LRCW 4 Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, 2 volume set: Archaeology and archaeometry. The Mediterranean: a market without frontiers 9781407312491, 9781407312507, 9781407312514, 9781407342184

Citation preview

LRCW 4 Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean Archaeology and archaeometry The Mediterranean: a market without frontiers Edited by

Natalia Poulou-Papadimitriou, Eleni Nodarou and Vassilis Kilikoglou

Volume I

BAR International Series 2616 (I) 2014

ISBN 9781407312491 (Volume I) paperback ISBN 9781407312507 (Volume II) paperback ISBN 9781407312514 (Volume set) paperback ISBN 9781407342184 (Volume set) e-format DOI https://doi.org/10.30861/9781407312514 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

PSYCHA FOUNDATION

Centre for Byzantine Research

CONTENTS VOLUME I INTRODUCTION ........................................................................................................................................................ ix PREFACE, by A. MENTZOS ......................................................................................................................................... xi

ARCHAEOLOGY AND ECONOMIC HISTORY L. PALMIERI Container and content in North Africa: a new glimpse of the relationship between foodstuffs production and amphorae production (4th-6th centuries AD) .................................................................................... 1 E. VACCARO Patterning late Roman ceramic exchange in southern Tuscany (Italy): the coastal and inland evidence, i.e. centrality vs. marginality ................................................................................ 11 P. MILLS, U. RAJALA Supply and distribution of late Roman coarsewares from the Nepi Survey Project................................................. 27 G. W. TOL, P. ATTEMA A road station on the Tabula Peutingeriana. Excavations at Astura ........................................................................ 39 C. S. MARTUCCI, G. F. DE SIMONE, G. BOEMIO, G. TROJSI Local production and trade patterns in the environs of Vesuvius: the pottery from Pollena Trocchia and Nola ............................................................................................................ 51 F. A. CUTERI, M. T. IANNELLI, P. VIVACQUA, T. CAFARO Da Vibo Valentia a Nicotera. La ceramica tardo romana nella Calabria tirrenica (Italia) ....................................... 63 V. CAMINNECI, A. AMICO, F. GIANNICI, R. GIARRUSSO, A. MULONE Ceramiche comuni e da fuoco dall’insediamento tardoantico di Carabollace (Sciacca, Sicilia, Italia): caratterizzazione tipologica e archeometrica .................................................................... 81 E. TZAVELLA, C. TRAINOR, M. MAHER Late Roman pottery from the Sikyon Survey Project: local production, imports, and urban evolution (4th-7th c. AD) (Greece) ......................................................................................................... 91

PRODUCTION CENTERS J. NACEF Nouveaux témoignages sur la production de la céramique antique du Sahel tunisien ............................................103 H. MÖLLER, A.-K. RIEGER Pottery from the Eastern Marmarica (Egypt): a semi-arid region as producer and mediator in Ptolemaic, Roman and late Roman time .............................................................................................................113 S. GUPTA-AGARWAL Cultural transmission and consumer demand: a case study using ceramics from Karanis, Egypt ..........................125 P. MILLS, P. REYNOLDS Amphorae and specialized coarsewares of Ras al Bassit, Syria: local products and exports ..................................133 S. JAPP Micaceous waterjars in Pergamon – a specific variant of the Late Roman 3 amphora ...........................................143 i

A. B. BIERNACKI, E. KLENINA Early Byzantine coarse ware from Novae: production and chronology ..................................................................151 E. PAPAVASSILIOU, K. SARANTIDIS, I. PAPANIKOLAOU A ceramic workshop of the early byzantine period on the island of Lipsi in the Dodecanese (Greece): a preliminary approach...........................................................................................................................................159 S. DIDIOUMI Local pottery production in the island of Cos, Greece from the early Byzantine period. A preliminary report................................................................................................................................................169 CH. DIAMANTI, K. KOUZELI, P. PETRIDIS Archaeology and archaeometry in late Roman Greece: the case of mainland and insular settlements, workshops and imports ...........................................................................................................................................181 E. GEROUSI A late Roman workshop at Dilesi in Boeotia ..........................................................................................................193 F. CANTINI, G. BOSCHIAN, M. GABRIELE Empoli, a late Antique pottery production centre in the Arno valley (Florence, Tuscany, Italy) ...........................203 M. S. RIZZO, L. ZAMBITO, F. GIANNICI, R. GIARRUSSO, A. MULONE Anfore di tipo siciliano dal territorio di Agrigento .................................................................................................213 S. AMARI Observations on the late Roman coarse wares production from the factory of Acium in Sicily (Italy) .......................................................................................................................225

DISTRIBUTION AND CONSUMPTION C. CAPELLI, M. BONIFAY (INVITED) Archéométrie et archéologie des céramiques africaines: une approche pluridisciplinaire, 2. Nouvelles données sur la céramique culinaire et les amphores...............................................................................235 F. KENKEL The late Roman pottery from Tall Zirāʿa in northern Jordan – common ware and trade material of a village in the area of influence of the Decapolis-cities ....................................................................................255 G. ZENONI New stratigraphical contexts for the study of the late pottery of Palmyra ..............................................................261 A. C. TURKER Roman and Early Byzantine amphorae from the archaeological survey in Pamphylia and Tracheia Cilicia ................................................................................................................................................271 E. M. DOKSANALTI, M. TEKOCAK Cooking pots: production in Knidos in the light of recent excavations ..................................................................281 S. AKRIVOPOULOU, P. SLAMPEAS Late Roman Coarse Ware from a rescue excavation in Thessaloniki .....................................................................287 K. FRAGOULIS, D. MINASIDIS, A. MENTZOS Pottery from the cemetery basilica in the early byzantine city of Dion ..................................................................297 V. D. KOROSIS Transport and storage vessels and coarse wares from Megara, Attica (Greece). The testimony of the pottery concerning the city during the early Byzantine period .............................................305 ii

M. CAVALAZZI, E. FABBRI Cooking ware from 5th to 7th century in Emiglia Romagna: two case studies ......................................................313 L. TONIOLO, A. M. DE FRANCESCO, E. ANDALORO, R. SCARPELLI Napoli tardo-romana: studio morfo-tipologico e analisi archeometrica preliminare del vasellame da fuoco e da mensa dal contesto dei Girolomini .............................................................................323 L. MIRABELLA, S. SIANO Cooking ware and amphoras from the late Roman Villa Rustica in Fisciano (Salerno, Italy)................................335 M. C. PARELLO, A. AMICO, F. GIANNICI, R. GIARRUSSO, A. MULONE La ceramica comune e da fuoco di età tardoromana dai siti di Verdura (Sciacca) e Canalicchio (Calamonaci) (Agrigento, Sicilia, Italia) ..........................................................................................343

TYPOLOGY AND CHRONOLOGY F. TRAPANI The coin hoard from Misurata (Libya): the containers ...........................................................................................355 R.-P. GAYRAUD, J.-CH. TREGLIA Amphores, céramiques culinaires et céramiques communes omeyyades d’un niveau d’incendie à Fustat-Istabl `Antar (Le Caire, Égypte) ..............................................................................................365 N. HAIDAR VELA Un contexte du VIIe siècle sur le sanctuaire de Saint-Syméon à Qal’at Sem’an (Syrie) ........................................377 P. BES, L. VANHECKE Leaf impressions on late Roman Sagalassos-amphorae ..........................................................................................387 A. F. FERRAZZOLI, M. RICCI Characterization and typology of a late Roman and proto-Byzantine lamp production center from Elaiussa Sebaste (Turkey)....................................................................................................................389 A. SAZANOV Cretan amphorae from northern Black Sea region: contexts, chronology, typology ..............................................399 A. SAZANOV The deposit of the end of the 6th century AD from Chersonesos (Crimea) ............................................................411 I. D. PARASCHIV Roman-byzantine pontic amphorae discovered at (L)Ibida, the province of Scythia .............................................425 P. TUŠLOVA, B. WEISSOVA Roman coarse wares and amphorae from 27 Metropolit Panaret Street, Plovdiv, Bulgaria ...................................433 A. OPAIŢ The baggy amphora shape: a new fashion? .............................................................................................................441 P. REYNOLDS, E. PAVLIDIS Nikopolis (Epirus Vetus): An early 7th century pottery assemblage from the ‘Bishop’s house’ (Greece) .................................................................................................................................451 A. FIOLITAKI Cooking wares from Agioi Asomatoi Vafe and Kefala Vrises in the region of Chania (Crete) .............................469 E. C. PORTALE The sunset of Gortyn: amphorae in 7th –8th centuries AD .....................................................................................477 iii

M. ALBERTOCCHI New approaches in the study of protobyzantine plain wares production in Gortys (Crete) ....................................491 R. PERNA, D. ÇONDI, C. CAPPONI, B. LAHI, C. MARTINELLI, S. SEVERINI, D. SFORZINI, B. SHKODRA RRUGIA, V. TUBALDI Ceramiche d’uso comune, anfore e materiali laterizi provenienti dall’Edificio Termale di Hadrianopolis (Sofratikë – Albania) ...................................................................................................................501 F. MASELLI SCOTTI Romans d’Isonzo (Gorizia-Italia). Ceramica comune dalla necropoli altomedioevale ...........................................515 E.-M. CIRRONE Nuovi dati sulla ceramica da fuoco proveniente dai contesti tardoantichi della villa di San Vincenzino a Cecina (Livorno) ....................................................................................................................521 M. FERRI, C. MOINE, L. SABBIONESI Ceramica tardo antica nella laguna di Venezia San Lorenzo di Ammiana .............................................................531 E. CIRELLI Typology and diffusion of the Amphorae in Ravenna and Classe between the 5th and the 8th century AD .................................................................................................................541 M. GIULIODORI, V. TUBALDI, E. PARIS, C. MARTINELLI I tegami da fuoco dal complesso Tempio-Criptoportico di Urbs Salvia: dati preliminari e analisi archeometriche ................................................................................................................553 F. DIOSONO, H. PATTERSON Some observations on the late Antique and early Medieval pottery from the central Apennines: the case of Villa San Silvestro of Cascia................................................................................................................563 C. BONANNO Ceramica romana tarda da cucina, da mensa e da dispensa e anfore nella Sicilia centrale .....................................573 M. S. RIZZO, L. ZAMBITO, F. GIANNICI, R. GIARRUSSO, A. MULONE Ceramiche da fuoco da contesti di eta’ tardoantica e bizantina nella Sicilia centro-meridionale: il caso di Cignana ...................................................................................................................................................581 E. ARIÑO, S. DAHÍ Ceramic contexts from late antiquity to the early Middle Ages (350-700c) in rural settlements of Salamanca (Spain) ...............................................................................................................589 S. DEMESTICHA Typology in the context: the case of late Roman amphorae ...................................................................................599 D. ASIMAKOPOULOS, A. OULKEROGLOU, O. KATSAVELI Depictions of late Roman transport amphorae. A preliminary report .....................................................................607 M. A. CAU ONTIVEROS, J. Mª. MACIAS, P. BERNI, P. REYNOLDS LRCW.NET: A web site with a virtual laboratory for the study of coarse and cooking wares in the late Antique Mediterranean ...................................................................................................617

iv

VOLUME II REGIONAL CONTEXTS: EAST MEDITERRANEAN J. POBLOME (INVITED) Shifting societal complexity in Byzantine Asia Minor and Dark Age pottery ........................................................623 A. CERUTTI Preliminary data for the Brittle Ware from the new excavations in the south-west Quarter of Palmyra (Syria) ......................................................................................................................................643 E. E. INTAGLIATA The White Ware from Palmyra (Syria): preliminary data from the new excavations in the south-west Quarter ........................................................................................................................................649 R. ROSENTHAL-HEGINBOTTOM Dating the Jerusalem rilled-rim and arched-rim basins ...........................................................................................657 Ö. ÇÖMEZOĞLU Coarse wares from the early Byzantine (episcopal) church of Rhodiapolis: cooking wares and amphorae ..................................................................................................................................665 N. SCHWERDT Utilitarian and cooking wares from the baths of Faustina in Miletus (Turkey) ......................................................677 C. ROCHERON, T. BLANCO Les céramiques communes de la fin de l’Antiquité à Xanthos: continuité ou innovation? .....................................685 C. MONDIN Common ware discovered during the archaeological excavations at Tyana/Kemerhisar (south Cappadocia, Turkey): preliminary remarks..................................................................................................693 S. ZACHARIADIS Pottery from the workshop building block in the early Byzantine city of Philippi, Greece ....................................705 J.-S. GROS Une forme particulière des abords de l’agora de Thasos: la bassine ovale à marli .................................................715 SP. VASILIOU, E. B. TSIGARIDA Late Roman pottery from recent excavations at Cassandra in Chalcidice ..............................................................723 E. DAFI Amphorae and cooking wares from the coastal site of Antikyra in Boeotia, Greece ..............................................737 A. KOUVELI Excavation for the New Acropolis Museum in Athens: Amphorae from a domestic well of a late Roman house ...................................................................................................................749 A. MARTIN A sixth-century context at Olympia (SW Building) ................................................................................................761 A. YANGAKI Late Roman pottery from Funerary Monument 18 and the adjacent monuments of the Arcadian gate (Ancient Messene, Peloponnese) ...........................................................................................769 A. PENTEDEKA, K. MORGAN, A. SOTIRIOU Patterns of local pottery production on late Roman Ithaca: preliminary remarks on the coarse and cooking ware fabrics ..................................................................................................................777 v

L. BOURNIAS Roman and Early Byzantine lamps from the island of Naxos in the Cyclades .......................................................787 V. KLONTZA-JAKLOVA Transport and storage pottery from Priniatikos Pyrgos – Crete: a preliminary study .............................................799 M. XANTHOPOULOU, E. NODAROU, N. POULOU-PAPADIMITRIOU Local coarse wares from late Roman Itanos (East Crete, Greece) ..........................................................................811

REGIONAL CONTEXTS: WEST MEDITERRANEAN D. BERNAL-CASASOLA, M. BUSTAMANTE-ALVAREZ, A. M. SÁEZ- ROMERO Contextos cerámicos tardorromanos de un ambiente haliéutico de la ciudad de Septem (Mauretania Tingitana) .........................................................................................................................819 A. PECCI, M. A. CAU ONTIVEROS Residue analysis of late Roman cooking pots and amphorae from Sa Mesquida (Mallorca, Balearic Islands) ......................................................................................................833 C. NERVI Local and imported pottery at Nora (Sardinia-Cagliari) from the 3rd to the 7th centuries. Comparison between urban and rural finds .............................................................................................. 843 M. CAVALIERI, B. MAGNI, G. BALDINI, E. BOLDRINI, S. RAGAZZINI The amphorae in the roman villa at Aiano-Torraccia Chiusi (San Gimignano, Siena, Italy)..................................855 T. MUKAI, M. AOYAGI Un contexte de la fin du IIIe s. à Somma Vesuviana (Campanie, Italie) ................................................................863

THE MEDITERRANEAN: A MARKET WITHOUT FRONTIERS N. POULOU-PAPADIMITRIOU, E. NODAROU (INVITED) Transport vessels and maritime trade routes in the Aegean from the 5th to the 9th century AD. Preliminary results of the EU funded project “Pythagoras II””: the Cretan case study ............................................873 S. ANTONELLI, O. MENOZZI Late Roman coarse ware and amphorae from Cyrenaica (Libya): the case of Lamluda .........................................885 J. LEIDWANGER A preliminary archaeometric analysis of the Late Roman 1 amphoras from the cargo of the seventh-century Yassıada shipwreck, Turkey ...............................................................................................897 K. W. SLANE, E. KIRIATZI Kythera al Fresco: Middle and late Roman cooking pots from the Aegean region.................................................907 S. Y. WAKSMAN, Y. MOROZOVA, S. ZELENKO, M. ÇOLAK Archaeological and archaeometric investigations of the amphorae cargo of a late Roman shipwreck sunk near the Cape of Plaka (Crimea, Ukraine).....................................................................................919 E. KLENINA Trade relations of Novae (Moesia II) according to ceramic evidence from an assemblage of the late 5th–early 6th century..............................................................................................................................931 A. CEAZZI, A. DEL BRUSCO La ceramica comune, la ceramica da cucina locale e importata, e le anfore dallo scavo di via Bolivia, Aquileia (Udine-Italia) .................................................................................................943 vi

P. VENTURA, C. DEGRASSI, P. RIAVEZ Finds from late Roman and early medieval Tergeste (I) .........................................................................................955 E. CIRELLI, A. CANNAVICCI A 6th century dump from Classe (Ravenna) ...........................................................................................................963 D. TINTERRI Trading amphorae in late Roman Liguria (AD 400-700) ........................................................................................975 A. COSTANTINI Pisa: importazioni tra tarda antichita’ e altomedioevo da Piazza Duomo ...............................................................985 S. GENOVESI Contenitori da trasporto dall'area del Portus Pisanus e rotte commerciali tirreniche tra IV e VI sec. D. C.: nuove evidenze dalla campagna 2009 .................................................................................993 F. FILIPPI, F. DELL’ERA, A. F. FERRAZZOLI Roma, complesso di San Francesco a Ripa: un esempio di uso secondario di anfore in età tardo antica ..................................................................................................................................................1005 M. T. IANNELLI, F. A. CUTERI, G. HYERACI, P. SALAMIDA Isca sullo Ionio (Calabria - Italia): scavi in loc. Zagaglie .....................................................................................1013 O. BELVEDERE, A. BURGIO Anfore e ceramiche comuni e da fuoco da Thermae Himeraeae (Sicilia) e dal suo hinterland ............................1023 F. OLIVERI, S. ZANGARA Deep water investigation off the Egadi Islands (western Sicily): Levanzo 1 wreck .............................................1035 M. A. CAU ONTIVEROS, C. MAS FLORIT, P. REYNOLDS, J. RIUTORT Two late Antique ceramic deposits from the city of Palma de Mallorca (Balearic Islands, Spain) ......................1049 R. JÁRREGA DOMÍNGUEZ Late Roman amphorae in the eastern Tarraconensis. Some chronological and quantitative approaches ..................................................................................................................................1061

NOTE CONCLUSIVE, by S. SANTORO ...................................................................................................................1069

vii

viii

INTRODUCTION The conference series on Late Roman Coarse Wares, Cooking Wares and Amphorae in Mediterranean (LRCW) started in Barcelona in 2002 and twelve years later it has been “institutionalized” as a major event bringing together every three years scholars working on late Antiquity. The resulting volumes1 set forth the most recent advances in late Roman/early Byzantine studies with the presentation of new data and, in many cases, an integrated approach of archaeology and archaeometry. The fourth LRCW congress was held at the Aristotle University of Thessaloniki between 7-10 April 2011 under the full title: 4th International Conference on Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A market without frontiers. The participation was overwhelming: 65 papers were presented as posters and 68 orally, by a total of 198 participants. The sessions of the conference were organized in a thematic rather than on a strictly geographical perspective: archaeology and economic history, production centers, distribution and consumption, regional contexts-east Mediterranean, regional contexts-west Mediterranean, and the Mediterranean: a market without frontiers. In this way we brought together researchers working on similar issues and facing similar difficulties across the Mediterranean, stressing the economic character of the sites and assemblages and culminating in the final session with the concept that the Mediterranean is after all an open way, a mare nostrum connecting the western to the eastern part. We are extremely happy that in the LRCW4 were equally well represented both ends of this route, with special reference to colleagues from countries of the east Mediterranean who attended for the first time. The conference and each session begun with invited speakers who presented key issues in the scholarship of late Roman archaeology. The LRCW4 conference was not just about long days of presentations in the Aula of the School of Philosophy. It also took in visits to pottery workshops, some sight-seeing at key points of the city and good food and wine. Taking things in chronological order, the first evening opened both gracefully with Greek folk songs by the “Jannis Mandakas” choir of the Aristotle University of Thessaloniki (directed by P. Bekiaridis) and tastefully with a reception bringing to the participants traditional Greek tastes and wine. The poster session took place in the historic building of Loutra Paradeisos/Bey Hamam, the first Ottoman bath of Thessaloniki, built in 1444. Those participants interested in ceramic petrography had the opportunity to attend a workshop on fabrics of late Roman pottery from Crete drawn from the collection of the INSTAP Study Center for East Crete. The last day was devoted to a hands-on pottery session: participants had the opportunity to visit the Museum of Byzantine Culture where they examined unpublished pottery assemblages from excavations in the city of Thessaloniki and Chalcidice. For this workshop we are grateful to S. Choulia-Kapeloni, director, and to D. Nalpantis and T. Antonaras, archaeologists of the Museum of Byzantine Culture; to D. Makropoulou, director, and S. Vassileiadou and M. Kagiadaki, archaeologists of the 9th Ephorate of Byzantine Antiquities; and to J. Kanonidis, director, and A. Tsanana, archaeologist of the 10th Ephorate of Byzantine Antiquities. The conference ended with an excursion to Dion, where Professor A. Mentzos assisted by a number of students guided the participants around the Roman-Early Byzantine site excavated by the Department of Archaeology. At the Museum of Dion we had a guided tour of the exhibition and one more hands-on pottery workshop examining the pottery from the excavation. We are grateful to the director of the university excavation at Dion, Professor S. Pingiatoglou, for making this visit possible. An excursion in Pieria could not have ended otherwise but in a local tavern with traditional flavors of northern Greece. At a time of financial hardship for Greece neither the conference nor the present volume would have materialized without the support of the sponsors; we are grateful to all of them. The Aristotle University of Thessaloniki and the National Centre for Scientific Research Demokritos undertook the organization of the conference. Professor A. Mentzos is gratefully acknowledged for contributing to the organization of the conference. We would also like to thank all the students who helped, namely M. Andrianaki, Th. Ioannidou, K. Fragoulis, S. Konstantinidou, G. Panelas, D. Papazoglou, E. Psoma, S. Thatharopoulou, K. Tsimbli, E. Tsiola, L. Tsiola, K. Tzimagiorgi, and A. Zafeiris; G. Aloupis and N. Tsolas operated the camera for the live streaming throughout the conference. D. Minasidis is gratefully acknowledged for his invaluable help in all organizational issues. 1

Gurt i Esparraguera, J. M., J. Buxeda i Garrigós, and M. A. Cau Ontiveros (eds.) 2005. LRCW1. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry. British Archaeological Reports, International Series 1340. Oxford, BAR Publishing. Bonifay, M. and J.-C. Tréglia (eds.), 2007. LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports, International Series 1662. Oxford, BAR Publishing. Menchelli, S., S. Santoro, M. Pasquinucci and G. Guiducci (eds.), 2010. LRCW3. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Comparison between Western and Eastern Mediterranean. British Archaeological Reports, International Series, 2185. Oxford, BAR Publishing.

ix

After three years of compiling articles, reviews and final texts we are happy to present the fourth LRCW volume: LRCW4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A market without frontiers. The structure of the book follows the thematic sessions of the conference, whereas the arrangement of the articles within the individual units is geographical, a journey of the mind from western to eastern Mediterranean (or vice versa). We take the occasion to thank all the contributors for entrusting us with their work and all the reviewers for undertaking the task of peer-reviewing: we are grateful to Paul Arthur, Michel Bonifay, Claudio Capelli, Miguel A. Cau Ontiveros, Piotr Dyczek, Josep M. Gurt Esparraguera , Philip M. Kenrick, John Lund, Archer Martin, Simonetta Menchelli, Jeroen Poblome, Elisa-Chiara Portale, Paul Reynolds, Sara Santoro, Jean-Christophe Tréglia, Agnès Vokaer, and Paul Williams who painstakingly read and commented upon the submitted articles. Doniert Evely undertook the task of editing the English. We would like to acknowledge the Psycha Foundation for partly funding this publication. Finally, many thanks are due to Niki Saridaki who was in charge of the Secretariat for all these years and to Sophia Didioumi who helped with the organization of the conference and then handled and processed the 102 papers of the present volume. The LRCW5 conference is already under way. We are looking forward to seeing everybody in Alexandria in April 2014. Natalia POULOU-PAPADIMITRIOU, Aristotle University of Thessaloniki, Dept. of History and Archaeology Eleni NODAROU, INSTAP Study Center for East Crete Vassilis KILIKOGLOU, NCSR Demokritos Thessaloniki, February 2014

x

Preface to the LRCW4 It has now become a tradition for groups and individual scholars to convene every three years under the auspices of the LRCW International Standing Committee, in a place near or on the Mediterranean, in order to exchange ideas and update their knowledge on matters concerning everyday and transport ceramics of the late Roman/early Byzantine period. The success of this conference series is proven by the increasing participation in the LRCW, the variability and the geographical breadth of the issues covered and the material presented in these conferences. The LRCW4 at Thessaloniki proved to be right in this line of evolution. The approaches to the material in the LRCW4 conference varies from the traditional typo-morphological ones with emphasis on pottery sequences and local production centers to more technological procedures involving archaeometric analyses investigating the clay fabric and its mineralogical and/or chemical composition. These syntheses tackle broad issues such as the commerce and the economy of the late Antique world and cover almost all the areas around the Mediterranean. Of particular interest is the growing representation of the Black Sea coast, the Balkan hinterland, Asia Minor, and countries from the Near East. In this context a great number of local and regional workshops were identified and studied. One particular field of interest was the investigation of patterns of exchange, on a local as well as on a regional scale. Special attention was given to the history of commerce and the trade mechanisms, with a focus on the trade of amphorae, but also of cooking pots. Research in the field of the differentiation and identification of the amphorae production centers continues, with a growing interest on the distribution of ceramics, the investigation of long range maritime relations and long distance exchanges, as well as in establishing the maritime routes followed by cargo ships of the era. Within this framework, a constantly rewarding branch of research is the study the earthenware found in shipwrecks and marine archaeology in general. Of growing importance is the study of the relation between container and content. Advancing this direction further is the special interest exhibited in the interrelation of ware types with culinary habits. The growing level of participation seen in LRCW4 reveals the increased attention among pottery specialists and archaeologists alike paid to the study of everyday ceramics and storage/commercial wares. It has become a widespread belief that in view of the multiplication and the diversification of the archaeological data, common and commercial ceramics offer a reliable tool for the dating of the archaeological strata and, most importantly, for a more detailed, indepth understanding of ancient society and the people of the late Roman and early Byzantine periods. I feel that it would be unjust to close without a word on the editors. Natalia Poulou-Papadimitriou, Eleni Nodarou, Vassilis Kilikoglou and their collaborators, Sophia Didioumi and Niki Saridaki, performed an admirable feat in handling a vast and varied material of 102 reports; they deserve our unconditional appreciation.

Aristoteles MENTZOS, Professor of Byzantine Archaeology, Aristotle University of Thessaloniki

xi

xii

ARCHAEOLOGY AND ECONOMIC HISTORY

CONTAINER AND CONTENT IN NORTH AFRICA: A NEW GLIMPSE OF THE RELATIONSHIP BETWEEN FOODSTUFFS PRODUCTION AND AMPHORAE PRODUCTION (4TH-6TH CENTURIES AD) LILIA PALMIERI Università degli Studi di Milano, Dipartimento di Beni Culturali e Ambientali via Festa del Perdono - 7, 20122 Milano, Italy [email protected] The paper focuses on the analysis of the distribution patterns of the olive oil presses and amphorae workshops located in the province of Byzacena between the 4th and the first half of the 6th centuries AD, in order to reconstruct the economic dynamics which characterized the province in the late Roman and Vandal periods. In particular, a preliminary investigation of the production topography reveals the deep transformation that involved the African economic landscape in the Vandal age and how the production system changed starting from the second half of the 5th century AD. Although the olive oil production still includes the exploitation of the presses installed on the territory of western Byzacena, the major African oleoculture region, the pottery workshops involved in the amphorae production are no longer installed near the coastal sites in eastern Byzacena, but are located to the south of the intensive oleoculture region. This transformation implies a reorganization of the olive oil routes for the intra- and interprovincial trade. KEYWORDS: TUNISIA, BYZACENA, ECONOMIC LANDSCAPE, VANDALS, PRODUCTION, OLIVE OIL PRESSES, POTTERY WORKSHOP, AMPHORAE Research on production sites in North Africa provides a major contribution to the interpretation of the transformation processes that have changed the African economic landscape throughout the ages. Comparison of the analyses of the production facilities of the main foodstuffs exported by Africa to the whole Mediterranean basin - olive oil, wine, garum - and of the pottery workshops involved in the production of containers allows us to focus on the production and trade dynamics, in order to understand both the transformations of the production topography and the changes in the intraand inter-provincial routes followed by these products. Among the analysed foodstuffs we could not include grain, although this produce is well known from historical sources as the first economic resource of the African provinces till the late Roman period, because the archaeological evidence is almost entirely lacking for both its production and trade. The research carried out in recent decades in North Africa, particularly in the provinces of Zeugitana and Byzacena, has revealed an intensive production, especially for the late Roman period, by identifying hundreds of olive oil presses and dozens of fish-salting factories and pottery workshops (Leone and Mattingly 2004; Slim et al. 2004). The identification of wine presses without archaeometric analyses is problematic, because of their great similarity to the olive oil presses. Some recent studies have also proposed a detailed typology of the African amphorae (Bonifay 2004) and the identification of the production sites of some containers. The archaeometric analyses performed on some amphorae samples have finally revealed the association

between certain types of amphorae and some foodstuffs (Bonifay 2007). The issue of production sites is the core of a doctoral research focused on the socio-economic transformations in North Africa between the 4th and the 7th centuries AD (Palmieri 2010-2011). Through a research process starting from the analysis of the settlements installed in the province of Africa (Palmieri 2004) and continued with the study of the olive oil production and trade in Vandal Africa (Palmieri 2008), we have finally expanded the investigation to the African economic landscape in the late Roman, Vandal and Byzantine periods, with the aim of focusing on the production and trade dynamics of the provinces of Zeugitana and Byzacena in their historical context. A first analysis in this direction has already been published for the territory of NeapolisNabeul (Palmieri 2012). With this new research we propose, therefore, a reversal of the investigation perspective: we consider the production process from the starting point (Africa) and not from the end point (the Mediterranean basin), the preferential archaeologists’ point of view when working in contexts of consumption rather than in contexts of production. In this short paper we will attempt only the examination of the distribution patterns of the olive oil presses and amphorae workshops located in the province of Byzacena and a preliminary investigation of the production topography compared with the socio-economic dynamics which characterize the province between the 4th and the first half of the 6th centuries AD.

LRCW4

continuous economic development from the 3rd to the 5th centuries AD, with an intensive exploitation of land and the existence of the associated production facilities.

1. Olive oil production in Byzacena Although many surveys have been carried out in North Africa in recent decades, the province of Byzacena is still far from being fully investigated and the field research is still at its beginning. Only 23 sectors out of a total of 77 topographic sheets of the Archaeological Map of Tunisia have been examined.

The surveys carried out in the region of Thala (Ben Baaziz 2005a), Ksar Tlili (Ben Baaziz 2005b) and in the upper valley of the Oued el Htab (Ben Baaziz 1985; 2000), in the northern area of western Byzacena (corresponding to the Tunisian Dorsal), have revealed many presses, which show the high agricultural specialization of the area, with a density of about one press per 2km2.

The surveys allow us to recognize two areas with different grades of archaeological evidence: the highest concentration of olive oil presses has been recorded in the High Steppe region in western Byzacena - particularly in the territories of Thelepte-Feriana, Cillium-Kasserine and Sufetula-Sbeitla; whereas, moving to the east coast ports, like Hadrumetum-Sousse, Leptiminus-Lamta and Sullectum-Salakta, evidence for these production facilities significantly decreases.

The High Steppe region, in the centre of western Byzacena, shows an even greater emphasis on olive oil production, particularly in the territories of ThelepteFeriana, Cillium-Kasserine and Sufetula-Sbeitla (Mattingly 1988). Although the surveys carried out near the ancient cities of Thelepte (Hermassi 2003) and Sufetula (Duval 1990) have revealed many presses, yet the most important investigation in the area has been performed by a joint Tunisian-American team around the city of Cillium-Kasserine. This research has collected information not only on the production facilities, but also about the agricultural methods and production systems (Hitchner 1988; 1989; 1990; 1992-1993). The area under investigation has revealed a high density of structural units, 0.8 press per km2: most of them show an exploitation in the late Roman and Vandal periods, from the 3rd to the 5th centuries AD, with some sites active until the 6th and the 7th centuries AD (Hitchner 1988, 8; 1989, 389).

In eastern Byzacena, which corresponds to the Tunisian Sahel, the field surveys carried out on only a third of the territory have identified 1219 sites, but only 60 sites with an olive oil press (Ben Baaziz 1998-99). The low density of the olive oil presses near the coast cannot be completely attributed to the lack of surveys, because the well investigated coastal areas like Mahdia and BararusRougga have confirmed this distribution pattern (Trousset 1977, 198-199; Ben Younes 1998). The territory of Segermes, a city on the border with Zeugitana, investigated by a Danish-Tunisian team, has also revealed presses only in eight rural sites (Dietz et al. 1995, II, 773799; Ørsted et al. 2000, 105-131). The same lack of production facilities is also attested in other territories located a little more inland, like Vicus Augusti-Sidi el Hani and Oued Cherita (Ben Baaziz 1998; Annabi 1998). Finally, it is significant that the practice of reusing the structural elements making up the olive oil presses as building materials is almost entirely absent from the coastal sites.

Three classes of olive oil facilities have been identified in the territory of Kasserine. The first type involves presses located within the villages, the second identifies the presses located within the farms and the third comprises the oileries equipped with many presses, usually between four and eight. This latest group is likely to be linked closely to the intensive olive oil production of western Byzacena, because the incidence of multiple presses converts these installations into ʽindustrial oileries’.

We could interpret this absence of evidence as a probable reflection of the lesser development of olive production in the coastal areas. The data suggest the development of a more varied agricultural system for the territory of eastern Byzacena: one not characterized by crop specialization, but based on diversification. Within this system other crops were also cultivated, like cereals and perhaps vines, according to the ancient sources and particularly Pliny the Elder, who describes a landscape given over to cereals (Pliny, Nat. Hist., V, 24; XV, 3; XVII, 41; XVIII, 94).

The most significant examples of multi-press buildings of the third category identified for this area are the large sites called KS223 and KS225 (Mattingly 1990), whose activity is attributable to a period between the 3rd and the 5th centuries AD. These two units are located only 350m from each other and it is very likely that they were part of the same rural estate. The two production centres are equipped with four presses, overlook courtyards and are surrounded by a number of outbuildings and enclosures. Mattingly has estimated that their production was as high as 40,000-80,000 litres of olive oil in peak years, requiring potentially as many as 10,000-20,000 mature olive trees spread out over 100-400ha (Mattingly and Hitchner 1993).

This area deficient in evidence related to olive oil production differs from the western part of the province, which appears as a region characterized by crop specialization. The analysis of the archaeological evidence allows us to demarcate a sub-region where the landscape is dominated by the presence of presses. It extends from the territory of Rohia Ksar Tlili (north) to the foothills of the territory of Capsa-Gafsa (south), and from the Djebel M’ghuila (east) to the territory of Theveste-Tebessa (west). From this evidence we can follow in this sub-region of western Byzacena a

The industrial oileries, like those described for the Kasserine area, are quite common in western Byzacena. The production centres of Henchir el Kousset, Henchir Bou Safa and Henchir Choud el Battal are each equipped with six presses, the building of Tamesmida with eight presses (Brun 2004, 227). At the present the largest oilery 2

PALMIERI

western Byzacena are, in fact, only two: Henchir el Guelel-Sidi Saad, to the east of Sufetula-Sbeitla, and Majoura, to the south of Cillium-Kasserine.

of Byzacena is the one installed at the site of Henchir El Begar, equipped with 12 presses (Sehili 2008, 786). These large installations, which can be compared only with those identified in Tripolitania, attest to the existence of large estates specializing in olive oil production, as is confirmed by some documents that play a key role in the interpretation of the late antique economic landscape, the so-called Albertini Tablets (Courtois et al. 1952). This is a set of archival documents inscribed on wooden tablets, discovered in 1928 to the south of Theveste-Tebessa and dated back to the late 5th century AD. In particular, all but four of the 33 surviving documents concern the sale of agricultural parcels planted for the most part with olive trees (= particellae agrorum), located on a fundus to the east of Theveste-Tebessa. The tablets, therefore, provide a further proof of the survival of the Roman economic system in North Africa during the Vandal age. By documenting the persistence of large estates with crop specialization, in fact, they demonstrate how the exploitation of the agricultural parcels, identified in the texts as culturae Mancianae, was still regulated by the Roman law (lex Manciana).

The archaeological data and the extant archaeometric analyses allow us to recognize three types of olive oil amphorae produced in Byzacena between the 4th and the first half of the 6th centuries AD. The first type of olive oil amphora is Africana 1, a medium-sized cylindrical container with a capacity of 3540l, produced between the end of the 2nd and the 4th centuries AD and widely distributed throughout the Mediterranean, particularly common at Rome and Ostia, but also in the eastern Mediterranean and even Britain (Beltrán Lloris 1970; Panella 1973; Manacorda 1977; Riley 1979). The production of this container is attested at a number of sites along or near the eastern coast at Hadrumetum-Sousse, Leptiminus-Lamta, Henchir Ben Hassine (?), Sullectum-Salakta, El Mokaida and El Hri, Acholla and Thaenae-Thyna. At the last site one of the two identified kilns exclusively produced the Africana 1, both in the variant B (end of the 2nd - middle of the 3rd centuries AD) and in the late variant C (second half of the 3rd - 4th centuries AD). Moving further south, on the border with Tripolitania, the medium-sized workshop identified near the fish salteries located close to the river Oued el Akarit, has also produced Africana 1C, a later production dated to the 4th century AD (Bonifay 2004, 31). In western Byzacena the type has only been recorded at the site of Henchir el Guelel-Sidi Saad, whose atelier was active between the end of the 2nd and the 5th centuries AD. Except for the site of Thaenae-Thyna, this production does not appear to be the only one conducted within the workshops, but it is always associated with the production of containers involved in the transport of other foodstuffs, such as wine or garum.

The integration of the evidence of the Albertini Tablets with the archaeological data from Kasserine has allowed scholars to interpret the agricultural organisation of the region - one based on groups of settlements located at some distance from the urban centres and with a concentration of the production activities in the countryside. The industrial oileries, like KS223 and KS225, were probably the pars rustica of large estates divided into small units maintained by tenants under the authority of a farmer (Mattingly 1989; Hitchner 1995). The data derived from the Kasserine survey attests for the period between the 3rd and the 5th centuries AD to an increase in small farms equipped with olive oil presses and the growth of the industrial production centres like KS223 and KS225. This picture thus suggests for the territory of Cillium-Kasserine an intensive exploitation of the agricultural system in the late Roman and Vandal periods, characterized both by a parcelling up of the rural landscape into small units (small farms), perhaps more involved in the intra-provincial production - Byzacena, and by large production centres (industrial oileries), dedicated to the inter-provincial production - Zeugitana and the Mediterranean basin.

The second type of olive oil amphora is Keay 59, a large cylindrical container produced between the end of the 4th and the first half of the 5th centuries AD and distributed in small quantities in both the western Mediterranean - in Spain (Keay 1984), France (Congés and Leguilloux 1991) and Italy, and in the eastern Mediterranean - Egypt. The type has only been identified in the kiln detected at the coastal site of Iunca, to the south of Thaenae-Thyna (Bonifay 2004, 31). The third type of olive oil amphora is Keay 8B, a large cylindrical container produced between the second half of the 5th and the first half of the 6th centuries AD, particularly frequent at Carthage (Peacock 1984) and attested to in both the western Mediterranean - Tarragona (Keay 1984), Marseille (Bonifay and Pieri 1995), Albenga (Pallarès 1987) - and in the eastern Mediterranean - in particular Alexandria (Bonifay and Leffy 2002) and Constantinople (Hayes 1992). The amphora should be interpreted as an evolution of Keay 59 (Bonifay 2004, 132). It is possible to associate two sites with this type, Iunca on the east coast and Majura in the High Steppe region, where the exclusive production of this container is indicated since the end of 5th century AD.

2. Amphorae production in Byzacena If the olive oil presses are the dominant element of the economic landscape of western Byzacena, the workshops of the olive oil amphorae identified by some surveys carried out in the last twenty years are located especially in eastern Byzacena (Peacock et al. 1989; 1990; Bonifay 2004, 8). Most of the workshops, which produce olive oil amphorae for the maritime trade, are concentrated in eastern Byzacena, in the suburbs of the Sahel ports, particularly in the city of Sullectum-Salakta. The ateliers producing olive oil amphorae recorded in the territory of 3

LRCW4

At the end of the 5th century AD the archaeological data therefore demonstrates a relocation of the olive oil amphorae production from the coast of eastern Byzacena to the interior region of western Byzacena, the most important area of olive production in the province.

The third economic route was the road CarthageTheveste, the main axis of Zeugitana, which covered a distance of 252km. It connected Carthage to western Byzacena, passing through Ammaedara-Haidra (Desanges et al. 2010, 46).

3. Comparative analysis

We should not, moreover, underestimate the role of the wadis, in particular the water-system formed by the Oued Miliane/Khebir, which linked directly the plateau of central Tunisia to the Gulf of Tunis.

By observing the distribution patterns of the amphorae workshops, we can see that most of the ateliers which produced olive oil amphorae for the maritime trade in the 4th century - the type Africana 1 - are concentrated in the major ports of the Sahel and in their suburbs (Hadrumetum-Sousse, Leptiminus-Lamta, SullectumSalakta, Thaenae-Thyna). The only exception to this pattern is represented by the interior site of Henchir el Guelel-Sidi Saad, to the east of Sufetula-Sbeitla.

Through these routes the foodstuffs produced in western Byzacena, and in particular the olive oil, could easily reach by land, and perhaps by river, the major ports of Zeugitana and Byzacena, and then could be exported to the Mediterranean. The use of oilskins for the transportation of olive oil in North Africa is attested not only by some representations on African pottery, but also by some documents of inestimable economic value. These are the ostraka discovered in 1911 during the excavations carried out on the Ilôt de l’Admirauté at Carthage: they confirm the use of oilskins for the land transport of olive oil in the African provinces. These lists, written on fragments of African amphorae, record the quantities of olive oil collected for the annona between early February and late July of AD 373 (Peña 1998), and among them nine attest to the transport by oilskins of a part of this oil. The documents distinguish between two types of leather containers: the first, called as(copera) or as(copa), with a capacity of 504lbs, that is 182.7l or 164.8kg, and the second, called as(copera) a b(yzacena), with a capacity of 72lbs of oil, that is 26.1l or 23.5kg. Half of the olive oil reached Carthage by land, in oilskins, from Byzacena and Vol Calassini, roughly 30km south-southwest of Carthage, a quarter came by sea, in amphorae, from Hippo Regius, 260km west of Carthage, and a quarter travelled in containers of unknown sorts. The oil was collected in storage buildings near the commercial harbour of the city before being sent to Rome.

Moreover, Bonifay emphasizes the total absence of this particular olive oil container in the town of Thysdrus-El Jem (Bonifay 2004, 478). At this site, and also in the pottery workshops of eastern Byzacena investigated by the Anglo-Tunisian team led by Peacock, the amphora Africana 1 is very rare: other containers prevail, which are not however so widely distributed in the Mediterranean. This economic bipolarity, olive oil presses in the High Steppe region and amphorae workshops on the east coast, necessarily involves a significant transport by land of olive oil from the interior region of the province to the major coastal ports, using other containers such as oilskins (Fig. 1). The main road of the Byzacena was the via Hadrumentina, which covered a distance of 205km. It was the shortest road between the port of HadrumetumSousse and the sites of Theveste-Tebessa and Lambaesis two major military and administrative centres - passing through Ammaedara-Haidra (Desanges et al. 2010, 46). From an economic perspective, this road played a role of great importance, because it connected the four geographic regions of Byzacena - High Tell, High Steppe, Lower Steppe and Sahel - and linked, therefore, the large estates of the High Steppe region to the east coast.

We can suppose that this collection system was in use also in the major ports of the Sahel, with a centralized depot for the product, where it was transferred into amphorae, stored and afterwards exported throughout the Mediterranean. In the 4th century AD the oil produced in western Byzacena, in part by those industrial oileries installed in the High Steppe region, was transported in oilskins on one hand to eastern Byzacena, by the routes which connect this area to the port of Hadrumetum and to the other ports of the province, and on the other to Zeugitana, in order to be sent from Carthage to Rome. This system continues until the end of the 4th century that is until the amphora Africana 1 is first produced.

Another road ran parallel to the via Hadrumentina, connecting the sites of Thelepte-Feriana, CilliumKasserine and Sufetula-Sbeitla to the major ports of the Sahel. It is on this road that we find the centre of Thysdrus-El Jem, whose territory had no particular agricultural proclivity: the excavations here have not recorded olive oil containers. The site, in fact, gained much of its wealth from its role as an important road junction: from here started the main roads leading to the major ports involved in the olive oil export (Hadrumetum-Sousse, Leptiminus-Lamta, SullectumSalakta). Thysdrus-El Jem was, therefore, a ‘distribution centre’ for moving foodstuffs produced in the interior regions on toward the ports of Byzacena - a link connecting the two areas of the province (Slim 1996).

Some changes take place in the first half of the 5th century AD, when the production of the large cylindrical amphora Keay 59 begins. The production area of this container should be identified in the southern Sahel, to the south of the major ports of eastern Byzacena, which are still involved in the export of other foodstuffs - like 4

PALMIERI

wine and garum, as is shown by the amphorae distribution in the Mediterranean.

transformations in Zeugitana and Byzacena between the 4th and the first half of the 6th centuries AD.

In the middle of the 5th century AD, in consequence of the Vandal conquest, the organization of the amphorae production changes, as well as the internal transportation route for the olive oil (Fig. 2). While the oil production in western Byzacena continues, as indicated by the presses installed on the territory of Cillium-Kasserine, the workshop involved in the production of the olive oil containers is no longer set near the coastal sites, but in western Byzacena. It is at Majura, in fact, located to the south of the intensive oleoculture area, where the large amphora Keay 8B is produced, a container which can be considered a ‘type-fossil’ for the Vandal age. This type of amphora is found in the olive oil presses of the Kasserine region, it was recorded in large quantities at Carthage in contexts of the late 5th - early 6th centuries AD, and it was also distributed in western Mediterranean, particularly in Tarraconensis, the preferred foreign market for North Africa in the Vandal age.

The organization of the olive oil production in the 4th century AD appears as a system in which the presses are mostly installed on the territory of western Byzacena, whilst the amphorae workshops are mostly installed at the suburbs of the ports of eastern Byzacena and at Carthage, in Zeugitana. After reaching the coastal sites by an overland transport within oilskins, the olive oil was transferred into amphorae and stored, in order to be exported into the Mediterranean, and in particular to Rome to supply the annona. The production of the amphora Keay 59 in the first half of the 5th century AD may allow us to glimpse the first changes in the system, when the olive oil amphorae production moves from the coastal sites of the central Sahel - Hadrumetum-Sousse, Leptiminus-Lamta, Sullectum-Salakta, Thaenae-Thyna - to a coastal site of the southern Sahel - Iunca. But it is at the middle of the 5th century AD, when the Vandals permanently control the provinces of Zeugitana and Byzacena after the conquest of AD 439, that a real break with the previous phase occurs. The amphorae production, in fact, is reorganized and the routes followed by the olive oil are redrawn. The new production and trade system of the Vandal age still includes the exploitation of the presses installed on the territory of western Byzacena, but the workshop involved in the amphorae production is located now in the same area, to the south of the intensive oleoculture region. In consequence of the destruction of the main commercial harbour of Byzacena, HadrumetumSousse, the product for exportation travels only to the north and no longer to the east, along the route ThevesteCarthage, and perhaps along the Oued Miliane/Khebir, probably in amphorae rather than in oilskins. Once it reached Carthage, the olive oil is stored before being exported mainly to the western Mediterranean and the Iberian Peninsula, the favoured economic axis of the Vandal trade.

The olive oil production of western Byzacena does not now get to the east coast anymore, to be stored and afterwards exported into the Mediterranean. Firstly this is because there is no longer the need of a centralized collection of the product in the major ports of the province after the ending of the annona system (AD 455), and then because the major port of the inter-provincial trade of Byzacena, Hadrumetum-Sousse, is no longer active, having been destroyed by the Vandals (Foucher 1964). The olive oil is now sent only to Carthage. It is possible that the transport does not go anymore by oilskins, but that a significant volume of olive oil is transferred into containers directly at the place of production and travels thenceforth in amphorae. In the early 6th century AD Carthage, which after the Vandal conquest had remained active as a maritime trade centre, assumes the role of major port for the interprovincial trade, both for Zeugitana and Byzacena. As the axis of trade moved to the Iberian Peninsula, because of the economic links between the Vandals and the Visigothic kingdom, the city is favoured by its position within an exchange system which gravitates towards the western Mediterranean, but which does not exclude the trade with the eastern (Palmieri 2008).

At the end of the 5th century AD we note a decrease in the olive oil production, documented by the number of presses in activity and in consequence by the volume of trade, - also because of the ending of the annona system and we therefore agree with Clementina Panella, who emphasizes a decline in the African production and exports at the end of the 5th century AD (Panella 1993). Even so, a degree of production and trade dynamism remains, enabling the export of olive oil into the Mediterranean between the end of the 5th and the first half of the 6th centuries AD, in the so-called second Vandal period.

4. Conclusions The analysis and comparison of the mechanics of the olive oil production and the amphorae production allow us to reconstruct the economic dynamics which characterized the province of Byzacena in the late Roman and Vandal period.

It is also necessary to underline the fact that the change in the siting of the pottery workshops and the introduction of a new type of amphora - the large cylindrical amphora Keay 8B - correspond with an alteration in the politicaladministrative system of the province, namely the settlement of the Vandals in the provinces of Zeugitana and Byzacena. This observation modifies in part Keay’s vision, shared also by Bonifay (Bonifay 2004, 481),

Although partial, because it is limited to the issues related to only one product of the provincial economy, the analysis gives us the opportunity to emphasize some aspects, by adding a further element to the interpretive hypothesis formulated by scholars on the socio-economic 5

LRCW4

Alexandriennes, 6, 39-84. Le Caire, Institut Français d’Archéologie Orientale.

which did not recognize any association between the typological changes of the amphorae and the sociopolitical transformation of the African provinces (Keay 1998).

Bonifay, M. and Pieri, D. 1995. Amphores du Ve au VIIe s. à Marseille: nouvelles données sur la typologie et le contenu. Journal of Roman Archaeology 8, 94-120.

The new organization of the harbour of Carthage in the late 5th - early 6th centuries AD, combined with a commercial liveliness which is an important sign of an economic revitalization (Fulford and Peacock 1984, 214; Reynolds 1995, 105), is the consequence partly of the elimination of the major port of the Sahel, HadrumetumSousse, destroyed by the Vandals, and partly of the lesser economic importance anyway of the other commercial ports of Byzacena. The impossibility of using the ports of the east coast for Mediterranean trading brought about a deep shift in the economic landscape of Byzacena: the overland roads towards north became preferred and in consequence the olive oil routes for the intra- and interprovincial trade were reorganized.

Brun, J. P. 2004. Archéologie du vin et de l’huile dans l’empire romain. Paris, Errance. Congès, G. and Leguilloux, M. 1991. Un dépotoir de l’antiquité tardive dans le quartier de l’Esplanade à Arles. Revue archéologique de Narbonnaise 24, 201234. Courtois, Ch., Leschi, L., Perrat, C. and Saumagne, Ch. (eds.), 1952. Tablettes Albertini. Actes privés de l’époque Vandale (fin du V siècle). Paris, Arts et Métiers Graphiques. Desanges, J., Duval, N., Lepelley Cl. and Saint-Amans, S. (eds.), 2010. Carte des routes et des cités de l’est de l’Africa à la fin de l’antiquité. Turnhout, Brepols Publishers.

Bibliography Annabi, M. K. 1998. Oued Cherita 072, Carte Nationale des Sites Archéologiques et des Monuments Historiques. Tunis, Institut National du Patrimoine.

Dietz, S., Ladjimi Sebaj, L. and Ben Hassen, H. (eds.), 1995. Africa Proconsularis. Regional Studies in the Segermes Valley of Northern Tunisia, I-II. Copenhagen, Aarhus University Press.

Beltrán Lloris, M. 1970. Las ánforas romanas en España. Zaragoza, Institución ʽFernando el Católico’.

Duval, N. 1990. Sufetula: l’histoire d’une ville de la Haute Steppe à la lumière des recherches récentes. In L’Afrique dans l’Occident romain (Ier siècle av. J.C. IVe siècle ap. J.C.), 495-535. Rome, Ecole Française de Rome.

Ben Baaziz, S. 1985. Les huileries de la Haute vallée de l’Oued el Htab. Africa 9, 209-215. Ben Baaziz, S. 1998. Sidi el Hani 064, Carte Nationale des Sites Archéologiques et des Monuments Historiques. Tunis, Institut National du Patrimoine.

Foucher, L. 1964. Hadrumetum. Tunis, Université de Tunis.

Ben Baaziz, S. 1998-99. Les établissements ruraux du Sahel antique. In A. Mrabet (ed.), Du Byzacium au Sahel. Itinéraire historique d’une région tunisienne. Actes du colloque sur le Sahel tenu à Sousse en décembre 1996, 31-50. Tunis, L’Or du Temps.

Fulford, M. and Peacock, D. P. S. 1984. The avenue du Président Bourguiba, Salambo. The Pottery and other Ceramic Objects from the site. Excavations at Carthage: the British Mission, vol. I,2. Sheffield. Hayes, J. W. 1992. Excavations at Saraçhane in Istanbul, vol. 2. The pottery. Princeton N.J., Princeton University Press.

Ben Baaziz, S. 2005a. Thala 067, Carte Nationale des Sites Archéologiques et des Monuments Historiques. Tunis, Institut National du Patrimoine.

Hermassi, M. 2003. Quelques données sur les vestiges oléicoles autour de Thelepte. In F. Bejaoui (ed.), Histoire des Hautes Steppes. Antiquité et Moyen-Age. Actes du Colloque de Sbeitla, session 2001, 81-92. Tunis, Institut National du Patrimoine.

Ben Baaziz, S. 2005b. Ksar Tlili 068, Carte Nationale des Sites Archéologiques et des Monuments Historiques. Tunis, Institut National du Patrimoine. Ben Younes, H. 1998. Mahdia 074, Carte Nationale des Sites Archéologiques et des Monuments Historiques. Tunis, Institut National du Patrimoine.

Hitchner, B. R. 1988. The Kasserine archaeology survey, 1982-1986. Antiquites Africaines 24, 7-41.

Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports, International Series 1301. Oxford, BAR Publishing.

Hitchner, B. R. 1989. The organization of rural Settlement in the Cillium-Thelepte region (Kasserine, Central Tunisia). In A. Mastino (ed.), L’Africa Romana. Atti del VI convegno di studio, Sassari, 1618 ottobre 1988, 387-402. Sassari, Edizioni Gallizzi.

Bonifay, M. 2007. Que transportaient donc les amphores africaines? In E. Papi (ed.), Supplying Rome and the Empire. Actes du séminaire international, Certosa di Pontignano, 3-4 mai 2004, 8-32. Porthmouth.

Hitchner, B. R. 1990. The Kasserine archaeology survey 1987. Antiquités Africaines 26, 231-260.

Bonifay, M. and Leffy, R. 2002. Céramiques du remplissage de la citerne du Sérapeum à Alexandrie. In J. Y. Empeur (ed.), Alexandrina 2. Études

Hitchner, B. R. 1992-1993. The Kasserine archaeology survey, 1982-1985. Africa 11-12, 158-198. 6

PALMIERI

Hitchner, B. R. 1995. Historical Text and Archaeological Context in Roman North Africa: The Albertini Tablets and the Kasserine Survey. In D. B. Small (ed.), Methods in the Mediterranean. Historical and Archaeological Views on Texts and Archaeology, 124-142. Leiden, E.J. Brill.

Sevilla, 14-17 dicembre 2006, 1081-1090. Roma, Carocci editore. Palmieri, L. 2010-2011. Processi di trasformazione economica e sociale in Africa settentrionale dal IV al VII secolo d.C. Analisi dei sistemi di produzione in Zeugitana e Bizacena. Unpublished PhD thesis, UCSC of Milan.

Keay, S. J. 1984. Late Roman amphorae in the Western Mediterranean. A typology and economic study: the British Archaeological Catalan evidence. Reports, Int. Series, 196. Oxford, BAR Publishing.

Palmieri, L. 2012. La trasformazione del paesaggio economico africano in età tardo-antica: analisi dei sistemi di produzione. Il caso di Neapolis-Nabeul. In M. B. Cocco, A. Gavini, A. Ibba (eds.), L’Africa Romana. Trasformazione dei paesaggi del potere nell’Africa settentrionale fino alla fine del mondo antico. Atti del XIX convegno di studio, Sassari, 1619 dicembre 2010, 499-514. Roma, Carocci editore.

Keay, S. J. 1998. African amphorae. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del Convegno in onore di John W. Hayes, 141-155. Florence, All’Insegna del Giglio. Leone, A. and Mattingly, D. J. 2004. Vandal, Byzantine and Arab Rural Landscapes in North Africa. In N. Christie (ed.), Landscapes of Change: Rural Evolutions in Late Antiquity and the Early Middle Ages, 135-162. Aldershot, Ashgate.

Panella, C. 1973. Appunti su un gruppo di anfore della prima, media e tarda età Imperiale. In A. Carandini and C. Panella (eds.), Ostia III. Le Terme del Nuotatore. Scavo degli ambienti III, VI, VII. Scavo dell’ambiente V e di un saggio nell’area SO. Studi miscellenei 21, 460-633. Roma, De Luca.

Manacorda, D. 1977. Anfore. In A. Carandini and C. Panella (eds.), Ostia IV. Le Terme del Nuotatore. Scavo dell’ambiente XVI e dell’area XXV. Studi Miscellanei 23, 117-254. Roma, De Luca.

Panella, C. 1993. Merci e scambi nel Mediterraneo tardoantico. In Storia di Roma. III, 2, 613-697. Torino, Einaudi.

Mattingly, D. J. 1988. Oil for export? A comparison of Libyan, Spanish and Tunisian olive oil production in the Roman Empire. Journal of Roman Archaeology 1, 33-56.

Peacock, D. P. S. 1984. The amphorae: typology and chronology. In Fulford and Peacock (eds.), 116-140. Peacock, D. P. S., Bejaoui, F. and Ben Lazreg, N. 1989. Roman amphora production in the Sahel region of Tunisia. In Amphores romaines et histoire économique: dix ans de recherches. Actes du Colloque, Sienne, 22-24 mai 1986, 179-222. Rome, École Française de Rome.

Mattingly, D. J. 1989. Olive cultivation and the Albertini Tablets. In A. Mastino (ed.), L’Africa Romana. Atti del VI convegno di studio, Sassari, 16-18 ottobre 1988, 403-415. Sassari, Edizioni Gallizzi. Mattingly, D. J. 1990. Appendix 1: Olive presses and olive oil cultivation. In Hitchner 1990, 248-255.

Peacock, D. P. S., Bejaoui, F. and Ben Lazreg, N. 1990. Roman Pottery Production in Central Tunisia. Journal of Roman Archaeology 3, 59-84.

Mattingly, D. J. and Hitchner B. R. 1993. Technical specifications for some North African olive presses of Roman date. In M. C. Amouretti and J. P. Brun (eds.), La production du vin et de l’huile en Méditerranée, 439-462. Athens, École Française d’Athènes.

Peña, J. T. 1998. The mobilization of state olive oil in Roman North Africa: the evidence of late 4th cent. ostraca from Carthage. In J. T. Peña, J. T. Rossiter, A. I. Wilson, C. Wells, M. Carroll, J. Freed and D. Godden, Carthage Papers. The early Colony’s Economy, Water Supply, A Public Bath, and a Mobilization of State Olive Oil, 117-238. Portsmouth.

Ørsted, P., Carlsen, J., Ladjimi Sebaj, L. and Ben Hassen, H. (eds.), 2000. Africa Proconsularis. Regional Studies in the Segermes Valley of Northern Tunisia III. Copenhagen, Aarhus University Press.

Reynolds, P. 1995. Trade in the Western Mediterranean AD 400-700: The Ceramic Evidence. British Archaeological Reports, International Series 604. Oxford, BAR Publishing.

Pallarès, F. 1987. Alcune considerazioni sulle anfore del Battistero di Albenga. Rivista di studi Liguri 53, 269306. Palmieri, L. 2004. La provincia d’Africa dal I al VI secolo d.C.: analisi dei modelli insediativi presenti sul territorio, Acme, Annali della Facoltà di Lettere e Filosofia dell’Università degli Studi di Milano LVII, III, settembre-dicembre 2004, 81-116.

Riley, J. A. 1979. The Coarse Pottery from Berenice. In J. A. Lloyd (ed.), Excavations at Sidi KhrebishBenghazi (Berenice). Lybia Antiqua, supplement II, 91-467. Tripoli. Sehili, S. 2008. L’huile dela Byzacène. In J. González, P. Ruggeri, C. Vismara and R. Zucca (eds.), L’Africa Romana. Le ricchezze dell’Africa. Risorse, produzioni, scambi. Atti del XVII convegno di studio, Sevilla, 14-17 dicembre 2006, 777-792. Roma, Carocci editore.

Palmieri, L. 2008. I Vandali e l’olio: produzione e commerci nell’Africa del V sec. d.C. In J. Gonzales, P. Ruggeri, C. Vismara and R. Zucca (eds.), L’Africa Romana. Le ricchezze dell’Africa. Risorse, produzioni, scambi. Atti del XVII convegno di studio, 7

LRCW4

Slim, H. 1996. El Jem. L’antique Thysdrus. Tunis, Alif. Slim, H., Trousset, P., Paskoff, R., Oueslati, A. 2004. Le littoral de la Tunisie. Étude géoarchéologique et historique. Paris, CNRS Editions. Trousset, P. 1977. Nouvelles observations sur la centuriation romaine à l’est d’El Jem. Antiquités Africaines 11, 175-207.

Fig. 1. Olive oil trade in Byzacena between the 4th and the first half of the 5th centuries AD.

8

PALMIERI

Fig. 2. Olive oil trade in Byzacena between the second half of the 5th and the first half of the 6th centuries AD.

9

LRCW4

10

PATTERNING LATE ROMAN CERAMIC EXCHANGE IN SOUTHERN TUSCANY (ITALY): THE COASTAL AND INLAND EVIDENCE, I.E. CENTRALITY VS. MARGINALITY EMANUELE VACCARO McDonald Institute for Archaeological Research (University of Cambridge) [email protected] A large sample area of southern Tuscany, including both coastal and inland territories, has been investigated since 2000 by integrating systematic field-walking surveys with excavations at key sites. Over ten years of research have produced a large quantity of data concerning patterns of production, trade and consumption of late Roman pottery. Consistency of data collection and quantification now allows for a comprehensive comparative study of the late Roman pottery circulating in the well-connected coastal areas and more marginal inland territories. This paper seeks to offer a broad picture of late Roman economies in the micro-region by integrating quantified ceramic assemblages from different site types (towns, sea-ports and rural settlements), with data extracted from ploughsoil assemblages. Traditionally, the late Roman period in southern Tuscany has been depicted as a time of progressive economic decline, mainly on the basis of decrease of settlement density as emerged from field surveys. This contribution represents an attempt to re-address the interpretation of late Roman economy in the study-area through a comparative use of pottery. KEYWORDS: SOUTHERN TUSCANY, FIELD SURVEY, EXCAVATION, OVERSEAS AND REGIONAL CONNECTIONS, REDISTRIBUTION PROCESSES, REGIONAL AMPHORAE, REGIONAL TABLEWARE AND COARSEWARE Introduction 1

the early 2000s as part of the larger project Carta Archeologica della Provincia di Grosseto whose aim was to cover the entire province over a few decades. Between 2000 and 2006 a further c.52km2 were systematically surveyed in the coastal area between the Alma and Osa valleys using large transects. If we add up the areas fieldwalked during earlier archaeological surveys and those surveyed more recently a total of c. 68km2 was covered within the broader context between the Alma and Osa valleys. Considering the new and preceding surveys which centered or touched on the coastal study area, a total of 750 surface archaeological sites, plus 114 offsites, were discovered, making a substantial contribution to the understanding of settlement patterns in the longue durée (Vaccaro 2008, 2011).

This paper compares trade patterns for late Roman pottery in coastal and inland areas of southern Tuscany (central Italy) to readdress the general interpretation of economies and settlement in this sub-region, often viewed in the light of models of decline of economic complexity and sophistication given the demographic regression emerging from field survey projects (Francovich and Hodges 2003, 31-43; Valenti 2009). To do so this paper will use and combine the large datasets provided by regional field-walking surveys and focused excavations (Fig. 1). Since the late 1980s the University of Siena has undertaken extensive and systematic field surveys in the province of Grosseto, concentrating on four major river valleys (Alma, Bruna, Ombrone and Osa) covering an area of 846km2. Field research in this territory has not been constant over time and we can identify two main phases: one in the late 1980s and a more recent one begun in the early 2000s. Initially, field work centred on the countryside between the Etruscan to medieval town of Roselle and that of Grosseto, whose urban status was recognized in AD 1138 after a long period of proto-urban development in the early medieval period (Citter and Arnoldus 2007, 134-152). During that pioneering phase of research some 16Km2 were systematically surveyed yielding a significant dataset of surface sites supporting a preliminary analysis of settlement patterns between the Etruscan and medieval periods. A second phase of more extensive and intensive field surveys was undertaken in

As mentioned, research initially centred mainly on coastal areas, although in 2006 the field project was extended towards the interior and specifically to the Cinigiano area, some 30km from the central Tyrrhenian Sea. In an area of about 160km2, as many as 54km2 were sampled using transects and a total of 16km2 were systematically surveyed between 2006 and 2008 yielding 334 surface sites and 137 off-sites (Ghisleni et al. 2011). Alongside field surveys, new site-based pottery studies were carried out to shed light on late Roman and early medieval ceramic trade and consumption patterns at a series of excavated sites, allowing us to better contextualize systems of economic connections, and compare and combine these with surface datasets. Since the relationship between regional and interregional exchange is a key theme in site-based research, this involved the large late Roman ceramic assemblages from two previously excavated areas in the town of Roselle, the late Roman excavated contexts from the cave-site at Scoglietto and from the large settlement at Paduline-

1

The research providing these results was partly funded by the [European Community’s] Seventh Framework Programme [FP7/2007-2013] under grant agreement n° 236093.

11

LRCW4

The coastal area

Serrata Martini as well as, thanks to a collaboration with the Soprintendenza per i Beni Archeologici della Toscana, those from two underwater sites recently discovered during rescue archaeology work in the northern sector of the coastal study area. Data from the underwater assemblages recovered at the sea-port of Portus Scabris and the site of Cala del Barbiere, where a new 4th century AD shipwreck was located, mostly composed of a Tunisian cargo, offer remarkable evidence for the cabotage routes which enabled the trade of overseas goods from major sea-ports to small coastal consumption sites (Vaccaro 2011).

