structure of Diocletian’s palace in Split, although it could be expected, because of the same architectural program. This disagreement is the consequence of the gradual construction of Romuliana, perceptible in the interrelation of the constructed entities and there is also evidence that architectural program for the whole complex was established gradually. This fact and the spatial relations between the peripteros and the neighboring structures (as well as their assumed purpose), encouraged Duval to compare Romuliana with some sites in Gaul, where in the rural centers were constructed the temple, some public buildings for theatrical performances, shops and economic structures called cinciliabula, where local rural population satisfied their cultural and economic needs. He assumed something analogous at Romuliana.158 This assumption was suggested after the discovery of the inscription FELIX ROMULIANA, corroborating that archaeological site near Gamzigrad is the imperial palace Romuliana, where Galerius intended to spend the rest of his life after the abdication in AD 312. So, this evidence opposes Duval’s assumption, besides the fact that the conclusion about the purpose of some buildings and the life organization at Romuliana is premature until all buildings around peripteros and the building with corridor (D4) are investigated and their purpose identified. It results from the comparison with the known imperial residences that the resemblance between the palaces, considering the spatial structure and the internal division, is less of formal nature, but it is discernible in the similar designs of some halls or spatial entities of the residential complex, while the greatest resemblance is conspicuous in the employed architectural program. It could be assumed that there was a distinct ideological program for building the tetrarchic palaces intended for the retired tetrarchs and which result from the system of tetrarchy, 154— ]ur~i} 1993, 67–90. 155— Duval 1965, 67–95; Duval 1987, 463–485, with complete earlier literature. 156— Strzygowski 1906, 325–335; Dyggve 1941, 3–55; Swoboda 1924; Swoboda 1961, 78–89; Azevedo, de 1959, 3 sqq. 157— Duval 1987, 489. 158— Duval 1987A, 82. 105 MILKA ^ANAK-MEDI], BRANA STOJKOVI]-PAVELKA whose creator was the emperor Diocletian. That system predicted the abdication of Augustus after twenty years of rule and, according to Diocletian’s decision, spending the rest of life in a specially built palace in his homeland. The same idea inspired Galerius’ decision to fortify again and to architecturally complete the interior of Romuliana.159 In both instances residences of the retired emperors were built as strong fortifications. This is the fact which is of greater importance for the opinion about the tetrarchic imperial palace than the division of its interior. The concept of sacrum palatium of that time was in fact inseparable from the image of the strong fortification, because it expressed the autocratic system of reign and the idea of everlasting state. It is confirmed that sacrum palatium in the Late Roman period really looked like that by Diocletian’s imperial palace in Antioch, Diocletian’s residence in Split and their identification as names turris and castrum in the written sources of that time. Later on, even in Theodosius’ and Justi- nian’s codices, representation of castrum had the meaning of divine imperial power.160 This concept of imperial palace in the late antique period is confirmed by one mosaic from Carthago, explained as the representation of (D)IV(INA) DOM(VS).161 In the mosaic the divine house is represented as a fortification with two towers at the corners and a large gallery in the upper section of the façade, and this was the external appearance of Diocletian’s palace in Split, and according to Libanius, also of the palace in Antioch,162 and the identical façade has the building represented on the ivory slab from Trier, assumed to represent the entrance to the imperial palace in Istanbul.163 Such façade became an archetypal model for the divine residence, and the fortification gates with towers flanking them were the architectural ideogram of the imperial residence.164 So, Romuliana, the palace of the emperor Galerius in Dacia Ripensis, was based on that visual representation and ideological concept. 159— Srejovi} 1985, 9–20. 160— Diepenbach 1921,36. 161— Smith 1956, 71 sq. 162— Swoboda 1961, 78–83. 163— Mango 1959, 100, Fig, 12. 164— Smith 1956, 38–50. 106 MAJA @IVI] ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE Color combinations of red, green and white are characteristic in fresco painting and architecture of the tetrarchic period as can clearly be noted in Gamzigrad. Tuffaceous sandstone, andesite, red brick and white lime were used for building ramparts and towers. Expensive stone in the aforementioned colors (red and green porphyry, white marble, green and red marble breccias, whitish and yellowish serpentine breccias and other combinations of stone) were extensively used in architectural decoration. The use of stones of specific colors can be interpreted through the symbolism of colors. Green symbolizes hope, strength but also the eternity, i.e. immortality. Both Diocletian and Galerius were first and foremost soldiers and conquerors, which is what led them to the imperial throne. This explains the lavish use of red in the buildings they erected, because red at the time symbolized warfare and conquests. It could be said that red and green also marked the course of their fate: their success in life is colored red, while green marks the pinnacle of their careers. The considerable use of red porphyry, which Romans called “purple stone” because of its color, elevated it during the tetrarchy to the position of imperial stone. The stone is of volcanic origin, of exceptional hardness, difficult to carve and very rare.1 Purple was also the color of death as many sarcophagi from that time bear witness, as do the columns in the Diocletian’s “mausoleum” of the palace in Split. Long after the tetrarchy, in Byzantium, emperors and nobility were using the purple stone.2 The ruling ideology of Galerius should be considered as the consistent implementation of tetrarchic media propaganda, the founder of which was his step father and father-in-law Diocletian. The very iconography of the decoration at Gamzigrad palace (Fig. 62), above all the mosaics, but also sculptures and architectural decoration, is the visual expression of that ideological-political concept of which Galerius was a faithful exponent. Architectural Decoration Architectural decoration in the imperial palace, temples and other representative structures in Romuliana was made of expensive stone. The most popular stone for making columns, architraves, Ionic and Corinthian capitals was 1 — There is just one deposit of this stone in Gebel Dokhan in the Nile delta. white marble from the Greek quarries. Along 2 — Jovi} 1998, 134–137. with the expensive stone obtained mostly 107 MAJA @IVI] FIGURE 62. Ariel view of the palace from Greece, local stone was also used for making various architectural elements. The wealth of ornaments on many architectural elements of Galerius’ Romuliana gives us a better understanding of the architectural decoration of imperial buildings from the time of tetrarchy in the Balkan provinces, considering that finds of that kind, either from the tetrarchy or from during the independent rule of Constantine, are rather infrequent. The only exceptions are the ornaments on Galerius’ principal palace in Thessalonica and Diocletian’s palace in Split, while only small fragments of the stone decoration of the imperial buildings in Sirmium, Naissus and Mediana have been preserved. The stone decoration of the buildings of the Gamzigrad complex is characteristic and in many aspects differs from the decoration of the buildings which Diocletian, Constantine and Licinius built in Rome, but at the same time also from the architectural decoration of Diocletian’s palace in Split. It could be said that only the capitals carved of highest quality marble from Greece (from Mount Pentelicus) and from Asia Minor and the wall veneer of superior stone (white marble and green porphyrite) are almost identical to those in Diocletian’s palace in Split and those in Villa del Cassale – Piazza Armerina in Sicily. If we analyze the architectural elements with carved decoration from Romuliana, we could conclude that one segment of that decoration is similar to the decoration of Galerius’ triumphal arch in Thessalonica, particularly the decoration of the so-called “column B” (Figs. 188, 189), while the remaining decoration is unique and we cannot find any analogies for it. The buildings of the Galerius palace complex at Gamzigrad were decorated with geometric, floral and figural motifs. The geometric decoration was of very simple design. This is the decoration we encounter on door lintels, door posts, window frames and parapet panels. Floral ornamentation executed in naturalistic or stylized manner was used on many of the architectural elements. Geometric ornamentation was executed to a simple scheme. Such ornamentation can be found on transoms, door jambs, window frames and parapets. Naturalistic or stylized floral moldings were applied on a wealth of architectural elements. Archivolts and cornices were always decorated with stylized floral motifs, including palmettes, rosettes, egg-and-dart motifs and acanthus leaves. These orna- 108 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE ments are common in the Roman art. In contrast to the archivolts and cornices, consoles, pilasters and string courses were mostly decorated with naturalistic representations of vines, mostly in combination with kantharoi or craters and sometimes with Sileni picking graves. Such iconography is indubitably related to the cult of Dionysus, the deity to whom Galerius paid special attention for ideological reasons. This decoration could be related to the decoration of the so-called “small triumphal arch” in Galerius’ palace in his capital Thessalonica and was most probably modeled after that arch. The authors of that decoration are most probably from Thessalonica itself or from Attica. It is most logical to assume that artisans from Thessalonica itself decorated most of the architectural elements in Romuliana, as well as in Galerius’ palace in Sirmium. It should be noted that identical motifs appear in the decoration carved into superior stone as well as in ornamentation made of local stone. Also, the technique of execution of architectural elements of local stone does not fall behind the technique applied on the precious stone. This speaks in favour of the claim that all elements of architectural decoration were made by artisans from the same workshop, most probably from Thessalonica or Attica. The best illustration of this claim is the aforementioned pilaster with elements of the Dionysus’ cult with representation of Sileni picking grapes, discovered in the course of investigation in the area of the east gate of a later Romuliana fortification (Fig. 63). Sileni picking grapes and two more figures, one of which is carrying a ram on his shoulder, are depicted on the frontal side of this pilaster. Sileni picking grapes without doubt directly suggest the cult of Dionysus. The ram could also be associated with Dionysus through the deities connected with it. We know that this animal, through the Egyptian Amon, is the symbol of syncretized deity Jupiter–Amon, but also of Attis as a weaker Jupiter. Also, the ram is connected with Hermes, who is in certain mysteries represented as Kriophoros (ram-bearer). All these deities are related to Dionysus in one way or another. The vine spirals over the entire frontal side of the pilaster, while bunches of grapes are depicted adjacent to the naked Sileni. The lateral sides of the pilaster are decorated with double-grooved lesenes. Pilaster with Sileni picking grapes FIGURE 64. a, b) Pilaster with Victoria FIGURE 63. a 109 b MAJA @IVI] FIGURE 65. Capital with head of Silenus and female head could be Apollo and thus the relationship with Dionysus would be established once more, which would entirely satisfy the ideological concept of Galerius’ palace. Visible on the frontal side of the pilaster are fragments of two standing figures above the figure of Victoria with spread wings and a laurel wreath in her right hand. Perhaps the rulers – Augusti and Caesars from the time of the so-called second tetrarchy (either all four of them or possibly only Diocletian and Galerius) – were represented on the frontal side of this pilaster. Victoria with the laurel wreath in her hand is a direct allusion to Galerius’ great triumph over the Persians, celebrated in Rome in AD 303 and glorified by the porphyritic statue from Romuliana. If we reconsider the iconography of this pilaster, this time as a whole, the assumption of Apollo as the represented deity seems once again plausible. Namely, if we place Apollo in the context of the observer of various events, as it is mostly the case in classical period of the Greek art and according to which were also made the sculptures in Romuliana, in concrete circumstances – observer of the celebration of Galerius’ triumph – his position on this Gamzigrad monument seems well-grounded. In addition to the decoration of the main gate façade that has been located with certainty, we would also like to mention two figural capitals discovered in the northwest tract of Romuliana, whose original position within the Gamzigrad The figural decoration is generally connected to the façade of the east (main) gate of Romuliana. The pilaster with representations of the tetrarchs in medallions, as well as the small pilaster with Victoria, are specific as they illustrate the ruling ideology of the tetrarchs and perhaps also the personal cult of Galerius and the apotheosis of Galerius and his mother Romula. This is an elaborate ornament, which must have been devised in some of the sculptural workshops in Thessalonica, but it was executed on the spot by the same artisans who executed ornaments on the other buildings in Romuliana. Only the central segment decorated on the frontal and lateral sides is preserved from the pilaster with Victoria (Fig. 64). The ornament is preserved just on one lateral side, where the palm tree is represented at the bottom, while an altar and some kind of aedicule showing a dressed figure on the throne is carved in the central section. It is assumed that this figure represents Jupiter. We will, however, suggest that, based on the analysis of some iconographic elements, the represented figure could also be Apollo. In other words, besides the basic symbolism of the palm or palm branch indicating victory, rise, rebirth and immortality,3 with the addition of the role of the Aeneas’ golden bough, as well as the one used in the Eleusinian mysteries, the palm is also associated with Apollo. According to the legend, on Delos Leto first gave birth to Artemis and then with her help also to Apollo, while embracing the palm tree.4 For that reason, the palm tree is one of the plants dedicated to the divine twins. If we take into account the connection of Apollo and Dionysus in the Delphic myth, it is not unrealistic to assume that the figure depicted on this Gamzigrad pilaster 3 — Chevalier–Gheerbrant 1983, 474. 4 — Srejovi}, Cermanovi} 1979, 38. 110 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE interpreted as Romula’s villa or Romula’s house. The epithet felix indicates that villa or the house mentioned in the name is in fact a metaphor. This epithet added to the proper names or the names of places, i.e. of the structures, was not used as an epithet of embellishment, but to denote gods, emperors, empresses and the members of their families, regions, cities and buildings connected with them and as the symbol of their divine nature, sacredness, glory, fertility and progress. Where the Gamzigrad archivolt is concerned, the word felix is added to the name of the place where the new Romulus, i.e. Galerius, was born and which is dedicated to his mother Romula, whose name was also taken from the legend, is direct association to the persons related to the foundation of Rome and its mythical history. For that reason the name Felix Romuliana should be understood as Roma nova or Roma secunda, i.e. a name suggesting a sacred place, an eternal and celestial city. The relief decoration of the archivolt with the carved inscription Felix Romuliana confirms that the first word of the inscription has a religious, i.e. an entirely ideological meaning, that it expresses the consecration of the person and the place, and indicates something eternal and sacred. If we analyze the decorative elements surrounding the inscription Felix Romuliana, the laurel wreath flanked with the peacocks and ivy leaves, we come to conclusion that they are all palace complex has not been established, as is unfortunately the case with most of the other architectural elements. A female head was carved next to the head of Silenus (Fig. 65) on one of the capitals. The face of the woman is round, surrounded by the wavy hair with a middle parting into which an ivy wreath is braided. The eyes are big, expressive with prominent pupils and modeled in the manner of “hard” style. The mouth is small and nose short. If this is the portrait of Galerius’ mother Romula, it would be logical to assume that the column with such capital must have been an element of some rather important building within the palace complex. We distinguish as the most important architectural element the archivolt with the inscription Felix Romuliana (Fig. 66a), mainly because it represents the key to the Gamzigrad conundrum. The fragmented archivolt made of tuffaceous sandstone was discovered in 1984 in the southwest tract of Romuliana, in the structure E (the so-called Romula’s triclinium). The inscription field is circular, surrounded by the laurel wreath flanked with the peacocks. The inscription Felix Romuliana is carved within the wreath. Above the inscription, an ivy leaf each is carved in the middle and to the left and right of the word Felix (Fig. 66b). As the emperor Galerius dedicated his memorial structure to his mother Romula, the name Romuliana could be literally FIGURE 66. Archivolt with inscription: a) photo; b) drawing a b 111 MAJA @IVI] 0 5 10 m a b FIGURE 67. a, b) Archivolt with tetrarchs also discernible fragment of one of two peacocks flanking the laurel wreath. Another fragment of the archivolt with representations of another two rulers and another peacock is missing. The laurel wreath is bound with ribbon with winding ends and in this detail it resembles an archivolt with Christogram from Sirmium.6 The tetrarchs are represented under the arches of a building, most probably a tetrapylon. The archaeological investigations carried out at Gamzigrad, following the discovery of the archivolt with the inscription Felix Romuliana, confirmed assumption about the purpose of Romuliana and shed light on the details of Galerius’ political, ideological and religious concept on which its construction had been based. Between 1985 and 1989, the area in front of the large temple (plan XXVI–XXXI) and the external section of the east gate of symbols of consecration and apotheosis, i.e. symbols of immortality and eternity. We think that it is not necessary to give a lot of examples where the same decorative elements, individually or together, symbolize just what the epithet felix denotes when added to the name of place built with the idea to glorify and make eternal one charismatic ruler and his mother. Therefore, the iconography and contents of the inscription Felix Romuliana do not relate literally to the Romula’s villa or house, but indicate the sacred structure built by the new Romulus for the eternal memory of his mother and himself, i.e. the sacred place intended for the immortals and gods.5 In 2007, a fragment of architectural sculpture was discovered by chance among the stones brought from the area in front of the east gate and which were earmarked for conservation. It was a fragment of archivolt made of tuffaceous sandstone smaller than the archivolt bearing inscription Felix Romuliana and with the representation of a tetrarch (Fig. 67). A small fragment of this monument was preserved, showing one of the rulers in armor and chlamys and the outline of another ruler. There is 5 — Srejovi} 1993 A, 38–39; Srejovi} 1993 B, 38–39; @ivi} 2005, 260, 89; @ivi} 2007, Kat. Nr. I. 5. 11. 6 — Jeremi} 1993, 196, kat. 34; Jeremi}, 1993, 196, cat. 34. 112 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE FIGURE 68. East gate when Diocletian adopted Galerius and bestowed on him the title of Caesar, Galerius became a member of Jupiter’s divine family, son of the supreme God. The colossal marble statue of Jupiter, which most probably stood in the large temple, located almost in the centre of Romuliana, did not, therefore, represent only the supreme God, but also the founder of tetrarchy, who gave the members of his family, the tetrarchs, power over the Empire. Thus the tetrarchs, as they were given power by the will of the supreme God, i.e. his earthly incarnation, were predestined to rule. The façade of the main entrance to Romuliana, the east gate of the later fortification, was decorated with pilasters on which relief decoration most convincingly bears witness to the fact that the entire Romuliana had been constructed to glorify the tetrarchs, members of the Jupiter’s family and their system of ruling the world. On the entirely preserved pilaster, in the medallions on the military standard (signum) all members of Jupiter’s family are depicted in pairs, in strict hierarchical order: the person represented on the left, honorary side, is higher on the earlier and later fortifications (Fig. 68) were investigated in the east section of the Gamzigrad complex. In the area in front of large temple fragments of a colossal white marble statue representing Jupiter with an eagle in his extended right hand have been found (Fig. 69). A large number of architectural elements with relief representations were found in the ruins of the façade of the east gate of the later fortification. These representations and the colossal statue of the supreme deity are a fraction of elements of a complex visual concept based on ideological, political and religious grounds, which by using symbols, allusions and mythological scenes, described the origin, life course and apotheosis of the emperor who was born and buried at Romuliana. It is understandable that Jupiter was given a central place within the concept as the system of tetrarchy, of which Galerius was one of the most significant exponents and most consistent followers, as best illustrated by the construction of Romuliana, is realized according to the will of the supreme Roman deity. After Diocletian, the creator of tetrarchy, proclaimed himself the earthly incarnation of Jupiter, in AD 293, 113 MAJA @IVI] FIGURE 69. Head of Jupiter, white marble FIGURE 71. FIGURE 70. Parapet panel with eagle 114 Pilaster with tetrarchs ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE the hierarchical ladder than the one depicted on the right (Fig. 70). In the medallion at the base of the standard both portrayed persons are simply dressed in a cloak and toga, while the persons depicted in the central medallion and in the medallion at the top of the standard are clothed in a paludamentum fastened with valuable clasp on the right shoulder. According to these iconographic characteristics it can be established with certainty that the persons depicted in the medallions are tetrarchs, represented in the following order: in the medallion at the base of the standard are the Augusti, who renounced the throne in AD 305 (seniors Augusti) – Diocletian (left) and Maximian Herculius (right). In two remaining medallions are actual Augusti and Caesars – Galerius (left) and Maximinus Daia (right) and Constantius Chlorus (left) and Severus II (right). The pilaster described above has on its front side not only the first reliably dated representation of the tetrarchs, but also the first complete visual representation of the tetrarchic hierarchy. In iconography that order is depicted in such a way that the senior Augusti (seniors Augusti), who are at the top of the hierarchy and who voluntarily renounced the throne, are given the most important position. They are followed by their adoptive sons – the invincible Augusti (invicti Augusti) and at the end the adoptive sons of their adoptive sons – the noblest Caesars (nobilissimi Caesares). The complete composition, as well as its details, clearly illustrate the idea on which the tetrarchy is based: the greatest harmony (concordia) and the closest relationship between the rulers, the uniformity of their images and destinies, and the four stages of their rule, i.e. the gradual rise of each of them from Caesar through Augustus to senior Augustus and God. The idea that the undertakings and fate of one member of Jupiter’s ruling family are shared by all the others equally, an idea emphasized on all monuments from the first tetrarchy, is even more strongly suggested on the portal of the main Romuliana gate. The decoration on the pilaster shows not only the rulers of the so-called second tetrarchy as equal participants in the construction of Galerius’ magnificent memorial, but also the founders of the tetrarchy – Diocletian and Maximian Herculius. The members of the “Gamzigrad sextet” are portrayed at the same age of life; their features are identical, as are their hairstyle and beards. It seems at first glance that this is a sixfold representation of one person, probably that of Galerius himself.7 The portal of the main Romuliana gate, besides representations directly associated with tetrarchy and its rulers, was also decorated with representations indicating the imperial apotheosis and the afterlife, i.e. the eternal life. The relief decoration on the archivolts, pilasters, consoles and parapet panels includes vine foliage, ivy leaves, laurel branches, the depiction of grape harvesting and an eagle. The fragmented parapet panel depicting the eagle carrying a floral wreath in his beak (Fig. 71) bears distinct symbolism. In Rome, the eagle was essentially the messenger of the god’s will, primarily the will of the supreme god, Zeus, i.e. Jupiter, and was sometimes identified with him. The eagle is also the symbol of a ruler, i.e. the symbol of the Roman Empire. As the eagle is an imperial bird, he is the celestial equivalent to the lion on earth. As a symbol of the highest power and supreme authority, of genius, heroism and every transcendent state, the eagle is often depicted on top of columns and obelisks, which are considered a substitute for the omphalos, i.e. the axis, or navel of the world. The belief in Greece was that eagles flew from the end of the world and stopped on the top of omphalos at Delphi; from dawn till dusk they follow the path of the sun as omphalos corresponds to the axis of the world.8 The eagle also plays a significant part in divination. It is well-known that oracles interpreted the flight of eagles in order to understand the will of the gods. Pindar says that eagle, King of birds, sleeps on the scepter of Zeus and informs the people of his will.9 The parapet panel associated with the façade of the main gate of the later Romuliana fortification, of whose decoration just the segment representing eagle with the wreath in its beak is preserved, certainly alludes to the triumph of the divine Galerius and his crowning with glory anticipating immortality. As all other architectural ornaments discovered in the course of investigations of the area of the east gate of the later fortification, this one is also directly related to the glorification and deification of the ruler and acclamation of the ruling system he consistently carried out. Among other ornaments from the portal of the main Gamzigrad gate rather important are representations of the laurel wreath flanked with peacocks, which indicate the deification 7 — Srejovi} 1993 A, 42; Srejovi} 1993 B, 42; @ivi} 2007, Kat. Nr. I. 4. 12. 8 — Chevalier–Gheerbrant 1983, 459. 9 — Chevalier–Gheerbrant 1983, 460. 115 MAJA @IVI] Head of Hercules, white marble FIGURE 73. Torso of Asclepius, white marble FIGURE 72. of the member of the imperial family. The apotheosis is also represented by the deities portrayed in sculptures and mosaics discovered in the imperial palace and in the area surrounding the large temple. These include Dionysus, Hercules (Fig. 72) and Asclepius (Fig. 73). The connection between these deities and the creator of Romuliana is more than obvious. The mother of Dionysus, Hercules and Asclepius, as well as of Galerius, was a mortal, while their father was supreme god. They all have a soteriological function, i.e. they are all saviors of humanity, who after great deeds performed on the earth were included among the gods. Galerius paid special attention to the cult of Dionysus, thus the entire Romuliana is in the spirit of this deity. There were many reasons for that. His triumph over the Persian king Narseus in AD 298 Galerius could compare only with Dionysus triumphal campaign in India. Also, the decoration of Galerius’ palace in the capital Thessalonica, where Dionysus had special place, confirm that since then the myth of Dionysus was used by Galerius to create his own myth. Archaeological excavations conducted at Gamzigrad in the 1990s speak in favor of the thesis that Galerius shaped the relationship with 116 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE FIGURE 74. Relief with Ariadne, white marble the emperor’s favorite deity and inspiration for the creation of his own myth, had a leading role. When Theseus, who promised to marry Ariadne after she helped him out of the labyrinth, left Minos’ daughter on the island of Naxos, Dionysus found a sleeping girl and took her with him to Mount Olympus.10 There he married her; it was a sacred marriage (hieros gamos), i.e. a mystical wedding. At the wedding Dionysus gave Ariadne a diadem made by Hephaestus, which in honor of their love was transferred among the stars as the constellation Corona borealis. As Ariadne was in Hellenistic time the symbol of human soul,11 Dionysus not only saved the soul from death but, united with her in the mystical marriage. his mother Romula taking Dionysus as model, because, according to the myth, after his triumph in the east Dionysus deified his mother Semele and included her among the gods under the name of Tiona. We would like to mention here one wall relief with a depiction of Sleeping Ariadne on a rectangular block of white marble (Fig. 74). Ariadne was modeled in bas relief and only partially as free standing sculpture. The relief was discovered in the vicinity of the thermae in the southeast tract of Romuliana; it was broken into three fragments and found in secondary position in the building rubble of the 6th century structure (most probably the basilica). Because of that we could not conclude with certainty in which of the buildings this relief was originally housed, but we could assume that it embellished one of the niches of Galerius’ thermae in the immediate vicinity. The back side of the marble block is rounded and very roughly worked, suggesting that the block was placed in a deep niche. There is no doubt that the relief representation of Ariadne was in harmony with the entire concept of Romuliana, where, as is the case with Galerius’ palace in Thessalonica, Dionysus, Funerary Monuments Even in the first years of investigations at Gamzigrad three tombstones were discovered, two of them in the area in front of 10 — According to another legend when Theseus left her in the island of Naxos she hung herself. 11 — Elijade 1991, 125. 117 MAJA @IVI] FIGURE 75. Tomb south of Romuliana ramparts 118 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE an arch resting on two short pilasters with Corinthian capitals. Between the pilasters is a composition depicting a horseman (to the left) and three more human figures (to the right). The horseman is dressed in short tunic and with mantle (chlamys) over his shoulders, waving in the wind. In his right hand he carries a double-headed axe (labrys) over his shoulder. A man and a woman are sitting in front of the horseman and between them is a child. This iconography once again points to the Thracian horseman, i.e. to the Dionysus himself. the west gate of the later fortification. Unfortunately, we could not tie these tombstones to any funerary architecture, i.e. any masonry tombs to which they could have belonged and which could have been expected within the section of the necropolis to the southwest of the west gate and to the south of the south wall of the later Romuliana fortification. During the 2005 campaign carried out, as part of the Serbian and German excavation project, to the south of the west gate of the later Romuliana fortification, a tomb (Fig. 75) with the Military equipment from the tomb south of Romuliana ramparts FIGURE 77. Gold fibula from the tomb south of Romuliana ramparts FIGURE 76. The second fragment, a molded rectangular inscription panel, is surrounded by vines growing out of the kantharos under the inscription panel. An inscription consisting of thirteen (13) lines (Fig. 79b) was skillfully carved into the panel.13 The finials of meticulously carved letters are accentuated, while the lines are separated by double leading lines. There are no ligatures or dividing marks. The inscription has been read and published complete military equipment (Fig. 76) and gold cruciform fibula (Fig. 77) was discovered.12 On the basis of the findings inside the graves, that burial took place at the very end of the 3rd century or the beginning of the 4th century. The deceased was no doubt a high-ranking officer, most probably belonging to Galerius’ personal guard. The integral parts of this tomb is formed of two limestone ashlars with an opening in the centre (Fig. 78), which could have been the bases for the funerary steles. We will now return to the very beginning of the investigations, to year 1953, when two tombstones made of white limestone were discovered in the area in front of the west gate. The first monument was found fragmented into two pieces (Fig. 79a) – the tympanum and the inscription panel. Judging by the discovered fragment, the tympanum was in the shape of 12 — Investigations as part of international cooperation with Römisch– Germanische Komission des Deutchen Archäologischen Instituts, Frankfurt/ Main, conducted between 2004 and 2007, included investigation of the area outside imperial palace walls by geophysical methods (geomagnetic and geoelectrical measurements) and test trenching. 13 — Srejovi} 1983, 165. 119 MAJA @IVI] FIGURE 78. Stone block with hollow (stela postament ?), from the tomb by P. Petrovic,14 while Professor M. Mirkovi} 15 suggested that instead of dignissimi viri, devotissimi viri could be read. Another monument discovered in the immediate vicinity of the one described above, is as interesting as the previous specimen, primarily because of its iconography. A fragment of the tympanum of this monument is preserved, while the inscription panel is missing. The tympanum is of semicircular shape and flanked by reclining lions with their paws resting on the head of a horned animal, most probably an ibex. A horseman with cap on the head and dressed in short tunic and chlamys waving in the wind is depicted on the tympanum. The chlamys is fastened on the right shoulder by the cruciform fibula. In his right hand over his shoulder the horseman is carrying a double-headed axe (labrys). A legionary with a lance in his right hand and shield in the left is depicted behind the horseman. The iconographic details, the clothes and particularly the cap of the horseman – identical to the caps worn by the tetrarchs, and the cruciform fibula – indicate that on this monument Galerius himself is depicted in the role of the Thracian hero, i.e. Dionysus. The last monument, made of tuffaceous sandstone, abounds in iconographic details. The front side of this fragmented tombstone is divided into two segments: the tympanum and the in- scription panel. The tympanum is also divided into two superposed zones. Two human figures are depicted on the lower preserved section of the tympanum: a woman with the wreath in the right hand and a large jug in the left hand, and a man with a torch in his right hand; a tripod with offerings (fruit, cakes) to the deity is depicted between them. An altar (ara), protruding in the lower section and recessed in the top section, can be discerned between the man and the tripod. The upper part of the tympanum is considerably damaged. The remains of the base of one of the two pilasters framing the scene can barely be made out on its right side. Also the lower section (feet and part of the drape) of one of the three depicted figures is preserved, while the feet of the other two figures are barely discernible. The figure depicted next to the pilaster was most probably seated, while the other two are standing. The molded rectangular panel prepared for the carving of inscription is framed by vines growing out of the kantharos placed in the lower zone, beneath the inscription panel. 14 — Petrovi}, 1975, 142–143, T. XV. 15 — Srejovi} 1983, 165. 120 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE b a FIGURE 79. Tombstone found near the tomb: a) attempted reconstruction by joining tympanum and inscription panel; b) drawing of inscription panel two horsemen. On reliefs of this type the woman is usually depicted feeding the horses from a scarf or giving them water from a vessel. On the Gamzigrad relief the hands of the woman cannot be seen, as they are completely hidden by the figures of the horses. Only the head and lower part of the woman’s body are clearly visible. The woman is dressed in a long, belted chiton. One of two horsemen is preserved and he is dressed in tight trousers and with chlamys over the shoulders, waving in the wind. His right hand is raised. On the basis of the similar reliefs we assume that another horseman was depicted in the same manner. The essence of the iconographic content of this votive relief is the Thracian horseman (hero), who represents the basis of the cult of the so-called Danubian horseman. If we take a look behind the scenography of this cult, we could see that at its basis is Dionysus, i.e. that there is an evident parallel between this pre-Hellenic deity and the Thracian hero. According to the preserved iconographic elements, as there is no inscription in the panel to give us the necessary information, this monument could relate to the cult of Apollo, i.e. the person for whom the monument was intended could be identified as a legionary originating from the east provinces of the Empire or some of the members of his immediate family originated from those parts. In addition to the funerary monuments we would like to mention one fragmented votive panel, because of its iconographic content, which is directly connected with the complete decoration of Galerius’ palace at Gamzigrad and which is the illustration of his ruling ideology. The panel of white marble was discovered in 1979 in the area in front of the large temple and with a depiction of the so-called Danubian horsemen on the panel. The composition is divided into two superposed friezes. The top frieze is wider and in its centre is a depiction of a female figure flanked by 121 MAJA @IVI] probably a sculptural group depicting Galerius and Diocletian. The porphyrite head of emperor Galerius found in 1993, during archaeological excavations next to the east façade of Galerius’ thermae, can be dated from AD 303 onward, as in November of that year Galerius had been crowned in Rome with the laurel wreath, because of his great victory over the Persians in AD 298,16 thus unofficially becoming the leading person in the Empire.17 The portrait of Galerius from Romuliana shows stylistic differences in relation to the stiff style, characteristic of the portraits from the first years of the tetrarchy, i.e. of the so-called first tetrarchy.18 In comparison with this statue of Galerius, the portraits from Ni{ and Tekija are even more removed from that style and show stylistically greater similarity with the statues of diarchs on Constantine’s triumphal arch in Rome. If we understand Galerius’ head and left hand with globe as fragments of one sculpture, the iconography of that image is entirely clear. Galerius is represented as Pantocrator – the ruler of the entire ecumene. Such a representation of Galerius is in accordance with the entire iconography of Romuliana and represents a step forward in relation to the iconography from Galerius’ capital Thessalonica. Thus, on the Sculpture Many fragments of sculptures carved from superior stone have been discovered at Gamzigrad. Most numerous are sculptures of white marble of various size of grain. One group of white marbles of the highest quality most probably originated from the Greek island of Paros. Other white marble used for the Gamzigrad sculptures most probably also came from Greece (Mount Pentelikos). A smaller number of sculptures was made of red porphyrite (porfido rosso antico), known as “imperial stone”, originating from Egypt (Gebel Dokhan). We could classify all the sculptures and their fragments discovered during the investigations at Romuliana in two groups, according to their subject matter. The first group includes portraits and the second sculptures of deities and heroes. a. Portraits In the course of the investigations at Gamzigrad carried out so far, fragments of two statues of emperors made of red porphyrite have been discovered. These are the head of the sculpture of Emperor Galerius and the left hand holding the globe (Fig. 80), and fragment of the neck of another emperor. This was most b a FIGURE 80. Head of Galerius (a, b) and hand with globe (c), red porphyry 122 c ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE triumphal arch in Thessalonica Galerius is represented as Pacator orbis and Pacator gentium, i.e. the representative of the ecumenical sovereignty of the tetrarchy, while in the iconography of Romuliana he is Pater orbis and Restitutor orbis, i.e. the cosmocrator and restitutor of the world. Although the soft oval shape of the face in the portrait of Galerius from Romuliana marks a shift from the “stiff” style of the so-called first tetrarchy, in the modeling of eyes and forehead it is entirely similar to it, merging into the deliberate impersonality of tetrarchic portraits. This complete impersonality of the portraits of tetrarchs was intended to emphasize their perfect similarity (similitudo) and concord (concordia). The explanation for the ample use of porphyry during the tetrarchy is not to be found only in the hardness of the stone as a guarantee of its longevity and a symbol of the eternal life of the tetrarchy, but also in the symbolism of the purple color. The iconography of the tetrarchy, expressed in the use of superior material of rich color, a deliberate move away from individuality and intentional stylization, which emphasized the super human, is basically of eastern character, and the Byzantine in its emergence followed from it. The porphyry and the symbolism of its color will become interwoven in the fabric of the Byzantium and in a manner of speaking become its trademark via Constantine’s capital, new Rome. Purple belonged essentially to the emperor: only basileus sat on a purple throne, wore purple boots and wrote his signature in purple ink. In the ecclesiastical cult the Gospels on the altar are of purple color, and a series of manuscripts on purple parchment was produced in workshops (scriptoria) in the capital in the 6th century. Considering the Byzantine symbolism of colors, the purple united eternal, celestial, transcendental (blue, dove-gray) with the terrestrial (red). Uniting in itself the opposites, the warm and the cold spectrum of colors, the active and the passive, the purple color acquired a particular meaning in the culture of antique, i.e. Byzantine thought. The culture of sight, however, was not ultimately determined by the culture of thought. Therefore, it is no accident that in the Byzantine art the purple color was practically not used beyond the capital. It was substituted by more simple colors – red and blue.19 of sculptures depicting, besides Galerius, certainly one, if not three more rulers of the tetrarchy, have been found. These fragments include fragments of wings, most probably those of goddess Victoria crowning Galerius (Fig. 81), as well as a foot and an arm also belonging to this goddess (Fig. 82). Other porphyritic fragments (Fig. 82) could have been fragments of small sculptures, most probably those of deities, which most closely resemble the sculptures from Mediana.20 Several fragments of porphyry were carved as ears of wheat and we can assume that they were parts of the sculpture (sculptures) of Liber or Libera, thus leading us back to Dionysus, the favorite god of Galerius. All other Gamzigrad sculptures were carved of white marble of top quality, originating from Greece (Pentelic or Paros marble) and from Asia Minor. What characterizes these sculptures is their stylistic uniformity, on the basis of which they could be dated to the period between AD 300 and AD 310. Exceptional skill in modelling and their captivating beauty point to superior craftsmanship – sculptors who emulated sculptures of the Greek sculptors from the 5th and 4th centuries BC, i.e. late Hellenistic copies of these masterpieces of classical Greek art (2nd–1st centuries BC), as well as their copies from the time of Hadrian. The sculptures of Romuliana were undoubtedly carved 16 — This year was crucial in Galerius’ reign. Two very important events for Galerius took place: triumph over the Persians and celebration of five years of his rule (quinquennalia). In the same year started construction of his triumphal arch in Thessalonica (that was certainly completed before great jubilee of the tetrarchs in Rome in November 303), and his mother was associated to the god Mars. By all appearances that was the time when building of the palace at Romuliana started. It concerns construction of earlier fortification and some structures within it. About symbolic meaning of insignia on Galerius’ head see Popovi} I. 2008, 105–119. 17 — In November 303 in Rome great jubilee of the tetrarchs was celebrated: vicennalia of Augusti and decennalia of Caesars. Diocletian proclaimed Galerius Augustus in spring of 305 and also Constantine on the 1st of May of the same year, when also Severus II and Galerius’ nephew Maximin Daia were proclaimed Caesars. At that time Galerius was considered the absolute ruler of the Empire. After the death of Constantius Chlorus in 306 Galerius officially became first Augustus. 18 — More on this find Srejovi} 1993 A, 232; Srejovi} 1993 B, 232; Srejovi} 1993 D, 4–8; Srejovi} 1993 E, 64–65; Srejovi} 1994 A, 41–47; Srejovi} 1994 B, 143–152; Srejovi} 1995 A, 14; Srejovi} 1995 B, 300; Srejovi} 1995 C; Srejovi} 1996, 20–29; Srejovi} 1998, 318–339, sl. 60; @ivi} 2005, 204–205, cat. 2; @ivi} 2007, Kat. Nr. I. 5. 12. b. Sculptures of Gods and Heroes All other sculptures which decorated the Romuliana buildings belong to the other group. Several dozen fragments of porphyritic sculptures, some of which could be elements of a group 19 — Bi~kov 1991, 124. 20 — Jovanovi}, 1975, 57–65. 123 MAJA @IVI] in workshops which relied on classic Greek models, but their authors at the same time intended to emphasize the importance and character of the portrayed images, using distinct artistic expression. This is best illustrated by the remains of two sculptures discovered near the so-called large temple, which dominates Romuliana. These are the head of Jupiter from the colossal statue of the supreme god and the head of Hercules from the statue of the greatest Greek and Roman hero. It was no accident that both heads were found in the area in front of the large temple, since Galerius as the incarnation of Hercules and Diocletian as the terrestrial incarnation of Jupiter were the very people to whom this temple with double cella was dedicated. Owing to its colossal size, the head of the supreme god bears witness to the supernatural power and divine character of the tetrarchic rule. At the same time, however, the Gamzigrad representation of Jupiter emanates substantial gentleness, which indicates departure from the classic manner in the representation of a supreme deity. This departure could also be related to the idea of tetrarchy: the ruler is god on earth, an embodiment of intransient cosmic powers and a pledge to eternity. He is a gentle and fair ruler, guaranteeing the safety of his subjects. The portrait of Hercules, work of exceptional beauty, was also not carved in the classical manner: the face of the greatest Greek hero is asymmetrical, the surfaces of the forehead and cheeks are not smooth, the curls of the hair and beard are emphasized by deeply etched lines, thus achieving a marvelous contrast of light and shade and lending the face the outstanding expressiveness. Judging by the position of the head, which is slightly inclined to the left, it is very probable that Hercules had been depicted in the moments of relaxation, resting his left arm on the rock covered with lion’s skin (Fig. 84a), or on his club. It is also possible, considering the given proportions and the quality and color of the marble used, that the right hand, holding the apples (Fig. 84b), was also part of this very sculpture. A fragment Fragments of the wings of Victoria (?), red porphyry FIGURE 82. Foot and arm of Victoria (?), red porphyry FIGURE 81. 124 Fragments of red porphyry, sculpture fragments FIGURE 83. ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE a b Sculpture of Hercules: a) postament with lion’s skin; b) hand with apples FIGURE 84. manner of carving, which illustrates the intensity of inner life, the Gamzigrad Hercules should without doubt be taken to relate to the models from the classical Greek art, i.e. those from the 4th century BC. The portraits carved in this way appear in Roman sculpture around the year 300. Among the sculptures of deities carved in white marble it is worth mentioning another few recognizable statues and some fragments which we think are important. These are portraits of Asclepius, Athena, three male torsos (one most probably represents Dionysus, the second torso could have been from the statue of Apollo, while the third might be the torso of Satyr), the head of a boy, most probably depicting Dionysus as a child (Fig. 85), an arm and a hand of the torchbearer and one exceptionally lifelike hand belonging to a colossal statue, perhaps of Galerius himself (Fig. 86). It is also important to mention a torso, one of the two almost identical pieces, representing the emperor in armor. The torso (Fig. 87) belonged to an oversized sculpture depicting the emperor, most probably Galerius, in full military garb. Fragments of leather armor are clearly discernible. This statue could perhaps be related to his triumph over the Persians, so it could be dated after AD 298. A fragment of the right upper arm also belongs to this torso (Fig. 88a). A of the sculpture of Hercules – the feet and the rock covered with lion’s skin, on the marble base, may have been part of the sculpture to which the head of Hercules also belongs, considering the position of the body and the iconography. The fragmented right hand with apples would therefore be part of the same statue of Hercules depicted with the apples of Hesperides, as illustration of one of his labors. The model for this statue may have been the copy of the statue of Hercules made by Lysippos for Sikyon, i.e. the famous Heracles Farnese from the Archaeological Museum in Naples. Also similar to the Gamzigrad head of Hercules is the head of Hercules at the Metropolitan Museum, which also copies the Hercules Farnese type.21 The head of Hercules from Sucidava, although greatly damaged, is almost identical to Hercules from Romuliana by the way of carving the beard, lips and wrinkles on the forehead and around the mouth.22 Considering the 21 — Lippold 1950, T. 101/1. 22 — Bordenache 1969, No. 126. 125 MAJA @IVI] FIGURE 85. Child Dionysus, white marble ribbon or a scarf is tied in a bow around the arm, under the armpit, on the inside of the arm. Also visible are the remains of a spiral (armila, or possibly a shield) to which most probably an oak or juniper leaf are attached (Fig. 89b). The oversized torso of the emperor was most probably that of Galerius himself, in military garb, while another almost identical torso was perhaps that of Diocletian. The statue of Asclepius was modeled on the Greek sculptures from the 4th century BC, which had often been copied in Roman times.23 The fragmented statue of Athena was most probably a copy of the Greek original from the 5th century BC. The first of the three male torsos could be that of Dionysus, or more likely that of Apollo depicted as leaning on a column, a palm tree, which entirely corresponds to this sculpture. Professor Srejovi} assumed that this was the torso of Hercules, and that it was modeled on the statue of Hercules made by Scopas for Sikion.24 The second torso most probably represents Apollo, Hand of colossal statue, white marble Torso of emperor in armor, white marble FIGURE 86. FIGURE 87. and by its posture (the weight is shifted to the right leg while the left one is slightly bent at the knee, the youth most probably has raised his right arm while the left one is resting on a support) it is very similar to the images of that deity from the mid 5th 23 — Balty 1969, T. III; Jovanovi} 1975, 57. 24 — Srejovi} 1983, 78, kat. 9. 126 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE b FIGURE 88. Hand, part of the torso of emperor in armor, white marble, a) complete; b) detail FIGURE 89. a Female hand holding torch, white marble Big hand holding torch, white marble FIGURE 90. plicas had been made during the Empire.25 The head of the boy was most probably part of a group of sculptures depicting a deity with a child in his hands, most probably Hermes with century BC. The same posture was often used for the statues of Apollo in the 4th century BC, but much more supple and flexible. The third torso is a fragment of a sculpture depicting a young man resting his lower arm on a column or a tree. The model for this sculpture could have been Praxiteles’ sculpture “The Resting Satyr”, made around 350 BC, of which many re- 25 — Srejovi} 1983, 78, kat. 8. 127 MAJA @IVI] Dionysus as a child, or Hercules with Telephos.26 The fragmented arm of a woman with a torch in her hand (Fig. 89) was a fragment of the sculpture which was most probably part of a large sculptural group. The group of sculptures was most probably a representation of the cult as it has been discovered in front of the large temple, as was the hand of the torchbearer. A large hand of the right arm, fragment of the colossal statue, was also found in front of the large temple. It may belong to Jupiter, or to the colossal statue of Hercules, i.e. Galerius as his earthly incarnation (Fig. 90). The artists who sculptured the statues for the Galerius’ palace near Gamzigrad and the artisans who carved the architectural decoration and decorated its floors with beautiful mosaics are most probably the same ones who built and decorated Galerius’ palace in Thessalonica. There, in the capital, under the influence of Attica workshops, a style influenced by the Greek art of the Classical period had been cultivated. Therefore, it seems most plausible that the authors of the sculptures in Romuliana were actually artists from Thessalonica, who had arrived there from Attica or Asia Minor. The purpose of the sculptures was not just sheer decoration of the Gamzigrad palace, as it is suggested by the repertoire of the depicted deities. Their purpose was primarily to glorify the idea of the tetrarchy, i.e. the divine Galerius as its most consistent supporter. Mosaic All the mosaics discovered so far at Gamzigrad are floor mosaics. That there were wall mosaics as well as mosaics in the vaults of certain buildings is established by a large quantity of mosaic cubes made of silver and gold glass paste discovered in different sections of Palace I, as well as in the zone of Galerius’ thermae.27 The appearance of the mosaic carpets is entirely classical: the floor surface bordered by the walls is divided into independent borders, most probably of concentric type, that frame and accentuate the field with figural compositions.28 The panels – emblems of the mosaic carpets in Palace I, which depict Dionysus at a banquet, hunting scenes and labyrinth – reveal all the characteristics of the mosaic art of Late Hellenistic period, creating the impression of pictures axially inserted into the floor. If we consider the architecture of Galerius’ palace at Gamzigrad, we come to the conclusion that it also relies on the examples from that time and that its builders and mosaic makers were brought in from the same artistic centre. 128 The geometric patterns on the Gamzigrad mosaics were also created after earlier, traditional models. On the floor mosaics with geometric patterns the surface was divided in several different ways. Most often used was the system of diagonal octagons with alternating large octagons and small squares creating trapezoid fields along the edges and triangular fields in the corners. This system appears in two styles: with simple squares (mosaic in hall G of palace I) and with meander designs instead of squares (mosaic in hall A in Palace I). The variant with simple squares had often been used on 2nd and 3rd century mosaic floors in Italy and the northern provinces. The other variant was more widely accepted only from the reign of Septimius Severus and mostly in Africa, although there are some examples from Pannonia and Greece.29 The north tract of Romuliana is denoted as tract D1 (plan V), i.e. as Palace I. It is in fact the private i.e. residential quarter of Galerius’ complex at Gamzigrad. It covers an area of around 3,260 square meters, of which 1,586 square meters are covered with mosaics. Most of the rooms in this section of the palace are arranged orthogonally, complying with cardinal points. The south section of tract D1, i.e. its centre, consists of five rooms: D1 – 1 (hall A in the earlier general plan of the palace), D1 – 2 (hall B), D1 – 3 (hall C), D1 – 4 (hall D) and D1 – 5 (hall E). The anteroom – vestibulum of the palace – is marked in the plan as room D1 – 1 and its dimensions are 42.7 x 7.5 meters. The area in front of the entrance was carefully paved with different kinds of marble (Fig. 91). The entire floor surface is covered in mosaic consisting of two lateral and one central mosaic carpet. Lateral carpets were created by diagonally placed octagons, filled with diverse patterns (Fig. 92). Even seemingly identical motifs are different in color (Fig. 93). The central mosaic carpet is somewhat narrower than the lateral ones, but it is much more 26 — Srejovi} 1983, 78, kat. 18. 27 — During the excavations of the apse of Galerius’ thermae in 2008, besides the remains of devastated floor mosaic there were also found many glass paste tesserae of cobalt blue and gold color, indicating that this room had ceiling mosaic, which most probably represented firmament with stars. 28 — Srejovi} 1983, 69; cf. Lavin 1963, 185. 29 — Salies 1974, 10 sqq; Thomas 1964, T. 114. 30 — Mosaic carpet depicting Labyrinth was removed and transferred to the National Museum in Zaje~ar in 1993. It is on permanent display in the exhibition Felix Romuliana – Gamzigrad. ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE colorful. It consists of a series of square and rectangular panels, some of which consist of swastikas in various combinations and varying degree of elaboration, starting from east to west (Fig. 94). In the centre of the central mosaic carpet was an FIGURE 91. almost square panel with a border (2.50 x 2.30 m before removal and 1.80 x 1.80 m after removal) with a representation of the labyrinth (Fig. 95).30 The central mosaic carpet leads to the west, to the wide door in front of which are three Palace entrance hall 129 MAJA @IVI] FIGURE 92. Side mosaic carpets stone steps as an entrance to the transversal hall (D1 – 3, in the earlier general plan – hall C), which is 7.59 meters wide and 30.5 meters long. Its floor of 231.8 square meters is entirely covered in mosaic. The mosaic carpet is decorated with a design of orthogonal radial rhombs. The composition consists of intersecting bands; at the points where they intersect, square fields are created, i.e. stars consisting of eight rhombs with sides equal to half of a square. Such decorative designs were often used in Italy as early as the middle of the 1st century and they were popular in Greece from the 2nd to the 5th century.31 At Gamzigrad, however, we can notice certain differences regarding the complexity of ornaments in comparison with Galerius’ palace in Thessalonica, which is considered as the closest analogy for both the geometric and the figural mosaics of Romuliana.32 At the far end of this room (D1 – 3, i.e. C), looking north, there is a wide door in the east wall, with a staircase descending to the big ceremonial hall (D1 – 4, in the earlier general palace plan – hall D). The hall is 36.06 m long and 10.93 m wide. The entire floor surface (394 square meters) is covered in mosaic. A long mosaic carpet extends from the wide entrance in the west to the semicircular apse in the east, consisting of eleven (11) panels with hunting scenes. The panels are framed with meander in perspective and lateral carpets of more modest execution. The lateral mosaic carpets were executed in a rather simple design of mosaic decoration: paratactic circles of identical diameter are arranged in an orthogonal system; cruciform flowers resembling patterns common in the opus sectile technique are created at the intersection of the circles. Mosaic floors decorated in this pattern are not frequent and could be connected to the workshops in the East, i.e. Galerius’ palace in Thessalonica.33 The mosaic panels in hall D1 – 4 (hall D) draw attention to the elevated niche, most probably intended for the marble throne. In the southeast corner of this room is a staircase leading to an octagonal room (D1 – 5, in the earlier general palace plan – hall E) with walls covered with green porphyrite veneer. Two large mosaic panels – Venatores (Fig. 96) and Leaping Leopard (Fig. 97) faced the door with the staircase.34 This means that these two scenes were completely visible from the staircase. 31 — Salies 1974, 6 sqq; Waywell 1979, 307. 32 — Kolarik 2006, 164–167. 33 — Kolarik 2006, 166–167. 34 — Both panels were removed and transferred to the National Museum in Zaje~ar, where they are on permanent display in the exhibition Felix Romuliana – Gamzigrad. 130 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE a b a–f) Details of side mosaic carpets FIGURE 93. c d e f a b c e f a–f) Central mosaic carpet FIGURE 94. d 131 MAJA @IVI] a b Mosaic depicting labyrinth, central mosaic carpet, in situ (a), after removal (b) FIGURE 95. Mosaic with representation of venator FIGURE 97. Mosaic with representation of leaping leopard FIGURE 96. 132 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE rectangular section of the room and 39.30 square meters in its apsidal section. A mosaic with the representation of Dionysus was discovered next to the threshold at the entrance (Fig. 99). The floor in the central section of the room (a rectangle 7.5 x 3.4 m in size) was made of variegated marble tiles in the opus sectile technique, arranged in twenty-one (21) square panels. This surface was bordered by two mosaic carpets. One was very narrow, with figural representations of which only the figure of a dog in a landscape (Fig. 100) is preserved, while the other was wider and decorated with diagonally arranged octagons. The walls of the triclinium were covered with white marble in the lower zones, while they were decorated with stucco ornaments and veneer of multicolored stone in the upper zones. A large atrium (D1 – 8, in the earlier general palace plan – hall H) was situated to the east of the triclinium (D1 – 7) and to the north of the peristyle with the fountain (D1 – 6). It is 18.40 m long and 14.10 m wide, covering 259.50 square meters in total. The court has a colonnade creating a portico along the east, south and north sides. Only the bases of the columns and the shafts of two large columns emphasizing the door in the east wall are still preserved. The open area of the atrium is paved with floor tiles, while the area under the portico is covered with mosaics. The pattern of the mosaic is very simple, consisting of diagonally arranged intersecting bands. Here the squares created at the intersections of bands are rather large and there are smaller squares in the areas between the intersecting bands. The mosaic floors made according to this pattern can also be related to Galerius’ palace in Thessalonica.35 The atrium with the colonnade is connected to a circular room in the east in a similar way as the peristyle with the fountain is connected to the triclinium. The circular room (D1 – 9, in the earlier general palace plan – hall N), whose decoration is emphasized by tall columns, is an anteroom to a quatrefoil room (D1 – 10, in the earlier general palace plan – hall R) and a trefoil shaped room (D1 – 11, in the earlier general palace plan – hall D), which are most probably stibadia. The diameter of the anteroom is 5.32 m and its entire 22.20 square-meter floor was covered with mosaic. The mosaic floor consists of an illusionistic design resembling a shield (Fig. 101), exceptionally popular in Greece and the west Balkan provinces in the 2nd The north section of tract D1 includes ten rooms in total ( D1 – 6 to D1 –15, in the earlier general plan of the palace – halls F, G, H, N, P, O, L, M, Q and R). The peristyle (D1 – 6, i.e. hall F) covers an area of 356.5 square meters in total, of which 112 square meters are beneath the porticos and covered in mosaics. There is a fountain in the central section of the peristyle, opening to the sky and paved with marble (Fig. 98). The floor of the portico was decorated with mosaic of simple, but multicolored geometric patterns. A large room with apse (hall D1 – 7, in the earlier general palace plan – hall G) – the triclinium – was most probably the audience hall. This room is 18.50 m long and 11.00 m wide. The mosaic floor covers an area of 206.60 square meters in the FIGURE 98. Atrium with fountain 35 — Kolarik 2006, 165–166. 133 MAJA @IVI] FIGURE 99. Mosaic with representation of Dionysus FIGURE 100. Mosaic with representation of dog and 3rd centuries.36 Such an uninterrupted spiral pattern can also be taken to relate to the cult of Dionysus, i.e. with continuous renewal of life. The floor of the room on quatrefoil plan (D1 – 10, i.e. P) was paved with precious stone in the opus sectile technique. The elaborate system of channels used for heating the entire room and its conchs in particular, were discovered under the floor. The room with the trefoil plan (D1 – 11, i.e. O) also had a floor paved with marble tiles, but their layout was much simpler. No heating installations were discovered under the floor of this room. In the east, a small room with an apse (D1 – 15, i.e. rooms Q and R) was also included within the section including halls D1 – 9, 10 and 11 (i.e. halls N, P and O). This hall has partially preserved mosaic flooring with figural and geometric motifs. The mosaic floor, as well as the heating installations underneath, bears witness to the importance of that room. The thermae in the southeast section of the Gamzigrad settlement have been intermittently investigated since 1982. A section of a mosaic carpet of very high quality of execution has been discovered in 1993, during archaeological investiga- tions in the room identified as the apodyterium I of the Galerius thermae. It is a small polychromatic fragment of simple ornamental design: fretwork and geometric designs – triangles, squares and rhombs. Fretworks and squares are arranged in various ways within larger rhombs, while between them are square panels, i.e. stars created by the touching bases of isosceles triangles. The fretwork also forms the border and above the section of the border on one side is a partially preserved representation of the tympanum of a building, possibly a temple (Fig. 102). Among the mosaics discovered in other structures, worth mentioning are those in the rooms with a cruciform base (structure E), better known as “Romula’s triclinium” (plan XXXIII).37 The building is situated in the southwest tract of Romuliana (plan XXXII) and its distinctive plan could indicate its special purpose in relation to the other structures within the walls of Romuliana. Hence also its name “Romula’s triclinium”, 36 — Waywell,1979, 304–305; Fiala–Patsch 1895, Bd. III, 275. 37 — More on this structure: Srejovi} 1985, 51–67; Vasi} ^. 1997, 46–47. 134 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE FIGURE 101. 0 Mosaic with shield representation: in situ (a), drawing (b) 1m gular floor plan (E 6 and E 7). The floor of room E 7 was paved with a geometric mosaic in the opus tessellatum technique (Fig. 103), while the floor in room E 6 has not been explored. Another room opened to the east from the central room (room E 3) with a rectangular floor plan and with floors paved with mosaic in the opus tessellatum technique. All the aforementioned mosaics in Galerius’ palace clearly indicate that Gamzigrad mosaic artists used traditional patterns, mostly those popular in the time of Hadrian and Antoninus Pius, for creating mosaics with geometric motifs. If, however, we consider the mosaics from the imperial palace of the Gamzigrad complex as a whole, we could say that by their composition, color and style, they mostly correspond to the mosaics from the 2nd and 1st centuries BC, e.g. to those from the island of Delos.38 Therefore, we may conclude that they denote the final rebirth of the mosaic art of the Hellenistic time. The manufacturing technique of the Gamzigrad mosaics complies with very high standards. The base layer of the mosaics regularly consisted of three and rarely of four layers of plaster, of which the surface layer has the finest texture. The mosaic as it is assumed that this structure took up a central position within the complex intended for Galerius’ mother. It is assumed that it was used for the purposes of the cult, perhaps to venerate one of the “mountain deities” (perhaps the goddess Cybele and in that case it was possible to communicate with the temple in the north tract of Romuliana through the single entrance in the north side). This structure, consisting of seven rooms, was enclosed by a separate wall and was located in the centre of the square courtyard. The only entrance to the building was in the middle of the north courtyard wall and it led to the rectangular vestibule (room 1). The floor of the vestibule was covered with mosaics of geometric design, executed in the opus tessellatum technique, partially preserved in the northeast and northwest corners of the room. The vestibule was connected to the east and west to symmetrically arranged rooms on a rectangular base (E 4 and E 5), and in the south to the main room with a square base. (E 2). The floors of the rectangular rooms were paved with mosaics which were better preserved than the ones in the vestibule, which were also of simple – geometric decorative design, executed in the opus tessellatum technique. The floor of the central room was paved with rectangular marble slabs in the opus sectile technique. In the east and west, the central room was connected to two other rooms with a rectan- 38 — Bruneau 1972. 135 MAJA @IVI] FIGURE 102. cubes (tesserae) of various stones,39 ceramic and glass paste, ranging in size from 2 mm to 1.5 cm, were laid into that layer. The mosaics with geometric motifs and monochromatic background of figural compositions were mostly made of larger cubes (8 mm to 1.5 cm) in the opus tessellatum technique. Images of people and animals (particularly the details of their heads, dress and fur), foliage on trees and bushes, but also some geometric patterns – for example the meander in the perspective of the hall D 1 – 4 (hall D) in palace I – were executed in the opus vermiculatum technique. These details were created with considerably smaller cubes (2 mm to 8 mm), often individually shaped to cover even the smallest space. Beside the first-class manufacturing technique, the Gamzigrad mosaics are also characterized by the multitude of colors used. Eight colors in twenty-nine shades were used in their manufacture. The prevailing hues are red and gray as well as white. The use of white differed for figural and geometric compositions. In figural compositions the background is light, created mostly with white, but also with light gray and light pink cubes. On the other hand, white was rarely used for the mosaics with geometric patterns. Instead of white, pastel shades of green, brown and gray were used. Glass was also often used, 136 Mosaic from thermae in building E (“Romula’s triclinium”) FIGURE 103. Mosaic in room 7 i.e. large quantities of glass paste tesserae would be inlaid. Thus, certain details of clothing were brilliant as were the borders of figural compositions (e.g. the meander in the perspective of hall D). In contrast to this brilliance, shadows were usually soft and outer contours discreet, which is why the Gamzigrad mosaics exude elegant restraint both in the form and in choice of color. The floor mosaics in Galerius’ palace at Gamzigrad probably represented one thematic unity, i.e. their role was not only decorative, but also narrative, with ideological background. Unfortunately, the story could only be partially reconstructed, because it remains uncertain what had been depicted in the panels in front of the apse of hall G (triclinium) and in the apse of hall R (a small structure with apse). However, three figural motifs are preserved, and walking through the palace one first encounters the labyrinth (hall D1 – 1, i.e. hall A), then the hunting scenes (hall D1 – 4, i.e. hall D), of which the scene 39 — For the floor mosaics in Galerius’ palace at Gamzigrad were used cubes of white, gray, pink and reddish marble. That marble, as well as the marble used for sculptures in Gamzigrad palace, mostly originates from the Greek quarries. ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE chaos, which like wild animals linger at the edges of the organized world. For the king it is a test of worthiness, continuous confirmation of his eternal youth.45 So, the hunt has a magical and religious meaning. The Romans took over the hunt of Alexander the Great46 and other divinized heroes as a model from Greek tradition.47 Dionysus himself, Galerius’ favorite deity, in one of his many roles is also the god of the dead, a wild hunter (like the Thracian horseman, Thracian hero), capturer and guide of souls.48 Despite the fact that the hunt of Dionysus Zagreos, the great hunter, symbolizes his insatiable desire for pleasure, at the core of Dionysian myth considered as a whole lies the aspiration to the spiritualization of the living being, from plants all the way to ecstasy. A mosaic depicting Dionysus, completely executed as a painting, differs from any known representation of that deity in the world. This is an eternally young god – a mortal who is continually reborn and captivates by his beauty; the perfect harmony of composition and broad palette of colors and their nuances, which perfectly portray nature and its lushness. More than in any other representations, the Gamzigrad mosaic with his vividness and sensuality: Dionysus – Bakhos – Iakhos,49 vines, with the venatores is the most prominent, and finally Dionysus at a feast (hall D1 – 7, i.e. hall G). Although seemingly different, the three motifs are essentially closely related. The distance between them has been created for ideological reasons. It suggests continuous movement, i.e. gradual approach to a hidden destination, only accessible to the chosen ones. The Mosaic with labyrinth40 was the middle panel of the central mosaic carpet in the vestibule of the palace (hall D1 – 1, i.e. hall A).41 The labyrinth is one of the best illustrations of Galerius’ ruling concept, based on the myth of Dionysus as the myth of the divine emperor himself. The interpretation of the labyrinth, which is by itself a poly-semantic symbol, could be even more complex: essentially the representation of a city, which is the sublimation of the world, i.e. the cosmos; it could be understood as the picture of the Universe – the mandala.42 The labyrinth also alludes to Theseus escaping from Minotaur with Ariadne’s help – i.e. to escaping death by finding the exit from the labyrinth. The wall relief depicting the sleeping Ariadne is just one segment of that story that together with the mosaic representations of the labyrinth and imperial hunters (venatores) visually represents the myth of Dionysus himself, i.e. of Galerius. Meaningful symbols are placed around the depiction of the labyrinth: a crater and pelta. The crater is a direct allusion to Dionysus, while the pelta could be associated with Hercules and Theseus and even Aeneas (Amazonians, Trojans), i.e. his son Romulus and the founding of the city of Rome, i.e. of the Roman state. All the elements of the mosaic representation of the labyrinth point to the central figure of the message suggested to us and it is Emperor Galerius, i.e. new (second) Romulus. The mosaic with the representation of the labyrinth could be, therefore, understood as an illustration of Galerius’ triumph over the Persian king Narseus in AD 298, which shaped his future fate. The Mosaic depicting imperial hunters – (venatores)43 also alludes to the divine nature of the emperor. The hunting scenes in late Roman funerary art symbolize agon, the heavenly contest. They are to suggest the virtue and bravery (virtus) of the emperor himself. Hence, hunting scenes in the late Roman art symbolize a spiritual challenge, attaining virtue, i.e. imperial power and sublimation. The spiritual quest is one of two aspects of the symbolism of hunt (the first, the killing of animals, symbolizes in fact the destruction of ignorance and evil tendencies).44 Back in ancient Egypt the hunt denoted the expansion of divine creation: hunting shifts the borders of 40 — @ivi} 2005, 261, cat. 90; @ivi} 2007, Kat. Nr. I. 5. 13. 41 — Panel with representation of labyrinth is housed in National Museum in Zaje~ar as exhibit at permanent display in the exhibition Felix Romuliana – Gamzigrad. 42 — Chevalier–Gheerbrant 1983, 385. 43 — @ivi} 2005, 300, cat. 151. 44 — Chevalier–Gheerbrant 1983, 357. 45 — Chevalier–Gheerbrant 1983, 357. 46 — Age of Spirituality 1979, 90. 47 — Age of Spirituality 1979 , 83. 48 — He liberated his mother Semele hit by Zeus’ lightning from the underworld and under the name Tiona, he took her to the Mt. Olympus among the gods. Because he took his mother (Earth) to the Olympus, Dionysus was considered to be liberator from the underworld, who wanted to bestow immortality to all children of the Earth. Here we should look for the origins of his soteriological function. 49 — Under this name Aristophanes describes underworld Dionysus, who directs the dances of the initiates and of the dead in the underworld fields of the kingdom of Hades. Iakhos is the god personification of joyful cry: festive cries echoed from the processions accompanying holy secrets from Athens to Eleusis along the sacred way. In chariot at the head of procession was the statue of god Iakhos taken for that occasion from the temple of Demeter and immediately after followed the priest of Iakhos (Iakhogogos). Young god, personification of joyful cry, represented Eleisinian incarnation of Dionysus. 