Mediterranean finewares In the late Roman period, the demand for imported Mediterranean finewares was almost exclusively met by ARS, both at the urban site of Roselle and in the countryside. So far, no evidence for trade in LRC or other Mediterranean finewares has been found in the sample coastal area, with the exception of the maritime villa site at Santa Francesca, near Talamone, and the fortified hilltop site at Talamonaccio, which both yielded a few forms of Dérivées-des-Sigillées Paléochrétiennes (henceforth DSP) fired under reducing conditions with a dark-gray slip (Vaccaro 2011, Plate XIX). Two large fragmentary bowls bear a stamped decoration in the shape of palm branches within concentric circles and may belong to form Rigoir 1, typical of Marseille in the 5th to early 6th century AD (Bonifay et al. 1998, 394) and ascribable to Provencal workshops. The presence of DSP only at these two sites may indicate that these materials were redistributed from the port of Cosa, given the significant presence of this class among the non-stratified materials from Brown’s excavations on the Arx and a minimum number of at least one individual in the construction phase of the 6th-century castrum (Fentress et al. 1991, 215).

Another crucial body of evidence has been produced by the Excavating the Roman Peasant Project, underway in inland southern Tuscany since 2009, specifically in the area largely surveyed between 2006 and 2008. This project, an international collaboration between the Universities of Pennsylvania, Cambridge and Siena, is the first in Italy designed to shed light on the architecture, material culture, lifestyle, social and economic networks of Roman and late Roman peasants through the systematic excavation of a sample of small and mediumsized rural sites recently identified during a field walking survey (Ghisleni et al. 2011). This paper combines the results of field surveys and excavations to shed new light on patterns of ceramic production and trade in this sub-regional territory and hence offer a reconsideration of late Roman economic complexity based on fresh data. The combined use of surface and excavation data is potentially useful although great caution is needed as surface materials cannot be used for quantitative analysis unless they are collected and processed using identical analytical methods. With this in mind, we used pottery from ploughed assemblages to define the distribution of some ceramic classes while we relied mainly on excavated materials to calculate relative proportions.

In the coastal study area, apart from these exceptions, the remainder of the Mediterranean finewares consisted of ARS. The study of this ceramic class, based on the use of individual weighted means (Fentress et al. 2004), draws on the combined use of surface data from the territory which yielded a total of 179 identifiable ARS vessels and excavated data from the North Hill in Roselle and late Roman Portus Scabris2 which yielded 190 and 128 identifiable vessels respectively (Fig. 2). The analysis of ARS distribution during the period AD 300-500 at rural and urban sites reveals more similarities than differences: the trend is generally positive, although with some brief periods of decline perhaps resulting from adverse political conditions. These occur in the 430s and at the end of the 5th century and coincide with the Vandal conquest of Carthage, which suddenly interrupted the flow of fiscal goods from Tunisia to Rome, and with the collapse of the western Empire (on the importance of the fiscal route connecting Carthage and Rome see Wickham 2005, 720-728 ). Later, the most interesting period, when the distribution curve for ARS in rural areas is out of line with that in the town, is between the 530s and 550s. At Roselle, attestations of ARS increased slightly from the 520s and 530s and remained constant until mid-century. Probably this was thanks to the site’s strategic and military role during the Gothic war, giving the state a vested interest in ensuring that supplies remained as constant as possible (Celuzza and Fentress 1994). The picture for rural areas at this time is different. From the 530s we see a sudden drop in ARS which appears to be closely linked to a general contraction in trade due to the

A necessary clarification concerns the application of the bipolar centrality versus marginality paradigm in this subregional area. Anyone familiar with late Roman southern Tuscany is well aware that no area here can be considered economically central in the broader Mediterranean context: southern Tuscany lost its centrality as early as the 1st or early 2nd century AD (Carandini and Cambi 2002, 196-217; Vaccaro 2008). Nevertheless centrality and marginality do apply to the common perception of the relationship between coastal and inland territories in the globalised late Roman world, where proximity to searoutes contributed significantly to determining how economically well-connected each region was. The chronological framework is the 4th to the 6th century AD, although in a comparative perspective the period AD 300500 is more promising given the shortage of 6th-century AD data from inland areas.

2

At Portus Scabris the study of ARS focused only on the 5th to 7th-century AD specimens.

12

VACCARO

Amphorae

widespread devastation caused by the Greek-Gothic war, which had a dramatic impact on economies and settlement in most of the Italian peninsula. From the mid6th century onwards both town and countryside again present similar trends. The progressive crisis in acquisitions of ARS in both countryside and town is a structural phenomenon. This trend began long before the 590s, when the Lombard conquest of the area merely interrupted links with the Mediterranean world that had already lost stability and continuity, and become merely episodic. After the 580s/590s ARS forms no longer reach rural areas, not even contexts in direct contact with the coast. ARS had become a rare and valuable product and although it continued to circulate along cabotage routes, as we know from the data from Portus Scabris, it was no longer redistributed to rural sites.

The amphora trend in the coastal area reflects that for ARS. The vast majority of amphorae are Tunisian, whilst other areas are represented to a limited extent with the exception of the Portuguese Almagro 51C, abundant in some 4th- to early 5th-century AD contexts in the town of Roselle, namely those from the North Hill and Room 21 in the Domus dei Mosaici. These assemblages come from two different areas, one to the south of the Forum (Domus dei Mosaici) and one to the north, which underwent similar functional transformations in the late Roman period. A large metal junk shop was established on the Domus dei Mosaici, probably in the late 3rd AD and at the same time some rooms, previously part of a now abandoned bath complex at the foot of the North Hill, were reused for a similar function, although on a smaller scale. Contexts 5147=5170, and to a far more limited extent context 5141, from the area near the North Hill, yielded a total of 24 identifiable amphorae (Vaccaro 2011, 57-59). Except for four individuals which are definitely residuals (two Dressel 14, a Pelichet 46 and a Dressel 7-13), the others are compatible with their presence in phase in a context of the 4th to early 5th century AD (Fig. 3). However, some doubts persist about six Tunisian amphorae of types Africana I/Keay 3B and Africana IIA/Keay 5, with three specimens each (Fig. 4, 5-7). The Africana IIA/Keay 5 variants reveal close affinities with those produced at the El-Assa workshop (Bonifay 2004, 39-41) whose chronology does not seem to go beyond the second half of the 3rd century AD (Bonifay 2004, 111). The specimens of Africana I/Keay 3B are more likely to be in phase as the late variants of this typology were produced up to the second half of the 3rd or 4th century (Bonifay 2004, 107; see also Keay 1984, 106, Fig.40, no. 3). Overall, given the large proportions with which both types are documented in these assemblages we may also be dealing with the largescale reuse of these containers in a period following that of their main circulation. The other Tunisian amphorae documented are three specimens of Keay 25, two of which are generically referable to this type and one (Fig. 4, 8) belonging to the variant Africana IIIC, dating to between the late 4th and mid-5th century AD (for the date Bonifay 2004, 122).

The evidence from Portus Scabris is particularly important as it provides information on the flows of pottery and amphora-borne foodstuffs which continued to use cabotage routes between AD 300 and 700. Though the underwater material was recovered by machine, not stratigraphic excavation, it nonetheless sheds light on the kind of economic connections touching on this stretch of the Tyrrhenian Sea and can be compared with the goods redistributed to the interior. Portus Scabris was not just a place where overseas products were redistributed to surrounding areas but may also have been used as a temporary mooring place for boats in difficulty whose cargoes were destined for other ports. The removal from the hull of materials broken during the voyage thus led to the presence in the underwater deposit of pots and transport containers not destined for this site or for inland contexts. The ARS trend at Portus Scabris reveals that trade flows recovered in the 530s and 540s, followed by another sudden decline from the 550s onwards. The increase in ARS in the Portiglioni harbour, as at Roselle, may be related to the events of the Greek-Gothic war. During military operations, the coastal areas of the Tuscia were of enormous strategic importance to the Byzantines. They used the ports which had fallen under their control to send the necessary supplies to troops deployed in central Italy. The Tuscan ports also played a fundamental role in providing foodstuffs to Rome, as we know from an episode in 544 related by Procopius (De Bello Gothico, Lib.III, X). The trade reaching the harbour at Portiglioni was not primarily commercial, as indicated by the significant decrease in volumes of ARS and transport containers reaching nearby areas during the Gothic war. Rather, trade was encouraged by the reorganization of the fiscal system decided by the Byzantine state (Zanini 1996). Essentially, there does not seem to have been a large-scale redistribution of overseas products towards nearby areas. After the mid-6th century, Portus Scabris also saw a fairly marked decline in volumes of ARS; however, this decrease was less dramatic than in the area between the Alma and Osa valleys and the town of Roselle itself. This suggests that the harbour retained a role of some importance in maritime trade through the central and northern Tyrrhenian at the end of the 6th century (Vaccaro 2011, 235).

Two types of amphorae reflect a peak in the consumption of Portuguese fish sauces: Almagro 51A/B and especially Almagro 51C, with one and five specimens respectively (Fig. 4, 1-2). This is a significant difference in the consumption of imported bulk commodities between the town of Roselle and the coastal countryside, as the Almagro 51C type is occasionally documented only at the large Roman and late Roman site of San Martino de plano. By contrast, this type is fairly common in the 4thto early 5th- urban assemblages as confirmed by the study of the late Roman assemblages from the Domus dei Mosaici currently underway. The analysis of a 4thcentury AD assemblage from Room 21 revealed a small amphora record: five out of a total of 52 in-phase vessels. Interestingly four are of the Portuguese Almagro 51C type and one is a Mauretanian Dressel 30/Keay 1 (Fig. 4, 3-4 and 10). Though mostly absorbed by urban 13

LRCW4

consumption, rather than redistributed to the countryside en masse, the coastal circulation of Portuguese amphorae must have been quite intense as revealed by the so-called “A” shipwreck at Punta Ala, which has been dated to the mid-3rd AD, and whose heterogeneous cargo consisted of Africana II amphorae, Portuguese Almagro 51A/B, Almagro 51C and Beltran 72, and finally Baetic Dressel 20 and Dressel 23 amphorae (Dell’Amico and Pallarés 2007; Vaccaro 2011, 164-165).

110 minimum number of individuals. Although the amphora record is tiny, with three vessels, it reveals persisting connections with at least three areas of supply: north-eastern Sicily or Bruttium with a Keay 52, Tunisia with a Keay 61D/B and finally Cilicia with a LRA 1. Interestingly, the attestation of Mediterranean amphorae not associated with late ARS is paralleled by another context dating to c. AD 550-600 yielded by the excavation at the villa of Paduline-Serrata Martini, at the mouth of the Bruna river (Vaccaro 2011, 105-110). Here a small assemblage composed of 35 vessels (22.8% are residual) mainly from regional and central/southern Tyrrhenian sources provided no certain evidence for inphase ARS, although in the amphora record, with 10 individuals, two vessels are represented by a spike of a generic Keay 62 large amphora and a small spatheion devoid of handles (Vaccaro 2011, 105-106), similar to Bonifay’s type 3A (Bonifay 2004, 128). Another possibly in-phase vessel is attested by some walls of a LRA 2, whilst the other containers are either residuals or unidentifiable given the absence of diagnostics.

The late Roman assemblage from the North Hill in Roselle provides interesting data about the variegated amphora supply to this urban site. Southern Spain contributed some oil, although in negligible quantities, as revealed by a single specimen of Dressel 23 (Fig. 4, 11). Context 5147=5170 also yielded a specimen of the fairly rare Keay 24B (Fig. 4, 9), recently suggested to be a Tripolitanian amphora introduced between the mid-4th and the mid-5th century, definitively replacing the wellknown types of Tripolitanian amphorae documented from the early imperial period (Macías-Remolà 2005, 127). Although the consumption of wine is documented only by three amphora specimens, these nonetheless reveal a fairly varied picture for the wines reaching the Rosellan market, with at least three sources: an Empoli type from the middle Arno valley or the Volterra area (Fig. 4, 14), a MRA1 variant can be assigned to an eastern Sicilian production (Fig. 4, 13), while a rim fragment belongs to either a late Dressel 24 or an early LRA 2 (Fig. 4, 12) thus suggesting occasional eastern Mediterranean connections.

As regards amphora-borne foodstuffs in the coastal area we shall also consider two other classes which shed light on patterns of redistribution from sea-ports to coastal consumption sites: the Empoli wine-amphora and LRA 1. The regional Empoli type, with 94 identified specimens out of a total of 207 transport containers dating to between the 4th – 7th/early 8th century AD, is by far the best documented type at Portus Scabris. This high proportion may be partly explained by the presence of a beached shipwreck carrying Empoli amphorae which has created an imbalance in the proportions of wares from different sources (Vaccaro 2011, 141-143). Clashing with the evidence from Portus Scabris is the very limited redistribution of this type not just to the countryside but also to urban sites. The Empoli amphora is documented with a handful of specimens in 4th- to early 5th-century AD contexts at the towns of Roselle and Vetulonia, whilst only three late Roman rural sites yielded one individual each. This low density of terrestrial finds and the strikingly high concentration at Portus Scabris supports a major use of this regional type to supply wine to Rome given its concentration here, especially in some 5thcentury AD contexts (see for example the mithraeum of the Crypta Balbi: Saguì and Coletti 2004, 244 and the Magna Mater: Panella et al. 2010, 66). As a result, its inland penetration in southern Tuscany is sporadic. The distribution of LRA1 tells a similar story to that of the Empoli type at least in a broad sense. The LR1 with its 5th- and 6th-century variants is documented in the underwater dump at Portus Scabris with c. 6% of the total late Roman amphorae, indicating that the site was used as an intermediate port of call for Tyrrhenian cargoes including eastern Mediterranean amphorae (Vaccaro 2011, 148-149). Again if we look at our study area, the re-distribution of this amphora type is minimal: only two surface and two excavated sites yielded a few fragments of this type. The presence of a LR1 amphora in a mid-6th- to early 7th- AD excavated context in Roselle (supra context 1260) testifies to the occasional import of

Moving on to the countryside we immediately note the predominance of Tunisian amphorae from c. AD 300 onwards. Between the 4th and mid-5th AD, at least 20 out of 32 rural settlements yielded fragments of Tunisian amphorae, mostly Keay 25 and the spatheion 1/Keay 26. The high availability of Tunisian amphorae during the 4th to mid-5th century is evident from the large numbers of Keay 25 amphorae at Portus Scabris and a possible 4thcentury AD shipwreck at Punta Ala, whose partly recovered cargo consisted almost exclusively of Keay 25 amphorae, Hayes 50 A/B dishes and Tunisian kitchenware (Vaccaro 2011, 157-166) The amphora trend changes between the mid-5th and 6th centuries when far fewer rural sites yielded large cylindrical Tunisian amphorae produced from the Vandal period onwards. Keay 35, 55 and 62 amphorae are documented only at about 36% of occupied sites (12 out of 33) (Fig. 5). The late Roman ceramic evidence from Roselle is still quite uneven as we are missing well-quantified contexts dating to between the second quarter and the end of the 5th century AD. Nevertheless it is worth noting that the excavation at the foot of the North Hill yielded two assemblages allowing us to track patterns of pottery trade and consumption until the late 6th and early 7th century AD (Vaccaro 2011, 63-70). One of these contexts appears to be particularly meaningful as it reveals the continuity of trade in overseas amphorae until c. AD 600, despite the absence of in-phase ARS. Context 1260 yielded a total of 14

VACCARO

this type after the Greek-Gothic war, coinciding in this area with the massive reduction of overseas trade.

general fineness of the clays makes it difficult to trace trade routes from production to consumption sites. The primacy of regional tablewares becomes even more overwhelming in the 6th century AD. As shown by the combined evidence of the 5th-century AD contexts from the inland study area and some 6th-century assemblages from Roselle, the production of sigillata chiara tarda dell’Italia centro settentrionale does not continue after the mid- or late 5th AD. However, a new class of regional tableware is introduced in the 6th century AD: the colature rosse ware, whose coating presents either red slip drippings or a more careful red brush-painting. Given the similarity of fabrics often encountered, these two subclasses likely belong to the same workshops and the same craft tradition, although the different degrees of care taken in applying the red coating may suggest an intentional differentiation. It is also worth considering that, whilst the colature rosse ware appears in the first half of the 6th century AD alongside the traditional colour-coated ware, the more carefully red painted ware is never found before AD 550 in this area.

Overseas commodities as a late Roman socio-economic marker? If we combine the distribution of ARS and Tunisian amphorae with the typologies of rural sites in two main phases (4th to mid-5th and mid-5th to mid-6th) we see that access to Tunisian products decreases markedly from one phase to the next, with access to overseas products becoming a prerogative of villa sites, villages and a few well-connected sites along the coast, while they are almost completely absent at farmsteads. The fact that all types of site accessed overseas products between the 4th to mid-5th AD while this access later became limited to larger and better connected sites requires further explanation. The supply of ARS and overseas foodstuffs still reaching major settlements, albeit in smaller quantities, may not indicate simply the continuity of the distribution network for these products. Above all it may suggest the presence of weak elites still active in the study area who expressed their socio-cultural status in part through the consumption of overseas goods, whereas the peasants living at the last few farmsteads were now completely dependent on the regional market for tableware and amphora-borne foodstuffs.

One small context near the North Hill in Roselle (US 1262) yielding only 27 vessels is significant as it reveals that regional products were even more overwhelming than in the 4th and early 5th century: apart from a total of 12 kitchenware vessels, the rest of the material is regional colour-coated ware (11 specimens), colature rosse ware (three specimens) and only one Hayes 99B bowl in ARS (Vaccaro 2011, 63-64). The functional repertoire had by then lost the series of large flat-based dishes common in the 4th and early 5th century AD and a new form, the flanged bowl, was introduced. At the same time, different-sized bowls and jugs continued to be produced. All this suggests some significant transformations in dining habits with a possible shift from the use of large collective plates to individual deep vessels more suited to the consumption of semi-liquid foods and small cuts of meat than large cuts (on this see also Fontana 1998, 96).

Regional Red Slip and other Tablewares Whilst trade patterns for overseas products show a progressive decline from the beginning of the 6th AD, the circulation of local and regional products suggests a wide variety of regional connections which, as we will see, linked coastal and inland areas especially in the 4th and 5th centuries AD and possibly beyond. In coastal areas the distribution of regional tableware is overwhelming compared to imports from as early as the 4th-early 5th AD, as the quantifiable excavated data from Roselle demonstrates. Considering the two groups of ceramic assemblages respectively from the area at the foot of the North Hill and Room 21 in the Domus dei Mosaici, the vast majority of table vessels consists of regional colour-coated ware characterised by a thin, nonsintered or semi-sintered red slip and a class of products with an higher-quality sintered red slip named sigillata chiara tarda dell’Italia centro settentrionale, sometimes bearing a brown over-painted decoration on the bottom of large open vessels (Vaccaro 2011, 53-56). These two classes together account for respectively five and over seven times the ARS specimens in the two datasets. The intensive supply of regional tablewares can be explained by the wide variety of functional forms in the repertoires, such as different-sized bowls, flat-based dishes, basins and closed vessels, better able than imported ARS vessels, mostly represented in the 4th and early 5th century AD by flat-based dishes and large bowls, to provide a complete set of table vessels. Colour-coated and slipped tableware circulated throughout the region (Cantini 2009; Menchelli and Pasquinucci 2012) and the

Two other contexts, from Roselle (US 1260) and Paduline-Serrata Martini, provide additional evidence for the long lasting production of regional tableware well into the second half of the 6th and early 7th century AD. Context 1260 from Roselle yielded a total of 110 minimum vessels, 88% of which are very likely in phase. Interestingly this context yielded no ARS apart from some residual forms whilst regional colour-coated and colature rosse wares accounted for 38 and eight vessels each. The variety of functional forms is still quite broad although less so than in the 4th and early 5th century AD: the absence of large flat-based dishes in the late contexts is corroborated by this data, whilst the presence of flanged-bowls, small bowls and jugs is consistent with the pattern observed in the first half of the 6th AD (Vaccaro 2011, 66-68). At the villa site of PadulineSerrata Martini a mid-6th to early-7th rubbish dump sealed the praefurnium of the bath complex and marked the final abandonment of the site. It yielded a total of nine vessels including colour-coated (four individuals) and colature rosse/painted (five individuals) wares accounting for c. 33% of all in-phase pottery. The persisting variety 15

LRCW4

mid-6th AD in association with as many as 66 small bronze coins whose chronologies ranges from the mid3rd to 5th centuries AD and two bronze coin weights corresponding respectively to the solidus aureus and perhaps the tremissis. The complexity of this ceramic assemblage, a total of 59 vessels with some 10% ARS and, to a lesser extent, Tunisian coarseware and amphorae, the presence of many nummi which continued to be used for daily transactions well after the date of their issue, the two bronze coin weights and the location of the site in a spacious cave directly on the Roman rivermouth suggests that it served as a small beach-site supporting coastal routes and intimately connected to them (Vaccaro 2011, 94-95).

of the ceramic repertoire offers a range of open and closed vessels, including different-sized flanged bowls, basins and occasional closed vessels. The latter are also common in the class of non-slipped regional tableware (Vaccaro 2011, 107-108). Kitchenware The kitchenware evidence has potential which will be further exploited in the continuation of this comparative study between coastal and inland territories of southern Tuscany. As part of the Excavating the Roman Peasant Project, we plan to extensively apply thin section analysis to cooking wares in the sub-region to define production areas and trace the routes which supported regional and local trade with specific emphasis on the rural sites of the interior. So far, the study has been mostly based on the chrono-typological and functional approach, of limited use in defining circulation patterns. Nevertheless it is worth outlining some trends. The 4th- and early 5thcentury evidence from Roselle shows that the functional repertoire of cooking vessels is very varied, with cooking pots, lids, casseroles/large cooking bowls, bowl/lids and cooking dishes/tegami (Fig. 6) and a very balanced ratio of open to closed vessels, complementary as they serve different cooking purposes, such as braising/frying and boiling food (Arthur 2007). Interestingly, the whole 4thand early 5th-century repertoire of kitchenware documented in Roselle was likely produced regionally given the capillary distribution of these types in late Roman urban and rural sites throughout Tuscany (Vaccaro 2011 with bibliography). The only possible exception is a total of three individuals of Tunisian kitchenware, two dish/lids and a casserole, which might be still in phase. The surface evidence from field-walked sites indicates a similar pattern, with the most common types documented in Roselle regularly yielded by the ploughed ceramic assemblages (Vaccaro 2011, 92-94).

The rich kitchenware evidence (44 of a total of 59 vessels) is particularly meaningful. In addition to a few Tunisian forms (a casserole Fulford CW 18.1, and two dish/lids similar to types Fulford CW lids 4.3 and Fulford CW dishes or lids 4 and 8), two generically extraregional/western Mediterranean vessels and eleven local/regional forms, a group which includes both open and closed vessels was assigned to workshops in Latium or Campania, with the former far more likely given the widespread presence of similar vessels in late Roman contexts in Rome (Fig. 7). The typological study combined with the minero-petrographic analysis, revealing the presence of volcanic inclusions in a sample of three particularly distinctive types, corroborated the presence of a large group of extra-regional kitchenware possibly from Latium, previously suggested in light of the morphological study of the vessels and macroscopic characterisation of the fabrics (Vaccaro 2011, 100-103). The wide repertoire documented in this group offers a variety of cooking pots. One type, with a thickened rim, of pseudo-rectangular section, squared or pointed on the outer face (Fig. 8, 10-12), is particularly common in late Roman contexts in Rome (Conservatory of San Pasquale: Fogagnolo 2004, 594, Fig. 9, no.73; Basilica Hilariana: Vatta and Bertoldi 2004, 466-465). Other types, also common in Rome (Vaccaro 2011, 100-103), are equally well documented in the Scoglietto deposit (Fig. 8, 13-14). The repertoire of closed vessels is completed by a cooking jug (Fig. 8, 15). Open vessels are also well documented with a wide variety of functional forms suitable for braising and frying food. A highly distinctive type of casserole with flaring walls and a thickened incurving rim (Fig. 8, 1-4) which finds many parallels in Rome in the 5th and 6th centuries (e.g. Schola Praeconum I: Whitehouse et al. 1982, 75, Fig. 9, no.116; Crypta Balbi: Saguì and Coletti 2004, 251, Fig. 6, no.31; Basilica Hilariana: Pacetti 2004, 451, Fig. 8, no.62) is widely attested. Another type of casserole, one type of cooking dish/tegame, two bowl/lids and seven lids are documented as well, all from the same area of production (Fig. 8, 5, 6-7 and 8-9). Particularly interesting is the absence of a significant distribution of these vessels towards coastal consumption sites, such as Roselle, where the distinctive types of cooking pot with a thickened rectangular rim and the casserole with incurving rim are apparently absent. Possibly these

Moving on to the 6th century, the picture of kitchenware variety offered by contexts 1262 (first half of the 6th) and 1260 (mid-6th to early 7th) does not radically change in Roselle, although the primacy of cooking pots becomes overwhelming among open vessels (Fig. 6). Casseroles, bowl/lids and cooking dishes/tegami continue to be documented but in far smaller quantities than a century before (Vaccaro 2011, 64-65 and 68-70). There is no specific evidence for extra-regional kitchenware. However, the combination of morphological and thin section 3 analysis applied to the kitchenware repertoire reveals that some interregional trade existed along the coast, probably taking advantage of cabotage routes and reaching two sites in particular: the cave site of Scoglietto, right on the ancient mouth of the Ombrone river, and the villa site of Paduline-Serrata Martini. Recent excavations at the cave site yielded a coherent ceramic assemblage dated to between the mid-5th and 3

Thin section analysis is carried out in collaboration with Claudio Capelli, to whom I express my thanks for the preliminary results presented here.

16

VACCARO

amphora 4, probably for wine. Although no whole vessels have hitherto been found it seems to combine the grooved handles of the early to mid-imperial Spello type, a rounded rim and the characteristic base of the late Roman Empoli type (Fig. 9, 14-15). The production of a similar amphora may have started at a local or sub-regional scale in the 3rd century AD, as a partially preserved individual from a mid-Roman dump at Case Nuove suggests (Fig. 10, 1).

vessels were traded sporadically from Rome to a few coastal sites to the north. The little kitchenware evidence from the mid-6th- to early 7th-century dump at the villa site of Paduline-Serrata Martini some 17km north of Scoglietto corroborates this tendency of Latium kitchenware to ciculate along the coast. Four vessels (of a total of 10 kitchenware forms) have fabrics with volcanic inclusions similar to those of Scoglietto and their morphology suggests close comparisons with Latium. Two cooking pots with thickened and rectangular more and less everted rims (Fig. 8, 16-17) are a variant of the type already seen at Scoglietto, whilst a cooking jug and a bowl/lid (Fig. 8, 18-19), find parallels respectively in the late Roman and early medieval vicus at Madonna del Passo in Sabina (Patterson and Roberts 1998, 431, Fig. 8, no.4) and in Rome in 6th- and 7th-century contexts (Crypta Balbi: Saguì and Coletti 2004, 266, Fig.15, no.83; Basilica Hilariana: Pacetti 2004, 448, Fig.6, nos 43-44 and 456, Fig.11, no.83).

At Case Nuove, overseas amphorae are only documented by walls belonging respectively to a Tunisian and a Portuguese specimen, while one identifiable individual is a regional Empoli type. Here the overall quantity of ARS is just over 4% of 72 minimum vessels, while at Pievina ARS accounts for over 8% across all the late Roman contexts analyzed so far (195 MNI). At Case Nuove, ARS is documented only with three MNI, two in C3 production and one in D2. C3 products are represented by a dish Hayes 50B and a small bowl of type Hayes 71B (Fig. 10, 2-3). The only D2 product is a large bowl or dish Hayes 64 (Fig. 10, 4). The lower proportion of overseas fineware, then, is counterbalanced by the higher quality of the regional tablewares dumped at Case Nuove. At both sites we found a very varied repertoire of regional table vessels with a broad range of functional types. While at Pievina tableware is represented by colour-coated ware with a low-quality slip which often makes it impossible to distinguish it from non-slipped productions, this is never the case at Case Nuove. Here, not only does the colour-coated ware have a better quality slip, but the site is also well supplied with Sigillata chiara tarda dell’Italia centro-settentrionale (about 15% of the total), with a lustrous sintered or semi-sintered slip, occasionally over-painted. Interestingly this class is absent at Pievina. The sigillata chiara tarda dell’Italia centro-settentrionale from Case Nuove includes a wide range of different open vessels of which the best documented are large flat-based dishes imitating the ARS prototype Hayes 61A/B3 and A/B4 respectively dated to the first half of 5th and first third of 5th AD (Fig. 10, 5-6) and above all the series of flat-based dishes and bowls deriving from Hayes 50B in ARS (Fig. 10, 7-8). Also documented in this class are small bowls like that illustrated in Fig. 10, 9 which refers to a production not influenced by ARS models but probably deriving from autonomous morphological repertoires. Interestingly this class occasionally has an over-painted decoration which in one case presents concentric circles surrounded by dots (Fig. 10, 10), deriving from the single or concentric circles with dot-fringe typical of styles A (ii) or A (iii) of the mid-4th through to mid-5th AD ARS repertoire (Hayes 1972, 236-237). Finally one base of a large open vessel is characterised by a peculiar ante-cocturam graffito in the shape of a small hand (Fig. 10, 11). Closed forms are documented by only one small vessel, possibly a table flask (Fig. 10, 12). The varied repertoire of table

The interior Our analysis of late Roman pottery trade and consumption systems in the interior is based on a combination of field survey data, a series of wellstratified contexts from the small farmstead at Pievina (late 4th-late 5th AD) and a large rubbish dump at Case Nuove (late 4th-mid-5th AD), related to a peasant site near a villa. ARS, amphorae and regional tablewares As regards overseas products, 4th and 5th centuries ARS was collected, in very low percentages, at almost every late Roman surface site while the evidence for Tunisian amphorae was limited to a smaller number of settlements (Fig. 1). In the 6th century only the villa site at Santa Marta indicated episodic links with overseas trade (Ghisleni 2009). The excavated ceramic contexts from the sites of Pievina and Case Nuove reveal that the vast majority of wares were manufactured regionally. At Pievina, 11.5% of imported pottery is documented across all the late Roman contexts from the late 4th to the end of the 5th century, with ARS, occasional Tunisian kitchenware, Tunisian amphorae Keay 25 and 26, Portuguese Almagro 51 A/B and C, and LR1 (Fig. 9, 1-13). Interestingly, all the Mediterranean amphora types documented are also attested in the town of Roselle, but more important is the high percentage of Portuguese amphorae in roughly contemporary urban contexts. This, alongside the presence at Roselle of all the ARS forms documented at Pievina may suggest that the town was a redistribution point for small inland farmsteads. Amphora finds at Pievina included a hitherto-unknown regional or local

4

Two samples were analysed in thin section: one revealed generic inclusions which do not exclude a local/regional source and one is again generically regional, although it may be tentatively referred to the Arno valley.

17

LRCW4

vessels provided by sigillata chiara tarda dell’Italia centro-settentrionale is further enriched at Case Nuove by the wide variety of forms in colour-coated ware. Compared to regional fineware, the repertoire of colourcoated ware reveals an even more varied range of functional open vessels, introducing a series of new forms such as a large deep basin (Fig. 10, 14), a large bowl with notched lip, possibly deriving from an original rethinking of the prototypes Hayes 67 and Hayes 68 (Fig. 10, 15), two tiny sauce-bowls (Fig. 10, 18-19) and finally an hemispherical grooved bowl (Fig. 10, 16). Also of interest is the larger number of closed vessels represented by large table jar and handled table jugs (Fig. 10, 21-22). On the other hand the very varied series of large bowls and dishes imitating the ARS prototypes Hayes 61 with variants A/B3 and A/B4 and Hayes 50B (Fig. 10, 13 and 17) unites colour-coated ware and sigillata chiara tarda dell’Italia centro-settentrionale.

kitchenwares documented at Case Nuove and the greater homogeneity of fabrics, compared with the wider variety of Pievina, tentatively suggest a more targeted and possibly centralized selection of local and regional products, also supported by the large quantities of sigillata chiara tarda dell’Italia centro-settentrionale, perhaps determined by proximity to the villa site and hence the possibility of accessing better quality pottery (Vaccaro and MacKinnon forthcoming). Conclusions The comparative analysis suggests that the relationship between coastal and inland areas may be closer than expected. During the late 4th to late 5th AD, when the evidence from our two areas is more homogeneous we observe intensive regional exchange and the supply of overseas products to sites far inland like Pievina, which could not even rely on proximity to major roads or large and wealthier sites. The overseas products widely available at coastal sites and especially in the town of Roselle were more or less regularly re-distributed towards the interior. Possibly the items to be shipped inland on overland routes were selected on the basis of size: the smaller the item the cheaper its transport cost. ARS vessels and small amphorae were particularly suited to this trade system. Probably it is no accident that heavier 5th- and 6th-century cylindrical Tunisian containers, documented at a series of coastal sites, are completely absent from the excavated peasant sites inland. Evidently size, weight and shape made this kind of products unsuited to overland transport, preventing their distribution inland except to the only Roman and late Roman villa attested in this inland area. This all tells us clearly that if inland peasants still engaged some widedistance trade connections, these were opportunistically limited to foodstuffs transported in small containers. At the same time the distribution of regional tablewares and kitchenwares reinforces the connections between coastal and inland territories, with some differences. In the interior, the comparative analysis of Pievina and Case Nuove revealed differing access to the higher quality regional tablewares which parallels the attestation in the coastal area of sigillata chiara tarda dell’Italia centrosettentrionale almost exclusively in the urban sites of Roselle and Vetulonia and a few villa sites. The sizable presence of this class at Case Nuove has been linked to its proximity to the villa site at Santa Marta.

The presence of high-quality regional fineware at Case Nuove and not at Pievina is remarkable as in the coastal area it has hitherto been identified only in towns, two major late Roman villa sites and one well-connected settlement on the sea. To explain this, we must consider the position of Case Nuove. The settlement is located just 500 metres as the crow flies from the only villa site identified in this inland area. The site served in the Augustan and late Roman periods as an agricultural facility and after the end of each phase of utilization, was intermittently used as a garbage dump. The presence of the villa site, used at least up to the 6th century, is likely to have encouraged the establishment of a more sophisticated demand for high quality regional pottery which consequently became more easily accessible to peasants using the site of Case Nuove and dumping their late Roman ceramics there. Thus the somewhat more sophisticated repertoire of regional tablewares at Case Nuove compared to Pievina should be interpreted in light of its proximity to the villa site. Kitchenware Both at Pievina and Case Nuove, kitchenware seems to come from local or regional workshops. As shown in Fig. 6 the ratio of closed to open vessels is totally in favour of the former (lids, given their size, are all compatible with the cooking pots) revealing a predilection for boiled food (Fig. 9, 27-34 and Fig. 10, 23-32). Although the morphological and functional repertoires do not support any marked distinction between the two sites, a closer analysis of cooking pot fabrics reveals a significant predominance at Case Nuove of vessels tempered with spathic calcite, which may have ensured higher resistance to thermal shock than cooking pots manufactured with other fabrics (Tite et al. 2001, 322). Interestingly about 55% of the cooking pots from this site are calcitetempered whereas only 24% of the cooking pots from the late Roman farmstead at Pievina present this technological feature. This difference requires further explanation. Were consumers at Case Nuove aware of the thermal shock resistance of different types of cooking pots? The possible higher technological quality of the

To sum up, the late Roman trade connections between coastal and inland areas of southern Tuscany suggest the existence of a still sophisticated exchange system which, well into the 5th century, involved both overseas and regional products and continued to make large-scale use of late Roman small denomination bronze coins, as the evidence from the small inland peasant sites reveals (Ghisleni et al. 2011).

18

VACCARO

Fulford, M.G. and Peacock, D.P.S. 1984. Excavations at Carthage: The British Mission, I.2. The Avenue du President Habib Bourguiba, Salammbo. The Pottery and other ceramic objects from the site. Sheffield.

Bibliography Arthu r, P. 2007. Pots and boundaries. On cultural and economic areas between late antiquity and the early middle ages. In M. Bonifay and J.-C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry, 15-27. British Archaeological Reports, International Series 1662 (II). Oxford, BAR Publishing.

Ghisleni, M., Vaccaro, E. and Bowes, K. 2011. Excavating the Roman Peasant I: Excavations at Pievina. Papers of the British School at Rome 79, 95145. Ghisleni, M. 2009. Il sito di Santa Marta nel contesto delle dinamiche insediative di età romana, tardo antica e medievale della media valle dell’Ombrone (Cinigiano, GR). In G. Volpe and P. Favia (eds.), V Congresso Nazionale di Archeologia Medievale (SAMI V), 243-247. Florence, All’Insegna del Giglio.

Bartolini, E. (ed.), Procopio di Cesarea, La Guerra Gotica (De Bello Gothico). Milan, Tea. Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports, Int. Series 1301. Oxford, BAR Publishing.

Hayes, J. W. 1972. Late Roman Pottery. London, British School at Rome.

Bonifay, M., Carre, M. B. and Rigoir, Y. (eds.), 1998. Fouilles à Marseille: les mobiliers (1er-VIIe siècles ap. J.-C.). Paris, Errance.

Keay, S. 1984. Late Roman Amphorae in the Western Mediterranean. A typology and economic study: the Catalan evidence. British Archaeological Reports, Int. Series 196. Oxford, BAR Publishing.

Cantini, F. 2009. Produzione, circolazione e consumo del vasellame decorato con ingobbio rosso in Toscana tra I-II e XIII secolo. In E. De Minicis (ed.), Le ceramiche di Roma e del Lazio in età medievale e moderna 6, 59-79. Rome, Kappa.

Macías, J. M. and Remolà, J. A. 2005. La cultura material de Tarraco-Tarracona (Hispania TarraconensisRegnum Visigothorum): cerámica común y áanforas. In J. Ma. Gurt i Esparraguera, J. Buxeda i Garrigós, M. A. Cau Ontiveros (eds.), LRCW1. Late Roman Coarse Wares, Cooking wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. British Archaeological Reports, International Series 1340, 125-135. Oxford, BAR Publishing.

Carandini, A. and Cambi, F. (eds.), 2002. Paesaggi d’Etruria. Valle dell’Albegna, Valle d’Oro, Valle del Chiarone, Valle del Tafone. Rome, Edizioni di Storia e Letteratura. Celuzza, M. G. and Fentress, E. 1994. La Toscana centromeridionale: i casi di Cosa-Ansedonia e Roselle. In R. Francovich and G. Noyé (eds.), La storia dell’altomedioevo italiano (VI-X secolo) alla luce dell’archeologia, 601-613. Florence, All’Insegna del Giglio.

Menchelli, S. and Pasquinucci, M. 2012. Ceramiche con rivestimento rosso nella Tuscia settentrionale. Rei Cretariæ Romanæ Favtorvm Acta 42, 229-237. Pacetti, F. 2004. Celio. Basilica Hilariana: scavi 19871989. In L. Paroli and L. Venditeli (eds.), Roma dall’antichità al medioevo II. Contesti tardoantichi e altomedievali, 418-434. Milan, Electa.

Citter, C. and Arnoldus Huyzendveld, A. (eds.), 2007. Archeologia urbana a Grosseto. Origine e sviluppo di una città medievale nella "Toscana delle città deboli". Le ricerche 1997 – 2005. I: la città di Grosseto nel contesto geografico della bassa valle dell'Ombrone. Florence, All’Insegna del Giglio.

Panella, C., Saguì, L., Casalini, M. and Coletti, F. 2010. Contesti tardoantichi di Roma: una rilettura alla luce di nuovi dati. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. Comparison between eastern and western Mediterranean, British Archaeological Reports, International Series 2185, 57-78. Oxford, BAR Publishing.

Fentress, E., Clay, T., Hobart, M and Webb, M. 1991. Late Roman and medieval Cosa I: the arx and the structure near the Eastern Height. Papers of the British School at Rome 59, 197-230. Fogagnolo, S. 2004. Trastevere. Conservatorio di San Pasquale: dal quartiere romano all’occupazione medievale. In L. Paroli and L. Venditeli (eds.), Roma dall’antichità al medioevo II. Contesti tardoantichi e altomedievali, 579-597. Milan, Electa.

Patterson, H. and Roberts, P. 1998. New light on dark age Sabina. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo, Atti del Convegno in onore di John W. Hayes, 421-435. Florence, All’Insegna del Giglio.

Fontana, S. 1998. Le “imitazioni” della sigillata africana e le ceramiche da mensa italiche tardo-antiche. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo, Atti del Convegno in onore di John W. Hayes, 83-100. Florence, All’Insegna del Giglio.

Saguì, L. and Coletti, C.M. 2004. Contesti tardoantichi dall’area a S-E della Crypta Balbi. In L. Paroli and L. Venditeli (eds.), Roma dall’antichità al medioevo II. Contesti tardoantichi e altomedievali, 242-277. Milan, Electa.

Francovich, R. and Hodges, R. 2003. Villa to Village. London, Duckworth.

19

LRCW4

Tite, M. S., Kilikoglou, V. and Vekinis, G. 2001. Strength, Toughness and Thermal Shock Resistance of Ancient Ceramics, and their Influence on Technological Choice. Archaeometry 43, 301-324.

Valenti, M. 2009. La Toscana rurale del V secolo. In P. Delogu and S. Gasparri (eds.), Le trasformazioni del V secolo. L'Italia, i barbari e l'Occidente romano, 495-529. Turnhout, Brepols.

Vaccaro, E. 2011. Sites and Pots: Settlement and Economic Patterns in Southern Tuscany (AD 300900). British Archaeological Reports, International Series 2191. Oxford, BAR Publishing.

Vatta, G. and Bertoldi, T. 2004. Celio. Basilica Hilariana: scavi 1997. In L. Paroli and L. Vendittelli (eds.), Roma dall’antichità al medioevo II. Contesti tardo antichi e altomedievali, 458-479. Milan, Electa.

Vaccaro, E. 2008. An overview of rural settlement in four river basins in the province of Grosseto on the coast of Tuscany (200 B.C.-AD 600). Journal of Roman Archaeology 21, 225-247.

Whitehouse, D., Barker, G., Reece, R. and Reese, D. 1982. The Schola Praeconum, I. The Coins, Pottery, Lamps and Fauna. Papers of the British School at Rome 50, 53-101.

Vaccaro, E. and MacKinnon, M., forthcoming. Pottery and animal consumption: new evidence from the 'Excavating the Roman Peasant Project'. HEROM. Journal on Hellenistic and Roman Material Culture.

Wickham, C. 2005. Framing the early Middle Ages: Europe and the Mediterranean 400-800. Oxford, Oxford University Press. Zanini, E. 1996, Ricontando la terra sigillata africana. Archeologia Medievale 23, 677-688.

20

VACCARO

Fig. 1. The coastal and inland sample areas in southern Tuscany (province of Grosseto).

10 8 6 4 2 0

70 100 130 160 190 220 250 280 310 340 370 400 430 460 490 520 550 580 610 640 670 Surveyed Area

Roselle North Hill

Portus Scabris

Fig. 2. The ARS trend in the sample coastal area, at Roselle (North Hill) and Portus Scabris. The ARS evidence from the latter site refers to the period AD 400-690. Amphora type "Empoli" type MRA 1 Early LRA 2 or late Dressel 24 Keay 13A Almagro 51C Almagro 51 A/B Keay 24B Africana III/Keay 25 Africana IIA/Keay 5 Africana I/Keay 3B Residuals

MNI 1 1 1 1 5 1 1 3 3 3 4

Fig. 3. Quantified amphorae (minimum number of individuals) from a selection of 4th to early 5th century AD contexts from the North Hill in Roselle.

21

LRCW4

Fig. 4. Selection of amphorae from 4th- and early 5th-century contexts in Roselle (all from the area of the North Hill, with the exception of numbers 3-4). Nos 1-4: Almagro 51C; no.5: Africana I/Keay 3B; nos 6-7: Africana IIA/Keay 5; no.8: Africana IIIC/Keay 25; no.9: Keay 24B; no.10: Dressel 30/Keay 1; no.11: Dressel 23; no.12: late Dressel 24 or early LRA2; no.13: MRA1; no.14: Empoli type.

4th to mid 5th c. AD sites 32

Sites with Tunisian amphorae 20

Sites with ARS

Mid 5th to mid 6th c. AD sites 33

27

Sites with Tunisian amphorae 12

Sites with ARS 16

Fig. 5. Relationship between rural sites, ARS and Tunisian amphorae in the two periods 4th-mid 5th and mid 5th-mid 6th AD.

22

VACCARO

Forms

Casseroles (TKW) Dish/lids (TKW) Cooking pots (KW) Casserole/cooki ng bowls (KW) Lids (KW) Bowl/lids (KW) Cooking dishes/tegami (KW) Cooking dishes/testi (KW) Cooking jugs (KW) Flanged bowl (KW)

Roselle. Room 21, Domus dei Mosaici (4th) MNI 0

Roselle. North Hill (4th-early 5th) MNI

Pievina SU 1019, 1026 (late 4th-mid 5th) MNI

Pievina SU 1018 (mid 5th?) MNI

Pievina SU 1006 (midlate 5th) MNI

0

Case Nuove (late 4thmid 5th) MNI 0

0

Roselle NH US 1262 (first half of the 6th) MNI 0

Roselle NH US 1260 (mid 6th to early 7th) MNI 0

1

0

0

2

0

0

0

0

0

0

7

15

11

19

11

7

9

24

4

13

1

1

2

1

2

3

7 0 3

13 2 4

2 0 0

5 0 2

7 0 1

3 0 1

0 0 0

2 1 2

0

0

1

1

0

0

0

0

0

0

0

0

1

0

1

0

0

0

0

0

0

0

0

1

Fig. 6. Kitchenware from a series of analyzed contexts (TKW=Tunisian Kitchenware; KW=regional or local kitchenware): forms and MNI

Forms/sources

Scoglietto cave (mid 5th to mid 6th) MNI 1 2 1

Casseroles (TKW) Dish/lids (TKW) Cooking pots (generic W Mediterranean?) Lids (generic W Mediterranean?) Cooking pots (Local/regional) Cooking lids (Local/regional) Cooking dishes/tegami (Local/regional) Cooking pots (Latium) Casseroles (Latium) Cooking dishes/tegami (Latium) Bowl/lids (Latium) Lids (Latium) Cooking jug (Latium)

1 6 4 1 9 8 1 2 7 1

Fig. 7. Kitchenware from the site at Scoglietto cave (mid 5th-mid 6th AD): forms and MNI. TKW=Tunisian kitchenware.

23

LRCW4

Fig. 8. Kitchenware from Latium documented at the two sites of Scoglietto cave and Pauline-Serrata Martini.

24

VACCARO

Fig. 9. The late Roman pottery “connections” of the farmstead at Pievina (Phase 2.2 = late 4th-mid 5th; Phase 2.3 = mid 5th; Phase 2.4 = mid to late 5th).

25

LRCW4

Fig. 10. Selected pottery from the late 4th-mid 5th AD rubbish dump at Case Nuove, with the exception of regional amphora no.1 from a context dated to the first half of the 3rd AD.

26

SUPPLY AND DISTRIBUTION OF LATE ROMAN COARSEWARES FROM THE NEPI SURVEY PROJECT PHILIP MILLS¹, ULLA RAJALA² ¹School of Archaeology and Ancient History, University of Leicester, University Road, Leicester, LE1 7RH, United Kingdom; [email protected] ²Stockholm University, Wallenberglaboratoriet, Lilla Frescati vägen 7, SE-106 91 Stockholm, Sweden; [email protected] The use of Lynch’s elements of urban form is a useful approach to the study of the ceramic material recovered during a field survey around Nepi, Italy. It has potential for functional, fineware and fabric analysis for investigating the urban hinterland around this Roman town. Our research shows that these are powerful tools by which site type and status can be defined, and that patterns of supply and marketing of different wares in the study region can be mapped, within a chronological framework, so that the local marketing networks through the urban centre as well as the wider Roman world can be studied. KEYWORDS: NEPI, ITALY, SURVEY, FIELD WALKING, FUNCTIONAL ANALYSIS, SUPPLY, SITE TYPE, SITE STATUS Introduction

The chronological framework

This paper is based on material recovered by the Nepi Survey Project of the University of Cambridge, directed by Dr Simon Stoddart and led in the field by Ulla Rajala under the auspices of the Tiber Valley project. The Nepi Survey focused on exploring the extent of pre-Roman settlement, whereas the Roman ceramic material was studied in 2008 by Philip Mills as part of ‘The Romanisation of a Faliscan Town’ project, funded mainly by a Small Grant supplied by the British Academy. The survey was carried out within the predicted territory of ancient Nepi in central Italy c. 45km NW of Rome. This hypothetical territory was defined by using Thiessen polygons over the town’s hinterland and transects were drawn inside this area (Fig. 1). The coherence of the survey area (see Fig. 2) based around a single urban centre has made it an ideal dataset for developing the concept of the ‘ceramiscene’ (Mills and Rajala 2011a; 2011b), a theoretical approach developed by the authors based on the ‘taskscape’ (Ingold 1993; also Rajala 2012). We also wanted to test how well the models used for characterizing urban form may be applied to the urban hinterland. A potentially useful model has been developed by Ellis (1995) using the elements of urban design defined by Lynch (1960). These elements comprise: Nodes, Pathways, Districts, Landmarks and Boundaries. The work of Hanson and Hillier (1984) incorporated hierarchical structures into this model. These simple classificatory tools allow characterizing the different contemporary components of the Roman rural and urban landscapes as a whole within the chronological framework defined by the pottery assemblage.

The sites identified, with their preliminary interpretations, are shown in Fig. 2. The date distribution of all pottery, by site type, by rim equivalent (RE), is shown in Fig. 3. This clearly shows that there was a large amount of Roman pottery deposited during the rural expansion, with fluctuations due to the civil wars and a large sustained peak in the Augustan period, with a sharp decline in the deposition of pottery to just the urban centre, cemeteries and larger villas from the late 2nd century AD, with a resurgence and reoccupation from the later 3rd century until the 6th century AD (Figs. 4-5). These fluctuations follow the patterns already established in earlier studies in the region (cf. e.g. Potter and King 1997; Di Giuseppe 2008; Witcher 2008). The spatial framework Nodes Lynch defines nodes as the strategic points around a city that the observer can enter and which are the foci to and from which they are travelling. The obvious analogues to Lynch’s nodes are the sites identified during the Nepi survey. The rich materials and history of landscape study in this region has bequeathed a robust methodology of defining site types based on extent and density of scatters, alongside other material traces (Kahane et al. 1968, 154; Potter 1992; Witcher 2006). Functional analysis (Evans 2001) has been applied to the pottery recovered from these sites, and the data showing a clear structured correlation (Tables 1 and 2, Fig. 6) with the predicted site type. Whilst further comparison with quantified Italian assemblages is important, it is clear that the method has strong potential as a powerful tool for site identification. 27

LRCW4

Missing sites

Coupled with fine ware figures (Fig. 7), the relative status between similar site types can also be compared.

The evidence for the presence or absence of a site can be missed for a number of reasons; including the lack of sharp differences in densities inside modern fields, low surface densities , large ‘halos’, not having access to land, crop cover, intensive agricultural activity or modern development. The data analysis for nodes and districts especially suggests a means of defining how some missing sites could be accounted for, in terms of location, type and status. Conversely these data also provide support for the absence of certain site types, so that apparent clustering of site types can be modelled and tested. For example, no villas in the NE sector of the survey area were recorded because of erosion and intensive modern land use. However, the clustering of amphorae and fine wares in the surveyed fields suggest possible locations where a villa may have been centred (Figs. 8–10).

Districts In a city these areas are identified by different common characteristics and perceived as having geographic extent. In the rural hinterland they can be defined by the geographical areas, which are bordered by edges and paths, and by the range of geographical tools, which have been developed for this purpose. It is hoped that more human districts, i.e. properties, could be recovered by the analysis of the pottery. The distribution of African red slip (Fig. 8) and North African amphora sherds (Fig. 9) both show a clustering, which would not be expected by unstructured deposition. This could indicate that they remain close to their site of primary use (in this case later villas) and would be a plausible indicator of property extent to be tested against districts defined by other means (e.g. weighted Thiessen polygons). The distribution of North African coarsewares (Fig. 10) show a very different pattern, albeit for a much smaller quantity of material, that suggests existence of further sites and underlines the effect of post-depositional processes.

Supply The detailed examination of fabrics allows supply to be mapped. The amphorae show an interesting pattern and a date distribution (Fig. 11) very different from the survey as a whole emphasizing an economic change with high status sites dominating the countryside and a change in the nature of imperial links and imported products. The distribution of fine wares and amphorae is also consistent with the remarketing of these products through the urban centre. This can be contrasted with the patterns exhibited by regionally produced oxidized wares. The most common fabrics are shown in Fig. 12. Most fabrics are common from the Republican/ Early Imperial period, with a hiatus in the 3rd century and a small resurgence in the 4th century AD. The exception is the late Roman organic tempered ware O51, which is only present in forms from the late 4th century onwards. Fig 13 shows the relative presence on the different sites of the fabrics, which show a variety of different marketing strategies, strikingly avoiding the urban centre in the case of the later O51.

Paths These are the channels through which the observer moves, and clearly fit into a hierarchical structure. The main information for this survey is supplied by the known courses and rank of the ancient Roman roads (Frederiksen and Ward Perkins 1957). Additional information about these routes and other pathways is also provided using such resources as Google Earth and can be tested against predicted lines based on the other surviving elements and terrain modelling. Edges Those linear elements not used or considered paths can be defined by the ravines and waterways in this highly dissected landscape (although rivers may also have acted as paths in their own right). Whilst it should be borne in mind that some edges may not be archaeologically recoverable, such as for example marshy ground or tree cover, these more permanent natural restrictions on human movement are still a useful element for characterizing the ancient landscape.

Conclusions This paper investigates the potential for pottery data to be used beyond constructing a chronological framework in field survey projects. The use of functional and fineware analysis is seen as a powerful tool for understanding site type and status. The study of quantified assemblages aids the reconstruction of ancient supply over time within a periurban environment. In this way we can see the development of the landscape around Nepi from an Etruscan farm based economy through the rise of the Roman villa economy, accompanied by a migration to the urban centre, with a gradual resettlement of the countryside in late antiquity and evidence in a sharp change in trading patterns. The theoretical framework which underpins this approach is hoped to provide an integrated approach to studying the interactions between town and country.

Landmarks These are defined as the reference points that the observer does not enter. Whilst many may not survive in the archaeological record, others, such as mountains and ravines will survive to the present day. It should also be remembered that the other elements may also act as landmarks.

28

MILLS-RAJALA

Acknowledgements

Ingold, T. 1993. The Temporality of the Landscape. World Archaeology 25, 152–174.

The authors are grateful to Simon Stoddart and the original Nepi Survey team. The survey was carried out under concession from and in close collaboration with Soprintendenza per l’Etruria meridionale under the supervision of Dott.ssa Daniela Rizzo. The survey was supported and assisted by the British School at Rome and its staff together with the mayor and personnel of Comune di Nepi. The authors want to thank the British Academy and McDonald Institute for Archaeological Research, Cambridge, for granting funding for the original fieldwork and the various funding bodies, including the British Academy, the Academy of Finland and New Hall College among the many that supported the PhD research of Ulla Rajala.

Kahane, A., Murray Threipland, L. and Ward Perkins, J. 1968. The Ager Veientanus, North and East of Veii. Papers of the British School at Rome 36, 1–218. Lynch, K. 1960. The Image of the City. Cambridge (Mass.), The MIT Press. Mills, P. J. E. and Rajala, U. R. 2011a. Interpreting a Ceramiscene Landscape - the Roman Pottery from the Nepi Survey Project. In D. Mladenović and B. Russell (eds.), Proceedings of the Theoretical Roman Archaeological Conference, 2010, 1–17. Oxford, Archaeopress. Mills, P. and Rajala, U. 2011b. The Roman ceramic material from the fieldwalkingin the environs of Nepi. Papers of the British School at Rome, 147–240.

‘The Romanisation of a Faliscan Town’ project was kindly financed by the British Academy. The study got assistance from the British School at Rome, Museo civico di Nepi and its director Stefano Francocci as well as the Museo archeologico dell’Agro Falisco (Civita Castellana) and its staff.

Potter, T. W. 1992. Reflections of twenty-five years’ fieldwork in the Ager Faliscus. Approaches to landscape archaeology. In M. Bernardi (ed.), Archeologia del paesaggio II, 637–666. Florence, All’Insegna del Giglio.

Bibliography

Potter, T. W. and King, A. C. 1997. Excavation at The Mola di Monte Gelato: a Roman and Medieval Settlement in South Etruria. Archaeological Monograph of the British School at Rome 11. London, The British School at Rome.

Di Giuseppe, H. 2008. Assetti territoriali nella media Valle del Tevere dall’Epoca orientalizzante a quella repubblicana. In F. Coarelli and H. Patterson (eds.), Mercator Placidissimus – The Tiber Valley in Antiquity: New Research in the Upper and Middle River Valley. Rome, 27–28 February 2004. Quaderni di Eutopia 8, 431–465. Rome, Quasar.

Rajala, U. 2012. Political landscapes and local identities in Archaic Central Italy – interpreting the material from Nepi (VT, Lazio) and Cisterna Grande (Crustumerium, RM, Lazio). In S. Stoddart and G. Cifani (eds.), Landscape, Ethnicity and Identity in the Archaic Mediterranean Area, 120–143 Oxford, Oxbow.

Ellis, S. P. 1995. Prologue to a Study of Roman Urban Form. In P. Rush (ed.), Theoretical Roman Archaeology: Second Conference Proceedings, 92– 104.

Witcher, R. 2006. Broken pots and meaningless dots? Surveying the rural landscapes of Roman Italy. Papers of the British School at Rome 74, 39–72.

Evans, J. 2001. Material approaches to the identification of different Romano-British site types. In S. James and M. Millett (eds.), Britons and Romans: advancing an archaeological agenda. CBA Research Reports 125, 26–35.

Witcher, R. 2008 The Middle Tiber Valley in the Imperial Period. In F. Coarelli and H. Patterson (eds.), Mercator Placidissimus – The Tiber Valley in antiquity: New Research in the Upper and Middle River Valley. Rome, 27–28 February 2004. Quaderni di Eutopia 8, 467–486. Rome, Quasar.

Frederiksen, M. W. and Ward Perkins, J. B. 1957. The ancient Road Systems of the central and northern Ager Faliscus (Notes on southern Etruria, 2). Papers of the British School at Rome 25, 67–203. Hillier, B. and Hanson, J. 1984. The social logic of space. Cambridge, Cambridge University Press.

29

LRCW4

Fig. 1. Site location (illustration by U. Rajala)

Fig. 2. Sites identified in the survey (with field-walked field units shaded; illustration by U. Rajala)

30

MILLS-RAJALA

Fig. 3. Stacked bar chart date distribution by RE by settlement type

Fig. 4. Sites with pottery dating from the 3rd and 4th century AD (illustration by U. Rajala)

31

LRCW4

Fig. 5. Sites with pottery dating from the 5th and 6th century AD (illustration by U. Rajala)

Fig. 6. Functional types at selected sites (illustration by U. Rajala)

32

MILLS-RAJALA

Fig. 7. Ware groups at selected sites and spreads (illustration by U. Rajala)

Fig. 8. ARS distribution (with 3rd-to-6th-century sites with late Roman rims; illustration by U. Rajala)

33

LRCW4

Fig. 9. North African amphora distribution (with 3rd-to-6th-century sites with late Roman rims; illustration by U. Rajala)

Fig. 10. North African coarseware distribution (with 3rd-to-6th-century sites with late Roman rims; illustration by U. Rajala)

34

MILLS-RAJALA

Fig. 11. Amphora date range and distribution

35

LRCW4

Fig. 12. The main oxidized wares (drawings by P. Mills) 36

MILLS-RAJALA

Fig. 13 Distribution of the main oxidized fabrics at selected site (illustration by U. Rajala) MNR

A

J

P

BK

Cu

M

B

D

L

N

O

6.7%

6.7%

20.0%

0.0%

6.7%

6.7%

0.0%

26.7%

6.7%

20.0%

0.0%

15

10.0%

0.0%

45.0%

5.0%

0.0%

0.0%

0.0%

30.0%

0.0%

10.0%

0.0%

20

2.2%

8.9%

51.1%

2.2%

2.2%

0.0%

0.0%

17.8%

6.7%

8.9%

0.0%

45

Building Burial/ Cemetery

F

Farm Halo

3.3%

7.3%

31.3%

0.0%

8.0%

0.7%

2.7%

33.3%

3.3%

10.0%

0.0%

150

Large Villa

7.5%

3.8%

33.8%

0.0%

0.0%

0.0%

1.3%

35.0%

3.8%

15.0%

0.0%

80

Minor

0.0%

25.0%

50.0%

0.0%

25.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

8

Roadside

0.0%

5.6%

22.2%

0.0%

0.0%

0.0%

0.0%

55.6%

5.6%

11.1%

0.0%

18

Settlement

0.0%

0.0%

33.3%

0.0%

0.0%

0.0%

0.0%

66.7%

0.0%

0.0%

0.0%

3

Villa

5.7%

5.7%

42.9%

0.0%

2.9%

0.0%

0.0%

34.3%

0.0%

5.7%

2.9%

35

Overall

4.5%

6.4%

35.6%

0.5%

4.5%

0.5%

1.3%

32.1%

3.5%

10.7%

0.3%

374

Table 1 Functional break down by settlement type by minimum number of rims (MNR) RE

A

F

J

P

BK

Cu

M

B

D

L

N

O

3.6%

24.1%

15.7%

0.0%

8.4%

6.0%

0.0%

20.5%

4.8%

16.9%

0.0%

83

32.0%

0.0%

44.8%

0.0%

0.0%

0.0%

0.0%

20.8%

0.0%

2.4%

0.0%

125

Farm

5.6%

9.3%

53.0%

1.9%

1.5%

0.0%

0.0%

15.7%

4.5%

8.6%

0.0%

268

Halo

12.3%

9.2%

27.8%

0.0%

6.6%

0.9%

2.2%

28.7%

3.8%

8.5%

0.0%

773

Large Villa

13.9%

4.2%

30.8%

0.0%

0.0%

0.0%

0.0%

32.1%

3.8%

15.1%

0.0%

548

Minor

0.0%

39.0%

40.3%

0.0%

20.8%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

77

Roadside

0.0%

14.3%

27.5%

0.0%

0.0%

0.0%

0.0%

51.6%

4.4%

2.2%

0.0%

91

Settlement

0.0%

0.0%

21.1%

0.0%

0.0%

0.0%

0.0%

78.9%

0.0%

0.0%

0.0%

19

Building Burial/ Cemetery

Villa

12.1%

9.3%

35.8%

0.0%

3.3%

0.0%

0.0%

34.0%

0.0%

3.7%

1.9%

215

Overall

11.6%

9.2%

33.3%

0.2%

3.9%

0.5%

0.8%

28.1%

3.2%

9.0%

0.2%

2199

Table 2 Functional type by settlement type by rim equivalent (RE)

37

LRCW4

38

A ROAD STATION ON THE TABULA PEUTINGERIANA. EXCAVATIONS AT ASTURA GIJS TOL, PETER ATTEMA University of Groningen, Groningen Institute of Archaeology, Poststraat 6, 9712 ER Groningen, the Netherlands [email protected]; [email protected] In the summers of 2007 and 2008 a team of the Groningen Institute of Archaeology (GIA) mapped and sampled a section with archaeological materials exposed by marine erosion near the mouth of the Astura river (Nettuno, Lazio, Central Italy). The principal aim of the study was to obtain insight in the pottery wares and shapes that circulated in the wider study area between the late Roman and the early Medieval period, a phase for which sites are few in the database of GIA’s long-running Pontine Region Project. This article provides an overview of the types of amphorae, coarse and cooking wares attested. These indicate that the site participated in long-distance trade networks, probably using the harbour at Torre Astura as a landing-stage. KEYWORDS: PONTINE REGION, PONTINE REGION PROJECT, LATE ROMAN, EARLY MEDIEVAL AMPHORAE, COARSE WARES, ASTURA summers, 2007 and 2008, the section was studied with two specific aims in mind.

Introduction Between 2004 and 2006 the Groningen Institute of Archaeology (GIA) compiled the archaeological map of the municipality of Nettuno, a middle-sized town located on the Tyrrhenian seaboard, approximately 60km south of Rome (Attema et al. 2009; 2011). Field surveys around the town recorded a persistent decline in the number of sites dating between the Mid Imperial and Early Medieval period and a strong decrease in the rate of pottery consumption (Tol 2012; Attema et al. 2010). The same phenomenon is noted throughout the entire Pontine region and more in general in Tyrrhenian Central Italy (e.g. Patterson et al. 2004 for the middle Tiber valley; Capanna and Carafa 2009 for the direct surroundings of Rome). However, the scale of this decline – at least for the Pontine region –is still hard to establish due to a poor knowledge of the coarse and cooking wares as well as the amphorae of this period. We therefore aimed at investigating a site that could provide insight in the variety of wares and shapes circulating in our study area between the mid Imperial period and the early Middle Ages.

1) To identify and date all features visible in the section 2) To establish a local sequence of pottery in circulation in our study area between the late Imperial and the Medieval periods. The excavation In the two campaigns the section was straightened, cleaned and drawn over a length of 95m, divided into nineteen sections measuring five meters each. From each section a sample of archaeological materials was collected. The remains of a wall structure (identified in section I) were studied in more detail. After cleaning, its extension landward was measured and the soil in the foundation trench was excavated horizontally (Fig.2). The stratigraphy Five different strata were recorded throughout the section (Fig. 3). The upper two layers, containing many roots of plants and trees (layers 1 and 2) were topsoil and consisted of sandy humus-rich soil, with layer 2 having a slightly lighter colour. Layer 3 was composed of relatively dense light grey sand of varying thickness, in sections A-I and P-R more than a meter thick, but elsewhere measuring no more than 20cm. The lowest part of this layer contained a band of archaeological materials, generally 10-20cm thick. The disposition of the materials (on top of each other) and the fact that they date from different periods suggests that this ‘finds layer’ either represents a secondary deposition or is formed by deflation (Given 2004, 18). Deflation is typical for sandy soils and occurs in environments prone

In 2004, between the large villa complex of Torre Astura and the mouth of the Astura river (Fig. 1), a section exposed by marine erosion was observed by students of the GIA. Its location corresponds to the southernmost edge of a large site that was tentatively identified by Fabio Piccarreta (1977) as the settlement Astura, depicted on the Tabula Peutingeriana. The section contained a thick band of archaeological materials with a substantial late Roman to early Medieval material component. Following a second visit, during which this observation was confirmed, permission was asked and obtained by the Soprintendenza per I Beni Archeologici del Lazio to conduct further investigations. In two consecutive 39

LRCW4

to strong eolic influences (as in coastal areas). Layers 4 and 5 were the so-called sabbie rosse, dune sands of Late Pleistocene origin that are characterized by their reddish colour (Feiken 2011, 2). Layer 4 is somewhat darker in colour and has many iron-manganese concretions. Layer 5, not present in all sections and always positioned below this layer, is paler in colour and contains less iron.

Hammamet 3 amphora (Fig. 5.13). Also products from the eastern Mediterranean found their way to the site. For the Kapitän 2 amphora an Aegean origin is assumed. The somewhat later ‘Late Roman amphorae’, identified on the basis of their typical body rilling, are also eastern products; the LRA 1 was certainly produced in Cyprus and Cilicia; the LRA 2 probably originated in the Argolid and the LRA 3 probably derives from Asia Minor (Figs. 5.7 and 5.12).