137 MAJA @IVI] wine, ivy,50 bull, male goat51 – in all of his many forms he continues the symbolism of the permanent renewal of the nature, the continuous cycle of passing and rebirth and the eternal creation of life from the primordial chaos. The Gamzigrad depiction of Dionysus is the visual representation of this god’s permanent aspiration to bring humans into the world of gods after making them immortal. Dionysus is the savior of souls and the one who bestows eternal life. The act of saving Ariadne, his future wife and his mother Semele is an allegory of the salvation of the human soul. By Dionysus’ gift and love Ariadne and Semele became figures of salvation.52 The epithet swth,r (swtwr) or plousiodo,thj (ploutodwth,r, ploutodwtw/n), i.e. ploutodw,teira, refers at the same time to Dionysus, Asclepius, Hermes, Zeus, Apollo, Hades and Demeter. This is the autonomous epithet of the god who donates – qeo.j swth, r. The holy marriage (hieros gamos) to Ariadne and her introduction among the gods on Mount Olympus alludes to the moment of blurring the borders between the “this” and “other” – worldly, between “life” and “afterlife”, i.e. it offers hope in the existence of the moment when the soul escapes death and attains the right to eternal life and permanent rebirth, and crosses over to the realm of immortality. Dionysus’ wish to remove the borders between the two worlds, the world of gods and the world of humans, in a simplified form represents him as the god who destroys all taboos and all borders.53 In essence, this striving for liberation has a far deeper meaning, which is why the Dionysian cult is at the core of Greek spiritualism in the way it sees the soul and contributes to its release and revelation. It is from the Dionysian rites that the idea of the soul related to the divine and the soul more material than the body was born.54 Galerius’ magical foundation at Felix Romuliana was created for the eternal memory of the divine emperor (ad memoriam aeternam Divi Imperatoris Galerii Maximini). The leading idea of Galerius’ grand architectural undertaking in his homeland was the idea of Galerius as the new Dionysus. And through Dionysus the parallel Galerius as the new Alexander and new Romulus was established. Belief in the triumph of Dionysus was the belief in the cyclical rebirth of the world, i.e. in the return to the Golden Age (Saturn as the divine sower – sator,55 was the first to give people food and was the first ruler of the world). His age was considered the Golden Age for humans. Diocletian as the founder of the tetrarchy, aside from Jupiter was also identified with Saturn, thus emphasizing his gentleness and impartiality for his subjects. Galerius’ inspiration for showing exceptional respect to Dionysus was at least partially due to his intention to deify his mother Romula and himself. Galerius intended his mother to have almost all divine powers of his time, so that she was identified with Cybele (Magna Mater), Demeter, Tyche and Fortuna, i.e. her form Fortuna Redux. The mosaic representation of Dionysus and the wall relief depicting a sleeping Ariadne symbolize the idea of death – and resurrection, that is, they indicate the two acts of the apotheosis whose impressive material evidence was discovered at Magura hill around one kilometer in a straight line from the main (east) gate of Romuliana. Despite the fact that the very act of apotheosis undoubtedly also had a political connotation, the apotheosis of Galerius and even more so of Romula reflected first of all personal religious faith. Like Dionysus and his mother Semele who joined the gods at the Mount Olympus after Dionysus’ triumphal expedition to India, Galerius, the new Dionysus, and his mother Romula ascended to heaven from the top of Magura hill. It is certainly also a question of whether Galerius’ adoration of Dionysus and 50 — Ivy is the favorite plant of Dionysus: it crowned the head of Dionysus and covered tirsos – staff crowned with pine cone (pine cone is the symbol not only of Dionysus, but also of Asclepius. It is like top one of the toys of Child Dionysus and same as top it symbolizes whirlpool or spiral, i.e. great generative powers). Ivy is except to Dionysus also dedicated to Attis, whose cult has many resemblances to the cult of Dionysus. 51 — Bull and male goat are zoomorphic manifestations of Dionysus. In the shape of these animals he discovers using his divine will the fountains of wine, honey and milk. There are data that Dionysus also appears in the shape of doe. Dionysus’ sacred animals are panther and lion and sacrificial animals are male and female goats. 52 — Jeanmarie 1951, 345. 53 — Main characteristic of Dionysus’ cult is orgiasm, falling into ecstasy (Dionysian purification brings to culmination something the soul should be liberated from), killing and dismembering of the child (sparagmos) – as Titans tore apart the horned child Dionysus – and eating of raw meat (omofagia), and drinking wine. This distinct kind of Communion is the main ritual of Dionysian mysteries: dismembering of live bull or kid (animals dedicated to Dionysus) – eating raw meat of sacrificial animal = eating the god; drinking wine the god of which is Dionysus = drinking god’s blood. Dedicated participants of mysteries (mystes), eating the body of god and drinking his blood, united with him – became equal to him and this is guarantee for their salvation after death. As parts of dismembered child come together by miracle and Dionysus is reborn over and over again – thus his followers will enter the spheres of immortality. Basic idea of Dionysus’ mysteries was: AS THE SEEDS ARE REBORN, THUS LIFE OF THE DEAD SHOULD BE RESTORED! 54 — Séchan, Lévêque 1966, 300. 55 — Ma{kin 1978, 85. 138 ARTISTIC ACHIEVEMENTS IN THE IMPERIAL PALACE tion included those deities whose mythical story is closely connected with Dionysus, be it that there are certain parallels in the cult, or that they exist in a distinct syncretistic form. After analyzing the fragments of sculptures, we identified a rather limited range of figures portrayed: Jupiter, Hercules, Dionysus, Apollo, Satyr, Athena and Asclepius. With the exception of Athena, whose presence is almost obligatory in the sculptural repertoire of the entire antique period, all the other depicted deities once again point to Dionysus/Galerius, i.e. to the ideological concepts of the tetrarchy as alpha and omega. Jupiter is without doubt present at Gamzigrad not only as the supreme god, but also as supreme god in a figurative sense, i.e. as supreme ruler. He is none other than the founder of the tetrarchy, Diocletian, earthly incarnation of Jupiter. Hercules is certainly the founder of the Herculii dynasty and his bravery makes him a worthy match for a ruler such as Galerius. There is no need to discuss Dionysus again, while Apollo, Satyr and Asclepius owe their presence in the rooms of the Gamzigrad palace primarily to him. The mosaic carpet on the floors of Galerius’ foundation, first of all the panels with figural representations and among them particularly representations of Dionysus at the feast, venatores and the labyrinth support the same idea in a distinctive way. Like the architectural elements and sculptures, but in a much more vivid manner, they are telling the story of the divine Galerius and represent the most powerful segment of the oversized and overly detailed scenography, created to show the cosmic origin of an ideology. Together with the unique and utterly uneconomically constructed defense system, all the analyzed decorative elements of Galerius’ Gamzigrad palace, and quite apart from the indisputable beauty of some of them, are striking because of their overemphasized character. And it is just this exaggerated insistence on power, this almost suffocating impression of omnipresent and restricting power, which leads us to the conclusion that even if his biological time had not come to an end, Galerius’ “ideological” time was inexorably on its way out. of other members of the classical pantheon who shared with him a semi-divine origin, including the deified Alexander the Great, also formed part of his political program. Alexander the Great, more than dei filius, was a symbol of the perfect warrior and the perfect prince for many Hellenistic kings, Republican potentates and emperors before, as well as after Galerius. Romulus as rex and dux is a similar image in the ideology of the Empire. The same could also be said for Dionysus, Bacchus and Liber from the aspect of both dux and rex, although his fame as the first triumphator threatened to overshadow his contribution to the ideal of the universal monarchy (Dionysus was also praised as freedom bearer; the Greeks portrayed Dionysus, whom they also called “the Legislator”, with two stone tablets inscribed with laws; Dionysus is also considered as the inventor of diadem). The political aspect of Dionysus will become particularly apparent when he and Alexander the Great are to become twin images in the Hellenistic and Roman times.56 The mosaic representation of the god Dionysus denotes the sublimation of the entire architectural and decorative concept of Romuliana: the divinized human enters the sphere of the immortals. Dionysus, a divine ephebe, basically mortal, mysteriously achieved victory over his own death, thus making his immortality the guarantee of the immortality of human soul. Dionysus, god of multiple forms, illusions and miracles, was also the inspiration for the creation of the image of Christ.57 An analysis of the decoration of architectural elements and several reliefs of votive and funerary character led to the conclusion that the applied iconographic scheme in most cases points directly to Dionysus. It is first of all established by the depiction of a vine with leaves and bunches of grapes on many architectural elements, as well as the figures from Greek and Roman mythology that are directly connected with the cult of Dionysus. Also, considering a number of architectural elements such as tombstones and votive monuments where the iconography points to another deity, we came to the conclusion that the choice of these deities was not at all accidental. This selec- 56 — Du{ani} 1995 A, 79. 57 — The divine sacrifice, best represented in Dionysus image, is the higher form of conceptual abstraction and introduction of symbolic sacrifice for common good. This will be particularly characteristic of Christ’s sacrifice in Christianity. Dionysus is related to the Christian god who dies and is resurrected also by zoomorphic symbols of these deities: fish and lamb as Christ’s symbols and kid and bull as those of Dionysus. 139 IVANA POPOVI] SACRED-FUNERARY COMPLEX AT MAGURA Magura is high, elongated ridge consisting of marl, sandstone and volcanic rocks, around 1,000 meters far from the east main gate of Romuliana (plan XLIV). Already the first investigators of Gamzigrad noticed at this hill the traces of Roman architecture, which were assumed to be the remains of towers and watchtowers because of the prominent position of Magura in comparison with the surrounding terrain. Nevertheless, local inhabitants thought of these ruins as sacred place and until recently they used to come to Magura and lit candles there on the days dedicated to the memory of the dead. Archaeological investigations conducted in the period between 1989 and 1993 revealed that Magura was since prehistoric Position of monuments at Magura in relation to the imperial palace: 1) tetrapylon; 2) building with central courtyard; 3) mausoleum 1; 4) consecration memorial 1; 5) mausoleum 2; 6) consecration memorial 2 PLAN XLIV 0 141 100 200 300 400 500 m IVANA POPOVI] FIGURE 104. Sacred complex at Magura before excavations, view through the east gate of Romuliana FIGURE 105. Arial view of structures at the top of Magura before excavations FIGURE 106. Mausoleum 2 and consecration memorial 2 (in the foreground) and mausoleum 1 and consecration memorial 1 (in the background) before excavations times the place where respect had been paid to the dead and where also a necropolis existed already by the end of 2nd – beginning of the 1st millennium BC.1 In the course of five years of archaeological investigations at Magura five structures were completely explored and published,2 while archaeological excavations have not been completed on one structure and it is still unpublished. At rather small plateau on the top of Magura, covering approximately 7,000 square meters, in the axis of the east Romuliana gate, two mausolea and also two tumuli (Fig. 104) that were approached through the 1 — Srejovi}, Vasi} 1994, 65–67. 2 — Srejovi} 1993 A, 45–52; Srejovi} 1993 B, 45–52; Srejovi}, Vasi} 1994; Srejovi}, Vasi} 1994 A, 127–141; @ivi} 2003, 37–41; Vasi} M. 2007, 46–50. 142 SACRED-FUNERARY COMPLEX AT MAGURA FIGURE 107. FIGURE 108. Bases of the pillars of tetrapylon Ideal reconstruction of tetrapylon façade tetrapylon built on hill slope some 230 meters to the northeast from the structures on the hill, have been investigated. Although demolished and devastated many years ago, these sacred monuments are partially still visible (Fig. 105), and the silhouettes of two giant tumuli are clearly discernible on the horizon even from a great distance. The very location of Magura and the structures on its top (Fig. 106) suggest the special significance of these monuments and their connection with the palace. 0 5 10 m the pillar crowns was the upper section of the tetrapylon façade with one niche each, left and right of the arch. The gateway ended in pyramidal roof around 3 meters tall (Fig. 108), meaning that the total height of the tetrapylon was 13.32 m or 45 Roman feet.3 As it was primarily the mark of cross-roads, the tetrapylon was not decorated with reliefs of program or ideological character. However, considering its intermediary role between the palace and sacred complex, this tetrapylon replicates the concept used in Thessalonica, where the road from palace to the circular building, rotunda (Fig. 189), which is assumed to have been the mausoleum of this emperor, was leading through the Galerius’ triumphal arch.4 Tetrapylon (1) The monuments on the top of Magura that give this area the character of distinct sacred hill (mons sacer) are reached by passing through the tetrapylon, of which bases of four pillars, 3.60 x 3.60 m in size and made of stone blocks are preserved (Fig. 107). These pillars created a gate, in the center of which was the intersection of two roads, one running northwest–southeast, towards the top of Magura, and the other running northeast– southwest and leading to the east gate of Romuliana. So, the tetrapylon was the cross-roads, the place where visitor was directed towards the monuments on the hill top or towards the main palace gate. The analysis of the pillars building technique, as well as the structural designs of tetrapylons in other parts of the Empire, suggested the conclusion that the gate façade was 10.65 m, i.e. 36 Roman feet high. All four pillars ended in the crowns from which at the height of around 6 meters from the ground sprung stone arch linking two pillars each, so the gate was covered with some kind of baldachin. Above Building with central courtyard (2) Going up the hill towards southeast and towards the top of Magura, on the right hand side of the road and on the wooded west slope of the hill, around 130 meters from the tetrapylon, 3 — Vasi} ^. 1993 A, 161–163; Vasi} ^. 1993 B, 161–163; Srejovi}, Vasi} 1994, 108–119. 4 — Wulf–Rheidt 2007, 78. 143 IVANA POPOVI] 0 0 PLAN XLV 50 100 m 5 10 m Location and plan of the building with central courtyard 144 SACRED-FUNERARY COMPLEX AT MAGURA FIGURE 109. FIGURE 110. Podium of mausoleum 1 Ideal reconstruction of mausoleum 1 is situated a structure investigated by test trench excavations in 1994 and 1995. In order to preserve the discovered walls before conservation works, the structure has been covered with earth and as, unfortunately, investigations has not been resumed, this structure remained unpublished.5 There were discovered upper sections of the walls of rectangular building oriented into the east–west direction. The site survey conducted in the spring of 2009 revealed that it is a rectangular structure, 23 m x 18 m in size, with the courtyard, 11 m x 9 m, in the center (Pl. XLV). The walls around 60 cm thick and built of half-dressed marl slabs rest on deep foundations consisting of stone rubble and mortar. Large quantity of discovered mosaic tesserae, some of which were gilded, suggest that walls of the rooms had been covered with mosaics. The portable archaeological finds have not been recorded in this structure, so its date and function could not be precisely established. 0 5 10 m tomb walls were supporting an arch spanning 0.86 m in the west section. As the upper sections of mausoleum are almost completely destroyed and stone blocks from this structure used for building Early Byzantine Romuliana, the reconstruction of this structure had been carried out on the basis of rather small preserved segments of the façade. This is the structure with square podium 3.60 meters high and with approximately 9.55 m long sides that had been built of limestone blocks. The structural elements above the podium have been reconstructed only hypothetically on the basis of fragments of architectural decoration found in the vicinity of mausoleum and, according to the proportional relations, common for the Roman architecture. The conclusion has been drawn that on the square podium was an octagonal structure with the door in the façade, leading to the burial chamber (cella). The walls of the cella, judging by the discovered mosaic tesserae, often gilded, were decorated with mosaics. The mausoleum had a pyramidal roof (Fig. 110) and Mausoleum 1 (3) The ruins of the structure identified as mausoleum 1 are situated at the end of the road starting from tetrapylon and around 230 meters to the southwest at the highest point of Magura. This structure of square plan is considerably damaged and only the core of high podium built of limestone blocks and destroyed tomb in its center are preserved (Fig. 109). The funerary structure, 0.96 m x 0.89 m in size, is oriented in east–west direction with slight deviation to the northeast. The walls made of five courses of bricks are preserved up to the height of 0.58 m. The 5 — Field documentation in National Museum in Zaje~ar and Office for Protection Cultural Monumnets in Ni{ is incomplete and the plan of building is lacking. 145 IVANA POPOVI] FIGURE 111. FIGURE 112. Mausoleum 1 and consecration memorial 1 before excavations Scorched earth with remains of burnt wooden structure, consecration memorial 1 the total height of the structure was approximately 11.54 m, i.e. 39 Roman feet.6 Judging by the large number of fragments of porphyry, white marble and limestone discovered in the area between the mausoleum and consecration memorial 1, the structure was lavishly decorated with sculptures and architectural decoration. Consecration Memorial 1 (4) Not far from the north side of the mausoleum there is a large tumulus denoted as consecration memorial 1. It consists of circular stone wall 30 meters in diameter, its preserved height is between 1.80 and 2.00 meters, while on the top of it where piled over 2,500 cubic meters of earth, reaching the height of 8.5 m (Fig. 111). The base of the tumulus is very damaged by subsequent trench digging, so the prehistoric and antique layers are totally disturbed. In the center of stone circle is preserved the scorched surface and few holes of vertically inserted posts. Within that surface (Fig. 112) were found the remains of the burned wooden structure of which are preserved some iron structural elements and fragments of the veneer of gold, gilded and silver-plated bronze.7 In addition to over 3 kg of amorphous pieces of cast silver, there were also recorded few fragments of silver objects of relatively small size (cca 2–10 cm), deformed and greatly damaged due to exposure to high temperature. Still, in most instances we are able to define what the fragments are and to reconstruct to a certain degree the appearance of the fragmented objects. The fragments which make possible the reconstruction suggest that there is a tripod and eight vessels: three plates, three vessels for water or wine – strainers (or paterae), two oinochoe and two bowls or beakers.8 In one of the bowls were found 15 gold coins of which two remained fused to the fragment of the vessel wall.9 Some smaller fragments of silver vessels recorded in the field documentation unfortunately could not be found today.10 Just the fragments of horizontal rims with hunting scenes executed in polishing, engraving, niello and gilding technique have been preserved of two plates. On larger fragment (Fig. 113), under the rim of cast spherical and oval beads is engraved male figure in short tunic, holding ellipsoid shield in one hand and horizontally pointing spear emphasized by niello in the other. The animals – lion and panther bellow it, are leaping toward the hunter, while the figure of lion turned to the right is shallow engraved in the right corner. The traces of gilding could be noticed on the depicted animals. On smaller fragment (Fig. 114), under the rim of cast spherical and ellipsoid granules, is engraved the motif of foliages, spirals and volutes and between the floral ornaments is depicted the leaping bull. Judging by dimensions of fragments, the diameter of the plates was around 6 — Srejovi} 1993 A, 46; Srejovi} 1993 B, 47; Vasi} ^. 1993 A, 150–152; Vasi} ^. 1993 B, 150–152; Srejovi}, Vasi} 1994, 71–81. 7 — Srejovi}, Vasi}, 1994, 82–86. 8 — Popovi} I. 2006, 55–68; Popovi} I. 2009, 315–342. 9 — Srejovi}, Vasi} 1994, 86, fig. 50; Bori}-Bre{kovi} 2009, 350–351, sl. 3–4. 10 — Popovi} I. 2009, 330–331, sl. 18–21. 146 SACRED-FUNERARY COMPLEX AT MAGURA a b FIGURE 113. Fragment of the silver plate rim: a, b) original and drawing; c) reconstruction of the vessel c a b FIGURE 114. Fragment of the silver plate rim: a, b) original and drawing; c) reconstruction of the vessel 147 c IVANA POPOVI] 0 1 2 3 cm a FIGURE 115. b Fragment of the border of central medallion on silver plate: a) original; b) reconstruction of the vessel a b 0 1 2 3 4 5 cm c FIGURE 116. Fragmented handle of silver strainer or patera: a, b) original; c) reconstruction of the vessel 148 SACRED-FUNERARY COMPLEX AT MAGURA a 0 Fragmented handle of silver jug: a) original; b) reconstruction of the vessel 1 cm FIGURE 117. 16.5 cm.11 The friezes with scenes of wild animals hunt, characterized by the pronounced colors achieved by niello and gilding, are typical of the rims of large plates over 40 cm in diameter and dated in the advanced 4th century. The plates from Magura are of smaller size than other plates of this kind and are the earliest so far known examples of this style, popular during the 4th century. Just a small border fragment of central medallion decorated with wavy ornament, probably enhanced with today missing niello, is preserved of the third plate (Fig. 115).12 This motif had been used for decoration of the border of central medallions of silver plates from the 3rd century Gaulish hoards, but also of large Late Roman plates.13 Three more or less fragmented handles are parts of various vessels for wine and water. The fragmented horizontal, solid cast handle (Fig. 116) was a part of a strainer or patera. The strap handle is expanded in the middle with segments shaped like antithetically placed triangles and it is the shape characteristic of the handles of bronze vessels, particularly the strainers. There is a semicircular protrusion in the middle of expanded handle section, while from one edge starts fragmented grooved cylindrical part. It is pos- 0 2 cm b sible that this part was fused to the handle edge because of high temperature and that it was originally under semicircular protrusion on the underside of the handle, like pin for fixing to the handle or the rim of the vessel below (Fig. 116c). To the right and left of the expanded section are horizontal fields creating two identical decorative friezes with deeply engraved spiral foliages. The foliages with leaves shooting from them end at every curl of the spiral in the stylized flower. Considering the size of preserved fragment and probable length of ornamental friezes, the handle could have been around 13 cm long. The ornamental frieze on handle is characteristic of the Gaulish 11 — Popovi} I. 1997, 85. 12 — Popovi} I. 2009, 320–322, sl. 9. 13 — Trésors 1989, cat. 89, 112–115, 130; Baratte 1993, 107, Fig. 36 b; Guggisberg 2003, 139–140, Abb. 95, 119, 120. 149 IVANA POPOVI] 0 1 cm a Fragmented handle of silver jug: a) original; b) reconstruction of the vessel FIGURE 118. 0 silver vessels. It appears on the rims of plates and situlae as well as on the “collars” of distinctive Gaulish bowls14 and was most frequently used on vessels from the hoards deposited around AD 260–270.15 The fragmented vertical, solid cast handle (Fig. 117), consisting of damaged leaf-like segment and lower segment of the handle body of octagonal section (dim. 5 x 1.6 cm), was part of the jug of the oinochoe (oenochoe) type. These vessels for pouring liquids appear among the silver vessels not before the second half of the 3rd century and came to more frequent use in the 4th century. The considerable mass production of these vessels started, as it seems, already by the end of the 3rd century as is suggested by the octagonal jug from Sisak hoard dated by the coins from AD 295/296.16 Unfortunately, it is not possible to determine the precise shape of the jug from Magura on the basis of preserved fragment. The vertical handle, probably curved at right angle in upper section and attached to the rim, transforms at lower end into the leaf-like extension terminating in spherical protrusion used to fix the handle to the vessel wall (Fig. 117b). This decorative design, known from the 3rd century bronze vessels, was encountered also on other silver jugs 2 cm b from somewhat later period, e. g. specimens from the burial at Tarane{ in Macedonia,17 dated in AD 326 on the basis of the fibula with inscription, and from Ni{ (Naissus),18 dating from approximately same time. However, the leaf-like extension on the Ni{ jug was not cast together with the handle, but it had been applied at the spot where handle near its end curves in an arch and terminates in the button-like protrusion. Just of that construction is also the third fragmented handle (Fig. 118) from the Magura find. This strap handle, which is made of polished silver and grooved and arched in the middle, is tapering in the lower section and terminates in the node of 14 — Trésors 1989, cat. 22, 90, 48, 74, 94 (with bibliography). 15 — Baratte 1993, 159. 16 — Jelo~nik 1961, 64, T. XV, 1. 17 — Ivanovski 1984, 221, T. III, 1; Ivanovski 1987, 83, no 2, fig. 5, 1. 18— Kondi} 1994, kat. 271. 150 SACRED-FUNERARY COMPLEX AT MAGURA a 0 FIGURE 119. 1 2 cm Fragment of rim and neck of silver bowl: a) original; b) reconstruction of the vessel b a 0 FIGURE 120. Fragment of rim and neck of silver bowl: a) original; b) reconstruction of the vessel 1 2 cm b 120b). However, the indistinguishable representations of human figures and unidentifiable animals on fragments of the Magura vessels do not make possible closer identification of relief representations while the egg-and-dart molding along the rims has the analogies on the Gaulish vessels. The massive solid cast fragment consisting of damaged segment shaped as an inverse cone with everted rim and cylindrical, onion bulb shape. Its top section was applied to the wall of the vessel, which was probably of the oinochoe shape. The handle of similar shape has also been recorded in the above mentioned burial at Tarane{.19 Only the fragments of rim and neck of two vessels with walls decorated with cast figural representations have been preserved (Figs. 119a, 120). The rims of these solid cast fragments are profiled and decorated with the eegg-and-dart molding. These are probably rather deep conical or slightly biconical vessels with upper sections turning concave near the base (Figs. 119b, 19 — Ivanovski 1984, T. V, 3. 151 IVANA POPOVI] a Fragment of the foot of liver tripod: a) original; b) reconstruction of the tripod FIGURE 121. 0 2 cm b profiled segment (Fig. 121), is most probably fragment of the leg of tripod on which silver vessels were set. Unfortunately, it is not possible to determine, on the basis of preserved fragment, the precise shape and structure of the tripod, which were usually made of bronze. Considering the shapes of preserved specimens it could be assumed that fragment from Magura is the top of one leg with molded segment, which supported circular table board (Fig. 121b). Silver objects from Magura, according to their fabric, could be classified in two groups: three plates, oinochoe with arched handle and small fragment of indistinguishable shape are made of polished silver of fine fabric, while tripod, two bowls and jug with leaf-like handle ending have thicker walls and were made of unpolished silver. The strainer handle is also of massive structure and only with ornamental friezes polished. Certain elements on these vessels, like expansion on the strainer handle and leaf-like attachment on the oinochoe handle, were copied from the bronze vessels of the 3rd century and some vessel forms and decorative patterns used on them indicate the elements characteristic of the Late Roman silver vessels. Na- mely, even two vessels for pouring liquid, that otherwise appear not before the final third of the 3rd century and were largely used in the 4th century, have been encountered in the Magura find. On the other hand, the plates with hunting scenes depicted on their horizontal rims and executed by incision, niello and gilding are not known in the 3rd century finds and that style of decoration was used only from the time of Constantine and on the plates of very large size. Therefore, we may conclude that silver vessels from Magura rely in some elements on the tradition of the 3rd century toreutics, while in some other elements they announce new elements characteristic of the 4th century. It is also important to mention that certain decorative elements, motifs of astragal or beads on the vessel rims and particularly friezes with spirally wound foliages with flower at the beginning of the volute, have the analogies on the Gallo-Roman vessels from the second half of the 3rd century. The mechanism on the strainer handle decorated with friezes with patterns typical of the vessels from Gaul is also an 152 SACRED-FUNERARY COMPLEX AT MAGURA ry in Gaul, Probus’ fifth consulate, establishing of diarchy, second Maximian’s consulate and meeting of two Augusti in Milan. The coins from the second group include the aurei from AD 293 and the beginning of AD 294.23 It is conspicuous that there are no coins issued by Constantius Chlorus and Galerius, who both became Caesars already on the 1st of March AD 293,24 and who celebrated their first consulate in the beginning of AD 294 with the issues from almost all mints.25 But, these specimens are not represented either as individual finds of gold coins from the territory of the diocese Moesia.26 On the other hand, if we disregard the diarchic quinarii, as the nominales not in everyday circulation and aureus from Iantinum mint, the coins from Magura neither by the years represented nor the mints and types do not differ from regular circulation, in the diocese Moesia, where the usual circulation of gold currency with certain oscillations lasted from AD 286 to the Diocletian’s reform in AD 294.27 The contents of numismatic segment of the find from Magura, where there were no reformed aurei usually encountered as individual finds in the diocese Moesia from AD 294 to AD 305,28 suggest the conclusion that the year 294 (295) is not only terminus ante quem non but at the same time also terminus post quem non for the ceremony taking place at Magura,29 when silver vessels and gold coins were placed on the tripod, which was on the pyre of which are preserved parts of wooden structure and metal plating. FIGURE 122. Aureus of Diocletian from AD 293, Iantinum mint: Obv. Head of Diocletian in profile; Rv. Hercules wrestling with Antaeus argument for assumption about the Gaulish provenance of this vessel.20 The results of analysis of the silverware from the memorial at Magura gain even more in importance if they are compared with conclusions reached by the study of numismatic segment of the find including 14 gold coins deposited in one of two bowls or beakers. The Gaulish component noticed on certain silver vessels is conspicuous also in the numismatic material, as one Diocletian’s aureus from AD 293 (Fig. 122) was minted in Iantinum (Meaux in Gaul),21 and that mint was established in that very year exclusively to finance war operations against local usurper Carausius.22 The contents of this find offer also other interesting information. It consists of two groups. The first one contains two quinarii of Probus, the triumphal and consular type from AD 281 and the beginning of AD 282, one quinarius of Diocletian issued on the occasion of establishing diarchy on the 1st of April 286, two consular quinarii of Maximian from AD 288, consular aureus of Diocletian from AD 290 and one aureus of Maximian, issued to commemorate the meeting of two Augusti in Milan in the beginning of AD 291. Six other specimens are the aurei minted in AD 293 or in the beginning of AD 294, including two consular aurei of Maximian, one Diocletian’s aureus from Iantinum mint, issued in the beginning of AD 293, and one Maximian’s and two Diocletian’s aurei, probably from the very beginning of AD 294. So, the coins from the first group are special issues, commemorating events from the years 281, 282, 286, 288, 290 and 291, i.e. the Probus’ triumphal return to Rome after the victo- Mausoleum 2 (5) Mausoleum 2 is situated around 45 meters to the southeast of mausoleum 1. Only lower sections of this building, i.e. the top of high podium with the crypt in the center and staircase on the west side, are preserved of this monument (Fig. 123). The 20 — Popovi} I. 2006, 55–63, fig. 1–24; Popovi} I. 2009, 336–338. 21 — Bori}-Bre{kovi} 1994, 172, nr. 12. 22 — Seston 1946, 102. 23 — Bori}-Bre{kovi} 1994, 160–179; Bori}-Bre{kovi} 2009, 343–354. 24 — Stein 1968, 68. 25 — Bori}-Bre{kovi} 1994, 178. 26 — Vasi} M. 2008, 59. 27 — Vasi} M. 2008, 57–59; Vasi} M. 2008, 57–59. 28 — Vasi} M. 2008, 59–62; Vasi} M. 2008, 59–62. 29 — Bori}-Bre{kovi} 2009, 357. 153 IVANA POPOVI] FIGURE 123. foundation of the mausoleum 2 is shaped as an elongated circle with interior diameter being 2.28 m and external diameter 5.65 m. On top of the foundation was erected the base of podium, circular on the inside and twelve-sides on the outside. In the center was circular crypt with half-dome vault made of bricks. Inside the closed crypt is the masonry tomb of rectangular shape, 2.26 x 3.18 m in size and oriented in the east–west direction. The interior vertical tomb walls made of four courses of bricks are preserved up to the height of 43 cm and they were supporting the barrel vault also built of bricks (Fig. 124). Despite the fact that mausoleum 2 was greatly damaged, it has been concluded after the analysis of its structure, preserved decorative elements and according to the analogous buildings from other parts of the Empire, that on the top of the podium was the base consisting of stone slabs and on that base were arranged 12 columns with Ionian capitals that created portico around the circular cella, 4.85 m in diameter. There was an arched door- Mausoleum 2 0 5 10 m Vaulted tomb in mausoleum 2 FIGURE 125. Ideal reconstruction of mausoleum 2 FIGURE 124. 154 SACRED-FUNERARY COMPLEX AT MAGURA Consecration memorial 2 and mausoleum 2 during excavations FIGURE 127. Ideal reconstruction of pyre from consecration memorial 2 FIGURE 126. the existence of wooden structure, i.e. the pyre platform 27.4 meters long and 17.4 meters wide. The longer sides of the pyre were oriented in the northeast–southwest direction. Judging by the number and disposition of holes, there was another slightly smaller wooden structure above the platform resting on 18 wooden posts and it was the second level of the pyre. The third, smallest platform of this structure covered an area of 9.00 x 7.20 m. The stepped wooden structure (Fig. 127) is the imperial pyre (rogus), analogous to those represented on Roman consecration coins since the middle of the 2nd century.31 The site of fire was crisscrossed with trenches dug by plunderers, hence very few objects placed on the pyre platform have been found. In addition to the carbonized logs of which one was 1.20 m long, animal bones, amorphous porphyry and marble, there were found small objects of iron and bronze – rings, hooks, rivets and wedges of diverse size, but also the pieces of way at the podium level of its façade (Fig. 125). The total height of the structure was around 13 meters.30 Consecration memorial 2 (6) Large tumulus, identified as consecration memorial 2 (Figs. 107, 126), was erected on previously leveled terrain, not far from the west façade of mausoleum 2 and around 15 meters to the south of consecration memorial 1. This space was initially surrounded by stone wall on top of which was piled over 5,000 cubic meters of earth and gravel, so the cone around 10.5 m high had been created. In the middle of the area, surrounded by the wall preserved up to the height of 1.5 to 2 m, was rectangular zone of scorched earth. The holes of vertically inserted posts and slanting half-logs have been encountered within that zone. These holes were arranged in five parallel rows suggesting 30 — Srejovi} 1993 A, 46–47; Srejovi} 1993 B, 47; Vasi} ^. 1993 A, 154–157; Vasi} ^. 1993 B, 154–157; Srejovi}, Vasi} 1994, 89–101. 31 — Hannestad 1988, 216, 262, Fig. 160. 155 IVANA POPOVI] a Fragments of chain mail armor, consecration memorial 2 FIGURE 128. military equipment 32 – fragments of chain mail armor (Fig. 128), short iron dagger with the horizontal strap on the tang, one lance and two oval iron buckles (Fig. 129). The chain mail armor (lorica hamata) was used in the time of early Empire by the horsemen and auxiliary troops and from the 2nd–3rd centuries also by the legionaries. It looked like a short-sleeve tunic made by interlinked iron or bronze rings. On Galerius’ triumphal arch in Thessalonica such armors are wearing the cavalrymen and the soldiers of auxiliary units, while Galerius himself is protected by the scale armor (lorica squamata).33 In any case, the fragments of armor and other elements of military equipment bear witness to a symbolic funeral of the person participating in military campaigns, but as these are the forms common for the 3rd – 4th centuries, they do not provide information about the identity of the deceased and the exact date of the burial. *** The very position of the mausolea and tumuli at Magura reveals that mausoleum 1 and consecration memorial 1 make one complex of structures, while mausoleum 2 and consecration memorial 2 are also one entity. Archaeological excavations of these monuments confirmed that first the mausoleum 1 had been erected on the north side of hill plateau and then next to its south side was constructed circular stone wall and tumulus was piled on top of it. Sometime later the mausoleum was constructed on the south side of the plateau and then next to 156 b FIGURE 129. Object of iron from consecration memorial 2: spearhead, dagger (a), buckles (b) its southwest side was erected circular wall, on the top of which the earth was heaped to create the tumulus. Both mausolea were demolished in the middle of the 5th century and stone blocks had been used for building the Early Byzantine structures in Romuliana. Probably at the same time the tumuli had been breached by the trenches. In both mausolea the crypt contained the burial chamber plundered in the period of mausolea destruction. The traces of wooden structure, i.e. of pyre, particularly conspicuous in the center of consecration memo- 32 — Srejovi} 1993 A, 47–48; Srejovi} 1993 B, 47–48; Vasi} ^. 1993 A, 158–159; Vasi} ^. 1993 B, 158–159; Srejovi}, Vasi} 1994, 102–107. 33 — Laubscher 1975, 27, 29. Taf. 31. SACRED-FUNERARY COMPLEX AT MAGURA 0 FIGURE 130. 50 Monuments at Magura, view from the palace (top) and from the main road (bottom) soleum 1 was well visible from the palace, mausoleum 2, was completely hidden behind the imposing tumulus. Therefore, view to the tumuli was more important than the view to relatively modest mausolea. On the other hand, both mausolea could have been seen from the distance, from the main road in the Timok Valley (Naissus – Timacum Minus – Aquae),35 which means that approaching visitor was supposed to see them first, and only after that the secluded palace (Fig. 130).36 The archaeological material discovered in the course of excavations is not abundant and objects of silver from consecration memorial 1 are of small size and deformed because of high temperature from the pyre. Nevertheless, the reconstruction of silver vessels of which the fragments were found, indicates that it was the silverware which could have been produced in the end of 3rd or the beginning of the 4th century, most probably in the Gaulish workshops. The gold coins deposited in one of the bowls included special issues of rulers from Probus to Diocletian, ending with Diocletian’s aurei from the beginning of AD 294. One aureus of this emperor was minted in AD 293 rial 2 confirm that act of cremation had been performed within these monuments. But, it was not the cremation of bodies of the deceased buried in the mausolea, but of their wax effigies (effigies).34 Important for understanding the meaning and function of four funerary-memorial monuments on the top of Magura is their distinct outline. In other words, while mau- FIGURE 131. 