The material remains

Several wall fragments bearing ‘combed’ decoration as well as fragments of fondi ombellicati can be attributed to globular amphorae, perhaps of the Castrum Perti type (Figs. 5.14-5.15; Murialdo 1995). These containers, probably carrying olive oil, date between the later 6th and the early 8th century AD. The main production region of these vessels is thought to have been modernday Tunisia, but the production of similar amphorae is attested in the eastern Mediterranean and in southern Italy as well (Arthur 1992; 1993; Petrone et al. 1994). Furthermore ten rim fragments remain unidentified (Figs. 5.16-5.17).

Sample areas in section D and H and the foundation trench of the wall structure in section I, yielded the bulk of the collected materials. This comprised a rich collection of building materials, fine wares (mainly African red slip ware), amphorae, coarse and cooking wares, oil lamps, glass, coins and faunal materials. The collection of fragments of the latter three categories was aided by the use of a small-grained sieve. The materials indicate occupation in two different periods, between the 4th and 7th century and the later 12th or early 13th century AD respectively. Below we focus exclusively on the amphorae, coarse and cooking wares belonging to the first occupation phase. A full report on the excavation is published elsewhere (Tol 2012).

Cook and coarse ware pottery Altogether 4716 fragments of utilitarian pottery were found, including 1166 diagnostic fragments. The sample of cooking wares comprises fragments of African cookware, Aegean cookware, Micaceous Ware and Pantellerian Ware, as well as a large variety of shapes that are thought to be of local or regional production. Among the coarse wares is a small number of African imports; the bulk of the collected fragments is however of local or regional manufacture.

Amphorae The sample contained 425 fragments of amphorae; 49 of these could be assigned to a specific type based on their morphological characteristics (Fig.4). 1 A small number of Late Republican to Mid Imperial amphora types was found (Dressel 1 and 2-4, Dressel 20, Africana 1 and 2, Cretoise 2; Figs. 5.1-5.3). Most identified types however date in the late Roman period. The most common amphora of this period is the Keay 52 amphora (Figs. 5.8-5.9). This small wine container was supposedly produced in the southern part of the Italian peninsula, in an area comprising eastern Sicily and western Calabria (Gasparetti and di Giovanni 1991; Pacetti 1998; Siena, et al. 1998). The consumption in this period of North African olive oil is attested by the presence of containers from both Tripolitania as well as modern day Tunisia (Tripolitanian 3, Africana 2d and 3b/c; Figs. 5.4-5.6 and 5.10), while a single fragment of a Spatheion tentatively indicates that also other commodities from this region found their way to our site (Fig. 5.11). The consumption of North-African olive oil continued at least until the late 5th century AD, or possibly even later, as is clear from the presence of a

A total of 39 diagnostic fragments of African cookware were collected, comprising both fragments of casseroles and associated lid fragments (Figs. 6.1-6.10). The dominant forms are Hayes 197, 23B and 196, products that are extremely common on inland sites as well (Attema et al. 2009/2010; Tol 2012). These types generally date in the Mid-Imperial period, but their production is thought to continue into the Late Imperial period (Bonifay 2004, 224). Some Late Imperial African cookware types are attested as well, but no fragments post-date the 5th century. The sampling yielded two fragments of Aegean cookware (Fig. 6.16). They both belong to a large dish that is not uncommon on coastal settlements in the western Mediterranean and date between the mid-5th and mid-6th century AD (Waksman and Tréglia 2007, 651, fig. 1.14; Fulford and Peacock 1994, 188, fig. 71.5.4). Two fragments, belonging to the same bowl type, are in so-called ‘micaceous ware’ (Fig. 7.1; Fulford and Peacock 1984, 12-13, fabric 1.6/1.7). This ware is thought to have been produced in Sardinia, north-east Sicily and/or Calabria and dates between 400 and 600 AD.

1

The typological table accompanying the amphora fragments does not include a bibliography. Most types were identified by using the USAP-database (http://ads.ahds.ac.uk/catalogue/resources.html?amphora2005) It must be stressed that amphora types with clear diagnostic features are easier to identify. Certain amphora types are only recognizable by their specific rims, whereas other types have also distinct handles, bases or bodies.

40

TOL-ATTEMA

The Astura excavations yielded seven diagnostic fragments of ‘Pantellerian ware’ (Fulford and Peacock 1984, 8-10, fabric 1; recently Santoro Bianchi et al. eds. 2003). The presence of this ware has recently been identified on various coastal sites in the western Mediterranean. Its production comprises the period between the mid-4th century and the late 6th century AD. In our samples four casseroles, similar in size and body form, but with variations in the shapes of the rim, belong to this ware as well as a single associated lid type (Figs. 7.2-7.6).

A large variety of jar types is recorded as well. Three examples with a concave rim interior have analogies with a type that is widely attested in contexts in both Latium and Campania between the 4th and the 6th century AD (Figs. 8.4-8.5; Arthur 1994, 237, type 32). Two fragments of a jar with a slightly thickened rim were collected (Figs. 8.6-8.7; see e.g. Ricci 1998, 354, fig.2.8-2.9); this shape is common in Latium between the 4th and the 7th century AD. A jar with a heavily thickened folded rim is known from various Roman contexts between the 4th and the first half of the 7th century AD (Fig. 8.8; Arena et al. 2001, 573, fig.V.3.8).

The dominant pottery shape of local or regional production is the casserole. The most common form attested, well-known from contemporaneous Central Italian contexts, is characterized by a marked triangular rim shape and dark red clay containing many augite particles (e.g. Calabria and Patilli 2005, 302, fig. 8.5; Ikäheimo 2010, 414, fig. 3.7; Bertoldi and Pacetti 2010, 440-441, types 6 and 7). Various variants of this shape disclose a typo-chronological development of this form, with seventh century examples having a longer rim and a slightly concave rim interior (Figs. 7.7.-7.12).

Among the coarse wares, six fragments can be identified as north-African imports. Five of these belong to basins (Figs. 6.11-6.14), whereas the sixth fragment is a twisted handle of a jug (Fig. 6.15). All these shapes are commonly found in (central) Italian contexts and cover a period between the early 3rd and the 7th century AD. The most commonly attested coarse ware shape of local/regional production is the bowl, having a wide variety of forms. Three of these are flanged bowls, a shape common between the 5th and the 7th century AD (Figs. 8.9-8.10 and 8.12). Another type, characterized by an almost horizontal rim, is found in Lazio as well as Campania in this same period (Fig. 8.11; Whitehouse et al. 1985, 183, fig.7.31). The sample also includes two fragments of a bowl type that is extremely common in 5th and 6th century contexts in Campania, characterized by a distinct thickening on the interior of the rim (Figs. 8.13-8.14; Arthur 1994, 192-193, types 61-62) and a single fragment of a type bearing a marked ridge on the interior of the rim (Fig. 9.1). The latter shape is similar in shape (although not in size) to a Mid-Imperial African bowl type (Fontana et al. 2009, 301, fig. 16.28). Another bowl is of possible African origin (Fig. 8.18; see also Bonifay 2004, 236), while for four other types no parallel was found (Figs .8.15-8.17 and 8.19).

Another large group of casseroles is characterised by a pronounced convex thickened rim; this shape is attested at Castrum Perti in Liguria and dated between the 4th and the 6th century AD (Figs. 7.13-7.14; Bonora et al. 1988, 370, TAV.XIV.1-8; Murialdo et al. 1998, 247, fig.8.9-8.10). Two casserole fragments have a marked triangular thickening on the exterior of the rim (Fig. 7.16). This type is frequently found in central-Italian contexts of the first half of the 5th century AD (e.g. Whitehouse et al. 1982), whereas two other fragments are close to a casserole type found during excavations near present-day Seville in Spain and date in the late 5th or early 6th century AD (Fig. 8.3; Amores et al. 2007, 159, fig. 3). Considering their absence in contemporaneous central-Italian contexts an Italian manufacture for these pieces can be excluded. Five other casserole types are attested by a single example. The first (Fig. 8.1) is in the same fabric as the earlier mentioned casseroles with triangular rim and therefore probably of local origin as well. A casserole with a convex thickened rim finds a parallel in a context in Rome, where it was dated between the later 5th and 6th century AD (Fig. 7.15; Fogagnolo 2004, 594, TAV.IX.76). Two examples with a stepped rim were retrieved. The first has a parallel at the Monte Gelato (Fig. 7.18; Potter and King 1997, 357, fig. 235.179), whereas a fragment similar to the second (Fig. 7.17; Pacetti 2004, 456, TAV.XI.86) was found during excavations at the Basilica Hilariana in Rome. They both date between the later 4th and the mid-6th century AD. A deep casserole with a thickening of the interior lip is identical to a shape of Campanian origin, common between the 4th and the 7th century AD (Fig. 8.2; Arthur 1994, 223, type 2). A casserole with a short overhanging rim is found in various Roman contexts and dates between the second half of the 6th or the 7th century AD (Fig. 7.19; Ricci 1998, 353, fig.1.3-1.4).

Two different types of basins/catini are included in our sample. The first of these, characterized by horizontal rilling on the exterior of the body, is commonly found in contexts in and around Rome between the second half of the 5th and the late 7th century AD (Fig. 9.2; see also various examples in Sagui and Coletti 2004; Ciarrocchi et al. 1993, 412, fig. 12.8). The other type could however not be dated (Fig. 9.3). The sample of jug types comprises a collection of welldocumented Central Italian types (Figs. 9.4-9.6); for two other types no clear parallel could be found (Figs. 9.79.8). Furthermore two different types of anforette were unearthed (Figs. 9.9-9.10). The first of these, with an upright triangular rim with broad thick handles attached to it, is identical to one of the typical shapes from the 7th century deposit of the esedra at the Crypta Balbi. This type was possibly imported from the eastern coastal area of Spain and is also documented in Sicily and Sardinia (Ricci 1998, 373, fig. 13.5). The second, characterized by a groove on the interior of the rim and the attachment of a handle on its exterior, is similar to a 41

LRCW4

Arthur, P. 1994. Il complesso archeologico di Carminiello ai Mannesi, Napoli (scavi, 1983-1984). Galatina, Congedo.

fragment from the Crypta Balbi, dating in the 6th century AD (Arena et al. 2001, 198, fig. 1.5.12a). Furthermore a small sample of fragments in ceramica acroma depurata was collected, mostly belonging to different types of basins/catini (Figs. 9.11-9.15). These fragments date between the late 5th and 7th century AD (and possibly even later).

Attema, P. A. J., de Haas, T. C. A. and Tol, G. W. 2009. Nettuno, il territorio dalla preistoria al medioevo: la carta archeologica. Pomezia, Arti Grafiche. Attema, P. A. J., de Haas, T. C. A. and Tol, G. W. 2010. The Astura and Nettuno surveys of the Pontine Region Project (2003-2005), 2nd and final report. Palaeohistoria 51/52, 169-327.

Preliminary conclusions and future work The principal aim of the study of the section at Astura was to expand our knowledge of materials circulating in the wider study area between the Late Roman to Early medieval period. Although at Astura no primary stratigraphy is preserved, the investigations have given insight in the range of pottery wares and shapes consumed on this site between the 4th and the 7th century AD. Although many of these shapes appear to be restricted to the settlement of Astura, other types have now been attested on sites surveyed by us resulting in the identification of an additional number of sites with evidence for post-Roman occupation. It is above all the large Roman villae that continue to be inhabited until the late 5th and early 6th century AD, a date that is in line with the latest villa occupation in other regions of Central Italy.

Attema, P. A. J., de Haas, T. C. A. and Tol, G. W. 2011. Between Satricum and Antium: Settlement dynamics in a coastal landscape in Latium Vetus. Babesch Supplementa Series 18. Leuven, Peeters. Attema, P., Derks, T. and Tol, G. W. 2010. The ‘Carta archeologica’ of Nettuno, evidence for Late Antique and Early Medieval settlement on the coast of South Lazio near Antium and Torre Astura (Italy). In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry, Comparison between western and eastern Mediterranean, British Archaeological Reports, International Series 2185 (I), 447-457. Oxford, BAR Publishing.

The many imported products attested on this site, suggest that the settlement of Astura used the harbor of Torre Astura as a landing place. It must have been one of the harbors of call for commercial ships travelling north, as can be inferred from similarities in both the ware types and the dominant pottery shapes attested in contemporaneous contexts at Portus, Ostia and Rome (Ciarrocchi et al. 1993; various contributions in Paroli and Venditelli 2004).

Bertoldi, T. and Pacetti, F. 2010. Materiali di V secolo dalla Basilica Hilariana sul Celio: analisi tipologica delle ceramiche comuni. In LRCW 3, 433-445. Bonifay, M. 2004. Etudes sur la céramique romaine tardive d’Afrique, British Archaeological Reports, Int. Series 1301. Oxford, BAR Publishing. Bonora, E., Falcetti, C., Ferretti, F., Fossati, A., Imperiale, G., Mannoni, T., Murialdo, G. and Vicino, G. 1988. Il “castrum” tardo-antico di S. Antonino di Perti, Finale Ligure (Savona): fasi stratigrafiche e reperti dell’area D. Seconde notizie preliminari sulle campagne di scavo 1982-198. Archeologia Medievale 15, 335-396.

Bibliography Amores, F., Garcia, E., González, D. and López, P. 2007. Cerámicas de cocina y comunes tardoantiguas en Hispalis (Sevilla, España). Primera tipología y sus contextos (La Encarnación, ss. V-VI). In M. Bonifay and J.-C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry. British Archaeological Reports, International Series 1662 (I), 147-164. Oxford, BAR Publishing.

Calabria, M. E. and Patilli, T. 2005. Nuove acquisizioni dallo scavo di Ferento. In E. De Minicis and A. Giuntella (eds.), Le ceramiche di Roma e del Lazio in età medievale e moderna 5, 298-304. Roma, Kappa Edizioni di Architettura e Psicologia. Capanna, M. C. and Carafa, P. 2009. Il progetto ‘Archeologia del suburbio di Roma’ per la ricostruzione dei paesaggi agrari antichi. In V. Jolivet, C. Pavolini, M. A. Tomei and R. Volpe (eds.), Suburbium II: Il suburbio di Roma dalla fine dell'età monarchica alla nascità del sistema delle ville (V-II secolo a.C.). Collection de l'École Française de Rome 419, 27-39.

Arena, M. S., Delogu, P., Paroli, L., Ricci, M., Sagui, L. and Venditelli, L. (eds.), 2001. Roma dall’antichità al medioevo: archeologia e storia nel museo nazionale romano Crypta Balbi. Milan, Electa. Arthur, P. 1992. Fornaci altomedievali ad Otranto: Nota preliminare. Archeologia Medievale 19, 91-122.

Ciarocchi, B., Martin, A., Paroli, L. and Patterson, H. 1993. Produzione e circolazione di ceramiche tardoantiche ed altomedievali ad Ostia e Porto. In L. Paroli and P. Delogu (eds.), La Storia economica di

Arthur, P. 1993. Early Medieval Amphorae, the Duchy of Naples and the food supply of Rome. Papers of the British School at Rome 48, 231-244. 42

TOL-ATTEMA

Roma nell’alto Medioevo alla luce dei recenti scavi archeologici, 203-246. Firenze, All’Insegna del Giglio.

Dall'antichità al medioevo vol. 2. Contesti tardo antichi e altomedievali, 435-457. Milan, Electa. Paroli, L. and Vendittelli, L. (eds.), 2004. Roma. Dall'antichità al medioevo vol. 2. Contesti tardo antichi e altomedievali. Milan, Electa.

Feiken, H. 2011. Geology and Geomorphology of the Nettuno Area. In P. Attema, T. de Haas and G. Tol (eds.), Between Satricum and Antium. Settlement Dynamics in a Coastal Landscape in Latium Vetus, 1-8. Leuven, Peeters.

Patterson, H., di Giuseppe, H. and Witcher, R. 2004. Three south Etrurian 'Crises': first results of the Tiber Valley Project. Papers of the British School at Rome 72, 1-36.

Fogagnolo, S. 2004. Un contesto ceramico di VIII secolo. In L. Paroli and L. Venditelli (eds.), Roma. Dall’antichità al medioevo vol.2. Contesti tardo antichi e altomedievali, 569-575. Milan, Electa.

Petrone, D., Siena, E., Troiano, D. and Verrocchio,V. 1994. Una fornace d'età bizantina a Castellana di Pianella (PE). Archeologia Medievale 21, 269-286.

Fontana, S., Ben Tahar, S. and Capelli, C. 2009. La ceramica tra l’età punica e la tarda antichità. In A. Drine, E. Fentress and R. Holod (eds.), An island through time: Jerba studies, volume 1, The Punic and Roman periods. Journal of Roman Archaeology, Supplementary Series 71, 241-327.

Piccarreta, F. 1977. Astura. Firenze, Leo Olschki. Potter, T. W. and King, A. C. 1997. Excavations at the Mola di Monte Gelato. Archaeological Monographs of the British School at Rome 11. Rome. Ricci, M. 1998. La ceramica comune dal contesto di VII secolo della Crypta Balbi. In L. Sagui (ed.), 351382.

Fulford, M. G. and Peacock, D. P. S. 1984. Excavations at Carthage: The British Mission I,2. The Avenue of president Habib Bourguiba, Salammbo: The pottery and other ceramic objects from the site. Sheffield.

Sagui, L. (ed.), Ceramica in Italia: VI-VII secolo, Atti del Convegno in onore di John. W. Hayes. Firenze, All’Insegna del Giglio.

Fulford, M. G. and D. P. S. Peacock 1994. Excavations at Carthage, The British Mission II, 2: The Circular Harbour, North Side. The Pottery. British Academy Monograph in Archaeology 5. Oxford.

Sagui, L. and Coletti, C. M. 2004. Contesti tardo antichi dall’area a S-E della Crypta Balbi. In L. Paroli and L. Venditelli (eds.), Roma. Dall'antichità al medioevo vol.2. Contesti tardo antichi e altomedievali, 242-277. Milan, Electa.

Gasparetti, G. and di Giovanni, V. 1991. Precisazioni sui contenitori calabresi della tarda antichità (le anfore tipo Keay LII). Mélanges de l’École Française de Rome. Moyen-Age, Temps Modernes 103.2, 875-885.

Santoro Bianchi, S., Guiducci, G. and Tusa, S. (eds.), 2003. Pantellerian Ware. Archeologia subacquea e ceramiche da fuoco a Pantelleria. Palermo, D. Flaccovio.

Given, M. 2004. Mapping and Manuring: Can We Compare Sherd Density Figures? In S. E. Alcock and J. F. Cherry (eds.), Side-by-side Survey: Comparative Regional Studies in the Mediterranean World, 13-21. Oxford, Oxbow.

Siena, E., Troiano, D. and Verrocchio, V. 1998. Ceramiche dalla Val Pescara. In L. Sagui (ed.), 665704. Tol, G. W. 2012. A fragmented history. A methodological and artefactual approach to the study of ancient settlement in the territories of Satricum and Antium, Groningen Archaeological Studies 18. Groningen.

Hayes, J. W. 1972. Late Roman pottery. London, The British School at Rome. Ikäheimo, J. P. 2010. Regional cookwares of the Rome area in AD 400-550: preliminary evidence from the Palatine Hill. In LRCW 3, 409-415.

Waksman, Y. and Tréglia, J.-C. 2007. Caractérisation géochimique et diffusion méditerranéene des céramiques culinaires «égéennes». In LRCW 2, 645657.

Murialdo, G. 1995. Alcune considerazioni sulle anfore Africane di VII secolo del “Castrum” di S. Antonino nel Finale. Archeologia Medievale 22, 433-454. Murialdo, G., Olcese, G., Palazzi, P. and Parodi, L. 1998. La ceramica comune in Liguria nel VI e VII secolo. In L. Sagui (ed.), 227-252.

Whitehouse, D., Barker, G., Reece, R. and Reese, D. 1982. The Schola Praeconum I: The coins, pottery, lamps and fauna. Papers of the British School at Rome 50, 53-101.

Pacetti, F. 1998. La questione delle Keay LII nell’ambito della produzione anforica in Italia. In L. Sagui (ed.), 185-208.

Whitehouse, D., Costantini, L., Guidobaldi, F., Passi, S., Pensabene, P., Pratt, S., Reece, R. and D. Reese. 1985. The Schola Praeconum II. Papers of the British School at Rome 53, 163-210.

Pacetti, F. 2004. Celio. Basilica Hilariana: scavi 19871989. In L. Paroli, and L. Venditelli (eds.), Roma.

43

LRCW4

Fig. 1. The location of Astura within the Pontine Region

Fig. 2. The structure (height 1.84m; width 2.00m) in section I.

44

TOL-ATTEMA

Fig. 3. The Astura section.

Type Dressel 1C Dressel 2-4 Haltern 70 Cretoise 2 Dressel 20 Africana IIa con gradino Africana Ib Tripolitana III Kapitän 2 Africana IId Africana IIIb LRA 2 LRA 1 Keay 52 Africana IIIc Spatheion 1 LRA 3 Hammamet 3 Globular amphora Unidentified rims

No. of frs 1 3 1 1 1 2 1 1 1 1 1 3 1 16 2 1 2 1 9 10

Date 140 – 10 BC 75 BC – AD 100 75 BC – AD 200 AD 25 - 225 AD 75 - 275 AD 150-300 AD 150 - 300 AD 200 – 400 AD 200 – 400 AD 250 – 300 AD 300 – 400 AD 325 – 650 AD 350 – 650 AD 350 – 700 AD 375 – 450 AD 375 – 450 AD 375 – 600 AD 475 – 675 AD 575 – 725 -

Figure

5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8/5.9 5.10 5.11 5.12 5.13 5.14/5.15 5.16/5.17

Fig. 4. Overview of identified amphora types at Astura.

45

LRCW4

Fig. 5. Amphorae from Astura (scale 1:4)

46

TOL-ATTEMA

Fig. 6. Coarse and cookwares from Astura (scale 1:3)

47

LRCW4

Fig. 7. Coarse and cookwares from Astura (scale 1:3)

48

TOL-ATTEMA

Fig. 8. Coarse and cookwares from Astura (scale 1:3)

49

LRCW4

Fig. 9. Coarse and cookwares from Astura (scale 1:3)

50

LOCAL PRODUCTION AND TRADE PATTERNS IN THE ENVIRONS OF VESUVIUS: THE POTTERY FROM POLLENA TROCCHIA AND NOLA CATERINA SERENA MARTUCCI1, GIROLAMO F. DE SIMONE2, GAETANA BOEMIO3, GIORGIO TROJSI3 1

Seconda Università degli Studi di Napoli; [email protected] 2

St. John’s College, University of Oxford; [email protected]

3

Università degli Studi Suor Orsola Benincasa Napoli; [email protected]; [email protected]

This article aims to provide a snapshot of local production and trade patterns in the territories of Neapolis and Nola in late Antiquity through the study of published data from Neapolis and pottery assemblages from two sites: a bath complex in Pollena Trocchia and a villa on the outskirts of Nola. Pottery assemblages from both sites are described and compared to the evidence from Neapolis and sites in the Apennines in order to reconstruct trade patterns from the coast to the mountains. Results suggest that the distribution of imported goods is determined more by site ranking and proximity to main routes, rather than by distance. Furthermore, archaeometric analyses are provided for samples selected from burnished ware and cooking ware from Pollena Trocchia and Somma Vesuviana. Results suggest a complex distributional pattern and the existence of several kilns, both around Vesuvius and on the Apennines, which shared the same vase typology. KEYWORDS: NORTH VESUVIUS, NEAPOLIS, NOLA, AD 472 ERUPTION, CAMPANIA, BURNISHED WARE Campania. The data presented in this paper originate from our flagship excavation in Pollena Trocchia (Fig. 1), the villa in via Saccaccio at Nola, and the so-called Villa of Augustus in Somma Vesuviana (thin sections only). Comparative figures from Neapolis’ theatre are drawn from published data (Ciarocchi et al. 2010).

1. Introduction The Bay of Naples is mostly known for its cities on the coast, e.g. Neapolis and Pompeii, while the sites of the hinterland have received less attention. The northern slope of Vesuvius has especially suffered from a lack of systematic field survey, leading to the incorrect assumption that this area was economically less important than the southern slope, up to the point that notwithstanding significant contributions (e.g. Pagano 1995; Soricelli 2001), still today scholars commonly associate the phrase ‘the Vesuvian area’ mostly with the territories of Pompeii and Herculaneum (on this specific problem, see De Simone 2008 with a list of previous references).

2. The Roman Baths in Pollena Trocchia The site in Pollena Trocchia, Masseria De Carolis, was built above the ashes of the AD 79 Pompeian eruption, probably by the end of the 1st century AD. The building was in use for five centuries; in AD 472 two thirds of the site were buried by the late antique eruption of Vesuvius commonly known as the ‘Pollena eruption’. Short afterwards people returned to the site and utilised the upper part of the building in a variety of ways, until another eruption buried the structures completely (AD 505/512). The site (Fig. 3) consists of a bath complex, which is likely part of a Roman villa. All contexts under the AD 472 debris belong to the last phase (c. AD 450472). During this period, pavements and walls were widely spoliated. In one corner of the praefurnium the burial of a four year old boy was discovered, and in the courtyard two newborns were buried in enkhytrismoi, one in an LRA 1, the other in a Keay LII. Above the ashes of the Pollena eruption, there were traces of a new settlement: a small oven or fireplace was built, the upper level of the buildings appears to have been inhabited, and a cistern of the baths was emptied from the ashes of the previous eruption and re-plastered, until the later eruption occurred. [GFDS]

The remarkable discoveries made in our partner excavation in Somma Vesuviana (De Simone 2009; Aoyagi et al. 2010; 2012; De Simone et al. forth.) encouraged us to commence a new project, the Apolline Project, the main aim of which is to investigate the territories of Neapolis and Nola through field survey, excavation of new sites, and the publication of data from legacy excavations. The northern slope has great potential to illuminate the history of Roman Campania from the 2nd to the 5th centuries AD, because its inhabitants recovered from the AD 79 eruption sooner, and more notably, than those on the other side of the volcano. Life continued until AD 472, when another eruption struck and sites on the slopes were buried by 3-5m of volcanoclastic debris. Through the study of pottery from new excavations and previously dug sites, the Apolline Project aims to use this sharp terminus ante quem to provide a clear snapshot of the late antique economy in 51

LRCW4

variant of this casserole had a plain rim and a handle on the rim decorated with digital impressions (Fig. 5, n. 2). Lids were attested in many shapes: with plain rim (more or less rounded), conical body, and knob handle. The position in which potsherds were discovered does not allow the matching of lids with vessels. Cooking pots mostly had ovoid bodies, plain rims, small collars, and flat or disk bases, sometimes with handles (Fig. 5, n. 4; see Mukai et al. 2010b, 228, fig. 5, n. 40).

2.1. Overview of the pottery This contribution deals with the pottery found in the latest phase before the AD 472 eruption of Vesuvius. Each context has been dated independently according to the latest identifiable sherd in it; nevertheless frequently in each of them there was also earlier pottery. In consideration of how the artefacts were sorted in the stratigraphic units, the widespread presence of minute charcoal remains, the organic nature of the soil, the high fragmentation of the pottery and the occurrence of matching sherds from different contexts (sometimes from rooms with no physical connection between each other), we concluded that the pottery was in secondary deposition. The contexts belonging to this phase have been interpreted as the result of dumping activities, which were likely carried out by the inhabitants/squatters who still lived in another part of the site or in its vicinity.

Local coarse ware (i.e. pots for food preparation and storage) and cooking ware were attested in equal proportions. In some cases the vessels had a partially or completely polished surface. Coarse ware was produced almost exclusively on the wheel, with more or less fine fabrics, commonly with uniform firing. Food was prepared in a flat-based dish with flaring walls and flattened rims on the outside, Carminiello type 1 (Arthur 1994, 183, fig. 80, n. 1), an imitation of Hayes 61 in ARS (De Simone et al. 2009, tab. 1, n. 8), or with a triangular profile rim thickened on the outside, with a shape similar to Carminiello type 8 (Fig. 5, n. 5; Arthur 1994, 183, fig. 89). Commodities were stored in small-sized amphorae, produced both in plain ware and, more frequently, in painted ware. One kind of storage amphora, in coarse ware, had a characteristic shape, which consists of a big globular body, with double moulding rim and a handle on the rim and the shoulder. Its fabric is characterised by firing in a reducing atmosphere, while the surface, with oxidized firing, shows a light orange slip (Mus. 10R 6/6; Fig. 5, n. 6). Both the shape and the kind of firing can be compared with some specimens from Benevento (Lupia 1998, 181-182, fig. 102).

The majority of the ceramics (total number of sherds: 19,034; NMI 4245, from which the percentages are calculated) were produced in Campania (68.71%), and for some of these (about one fourth), a more local origin may be hypothesised, i.e. Neapolis and the environs of Vesuvius. Imported vessels mostly came from Africa (17.22%), the rest from Pantelleria (0.98%), Spain (0.11%), the Levant (0.07%), Calabria (0.02%), unidentified (12.89%). 2.2. Amphorae The study of the amphorae is still in progress therefore the following is only a broad outline. Most of the latest amphorae come from Tunisia (Keay XXXV A, Keay LV A) and Tripolitania (Keay X); among the scant imports from the Levant, noteworthy are the Egyptian LRA7, the LRA2 from Chios, and the LRA1 from Syria or Cyprus.

In summary, the coarse and cooking wares in Pollena Trocchia were of quite good quality, mostly produced locally, with a varied range of shapes and good fabrics. Section 5 will describe some micro-regional productions with specific characteristics in technique used and typology. [CSM]

2.3. Wares for food processing, cooking, and storage Cooking ware made up 49.7% of the total NMI and it almost exclusively consisted of local products, while the imports were predominantly residual in each context. Local imitations included the Pantellerian ware casseroles, attested as being both hand-made and produced on the wheel. Pantellerian imports mostly consisted of shallow casseroles with rounded rims, more or less flaring walls, and squared or ear-shaped handles (Fig. 5, n. 1, Ciarocchi et al. 2010, 130, fig. 67, n. 1). Local typology includes imitations of casseroles with rounded walls and thickened inward rims, with earshaped handles (Fig. 5, n. 3, Mukai et al. 2010a, 472, fig. 6, n. 31).

2.4. Tableware Tableware encompasses here ARS (10.8%), painted ware (4.88%), and burnished ware (1.53%). The ARS was found with bowls and plates, mostly in D production. The flat-based dishes are present with broad flat rims (Hayes 58 and 59), or with a triangular profile (Hayes 61); the large dishes have plain rims with straight walls (Hayes 50A); the bowls have rolled rims and broad flanges on the exterior below the rim, or hooked at the edge (Hayes 91). Of particular interest is the bowl from D2 production with a broad horizontal rim, similar to Hayes 78, but differing from it in that the base was decorated with feather rouletting (Fig. 6, n. 10).

The cooking ware was, on average, of good quality, with standardized shapes; it was wheel-made by professional potters, using well-compacted fabrics with large amounts of inclusions. The most commonly attested shape was the casserole with convex walls, a thickened and inward folded rim, and an ear-shaped handle on the rim or on the wall (Carminiello 2-3 type, Carsana 1994, 224-228). The rim is thickened, with triangular or rounded profile. A

The painted ware encompasses among the open vessels mostly bowls or basins, while among the closed shapes noteworthy are the jugs or small amphorae. The most representative types are the bowls/basins Carminiello 62 (Fig. 6, n. 8), 67 (Fig. 6, n. 9), and 69 (Fig. 6, n. 7); they have a diameter of 30-50cm and were used for liquid 52

MARTUCCI-DE SIMONE-BOEMIO-TROJSI

Most of the pars urbana of the villa (35 rooms) was excavated, although the pottery which has been preserved comes from only 11 of these rooms. It is unclear whether the artefacts from the other rooms were lost in the past while being transported between several depots, or simply were never collected. There were 276 NMI in total, of which 107 are Late Antique, and these are reported here. For this phase, tableware is attested in a higher concentration than cooking ware and amphorae. ARS is scarcely represented (11%), and mostly consists of Lamboglia Ib bowls and fragments in D production, among which were the bowl Hayes 91A and the flatbased dish Hayes 59 type. Painted ware is the most frequent class (29%; for a discussion of the class, see Sugiyama et al. 2010) and it mostly consisted of jugs and small amphorae with moulded rims and ovoid bodies, which are also attested in Pollena Trocchia (Martucci et al. 2012, 115, tav. 3, n. 17), Carminiello ai Mannesi (Arthur 1994, 204, fig. 95, n. 94.8), and Ordona (Turchiano 2000, 353, n. 6.1). Slipped ware included a few bowls, with diameters of 22-30cm, found in various shapes: with rounded rim with flange; squared everted rim (Fig. 7, n. 14); slightly hooked rim; and rounded rim bearing grooves (Fig. 7, n. 15). There were few burnished ware sherds and these pertained to jugs (Fig. 7, n. 16). Cooking ware (15%) included cooking pots and casseroles (Carminiello 2.5, 12.6, 12.7, 17.1), with external surfaces blackened by fire, and sometimes polished. Of particular interest is a dish which has a rounded rim, and was probably produced in the environs of Vesuvius. In fact quite a few sherds of this type have been found both in Pollena Trocchia and in Somma Vesuviana. The fully preserved items of this type (Fig. 7, n. 20, from Somma Vesuviana) show a central hole and are generally interpreted as bread cooking dishes. Amphorae (16%), were present exclusively as African types. Noteworthy among these are the Keay XXV, the ‘Africana grande’ and one ‘Africana piccola’ with CAVS.T stamp (Ostia I, 11, tab. XLVII, n. 603). The general assemblage of pottery shapes at villa Saccaccio does not differ substantially from that published for the sites in Naples, Pratola Serra, Pollena Trocchia, and Somma Vesuviana. An interesting similarity may be noted between the typologies of jugs and small amphorae in painted ware attested in Campania and at Ordona. The close relationship between Late Antique Apulia and Campania is also confirmed by the considerable presence of Campanian cooking ware with polished surfaces in Apulian contexts (Turchiano 2000, 383-384). The hypothesis of a strong commercial link between the two regions is also corroborated by literary sources, and in particular by Paulinus of Nola (AD 354-431), who mentions the migration of Campanian farmers to Apulian fields where workers were needed, as well as a flow of pilgrims from Beneventum and Apulia to the tomb of St. Felix in Cimitile in the cemetery of Nola (Carmen 20.312-7). [GB]

food. Among the closed shapes, noteworthy are the jugs or small amphorae in painted ware, especially the Carminiello type 94 with moulded rim and ovoid body (Fig. 6, n. 12), which is also attested in Somma Vesuviana (Mukai et al. 2010b, 232, fig. 9, n. 64), Carminiello ai Mannesi (Arthur 1994, 203, fig. 94; Mukai et al. 2010a, 472, fig. 6, n. 29). Another type (Fig. 6, n. 11), with plain rounded rim, is also attested in Ordona (Leone 2000, 407, tab. 6, n. 23.1). They were wheelmade, with uniformly thick walls, quite fine fabrics, and they were well-fired. Surface colours range from rosy (Mus. 5YR 7/4, 7.5YR 8/4) to beige-orange (Mus. 5YR 7/6, 6/8; 7.5YR6/8). The colour on painted ware jugs is applied with a cloth (a straccio) or a brush with thin layers. The colour was irregularly applied on the surface, with drippings either on the internal surface and the rim or on the external surface. In Pollena Trocchia, burnished ware is mostly represented by the jug, which was found with various rims (Fig. 6, n. 13). The exterior surface clearly shows polishing of the slip with a tool to create an external impermeable film. Paul Arthur identified one of the productions in burnished ware, named it “fabric A”, and dated it to the mid-5th century AD (Arthur 1994; id. 1998). Among the sherds in burnished ware found in Pollena Trocchia, the macroscopic analysis reveals the existence of at least two productions, thus we carried out archaeometric analyses in order to characterise them (see §6). In summary, while in the 4th century AD the local pottery still provided substitutes for, or was integrated with, imported fine ware, commonly through imitation of forms, from the mid-5th century AD the repertoire shows more original shapes. A similar situation has been reported in inner Apulia, especially at Ordona (Annese 2000, 339-340), where pots in plain ware and painted ware are attested in higher percentages than imported ARS. The predominance of local pottery can be explained either as poor penetration of imported goods into inner areas, or as self-sufficiency of the regional market (which competes with the African imports but still imitates their most common shapes). Data are currently insufficient to resolve this question without further investigation, which is currently underway. 3. The Roman Villa at Nola, via Saccaccio In the mid-1980s, the remains of a Roman villa (Fig. 4) were discovered in via Saccaccio, on the outskirts of ancient Nola, during some works on the modern sewer. Two short reports were published (Sampaolo 1986; 1991) and an outline of five phases, mostly based on the architecture, was provided. During the first three phases, from the 2nd century BC to the 1st century AD, the width of the passage ways was narrowed, the floor level rose, and the walls were reinforced. During the subsequent two phases, between the 4th and the 5th centuries AD, new walls in rough opus incertum with tufa blocks and abundant mortar were set against the earlier walls or even obliterated them. 53

LRCW4

pitch), suggests that probably wooden barrels were used (see Palladius 1.18). [GFDS]

4. Patterns of regional trade For the study of the distribution of Late Antique pottery in Campania we use here the Complex Systems’ model (Bentley and Maschner 2003), whose key-concepts are open networks, interconnectivity, and agent-based modelling (as opposed to positivistic models such as costefficiency and city-countryside dependency). Within this framework, we hypothesise that the distributional patterns of ARS in the hinterland might not necessarily be dependent on distance and might differ from trade patterns of other pottery classes (e.g. burnished ware) in the same area (a similar interest in the pottery distribution from the coast to the hinterland is in Martin 2005). In the discussion below we compare the percentages of pottery classes from Neapolis’ theatre (Ciarocchi et al. 2010), with the villas in Pollena Trocchia and Nola Saccaccio, which share a similar use as dumping areas in the 5th century. Furthermore, these are the only sites in the region for which the quantitative data are available. At Nola Saccaccio, the low NMI (107) testifies to the fact that the site was almost completely abandoned, while at the theatre in Naples (total number of artefacts – including glass, metals, etc. – circa 25,000, total NMI for pottery 3123) and in Pollena Trocchia (total number of artefacts 27,068, total number of pottery sherds 19,034, total NMI for pottery 4245) the activities were still intensive.

5. Markers of local workshops The previous section emphasised the importance of local productions among the pottery circulating in the countryside and hinted at the existence of micro-regional products. This section focuses on two shapes, which are not attested to Neapolis, but rather seem to be typical of the environs of Vesuvius. The first is a ‘small amphora’ in burnished ware (Italian: steccata, French: céramique commune polie). The term ‘small amphora’ was given in previous scholarship in consideration of the two handles: we have kept the original name for this vessel, although it is probably a jug. The small amphora has two handles and ovoid/globular shape, foot ring, and flaring grooved neck. The specimens taken here into consideration come from: Sant’Anastasia Via Rosanea (Fig. 7, n. 18, Parma, Gifuni 1988, 158-159, tav. 2), from a funerary context sealed by the AD 472 eruption and dated by Parma to the second half of the 5th century AD; Pompei località Cimitero (De Carolis 1997, 17-19, fig. 1), from a funerary context and dated by the author to the second half of the 5th century AD; Somma Vesuviana località Starza della Regina (Fig. 7, n. 17, 19, Mukai et al. 2010b, 227, fig. 4, n. 34; fig. 5, n. 36), from contexts of demolition and partial abandonment of the site dated by the authors to the second half of the 5th century AD.

The comparison between percentages of pottery categories (Fig. 2) shows a similar pattern between Naples’ theatre and Nola Saccaccio, where the ratio among the classes is quite the same. In Pollena Trocchia, although the percentage of the ARS is similar to the other sites (10.79%), other imports, like amphorae (3.55%) and lamps (3.41%), are considerably less than in the cities. The predominance of local vessels in Pollena Trocchia is even clearer if we focus on the cooking ware, where on a total NMI of 2109, local products are 89.47%, African cooking ware is 8.53%, Pantellerian ware is 2%. On the contrary, in Neapolis’ theatre, on a total NMI of 687, local products are 57.2%, African cooking ware is 23.7%, Pantellerian ware is 16.7%, Oriental imports are 0.9%, unidentified 12% (Ciarocchi et al. 2010, 124).

Considering that this specific shape is absent in Neapolis, in order to locate possible production areas in the Vesuvian countryside, we sampled the items from Somma Vesuviana for archaeometric analyses (see §6.2, samples B, C. We thank the University of Tokyo team and Drs. T. Mukai and C. Sugiyama). The sampling of burnished ware has been enlarged with a fragment of jug from Somma Vesuviana (sample A: Mukai et al. 2010b, fig. 5, n. 35, in the same context of the previous samples) and with three walls (57.1; 84.29; 267.2) and one rim of jug (82.8, a residual sherd dated to the end of the 4th century, similar to Aoyagi et al. 2007, 440, fig. 4, n. 18) from Pollena Trocchia, from contexts dated to the second half of the 5th century AD (see §6.1). Three different fabrics were identified. The matrix of these fabrics seems coarser than the Neapolitan burnished ware (as described in Arthur 1994, 210) and is characterised by a less glossy slip. Petrographic examination of the samples from Somma Vesuviana and one from Pollena Trocchia (57.1) showed inclusions from Mt. Vesuvius. The other three samples from Pollena Trocchia, on the other hand, consisted of two different fabrics. The first had a very fine matrix with no lithic components (84.29), and therefore could not be characterised. The second had metamorphic inclusions (82.8; 267.2), which might have originated outside Vesuvius, possibly the Apennines.

This evidence seems to suggest that the distribution of imports mostly followed the main routes, thus it reached in a similar fashion the cities, while small settlements in the countryside, like that in Pollena Trocchia, were mostly supplied by local producers. One possible explanation for this evidence is that in the mid-5th century the reduction of goods traded led to the creation of a new distributional pattern, in which the hierarchy of the sites played a more prominent role than their distance from the source. Thus although Pollena Trocchia is only 7km from Neapolis, and Nola 20km, Nola, the much larger city, received the greater share of (presumably) more expensive imported wares, while oil and wine might have been supplied by a local source. The discovery of the remains of a sizeable cella vinaria in the site at Somma Vesuviana (Aoyagi et al. 2010, 212-213), on a higher terrace with cocciopesto floor and low kerbs (associated with many charcoal fragments, resins, and

A second marker of micro-regional production has been identified in a bread-cooking dish in cooking ware, handmodelled, which is characterised by a central hole (Fig. 7, n. 20, from Somma Vesuviana, already published in Aoyagi et al. 2007, 441, fig. 5, n. 30, dated to the end of 54

MARTUCCI-DE SIMONE-BOEMIO-TROJSI

6.1. Pollena Trocchia, Masseria De Carolis

the 4th – early 5th centuries AD). According to Mukai (Aoyagi et al. 2007, 441, fig. 7, n. 51) the type evolved in the second half of the 5th century with higher and more undulated walls. As far as we know, this shape is not attested in Neapolis. Entire vessels of this shape were found only in Somma Vesuviana, but many fragments were discovered also in Pollena Trocchia and in Nola Saccaccio. In order to characterise this production, we took three samples from one context of the second half of the 5th century in Pollena Trocchia (280.5; 280.6; 280.7).

Transverse sections of almost all samples showed a brownish-reddish matrix (Mus. 2.5YR 4/6, 5/6, 5/8), isotropic textures, overall a mid-high porosity (except for the sample 57.1, where it was low), quite rough fabric, characterised by frequent clay pellets, and abundant fragments of vesiculated glass and chamotte. The mineralogical component was characterised by Kfeldspar crystals and plagioclase, biotite lamellae, abundant pyroxenes, common olivine and iron oxides. The lithic component was abundant and included fragments of lava rocks (also with leucite) and vesiculated glass.

The fabric is coarse and porous, with abundant lithics of varying sizes. To the naked eye, the fabrics appeared quite different and so we divided them into three groups. Petrographic analyses subsequently revealed rather that they were very similar, therefore the differences might be related to poorly standardised production practices (possibly on a domestic scale), or maybe attributable to the presence of multiple kilns in the same area. [CSM]

The matrix of the samples 267.2 and 82.8 was different. It was reddish (Mus. 2.5YR 5/4) and the particle size was mid-fine with isotropic texture. Its porosity was low, with sparse clay and chamotte pellets. The mineralogical component was characterised by K-feldspar crystals, subordinate plagioclase and quartz, and sparse lamellae of muscovite and iron oxides. The lithic component was abundant and was represented by metamorphic rocks. In general, XRD analysis confirmed the TS observations: in all samples it revealed the presence of quartz, potassium feldspars (anorthoclase, sanidine), more abundant plagioclases (anorthite, albite), pyroxenes (diopside, augite), and micas (biotite, muscovite). Iron oxides (hematite) were sparse. Calcite, when observed, seemed tied to secondary depositional phenomena.

6. Archaeometric analyses Archaeometric analyses were carried out on seven samples of burnished ware (three from Somma Vesuviana, four from Pollena Trocchia, see §5) and on three samples of bread-cooking dish in hand-modelled cooking ware from Pollena Trocchia. X-Ray Diffraction (XRD with Miniflex Rigaku) and optical microscopy in polarised light (with Nikon Eclipse E400 Pol) on thin sections (TS) were carried out.

Summary table of XRD analyses on Pollena Trocchia samples. Inventory number Quartz Calcite K-Feldspar 57.1 Burnished ware +++ + ++ 82.8 Burnished ware +++ ++ 84.29 +++ + ++ Burnished ware 267.2 +++ ++ Burnished ware 280.5 +++ ++ Bread cooking dish 280.6 +++ ++ Bread cooking dish 280.7 ++ +++ Bread cooking dish Legend: +++ abundant; ++ common; + sparse; + traces.

Plagioclase + + +

Pyroxene + +

Fe oxide + + +

Mica + + +

+

+

+

+

+

+

+

+

+

+

+

++

+

+

+

+

sparse. Among the lithics, there were abundant fragments of basic volcanic rocks.

6.2. Somma Vesuviana, Starza della Regina (so-called Villa of Augustus)

Sample A was characterised by an isotropic texture, brownish-reddish matrix (Mus. 2.5YR 4/6), marked porosity and sparse clay pellets. The added component was medium-fine and was characterised by abundant Kfeldspar, plagioclases, quartz, sparse muscovite and abundant iron oxides. XRD results for all samples showed quartz, abundant potassium feldspars (anorthoclase, sanidine) and plagioclases (anorthite), fewer pyroxenes (diopside), sparse micas (muscovite, biotite) and iron oxides (hematite). Calcite was also

Analyses were carried out on three samples of burnished ware. TS examination revealed homogeneity and similarity between the latter two samples (B and C; see §5). They were both characterised by a brownish-blackish matrix (Mus. 2.5YR 4/3, 4/4), isotropic texture, abundant moderately rough to rough inclusions. They had medium porosity, abundant clay and chamotte pellets. The added mineralogical component was quartz-feldspar with pyroxenes, while biotite lamellae and iron oxides were 55

LRCW4

attested in sample A, and as for B and C, was also likely as a result of secondary depositional phenomena

.

Summary table of XRD analyses on Somma Vesuviana samples Inventory number Quartz Calcite K-Feldspar Somma V. US 224 A +++ ++ + Somma V. US 224 B +++ ++ Somma V. US 224 C +++ ++ Legend: +++ abundant; ++ common; + sparse; + traces.

Plagioclase + ++ +

Pyroxene + +

Fe oxide + + +

Mica + + +

medievale. In P. Arthur (ed.), Il complesso archeologico di Carminiello ai Mannesi (scavi 19831984), 221-258. Galatina, Congedo.

6.3. Discussion

Ciarocchi, B. et al. 2010. I materiali dai livelli tardoantichi a quelli moderni. In I. Baldassarre, D. Giampaola, F. Longobardo, A. Lupia, G. Ferulano, R. Einaudi, F. Zeli (eds.), Il teatro di Neapolis: scavo e recupero urbano, 105-150. Napoli, Università Orientale.

The analyses highlighted quite homogeneous and uniform mineralogical and lithic components. The inclusions, the characteristics of the texture, and the minero-petrographic characterisation point to a trachytic-leucitic volcanic area; while the presence of scoria and of vesicular glass fragments narrow the options to the Somma-Vesuvius volcanic complex. Overall the samples were very similar to those analysed in the Late Antique site at Pompeii, via Lepanto (De Carolis et al. 2009), for which a Vesuvian origin has been already hypothesised. Only two samples in burnished ware (82.8 and 267.2) differed from the rest. They showed a demonstrably different lithic profile, which was characterised by metamorphic rocks. This suggests the presence of other centres of production. [GT]

De Carolis, E. 1997. Testimonianze archeologiche in area vesuviana posteriori al 79 d.C. Archeologia, Uomo, Territorio 16, 17-32. De Carolis, E., Soricelli, G., Grifa, C., Langella, A., Morra, V. 2009. Ceramiche tardoantiche da Pompei. In M. Pasqualini (ed.), Les céramiques communes de Campanie et de Narbonnaise, 651-672. Napoli, Centre Jean Bérard. De Simone A. 2009. Ricerche e scavi a Somma Vesuviana. In G. F. De Simone and R. T. MacFarlane (eds.), Apolline Project vol. 1. Studies on Vesuvius’ North Slope and the Bay of Naples, 157-171. NapoliProvo (UT), Università Suor Orsola Benincasa.

Bibliography Annese, C. 2000. Le ceramiche tardoantiche della domus B. In G. Volpe (ed.), Ordona X. Ricerche archeologiche a Herdonia (1993-1998), 285-340. Bari, Edipuglia.

De Simone, G. F. 2008. Il territorio nord-vesuviano e un sito dimenticato di Pollena Trocchia. Cronache Ercolanesi 38, 329-349.

Aoyagi, M., Angelelli, C., Matsuyama, S. 2010. La cd. Villa di Augusto a Somma Vesuviana (NA) alla luce delle più recenti ricerche archeologiche (campagna di scavo 2002-2008). Amoenitas 1, 177-219.

De Simone, G. F. and Macfarlane, R. T. (eds.) 2009. Apolline Project vol. 1. Studies on Vesuvius’ North Slope and the Bay of Naples. Napoli-Provo (UT), Università Suor Orsola Benincasa.

Aoyagi, M., Angelelli, C., Matsuyama, S. 2012. Somma Vesuviana, cd. Villa di Augusto. Aggiornamenti delle indagini 2009-2010. Amoenitas 2, 219-240.

De Simone, A., Aoyagi, M., De Simone, G. F. forthcoming. Dionysos and the late Antique wine produce in the so-called Villa of Augustus on the North Slope of Vesuvius. In G. Métraux and A. Marzano (eds.), Roman Villas in the Mediterranean Basin. Boston-Leiden, Brill.

Arthur, P. 1994. Ceramica comune tardoantica e altomedievale. In P. Arthur (ed.), Il complesso archeologico di Carminiello ai Mannesi (scavi 19831984), 181-220. Galatina, Congedo. Arthur, P. 1998. Local pottery in Naples and northern Campania in the sixth and seventh centuries. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo, Atti del Convegno in onore di John W. Hayes, 491-510. Firenze, All’insegna del giglio.

Leone, D. 2000. Le ceramiche tardoantiche della fattoria di Posta Crusta. In G. Volpe (ed.) Ordona X. Ricerche archeologiche a Herdonia (1993-1998), 387-436. Bari, Edipuglia. Lupia, A. 1998. Testimonianze di epoca altomedievale a Benevento. Napoli, Maetzke.

Bentley, R. A. and Maschner, H. D. G. (eds.), 2003. Complex Systems and Archaeology. Salt Lake City, University of Utah Press. Carsana, V. 1994. Ceramica da cucina tardo antica e alto 56

MARTUCCI-DE SIMONE-BOEMIO-TROJSI

Martin, A. 2005. Variation in Ceramic Assemblages as an Indicator of Openness to Trade. In J. Pollini (ed.), Terra marique, 61-76. Oxford, Oxbow.

Pagano, M. 1995. L’area vesuviana dopo l’eruzione del 79 d.C. Rivista di Studi Pompeiani 7, 35-44. Parma, A., Gifuni, A. 1988. Prime indagini su una necropoli tardo-romana in via Rosanea a Sant’Anastasia. In Primo convegno dei gruppi archeologici dell’Italia Meridionale, 151-173. Prata Sannita, Gruppi Archeologici dell’Italia Meridionale.

Martucci, C. S., Boemio, G., Trojsi, G., De Simone, G. F. 2012. Pollena Trocchia (NA), località Masseria De Carolis. L’analisi dei reperti per la ricostruzione del contesto economico e sociale del sito. Amoenitas 2, 87-117.

Sampaolo, V. 1986. Dati archeologici e fenomeni vulcanici nell’area nolana. Nota preliminare. In C. Albore Livadie (ed.), Tremblements de terre, éruptions volcaniques et vie des hommes dans la Campanie antique, 113-119. Napoli, Centre Jean Bérard.

Mukai, T., Sugiyama, C., Aoyagi, M. 2010a. Une contribution pour la datation des céramiques tardives. Le contexte avec terminus ante quem de 472 apr. J.-C. donnée par l’éruption du Vésuve sur le site romain de Somma Vesuviana, Italie. In S. Menchelli, S. Santoro, M. Pasquinucci, and G. Guiducci (eds.), LRCW3. Late Roman Coarse Ware, Cooking Ware and Amphorae and in the Mediterranean. Archaeology Archaeometry. Comparison between western and eastern Mediterranean. British Archaeological Reports, International Series 2185, 471-478. Oxford, BAR Publishing.

Sampaolo, V. 1991. Località Saccaccio. Villa romana. Il pavimento della fase repubblicana. Bollettino di Archeologia 11-12, 162-165. Soricelli G. 2001. La regione vesuviana tra secondo e sesto secolo d.C., in E. Lo Cascio and A. Storchi Marino (ed.), Modalità insediative e strutture agrarie nell’Italia meridionale di età romana, 455-472. Bari, Edipuglia.

Mukai, T., Sugiyama, C., Watanabe, K., Hirose, I. 2010b. Nota preliminare sui materiali ceramici rinvenuti nel corso delle campagne di scavo 2002-2007 nella “Villa di Augusto” a Somma Vesuviana. Amoenitas 1, 221235.

Sugiyama, C., Mukai, T., Aoyagi, M. 2010. The colored ware of the fourth and fifth centuries at Somma Vesuviana. In LRCW3, 479-485.

Ostia I = Carandini, A., Fabbricotti, E. Gasparri C. (eds.), 1968. Ostia I. Le terme del nuotatore, scavo dell’ambiente IV. Roma, De Luca.

Turchiano, M. 2000. La cisterna e il suo contesto. Materiali tardoantichi dalla domus B. In G. Volpe (ed.), Ordona X. Ricerche archeologiche a Herdonia (1993-1998), 343-387. Bari, Edipuglia.

57

LRCW4

58

MARTUCCI-DE SIMONE-BOEMIO-TROJSI

59

LRCW4

60

MARTUCCI-DE SIMONE-BOEMIO-TROJSI

61

LRCW4

62

DA VIBO VALENTIA A NICOTERA. LA CERAMICA TARDO ROMANA NELLA CALABRIA TIRRENICA FRANCESCO A. CUTERI1, MARIA TERESA IANNELLI2, PAOLA VIVACQUA3, TIZIANA CAFARO3 1

Università Mediterranea di Reggio Calabria; [email protected]

2

Soprintendenza per i Beni archeologici della Calabria; [email protected] 3

Collaboratori Soprintendenza per i Beni archeologici della Calabria

In this study we present new archaeological evidence along with a more detailed analysis of already studied contexts in the area of Vibo V. This area is characterized by several villae as well as by smaller rural settlements and by ancient towns: Vibo Valentia,Tropea and Nicotera. Between the end of the 6th and early 7th c. AD the northern area of Vibo V., including Soriano’s, Buccarelli’s properties and mainly the S. Aloe area, was still inhabited with urban settlement modalities not yet clearly defined. The types of amphorae found, almost exclusively of African production with only a very small amount of oriental types, is consistent with an intense commercial activity between cities in Bruttium and Northern Africa. Large amounts of African pottery are found up to the 7th c. AD in Piscino, which is a Center linked to the ecclesiastic Authority with an important administrative activity. The intense commercial activity might have induced a local production imitating the African one. Indeed in S. Aloe and Buccarelli’s property local pottery has been found resembling the typical African cooking ware as well as African sigillata D of the 5th and 6th c. AD; The finding of pottery kilns and ceramic refuse deposits suggest that in Calabria during the late Roman period the commercial activities were intense, most likely because of the strong links with Africa and that the region was very well coping with the production system of latifundia. KEYWORDS: ITALIA, CALABRIA, VIBO VALENTIA, ANFORE AFRICANE, ANFORE ORIENTALI, KEAY LII INTRODUZIONE

1. VIBO VALENTIA

Il quadro qui prospettato è da intendersi come un approfondimento di quello presentato in LRCW2 e il territorio preso in esame, la provincia di Vibo V., si caratterizza per la presenza di numerose villae, piccoli insediamenti rurali e centri abitati di antica fondazione (Fig. 1) (Iannelli 1989).

1.1 Sintesi storica Recenti studi hanno riconsiderato il processo di romanizzazione nella colonia latina di Vibo Valentia e nell’ager vibonensis, proponendo una continuità tra lo stanziamento della colonia, nel 192 a. C., e la precedente occupazione brettia della fine del IV sec. a. C. Dopo la guerra sociale, il processo di municipalizzazione ha inizio intorno all’anno 89 a. C. e la città diventa municipium autonomo con ordinamento quatuorvirale. Costruito in continuità sui resti della polis greca, il municipium ebbe un grande impulso con la costruzione della via Annia Popilia e il periodo imperiale è caratterizzato dalla costruzione di edifici pubblici quali il teatro, l’odeum, le terme, e di domus patrizie. I mosaici pavimentali e la ceramica d’importazione rimandano al’Africa settentrionale, regione con la quale sembrano essere intercorsi intensi scambi ben oltre il V sec. d.C. Nel VIVII secolo è stata riscontrata una generale ruralizzazione della città, con la frequentazione di precise aree come ad esempio quella di Piscino di Piscopio dove le importazioni di ceramica, soprattutto spatheia nord africani, sono state messe in relazione con la committenza di una istituzione ecclesiastica (Arthur, Peduto 1991; Saguì 2002).

Molte novità riguardano Vibo Valentia dove è stata riscontrata, nell’area urbana ed in quella portuale di Bivona, un’alta percentuale di vasellame da mensa e di anfore d’importazione africana. In città, il cantiere più rappresentativo è Sant’Aloe, ma materiali significativi provengono anche dai cantieri Soriano, Buccarelli, Miceli e da Piscino di Piscopio. Quanto recuperato nelle diverse aree, per via del diverso stadio degli studi, non è presentato con criteri di omogeneità. Importanti dati sulla circolazione di merci sono stati ottenuti anche in riferimento al territorio costiero. Qui, il repertorio delle anfore si presenta assai articolato e le diverse attestazioni rivelano traffici commerciali ad ampio raggio soprattutto con l’Africa settentrionale ma anche con la Penisola Iberica e l’Oriente. Ben attestate sono anche le produzioni locali. Qualche dato in più si presenta, inoltre, su Nicotera e Tropea. In tutti i contesti esaminati è stata rivolta maggiore attenzione alle produzioni anforiche, anche se non mancano correlazioni con le altre classi ceramiche; quanto presentato, nonostante l’evidente parzialità, ci sembra di notevole interesse per una più precisa definizione dell’intero quadro regionale.

Fin dalla deduzione della colonia appare fondamentale il ruolo del porto, già attivo in età greca, e lo studio dei materiali ceramici di Bivona conferma gli intensi traffici commerciali attivi fino alla tarda antichità in via prioritaria con l’Africa, ma anche con l’Oriente e con il resto del Mediterraneo. 63

LRCW4

1998; Ostia I; Sangineto 2006; Carsana 1994; Small, Buck 1986; Olcese 1993 e 2003; Semeraro 1992). Tra le forme individuate si registrano: 47 esemplari di olle globulari; 41 esemplari di pentole/casseruole; 30 esemplari di tegami; 61 esemplari di piatti/coperchi; 7 esemplari di coppe; 20 esemplari di vasi a listelli; 15 esemplari di brocche; 8 esemplari di bicchieri; 3 esemplari di grandi contenitori e uno di mortarium.

1.2 I rinvenimenti in area urbana Gli interventi di scavo archeologico si sono limitati all’indagine nei cantieri avviati per le moderne esigenze edili, anche se non mancano esempi, come a Sant’Aloe, dove è venuto alla luce un importante quartiere con ricche strutture residenziali e un edificio termale a carattere pubblico (II-III sec. d.C.), in cui le indagini sono state più estese e ripetute nel tempo.

Tra le forme in ceramica africana da cucina sono state riconosciuti piatti/coperchio (tipo Ostia II, fig. 302, Ostia I, fig. 261, Ostia IV, fig. 60 e Ostia III, fig. 332), casseruole e scodelle (Ostia I, fig. 15 e Lamboglia 9A) che permettono di definire un notevole afflusso di ceramiche comuni utilizzate in cucina e prodotte in Africa fra III e IV sec. d.C.

A S. Aloe i reperti analizzati rimandano alle seguenti classi: anfore, ceramica comune, terra sigillata africana e ceramica africana da cucina. Lo studio delle anfore, sebbene riferito ad un campione modesto, ha permesso di approfondire la conoscenza in rapporto agli scambi che hanno interessato la città tra IV a.C. e VII d.C., confermandone il pieno inserimento nei traffici commerciali che riguardano soprattutto la penisola italica ed in particolare il versante tirrenico.

Per quel che concerne la ceramica sigillata africana, sono state riconosciute coppe, piatti/coperchio, scodelle e vasi a listello (produzioni A, C, D, E). La massima diffusione della TSA si ha tra la fine del III e gli inizi del IV secolo d.C., momento in cui le officine della Tunisia settentrionale hanno una ripresa e in cui la produzione A viene sostituita dalla D.

Tra le anfore, destinate al trasporto di salsa di pesce, olio e vino, sono presenti le seguenti tipologie: Anfore africane (metà II sec. - metà VII sec. d.C.): Africana I (piccola); Keay III; Keay IV; Africana II (grande); Keay VI; Keay XI; Anfora cilindrica di medie dimensioni; Keay XXIII; Keay XXXIX; Anfora cilindrica del basso impero; Keay XXVIIB; Spatheia; Keay LXII; Keay LXI A; Anfora cilindrica di grandi dimensioni.

Il quadro generale del quartiere di S. Aloe è quello che vede la presenza di manufatti africani: tra il 350 e il 450 d.C. la merce più diffusa è la TSA in C (Piatti Hayes 49 e Ostia I) e in D (Scodelle Lamboglia 54). Fino al V secolo le attestazioni di TSA in Calabria, e nello specifico a Vibo Valentia, sono cospicue, mentre cominciano a diventare meno frequenti nel corso del VI e del VII secolo d.C.

Anfore orientali (IV sec. – VII sec. d.C.): Late Roman 1, di produzione siriana; Late roman 4, di produzione palestinese; Keay LIV, dalla Palestina; Late roman 5/6, palestinese; Keay LIII, dalla Siria.

Per la proprietà Soriano, ubicata nelle immediate vicinanze dei resti del teatro romano, i dati presentati sono frutto di una revisione di quanto emerso con gli interventi di scavo effettuati in più anni (1977, 19861989) sulle emergenze archeologiche distrutte dalle ruspe. Con tali interventi si è riusciti a definire meglio la cronologia delle strutture romane prima rinvenute e la tipologia dell’insediamento (Sangineto 1989). Le strutture portate alla luce appartengono ad un unico complesso, probabilmente una lussuosa domus privata databile a dopo il I sec. d. C., con prosecuzione di vita probabilmente fino al IV-V sec. d.C.

Anfore di produzione locale (IV sec. – inizi VII sec. d.C.): Keay LII, prodotte in Calabria e Sicilia. L’analisi del materiale anforico ha dunque permesso di riconoscere 37 tipi diversi di anfora e di evidenziare che vi sono alcuni contenitori, prodotti con un impasto “locale”, che replicano quelli africani “piccoli” e “grandi”. Inoltre, molte delle Keay LII individuate presentano un impasto di colore giallo-grigiastro che potrebbe essere caratteristico dell’ager Vibonensis, tenendo conto che identifica una notevole produzione locale di anfore greco-italiche, Dressel 1 e 2-4.

Le anfore sono riferibili a produzioni africane ed iberiche. Quelle più antiche sono documentate da alcuni orli e anse di Dressel 1 A, Dressel 1B e Dressel 2-4; alla prima età imperiale sono da riferirsi due orli ed un’ansa di Beltràn IIB ed una Dressel 20 nella versione di III d. C. Le produzioni italiche sono rappresentate da due anse con sezione “a fiorellino” che presentano un impasto molto poroso di colore rosato nocciola con inclusi bianchi, neri, grigi e brillanti con superficie esterna ingobbiata di colore giallastro (Fig. 2, n. 4). Queste anse sono attestate anche nel territorio a Francica S. Petru, con le medesime caratteristiche morfologiche e d’impasto. Esse sono attribuite ai tipi Ostia IV, fig. 166 e sembrano relative a produzioni proprie dell’area nord-siciliana diffuse tra il I e V d. C. ma prodotte probabilmente anche in Calabria. In effetti, gli esemplari attestati nel nostro territorio sono risultati di origine locale (Bernardi 2010-2011). Alle

La percentuale delle anfore rinvenute e le indicazioni sulla cronologia che ne derivano consentono di affermare che l’area pubblica di S. Aloe subì una netta flessione nel corso del III secolo d.C., analogamente a quanto riscontrato in altri centri della regione. Una sensibile ripresa si registra per tutto il V secolo, con una continuità di frequentazione che, attraversando il VI secolo, raggiunge le fasi iniziali del successivo, come mostrano la presenza di anfore vinarie Keay LII e le importazioni di anfore africane. Lo studio preliminare della ceramica da cucina permette di riconoscere materiali che trovano confronto con tipi in uso tra III e VII secolo d.C. a Cartagine, Roma, Ostia, Albintimilium, Carminiello ai Mannesi, San Giovanni di Ruoti, Otranto e Tortora (Fulford, Peacock 1984; Ricci 64

CUTERI-IANNELLI-VIVACQUA-CAFARO

con collo cilindrico e orlo svasato; un tipo di bacinella o catino con orlo svasato, un tipo di contenitore con fondo costituito da base cilindrica rastremata; tutte tipologie documentate con leggere varianti ad Ostia in stratigrafie relative ad un arco cronologico che va dalla seconda metà del I d.C. al IV sec.d.C.

produzioni africane di età repubblicana e prima età imperiale sono da riferirsi un orlo di Mañà C2 ed uno assimilabile alla Tripolitana I. Alla fine del II d. C. sono presenti le anfore tipo Africana II A con due orli (Fig. 2, n. 2); un’anfora simile al tipo Africano I B (Fig. 2, n. 1) ed una Keay XXVIC= spatheion 1B (Fig. 2, n. 3) secondo la tipologia di Bonifay. I rinvenimenti delle anfore collocate in un ambito cronologico di fine II a.C. V d.C. mettono in evidenza la capacità dell’abitato urbano di Vibo Valentia di recepire, in maniera costante e per un lungo periodo, merci provenienti in particolare dall’Africa e dalla Spagna. Questo rapporto non sembra interrompersi neanche nei secoli successivi, come dimostrano i dati di scavo dalla città e dal territorio.