100 m Bronze fibula from the outside of wall circle of consecration memorial 2 34 — Srejovi}, Vasi} 1994, 125–126. 35 — Srejovi}, Vasi} 1994, 118–119; Wulf-Rheidt 2007, 78. 36 — Wulf-Rheidt 2007, 78. 157 IVANA POPOVI] fibulae with short foot and bulbs shaped as pine cones, that appear in the final decades of the 3rd century.38 The investigator of these monuments, Dragoslav Srejovi}, identified the persons buried at Magura and then divinized in the act of symbolic cremation as Romula and her son, Emperor Galerius. In the following pages we present the work of this author, where he backed up his conclusions with the analysis of written sources and the symbolism of the monuments in the Felix Romuliana palace. in the Gaulish mint Iantinum, that started its short activity first of all to finance the war operations of Constantius Chlorus in Gaul. These facts suggest that the ritual carried out within the consecration memorial 1 took place during the last decade of the 3rd century. For more precise chronological determination of the funerals at Magura rather significant are the finds of one complete and one fragmented fibula37 (Fig. 131), discovered on the outside of the wall of stone circle of the consecration memorial 2. They are of the type of early cruciform 37 — @ivi} 2003, kat. 413. 38 — Keller 1971, 32–35; Pröttel 1991, 349–353 (type 1 A). 158 DRAGOSLAV SREJOVI] DIVA ROMULA – DIVUS GALERIUS East of the main gate of Galerius’ palace in Gamzigrad (Felix Romuliana), at a distance of about 1000 m as the bee flies, is the ridge of Magura on the top of which remains of two mausolea and two consecration memorials were discovered between 1989 and 1993.1 The spatial arrangement of these monuments and the archaeological material found in them show that Mausoleum 1 and Consecration Memorial 1 were built first, and that Mausoleum 2 and Consecration Memorial 2 were erected shortly afterwards, i.e. that they were all built between 294 and 313.2 The dating of the mausolea and consecration memorials on Magura into the period between 294 and 313 helps us to establish the identity of the persons is buried and elevated to the rank of gods in Romuliana. The identity of one of these persons quite certain, for historical sources record that Galerius was buried in Romuliana in the spring of 311.3 The identity of the other person may be established in an indirect way. Since that person must have been very closely related to Galerius, it is obvious that he or she should be sought among the members of Galerius’ nuclear family. The possible candidates are not numerous: apart from Galerius’ parents, all the other members of his family died under circumstances which exclude not only the possibility of an apotheosis, but even that of the usual burial. In order to reconstruct as accurately as possible the events which led to the building of Felix Romuliana – both the palace in Gamzigrad and the mausolea and consecration memorials on Magura – it is necessary to discuss first the principal facts concerning Galerius’s family and private history. T. D. Barnes fixed the main dates in Galerius’ career, established where his principal residences were and traced his journeys.4 In order to understand fully Galerius’ building undertaking and, particularly, to interpret accurately the sacred 1 — Srejovi} 1993 A, 45–50; Srejovi} 1993 B, 45–50; Vasi} ^. 1993 A, 148–160; monuments on Magura, it is necessary, howVasi} ^. 1993 B, 148–160. ever, to take also into account the evidence of 2 — A detailed discusion of the mausolea and the consecration memorials Galerius’ character, of his relationship to his can be found in Srejovi}, Vasi} 1994. family and friends, and of his political am3 — Ps. Aur. Vict., Epit., 40.16: “Ortus Dacia Ripensi, ibique sepultus est; quem locum Romulianum ex vocabulo Romulae matris appelarat.” bitions. It is also essential to reconstruct the 4 — Barnes 1982, 2–6; 37–39; 61–64. biographies of the members of his family, 159 DRAGOSLAV SREJOVI] and especially that of his mother Romula, after whom the place in which he was born and buried was named. There is little evidence of Galerius’ origin and youths.5 The date of his birth is not known. He was born in a place not far from Serdica in Dacia Ripensis and his original name was Maximinus. His parents were peasants. His father’s name is not recorded. His mother Romula fled from the Carpi from the left to the right bank of the Danube, i.e. to Dacia Ripensis. Galerius had a sister, probably somewhat younger, who became the mother of emperor Maximinus Daia. In his early youth Galerius was a herdsman and his nickname was Armentarius. It is not precisely known when his military career began. He must have joined the army at an early age, for under Diocletian he was promoted to very high ranks, presumably even that of the praetorian praefectus.6 The first known date in Galerius’ career is the 1st of March 293, when he was proclaimed caesar as C. Galerius Valerius Maximianus. From that moment on, his career can be reconstructed with fair accuracy. In the same year he left his first wife who had born him a daughter named Maximilla7 (later Maxentius’ wife) and married Diocletian’s daughter Valeria.8 It is not known where his principal residence was before 299. It may have been in Sirmium, where Diocletian stayed on several occasions in the course of 293 and 294 and from where Galerius might have conducted his military operations to protect the Danube frontier from the Sarmatians in 294 and from the Carpi and the Bastarnae in 295–296. In 296, after the successful termination of these wars, Galerius had the forests of Pannonia cleared up and the Pelsonian (Balaton) lake drained. In the same year he named that province after his wife Valeria, probably wishing to express in this way his gratitude to her for having adopted, as her own child, his son Candidianus, born to him by a concubine in that year. In 297–298 the most important event in Galerius’ life took place: the war he waged against King Narseus of Persia and his great victory over the Persians. After that triumph he was eulogized throughout the Empire as a second Romulus or Alexander, and Diocletian showered him with great honours in Antioch early in 299. From that time on, Galerius began to create an ideological programme of his own and launched an intensive propaganda campaign in its support: he claimed that he was Mars’s son and Romulus’s brother, and that he was begot, like Alexander the Great, by the god himself, who approached his mother Romula in the form of a dragon.9 In the spring of 299 he hastened from Anatolia to his portion of the Empire, from where he led, in that and the following two years, successful campaigns against the Marcomanni, the Sarmatians and the Carpi. Romula was constantly with him at that time, presumably in his principal residence in Thessalonike, and it is thought that she influenced the developments which took place in Nicomedia in the winter of 302/303 – Galerius’ pressure on Diocletian to launch the persecution of the Christians. In mid-March 303 Galerius came to the Danube, where he fought against the Carpi again. He remained in his part of the Empire, probably in his residence in Serdica, until the spring of 305, when Diocletian proclaimed him augustus in Nicomedia. On the same day, the 1st of May, Constantius, too, was proclaimed augustus, while Severus and Galerius’ nephew Maximinus were nominated caesares. At that time, Galerius had already realized that he was the absolute master of the Empire. In the same year he decided to retire from the throne as soon as he celebrated his vicennalia and to install as rulers his old friend Licinius and his son Candidianus.10 After the death of Constantius I in July 306 Galerius became – in the formal sense, too – the first augustus. However, from that moment on there ensued a series of developments, which he had not envisaged and which did not suit him: Constantine, the son of Constantius I, was proclaimed emperor after his father’s death; Maxentius, the son of Maximian and Galerius’ son-in-law was invested with the purple at Rome on 5 — The following survey of the principal evidence for the family and private career of Galerius is based on the work of Ensslin 1930 and Barnes 1982, loc. cit., who cite and discuss all the relevant historical sources. 6 — Barnes 1982, 156 and note 5 thinks that Diocletian may have given his daughter Valeria in marriage to Galerius, as the praetorian praefectus before 293. 7 — Maximilla may have been the daughter of Galerius and Valeria, cf. Barnes 1982, 38. 8 — Cf. note 4. 9 — Lactant., De mort. pers., IX, 9: “Exinde insolentissime agere coepit, ut ex Marte se procreatur et videri et dici vellet tamquam alterum Romulum maluitque Romulam matrem stupro infamare, ut ipse diis oriundus videratur”; Ps. Aur. Vict., Epit., 40.16: “Is insolenter affirmare ausus est, matrem, more Olympiadis, Alexandri Magni creatricis, compressam dracone semet concepisse”. 10 — Lactant., De mort. pers., XX, 4: “… ita cum imperii summam tenerent Licinius ac Severus et secundum Caesarum nomen Maximinus et Candidianus, inexpugnabili muro circumsaeptus securam et tranquillam degeret senectutem”. 160 DIVA ROMULA – DIVUS GALERIUS the 28th of October 306. Not long afterwards Maxentius nominated his father Maximian “augustus for the second time”. Galerius reluctantly accepted Constantine as his co-ruler and moved, in 307, against his loathed son-in-law Maxentius, bringing his army under the walls of Rome. Finding that he could not defeat Maxentius, he ceded Italy to him. The war with Maxentius had already cost Galerius’ co-ruler Severus II his life. In November 308 Galerius considerably strengthened his position when he managed, with Diocletian’s support, to nominate Licinius augustus in Carnuntum. Galerius’ wife Valeria was proclaimed augusta on that occasion, too. Towards the end of 308 Galerius was in Serdica. His movements from 309 to his death in the spring of 311 are little known. At the beginning of 309 he awarded the title of filius Augustorum to Maximinus. He imposed heavy taxes in order to accumulate funds necessary for the celebration of his vicennalia on the 1st of March 312. In the spring of 310 he fell seriously ill. It is supposed that he proclaimed his son Candidianus caesar at the end of 310 or the beginning of 311.11 Late in April 311 he published the edict on the tolerance of the Christians, and he died a few days afterwards. On his deathbed he committed his wife Valeria and his son Candidianus to the protection of Licinius. He was buried in Romuliana. Valeria enjoyed Licinius’ protection for a short time only. A few months after her husband’s death she left Licinius’ court and went to Maximinus’ portion of the Empire. Maximinus immediately offered to marry her, because he wanted to establish through her a tie of kinship with Diocletian, the founder of the tetrarchy.12 When Valeria refused his offer, she lost all protection: all her goods were confiscated and she was banished, together with Prisca, her mother and Diocletian’s wife, to an obscure place in Syria. After Maximian’s death in Tarsus in July 313 she tried to re-establish friendly relations with Licinius so that she could watch over the fate of Candidianus, who was seemingly showered with great honours. When she heard, however, that Licinius had ordered that Candidianus should be assassinated – an order carried out in Nicomedia in 313 – she fled the country. She remained in hiding in Thessalonike for fifteen months, and then she was found out and decapitated together with Prisca. Their bodies were thrown into the sea. That was Licinius’ final settling of scores with the tetrarchy and with the memory of Galerius. After these events, all building activity in Romuliana must have ceased. Many important details concerning the family and private career of Galerius are passed over in historical sources. The above review of the dates from the biographies of Galerius and the members of his family makes it possible for us at least to surmise some other dates, especially since we have monuments directly associated with Galerius, such as the palace and arch in Thessalonike and the palace with the tetrapylon, mausolea and consecration memorials in Gamzigrad and on Magura. Thus the available historical and archaeological evidence suggests that 298/299 was a crucial year in Galerius’ biography. In this year Galerius’ triumph over the Persians and the celebration of the quinquennalia of his rule happily coincided. It was in this year that the construction of Galerius’ arch in Thessalonike began13 and that Romula was associated with Mars, so it may be assumed that it was then that Galerius decided to mark the place of his miraculous conception by a vast edifice called Romuliana after his mother.14 Romula must have been accorded special tributes from 299 onwards as the mother of Mars’ son. The building of Romuliana was probably just one of the many honours bestowed on her, which happens to be recorded in historical sources. After Galerius’ departure from Antioch in the spring of 299 and arrival in his part of the Empire, he began to devise his own political programme, which is archaeologically documented by the monuments in Thessalonike and Romuliana. The relief decoration on Galerius’ arch in Thessalonike glorifies not only the tetrarchy, but also Galerius as the vanquisher of the Persians.15 The same ideas are expressed in the iconography of the monumental porphyry figure of Galerius in triumph which adorned one of the halls of Romuliana. The Thessalonike Arch was certainly completed before the great jubilee of the tetrarchs in Rome in November 303. The iconography of the porphyry image also dates from approximately the same time.16 11 — On this problem see Barnes 1982, 6 and note l8, where the relevant literature is cited. 12 — Licinius probably had the same wish, which might explain Valeria’s withdrawal from his part of the empire. Cf. Moreau 1954, 41. 13 — Laubscher 1975, 107–108. 14 — On the building of the palace in Gamzigrad see: Srejovi} 1983 C, 61–66, 198–199; Srejovi} 1986 A, 102; Srejovi} 1993 B, 45, 50–51. 15 — Seston 1946, 248 ff.; Laubscher 1975, 95 ff. 16 — Srejovi} 1993 B, 232–233. 161 DRAGOSLAV SREJOVI] The archaeological excavations in Gamzigrad have shown, however, that the building of Romuliana was suspended at one time, or, more precisely, that the original design was abandoned and that immediately afterwards another, considerably more ambitious building project was launched.17 It seems that it is possible to discover which developments caused the change of the original building design. Since only the fortifications were constructed according to the original project, the time required for their building may be estimated at two or three years. On the other hand, since the new, considerably more monumental fortifications of Romuliana had been built by the end of 306, it may be assumed that it was in 303 that Galerius decided to extend Romuliana and make it the most stately edifice in his part of the Empire. It was precisely in the winter of 302/303 that Galerius stayed in Diocletian’s palace in Nicomedia, reportedly to persuade him, under the influence of his mother, to persecute the Christians, and this is the last reference to Romula’s name in historical sources. The Persian triumph and the great jubilees of the decennalia of the caesares and the vicennalia of the augusti were celebrated in Rome in November 303. It is obvious that these events can be hardly taken as an explanation of Galerius’ decision to re-design Romuliana. Consequently, one should probably assume that the real reason for this decision has remained unrecorded in historical sources. The archaeological excavations carried out in Gamzigrad and on Magura in recent years seem to provide an answer. Mausoleum 1 and Consecration Memorial 1, built in the manner of the earlier fortifications of Romuliana, were built on the top of Magura; on the other hand, architectural elements with relief decoration clearly alluding to the imperial apotheosis, including the apotheosis of a female member of the imperial family, have been found near the main gate of the earlier fortifications in Gamzigrad. They include two fragmented archivolts with relief representations of a laurel wreath flanked by peacocks. Within each of the wreaths is a carved inscription, which has been preserved on one of the archivolts and deliberately erased on the other one. Three ivy leaves are carved round the extant inscription, which is FELIX ROMULIANA. The peacocks are the traditional symbols of the apotheosis of the female members of the imperial family, the wreath which they flank should be interpreted as the corona laurea funeraria, while the predict FELIX in the inscription belongs to the charismatic and ritual sphere.18 The other architectural elements associated with the earlier fortifications of the Gamzigrad palace are also decorated with the relief ornaments symbolizing immortality: the picking of grapes, the kantharoi and the intertwined vine and ivy twigs. All this suggests that the principal reason for the radical alteration of the original project of Romuliana was the death of one of the female members of Galerius’ family – his mother Romula, his wife Valeria or his daughter Maximilla. This establishes the identity of the person buried near Galerius and deified in Romuliana. The first mausoleum built in Romuliana, marked Mausoleum 1, is Romula’s mausoleum, and the consecration memorial, marked Consecration Memorial 1, commemorates the site of Romula’s apotheosis. Galerius’ devotion to his mother, frequently referred to in historical sources, must have found a particularly intense expression at the time of her death. It was quite natural that the son should bury his mother with greatest solemnities in the place which he had named after her. He chose the top of Magura as her resting place and the scene of her apotheosis. It was certainly not a random choice, for this place dominates the entire surrounding area and resembles a large garden, which seems to have been considered hallowed from times immemorial – as indicated by the prehistoric cemetery discovered on this site. It is not known where, how and when Romula died. On the basis of the hypothetical year of Galerius’ birth (c. 260 at the latest),19 it may be assumed that she was born in Dacia not later than 240 and that she fled to the right bank of the Danube c. 250, where she married a peasant from a farm in the neighbourhood of Gamzigrad, to whom she bore a son and a daughter. She probably lived there until her son was promoted to the highest military ranks and was awarded the title of caesar. She is known to have been an ardent worshipper of “mountain deities”, presumably Liberus and Libera, to whom she offered daily sacrifices and made ritual feasts.20 Since it is known that Galerius was devoted to his mother and closely 17 — Srejovi}, Jankovi}, Lalovi} 1981, 65–80; Srejovi} 1983 C, 53; Vasi} ^, 1993 D, 118 ff. 18 — Srejovi} 1985, 66–67; Lalovi} 1993 B, 204–208. 19 — Barnes 1982, 37 and note 43. 20 — Lactant., De mort. pers., XI, 1 ff.: “erat mater eius deorum montium cultrix. Quae cum esset mulier admodum superstitiosa, dapibus sacrificabat paene cotidie ac vicanis suis epulas exhibebat”. Cf. the commentary to this passage in Moreau 1954, 267–268. 162 DIVA ROMULA – DIVUS GALERIUS attached to her, it is presumed that she lived in the residences of her son. About 303 Galerius’ principal residence was probably either Thessalonike or Serdica.21 However, Lactantius, describing Romula’s ritual feasts in 303, refers to their participants as vicani (peasants, countrymen), which indicates that Romula lived at that time in a place surrounded by villages inhabited by her countrymen. This could not have been Romuliana, for it had not been built yet, but Romula may have lived in its neighbourhood, possibly at [arkamen, where buildings similar to those in Romuliana have been discovered.22 Galerius lived in his part of the Empire, possibly in the vicinity of his mother, from mid-March 303, and his expedition against the Carpi in the autumn of that year was undertaken probably only after his mother’s death. Romula probably died in her seventies in the summer of 303. It was not at all unusual for a Roman emperor to bury his mother with the greatest honours and to incluide her in the rank of divae. This was in accordance with the best Roman traditions, and particularly with the ideology of the tetrarchs, whose political propaganda required that they should be silent about their fathers and that they should glorify their mothers. Atia, Agustus’s mother, had a public funeral (funus publicum).23 On the other hand, the deification of the female members of the imperial family became the usual practice after the end of the 1st century. Suffices it to mention the consecrations of Domitilla, Marciana and Matidia.24 Romula’s apotheosis is not attested by historical sources or numismatic and epigraphic finds,25 but Consecration Memorial 1 on Magura is a sufficient testimony; besides, all that we know of Galerius’ attitude to his mother is strongly in favour of this hypothesis. Since the form of the grave in Mausoleum 1 indicates inhumation, it may be assumed that the ceremony of Romula’s apotheosis was enacted on the site of Consecration Memorial 1 only after her body had been laid in the mausoleum. Archaeological finds from that memorial indicate that very distinguished persons took part, either directly or indirectly (by contributing gifts to be laid on Romula’s consecrative pyre), in this ceremony. The construction of Romula’s mausoleum and the monument marking the site of her apotheosis on Magura determined the entire further building activity in Romuliana. Diva Romula got a temple in the north part of the palace probably as early as 303. At the end of that year, when the caesares celebrated their decennalia, and the augusti their vicennalia, 163 Galerius must have already had in mind the celebration of the twentieth anniversary of his rule, and he intended to complete the building of Felix Romuliana by that great jubilee. The fact that the main gate of the later fortifications of the palace had been finished by 305/306 shows that the construction of Romuliana was kept at a brisk pace. All the architectural monuments built in Romuliana between 305 and 312 show that by the end of 305 Galerius had decided not only to renounce the throne on the 1st of March 312, but also to retire to Romuliana as senior augustus. Whatever was built in Romuliana in that period was associated with Galerius’ person and his ideological programme. That programme left nothing to chance: the position of the individual buildings, their size, their appearance and decoration – everything was in the service of the tetrarchy and of Galerius as its absolute head. Everything weas conceived as part of a great spectacle, a grandiose theatron for the ceremony of the imperial apotheosis and the establishment of the cult of Divus Galerius. Galerius died a year before the planned completion of Felix Romuliana. At the time of his death the entire north part of the palace and some structures in its south part (the baths, the four-aisled building, the porched building, the building with the cruciform ground plan and the building located between the baths and the porched building) had already been completed. It took another year to construct the communication lines and to set up the temenos of the large temple. It is not known where Galerius expired. Lactantius mentions that the odour of his sickness spread not only all over the palace, but throughout the town.26 It is consequently assumed that he died in Serdica, although there is no reliable evidence in support of this hypothesis has been found.27 The decision that he should be buried in Romuliana was certainly not unpremeditated. There is no reason to doubt that this was 21 — Barnes 1982, 61–62. 22 — Vasi} ^. 1993 F, 189. 23 — Dio Cassius 47.17. 24 — Cf. Price 1987, 92 ff., with a full bibliography. 25 — It is not improbable that the deliberately destroyed inscription on the archivolt from Romuliana was DIVA ROMULA. 26 — Lactant., De mort. pers., XXXIII,25. 27 — Barnes 1982, 64, notes that Galerius spent the last days of his life in the province of Dardania. DRAGOSLAV SREJOVI] Galerius’ own choice, a choice he had made when he buried his mother in Romuliana and when he had decided to make this place his residence in his old age. It should be, therefore, surmised that before his death he was transferred to some place near Romuliana, possibly to the same place in which Romula had died. Licinius, Valeria and Candidianus were probably with him during his last days.28 In case he passed away in Serdica, his body might have been transported to Romuliana in a comparatively short time and laid to rest in the crypt of Mausoleum 2 on the top of Magura.29 The funeral and the enactment of Galerius’ apotheosis were probably attended by Licinius only, for the other two augusti were far away from Romuliana – Maximin in Syria, and Constantine probably in Autun.30 Galerius’ apotheosis was enacted in the same way as that of Romula, but the site was more spacious, the pyre was larger and furnished with adjuncts becoming a soldier-emperor. While the gigantic mound that was to mark permanently the scene of his apotheosis was being heaped up, the large temple dedicated to the worship of Divus Galerius was being completed in the south part of the palace. Fragmented statues of tetrarchic deities – Jove and Hercules – were found in the ruins of this temple. Sculptures, reliefs and mosaics alluding to Galerius’ apotheosis, particularly representations associated with Dionysus, Hercules and Asclepius have been discovered in the other parts of the palace. The link between these three deities and the ideological author of Romuliana can not be more obvious: Dionysus, Hercules and Asclepius were, like Galerius, begotten by a god upon a mortal woman. They were all saviours of mankind, admitted to the rank of gods after having accomplished great tasks on earth. Dionysus is, however, privileged: the entire Romuliana is in the sign of this deity. There were several reasons for this. Galerius could compare his great victory over the Persians only to Dionysus’ triumphant expedition to India. The decoration of Galerius’ palace in Thessalonike, in which Dionysus is accorded a very prominent place, also shows that, from that time on, Galerius used the Dionysus myth as a prototype for the creation of his own myth. The discovery of Romula’s mausoleum on Magura shows that Galerius also modelled his relationship with his mother upon that of Dionysus, who, after his victories in the East, deified his mother Semela. It is difficult, however, to establish whether the cult of Divus Galerius took root in Romuliana, but even if it did, it must have been discarded already in the course of 313. Galerius’s apotheosis was probably the last rite of this kind enacted in the traditional way in the Roman world. It seems that even Diocletian, who probably died in 313, was denied an apotheosis,31 and the first emperor who died a natural death after him was Constantine, whose consecration marks a radical break with tradition.32 It has been argued that the imagery of consecration coinage shows that tetrarchic imperial funerals differed from those preceding them.33 The archaeological evidence from Magura does not support this view. It is a well known fact that the tetrarchs were particularly anxious to revive the traditional Roman religion and cults, and this in itself makes it unlikely that they wished to change the ceremonial of the imperial funeral and apotheosis. The mausolea and consecration memorials on Magura conform in all their elements to the requirements for the traditional imperial funerals, particularly those documented in the period from Trajan to Septimius Severus. The omission of the term consecratio and of the representation of the pyre on Galerius’ consecration coinage, and the introduction of a new legend “for his eternal memory” does not mean that the pyre had lost its importance in the consecration rite. This is very clearly shown by the remains of the monumental pyres uncovered near Romula’s and Galerius’ mausolea in Romuliana. The choice of the site for the erection of Romula’s and Galerius’ mausolea also shows that the tetrarchic imperial funerals did not differ from the traditional ones. Care was taken to separate the mausolea from the palace, which means that the ancient Roman rule prescribing burial extra muros was strictly observed. This fact brings into question the commonly accepted view that Diocletian, the founder of the tetrarchy, departed from the tradition when he built a mausoleum within his palace in Split, thus giving a new direction to the development of imperial funerary architecture.34 It is only on the assumption that the octagonal building in Diocletian’s palace in Split is a mau- 28 — Lactant., De mort. pers., XXIII, 6. 29 — Serdica is about 220 kilometres distant from Romuliana. 30 — Cf. Barnes 1982, 66, 70; Lactant., De mort. pers., XXXVI, 1 ff. 31 — Cf. Barnes 1982, 35. 32 — Koep l958. 33 — Price 1987, 99–103. 34 — Frazer 1966; Waurick 1973, 124 ff. 164 DIVA ROMULA – DIVUS GALERIUS soleum that all the later similar structures – including the Rotonda in Galerius’ palace in Thessalonike and the so-called Maxentius’ Mausoleum in the Via Appia in Rome – have been classed in the same category of monuments.35 Thus a mere conjecture is responsible for the view that a mausoleum was a common feature of late classical imperial residences. The mausolea in Romuliana show that this is not true. As the only verified tetrachic imperial mausolea, they make it possible for us to re-interpret some other monuments of late classical sacred and funerary architecture. Romula’s and Galerius’ mausolea in Romuliana were erected without the walls of the palace, as it was prescribed by Roman laws. They were located in an area resembling the gardens of Eden and similar to the funerary gardens without the walls of Rome. Conceptually, the imperial mausolea in Romuliana have their closest parallels in Hellenic funerary architecture, and it is possible that they were modelled after some unknown mausolea of Alexander the Great’s successors. The tetrarchs must have found the view of the ruler as a heros, as a demigod even during his earthly life, very congenial and convenient for the tetrarchic ideology. This applies particularly to Galerius, who considered himself a protohero, another Romulus (alterus Romulus) and Alexander (Alexander redivivus). The consecration memorials in Romuliana are associated with the same ideology, for tumuli also suggest the cult of the heros and may be understood – like the toloid structures represented on Galerius’ consecration coinage – as symbols of apotheosis.36 The traditional model for the imperial apotheosis – the deification of Romulus and Hercules – was enriched in Romuliana by the identification of Galerius with Alexander the Great and Dionysus. Just as Dionysus and his mother Semela joined the gods after his triumphant expedition to India, Galerius – Neos Dionysos – and Romula ascended to the sky from the top of Magura. After their apotheoses Romuliana was left to the mortals, who, after a brief period of strife and religious intolerance, radically changed its function and its appearance.37 (Diva Romula – Divus Galerius, D. Srejovi}, ^. Vasi}, Imperial Mausolea and Consecration Memorials in Felix Romuliana (Gamzigrad, East Serbia), Belgrade 1994, 141–156; Diva Romula – Divus Galerius, The Age of Tetrarchs (ed. D. Srejovi}), Belgrade 1995, 295–310; Diva Romula – Divus Galerius. Posledwe apoteoze u rimskom svetu, Sun~ani sat 5, Sremska Mitrovica 1995, 17–30. 35 — Cf. Frazer 1979, 110; Rasch 1984, 75 ff. 36 — Schulten 1979, 135 ff. 37 — On Byzantine and early mediaeval Romuliana see Jankovi} 1983, 99–119; Jankovi} 1983 A, 120–141 and Jankovi} 1983 B, 142–160. 165 SOFIJA PETKOVI] ROMULIANA IN THE TIME AFTER THE PALACE Professor Dragoslav Srejovi}, who was director of the SASA project from 1970 to 1996, contributed mostly to the investigation and interpretation of Gamzigrad. He was the first to identify Late Roman fortification at Gamzigrad as tetrarchic imperial palace1 and this has been confirmed by archaeological excavations in the following years. So it happened that uniqueness and importance of Romuliana overshadowed archaeological finds from the period preceding the activities of Galerius and from the times that followed. Excavations 2002–2007 PLAN XLVI Recent archaeological excavations at Romuliana (2000–2008) 167 SOFIJA PETKOVI] FIGURE 132. Tower 19, plan and entrance to the tower However, during half a century of investigations, Gamzigrad proved to be rather complex multi-layered site. Archaeological investigations revealed that after brief splendor in the first decade of the 4th century the imperial palace Felix Romuliana was transformed into fortified settlement, which lived intensively from the end of the 4th to the end of the 6th / beginning of the 7th century. The stratigraphic data and interpretation of the reconstructions of earlier structures of Galerius’ Romuliana and new structures, dating from the second half of the 4th to the beginning of the 7th century, are most thoroughly published in the catalogue of the exhibition Gamzigrad. Kasnoanticki carski dvorac / Gamzigrad. An Imperial Palace on the Late Classical Times, organized in 1983 in the Gallery of SASA.2 The cultural stratigraphy of Gamzigrad, presented by Dragoslav Srejovi} and Djordje Jankovi} in the first publication in 1983, was developed and modified to a certain extent in the monograph on memorial complex at Magura, published in 1994 in connection with the exhibition Rimski carski gradovi i palate u Srbiji / Roman Imperial Towns and Palaces in Serbia, organized in 1993 in Gallery of SASA in Belgrade.3 New archaeological investigations at Gamzigrad that are have been in progress since 1997 until today: 1997–1998 and 2002 in the south tower (tower 19) of west gate of later fortification, 2004–2007 in the southeast section of fortification in the sector of thermae and 2005–2007 outside the walls of fortified palace, extra muros, completed stratigraphic picture of Romuliana from the end of 4th to the end of 6th / beginning of the 7th century (plan XLVI). Five horizons of living, grouped in two phases, have been established: 1. First phase, dated from the final quarter of the 4th century to the middle/second half of the 5th century, and 2. Second phase, dated in the end of 5th – end of 6th / beginning of 7th century 1 — Srejovi} 1983 C, 53–66, 195–199. 2 — Srejovi} 1983, 14–16, 193–194; Jankovi} 1983; Jankovi} 1983 A; Jankovi} 1983 B. 3 — Srejovi}, Vasi} 1994, 56–59. 168 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 133. Northeast section of cultural layers in segment II of tower 19 Archaeological investigations in the south tower of west gate of later Romuliana fortification (tower 19) started in 1996 in order to prepare this structure for conservation and presentation and to gather relevant data about the construction of the fortification. Systematic excavations of tower 19 started in 1997 and in the course of field works tower was divided into four identical segments (Fig. 132): segment I in the south, segment II in the west, segment III in the north and segment IV, including the tower entrance, in the east. The objective of the division of internal tower space was to provide comprehensive stratigraphic data, which could be observed on the profiles of cross-section of cultural layers AA' and BB'. During 1997 excavation campaign segment II was investigated, segment III was investigated in 1998 and reports of the investigations are published in the excavations chronicles in Starinar in 1997 and 2000.4 The excavations in tower 19 were resumed in 2002 and segments II and IV have been investigated.5 The stratigraphy of cultural layers in tower 19 is mostly based on the excavations in segment II. It concerns the layers accumulated from the beginning of the 4th to the second half of the 6th century on top of the culturally sterile layer of gray-yellow clay, layer H (Fig. 133).6 Layer G, 20–30 cm thick and consisting of yellow sandy soil with large amount of broken tegulae and imbrices, is the drainage layer under the floor substructure in tower 19. Layer F, around 30 cm thick, is the substructure of tower floor and consists of lime mortar with smaller pieces of stone rubble and gravel. Horizon f is the floor of white lime mortar that was most probably paved with tegulae and covered top surface of the bases of stone pillars 1, 2 and 3. The pillar bases were resting on the drainage layer G. Horizon e (Fig. 134) is rammed earth of reddish-brown color, immediately on top of mortar floor of tower 19 dating from the period of restoration of the settlement at Romuliana during last quarter of the 4th century. In the interior of tower 19 was the workshop for metalworking, i.e. the smithy.7 Many kilns of rectangular or circular plan paved with tegulae and with calotte-shaped roof of pieces of stone and tegulae bonded with clay have been recorded at this level. Considering the type of structure and finds of iron objects and slag, the kilns investigated at horizon e in segment II of tower 19 were features of the blacksmith’s shop. Some of the kilns had been reconstructed many times and the workshop was destroyed in a big fire (Fig. 135). 4 — Biki}, [ari} 1997, 203–208; Biki}, [ari} 2001, 280–282. 5 — Petkovi} et al. 2004, 295–301; Petkovi} 2004, 127–153. 6 — Petkovi} 2006, 32–40. 7 — Petkovi}, @ivi} 2006 A, 111–128; Petkovi}, @ivi} 2006 B, 135–140. 169 SOFIJA PETKOVI] 0 FIGURE 134. 1m Ground plan of horizon e in segments II and III of tower 19, final quarter of 4th century 170 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 135. Blacksmiths’ kilns from horizon e in tower 19 (kilns 4/02, 5/02, 6/02 and 7/02), final quarter of 4th century cates the effect of high temperature, so it is assumed that this kiln was used for metal smelting. In the kiln was also found Late La Tène lanceolate fibula, most probably prepared for remelting (Fig. 138). Horizon d, which also perished in conflagration, is overlaid by the leveling layer (layer D) containing the finds similar to those found in the previous layer (Fig. 139). In layer D was also found the completely preserved antler-made comb with case and decorated with the horse protomes, that rather precisely date layer D in the end of 4th – beginning of the 5th century. Such combs had been worn as status symbol by the soldiers of auxiliary cavalry units of the Roman army (fig. 140)8. This horizon of living in tower 19 is dated, according to the finds from the leveling layer just on top of it (layer E), in the final third or quarter of the 4th century (Fig. 136). Two workshops have been investigated in the next horizon of living in tower 19, horizon d; one in segment II and central section of the tower containing the blacksmith’s kiln 3/02 and the other was in segment III and central section of the tower between the pillars, and it has walls of broken stone, bonded with clay, and two blacksmith’s kilns of rectangular plan (hearths 7 and 9) (Fig. 137). The kiln of circular shape, that was twice renovated, was investigated outside the described structures. Inside the kiln was found the large amount of metal slag, ash and carbonized wood. Intensely burnt soil under and around this kiln indi- 8 — Petkovi} 1999, 219. 