Nel cantiere Buccarelli, dove furono documentate più fasi costruttive databili tra II sec. a.C e II sec. d.C. ed il successivo, anche se temporaneo, abbandono dell’area, la maggior parte dei materiali ceramici proviene dallo sterro eseguito per la costruzione di un nuovo edificio (Sangineto 1989). I primi reperti qui presentati sono inquadrabili in un arco cronologico e produttivo ben definito e documentano una frequentazione che va ben oltre il II sec. d.C.

Il rapporto privilegiato con l’Africa è ben attestato anche dalla sigillata africana che nel cantiere Soriano arriva abbondante con una diversità di forme soprattutto nel corso del IV-VI d. C. La produzione C è attestata dalla scodella Hayes 53B (IV - prima metà del V d. C.) prodotta nell’atelier di Sidi Zahruni (Bonifay 2004, 197). La forma maggiormente attestata è la Hayes 67, datata tra la metà del IV-V d. C., proveniente probabilmente dalle officine di El Mahrine, come parrebbero mostrare alcuni fondi decorati con motivi geometrici. Ben attestata è anche la Hayes 76 (seconda metà del V d. C.), ed assai significativa è la presenza della Hayes 99, scodella con orlo a mandorla di cui di recente si è proposta un affinamento tipologico in base al materiale della Crypta Balbi e del Castrum Perti. Questa forma è attestata a Piscino di Piscopio con numerosi esemplari e varianti e la sua datazione è collocata tra la metà del V-VII d. C. E’ attestata, infine, la forma Hayes 105.

Infatti, gli esemplari di Sigillata Africana campionati coprono un arco di tempo che va dal III alla seconda metà del VI, con una maggiore concentrazione per i secoli IVV d.C.: forma Hayes 61; Atlante pp. 83-84, tav. XXXIV; Atlante p. 86, tav. XXXV, 4; Lamboglia IB= Hayes 80; Atlante p. 26 tav. XIV, 4-5; forma Lamboglia 38; Atlante tav. XVII, 6; Atlante p. 94, tav. n. XLIII, 3. L’africana da cucina (metà II- V sec. d.C.) è schematicamente rappresentata dalle seguenti forme: tegame Ostia III; Atlante p. 218, tav. CVII, 5; forma Hayes 23 A= Lamboglia 10 B; Atlante p. 217, Tav. CVI, 12; tegame Lamboglia 10 A; Atlante p. 217, tav. CVI, 10; piatto coperchio; Atlante p. 208, Tav. CIV-CV; piatto coperchio; forma Ostia IV, Atlante Tav. CV, 5. La ceramica comune è rappresentata con diverse forme di olle, tegami, coperchi e casseruole che si diffondono a partire dal IV-V subendo una certa standardizzazione che per alcuni tipi si protrae, come dimostrano sia i confronti regionali che quelli extra-regionali, fino alle soglie dell’VIII secolo. Tra le olle (Fig. 4), presenti anche con esemplari ansati, si registrano molti tipi con orlo verticale che potrebbero essere sia di produzione egea che locale. La forma, già attestata nel IV, è frequente sia a Roma che a Napoli in contesti che vanno dal VI (Carminiello ai Mannesi) al VII-VIII secolo (Crypta Balbi) (Carsana 1994; Olcese 2003). Anche i tegami e le casseruole sono attestati con tipi ampiamente diffusi che si datano tra IV e VII secolo. Segnaliamo, infine, che tra la ceramica comune è presente un frammento di brocca o olpe con decorazione ad ingobbio rosso. Questa classe è molto comune nell'area campano laziale, dove sono presenti più centri di produzione in contesti che vanno dal V al VII secolo. Per quel che concerne le anfore si rimanda alle considerazioni già formulate (Sangineto 1989, 841-842).

L’africana da cucina è ben rappresentata anche se lo strumentario per la cottura dei cibi era costituito soprattutto dai recipienti di produzione locale quali le olle. Sono presenti i piatti coperchio, tipo Ostia I, fig. 261, Ostia III, fig. 332, Ostia IV, fig. 60, e i tegami Lamboglia 10 e 9 A. Tra la ceramica locale di uso domestico, le olle con gradino incavo sulla parte interna dell’orlo per l’alloggiamento del coperchio sono tra le più documentate e con diverse varianti (Fig. 3, nn. 1-5). Sono datate tra II e VI d. C. e il modello morfologico è forse da ricercarsi nelle produzioni da cucina tipiche dell’area egea che vengono imitate anche ad occidente. Tuttavia, non è da escludere una loro origine peninsulare, vista la particolare funzionalità e la notevole diffusione del tipo nella lunga durata. A Vibo Valentia, considerato il rinvenimento in città di diverse fornaci, è possibile ipotizzare una produzione di ceramica di uso comune ad imitazione delle forme più in voga. Le stesse produzioni di ceramica da fuoco sono presenti in area urbana in via Protettì (Fig. 3, nn. 7-11), dove è stato rinvenuta una monumentale struttura in pietra e laterizi identificabile con l’odeum della prima età imperiale. Qui sono stati recuperati anche pochi frammenti di ceramica steccata a stralucido.

In riferimento alla proprietà Miceli, posta in un’area centrale della città di Vibo Valentia, la ceramica presentata, in questa fase, solo le anfore, proviene prevalentemente dallo strato più superficiale, composto da materiale molto eterogeneo (scavo 2009-2010). Tra le produzioni locali, la tipologia maggiormente attestata è quella delle Keay LII. All’interno di questo gruppo si registrano caratteristiche diverse che permettono di individuare due distinti gruppi morfologici:

La ceramica comune da mensa, particolarmente diffusa, è rappresentata da circa 1.300 frammenti. Si è notata la presenza di alcune forme: un tipo di brocca monoansata 65

LRCW4

Poco attestate le anfore iberiche: un orlo di Beltràn II A di I-II d. C., ed una Dressel 7-11 di I d. C.; non sembrano essere attestate le produzioni di III-V d. C.

il primo (Fig. 5, n. 1), riferibile alla prima fase di produzione (IV-V d. C.), presenta il caratteristico profilo dell’orlo a sezione triangolare, con listello poco pronunciato e ingrossato, che trova confronti con esemplari rinvenuti sia in area urbana (quartiere S. Aloe) e sia nel territorio (Cessaniti Mantineo) e con esemplari siciliani di Naxos; il secondo (Fig. 5, n. 2), databile al VIVII sec. d.C., è caratterizzato dal semplice profilo triangolare poco ingrossato e rimanda ad esemplari ritrovati a Scolacium e lungo la costa adriatica. Gli impasti corrispondono a quelli ritenuti caratteristici delle produzioni vibonesi: colore giallino-rosato e pochi inclusi di colore bianco, nero, grigio e brillante. E’ attestata, inoltre, un’ansa con la caratteristica sezione a fiorellino riferibile alle anforette tipo Ostia I, fig. 455-456 e Ostia IV, fig. 166. Le analisi archeometriche (Bernardi 20102011), condotte su alcuni esemplari del territorio, hanno rivelato un’origine locale della produzione e tale attribuzione può essere forse estesa anche agli esemplari rinvenuti in città dal momento che gli impasti trovano corrispondenza con quelli che caratterizzano le Keay LII. Questo dato si rivela di grande interesse lasciando intravedere una produzione calabrese, oltre che siciliana, di questi contenitori.

Sono state rinvenute una parete di LRA 1 ed una di LRA 2 che indicano un minoritario approvvigionamento anche dalla pars orientalis dell’Impero tra VI e VII sec. d. C. Questo trend si registra anche nei siti del territorio di Vibo Valentia ed in generale lungo tutta la costa tirrenica calabrese, dove le produzioni orientali, per i secoli VIVII, sono attestate con indici molto bassi; sullo Jonio, invece, le evidenze archeologiche rivelano una maggiore attestazione di anfore orientali che risultano comunque numericamente minoritarie rispetto a quelle africane (Cuteri, Salamida 2010). Nel complesso, il cantiere Miceli rappresenta un campione significato all’interno dell’abitato di Valentia. Le produzioni anforiche rivelano un panorama assai variegato e restituiscono uno spaccato importante della città in epoca tardo antica. La presenza, infatti, di produzioni africane di V-VI d. C. rivela continui traffici commerciali con l’Africa, ed un’esportazione del vino all’interno delle Keay LII, prodotte almeno fino al VI – VII d. C.

Le importazioni sono caratterizzate da: Tripolitana I, Africana II D, Keay XXV= Africana III A e B; Keay XXVI e Keay VIII B. La Tripolitana I, sebbene rappresentata da un solo esemplare (Fig. 5, n. 3), mette in evidenza come la città usufruisse durante il I-II d. C., sebbene in percentuali non definibili, dell’olio prodotto in questa provincia dell’Impero. Tuttavia, il monopolio del commercio di questa merce, sia nella città che nel territorio di Vibo Valentia, era appannaggio della provincia spagnola della Betica, vista la discreta quantità di anfore Dressel 20 attestata in area urbana. In ogni caso, l’importazione di olio risulta più marginale rispetto ad altri prodotti, come le salse di pesce, probabilmente perché poteva essere prodotto in loco.

1.3 L’area portuale di Bivona In maniera sintetica, integrando il quadro presentato in LRCW2, si propongono alcune annotazioni. Per quanto riguarda la ceramica africana e quella comune ricordiamo che gli esemplari presi in considerazione coprono prevalentemente un arco cronologico che va dal IV al VI sec. d.C. Tra le sigillate si segnalano alcuni frammenti appartenenti ad indicatori tipici della tarda antichità. Si tratta delle diffusissime Hayes 61b, del vaso a listello Hayes 91, del piatto di grandi dimensioni Hayes 87b e Hayes 104, della coppa Hayes 94 e della ciotola Hayes 99. Tutte queste forme appartengono alla produzione D. Per quanto riguarda la cronologia la Hayes 61b e la H94 potrebbero essere riferite al quinto secolo mentre le altre forme sono più tarde e coprono un arco cronologico che si spinge fino al 600-620 d.C. (limite rappresentato dalle Hayes 99 e Hayes 104, a cui aggiungiamo un frammento di Hayes 107 del 600-650). Un frammento di fondo in sigillata D2 presenta uno stampo frammentato riconducibile ad una croce con bracci decorati da punzonature circolari e uno stampo analogo è presente su un grosso piatto da portata (forma Hayes 104 o 105), inquadrabile approssimativamente tra il 525 e il 650 d.C., esposto al museo di Scolacium. Gli altri stampi circolari riscontrati sono molto frequenti e tipici della produzione di sigillata D. Per quanto riguarda l’africana da cucina, attestata con numerosi frammenti, è da segnalare la presenza di piatti coperchio sia con orlo indistinto che appena ingrossato.

Le produzioni medio imperiali di metà III- inizio IV d. C. sono rappresentate da due esemplari di Africana II D (Fig. 5, n. 4) le cui caratteristiche morfologiche sono accostabili al tipo Ostia IV, fig. 131 che Bonifay considera una forma di transizione tra l’Africana II D e la Keay XXV. 1 (Bonifay 2004, 115). Di queste anfore è incerto il contenuto, anche se è probabile che contenessero olio. Dopo le Keay LII, le forme maggiormente attestate sono le Keay XXV che appartengono, secondo la tipologia proposta da Bonifay, alla variante III B datata al IV d. C. (Fig. 5, n. 5). Sono attestati 2 orli di Keay XXVI=spatheion tipo 1, datato al V d. C. (Fig. 5, n. 6), contenitore per il quale è incerta la merce trasportata, anche se la presenza di pece all’interno di alcuni esemplari porta ad escludere l’olio lasciando supporre un contenuto ricavato dai prodotti agricoli o dalla lavorazione del pesce, o anche il vino (Bonifay 2004, 129). Significativa è la presenza di un orlo di Keay VIII B originaria del sud della Bizacena, adibita probabilmente al trasporto dell’olio, datata tra la metà del V e la prima metà del VI d. C. (Fig. 5, n. 7)

Tra la ceramica comune si evidenzia la presenza di due frammenti pertinenti a bacini, uno con orlo ingrossato a sezione triangolare decorato con solcature graffite a crudo, l’altro con orlo a tesa dipinta con fascia in bruno, databili tra V e VIII secolo. Esemplari analoghi sono 66

CUTERI-IANNELLI-VIVACQUA-CAFARO

Mediterraneo occidentale e orientale. Una distribuzione così diffusa ha indotto a prospettare la coesistenza di possibili produzioni locali (Spagna meridionale, Italia meridionale, area pontico-danubiana), che sono però ancora da dimostrare; ciò in aggiunta alla produzione africana che sembra esordire a Cartagine dopo il 535 e che sarebbe confermata dalle analisi delle argille condotte alla Crypta Balbi ed al Castrum Perti (Capelli 1998; Murialdo 2001) ed avvalorata dalla caratteristica rifinitura a spatolature verticali, peculiare della tradizione africana. Per questi contenitori si ipotizza anche una produzione nella pars orientalis dell’impero (Bonifay 2005).

molto comuni in tutta l'Italia meridionale e in special modo nell'area campano- laziale (Benevento, Capua, Alife, Roma). Per quanto riguarda la ceramica da fuoco, la presenza di un altro frammento di testo, in aggiunta a quello segnalato in LRCW2 (Cuteri et al. 2007, fig. 5, n. 3), conferma la pratica della panificazione domestica. Infatti, com’è noto, i testi sono bassi recipienti di diametro variabile, prodotti in ceramica acroma grezza refrattaria, nel nostro caso al tornio lento, impiegati prevalentemente per la cottura di focacce. In contesti romani e tardo romani sono attestati a Luni e Ansedonia. Per quanto riguarda le anfore ricordiamo la presenza di prodotti giunti prevalentemente dall’area nord africana: anfore Africana I, Africana II A, Africana II C, contenitori cilindrici di media taglia tipo Keay XXV (varianti B, C e G), diffusi dalla fine del III alla metà del V sec. Degli spatheia è attestata anche la tardissima versione miniaturistica che troviamo ugualmente a Piscino di Piscopio. Troviamo ancora i tipi Keay XXVII B, XXXV B e XXXVI (prima metà del V sec. d.C), l’anfora Keay LIX, di V secolo, e le più tarde Keay LV A, Keay LXII e Keay LXI/C, quest’ultima di fine VI-VII sec. d.C. Le anfore orientali, meno numerose delle precedenti, sono così rappresentate: numerose LRA 1 (versione Kellia 164 di VI-VII secolo); LRA 2 di fine IVVI d.C.; Ánforas tardías tipo G, anfora tipo Agora M 273, caratteristica del V sec. ed, infine, Samos cistern type di seconda metà VI-VII sec. Le scarse produzioni iberiche, destinate al trasporto d olio e salse di pesce, sono documentate da Dressel 20 e l’Almagro 50 / Keay XVI B-C, di III-metà V sec. d.C.; le anfore italiche sono rappresentate dalle Keay LII di fabbricazione regionale.

Il campione di Piscino risulta assai importante e particolare; gli impasti sono di colore rosato o rossoarancione con inclusi bianchi e presentano uno strato di ingobbiatura biancastra sulla superficie esterna. Altri invece presentano un impasto giallastro con piccolissime inclusioni. Le analisi archeometriche eseguite su 5 campioni di impasto hanno escluso una produzione locale di questi contenitori (Fig. 6, nn. 1,2,4,6,8). Quattro di essi sono riferibili alle produzioni africane (nn. 1,2,4,6), mentre un campione sembra di più incerta provenienza e potrebbe essere riferito ad una produzione orientale (n. 8). Nel contesto calabrese, gli spatheia di piccole dimensioni sono stati nella maggior parte dei casi ritenuti di produzione africana o, come a Bivona e Crotone, di incerta produzione; nel caso della necropoli di Scolacium, su nove esemplari solo due sembrano essere riferibili ad una produzione di carattere locale/regionale. Rimane ancora ignoto il loro contenuto. L’elevata incidenza in vari contesti lascia peraltro presupporre che potesse trattarsi di un genere (vino, olive, salsa di pesce, unguenti) anche alla portata di strati sociali non privilegiati e destinato ad un uso diretto dal contenitore nel quale veniva trasportato. In ogni caso, le ipotesi formulate riguardo al contenuto sono varie e si rimanda, al momento, agli studi di maggior dettaglio (Murialdo 2001, 277-278; Reynolds 1995, 53; Carignani 1989, 7880; Bonifay 1998; Bonifay, Pieri 1995). Si vuole però segnalare che anche per la Crypta Balbi, così come per Piscino di Piscopio, si prospetta un uso legato a funzioni liturgiche, senza comunque escludere un contenuto d’olio, in questo caso certamente di alta qualità, o di miele (McCormick 1998) se, come sottolinea la Sagui’, i κεράμια μικρά menzionati dalle fonti possono essere identificati con questi piccoli contenitori (Saguì 2002, 14).

1.4 Piscino di Piscopio: un complesso extra muros Lo scavo nel sito tardo-antico di Piscino di Piscopio ha restituito alcune strutture murarie ed un ambiente mosaicato. Il contesto, interpretato come di natura ecclesiastica o martyrion, per la presenza di uno spazio lasciato vuoto nel tappeto musivo pertinente probabilmente ad una tomba monumentale (Arthur, Peduto 1991), è da altri considerato pertinenza vescovile o di un pretorium trincerato sulle alture del castello. Solo di recente è stata effettuata un’analisi completa della ceramica tardo romana e ciò ha permesso di ricostruire la storia economica del sito grazie all’individuazione sistematica delle diverse produzioni di anfore, della ceramica fine di importazione africana e della ceramica di uso comune, prodotta probabilmente in loco (Vivacqua 2008).

Oltre agli spatheia, a Piscino sono presenti: anfore Keay LXII; una Tripolitana II nella versione tarda, che trova confronti con esemplari del Castrum Perti databile al VIVII d. C.; LRA1 nella versione della variante LRA 1B tarda, attestata anche nei siti della Crypta Balbi e del Castrum Perti; una forma di transizione, assimilabile alla famiglia delle LRA 1 e datata a partire dalla metà del V d. C., che trova un confronto con esemplari di Marsiglia. Le produzioni egee sono attestate con alcune pareti di LRA 2. Un puntale assimilabile alla famiglia delle LRA3 attesta contatti anche con l’area microasiatica. E’ presente un solo orlo riferibile alla famiglia delle LRA 5/6 e

Tra il materiale anforico, i contenitori di piccole dimensioni, meglio noti come spatheia, sono attestati con un numero massimo di 24 esemplari (Fig. 6, nn. 2-13). Essi rappresentano il tipo di contenitore più attestato costituendo il 72% del materiale anforico rinvenuto. Queste piccole anfore cilindriche raggiungono una diffusione capillare, forse per la maggiore trasportabilità che le porta a toccare i siti fortificati prealpini e alpini, i nuclei rurali dell’entroterra ed anche i principali centri del 67

LRCW4

2. IL TERRITORIO

l’impasto, di colore arancione sabbioso con superfici giallo pallido, rimanda alla produzione considerata di Cesarea Marittima. Non sono state rinvenuti frammenti di Keay LII.

2.1 I contesti archeologici Il territorio di Vibo Valentia (Fig. 1), anche per l’età imperiale e tardo antica, si rivela assai ricco e pienamente inserito nei circuiti commerciali mediterranei. I siti presi in esame in questa sede sono 15: 11 sono interpretabili come abitati e 2 come aree di necropoli, mentre i rinvenimenti subacquei di Vibo Marina e Briatico sono pertinenti a probabili relitti e quello di Trainiti corrisponde all’antemurale dell’antico porto. Gli abitati, probabilmente in villa, sono i seguenti: Pizzo S.Giorgio, Briatico Scrugli-lavori RFI, Briatico Case Scrugli, Briatico Buccarelli, Briatico proprietà Satriani, Tropea, S. Onofrio Lanzaro, Triparni, Pannaconi Cava, Macrone, Francica - località S. Petru. Le necropoli, con tombe a cassa, alla cappuccina o ad echytrismon, sono presenti a Vibo Marina S. Venere e Briatico Trainiti. Rimane d’incerto riferimento il materiale rinvenuto a Cessaniti Mantineo dove sono stati recuperati fuori contesto alcuni frammenti di anfore Keay LII.

Accanto alle anfore, anche la ceramica comune e da fuoco si colloca in un ambito cronologico di V-VII d. C. e presenta una tipologia assai variegata (Fig. 7, nn. 1-7). Ben attestati sono i bacini e caratteristici sono alcuni tipi che richiamano esemplari attestati e ben diffusi in Campania nel V-VI d. C. (Fig. 7, nn. 1-3) e presenti anche nella variante dipinta. Non manca la ceramica steccata a stralucido con un bacino dall’orlo pendente con estremità arrotondata (Fig. 7, n. 5) che richiama esemplari prodotti in africana da cucina tipo Ostia I, fig. 261 ed un vaso a listello che richiama la forma in sigillata africana Hayes 91 (Fig. 7, n. 6). Tra la ceramica da fuoco è importante segnalare la presenza di alcune forme che imitano le olle egee con orlo estroflesso con gradino ed incavo nella parte interna per l’alloggio del coperchio, ben diffuse in tutto il Mediterraneo (Fig. 7, nn. 11-12). Un tipo particolare è rappresentato da alcune olle che si avvicinano ad esemplari “tipo Ventimiglia” (Fig. 7, nn 13-14), mostrando però un impasto locale/regionale.

2.2 Le anfore Il quadro del materiale anforico che viene fuori dallo studio del territorio si presenta assai articolato. Infatti, la presenza dell’antico porto nell’area di Trainiti e la ricchezza dell’altopiano del Poro hanno determinato una vitalità economica caratterizzata anche da commerci ad ampio raggio con diverse regioni dell’Impero. Inoltre, le ricche produzioni agricole hanno permesso di immettere nei circuiti commerciali le eccedenze, attraverso l’esportazioni di anfore adibite al trasporto di vino, garum, frutta e pece, già a partire dall’età greca e per tutta l’epoca tardo antica.

Da quanto qui sintetizzato possiamo trarre delle prime indicazioni sulla città e su produzioni e commerci di epoca tardo antica: 1) tra fine VI e inizi VII secolo d.C., la parte settentrionale della città, nelle aree Soriano, Buccarelli e, soprattutto, S.Aloe, continua ad essere abitata con forme ancora tutte da definire; infatti, anche se in gran parte della città le abitazioni di età imperiale erano abbandonate, questi settori, anche se in maniera puntiforme, mostrano una continuità di frequentazione; 2) sono state rinvenute significative quantità di anfore di provenienza africana a fronte di una bassa percentuale di anfore di provenienza orientale, a testimonianza degli intensi rapporti commerciali che le città del Bruttium intrattenevano con l’Africa settentrionale; 3) la presenza delle produzioni africane è fortemente presente e si attarda per tutto il VII sec. a Piscino di Piscopio, centro legato probabilmente all’autorità ecclesiastica che svolge anche importanti compiti amministrativi; 4) l’intensità degli scambi commerciali può aver indotto artigiani del luogo a produrre ceramica comune ed anfore ad imitazione di quelle africane; del resto ci sono, a S. Aloe, Buccarelli ed altrove, imitazioni in ceramica di produzione locale di forme tipiche dell’africana da cucina e della terra sigillata africana D di V e VI d.C.; 5) l’insieme dei rinvenimenti sottolinea la floridezza economica e produttiva della regione in epoca tardoantica; ciò è dovuto, probabilmente, agli stretti legami mantenuti con l’Africa, che continuò ad essere la regione economicamente più attiva del Mediterraneo, ed al fatto che la Calabria mostrava una maggiore adattabilità al sistema del latifondo.

Le anfore attestate sono principalmente di produzione locale (47%) e sono rappresentate dalle Keay LII. Tra le importazioni troviamo in maggioranza quelle di provenienza africana (38%): Africane I-III, tripolitane, Keay XXV, Keay XXVI, anfore cilindriche di grandi dimensioni; seguono le iberiche attestate con il 3%: anfore da garum come le Almagro 51C e le Keay XIX; le produzioni orientali (1,8%) sono rappresentate da alcune pareti e da un solo orlo di LRA 2=Yassi Ada I. Anche le anfore Tripolitane II e III sono poco rappresentate (4%). Le produzioni Tripolitane cominciano ad arrivare già nel I d. C. con la Tripolitana I, ed è soprattutto attestata la successiva Tripolitana II (Fig. 8, nn. 1-2). Un rapporto privilegiato con le produzioni tripolitane era stato già messo in evidenza per la Calabria ionica centromeridionale, dove questa tipologia è ben rappresentata non solo nelle canoniche forme dei tipi II e III ma anche in quelle assimilabili alle produzioni finali della Tripolitana II, esito di un’evoluzione morfologica articolata in più fasi. In realtà, le anfore Tripolitana II e III, come mostra lo studio dei contesti di Blanda e Bivona, non sembrano particolarmente attestate sul versante tirrenico. I pochi esemplari rinvenuti nel territorio di Vibo Valentia, in particolare a Briatico, Pannaconi e Vibo Marina S. Venere, insieme a quelli di 68

CUTERI-IANNELLI-VIVACQUA-CAFARO

garum, si richiedeva l’uso di più piccoli contenitori (Etienne, Mayet 2007).

Nicotera Casino Mortelleto, rappresentano una testimonianza importante per il traffico di merci, presumibilmente olio o salse di pesce, provenienti dalla provincia della Tripolitania (Fig. 8, nn. 1-3). Ma il rapporto commerciale più proficuo e duraturo nel tempo è con le regioni della Bizacena e della Zeugitania, che nel Vibonese rappresentano il maggior bacino di approvvigionamento di merci tra la fine del II sec. d. C. ed il VII d. C.. Questo aspetto si coglie anche nei siti del territorio di Vibo Valentia dove appaiono le africane I A e B e II A, C, D (Fig. 8, nn. 4-12). Esse sono particolarmente attestate sulla costa a Vibo Marina S. Venere, a Briatico Satriani e Scrugli, a Macrone contrada Mortà, nei siti di mezza costa di S. Onofrio Lanzaro e Triparni ed, infine, nella città di Vibo Valentia. Incerto è il contenuto di queste anfore.

È con il IV secolo d. C. che si registra in Calabria e Sicilia una nuova stagione produttiva segnata dalla fabbricazione delle Keay LII e dalla concomitante esportazione del surplus del vino. Questi contenitori si inseriscono all’interno di una linea produttiva che aveva interessato il Bruttium già in età greco-ellenistica, dimostrando l’attitudine di questa regione a produrre per l’esportazione. Tale capacità è documentata in particolar modo a partire dal IV sec. d. C., quando le dinamiche di occupazione della terra sembrano essersi stabilizzate nella strutture del grande latifondo, sul modello di quello africano, grazie alla concentrazione di estese proprietà nelle mani di alcuni domini. Il latifondo risultò particolarmente vitale poiché, pur non soffocando altri sistemi di conduzione agricola, come la piccola proprietà, riuscì probabilmente a concentrare in sé una struttura economica polifunzionale, compatibile con i diversi modi e rapporti di produzione che caratterizzavano il territorio.

Per quanto riguarda la produzione, per alcuni esemplari di Africana I provenienti da Blanda, datati al IV-V d. C., si è recentemente parlato, in base alle caratteristiche dell’impasto, di una imitazione locale/regionale (Sangineto 2006a). Il fenomeno dell’imitazione delle forme ceramiche africane è molto attestato e non è da escludere che l’intensità dei traffici commerciali possa aver indotto artigiani del luogo a produrre ceramiche ed anfore ad imitazione di quelle d’oltremare. Tra i nostri esemplari, le Keay XXV B presentano un impasto di colore giallino molto micaceo che esclude l’Africa come centro produttore prospettando un’imitazione anche locale/regionale. Accanto a queste ultime, è stato rinvenuto un solo frammento di puntale, di incerta provenienza, assimilabile alla forma Dressel 30=Keay I (Fig. 8, n. 13), datata al III-IV d. C. e normalmente ritenuta di produzione mauretana, come dimostrano i bolli su di essa rinvenuti, o anche di produzione tunisina (Bonifay 2004, 148-151; Ghalia, Bonifay, Capelli 2005).

Nel Vibonese, le Keay LII (Fig. 10, nn. 1-6) sono attestate lungo la costa a Briatico Trainiti, nello scavo della necropoli romana, a Briatico Case Scrugli, all’antemurale di Trainiti, a Vibo Marina e a Santa Venere e, a mezza costa, a S. Onofrio Lanzaro, a Triparni e a Cessaniti Mantineo. Inoltre, numerosi frammenti sono stati recuperati in città, come a S. Aloe, o in via XXV Aprile, dove due frammenti presentano il bollo con la menorah (Cuteri 2008-2009). Ciò che è importante sottolineare è la forte analogia tra gli impasti delle Dressel 2-4, Dressel 21-22 e Keay LII, confermata dalle analisi archeometriche che hanno rivelato una produzione in loco di queste diverse tipologie anforiche. Altra produzione attestata nel territorio è quella delle forme simili ad Ostia I, fig. 455-456/Ostia IV, fig. 166, normalmente attribuite ad una produzione siciliana, provenienti probabilmente da un relitto affondato nei pressi di Vibo Marina e datate alla metà del IV-V d. C. (Fig. 10, nn. 7-9). Queste anfore, simili alle produzioni di Termini Imerese tipo 151 (Belvedere et al. 1993) e Capo Orlando tipo 2 (Ollà 2004), rappresentano un ritrovamento importante nell’ambito delle produzioni dirette a Roma. La qualità dei frammenti non ha permesso un’analisi dell’impasto che però sembra da riferire alla produzione di Capo Orlando. Interessanti appaiono anche alcune anse con particolare sezione a fiorellino (Fig. 10, n. 10), attestate anche nella città di Vibo Valentia, in Sicilia ed in Corsica. Le analisi eseguite su alcuni frammenti di Mariana in Corsica indicano un’origine siciliana (Menchelli et al. 2007) mentre i nostri esemplari, che presentano un impasto simile a quello delle Keay LII, risultano di produzione locale. Si datano tra I e V d. C.

Le analisi condotte sulle produzioni anforiche consentono di mettere in evidenza come vi sia una continuità di importazioni africane a partire dal II d. C., destinata a crescere dalla fine del III-V d. C. con la presenza di anfore di tipo Africana II e Africana III=Keay XXV. Continua questo rapporto anche nella tarda antichità con una buona percentuale di anfore cilindriche di grandi dimensioni. Rimangono più marginali gli spatheia che nel territorio sono poco diffusi, mentre sembrano concentrarsi soprattutto nella versione miniaturistica a Piscino di Piscopio. Nel territorio preso in esame, anche l’approvvigionamento di salse di pesce dalla penisola spagnola è ben attestato già in età imperiale con l’arrivo delle Dressel 7-11 e Beltran II A e B e continua soprattutto nel III-V d. C. con la presenza delle Keay XIX A e C (Fig. 9, nn. 1-3) e Alamgro 51C (Fig. 9, nn. 4-7). È possibile che, in età tardo antica, vi furono dei cambiamenti nelle modalità della pesca e della lavorazione del pesce, trasportato poi in piccole anfore come le Almagro 51C. Se nel I e II d. C. c’era bisogno di anfore con grande capacità per esportare sardine e sgombri, in questo periodo, invece, con la produzione delle salse tipo l’hallec, residuo della fabbricazione del

Nel corso del IV sec. d. C., accanto alle Keay LII non mancano le merci provenienti dall’Africa. Esse sono rappresentate, fino alla metà del V d. C., da anfore cilindriche di medie dimensioni, da anfore Keay XXVI o da spatheion (Fig. 11). Le Keay XXV o le Africane III sono tra i contenitori maggiormente diffusi in Calabria. La varietà morfologica delle anfore cilindriche e 69

LRCW4

Queste anfore sono attestate a Vibo Marina Santa Venere, sito che si conferma anche per la sua posizione topografica particolarmente aperto ai traffici commerciali. Le attestazioni di anfore di origine orientale appaiono assai meno concentrate sulla costa tirrenica calabrese rispetto a quella jonica. Tuttavia, i dati desunti dallo studio del territorio dimostrano come anche il versante tirrenico centro meridionale, storicamente legato alla chiesa di Roma e articolato nelle massae pontificie, fosse interessato, seppure in misura percentualmente minore, dalla circolazione di merci prodotte nell’oriente (Cuteri-Salamida 2010). In conclusione, dunque, possiamo annotare che nei secoli tra il IV ed il VII d. C. il territorio di Vibo V., grazie anche al suo porto, si rivela particolarmente vitale e aperto ai traffici commerciali, sebbene con una netta preponderanza, sui siti costieri, di merci africane rispetto a quelle orientali.

affusolate del tardo impero potrebbe essere il riflesso di una moltiplicazione degli impianti ceramici verificatasi nella Tunisia tra la fine del III e gli inizi del IV. Ciascuno di tali impianti potrebbe aver inventato o adottato uno o più modelli; contrariamente a quanto avviene nel I e II d. C., quando a fronte di una certa quantità di tipi prodotti solo pochi erano quelli destinati al commercio a lungo raggio e venivano smerciati in quantità significative con le Africane I e II (tipi classici del III d. C.) ancora fabbricate nel corso del IV. Gli spatheia (Keay XXVI) sono attestati nel territorio a Briatico Scrugli e Trainiti e a Vibo Marina S. Venere. Non sono stati rinvenuti frammenti del tipo miniaturistico (spatheion 3, classificazione Bonifay), datati al VI-VII d. C. Questi ultimi sono comunque attestati con elevati indici nel sito di Piscino di Piscopio e a Bivona; altri esemplari sono attestati anche a Crotone e Scolacium. Per le Keay XXVI è possibile ipotizzare un contenuto vinario, anche se vi sono numerose incertezze. Tra i frammenti di Keay XXV rinvenuti nel territorio di Vibo V., il tipo della Fig. 11, n. 4 (Keay XXV S- Africana III B), forma assai imitata, presenta un impasto micaceo (Bonifay 2004, 458) che lascia supporre un’origine diversa dall’Africa settentrionale, senza far escludere un contesto locale/regionale come si verifica in altri siti mediterranei.

2.3 Tropea Tropea, sede della massa trapeiana, già nel VI secolo era un importante centro economico e portuale, nonché sede di diocesi. In piazza Duomo, la necropoli ha restituito sigillata africana e numerose brocchette e anforette, anche di produzione africana, che hanno permesso di inquadrare il contesto tra V e VII sec. d.C. Le anfore tardo antiche sono rappresentate da un frammento di spatheion e da un orlo di Keay LII. Altri materiali provengono dall’are di frequentazione tardo-antica di via Tondo (Fig. 13), dove sono state individuate, in un contesto domestico o residenziale rinvenuto fortuitamente nel 1970, pentole e casseruole in ceramica grezza con evidenti segni d’uso al fuoco; anche alcune brocchette presentano tracce di annerimento. Per quanto riguarda le pentole, esse rimandano a tipi di derivazione orientale in uso tra VI e VIII secolo, mentre per le casseruole possiamo evidenziare una certa similitudine con esemplari rinvenuti a Napoli, Roma, Abruzzo e Sardegna che si rifanno al tipo 115 di Pantelleria. Per questi esemplari la datazione oscilla tra il V e l’VIII secolo.

A partire dal V sec. d. C. cominciano ad arrivare le anfore cilindriche di grandi dimensioni, con le Keay LXI e LXII e LVII A (Fig. 12, nn. 1-7). Queste tipologie di anfore appartengono alla seconda generazione dei contenitori di grandi dimensioni. Le anfore Keay LXII e Keay LXI, attestate a Vibo Marina Santa Venere, appartengono all’ultima generazione della famiglia dei grandi contenitori cilindrici, datati al VI-VII d. C. La loro diffusione in Occidente inizia a partire dal VI d. C. ed è stata messa in relazione con la riconquista bizantina del Nordafrica (533 d. C.) e con la conseguente riorganizzazione delle risorse agricole della provincia. Ciò che colpisce è la presenza di queste anfore in contesti posti lungo la costa, che rivelano la prolungata attività del porto di Vibo Valentia. Lo dimostrano, con chiarezza, anche l’area portuale di Bivona, dove il materiale importato più tardo arriva al VII d. C. e proviene dall’Africa e dall’Oriente e l’insediamento di Piscino, posto a mezza costa, che presenta produzioni anforiche di VI-VII d. C., costituite dagli spatheia di piccole dimensioni, dalle Keay LXII e dalle LRA 1, 2, 3 e 5/6. Al Castrum Perti, sulla base dell’elevata presenza di alcune anfore tunisine, si è ipotizzato un ruolo determinante dell’annona, direttamente controllata dall’autorità centrale, con una prevalente destinazione dell’esportazione di derrate alimentari da alcune regioni del nord Africa verso aree poste sotto il diretto controllo dell’Impero bizantino e legate allo stanziamento di guarnigioni militari (Murialdo 2001, 256). La stessa cosa può essere ipotizzata per il territorio di Vibo V., interessato nel VI sec. d.C. dallo svolgersi di alcuni atti della guerra greco-gotica, dove probabilmente vi erano stanziate truppe che si rifornivano con le merci africane.

2.4 Nicotera Il centro è menzionato come massa solo nel VI secolo. Per integrare i dati già presentati in LRCW2 (Cuteri et al. 2007), che sottolineano una frequentazione del territorio tra II e metà del VII sec. d.C. ed una notevole presenza di anfore provenienti dal Nord Africa, oltre alle .Keay LII di origine bruzia, proponiamo alcuni frammenti di ceramica recuperati in località San Pietro (Corrado 2002-2003). Si tratta di forme aperte in sigillata africana C3 e C4, decorate a rilievo applicato e a matrice, inquadrabili tra il 300 e il 430 d.C. Alla classe C3 sono riferibili: frammenti di coppe Salomonson a/Hayes 53A e Salomonson b, attestate quest’ultime con varianti prossime alla Hayes 52B, nn. 4 e 19 (Atlante, pp.158-159; 162-163); frammenti di piatti del tipo Salomonson e/Hayes 55 (Atlante, p.158). Alla classe C3-C4 sono invece da riferire i resti di grandi vassoi rettangolari Hayes 56 (Atlante, pp.160-161) realizzati ad imitazione delle coeve

Alle importazioni africane bisogna aggiungere alcuni frammenti di parete di LRA 1 e LRA 2, ed un orlo assimilabile al tipo LRA 2=Yassi Ada I (Fig. 12, nn. 8-9). 70

CUTERI-IANNELLI-VIVACQUA-CAFARO

établissement rural de l'antiquité tardive: fouilles récentes (1981-1985), 75-89. Paris, Maison des Sciences de l’Homme.

produzioni in metallo. Le decorazioni a rilievo, ottenute a matrice, mostrano i seguenti motivi: fitomorfi e zoomorfi (Fig. 14, nn. 1-2), con richiami a scene mitologiche, rappresentazioni circensi ed episodi delle fatiche di Ercole o del Vecchio e Nuovo Testamento(Fig. 23, nn. 3, 4) quali il sacrificio di Abramo e la guarigione del paralitico. Due frammenti di vassoio rimandano al matrimonio di Achille (Fig. 14, nn. 5-6), ed altri propongono scene con i Dioscuri a cavallo o dignitari seduti nella tribuna dell’anfiteatro.

Carsana, V. 1994. Ceramica da cucina tardo-antica ed altomedievale. In P. Arthur (ed.), Il complesso archeologico di Carminiello ai Mannesi, Napoli (scavi 1983-1984), 181-220. Galatina, Congedo Ed. Corrado, M. 2002-2003. Ceramica sigillata africana da Nicotera, località San Pietro. Studi Calabresi 3-4, 245-246. Cuteri, F. A. 2008-2009. Ebrei e samaritani a Vibo Valentia in età tardo antica: le testimonianze archeologiche. Sefer Yuhasin XXIV-XXV, (Anno 5768-5769), 17-38.

BIBLIOGRAFIA Atlante I = AA. VV. 1981. Atlante delle forme ceramiche, I, Ceramica fine romana nel Bacino mediterraneo (medio e tardo impero), supplemento. Enciclopedia dell'Arte Antica, Classica e Orientale, Roma, Treccani.

Cuteri, F. A., Corrado, M., Iannelli, M. T., Paoletti, M., Salamida, P. e Sangineto, B. A. 2007. La Calabria fra Tarda Antichità ed Alto Medioevo attraverso le indagini nei territori di Vibona Valentia, della Massa Nicoterana, di Stilida-Stilo. Ceramiche, commerci, strutture. In M. Bonifay and J.-Ch. Tréglia (eds.) LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. British Archaeological Reports, International Series 1662, 461-476. Oxford, BAR Publishing.

Arthur, P. e Peduto, P. 1991. Un edificio bizantino extra moenia a Vibo Valentia. In Calabria bizantina: testimonianze d’arte e strutture di territorio, Atti dell’VIII e IX Incontro di Studi Bizantini (Reggio Calabria 1985/1988), 7-20, Soveria Mannelli, Rubbettino Ed. Belvedere, O., Bergio, A., Macaluso, R. e Rizzo, M. S. 1993. Termini Imerese. Ricerche di topografia e di archeologia urbana, Palermo.

Cuteri, F.A. e Salamida, P. 2010. Il litorale jonico calabrese da Crotone a Reggio (Calabria – Italia). Circolazione di menufatti ceramici tra V e VII secolo. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Comparison between western and eastern Mediterranean, British Archaeological Reports, International Series 2185, 507-513. Oxford, BAR Publishing.

Bernardi A. 2010-2011. Indagine chimico fisica delle anfore del territorio di Vibo Valentia, Tesi di Laurea Università Cà Foscari Venezia, Facoltà di Scienze Matematiche Fisiche e Naturali. Bonifay, M. 1998. Sur quelques problèmes de datation des sigillées africaines à Marseille. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del Convegno in onore di John W. Hayes, 71-81. Firenze, All’Insegna del Giglio.

Etienne R., Mayet, F. 2007. L’industrie des salaisons et sauces de poisson dans la Péninsule Ibérique. Etat de la question. In L. Lagóstena, D. Bernal e A. Arévalo (eds.), Cetariae 2005. Salsa y Salazones de Pescado en Occidente durante la Antigüedad, Actas del Congreso Internacional (Cádiz, 7-9 novembre 2005). British Archaeological Reports, International Series 1686, 5-20. Oxford, BAR Publishing.

Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique, British Archaeological Reports, International Series 1301. Oxford, BAR Publishing. Bonifay, M. 2005. Observations sur la typologie des amphores africaines de l’Antiquité tardive. In J. M. Gurt I Esparraguera, J. Buxeda I Garrigòs and M. A Cau Ontiveros (eds.), LRCW1. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. British Archaeological Reports, International Series 1340, 451-472. Oxford. BAR Publishing.

Fulford, M. e Peacock, D. P. S. 1984. Excavations at Carthage: the British mission, I, 2. The pottery and the other ceramic objects from the site. Sheffield. Ghalia, T., Bonifay, M. e Capelli, C. 2005. L’atelier de Sidi Zahruni: mise en evidence d’une production d’amphores de l’Antiquité tardive sur le territoire de la cité de Neapolis (Nabeul, Tunisie). In LRCW1, 495507.

Bonifay, M. e Pieri, D. 1995. Amphores du V au VII s. à Marseille: nouvelles données sur la typologie et le contenu. Journal of Roman Archaeology, 8, 94-119. Capelli, C. 1998. Analisi minero-petrografiche. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del Convegno in onore di John W. Hayes, 331-333. Firenze, All’Insegna del Giglio.

Hayes, J. W. 1972. Late Roman Pottery, London, The British School at Rome. Iannelli, M. T. 1989. Hipponion-Vibo Valentia, Documentazione archeologica e organizzazione del territorio. Annali della Scuola Normale Superiore di

Carignani, A. 1989. Amphores. In P. Pergola e C. Vismara (eds.) Castellu (haute-corse). Un 71

LRCW4

Pisa, classe di Lettere e Filosofia, Ser. III, XIX, 2, 683-764.

Reynolds, P. 1995. Trade in the Western Mediterranean, AD 400-700: the ceramic evidence. British Archaeological Reports, International Series 604. Oxford, BAR Publishing.

McCormick, M. 1998. Bateaux de vie, bateaux de mort. Maladie, commerce, transports annonaires et le passage économique du bas–empire au moyen age. In Morfologie sociali e culturali in Europa tra tarda antichità e alto medioevo XLV, 35-122. Settimana di Studio del CISAM, Spoleto.

Ricci, M. 1998. La ceramica comune dal contesto di VII secolo della Crypta Balbi. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del Convegno in onore di John W. Hayes, 351-382. Firenze, All’Insegna del Giglio.

Menchelli, S., Capelli, C., Pasquinucci, M., Picchi, G. Etienne R., Mayet, F. 2007. Corsica tardo antica: anfore italiche e ceramica comune da Mariana. In LRCW2, 313-328.

Saguì, L. 2002. Roma, i centri privilegiati e la lunga durata della tarda antichità. Archeologia Medievale XXIX, 7-42. Sangineto, A. B. 1989. Scavi nell’abitato romano di Vibo Valentia. Annali della Scuola Normale Superiore di Pisa, classe di Lettere e Filosofia, Ser. III, XIX, 2, 833-843.

Murialdo, G. 2001. Le anfore da trasporto. In T. Mannoni e G. Murialdo (eds.), S. Antonino: un insediamento fortificato nella Liguria bizantina Collezione di monografie preistoriche ed archeologiche XII, 255296. Bordighera.

Sangineto, A. B. 2006. Ceramica comune. In G. F. La Torre (ed.), Blanda Julia sul Palecastro di Tortora. Scavi e ricerche (1990-2005), 336-369. Messina, DiScAM.

Olcese, G. 1993. Le ceramiche comuni di Albitimilium. Indagine archeologica e archeometrica sui materiali nell’area del Cardine. Firenze, All’Insegna del Giglio.

Sangineto, A. B. 2006a. Anfore. In G. F. La Torre (ed.), Blanda Julia sul Palecastro di Tortora. Scavi e ricerche (1990-2005), 310-335. Messina, DiScAM.

Olcese, G. 2003. Ceramiche comuni a Roma e in area romana: produzione, circolazione e tecnologia. Mantova, S.A.P.

Semeraro, G. 1992. La ceramica comune romana e preromana. In F. D’Andria and D. Whitehouse (eds.), Excavations at Otranto, volume II: the Finds, 63-78. Galatina, Congedo Ed.

Ollà, A. 2004. Ceramica da Bagnoli-S. Gregorio: importazioni e produzioni locali. In U. Spigo (ed.), Archeologia a Capo Orlando. Studi per l’Antiquarium, 109-132. Milazzo.

Small, A. M. e Buck, R. J. 1994. The Excavations of San Giovanni di Ruoti. I. The villas and their Enviroment, Phoenix, Supplement 33, Toronto, University of Toronto Press.

Ostia I = AA. VV. Ostia I. Le terme del nuotatore, Studi Miscellanei 13, Roma 1968. Ostia II = AA. VV. 1970. Ostia II. Le terme del nuotatore, Studi Miscellanei 16. Roma.

Vivacqua, P. 2008. Nuove ricerche sul complesso di Piscino di Piscopio (Vibo Valentia): le evidenze ceramiche. In Vivarium Scyllacense – Bollettino dell’Istituto di Studi su Cassiodoro e sul Medioevo in Calabria XVII, 1-2, 65-94.

Ostia III = AA. VV. 1973. Ostia III. Le terme del nuotatore, Studi Miscellanei 21. Roma. Ostia IV = AA. VV. 1978. Ostia IV. Le terme del nuotatore, Studi Miscellanei 23. Roma.

72

CUTERI-IANNELLI-VIVACQUA-CAFARO

Fig. 1. Carta del territorio di Vibo Valentia con l’indicazione delle aree d’interesse archeologico

Fig. 2. Vibo Valentia, Cantiere Soriano. Anfore: 1. Africana I B; 2. Africana II A; 3. Keay XXVI C; 4. anse a fiorellino

Fig. 3. Vibo Valentia. Cooking ware: 1-6. (Cantiere Soriano); 7-11. (Via Protettì); n. 12,13. coarse ware (Cantiere Soriano)

73

LRCW4

Fig. 4. Vibo Valentia, Cantiere Buccarelli. Ceramica da cucina

Fig. 5. Vibo Valentia, Cantiere Miceli. Anfore: 1-2. Keay LII; 3. Tripolitana 1; 4. Africana IID; 5. Keay XXV=Africana III A e B; 6. Keay XXVI= Spatheion tipo 1; 7. Keay VIIIB

74

Fig. 6. Piscino di Piscopio (VV). Spatheia di piccole dimensioni: 1. campione 4dPP; 2. campione 3cPP; 4. campione 2bPP; 6. campione 5ePP

CUTERI-IANNELLI-VIVACQUA-CAFARO

Fig. 7. Piscino di Piscopio (VV). Ceramica comune (1-6) e da fuoco. Tipi particolari: 1-3. simili a forme campane; 6. imitazione di forme africane; 13-14. simili al “tipo Ventimiglia”

Fig. 8. Territorio di Vibo Valentia. Anfore tripolitane e africane I e II: 1-2. Tripolitana II; 3. Tripolitana III; 4. Africana 1A; 5-6. Africana IB; 7-9. Africana IIA; 10. Africana IIC; 11-12. Africana IID; 13. Dressel 30

75

LRCW4

Fig. 9. Territorio di Vibo Valentia. Anfore iberiche: 1. Keay XIX A; 2-3. Keay XIX C; 4-7. Almagro 51C

76

CUTERI-IANNELLI-VIVACQUA-CAFARO

Fig. 10. Territorio di Vibo Valentia. Anfore di produzione calabrese e siciliana: 1-6. Keay LII; 7-9. Ostia I, figg. 455456/Ostia IV, fig. 166; 10. anse “a fiorellino”

77

LRCW4

Fig. 11. Territorio di Vibo Valentia. Anfore Africana III: 1. Anfora Keay XXV B=Africana III A; 2. Anfora Keay XXV C=Africana III A; 3. Keay XXV Q=Africana III B; 4. Keay XXV S= Africana III B (impasto non africano); 5. Keay XXV G=Africana III C

Fig. 12. Territorio di Vibo Valentia. Anfore di grandi dimensioni: 1. Keay LVII B; 2. Keay LVII A; 3. Keay LIX; 4. Keay LXII Q; 5-6. Keay LXII; 7. Keay LXI; 8. LRA2=Yassi Ada I; 9. LRA2 78

CUTERI-IANNELLI-VIVACQUA-CAFARO

Fig. 13. Tropea, Via Tondo. Frequentazione tardo antica: ceramica da fuoco, brocchette ed anforette

Fig. 14. Nicotera, Località San Pietro: frammenti in sigillata africana C3 e C4 decorata

79

LRCW4

80

CERAMICHE COMUNI E DA FUOCO DALL’INSEDIAMENTO TARDOANTICO DI CARABOLLACE (SCIACCA, SICILIA, ITALIA): CARATTERIZZAZIONE TIPOLOGICA E ARCHEOMETRICA VALENTINA CAMINNECI1, ANNALISA AMICO2, FRANCESCO GIANNICI3, RENATO GIARRUSSO4, ANGELO MULONE4 1

Soprintendenza BB.CC.AA. Agrigento; [email protected] 2 Università di Messina; [email protected] 3 Dipartimento di Fisica e Chimica, Università di Palermo 4 Geolab s.r.l. Palermo; [email protected]

Some selected samples of coarse and cooking ware found at Carabollace, a late Roman settlement near the coast of Sciacca, in western Sicily, at the mouth of the Carabollace river, have been analyzed by thin section petrography. This research intends to point out the circulation and the importation networks as well as the local productions of pottery, with the aim to provide an answer to several historical questions. The location near the river and the sea suggests that this settlement may have been an emporium for transmarine commerce and for goods from the local markets of the hinterland. Ceramic evidence attests that Carabollace was strictly linked to the African trade in the 5th and 6th century AD. Thin section analysis helped us also to recognize some local and regional products that present interesting similarities with coarse ware attested in the near site of Verdura whose samples can be grouped in the same petrographic group. KEYWORDS: SICILY, CARABOLLACE, LATE ANTIQUITY, COARSE WARE, COOKING WARE, ANCIENT VIABILITY, THIN-SECTION ANALYSIS. Le strutture rinvenute a Carabollace, pertinenti a due edifici rettangolari ripartiti in più vani, databili tra la seconda metà del V ed il VI secolo d.C., che obliterano ed in parte riutilizzano i muri di una fase immediatamente precedente sempre del V secolo d.C., potrebbero essere interpretate come magazzini per lo stoccaggio di merci. L’edificio A (Fig. 1), in particolare, presenta analogie planimetriche con la casa XVII dell’abitato di Punta Secca, per cui si è supposta un’eguale destinazione (Di Stefano 2002). La presenza del fiume assicurava la deportatio ad aquam e la circolazione dei prodotti importati anche nei siti più lontani dalla costa, come sembrerebbe suggerire l’etimo carabus, barchetta, degli idronimi Carabollace e Carboj, quest’ultimo, ricordato da Edrisi nel XII secolo come al qawarib, il fiume delle barchette, con tracce di insediamento su entrambi i lati della foce, per sfruttare il vento favorevole (Fig. 2, 9-10).

In attesa di poter proseguire le ricerche archeologiche nell’insediamento tardo antico di contrada Carabollace, sulla costa saccense, oggetto di una prima presentazione negli Atti del congresso precedente (Caminneci, Franco e Galioto 2010), si è cercato di chiarire, attraverso l’ausilio di indagini archeometriche, la provenienza del materiale ceramico rinvenuto. Allo studio morfologico dei reperti si è affiancata, dunque, l’analisi degli impasti al fine di ricostruire con maggiore attendibilità i flussi commerciali che toccarono questo centro. A due passi dal mare, Carabollace sembra avere svolto, tra l’età tardo imperiale e quella alto bizantina, nei confronti dell’entroterra circostante, un importante ruolo di riferimento, come parte di un sistema di approdi costieri, posti alla foce dei fiumi, probabilmente collegati attraverso il cabotaggio con un caricatore più importante, forse Lilibeo, la città che in età vandalica assume una notevole importanza politica e strategica, quale testa di ponte del collegamento con la Tunisia (Caminneci 2010).

Partendo dal porto-caricatore di Nabeul attraverso una rotta nota ancora ai tempi di Edrisi, ad appena due giorni di navigazione, la costa occidentale siciliana rappresentava la tappa “naturale” della navigazione dall’Africa verso Roma. Di tale rotta sarebbero testimonianza, per il periodo in questione, i relitti di Cefalù, Filicudi1, ed Ustica, i cui carichi mostrano significative rispondenze con le produzioni presenti a Carabollace, così come i relitti di Dramont E e la Palud, entrambi di provenienza neapolitana (Volpe 2002; Bonifay 2004, 453). La presenza delle stesse classi di materiali a Carabollace e nei relitti francesi potrebbe essere il segnale di un’unica rotta di percorrenza che da Nabeul, da cui provenivano i contenitori anforari, toccando la Sicilia, giungeva alla Narbonese, dove fu

La carta archeologica dei siti individuati nel territorio comunale di Sciacca e di Menfi rivela un quadro definito delle modalità del popolamento nel periodo tardoromano con insediamenti sulla costa, proprio alle foci dei fiumi e, nell’immediato entroterra, nelle piane in prossimità dei valloni fluviali (Fig. 2). In questo tratto, sulla strada costiera che collegava Agrigentum a Lylibaeum, l’Itinerarium Antonini segnala la statio Ad Aquas, la cui importanza viene evidenziata nella Tabula Peutingeriana, che le dedica l’icona imponente di un edificio termale, con la didascalia Aquae Alabodes (Uggeri 2004, 163).

81

LRCW4

devono aver giocato un ruolo determinante nelle dinamiche distributive delle merci locali e di importazione (Caminneci 2010).

esportato grano siciliano fino ad età teodoriciana (Caminneci e Franco 2012). In particolare, il carico misto de La Palud con la presenza di anfore orientali accanto a quelle africane potrebbe essere la prova della funzione di tramite assicurata dall’Africa nella diffusione delle merci dall’oriente. Le Anfore LRA1, LRA2, LRA3 rinvenute a Carabollace, unica classe di materiale di provenienza orientale, potrebbero, perciò, essere giunte dall’Africa, tappa intermedia verso l’occidente contemplata anche negli itinerari seguiti dai santi di ritorno dall’Oriente (De Salvo 1997-1998).

Potrebbe essere comunque interessante ricordare, nell’ambito di una riflessione sulle produzioni locali in Sicilia, che nell’Isola in età tardo romana non si tengono quelle fiere che altrove (ad esempio in Puglia, Volpe 1996, 336 ss.) costituiscono una proficua occasione di scambio dei prodotti locali, circostanza forse dovuta al fatto che la produzione isolana venisse rastrellata dai grossi proprietari e convogliata verso i porti (Cracco Ruggini 1980, 16). La mancanza di questi appuntamenti e, quindi, di dinamiche di scambio a raggio limitato, se non ci smentisce il progresso delle conoscenze, potrebbe, a nostro avviso, avere inciso sull’entità delle produzioni locali di ceramica e favorito, di contro, la diffusione dei prodotti importati. Sembra, comunque, ormai accertata una produzione di anfore su scala regionale, il cosiddetto tipo Benghazi MRA 1, rinvenuto anche a Carabollace, come pure risultano ampiamente documentate in più punti dell’Isola officine che produssero anfore morfologicamente affini, dal corpo rastremato, fondo piatto e anse a maniglia, alcune delle quali, quelle prodotte a Termini, destinate all’esportazione, come dimostrano i rinvenimenti in contesti romani (Panella, Saguì e Coletti 2010, 66). Di queste anfore la Rizzo, in questo volume, offre un primo inquadramento cronotipologico (Rizzo et alii nel presente volume).

IMPORTAZIONI E PRODOTTI LOCALI Lo stretto collegamento con il Nord Africa è testimoniato dalla percentuale più numerosa della ceramica rinvenuta a Carabollace, ascrivibile agli ateliers attivi nell’area del Golfo di Hammamet. Campioni rappresentativi di sigillate, ceramica da fuoco ed anfore di produzione africana, analizzati al microscopio polarizzatore presso il DipTeRis dell’Università di Genova nell’ambito di un vasto progetto di ricerca archeometrica condotto dal CNR/IBAM e il CNRS francese-/Centre Camille Jullian mirante alla individuazione delle fabbriche africane in Sicilia (Malfitana, Bonifay e Capelli 2008), hanno confermato la presenza di prodotti dalla regione di Hammamet, in particolare da Sidi Zahruni, a cui si attribuisce il maggior numero di esemplari (Keay 35 A-B, 55,56,57 ed il mortaio Bonifay Commune 13), di Oudhna (Hayes 91 e 99), di Nabeul (spatheia, Bonifay Commune 61 e 62, ma anche da centri della Tunisia centrale (sigillata C4 con le forme Hayes 73, 76 e C5 con le Hayes 82, 84), come Salakta, Sidi Khalifa (sigillata C/D, varianti Hayes 68, 88 A, 91) e la zona di Rougga, da cui oltre l’anfora Keay 34, proviene la forma Hayes 103/ Rougga gruppo 3, documentate insieme nello strato di crollo US 60, riferibile al contesto più tardo della seconda metà V-VI sec. d.C. La forma Hayes 103 caratterizzata da un impasto ricco di microfossili, ampiamente documentata nella Byzacena, risulta sinora un unicum in Sicilia (Caminneci e Franco 2012).

CERAMICHE DA FUOCO Il contesto ascrivibile al pieno V sec.d.C. ha restituito frammenti residuali di ceramica da cucina africana (nmi 47) riferibili alla produzione a patina cenerognola e ad orlo annerito, con le casseruole Hayes 197, i coperchi Hayes 185 e 196 e loro varianti, e la scodella Hayes 181, polita a strisce (Bonifay 2004, figg. 120, 118, 121, 114). La maggior parte dei frammenti di ceramica da fuoco (nmi 103) appartengono, invece, alla Pantellerian Ware, attestata in entrambi i contesti cronologici con le stesse forme. Il colore dell’impasto varia dal rosaceo al rossorosso mattone; sono presenti inclusi vetrosi bianchi, grigi, neri, anche vulcanici. Le superfici, annerite, sono in genere lisciate e rifinite a stecca.

Se le importazioni, con indici che coprono quasi la totalità delle attestazioni per le anfore, risultano preponderanti, alcuni impasti, campionati tra le ceramiche comuni e da fuoco, che, attraverso il semplice esame macroscopico, non è stato possibile attribuire a precise produzioni, fanno ragionevolmente supporre un’origine locale o regionale di alcune ceramiche, sebbene ad oggi non siano note fabbriche in questo territorio. La circostanza fortunata delle ricerche archeologiche condotte quasi contemporaneamente nel vicino sito costiero di Verdura (Parello, Amico e D’Angelo 2010), ha consentito di inquadrare i dati archeometrici dei reperti da entrambi i siti, grazie alle analisi condotte presso il laboratorio Geolab s.r.l. di Palermo, nell’ambito di un più ampio contesto geografico di riferimento. Le affinità riscontrate tra gli impasti ci fanno intravedere la circolazione di questi prodotti in quello che potremmo definire un “distretto culturale”, un comprensorio esteso tra costa ed entroterra, in cui la viabilità di terra e, soprattutto, il trasporto su fiume

Le forme più attestate sono il tegame (nmi 28) e la teglia (nmi 24), nelle loro varianti (Santoro, Guiducci e Tusa 2003, figg. 4-5): la preferenza accordata alle forme basse potrebbe essere messa in relazione a particolari scelte alimentari (Arthur 2007). Tra i tipi riconosciuti, la teglia con orlo a tesa ed estremità ingrossata, attestata nel Quartiere Ellenistico Romano di Agrigento (Fiertler 2003, B1/1), così come la teglia con orlo a tesa inclinata (Fiertler 2003, tipo B1/2,2) documentata ancora ad Agrigento, nella necropoli paleocristiana (Bonacasa Carra1995, fig. 70, tipo 86.415). Anche i piatti-coperchio (nmi 32) presentano una varietà nel profilo dell’orlo, da quello quasi indistinto del campione CAR 61/84 (Gruppo 1, Fig. 1,1), a quello ingrossato, mentre, generalmente, i frammenti della presa 82

CAMINNECI-AMICO-GIANNICI-GIARRUSSO-MULONE

inclusi quarzosi, che avevamo in precedenza attribuito, sulla base dell’osservazione macroscopica, ad una calcitic ware, forse locale (Santoro 2007, 369), ma che le analisi indicano invece proveniente dall’Africa settentrionale (Gruppo 3. CAR 125/1). Lo stesso impasto è stato riscontrato anche in alcuni campioni di ceramica comune.

mostrano una lieve depressione al centro. I confronti, oltre che ad esemplari rinvenuti a Pantelleria (Forma L.1, e tipi L2.1 e L2.2 e Sami 2005, tav.I, 1.4, dal relitto di Scauri) rimandano a varianti documentate nei contesti agrigentini citati (Fiertler 2003, Forma D1 e tipo D2/1, D2/2.2, D3/1,2; Bonacasa Carra 1995, figg.73-74) e della villa di contrada Saraceno, presso Agrigento (Castellana e Mc Connell 1990, fig.8, 86/5 e 86/10).

[V. C.]

E’ da rimarcare la presenza di Pantellerian Ware nel contesto di pieno VI secolo d.C., così come avviene in altri siti coevi dell’agrigentino (Calamonaci: Parello et alii in questo volume; Cignana: Rizzo e Zambito 2010), a differenza di quanto è attestato in altre aree della Sicilia, dove il trend delle importazioni pantesche subisce un arresto nella seconda metà del V secolo (area iblea: Cacciaguerra 2010). La continuità delle relazioni commerciali fu forse determinata dalla posizione geografica dell’isola, antistante la costa centromeridionale siciliana, tappa intermedia ed equidistante, appena un giorno di navigazione, lungo la rotta da Nabeul alla Sicilia. Nel 698 i musulmani, che avevano preso Cartagine, muoveranno proprio da Pantelleria per la spedizione di conquista, che approderà sulla costa di Mazara, circa quaranta Km a Nord di Sciacca. La diffusione della ceramica da fuoco dimostra il coinvolgimento dell’isola nei flussi commerciali del Mediterraneo in età tardoantica, che condizionarono anche i ritmi produttivi ad andamento stagionale (Santoro 2003). L’Itinerarium maritimum, se è vera l’ipotesi di attribuzione all’età vandalica (Uggeri 1998), dedicando un’intera sezione alle isole ne rimarca il ruolo strategico ed economico nel nuovo assetto determinato dalla perdita dell’Africa.

CERAMICA COMUNE Lo studio della ceramica comune di Carabollace ha evidenziato la consistente presenza (almeno il 40%) di reperti importati dall’Africa databili fino agli inizi del VII secolo. Molti impasti sono confrontabili con campioni ceramici del vicino sito di Verdura con ogni probabilità di origine africana (Gruppo 1: Parello et alii, nel presente volume); alcuni impasti poco caratterizzati sono di origine incerta; di alcuni campioni di ceramica comune del sito sono state invece effettuate le analisi in sezione sottile (CAR C1/C7). I dati quantitativi dei reperti sono calcolati sul numero minimo dei reperti diagnostici di ogni forma (nmi). Tra le brocche (nmi 50), i tipi Bonifay Communes 47 (nmi 3: Caminneci, Franco e Galioto 2010, fig. 3,10), 61 (nmi 2) e 62 (nmi 23, Fig. 3,1; Caminneci, Franco e Galioto 2010, fig. 3,7-8;) sono riconducibili a centri produttivi dell’area di Nabeul (Bonifay 2004, 73, 282, 290, 293); africana può essere una brocca con orlo a sezione triangolare (Fig. 3,2) simile al tipo Fulford Closed form 20.4 (Fulford 1984, 209). In base alle analisi archeometriche anche le brocche alla fig. 3,3-4 (CAR C6, Fig. 3,3) sono compatibili con i materiali africani. Due brocche (Caminneci, Franco e Galioto 2010, fig. 3,9 e 12, CAR C7) presentano impasto simile al campione VE-C13 di contrada Verdura, di probabile origine locale (Parello et al. nel presente volume).

Dallo strato di crollo della seconda metà del V secolo proviene un tegame della Fulford, Peacock 1984 Fabric 1.2 (Fulford e Peacock 1984, 8.5: Caminneci 2010a, fig.2,14), dall’impasto ricco di vetro vulcanico prodotta forse nelle Eolie o, come vorrebbero gli studi più recenti, in Sardegna (Cau, Iliopoulos e Montana 2002).

Interessante è la presenza di un frammento di parete di vaso chiuso decorato a stecca con un ramo stilizzato e con spesso ingobbio biancastro, simile ad un esemplare di contrada Verdura sottoposto ad analisi (Parello, Amico e D’Angelo 2010, 284, fig. 4,1; Parello et alii nel presente volume) e riferibile ad un peculiare gruppo di vasi, diffusi a livello sub-regionale, forse prodotti in età tardoantica in diversi centri della Sicilia sud-occidentale.

Gli inclusi vulcanici sono tipici anche della Fulford e Peacock 1984 Fabric 3.7, che presenta un’argilla nerogrigiastra, cioccolato al nucleo, ricca anche di quarzo e di mica dorata, riconosciuta qui in tre esemplari, rinvenuti negli strati di seconda fase, varianti di una stessa olla al tornio con orlo svasato e profilatura più o meno accentuata, con o senza anse (Fulford e Peacock 1984, fig.70, 35,1; 35,2: Caminneci, Franco e Galioto 2010, fig.2,15), la cui provenienza è ad oggi controversa (Reynolds 1995, 104).