171 SOFIJA PETKOVI] FIGURE 136. Objects from layer E in tower 19, final quarter of 4th century Above the layer C was accumulated the horizon of yellowbrown rammed earth (horizon b1), where the traces of intense fire were recorded, and from that level was dug an ellipsoid pit (3 x 0.75 m), filled with ash and soot. This horizon of living in tower 19 was destroyed in rather large fire. The overlaying layer (layer B1), brown soil with traces of burning – lenses of soot, ash and burnt earth mixed with building rubble, indicates the destruction of habitation in tower 19 and its abandoning for rather long period of time. Relevant for dating layer B1 in the second half of the 5th century are the lamps made on potter’s On the top of layer D in the south half of segment II of tower 19 have been found sections of the floor of green-yellow rammed earth (horizon c) and kiln of rectangular plan and fragmented ceramic pot inside it. Cultural layer C, gray-brown soil with soot, overlaying horizon c, abounded in fragments of pottery and glass vessels from the first half of the 5th century and in animal bones indicating many years of living in tower 19. The tower 19 was used as residential structure or as temporary shelter in the first half of the 5th century. 172 ROMULIANA IN THE TIME AFTER THE PALACE 0 FIGURE 137. 1m Ground plan of horizon d in segments II and III of tower 19, end of 4th – beginning of 5th century 173 SOFIJA PETKOVI] FIGURE 138. Furnace for smelting metal at horizon d in segment II of tower 19 and bronze lanceolate La Tène fibula ready for remelting, found inside the furnace Objects from layer D in tower 19, end of 4th – beginning of 5th century FIGURE 140. Antler comb with case, decorated with horse’s protomes, end of 4th – beginning of 5th century FIGURE 139. wheel, iron fibula with backward turned trapezoid foot characteristic of the “barbarian” culture Chernyahov – Sîntana de Mures and the double-filed and single-filed combs of antler (Fig. 141). This layer reflects the deterioration of tower 19 in the ensuing decades. The habitation at horizon b1 in tower 19, established in the middle of the 5th century, was most probably destroyed in the Hunnish invasion of Dacia Ripensis in AD 441, when Romuliana had also bore the brunt of the “barbarians”. 174 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 141. Objects from layer B1 in tower 19, middle / second half of 5th century Grave from 5th century buried in central zone of tower 19 (grave 1/02) and grave goods FIGURE 142. Nevertheless, the life continued at Gamzigrad even in the second half of the 5th century, as it is confirmed by two graves buried in the central section of tower 19. Small cist made of Roman tegulae without any finds was discovered in 1996 and in 2002 was discovered the grave of identical type, where an adult woman with bone spindle whorl on the chest and double-lined comb of antler by her head had been buried (Fig. 142). The next horizon of living in tower 19 (horizon b) could have been identified only partially, on the basis of floor segments 175 SOFIJA PETKOVI] FIGURE 143. Objects from layer B in tower 19, first half of 6th century of light yellow rammed earth and two hearths of ellipsoid plan in segments I and II. This horizon was covered with layer of brown earth with lenses of soot and larger fragments of building rubble (layer B) that contained just few finds, except for the 6th century fragments of pottery and glass vessels. Important for more precise dating of horizon b and associated layer B is the cast bronze fibula with backward turned foot and broad decorated bow from segment III, that is of the type originating in the Lower Danube basin during the first half of the 6th century, and the coins of the ruler of Eastern Roman Empire, Justin I (518–527). Generally, this horizon of living in tower 19 could be dated in the first half of the 6th century (Fig. 143). The remains of floor of yellowish-white mortar, identified as horizon a, have been recorded above the layer B in segment II of tower 19. This horizon is overlaid by the layer of building rubble from the collapsed tower (layer A) containing also few iron clamps and insignificant amount of the 6th century pottery fragments. Horizon a most probably corresponds to the restoration of Romuliana in the middle of the 6th century, during the reign of emperor Justinian I (527–565). The restoration is mentioned in the Procopius’ work De aedificiis and could be noticed in the repairs of west and south rampart. At that time the monumental towers of fortification of Galerius’ palace had been reconstructed for the last time as elements of the defensive system of Early Byzantine settlement at Gamzigrad, mentioned in the area of the town Aquis (VAkue,j), as ~Rwmuli,ana.9 Systematic archaeological excavations were conducted between 2004 and 2007 next to the south and east façade of the Roman baths, which were dated in the later phase of construction of Galerius’ palace. The objective of these works was to connect “Large temple”, east gate of fortification of Galerius’ palace and its southeast corner with the earlier square tower V and later polygonal tower 5, in order to prepare the site for conservation, restoration and presentation.10 These investigations are still in progress and an area of 264 square meters from the present day ground level to the level of Galerius’ architecture from the end of 3rd / beginning of 4th century (horizon g) has been explored (Fig. 144). The total thickness of cultural layer to the virgin soil (layer J) in this area is around 3.70 m and the following stratigraphy of cultural layers accumulated after construction of the imperial palace at Romuliana (horizon f) (Fig. 145) has been encountered. Horizon f, floor of white mortar, is dating from the later phase of construction of the Galerius’ palace, i.e. from the beginning of the 4th century. This horizon, containing the 4th century finds, overlays cultural layer F, consisting of 10 cm of soot, ash and burnt earth, 15–20 cm of the leveling layer of sand with small pieces of rubble and 15–30 cm thick substructure of mortar floor of the next horizon (horizon e). 9 — Procop., De aedif., IV,4. 10— @ivi}, Petkovi} 2004, 19–28; Petkovi}, @ivi} 2005 A, 32–37; Petkovi} 2008 B, 61–63; Petkovi} 2008 C, 64–67. 176 ROMULIANA IN THE TIME AFTER THE PALACE a 0 5 10 m Plan of Galerius’ thermae (a) and situation in 2002, before conservation and restoration works (b) FIGURE 144. large quantity of iron slag and iron objects, including also an anvil. Also, two metallurgical kilns with fragments of iron objects and slag around them (kiln 1/07 and kiln 2/07) have been discovered to the east of Galerius’ baths. Like in tower 19 the smithies and other shops for metalworking were built in the area of the imperial baths during the final decades of the 4th century. The remains of collapsed walls of some building and segment of damaged mosaic floor of large stone cubes of white, gray and black color have been found east of the thermae at horizon e. This structure had been pulled down in the end of the 4th century and the floor was destroyed by digging large waste pit (pit 4/04) from the later settlement horizon (horizon d). b Horizon e is the white mortar floor on the top of layer F and it dates from the time of life restoration at Romuliana after abandoning construction of Galerius’ palace. At this level was constructed the blacksmith’s kiln of rectangular plan, paved with tegulae (kiln 1/04), to the west of Galerius’ thermae, and next to the south façade of the thermae was built the room, 8 x 4 m, with walls of broken stone and leveling courses of tegulae bonded with yellow clay and with floor of white mortar. In this room was built the blacksmith’s kiln (kiln 2/04) next to the south wall of thermae and inside the kiln was found rather 177 SOFIJA PETKOVI] a b Section of cultural layers in the thermae sector, east of Galerius’ structure (a), south and north profile (b) FIGURE 145. The holes of rather large posts filled with soot and carbonized wood have been discovered in the mortar floor of horizon e. These are most probably the remains of columns of wooden portico in the southeast corner of fortification. The mortar floor in the area southeast of thermae was at the street level, i.e. according to the results of investigations it is outside the structures. Despite the remains of smithies, it seems that thermae had been still in use in this period, or the thermae building was partially also used for other purpose. The settlement at Gamzigrad, established in the final quarter of the 4th century, was destroyed by conflagration, as it is confirmed by the layer of burning with lumps of large carbonized posts, possibly the remains of burnt structure of wooden portico. Also, the traces of collapsed walls are evident, so life in the settlement at this level came to an end in a destruction of rather large scale. Horizon e is covered with 40–60 cm thick layer E, consisting of gray-brown soil with traces of burning and building rubble and containing many finds from the end of 4th and the first half of the 5th century (Fig. 146). This layer accumulated during the first half of the 5th century as a result of destruction and burning of structures from horizon e and leveling of ruins prior to the construction of the new settlement. The destruction of the settlement at Romuliana established during the final quarter of the 4th century could be related to the series of “barbarian” attacks on the territory of the Empire after the battle of Adrianople, between 379 and 382, or the incursion of Uldis’ Huns from the left Danube bank to Dacia Ripensis in 409.11 The settlement at Romuliana was restored on the top of the leveling layer (layer E) in the first half of the 5th century over the entire investigated zone southwest and southeast of the thermae complex and it is identified as the horizon of yellow rammed earth and here and there of the low quality yellowishwhite mortar floor (horizon d). On top of the leveled layer of fire (layer E), in the smithy the room added to the south, façade of the thermae in the previous horizon was reconstructed. It was expanded towards the west (dimensions are 8.50 x 4.00 m), so its west wall is parallel with the west wall of thermae and in the north it is leaning on the second pilaster of west façade. The layer of carbonized grain, up to 30 cm thick, and fragments of carbonized beams of the roof structure were found inside the room on the floor of rammed earth. The finds from this layer (layer D) are distorted because of high temperature, and considering the layer of carbonized grain and few fruits, it had most probably been the food storehouse, granary, burnt down in conflagration (Fig. 147). The remains of collapsed house walls, built of larger stone blocks and fragments of tegulae bonded with clay, were found at horizon d, southeast of the thermae. One structure situated 11 — Zosim., Hist., IV, 34; Jankovi} 1983, 107–109. 178 ROMULIANA IN THE TIME AFTER THE PALACE to the southeast of the thermae and oriented in the north–south direction had a kiln of rectangular plan (kiln 3/05) and floor paved with tegulae. Somewhat better preserved structure has been investigated to the east of Galerius’ thermae. The house of rectangular plan, 6 x 4.5 m in size, oriented in the north–south direction (house 1/07), has been completely destroyed by fire. Under the collapsed roof structure of tegulae and imbrices was encountered 10–15 cm thick layer of soot with pieces of carbonized wood, covering the floor of light yellow rammed earth. The remains FIGURE 146. of two earlier floors of lime mortar, leveled with sand and small gravel (horizon d1and horizon d2), were discovered under that floor. The preserved sections of east and south wall are of broken stone and tegulae bonded with loose yellow mortar. Many reconstructable pottery vessels, pots, bowls, jugs and lids, as well as pottery lamps, comb, knife handle and antler-made pyxis, bronze fibulae and many bronze coins of Valentinian I (364–375), Valens (364–378), Gratian (375–383) and Valentinian II (375–392), iron objects and tools, fragments of glass vessels were found inside the house. These finds date the Objects from layer E in the thermae sector, end of 4th – beginning of 5th century 179 SOFIJA PETKOVI] FIGURE 147. FIGURE 148. Granary at horizon d in room added to the south façade of Galerius’ thermae, middle of 5th century Objects from the house at horizon d east of Galerius’ thermae (House 1/07), middle of the 5th century building construction in the first half of the 5th century and its destruction in the middle of that century, most probably during the Hun invasion in AD 441 (Fig. 148). Two large pits – silos dug from the horizon d have been discovered in the area southwest and southeast of the thermae. Both silos, over 2 meters in diameter and approximately of identical depth, were later used as waste pits and contained finds from the end of the 4th and the first half of the 5th century. The settlement in the southeast section of Romuliana, established during first half of the 5th century, was razed to the ground in the Hun attacks between 441 and 447.12 It is confirmed by substantial layer, 50–75 cm thick, of building rubble over the destroyed horizon d, and consisting of fragments of 12 — Jankovi} 1983, 109. 180 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 149. Objects from layer D in the thermae sector, 5th century 0 FIGURE 150. 1 2 3 4 5 10 m Plan of artisan-metallurgical complex at horizon c in the thermae sector, end of 5th – beginning of 6th century 181 SOFIJA PETKOVI] b Large furnace for smelting iron ore in artisan-metallurgical complex (kiln 7/04) (a) and section of slag deposit from the furnace (b), first half of 6th century FIGURE 151. center, have been investigated southwest and southeast of the thermae (Fig. 150). The structure in the room added to the south façade of the thermae had been restored for the second time in horizon c. The reconstructed building, 7.90 x 3.55 in size, built of large rubble stone and fragments of tegulae bonded with clay, was reinforced at the corners with large dressed limestone blocks brought from the demolished podium of “Large temple”. Considering large quantity of fragments of amphoras and pithoi, vessels for storing wine, oil and wheat inside this room, it had been used as the provisions storehouse. It is not possible, at this level of investigation, to speak with certainty about size and function of the structures investigated at horizon c to the southeast of the thermae in 2004–2005. It was most probably the metallurgical complex, judging by remains of many kilns and large quantity of metal slag and dross found at this level.13 The entire structure was situated in the furthest southeast corner of fortification, between the thermae a tegulae, stones and lime mortar mixed with lenses of soot, ash, carbonized wood, burnt earth and brown-yellow clayey sand soil (layer D). The substantial layer of building rubble, around 1 meter thick and resulting from destruction of the southeast towers of later and earlier fortification (tower V and tower 5) and the east rampart, has been encountered to the southeast of Galerius’ thermae. Layer D, accumulated as a result of destruction of structures from horizon d in the middle of the 5th century and of later collapsing of south and east rampart, associated towers and thermae during second half of the 5th century, contains the finds from the end of the 4th and the first half of the 5th century (Fig. 149). Most probably in the end of 5th or in the beginning of the 6th century the building rubble in the southeast section of fortified Romuliana had been leveled and new settlement (horizon c) was established. Many structures of that settlement, most probably the workshops organized as metallurgical – artisan 13 — Petkovi}, @ivi} 2005, 101–108; Petkovi}, @ivi} 2006 A, 111–128; Petkovi}, @ivi} 2006 B, 140–147. 14 — @ivkovi} et al. 2005, 125–129; @ivkovi}, [trbac 2005, 43–46. 182 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 152. and east and south rampart. It consisted of many rooms parallel to the south and east rampart and surrounding an open central area, some kind of the courtyard paved with large stones and pebbles. In the courtyard was a large horse-shoe shaped kiln, 5 x 2.5 m in size, for smelting iron ore (kiln 7/04), and in front of it were two waste pits for slag (pit 2/04 and 3/04) and large slag heap, in places up to 1 meter thick, that extended 12 meters to the southeast along natural slope of the terrain (Fig. 151). The kiln 7/04 was certainly used according to archeometallurgical analyses for smelting iron ore,14 while for the time being we can not speak with certainty about the purpose of other kilns (1/05, 2/05, 6/05 and 7–2005) found within the Gold tremissis of Emperor Justin I, 518–527 FIGURE 153. Objects from layer C in the thermae sector, first half of 6th century 183 SOFIJA PETKOVI] structure at horizon c. It is not certain whether they had been used in iron metallurgy, for processing non-ferrous metals or glass. Still, some of them were probably founder’s or blacksmith’s kilns, especially those with large quantity of iron slag found inside them. However, some of the investigated rooms could have been the workshops for working nonmetals. Namely, grouping of artisans’ workshops in one “quarter” within settlement, next to the rampart, was usual in the Late Roman period. The stratigraphic data and archaeological finds date rather reliably establishing of the artisans’ complex in the end of 5th / beginning of the 6th century, i.e. in the reign of emperors Anastasius (491–518) and Justin I (518–527). Particularly interesting is the discovery of one gold coin of Justin I on the floor of one of the rooms of this artisan center (Fig. 152). FIGURE 154. The metallurgical-artisan complex from horizon c is overlaid by 45–50 cm thick layer of light brown earth with traces of burning and substantial quantity of iron slag and with lenses of sand and clay (layer C). It contains archaeological finds from the Early Byzantine period, from the 5th and first half of the 6th century: coins, fibulae, pottery and glass vessels (Fig. 153). Large number of iron objects, whetstones, and fragments of antlers with traces of working, pottery and stone casting molds indicate the metallurgical activities – smelting, casting and forging iron (Fig. 154). The new settlement (horizon b) was established on the leveled ground in the southeast section of Romuliana. It consisted of many structures with foundations built in dry masonry technique of broken stone and bricks without lime mortar and Objects from artisan-metallurgical complex in the thermae sector related to iron metallurgy, first half of 6th century 184 ROMULIANA IN THE TIME AFTER THE PALACE a clay as bonding material. A vessel of unusual shape and technology of manufacture was found in a destroyed kiln (kiln 8/04), on the floor of yellow rammed earth in a house built against the wall of earlier metallurgical structure from horizon c. The surface of the tall jug of almost cylindrical shape and with one handle was decorated with applied bands with rectangular and crescent shaped impressions. Under the handle is a triangular tongue-like projection with impressed Latin letter V with “eyelets” at the ends. The vessel was made of well refined clay with large amount of mica and fired to the yellowish-gray color, so the vessel surface has metallic, silver-like luster. The lower section of b Jug found in the kiln (kiln 8/04) from horizon b in the thermae section: complete (a), handle detail (b), second half of 6th century FIGURE 156. Objects from layer B in the thermae sector, second half of 6th century FIGURE 155. 185 SOFIJA PETKOVI] Necropolis extra muros south of Romuliana fortification, final quarter of 4th – first half of 5th century – jewelry from child’s grave 1/06, grave 2/06 and bronze buckle of military belt from that grave FIGURE 157. FIGURE 158. Grave of military commander next to foundation zone of south rampart of Romuliana (grave 6/06), necropolis extra muros, end of 4th – beginning of 5th century FIGURE 159. Objects from the grave of military commander (grave 6/06) – cruciform gilded fibula with imperial portraits and bronze buckle of military belt 186 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 160. Grave of Roman soldier from the bank of Draganov potok, necropolis extra muros, northeast of Romuliana fortification, end of 4th – first half of 5th century 0 to the medieval settlements (10th–11th centuries), i.e. by the waste pits and semi-subterranean houses. Layer B, 55–75 cm thick, consists of gray-brown loose soil with traces of fire and large pieces of building rubble. It contains archaeological finds from the 6th century, including fragments of pottery and glass vessels, iron tools and weapons, whetstones, bronze and iron fibulae, two-filed combs of antler and coins of emperors Justinian I and Justin II. The thickness of cultural layer and abundance of diverse finds witness to the intensive life in the settlement at Romuliana during second half of the 6th century (Fig. 156). The archaeological test trenching in 2006 as part of SerbianGerman joint project of investigations of the area outside palace walls to the south of Romuliana fortification brought to light three habitation horizons dating from the middle of the 3rd century to the end of 6th / beginning of 7th century. Second and third horizon date from the period after abandoning construction of Galerius’ palace Felix Romuliana. Second horizon, investigated south of Romuliana fortification, is in fact the Late Roman necropolis15 with cist graves made of tegulae (Fig. 157) and inhumations in burial pits without any structure. The necropolis extended directly from the south rampart to the south and southwest, as it is confirmed by the grave discovered next to the wall foundations (Fig. 158). In this grave was buried an adult male holding high ranking position in the military hierarchy, as it is suggested by the discovery of gilded cruciform fibula decorated with imperial portraits in the niello technique, used to fasten military cloak (paludamentum) on the right shoulder and the bronze buckle of military belt (cingulum militae) (Fig. 159). The small bag placed next to the right shinbone contained besides the military belt also iron steel and flint and seven bronze coins, one of emperor Constans, minted in the middle of the 4th century, and six of emperor Valens, minted between the years 364 and 378. On the basis of these finds the grave is rather precisely dated to the last quarter of the 4th century. 1m the vessel was made on a slow wheel and the upper was made by hand. This jug is so far a unique find and its purpose is not clear. Considering that it has been found inside the kiln, it is possible to relate it with metallurgical activities (Fig. 155). According to the settlement remains from horizon b in the southeast corner of Romuliana fortification, it could be concluded that artisans’ workshops were situated in this area in the second half of the 6th century. They were built of light material (wattle, daub, rammed earth and the like) on the foundations of broken stone and tegulae. Horizon b and corresponding overlying layer B are considerably damaged by the diggings related 15 — Petkovi} 2009, 251–275. 187 SOFIJA PETKOVI] 0 Granary – storehouse from 6th century, extra muros, south of Romuliana fortification FIGURE 161. 188 2m ROMULIANA IN THE TIME AFTER THE PALACE The boundaries of the necropolis in the south and west are not discernible, while in the east the natural border is the bed of Draganov potok (stream) and on the left stream bank were four cist graves investigated earlier. One of the graves belonged to the higher ranking Roman soldier, who was buried with the military belt (cingulum militae), and the bronze coin of emperor Valentinian (364–375) was found on the belt buckle (Fig. 160). This necropolis is dated, according to the type of funerary structures and grave goods, to the last quarter of the 4th – first half of the 5th century. The next horizon in the same area is represented by the Early Byzantine restoration in the 6th century, i.e. by reconstructed economic structure from the end of 3rd – beginning of the 4th century, explored in trench 06/1. The structure is divided by partition wall, and one pit – silo with remains of pithos for storing wheat was dug in one of the rooms. In the other room was encountered the layer of carbonized wheat and one fragmented iron plowshare. This structure is covered by layer A, the layer of destruction with large quantity of building rubble containing pottery fragments and other objects from the 6th century (Fig. 161). It could be concluded that stratigraphy of cultural layers published in the mentioned works of Srejovi} and Jankovi} in 1983 and 1994 is generally confirmed by new investigations in tower 19, in the area of thermae and outside the palace fortification. However, it could be completed with the results of new archaeological excavations in particular concerning the horizons of living after abandoning palace construction. Within this context it is possible to connect certain structures investigated earlier in the fortification interior with the phases of life of Romuliana after abandoning the palace, identified in the course of recent investigations at Gamzigrad.16 The structures originating from phase I (plan XLVIII), period from the last quarter of the 4th century to the middle of the 5th century, include: building 1 – residential structure in adapted hall R and workshop for manufacturing or dying textile in halls L and M of palace I, building 2 – glass workshop south of the “Large temple”, building 3 – glass workshop in the stibadium of palace I (rooms N, O and P), building 4 – adapted hall with apse in palace II and basilica I. The triclinium of palace I (building 6), with floor paved with tegulae, had also been in use in that period. In the first half of the 5th century the room with apse of building 1 was transformed into church and perhaps the bap- tistery of cruciform plan was built in room L of palace I. However, the house built of broken stone bonded with mud, building 5, was added next to the west wall of triclinium in the middle of the 5th century, and inside triclinium was constructed the hut of wattle and mud, building 6. Building 2, glass workshop destroyed by fire was rebuilt in the middle of the 5th century as residential structure. Basilica I was in use during this entire phase, i.e. until the middle/second half of the 5th century. The structures originating from phase II (plan XLVIII), the end of the 5th – beginning of the 7th century, include: early Byzantine building 1 – within the church complex in east section of palace I, small sacred structure with apse and cruciform baptistery from previous phase, buildings 2 and 3 in the southeast corner of palace I, building 4, with the hoard of iron tools, was leaning to the west wall of palace I triclinium, building 6, south of the “Small temple”, building 8, hut – barn, west of the “Large temple”, building 9, storehouse – granary in the restored Late Roman structure 2 and basilica II.17 It is possible to distinguish two horizons in this phase: 1. Earlier horizon, dating from the end of 5th – beginning of the 6th century, includes building 1 with church, atrium and baptistery in the single-aisled structure with apse and the storehouses – granaries, buildings 8 and 9 in the area of “Large temple”. Their date is confirmed by the discovery of monetary hoard under the floor of building 918 and the fact that these structures perished in large fire. 2. Later horizon, dating from the middle of the 6th century, includes buildings 2–4 and 6, that are residential structures, and basilica II with tetraconchal baptistery and most probably also large basilica in the area west of Galerius’ thermae.19 16 — The structures from Late Roman and Early Byzantine period are marked as in Jankovi} 1983, and Jankovi} 1983 A. 17— Buildings 5 and 7 because of ambiguous stratigraphy and the fact that they contained material from the 4th to the 11th century have not been taken into consideration. 18 — Under the floor of rammed earth was found the hoard containing 45 bronze coins from the second half of the 4th–5th century, including coins of emperors Marcian, Leo II, Zenon and Anastasius I (Lalovi} 1983 A, 171, kat. 379). 19 — Dr Dj. Jankovi}, to whom I am grateful for the information, investigated by test trenches the apse of this basilica and established its perimeter. It was concluded that it was three-aisled structure, around 20 meters long and wide, built in the 6th century. These results were confirmed by geophysical prospection in 2007. 189 SOFIJA PETKOVI] Settlement at Romuliana dating from the final quarter of the 4th to the second half of 5th century PLAN XLVIII Settlement at Romuliana from the end of 5th to the end of 6th / beginning of 7th century PLAN XLVII Two building phases at Romuliana, existing in the area of fortified palace from the last quarter of the 4th to the beginning of the 7th century, could be considered from many aspects. Relatively small number of structures, discovered completely or partly, corresponds to the cultural layers abounding in archaeological finds: large quantity of fragments of glass and pottery vessels, coins, metal objects – fibulae, elements of costume and jewelry, iron tools and weapons, many objects of bone and antler. Considering numerous chronologically relevant archaeological objects and the fact that structures had been destroyed by conflagration, the cultural layers are regarded as closed associations between the floor levels and could be rather precisely dated on the basis of stratigraphy and finds. It is also possible to determine the function of certain structures. Archaeological excavations of the 4th–6th century horizons at Romuliana indicate that most of the structures investigated within the fortification are of economic character and just a few are residential structures – houses. The economic structures are storehouses for provisions, granaries and artisans’ workshops. The storehouses for provisions mostly contained cereals. The archaeobotanical analyses, which have been performed, offer the picture of agriculture of the Late Roman Romuliana.20 Main agricultural crop in phase I, from the end of 4th to the middle of the 5th century, was wheat, while the vine was also cultivated. The most popular was bread wheat (Triticum aestivum), while millet (Panicum miliaceum) and barley (Hordeum vulgare) were cultivated to a smaller extent. In contrast to this, many kinds of cereals, leguminosae, vine, fruit, oil seed crop known as false flax (Camelina sativa) and also mangel (Beta vulgaris) and coriander (Coriandrum sativum) were cultivated in phase II, from the end of 5th to the end of the 6th / beginning of the 7th century. There were recorded millet, oats (Avena), einkorn (Triticum monococum), barley and bread wheat, many sorts of leguminosae, including lentil, horse bean (Vicia fabia), grass pea (Lathyrus sativus), chick-pea (Cicer arietinum), peas (Pisum sativum) and many sorts of fruit like vine, pear, nut and peach. It could be concluded that agriculture was less important in the economy of Romuliana in phase I settlement than in the Early Byzantine settlement of phase II. Unfortunately, large amount of animal bones discovered in the layers of phase I and II has not been analyzed so far. 20 — Medovi} 2008, 151–173. 190 ROMULIANA IN THE TIME AFTER THE PALACE Parts of steelyard and weight for weighing grain and other goods (a), with inscription Roustikiou (b), 6th century FIGURE 162. a FIGURE 163. b Bronze fibulae produced in Romuliana, end of 4th – first half / middle of 5th century century with large furnace for smelting iron ore.21 The smithies were encountered in the south tower of west gate, tower 19, in the area of “Large temple” and in the area of thermae in the horizons accumulated from the end of 4th to the middle of 6th century. The preliminary archeometallurgical analyses confirmed for the time being the iron metallurgy, but it is probable that some of the furnaces were used for processing nonferrous metals as well. Nevertheless, it was possible to identify as frequent finds the bones of cattle, pig and sheep/goat, but there was also found considerably amount of wild animal bones – wild boar, doe and deer. It could be assumed that along with the stock breeding hunting was also rather important in the Late Roman Romuliana for providing meat for everyday diet of the inhabitants (Fig. 162). Most frequent among the craftsmen’s workshops are the metallurgical structures, smithies and foundries, and in the southeast corner of fortification is partly discovered artisanmetallurgical complex from the end of 5th / beginning of the 6th 21 — Petkovi}, @ivi} 2006 B, 135–148. 191 SOFIJA PETKOVI] FIGURE 164. Pottery vessels produced at Romuliana from the end of 4th to the end of 6th century The molds for casting belt buckles and other small iron objects (tools and weapons), discovered in the mentioned metallurgical complex, witness to the production of metal objects. Also, distinct variant of bronze fibulae with decorated flat bow and backward turned foot had most probably been produced at Romuliana (Fig. 163).22 Considering other crafts practiced at Romuliana it is certain that there were pottery workshops, despite the fact that they have not been recorded so far in the course of archaeological excavations. This is indicated by large quantity of pottery vessels of local manufacture, characterized by distinct fabric, color and shapes. It should be emphasized that continuity in pottery production from the end of the 4th to the end of the 6th century is evident in the evolution of certain shapes of pots, bowls and jugs (Fig. 164). The production of glass vessels and window panes is confirmed by the discovery of glass workshop from the end of the 4th century in the area south of “Large temple” (building 2) and glass furnace in the structure built at the same time in the stibadium of palace I (building 3).23 The pieces of raw glass have been found in this structure besides the furnace, large number of semi-finished objects, finished pieces and broken glass prepared for remelting. Comparative analyses of the material from this workshop and glass vessels fragments from tower 19 and the thermae area are in progress and we expect it to solve the problem of the local production and import (Fig. 165). There were certainly many workshops for production of objects of bone and antler, as it is suggested by the large quantity of diverse objects, including combs, pyxides, platings, handles, 22 — Petkovi} 2008 C, 463–465. 23 — Jankovi} 1983, 103, sl. 72–77. 192 ROMULIANA IN THE TIME AFTER THE PALACE Glass vessels produced at Romuliana from the end of 4th to the end of 6th century FIGURE 166. Objects of bone and antler produced at Romuliana from the end of 4th to the end of 6th century FIGURE 165. 193 SOFIJA PETKOVI] spindle whorls, pins and the like.24 Also, the large quantity of raw materials and semi-finished products, first of all deer antlers with traces of working, has been found in the horizons from the end of the 4th to the end of the 6th century. Two large pits – deposits of semi-finished antler objects have been found in the area of east gate, but, unfortunately, more precise information about possible workshop are lost, due to activity of the machines in the course of removing building rubble (Fig. 166). The workshop for making and dying textile (Late Roman building 1), existing in the area of palace I in the 4th century, has also been discovered.25 The traces of commercial connections of Late Roman Romuliana are not particularly conspicuous. Foreign influences, which could be observed in the archaeological material, indicate immigration of population rather than highly developed commerce. The most prominent influences in the phase I of the settlement are those from the Chernyahov – Sîntana de FIGURE 167. Mures culture, conspicuous on the objects for personal use: jewelry, combs, fibulae, belt buckles and the like (Fig. 167). As the glass workshops were active at Romuliana at that time, the raw material, consisting of thick fragmented glass slabs produced in the Near East (Syria, Palestine), was imported. The raw material for glass production has been found except in the workshops also in the layers dating from the end of 4th – first half of the4 5th century in the thermae area. The evidence for trade activities in the phase II settlement includes amphoras of the Danubean–Black Sea type for transportation of wine and oil and rather large pots – pithoi. It is interesting to mention that also certain kind of fruits, like figs and dates, was also imported, and their remains were recorded in the layers dating from the end of 5th – beginning of the 7th century. It seems that Romuliana (Romulianum, ~Rwmuli,ana) was from the end of 4th to the end of 6th century rather economi- Objects found at Romuliana with characteristics of “barbarian” Chernyahov – Sîntana de Mures culture, 4th–5th centuries 194 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 168. Bronze fragments of military belts and bronze cruciform “officers” fibulae, final quarter of 4th – first half of 5th century The unit of auxiliary cavalry, equites pseudocomitatenses Timacenses, was, most probably, stationed at Romuliana in the end of 4th – beginning of the 5th century, as it is suggested by the finds of horse harness and combs with three-sided handle, decorated with the horse’s protomes.26 Also, the grave goods from male burials in the Late Roman necropolis extra muros, indicate military population. Particularly significant are the signs of military commanders, fragments of military belts (cingulum militae) and many cruciform fibulae, discovered in the horizon I of the Late Roman Romuliana.27 Also, the projectiles, iron arrowheads (sagittae) and spearheads of short spears reinforced by lead (plumbatae) (Fig. 168) were also found. cally independent and self-sufficient settlement, as a consequence of extensive ruralization. The strong walls of Galerius’ palace Felix Romuliana at Gamzigrad, that were of representative character, in the Late Roman times got military-defensive function, from the last quarter of the 4th to the end of the 6th century. The phase I of the settlement originates from the final decades of the 4th century. It was, most probably, established after the battle of Adrianople, around AD 382, when emperor Theodosius I granted reception (receptio) to the groups of barbarians from the left Danube bank, including Goths, Huns and Alani. The devastated Dacia Ripensis was settled by “barbarian” population on condition to cultivate deserted fields of the imperial domain and to provide specified quota of auxiliary military units, first of all the light cavalry. The evidence for the settling of “barbarian” allies at Romuliana is provided by the finds from tower 19 and from the thermae sector, first of all the combs made of antler with bellshaped handle, bronze and iron fibulae with backward turned foot and iron T-shaped fibulae. 24 — Petkovi} 2003, 35–40; Petkovi} 2006 A, 78–81; Petkovi} 2008 A, 353–366. 25 — Jankovi} 1983, 100, sl. 68. 26 — Petrovi} 1995, 56; Petkovi} 1999, 226–228; Petkovi} 2008 A, 361–363, sl. 2, 1–2, sl. 13–14. 27 — Petkovi} 2008, 395–400. 