Tra i contenitori da dispensa (nmi 14) un catino del tipo Bonifay Communes 21, uno del tipo 22 ed un altro affine al tipo Bonifay Communes 27B (CAR C3, Fig. 3,5) sono riconducibili al Nord Africa (Bonifay 2004, 263). Per il resto, sono ben attestati (con almeno dieci esemplari) recipienti in impasti per la maggior parte compatibili con materiali di probabile origine locale provenenti dal vicino sito di contrada Verdura (CAR C2, Fig. 3,6) e morfologicamente affini ai cosiddetti cantarelli di produzione agrigentina (Bonacasa Carra 1995, 143); un catino dello stesso tipo, tuttavia, presenta impasto diverso (CAR C4) di incerta origine, forse africano.

Una probabile origine dall’area peloritana è stata proposta per il coperchio con orlo a sezione triangolare (Gruppo 2. Campione CAR 22/115, Fig. 1,2) per le caratteristiche petrografiche dell’impasto, che riconducono alle rocce metamorfiche dei Monti Peloritani, anche se non si può escludere la Sardegna o la penisola iberica. Un’altra produzione al tornio lento è documentata da quattro esemplari -un tegame (Fig.1, 3) e tre frammenti di coperchio- dall’impasto rosato, granuloso e con molti 83

LRCW4

considerare di importazione. Il gruppo 1, rappresentato da un campione di ceramica da fuoco (CAR 61/84) è da ascrivere alla ben nota produzione di Pantellerian Ware. Il secondo gruppo, rappresentato anch’esso da un campione di ceramica da fuoco (CAR 22/115), ha caratteri mineralogici e tessiturali di un’argilla alluvionale, sedimentata in un ambito territoriale litologicamente dominato da rocce metamorfiche. Un tale contesto geolitologico può essere individuato, limitatamente alla Sicilia, nell’area dei Monti Peloritani; non può essere tuttavia esclusa un’importazione da altre regioni del mediterraneo quali ad esempio la penisola Iberica o la Sardegna. Il gruppo 3 è rappresentato da campioni (CAR C4, CAR C6, CAR 125/1) con peculiarità composizionali e tessiturali (scheletro a granulometria bimodale costituito quasi esclusivamente da quarzo con molti granuli di forma arrotondata e a superficie smerigliata) che inducono a ipotizzare come possibile area di provenienza il Nord Africa. I rimanenti quattro gruppi costituiscono altrettante tipologie di impasto molto diverse tra loro potenzialmente compatibili con le produzioni africane (Gruppi 4, 5 e 6) e locali (Gruppo 7), come supposto su base archeologica. Di seguito vengono descritte le peculiarità principali che contraddistinguono i singoli gruppi.

Anche tra i bacini (nmi 10) sono presenti manufatti africani: tre sono riferibili al tipo Bonifay Communes 31 (Fig. 3,7, CAR C5, Fig. 3,8; Bonifay 2004, 267-272); due del tipo Bonifay Communes 23 (2004, 265) presentano in particolare stesso impasto di un esemplare del tipo Bonifay Communes 38 prodotto nella regione del Sahel (CAR C1, Fig. 3,10; Bonifay 2004, 71, 275). Almeno due bacini caratterizzati da orlo a gancio (Caminneci, Franco e Galioto 2010, fig. 3,13) ed un altro con motivo ad onda inciso sotto l’orlo (Caminneci, Franco e Galioto 2010, fig. 3,2) sono invece riferibili al gruppo del campione VE-C13 di contrada Verdura. I vasi a listello (nmi 21) sono interamente di fabbrica africana: almeno dieci sono del tipo Bonifay Communes 13 (Bonifay 2004, 255-258), attestato sia nelle varianti A, B e C del V secolo (Caminneci , Franco e Galioto 2010, fig. 3,4-6), che nella variante D databile fino al VI secolo (Fig. 3,11); uno è affine al tipo Bonifay 14 (Fig. 3,12); sei sono gli esemplari del tipo Carthage Class I (Hayes 1976, 88-89; 1978, 68-69), nelle varianti B e C (Fig. 3,14 e 13), e Carthage Class IIB (Hayes 1976, 88-89; 1978, 68-69, 74) databili tra il VI e il VII secolo (Fig. 3,15). Oltre alla consistente porzione di reperti di origine africana, che suggerisce l’esistenza di stretti contatti tra la costa sud-occidentale della Sicilia e la regione di Nabeul almeno fino agli inizi del VII secolo, pare di un certo interesse l’individuazione di brocche, catini e bacini in impasti simili ad anfore e ceramiche comuni del vicino sito di contrada Verdura, per i quali le indagini archeometriche hanno evidenziato origine locale. Comincia di conseguenza a delinearsi il quadro riguardante la circolazione a livello territoriale dei prodotti locali, che attraverso la viabilità principale interessava anche i centri costieri, ricettori delle merci sia in entrata che in uscita dai latifondi. Per Agrigento ed il territorio limitrofo le ricerche evidenziano infatti una generale vitalità produttiva nel corso dell’età tardoantica. Ricordiamo a tal proposito le produzioni ceramiche del centro urbano di Agrigento e quelle di Campanaio di Montallegro (Bonacasa Carra 1995, 373; Wilson 2000, 361-362). A questi centri si aggiunge inoltre il sito recentemente individuato in contrada Giammaritaro di Aragona, nell’entroterra agrigentino (Parello, Amico e D’Angelo 2010, 286; Rizzo e Zambito 2010, 294).

Gruppo 1. Campione CAR 61/84 Impasto caratterizzato da abbondante scheletro smagrante (addensamento circa 30%) a granulometria eterogenea con prevalenza della frazione sabbioso grossolana e molto grossolana (0,5–2mm), costituito da minerali e frammenti di rocce vulcaniche peralcaline: feldspato alcalino (anortoclasio), subordinati plagioclasio, clinopirosseno (egirinaugite) enigmatite, ossidi opachi, rari granuli di ossidiana e pomice. Gruppo 2. Campione CAR 22/115 L’impasto è caratterizzato da scheletro smagrante mediamente abbondante (addensamento circa 20%) a granulometria bimodale con prevalenza della frazioni siltosa e sabbiosa molto grossolana (max 2,5mm). La frazione siltosa è costituita da clasti di quarzo, feldspati e mica (muscovite e biotite); quella sabbiosa da clasti angolosi di rocce metamorfiche e granuli monomineralici da esse derivate. I litoclasti metamorfici sono rappresentati da gneiss eteroblastici composti da quarzo, k-feldspato, biotite e muscovite; gneiss quarzoso muscovitici a grana minuta; micascisti a granato, biotite e muscovite. Tra i clasti monomineralici oltre a quarzo, plagioclasio e k-feldspato, biotite e muscovite si riscontra anche la presenza di orneblenda. La matrice ceramica ha l’aspetto di un’argilla calcarea a basso grado di cottura.

[A.A.] INDAGINI PETROGRAFICHE Le indagini petrografiche sono state effettuate su 10 campioni di ceramica da fuoco - CAR 22/115, CAR 61/84, CAR 125/1 (Fig. 2) e di ceramica comune - CAR C1, CAR C2, CAR C3, CAR C4, CAR C5, CAR C6, CAR C7 (Figg.4-5). Le indagini, condotte mediante osservazione di sezioni sottili al microscopio polarizzatore, hanno consentito di suddividere i reperti ceramici indagati in sette gruppi. I primi due gruppi (gruppo 1 e 2) sono rappresentati da campioni che hanno caratteristiche composizionali e tessiturali non compatibili con le potenziali materie prime disponibili in loco e, conseguentemente, sono da

Gruppo 3. Campioni CAR C4, CAR C6, CAR 125/1 Impasto caratterizzato da scheletro smagrante abbondante, con addensamento intorno a 25%, a granulometria bimodale, con prevalenza della frazioni 84

CAMINNECI-AMICO-GIANNICI-GIARRUSSO-MULONE

molto fine e medio - grossolana, massimo 1 mm. Entrambe le frazioni sono costituite da granuli di quarzo di forma angolosa nella frazione fine da subarrotondata ad arrotondata nella frazione grossolana. La matrice argillosa di tipo calcareo presenta abbondanti microfossili bentonici e planctonici, sporadici bioclasti e litoclasti calcarei parzialmente calcinati.

BIBLIOGRAFIA Arthur, P. 2007. Pots and boundaries. On cultural and economic areas between late antiquity and the early middle ages. In M. Bonifay and J.-C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry, 15-27. British Archaeological Reports, International Series 1662 (I). Oxford, BAR Publishing.

Gruppo 4. Campione CAR C3

Bonacasa Carra, R. M. 1995. La ceramica comune. In R. M. Bonacasa Carra (ed.), Agrigento. La necropoli paleocristiana sub divo, 141-190. Roma, L’Erma di Bretschneider.

Impasto povero di scheletro smagrante (addensamento circa 5%) a granulometria da fine a molto fine (max. 0,2mm) costituito da clasti angolosi di quarzo, raro feldspato e mica. L’impasto ceramico ha l’aspetto di un’argilla calcarea ricca in ossidi di ferro a medio grado di cottura, con numerosi pori da decomposizione termica di bioclasti, microfossili e rari calcinelli.

Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports, Int. Series 1301. Oxford, BAR Publishing. Cacciaguerra, G. 2010. La ceramica da fuoco nella Sicilia tardoantica e altomedievale: l’evidenza dell’area iblea orientale. In S. Menchelli, S. Santoro, G. Pasquinucci e G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. eastern Comparison between western and Mediterranean. British Archaeological Reports, International Series 2185 (I), 301-310. Oxford, BAR Publishing.

Gruppo 5. Campione CAR C1 Impasto caratterizzato da scheletro abbondante a granulometria prevalentemente compresa nell’intervallo della sabbia fine con sporadici granuli nella sabbia media (dimensione massima pari a 0,6mm), costituito da granuli di quarzo di forma prevalentemente da sub-angolosa a sub-arrotondata con qualche granulo relativamente più grossolano di forma arrotondata. Presenti anche rari granuli di feldspato (ortoclasio). La pasta di fondo ha l’aspetto di un’argilla calcarea sottoposta a basso grado di cottura. Con frequenza rara si riscontrano microfossili bentonici e planctonici con cavità interne riempite da pirite ematitizzata.

Caminneci, V. 2010. Tra il mare ed il fiume. Dinamiche insediative nella Sicilia occidentale in età tardoantica: il villaggio in contrada Carabollace (Sciacca, Agrigento, Sicilia, Italia). Folder 213, 1-16. Caminneci, V., Franco, C. e Galioto, G. 2010. L’insediamento tardoantico di contrada Carabollace (Sciacca): primi dati sui rinvenimenti ceramici. In LRCW3, 273-282.

Gruppo 6. Campione CAR C5 Impasto caratterizzato da scheletro abbondante, mal distribuito, a granulometria da fine a molto fine con qualche granulo di dimensioni medie e grossolane (massimo: 0,7mm) costituito prevalentemente da clasti da sub-angolosi a sub-arrotondati, di quarzo e rare quantità di feldspato. Presenti rari calcinelli. Matrice omogenea, di colorazione bruna, con aspetto di argilla poco calcarea sottoposta a medio grado di cottura.

Caminneci, V. e Franco, C. 2012. L'insediamento costiero di Carabollace e le relazioni commerciali della Sicilia occidentale con l'Africa in età tardo antica. In M. B. Cocco, A. Gavini, A. Ibba, L’Africa Romana XIX, Trasformazione dei paesaggi del potere nell’Africa settentrionale fino alla fine del mondo antico. Atti del Convegno di Studio, Sassari 15-19 dicembre 2010, 3041-3050. Roma. Castellana, G. e Mc Connell, B.E. 1990. A rural settlement of imperial and byzantine date in Contrada Saraceno near Agrigento, Sicily. American Journal of Archaeology 94, 1, 25-44.

Gruppo 7. Campione CAR C2, CAR C7 Impasto caratterizzato da scheletro mediamente abbondante a granulometria da fine a molto fine con dimensione massima pari a 0,2mm, costituito da granuli angolosi e sub-angolosi di quarzo e raro feldspato (ortoclasio). La pasta di fondo ha l’aspetto di un’argilla calcarea sottoposta a medio grado di cottura. In essa sono presenti numerosi calcinelli di dimensioni anche molto grossolane (fino a 2 mm) e numerosi resti di microfossili e bioclasti parzialmente calcinati.

Cau Ontiveros, M. A., Iliopoulos, I. e Montana, G. 2002. Pots and volcanoes: provenance of some Late Roman Cooking Wares in the western Mediterranean. Archaeometry 2000. México (cd rom interactive). Cracco Ruggini, L., 1980, La Sicilia tra Roma e Bisanzio. In G. Gabba e G. Vallet (eds.), Storia della Sicilia III, 1-96. Napoli.

[F. G., R. G., A. M.] Rilievi: Annalisa Amico, Manola Cotroneo Restauri: Anna Nativo, Francesco Termine. 85

LRCW4

De Salvo, L. 1997-1998. Negotiatores de Oriente venientes (V. Hilar. 25,8). Kokalos 43-44, 85-105.

Parello, M. C., Amico, A., Giannici, F., Giarrusso, R. e Mulone, A., in this volume. Ceramiche comuni e da fuoco di età tardo romana dai siti di Verdura (Sciacca) e Canalicchio (Calamonaci) (Agrigento, Sicilia, Italia).

Di Stefano, G. 2002. Il villaggio bizantino di Kaukana. Spazi urbani, monumenti pubblici ed edilizia privata. In Byzantino – Sicula IV, Atti del I congresso internazionale di archeologia della Sicilia bizantina, 173-190. Palermo.

Reynolds, P. 1995. Trade in the Western Mediterranean AD 400-700: the ceramic evidence. British Archaeological Reports, International Series 604. Oxford, BAR Publishing.

Fiertler, G. 2003. La Pantellerian Ware dal quartiere ellenistico-romano di Agrigento aspetti della problematica e proposta per una tipologia. In G. Fiorentini, M. Caltabiano, A. Calderone (eds.), Archeologia del Mediterraneo. Studi in onore di Ernesto De Miro, 321-337. Roma, L’Erma di Bretschneider.

Rizzo, M. S. e Zambito, L. 2010. Ceramiche comuni e anfore dal Villaggio tardoantico di Cignana (NaroAgrigento, Sicilia, Italia). In LRCW3, 293-300. Rizzo, M. S., Zambito, L., Giannici, F., Giarrusso, R. e Mulone A., in this volume. Anfore di tipo siciliano nel territorio di Agrigento.

Fulford, M. G. e Peacock, D. P. S. 1984. Excavations at Carthage: the British Mission, vol. I, 2. The Avenue du Président Habib Bourguiba, Salammbo: the pottery and other ceramic objects from the site, Sheffield.

Sami, D. 2005. La ceramica di Pantelleria. inquadramento tipologico e primi dati quantitativi dallo scavo subacqueo al porto di Scauri. Archeologia Medievale 32, 401-408.

Hayes, J. W. 1976. Pottery: Stratified Groups and Typology. In J.H. Humphrey (ed.), Excavations at Carthage 1975 conducted by the University of Michigan I, 47-123. Tunis.

Santoro, S. 2003. Ceramic and Society in Late Antiquity: Pantellerian Ware, a Case History. In V. Sermels, M. Maggotti (eds.), Ceramic in the society, Proceedings 6th European Meeting on ancient ceramics, 3-6 Ottobre 2001, Fribourg, Institute of Mineralogy and Petrography, 255-260. University of Fribourg.

Hayes, J. W. 1978. Pottery report-1976. In J. H. Humphrey (ed.), Excavations at Carthage, 1976, conducted by the University of Michigan IV, 23-98. Ann Arbor.

Santoro, S. 2007. Le ceramiche da cucina prodotte in Italia ed esportate nel Mediterraneo:un primo panorama archeometrico ed archeologico sulla base di una banca dati. In LRCW2, 365-371.

Long, L. e Volpe, G. 1998. Le chargement de l'épave de la Palud (VIe s.) à Port-Cros (Var). Note préliminaire. In M. Bonifay, M. B. Carre, Y. Rigoir, (eds.), Fouilles à Marseille. Les mobiliers (1er-VII siècle ap. J.C. Études Massaliètes 5, 317-342. Paris.

Santoro, S., Guiducci, G. e Tusa, S. (eds.), 2003. Pantellerian Ware, Archeologia Subacquea e ceramiche da fuoco a Pantelleria. Palermo.

Malfitana, D., Bonifay, M., e Capelli, C. 2008. Un progetto italo francese (CNR-CNRS) per lo studio delle importazioni di ceramiche africane nella Sicilia romana, vandala, ostrogota e bizantina. Problemi archeologici e archeometrici. Status quaestionis, metodologie e percorsi di indagine. Rivista di Archeologia 31, 227-235.

Uggeri, G. 1998. Relazioni tra Nord Africa e Sicilia in età vandalica. In M. Khanoussi, P. Ruggeri, C. Vismara (eds.), L’Africa romana XII, Atti del XII Convegno di studio, Olbia 12-15 dicembre 1996, 1457-1467. Università degli Studi di Sassari. Uggeri, G. 2004. La viabilità della Sicilia in età romana. Galatina.

Montana, G., Iliopoulos, I., Giarrusso, R. 2005. Pantellerian Ware: new data on petrography, chemistry and technological properties. In J. M. Gurt i Esparraguerra, J. B. Buxeda i Garrigos, M. A. Cau Ontiveros (eds.), LRCW1. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. British Archaeological Reports, International Series 1340, 425-436. Oxford, BAR Publishing. Panella, C., Saguì, L., Coletti, F. 2010. Contesti tardo antichi di Roma: una rilettura alla luce dei nuovi dati. In LRCW3, 57-78.

Volpe, G. 1996. Contadini, pastori e nell’Apulia tardoantica. Bari, Edipuglia.

mercanti

Volpe, G. 2002. Relitti e rotte commerciali nel Mediterraneo occidentale tardo antico. In M. Khanoussi, P. Ruggeri, C. Vismara, (eds.), L’Africa Romana XIV, Lo spazio marittimo del Mediterraneo occidentale, Atti del XIV Convegno di Studio, Sassari 7-10 dicembre 2000, 239-250. Roma. Wilson, R. J. A. 2000. Rural settlement in Hellenistic and Roman Sicily: Excavations at Campanaio (AG), 1994-98. Papers of British School at Rome LXVIII, 337-369.

Parello, M. C., Amico, A. e D’Angelo, F. 2010. L’insediamento alla foce del Verdura in territorio di Sciacca (Agrigento-Sicilia). In LRCW3, 283-291.

86

CAMINNECI-AMICO-GIANNICI-GIARRUSSO-MULONE

Fig. 1 Ceramica da fuoco (Scala 1:3)

Fig. 2 Microfotografie (nicols incrociati) dei campioni 22/115, 64/81, 125/1 (0.25 mm)

87

LRCW4

Fig. 3 Ceramica Comune. Brocche (1-4), catini (5-6), bacini (7-10), vasi a listello (11-15)

88

CAMINNECI-AMICO-GIANNICI-GIARRUSSO-MULONE

Fig. 4 Microfotografie (nicols incrociati) dei campioni CAR C1, CAR C2, CAR C3 (0.25 mm)

Fig. 5 Microfotografie (nicols incrociati) dei campioni CAR C4, CAR C5, CAR C6, CAR C7 (0.25 mm)

89

LRCW4

90

LATE ROMAN POTTERY FROM THE SIKYON SURVEY PROJECT: LOCAL PRODUCTION, IMPORTS, AND URBAN EVOLUTION (4TH - 7TH C. AD) (GREECE) ELLI TZAVELLA¹, CONOR TRAINOR², MATTHEW MAHER³ ¹University of Leiden; [email protected] ²Trinity College Dublin; [email protected] ³University of Winnipeg; [email protected] The Sikyon Survey Project (University of Thessaly, 2004-2013) is a fully integrated multi-disciplinary research programme designed to study the human presence and activity on the plateau of Ancient Sikyon. The main aim of the project is to address questions concerning the Hellenistic and Roman city of Sikyon. The intensity of the survey has allowed us to address a broad set of diachronic issues, including use of space, use of resources, hierarchy of habitation, technology, and trade. The Sikyon Survey Project employs an integrated approach to ceramic studies, including typological shape analysis and both macroscopic and petrographic fabric analysis. One of the most significant aspects of our dataset is the late Roman ceramics. This 4th-7th c. material appears to indicate an advance in local pottery technology, a greater variation in locally produced ceramic types, and a distinction between local / regional and imported ceramic products. Moreover, as these types occur commonly in the northeast Peloponnese, the refined typology of the Sikyon material will contribute to our understanding of the broader regional ceramic assemblages. Finally, this material has shed light on the evolution of the urban topography of ancient and post-antique Sikyon. KEYWORDS: SIKYON, CORINTH, ROMAN, LATE ANTIQUE, LATE ROMAN, EARLY BYZANTINE receive much attention in the historical records. It does appear, however, as a station in the Peutinger Table (Müller 1916, l. 581). A bishop of Sikyon, Hermogenes, is listed in the Council of Serdica (343/4) (Mansi, III, l. 46). The town is also attested among the civitates in the 6th century Synekdemos (Honigman 1939, 646.8). In the 13th century it reappears under the name ‘Vasilika’ in the Aragonese version of the Chronicle of the Morea (MorelFatio 1885, 49), which narrates the Frankish conquest of the Peloponnese. Fortunately, the sizable Roman and early Byzantine ceramic dataset collected during the course of the survey has provided a wealth of information which can help fill in these relatively unknown periods of settlement on the plateau.

Introduction The ancient city of Sikyon, located some 17km northwest of Corinth, stands on a distinctive plateau which rises from the coastal plain on the south shore of the Corinthian Gulf (Fig. 1). The commercial and strategic importance of its location cannot be overstated, as it was instrumental to the prosperity of the Hellenistic and the later Roman period settlements. Geographically, the plateau occupies a central position in the northern Peloponnese, commanding the coastal roads to Achaia and Elis towards the west, and straddling some of the most accessible and politically significant routes into the Arcadian mountains and the southern Peloponnese. Moreover, its location ensured access by sea to both the Corinthian Gulf and ultimately the western provinces beyond.

Survey Methodology

The limited historical record suggests that Sikyon grew in its new setting during the Hellenistic period and appears to have especially prospered during the 3rd century BC under the general Aratos (Skalet 1928; Griffin 1982). With the decline of the Achaian League and the subsequent rise of Rome, however, though the settlement survived for many centuries and habitation of some form or another on the plateau continued to the present day, Sikyon never regained its old splendour. Under the Early Roman Empire, for example, Sikyon lived in the shadow of the restored city of Corinth, the capital of the Roman province of Achaia. Besides Pausanias’ (2.5.6-2.11.2) visit in the second half of the 2nd c. AD, Sikyon does not

The Sikyon Survey Project is an intensive urban survey among the largest of its kind conducted in the Mediterranean. It is a fully integrated multidisciplinary research program designed to study the human presence and activity at ancient Sikyon. The survey began in 2004, with collection seasons between 2004 and 2008, and study seasons from 2006 to 2011. The goals of the on-going research are threefold: the primary aim is to produce a multidisciplinary study of the intra-mural area across the ages, and to trace human presence and activity from prehistoric times to the early modern era. The second and more broad-ranging aim is to 91

LRCW4

represent types dating to the Hellenistic, Early, and Middle Roman periods, have been recovered in the southern part of the South Plateau (Fig. 2). These include wasters of a local amphora (Fig. 5) which are typologically closest to the Brindisi type amphora (Peacock and Williams 1986, 82 [Class 1]; Palazzo 1989, 548-549; Bezeczky 2002; 2013, 110-114). The Sikyonian amphora type also closely resembles the Dressel 25 type, for which a Greek origin has been already proposed by Van der Werff (1986, 115-116). The main variation of Dressel 25 is thought to have been produced in the area of Guadalquivir, Spain, since it is considered as an antecedent of the Spanish amphora Dressel 20 (Peacock and Williams 1986, 134-136). Van der Werff (1986, 115116) proposed that the variation of Dressel 25 found in Ostia may have been produced in Southern Italy or Greece, an argument repeated by Bezeczky (2004, 87). Wine production was a significant component of ancient Sikyon’s economy (Lolos 2009, 120-122), and thus it seems probable that the Sikyonian amphorae once contained wine.

investigate the plateau in its context within the landscape and thus build upon the framework of the previous extensive survey of the territory of Sikyon, conducted between 1996 and 2002 (Lolos 2011). Finally, the project aims to set the standard for addressing the surface exploration and archaeology of sizable Mediterranean town centres. The methodology adopted for the survey corresponds to both the aforementioned research goals as well as the topography of the study area (for a detailed discussion of the methodology employed see Lolos et al. 2007). The plateau has a surface area of approximately 250 hectares, 180 of which are available for survey, and is divided into three sections: the Upper Plateau, or the acropolis which rises west of the theater, the Northern Plateau, and the Southern Plateau, the latter two being separated by the main road of the modern village Vasiliko (see Fig. 2). In each of these three areas, the modern field boundaries were plotted within the larger digital Sikyon map; these modern fields define the tracts which the teams actually walk. Each individual tract was then divided into squares of more or less standard manageable size of 20 x 20m. These were walked by teams of five members who walked in parallel lines approximately 4m apart. Each team member counted all the sherds and tiles along his/her line while having collected the ‘diagnostic’ pottery at a representative ratio. In every fifth square of a given tract, to get a representative sample, every artifact was collected for study.

By the late Roman period, here categorized as the 4th to 5th century, pottery production on the plateau appears to have ceased, moved, or at least changed very significantly in scale. For example, we no longer see a high concentration of ceramic wasters, and the single identifiable waster which may be late Roman does not come from the earlier attested production area, but rather from the nucleated late Roman settlement around the area of the agora (see Fig. 2). The extremely homogenous appearance of earlier Sikyonian ceramic fabrics also changes in late Roman times. On the whole ceramic fabrics become much harder-fired and the once most prominent matte white inclusions now appear to be replaced by angular, glassy, and milky inclusions.

In the five seasons of intensive surface reconnaissance a total of 2839 (20 x 20m) squares, totaling 114 hectares, were covered (Fig. 2). The intensity of the survey has allowed us to present a highly nuanced picture of the ancient city as an ever-evolving entity with shifting patterns of economy, technology, and delineation of civic space.

During the early Byzantine period (6th – 7th centuries, we can see yet another shift in pottery production at Sikyon. At this time, cookwares appear in a hard-fired, dark grey fabric, which has been blackened throughout, possibly as a result of being fired in a reducing atmosphere – a development also noted in the late Roman cookwares from the Nikopolis Survey (Moore 2001, 84.). Upon initial macroscopic inspection, the fabric of these vessels is much more reminiscent of Early Roman cookware production at Sikyon and exhibit frequent matte white inclusions. A similar trend has been observed in survey material of other areas of the Mediterranean which show uninterrupted urban activity from the Roman to the early Byzantine period, where wheelmade wares of the 6th and 7th centuries are characterized by hard firing (Hayes 2000, 106-7). Indeed, two early Byzantine stewpot wasters recovered from the Sikyon agora area (Fig. 2) also indicate continued cookware production on the plateau, albeit in a small scale (Trainor 2013).

Ceramic production in Roman Sikyon Sikyon is geologically more or less identical to Corinth, and indeed much of the rest of the coastal northeast Peloponnese (Fig. 3). In Hellenistic and Roman times, Sikyonian pottery was produced in a silicate fabric which contains inclusions of chert, monocrystalline and polycrystalline quartz, occasional lumps of micritic limestone, as well as some dark red to black clay pellets, and rarely mudstone fragments. This mineralogical composition is consistent with the local geology that is made up predominantly of Pliocene marls and carbonaterich Pleistocene sediments (Tataris et al. 1970, Hayward 2003). Ceramic fabric analysis and the presence of large quantities of ceramic kiln wasters within the Sikyon survey area have allowed us to determine that the vast majority of recovered ceramics were produced locally (Fig. 4). Approximately 90% of the approximately 90,000 sherds that were included in our program of ceramic fabric analysis appear to have been produced locally in a silicate fabric (Lolos et al. 2012). Ceramic wasters, which

As the Roman and Byzantine ceramics from Sikyon had never received the attention of earlier scholars, the survey team had to recognize and understand the typological and technological range of local ceramic production. 92

TZAVELLA-TRAINOR-MAHER

these vessels is Reddish-Yellow (5YR 6/8) / YellowishRed (5YR 5/8), followed by Light Red (2.5YR 6/8-6/6), while less common colours occur in the ranges of Pale Brown (10YR 6/3) to Brown (10YR 5/3) and rarely, Grey (2.5YR 5/1) to Dark Grey (2.5YR 4/1) (colours taken from Munsell Soil Chart). The inclusions which feature most prominently in this fabric are clear and milky angular ones, as well as glassy red, small dark glassy inclusions, small dark red ones, and finally rare silver glitter (possibly mica) may also occur. Compared to earlier pottery from Sikyon, although this is not a typical local fabric, the inclusions do appear to be consistent with the local/regional geology. We postulate that these table amphorae or pitchers may represent a new local form, but further research is required in order to determine this with more certainty. Finally, some of these amphora rims bear a distinctive groove on the upper surface (Fig. 10). Two similar amphorae, bearing a rim with a groove on the upper surface, have been found in Corinth, where one of them is considered to be ‘regional’ (Slane and Sanders 2005, 255, No. 1-26; 264, No. 2-30).

Amphorae The late Roman and early Byzantine assemblage at Sikyon shows a striking quantitative predominance of cookware and utility shapes (like basins and large bowls) over amphorae. This is in contrast with the earlier material, where amphora and cookware numbers are roughly equal. Consequently, the late Roman amphorae provide limited evidence for the history of the settlement. Some of the local amphorae are imitations of the LRA 2, which appear in a local or regional silicate fabric (Fig. 6). These pieces range in colour from Gray (5Y 5/1) – Dark Gray (2.5Y 4/1) to Reddish Yellow (5YR 6/8) – Yellowish Red (5YR 5/8), and exhibit the angular, opaque glassy and milky coloured inclusions, as well as occasional glassy red ones. A further LRA 2 rim (from NP 50.15), made of cooking fabric, also occurs which may have a Corinthian origin (Marc Hammond, pers. comm., based on the petrographic analysis of an identical rim fragment from Corinth). A plain conical rim with carinated outer surface (from NP 103.03), also made in a local or regional cooking fabric, has a parallel in a late 4th century context from the theatre at Sparta (the profile closely resembles that published by Pickersgill and Roberts 2003, 570, No. 59, Fig. 10).

Pitchers Late Roman pitchers typically bear a ring-rim which is semi-cylindrical in section (Fig. 11). These appear at Argos in 4th-5th century contexts (Abadie-Reynal 2007, 230, No. 399, Pl. 63; Aupert 1980, 428, Nos. 221-22, Fig. 40), while their later versions, with an elongated rim and a sturdy D-shaped handle, are typical of 6th-7th century contexts in Corinth (Slane and Sanders 2005, 272, No. 343, Fig. 9) and Berbati (Hjohlman 2005, 139, No. 14; 152, No. 62; 221, No. 258). 4th-5th century pitchers of this type found in Sikyon are often made in a light coloured, buff, semi-fine fabric, while the later versions appear in a typically dark grey cooking fabric. Smaller jugs bear a finely made, funnel-shaped rim, stepped on the outer surface (Fig. 11) – these are common at Argos (Abadie-Reynal 2007, 232, No. 407, Pl. 64), Corinth (Slane 1990, No. 218 [but in a larger version]), Sparta (Pickersgill and Roberts 2003, 570, No. 58, Fig. 10), and elsewhere in 4th and 5th century contexts.

Imported amphorae of this period include mainly LRA 2 body fragments, most of which are made of the Southern Argolid fabric (on which see Rudolph 1979; Slane and Sanders 2005, 286f, note 65). Examples of Corinthian versions of the LRA 2 amphora have already been published (Slane and Sanders 2005, 287), and point to a regionalisation of amphora production in the 6th century. It is noteworthy, however, that the slim versions of the LRA 2 amphora produced in Corinthian cooking fabric, which are typical of Corinthian 6th and 7th century contexts and which appear also in Athens, are absent from Sikyon (on these amphorae, see Slane and Sanders 2005, 287 for Corinth, and Robinson 1959, 115, M325 for Athens). Only two fragments of LRA 1 amphorae were found (Fig. 7, SP 97.05). Two examples of an unidentified imported amphora with a vertical rim appear in a semi-fine reddish-yellow/light yellow non-local powdery fabric (Fig. 7, NP 80.02). An amphora with a similar rim has been found in Corinth (Slane and Sanders 2005, 278, No. 4-23, Fig. 11). In general, the amount of late Roman and early Byzantine amphorae imported from abroad is surprisingly small for a settlement which lies only ca. 4km from the coast.

Cookware A typical form of late Roman stewpot has an upper body with straight walls and an oblique rectangular or square rim with a pronounced ridge on the outer surface (Fig. 12). The same shape of rim is also attested on casseroles. This type of stewpot appears in Otranto and Olympia in the 3rd century (Walter 1958, 61, Fig. 50a; 66, Fig. 54), in Butrint in 3rd-4th century contexts (Reynolds 2004, 227, No. 75, Fig. 13; 250, No. 317, Fig. 13), in Sparta in a late 4th century context (Pickersgill and Roberts 2003, 570, No. 55, Fig. 9), and in Corinth it turns up as an import in a 5th century context (Slane and Sanders 2005, 256, No. 1-34, 1-35; Slane 2008, 240). In less securely dated contexts, this type also appears at Saranda (Budina 1971, 288, Pl. III.4), Dürres (Guiducci and Montana 2007, 542, No. 15, Fig. 2), in the Nikopolis survey

The largest group of late Roman amphorae found in Sikyon is a questionable one (Figs. 8 and 9). These have rims with a small diameter (7.5-10cm) and therefore likely belong to table amphorae or even large pitchers. Most of these rims present formal similarities with some well-known African amphorae forms, such as Keay 57 and sub-types of Keay 61 and 62 (Keay 1984, 298-299 [Type 57]; 303-309 [Type 61]; 309-350 [Type 62]. Bonifay 2004, 135 [Type 57], 137-141 [Types 61-62]). Their diameter, however, is far too small for them to be identified as such. The most common colour range for 93

LRCW4

65.01) adds further weight to the hypothesis for smallscale production of cookwares on the plateau.

(Moore 2001, 84, Fig. 6.1-2), in Elis (Mitsopoulou-Leon 1972-75, 223, Fig. 35), and in Berbati (Hjohlman 2005, 209 and 211, No. 222, Fig. 75). In Benghazi it was attributed an Aegean origin, while it occurs also in Spain (Reynolds 1995, 102). The high numbers and varied profiles in which it appears at Sikyon suggest that it was probably in use during most of the late Roman period, and possibly as early as the 3rd century. Moreover, at Sikyon this stewpot appears in several fabric variations, which may indicate different locations and/or periods of production. The most common variation is a ReddishYellow (5YR6/6-6/8) / Yellowish-Red (5YR5/8) medium coarse to coarse fabric that exhibits frequent opaque glassy inclusions, some evidence of degraded limestone, as well as some glassy red inclusions. Although this fabric is not identical to our earlier local fabrics, it is our current hypothesis that it is of local or regional origin owing to both its geological composition and the high numbers of these vessels recovered during the survey. A few other sherds of this type of stewpot exhibit rare, green glassy inclusions, which may suggest non-local production. K. Slane has suggested that the origin of stewpots of this shape that appear in Corinth is in Epirus or Elis (Slane 2008, 240).

Stewpots of the 7th century appear in high numbers, demonstrating that occupation on the Sikyon plateau continued during this century (Fig. 14). They belong to types known mostly from Corinth, and thus conform to the pattern of regionalised production and circulation. The stewpot with convex rim (Fig. 13) evolves to a type with curved rim with pointed edge (Fig. 14, SP 93.01) during the first third of the 7th century (cf. Hjohlman 2005, 149, No. 42); the same type later has a longer, undercut rim (Fig. 14, NP 65.07) (Guy Sanders, pers. comm.). Finally, stewpots with a plain vertical rim (Fig. 14, SP 76E.i.c., SP 88.01) are paralleled in 7th century Corinthian assemblages (Slane and Sanders 2005, 279, No. 4-27). All of these later types listed above occur in the Sikyonian assemblages. Bowls Bowls of the 4th century have a heavy horizontal rim, ridged on the upper surface, and the inner edge of the rim sometimes bears piecrust decoration (Fig. 15). These are large bowls, with diameters of 25-37cm. They appear in two main fabric groups: the first in a darker grey fabric with orange core that exhibits large angular, opaque glassy inclusions, some matte white, occasional small glassy red inclusions, and rare tiny glitter. This fabric looks very similar to that of the local Sikyonian material. The second group has a reddish yellow colour and contains large angular glassy and matte white inclusions, as well as tiny dull powdery red ones – this second type could be regional or imported from further away. Parallels have been found at Corinth (Slane 1990, 126, No. 274, Fig. 33) and Sparta (Pickersgill and Roberts 2003, 570, No. 50, Fig. 9) in later 4th century contexts.

An imported late Roman casserole (Fig. 12, SP 70.05) bears a flange on the upper wall and is made of very hard, gritty, dark grey fabric, with tiny silver glitter. Parallels have been published from Argos, Sarachane and elsewhere, and date to the late 3rd and 4th centuries (Abadie-Reynal 2007, 221-22, No. 374, Pl. 59; Hayes 1992, 91-2, Deposit 2, No. 1; Deposit 4, No. 6 and 151, Fig. 30). By the 6th century, the stewpot with the oblique ridged rim has been replaced by a pot of similar shape and size, but with a non-ridged rim (Fig. 13, SP 76E.i.c.) (Hjohlman 2005, 170, No. 96, Fig. 39; Williams and Zervos 1983, 29, No. 79, Fig. 14). A single waster rim of this type was found in the survey (Fig. 13, SP 76E.i.c.), just north of the early Byzantine settlement area, suggesting that small-scale production of cooking pots may have taken place there. The fabric survives relatively intact and includes primarily matte white and opaque glassy inclusions.

Basins and storage jars Basins occur in very high numbers in late Roman and early Byzantine period Sikyon. They therefore continue a tradition of earlier Roman times, and corroborate the evidence from Corinth, where basins also represent a common category in late Roman - early Byzantine households (Slane and Sanders 2005, 288). Basins used at Sikyon during this period have a distinct in-turned upper part of the body and a heavy rim which folds over horizontally (Fig. 16, NP 105.01, SP 93.02). The inturned upper part of the body allows the hypothesis that they may have been used additionally as storage jars. The closest parallels known to us have been found at Corinth (Slane and Sanders 2005, 257, No. 1-42) and Berbati (Hjohlman 2005, 209, No. 218 is a larger version, thus characterized as a ‘pithos’). At Sikyon these forms of basin rims and variations between them, occur in high numbers in all assemblages which contain ceramics of the 5th to the 7th centuries, thus suggesting that they were in use for most of this period. A typological evolution with time is probable, but to establish such a sequence excavated stratified parallels are required. These basins

The type of stewpot that prevails in the 6th century has a similarly deep body and a rim with a convex outer surface, which is sometimes undercut (Fig. 13, SP 88.02, SP 93.02, SP 88.01) – this is a very common shape at Corinth (Sanders 1999, 472, No. 18), Argos (Aupert 1980, 433, Nos. 269, 285, Fig. 43), Berbati (Hjohlman 2005, 132, No. 8; 223, Nos. 265-67), and is also found as an import at Butrint (Reynolds 2004, 235, Nos. 211-12, Fig. 13). At Sikyon this type appears in two fabrics, one in our local early Byzantine cooking fabric, and the other in a finer grey-beige fabric with common angular opaque glassy inclusions, silver glitter, and tiny black inclusions. A very similar stewpot with a long flat oblique rim, with later 6th century parallels from Corinth is also common (Fig. 13, SP 88.01, below) (Slane and Sanders 2005, 272, Nos. 3-30 and 3-31). A waster of this type (Fig. 13, NP

94

TZAVELLA-TRAINOR-MAHER

occur in a local (or possibly regional) well-fired fabric, are Reddish-Yellow (5YR 6/8) / Yellowish- Red (5YR 5/8) and contain angular opaque glassy, matte white, and small glassy red inclusions.

habitation, but they do not disappear entirely. At the same time, the areas located to the north and the south of the agora demonstrate consistently significantly higher ceramic densities than in the earlier centuries.

It is noteworthy that the late Roman basins which appear commonly both in Corinth and Argos are not typical of Sikyonian assemblages; specifically, the easily distinguishable basins with hooked grooved rims (Slane and Sanders 2005, 256, No. 1-39) and those with hammerhead rims, common in Corinth (Slane and Sanders 2005, 265, Nos. 2-44 and 2-45) and Berbati (Hjohlman 2005, 190, Nos. 171-72; 205, No. 202), occur in Sikyon only in one instance each, in Corinthian fabric (Fig. 16, SP 75.12). Basins with rolled rims (Sanders 1999, 472, Nos. 15 and 16) are also not very common (Fig. 16, NP 76.03).

The process of settlement nucleation continues into the early Byzantine period, when the radius of the inhabited area becomes smaller than in the late Roman period, but is still centered around and to the east of the agora. Several rescue excavations have been conducted by the 37th Ephorate of Prehistoric and Classical Antiquities in the modern village, but only a few of them have uncovered early Byzantine material. It is likely that the early Byzantine settlement extended into part of the area occupied by the modern village, which was necessarily excluded from the survey. This settlement pattern is confirmed by the distribution of the finewares. The vast majority of the late Roman and early Byzantine imported finewares – that is, African Red Slip and Late Roman C ware – were found in the area immediately east of the agora towards the outskirts of the modern village.

Settlement pattern The late Roman and early Byzantine periods are generally characterized by long distance trade and wide circulation of amphorae and finewares, and yet at Sikyon, we see only a few examples of imported wares. Indeed, among the approximately 3000 diagnostic sherds of the late Roman and early Byzantine periods, only 20 belong to imported finewares. Since local tableware, cookware, and utility vessels appear in high numbers until the 7th century – thus demonstrating habitation continuity until then – the near absence of imported ceramics between the 5th and 7th centuries is probably to be explained through a shift in the character of the settlement. (The small number of fineware sherds collected is also certainly a reflection of preservation conditions on the ground). Regionalization of ceramic production is characteristic of 6th and 7th century assemblages at major sites, as recent studies have shown (e.g., Slane and Sanders 2005, 287), and is therefore to be expected at Sikyon as well. There are two further indications, however, which lead us to believe that the settlement character changed during the late Roman period: the predominance of cookware and utility vessels (e.g., basins and large bowls) over amphorae (as discussed above), and the nucleation of the inhabited area of the plateau between the 3rd and 7th centuries. Concerning the latter, the distribution of ceramic wares across the plateau offers evidence for a changing settlement pattern from the Hellenistic to the early Byzantine period. Although this is another area of on-going research, an emerging picture can already be described.

The ceramic evidence therefore suggests that the settlement on the plateau changed character in the late Roman and especially the early Byzantine period. It should not be necessarily characterized as a small village or hamlet, since the early Byzantine pottery still covers a relatively broad area and appears in considerable quantities. Moreover, the plateau is known to have housed three early Christian basilicas: one on the ruins of the Archaic temple in the Agora (Philadelpheus 1926, 47; Krystalli-Votsi 1984, 242), and the other to its northeast, traced through geophysical research. Evidence for a third one offer architectural ruins and two marble members, a screen slab and a column capital, which were discovered at the Agia Triada church in the centre of the modern village (Philadelpheus 1926, παράρτ. 23, Nos. 16-17). Furthermore, a tower built of spolia, which appears to belong to a late Roman or early Byzantine fort, is preserved at the south-east outskirt of the plateau. Despite evidence for continued occupation on the plateau, it appears that the main centre of habitation significantly changed in character around this time. It appears to have shrunk, nucleating around the Agora in the main, with another possible locus underlying part of the modern village of Vassiliko. The creation of a new settlement core in the coastal area of Sikyon, beneath modern Kiato, is also an attractive possibility. A large and lavish basilica, dated to around AD 500 (Orlandos 1954), has led some scholars to believe that this was the location of the main settlement of the Sikyonia during the early Byzantine period. The latter hypothesis was strengthened by the critical edition of Synekdemos, which lists Sikyon as ‘Nea Sikyon’ (Honigman 1939, 18, 646.8, column II). Some scholars (e.g. Avramea 1997, 233, 351) used this literary testimony in order to support a relocation of the city. Yet, the newest editor, Honigman (1939, 7 and 18, comment on 646.8a), believes that ‘Nea’ is a false copy of ‘Nemea’ and that the name of Sikyon remained unchanged.

In Early Roman times, for example, habitation and industrial activity appear to be spread widely over the plateau, following the pattern which predominated during Hellenistic times. The habitation pattern during Middle Roman times differs only slightly from the preceding period in that it appears more concentrated in the North Plateau. The late Roman period, however, shows a distinctive concentration of activity in the central part of the plateau in the area around the agora. Ceramic find densities become lower in the north and northeast part of the North Plateau, which was traditionally used for 95

LRCW4

It would appear that archaeological research has a crucial role to play in the clarification of these topographical issues. The ongoing excavation in the Agora area which started in 2013 by Y. Lolos (University of Thessaly), as well as future excavations in and around the modern town of Kiato will almost certainly help shed some light on the settlement history of Sikyon both in ancient and medieval times.

Bezecky, T. 2013. The amphorae of Roman Ephesus, Forschungen in Ephesos vol. 15/1. Vienna, Österreichische Akademie der Wissenschaften.

Acknowledgements

Budina, D. 1971. La carte archéologique de la cote jonienne et du bassin de Delvine. Iliria I, 276-342.

Institute at Athens 5, 85-98. Århus, Aarus University Press.

Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports, International Series 1301. Oxford, BAR Publishing.

The Sikyon Survey Project was conducted under the auspices of the University of Thessaly under the direction of Dr Yannis Lolos, in collaboration with the University of York and the Institute of Mediterranean Studies. We would like to thank the 37th Ephorate of Prehistoric and Classical Antiquities and the 25th Ephorate of Byzantine Antiquities for issuing study permissions and for their continued support. We would also like to thank the staff of the Sikyon Survey team, Dr Yannis Lolos, Ben Gourley, and Dr Daniel Richard Stewart; Daniel Stewart kindly offered valuable advice prior to publication of the present article. Dr Kathleen Slane and Dr Guy Sanders generously offered useful advice. Dr Ioulia TzonouHerbst enabled our repeated visits in the Corinth Museum Collections. Dr Michel Bonifay kindly offered advice in regard to some late Roman amphora shapes. Pottery drawings were made by Anastasia Douka and Christina Kolb. Finally, we wish to thank the conference organizers, Professor A. Mentzos, Dr. N. PoulouPapadimitriou, Dr. V. Kilikoglou, and the anonymous reviewers for their helpful and insightful comments and suggestions.

Griffin, A. 1982. Sikyon. Oxford. Guiducci, G. and Montana, G. 2007. Ceramiche da fuoco tardoantiche da Dyrrachion (Albania). In M. Bonifay and J-C Tréglia (eds.), LRCW2. Late Roman coarse wares, cooking wares and amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports, International Series1662 (II), 541-549. Oxford, BAR Publishing. Hayes, J. W. 1992. Excavations at Saraçhane in Istanbul, vol. 2: The Pottery. Princeton N.J., Princeton University Press. Hayes, J. W. 2000. Roman Ceramic Studies in Mediterranean Survey. In R. Francovich and H. Patterson (eds.), Extracting Meaning from Ploughsoil Assemblages. The Archaeology of Mediterranean Landscapes 5, 105-109. Oxford, Oxbow. Hayward, C. 2003. The geology of Corinth: study of a basic resource. In C.K. Williams and N. Bookidis (eds.), Corinth XX, The Centenary, 1896-1996, 15-42. Princeton. Hjohlman, J., Penttinen A., Wells, B. 2005. Pyrgouthi. A rural site in the Berbati valley from the early Iron Age to Late Antiquity. Excavations by the Swedish Institute at Athens, 1995 and 1997. Stockholm.

BIBLIOGRAPHY Abadie-Reynal, C. 2007. La céramique Romaine d'Argos (fin du IIe siècle avant J.-C. - fin du IVe siècle après J.-C.). Études Péloponnésiennes 13. Athens.

Honigmann, E. (ed.) 1939. Le Synekdèmos d'Hiéroklès et l'Opuscule géographique de Georges de Chypre: Texte, introduction, commentaire et cartes. Bruxelles.

Aupert, P. 1980. Objets de la vie quotidienne à Argos. Bulletin de Correspondance Hellénique Supplément VI, Études Argiennes, 395-458. Paris.

Keay, S. 1984. Late Roman Amphorae in the Western Mediterranean: A Typology and Economic Study. The Catalan Evidence. British Archaeological Reports, International Series 196. Oxford, BAR Publishing.

Avramea, A. 1997. Le Péloponnèse du IVe au VIIIe siècle: changements et persistances. Paris, Publications de la Sorbonne.

Krystalli-Votsi, K. 1984. Ανασκαφή Σικυώνος. Praktika tes en Athinais Archaeologikes Etaireias 1984, 239240.

Bezeczky, T. 2002. Brindisian olive oil and wine in Ephesus. In L. Rivet and M. Sciallano (eds.), Vivre, produire et échanger: reflets Méditerranéens, Mélanges offerts à Bernard Liou. Archéologie et Histoire Romaine 8, 355-358. Montagnac, Monique Mergoil.

Lolos, Y. 2009. Οινοπαραγωγή στην Κορινθία κατά την Αρχαιότητα (Κόρινθος – Σικυών – Κλεωναί). In G. Pikoulas (ed.), Πολυστάφυλος Πελοπόννησος, Οἶνον Ἱστορῶ IX, 115-132, Athens.

Bezeczky, T. 2004. Early Roman food import in Ephesus: Amphorae from the Tetragonos Agora. In J. Eiring and J. Lund (eds.), Transport Amphorae and Trade in the Eastern Mediterranean, Acts of the International Colloquium at the Danish Institute at Athens, September 26-29 2002, Monographs of the Danish

Lolos, Y. 2011. Land of Sikyon: The Archaeology and History of a Greek City-State. Hesperia Supplement 39. Princeton. Lolos, Y., Gourley, B., Stewart D. 2007. The Sikyon Survey Project: A Blueprint for Urban Survey? Journal of Mediterranean Archaeology 20, 267-296. 96

TZAVELLA-TRAINOR-MAHER

Lolos, Y., B. Gourley, A. Sarris, C. Hayward, C. Trainor, E. Kiriatzi and N. Papadopoulos. 2012. Surveying the Sikyonian Plateau. Integrated Approach to the Study of an Ancient Cityscape. In N. Zacharias, M. Georgakopoulou, K. Polikreti, G. Fakorellis and Th. Vakoulis (eds.), Proceedings of the 5th Symposium of the Hellenic Society for Archaeometry, 305-326. Athens.

Reynolds, P. 1995. Trade in the Western Mediterranean, AD 400-700. The Ceramic Evidence. British Archaeological Reports, International Series 604. Oxford, BAR Publishing. Reynolds, P. 2004. Catalogue of Roman ceramics and selected medieval pottery from Butrint 1994-99. In R. Hodges, W. Bowden and K. Lako (eds.), Byzantine Butrint. Excavations and Survey 1994-1999, 327-395. Oxford, Oxbow.

Mansi, J. D. 1692-1769 (re-published 1960-61). Sacrorum conciliorum nova et amplissima collection. Graz, Akademische Druck- und Verlagsanstalt.

Rudolph, W. 1979. Excavations at Porto Cheli and vicinity. Perliminary report V: the Early Byzantine remains. Hesperia 48, 294-324.

Mitsopoulou-Leon V. 1972-1975. Zehnter vorlaeufiger Bericht ueber die Grabungen in Alt-Elis. Jahreshefte des Oesterreichischen Archaeologischen Instituts 50, Beiblatt (181-223), 223, Abb. 35.

Skalet, C. H. 1928. Ancient Sicyon, prosopographia Sicyonia. Baltimore.

a

Slane, K. 1990. Corinth XVIII, ii. The Sanctuary of Demeter and Kore: The Roman Pottery and Roman Lamps. Princeton.

Moore, M. 2001. Roman and Late Antique Ceramics of Southern Epirus: Results of the Nikopolis Survey Project. In J. Isager (ed.), Foundation and Destruction: Nikopolis and Northwestern Greece. Monographs of the Danish Institute at Athens.

Slane, K W. 2008. Corinth’s Trade with the Adriatic. Rei Cretariae Romanae Fautorum Acta 40, 237-241.

Morel-Fatio, A. 1885. Chronique de Morée. Geneva.

Slane, K. and G. D. R. Sanders. 2005. Corinth: Late Roman Horizons. Hesperia 74, 243-29.

Müller, K. (ed.), Itineraria romana: Römische Reisewege an der Hand der Tabula Peutingeriana dargestellt. Stuttgart.

Tataris, A. N. Mouragoudakis, G. Kounis, G. Christodoulou and St. Tsalia-Monopoli. 1970. Geological Map of Greece, 1:50 000 Nemea Sheet. IGSR, Athens.

Orlandos, A. 1954. Ανασκαφή Σικυώνος. Praktika Archaeologikes Etaireias 1954, 219-231.

Trainor, C. P. 2013. Ceramics, Clays, and the Technological Landscape of Urban Sikyon: 2nd century BC-7th century AD. PhD Thesis, Trinity College, Dublin.

Palazzo, P. 1989. Le anfore di Apani (Brindisi). In Amphores romaines et histoire économique: dix ans de recherche, 548-549. Rome, École Française de Rome.

Van der Werff, J. H. 1986. III. The amphora wall in the house of the porch, Ostia. In J. Boersma, D. Yntema and J. H. Van der Werff (eds.), Excavations in the House of the Porch (Vii.4-5) at Ostia. Bulletin Antieke Beschaving 61, 77-137.

Peacock, D. and Williams, D. 1986. Amphorae and the Roman Economy: an Introductory Guide. London, Longman. Philadelpheus, Α. 1926. Ανασκαφαί Archaiologikon Deltion 10, 46-50.

with

Σικυώνος,

Walter H. 1958. Der archaeologische Befund. In E. Kunze (ed.), VI Bericht über die Ausgrabungen in Olympia, 41-67. Berlin.

Pickersgill, C. and Roberts, P. 2003. Roman Pottery from the Sparta Theatre and Stoa. The Annual of the British School at Athens 98, 549-597.

Williams, C. K. II and Zervos, O. H. 1983. Corinth, 1983: East of Theater. Hesperia 52, 1-47.

97

LRCW4

98

TZAVELLA-TRAINOR-MAHER

99

LRCW4

100

TZAVELLA-TRAINOR-MAHER

101

LRCW4

102

PRODUCTION CENTERS

NOUVEAUX TEMOIGNAGES SUR LA PRODUCTION DE LA CERAMIQUE ANTIQUE DU SAHEL TUNISIEN JIHEN NACEF Université de Monastir Institut Supérieur des Etudes appliquées aux Humanités de Mahdia, Tunisie Membre de l’Unité de recherche sur la céramique antique du Sahel tunisien [email protected] In this paper we examine the ceramic production of two late Roman pottery workshops; Ghar draj and Bir El Hammam situated in Byzacena. The ancient site Ghar Draj was progressively devastated by modern construction. The main material collected is composed of amphora sherds, kitchen ware and coarse ware. Amphorae types comprise Africana I, Africana II and transitive type for late amphorae. We should note that the pattern of pottery production of Ghar Draj is similar to that encountered in the Byzacena coastal sites. The second site Bir El Hammam was identified in the late 1990s. However, during that period research had focused on the remarkable bath structures that preceded chronologically the late Roman pottery workshop. The ceramics produced at the site included a group of late amphorae, mainly Keay 62 with a large number of variants, Keay 61A and Keay 8A. The coarse ware comprised basins, bowls and closed forms. The workshop of Bir El Hammam indicates the economic vitality of the region just before the Arab conquest. It is the ultimate material evidence of human occupation in the end of 7th or probably the early 8th centuries. This site provides new information on the origin of a later African amphora type, Keay 8A. KEYWORDS: AFRICAN AMPHORAE, LATE AMPHORA, COARSE WARE, RURAL WORKSHOPS, COASTAL WORKSHOPS BYZACENA. céramique et quelques poches de cendres. L’essentiel du matériel ramassé se compose de fragments d’amphores, de tessons de vaisselle culinaire et de vaisselle commune. Les amphores identifiées se rapportent à des Africaines classiques et au type de transition avec le type Keay 25. Le type Africaine I est représenté surtout par des variantes tardives alors que les variantes précoces sont quasiment absentes. C’est la variante B (Panella 1973, 629, fig. 26) qui est la mieux représentée avec son profil de bord semi-circulaire. La surface interne est marquée par un gradin (Fig. 2, n° 2). Plusieurs tessons se rattachent à une variante tardive qui se caractérise par un profil de bord plus haut et moins arrondi. Un autre tesson présente un profil plus convexe et dont la section est crochue. La face interne comporte deux ressauts (Fig. 2, n° 3). Le deuxième type, Africaine IID, est attesté par deux variantes. La première présente un bord haut avec le sommet de la lèvre arrondi ou légèrement biseauté (Fig. 2, n° 4). Le col est tronconique. Un fragment comporte un timbre incomplet placé sur le col et inséré dans un cartouche en creux (Fig. 2, n° 5). Les lettres sont aussi en creux: [---] RI. La deuxième variante correspond à la variante II D2 de Bonifay (2004, 115). La séparation entre le bord et le col est moins marquée que dans la variante précédente (Fig. 2, n° 6). Enfin un dernier tesson est peu commun, avec un sommet de la lèvre aplati et la face interne du bord rectiligne. Apparemment, le col est extrêmement évasé (Fig. 2, n°7). Le troisième type est un type de transition entre Africaine IID et Keay 25. Il s’agit

La connaissance des céramiques d’Antiquité tardive des côtes orientales de la Byzacène a observé une progression rapide ces dernières décennies. Les prospections réalisées dans le cadre de l’unité de recherches sur la céramique antique du Sahel tunisien ont révélé de nouveaux ateliers de potiers. Le présent travail est un rapport préliminaire sur une production de céramique tardive dans la province de Byzacène fortement ponctuée d’ateliers, notamment à Leptiminus, à Moknine, dans la région de Salakta et de Ksour Essef. Deux ateliers, Ghar Draj et Bir El Hammem, ont été identifiés sur la commune de Teboulba, à mi-chemin entre deux cités antiques importantes de Byzacène, à savoir Leptiminus (9/9.5km) et Thapsus (6/6.5km) (Fig. 1). La nature du terrain et l’état fragmentaire du matériel sur le site de production ne permettent pas de fournir des exemplaires entiers. Par conséquent, mon étude typologique sera basée uniquement sur la section des bords d’amphores (Bonifay, 2004, 445). Cependant, l’étude minutieuse du matériel ramassé et les comparaisons avec les contextes publiés nous permettent d’appréhender le faciès des types produits in situ. Le site de Ghar Draj est localisé à quelques centaines de mètres du rivage, sur un affleurement rocheux de type calcaire, à l’intérieur d’une oliveraie. Il est dévasté progressivement par les constructions modernes. En surface, on n’observe que des amas de tessons de

103

LRCW4

d’une amphore de moyennes dimensions. Le bord haut et effilé se termine par une lèvre légèrement retroussée vers l’extérieur. La séparation entre le bord et le col est marquée par un sillon comme c’est le cas de l’Africaine IID (Fig. 2, n° 8). La dernière production de cet atelier se rattache au type Keay 25 sous-type 1. Le sommet de la lèvre est bien retroussé vers l’extérieur ou arrondi (Fig. 2, n° 9).

Bonifay a déjà souligné (2004, 140) la difficulté de cerner les types intermédiaires. Toutefois, le matériel de l’atelier de Bir El Hammam comporte à priori 4 variantes. La première est la variante classique A (Fig. 3, n° 13). Elle se caractérise par son bord biparti. La section de la lèvre est triangulaire et la partie inférieure du bord dessine un arc de cercle convexe à l’encontre du modèle des ateliers de Nabeul en Zeugitane, à profil plutôt rectiligne. La variante B est bien attestée avec une lèvre retroussée vers l’extérieur et qui touche la surface externe du bord. La face interne de la lèvre est rectiligne, contrairement à la variante précédente, ou légèrement bombée. On signale un décor peigné sous forme de lignes horizontales placées dans la partie supérieure du col (Fig. 3, n° 14-15). La troisième variante est la variante C, avec un profil de bord nettement convexe dans sa partie inférieure et dont la surface interne est nettement convexe. L’extrémité de la lèvre se détache vers l’extérieur (Fig. 3, n° 16-17). Enfin, des tessons se rattachent à la variante E du type Keay 62. C’est la variante la plus tardive, avec une lèvre à section triangulaire (Fig. 3, n° 18). Par contre, on observe le profil biparti du bord présentant un changement au niveau de la partie inférieure qui devient plus rectiligne. Au départ du col et au niveau de l’attache supérieure de l’anse, on remarque quelquefois un décor peigné en lignes horizontales (Fig. 3, n° 19). Un tesson de bord parait particulier avec la partie inférieure qui est extrêmement haute et à renflement léger (Fig. 3, n° 20). En somme, il semble donc plausible, d’après ce matériel, d’attribuer la variante tardive E à la Byzacène comme l’a suggéré M. Bonifay (2004, 140).

La vaisselle culinaire est attestée en faible quantité et représentée par les tajines Hayes 181 (Fig. 2, n° 10) et les couvercles Hayes 182 (Fig. 2, n°11), ainsi que par les marmites Hayes 183 (Fig. 2, n° 12). Il se pourrait également que la vaisselle commune soit produite sur ce site. Le faciès de production de l’atelier de Ghar Draj rappelle certains ateliers côtiers de Byzacène, notamment Leptiminus et Sullecthum. Ces derniers ont produit essentiellement des amphores Africaines classiques. Quant au type de transition, il a été produit par les ateliers de Sullecthum et correspond au type Sullecthum 10. On reconnait déjà une épave à cargaison composée en partie de ce type provenant des ateliers de cette dernière (Joncheray 1997, Fig. 1, 3). La céramique culinaire engobée est aussi présente sur cet atelier comme c’est le cas des ateliers de Leptiminus où la forme Hayes 181 constitue la principale forme (Carr 2009, 110-111). On peut placer la date de fonctionnement de l’atelier de Ghar Draj entre le début du IIIe siècle et le début/milieu du IVe siècle.

Le deuxième type se rattache au Keay 61 et comporte trois variantes produites sur place. La première est la variante D caractérisée par un bord en bandeau épais et à section triangulaire. Le passage du bord vers le col se remarque par un léger renflement (Fig. 3, n° 21-22). La deuxième variante est la variante A présentant un bord également en bandeau, plus haut que le précédent (Fig. 3, n° 23-24). Enfin la dernière variante est la variante A/D (Fig. 3, n° 25-26). Les fragments ramassés se caractérisent par un bord dont le profil se rapproche de Keay 61 A mais plus élancé et mince comme s’il s’agissait de Keay 8A. On ignore la forme du reste du corps d’amphore, mais il est fort probable qu’elle soit similaire aux autres amphores tardives. Une pointe à renflement annulaire et à pivot axial interne mais de profil moins pointu que de celui de Keay 8A pourrait être attribué à ce type (Fig. 3, n° 27). Vraisemblablement, il s’agit d’un type de transition entre Keay 61A et Keay 8A.

Le deuxième site, Bir El Hammem, a été signalé pour la première fois par l’archéologue tunisien N. Ben Lazreg (Stirling et al. 2001, 69, note 56) à la fin des années 1990 du XXe siècle alors que la problématique des recherches portait sur des structures relatives à des thermes romains encore visibles à ce moment-là. Toutefois, en surface, on ne peut voir actuellement que des ratés de fours, des quantifiés considérables de tessons de céramique non usagés ainsi que des briques de fours. Les constructions modernes ont, sans doute, balayé les structures relatives à l’atelier. Par ailleurs, les habitants ont mis au jour des structures de plusieurs fours de potiers actuellement enfouis sous les constructions modernes. La céramique ramassée en surface comporte deux groupes. Le premier, majoritaire, se compose de fragments d’amphores et de vaisselle commune produits in situ. Le deuxième groupe représente la céramique d’occupation provenant du site voisin de Ghar Draj et de la vaisselle culinaire et fine.

Le troisième et dernier type d’amphore est le type Keay 8A (Fig. 3, n° 28-33). Sans faire appel à une description détaillée de ce type déjà faite (Bonifay 2004, type 50) il pourrait constituer la dernière évolution des amphores africaines cylindriques de grandes dimensions et l’une des les plus tardives (Murialdo 2001, 267; Bonifay et al. 1998, 372).

Le type Keay 62, abondamment produit sur des ateliers de Zeugitane et de Byzacène (Bonifay 2004, type 46; Nacef 2007, type Ech Chekaf I), est bien présent sur notre atelier mais il s’agit surtout de variantes tardives. On doit noter tout d’abord que ce type, Keay 62, pose un problème de typologie et d’évolution de ses variantes. M. 104

NACEF

Par ailleurs, l’association des types amphoriques produits sur le site de Bir El Hammam trouve des équivalents sur les sites urbains de consommation dans le bassin occidental méditerranéen. La vaisselle associée aux amphores produites et attestée sur le site de Bir El Hammam, se rattache aux formes Hayes 88A/Bonifay type 46. C’est une variante classique des contextes du deuxième quart du VIe s. (Bonifay et al. 1998, 417, Tabl. CIII ). Des contextes urbains, comme à Luni, ont livré des fragments d’amphores rattachées aux types Keay 8A (Frova 1977, Tav. 280. 25; Lusuardi Siena et Murialdo 1991, 128, Tav. 4, n°2) et Keay 61 (Lusuardi Siena et Murialdo 1991, Tav. 4, 164, 124). Par ailleurs, les types 62 et 61 ont été trouvés dans des contextes à Luni en association avec les formes Hayes 102 et 104, ce qui fait penser que leur exportation a continué jusqu'à la fin du VIe s. - début du VIIe s. non seulement vers des centres côtiers comme Luni mais aussi vers l’intérieur. En Ligurie on les retrouve dans un niveau du VIIe s. à San Antonino di Perti (Murialdo 2001, 261-262 et 264-267), ainsi que dans la tombe de l’église de Finale dans un niveau de la fin du VIe s. (Murialdo 1988, 237, Fig. 7, n° 6; Panella 1993, 675).

trouver des inscriptions. Les plus communes sont les timbres, les marques ou même les croix, surtout sur le bas de la panse ou sur les pointes, appliqués ou tracées avant cuisson. Les croix avaient le rôle de marquage au cours de comptage des amphores pour une fournée comme ce qui a été observé pour les amphores africaines classiques produites par les ateliers de l’antique Sullecthum. Par contre, les amphores africaines tardives ne portent pas de timbres mais dans certains cas elles présentent des graffiti incisés avant cuisson. La cargaison de l’épave de la Palud comportait des amphores de type Keay 62A présentant des graffiti incisés avant cuisson sur le col (Long et Volpe 1998, 319, Fig. 282). Elles proviennent des ateliers de la région de Nabeul (Bonifay 2004, 453). En dernier lieu, on doit noter que cet atelier a produit de la vaisselle commune. Les bassins attestés (Fig. 4, n° 3638) rappellent les types Ech Chekaf II et IV (Nacef 2007, 584). Les bols présents in situ présentent des formes variées (Fig. 4, n° 39-41). Les jarres (Fig. 42-43) se rattachent au type 65 C de Bonifay. Enfin les formes fermées sont bien attestées. Ces cruches se caractérisent par un bord à section en bourrelet (Fig. 4, n° 44-47).

A Marseille, les contextes tardifs ont livré une quantité importante de matériel africain. Le faciès C3, par exemple, est dominé par le type Keay 62 avec une multitude de variantes. Ce faciès semble pouvoir être placé dans le deuxième tiers du VIe siècle (Bonifay et al. 1998, 357) alors que le faciès D, daté de la fin du VIe et la première moitié du VIIe siècle, se caractérise par des variantes tardives du type Keay 62, éventuellement la variante E associées aux types Keay 61C et Keay 61A (Bonifay et al. 1998). D’un autre côté, on remarque dans le faciès E daté de la deuxième moitié du VIIe et du début VIIIe siècles, que les amphores africaines comportent, en plus des spatheia miniatures attestés dans la production d’atelier de Moknine (Bonifay 2004, type 33D), les types Keay 61, Keay 8A et des amphores à corps globulaire. En somme, on doit signaler que dans le facies D de Marseille l’association de plusieurs types d’amphores africaines tardives, dont l’origine commune est à attribuer à la Byzacène n’est pas à mettre en doute. En plus des productions de l’atelier de Henchir Ech Chekaf (Type II/ Keay 61C), on note l’omniprésence des amphores des ateliers tardifs de Bir El Hammam et de Leptiminus (Dore 2001). Sur le site de Sidi Jdidi le type Keay 61 est attesté dans un contexte daté de la seconde moitié du VIIe s. (Bonifay et al. 2004, 306).

Le fonctionnement de l’atelier de Bir El Hammam semble coïncider chronologiquement avec celui de l’atelier de Henchir Ech Chekaf. Ce dernier a livré la variante A, classique, du type Keay 62 ainsi que la variante D. Par contre, la variante E la plus tardive, bien que produite sur ce site en faible quantité, est bien attestée sur l’atelier de Bir El Hammam, ainsi que les types Keay 61 A et Keay 8 A, ce qui plaide pour une datation jusque dans la deuxième moitié du VIIe s., voire même le début du VIIIe siècle. D’un autre côté, la production des amphores globulaires de tradition byzantine, type Ech Chekaf IV (Nacef, 2007) rejoint la même datation. L’atelier de potiers à Bir El Hammam a été installé, apparemment, sur des thermes romains. On ignore la datation du fonctionnement de ces derniers, ni à quelle date ils ont été désaffectés. Le toponyme du site Bir El Hammam, qu’on peut traduire littéralement par ‘puits des thermes’, est lui-même révélateur. En effet, le personnel de l’Institut National de Patrimoine, évoquait autrefois, avant l’avancée des constructions modernes, la présence en surface du site de mosaïques, de bassins et de structures importantes. Par ailleurs, N. Ben Lazreg signalait que le four était construit au sein des thermes (Stirling et al. 2001, 69, note 56).

Nous avons récolté en surface à Bir El Hammam deux fragments de parois d’amphores tardives qui portent des lettres incisées avant cuisson (Fig. 3, n° 34-35). Le premier comporte des traces de lissage et une seule lettre, probablement un A ou un alpha avec la traverse croisée. Le deuxième présente une seule lettre, également un alpha avec la traverse brisée, flanquée à gauche d’une spirale, qui devait former un mot entier. On voit de part et d’autre de la lettre alpha les restes d’autres lettres sous forme d’une haste à droite et d’un trait incliné à gauche. Généralement, sur des amphores africaines, il est rare de

Toutefois, cette situation est courante sur le sol africain. Elle nous rappelle, à quelques kilomètres vers le nord, le cas de Leptiminus. Une zone d’ateliers tardifs a été mise au jour, dans les thermes publics à l’est du site (Stirling et al. 2001). On peut voir encore dans le jardin du musée du site les mosaïques des salles des thermes qui sont percées 105

LRCW4

Tubulbaca. Le passage de Tubulbaca à Thubulba avec la disparition de la désinence CA, est banal à l'exemple de Damascos devenue Damas.

par des fours et des amas de déchets de cuissons et de casse de céramique. Ces ateliers ont produit les mêmes types que l’atelier de Bir El Hammam (Dore 2001). Le même phénomène est observé à Oudhna, où des ateliers produisant des sigillées africaines tardives ont été installés dans des thermes des Laberii (Barraud et al. 1998). De même à Tripoli, un atelier de potier d’époque romaine et qui a produit essentiellement des amphores, a été localisé sur un petit complexe thermal (Faraj Shakshuki et Shebani 1998, 282).

Cela dit, nous avons un évêché dans la province de la Byzacène (Notitia de 484, Byzacène, 5). La localité était représentée à la Conférence de Carthage en 411 par l'évêque catholique Januarius (Lancel 1991, Actes, t. IV, 1502) et était encore un évêché sous la domination vandale (Lancel 2002, Notitia, 260 et 356). Pour S. Lancel, l'évêché est localisable. Or il n'y a pas lieu de rejeter l'hypothèse avancé naguère avec prudence par P. J. Mesnage en se fondant sur la filiation toponymique (Mesnage 1912, 230), mais à l'époque les témoignages archéologiques font défaut, ce qui n'est pas le cas aujourd'hui où la présence d'un évêché dans les environs de la ville actuelle de Téboulba ne fait pas de doute. Les structures d’une église ont été identifiées à Henchir Soukrine à quelques kilomètres au nord des deux ateliers de potiers (Béjaoui 1988). On doit signaler, également, une deuxième église présentant un baptistère aussi intéressant et de forme originale (Warland 2009, 292) mise au jour à Galâa, à environ deux kilomètres au sud de zone d’ateliers de potiers. Cette église a été en partie édifiée sur des tombes puniques avoisinant un atelier de potier d’époque punique tardive. Le baptistère est daté de la seconde moitié du VIe s. (Ben Lazreg 2003, 494-495). Les marques gravées sur des tessons d’amphores de Bir El Hammam signalées ci-dessus pourraient être des marques de fabrique dépendant d’une propriété ecclésiastique.

En définitive, on pourrait comprendre les raisons de l’installation de ces ateliers sur le domaine des édifices importants, fort probablement publics, comme les thermes. En premier lieu, il pouvait s’agir de raisons techniques, comme l’infrastructure fournie par ces établissements, essentiellement l’eau et l’accessibilité (Stirling et al. 2001, 69) à des axes routiers pour diffuser la marchandise. En deuxième lieu, on doit noter l’emplacement des ateliers de Ghar Draj et de Bir El Hammam à proximité du rivage, où on observe encore les restes de mouillages utilisés à l’époque moderne pour le transport de la pierre extraite des carrières avoisinantes. En contrebas de l’atelier de Ghar Draj, des structures d’époque romaine ont été mises au jour, dont on ignore la fonction. Plusieurs interrogations se posent au terme de ce travail. Tout à bord, la production de la céramique dans cette région était-elle développée sous l’égide d’une autorité quelconque? S’agit-il d’ateliers ruraux ou suburbains ? Pour répondre à ces questions, on dispose de peu d’éléments. Les cités reconnues et avoisinantes aux deux sites sont Leptiminus au nord et Thapsus au sud. Il s’agit de cités portuaires importantes et dont l’origine remonte à l’époque punique. Il n’est pas lieu ici d’évoquer les problèmes de territoires mais plutôt de marquer des points de repère pour notre atelier.

Les renseignements fournis par les ostraca de Carthage datées du IVe siècle pourraient être d’une grande utilité. En effet, le texte de l’un des ostraca signale une quantité d’huile provenant de Byzacène et exactement de "Tebelbucitan". J. T Peña pense que le "plebs Tubulbacensis" pourrait constituer une référence topographique qui s’inscrit en Byzacène. Il s’agissait selon lui d’une sorte de district probablement de nature ecclésiastique et les conteneurs qui transportaient cette quantité d’huile vers Carthage étaient à 84% des outres (Peña 1998, 194). Si on accepte l’idée que le Tebelbucitan des ostraca de Carthage est le même Tubulbacensis du concile de Carthage et qu’il correspond à l’actuelle localité de Teboulba, une question se pose alors: pourquoi transportait-on cette huile d’olive dans des outres jusqu’à Carthage alors qu’on est à mi-chemin entre deux ports de Byzacène? En l’absence de toute preuve épigraphique, on ne peut confirmer cette hypothèse de localisation.

De prime abord, le toponyme actuel de Téboulba pourrait traduire une survivance d’un exemple des toponymes mentionnés dans les sources antiques. Nous avons la liste des évêques présents au concile de 411 à Carthage, dans le cadre du schisme donatiste. Elle mentionne un seul évêque catholique, un certain Ianuarianus, qui représente son église localisée au Terentianus Tubulbacensis. Les spécialistes indiquent qu’elle siège en Byzacène et P. J. Mesnage propose que cet évêché ait pu donner son nom au village de Teboulba (Mesnage 1912, 230). Toutefois, mon collègue et amie Ridha Ghaddhab spécialiste dans la question, précise (une communication orale) que nous n'avons pas le nom antique de la localité, mais le nom ethnique : Tubulbacensis. Or, les Latins usaient la dérivation par le suffixe pour former des adjectifs à partir des toponymes. Ici nous avons le suffixe ensis qui est accolé dans la majorité des cas des noms se terminant en –a (Galand 1951), ce qui pourrait donner

Conclusion Cette découverte permet de compléter utilement la carte des ateliers en Byzacène. Cette dernière comporte un héritage artisanal important dans ce domaine, remontant à l’époque punique: l’atelier d’El Maklouba (Peacock et al. 1989; Nacef 2008-2009) et l’atelier de Thapsus (Ben 106

NACEF

Younes 2002). À l’époque romaine, l’essor de cette activité a connu un large développement qui ne rompt pas avec la tradition punique. La majorité des ateliers sont en fait spécialisés dans la production des amphores, la céramique commune et la vaisselle culinaire.

other Studies, 75-98. Journal of Roman Archaeology Supplementary series no. 41. Portsmouth, Rhode Island. Faraj Shakshuki, M. et Shebani, R. 1998. The Roman Kilns of Hai-el-Andalus, Tripoli. Libya Antiqua n. s. 4, 279-282.

En somme, l’absence des structures archéologiques relatives à une organisation urbaine et la faible densité des vestiges archéologiques d’une manière générale plaident en faveur d’ateliers de potiers de type rural, que ce soit pour le site de Ghar Draj ou pour le site de Bir El Hammam. Il est fort probable que l’ancienne Tubulbaca constituait une périphérie de la ville de Thapsus. A l’époque byzantine, l’atelier de Bir El Hammam dépendait probablement d’un domaine rural, mais on ignore s’il était de caractère ecclésiastique. Quoi qu’il en soit, l’atelier de Bir El Hammam témoigne de la vitalité économique de la région à la veille de la conquête arabe, dont les traces sont attestées dans le bassin occidental méditerranéen, notamment sur le Castrum de S. Antonino di Perti. C’est l’ultime témoin matériel à la fois d’une occupation humaine et d’une activité économique à la fin du VIIe s. dans la région.

Frova, A. (ed.), 1977. Scavi di Luni II. Relazione delle campagne di scavo 1972-1973-1974. Rome, Bretschneider. Galand, L. 1951. La formation des ethniques dans l'Afrique du Nord romaine. Problème de méthode et observations. In 3e congrès international de toponymie et d'anthroponymie, Bruxelles, 15-19 juillet 1949, Actes et mémoires, 778-786. Louvain. Joncheray, J. P. 1997. Deux épaves du Bas-Empire romain première partie: l’épave Chrétienne D. Cahiers d’archéologie subaquatique XIII, 121-164. Lancel, S. (ed.), 1991. Actes de la conférence de Carthage en 411, t. IV, Notices sur les sièges et les toponymes, notes complémentaires et index. Paris, les Editions du Cerf. Lancel, S. (ed.), 2002. Notitia de 484, Byzacène, 5= Victor de Vita. Histoire de la persécution vandale suivie de La passion des sept martyrs, Registre des provinces et des cités d'Afrique. Édition critique, traduction et commentaire. Paris, Les Belles Lettres.

Bibliographie Barraud, D., Bonifay, M., Dridi, F. et Pichonneau, J.-F. 1998. L’industrie céramique de l’Antiquité tardive. In H. Ben Hassen et L. Maurin (ed.), Oudhna (Uthina), la redécouverte d’une ville antique de Tunisie, 139167. Mémoires 2, Bordeaux, Ausonius Publications.

Long, L. et Volpe, G. 1998. Le chargement de l’épave de la Palud (VIe s.) à Port-Cros (Var). Note préliminaire. In Bonifay et al. 1998, 317-342.

Béjaoui, F. 1988. Note préliminaire sur l’église et le baptistère de Henchir Soukrine. Africa X, 98-104. Ben Lazreg, N. 2003. Découvertes dans deux villes portuaires romaines du moyen Byzacium, In A. Ben Abed-Ben Khader (ed.), Images de pierre. La Tunisie en mosaïque, 494-495. Paris, Ars Latina.

Lusuardi Siena, S. et Murialdo, G. 1991. La ceramiche mediterranee in Liguria durante il periode bizantino (VI-VII secolo). In A cerâmica medieval no Mediterrâneo Ocidental, IV Congresso Internacional (Lisbonne, 16-22 novembre 1987), 123-146. Mértola.

Ben Younes, H. 2002. El Gaala site périphérique de la ville de Thapsus (Ras Dimas) à l’époque préromaine. Note préliminaire. Reppal XII, 9-15.

Mesnage, P. J. 1912. L’Afrique chrétienne. Évêchés et ruines antiques. Paris. Murialdo, G. 1988, Necropoli e sepolture tardo-antiche del Finale. Rivista di Studi Liguri LIV, 221-242.

Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports, Int. Series 1301. Oxford, BAR Publishing.

Murialdo, G., 2001. Le anfore da trasporto. In T. Mannoni et G. Murialdo (eds.), S. Antonino. Un insediamento fortificato nella Liguria bizantina, 255296. Collezione di Monografie Preistoriche ed Archeologiche XII, Bordighera.

Bonifay, M., Carre, M.-B. et Rigoir, Y. 1998. Fouilles à Marseille les mobiliers (Ier –VIIe siècles ap. J.-C.). Études Massaliètes 5, Paris, France. Carr, K. 2009. Strong local production in Tunisia: supplementing Bonifay from the case of Leptiminus. In J. Humphrey (ed.), Studies on Roman Pottery of the Provinces of Africa Proconsularis and Byzacena (Tunisia), 105-126. Journal of Roman Archaeology Supplementary series 76. Portsmouth, Rhode Island.

Nacef, J. 2007. Nouvelles données sur l’atelier de Henchir Ech Chekaf (Ksour Essef-Tunisie). In M. Bonifay et J.-C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry. British Archaeological Reports, International Series 1662, 581-591. Oxford, BAR Publishing.

Dore, J. 2001. The major pottery deposits following the disuse of the East Baths. In L.M. Stirling, D.J. Mattingly, N. Ben Lazreg (ed.), LEPTIMINUS (LAMTA). REPORT NO. 2. The East Baths, Cemeteries, Kilns, Venus Mosaic, Site Museum, and

Nacef, J. 2008-2009. Production de la céramique antique et ateliers dans la région de Salakta et Ksour Essef. Thèse de Doctorat non publiée, Faculté des sciences humaines et sociales de Tunis. 107

LRCW4

Panella, C. 1973. Le anfore. In A. Carandini et C. Panella (eds.), Ostia III, Le Terme del Nuotatore, Scavo degli ambienti III, VI, VII, Scavo dell'ambiente V e di un saggio nell'area SO, 464-696. Studi Miscellanei 21. Rome, De Lucca.

from Carthage. In Carthage Papers: the Early Colony’s Economy, Water supply, a Public Bath, and the Mobilization of State Olive Oil, Journal of Roman Archaeology Supplementary series 28, 116-238. Portsmouth, Rhode Island.

Panella, C. 1993. Merci e scambi nel Mediterraneo tardoantico. In Storia di Roma III, 2, 613-697. Turin.

Stirling, L. M., Mattingly, D. J., Ben Lazreg, N. 2001. LEPTIMINUS (LAMTA). REPORT NO. 2. The East Baths, Cemeteries, Kilns, Venus Mosaic, Site Museum, and other Studies. Journal of Roman Archaeology Supplementary series 41. Portsmouth, Rhode Island.

Peacock, D. P. S., Bejaoui, F. et Belazreg, N. 1989. Roman amphora production in the Sahel region of Tunisia. In Amphores romaines et histoire économique, Dix ans de recherche, 179-222. Collection de l’Ecole Française de Rome 114. Rome.

Warland, R. 2009. Frühes Christentum in Nordafrika. In Das Königreich der Vandalen. Catalogue d’exposition, 285-294. Karlsruhe, Badisches Landesmuseum.

Peña, J. T 1998. The mobilization of state olive oil in Roman Africa: the evidence of late 4th-c. ostraca

108

NACEF

109

LRCW4

110

NACEF

111

LRCW4

112

POTTERY FROM THE EASTERN MARMARICA (EGYPT) - A SEMI-ARID REGION AS PRODUCER AND MEDIATOR IN PTOLEMAIC, ROMAN AND LATE ROMAN TIME HEIKE MÖLLER,1 ANNA-KATHARINA RIEGER2 1

University of Cologne; [email protected]

2

Martin-Luther University of Halle – Wittenberg; [email protected]

The Marmarica is a semi-arid and arid region situated in today’s northwestern Egypt and northeastern Libya. The area is archaeologically almost unexplored and its material culture is therefore nearly unknown. However, it is a culturally rich area that played a considerable role, geographically as well as economically, as a mediator between the Mediterranean Sea, Sahara, Cyrenaica and the Nile Valley. A surprisingly high number of pottery production sites could be identified in the coastal zone of the Eastern Marmarica and many of them were surveyed over the past years. The majority of the discovered sites show amphora production on a large scale (following Empereur´s amphores égyptiennes series AE 2-3) and, on a lower scale, coarse ware production dating from Ptolemaic times to the 4th century AD. Only one site stands out and revealed vessels of the 6th and 7th century AD – mainly Late Roman cooking ware. Therefore an obviously enormous hiatus is notable in between the 4th and the 7th century AD in an area with a high – almost “industrial” – output of amphorae in Ptolemaic and Roman times dropping to a very reduced or nonexisting level in Byzantine times. This fact correlates with a remarkable decrease of local ware and vice versa an increase of imported containers in the region mainly from the Eastern Mediterranean in Late Roman times. KEYWORDS: EASTERN MARMARICA (NORTHWESTERN EGYPT), ANCIENT WAYPOINTS AND ROUTES, PTOLEMAIC AND ROMAN POTTERY PRODUCTION, AMPHORAE, LATE ROMAN COOKING WARE Around 30 more could be identified on satellite imagery. They can be distinguished into large, medium and smallscale production sites (Rieger and Möller 2011), (Fig. 1) with a chronological frame from Ptolemaic to Roman time. Only one production site stands out and is of later date: Byzantine or even early Arab.

Introduction1 The Marmarica is located on the northern fringe of the Libyan Desert, north of the Qattara Depression, between the Nile Valley and the Cyrenaica. The region is a marginal area, in political and economical terms, which is underlined by its precarious climatic conditions.

A large-scale production site therefore means a huge amount of wasters and sherd heaps of up to 6 m in height, which are very easily recognisable in the landscape as round or oval mounds consisting of two, sometimes three hills. One dump is reddish in colour due to the fabric of the sherds, one is often more greenish in colour because it contains slags, wasters and overfired bricks. The sherd density is very high (Fig. 2). Medium-scale production sites appear as more or less circular mounds consisting of only one single heap, which never reaches the dimension and height of the large-scale production sites. The sherd density is lower, the reddish coloured pottery heaps are missing, instead they appear greyish/greenish in colour due to their mixture of slags and overfired pottery and bricks. Small-scale production sites and/or greyish areas are silhouetted against the ground as little round greyish spots. The sherd density is very low, the degree of fragmentation of the sherds in contrast to the other categories is very high. But even though there is no evidence like slag or wasters that could point to pottery production, the excavation of one of these greyish spots was able to prove that those marks can be seen as indicators for pottery production since a small pottery kiln was exposed (Fig. 3).

Recent research in that semi-arid and arid region based on hydrological, pedological and archaeological surveys and analyses revealed considerably rich anthropogenic remains in the eastern part of the Marmarica. The data collected so far allows us to reconstruct settlement as well as land use patterns in an ancient arid landscape with an agricultural productivity that reaches surplus character (Rieger 2009; Vetter et al. 2009). Furthermore, it provides an insight into a level of pottery production that partly reaches an “industrial” output level (Rieger and Möller 2011). Pottery production in the Eastern Marmarica As yet, more than 55 pottery production sites could be located along the coast from Ras Abu Laho to Ras elHekma in an investigation area of about 100 km x 10 km. 1

The research was conducted in the frame of the project “Marmarica-Survey” of A.-K. Rieger, financed by the German Research Foundation (DFG) in collaboration with the Egyptian Ministry of State for Antiquities (MSA). The Gerda Henkel Stiftung supports H. Möller with a scholarship. Our thanks go to these institutions for funding and supporting the work.

113

LRCW4

harbours and urban centres amphorae with the goods could be traded (Vetter et al. 2013; Rieger and Möller 2011).

While medium and small kiln sites were situated within settlements, large-scale sites seem to be more separated without any close association to a settlement. Generally they are placed along the coast and on the Northern Tableland in a zone where permanent habitations prevail due to an effective system of rainwater and soil harvesting, that allowed people to base their livelihood on agriculture (dry-farming) (Vetter et al. 2009).

Amphorae produced in the Eastern Marmarica in Ptolemaic and Roman times The amphora types produced in the Eastern Marmarica resemble the Amphores Egyptienne series – AE 2 and AE 3 (Empereur 1986a). Same and similar forms were produced further east, not only around Lake Mariout, but also in the Delta and the Nile Valley (Dixneuf 2011; Ballet and Dixneuf 2004) and also in the Western Marmarica in Roman times (Hulin et al. 2009, 181; cf. Rieger and Möller 2011). The slight differences in form seem to substantiate regional differences (Dixneuf 2011, 97-128; cf. Möller in prep. for the Eastern Marmarica). It is remarkable that there is no production of the AE 4 (Dressel 2-4) in the Eastern Marmarica, since it was imitated all around the Mediterranean. In Egypt the production sites are situated around the Lake Mariout (Empereur 1986b). But probably other kiln sites do exist since some fragments of alluvial clay have been found that due to their fabric cannot have been produced in the same region, but maybe close-by (Dixneuf 2011, 129ff.). Nevertheless, the area of production seems to be very limited. This could be due to the fact that this type of amphora was probably the common packing unit for shipping and therefore the production sites were located next to the big central markets for overseas trade (Möller in prep.). This could explain the lack of production of AE 4 amphorae in the Eastern Marmarica.

Taking a closer look at the setting in their natural environment shows that the production sites are located according to the raw materials needed for the production process, namely close to water and clay sources, near the wadi beds and next to manmade water-supplies – the cisterns, a guaranteed and constant access to water in the dry region. In addition, clay deposits can be found in the wadi beds close to the sites. The pedological analyses show the rich loamy nature of the soil for this region and deposits of highest-quality primary material are visible all over the region (see Möller and Rieger in prep.). The fabric2 of the pottery shows a silty, quartzitic more or less carbonaterich, slightly ferrous clay. Sometimes organic inclusions occur. The section colours vary from red/brown to brownish, reddish and reddish-orange. The characteristics of the clay do not differ from one site to another in the Eastern Marmarica and resemble the clay of the Mareotis region (Rodziewicz 1998, 245 ff.; Rieger and Möller 2011), because of similar geological conditions (Fig. 4). The construction and spatial organisation of the largescale production sites seems similar or even identical for all the places examined (Rieger and Möller 2011, Ashmawi 1998): Adjacent buildings - a rectangular elongated building for producing and storing the amphorae – is situated in close proximity to the kiln. The kiln itself is an updraught kiln of about 5m in diameter with an estimated height of about 4m, as it could be reconstructed based on results from the excavation of a kiln-site (Fig. 5) in Wadi Qasaba (Möller in prep.). At many sites pressing installations with large basins for wine processing could be identified in close proximity to the kilns. This also is a common feature further east along the coast and in the Mareotis region (cf. Ashmawi 1998; Empereur 1993; Szymanska and Babraj 2004). Finds of resin to coat the inner wall of the amphorae can be seen as a further indicator that wine and its derivates were some of the products, and probably the primary ones. Written sources complement this hypothesis (cf. Norsa and Vitelli 1931; Strabon, Geographika, 17.I.14). Archaeobotanical analysis and planting mounds - likely for grapes - can verify the assumption of extended viticulture (cf. Rieger et al. in prep.(a); Vetter et al. in press).

The earliest type produced in the Marmarica is the AE 2 (2nd/1st century BC) with its subtypes (Fig. 6). The form is completely different to the later Roman ones and resembles the Greek amphora shape. The rim is plain, but sometimes slightly incurved on top and markedly undercut. The diameter is generally around 15cm, sometimes less. The handles are long and straight and set on the shoulder. They are bent in an almost rectangular fashion there; where they are attached to the neck slightly below the rim, the section is flat and oval. The neck is quite long, the body rounded and narrow and tapers into a solid conical spike (for the production of AE 2 amphorae see: Dixneuf 2011, 87ff.; Coulson et al. 1986, 548; Empereur and Picon 1998, 77; Majcherek and Shennawi 1992, 133; Rieger and Möller 2011; Möller in prep.). Roman AE 3 amphorae have one general type with variations, which can be divided into numerous subtypes (Fig. 7) (cf. Dixneuf 2011, 97ff.). In general the form of the bicontronical amphora is distinguished by a long cylindrical neck, striated or even, and differing in height with handles that vary in shape from bigger, almost rectangular ones that are always fitted in the upper part of the neck, attached slightly underneath or directly onto the rim. The latter always appears with little looped handles. The shoulder is rounded and leads into a cylindrical, long body that tapers into a solid spike. Earlier forms occur

As soon as the amphorae were fired, the commodities could be filled in the containers and be transported to their destination, since the production sites are placed nearby to roads and tracks. Thus closely connected to 2

Our thanks go to G. Schneider, Berlin for a first estimation of the local fabric of the Eastern Marmarica.

114

MÖLLER -RIEGER

carried out within the next few years. If successful and the production of LRA 5-6 amphora can be proven, this would be so far the only known spot along the western coast of Egypt between Abu Mena and Sollum that produced late Roman amphorae (cf. Dixneuf 2011, Fig.181; Möller in prep.).

from the 1st to the 3rd century AD, later ones to the 4th century AD. (For the forms and their dating in the Mareotis area and the Nile valley cf. Dixneuf 2011; for the Marmarica cf. Möller in prep.). The production of different forms within the type AE 3 at one production site is possible and can be seen at one excavated kiln-site in Wadi Qasaba (Rieger and Möller 2011; Möller in prep.). Here the transition or replacement of forms and the coexistence of different subtypes was visible. The scientific analysis of a charcoal layer that separates the two thick pottery layers is expected soon and should give a better idea of the production period and more information concerning better dating of the amphora-forms produced in the Eastern Marmarica.

Within the 5th century AD the production centres in Egypt seemed to cluster around the Delta area and the Nile Valley; and monastery sites (cf. Abu Mena – Engemann 1992) seem to have played an important role in amphora production. Conclusion It remains unclear why the demand of local products in the Eastern Marmarica, and therefore the need for a high quantity of production sites, decreased within the 5th century AD. But it is certain that this phenomenon - a decrease of local amphorae from the coast and an increase of imported ware in late Roman times - is not an isolated phenomenon in Egypt (cf. Martin 2010, 947). Containers of “non-alluvial“ clay coming from the Mareotis or further west from the Eastern Marmarica seem almost to disappear in the 5th century AD (cf. Dixneuf 2011, Fig. 181). In contrast, Egyptian amphorae made of Nile clay still occur, while the import of mainly Eastern imports increases, as shown by statistics of amphorae finds in Alexandria (Majcherek 2004).

Marmarican networking in Roman and late Roman times and its differences To reconstruct the ancient road-network in the Eastern Marmarica, a survey was conducted in 2007 (Vetter et al. 2013), concentrating mainly on cisterns and their archaeological remains. The finds of these ancient sites were mainly amphorae – Ptolemaic and Roman – type AE 2 and AE 3 that were most likely produced at one of the kiln-sites at the coast judging by their similar form and fabric. So here we can follow the trail of the locally produced amphorae from the coast down to the oasis following ancient routes in Ptolemaic and Roman times, trading the foodstuffs that were produced at the sites (see above).

Nevertheless, the Eastern Marmarica as a producer of local pottery in an almost “industrial“ sense in Ptolemaic and Roman times still keeps its importance as mediator between East and West, North and South: In Byzantine times, or in other words during the 5th century, its function just changes and is focused on imports mainly from the Eastern, but also from the Western Mediterranean.

In late Roman times, at the beginning of the 5th century AD, the picture changes at the cistern sites: Only imported amphorae can be recorded, or in other words: finds that were not locally produced, with the terminus local meaning made at one of the coastal production sites in the ancient Marmarica. Wares produced in Egypt occur in very low quantities. Some examples from Abu Mena, judging by their fabric, do exist, next to singular finds from the Nile valley (Dixneuf 2011, Fig. 181). Most of the finds are imported from further away, from the Western Mediterranean, as Keay LII amphorae, produced in South-Italy, show (Pieri 1992, Fig. 9.13) as well as – and in comparison to other imported types occurring in high quantities – LRA 1 Amphorae from the Eastern Mediterranean, mainly from Cyprus and Cilicia (Pieri 2007; cf. Rieger et al. in prep.(b); Vetter et al. 2013) (Fig. 8). This assertion suits the fact that the production of amphorae in the Eastern Marmarica seems disrupted in the course of the 4th century AD, since the large-scale production sites were abandoned latest at that time.

Bibliography Ashmawi, F. el. 1998. Pottery Kiln and Wine-Factory at Burg el-Arab. In J.–Y. Empereur (ed.), Commerce et artisanat dans l´Alexandrie hellénistique et romaine. Bulletin de Correspondance Hellénique, Supplément 33, 56–64. Ballet, P. and Dixneuf, D. 2004. Ateliers d´amphores de la chôra égyptienne aux époques romaine et byzantine. In J. Eiring and J. Lund (eds.), Transport Amphorae and Trade in the Eastern Mediterranean, Acts of the International Colloquium at the Danish Institute at Athens, September 26–29, 2002. Monographs of the Danish Institute at Athens 5, 67– 72. Århus. Aarhus, University Press.

Only one production sites stands out: a medium-scale production site in Wadi Hashefi, where late Roman cooking ware was produced, as indicated by wasters on the site (Fig. 9). Until now it remains unclear if even a variation of LRA 5-6 amphora was made on that site according to the high quantity of finds of this type within this area. Since neither wasters nor scientific analyses of the clay could prove the production of these bag shaped amphorae on this site, further investigations have to be

Coulson, W. D. E., Wilkie, N. C. and Rehard, J. W. 1986. Amphoras from Naukratis and Environs. In J.-Y. Empereur and Y. Garlan (eds.), Recherches sur les amphores grecques. Bulletin de Correspondance Hellénique, Supplément 13, 535–550. 115

LRCW4

Dixneuf, D. 2011. Amphores égyptiennes. Production, typologie, contenu et diffusion (IIIe siècle avant J.–C. – IXe siècle après J.–C.). Alexandrie, Centre d’Études Alexandrines.

Möller, H., in prep. Ptolemäische und römische Keramik in der antiken Marmarica – Nordwestägyptens Wüstenrandgebiet als Produktionsort und Mittler. PhD thesis, University of Freiburg im Breisgau.

Empereur, J.–Y. 1986a. À la recherche des fours d’amphores. In J.-Y. Empereur and Y. Garlan (eds.), Recherches sur les amphores grecques. Bulletin de Correspondance Hellénique, Supplément 13, 103126.

Möller, H. and Rieger, A.–K. in prep. Locally made pottery along the routes of the Libyan Plateau in Graeco-Roman times - first results of petrographical and geochemical observations on calcareous clayfabrics in the Eastern Marmarica, Cahiers de la céramique égyptienne 10.

Empereur, J.–Y. 1986b. Un atelier de Dressel 2-4 en Egypte au IIIe siècle de notre ère. In J.-Y. Empereur and Y. Garlan (eds.), Recherches sur les amphores grecques. Bulletin de Correspondance Hellénique, Supplément 13, 599–608.

Pieri, D. 1992. Amphores de l´antiquité tardive à Marseille (Corne du port antique, Sondages DII 11 et DII 15). Analyse stratigraphique et typologique. Unpublished Maîtrise d’Histoire, Université de Provence.

Empereur, J.–Y. 1993. La production viticole dans l’Egypte ptolémaïque et romaine. In M.–C. Amouretti and J.–P. Brun (eds.), La production du vin et de l’huile en Méditerranée. Bulletin de Correspondance Hellénique, Supplément 26, 9–47.

Pieri, D. 2007. Les centres de production d’amphores en Méditerranée orientale durant l’antiquité tardive: quelques remarques. In M. Bonifay and J.–C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports Series, 1662, 611–625. Oxford. BAR Publishing.

Empereur, J.–Y. and Picon, M. 1998. Les ateliers d´amphores du Lac Mariout. In J.–Y. Empereur (ed.), Commerce et artisanat dans l´Alexandrie hellénistique et romaine. Bulletin de Correspondance Hellénique, Supplément 33, 75–91.

Rieger, A.–K. 2009. Archäologie eines ariden Raumes: die Strukturierung einer ressourcenarmen Landschaft durch den Menschen am Beispiel der antiken Marmarica (Nordwestägypten). In R. Kath and A.–K. Rieger (eds.), Raum – Landschaft – Territorium. Zur Konstruktion physischer Räume als nomadischer und sesshafter Lebensraum. Nomaden und Sesshafte 11, 71–99. Wiesbaden, Reichert.

Engemann, J. 1992. A propos des amphores d’Abou Mina. Cahiers de la céramique égyptienne 3, 153– 160. Hulin, L., Timby, J. and Mutri, G. 2009. The Western Marmarica Coastal Survey 2009: Preliminary Report, Libyan Studies 40, 179–187.

Rieger, A.–K. and Möller, H. 2011. Kilns, Commodities and Consumers. – Greco-Roman Pottery Production in Eastern Marmarica (Northwestern Egypt). Archäologischer Anzeiger 2011/1, 141-170.

Norsa, M. and Vitelli, G. 1931. Il Papiro Vaticano Greco 11. Studi e Testi 53. Vatican City, Biblioteca Apostolica Vaticana. Majcherek, G. and Shennawi, A. el- 1992. Research on Amphorae Production on the Northwestern Coast of Egypt, Cahiers de la céramique égyptienne 3, 123– 136.

Rieger, A.–K., Vetter, T., Möller, H. and Nicolay, A. in prep. Water, Soil and Agriculture in Ancient Eastern Marmarica (northwestern Egypt) – a Geoarchaeological Approach to an Unknown (semiarid) Landscape.

Majcherek, G. 2004. Alexandria´s Long-Distance Trade in Late Antiquity – the Amphora Evidence. In J. Eiring and J. Lund (eds.), Transport Amphorae and Trade in the Eastern Mediterranean, Acts of the International Colloquium at the Danish Institute at Athens, September 26–29, 2002. Monographs of the Danish Institute at Athens 5, 229–237. Århus, Aarhus University Press.

Rieger, A.–K., Möller, H., Vetter, T., and Großkopf, B. in prep for 2012. On the route to Siwa: the ancient roadhouse of Abar el-Kanayis on the MarmaricaPlateau. Mitteilungen des Deutschen Archäologischen Instituts Kairo 68.

Rodziewicz, M. 1998. Experimental Identification of Pottery from Mareotis. In J.-Y. Empereur (ed.), Martin, A. 2010. The pottery from a late-antique Commerce et artisanat dans l´Alexandrie settlement at Schedia. In S. Menchelli, S. Santoro, M. hellénistique et romaine. Bulletin de Correspondance Pasquinucci and G. Guiducci (eds.), LRCW3. Late Hellénique, Supplément 33, 245–260. Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Strabon. Geographikae, 17.I.14 Archaeometry. Comparison between Western and Szymanska, H. and Babraj, K. 2004. Marea Fourth Eastern Mediterranean. British Archaeological Season of Excavations. Polish Archaeology in the Reports, International Series, 2185 (II), 945–949. Mediterranean 15, 53–62. Oxford. BAR Publishing.

116

MÖLLER -RIEGER

Vetter, T., Rieger, A.–K., Möller, H. and Klammer, O. 2013. Water, Routes and Rangelands: Ancient Traffic and Grazing Infrastructure in the Marmarican Desert Areas (Northwestern Egypt). In H. Riemer and F. Förster (eds.), Desert Road Archaeology in the Eastern Sahara, 455–484.

Tomber, R. 2006. The Pottery. In V. A. Maxfield and D. P. S. Peacock (eds.), Survey and Excavations, Mons Claudianus, vol. 3: Ceramic Vessels and Related Objects. Fouilles de l´Institut Français d´Archéologie Orientale 54, 1–236. Vetter, T., Rieger, A.–K. and Nicolay, A. 2009. Ancient Rainwater Harvesting System in the Northeastern Marmarica (Northwestern Egypt). Libyan Studies 40, 9–23.

Vetter, T., Rieger, A.–K. and Nicolay, A. in press, Disconnected contributing areas and adapted ancient watershed management in the semi-arid north-eastern Marmarica (NW-Egypt). Geomorphology.

117

LRCW4

Fig. 1. Pottery production sites in the Eastern Marmarica in Ptolemaic, Roman and Byzantine Times (Map: A.-K. Rieger, H. Möller)

Fig. 2. Large Scale Production Site in Wadi Kharouba (Photo: H. Möller)

118

MÖLLER -RIEGER

Fig. 3. Small Scale Production Site in Wadi Umm al Ashdan (Photo: H. Möller, A.-K. Rieger)

119

LRCW4

Fig. 4. Eastern Marmarican Amphora-Fabric Group A (Photo: IFAO Kairo)

Fig. 5. Wadi Qasaba Pottery Production Site

120

MÖLLER -RIEGER

Fig. 6a. Eastern Marmarican AE 2-1.1 (Typology: Dixneuf 2011) produced at kiln sites in Wadi Kharouba (Drawing: B. Böhm, H. Möller) Fig. 6b. Eastern Marmarican AE 2-2.1 (Typology: Dixneuf 2011) produced at a kiln site in Alam el Rom (Hashima) (Drawing: B. Böhm, H. Möller)

121

LRCW4

Fig. 7a. Eastern Marmarican AE 3-1 Fig. 7b. Eastern Marmarican AE 3-1 Fig. 7c. Eastern Marmarican AE 3 local variant Fig. 7d. Eastern Marmarican AE 3-1 produced at a Small Scale Production Site in the settlement of Wadi Umm al Ashdan. Fig. 7a-c produced in Large Scale Production Sites in the Eastern Marmarica. Fig. 8d produced at a Small Scale Production Site in the settlement of Wadi Umm al Ashdan. Typologie (Dixneuf 2011) with local variants (see Möller in prep.). Drawings: B. Böhm, H. Möller)

122

MÖLLER -RIEGER

Fig. 8a. Keay LII Amphora of South-Italian origin found in Abar el Kanayis (Drawing: B. Böhm, H. Möller) Fig. 8b. LRA 1 Amphora from Cyprus found in Abar el Kanayis. (Drawing: B. Böhm, H. Möller)

Fig. 9. Late Roman Cooking Ware produced at a Medium Scale Production Site in Wadi Hashefi (Drawing and Photo: H. Möller)

123

LRCW4

124

CULTURAL TRANSMISSION AND CONSUMER DEMAND-A CASE STUDY USING CERAMICS FROM KARANIS, EGYPT SONALI GUPTA-AGARWAL Archaeology Program, Cotsen Institute of Archaeology, UCLA [email protected] / [email protected] Karanis an ancient Greco Roman settlement situated in the Fayum, Egypt is taken as a case study to explore transmission of knowledge from skilled to unskilled potters and the effect of consumer demand in this transmission using locally manufactured cooking ware. Apprenticeship can be rigid and repetitive or flexible and variant depending on the degree of freedom given by the skilled potter to the student. The models of interaction between the potter and the student have an effect ranging from continuity to change in ceramics. It has been hypothesized that certain attributes of raw material availability such as abundance and quality of clay, condition the production of local ceramic form types. Cultural transmission at Karanis can be studied by measuring the dimensions of these local forms (cooking forms in particular) to test variability or standardization. Preliminary statistical analysis of the excavated pottery, using a coefficient of variation, indicates clear variability within a particular cooking ware type. The paper examines the observed variability being a result of apprentice-master interaction, consumer demand as also the dialogue between the individual and the collective using Bourdieu’s concept of habitus. KEYWORDS: CULTURAL TRANSMISSION, STANDARDIZATION, VARIABILITY, HABITUS, APPRENTICESHIP. importantly defines chronology and gives a temporal dimension to human lives. This paper seeks to shed light on underlying micro variables that prompt this change and continuity albeit in a preliminary way. Social interaction and the processes of learning have an effect on “conservation of” and “change in” material cultural traditions (Minar and Crown 2001). ‘The limits of cultural transmission involve ethnocultural variability which promotes and sustains ways of life’ (Schonpflug 2009, 3). Thus, evolution is adaptive, as it requires variability and allows the creation of differences to ever changing demands (Schonpflug 2009). The source of variation is important in understanding continuity and change. Some of the sources of variation in this process include human error in copying, cognition and social processes (Stark et al., 2008). The variability of objects as well as the variation of objects is significant as behind them lies a chance for us to gaze at the causative minute processes. It is anticipated that a deeper understanding of the artifactual variability, variation and standardization can also contribute toward an understanding of the learning mechanisms.

The idea of transference or transmission of skill has always been a fascinating concept. Why at all certain ideas and behaviors are transmitted and learnt, changed or abandoned, covers the study of cultural transmission. Culture can affect individuals’ behavior, which they acquire from other individuals through teaching, imitation, and other forms of societal based transmission. The term ‘cultural’ is said to apply to traits that are acquired by any process of non genetic transmission such as imprinting, conditioning, observation, imitation or direct teaching (Schonpflug 2009, 2). The skill of craft is not only learnt and transmitted from one person to another who may be genetically related but who may not be related at all. Thus, one cannot categorize something as fluid as transmission in watertight categories. Factors such as society, economy, politics, geography and people all have to be taken into account to understand what ‘really’ influences the intricate working of transmission. These influencing themes are reassessed periodically to feed the ‘whole’ of what is eventually transmitted. Transmission is then more dynamic being influenced by social and the historical contexts in which it takes form. Hence, cultural transmission needs to be approached in a realistic way. This realism is seen in the studies of situated learning (Lave and Wenger 1991; Minar and Crown 2001; Wallaert-Petre 2001); craft skill learning (Shennan and Steele 1999) and the use of habitus and practical knowledge (Bourdieu 1977) amongst many others.

To get to the micro nature of research an apt example is the ethnographic work done in the Cameroon by Wallaert-Petre (2001) involving the Dii, Duupa, and Doayo people on one hand and the Fali people on the other. This work helped understand the nature of interactions. Amongst the Dii, Duupa, and Doayo, learning is essentially taught over a period of seven years in a segmented fashion. This is a stage-by-stage instruction of the production process in this seven-year

It has often been said that “Change is the only constant”. In ceramic repertoires around the world, this change most 125

LRCW4

similar sized Khashboohas’ made at another workshop in the vicinity, the metric dimensions differed.

period. Learning is by observation and imitation with no room for the independent questioning mind. On the other hand the Fali people encapsulate the entire production process in one go. Adherents take two years to fulfill the apprenticeship period. Learning here too is encouraged through observation and imitation coupled with questioning and experimentation. The interpretation being that where transmission of learning involved careful guidance by the skilled teacher with a thrust on replication, one witnessed less variation in finished products. This difference in the instructional framework shows the ceramic evidence of continuity in tradition of the Dii, Duupa and Doayo people (Wallaert-Petre 2001). Here, no questioning was permitted and learning concentrated on perfect error free replication. Where transmission of learning involved less direction and more experimentation, greater innovation and hence variation is seen in the end products. Rapid changes are visible in the ceramic style of the Fali people prompted by the particular instructional framework. As has been observed in other ethnographic contexts, the more control the instructor has over the novice during the process of learning and creation of a material item, the more similar the objects they produce will look (Roe 1995, 51).

Metric attributes are related to the mechanical actions of the potter. These relate to motor habits and skills. Thus, if there is more standardization in metric attributes, more tight control, it will indicate a high degree of craft specialization and more control of the potter vis –a –vis the learner; and if there is more variability within the same type in the same workshop, it will show the absence of such a high degree of specialization and less control. The postulate being that teaching and consequent learning can be rigid or flexible depending on the degree of freedom given by the skilled potter to the student. Also, the degree of freedom or rigidity has an effect on the development of ceramic styles as one of two major influences of long-term continuity and change in a specific ceramic tradition. One of the important influential aspects is consumer behavior. Consumers not only impact ceramic changes, but also potentially the learning process for a specific ceramic functional type. If the consumer looks for a specific type of pottery without demanding change, the learning process is impacted because the producer (master potter) will require a standardized output. It prompts rigidity in the production, which allows a high production rate of the specific ceramic type satisfying the needs of the consumer. It saves time and effort and behind it has a history of experimentation deeming it successful. Standardized production implies that the craft is carried out by a group of individuals utilizing a limited range of materials and somewhat formalized or routinized techniques that result in identical procedures focused on replication such as mass production (Rice 1987, 202). As a result, standardized and thus specialized ceramics reflect little heterogeneity in composition and appearance within each category of pottery. Thus, in specialized pottery production, the variability of ceramic vessels is presumably reduced (Rice 1987, 202). The degree of standardization may be assessed through raw material composition, manufacturing techniques, form, dimensions, and surface decoration. Then does standardization lead to long-term continuity? The production of amphorae in the Mediterranean is a good example of continuity of a particular form that stays virtually identical in shape and size for long periods.

My work is greatly influenced by this anthropological approach of observing living potter communities. One of these communities of potters is at Nazla, a village in the Fayum. Potters live on the edge of the settlement near a stream, well equipped with kilns and make pottery vessels for consumers. The potters here manufacture a vessel called misa’a (Photo 1). It is mostly used as a container of grain or receptacles of food for animals. One of the potters namely Hassan revealed that he followed strict rules of measurement making this vessel. He professed that he was particular about the rim thickness. To test what Hassan said, I measured the dimensions of about 15 misa’as made by Hasan and 45 others made by his extended family. These vessels were found to be metrically standardized i.e., there was very little variation in the dimensions of the vessel. Fustat is another pottery manufacturing area not far from Coptic Cairo. Here, in one of the many workshops, the potters specialize in manufacturing the upper part of the shisha (Photo 2) or traditional hookah also called the Kursi or seat. Here too, it was noticed that the metric dimensions of the shisha were very standardized. However, when the potters here were asked to make the misa’a which the potters of Nazla specialized in, they were successful in making the form but when it came down to the metric dimensions, those were extremely varied. The important point is that they had the knowledge of making the form of the misa’a but they did not specialize in them. This illustrates why we ceramicists focus on both form and metric attributes. The form attributes of similar-sized vessels appear to be highly standardized i.e. they look uniform. However, the metric attributes of these vessels differed. Another example is that of the Khashbooha, a flower pot (Photo 3). Its dimensions were highly standardized when made by potters of one workshop. However, when compared to

The second case may be where the consumer desires change. Here, the learning process is impacted in a different way, prompting flexibility in learning. This flexibility while allowing the production of a variety of types also will result in greater variability in a specific ceramic type. This will slow down the production and thus directly impact its scale. Perhaps this metric standardization or metric variability is related to specific functional types that can help in discerning the learning interaction between the potter and the learner. With that said, there are many other key variables which influence transmission. For example, when talking of consumer demand, it is important to understand the nature of that demand which defines the function of the vessel, the level of demand and the scale of production informing us on the quantity of vessel types in circulation, the scale of 126

GUPTA-AGARWAL

a flexible relationship wherein the learner is free to experiment, not be tied down to strict rules of measurement and explore newer dimensions in the skill. This kind of interaction in the long run allows innovation consequently leading to change at a faster pace than the previous model. Where transmission of learning involves less direction more metric variability would be seen in similar types as also more variety of types.

production, the proximity, abundance and quality of clay and the distribution of the product. The production and the distributional area of the vessel help in inferring the catchment area (Hodder 1979). The shape of the distribution pattern would also help infer the social constraints affecting this distribution (Hodder 1979). The changing pattern of the ceramic assemblage will shed further light on the politics and economics of the period. Thus, production of vessels does not occur in isolation but is embedded in political, social and/or economic systems (Costin 1991).

I conducted an experiment at Fustat, where potters of three workshops close together were asked to identify 26 random broken rim sherds picked up from around the activity areas. It was interesting to note that all 8 potters from these workshops could correctly identify the sherds that they had made as also the ones made by the others. The potters’ ability to identify these key features of the rim by virtue of practice and working closely enables one to understand the underlying subtleties of workshops and their contribution to learning. It potentially allows us to gauge whether the bent of the interaction in this workshop is more towards metric standardization with slight metric variability within the ‘accepted range’ or metric variability beyond the ‘accepted range’. In my view, these interactive models can be systematically studied. These subtle timed markers of habit, behavior, action and gesture or body movements (Wendrich 2002) leave markers on the pottery that should be discernible in the archaeological record. Skill in pottery making resides as much in the hands as in the brain. Learning through repeated practice allows internalization leading to automatic responses and replication. The workshop tradition then can be seen as a unit that bears a signature which cannot be replicated beyond this specific domain, unless one is a part of the specific enculturation or ‘learning tree’ (personal observation).

There are two types of attributes: one, intentionalcontrolled by the craftsperson (technological, morphological, and stylistic) (Costin and Hagstrum 1995) and the other mechanical, relating to motor habits and skills. However, these attributes do not transmit in thin air. Daily practice, a place where this diverse knowledge is transmitted better understood by Bourdieu’s habitus. It provides a means of thinking relationally about those significant issues (Grenfell 2008). Bourdieu (1994, 65; Grenfell 2008) once stated “all of my thinking started from: how can behavior be regulated without being the product of obedience to rules”. This is the crux as to what the earlier mentioned Cameroon example illustrates. How does social structure and agency come together to shape the ‘outer’, the ‘social’ and the ‘inner’ (Grenfell 2008)? For this research, the social structure becomes the space where the potters work, where they interact and are in constant dialogue with the outer and the inner, shaping each other. It then becomes the primary unit that allows transmission to take place. The exchange of knowledge and experience is a dynamic process where knowledge is learnt, unfolded and developed in a social space (Bourdieu 1977, 1987). Therefore this collective group represents a workshop or a community of practice, transmitting a certain tradition, in a manner which is somewhere on the scale from rigid to flexible. This expression of the individual finds meaning in the collective.

For the purposes of my preliminary research, the focus was on domestic utilitarian wares, which have ready market of consumers. The aforesaid wares are not influenced by notions of value defined by aesthetics or exclusivity. It is also important to recognize the importance of the consumer who not only impacts ceramic changes but also potentially the learning process. The demand for a certain type, here the cooking pots can be made visible in the archaeological record by quantifying the types in different contexts. These could then be compared to the types that are quantitatively less, thereby allowing us to know whether there is high demand or low demand. This demand affects the specialized production of a type or a non-specialized production, which has a role to play in the interaction between the potter and the learner.

Learning then from its inception is made possible through interaction between a potter and a learner. I call this the potter-learner model of interaction. Here, the learner is someone who works under the supervision of the master potter and can be at any stage of learning or even a seasoned potter himself. When there is metric standardization or tight control, the potter is in more control than the learner. This interaction is rigid and rooted in traditions. There is no room for innovation as the learner is bound by strict rules of duplication. If seen vertically through time, this learning framework also maintains continuity of a tradition. That is not to say that there is no change, it does take place but very slowly. There would be less metric variability within an accepted range and more metrically standardized vessels of a particular type within one period, perhaps indicative of specialized workshops. In this set up, one would often find the potter reprimanding the learner with the fear of getting hit by a stick in order to get small hands trained the right way. The other extreme of the interactive model is where we would find metric variability. Here the Learner will be more in control than the potter. It prompts

To verify the plausibility of theories, it is important to test them. Hence, the site of Greco-Roman Karanis in Fayum, Egypt was chosen, to verify, whether the Potter Learner interaction could be discerned through archaeological ceramics. The first step was to conduct a quantification of sherds from a survey of kilns to see indications of what was locally manufactured by the potters at Karanis. A series of 8 kilns (Fig. 3) were chosen from this 60 hectare site. This was compared with an excavated kiln (Fig. 4) to see the indications of what 127

LRCW4

Though the cooking pots appear similar in form but there is clear variability in metric attributes. This could possibly suggest an absence of craft specialization for a particular functional type, thereby suggesting more freedom to the learner where a particular workshop is making a variety of types and hence the variability within a type is due to lack of specialization. However, this preliminary analysis could also indicate that this metric variability within a certain type maybe due to intra kiln variability. It is clear that local availability and abundance show standardization in the technique of manufacture but there is variability in subtle nuances of pottery making which could be a ‘signature’ of different workshops manufacturing the same type of cooking pot or lack of specialization in all of these workshops for the particular cooking pot.

was locally manufactured (Figs. 3 and 4). The results clearly indicated that in both the investigations, the local types dominated the corpus. The local fabric being made from local silts mixed with marl. The clay type used at Karanis allowed the handling of the next question regarding local form types manufactured at Karanis. Using ethnographic examples in Egypt regarding quality of clays and their abundance, defining features were placed in a contingency table to generate information on the forms manufactured and found in Karanis. The different categories of clay were also reflective of different functional activities. The data was tested to see what form types were produced by the potters under four different propositions. 1) High quality clay is also highly abundant 2) High quality clay is less abundant

This data set till now has helped in illustrating the ideas involved and has helped in mapping the road to further research. To delve deeper into the understanding of transmission, in the future it is intended to investigate metric variability or standardization within each of the kilns and then doing an intra kiln comparison of similar types. It is quite clear that by using the right approach, understanding ‘transmission’ at Karanis is not a distant dream but an enticing possibility.

3) Low quality clay is highly abundant 4) Low quality clay is less abundant The Kiln survey collection was used to populate the model. The vessels were divided in accordance to rim types. A count of the forms occurring in each fabric type was taken. This allowed an assessment of relationships between form on one hand and quality and abundance on the other. The results showed a clear pattern, that at Karanis, everyday utilitarian form types were occurring in local fabric while the non-utilitarian forms were occurring in other fabrics. A total of 1550 diagnostic sherds from the kilns were analyzed and grouped to understand the abundant rim types. Six cooking pot rims were most common. The kiln pottery represented by Trench 8 (Fig. 2) was compared to the similar types found in the habitation areas represented by Trench 4 in the north and trench 15 in the east (Fig. 2). The entire corpus of pottery analyzed was late roman in context and hence could be considered contemporaneous.

Bibliography Bourdieu, P. 1977. Outline of a Theory of Practice, Cambridge Studies in Social and Cultural Anthropology. Cambridge, Cambridge University Press. Bourdieu, P. 1987. What makes a social class? On the theoretical and practical existence of groups. Berkeley Journal of Sociology 32, 1-17. Bourdieu, P. 1994. In Other Words: Essays Towards a Reflexive Sociology. Translated by M. Adamson. Cambridge, Polity.

To test standardization or variability of the form types at Karanis, the Coefficient of Variation test or C.V. was used. It is a powerful tool for assessing variability or standardization in a single site or within the same cultural setting. Karanis qualified.

Costin, C. L. 1991. Craft Specilization: Issues in Defining, Documenting and Explaining the Organization of Production. Archaeological Method and Theory 3, 1-56.

As per the ethnographic experiment conducted at the workshop with 26 sherds (mentioned earlier), it was clear that amongst potters, the rim plays an important role in differentiating the identity of the maker, then for pottery in the archaeological context, the rim too is crucial for understanding differences between individual potters and school of potters representing a collective.

Costin, C. L. and Hagstrum, M. B. 1995. Standardization, Labor investment, Skill and the Organization of Ceramic Production in Late Prehispanic Highland Peru. American Antiquity 60, 619 - 639. Grenfell, M. (ed.), 2008. Pierre Bourdieu: Key Concepts. Stocksfield, Acumen.

The analysis based on measuring metric attributes (Fig. 5) reveals that for a particular type of cooking form rim types, the rim diameter is quite varied between the trenches. However, the ones from trench 4 appear to be closer to the types from trench 8 (the kiln), while the types in trench 15 appear to be quite varied. For the rim thickness, the types in trench 15 appear to lie in between the types in trench 4 and 8. The neck thickness on the other hand shows less variability. Tests show that there is variation in the local ceramic repertoire at Karanis.

Hodder, I. 1979. Pottery distributions: Service and tribal areas. In M. Millet (ed.), Pottery and the Archaeologist, Vol. 4, 7-23. London, Occasional Publication, Institute for Archaeology. Lave, J. and Wenger, E. 1991. Situated Learning: Legitimate Peripheral Participation. Cambridge and New York, Cambridge University Press. 128

GUPTA-AGARWAL

Stark, M. T., Bowser, B. J. and Horne, L. (eds.) 2008. Cultural transmission and material culture: breaking down boundaries. Arizona, University of Arizona Press.

Minar, J. and Crown, P. 2001. Learning and craft production: An introduction. Journal of Anthropological Research 4 (57), 369-80. Rice, P. 1987. Pottery Analysis: A sourcebook. Chicago, University of Chicago Press.

Wallaert-Petre, H. 2001. Learning how to make the right pots: Apprenticeship strategies and material culture, a case study in handmade pottery from Cameroon. Journal of Anthropological Research 57, 471-93.

Roe, P. G. 1995. Style, Society, Myth and Structure. In C. Carr and J. E. Neitzel (eds.), Style, Society, and Person: Archaeological and Ethnological Perspectives, 27-76. New York, Plenum Press.

Wendrich, W. 2002. Moving Images in Ethnoarcheology: Using Video in the Study of Ancient Production Processes. In W. Wendrich and G. van der Kooij (eds.), Moving Matters-Ethnoarchaeology in the Near East, 117-124. Leiden/The Netherlands: Research School of Asian, African and Amerindian Studies (CNWS)/ Universiteit Leiden.

Schonpflug, U. (ed.) 2009. Cultural Transmission: Psychological, Developmental, Social and Methodological Aspects. Cambridge, Cambridge University Press. Shennan, S. J. and Steele J. 1999. Cultural Learning in Hominids: A Behaviorial Ecological Approach. In H. O. B. and K. R. Gibson (eds.), Mammalian Social Learning: Comparative and Ecological Perspectives, 367-388. Cambridge, Cambridge University Press.

129

LRCW4

Photo 2. Shisha being dried in the sun

Photo 1. Misa'a displayed on a hearth

Photo 3. A potter making the Khashbooha

130

GUPTA-AGARWAL

Fig. 1. Sayyid the potter with his son making the Khashbooha in the workshop at Fustat, Egypt. Fig. 2. A Google earth map of the Greco Roman site of Karanis showing the excavated areas. Fig. 3. Results quantifying the sherd count from kiln survey which included 8 kilns at Karanis. Fig. 4. Results quantifying the sherd count from the excavated kiln (Trench 8) at Karanis.  

 

Fig. 5. Analysis chart showing the coefficient of variation for rim diameter, rim thickness and neck thickness of a particular cooking pot type manufactured at Karanis.

131

LRCW4

132

AMPHORAE AND SPECIALIZED COARSEWARES OF RAS AL BASSIT, SYRIA: LOCAL PRODUCTS AND EXPORTS PHILIP J.E MILLS1, PAUL REYNOLDS2 ¹School of Archaeology and Ancient History, University of Leicester, University Road, Leicester, LE1 7RH, United Kingdom; [email protected]  2 ICREA Research Professor, ERAAUB, University of Barcelona, C. Montalegre 6-8, Barcelona 08001, Spain [email protected] This paper presents a preliminary typology of the specialized amphorae, mortaria, basins and dolia that were manufactured and exported from Ras el Bassit from the 1st to the 5th centuries AD. Their appearance in stratified deposits in Beirut is used to suggest date ranges as well as their role in interregional trade. KEYWORDS: BASSIT, SPECIALIST PRODUCTION, MORTARIA, DOLIA, AMPHORAE, BASINS Much of this material is made up of the by-products of the extensive local production of mortaria, basins, dolia, amphora and kitchen wares. In addition much can be learnt from the large quantities of unstratified material found all over the settlement, as well as a number of type pieces recovered from Courbin’s excavations in the 1970s.

Introduction Ras el Bassit is located on the Syrian coast, some 50km south west of Antioch and 220km north of Beirut (Fig. 1). It is well known as the source of a distinctive 3rd century mortarium which was commonly exported to sites on the Levantine coast – with some examples even reaching the West (Hayes 1967; Vallerin 1994). The current project, begun in 2000, is concentrated on the standing remains of a church and two buildings outside the church complex (Fig. 2). The site is of a 5th century synagogue, which was remodelled as a church in the mid 6th century (Mills and Beaudry 2010; Beaudry 2005a; 2005b; 2004; Courbin 1986; 1978). The church went through significant refurbishments, probably in the later 6th century. There seems to be a general abandonment in the early 7th century, with partial squatter occupation of the semi ruinous structure before its final collapse. The ruins of the apse were cleared and developed as a small chapel during the Frankish period, according to the glass and coin finds.

Local amphorae Figure 3, nos 1-4, are Beirut examples of Bassit Form 1. These have a pale red, sometimes yellowish fabric, differing substantially from the typical dense, heavy, dark red or orange-brown fabric of the 4th to 7th centuries. Inclusions of quartz, basalt and lime are common in both fabrics, but sparser and finer in the earlier examples. John Hayes (1991) correctly suggested that this small flagon-like vessel with a ring foot and band rim was from Bassit. Two of many examples recovered from Bassit are shown in Figure 3, nos 4 and 5. It is the most commonly exported Bassit amphora to Beirut in the 2nd century, though is still relatively rare in Beirut. Note that two sizes of the shape were produced.

This paper presents new evidence for the range of amphorae and specialized forms produced in Bassit from the early Imperial to late Roman periods, many of which are attested as exports in Beirut, which has been used as a source of good supplementary dating evidence for these Bassit products.

We can now add to this a good Bassit version of the Koan type (Fig. 3, no. 6: Bassit Form 3), from the Courbin excavations. Another new type is a large amphora, Figure 3, no. 5, also in the early Imperial fabric, which has the same handle type as the flagon form, though obviously the handles are larger (Bassit Form 2).

The pottery To date, some 43,100 sherds including over 3,416 rims have been catalogued at Bassit. The material derives mainly from makeup layers, derived from rubbish middens from outside the settlement but also a mixture of contemporary pottery, make up and hardcore used for the first floor balconies contemporary with the early 6th century remodeling.

Two other vessels, Figure 3, nos 7 and 8, also from the Courbin excavations, may be two variants of another new form, and are perhaps mid Roman on the basis of their fabrics (Bassit Form 4A-B). Figure 4, no. 9 shows the already published vessel from Beirut, dating to the late 4th century, in the classic late fabric, with a rather globular body (Reynolds 2005, Plate 6, fig. 38) (Bassit 133

LRCW4

Mortaria

Form 5). Figure 4, no. 10, from Bassit, now provides the rim for this type, but its base is still unknown. The handles are distinctive in the way they are folded, leaving a deep central groove, a feature equally typical of Cilician early LRA 1s of the 3rd and 4th centuries (Reynolds 2005, 565-567, Plate 3; 2009, fig. 3). Handles of this type litter parts of the site. It was clearly a very common form.

Figure 6, nos 22 and 23, show rare Beirut examples of 2nd century Bassit mortaria, none of which have been recovered from the Bassit excavations to date. Typologically these stand apart from the far more important production of mortaria in Bassit that begins in the early 3rd century in the pale red Imperial fabric (Fig. 6, no 24). The fabric is dark brown, although earlier examples tend to be paler. It is hard with an irregular fracture and harsh sandy feel. It has inclusions of abundant basalt at 0.6mm and common subrounded lime at c. 0.3mm in a sandy matrix. The shape, the existence of Latin stamps, and Roman names in Greek are highly unusual in the East. Early 4th century examples have smaller rims (Fig. 6, no. 25). By the late 4th century the type, now in classic late fabric, is only a rare import in Beirut, sometimes being decorated with stamped circles (Fig. 6, nos 26 and 27).

Finally, the amphora production of Bassit appears to end with a version of LRA 1 (Fig. 4, nos 11 and 12), a previously unattested product that is clearly modelled on the early 5th century version of the Cilician form (Reynolds 2005, fig. 31). Unlike the latter, the Bassit version appears not to have been made for export. There are no examples in Beirut. So far no later forms of the LRA 1 type, of late 5th to 7th century date, nor LRA 13s have been identified in Bassit (or Beirut) and this would seem significant given the large excavated sample covering this period. The production for export of LRA 1 was confined to Cilicia and Cyprus and not Syria (a conclusion already proposed in Reynolds 2005, 565-567).

In the early 5th century these were replaced by another Bassit mortar shape with its distinctive bevelled squarish rim, straight walls and flat base (Fig. 7, nos 28 and 30). A small spout is preserved on some examples. The rims become thinner through the 5th century (Fig. 7, no. 29). Whereas imports of this form are common in Beirut throughout the 5th century, but cease after c. 500, the shape would seem to continue in Bassit. The larger vessels with squared rim illustrated here (Fig. 7, nos 32 and 33) may be large mortars or basins.

Transport dolia We now come to another class of vessel we have named ‘transport-dolia’. A complete example of this large, thick-walled form, with Koan-style rim, stepped neck and large double rod handles, is on display in the Antioch Museum. Vessel top and bases were recovered from the Fig Tree Bay Ship Wreck, off eastern Cyprus (Fig. 5, nos 13 to 15: Leidwanger 2010, fig. 7). The Koan-style top and handles suggests the form carried wine, for export in this case. Perhaps wine was fermented in these vessels prior to shipment. The latest date one could attribute to this wreck is the early 3rd century. A vessel published from Sinai is not an amphora as stated, but rather a transport dolium of this type (Fig. 5, no. 16 = Arthur and Oren 1998, 203, fig. 4, no. 6). These occur in the late Roman period in Sinai. The neck is shorter than the 3rd century form.

Basins Figure 8, no. 34 shows a deep basin with large strap handles, which appears in Beirut mid 3rd century contexts. Smaller, more open vessels were particularly common in the 5th century (Fig. 8, nos 35-38), recalling the pattern of Bassit mortar imports. Figure 8, no. 39 (early 6th century), a large form with a deep lid seat, is similar to the 3rd century form. From the late 5th century, however, it was Amrit that provided the majority of basins, to not only Beirut but also to Homs and North Palestine, in turn reflecting the supply of Amrit dolia to the same sites (Fig. 8, no. 40).