195 SOFIJA PETKOVI] FIGURE 169. Gold solidi from the hoard deposited in the foundation of mausoleum 1 at Magura, first half of 5th century It is still unknown whether military garrison was stationed at Romuliana already in the time of Valentinian I and Valens, when the restoration of the Danube limes and fortifications in the diocese Dacia had started in AD 364, or after AD 382, in the time of emperor Theodosius I. The graves of military commanders, containing coins, support the former assumption.28 The discovery of the hoard of gold coins, containing 99 solidi, found during the excavations of the mausoleum 1 at Magura, speaks in favor of the military garrison being stationed at Romuliana in the end of the 4th century (Fig. 169). The hoard contains the coins of emperors Gratian, Valentinian II, Theodosius I and Arcadius. It is quite plausible that money from the hoard was deposited in the moment of sudden danger and that it was imperial donatio to the Romuliana garrison from AD 388/389 on the occasion of celebration of decennalia of Theodosius I and five years of reign of his son Arcadius.29 It is difficult to estimate the size of the garrison according to the number of discovered solidi, but it was in any case rather small cavalry unit, detachment of already mentioned equites pseudocomitatenses Timacenses.30 However, the towers and ramparts of the fortification had been reconstructed, as it seems, not before the middle of the 6th century, in the time of emperor Justinian I. It is suggested by the mortar floor of horizon a in tower 19, when it once again became part of the defense system, and by the repairs visible on the curtain walls of south and west rampart. Also, the geophysical prospection and test trench excavations extra muros, between 2004 and 2006, brought to light the defensive moat 28 — In the grave on the bank of Draganov potok one coin of Valentinian I was attached to the buckle of military belt (Jankovi} 1983, 105, sl. 83), while in grave 6/06, buried next to the south rampart (Petkovi} 2009, 266–273), were found, next to the left sheen bone, the remains of a bag containing one coin of Constans, from the midle of the 6th century, and 6 coins of Valens, minted in 367–378 (Vasi} M. 2009, 309–314). 29 — Jovanovi}, Lalovi} 1993, 61–78. 30 — A. Jovanovi} and A. Lalovi} suggest the assumption that it was an imperial donation of 10 solidi for 10 soldiers, amounting to a total number of 100 solidi. However, as only 99 pieces were found, it is certain that this hoard was just a portion of the supposed donation (Jovanovi}, Lalovi} 1994, 182). 196 ROMULIANA IN THE TIME AFTER THE PALACE End of the 4th – begenning of the 5th century 6th century 0 FIGURE 170. 5 10 m Plans of basilica I and basilica II built in south section of palace I along the south rampart that could be dated in the 6th century, on the basis of material discovered in the moat infill.31 The religious aspect of the Late Roman settlement at Gamzigrad from the second half of the 4th to the 6th century is indicated by church buildings discovered at the horizons of phase I and II. Large three-aisled churches, basilica I and basilica II with tetraconchal baptistery, had been built one on top of the other in the south entrance halls of palace I (halls A–D). Basilica I, built in the end of the 4th or in the beginning of the 5th century, had most probably been in use until the restoration, i.e. construction of the new church at the same location.32 Basilica II was built in the 6th century, most probably in the period of Justinian’s restoration of Romuliana (Fig. 170).33 The single-aisled church, west of described basilica, was created by second adaptation of the hall R, within palace I. First adaptation of hall R resulted in creation of Late Roman building 1 with atrium and many connected rooms surrounding it. Second adaptation of hall R included walling up of earlier made entrance in the east wall and building of subselium in the apse, most probably in the beginning of the 5th century. Next to this church was later (in the 5th century) added the baptistery of cruciform plan. Rather small sacred structure with apse 31 — Bülow, Schüler 2009, 231–249. 32 — Jankovi} 1983, 99–100, sl. 66. 33 — ^anak–Medi} 1978, 138; Jankovi} 1983 A, 120–121, sl. 90. 197 SOFIJA PETKOVI] Plan of church buildings in east section of palace I FIGURE 172. Tombstone of Gaudentius, church or military dignitary at Romuliana, end of 4th – beginning of 5th century FIGURE 171. End of the 4th–5th century 6th century 0 5m built of dry masonry and encompassing the mentioned baptistery was built to the south of described basilica in the end of the 5th or in the beginning of the 6th century.34 Another three-aisled basilica, which is partly investigated, was built in the 6th century in the area west of Galerius’ thermae, on top of the “Building with corridor”. Because of the methodology of earlier excavations at Gamzigrad, focused on the investigation of Galerius’ palace Felix Romuliana, the stratigraphic and other archaeological data for precise dating of the described basilicas are lacking. Also, detailed architectural analysis of the buildings and chronological distinguishing of the building phases of the churches at Gamzigrad has not been carried out so far. In any case, seven so far known basilicas at Romuliana dated from the end of 4th to the end of 6th / beginning of the 7th century, with two of them having specially built baptisteries, indicate an intensive Christianization of population. According to the written sources Romuliana was not the Episcopal seat in the mentioned period, but obviously an increased necessity for converting to Christi- anity local and immigrant “barbarian” population resulted in construction of many churches at Gamzigrad. The fragment of funerary stele of certain Gaudentius, dated in the end of the 4th / beginning of the 5th century, also originates from the Late Roman period. This white marble stele, which was erected for the high ranking person in military or church hierarchy of the Late Roman Romuliana (ac positione decorus), has the characteristics of an early Christian epigraphic monument (Fig. 172).35 Also, large number of objects from the phase II settlement of Late Roman Romuliana, from the end of the 5th to the end of 6th century, has Christian symbols, and they are associated with the liturgical practice of that time (Fig. 173). 34 — Jankovi} 1983 A, 122, sl. 91; Ili} 2008, 229, Fig. 5, thinks that this is the baptistery of “basilica II”, single-aisled church built within Late Roman building 1. However, it seems that these two structures are not contemporaneous. 35 — Lalovi} 1983, 163, kat. 335, sl. 166; Petrovi} 1995, 135, No. 113. 198 ROMULIANA IN THE TIME AFTER THE PALACE FIGURE 173. Early Christian objects from Romuliana Both phases of Romuliana from the period after construction of the imperial palace reveal common characteristics apparent in the “barbarization” of population, ruralization of settlements and economy, advanced craftsmanship, pronounced military-defensive aspect and Christianization. The ruralization of settlements is confirmed except by architectural remains of the “houses” built of wattle and daub with floors of rammed earth from the 5th and 6th centuries, also by barns and pits-silos for the storage of agricultural products grown outside the walls. The interments within Romuliana walls in the second half of the 5th century, like in the triclinium of palace 1, within Galerius’ thermae and in tower 19, suggest the decline of the settlement, which was used as refuge (refugium) by the local rural population. The settlement restoration in the 6th century did not change significantly the way of life at Romuliana. Although the traces of metallurgy, even intense iron processing, were discovered in the artisans-metallurgical complex in the southeast part of fortified settlement, local agriculture was the basis of the economy. Romuliana after the imperial palace was the fortification surrounded by country estates (villae rusticae) and villages (vici, pagi), with artisans workshops, storehouses for food and other goods, churches, administration and military garrison of small size, sufficient for the functioning of that system. Such settlements could have survived as independent, isolated unities with minimal commercial connections. They indicate the concept of self-sufficient fortified medieval towns surrounded by the country estates. 199 \OR\E JANKOVI] GAMZIGRAD IN THE MIDDLE AGES Gamzigrad was abandoned sometime in the early years of reign of emperor Heraclius (610–641). The reason was certainly the Avars, who lived in Pannonia from 567 and even endangered Constantinople in 619 and 626. The Avars devastated the Timok region and the lower Danube Basin already in mid-eighties of the 6th century. This resulted in the beginning of Slavic settling in the Byzantine fortifications. The early Slavic pottery mixed with the Byzantine material has been encountered at many locations at Gamzigrad.1 The house of square plan, built of stone, bonded with mud and belonging to the Slavs, was discovered in the very center of the town.2 It could be concluded, according to the hearth made of bricks in the center of the house floor and placed diagonally to the house walls. The hearth had been used for cooking and smoke escaped through the roof opening. The fragments of the Slavic pottery and also the spindle whorl and clay firedog were found next to the hearth. The hearths were not common neither for the Byzantines nor the south Slavic tribes, as both of them used ovens, but they are common features among the Serbs and some other north Slavic tribes. It is not clear whether the Avars in some of their campaigns in the 7th century conquered Gamzigrad or its population escaped to the south. In any case, the ancient Romuliana remained deserted for around three centuries and its name was forgotten. The modern Slavic name Gamzigrad originated in the medieval period. Petar Skok3 thinks that Gamzigrad is an imperative compound word, consisting of the basis “gam” from the Proto-Slavic language, from which the Serbian words “gamziti” and “gmizati” (both meaning to crawl) were derived. Second segment of the compound word “grad” (meaning town or even fortress) also originates from the Proto-Slavic lexis. This name dates from the times when the ancient town was in ruins and the snakes were crawling around. The name could have been given sometime in the 8th or 9th century by the neighboring Slavs, settled here in the late 7th century. When life at Gamzigrad had been restored in the 10th–11th century, it already must have had that name. For its inhabitants the newly established settlement within ancient walls was town (grad in Serbian), as could be understood from the 1 — Jankovi} 1997, 134. T. II. 2 — Jankovi} 1983 A, 125, 127, building 7. name Gamzigrad. The town of that time 3 — Skok 1971, 548. implicates fortified settlement, where was 201 \OR\E JANKOVI] the administrative, religious and military center of the given area. Gamzigrad of the 11th century fulfilled all these preconditions. In the Crni Timok area, as well as in the Beli Timok Basin, there was no other known fortified settlement which could be compared in size with Gamzigrad. The larger in size and more important in east Serbia was only De~ at the location of Roman-Byzantine town Aquis at Prahovo.4 Generally, the towns of the 9th–11th centuries inherited the ramparts of larger fortifications of the Early Byzantine Illyricum. The settling once again within the strong walls of Romuliana was the consequence of historical circumstances in the 10th and 11th century. Namely, the Bulgarian empire in the time of energetic Simeon covered most of the southeast Europe. After migrations of the Hungarians in Pannonia and the Pechenegs in the Danube Valley Bulgaria mostly lost its territories north of the Danube. Further decline of Bulgaria in the time of czar Petar resulted before long in its destruction. It was first devastated by the Russians under knez Svyatoslav and after that it was annexed by Byzantium under the emperor John Cimiskis. The Slavs restored the Bulgarian empire in the time of Samuil and the center of the state was in the area of Ohrid and Prespa. Byzantine emperor Basil II Bulgaroktonos subjugated the empire of Samuil in 1018 and established the Byzantine border along the Danube in the north and in the west along the line connecting mouth of the Drina River and Skoplje. These occupied territories were soon endangered by the Pechenegs, the nomadic people of Turkish descent. They reached in their plundering raids as far as Thessalonica and Constantinople, and for certain period of time they ruled over the right Danube bank in the lower Danube Basin. At that time Gamzigrad, as the fortress on the route from the Danube Valley towards Ni{ and further towards Thessalonica and Constantinople, must have had an important role in protecting population of that part of the Timok Valley. The 11th century was also the time of unsuccessful rebellions of the Slavs against the Byzantine rule. Gamzigrad had ramparts (Roman), some kind of administration, church in the main square, dwellings along the streets and inhabitants engaged in farming and handcrafts. Some of them were by all appearances the soldiers. We know of many such towns within the antique walls.5 The towns of that type investigated in our territory include Veliki Gradac (Taliatae) underneath modern Donji Milanovac, Kladovo – Trajanov Most (Pontes), ]uprija (Horreum Margi), Bela Palanka (Remesiana) and the others. Gamzigrad is considered to be the strongest of them all. These towns were established when the external dangers threatened the population and the state. The same situation was also with Gamzigrad. The strong walls of ancient imperial palace first protected the population of the Timok region during conflicts between Russia and Byzantium for Bulgaria. Then, under the Byzantine rule, the town having certain role in the border region was established. Some small fortifications from the Early Byzantine time had been restored on the right Danube bank in the time of Bulgarian state and when the Hungarian attacks started. Besides the mentioned fortified settlements of defensive character, there were also large cities existing continuously, like Belgrade, Brani~evo or Ni{. They were situated at key strategic positions, what was not the case with Gamzigrad. They were also the centers of large regions or principalities. According to the second charter of Basil II from 1019, issued to the Ohrid archbishopric, the Vidin eparchy was under its jurisdiction. Unfortunately, while in the first charter were mentioned the settlements (towns) within the eparchies, there are no such data in the second charter, which mentioned Vidin.6 Therefore, we do not know with certainty whether Gamzigrad was in the Vidin eparchy or what its position among other towns was. Still, judging by the areas of Ni{ and Brani~evo eparchies, there is no doubt that population of the Timok Valley was included in the Vidin eparchy. Some of the towns in Ni{ and Brani~evo eparchy have been archaeologically confirmed. Brani~evo eparchy had 6 towns and among them Brodarevo was farthest to the east in the Iron Gates and could be identified with Veliki Gradac, underneath modern Donji Milanovac.7 In the north of Ni{ eparchy, including 4 towns, was the town Svrljig, and its area possibly encompassed present day Knja`evac. According to this, we could conclude that most of the Timok Valley with the Danube Valley, as far as the Iron Gates, was not within these two eparchies, so it was within the Vidin eparchy. By comparison with other identified towns of these two eparchies it could be assumed that Gamzigrad was the center of the Crni Timok area and of the section of the Beli Timok Valley around Zaje~ar. 4 — Jankovi} M., Jankovi} \. 1978, 52, sl. VIII. 5 — Jankovi} M., Jankovi} \. 1978, 41–52 6 — Novakovi} 1908, 56. 7 — Jankovi} M. 1981, 63–64. 202 GAMZIGRAD IN THE MIDDLE AGES The restored life at Gamzigrad in the medieval period has been identified in one and the latest archaeological layer overlying the last Early Byzantine layer and it is also partially dug into Late Roman layers even to the floors from the time of Galerius’ palace.8 The medieval entities and layers have been reliably distinguished only after the most recent excavations. The excavations confirmed earlier assumptions that Gamzigrad had been inhabited in the second half of the 10th and in the 11th century. This layer could be divided into earlier and later segment, which also seems to have two phases. Usually we take coins as objects most reliable for dating. There were found three medieval bronze follises. Two earlier specimens date from the coinage of emperor Leo VI (886–913), when the use of Byzantine coins gradually started to spread in Bulgaria. One of them is pierced, i.e. it had been worn on a necklace as a pendant. They could have been deposited during the wars for Bulgaria. The later specimen dates from the reign of emperors Basil II and Constantine VIII (976–1028), when Gamzigrad had already Cross-pendant, bronze, necropolis in front of east gate, grave 7, 11th century FIGURE 174. been within the Byzantine borders. The layer in which it has been found could be dated even after the 11th century. Certain amount of discovered objects is well-dated at other sites, so they make possible more precise dating of the medieval horizon at Gamzigrad. Such are small crosses of the Greek type with circularly emphasized corners (Fig. 174), used as pendants, and similar specimens are characteristic of the 11th century.9 The same situation is with finger rings, which all date mostly from the 11th century.10 The fact that earrings are less frequent when the jewelry is concerned, also indicates the Byzantine period. Particularly interesting is a pair of earrings of the Timok type, consisting of two different specimens. They had been produced in the time after Christianization of Bulgaria in 864 and mostly in the 10th century.11 However, when these Gamzigrad specimens are concerned, they had been in use after the period of their production, i.e. under the Byzantine rule, as is suggested by the fact that they differ from each other. It could be seen also by the pendant of such earring worn on a necklace.12 The 11th century date is also indicated by few fragments of glass bracelets of dark blue color, of which one is twisted and the other decorated with glass of white, red and dark blue color. Similar bracelets, but more diverse, are common find in the settlements from 11th and mostly from the 12th century.13 Just one complete specimen has been found in a grave together with the bracelet of entwined wire. The finds of jewelry do not correspond either to those from the necropoles dating from the time of Bulgarian rule (Ravna near Knja`evac, Grabovica near Brza Palanka) or the time of later Byzantine rule (Ni{ or Veliki Gradac). 8 — About Gamzigrad in 10th–11th centuries cf. Jankovi} 1983 V. 9 — Jankovi} M. 1981, 60, quotes the specimen from Veliki Gradac, and Marjanovi}–Vujovi} 1987, 56, the small cross from the vicinity of Belgrade. 10 — @ivi} 1997, 329–331, presents also Gamzigrad specimens, cf. dating in Manev 1992, 83–94. 11 — Jankovi} M. 1983, 104–105, T. III 2–19, published largest quantity of chance finds of these earrings; S. Ercegovi}-Pavlovi} and D. Mini} investigated the cemetery with these earrings, as well as with more valuable ones with hollow pendants, – (Ercegovi}-Pavlovi}, Mini} 1986, 353, sl. 4. 5 i 5. 8–9, 11 (plain), 4. 7–10 (valuable)). Cast nodes of these more valuable earrings resemble the knobs on the pair of earrings from Gamzigrad. 12 — Jovanovi} C. 1997, 503, T. I 1. 13 — Cf. neighboring finds of glass bracelets from the vicinity of Popovica between Zaje~ar and Negotin, @eravica 1975, or from Lazareva pe}ina near Zlot, Vuksan 1997, 294–295, T. III, that should all be dated to the 12th century. 203 \OR\E JANKOVI] Two phases could be noticed in the pottery assemblage, from the late 10th century and from the 11th century.14 The earlier layer is characteristic by pots with pronounced neck always made on slow wheel. They were decorated with comb, with horizontal lines on the body and with oblique comb incisions on the shoulder, with slanting sheaves and with one or few wavy lines (Fig. 175). The conical bowls are decorated in the same manner. The pottery from this layer is usually preserved in small fragments. The pottery from the later layer is much more abundant and more diverse. The pots from the later layer are of various shapes. The pots of smaller or larger size, with stout body, high shoulder and narrow base and made on faster wheel are assumed to be the products of one potter. The rim is everted and emphasized by the side groove. They are decorated with single wavy line on the neck and with wavy lines and horizontal lines executed by comb on the shoulder and body. Considerably smaller quantity of pots, also made on faster wheel but in a different way, were the products of another potter. They are of rather elongated shape, with shorter shoulder and simple everted rim, and mostly of brown color. They are decorated with horizontal lines and series of notches. Most of other pots, made probably in some village workshops, are decorated with horizontal lines and sometimes with another ornament, mostly multiple wavy line and notches above the lines. Most frequent among other pottery shapes are the bowls. Those of local origin are always conical, with ring-shaped base or with plain base. Few bowls, made on potter’s wheel and with inverted rim and simple decoration, had been made in the workshop of some big city. The pottery for everyday use included also the cups. The pots having two small handles and big pots with vertical rim were very rarely used (Fig. 176). Most frequent of the amphoroid jugs are those of brown color and decorated by polishing. The fragments of amphoroid jugs decorated with shallow ribs and orange glaze are very rare. Particularly important are group finds, which bear witness to certain violent events and could be associated with the data Cooking pot decorated with wavy lines, large temple sector, 10th century FIGURE 176. Pot with vertical rim for storing provisions, 11th century FIGURE 175. 204 GAMZIGRAD IN THE MIDDLE AGES from the written sources and thus precisely dated. Two blacksmith’s hoards have been found – in one of them were blacksmith’s tools and in the other complete or broken iron objects. These hoards indicate that Gamzigrad had been conquered immediately after hiding of these hoards. Their owner, the blacksmith who had hidden them, was not able to retrieve them, because he was probable taken prisoner, killed or exiled. Two swords found in ambiguous circumstances provide also the evidence for the conquest of Gamzigrad. By all appearances they were also buried or hidden in some other way as it is not common just to abandon such weapons. The swords and other weapons were valuable booty at that time and the enemy would have not left it. All available data indicate that Gamzigrad had been abandoned after the rebellion against Constantinople in 1072, at the same time as the neighboring Danube towns.15 In the time when Gamzigrad had been restored and resettled, the walls of ancient Romuliana were in very good state of preservation, certainly much better than they are today. It was necessary just to repair the damages inflicted when the Avars conquered Romuliana and damages resulting from the ravages of time. These included the battlements with arrow-holes on the ramparts and towers. It was also necessary to repair wooden storey structures in the towers, but there is still no evidence for that. It has not been established whether the tower roofs had been repaired. The necropolis in front of the east gate and layers on the outside of ramparts reveal that they crumbled and collapsed after the medieval period. The old west gate was walled up, but there was certainly some kind of passage. The main entrance to the town was the east one, through lower gate, facing Draganov potok. In comparison with other towns of the same date also enclosed within antique walls, Gamzigrad was of exceptional strength. The newly built Bulgarian and even Byzantine towns like Belgrade and Skopje had less substantial ramparts and towers. Nevertheless, as a fortress which should have special tasks, Gamzigrad had certain shortcomings. It was not located directly on the communication line and did not prevent passing along the route, so the enemy was not forced to conquer it in order to penetrate inside the country. It was just used as spacious already existing fortified refuge for establishing the administrative center of the region. So, Gamzigrad must have been the seat of `upan, i.e. of the archont. The newly established town understandably retained old, inherited spatial organization. The old buildings, as well as ancient passages, had been used. The main street, running through Gamzigrad in the east–west direction, also remained in use. The cathedral church, built on the palace ruins, was erected within newly created square. The ruins of large temple on the other side of the square must have also been used for some important purpose, but there are no archaeological data about that. All the surrounding buildings and open areas were used for residential structures or workshops. The distribution of individual burials within the town indicates the areas without residential or economic structures, i.e. the courtyards and gardens. The most prominent place in the settlement was also occupied by the church. The location of ancient Late Roman basilica, incorporated once in the throne hall of Galerius’ palace, was used to build the new basilica with baptistery of the quatrefoil plan.16 Only the church foundations are discovered that supported aboveground section, which is not preserved. The church, which was over 31 m long, seems monumental. It was three-aisled basilica with one large central apse, semicircular on the inside and pentagonal on the outside. The deeply set foundations of side walls and stylobate could have supported the galleries above the side aisles. Not a single stone ornament of this church, otherwise common for such big churches of the 10th–11th centuries, has been preserved. There were not even found the antique columns with bases and capitals, that were, very probably, used. This cathedral church was situated in the square in the upper, northwest section of the town and rather small cemetery was established around it. For the time being it remains unclear whether this church had been completed (as the stone decoration is lacking), or whether there was an earlier church (as the cemetery suggests the existence of the temple). It is certain that liturgies took place in the Gamzigrad church, as is confirmed by broken procession cross with haft for mounting 14 — Similar situation with two layers, dated to the 10th and 11th centuries, was encountered at Veliki Gradac, although the pottery repertoire there is much more modest – Jankovi} M. 1981, 47–55. It is worth mentioning here that pronounced differences between contemporaneous pottery from distinct towns is the consequence of the activity of town pottery workshops and not of different dating. 15 — Jankovi} M., Jankovi} \. 1997, 52–55. 16 — Sometime ago I presented this church as the Early Byzantine structure whose building could not have started before the eighties of the 6th century and emphasized that it was certainly in use in the 11th century (Jankovi} 1983 A, 120–121). 205 \OR\E JANKOVI] FIGURE 177. Liturgy cross, bronze, decorated with incisions and engravings, 11th century on the wooden handle (Fig. 177).17 Similar basilicas with triangular or pentagonal apses on the outside18 have been discovered in Bulgarian capitals Pliska and Preslav, as well as in Prizren and other towns of the 9th–11th centuries. They also have pairs of pilasters on the west façades, as it was the case with Gamzigrad basilica. However, the baptistery structure leaning on the church has not been identified elsewhere, although there are baptistery basins of the similar ground plan. The nearby Zanjeva~ka church from the 14th–15th century,19 that also has an earlier phase,20 resembles in plan the Gamzigrad baptistery. Two building phases have been identified at Gamzigrad medieval settlement. It seems that habitations more or less deeply (0.2–1 m) dug into the ground are characteristic of the earlier settlement, dating from the 10th century. They are rather small, of square plan and 4 x 4 meters in size. In one of the corners was the oven, so-called kamenka, made of layers of stones (kamen meaning stone in Serbian), of rectangular shape and with an opening on the narrow side. It was also covered with stones, but in such a way to leave openings for heating the cooking vessels. The aboveground structure of the semi dug-in houses was of wood sometimes plastered with daub. Such habitations were sometimes adapted to the already existing old walls. Thus narrow crypt under the pronaos of small temple was used as a dwelling place and the naos was also used in a similar way. The layer of soot and ash was discovered at the depth of 1 meter and was covering entire room under the pronaos. The stone oven with south facing opening was found under that layer, next to the west wall. The oven was made of stone slabs and broken stone and it is separated from other part of the room by two slabs, 2.2 m long in total. There was found small quantity of pottery, including fragments of the jug deformed in fire and vessel with vertical rim and horizontal rib on the inside of the neck and without decoration. There were also found two spindle whorls and bronze plating. The pottery dating from the earlier period, from the end of the 10th century, has been discovered in the naos. The semi-dug in habitation with considerable quantity of finds has been encountered to the north of small temple. As it is the case with other semi-subterranean houses outside the ancient walls, it was difficult to establish its outline. Judging by the area covered with archaeological objects, it could be assumed that it was of square plan, with sides around 3.5 meters long. The oven constructed of broken stone, marble slab and sandstone ashlars, was discovered in the northeast corner. Its dimensions are 1 x 0.9 x 0.5 m. The fragments of pots of various size, bowls without pronounced rim and fragments of polished amphoras have been found in this house. The later settlement is characterized by aboveground habitations of diverse size. They had walls of wood or stone, bonded with clay, and walls of the older building were also often used. In the course of current archaeological excavations conducted in the southeast corner of Gamzigrad, the floors of above- 17 — This cross was earlier wrongly dated in the Early Byzantine period, cf. Jankovi} \. 1983 A, 135–136, br. 182. 18 — Deep, three-sided on the outside, apse of large 9th century basilica in Pliska (Mihailov 1993), must have been five-sided above the pastophoria. S. Nenadovi} reconstructs the apse of earlier basilica of Bogorodica Ljevi{ka as the three-sided (Nenadovi} 1963, T. 3–9), but it could have been fivesided as today. 19 — Stri~evi}, Suboti} 1959. 20 — Jankovi} 2007, 146. 206 GAMZIGRAD IN THE MIDDLE AGES Button of silver foil, necropolis in front of east gate, 11th century FIGURE 179. Earring cast of bronze, 10th century FIGURE 178. FIGURE 180. Earring cast of silver, large temple sector, 10th–11th centuries FIGURE 181. Earring cast of silver in openwork technique, area of thermae, 10th–11th centuries ground structures were discovered, oriented in different directions in relation both to each other and to the antique structures. This indicates that remains of aboveground walls of some antique structures had been completely demolished and that they had no impact on orientation and position of new structures. Different orientation of medieval buildings suggests the changes in the directions of streets, i.e. it implies thorough reconstruction and rearrangement of the settlement. It could not be established what were the reasons for different orientation in relation to the older walls and at least one street. The uniform distribution of the 11th century pottery finds throughout entire Gamzigrad indicates that most of the 11th century buildings were aboveground structures with wooden walls. Some of them might even have an upper storey. Approximately in the middle of north hall of palace I the square area covered with bricks was discovered at the depth of around 0.35 m. This area is the floor of a room 2.4 x 2.3 in size. It was perhaps the only room or just one of the rooms of a house with walls made of logs, planks or boards. The pottery from the 11th century was discovered in the first arbitrary layer, corresponding to the layer above that floor of bricks. The fact that cemetery has been discovered next to the west wall of north hall indicates special purpose of this structure. Other aboveground structures were of much larger size. Rather large section of the floor of some building has been discovered near the thermae in the southeast corner of Gamzigrad. It seems that building was of square plan and around 7 x 7 m in size. There was circular hearth, 1.5 m in diameter, in the south corner. The walls of this building were without doubt made of planks and logs. The walls of some buildings were made of layers of stone once probably bonded with mud mortar, as it was also common in the early Byzantine time. One of such buildings with stone walls was leaning to the west wall of north atrium of palace I, i.e. the east wall of north hall was separating this building from the cemetery. The south and partially also east wall of this building, around 4 x 4 m in size, are preserved. The walls were built of stone rubble and few marble spolia, but bonding agent is not preserved. The discovered pottery dates from the 11th century. It seems that hall with apse of palace II had been used in its full capacity. The west door were sealed by stone wall preserved up to the height of 0.65 m, while northwest section of the hall was separated by another stone wall, 8 meters long and running in the west–east direction. The bonding material of these walls is not preserved. The transversal wall, supposedly closing this room, has not been identified. The east entrance to the hall remained in use. The pottery and other finds dating from the 11th century were discovered in the smaller room (separated by the wall) and in west section of the hall, where also the floor of rammed earth has been encountered. The partition wall of the west–east direction suggests the existence of an upper storey. In that case west section of the hall was covered with gable roof of north-south direction. It means that house with upper storey and attic was located in the west section of antique hall, while in the east section was courtyard with the gate in the east wall. The new still not completed 207 \OR\E JANKOVI] Bronze finger ring with Greek inscription, east gate, 11th century FIGURE 183. Bronze finger ring with bird representation, 11th century FIGURE 182. investigations indicate that there were also detached buildings built of stone. Gamzigrad is unique, considering the unusual distribution of graves. Namely, except at least three cemeteries, numerous burials have also been discovered at different places within the town. The main town cemetery was located immediately in front of the east gate. Over 100 graves were discovered on both sides of the road.21 The dead had been buried according to the Christian ritual, in supine position with head to the west and arms or hands crossed. They were buried in their traditional dress. This is most conspicuous in the burials of women buried in dresses buttoned up under the neck. The pair of light blue glass buttons decorated with white paste was found in one grave and one silver button was discovered in the other burial (Fig. 178). The jewelry, usually including earrings, strings of glass beads, finger rings and rarely bracelets, was discovered in other female graves. The earrings at Gamzigrad are not as frequent as in the 10th century cemeteries. Besides already mentioned earrings of the Timok type, the earrings with grapelike pendant are even of an earlier date (Fig. 179). The earrings with four knobs, of which the bottom one is larger and central smaller (Fig. 180), or with pinecone-like pendant instead of the bottom knob (Fig. 181) are throughout the southeast Europe dated to the 10th–11th centuries. They were all cast in two-piece bronze molds. The discovered specimens, except the Timok ones, which are indubitably earlier, belong to the group of grape-like earrings, or to those with four knobs cast in two-piece molds. Similar specimens are dated to the 10th–11th centuries throughout the southeast Europe. The variegated glass beads used for strings are of the same date. The most of finger rings are quite simple and of extensive date. These are strap-like rings, decorated with engraved angular line, and rings with rhomboid head with two “small leaves” on each shoulder. They are dated from the 9th to the 11th century, as well as few finger rings with circular head with engraved Greek inscription (Fig. 182), pentagram (Fig. 183) or bird. The bracelets, usually frequent finds in the 12th–13th century graves in east Serbia, are rather exceptional finds at Gamzigrad, as we mentioned earlier. Some of the graves were found next to the north tower of east gate. They did not have any grave goods, so it is not clear whether they should be included in the same cemetery. There were two cemeteries next to the cathedral church. First group of graves was found north of the church, in the course of palace investigation. In nine graves discovered in 1963, the dead with arms crossed on the chest or pelvis and with head to the east had been buried without grave goods. These graves perhaps belong to the northwest cemetery (in relation to the church), with fourteen graves discovered in 1970, 1971 and 1973. All individuals were buried in supine position, with head to the west and arms crossed on the chest or pelvis. The grave structure consisted of stone or brick placed 21 — Jovanovi} S. 1997, 503–509; Jovanovi} S. 2000, 203–204. 208 GAMZIGRAD IN THE MIDDLE AGES FIGURE 184. Reliquary cross, bronze, th 11 century, found in church area in 1954 on edge above the head, or by the feet or the brick under the head. Next to the skull of one skeleton was discovered the skull of a horse (accidentally?). The grave encircled and covered with bricks and stone, containing the deceased with arms crossed on the chest and with reliquary cross on the chest, was discovered around 30 meters north of the church (Fig. 184). The other group, consisting of nine graves, was discovered to the northeast of the church. In some of the graves the dead were in supine position, but in three graves they were in crouched position. The arms of the dead in supine position were in same position as in the graves of first group. The skeletons were sometimes surrounded by bricks and one crouched skeleton was covered with bricks. The crouched skeletons in a double burial were covered with two rows of bricks. In two graves the knives were found, indicating that the deceased had been buried in traditional dress, with belt from which sheath for the knife was suspended. It is most obvious in the graves of women buried in dresses buttoned at the neckline. The pair of light blue glass buttons decorated with white paste was found in one grave and in the other one silver button was discovered. Around ten more graves have been encountered at different places at Gamzigrad, around large temple, in towers and within other investigated areas. These graves are buried to the depth of around 1 m in the early Byzantine and Late Roman layer and are mostly without grave goods. One of these burials is worth mentioning, as it had grave goods. The deceased person was in supine position with arms crossed on the chest and grave structure consisted of bricks placed on edge by the head and feet. The finger ring with rhomboid head and small pot, most probably for the food offering to the dead person, has been encountered. This is the example of differences existing between those buried around the church and the majority of Gamzigrad inhabitants. All mentioned cemeteries and graves provide important information about Gamzigrad. First, anthropological analysis of skeletons discovered in front of east Gamzigrad gate revealed that besides the Slavs there were interred also the members of the Mongol race.22 Second, the social differences are also apparent. In the cemetery in front of east gate had been buried common inhabitants of Gamzigrad, in traditional dress and women and girls with jewelry. In the graves next to the church the jewelry had not been found. This, as well as the cemetery location in the center of town around cathedral church, suggests that this was the burial ground for the town elite. There had been buried the priests and their families, governor of the town, `upan or strategos and his family and probably some other respectable citizens of Gamzigrad. There was perhaps even certain division between those buried to the north and those buried to the east of the church. Third, rather unusual are single graves outside these two cemeteries, in the settlement, as well as many crouched burials discovered within the ramparts, in the vicinity of the church and in the settlement. Such individual interments were not possible under regular circumstances, when the dead had been buried in cemeteries and not in the yards. The only explanation is the siege of Gamzigrad, when killed or deceased could have not been buried in the cemetery outside town walls. The crouched position of the dead could be the indication of death caused by disease (pestilence) or starvation. 22 — Miki} 2009, 115–121. 209 \OR\E JANKOVI] FIGURE 185. Hoard of blacksmith’s tools, tongs and hammers, iron, 11th century Main occupation of most of the inhabitants was agriculture, as it was the case in all rather small medieval towns. The inhabitants of Gamzigrad had been farming the land, sawing cereals, growing vegetables and fruit and raising cattle and smaller animals. This is confirmed by the iron objects like sickles, scythe, plowshare, discovered in the blacksmith’s hoard. Among the sickles is one broken and repaired piece, indicating that new sickles were expensive and not easily obtainable. Also, there are snaffle bits combined of different elements. Because of high value of iron, the blacksmiths gathered broken and damaged iron tools. That hoard contained 19 objects in total. There were found, besides the farming tools and plain knives, also fragments of scissors. One fragment is the handle of twopiece scissors with pivot, i.e. so-called tailor’s scissors, and other fragment is the blade of smaller one-piece scissors, probably used for shearing sheep or cutting leather. Three different snaffle bits and saber hand-guard could be associated with cavalrymen. Another blacksmith’s hoard contained 9 different tongs, four hammers and one anvil (Fig. 185). The pottery workshops have not been discovered, but two, dating from the later period, could be identified on the basis of their products. In both of them the hand wheel had been used with great skill. The pottery had also been produced in individual households as additional activity, also on turntable, but it was of rather crude manufacture. Most of earlier pottery, first of all pots, were produced in the individual households. Some of the vessels had been acquired from other regions. These include single or double-handled jugs, glazed and polished,23 bowls, some pots of distinct shape, but also simple specimens, recognizable by the potter’s stamp on the base. These vessels had been bought in some bigger town, or from the potters who brought their goods to the markets or to some shop in Gamzigrad. Some of the jugs, first of all those decorated by polishing, originate from the lower Danube Valley (Prahovo, Vidin or some even more distant town). One distinctive amphoroid jug, decorated with ribs, stamps and polishing (Fig. 186), comes, by all appearances, from the Velika Morava Valley or from Ni{.24 Many other objects, including combs and plating of bone and antler, bronze jewelry and other bronze objects, indicate the activities of other craftsmen or tradesmen. Of course, many things had been produced within the household. Many spindle whorls indicate, for instance, that wool spinning was common and everyday occupation. One of rare witnesses of writing is the discovered bronze writing implement. It is a staff pointed at one end and shaped as 23 — Biki} 1997, 319–328. 24 — In the Regional Museum in Para}in I have seen few fragments of similar jugs decorated with stamps. I wish to thank M. Brmboli} for his help. 210 GAMZIGRAD IN THE MIDDLE AGES Two-handled jug found on the south side of the church, 11th century FIGURE 187. Tongue from the belt, bronze, east gate, 11th century FIGURE 186. could have been engraved before the firing of the vessel.25 Perhaps this jug had been used for collecting the taxes in wine.26 There is conspicuously large number of snaffle bits discovered not only in the blacksmith’s hoard, of stirrups and other metal parts of horse harness, all indicating horse riding. Some of these cavalrymen were armed with swords and other with sabers, as suggested by the hand-guard in the blacksmith’s hoard. Therefore, it could be assumed that the cavalry unit was stationed at Gamzigrad and that in that unit were the soldiers with weapons typical for the steppic regions. The bronze tongues (Fig. 187), plates and buckles from the leather belts also belonged to the soldiers. The importance of Gamzigrad as fortified place became evident already in the first half of the 11th century. Namely, the nomadic tribe of the Pechenegs settled on the left Danube bank and started to attack the Byzantine territories. Thanks to their cavalry, they were very mobile, and, according to records of John Scilices, they reached as far as Thessalonica in 1034 and 1035/36.27 As they had to pass through the Timok Valley, they might have endangered also Gamzigrad. The same route also used sometime later the Uzes who reached in 1064 as far as triangle at the other. It was used for writing on the wax-coated wooden tablets. The text was inscribed using sharp point and the triangular end was used for erasing. This way of writing on the tablets was used for short notes, accounting and learning, as it was not permanent. It could be surmised that priest was the person who taught children to write. The evidence of archives with the documents written on parchment, seals from the letters or metal plating of the book covers have not been found. There is no doubt that there was some kind of written administration. The graffiti were encountered on the single or double-handled jugs, which were decorated by polishing, because their smooth surface was suitable for engraving. On some of them are simple patterns or signs. But, on the mentioned amphoroid jug decorated with stamps, there is an engraved six-letter word near the base. Unfortunately, the language and the alphabet have not been identified. There is an impression that signs or inscription on the shoulder of jug decorated with ribs and by polishing 25 — Biki} 1997, 322–333, sl. 1. 3. 26 — The same inscription is on jug from ^e~an, about which, as well as about other marks on jugs, had written Tomovi} 1991; cf. Jankovi} 2007, 214. 27 — VIINJ III 1966, 163. 211 \OR\E JANKOVI] Macedonia and Greece, as it is recorded by Scilices’ Successor.28 It seems that Gamzigrad did not experience the attacks of the nomadic tribes, as in the course of archaeological excavations the layers or finds that could be related with certainty to the conquest of Gamzigrad at that time and with the conflicts have not been encountered. For example, if Gamzigrad experienced the attack, there must have been found the nomadic arrow-heads. It seems that rebellions of the Middle Balkan Slavs were crucial for the fate of Gamzigrad. The first rebellion, which is most exhaustively described by John Scilices,29 broke out in the towns Belgrade and Morava in 1040 and spread to the south. It is not known whether the towns to the west of Vidin took part in that rebellion, which was crushed already in 1041. After the peaceful period of around thirty years, the Middle Balkan Slavs, oppressed by high taxes and striving for independence, had begun the rebellion once again in 1072. The rebels asked for help the Serbian state, so Mihailo, prince of Zeta (around 1052–1081), sent them his son Konstantin Bodin, who was proclaimed Bulgarian emperor in Prizren. The rebels liberated Skopje and Ni{ and reached to the south as far as Kastoria. The information by Nicephoros Vrienius that towns along the Danube between Belgrade and Vidin suffered great losses also dates from that period.30 It is assumed that these towns also took part in the rebellion, although there are opinions that they were attacked by the Hungarians, who at the same time conquered Belgrade and some other towns. The life at Prahovo, Korbovo, Tekija and Veliki Gradac died out at that time.31 The fact that most of these towns had not been restored suggests that towns between the Timok and the Morava took part in the rebellion, because, if they had been destroyed by the foreign enemy, they would have been restored. In order to prevent future uprisings of unreliable population, the imperial government probably evacuated their towns or they just escaped in the regions of Hungary. After that, all towns upstream of the Iron Gates have been restored as necessary for frontier defense against the neighboring Hungary. Downstream of the Iron Gates and in east Serbia was restored only Kladovo (fortification Pontes), also because of Hungary, which had its southernmost strongpoint at present day Turnu Severin, on the other side of the Trajan’s bridge. The described archaeological finds from Gamzigrad reveal that life was abruptly interrupted, that it underwent tiring siege and that it was conquered either by surrender or by assault. The abandonment of Gamzigrad did not mean the end of its medieval history. There were found one bowl decorated with sgraffito technique (the kitchenware of the same date was not found), three finger rings and fragment of a stirrup, dating from the end of 14th – beginning of the 15th century. Therefore, it could be assumed that ruins of Gamzigrad had been used as the place of refuge in the time of Turkish conquest. 28 — VIINJ III 1966, 175–176. 29 — VIINJ III 1966, 141–155. 30 — VIINJ III 1966, 237–239 31 — These events are more thoroughly investigated on the basis of archaeological evidence by Jankovi} M., Jankovi} \. 1978, 41–58. 212 ALEKSANDAR V. POPOVI] WRITTEN SOURCES ON GALERIUS We know about Emperor Galerius primarily from the accounts of two of his contemporaries, Lactantius and Eusebius. Although they belong to different cultural milieus, Lactantius to the Latin cultural circle and Eusebius to the Greek one, they are connected by mutual Christian faith and pronounced negative attitude towards Galerius. Because of the open enmity which they show for this emperor, we can suspect, and with reason, the objectivity of their accounts. Other sources which mention Galerius date from later times, so the time distance raises the question of their credibility. All in all, we may conclude that our knowledge about Galerius is not quite reliable. Even if we accept that everything we can read in the sources is correct and reliable, there are no many concrete data, so we must add that our knowledge is not only unreliable, but also rather scarce. According to the reports of antique and medieval writers, the biography of Galerius could be presented in brief as follows: Galerius (Caius Galerius Valerius Maximianus) was born in the family of peasants, in the vicinity of Serdica, around AD 250. Because of such background, malicious contemporaries gave him the nickname Armentarius – Herdsman. He started his military career under emperors Aurelian and Probus and he was promoted to the higher officer rank under Diocletian. In spring of AD 293 in Nicomedia Diocletian proclaimed him Caesar. He divorced his first wife, with whom he had a daughter, in order to marry Diocletian’s daughter Valeria. He was in charge of the Danube limes, so he had chosen Sirmium as the center of military operations. Following the Diocletian’s orders, he moved to Syria in AD 296, in order to wage war against the Persians. After the initial failures and the additional recruitment of soldiers in Illyricum, he defeated the Persian king Narseus in AD 298, and Romans made favorable peace with Persia.Galerius was named Persicus and celebrated his triumph in Antioch. This victory is commemorated on Galerius’ triumphal arch in Thessalonica, where he had the official residence. In the winter of AD 303 he started the great persecution of the Christians, no doubt in agreement with Diocletian. His triumph over the Persians was officially celebrated in Rome that very year, but as a part of celebration of vicennalia and decennalia of all tetrarchs, so his victory was treated equal to the victories of his co-rulers. Because of that, he considered the triumph in Antioch more important, and it had an impact on his ruling ideology and propaganda. He was proclaimed Augustus on the 1st of May 305, and after that 213 ALEKSANDAR V. POPOVI] Triumphal arch of Galerius, Thessalonica, AD 298/299–303 FIGURE 188. he started the long-lasting conflict with other tetrarchs. He became terminally ill in AD 310. He issued the edict of tolerance for the Christians on the 30th of April 311 in Nicomedia. He died few days later, probably in Serdica, and, according to PseudoAurelius Victor, he was buried at his birthplace Romulianum, although his mausoleum had been prepared in Thessalonica.1 From the historical sources about Galerius we selected and translated those passages, which could be relevant for the questions posed by investigations at Gamzigrad. Because of that, the reader would not found here, for example, the passages from Eusebius. We also provided original text along with the translation. The texts are accompanied with essential notes and short biographies of the authors. Before every passage the year associated with the given text is mentioned, when it was possible to determine. and oratorical skill that, despite he was a Christian emperor, Diocletian called him to teach rhetoric in one of the imperial capitals, Nicomedia. He maintained his popularity also at the court of Constantine, so sometime around AD 317 he became personal tutor of his son Crispus. Contemporaries gave him credit by giving him the nickname “Christian Cicero”. He earned such name because of his exceptional connection of pagan education and Christian spirituality. Lactantius’ most important work is The Divine Institutions (Divinae institutions) in seven volumes. His other preserved books include On the God’s Creation (De opificio Dei), On the Wrath of God (De ira Dei) and On the Death of Persecutors (De mortibus persecutorum). His biographer Hieronymus mentions also some other works, nowadays lost. Most interesting for us is certainly his work On the Death of Persecutors, one of the most important historical sources for the period of tetrarchy. There are certain doubts that Lactantius is its true author, but it is today generally accepted that he is the author of that text. It is assumed that it had been written in AD 314–315. The devoted Christian Lactantius speaks in this text with lots of hate and partiality about the emperors who persecuted Christians. The death of each one of them is central topic of the text and it is described in detail and explained as God’s punishment. The work is generally concerned with the period of tetrarchy, i.e. the years between AD 303 and AD 313. It was dedicated to certain Donatus the confessor.2 It is considered that when writing this text Lactantius took as a model Second Book of Macabeans. This work is preserved in a single manuscript from the 11th century with lots of lacunae and rather damaged. The manuscript was discovered in 1678 in the Benedictine Abbey of Moissac, and is housed today in the National Library in Paris (Colbertinus, BN 2627). Edition used: Lactantius, De mortibus persecutorum, edited and translated by J.L. Creed, Oxford 1984. LACTANTIUS Lucius Caecilius Firmianus Lactantius, or, according to some sources, Caelius Firmianus Lactantius (c. 250 – c. 325) was, according to the unanimous opinion, one of the most important early Christian writers. Being from the North African descent, he acquired his education as rhetor from his famous countryman, Christian orator Arnobius. He gained such fame for his education 1 — For basic data about Galerius, with the additional bibliography, see PLRE I, 574–575, s.v. C. Galerius Valerius Maximianus 9; Ensslin 1930; Stein 1968, 65–93; Seston 1946. 2 — In church terminology the word confessor does not denote just the priest administering the holy secret of confession, but also the believer who suffered persecution and torture, but survived, in contrast to the martyr. 214 WRITTEN SOURCES ON GALERIUS De mortibus persecutorum On the Death of Persecutors cotidie ac vicanis suis epulas exhibebat. Christiani abstinebant, et illa cum gentibus epulante ieiuniis hi et orationibus insistebant. Hinc concepit odium adversus eos ac filium suum non minus superstitiosum querelis muliebribus ad tollendos homines incitavit. IX, 1–4 Alter vero Maximianus, quem sibi generum Diocletianus asciverat, non his duobus tantum quos tempora nostra senserunt, sed omnibus qui fuerunt malis peior. Inerat huic bestiae naturalis barbaries, efferitas a Romano sanguine aliena; non mirum, cum mater eius Transdanuviana infestantibus Carpis in Daciam novam traiecto amne confugerat. Erat etiam corpus moribus congruens, status celsus, caro ingens et in horrendam magnitudinem diffusa et inflata. Denique et verbis et actibus et aspectu terrori omnibus et formidini fuit. His mother venerated the mountain deities 8 and was a very superstitious woman. When she was…9 she almost daily organized sacrificial feasts, where she hosted her neighbors.10 The Christians abstained from them, and while she celebrated with the pagans, they persevered with fasting and praying. That is how her hate for them started, and as cantankerous women do, she prompted her son, no less superstitious, to destroy these people.11 And the other Maximian, whom Diocletian took for his son-in-law,3 was not only worse than the other two whom our time experienced,4 but than all evildoers who had ever lived. In him existed the natural savagery of the beast, rudeness foreign to the Roman blood. It is small wonder, as his mother was from the other side of the Danube and she escaped, by crossing the river, to Dacia Nova (New Dacia),5 facing the attack of the Carpi.6 His body was also in accordance with his character, tall and meaty and grown and bloated to a terrifying size. Finally, by his words, deeds and looks he inspired fear and terror in everyone. (AD 305) XX, 3–5 Habebat ipse Licinium veteris contubernii amicum et a prima militia familiarem, cuius consiliis ad omnia regenda utebatur; sed eum Caesarem facere noluit, ne filium nominaret, ut postea in Constantii locum noncuparet Augustum atque fratrem, tunc vero ipse principatum teneret ac pro arbitrio suo debacchatus in orbem terrae vicennalia celebraret, ac substituto Caesare filio suo, qui tunc erat novennis, et ipse deponeret, ita (AD 298) IX, 9 Exinde insolentissime agere coepit, ut ex Marte se procreatum et videri et dici vellet tamquam alterum Romulum maluitque Romulam matrem stupro infamare, ut ipse diis oriundus videretur. 3 — Galerius married Diocletian’s daughter Valeria. 4 — That is Maximian Herculius and Diocletian. 5 — New Dacia (Dacia Nova) or Dacia Ripensis was established on the right Danube bank after Aurelian abandoned Dacia in AD 272. 6 — Carpians were the tribe of Gothic–Dacian origin. 7 — Since the victory over Persians in 298. Since that time he started to behave very haughty,7 so he wanted to look like and to be talked about as he is the Mars’ descendant, as he is another Romulus, and he preferred his mother Romula to be disgraced as adulteress, only to make himself look like being the offspring of the gods. 8 — Probably the vegetative deities like Silvanus, Diana, Liber Pater. These cults are confirmed in the inscriptions from Moesia and Dacia. 9 — Lacuna in the text. 10 — Vicani – neighbors, would literally denote the inhabitants of the same village (vicus). 11 — According to this, we may conclude that mother prompted Galerius to start the great persecution of Christians in 302/3. However, it is hardly possible to use this fragment as a convincing argument that she was alive when the persecution started. Period of which Lactantius is speaking is rather vague. Question could be raised whether Galerius himself could have made the decision about persecution, as he was Caesar at that time, or it was the ingerency of Diocletian as Augustus. (before AD 303) XI, 1–2 Erat mater eius deorum montium cultrix, mulier admodum superstitiosa. Quae cum esset …, dapibus sacrificabat paene 215 ALEKSANDAR V. POPOVI] morem nefarius homo in Romanam terram voluit inducere, quem ex illo tempore victoriae sine pudore laudabat. Et quia id aperte iubere non poterat, sic agebat, ut et ipse libertatem hominibus auferret. So, after he seized the supreme power, he turned to abusing the world he conquered. Namely, after he defeated the Persians who have such custom and tradition to give themselves up as slaves to their kings and kings to treat their people as their home slaves, this evildoer wanted to introduce this custom, which he shamelessly praised since the time of his victory, into the land of Romans. And as he could not openly order this, he behaved in a way to permanently arrest people.13 Triumphal arch and mausoleum (?) of Galerius, (Rotonda), Thessalonica (AD 311) cum imperii summam tenerent Licinius ac Severus et secundum Caesarum nomen Maximinus et Candidianus, inexpugnabili muro circumsaeptus securam et tranquillam degeret senectutem. XXXIII, 7 Lactantius describes how Galerius died in great pains resulting from decaying of his body (XXXIII, 1–1). Odor it autem non modo per palatium, sed totam civitatem pervadit. FIGURE 189. He (Galerius) himself had for a friend Licinius, who was allied to him by the fact that they lived for a long time in the same tent and were close from the very beginning of his military career, and he followed his advices in making all the decisions, but he did not want to make him Caesar, because he did not want to call him his son, and because he wanted to make him later Augustus and brother instead of Constantius; so he himself would have the supreme power and celebrate vicennalia as he wanted, being furious with the world, and after his son, who was nine at the time, becomes Caesar instead of him, he would abdicate; so, as Licinius and Severus would have the supreme power and Maximin and Candidianus would have the title of Caesar, second in ranking, he would spend secure and peaceful old age, surrounded by unconquerable walls.12 (after AD 298) XXI, 1–2 Adeptus igitur maximam potestatem ad vexandum orbem, quem sibi patefecerat, animum intendit. Nam post devictos Persas, quorum hic ritus, hic mos est, ut regibus suis in servitium se addicant et reges populo suo tamquam familia utantur, hunc The stench not only pervaded the palace, but also the entire city.14 (AD 311) XXXV, 3–4 Nec tamen ille hoc facto veniam sceleris accepit a deo, sed post dies paucos commendatis Licinio coniuge sua et filio atque in manum traditis, cum iam totius corporis membra diffluerent, horrenda tabe consumptus est. Idque cognitum Nicomediae … 12 — Srejovi} 1983 C, 63, uses this place as evidence that Galerius intended to renounce the throne after twenty years and to spend peaceful old age in the fortified palace built at Gamzigrad. This also accepts M. ^anak-Medi} in the text in this book. We think that, considering the context and Lactantius’ rhetoric style, it is a metaphor, and that Galerius’“unconquerable wall” are in fact Augusti and Caesars. 13 — Lactantius emphasizes that after the victory over Persians Galerius’ reign got many characteristics of the oriental tyranny. Many elements of Galerius' rulling ideology and iconography could be explained by these oriental influences. 14 — This corroborates the assumption that Galerius died in the town with the palace, by all appearances in Serdica. 216 WRITTEN SOURCES ON GALERIUS Because of that,18 Julius Constantius and Galerius Maximian, nicknamed Herdsman, were proclaimed Caesars and invited to the family.19 First was married to Herculius’ stepdaughter and second to Diocletian’s daughter, after their previous marriages were dissolved, as Augustus did it once in case of his daughter Julia and Tiberius Nero. The homeland of all of them was Illyricum: despite their poor education, thanks to the fact that they experienced hardships of life in the village and in the army, they were the best in the state. mensis eiusdem, cum futura essent vicennalia kalendis martiis impendentibus. Nor after he did that,15 he was pardoned by the God for his crime. So, after few days he committed his wife and son to Licinius and put them under his protection, and while all parts of his body were decomposing, he died of terrible putrefying. When this became known in Nicomedia…16 in the same month when the celebration of his vicennalia were planned for the ensuing March Calends.17 (AD 308–311) SEXTUS AURELIUS VICTOR 40, 8–11 Hoc acrior Galerius ascito in consilium Iovio Licinium vetere cognitum amicitia Augustum creat; eoque ad munimentum Illyrici ac Thraciae relicto Romam contendit. Ibi cum obsidione distineretur, militibus eadem, qua superiores, via attentatis, metu ne desereretur, Italia decessit; pauloque post vulnere pestilenti consumptus est, cum agrum satis reipublicae commodantem caesis immanibus silvis atque emisso in Danubium lacu Pelusone apud Pannonios fecisset. Cuius gratia provinciam uxoris nomine Valeriam appellavit. Huic quinquennii imperium, Constantio annuum fuit, cum sane uterque potentiam Caesarum annos tredecim gessissent. Sextus Aurelius Victor lived in the 4th century. He was of the North African descent. He was governor of the province Pannonia Secunda during the reign of Julian the Apostate. His main work is The Book on Caesars (De caesaribus). This is a short history of the Roman emperors from Augustus to Constantius II (AD 360). The main source of Victor was Suetonius. He resembles Salustius and Tacitus in his moralistic attitude to history. Few works, which he did not write, have also been ascribed to him. The best known is Epitome de caesaribus, which has many passages from the text De caesaribus and follows the Roman history until the death of Theodosius I. Edition used: Sexti Aurelii Victoris Liber de Caesaribus, praecedunt Origo gentis Romanae et Liber de viris illustribus urbis Romae, subsequitur Epitome de Caesaribus, recensuit F. Pichlmayr, Lipsiae 1911. Because of that,20 Galerius became even more energetic and after consulting Iovius, he raised Licinius, who was his long time friend, to the rank of Augustus. He left him to guard Illyricum and Thrace and hurried to Italy. There he was engaged De caesaribus The Book on Caesars 15 — After issuing the edict of tolerance for the Christiansa. 16 — Lacuna in the text. 17 — This means that Lactantius dates death of Galerius in AD 312. (AD 293) 18 — Because the Empire was, as suggested by Aurelius Victor in the preceding text, at the same time endangered from the outside by the Persians and from the inside by various usurpers: Carausius in Gaul, Julian in Africa and Achileus in Egypt. 39, 24–26 His de causis Iulium Constantium, Galerium Maximianum, cui cognomen Armentario erat, creatos Caesares in affinitatem vocant. Prior Herculii privignam, alter Diocletiano editam sortiuntur diremptis prioribus coniugiis, ut in Nerone Tiberio ac Iulia filia Augustus quondam fecerat. His sane omnibus Illyricum patria fuit: qui, quamquam humanitatis parum, ruris tamen ac militiae miseriis imbuti satis optimi reipublicae fuere. 19 — According to the tetrarchic ruling ideology, Augusti and Caesars were one divine family of Iovii and Herculii, where Augusti were the incarnations of Jupiter (Diocletian) and Hercules (Maximian Herculius), and Caesars are considered to be their sons. One of best studies of this subject is Seston 1946, 211–230. 20 — Because of the death of Caesar Severus during war against Maxentius in Italy. 217 ALEKSANDAR V. POPOVI] in a siege and then left Italy, afraid that the soldiers might dessert him, prompted to do that in the same way as did those before them.21 Sometime later he died of infected wound, after he made land in Pannonia suitable for the state, because he cut out vast forests and water from the lake Peluso22 drained into the Danube. Because of that, he named the province Valeria, after his wife.23 At that time he was already emperor for five years and Constantius for a year, while both enjoyed the power of the Caesars for thirteen years.24 Pseudo Aurelius Victor: Epitomae de Caesaribus (AD 305) 39, 2 Is Maximianum Augustum effecit; Constantium et Galerium Maximianum, cognomento Armentarium, Caesares creavit… He25 appointed Maximian26 Augustus; Constantius27 and Galerius Maximianus, nicknamed Herdsman, he made Caesars… AMMIANUS MARCELLINUS Ammianus Marcellinus was born in the prosperous family in Antioch, around AD 330. He served as an officer in the army during the wars in Gaul and Mesopotamia, thus gaining the personal insight into many events he later described in his literary works. He was the great admirer of the emperor Julian the Apostate (361–363) and took part in his unsuccessful campaign against Persia. After that he withdrew to his hometown. He moved to Rome around AD 380, where he moved in the circles of senatorial aristocracy. He died around AD 395. His main work is History (Res gestae) in 31 volumes, that continued the work of the Roman historian Tacitus. He described the history of the Roman Empire from the reign of Nerva until the death of emperor Valens in AD 378. Considerable part of Ammianus’ History is nowadays lost. The preserved volumes 14–31 cover the period from AD 353 to AD 378. Edition used: Ammianus Marcellinus in Three Volumes, with an English translation by J.C. Rolfe, London – Cambridge Mass., 1964. Res gestae History 40, 15–17 Galerius autem fuit (licet inculta agrestique iustitia) satis laudabilis, pulcher corpore, eximius et felix bellator, ortus parentibus agrariis, pastor armentorum, unde ei cognomen Armentarius fuit. Ortus Dacia Ripensi ibique sepultus est; quem locum Romulianum ex vocabulo Romulae matris appellarat. Is insolenter affirmare ausus est matrem more Olympiadis, Alexandri Magni creatricis, compressam dracone semet concepisse. (AD 363) XXIV, 10, 1–2 His hoc modo peractis, discursisque itineribus, Antiochiam venimus, ubi per continuos dies, velut offenso numine multa 21 — Previously, soldiers during the siege of Rome left Caesar Severus and deserted to Maxentius. 22 — Modern Balaton Lake in Hungary. Galerius was (it could be said of primitive and fierce temperament) rather praiseworthy, well built, outstanding and successful soldier, born in the family of peasants, tending cattle and therefore nicknamed Herdsman (Armentarius). He was born in Dacia Ripensis, where he was also buried; he called that place Romulianum after his mother’s name Romula. He arrogantly claimed that his mother, like Olympias, the mother of Alexander the Great, conceived him after being raped by a dragon.28 218 23 — Province Valeria encompassed part of present-day Hungary, on the right Danube bank. 24 — If we count the time since Galerius became Caeasr in 293, it would be the year 306. But, if we consider his inauguration as Augustus in 305, then the year in question is 310. Obviously we can not rely on Victor’s chronology. 25 — I.e. Diocletian. 26 — I. e. Maximian Herculius. 27 — I.e. Constantius Chlorus. 28 — About this question see Du{ani} 1995. Srejovi} 1983 C, 63, uses this place to support his claim that Diocletian proclaimed Galerius “second Romulus and Alexander”. As we can see Diocletian and Romulus are not mentioned in this fragment. WRITTEN SOURCES ON GALERIUS visebantur et dira, quorum eventus fore luctificos, gnari rerum prodigialium praecinebant. Nam et Maximiani statua Caesaris, quae locata est in vestibulo regiae, amisit repente sphaeram aeream formatam in speciem poli, quam gestabat, et cum horrendo stridore sonuerunt, in consistorio trabes, et visa sunt interdiu sidera cometarum, super quorum natura ratiocinantes physici variant. After this had been done in such a way, and after the journey, we came to Antioch, where day after day, as the deity had been offended, many terrible things were seen, for which the good judges assumed that would have miserable end. Namely, the statue of Caesar Maximian29 that was standing in the vestibule of the imperial palace suddenly dropped the bronze sphere shaped as firmament, which it was holding, and in the council hall the beams creaked with terrifying noise, while the comets were seen in the broad daylight, and the natural scientists disagree in the explanation of these phenomena. EUTROPIUS Eutropius is the Roman historian from the 4th century. At the demand of emperor Valens he wrote the text Abridged history from the Foundation of Rome (Breviarum ab urbe condita) in ten books, covering the period from the foundation of Rome to the death of emperor Jovian (AD 364). He used as his sources earlier Roman historians: epitomes of Tit Livy, Suetonius and others. Because of its digested form his historical text was often used in the Middle Ages when it even was the school textbook. Edition used: Eutropii Breviarium historiae Romanae, edidit H.R. Dietsch, Lipsiae 1883. (AD 311) X, 4 Per hoc tempus a Galerio Licinius imperator est factus, Dacia oriundus, notus ei antiqua consuetudine et in bello, quod adversus Narseum gesserat, strenuis laboribus et officiis acceptus. Mors Galerii confestim secuta est. At that time Galerius made Licinius the emperor.30 He was of Dacian origin and Galerius had known him for a long time. Because of the great strains and duties that Licinius experienced in the war against Narseus, he accepted him as a friend. Immediately after that Galerius died. CONSULARIA CONSTANTINOPOLITANA This is the list of the Roman consuls until the year AD 468. It is divided into three segments, and the first one ends with foundation of Constantinople in AD 330. This part encompasses also the period of tetrarchy, i.e. the years of Galerius’ reign. The integral text is preserved in just one manuscript: Claromontanus, hodie Berolinensis Cheltenhamensium n. 1829, from the 9th century. Edition used: Consularia Constantinopolitana ad a. CCCLXVIII, cum additamento Hydatii ad a. CCCLXXVIII, accedunt Consularia Chronici paschalis, in: Chronica Minora saec. IV. V. VI. VII., ed. Th. Mommsen, vol. I, Berolini 1892, 196–247. (AD 311) s.a. 311 311. Maximiano VIII consule. 1. His conss., quod est Rufino et Volusiano, diem functus Maximianus iun. Breviarium ab urbe condita Abridged History from the Foundation of Rome (around AD 250) IX, 22 … Maximianus Galerius in Dacia haud longe a Serdica natus. … Maximian Galerius was born in Dacia, not far from Serdica. 29 — Bronze statue of Galerius with globe in his hand has its pendant in fragmentary preserved porphyritic statue from Gamzigrad. It was, probably, erected in 298 in the palace in Antioch, on the occasion of Galerius' triumph after campaign in Persia. There is certain doubt that Galerius as Caesar was entitled to the statue with such symbols, reserved exclusively for his senior co-rulers, Augusti, so that it was erected later, i.e. only after AD 305, when he also became Augustus. On this issue see Seston 1946, 182, n. 4. 30 — I.e. he proclaimed him Augustus. 219 ALEKSANDAR V. POPOVI] AD 311 311. During Maximian’s eighth consulate.31 1. In time of these consuls, i.e. Rufinus and Volusianus, Maximian the Younger had died.32 EXCERPTA VALESIANA Excerpta Valesiana got their name after their first publisher, Henricus Valesius – Henri de Valois, who published them in Paris in 1636. The text consists of two parts. First part dates from around AD 390, and second from around AD 550. For our subject is interesting the first part, which is in a certain way kind of introduction to History by Ammianus Marcellinus, and it is usually published together with this work. It covers the period from AD 305 and AD 337, and it is titled Origin of Emperor Constantine (Origo Constantini imperatoris). Edition used: Excerpta Valesiana, in Ammianus Marcellinus in Three Volumes, with an English translation by J.C. Rolfe, Vol. I, London – Cambridge Mass. 1964, 506–569. (AD 311) 3, 8 Tunc Galerius in Illyrico Licinium Caesarem fecit. Deinde illo in Pannonia relicto, ipse ad Serdicam regressus, morbo ingenti occupatus sic distabuit, ut aperto et putrescenti viscere moreretur, in supplicium persecutionis iniquissimae ad auctorem scelerati praecepti iustissima poena redeunte. Imperavit annos XVIIII. Then Galerius appointed Licinius as Caesar in Illyricum.33 So, after he left him34 in Pannonia, he returned to Serdica, where horrible illness came over him. He was so disfigured that he was dying with putrefied and open bowels, as the fair punishment ensued to torture the creator of criminal instruction to carry out the most unjustly persecution. He had reigned for 18 years. INSCRIPTIONES LATINAE SELECTAE 633 Genio virtutum, | Marti Aug. con|servatori [Galer.] | Va[leri Maximi|ani]1 nobilissimi | et fortissimi Caes. | Valerius Florus | v. p. p. p. Num., nu|mini maiesta|tique eorum | dicatissimus posuit, | curante Iulio | Lambesio cur. | rei publicae. To the genius of virtues, Mars Augustus, protector of Galerius Valerius Maximianus,35 the most noble and most courageous Caesar, dedicated Valerius Florus,36 the most noble man,37 governor of the province Numidia, the most devoted to their deity and majesty, thanks to the care of Julius Lambesius the state curator. 38 31 — It means Galerius’ consulate. 32 — At first sight it seems confusing that there are three consuls in that year. We think that explanation is that Rufinus or Volusian were elected after the death of Galerius. 33 — It is considered today that he proclaimed him Augustus. 34 — It is not quite clear whom this refers to, but if we compare this text with above quoted fragment from De caesaribus by Aurelius Victor (40, 8–11), it seems that it is Licinius. 35 — Inscription is partially reconstructed, as Galerius’ name, now in brackets, was erased. It is obvious that Galerius, because of his attitude to Christianity, suffered damnatio memoriae in later times. 36 — About Valerius Florus see PLRE I, 368, s.v. Valerius Florus 3. 37 — Vir perfectissimus – title available only to the members of knight class. 38 — Curator rei publicae – title in Roman administration. Mention of god Mars as Galerius’ guardian in this inscription confirms Lactantius’ claim (De mort. persec., IX,9) that he proclaimed himself the son of Mars. Inscription was found at Thamugadi in Numidia. 220 SOURCES Ammianus Marcellinus in Three Volumes, with an English translation by J.C. Rolfe, vol. I–III, London – Cambridge Mass. 1963–1964. Sexti Aurellii Victoris Liber de Caesaribus, praecedunt Origo gentis Romanae et Liber de viris illustribus urbis Romae, subsequitur Epitome de Caesaribus, recensuit F. Pichlmayr, Lipsiae 1911. Chronica minora saec. IV. V. VI. VII., edidit Th. Mommsen, volumen I, Berolini 1892. Constantin VII Porphyrogénète, Le livre des cérémonies, tome I, livre 1, texte établi et traduit par A. Vogt, Paris 1935. Dionis Cassii Cocceiani Historia Romana, editionem primam curavit L. Dindorf, recognovit I. Melber, vol. II, Lipsiae 1894. Eutropii Breviarium Historiae Romanae, edidit H.R. Dietsch, Lipsiae 1883. Inscriptiones Latinae selectae, ed. H. Dessau, vol. I, Berolini 1892. Lactantius, De mortibus persecutorum, edited and translated by J.L. Creed, Oxford 1984. Procopii De aedificiis, edidit J. Haury, Lipsiae 1913. Vitruvius, Ten Books on Architecture, with an English translation by M. Hicky Morgan, London – Cambridge Mass. 1960. Ioannis Zonarae Epitome historiarum, cum C. Ducangii suisque annotationibus edidit L. Dindorfius, vol. III, Lipsiae 1870. Zosime, Histoire nouvelle, tome I, texte établi et traduit par F. Paschoud, Paris 1971. 221 BIBLIOGRAPHY Age of Spirituality 1979 Catalogue in: Age of Spirituality. Late Antique and Early Christian Art, Third to Seventh Century (ed. K. Weitzmann), New York 1979, cat. 83, 90. Azevedo de 1959 M. Cagiano de Azevedo, Admiranda Palatia, I palazzi imperiali e le residence tetrarchiche, Bolletino del Centro di studi per la storia dell’ architettura 14, 1959, 3–15. Azevedo de 1970 M. Cagiano de Azevedo, Policromia e polimateria nelle opere d’arte della tarda antichita e dell’ alto medioevo, Felix Ravenna. Rivista di antichita ravennati christiane e byzantine fasc. I (CI), Ravenna, 1970, 223–259. Baratte 1993 F. Baratte, La vaisselle d’argent en Gaule dans l’Antiquité tardive, Paris 1993. Balty 1969 J. Ch. Balty, Apamée 1965 – 1968, Colloque Apamée de Syrie, Bruxelles 1969. Barnes 1982 T. D. Barnes, The New Empire of Diocletian and Constantine, Cambridge Mass, London 1982. Bechert 1971 T. Bechert, Römische Lagentore und ihre Bauinschriften, Boner Jahrbücher der Rheinischen Landesmuseums in Bonn 171, Bonn 1971, 210–285. Berchem 1954 D. van Berchem, Recherches sur la chronologie des enceintes de Syrie et de Mesoportamie, Syria XXXI, Paris 1954. Berciu 1961 D. Berciu, Die Verbicioara-Kultur. Vorbericht über eine neue, in Rumänien entdeckte bronzezeitliche Kultur, Dacia V, Bucureºti 1961, 123–161. 