Two examples from Bassit, in the classic late fabric, are of the later form (Fig. 5, nos 17 and 18). A large number of similar rims and handles scattered across the site should also be classified as dolia rather than amphorae, again perhaps evidence for the large scale fermentation of wine.

Conclusion In conclusion, alongside the more widely acknowledged mortaria, Bassit was producing a large number of other specialized vessels. Amphorae, apparently for wine, were produced for export from the Imperial period until the late 4th century. The importance of locally produced wine is underlined by the transport dolia defined here, these vessels probably serving both for the fermentation of the wine on site and for its export. Whereas Cilicia exported its wine in ever increasing quantities in LRA 1 from the late 4th century onwards, Bassit only produced its version of the type in the early 5th century and this was not for export.

Dolia Bassit dolia with large collar rims (primarily for storage of water?) are occasionally found in Beirut in the 3rd and 4th centuries (Fig. 5, nos 19-21) but were superseded in the 5th to 7th centuries by Amrit-Tartus dolia, which are, in contrast, quite common in Beirut, Homs and North Palestinian sites.

134

MILLS-REYNOLDS

Instead we find that from the 3rd to the 7th centuries Bassit was engaged in the export to Levantine coastal sites of mortaria, first, and then basins. This reached its peak in the 5th century, after which Bassit lost this market to Amrit.

Courbin, P. 1986. Bassit. Syria 63, 175-220. Hayes, J. W. 1991. Paphos III. The Hellenistic and Roman pottery. Department of Antiquities, Cyprus, Nicosia. Hayes, J. W. 1967. North Syrian mortaria. Hesperia 36.4, 337-347.

Acknowledgements

Leidwanger, J. 2010. Amphoras from an early Imperial shipwreck at Fig Tree Bay, Cyprus. International imports and local imitations. Rei Cretariae Romanae Fautorum Acta 41, 9-15. Bonn.

Research on late antique Bassit is currently supported by a grant from the Social Sciences and Humanities Research Council from Canada, and excavation in the church was funded in 2007 by a project grant from Dumbarton Oaks. We would like to thank Professor Nichoas Beaudry for inviting PM to study the pottery and inviting PR as a consultant in the 2010 season. We wish to thank the Directorate General of Antiquities (Damascus), the Museum of Latakia, and Prof. Jacques Y. Perreault (Université de Montréal) for their interest and active support in the excavation and study of late antique Ras el Bassit. Finally, a word of thanks to Patricia Francis for casting her eye over the English of this text.

Mills, P. J. E., and Beaudry, N. 2007. The ceramic coarse wares from the basilica excavations at Ras el Bassit, Syria: a preliminary assessment. In M. Bonifay and J.-C. Tréglia (eds.), LRCW 2. Late Roman Coarse Wares, Cooking Wares and amphorae in the Mediterranean. Archaeology and Archaeometry. British Archaeological Reports, International Series 1662, 745-754. Oxford, BAR Publishing. Mills, P. J. E. and Beaudry, N. 2010. Pottery from Late Roman Ras el Bassit, Syria. In S. Menchelli. S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, Cooking Wares and amphorae in the Mediterranean. Archaeology and Archaeometry. Comparison between Western and Eastern Mediterranean. British Archaeological Reports, International Series 2185, 857-866. Oxford, BAR Publishing.

Bibliography Arthur, P. and Oren, E. R. 1998. The North Sinai survey and the evidence of transport amphorae for Roman and Byzantine trading patterns. Journal of Roman Archaeology 11, 193-212. Beaudry, N. 2004. La basilique de Ras el Bassit, une église paléochrétienne sur la côte de Syrie du Nord. PhD diss., Université de Paris-I Panthéon-Sorbonne and Université de Montréal.

Reynolds, P. 2005. Levantine amphorae from Cilicia to Gaza: a typology and analysis of regional production trends from the 1st to 6th centuries. In J. M. Gurt i Esparraguera, J. Buxeda I Garrigós, and M. A. Cau Ontiveros (eds.), LRCW1. Late Roman Coarse Wares, Cooking Wares and amphorae in the Mediterranean.Archaeology and Archaeometry. British Archaeological Reports, International Series 1340, 563-612. Oxford. BAR Publishing. Reynolds, P. 2009. Linear typologies and ceramic evolution, FACTA 2 (2008), 61-87.

Beaudry, N. 2005a. Formes architecturales et géographie historique: l'église de Bassit et le corpus nord-syrien. In F. Baratte, C. Jolivet-Lévy, and C. Morrisson (eds.), Mélanges offerts à Jean-Pierre Sodini, Travaux et Mémoires 15, 119-136. Paris, Centre d'histoire et de civilisation de Byzance. Beaudry, N. 2005b. Un autel et son reliquaire à Ras el Bassit (Syrie du Nord). Hortus Artium Medievalium 11, 111-122.

Vallerin, M. 1994. Pelves estampillés de Bassit, Syria 71, 171-204.

Courbin, P. 1978. Une nouvelle fouille française sur la côte syrienne: a-t-on retrouvé l'antique Posideion à Ras el Bassit? Archeologia (Paris) 116, 48-62.

135

LRCW4

Fig. 1. Sites mentioned in the text

Fig. 2. Site plan

136

MILLS-REYNOLDS

Fig. 3. Early Imperial Bassit amphorae 137

LRCW4

Fig. 4. Early Byzantine Bassit amphorae

138

MILLS-REYNOLDS

Fig. 5. Transport and storage dolia 139

LRCW4

Fig. 6. Bassit mortaria, 2nd to 4th centuries

140

MILLS-REYNOLDS

Fig. 7. Bassit mortaria and mortaria/basins, 5th to 6th centuries

141

LRCW4

Fig. 8. Bassit and Amrit-Tartus basins 142

MICACEOUS WATERJARS IN PERGAMON – A SPECIFIC VARIANT OF THE LATE ROMAN 3 AMPHORA SARAH JAPP Orient Department, German Archaeological Institute [email protected] In Pergamon different types of amphorae reflecting both local production and imports are common in Roman contexts. Additionally, a specific type, the micaceous Pergamenian Waterjar was present in large numbers. Its general shape is reminiscent of the early LR 3 Amphora, but there are some distinct differences with the most prominent being a kind of bloated neck (“Blähhals”). Two different versions of the Pergamenian Waterjar can be distinguished regarding the form and the surface treatment. These vessels can be dated to the 2nd and 3rd centuries AD contextually. They might even continue into the 4th century but the stratigraphy is not diagnostic enough to confirm this. Four more types of similar shape have been found in Pergamenian complexes. Together these seem to build a kind of development line towards the two main Pergamenian Waterjar types. Comparisons providing a reliable basis for the dating of the Pergamenian variant could not be found in publications of Asia Minor or other Mediterranean sites. Therefore the question arises whether these waterjars were produced in Pergamon, a site known for its capacious fine ware production in Hellenistic and Roman times. A tentative positive conclusion to this question seems to be possible, given current archaeometric analysis. KEYWORDS: PERGAMON, LOCAL POTTERY PRODUCTION, MICACEOUS WATERJARS, LR 3 AMPHORA, PERGAMENIAN WATERJAR material, which offer the possibility of a chronological framework. However, the living quarters were constantly used from Hellenistic through Roman to late Byzantine times and therefore the stratigraphy is difficult. Some of the pottery stems – due to the settlement development – from Bronze Age, Archaic and Classical times but most can be dated between the 3rd century BC and the 3rd/4th century AD. Finds from the late 4th to 6th century AD are rare as the hillside was mostly abandoned. A little more is known from the 7th to 8th century AD, as in the 7th century AD a defensive wall was erected to surround the upper city on the acropolis and serve as a refuge area for the inhabitants. In AD 716 Pergamon was conquered by the Arabs and its residents enslaved which depopulated the area. From the 8th until the 10th century AD there is a gap with re-settlement not taking place before the 10th century AD. From the 12th to 14th century AD several building activities can be observed, a development which is reflected in the finds (Byzantine living quarters in the Stadtgrabung: Rheidt 1991; Byzantine pottery: Spieser 1996).

The ancient city of Pergamon is situated on the western coast of Turkey not far from modern Izmir (historical and archaeological overview given by Radt 1999). The area was settled from prehistoric times onwards. During the Hellenistic period it was the capital of the Attalid kingdom. After the death of the last Attalid ruler the realm became part of the Roman province of Asia and Pergamon lost its function as a capital but stayed an important and wealthy city. During the early Byzantine period it declined but experienced a short revival in the late Byzantine period. The Hellenistic city extends over a natural hill with the main buildings concentrated on the acropolis, while several newly built Roman structures such as living quarters and entertainment structures were located in the southern plain. In Pergamon not only was the acropolis uncovered with its temples and other cultic buildings, the theatre, the arsenals and the palaces of the Hellenistic kings (Bohn 1885; Stiller 1895; Bohn 1896; Schrammen 1906; Conze 1912; Kawerau and Wiegand 1930; Boehringer and Krauss 1937; v. Szalay and Boehringer and Krauss 1937) but also parts of the southern hill slope with important public structures such as the Great Gymnasium (Schazmann 1923; Bohtz 1981). Additionally parts of the ancient living quarters were also uncovered, the latter coming to light mainly in the so called Stadtgrabung. The Stadtgrabung was conducted by the German Archaeological Institute between 1973 and 1993. Since 2005 sondages were sunk on the eastern hill slope (see preliminary reports of W. Radt and F. Pirson in Archäologischer Anzeiger since 1974). These excavations provide several closed complexes with associated

In Hellenistic as well as in Roman contexts several different types of jugs, jars and amphorae came to light. A lot of them were produced locally or regionally, and investigations are taking place currently. During the Hellenistic period imports from the Greek islands such as Thasos, Cos, Rhodos, Chios as well as from Knidos and the Black Sea Region are known (Börker and Burow 1998). In the Roman period the range of imports seems to be more extensive as examples from northern Africa and several sites in the eastern Mediterranean appeared such as specimens with funnel-like rims, Late Roman 3 Amphorae and Hollow-Foot Amphorae. 143

LRCW4

The specific type of amphorae to be presented in this article was found in middle to late Roman contexts. Due to an obvious similarity to amphorae published by Lang and Robinson from excavations in Athens (micaceous waterjars; Lang 1955; Robinson 1959) the term Pergamenian Waterjar was established (see MeyerSchlichtmann 1992, 300, no. 10). Its general shape is reminiscent of the early types of the Late Roman 3 Amphora: an oval-shaped body tapering downwards, a rounded shoulder and a cone-shaped neck, a rounded rim, ear-shaped handles with a groove and a short small foot. A broad shallow ribbing covers most of the vessel’s exterior surface. Several versions can be differentiated in Pergamon with two main types distinguished, which are almost without (published) comparisons in sites of the eastern Mediterranean. Therefore a Pergamenian production was assumed (Meyer-Schlichtmann, 1992, 300, no. 10; 1999, 219, no. 13. 224, no. 32. 228, no. 45; Japp 2008, 266, K 78; 2009, 204-5). The other types are extant in smaller numbers and not all of them were necessarily manufactured in Pergamon, but nevertheless they belong to the same type of amphora and together form an independent and closed group.

and the typical collar of the feet of Type 1 is also missing. Type’s2 foot can be described as knob-like (Fig. 4). The neck is much shorter than the Type 1 neck although still bulbous in a bloated manner. The rim is out-turning and thickened (Fig. 5). The handle resembles Type 1 handles where the central ridge and the location are concerned. The body shows a broad horizontal ribbing, more pronounced than Type 1, and a white to beige-coloured coat. This is matt but dense and covers the whole exterior surface. The clay is tempered with a large amount of fine silvery micaceous particles and some other small-sized inclusions. Some of the inclusions are of a whitishyellowish, some of a red and some of a grey to black colour (macroscopically they can be interpreted as lime and chamotte particles as well as sandy inclusions). The colour of the clay may be described as orangish brown (5 YR 6/6 to 7/8). The body shards are not as thin-walled as Type 1 shards and the outer surface seems to be rougher. The average height measures 50–57cm, the rim diameter 7cm and the foot diameter 4–5cm. The following four types appear in smaller numbers in Pergamon and designation as a Pergamenian production is not mandatory. 1 However, all these types belong together in a kind of development line concerning the general idea of the vessel and its shape, and they are presented here in a proposed chronological order.

Type 1 (Fig. 1 and 2) comprises the majority of fragments (Meyer-Schlichtmann 1999, 224, no. 32, fig. 87. 229, no. 46, fig. 89; Japp 2008, 267, K 80, fig. 64. 278, K 148, fig. 72. 278, K 151, fig. 72). The body is oval-shaped but rather long and the widest diameter is situated in the middle of the body. There is no pronounced shoulder. The neck is separated from the body by a ridge; it is cone-shaped in principle but looks bloated (“Blähhals”). The rim is out-turning and the lip is thickened and rounded on the upper part with a sharp ridge at the transition of rim and neck in most cases. The vessels are one-handled with the handle starting at the junction between neck and body and ending shortly below, giving the handle an ear-like appearance. It shows a deep groove on the exterior. The body tapers towards the hollow foot which is short and equipped with a collar on the exterior. The outer surface of the belly is characterized by a broad horizontal ribbing and a dark red, violet or brown matt coat. This coat does not cover the whole body mostly excluding the lower part and the foot. Sometimes drops running down to the foot are visible. The clay is of a reddish brown or an orangish brown colour (7.5 YR 6/46). The visual characteristics can be described as hard, fairly thin-walled and showing a lot of fine silvery micaceous inclusions. In addition a small to modest amount of fine, whitish and black particles is visible (macroscopically these could be interpreted as lime particles and sandy inclusions). The clay is hard-fired and the vessels are thin-walled. The usual height is about 5055cm, while the average rim and foot diameter are 3-4cm.

Type 3 (Fig. 6) is not oval-shaped but more roundbodied. The diameter is largest around the upper part of the body. The shoulder is rounded and more pronounced in comparison to Types 1 and 2. The body tapers towards a foot much bigger in diameter than those of either Types 1 or 2. However, the foot itself is short and shows a thickened collar. The neck is narrow in relation to the body size and straight with a transitional groove between rim and neck. A curve is not visible on examples of this type. The lip tapers and is flattened on the exterior. The vessels are one-handled; however, the appearance of the strap-handle is different as it is flat and displays one central ridge. The handles of Types 1 and 2 are comparable in appearance and location. Type 3 handles were attached at the junction between rim and neck and ended on the shoulder shortly below. The body displays a light brown to beige brown colour and was probably not covered by any kind of wash. The broad ribbing is much shallower than on Types 1 and 2. The clay is of a light brown to ochre colour (7.5 YR 5/6) and again tempered with numerous micaceous particles. Moreover some fine white inclusions (possibly lime) are visible. The body is fairly thin-walled. The average height amounts to approximately 47–50cm, the average rim diameter measures 5cm and the foot diameter 7–8cm. 1

It seems highly plausible that most of the examples of types 3, 4 and 5 found in Pergamon stem from the Ephesos production. This assumption was supported by T. Bezeczky (Wien) who directed the FWF project "Identification of Amphora types of Ephesos” and by the ongoing fabric analysis conducted by R. Sauer (Wien). Nevertheless, some of the pieces seem to have a Pergamenian origin due to the current research. See for comparisons Bezeczky 2005, 203-8; Ladstätter 2008, 180, fig. 26; 181, fig. 27.

The second most frequent type, Type 2 (Fig. 3), again displays an oval-shaped body, though it is more slender than the one of Type 1 (de Luca 1984, 37, no. 331-2, pl. 15; Meyer-Schlichtmann 1992, 300, no. 10, fig. 3; 1999, 219, no. 13, fig. 85. 228, no. 45, fig. 89; Japp 2008, 266, K 78, fig. 64. 267, K 81, fig. 64. 277, K 145, fig. 72). It tapers towards the hollow foot although less than at Type1. The foot is a little smaller than the Type 1 foot 144

JAPP

Type 4 (Fig. 7) shows some similarities to Type 3. However, the body is more slender and ovoid and tapers distinctly towards the foot, which is smaller than that of Type 3 but still bigger than that of the two main types. Nevertheless, the foot is not marked by any kind of collar or groove. The neck is still narrow in relation to the body size. It is short, cone-shaped and narrowing towards the rim. The lip is thickened and pointed on the exterior. The strap-handle is irregularly oval-shaped with a shallow groove on the upper part. Its location is comparable to the other types. Examples with a shallow ribbing and a dark reddish brown coat exist as well as examples without ribbing and coating. The fabric shows less micaceous particles than the other types and some medium-sized white inclusions (perhaps lime). The colour of the clay can be described as orangish red (2.5 YR 4/8-5/8), while the exterior surface looks reddish brown. The body walls are not as thin-walled as those of Type 3. The average height is 45–47cm, the rim diameter measures 4.5–5cm and the foot diameter 5.5–6cm. The main difference to Type 3 is the more oval-shaped body and the neck which had already started to show a kind of curve.

treatment. The neck is short and straight, while the rim is out-turning and convex-shaped; the lip is in-curving and rounded. Within the excavations in Pergamon one single vessel (Fig. 11) came to light which also belongs to this group of amphorae (comparable pieces in Ladstätter 2008, 181, no. 10. 17. 20, fig. 27). The body is long and oval-shaped while the largest diameter is located in the upper part. The shoulder is sloping. The neck is definitely longer than those of the other types and is clearly cone-shaped. The lip is out-turning and thickened on the exterior. The body tapers towards the foot which is thin-walled and hollow. The handle starts at the junction of neck and shoulder but the lower part is not incurving and ends on the shoulder. The body shows a ribbing but it is shallow and irregular. A coat is missing. The surface and the clay are of a light brick-red colour (2.5 YR 5/8), while the clay is tempered with a lot of fine silvery micaceous particles and some fine yellowish particles (possibly lime). The height is 49cm, the rim diameter measures 3.6cm and the foot diameter 4cm. This piece was most likely produced in Ephesos (assumption by T. Bezeczky, Wien; supported by the preliminary archaeometric results of R. Sauer, Wien).

Type 5 (Fig. 8) appears to be a mixture of Types 3 and 4 and displays the next step on the development line towards the Pergamenian Types 1 and 2. The body is even more slender and oval-shaped with the largest diameter in the upper third of the body. The neck is short, cone-shaped and tapering towards the rim. The lip is thickened on the exterior and slightly in-turning. Its upper part is flattened. The lower part of the body clearly tapers towards the short and hollow foot which is pronounced with a shallow collar. The strap-handle with the small but deep groove in the centre already shows close resemblances to Types 1 and 2. The distinct ribbing covers the whole body, not only the central part. Moreover, the body is covered by a dark irregular coat (the example shown in the photograph is discoloured because of firing). The clay is normally of a reddish brown colour and tempered with a lesser amount of micaceous particles than the other types. Additionally some light-coloured fine inclusions (perhaps sand and/or lime) are visible. The walls are not as thin-walled as other examples. The height measures around 50–55cm, the rim diameter around 4–5cm and the foot diameter 5–6cm.

In Pergamon it is known that Hellenistic and Roman Table Ware was produced locally. In rescue excavations in the 1970s a potters’ quarter situated in the valley of the river Ketios at the foot of the citadel hill was excavated. Unfortunately the excavations were published only in short preliminary reports and only Table Ware is mentioned (Erdemgil 1980; 1981a; 1981b; Erdemgil and Ozenir 1982; Bounegru and Erdemgil 1998; Poblome et al. 2001). Therefore it is still unknown whether coarse ware, jugs or amphorae were produced also. Furthermore it is possible that Utilitarian Ware was manufactured somewhere else in or near Pergamon. In 2006 an archaeological - archaeometric programme was commenced with the aim of gaining more confidence in the identification of locally produced Pergamon wares (Japp 2009; Mommsen and Japp 2009; Schneider and Japp 2009). Different types of Hellenistic and Roman Table Ware as well as some examples of Utilitarian Ware from Pergamon were sampled and examined by X-ray fluorescence (WD-XRF) and neutron activation analysis (NAA). The NAA analysis was undertaken at the Rheinische Friedrich–Wilhelms–Universität Bonn, Helmholtz-Institut für Strahlen- und Kernphysik under the direction of H. Mommsen. The WD-XRF analysis was carried out at the Freie Universität Berlin under the direction of G. Schneider. Heavy mineral analysis was chosen additionally to put this thesis on a firmer footing. The samples were taken by T. Bezeczky (Wien). As a result several chemical groups became apparent although two main groups emerged, which can be assigned to the Pergamenian production. Samples were taken from the Pergamenian Waterjar Types 1 and 2 and Types 5 and 6. 2

Type 6 (Fig. 9) has a long and oval-shaped body that tapers distinctly towards the ring foot, which is marked by a ledge. The handle is characterized by a deep groove on the exterior. The body displays a regular and pronounced ribbing which is confined to the central body and missing on the shoulder and the foot area. The exterior surface is only partly covered by a dark brown coat that runs down in long drops. The clay has a dark reddish brown colour and shows a lot of fine silvery micaceous particles. Other inclusions cannot be detected so far. This type of waterjar is very thin-walled and the fabric seems to be different as the surface gives a soapy impression.

2

One example of each type has already been published. Type 1: Japp 2009, 244, Perga 106, fig. 20; Type 2: Japp 2009, 241, Perga 102, fig. 12; Type 5: Japp 2009, 241, Perga 105, fig. 20; Type 6: Japp 2009, 241, Perga 104, fig. 20. The other samples

Among the known body fragments no neck and rim section is preserved. However, there is one rim type (Fig. 10) which fits well concerning fabric and surface 145

LRCW4

while Type 4 can be compared to her types 1 (Lang 1955, pl. 79e) and 2 (Lang 1955, pl. 79f-g) even though Lang’s type 1 is not micaceous. Here the general shape is of an ovoid body, short neck with out-turning lip and similar handles and foot. Both types are dated to the 1st century AD. The Pergamenian Type 5 may be related to Lang’s type 3 (1955, pl. 79h) even if the neck is longer on the Athenian piece and the lower part of the body not so tapering downwards. It belongs to the early 2nd century AD. One Athenian specimen (Lang 1955, pl. 79p: late type 2) looks like a mixture of the Pergamenian Types 1 and 2. With its bloated neck it resembles Type 2 but in its body and foot shape and dark coat it is closer to Type 1. In general Lang’s late type 2 (1955, 278) belonging chronologically to the 2nd and 3rd centuries AD seems to be the best comparison for the Pergamenian Waterjars.

The NAA classified three pieces (Types 1, 5 and 6) as singles, while two fragments belonging to Type 1 comprised a couple. Fortunately the XRF assigned two fragments of Type 1 and one fragment of Type 2 to the main Pergamon group while one piece of Type 2 could perhaps belong to the Pergamon group. One piece was a single. These results support the thesis, that Types 1 and 2 were produced in Pergamon. In principle these vessels can be dated in Pergamon between the 1st and the 3rd or even 4th centuries AD contextually. The chronological frame of the early Late Roman 3 Amphorae or Micaceous Waterjars can be used as a model as is shown in the Agora publication of Robinson (1959). Type 3 seems to be the earliest variant, which could be dated to the 1st century AD, probably to the Augustan period. The Agora examples F 65 (Robinson 1959, 17, pl. 2) and especially F 66 (Robinson 1959, 17, pl. 2) are comparable, which were dated according to the find complex into the last three quarters of the 1st century BC. Typical of these examples are the rounded pronounced shoulder, the short broad foot, the straight slightly cone-shaped neck with the thickened rim, and ear-shaped handles with a central groove. These vessels were probably made at a number of production centres in the coastal areas of Asia Minor, including for example Ephesos (Ladstätter 2008, 180-2). Type 4 could be the first Pergamenian transformation, perhaps from the 1st or early 2nd century AD. Here the shoulder is not as pronounced anymore and the body is more ovoid and the foot much smaller, although the rim and neck remain the same. Similar to Type 4 are Agora J 46 (Robinson 1959, 55-56, pl. 11) placed chronologically between the middle of the 2nd and the early 3rd century AD and Agora M 45 (Robinson 1959, 88, pl. 20) dated to the beginning of the 2nd century AD. In both comparisons the lower body part is more tapering than Type 4. Type 5 seems to follow Type 4 belonging to the 2nd century AD. Here the body is slender and more cylindrical, but the small foot and the cone-shaped neck with the ear-shaped handle are comparable to Type 4. It seems to be a mixture of Agora J 46, M 45 and M 125 (Robinson 1959, 95, pl. 23), dated between the early and mid-2nd century AD. The body of Type 5 is longer as the one of J 46. The necks of the other two examples M 45 and M 125 are longer and narrower compared to Type 5. Neither Types 1 and 2 nor the single have comparisons in the Agora material presented by Robinson. Moreover, the Agora specimens with two handles dated between the 4th and the early 6th century AD are missing in the Pergamenian find complexes – which is comprehensible due to the settlement development. It is highly likely that similar types could be found in the partially excavated sanctuary of Asklepios and in the Roman city of Pergamon situated in the southern plain as both areas had been settled during that period.

For the Pergamenian Types 1 and 2 a chronological frame between the first half of the 2nd and the whole 3rd century AD was assumed, but they could even come down into the 4th century AD. It may be that specimens of Types 1 and 2 with less bulbous necks can be dated earlier than vessels with more bulbous necks. But the find complexes in Pergamon are mostly not differentiated enough to allow such fine chronology at the moment. The assignation of Types 1 and 2 to local production in Pergamon is based on their occurrence in Pergamon and recent archaeometric analysis, but also on the nonexistence of parallels in scientific publications, so new information could change the picture. Nevertheless, it is possible that via trade and imports the idea of the Micaceous Waterjars/early Late Roman 3 Amphorae arrived in Pergamon and convinced the Pergamenian potters to produce a similar kind of vessel locally. Unfortunately no information concerning the content can be given at present. Manufacturing started perhaps in the 1st century AD with some of the earlier Types 3–5, but certainly in the 2nd century with Types 1 and 2 and lasted until the 3rd or even the beginning of the 4th century AD. Bibliography Bezeczky, T. 2005. Late Roman Amphorae from the Tetragonos - Agora in Ephesus. In: F. Krinzinger (ed.), Spätantike und mittelalterliche Keramik aus Ephesos, Archäologische Forschungen 13, 203-229. Wien, Verlag der Österreichischen Akademie der Wissenschaften. Boehringer, E. and Krauss, F. 1937. Das Temenos für den Herrscherkult, Altertümer von Pergamon IX. Berlin, Walter de Gruyter. Börker, Ch. and Burow, J. 1998. Die hellenistischen Amphorenstempel aus Pergamon, Pergamenische Forschungen 11. Berlin – New York, Walter de Gruyter.

In the publication of M. Lang on dated jars from Athenian wells her predecessors of the 1st century AD (Lang 1955, pl. 79b-d) seems to be similar to Type 3,

Bohn, R. 1885. Das Heiligtum der Athena Polias Nikephoros, Altertümer von Pergamon II. Berlin, Walter de Gruyter.

mentioned will be published in the next report of the archaeometric analysis in Istanbuler Mitteilungen.

146

JAPP

Bohn, R. 1896. Die Theater-Terrasse, Altertümer von Pergamon IV. Berlin, Walter de Gruyter.

Meyer-Schlichtmann, C. 1992. Neue Erkenntnisse zum ‘Heroon des Diodoros Pasparos’ in Pergamon. Istanbuler Mitteilungen 42, 287-306.

Bohtz, C. H. 1981. Das Demeter-Heiligtum, Altertümer von Pergamon XIII. Berlin, Walter de Gruyter.

Meyer-Schlichtmann, C. 1999. Keramik aus datierenden Befunden im Stadtviertel zwischen Hauptstrasse, Mittelgasse und Ostgasse. In U. Wulf, Die Stadtgrabung. Die hellenistischen und römischen Wohnhäuser von Pergamon, Altertümer von Pergamon XV 3, 215-229. Berlin – New York, Walter de Gruyter.

Conze, A. 1912. Stadt und Landschaft, Altertümer von Pergamon I. Berlin, Walter de Gruyter. Bounegru, O. and Erdemgil, S. 1998. Terra Sigillata Produktion aus den Werkstätten von PergamonKetiostal. Vorläufiger Bericht. Istanbuler Mitteilungen 48, 263–277.

Mommsen, H. and Japp, S. 2009. Neutronenaktivierungsanalyse von 161 Keramikproben mit Fundorten aus der Region Pergamons. Istanbuler Mitteilungen 59, 269-286.

Erdemgil, S. 1980. Kestel kurtama kazısı. In II. Kazı Sonuçları Toplantısı 1980, 103-107. Turkish Ministry of Culture, Ankara. Erdemgil, S. 1981a. Kestel kazısı 1980 yılı calıçmaları. In III. Kazı Sonuçları Toplantısı 1981, 63-66. Turkish Ministry of Culture, Ankara.

Poblome, J., Bounegru, O., Degryse, P., Viaene, W., Waelkens, M. and Erdemgil, S. 2001. The Sigillata manufactories of Pergamon and Sagalassos. Journal of Roman Archaeology 14, 143-165.

Erdemgil, S. 1981b. Kestel, 1980. Anatolian Studies 31, 193.

Radt, W. 1999. Pergamon, Geschichte und Bauten einer antiken Metropole. Darmstadt, Wissenschaftliche Buchgesellschaft.

Erdemgil S. and Ozenir, S. 1982. Preliminary Report on the Kilns excavated in Ketios Valley. Rivista di Archeologia 6, 109.

Rheidt, K. 1991. Die Byzantinische Wohnstadt, Altertümer von Pergamon XV 2. Berlin – New York, Walter de Gruyter.

Japp, S. 2008. Datierende Keramik aus dem Podiensaalgebäude. In H. Schwarzer, Das Gebäude mit dem Podiensaal in der Stadtgrabung von Pergamon. Studien zu sakralen Banketträumen mit Liegepodien in der Antike, Altertümer von Pergamon XV 4, 251-307. Berlin – New York, Walter de Gruyter.

Robinson, H. S. 1959. Pottery of the Roman Period. Chronology. The Athenian Agora V. Princeton New Jersey, The American School of Classical Studies of Athens. Schazmann, P. 1923. Das Gymnasion. Der Tempelbezirk der Hera Basilea, Altertümer von Pergamon VI. Berlin, Walter de Gruyter.

Japp, S. 2009. Archäometrisch-archäologische Untersuchungen an Keramik aus Pergamon und Umgebung. Istanbuler Mitteilungen 59, 193-268.

Schneider, G. and Japp, S. 2009. Röntgenfluoreszenzanalysen von 115 Keramikproben aus Pergamon, Çandarlı, Elaia und Atarneus (Türkei), Istanbuler Mitteilungen 59, 287-306.

Kawerau, G. and Wiegand, Th. 1930. Die Paläste der Hochburg, Altertümer von Pergamon V 1. Berlin, Walter de Gruyter.

Schrammen, J. 1906. Der Große Altar. Der Obere Markt, Altertümer von Pergamon III 1. Berlin, Walter de Gruyter.

Ladstätter, S. 2008. Überlegungen zur lokalen Amphorenproduktion spätantik - byzantinischer Zeit in Ephesos. In M. Steskal and M. La Torre (eds.), Das Vediusgymnasium in Ephesos, Forschungen in Ephesos XIV/1, 180-183. Wien, Verlag der Österreichischen Akademie der Wissenschaften.

Spieser, J.-M. 1996. Die byzantinische Keramik aus der Stadtgrabung in Pergamon, Pergamenische Forschungen 9. Berlin – New York, Walter de Gruyter.

Lang, M. 1955. Dated Jars of Early Imperial Times. Hesperia 24, 277-285.

Stiller, H. 1895. Das Trajaneum, Altertümer von Pergamon V 2. Berlin, Walter de Gruyter.

De Luca, G. 1984. Das Asklepieion. Via Tecta und Hallenstraße. Die Funde, Altertümer von Pergamon XI 4. Berlin, Walter de Gruyter.

v. Szalay, A., Boehringer, E. and Krauss, F. 1937. Die hellenistischen Arsenale, Altertümer von Pergamon VIII 3. Berlin, Walter de Gruyter.

147

Fig. 1 Pergamenian waterjar type 1 (photograph by the author)

Fig. 2 Pergamenian waterjar type 1 (drawing by the author; 1:10)

Fig. 4 Foot of a Pergamenian waterjar type 2 (drawing by the author, 1:4)

Fig. 3 Pergamenian waterjar type 2 (photograph by the author) Fig. 5 Upper body part of a Pergamenian waterjar type 2 (drawing by the author, 1:4)

Fig. 6 Waterjar type 3 (photograph by the author)

148

JAPP

Fig. 7 Waterjar type 4 (photograph by the author)

Fig. 8 Waterjar type 5 (photograph by the author)

Fig. 10 Upper body part of a waterjar type 6 (drawing by the author, 1:4)

Fig. 9 Waterjar type 6 (photograph by the author)

Fig. 11 Single specimen of a waterjar (photograph by the author) 149

LRCW4

150

EARLY BYZANTINE COARSE WARE FROM NOVAE: PRODUCTION AND CHRONOLOGY ANDREJ B. BIERNACKI¹, ELENA KLENINA² ¹Adam Mickiewicz University in Poznań, Sw. Marcin, 78, Poznań 61-809, Poland; [email protected] ²National Preserve “Chersonesos of Taurica”, Drevnaja, 1, Sevastopol 99045, Ukraine; [email protected] In the late 5th–6th centuries AD the pottery industry in Novae was apparently strictly controlled by the bishopric. The local workshops produced coarse ware and supplied mainly the local/regional market, i.e. the people of Novae and the immediate vicinity of the city. Macroscopic study of the pottery distinguished several groups of vessels. The most numerous group of ceramics manufactured in Novae is cooking pots. These vessels are very resistant to high temperature. One of the ceramic workshops in Novae additionally produced earthenware oil lamps. KEYWORDS: COARSE WARE, NOVAE, BYZANTINE, CHRONOLOGY, TYPOLOGY, LOCAL PRODUCTION of a catastrophe caused by natural or political events. According to the stratigraphical examination and analysis of the materials found in layers contemporary with the kiln, the assemblage should be dated to the 6th century.

The Polish-Bulgarian joint research project for the archaeological exploration of the Roman and early Byzantine city of Novae (Svištov, Bulgaria) (Fig. 1/1), which has been pursued for more than forty years now, has provided abundant source material on pottery vessels. In the early Byzantine times, the local production of ceramic tableware and kitchenware continued. Several pottery workshops from that period have been identified in the city (Fig. 1/2).

Ceramic material from the kiln is represented by saucepans, pots, oinochoai, and a fragment of a lamp (Klenina 2006, 85-88). The most numerous group of vessels is the saucepans. After firing their fabric is mainly a dark grey colour. In some cases, irregular burning resulted in them having stripes of reddish-brown and light grey colouring. In rare cases, the fabric became greyyellow. Probably, the high temperature of firing or an over-long stay in the kiln caused the appearance of some stripes of blue and green colour. The impurities in the clay consist generally of sand: in the large vessels this is coarse. The surface of the vessels is grey or light grey. The workshop used clay rich in iron and lime that fired brownish-green, greenish and yellowish-greenish at a temperature reaching about 1000–1100 °C.

One of these, dated to the 6th century, was owned by the eparchy and controlled by the bishop. A two-chambered kiln of the ceramic workshop has been discovered northwest of the episcopal residence, right beside the via principalis (Klenina 2006, 27). The bottom of the lower chamber is situated 1.75m beneath the modern surface. The kiln was constructed in a layer of yellow clay. The inner surface of the kiln was plastered with clay, applied to a thin film of mortar. In the course of kiln’s functioning, its inner surface was coloured red. The west part of the kiln did not survive. In the east part, there is a lime layer 0.035–0.040m in width probably the result of the destruction of the wall situated 0.07m eastwards of the outer kiln surface. The pottery kiln is two-chambered. Remains of a ceramic partition in the form of a grate that divided the kiln into two chambers were found in the lower one. Traces of the grate attachment are noticeable on the east wall. The kiln is preserved 1.20–1.25m high. The maximum inner diameters of the upper and lower chambers are respectively 1.85 and 1.58m; the height of the partition is set at 0.65m above the kiln base.

Another evidence of ceramic production in Novae is the existence of a cylindrical stand to assist in the firing process. A mid-sized cylindrical stand with a maximum diameter 0.14 m was discovered during the excavations (in squares 153/154: sector X, 16 hectares) west of the ceramic workshop in the layer above one of demolished walls of the thermae of the 3rd–4th centuries. In the firing process, the fabric of the cylindrical stand became grey. Special cylindrical stands with lateral apertures for the free flow of hot air were used for the preliminary drying of ware and directly for firing. Similar stands were in use in the ceramic workshops of Panticapaion, Phanagoreia, Cos (Gajdukevich 1934, 49, 89; Poulou-Papadimitriou, Didioumi 2010, 747). They were placed on the holes of the kiln-grate and the vessels to be fired were stood on them. Such supports are often found in the immediate vicinity or inside the kilns.

The kiln was filled with soil of a dark brown colour with lots of charcoal and ceramic vessels, intact and in fragments (36 in total). The vessels remained in situ after the destruction of the kiln in the late 6th–early 7th centuries (Kotecki 1977, 193, 197). A similar kiln was discovered in Iatrus (Conrad 1996, 109-110).

Comparative studies of the materials from the kiln and the ceramics found during the excavations let us distinguish several groups of coarse ware manufactured

Deformation of some vessels in the kiln points to a high temperature at the firing of this lot. Evidently, the producer could not properly complete the firing because 151

LRCW4

The next type of jug is a small oinochoe with a wide ribbed cylindrical neck, globular body, and a flat foot (Fig. 2/9-10, 4/1). The outturned rim is slightly thickened. An oval flattened handle, double rolled in profile, is attached directly to the rim and upper body. The rim resembles a trefoil in plan. The rim diameter is between 9.5 and 12cm, the height is 13–14cm, the body diameter is 12–12.5cm, and the foot diameter is 5-8cm. Such vessels date back to the 6th century. Jugs of this type occur in layers of the 6th century in the cities of Lower Danube region (Kuzmanov 1992, 213, type 9; Opaiţ 2004, 59-60).

in Novae in the late 5th–6th centuries. The fabric of the local ware is reddish-brown, reddish-yellow, dark grey, and grey-yellow in colour. It contains coarse sand, mica, and lime. The surface of the vessels is grey or light grey and brownish-yellow. Drinking cups of the local production, small in size, have a vertical double rolled rim, short cylindrical neck, and roundish body on a flat foot (Fig. 2/1-2, 4/6). A loopshaped oval handle is attached to the upper part of the vessels. The rim diameter is 6.8–8.5cm, the foot diameter is 4.5–5.0cm, and the height is 8.6–9.3cm. Vessels of this type are known in Novae in ceramic assemblages of the late 5th–6th centuries (Klenina 2006, 105-106, type 3). Cups of the same shape were discovered in assemblages of the first half of the 5th–6th centuries in Iatrus (Böttger 1982, 58, type 1).

In Novae, they also produced double handle jugs, or table amphorae. A big two-handled jug with a rolled rim flattened on top, ribbed cylindrical neck, globular body and a flat convex foot (Fig. 2/6) is rather characteristic. The rim sometimes has a groove in the upper part. Oval flattened handles are attached to the neck and shoulders (Klenina 2006, 108, type 7). The rim diameter is 7.6–10cm, the height is 28.7–29.5cm. Fragments of such vessels were found in assemblages of the mid 5th–6th centuries.

The second type of drinking vessels is a cylindrical bowl without handles on a flat foot, slightly concave in the centre (Fig. 2/3-4, 4/5). The foot is separated from the body by a sharp rib. The rim being a thickened continuation of the wall is cut flat on the top, with a slight inclination inwards. The rim diameter is 11 and 12cm, the height is between 16.2–17.5cm, the maximum body diameter varies from 10 to 11.2cm. The vessels are likely to have been used for heating up drinks and that is the reason why they are greatly sooted both outside and inside. The bowls were found in the layers of the late 5th–6th centuries. Similar vessels were discovered during the excavations of the Early Byzantine fortress near the modern town of Sadovec (Bulgaria) (Kuzmanov 1992, 211, type 4).

The next type of double-handled jugs is represented by vessels of a smaller size with a high cylindrical neck and oval body on a flat concave foot (Fig. 2/7, 4/4). The upper body is ribbed. The rim is not differentiated on the outside from the neck; inside, there is a projection for the lid. Oval handles with a small, asymmetrically placed groove are attached under the rim and on the shoulders. Sometimes these vessels bear dipinti in red, including cross-monograms. The rim diameter is 10–10.8cm, the height is 20.5cm. In Novae, such jugs are common in assemblages of the 6th century.

The next numerous group of the local table wares is represented by jugs of several types.

The most numerous group of ceramics manufactured in Novae is cooking pots. pots.

The commonest type is jugs-oinochoai with an everted stepped rim, short cylindrical neck with ribbed surface, globular body, and a flat foot, convex in the centre (Fig. 2/5, 4/2). An oval flattened handle, double rolled in profile, is attached to the body and shoulders. The rim has the shape of a trefoil. The upper body is sometimes ribbed (Klenina 2006, 108, type 6). The rim diameter is 12.0–19.7cm, the foot diameter is 9.2cm, the height is 19.5–20cm. The vessels were found in assemblages of the late 5th–6th centuries. The jugs were produced in the Episcopal workshop in the Early Byzantine period and apparently they were not widespread outside the city.

One of the most popular shapes of the cooking pots is a vessel with a vertical rim, grooved outside, a short cylindrical neck, roundish body, and a foot concave in the centre (Fig. 3/1). Two oval handles, double rolled in profile, are attached to the rim and shoulders (Klenina 2006, 113, type 21). The rim diameter varies from 912.5cm, the height is 18.5cm, the foot diameter is 7cm. Fragments of the vessels of this type occur in layers of the 6th century. The next type is represented by double-handled vessels that have a roundish body with decoration of several shallow grooves on its upper part, and a flat, slightly concave foot. The rim is greatly outturned (Fig. 3/2). The upper part of the rim is flattened and has a shallow groove. Inside, there is a projection for the lid. The cylindrical neck is sometimes decorated with grooves. Oval flattened handles, triple rolled in profile, are attached directly to the rim and the point of transition from the upper to lower body, forming dips slightly ribbed. The rim diameter is 12.4–23cm, the foot diameter is 5.5–10cm, the height is 15.4–25cm. Fragments of such vessels were discovered in Novae in assemblages of the late 5th–6th centuries. This type was common in

Oinochoai with a globular body, ribbed cylindrical neck smoothly widening upwards, and a flat concave foot also belong to the group of jugs (Fig. 2/8). Their neck ends in a funnel-shaped trefoil rim. An oval flattened handle, double rolled in profile, is attached to the rim and centre of the body. The rim diameter is 6cm, the height is 18.2cm, the body diameter is 16cm, and the foot diameter is 8cm. Such vessels date back to the 6th century. Jugs of this type occur in layers of the late 5th–6th centuries in the cities of the provinces of Moesia and Scythia, Constantinople, Athens, Corinth, and Thasos (Opait 2004, 62-63).

152

BIERNACKI- KLENINA

Nikopolis-ad-Istrum and Iatrus in the Early Byzantine period (Klenina 2006, 113-114, type 22). A great number of these pots were found during the excavations of the Early Byzantine fortress near the modern town of Sadovec (Bulgaria) (Kuzmanov 1992, 213). Vessels of similar shape are known from the sites of Thrace and Constantinople in layers of the 6th–early 7th centuries (Hayes 1992, 158).

Bibliography Böttger, B. 1982. Die Gefäßkeramik aus dem Kastell Iatrus. Iatrus-Krivina. Spätantike Befestigung und frühmittelalterlich Siedlung an der underen Donau II, 33-148. Berlin. Conrad, S. 1996. Zur Keramikproduktion im Kastell Iatrus (Moesia Secunda). Rei Cretariae Romanae Fautorum Acta 33, 109-117.

Cooking pots-oinochoai are vessels of local production with a smoothly outturned and rolled rim, a wide cylindrical neck, roundish body, and a flat foot (Fig. 3/34, 4/3). The transition from neck to body is rounded. Oval flattened handles, double rolled in profile, are attached to the rim and the point of transition from neck to body. The rim is designed as a trefoil. The rim diameter is 17.6– 24cm, the height is 22–23.5cm, the foot diameter is 11.4– 13cm, and the maximum body diameter is 22.8–27.5cm. All the pots were discovered in the destruction layer of the kitchen of the pilgrim house that was part of the episcopal complex dating back to AD c. 556. Similar vessels were found in the fortress near the modern town of Sadovec (Bulgaria) (Kuzmanov 1992, 213).

Gajdukevich, V. F. 1934 Antichnye keramicheskie obzhigatel’nye pechi. Izhvestiya Gosudarstvennoj akademii istorii material’noj kul’tury 80. MoscowLeningrad. Hayes, J. 1992. Excavations at Saraçhane in Istanbul, vol. 2: The Pottery. Princeton N.J., Princeton University Press. Klenina, E. 2006. Ceramic tableware and kitchenware of the 3rd – 6th century from Novae (Northern Bulgaria). In A. B. Biernacki (ed.), Novae. Studies and Materials II, 1-188. Poznań-Svastopol. Kotecki, J. 1977. Piec garncarski w "sondażu wodociągowym". In K. Majewski (ed.), Novae-Sektor Zachodni, 1975. Archeologia 28, 193-199. Warszawa.

The local workshops also produced moulded lamps of an oval shape with an oblong nozzle and low ring foot. The discus often bears a decoration of Christian symbols; the handles are formed as crosses (Stawoska-Judziłł 1995, 47-59). Fragments of lamps of this type occur in the assemblage of the 6th century (Klenina 2006, 87).

Kuzmanov, G. 1992. Die local Gefässkeramik. In S. Uenze (ed.), Die spätantiken Befestigungen von Sadovec (Bulgarien), 201-221. München. Opaiţ, A. 2004. Local and Imported Ceramics in the Roman Province of Scythia (4th – 6th centuries AD). British Archaeological Reports, International Series 1274. Oxford, BAR Publishing.

Overall then, in the late 5th–6th centuries the ceramic workshops of Novae produced coarse wares whose quality of production was quite good. Producers had to omit some technological procedures that would have raised the price of the ware, though. Usability became the main determining factor. The design of the table ware receded into the background. Most of the vessels were simply covered with a clay coating of the same colour as the fabric. The standardization of the cooking ware points to a simple way of cooking and a repetitive repertoire of culinary dishes. It should be noticed that for the most part the plain table ware of local production was also used for heating up food. The evidence consists of traces of soot on the outer and sometimes on the inner surface of the vessels. The absence of any local red slip ware was probably connected with the mass provisioning of table ware from Phocaea (Asia Minor) to the Black Sea and the Lower Danube regions in the 6th century. It must have travelled alongside the foodstuffs also being delivered from Asia Minor.

Poulou-Papadimitriou, N., Didioumi, S. 2010. Nouvelles données sur la production de l’atelier céramique protobyzantin à Kardamaina (Cos-Grèce). In S. Menchelli, S. Santoro, M. Pasquinucci, G. Guiducci (eds.), LRCW3. Late Roman Coarse Ware, Cooking Ware and Amphorae in the Mediterranean. Comparison between Western and Eastern Mediterranean, Comparison between western and eastern Mediterranean. British Archaeological Reports, International Series 2185 (II), 741-749. Oxford, BAR Publishing. Stawoska-Judziłł, B. 1995. Symbolism of the cross on clay-lamps from Novae. In A. B. Biernacki (ed.), Novae. Studies and Materials, I. 47-60. Poznań.

153

LRCW4

Fig. 1. Map of Eastern Mediterranean (1); location scheme of the kilns on the territory of Novae (2).

154

BIERNACKI- KLENINA

Fig. 2. Novae, tableware from the contexts of the late 5th - 6th century: cups (1-2), bowl (3-4), jugs-oinochoai (5, 8-10); two-handled jugs (6-7), lamp (11).

155

LRCW4

Fig. 3. Novae, kitchenware from the contexts of the late 5th - 6th century: pots (1-2), pots-oinochoai (3-4).

156

BIERNACKI- KLENINA

Fig. 4. Novae, contexts of the late 5th - 6th century: jugs-oinochoai (1-2), cooking pot-oinochoai (3), two-handled jug (4), bowl (5), cup (6). 157

LRCW4

158

A CERAMIC WORKSHOP OF THE EARLY BYZANTINE PERIOD ON THE ISLAND OF LIPSI IN THE DODECANESE (GREECE): A PRELIMINARY APPROACH

ELENI PAPAVASSILIOU, KONSTANTINOS SARANTIDIS, EIRINI PAPANIKOLAOU Archaeological Service of Dodecanese, Ippoton Str, Rhodes, Greece; [email protected] At Lipsi, in the Dodecanese Islands, archaeological excavation on the Colombini property brought to light a pottery workshop with four kilns, four storage areas, intact amphorae and large quantities of potsherds. Twenty complete amphorae of uniform type (5 intact and 15 in pieces) constitute an important discovery. They have a cylindrical body, decorated with grooves, and imitate the commonest early byzantine amphora type LRA1, generally used to transport wine and oil. Conclusions: 1) The potters of the provincial workshop were familiar with the standard utility forms of the period, copying its models with considerable clumsiness, so suggesting a late dating for their manufacture, around the 7th century. 2) The coastal location of the workshop would facilitate transport of its output, since the island lies on the shipping lane linking Constantinople with Egypt and Syria. 3) The incised monograms (Λ, ΓΕ, Β) on plugs or amphora shoulders are of doubtful significance; they could be the marks of potters, workshops, merchants, contents or indications of customs control. 4) The cause of the workshop's destruction remains unclear, as does its workingduration. KEYWORDS: DODECANESE, LIPSI, CERAMIC WORKSHOP, KILNS, LRA1 IMITATION OR SURVIVAL, WINE TRADE, BYZANTINE TRANSITIONAL PERIOD. The island of Lipsi belongs to the Dodecanese complex and lies between Leros and Patmos, opposite the coast of Ionia in Asia Minor (Fig. 1). According to the inscriptions discovered, it was known in Antiquity as ‘Lepsia’ (Manganaro 1963-4, 317-329; Volanakis 2002, 21) and belonged to the ‘Ionian’ or ‘Milesian’ islands (Triantafillidis 2006, 178, note 15, with relevant bibliography), forming along with Agathonisi, Farmakonisi, Patmos, Arkioi, Marathi and Leros, a string of outposts for ancient Miletus. The early Byzantine period lacks written testimonies on the island, which followed a common historical course with Patmos from the 11th century onwards.

Four rectangular kilns were unearthed at a depth varying from 1 to 2m. They were constructed with limestone and mortar; their foundations’ depth in most cases corresponds to the modern sea level. The kilns whose entrance is located to the south, share a common northwest-southeast orientation. Four rectangular buildings complement the workshop’s facilities. These structures served most likely to house the production process (storerooms, dryers etc.), as indicated by the in situ discovery of a pile of clay in the south-west corner of the northeastern one (Fig. 3). Kiln 1 (dimensions: 5.30x5.20m). The walls, preserved to a height of 1-1.50m, are constructed of medium and small-sized rubble bound together with pebbles, pottery sherds and clay mortar. Benches along the long sides and a central pillar would have supported the grid. The clay floor was traced at a depth of 2.30m, having a slight inclination towards its centre. Six layers of landfill were removed, containing a multitude of sherds, ash and burnt earth. Ten amphorae were discovered at the entrance, nine of which were broken and one still intact (Fig. 3, 4).

Research on the D. Colombini building plot began in the summer of 2008 as a rescue excavation, with trial trenches conducted by the 22nd Ephorate of Prehistoric and Classical Antiquities: work was initiated by the presence of surface pottery fragments in large quantities and of a visible kiln. The plot has a surface area of 5000m2, and lies at the site of ‘Kambos’, north-west of the port of Lipsi (Fig. 2). Pottery found in the trenches was dated to the Byzantine period. Therefore, the 4th Ephorate of Byzantine Antiquities undertook the impending systematic excavation (December 2008- May 2010).

Kiln 2 (dimensions: 4.20x2.80m). The kiln founded on the natural argillaceous landfill is preserved to an approximate height of 0.50m. Its masonry was constructed as the above example. The kiln’s excavation revealed traces of a destroyed grid resting upon benches along the long sides and on a rectangular pillar (0.42x0.28m) discovered underneath it. The grid was constructed of clay plaques and of rectangular blocks coated with clay. The floor was made of clay and pumice stone, obviously in order to maintain the high levels of heat. The landfill was dominated by burnt soil and a considerable quantity of ash. Two fragments of amphora shoulders (ΑΕ 230, ΑΕ 2155) bear graffiti of an incised

Since the venture of fully presenting the archaeological material requires the working through of a vast amount of sherds, it becomes clear that only a preliminary approach is attainable at the moment. This paper will therefore deal solely with the study of the kilns and the intact amphorae, thus conducting a rough classification of these products and moreover enabling a chronology for the workshop to be formed.

159

LRCW4

Β, another disc-shaped clay object (ΑΕ 2460) bears an incised Μ (?), while an incised Λ occurs in the exterior surface of a firing ring (AE 2427) (Fig. 6).

0.60m; amphorae no. 4-13 were discovered at the entrance of kiln 1, under a fill of 1-1.50m depth; amphorae no. 14-16 were found scattered in the landfill and finally, amphorae no. 17-20 come from the entrance of kiln 4 (see table of amphorae).

Kiln 3 (dimensions: 2.10x1.20m). The walls, constructed as those above, rest on the natural rock and are preserved to a height of up to 0.40m. Underneath the destroyed (due to the roof collapse) grid, the clay floor was revealed, partially constituted of the natural rock. The kiln’s investigation revealed also a deposit composed of burnt red soil, some amphora wasters, undecorated pottery fragments and trefoil rims deriving mostly from oenochoai and an intact small oinochoe as well (AE 1389) (Fig. 8). The closest parallels for this oinochoe are found in tombs in Eleutherna (Crete) and in Boeotia and they date in the 7th c. (Yangaki 2004, 122, 144, fig. 36a-b and 123, fig. 37a-b; Chamilaki 2010, 597-8, fig. 5, no 3527). Similar oinochoe dated in the second half of the 7th c. are also found in Emphorio (Chios) (Boardman 1989, 100-103, pl. 22, no 178).

Features-Typology-Dating The five intact amphorae from the excavations at Lipsi share some common features concerning their decoration. They slightly differ in height and maximum diameter of the body (Fig. 9). The neck is cylindrical and stumpy and rims are marked by a thick convex band (0.015m wide). The interior of the rim, whose external diameter ranges between 0.07-0.08m, preserves wheel ridges. The unbalanced Γ-shaped handles commence exactly below the rim. They are 0.03-0.04m wide, with a rounded light ridge running down them at the centre. The body displays parallel grooves (0.005-0.01m wide), which are light and indiscernible from the base of the neck to the base of the handles, but then become denser, more visible and finally fade out at a small distance from the foot. The foot is convex and rounded with no buttons. The clay is reddishyellow, mediocre, clean, incorporating whitish small inclusions and abundant golden mica (Munsell 5YR, 6/6 or 6/4). From the aforementioned typological features, the Lipsi amphorae can be classified among the imitations or survivals of the LRA1 (Bonifay and Pieri 1995, 108-9, fig. 49-51, with relevant bibliography; Pieri 2005, 69-84, fig. 25, 30; Pieri 2007, 615, 616-7, fig.15,1 about the Egyptian imitations of the LRA1 type).

Among the significant finds of the landfill are listed clay lids (for parallels see Poulou-Papadimitriou and Didioumi 2010, 742, fig. 6h.), one of which preserves the incised letters ΓΕ (ΑΕ 1215) (Fig. 5), potter’s wheels, firing rings, one of which bears the incised letters ΕΛΘ (ΑΕ 1175) and numerous malformed and distorted sherds, witnesses to unsuccessful firings. The furnace’s limited dimensions (almost half the size of its counterparts) and the multitude of oinochoai sherds indicate that this specific kiln served mostly for the firing of small vessels. Kiln 4 (dimensions: 4.80x3.90m). The kiln is founded on the natural argillaceous landfill. The masonry, coated with mortar, is preserved to a height of 1.10m: it is constructed with medium and small-sized rubble held together by pebbles, pottery fragments and lime mortar. No grid was found. The clay floor, with a slight inclination towards its centre, was revealed at a depth of 2.20m.The stratigraphy displayed four layers: the first with clay masses, the second with yellow soil (kaolinitic clay) and sherds, the third with burnt soil and sherds, and the fourth with ash, charcoal and sherds. The destruction layer contained numerous sherds from broken amphorae and four intact amphorae (ΑΕ 3031-3034) (Fig. 7). Among the sherds were fragments up to 0.02m thick, probably from pithoi, as well as lekanides and pots.

The prevailing amphora type, LRA1, which has been systematically and thoroughly studied, occurs par excellence in the Aegean and on coastal sites of the Mediterranean basin (Empereur and Picon 1989, 239, fig. 18) and was traded from the 5th to the 7th centuries. Having as place of origin Syria, Cilicia and Cyprus (Hayes 1992, vol. 2, 64, 434, notes 8-9; Pieri 2005, 6146; Reynolds 2005, 565-6), this particular amphora type was distributed in Egypt, Spain, France, on the Greek mainland (Athens, Corinth, Argos, Olympia) and the Aegean islands (Crete, Rhodes, Cos, Samos, Chios, Thasos), the coast of Asia Minor, Constantinople and reached the Black Sea and Lower Danube (Pieri 2005, 69-81, with further bibliography).

The landfill of the rest of the excavation plot consisted of a deposit of ashes and burnt soil containing bases, rims, vessel walls, large numbers of handles, as well as parts of clay potter’s wheels, firing rings and covers. Individual finds comprise approximately 3060 objects, according to the excavation diaries. Moreover, various sections of floors were discovered, providing evidence for different construction techniques (beaten soil, clay plaques, stone plaques and pebbles with traces of plaster), as well as four charcoal pits containing burnt masses of clay and charcoal too.

The Lipsi amphorae are not well made nor symmetrical: they are not high-quality vases, nor do they present any typological variety. The exterior surface is coarse with protruding grains of clay; visible irregularities slightly differentiate their component parts. In particular, the handles are placed askew; the rim of amphora 8 is not circular but of ellipsoid form, while that of amphora 20 is not horizontal, but slanted in section. On amphora 8, the base of the handles bears a finger imprint, presumably the potter’s. Grooves manifest a certain degree of offhandedness in their execution, except for those of amphorae 18 and 19, where they are well-made and distinct.

This paper deals with the five intact amphorae and the other fifteen fragmentarily preserved, but demonstrating their shape nevertheless. In detail, amphorae no. 1-3 were found inside the deposit, at a depth of approximately 160

PAPAVASSILIOU-SARANTIDIS-PAPANIKOLAOU

There are two groups of amphorae with similar dimensions (in terms of height and body/rim diameter) as those from Lipsi: one comes from the Yassi Ada shipwreck (625/6 AD terminus post quem) (Bass 1982, I, fig. 8-3, CA 5, 155-7, 161-3), the other from a recently discovered workshop producing imitations of the LRA1 amphora in Kardamaina (Cos). The amphorae from Cos are dated in the 7th c. (early 7th c: Diamanti 2010a, 68; mid-7th century: Poulou-Papadimitriou and Didioumi 2010, 742).

measured. Thus and particularly, kiln 2 at the north has a distance of 44 slats (πήχεις) from kiln 1 (meaning 28.16m), while kiln 4 which is more to the south of kiln 1 is located at a distance of 35 slats (22.4m), and kiln 3 west of kiln 1, is located at a distance of 9.5 slats (6.08m); c) because one could procure clay from the landfills on the banks of the torrent that – according to information from the local inhabitants – used to cut through the site (we cannot yet endorse this opinion without laboratory analysis of the amphorae and the raw materials) (on the digging and manipulation of the clay, see Giannopoulou 2010, 68-71, with relevant bibliography).

The Lipsi amphorae find typological parallels in the amphorae assemblage from Emporio in Chios dating in the 9th c. (Boardman 1989, 114, fig. 43, pl. 25, no 280). Amphorae similar to those from Lipsi have been found also in Paphos dating in the 8th and probably the beginning of the 9th c. (Hayes 2003, 506, fig. 32, no 241), whereas other examples derive from Cherson and are dated in the 9th/10th c. (Ivaschenko 1997, 77, fig. 10b).

3. The twenty amphorae, intact and broken, share homogenous features and are classified typologically as among the imitations or survivals of the widespread LRA1 type (concerning the analysis of the type, the clay, the distribution and dating of the LRA1 type, see Pieri 2005, 69-81 with previous bibliography). In terms of chronology and due to the lack of other finds, they are probably dated from the middle 7th century onwards.

The above evidence and the typological study of the amphorae lead to a date after the middle of the 7th c. The examples from Chios and Cyprus indicate circulation of these amphorae during the 8th c. or even later. However, the lack of other dating evidence does not allow a more secure identification of the period during which the Lipsi production center was active. We believe that further study of the archaeological material will allow a better dating. On the basis of the existing evidence a date of the mid-7th c. onwards is suggested.

4. Many different opinions have been expressed on the incised letters/graffiti found on amphora plugs or shoulders (van Alfen 1996, 201-2). These provide information on proprietors, names of potters, workshops, regions, merchants, the vessel content. May they be interpreted as numerical indications or customs controls, in the context of annona militaris? (Goutzioukostas and Moniaros 2009, 127-145, 236; Diamanti 2010a, 61-2; Didioumi 2010, 811). 5. The lifespan of the workshop, as well as the cause of its destruction or abandonment, remain unknown (earthquake, flood or enemy invasion?). Whatever the case, it seems that even such a small island (of just 16 square kilometers) was not only inhabited during the early Byzantine period, but also displayed commercial activity involving pottery. The small, rough but rugged amphorae, with a capacity of 9.5 litres, probably: a) conveyed the wine of Lipsi to areas lacking such produce, even to rich centres, even to Constantinople, remembering that limited local production of a fine product is the more coveted in the market and commands a higher price when compared with mass produced quantities. It should be noted that, given the limited nature of oil production on the island, these amphorae are most likely linked with the wine trade. Their capacity is compatible with this suggestion, according to research (for viniculture and its importance in the Byzantine period, see Lefort 2006, 403-4, 410-3; van Alfen 1996, 208); b) were requested by the state, as a form of tax, that required amphorae with or without wine in order to cover the army’s needs, at a time when the large production centres of the Levant had been lost and when small, local workshops were established in their stead and c) catered for the everyday needs of the island’s population. The renown brusco-like wine of Lipsi (i.e. a varietal of wine known as ‘Phokiano’ deriving from Phocea and cultivated until nowadays on the nearby islands of Samos and Ikaria), hitherto mentioned only in oral tradition and folk poetry (Volanakis 2002, 193), now becomes a

Conclusions 1. The presence of a workshop at the coastal site of ‘Kambos’ at Lipsi is attested by: a) the discovery of four rectangular kilns, of which only the outline of the combustion chamber is preserved (for the kilns’ typology, see Hasaki 2002, 139-177; Theocharidou 1985-6, 106, fig. 15; Le Ny 1988, 39-45, 22a-23; Cuomo di Caprio 2007, 524, fig. 169, II. For the operation of the kilns, see Theocharidou, loc.cit. 100-4; Le Ny, loc.cit. 19-36), b) the discovery of five almost identical amphorae of the same type and numerous warped, malformed and badly fired sherds. 2. The location of the workshop is noteworthy: a) from its closeness to the sea, which is an important factor for the immediate and relatively low-cost transportation of amphorae; b) because of its considerable distance from the settlement which during the early Byzantine period was located on the opposite side of the island bay, at Kouselio, where remains of a basilica and baths have been discovered (Volanakis 2002, 27-8; 2007, 68-72). It should be noted that according to the rules governing the construction of a tile workshop of the 14th century, a necessary condition for the erection of a ceramic workshop was that the chosen spot ought to be located outside the settlement (Theocharidou 1985-6, 98, 99, note 12). Moreover, the information mentioned in the ‘Exabiblos’ by Armenopoulos about the distance that should be left between kilns is approximately correct, as 161

LRCW4

BIBLIOGRAPHY

palpable reality through the discovery of this pottery workshop. Production dates probably to the 7th-8th centuries, a time when the countryside acquired a primary role because cities were smaller and less numerous and society relied more upon the peasant. One should note that archival documents reveal that during the Italian rule in the Dodecanese islands, the Vatican was procuring for its Holy Communion the red semi-sweet wine from Lipsi, which comes from sun-dried red grapes. Over the last decades, the cultivation of this specific grape has been restarted: today the island is a member of the viniculture association of European cities.

van Alfen, P. G. 1996. New light on the 7th-c. Yassi Ada shipwreck: capacities and standard sizes of LRA1 amphora. Journal of Roman Archaeology 9, 189-213. Bass, G. F. 1982. The pottery. In Bass G. F. and van Doornick F. H., Yassi Ada, A seventh century byzantine shipwreck I. Texas, Texas University Press. Boardman, J. 1989. The Finds. In M. Ballance, J. Boardman, S. Corbett, S. Hood, Byzantine Emporio. Excavations in Chios 1952-1955. The British School at Athens Supplement 20, 86-142. Oxford.

6. Finally, the workshop’s discovery, even though only partially excavated, is considered a fortunate incident, given the fact that the continuous agricultural activity in the plot has destroyed most of the constructions. The foundations of the four kilns’ combustion chambers alone remain almost intact. The excavation at Lipsi provides important evidence for a new peripheral place of pottery production – thus this small island is added to the topographic map of the insular pottery workshops of LRA1 imitations or survivals in the Aegean during the transitional period of Byzantium, well located on the shipping route to and from Constantinople (Diamanti 2008, 11-37; 2010a, 164-5; 2010b, 143-152; Didioumi 2010, 795-828; Poulou-Papadimitriou and Didioumi 2010, 741-9).

Bonifay, M. and Pieri, D. 1995. Amphores du Ve au VIIe s. à Marseille: nouvelles données sur la typologie et le contenu. Journal of Roman Archaeology 8, 94-120. Cuomo di Caprio, N. 2007. Ceramica in Archeologia 2. Antiche tecniche di lavorazione e moderni metodi di indagine. Roma. Chamilaki, K. 2010. Ταφικά σύνολα υστερορωμαϊκών χρόνων από νεκροταφείο στο Δήλιον Βοιωτίας. Πρώτες παρατηρήσεις. In D. Papanikola-Bakirtzi and N. Kousoulakou (eds.), Κεραμική της Ύστερης Αρχαιότητας από τον ελλαδικό χώρο (3ος-7ος αι. μ.Χ.), Επιστημονική Συνάντηση, Θεσσαλονίκη, 12-16 Νοεμβρίου 2006, Β΄, 580-609. Publication of the Archaeological Institute of Macedonian and Thracian Studies 8. Thessaloniki.

Acknowledgements

Diamanti, Ch. 2008. Η σημασία των παραλίων της Μικράς Ασίας στην παραγωγή και διακίνηση των υστερορωμαϊκών/πρωτοβυζαντινών αμφορέων 1 (=LRA1). Αρχαιολογικές μαρτυρίες και επιγραφικές συμβολές, Deltion of the Centre for Asia Minor Studies 15, 11-37. Athens.