222 BIBLIOGRAPHY Bi~kov 1991 V. V. Bi~kov, Vizantijska estetika, Beograd 1991. Breithaupt 1861 A. Breithaupt, Timazit, eine neue Gesteinsart, und Gamsigradit, ein neuer Amphibol, Berg- und Hüttenmännische Zeitung 5. II 1861, 51–54. Biki} 1997 V. Biki}, Sredwevekovna trpezna keramika iz Gamzigrada: poreklo i radionice (Summary: Medieval Table Pottery from Gamzigrad: Origin and Workshops), Arheolo{ka istra`ivawa Isto~ne Srbije / Archaeology of Eastern Serbia (urednik M. Lazi}), Beograd 1997, 319–328. Brukner 1987 O. Brukner, Importovana i panonska kerami~ka produkcija sa aspekta dru{tveno-ekonomskih promena, Po~eci romanizacije u jugoslovenskom delu provincije Panonije (urednik M. Stojanov), Novi Sad 1987, 25–44. Biki}, [ari} 1997 V. Biki}, J. [ari}, Kula 19, Romulijana – Gamzigrad, kasnoanti~ka palata, Starinar XLVIII, Beograd 1997, 203–208. Brukner 1995 O. Brukner, Domoroda~ka naseqa, Arheolo{ka istra`ivawa du` autoputa kroz Srem (urednik Z. Vapa), Novi Sad 1995, 91–136. Biki}, [ari} 2001 V. Biki}, J. [ari}, Izve{taj o arheolo{kim iskopavawima na lokalitetu Romulijana – Gamzigrad u 1998. godini. Kula 19, Starinar L (2000), Beograd 2001, 280–282. Bruneau 1972 Ph. Bruneau, Exploration archéologique de Délos, Paris 1972. Bob~ev 1952 S. Bob~ev, Smesnata zidariÿ v rimskite i ranovizantiéskite stroe`i, Sofiÿ 1952. Buli}, Karaman 1927 F. Buli}, Lj. Karaman, Pala~a cara Dioklecijana u Splitu, Zagreb 1927. Bob~ev 1961 S. Bob~ev, Nova seriÿ ot kasnoanti~ni zidove s smeseni zidariÿ sled 1951. g., Izvestiÿ na Bülgarski Arheologi~eski institut XXIV, Sofiÿ 1961, 153–201. Bülow von 2007 G. von Bülow, Neue Untersuchungen in Palast des Kaisers Galerius / Felix Romuliana (Gamzigrad), Roms Erbe auf dem Balkan. Spätantike Kaiservillen und Stadtanlagen in Serbien (Hrsg. U. Brandl, M. Vasi}), Mainz 2007, 54–58. Bob~ev 1970 S. Bob~ev, Rimsko-korintski kapiteli od Ygozapadna i Severna Búlgariÿ, Izvestiÿ na Bülgarskiÿ Arheologi~eski institut XXXII, Sofiÿ 1970, 113–126. Bülow von, Schüler 2009 G. von Bülow, T. Schüler, Geophysical and Archaeological Research at Gamzigrad. Report of the 2004–2007 Campaigns (Rezime: Geofizi~ka i arheolo{ka istra`ivawa u Gamzigradu – izve{taj sa kampawa u 2004–2007. godini), Starinar LVII (2007), Beograd 2009, 231–249. Boëthius 1948 A. Boëthius, Roman and Greck Town Architecture, Götenborges Högskalas Arssrift LIV (1948, 3), Götenborg 1948, 3–22. Bordenache 1969 G. Bordenache, Sculture greche e romane I, Bucarest 1969. Castagnoli 1956 F. Castagnoli, Ippodamo di Mileto e l’ urbanistica a pianta ortogonale, Roma 1956. Bori}-Bre{kovi} 1994 B. Bori}-Bre{kovi}, Gold Coins from the Consecration Memorial 1, in: Srejovi}, Vasi} 1994, 160–179. Chevalier – Gheerbrant 1983 J. Chevalier – A. Gheerbrant, Rje~nik simbola, Zagreb 1983. Bréhier 1936 L. Bréhier, La sculpture et les arts mineurs byzantine, Paris 1936. ]ur~i} 1993 S. ^ur~i}, Late-Antique Palaces. The Meaning of Urban Context, Ars Orientalis 23, 1993, 67–90. 223 BIBLIOGRAPHY ^anak-Medi} 1975 M. ^anak-Medi}, ^etvorolisni kapiteli iz Gamzigrada (Résumé: Les chapitaux à quatres feuilles d’achante), Zbornik Narodnog muzeja VIII, Beograd 1975, 247–254. ^anak-Medi} 1976 M. ^anak-Medi}, Le palais de l’époque de la base antiquité près de Gamzigrad, Actes du XIV e Congrés international des études byzantines, Bucarest 1976, 357–362. ^anak-Medi} 1978 M. ^anak-Medi}, Gamzigrad, kasnoanti~ka palata. Arhitektura i prostorni sklop (Résumé: Gamzigrad, palais bas-antique. Architecture et sa structuration), Saop{tewa XI, Beograd 1978. ^anak-Medi} 1995 M. ^anak-Medi}, Spatial develepment of Romuliana within the late roman Court architecture, The Age of Tetrarchs (ed. D. Srejovi}), Belgrade 1995, 52–63. ^ovi} 1987 B. ^ovi}, Grupa Donja Dolina – Sanski Most, Praistorija jugoslovenskih zemalja V, Sarajevo 1987, 260–266. Diepenbach 1921 W. Diepenbach, Palatium, Dissertation, Giessen 1921. Du{ani} 1995 S. Du{ani}, Late Roman Mining in Illyricum: Historical Observation, Ancient Mining and Metallurgy in Southeast Europe, International Symposium, Donji Milanovac, May 20th–25th, 1990, Bor – Belgrade 1995, 219–225. Du{ani} 1995 S. Du{ani}, Imitator Alexandri and Redditor Libertatis, Two controversial themes of Galerius’ political propaganda, The Age of Tetrarchs. A Symposium held from the 4th to the 9th October 1993 (ed. D. Srejovi}), Beograd 1995, 79–98. Duval 1965 N. Duval, La place de Split dans l’ architecture antique du Bas-Empire, Urbs 4 (1961–1962), Split 1965, 67–95. Duval 1971 N. Duval, Palais et forteresses de Yougoslavie, Bulettin de la Société nationale des antiquaires de France, Paris 1971, 115–124. Duval 1987 N. Duval, Existe-t-il une “structure palatial” propre à l’ antiquité tardive?, Le systeme palatial en Orient, en Grece et à Rome. Traveaux du Centre de recherches sur le Proche-Oriente et la Grece Antique 9 (1985), Strasbourg 1987, 463–490. Duval 1987 A N. Duval, Le site de Gamzigrad (Serbie) est-il le palais retraite de Galere?, Bulletin de la Societé nationale des antiquiaires de France, Paris 1987, 61–84. Dyggve 1941 E. Dyggve, Ravenatium Palatium Sacrum. La basilica imperiale per ceremonie?, Studi sill’architecture dei palazzi delle tarda antichita, Kóbenhavn 1941, 3–55. Elijade 1991 M. Elijade, Istorija verovawa i religijskih ideja, Beograd 1991. Ensslin 1930 W. Ensslin, Galerius Maximianus (2), Paulys Real-Encyclopädie der classischen Altertumswissenschaft, vierzehnter Band, Stuttgart 1930, 2516–2528. Ercegovi}-Pavlovi}, Mini} 1986 S. Ercegovi}-Pavlovi}, D. Mini}, Le site d’habitation et la nécropole de Pozajmi{te (Rezime: Sredwovekovno naseqe i nekropola na lokalitetu Pozajmi{te u Grabovici), \erdapske sveske / Cahiers des Portes de Fer III, Beograd 1986, 353–361. Fiala – Patsch 1895 F. Fiala – K. Patsch, Römische Funde in der Hercegovina, Wissenschaftliche Mitteilungen aus Bosnien und der Hercegovina III, Wien 1895, 275–308. Frazer 1966 A. Frazer, The Iconography of the Emperor Maxentius’ Buildings in Via Appia, The Art Bulletin XLVIII, London 1966, 385–392. Formigé 1943 J. Formigé, Proportion et traces harmonique, Paris 1943. Formigé 1949 J. Formigé, La Triomphé des Alpes (La Tourbie), Galia 11 (Suppl.), Paris 1949, 110 sqq. Frigerio 1934–35 F. Frigerio, Antiche porte de citta italiche e romane, Rivista archeologica della provincia antica diocesi di Como 108–110, Como 1934–35, 1–285. 224 BIBLIOGRAPHY Gall 1958 I.Z. Gall, Modes de construction et technique dans l’architecture romaine, Revue Archéologique, avril –juin, Paris 1958, 181–202. Gerke 1973 F. Gerke, Kasna antika i rano hri{}anstvo (prevod A.Cermanovi}-Kuzmanovi}), Novi Sad 1973. Grabar 1963 A. Grabar, Sculptures byzantines de Constantinople, IVe–Xe siècle, Institut français d’archéologie d’Istanbul XVIII, Paris 1963. Grenier 1934 A. Grenier, Manuel d’archéologie préhistorique celtique et gallo-romaine, Paris 1934. Guggisberg 2003 M. Guggisberg, Der spätrömische Silberschatz von Kaiseraugst. Die neuen Funde, Augst 2003. Hannestad 1988 N. Hannestad, Roman Art and Imperial Policy, Aarhus 1988. Hoddinott 1963 R.F. Hoddinott, Early Byzantine Churches in Macedonia and Southern Serbia, London 1963. Homo 1951 L. Homo, Rome impérial et l’urbanisme dans l’antiquité, Paris 1951. Hunyady 1942/1944 I. Hunady, Kelták a Karpat-medencében, Dissertationes Pannonice II, Budapest 1942–1944. Ili} 2008 O. Ili}, Early Christian Baptisteries in Northern Illyricum (Rezime: Ranohri{}anske krstionice u severnom Iliriku), Starinar LVI (2006), Beograd 2008, 223–244. Ivanovski 1984 M. Ivanovski, Eden docnoanti~en grob od Tarane{ (Summary: A Late-antique Grave from Taranesh), @iva antika / Antiquité Vivante 34/1–2, Skopje 1984, 219–226. Ivanovski 1987 M. Ivanovski, The Grave of a Warrior from the Period of Licinius I found at Tarane{, Archaeologia Iugoslavica 24, Beograd 1987, 81–90. Jankovi} 1983 \. Jankovi}, U sutonu antike, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of the Late Classical Times), Galerija SANU 45, Beograd 1983, 98–119. Jankovi} 1983 A \. Jankovi}, Ranovizantijski Gamzigrad, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of the Late Classical Times), Galerija SANU 45, Beograd 1983, 120–141. Jankovi} 1983 B \. Jankovi}, Slovenski grad, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of the Late Classical Times), Galerija SANU 45, Beograd 1983, 142–160. Jankovi} 1997 \. Jankovi}, Sloveni u Isto~noj Srbiji u VI i VII stole}u (The Slavs in East Serbia in VI and VII Century), Arheolo{ka istra`ivawa Isto~ne Srbije / Archaeology of Eastern Serbia (urednik M. Lazi}), Beograd, 1997, 133–149. Jankovi} 2007 \. Jankovi}, Srpsko Pomorje od 7. do 10. stole}a, Beograd 2007. Jankovi} M. 1981 M. Jankovi}, Sredwovekovno naseqe na Velikom Gradcu u X– XI veku (Résumé: Le site d’habitation de Veliki Gradac aux Xe/XIe siècles, Beograd 1981. Jankovi} M. 1983 M. Jankovi}, Neki podaci o izradi predmeta od obojenih matala na Kqu~u Dunava u IX–X veku (Summary: Some Informations on Production of Non-ferrous Metals in the Region of Klju~ in the Danube Valley from 9th to 11th Century), Zbornik radova Narodnog muzeja XI–1, Beograd 1983, 99–118. Jankovi} M., Jankovi} \. 1978 M. Jankovi}, \. Jankovi}, Podunavski gradovi pomenuti kao postradali 1072 godine (Résumé: Les villes danubiennes citées en 1072), Godi{wak grada Beograda 25, Beograd 1978, 41–57. Jankovi} M., Jankovi} \. 1990 M. Jankovi}, \. Jankovi}, Sloveni u jugoslovenskom Podunavqu (Résumé: Les Slaves du bassin danubien yougoslave), Beograd 1990. Jeanmarie 1951 H. Jeanmarie, Dionysos. Histoire du culte de Bacchus, Paris 1951. 225 BIBLIOGRAPHY Neighbouring Regions in the 2nd Millennium B.C., Beograd 1996, 185–201. Jelo~nik 1961 A. Jelo~nik, Najdba argenteusov zgodnje tetrarhije v Sisku / The Sisak Hoard of Argentei of the Early Tetrarchy, Situla 3, Ljubljana 1961. Jovanovi} S. 1987 S. Jovanovi}, Praistorijski lokaliteti kwa`eva~kog kraja, Glasnik Srpskog arheolo{kog dru{tva 4, Beograd 1987, 120–124. Jeremi} 1993 M. Jeremi}, Katalog, Rimski carski gradovi i palate u Srbiji (priredio D. Srejovi}), Beograd 1993, 196, kat. 34. Jovanovi} S. 1997 S. Jovanovi}, Sredwovekovna nekropola na Gamzigradu (Summary: Medieval Cemetery at Gamzigrad), Uzdarje Dragoslavu Srejovi}u / ANTIDWRON Dragoslav Srejovi} (urednik M. Lazi}), Beograd 1997, 503–509. Jeremi} 1993 M. Jeremi}, Catalogue, Roman Imperial towns and palaces in Serbia (ed. D. Srejovi}), Belgrade 1993, 196, cat. 34. Jevti} 2004 M. Jevti}, Gvozdeno doba u okolini Bora (Summary: The Iron Age in the Bor Area), Bor i okolina u praistoriji antici i sredwem veku / The Bor Area in Prehistory, Antiquity and the Middle Ages (urednik. M. Lazi}), Bor – Beograd 2004, 127–163. Jobst 2002 H. Jobst, Il culto delle divinità orientali a Carnuntum, M. Buora, W. Jobst (Hrsg.), Roma sul Danubio. Da Aquileia a Carnuntum lungo la via dell’ambra, Ausstellungskatalog, Udine 2002, 133–138. Jovanovi} A. 1975 A. Jovanovi}, Neki aspekti problema skupnog nalaza skulptura sa Medijane kod Ni{a (Résumé: Certains aspects du problème de la trouvaille collective de sculptures à Mediana près de Naissus), Starinar XXIV–XXV (1973–1974), Beograd 1975, 57–65. Jovanovi} A., Lalovi} 1993 A. Jovanovi}, A. Lalovi}, Ostava solida iz Gamzigrada (Romuliana) (Summary: A Hoard of Coins from Gamzigrad (Romuliana)), Numizmati~ar 16, Beograd, 61–78. Jovanovi} A., Lalovi} 1994 A. Jovanovi}, A. Lalovi}, Hoard of Gold Coins Discovered next to the Foundation of Mausoleum 1, in: Srejovi}, Vasi} 1994, 181–185. Jovanovi} B. 2004 B. Jovanovi}, Neolit i rani eneolit / The Neolithic and Early Eneolithic, Bor i okolina u praistoriji, antici i sredwem veku / The Bor Area in Prehistory, Antiquity and the Middle Ages (urednik. M. Lazi}), Bor – Beograd 2004, 33–55. Jovanovi} B., Jankovi} 1996 B. Jovanovi}, N. Jankovi}, Die Keramik der Nekropole der Para}in-Kultur – Trnjane bei Bor, The Yugoslav Danube Basin and Jovanovi} S. 2000 S. Jovanovi}, Romulijana. Sredwovekovna nekropola ispred isto~ne kapije, Razvitak 40, Zaje~ar 2000, 203–204. Jovi} 1998 V. Jovi}, Kamen u tetrarhijskoj arhitekturi – primena, poreklo i simboli~ko zna~ewe, Razvitak 38, br. 200, Zaje~ar 1998, 134–137. Kanitz 1868 F. Kanitz, Reise in Süd-Serbien und Nord Bulgarien, Wien 1868. Kanitz 1868 A F. Kanitz, Serbien, Leipzig 1868. Kautzsche 1936 R. Kautzsche, Kapitellstudien, Beiträge zu einer Geschichte des spätantiken Kapitels im Osten von virten bis ins siebende Jahrhundert, Berlin – Leipzig 1936. Käler 1965 H. Käler, Split i Piazza Armerina, rezidencije dvaju careva-tetraha, Urbs 4 (1961–1962), Split 1965, 97–109. Keller 1971 E. Keller, Die spätrömischen Grabfunde in Südbayern, München 1971. Koep 1958 L. Koep, Die Konsekrationmünzen Kaiser Konstantins und ihre religions-politische Bedeutung, Jahrbuch für Antike und Christentum 1, 1958, 94–104. Kolarik 2006 R. Kolarik, Late antique floor mosaics in the Balkans (Rezime: Kasnoanti~ki podni mozaici na Balkanu), Ni{ i Vizantija / Ni{ and Byzantium IV, Ni{ 2006, 159–178. 226 BIBLIOGRAPHY Kondi} 1984 V. Kondi}, Les formes des fortifications protobyzantines dans la région des Portes de Fer, Villes et peuplement dans l’Iliricum protobyzantin, Rome 1984, 132–161. Lavedan 1926 P. Lavedan, Histoire de l’ urbanisme, Antiquité – moyen-âge, Paris 1926. Lavin 1963 I. Lavin, The Hunting Mosaics of Antioch and their Sources, Dumbarton Oaks Papers 17, Washington 1963, 179–286. Kondi} 1994 J. Kondi}, Katalog, Anti~ko srebro u Srbiji (priredila I. Popovi}) / Antique Silver from Serbia (ed. I. Popovi}), Beograd 1994, 327, kat. 271. Lavin 1967 I. Lavin, The Ceiling Frescoes in Trier and Illusionisme in Constantinian Painting, Dumbarton Oaks Papers 21, Washington 1967, 99–113. Krautheimer 1965 R. Krautheimer, Early Christian and Byzantine Architecture, Baltimore 1965. Laubscher 1975 H. P. Laubscher, Die Reliefschmuck des Galeriusbogens in Thessalonike, Berlin 1975. Krencher-Zschietzschmann 1938 D.W. Krencher-Zschietzschmann, Römische Tempel in Sirien, Berlin – Leipzig 1938. Lalovi} 1983 A. Lalovi}, Epigrafski spomenici, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of the Late Classical Times), Galerija SANU 45, Beograd 1983, 163–177. Lalovi} 1983 A A. Lalovi}, Novac, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of the Late Classical Times), Galerija SANU 45, Beograd 1983, 171–177. Lazi} 2004 M. Lazi}, Bor i okolina u bronzano doba (Summary: The Bor Area in the Bronze Age), Bor i okolina u praistoriji, antici i sredwem veku – The Bor Area in Prehistory, Antiquity and the Middle Ages (urednik M. Lazi}), Bor – Beograd 2004, 101–125. Lalovi} 1993 A A. Lalovi}, Katalog, Rimski carski gradovi i palate (priredio D. Srejovi}), Beograd 1993, 204–208, kat. 48–50. Lalovi} 1993 B A. Lalovi}, Catalogue, Roman Imperial Towns and Palaces in Serbia (ed. D. Srejovi}), Belgrade 1993, 204–208, cat. 48–50. Levi 1974 D. Levi, Antioch Mosaic Pavements I, Princeton – Londra 1947. Lalovi} 1998 A. Lalovi}, Arheolo{ka istra`ivawa na Gamzigradu – Romuliana u 1996. i 1997. godini, Razvitak 200, Zaje~ar 1998, 126. Lalovi} 2000 A. Lalovi}, Arheolo{ka istra`ivawa na Gamzigradu – Romuliana u 1998 i 1999. godini, Razvitak 203–204, Zaje~ar 2000, 125. Lalovi}, Ru`i}, Jovanovi} 1997 A. Lalovi}, M. Ru`i}, S. Jovanovi}, Romulijana – Gamzigrad, Kasnoanti~ka utvr|ena palata, Terme, Starinar XLVIII, Beograd, 1992, 199. Lazi} 1998 M. Lazi}, Gamzigradska kultura – posledwe otkri}e Dragoslava Srejovi}a (Summary: Gamzigrad Culture – the Last Discovery of Dragoslav Srejovi}), Radovi Dragoslava Srejovi}a na istra`ivawu praistorije centralnog Balkana (urednik N. Tasi}), Kragujevac 1998, 147–158. Lippold 1950 G. Lippold, Die griechische Plastik, München 1950. Lugli 1957 G. Lugli, La tehnica edilicia romana con particolare reguarde a Roma e Lazio, Roma 1957. Ma~aj 1882 S. Ma~aj, Crnore~ki okrug, Glasnik srpskog u~enog dru{tva 73, Beograd 1882. Madas 1970 D. Madas, [avac, Lapovo – anti~ko nalazi{te, Arheolo{ki pregled 12, Beograd 1970, 133–134. 227 BIBLIOGRAPHY Majnari}-Pand`i} 1996 N. Majnari}-Pand`i}, Nekoliko napomena o uvo|enju ranolatenskog stila u severnu Hrvatsku i Bosnu (Summary: Several remarks on the introduction of the Early La Tène style in northern Croatia and Bosnia), Arheolo{ki radovi i rasprave 12, Zagreb 1996, 31–53. Mamford 1988 L. Mamford, Grad u historiji (preveo V. Ivir), Zagreb 1988. Mango 1959 C. Mango, The Brazen House; A Study of the Vestibule of the Imperial Palace at Constantinople, Kôbenhavn 1959, 98–107. Maneva 1992 E. Maneva, Srednovekoven nakit od Makedonija (Summary: The Medieval Jewellery of Macedonia), Skopje 1992. Mano-Zisi 1959 \. Mano-Zisi, Prologomena uz probleme kasnoanti~kih mozaika u Ilirikumu (Résumé: Mosaïques de l’epoque du bas-empire en Illyricum), Zbornik radova Narodnog muzeja II, Beograd 1959, 83–109. Marasovi} 1968 T. i J. Marasovi}, Dioklecijanova pala~a, Zagreb 1968. Marasovi}, Mcnally 1972 J. Marasovi}, Sh. Mcnally, Izve{taj o jugoslovensko-ameri~kom projektu istra`ivanja jugoisto~nog dijela Pala~e, I dio, Urbs, Split 1972, 21–23. Marasovi}, Mcnally 1977 J. i T. Marasovi}, Sh. Mcnally, Dioklecijanova pala~a. Izve{taj o jugoslovensko-ameri~kom projektu istra`ivanja, II dio, Urbs, Split 1977, 3–68. Marasovi} 1982 T. Marasovi}, Dioklecijanova pala~a. Svjetska kulturna ba{tina, Beograd 1982. Martin 1968 R. Martin, Sculpture et peinture dans les façades monumentales au IVe siècle avant J.C., Revue Archéologique 1, Paris 1968. Ma{kin 1978 N. A. Ma{kin, Istorija starog Rima, Beograd 1978. Mari} 1963 Z. Mari}, Keltski elementi u mla|em `eleznom dobu Bosne i Hercegovine (Résumé: Eléments celtiques de la fin de l’ âge du fer en Bosnie-Herzégovine), Glasnik Zemaqskog muzeja XVIII, Sarajevo 1963, 67–70. McKay 1984 A. G. McKay, Römische Häuser, Villen und Paläste, Feldmeilen 1984. Medovi} 2008 A. Medovi}, Gamzigradski ratari – dva koraka napred, jedan korak nazad (Summary: Gamzigrad farmers – two steps forward, one step back), Rad Muzeja Vojvodine 50, Novi Sad 2008, 151–173. Mihailov 1993 S. Mihailov, Novi danni za Golÿmata bazilika v Pliska, Pliska – Preslav 6, Sofiÿ 1993, 22–32. Miki} 2009 @. Miki}, Paleodemografska interpretacija sredwovekovne nekropole izvan isto~ne kapije Gamzigrada (Felix Romuliana) (Summary: Paleodemographic Interpretation of the Medieval Cemetery Outside the Eastern Gate of Gamzigrad (Felix Romuliana)), Balcanica 39, Beograd 2009. Mili} 1988 B. Mili}, Razvoj grada kroz stole}a, Zagreb 1988. Mirkovi} 1968 M. Mirkovi}, Rimski gradovi na Dunavu u Gornjoj Meziji (Zusammenfassung: Römische Städte an der Donau Oberrmösien, Dissertationes VI, Beograd 1968. Marjanovi}-Vujovi} 1987 G. Marjanovi}-Vujovi}, Krstovi od VI do XII veka iz zbirke Narodnog muzeja / Crosses 6th–12th C. from the Collection of National Museum, Beograd 1987. Mir~eva 2004 E. Mir~eva, Fibula süs zoomorfen prüsten ot Velikot Trnovsko (Summary: A fibula with zoomorphic ring from the Veliko Turnovo Distrikt), Izvestiÿ na Nerodni Muzeé 36/37, Varna 2004, 164–170. Martin 1956 R. Martin, L’urbanisme dans le Grece antique, Paris 1956. Mocsy 1974 A. Mocsy, Pannonia and Upper Moesia, London 1974. 228 BIBLIOGRAPHY Moreau 1954 J. Moreau, Lactance, De la mort des persécuteurs I–II, Commentaire de J. Moreau (Sources chrétiennes N° 39), Paris 1954. Petkovi} 2004 S. Petkovi}, Arheolo{ka iskopavawa na lokalitetu Romuliana – Gamzigrad u ju`noj kuli zapadne kapije mla|eg utvr|ewa u 2002. godini, Glasnik SAD 20, Beograd 2004, 127–153. Nenadovi} 1963 S. Nenadovi}, Bogorodica Qevi{ka, wen postanak i weno mesto u arhitekturi Milutinovog vremena, Beograd 1963. Petkovi} 2008 A S. Petkovi}, Iskopavawa na lokalitetu Gamzigrad – Romuliana 2004. godine, Arheolo{ki pregled, nova serija 2/3 (2004/2005) / Archaeological Excavations in Gamzigrad – Romuliana 2004, Archaeological Reports, New Series 2/3 (2004/2005), Beograd 2008, 61–63. Nikolajevi}-Stojkovi} 1957 I. Nikolajevi}-Stojkovi}, Ranovizantijska arhitektonska dekorativna plastika u Makedoniji, Srbiji i Crnoj Gori, Beograd 1957. Nikolajevi} 1965 I. Nikolajevi}, Chapiteaux d’ordre corinthien de Sirmium, Acten des VII Internationalen Kongresses für Christliche Archäologie, Trier 1965, 653–660. Nikoli}, \uri~i} 1997 D. Nikoli}, S. \uri~i} Rezultati sonda`nog istra`ivawa eneolitskog naseqa Beligovo (Summary: Results of the Sondage of the Aeneolithic Settlement of Beligovo), Glasnik Srpskog arheolo{kog dru{tva 13, Beograd 1997, 79–88. Novakovi} 1908 S. Novakovi}, Ohridska arhiepiskopija u po~etku XI veka, Glas Srpske kraqevske akademije 76, Beograd 1908, 1–62. Ostrogorski 1959 G. Ostrogorski, Istorija Vizantije, Beograd 1959. Papazoglu 1969 F. Papazoglu, Srednjobalkanska plemena u predrimsko doba, Centar za balkanolo{ka ispitivanja, knj. 1, Sarajevo 1969. Paribeni di 1940 R. di Paribeni, Le dimore dei potentiores nel basso impero, Mitteilungen des Deutschen archeologischen Instituts, Römische Abt. 55, 1940, 131–148. Petkovi} 2008 B S. Petkovi}, Iskopavawa na lokalitetu Gamzigrad – Romuliana 2005. godine, Arheolo{ki pregled, nova serija 2/3 (2004/2005) / Archaeological Excavations in Gamzigrad – Romuliana 2005, Archaeological Reports, New Series 2/3 (2004/2005), Beograd 2008, 64–67. Petkovi} 2008 C S. Petkovi}, Fibule u rimskim provincijama na tlu Srbije od I do IV veka n.e. (Fibulae in Roman Provinces on the Territory of Serbia, I–IV Century A.D.), rukopis doktorske teze, Filozofski fakultet, Beograd 2008. Petkovi} 2009 S. Petkovi}, Late Roman Necropolis of Romuliana, Area South to the Fortified Palace (Research 2005–2006) (Rezime: Kasnoanti~ka nekropola Romulijane, prostor ju`no od utvr|ene palate (istra`ivawa 2005–2006)), Starinar LVIII (2007), Beograd 2009, 251–275. Petkovi}, @ivi} 2006 S. Petkovi}, M. @ivi}, Grupa bronzanih fibula sa Romulijane (Summary: The Group of the Bronze Fibulae from Romuliana), Zbornik radova Narodnog muzeja XVIII/1, Beograd 2006, 439–458. Petkovi} 1999 S. Petkovi}, Meaning and Provenance of Horses’ Protomes Decoration on the Roman Antler Combs, Starinar XLIX (1998), Beograd 1999, 213–228. Petkovi}, @ivi} 2006 A S. Petkovi} – M. @ivi}, Tragovi metalur{kih aktivnosti u kasnoanti~koj Romulijani: istra`ivawa 2002–2005 (Summary: Traces of Metallurgical Activities in Late Roman Romuliana: Research 2002–2005), Glasnik SAD 22/2006, Beograd 2006, 135–148. Petkovi} 2003 S. Petkovi}, ^e{qevi od jeleweg roga iz ju`ne kule zapadne kapije mla|eg utvr|ewa Romulijane, Razvitak XLIII, br. 211–212, Zaje~ar 2003, 35–40. Petkovi}, @ivi}, Kapuran 2009 S. Petkovi}, M. @ivi}, A. Kapuran, Iskopavawa na lokalitetu Gamzigrad – Romuliana u 2007. godini, Arheolo{ki pregled 5 (2007), Beograd 2009 (u {tampi). 229 BIBLIOGRAPHY Petrikovits von 1971 H. von Petrikovits, Fortifications in the North–Western Roman Empire from the Third to the Fifth Centuries A.D., Journal of Roman Studies 61, London, 1971, 178–218. Pröttel 1991 P. M. Pröttel, Zur Chronologie der Zwiebelknopffibeln, Jahrbuch des Römisch–Germanischen Zentralmuseums Mainz 35/1 (1988), Mainz 1991, 347–372. Petrovi} 1975 P. Petrovi}, Paleografija rimskih natpisa u Gorwoj Meziji (Résumé: Paléographie des inscriptions romaines en Mésie Supérieure), Beograd 1975. Reasch 1984 J. J. Reasch, Das Maxentius-Mausoleum an der Via Appia in Rom, Mainz 1984. Picard 1962 Ch. Picard, Les portes sculptée à images divines, Paris 1962. Picard 1962 A Ch. Picard, La sculpture antique. Origines de la sculpture byzantine II, Paris 1962. PLRE I A.H.M. Jones – J.R. Martindale – J. Morris, The Prosopography of the Later Roman Empire, Volume I: A.D. 260–395, Cambridge – London – New York – Melbourne 1975. Reusch 1966 W. Reusch, Wandmalerei und Mosaikboden eines Peristylhauses in Bereich der Trierer Kaiserthermen, Trier Zeitschrift für Gesch. und Kunst des Trierer Landes und seiner Nachbargebiete 29, Trier 1966, 187–235. Richmond 1955 J.A. Richmond, Britain and Roman military antiquities, Proceedings of the British Academy V, XLI, London 1955. Roman, Dodd-Opriþescu, János 1992 P. Roman, A. Dodd-Opriþescu, P. János, Beiträge zur Problematik der schnurverzierten Keramik Südosteuropas, Internationale Interakademische Kommission für die Erforschung der Vorgeschichte des Balkans, Monographie III, Mainz 1992. Popovi} I. 1997 I. Popovi}, Miscellanea Argentea, Starinar XLVIII, 1997, 73–90. Popovi} I. 2006 I. Popovi}, Trépied et vaisselle en argent provenant du monument consécrateur n°1 à Magura, Felix Romuliana. 50 Years of Archaeological Excavations. Papers from the International Conference. Zaje~ar, 27th–29th October 2003 (Ed. M. Vasi}), Belgrade 2006, 55–68. Salies 1974 G. Salies, Untersuchungen zu den geometrischen Gliederungsschemata römischer Mosaiken, Bonner Jahrbücher 174, Bonn 1974, 4–34. Schlibach 1970 E. Schlibach, Byzantinische Metrologie, Handbuch der Altertumswissenschaft XII, München 1970. Popovi} I. 2008 I. Popovi}, Galerijev portret iz Romulijane u svetlu predstava carskih i sve{teni~kih insignija u rimskoj umetnosti (Summary: Galerius’ Portrait from Romuliana in the Light of the Representations of the Emprerors and Priests in Roman Art), Dragoslav Srejovi} i umetnost, Memorijal Dragoslava Srejovi}a, Zbornik radova 3 (urednik V. Jovi}), Kragujevac, 2008, 105–119. Schlunk 1970 H. Schlunk, Beiträge zur Kunstgeschichtlichten Stellung Toledos im 7. Jahrhundert, Madrider Mitteilungen 13, Madrid 1970, 161–169. Schulze 1909 R. Schulze, Die römische Stadttore, Bonner Jahrbücher 118, Bonn 1909, 280–352. Popovi} P. 1996 P. Popovi}, Early La Tène Between Pannonia and the Balkans, Starinar XLVII, Beograd 1996, 105–125. Price 1987 S. Price, From Noble Funerals to Divine Cult: the Consecration of Roman Emperors, Rituals of Royality (ed. D. Cannadine, S. Price), Cambridge 1987, 56–105. Schulten 1979 P. N. Schulten, Die Typologie der römischen Konsekrationsprägungen, Frankfurt 1979. Séchan, Lévêque 1966 L. Séchan, P. Lévêque, Les grandes divinités de la Grèce, Paris 1966. 230 BIBLIOGRAPHY Seston 1946 W. Seston, Dioclétien et la Tétrarchie. I Guerres et réformes (284–300), Paris 1946. the Late Classical Times), Galerija SANU 45, Beograd 1983, 19–21; 194. Srejovi} 1983 B D. Srejovi}, Rimsko poqsko imawe, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of the Late Classical Times), Galerija SANU 45, Beograd 1983, 21–23; 194. Seston 1950 W. Seston, Jovius et Herculius ou l’épiphanie des Tétrarcques, Historia I 1950, 257–266. Skok 1971 P. Skok, Etimologijski rje~nik hrvatskog ili srpskog jezika, Zagreb 1971. Srejovi} 1983 C D. Srejovi}, Carski dvorac, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of the Late Classical Times), Galerija SANU 45, Beograd 1983, 24–95; 194–199. Smith 1956 B. E. Smith, Architectural Symbolism of Imperial Rome and the Middle Ages, Princeton Monographs in Art et Archaeology XXX, Princeton – New York, 1956. Srejovi} 1985 D. Srejovi}, Felix Romuliana, Galerijeva palata u Gamzigradu (Résumé: Felix Romuliana. Le palais de Galère à Gamzigrad), Starinar XXXVI, Beograd 1985, 51–67. Sladi} 2003 M. Sladi}, Tragom ranih keltskih uticaja na prostoru Timo~ke Krajine (Summary: On tracing early celtic influences in Timo~ka Krajina, Balcanica XXXII/XXXIII (2001/02), Beograd 2003, 37–47. Srejovi} 1986 D. Srejovi}, Dve kasnoanti~ke carske palate (Summary: Two Late Classical Palaces in Yugoslavia: Diocletianus’ Palace in Split and Galerius’ Palace in Gamzigrad), Glas SANU CCCXLIII, Odeqewe istorijskih nauka, kw. 5, Beograd 1986, 9–25. Sladi} 2005 M. Sladi}, Vi{icina ba{ta – selo Gamzigrad: naseqe poznog latena u dolini Crnog Timoka (Summary: Vi{icina ba{ta – Gamzigrad village: Settlement of La Tène Culture in Valley of the Black Timok River), Glasnik Srpskog arheolo{kog dru{tva 21, Beograd 2005, 211–232. Srejovi} 1986 A D. Srejovi}, Felix Romuliana: Carska palata ili …? (Résumé: Felix Romuliana: Palais impériale ou …?) Starinar XXXVII, Beograd, 1986, 87–102. Sladi}, Jovanovi} 1977 M. Sladi}, S. Jovanovi}, Ostaci starijeneolitskih naseqa sa podru~ja Kwa`evca (Summary: Traces of the Early Neolithic Settlements in the Area of Knja`evac), Arheologija isto~ne Srbije / Archaeology of Eastern Serbia (urednik M. Lazi}), Beograd 1977, 167–175. Srejovi} 1993 A D. Srejovi}, Felix Romuliana. Galerijev ideolo{ki testament, Rimski carski gradovi i palate (priredio D. Srejovi}), Beograd 1993, 31–53, kat. 232. Sladi}, Ru`i} 2001 M. Sladi}, M. Ru`i}, Miletov bunar, novo naseqe gamzigradske kulture (Summary: Miletov Bunar, New Settlement of Gamzigrad Culture), Glasnik Srpskog arheolo{kog dru{tva 17, Beograd 2001, 153–168. Srejovi} 1993 B D. Srejovi}, Felix Romuliana. The Ideological Testament of Emperor Galerius, Roman Imperial Towns and Palaces in Serbia (ed. D. Srejovi}), Belgrade 1993, 31–53, cat. 232. Srejovi} 1983 D. Srejovi}, Uvod, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace in the Late Classical Times), Galerija SANU 45, Beograd, 4–16; 193–194. Srejovi} 1993 C D. Srejovi}, Catalogue, Roman Imperial Towns and Palaces in Serbia, Belgrade 1993, 232, cat. 71. Srejovi} 1983 A D. Srejovi}, Gamzigrad u praistoriji, Gamzigrad. Kasnoanti~ki carski dvorac (Gamzigrad. An Imperial Palace of Srejovi} 1993 D D. Srejovi}, Carski mauzolej u Gamzigradu (Felix Romuliana), Razvitak XXXIII/1–2, Zaje~ar 1993, 4–8. 231 BIBLIOGRAPHY Srejovi} 1993 E D. Srejovi}, U Gamzigradu (Romulijani) prona|ena glava statue rimskog imperatora Galerija, Razvitak XXXIII/1–2, Zaje~ar 1993, 64–65. Srejovi}, Vasi} 1994 A D. Srejovi}, ^. Vasi}, Emperor Galerius’ Buildings in Romuliana (Gamzigrad, Eastern Serbia), Antiquité Tardive 2, 123–141. Srejovi} 1994 A D. Srejovi} A Porphyry Head of a Tetrarch from Romuliana (Gamzigrad), Starinar XLIII–XLIV (1992–1993), Beograd 1994, 41–47. Srejovi}, Lazi} 1997 D. Srejovi}, M. Lazi}, Naseqa i nekropole bronzanog doba u Timo~koj Krajini (Summary: The Bronze Age Settlements and Cemeteries in Timo~ka Krajina), Arheologija isto~ne Srbije / Archaeology of Eastern Serbia (urednik M. Lazi}), Beograd 1997, 225–247. Srejovi} 1994 B The representation of Tetrarchs in Romuliana, Antiquité Tardive 2, Paris 1994, 143–152. Stein 1968 E. Stein, Histoire du Bas-empire I, Amsterdam 1968. Srejovi} 1995 A D. Srejovi}, Imperial Head Unearthed in Serbia, Archaeology 48 (1), New York 1995, 14–15. Stojkovi}-Pavelka 1988/89 B. Stojkovi}-Pavelka, Veliki hram u Gamzigradu. Predlog za rekonstrukciju (Résumé: Le grand temple à Felix Romuliana (Gamzigrad)), Saop{tewa XX–XXI, Beograd 1988/89, 135–145. Srejovi} 1995 B D. Srejovi}, Diva Romula, Divus Galerius, The Age of Tetrarchs. A Symposium held from the 4th to the 9th October 1993 (ed. D. Srejovi}), Belgrade 1995, 297–310. Stri~evi}, Suboti} 1959 \. Stri~evi}, G. Suboti}, Iskopavawe Zaweva~ke crkve (Résumé: Fouilles de l’église de Zanjevac), Starinar IX–X (1958–1959), Beograd 1959, 307–315. Srejovi} 1995 C D. Srejovi}, Felix Romuliana (Catalogue), Belgrade 1995. Srejovi} 1995 D D. Srejovi}, Diva Romula – Divus Galerius, Sun~ani sat 5, Sremska Mitrovica 1995, 17–30. Strzygowski 1906 J. Strzygowski, Spalato, ein Markstein der romanischen Kunst bei ihrem Übergange von Orient nach Abendlande, Studien aus Kunst und Geschichte Fr. Schneider gewidmet, Freiburg in Breisgau 1906, 325–335. Srejovi} 1996 D. Srejovi}, Felix Romuliana, la résidence de l’empereur Galère, Dossiers d’archéologie 319, Dijon 1996, 20–29. Sui} 1976 M. Sui}, Anti~ki grad na isto~nom Jadranu, Zagreb 1976. Srejovi} 1998 D. Srejovi}, Ogledi o drevnoj umetnosti, Beograd 1998. Swoboda 1924 K.M. Swoboda, Römische und romanische Paläste, Wien 1924. Srejovi}, Cermanovi} 1979 D. Srejovi}, A. Cermanovi}, Re~nik gr~ke i rimske mitologije, Beograd 1979. Srejovi}, Lalovi}, Jankovi} 1981 D. Srejovi}, A. Lalovi}, \. Jankovi}, Gamzigrad (Résumé: Gamzigrad), Starinar XXXI (1980), Beograd 1981, 65–80. Srejovi}, Vasi} 1994 D. Srejovi}, ^. Vasi}, Imperial Mausolea and Consecration Memorials in Felix Romuliana (Gamzigrad, East Serbia), Belgrade 1994. Swoboda 1958 E. Swoboda, Carnuntum, Römische Forschungen in Niederösterreich I, Graz – Köln 1958. Swoboda 1961 K. M. Swoboda, The Problem of the Iconography of Late Antique and Early Medieval Palaces, Journal of the Society of Architectual Historians XX, 1961, 78–89. Szabo 1971 M. Szabo, Kelták Niomában Magiarországon, Budapest 1971. 232 BIBLIOGRAPHY Szabo 1971 A M. Szabo, Sur les Celtes en Hongrie, Budapest 1971. Truhelka 1904 ]. Truhelka, Der vorgeschichtliche Pfahlbanin Savabete bei Donja Dolina, Wissenschaftliche Mitteilungen aus Bosnien und der Hercegovina IX, Wien 1904, 3–27. Szabo 1975 M. Szabo, Contribution a l’étude de l’art et de la chronologie de La Tène ancienne en Hongrie, Folia Archaeologica XXV, Budapest 1975, 71–86. Underwood 1948 A. Underwood, Some Principles of Measure in the Architecture of the Period of Justinian, Cahiers archéologique III, Paris 1948, 64–74. Szabo 1992 M. Szabo, Les Celtes de l’est, Paris 1992. Vasi} ^. 1993 A ^. Vasi}, Felix Romuliana. Carski mauzoleji i konsekrativni spomenici na lokalitetu Magura (Karaula), Rimski carski gradovi i palate (priredio D. Srejovi}), Beograd 1993, 148–163. Szabo 2001 M. Szabo, Celtes de Hongrie Xe–Ie siècle av. J.-C., Paris 2001. Tasi} 1979 N. Tasi}, Bubanj – Salkuþa–Krivodol kompleks, Praistorija jugoslavenskih zemalja III – Eneolit (Urednik A. Benac), Sarajevo 1979, 87–114. Vasi} ^. 1993 B ^. Vasi}, Felix Romuliana. The Imperial Mausolea and Consecration Memorials on the Site Magura (Karaula), Roman Imperial Towns and Palaces in Serbia (ed. D. Srejovi}), Belgrade 1993, 148–163. Tasi} 1997 N. Tasi}, Eneolit i bronzano doba severoisto~ne Srbije (Summary: Eneolithic and Early Bronze Ages in the North–East Serbia), Arheologija isto~ne Srbije / Archaeology of Eastern Serbia (urednik M. Lazi}), Beograd 1997, 79–89. Vasi} ^. 1993 S ^. Vasi}, Felix Romuliana. Galerijeva palata u Gamzigradu, Felix Romuliana. Galerijev ideolo{ki testament, Rimski carski gradovi i palate (priredio D. Srejovi}), Beograd 1993, 118–147. Tasi} 2004 N. Tasi}, Nalazi{ta i kulture iz eneolitskog perioda / Eneolithic Cultures and Sites in the Bor Area and its Environs, Bor i okolina u praistoriji, antici i sredwem veku / The Bor Area in Prehistory, Antiquity and the Middle Ages (urednik M. Lazi}), Bor – Beograd 2004, 57–99. Vasi} ^. 1993 D ^. Vasi}, Felix Romuliana. Galerius’ Palace at Gamzigrad, Roman Imperial Towns and Palaces in Serbia (ed. D. Srejovi}), Belgrade 1993, 118–147. Thomas 1964 E. Thomas, Römische Villen in Pannonien, Budapest 1964. Vasi} ^. 1993 E ^. Vasi}, Katalog, Rimski carski gradovi i palate (priredio D. Srejovi}), Beograd 1993, 189. Tomovi} 1990 G. Tomovi}, Glagoqski natpis sa ^e~ana (Résumé: Inscription glagolitique), Istorijski ~asopis / Revue historique XXXVII, Beograd 1990, 5–18. Vasi} ^. 1993 F ^. Vasi}, Catalogue, Roman Imperial Towns and Palaces in Serbia (ed. D. Srejovi}), Belgrade 1993, 189. Tóth 1973 E. Tóth, Imperial Palace at Savaria, Acta Archaelogica Academiae Scientiarum Hungaricae 25, Budapest 1973, 117–137. Vasi} ^. 1995 ^. Vasi}, Chronological Relations of Palace and Fortification System of Gamzigrad, The Age of Tetrarchs (ed. D. Srejovi}), Belgrade 1995, 315–323. Tóth 1978/79 E. Tóth, Über einen spätantiken Prunksaaltyp, Schrift von Steiner 15/16, Festschrift Modrijan, Graz 1978/79, 189–195. Trésors 1989 Trésors d’orfèvrerie gallo-romains (ed. F. Baratte, K. Painter), Paris 1989. Vasi} ^. 1997 ^. Vasi}, Odbrambeni sistemi Gamzigrada, rukopis doktorske teze, Filozofski fakultet, Beograd 1997. 233 BIBLIOGRAPHY Vasi} M. 1997 M. Vasi}, Late Roman Bricks with Stamps from the Fort Transdobeta, Mélanges d’histoire et d’épigraphie offerts à Fanula Papazoglu, Beograd 1997, 149–177. Ward-Perkins 1958 J. B. Ward-Perkins, Notes on the Structure and Building Methods of Early Byzantine Architecture, D. Talbot Rice, The Great Palace of the Byzantine Emperor, Edinburgh 1958, 52–104. Vasi} M. 2006 M. Vasi}, Stibadium in Romuliana and Mediana, Felix Romuliana. 50 Years of Archaeological Excavations. Papers from the International Conference. Zaje~ar, 27th–29th October 2003 (Ed. M. Vasi}), Beograd 2006, 69–75. Ward-Perkins 1974 J. B. Ward-Perkins, Architettura Romana, Torino 1974. Waywell 1979 E. Waywell, Roman Mosaics in Greece, American Journal of Archaeology 83/3, Princeton, New Jersey 1979, 293–321. Vasi} M. 2007 M. Vasi}, Felix Romuliana (Gamzigrad) – Palast und Gedenkmonument des Kaisers Galerius, Roms Erbe auf dem Balkan. Spätantike Kaiservillen und Stadtanlagen in Serbien (Hrsg. U. Brandl, M. Vasi}), Mainz 2007, 33–58. Vasi} M. 2008 A M. Vasi}, Zlatni i srebrni novac kasne antike (284–450. godine) iz zbirke Narodnog muzeja u Beogradu, Beograd 2008. Vasi} M. 2008 B M. Vasi}, Gold and Silver Coins of Late Antiquity (284–407 AD) in the Collection of National Museum in Belgrade, Belgrade 2008. Vasi} M. 2009 M. Vasi}, Findings of Coins from Romuliana, Unearthed in 2005 and 2006, Outside the Fortified Palace (Rezime: Nalazi novca iz Romulijane iskopani 2005. i 2006. godine izvan utvr|ene palate), Starinar LVIII (2007), Beograd 2009, 309–314. Vasi} R. 1997 R. Vasi}, Starije gvozdeno doba na podru~ju isto~ne Srbije (Summary: Early Iron Age in the Eastern Serbia), Arheologija isto~ne Srbije / Archaeology of Eastern Serbia (urednik M. Lazi}), Beograd 1997, 91–100. VIINJ III 1966 Vizantijski izvori za istoriju naroda Jugoslavije III (urednici G. Ostrogorski, F. Bari{i}), Beograd, 1996. Wulf-Rheidt 2007 U. Wulf-Rheidt, Residentien in Rom oder in der Provinz? Der Kaiserpalast Felix Romuliana im Spiegel der tetrarchischen Residenzbaukunst, in: Roms Erbe auf dem Balkan. Spätantike Kaiservillen und Stadtanlagen in Serbien (Hrsg. U. Brandl, M. Vasi}), Mainz 2007, 59–79. Yegül 1992 F. Yegül, Baths and Bathing in Classical Antiquity, New York, Cambridge, Massachusetts 1992, 43. @eravica 1975 Z. @eravica, Glasgegenstände der Lokalität Popovica (Ostserbien) aus dem XI–XII Jahrhundert, Sredwevekovno staklo na Balkanu (V–XV vek), Beograd 1975, 53–62. @ivi} 1997 M. @ivi}, Prstewe iz sredwovekovne zbirke nakita Narodnog muzeja u Zaje~aru (Summary: Rings from Medieval Collection of the National Museum at Zaje~ar), Arheolo{ka istra`ivawa Isto~ne Srbije / Archaeology of Eastern Serbia (urednik M. Lazi}), Beograd 1997, 329–338. @ivi} 2003 M. @ivi}, Felix Romuliana. 50 godina odgonetawa / M. @ivi}, Felix Romuliana. 50 Years of Solving, Beograd / Belgrade 2003. Vitruvius, Marcus Polio 1990 Marcus Vitruvius Pollio, Deset knjiga o arhitekturi (prevod M. Lopac), Sarajevo 1990. @ivi} 2005 M. @ivi}, Catalogo, Constantino il Grande, La civiltà antica al bivio tra Occidente e Oriente, Milano 2005, 204/205, 260, 261, 300, cat. 2, 89, 90, 151. Vuksan 1997 M. Vuksan, Sredwovekovni nalazi iz Lazareve pe}ine u Zlotu (Summary: Medieval Finds from the Lazareva Cave at Zlot), Arheolo{ka istra`ivawa Isto~ne Srbije / Arhaeology of Eastern Serbia (urednik M. Lazi}), Beograd 1997, 293–301. @ivi} 2007 M. @ivi}, Begleit – CD zum Katalog, Imperator Caesar Flavius Constantinus. Konstantin der Große (Augsstellungskatalog, Landesausstellung Trier 2007), Mainz 2007, Kat. I. 4. 12; I. 5. 11; I. 5, 12; I. 5. 13; IV. 1. 47. 234 The book was published with a support of the Ministry of Education and Science of the Republic of Serbia (project numbers 177007, 177032, 147034, 147041), Ministry of Culture, Media and Information Society of the Republic of Serbia and Ministry of Religions and Diaspora of the Republic of Serbia Serbian edition of this book has received at the 55th International Book Fair in Belgrade (October 2010) the plaque of the Ministry of Science and Technological Development for special contribution in the scientific-research work in the field of humanities. CIP - Katalogizacija u publikaciji Narodna biblioteka Srbije, Beograd 904"652"(497.11)(082) FELIX Romuliana – Gamzigrad / authors Ivana Popovi} ... [i dr.] ; urednik Ivana Popovi} ; [translation Mirjana Vukmanovi} ; photographs Neboj{a Bori}, Vladimir Popovi}, Branislav Strugar ; site plans Aleksandar Kapuran ; drawings Milovan Risti}, Aleksandar Kapuran]. – Belgrade : Institute of Archaeology : Faculty of Orthodox Theology University, Institute for Theological Research ; Zaje~ar : Diocese of Timok, 2011 (Beograd : Glasnik). – 234 str. : ilustr. ; 29 cm Tekst {tampan dvostuba~no. – Tira` 600. – Str. 7–9: Foreword / Ivana Popovi}. – Napomene i bibliografske reference uz tekst. – Bibliografija: str. 221–234. ISBN 978-86-80093-73-4 (IA) ISBN 978-86-7405-102-3 (ITR) 1. Popovi}, Ivana, 1955– [autor] [urednik] [autor dodatnog teksta] a) Arheolo{ka nalazi{ta, rimska – Gamzigrad – Zbornici b) Arheolo{ki nalazi, rimski – Gamzigrad – Zbornici c) Romulijana – Zbornici COBISS.SR-ID 186547724