We would like to express our heartfelt thanks to Professor Natalia Poulou-Papadimitriou (Aristotle University of Thessaloniki) who instigated the study of these amphorae from Lipsi. Also, we owe the completion of this paper to two colleagues who work on the early Byzantine pottery of Cos par excellence, to Sophia Didioumi and Charikleia Diamanti: they have communicated the results of their lengthy research in the pottery field to the scientific community. Moreover, we would like to extend our thanks to the photographer M. Mitton and the artist Α. Katsouri, as well as to the Director of the Ephorate, Mrs Michailidou for granting us the permission to study and publish the excavation material.

Diamanti, Ch. 2010a. Εντόπια παραγωγή και εισαγωγή αμφορέων στην Αλάσαρνα της Κω (5ος-7ος αι.), S. Saripolos Library 115. Athens. Diamanti, Ch. 2010b. Εντόπιοι υστερορωμαϊκοί αμφορείς από την Αλάσαρνα της Κω. In Papanikola-Bakirtzi and Kousoulakou (eds.), 143-152. Didioumi, S. 2010. Κεραμική παλαιοχριστιανικών χρόνων από την Κω: στρώμα καταστροφής σε οικόπεδο της πόλης της Κω. In Papanikola-Bakirtzi and Kousoulakou (eds.), 795-828.

The costs of the excavation and conservation of the architectural finds were undertaken in full by the plot owner, to whom we owe thanks for his patience and understanding during the entire research, as well as for the interest he displayed in the archaeological remains discovered in his building plot. Therefore, we dedicated the poster to the owner Mr D. Colombini.

Empereur, J.-Y. and Picon, M. 1989. Les régions de production d’amphores impériales en Méditerranée orientale, 239. In Amphores romaines et histoire économique: Dix ans de recherche, Sienne 1986. École Française de Rome 1989, 223-248.

The fieldwork begun, on behalf of the 22nd Ephorate of Prehistoric and Classical Antiquities, by the archaeologist K. Sarantidis and the excavation project was undertaken by the archaeologists E. Papanikolaou and K. Sarantidis alternately.

Giannopoulou, M. 2010. Περί κατασκευής πίθων. In Papanikola-Bakirtzi and Kousoulakou (eds.), 65-80. Goutzioukostas, A. and Moniaros, X. 2009. Η περιφερειακή διοικητική αναδιοργάνωση της βυζαντινής αυτοκρατορίας από τον Ιουστινιανό Α΄ (527-565): η περίπτωση της quaestura Iustiniana exercitus. Thessaloniki. 162

PAPAVASSILIOU-SARANTIDIS-PAPANIKOLAOU

Hasaki, E. 2002. Ceramic kilns in ancient Greece: Technology and organization of ceramic workshops. University of Cincinnati.

Archaeological Reports, International Series 1662 (II), 611-25. Oxford, BAR Publishing. Poulou-Papadimitriou, N. and Didioumi, S. 2010. Nouvelles données sur la production de l’atelier céramique protobyzantine à Kardamaina (Cos-Grèce). In S. Menchelli. S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, cooking wares and amphorae in the Mediterranean. Archaeology and Archaeometry. Comparison between western and eastern Mediterranean. British Archaeological Reports, International Series 2185 (II), 741-749. Oxford, BAR Publishing.

Hayes, J. W. 1992. Excavations at Saraçhane in Istanbul, vol. 2. The Pottery. Princeton N.J., Princeton University Press. Hayes, J. W. 2003. Hellenistic and roman pottery deposits from the “Saranda Kolones” Castle site at Paphos. The Annual of the British School at Athens 98, 449-519. Ivaschenko, Y. 1997. Les ateliers de céramiques du VIe au XVe s. au Nord de la Mer Noire:le problème de la continuité. In G. D. d’Archimbaud (ed.), La céramique médiévale en Méditerranée. Actes du VIe Congrès de l’AIECM2, Aix-en-Provence 1995, 73-85. Aix-en-Provence, Narration.

Reynolds, P. 2005. Levantine amphorae from Cilicia to Gaza: a typology and analysis of regional production trends from the 1st to 7th centuries. In J. M. Gurt i Esparraguera, J. Buxeda i Garrigós, M. A. Cau Ontiveros (eds.), LRCW1. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. British Archaeological Reports, Int. Series 1340, 513-611. Oxford, BAR Publishing. Theocharidou, P. 1985-6. Συμβολή στη μελέτη της

Le Ny, Fr. 1988. Les fours de tuiliers gallo-romains. Documents d’Archéologie Française, 12, 19-45. Paris. Lefort, J. 2006. Η αγροτική οικονομία (7ος-12ος αι.). In A. Laiou (ed.), Οικονομική ιστορία του Βυζαντίου, Α΄, 377-494. Athens. Publication of the National Bank of Greece Cultural Foundation.

παραγωγής οικοδομικών κεραμικών προϊόντων στα βυζαντινά και μεταβυζαντινά χρόνια. Deltion tes Christanikis Archaeologikis Etaireias ΙΓ΄, 97-111.

Manganaro, G. 1963-4. Le inscrizioni delle isole Milesie. Annuario della Scuola Archeologica di Atene, XLIXLII, 317-329.

Triantafillidis, Π. 2006. Ιστορικά και αρχαιολογικά Αγαθονησίου. Athens Annals of Archaeology 39, 175192.

Pieri, D. 2005. Le commerce du vin oriental l’époque byzantine (Ve-VIIe s.). Le témoignage des amphores en Gaule, Bibliothèque archéologique et historique 174. Beyrouth, Institut Français du Proche-Orient.

Volanakis, J. 2002. Ιστορία και Μνημεία των Λειψών Δωδεκανήσου. Rhodes. Volanakis, J. 2007. Τα παλαιοχριστιανικά μνημεία των Λειψών της Δωδεκανήσου. Dodekanesiaka Chronika ΚΑ΄, 64-76.

Pieri, D. 2007. Les centres de production d’amphores en Méditerranée orientale durant l’antiquité tardive: quelques remarques. In Bonifay and J.–C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. British

Yangaki, A. 2004. Οι τάφοι. In P. Themelis (ed.), Πρωτοβυζαντινή Ελεύθερνα. Τομέας Ι, v. 1, 115-183. Athens, Publications of the University of Crete.

163

LRCW4

TABLE OF AMPHORAE no 1

Type (cat. no.) Amphora (ΑΕ273)

Height 38 cm

Max. Ø 24 cm

Outer rim Ø ------

2 3 4

Amphora (ΑΕ697) Amphora (ΑΕ698) Amphora (ΑΕ855)

23 cm 38 cm 37 cm

28 cm 26 cm ------

-----8 cm ------

5 6 7 8 9 10 11 12 13

Amphora (ΑΕ858) Amphora (ΑΕ859) Amphora (ΑΕ860) Αmphora (ΑΕ861) Amphora (ΑΕ862) Amphora (ΑΕ930) Amphora (ΑΕ929) Amphora (ΑΕ931) Amphora (ΑΕ932)

12 cm 14 cm 20 cm 42 cm 25 cm 28 cm 32 cm 28 cm 19 cm

---------------23 cm 21 cm ---------------22 cm

7.5 cm ----------8 cm 7.5 cm 7.5 cm ----------------

14

Amphora (ΑΕ1894)

43 cm

23 cm

------

15 16 17 18 19 20

Amphora (ΑΕ2551) Amphora (ΑΕ2929) Amphora (ΑΕ3031) Amphora (ΑΕ3032) Amphora (ΑΕ3033) Amphora (ΑΕ3034)

30 cm 30 cm 42 cm 42 cm 42 cm 42 cm

-----25 cm 25 cm 25 cm 25 cm 25 cm

----------7 cm 7 cm 7 cm 7 cm

Comments-Location Incomplete, in pieces. Found in a pit deposit. Ditto Ditto Incomplete, in pieces. Found at the opening of kiln no. 1 Ditto Ditto Ditto Almost intact. Same location. Ditto Ditto Ditto Ditto Ditto. Possibly the same as no. 12. Incomplete, in pieces. Found in the earth-fill. Ditto Ditto Intact. Found in kiln no. 4. Ditto Ditto Ditto

Fig. 1. Map of the Dodecanese Islands and Lipsi

164

PAPAVASSILIOU-SARANTIDIS-PAPANIKOLAOU

Fig. 2. General view of the excavation site

Fig. 3. Plan of the excavation

Fig. 4. Aspect of kiln 1 with its amphorae

165

LRCW4

Fig. 5. Amphorae lids

Fig. 6. Potsherds with graffiti

Fig. 7. Aspect of kiln 4 with its amphorae

166

PAPAVASSILIOU-SARANTIDIS-PAPANIKOLAOU

Fig. 8. Intact small oinochoe from the deposit of kiln 3

Fig. 9. a) Comparative photos of the five intact amphorae

167

LRCW4

Fig. 9. b) Comparative drawings of the five intact amphorae

168

LOCAL POTTERY PRODUCTION IN THE ISLAND OF COS, GREECE FROM THE EARLY BYZANTINE PERIOD. A PRELIMINARY REPORT SOPHIA DIDIOUMI Archaeologist, University of Thessaloniki [email protected] Recent excavations by the Archaeological Service of the Dodecanese in the island of Cos, Greece have revealed pottery workshops set on the long shore-lines of the island, specializing mainly in the production of various types of commercial amphorae. A new type of globular amphorae is produced on the southwest side of the island, in Cardamaina, dated to the 7th century. On the northwest side of the island, in the Mastichari region, are produced imitations of Ephesian amphorae (Pieri type LRA3 B2 and B4, Bezezcky type Ephesus 56), dated from the middle of the 6th to the first half of the 7th centuries. Imitations of the LRA1 type are found at both sites and late imitations or survivals of LRA1, dated probably to the 8th/beginning of the 9th centuries, were found in a coastal settlement at Kephalos. The workshops and the storage areas are situated near the sea, to assist transportation of the agricultural products of the island by the sea routes. The mass production of the commercial amphorae in Cos allows us to suppose that the island participated in the system providing the army (annona militaris) and the capital, especially in the 7th century, after the loss of the east provinces of Syria and Palestine. KEYWORDS: POTTERY WORKSHOPS, LRA1 IMITATION AND SURVIVALS, GLOBULAR AMPHORAE, PIERI LRA3 B2 AND B4, BEZEZCKY EPHESUS 56, ANNONA MILITARIS. 22nd Ephorate of Prehistoric and Classical Antiquities and the 4th Ephorate of Byzantine Antiquities. More than 330 rescue excavations have taken place throughout the island since 1970, producing important new data about the history of the island in the early Byzantine and the transitional period (Didioumi 2011). More than 28 basilicas were erected in the 5th and the first half of the 6th centuries and, apart from the capital city of Cos, coastal settlements were located along its shores (Baldini, Contò and Marsili 2011; Brouscari 2011; Didioumi 2011).

The island of Cos, situated in the southeast Aegean, lies 2 to 5 miles off the coast of ancient Halicarnassus (Bodrum) in Asia Minor and the Cnidian or Doris peninsula (Datça peninsula). It belongs to the insular complex of the Dodecanese, the so-called Κυκλάδες νήσοι in ancient times (Fig. 1). The island was prosperous for a long time, especially during Hellenistic, Roman and early Byzantine periods, being well placed for the maritime trade route which led from Egypt, Cyprus, along the coasts of Palestine and Asia Minor to the capital and thence to the Black Sea. The island has enjoyed a rich agricultural production (olive oil, wine, grapes) from the ancient times to the present day, due to the fertile volcanic plains and the abundant water sources. In the Hellenistic period it was famous for Coan wine, the socalled τεθαλασσωμένον οίνον, which was exported throughout the Mediterranean and the Black Sea (Georgopoulou 2005, 159-177).

From the studies of the excavation materials conducted by the Archaeological Service more than four pottery workshops were excavated or located in the coastal settlements of the island, dated from the middle of the 6th to the 7th centuries (Poulou-Papadimitriou and Didioumi 2010). Yet another one was excavated in ancient Halasarna by the University of Athens (KokkorouAlevras et al. 2006, 61).

In the early 6th century Cos belonged to the Province of the Islands of the Prefecture of the East and after 536 AD to the Queastura Iustiniani Exercitus, the new maritime administrative unit established by emperor Justinian I, which comprised the Aegean islands, Caria, Cyprus, Scythia Minor and Moesia Inferior (Goutzioukostas and Moniaros 2009, 127-145, 236).

1. City of Cos The city of Cos (Fig. 1) flourished during the Hellenistic and Roman periods. Moreover, research proves that the town was prosperous also during the late Roman and early Byzantine period, mainly from the 4th to the 7th centuries AD. The city of Cos has not yet yielded a pottery workshop, although it was the capital and the most important city of the island with its closed commercial harbor, rich in civil and religious monuments (Brouscari 2006; Didioumi 2011, 89-94; Livadiotti 2011, 72-79). However, some late Roman and early Byzantine

A great number of sites and monuments from the late Roman and early Byzantine period has been revealed on the island by the research of the Italian Archaeological Mission (1912-1940) (Livadiotti 2011) and especially through the surveys and the rescue excavations conducted by the Archaeological Services of the Dodecanese, the 169

LRCW4

51, with bibliography; Roumeliotis 2001; 2003; Diamanti 2010a; 2010b; 2012).

pottery, mainly from recent rescue excavations, could be attributed to local production (Didioumi 2010, 810).

Two types of amphorae are indentified by Diamanti as local, the LRA1 (2010a, 58-61, figs. 38-83, pl. 5-10) and LRA13 type (2010a, 80-92, figs 87-139, pl. 11-19), according to Riley’s typology; they are dated to the second half of the 6th until the first half of the 7th century (Diamanti 2010a). The local amphorae of Halasarna bear different types of stamps, depicting according to Diamanti busts of emperors or cross-like monograms with inscriptions (Diamanti 2010a, 92-107; 2010b; 2012). The workshop of Halasarna also produced lamps and common pottery (Roumeliotis 2001; 2003).

2. Cardamaina The modern village of Cardamaina is situated in the south part of the island, opposite the Knidian peninsula, close to the island of Nisyros (Fig. 1). More than three early Byzantine settlements are spread out on the coast over 9.5km (Didioumi 2011, 101-103); four or five basilicas are known (Brouscari 2011, 5-6). An important fortified site is located on the west side here on a low hill, dated by coins of Constans II to AD 659-664 (Brouscari and Didioumi 2006, 311). In the area of Cardamaina are situated two different workshops, with indications of a third.

C. Local pottery at basilica of presbyter Photeinos In the centre of the modern village of Cardamaina the excavations of the 4th Ephorate revealed the basilica of presbyter Photeinos (Πρεσβύτερος Φωτεινός), founded at the middle of the 5th century. The destruction and the abandonment of the basilica are dated by coins of the middle or the second half of 7th century (Brouscari 2011). Although the material is still unpublished, according to the preliminary reports and the kind information of the excavator Ersi Brouscari, there are indications for the production of local amphorae near the basilica. The amphorae were found in a room of the auxiliary space of the basilica and they are identified by Brouscari as LRA1 and LRA2/13 according to Riley’s typology, but they are dated late, in the first or second half of the 7th century (Brouscari 2011, 13). They were found in a storage room of the basilica, which led Brouscari to the theory that the church had an active role in the commerce of this period.

A. Workshop at the settlement of Aghia Theotis (K. Roussou’s property) At the west end of the village a pottery kiln and parts of auxiliary spaces were excavated recently by the 4th Ephorate of Byzantine Antiquities at the property of K. Roussou (Poulou-Papadimitriou and Didioumi 2010). The workshop is 300m north from the basilica of Aghia Theotis (Αγία Θεότης) (Baldini, Contò and Marsili 2011, 128) (Fig. 2). The kiln belongs to the vertical or updraft type, circular in section, with three arches for the support of the griddedfloor (Poulou-Papadimitriou and Didioumi 2010, fig. 5 a, b; in press). The main production of the workshop consists of two types of amphorae: 25% to the family of Byzantine Globular amphorae (Fig. 3) and 67% to the late imitation of the LRA1 amphora (Fig. 4), dated to the middle 7th century or later. The workshop also produced cooking pots, LR2 unguentaria and wheel-made lamps (PoulouPapadimitriou and Didioumi 2010,742, fig. 6 d, e, g).

3. Mastichari The village of Mastichari is situated at the northwest part of the island, opposite the island of Kalymnos (Fig. 1). Early Byzantine coastal settlements and workshop installations have been excavated by the 4th Ephorate of Byzantine Antiquities along the coastline of Mastichari for some 5.5km (Didioumi 2011, 98-101).

The fabric of the vessels is reddish (Munsell, 5YR 6/6); some of them preserve a lighter slip (5YR 7/6), or a yellowish or whitish colouring (5Y/R 7/8). It is of medium hardness, with a few whitish and gray inclusions and a large quantity of golden mica.

A. Workshop at ‘Costaina’ (AKTI’ s plot) B. Workshop at ancient Halasarna

A large workshop installation used in the 3rd century BC and again from the second half of the 6th until the first half of the 7th century is located to the east of the village of Mastichari, in the region of Costaina (Κώσταινα). The excavation was conducted by the 4th Ephorate of Byzantine Antiquities and revealed olive oil storage installations and wasters of Coan amphorae dated to the Hellenistic period (3rd to 2nd centuries BC) and a large building complex with four kilns, dated to the late Roman and early Byzantine period (Didioumi 2011, 97-98, fig. II.2.12). The kilns are of various types (PoulouPapadimitriou and Didioumi, in press), producing mainly amphorae, common pottery, lamps and tiles. One of the

This is located c. 250m west of the kiln at K. Roussou’s property and 500m from the basilica of Aghia Theotis (Fig. 2). The excavation, conducted by the University of Athens in the remains of the religious centre of ancient Halasarna (Δήμος Αλασσαρνιτών), revealed part of the early Byzantine settlement (Kokkorou-Alevras 2009). Remains of structures, pottery wasters and moulds for clay lamps are evidence for the existence of a pottery workshop, dated to Hellenistic period (Georgopoulou 2005) and again to the second half of the 6th to the first half of the 7th century (Kalopissi-Verti 1998, 251, note

170

DIDIOUMI

plot (Didioumi 2011, 99-100, fig. II.2.14), 630m west from the basilica, a half lamp mould was found (Fig. 15), dated to the 6th century. The mould bears in the centre of the disc an arch on two columns framing a human figure. Angeliki Katsioti (2012) who studied this object associates the decoration with the pilgrimage of St. John Theologos in Ephesos, based on similar depictions on lamps and pilgrim flasks.

kilns (Kiln D) (Fig. 5) revealed amphorae and wasters of LRA3 B4 (Fig. 6). The destruction layers contained wasters of both LRA1 (Fig. 7) and globular amphorae (Fig. 8), amphorae pot stands (Fig. 9 and 10) and common pottery to a lesser degree. Intact amphorae imitating examples from Ephesus were found: Pieri LR3 B2 (Pieri 2005, 98, fig. 61) (Fig. 11 and 12) and Pieri LRA3 B4 (Pieri 2005, 98, fig. 61; Bezeczky 2005, 204, pl. 1, nos 1-10; 2010, 365) and Bezeczky Ephesus 56 (2005, 204-205, pl. 2, nos 11-20; 2010, 365) (Fig. 14).

Amphorae of Ephesus 56 type (Bezeczky 2005, 204-205) were found in the excavations (Fig. 16 a, b, c), dated probably from the second half of the 6th century to the first half of the 7th century, judging by their shape and their resemblance to the amphorae found in Ephesus and in Costaina.

The LRA 3 amphorae from Costaina present variations, according to the formation of the lip and the base. The LRA3 B4 are of 40cm height with a low capacity, no more than 2lt, similar to the examples of Pieri’s typology (2005, 97) and rim D. 3.1cm; There is no intact vessel of the Ephesus 56 type, but they too are of small dimensions, approximately 25 to 29cm in height and rim D. 5.7cm.

4.Kephalos The extensive shore-line of Kephalos at the southwest part of the island lies opposite Nisyros island and the Cnidian peninsula. Along 3km of the coast six early Christian basilicas were found (Baldini, Contò and Marsili 2011, 126-127; Livadiotti 2011, 74, fig. II.1.2 and II.1.3) and costal settlements (Didioumi 2011, 103-108), dated from the 5th to 9th centuries (Fig. 1).

The fabric of Costaina’s vessels are sandy, with a lot of golden and silver mica, gray inclusions, soft, and reddishyellow in colour (Munsell 5YR 6/6, 7/6 and 8/4). The slip on the LRA3 B4 type is brown-red (Munsell 5YR 5/6), often preserved also on the inside of the neck. The wasters preserve traces of burning (Munsell 5YR 5/2, weak red 5YR 5/4).

The excavation at Perou’s plot revealed a large part of a settlement near the shore, with five two-storied buildings separated by cross-roads. A number of coins of Heraclius, dated from 629/30 to 639/40 and a coin of Leo V, from AD 813 (Didioumi 2011, 106-108, fig. II.2. 25 and 26), are strong evidence for the continuity of the life in the settlement to the transitional period of Byzantium.

The excavation data give a chronology from the second half of the 6th century to the first half of the 7th century (Poulou-Papadimitriou and Didioumi, in press). The LRA3 amphorae are distributed after the 5th century all over the Mediterranean and the Black Sea (Pieri 2005, 99, fig. 62, with bibliography); as Pieri has suggested they are probably produced in different places (2005, 100-101). On the island of Cos, however, it is not a common find; one LRA3 B4 amphora came from the city of Cos (Skerlou 1994) – it is probably of Ephesian origin, according to macroscopic observation. LRA3 B1 amphorae of local origin were found in the recent restoration works in the Insula II near the Agora of Cos (Giannikouri, forthcoming), probably local production.

Imitations or survivals of LRA1 amphorae (PoulouPapadimitriou and Nodarou, in this volume), were found in a storage room, dating probably in the 8th or the beginning of the 9th centuries (Fig. 17 a, b). We can assume that some of the amphorae found in the later layers are of local production, based on the quality of the fabric. They are of small shape (H. 38 to 44cm, lip D. 6.3 to 7.3), with a narrow neck shape; their reddish-yellow fabric (Munsell 5YR 7/6 and 8/4) is sandy with small grey inclusions and a lot of golden and silver mica (Fig. 1 and b). They have close parallels with amphorae found in different sites, like Grete and Cyprus (PoulouPapadimitriou 2013, 406 with bibliography, figs. 30 α, β from Pseira and fig. 31 from Paphos), dated at the end of the 8th to the beginning of the 9th century and the amphorae produced at the workshop of Leipsoi (Papavasileiou et al., in this volume).

In Rhodes a find of great significant came to light in 1960 in the excavation of the hotel Helios. In a warehouse used during the Hellenistic and late Roman/early Byzantine periods were found micacaeous LRA3 B4 amphorae together with unguentarium (Kontis and Konstantinopoulos 1960, 280, pl. 227). They are not from Costaina’s workshop. No published data from the excavation can yet attribute them to a Rhodian workshop, although the excavators have suggested so.

Imitations or survivals of LRA1 amphorae are produced in the Dodecanese at the workshop of the small islet Leipsoi (Papavasileiou et al., in this volume), dated by the excavators in the 8th century. Another workshop has been recently excavated by the 4th Ephorate of Byzantine Antiquities in Lefko (Λευκό) in the island of Carpathos. The excavation revealed a kiln producing survivals of LRA1 amphorae, dated probably in the 9th century (Papavasileiou and Didioumi, in press). Finally, production of LRA1 amphorae is recorded by Empereur and Picon (1989, 242-243) in the island of Rhodes.

B. Local pottery at the settlement near basilica of Mastichari West of the village of Mastichari excavations in four different plots revealed the basilica of Aghios Ioannis (Άγιος Ιωάννης), dated around AD 500 (Orlandos 1966) and a settlement of the early Byzantine period (Didioumi 2011, 98-100) (Fig. 1). In one of them, the Poursanidis’ 171

LRCW4

workshops continued to be a kind of transit station for first collecting the agricultural produce and then disseminating it.

According to the results of their survey, the production is situated in an area near the village of Apollakia (Απολλακιά), but there are still no published vessels from the area to help us classify the finds (Didioumi, in press).

Acknowledgments 5. Conclusions

I would like to thank my friends and colleagues in the Archaeological Service in Cos island, with whom I worked together for several years, and especially the archaeologists Elpida Skerlou, Dora Grigoriadou, Ersi Brouscari, Dimitris Bosnakis. A great thank-you to the director of the 4th Ephorate Mania Michailidou; the former director of Archaeological Institute of Aegean Studies Aggeliki Giannikouri and the former director of the 22nd Ephorate Melina Philimonos; the colleagues in the Archaeological Service in Rhodes island, Angeliki Katsioti and Eleni Papavasileiou. Special thanks are due to Professor Natalia Poulou-Papadimitriou, my supervisor for my forthcoming PhD on early Byzantine Coan pottery. A great thank-you, finally, to those staff of the Archaeological Service in Cos island who worked in the excavations and the conservation of the pottery, especially Odysseas Markostamos.

We would like to emphasize the need to establish a common typology for the local pottery of Cos island, especially the amphorae. We think that the LRA13 amphorae produced in the workshop of Halasarna and dated in the second half of the 6th to the middle of the 7th century by Diamanti (2010 a) could also be attributed to the family of the Byzantine Globular Amphorae, as Poulou-Papadimitriou has already suggested (PoulouPapadimitriou and Nodarou, in this volume). This could be the case also for the amphorae found in the basilica of Cardamaina and described as LRA2/LRA13 (Brouscari 2011). The finds from Cardamaina are of great importance because they are evidence for both the mass production of new types of amphorae (Byzantine Globular Amphorae) and imitations of the common commercial amphorae of the Mediterranean, the LRA1. The imitation of Ephesian types in Mastichari, taken with other evidence such as the lamp mould of Mastichari, a number of Ephesian pilgrim flasks found in the island (Katsioti 2009) and bread stamps in Cardamaina (Brouscari 2010), shows that the island of Cos had a strong connection and influence with the great pilgrimage of St. John the Theologian in Ephesos, as Brouscari (2010, 277-278) and Katsioti (2009, 281) have already suggested.

Bibliography Baldini, I. and Livadiotti, M. (eds.), 2011. Archeologia Protobizantina a Kos: la basilica di S. Gabriele, Universita di Bologan, Studi e Scavi NS 28. Bologna. Baldini, I., Contò, G. and Marsili, G. 2011. L’architettura religiosa I Kos in età Protobyzantina e gli scavi Italiani. In Baldini and Livadiotti (eds.), 117-149. Bezeczky, T. 2005. Late Roman Amphorae from the Tetragonos-Agora in Ephesos. In Fr. Krinzinger (ed.), Spätantike und mittelalterliche Keramik aus Ephesos, 203-223. Wien, Österreichischen Akademie der Wissenschaften.

Cos, as well as Carpathos and Rhodes had an important role in the maritime trade in the 5th and 6th century, using their fleets to transport local produce to big markets, but also offering services to passing ships; they were also serving the system of annona militaris (Deligiannakis 2008). The fact that the production of the commercial amphorae in Cardamaina and Mastichari continues into the 7th century allows us to consider that the island of Cos also participated in the system providing the army (annona militaris) and the capital, a practice continued in the 7th century, as Pieri (2007, 614615) has already suggested. It seems that the island of Cos, an island producing wine and olive oil since antiquity, was included in the effort made by the Byzantine state to reorganize the supply of the army and the capital, especially after losing the east provinces, like Syria and Palestine (Poulou-Papadimitriou and Didioumi 2010, 744).

Bezeczky, T. 2010. Trade connections between Ephesus and Adriatic Region. Histria Antiqua 19/2010, 351358. Brouscari, E. 2006. La città di Coo nella tarda antichità. In F. Costabile (ed.), Polis. Studi interdisciplinari sul mondo antico II 2, 317-322. Brouscari, E. 2010. Σφραγίδα άρτου παλαιοχριστιανικών χρόνων από την Κω. In Ανταπόδοση, Μελέτες Βυζαντινής και Μεταβυζαντινής Αρχαιολογίας και Τέχνης προς τιμήν της καθηγήτριας Ε. ΔεληγιάννηΔωρή, 267-280. Athens. Brouscari, E. 2011. Συμβολή στην ιστορία και την αρχαιολογία της Κω κατά την παλαιοχριστιανική περίοδο: η βασιλική του πρεσβυτέρου Φωτεινού. Unpublished PhD Thesis, University of Athens.

The finds in Kephalos, although their study is still at a preliminary stage, are of great importance for the history of the island in the transitional centuries from 7th to the beginning of the 9th centuries. We can assume that the life did not stop suddenly at the middle of the 7th century because of the Arab invasions; there were commercial activities in the area and probably overseas. It seems that the coastal settlements with their warehouses and

Brouscari, E. and Didioumi, S. 2006. Υστερορωμαϊκή και Βυζαντινή Κως. Η συμβολή των νομισμάτων (4ος12ος αι.). In Το Νόμισμα στα Δωδεκάνησα και τη Μικρασιατική τους περαία. Πρακτικά Συνεδρίου της Δ’

172

DIDIOUMI

Επιστημονικής Συνάντησης, Κως, 30 Μαΐου – 2 Ιουνίου 2003, OVOLOS 8, 297-324. Athens.

Kalopissi-Verrti, S. 1998. Παλαιοχριστιανική λίθινη μήτρα κοσμημάτων από ανασκαφή στην Καρδάμαινα της Κω. Deltion tes Christianikes Archaeologikis Etaireias Δ΄, vol. 20 (1998), 245-252.

Deligiannakis, G. 2008. The economy of the Dodecanese in late antiquity. In Ch. Papageorgiadou-Banis and A. Giannikouri (eds.), Sailing in the Aegean. Readings on the Economy and Trade Routes, Μελετήματα 53, 219-223. Athens.

Katsioti, A. 2009. Πήλινα φιαλίδια - ευλογίες από τα Δωδεκάνησα. Deltion tes Christianikes Archaeologikis Etaireias Λ΄ (2009), 271-282.

Diamanti, Ch. 2010a. Local production and Import of Amphoras at Halasarna of Kos island (5th – 7th c.). Contribution to the Resaerch of the Production and Distribution of the Late Roman/Proto-Byzantine Amphoras of the Eastern Mediterranean. S. Saripolos Library 115. Athens.

Katsioti, A. 2012. Πήλινη ημίτομη μήτρα λυχναριού πρωτοβυζαντινών χρόνων από την Κω και ζητήματα εργαστηρίων κατασκευής λυχναριών στα Δωδεκάνησα In P. Adam-Veleni and K. Tzanavari (eds.), Δινήεσσα, Τιμητικός τόμος για την Κατερίνα Ρωμιοπούλου, 555-560. Thessaloniki.

Diamanti, Ch. 2010b. Stamped Late Roman/ProtoByzantine Amphoras production from Halasarna of Kos. Rei Cretariae Romanae Fautorum Acta 41, 1-8.

Kokkorou-Alevras, G, Kalopissi-Verrti, S. and Panagiotidi-Kessisoglou, M. 2006. The Sanctuary of Apollo and the Early Christian Settlement at Kardamaina (Ancient Αλάσαρνα) on the Island of Kos. A Guide, Athens.

Diamanti, Ch. 2012. Byzantine Emperors on Stamped Late Roman/Early Byzantine Amphoras. Rei Cretariae Romanae Fautorum Acta 42, 1-5.

Kokkorou-Alevras, G. 2009. Συστηματική ανασκαφική έρευνα και έρευνα επιφανείας στην αρχαία Αλάσαρνα τη Κω. Απολογισμός εικοσαετούς δραστηριότητας. In Chr. Loukos, N. Xifaras and Kl. Pateraki (eds.), UBI DUBIUM IBI LIBERAS, Τιμητικός Τόμος για τον Καθηγητή Νικόλα Φαράκλα, 135-155. Rethymno.

Didioumi, S. 2010. Κεραμική παλαιοχριστιανικών χρόνων από την Κω: Στρώμα καταστροφής σε ανασκαφή οτην πόλη της Κω. In D. PapanikolaBakirtzi and N. Kousoulakou (eds.), Κεραμική της Ύστερης Αρχαιότητας από τον Ελλαδικό Χώρο (3ος7ος αι. μ.Χ.), Πρακτικά Επιστημονικής Συνάντησης, Θεσσαλονίκη 12-16 Νοεμβρίου 2006. Publications of the Archaeological Institute of Macedonian and Thracian Studies 8, 795-827. Thessaloniki.

Kontis, I. and Konstantinopoulos, G. 1960. Ανασκαφικαί έρευναι εις την πόλιν της Ρόδου. Η αρχαία οδός Ρ29. Praktika Archaeologikis Etaireias 1960, 279-280. Livadiotti, M. 2011. Le ricerche Italiane della Kos protobizantina. In Baldini and Livadiotti (eds.), 71-85.

Didioumi, S. 2011. Νεότερα ευρήματα για την παλαιοχριστιανική Κω από τις σωστικές ανασκαφές (with an abstract in Italian a cura di E. Chrysafi). In Baldini and Livadiotti (eds.), 87-116.

Orlandos, A. 1966. Δύο παλαιοχριστιανικαί βασιλικαί της Κω. Archaeologiki Ephemeris 1966, 1-98. Papavasileiou, E., Sarantidis K., Papanikolaou, E., in this volume. A ceramic workshop of the early Byzantine period on the island of Lipsi in the Dodecanese (Greece): a preliminary approach.

Didioumi, S., in press. Εργαστήρια κεραμικής στα Δωδεκάνησα κατά την πρωτοβυζαντινή περίοδο: το παράδειγμα της Κω. In Αρχαιολογία και τέχνη στα Δωδεκάνησα κατά την ύστερη αρχαιότητα, Ημερίδα στη μνήμη Ηλία Κόλλια, Ρόδος 2011, University of the Aegean, Department of Mediterranean Studies.

Papavasileiou, E. and Didioumi, S., in press. Εργαστηριακές εγκαταστάσεις στα Δωδεκάνησα κατά την ύστερη αρχαιότητα - πρωτοβυζαντινή περίοδο. In Το Αρχαιολογικό έργο στα νησιά του Αιγαίου, Ρόδος 2013, Διεθνές Επιστημονικό Συνέδριο, Archaeological Institute of Aegean Studies.

Empereur, J.-Y. and Picon, M. 1989. Les régions de production d'amphores impériales en Méditerranée orientale. In Amphores romaines et histoire économique: dix ans de recherché. Actes de colloque de Sienne (22-24 mai 1986), Collection de l'Ecole Française de Rome 114, 223-248. Rome.

Pieri, D. 2005. Le Commerce du vin Oriental à l’Epoque Byzantine (Ve-VIIe siècles). Le témoignage des amphores en Gaule. Bibliothèque archéologique et historique 174. Beyrouth, Institut Français du ProcheOrient.

Giannikouri, A., forthcoming. Εργασίες στην αρχαία Αγορά της Κω. Archaeologicon Deltion 59-61 (20042006), Chronika.

Pieri, D. 2007. Les centres de production d’amphores en Méditerranée Orientale durant l’Antiquité Tardive: quelques remarques. In M. Bonifay and J.-C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry, British Archaeological Report, International Series, 1662, 611-625. Oxford, BAR Publishing.

Georgopoulou, V. 2005. Κωακοί ελληνιστικοί εμπορικοί αμφορείς. Τυπολογία, χρονολόγηση, δισπορά. Unpublished PhD Thesis, University of Athens. Goutzioukostas, A. and Moniaros, X. 2009. Η περιφερειακή διοικητική αναδιοργάνωση της βυζαντινής αυτοκρατορίας από τον Ιουστινιανό Α΄ (527-565): η περίπτωση της quaestura Iustiniana exercitus. Thessaloniki.

173

LRCW4

Poulou-Papadimitriou, N. 2013. Τεκμήρια υλικού πολιτισμού στη Βυζαντινή Κρήτη: από τον 7ο έως το τέλος του 12ου αι. In Πεπραγμένα Ι’ Διεθνούς Κρητολογικού Συνεδρίου, Χανιά 2006 (2011), 381487. Chania.

Poulou-Papadimitriou and Nodarou, in this volume. Transport vessels and maritime trade routes in the Aegean from the 5th to the 9th century AD. Preliminary results of the EU funded project “Pythagoras II”: the Cretan case study.

Poulou-Papadimitriou, N. and Didioumi, S. 2010. Nouvelles données sur la production de l’atelier céramique protobyzantin à Kardamaina (Cos – Grèce). In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Comparison between Western and Eastern Mediterranean. British Archaeological Reports, 741-749. International Series 2185, Oxford, BAR Publishing.

Roumeliotis, N. Ch. 2001. Παλαιοχριστιανικά λυχνάρια από την Αλάσαρνα (σημ. Καρδάμαινα) της Κω. In G. Kokkorou-Alevras, A. A. Lemou and E. SimantoniBournia (eds.), Ιστορία – Τέχνη – Αρχαιολογία της Κω. Α’ Διεθνές Επιστημονικό Συνέδριο (Κως, 2-4 Μαΐου 1997), 255-276. Athens, University of Athens. Roumeliotis, Ν. 2003. La céramique commune d’Halassarna (Cos) et sa place dans la production et les échanges du bassin Egéen à l’époque paléochrétienne (IVe – VIIe siècles). Thèse de doctorat de l’Université Paris I Panthéon - Sorbonne.

Poulou-Papadimitriou, N. and Didioumi, S., in press. Two pottery workshops of the Early Byzantine period in the island of Cos, Greece. In Tourner autour du pot. Les ateliers et potiers médiévaux du Ve aux XIIe siècles dans l’espace Européen, Colloque International, Douai, 5-8 octobre 2010.

Skerlou, E. 1994. Πόλη Κω Ανασκαφή στο οικ. Τσιμισίρη, οδ. Μακρυγιάννη. Archaeologikon Deltion 49 (1994), 784-786.

Fig. 1. Map of Cos island 174

DIDIOUMI

Fig. 2. Cardamaina, ealry Byzantine settlement near Aghia Theotita: no1. Kiln in Roussou’s plain. no. 2. Basilica of Aghia Theotita. no. 3. Hellensitic Theater. no 4. Excavation at ancient Halasarna

175

LRCW4

Type 2

Type 1

Type 3 Fig. 3. Cardamaina. Excavation at Rousou’s plot. Globular Byzantine amphorae

176

DIDIOUMI

Fig. 4. Cardamaina. Excavation at Rousou’s plot. LRA1 imitation

Fig. 5. Mastichari, Costaina: Kiln D

Fig. 6. Kiln D. LRA B wasters

177

LRCW4

Fig. 7. Kiln D. LRA 1 imitation Fig. 8. Kiln D. Waster of Globular amphora

Fig. 9. Pot stand for LRA1 or Globular amphorae Fig. 10. Pot stand for LRA3 B amphorae

Fig.12. LRA3 B2 amphora. Bases from LRA3 B amphorae used as lids

Fig. 11. LRA3 B2 amphora

178

DIDIOUMI

Fig. 13 a and b . LRA3 B4 amporae

Fig. 14. Ephesus 56 amphora

Fig. 15. Mastichari, Poursanidi’ s plot. Lamp mould. 6th century (Katsioti 2012)

179

LRCW4

Fig. 16 a, b, c. Mastichari, Poursanidi’ s plot. Imitation of Ephesus 56 amphorae. Local production. Middle of 6th/middle of 7th century

Fig. 17 a and b. Kephalos, Survivals of LRA 1 amphorae

180

ARCHAEOLOGY AND ARCHAEOMETRY IN LATE ROMAN GREECE: THE CASE OF MAINLAND AND INSULAR SETTLEMENTS, WORKSHOPS AND IMPORTS CHARIKLEIA DIAMANTI¹, ΚALLIOPΙ KOUZELI², PLATON PETRIDIS³ ¹Ministry of Culture, 2nd Ephorate of Byzantine Antiquities, 2 Klepsydras Str., 10555 Athens, Greece [email protected] 2

Ministry of Culture, Stone Conservation Center, 79 Piraeus Str. Athens; [email protected]

3

University of Athens, Department of Archaeology and the History of Art, University Campus Zographos [email protected]

In the last few years interdisciplinary scientific collaboration has successfully combined archaeological information and the results of archaeometric investigation revealing a number of new workshops, widening thus our knowledge about the late Roman productions in Greece. Nevertheless, a comparative study of the different production centers has not yet been suggested. Into the framework of this paper the results of the archaeological study and archaeometric analysis of the pottery produced or imported on some sites in Central Greece and islands of the Aegean Sea are presented. The information gathered will contribute to the elaboration of a methodological tool for the study of newly discovered production centers on the basis of petrographic and chemical analyses. KEYWORDS: MAINLAND AND INSULAR LATE ROMAN GREECE, POTTERY WORKSHOPS, ARCHAEOLOGICAL DATA, ARCHAEOMETRIC ANALYSES, DELPHI, KOS, PAROS, THASOS route to and from Constantinople, while Delphi is a town situated in the relatively “marginal” (at that time) area of Central Greece. They also belong to two different spheres of influence: Delphi mainly imported pottery from North Africa all over the late Roman period (Petridis 2010, 139), while Halasarna imported mainly from Asia Minor.

Introduction The late Roman pottery found in mainland and insular regions of modern Greece has not yet contributed to the economic history of early Byzantine times as it could, for two well known reasons: the published archaeological remains of pottery workshops in Greece are rather rare and pottery examination has not yet been carried out in large scale. The existing knowledge about the local productions and the financial activities they witness remain therefore incomplete.

In both cases laboratory examination of tiny pieces of sherds has been carried out. All the analytical techniques available at Stone Conservation Centre of the Hellenic Ministry of Culture for fabric examination have been used. More specifically:

The aim of this paper is to present comparatively the archaeometric results and the archaeological interpretation of two characteristic cases of recently published pottery workshops situated in insular and mainland Greece respectively. The material from two more insular sites, in Paros and Thasos, in Central and Northern Aegean respectively, is currently under interdisciplinary examination and will soon be compared with the results from the first two cases.

 Observation of fresh fractures and smooth (polished) sections under the stereoscope.  Petrographic examination (examination of thin sections under the polarizing microscope) that provides information on the nature, size, shape, number and distribution of the mineral grains present.

These two cases are the mainland site of Delphi (Fig. 1) in Central Greece and one of Kos’ harbours, Halasarna (Fig. 2). The two sites share some common historical features like the progressive expansion of their urban plan in the 5th and mainly the 6th century and their abandonment during the 7th century. But, against a widely spread opinion, the data from the study of pottery show that the last decades of the cities’ life are not characterized by poverty and isolation (Petridis 2007, 53; 2010, 141-143).



X-Ray Diffraction



FT-IR Spectroscopy

 Chemical analysis for major and minor elements by Flame Atomic Absorption Spectroscopy (fusion with lithium metaborate-mixed standards). Delphi-Archaeological Data The late Roman ruins of Delphi have been unearthed mainly during the “Grande Fouille” conducted from 1892 to 1903 by the French School of Archaeology at Athens; at that time, pottery was not preserved. Almost a century later, systematic excavations by the French School have

One could also observe an important difference between these two medium-sized cities: the harbour of Halasarna on the island of Kos might be a crossroad on the main 181

LRCW4 investigated thoroughly, from 1990 to 1997, two secular buildings of the archaeological site; one of them was public, the Roman Forum (usually called Roman Agora) and the other one was private, the so-called South Eastern Villa (Petridis 2009, n. 18). From both these excavations derives the main corpus of the studied pottery, although a smaller quantity also comes from the Xystos of the ancient Gymnasium, excavated in the late 1980’s by the French School and the Greek Archaeological Service. Finally, in the necropolis of the late Roman town and in some of its houses (for the late Roman houses of Delphi see Petridis 2005; 2008, 251-252) lamps and tableware of great interest have been found (Fig. 3).

manufactured at the same time and were destined for various domestic uses, classification in two groups according to the analysis was proposed:

Pottery workshops have been discovered at two points of Delphi. The most ancient among them came partially into light in the Xystos area of the ancient Gymnasium (for the pottery workshops discovered at Delphi see Petridis 2004; 2007; 2010, 35-38). Coins and typological analysis of the pottery date them to the second half of the 4th c. AD. Rests of two kilns in a very bad condition have been discovered there, along with a rectangular pit and two small cisterns. One of the kilns, 4m long, is constructed with bricks; the other is smaller and contains spolia from earlier buildings (Fig. 4). Kiln-wasters have been discovered around as well as a small number of kiln supports.

GROUP I (Figs. 9-10)

The Southeastern Villa, the largest to-date excavated architectural complex in late Roman Delphi, was used as a residence from the beginning of the 5th c. until the last quarter of the 6th c. Stratigraphy shows that, around 590 AD workshops of pottery, metalwork and tannery or dyeworks were installed inside the quite recently abandoned villa. The most important installations belong to pottery workshops (Fig. 5). Six kilns have been unearthed. They are all of small dimensions (the larger measures 2x2m2). They are square or rectangular in shape and they are built with bricks and small stones; they lay partly on the walls of the villa (even the apse of a triclinium) or on the rock; bricks or half-broken pithoi necks have been used to support their grills. Three of the kilns have been constructed inside the same room, side by side, but they must have not functioned simultaneously. Additional installations belonging to the workshops have been also discovered around the kilns: a cistern, basins, water-pipes and deep pits serving in a second use, probably, as clayworking pits; two of the rooms of the villa, were transformed into deposits (Fig. 6). Thousands of kilnwasters, moulds of lamps or moulds of metal objects and kiln supports have been found into the deposits (Fig. 7).

GROUP II (Figs. 11-12)

Delphi-Archaeometric Data The forty one (41) sherds from the late Roman excavations at Delphi that have been examined at the Stone Conservation Centre (for detailed results see Kouzeli 2010a, 175-180), can be classified in three groups, I-III.

Nine (9) samples belong to this group. Their fabric is very fine-grained, a texture that imposes chemical analysis. The results of analysis by Flame Atomic Absorption Spectroscopy for major and minor elements confirmed the classification in this group (Fig. 33). Fragments of feldspars, quartzitic rock, older ceramic material and pyroxenes usually smaller than 30μm are present. Occasionally tiny mica particles have been observed, while calcite is always absent.

Group II consists of eighteen (18) samples, including clay sample from the floor of local late Roman kiln from the area of the Southeastern Villa. The samples of this group are coarser-grained since they contain numerous particles larger than 30 microns that imposed petrographic examination. Feldspars, micas, quartzitic rock fragments and grains of older ceramic are systematically observed in analogous grain densities. GROUP III (Figs. 13-14) The samples of group III are also coarse-grained since they contain numerous particles larger than 30 microns. Feldspars, micas, quartzitic rock fragments and grains of older ceramic are systematically observed in grain’s densities similar to group II. On the other hand, calcareous microfossils are systematically observed in the samples of group III which is in accordance with higher calcium content in this batch of samples (Fig. 33).

The local production comprises all kind of pottery destined to cover many indoor and outdoor needs (Fig. 8). The clay is of quite good quality, generally light red to pink, well fired with a lightly soaped external surface of light red, light brown or reddish yellow colour. A slip of red or weak red colour covers the upper part of the vases leaving intentionally large tears descending vertically or obliquely to the bottom.

A significant number of samples are classified in the previously described three groups. Although very finegrained, two samples (No 34 and 35) are not classified in group I because of their texture and chemical composition. On the basis of the archaeological information these two samples belonged to imported Central Greek Painted Ware (a detailed study of the Central Greek Painted Ware in Petridis 2009b) manufactured most probably at Nea Anchialos (or Thessalian Thebes).

Taking into account that the samples from the late Roman excavations at Delphi belong to objects that were not 182

DIAMANTI-KOUZELI-PETRIDIS amphoras is covered by a white-pink wash, which in some cases, because of the fire of the vase, is more green or yellow.

The samples 5 and 6 are coarse-grained and differ from all samples examined in texture as well as in mineralogical and chemical composition. They are not classified in any group, which is in agreement with the archaeological examination according to which these samples belonged to an African Red Slip plate and a lamp respectively.

Kos-Halasarna-Archaeometric Data Forty two (42) samples of amphora sherds from the late Roman excavation at Halassarna have been examined.

Consequently, local products are clearly distinguished from the imported ones like the Red Slip Wares, mainly from Tunisia or the more close geographically Central Greek Painted Wares the interdisciplinary analysis of which showed the non Delphic origin of the type.

The systematic presence of inclusions in the matrix of all Halasarna samples examined, imposed petrographic examination. The following classification in three main groups on the basis of the laboratory data is proposed (for detailed results see Kouzeli 2010b, 137-149, 218-219):

Kos-Halasarna-Archaeological Data Since 1985, the University of Athens has carried out excavations and surface research at Halasarna of Kos (Figs. 2, 15) (Kokkorou-Alevras et al. 2006, 46-67; Diamanti 2010c, 23-29, 193-196). The evidence provided by this research indicates that from the beginning of the 5th c., a flourishing settlement, extending over the old pagan sanctuary of Apollo, developed at Halasarna. The settlement was occupied until the Arab invasions (middle of the 7th c. AD). As demonstrated by the ceramic evidence, Halasarna became deeply enmeshed in the economic life of the late Roman state as a commercial node on the East Aegean maritime routes. In addition to amphora, fineware and lamp imports (from Asia Minor, Cyprus, Aegean Sea centres, North Africa and Palestine), local production took place in the late Roman settlement (Figs. 16-23). The waste deposits of distorted and misfired examples implied local production of amphoras. According to Ch. Diamanti’s PhD study results on the koan amphoras of late Roman Halasarna, the settlement was an important production centre of the cylindrical LR 1 and ovoid LR 13 amphora types from the end of the 6th until the first half of the 7th c. (Diamanti 2010a, 146-7; 2010c, 80-114, 207-216, 331-399: S/N 359-617, fig. 87139, pl. 11-19. For a nearby late Roman kiln located by the Αrchaeological Service, see Poulou and Didioumi 2010). Production of these amphoras on a large scale as well as their stamps depicting the emperor provide evidence that they were destined for export of the settlement's agricultural goods as part of a tax to the state in kind. This is the first time that a workshop of stamped late Roman amphoras can be identified with certainty, at least as far as the Aegean is concerned. We have suggested elsewhere that the stamps were part of a state mechanism of financial control performed by Kommerkiarioi (commercial inspectors), (Diamanti 2010b, 1-6; 2010c, 106-7, 214-5; 2012; forthcoming b). The standardization and state-involved production of Κoan amphoras is supported by the fact that morphologically they follow strictly the well-known and widely distributed shapes of late Roman Amphoras 1 and 13 which is a different case of the locally produced late Roman amphoras from Delphi presenting strong individual characteristics (Pétridis 2010, 45-53).

GROUP I (Figs. 24-25) The majority of the samples examined (twenty-six samples) presents strong similarities concerning the nature, size, distribution and density of inclusions, consisting thus one group (Diamanti 2010c, 142-149, 219 “group B”, the samples of which on the basis of the archaeological information belong to the local amphora types, LRA 1 and 13). More specifically, angular plagioclase particles, angular quartz grains, muscovite in elongated thin needles and thicker flakes as well as rather rare calcareous grains of various uneven shapes (often weathered fossils) are distributed in the fine-grained matrix. Despite the fact that in the same amphoras the mineral grains’ density varies from place to place, that reflects to the chemical composition, the density of the inclusions in the samples examined was similar. Plagioclase, quartz and mica grains of similar size, distribution and density are present in the glassy matrix of the dull grey-green misfired pottery samples, classifies them to group I, confirming that are of local production. Chemical analysis is in accordance with the results of petrographic examination (Fig. 34). Furthermore, sherds belonging to this group present similarities to Hellenistic fabric from the same site and local clay samples (after the appropriate laboratory treatments), concerning the nature and amount of the mineral phases present, the Hellenistic being the finergrained. The members of group I are clearly distinguished from a number of samples from the same excavation also examined here: GROUP II (Figs. 26-27) The six samples of Group II bear a much higher number of inclusions compared to Group I: calcareous rounded particles, very short flakes of mica, feldspars, silicate rock fragments, quartz and older pottery grains in minor quantities are evenly distributed. The sherds of this group are archaeologically considered as belonging to amphoras of type LR 1 imported from Asia Minor and Cyprus (Diamanti 2010c, 141-142, 218 “group A1”). The chemical composition differentiates them from those of Group I (Fig. 34).

Macroscopically the local fabric is reddish orange, bearing inclusions. Often the outside surface of the 183

LRCW4 supported by the archaeometric examination of 94 samples. The correlation of archaeological, chemical and petrographic data reinforce the fact that two sites exhibited local production as well as import activity: According to the results of petrographic examination and chemical analysis for major and minor elements the sample from the local kiln at Delphi belongs to one of the three main groups of samples (Delphi - group II), while the distorted pottery from Halasarna belong to the same group with the majority of the samples examined (Halasarna - group I) that implies local production. On the other hand in both cases a significant number of samples are clearly distinguished from the local ones and according to the archaeological interpretation belonged to imported wares.

GROUP III (Figs. 28-29) Although also characterized by the presence of numerous inclusions, the samples of this group present differences from group II, concerning the shape of quartz grains, the number of calcite particles as well as in chemical composition (Fig. 34). According to the archaeological classification the sherds of group III belonged to imported Palestinian amphoras of types LR 4 and 5 (Diamanti 2010c, 142, 218-219 “group A2”). Paros (Zoodhochos Pighi) and Thasos In Paros island, in the Zoodhochos Pighi inlet, the second, after Halasarna, workshop, which produced both LR 1 and LR 13 (Fig. 30) amphora types has been located (Diamanti forthcoming c), indicating that this coproduction of the two famous late Roman amphora types could be a common practice in the Aegean Sea area of the second half of the 6th – 7th c. The amphoras were produced in a large scale and were obviously destined, in a big part, for export of parian agricultural goods. According to the preliminary results of observation of thin sections under the polarizing microscope this production is characterized by the presence of numerous inclusions up to 700μm (0.7 mm) unevenly distributed, calcareous, often heavily weathered particles among which are discerned bioclasts, plagioclase, orthoclase and quartz and mica grains with preferred orientation as well as silicic rock fragments are observed (Fig. 31).

Bibliography Bimson, M. 1969. The Examination of Ceramics by Xray powder Diffraction. Studies in Conservation 14, 83-89. Blondé F., Gros J.-S. and Pétridis P. 2011. La céramique au quotidien à Thasos de l’époque archaïque à l’époque protobyzantine. Revue des Études Grecques 124, 193-204. Catling H. W., Richards E. E. and Blin-Stoyle A. E. 1963. Correlations between composition and provenance on Mycenaean and Minoan pottery. The Annual of the British School at Athens 58, 94-115. Catling H. W. and Millet A. A. 1965. Study of the compositional patterns of Mycenaean pictorial pottery from Cyprus. The Annual of the British School at Athens 60, 212-224.

This material as well as the pottery coming from an excavation of a late Roman villa in the island of Thasos (Fig. 32) are under investigation (joint project of the Greek Ministry of Culture, the French School of Archaeology of Athens, the National and Kapodistrian University of Athens, the University of Lille 3-Charles de Gaulle and CNRS-France). The thasian material studied so far is typologically and ceramologically very rich as one would expect from an island with such a strong pottery making tradition, situated near the Via Egnatia and on the maritime routes to Constantinople. The imported amphorae cover most of the well documented types of the Western and Eastern Mediterranean with a strong presence of the LR 2 type. Among the numerous finds of imported tableware, the Asia Minor productions are the best represented. Of great importance is the discovery of glazed pottery in surely dated 6th century contexts. Its presence contradicts the formerly expressed hypothesis of its first appearance in Greece during the 7th century (Blondé et al. 2011, 200; Petridis 2013, 16).

Diamanti, Ch. and Kouzeli, K. 2009. Νέα αρχαιολογικά και αρχαιομετρικά στοιχεία για την παραγωγή και εισαγωγή πρωτοβυζαντινών αμφορέων στο Αιγαίο: Η περίπτωση της Αλάσαρνας στην Κω. Book of abstracts of the 29th Symposium of Byzantine and Post-Byzantine Archaeology and Art. Christian Archaeological Society, 36-37. Athens. Diamanti, Ch. 2010a. Εντόπιοι υστερορωμαϊκοί/ πρωτοβυζαντινοί αμφορείς από την Αλάσαρνα της Κω. In D. Papanikola-Bakirtzi and Ν. Kousoulakou (eds.), Κεραμική της Ύστερης Αρχαιότητας στον Ελλαδικό χώρο (3ος-7ος αι. μ.Χ.), Proceedings of a Scientific Meeting, Thessaloniki 12-16 November 2006, Publications of the Archaeological Institute of Macedonian and Thracian Studies 8, 143-152. Thessaloniki. Diamanti, Ch. 2010b. Stamped Late Roman/ProtoByzantine amphoras production from Halasarna of Kos. Rei Cretariae Romanae Fautorum Acta 41, 1-8.

Conclusion The comparative study of two late Roman sites the rather isolated marginal site of Delphi and the harbor of Halasarna has been attempted. The first results of a new project on the systematic interdisciplinary examination of Zoodochos Pigi at the isle of Paros have been also considered. The archaeological documentation has been

Diamanti, Ch. 2010c. Local Production and Import of Amphoras at Halasarna of Kos Island (5th-7th c.). Contribution to the Research of the Production and Distribution of the Late Roman / Proto-Byzantine 184

DIAMANTI-KOUZELI-PETRIDIS Amphoras of the Eastern Mediterranean. S. Saripolos Library 115, Athens.

Petridis, P. 2005. Un exemple d’architecture civile en Grèce: les maisons protobyzantines de Delphes (IVe– VIIe s.). Mélanges Jean-Pierre Sodini. Travaux et Mémoires 15, 93-204. Paris.

Diamanti, Ch. 2012. Byzantine Emperors on Stamped Late Roman/Early Byzantine Amphoras. Rei Cretariae Romanae Fautorum Acta 42, 1-5.

Petridis, P. 2007. Relations between pottery workshops in the Greek Mainland during the Early Byzantine Period. In B. Böhlendorf-Arslan, A. O. Uysal and J. Witte-Orr (eds.), Çanak. Late Antique and Medieval Pottery in Mediterranean Archaeological Contexts. International Symposium, Çanakkale 31 May - 3 June 2005, Byzas 7, 43-54.

Diamanti, Ch., forthcoming a. Υστερορωμαϊκή/ Πρωτοβυζαντινή Κεραμεική. In G. Kokkorou-Alevras and Κ. Kopanias (eds.), Η επιφανειακή έρευνα στον Αρχαίο Δήμο των Αλασαρνιτών της Κω. Athens, University of Athens. Diamanti, Ch., forthcoming b. Halasarnian Stamped Late Roman Amphoras-New Light on the Mechanism of Production and Distribution of Koan Stamped Amphoras during Late Antiquity. In N. Badoud and A. Marangou (eds.), Analyse et exploitation des timbres amphoriques grecs. Colloque international sous le haut patronage de l'Académie des Inscriptions et Belles-Lettres, École Française d'Athènes, Université de Rennes 2 – Haute Bretagne (Athènes, 3 - 5 février 2010).

Petridis, P. 2008. Παρατηρήσεις στις πόλεις και τις αστικές οικίες της Ύστερης Αρχαιότητας στον ελλαδικό χώρο. Deltion tes Christianikis Archaeologikis Etaireias 29, 247-258. Petridis, P. 2009. A New Approach to an Old Archaeological Site: the Case of Delphi. In J. Bintliff and H. Stöger (eds.), Medieval and Post-Medieval Greece: The Corfu Papers. 101-105. Oxford. Pétridis, P. 2009b. Les productions protobyzantines de céramique peinte en Grèce Continentale et dans les îles. In J. Zozaya, Μ. Retuerce, M. A. Hervás, and A. de Juan (eds.), Actas del VIII Congreso Internacional de Cerámica Medieval en el Mediterráneo, tomo I, 39-48. Ciudad Real.

Diamanti, Ch., forthcoming c. Amphoras production at the Aegean Sea during the 5th-7th c. The case of a workshop at Paros Island - Preliminary results. In Tourner autour du pot. Les ateliers de potiers médιévaux du Ve au XIIe siècles dans l’espace européen (Douai 5-8 Οctobre 2010).

Pétridis, P. 2010. La céramique protobyzantine de Delphes. Une production et son contexte. Fouilles de Delphes V, Monuments figurés 4. École Française d’Athènes, Paris.

Kokkorou-Alevras, G., Kalopissi-Verti, S. and Panayotidi-Kesisoglou M. 2006. The sanctuary of Apollo and the Early Christian settlement at Kardamaina (ancient Halasarna) on the island of Cos: A guide, Athens.

Petridis, P. 2013. Η Πρωτοβυζαντινή κεραμική του ελλαδικού χώρου, Athens.

Kouzeli, K. and Zachos, G. A. 2006. Middle/Late Roman Elateian Ware (Phocis, Greece). Archaeological – physicochemical - petrological analysis. Rei Cretariae Romanae Fautorum Acta 36, 555-560.

Poulou, Ν. and Didioumi, S. 2010. Nouvelles données sur la production de l'atelier céramique protobyzantin à Kardamaina (Cos-Grèce). In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Comparison between western and eastern Mediterranean, British Archaeological Reports, International Series ΙΙ, 741-750. Oxford. BAR Publishing.

Kouzeli, K. 2010a. Les analyses de pâte. In P. Petridis 2010, 175 – 180. Kouzeli, K. 2010b. Petrograpic and chemical analysis of the Halasarna amphoras. In Ch. Diamanti 2010c, 137149, 218-219. Magou H. and Kοuzeli, K. 1986. Hydriae Hadra: Composition and Provenance, 25th International Symposium on Archaeometry, Athens 1986, Book of Abstracts, 111.

Roumeliotis, Ν. 2001. Παλαιοχριστιανικά λυχνάρια από την Αλάσαρνα (σημ. Καρδάμαινα) της Κω. In G. Kokkorou-Alevras, A. Laimou and E. SimantoniBournia (eds.), Ιστορία - Τέχνη - Αρχαιολογία της Κω, 1ο Διεθνές Επιστημονικό Συμπόσιο, Κως 2-4 Μαΐου 1997, Archaeognosia Supplement Series 1, 255-276. Athens.

Peacock, D. P. S. 1970. The scientific analysis of ancient ceramics. A review. World Archaeology 1.3, 375-389. Petridis, P. 2004. Βιοτεχνικές εγκαταστάσεις της πρώιμης βυζαντινής περιόδου στους Δελφούς. In Αρχαιολογικά Τεκμήρια Βιοτεχνικών Εγκαταστάσεων κατά τη Βυζαντινή Εποχή 5ος–15ος αιώνας, 243-256. Athens, Piraeus Bank Group Cultural Foundation.

Tite M. S. 1972. Methods of Physical Examination in Archaeology. London, Seminar Press.

185

LRCW4

 

Fig. 1. Delphi, residential remains of the late Roman period.  

     

             

       

Fig. 2. Halasarna of Kos, Part of the late Roman settlement.      

186

DIAMANTI-KOUZELI-PETRIDIS

                                        

Fig. 3. Imported pottery found at Delphi: “Central Greek Painted Ware” plate.                                                                                                                                               

   Fig. 4. One of the kilns discovered at the Gymnasium area of Delphi.              

Fig. 5.

Fig. 6.  

Fig. 5. Pottery workshop inside a triclinium of the Southeastern Villa, Delphi. The entrances of two kilns and a levigating basin. Fig. 6. Deposit next to the kilns of the pottery workshop, Southeastern Villa, Delphi.      

Fig.7. Clay mould for the manufacture of ‘North African type’ lamps locally produced at Delphi. 187

Fig. 8. Vase from the 6th–7th c. local production of Delphi

LRCW4 .

            Fig. 9.

Fig. 10.  

Fig. 11.

Fig. 12.  

           

              Fig. 13.

Fig. 14.  

Fig. 9. Delphi-Group I –fresh fracture under the stereoscope. Fig. 10. Delphi-Group I –Thin section under the polarizing microscope (crossed prisms). Fig. 11. Delphi-Group II – smooth (polished) section under the stereoscope. Fig. 12. Delphi-Group II – Thin section under the polarizing microscope (crossed prisms). Fig. 13. Delphi-Group III – fresh fracture under the stereoscope. Fig. 14. Delphi-Group III – Thin section under the polarizing microscope (crossed prisms).

 

188

DIAMANTI-KOUZELI-PETRIDIS

   

 

       

Fig. 15. Halasarna of Kos, General plan - Sanctuary of Apollo and Late Roman Settlement (Drawing by G. Antoniou).

               

Fig. 16. Halasarna of Kos, Koan amphora of the LR 1 type (last quarter of the 6th -first quarter of the 7th c.). Fig. 17. Halasarna of Kos, Koan amphora of the LR 13 type (last quarter of the 6th -first quarter of the 7th c.). Fig. 18. Halasarna of Kos, imported (Aegean) amphora of the LR 2 type (late 5th – early 6th c.). Fig. 19. Halasarna of Kos, imported amphora of the LR 1 type (middle of the 6th-first half of the 7th c.). Fig. 20. Halasarna of Kos, imported (Palestinian) amphora of the LR 5/6 type (6th-7th c.). Fig. 21. Halasarna of Kos, imported (Aegean) amphora of the “Samos cistern” type (middle of the 6th c.).

        

Fig. 22. Koan amphora of the LR 1 type (Drawing by Ch. Diamanti).

Fig. 23. Koan amphora of the LR 13 type (Drawing by K. Mpairaktaris). 189

LRCW4

           

Fig. 24.

Fig. 25. 

              Fig. 26.

Fig. 27. 

            Fig. 28. Fig. 29 

Fig. 24. Halasarna of Kos -Group I – smooth (polished) section under the stereoscope. Fig. 25. Halasarna of Kos -Group I – Thin section under the polarizing microscope (crossed prisms). Fig. 26. Halasarna of Kos -Group II – smooth (polished) section under the stereoscope. Fig. 27. Halasarna of Kos -Group II – Thin section under the polarizing microscope (crossed prisms). Fig. 28. Halasarna of Kos -Group III – smooth (polished) section under the stereoscope. Fig. 29. Halasarna of Kos -Group III – Thin section under the polarizing microscope (crossed prisms). 190

DIAMANTI-KOUZELI-PETRIDIS

              Fig. 30. Naoussa of Paros – Parian amphora of LR 13 type.

Fig. 31.Naoussa of Paros- polished section under the stereoscope.

             

Fig. 32. Imported and locally produced tableware from Thasos (photo: G. Naessens, Halma-Ipel)          

            Fig. 33. Aluminum, iron, calcium, magnesium, manganese, potassium, sodium, chromium and nickel oxides (%) in Delphi samples (groups I, II and III). 191

LRCW4

                    Fig. 34. Aluminum, iron, calcium, magnesium, manganese, potassium, sodium, chromium and nickel oxides (%) in Halasarna samples (groups I, I Distorted, II and III).

 

192

A LATE ROMAN WORKSHOP AT DILESI IN BOEOTIA EUGENIA GEROUSI Hellenic Ministry of Culture and Sports, 20-22 Bouboulinas Str., 106 82 Athens, Greece [email protected] In the town of Dilesi on the east coast of Boeotia, where a settlement in use until the late Roman period has been excavated, a pottery workshop was revealed during a recent rescue excavation. The workshop included five kilns: two of them are dated in the Hellenistic era while the other three are dated in the late Roman period. This paper is a preliminary study since only 60% of the total of ceramic finds of the excavation has been sorted. A ‘carrot-shaped’ amphora type represents an important percentage of the late Roman workshop’s products. This type which looks like that known as Schöne-Mau XV/Camulodunum 189/Peacock and Williams Class 12, is almost unknown in the Aegean regions, whereas it is found in the northern parts of Europe. The second type of amphora produced in the kilns of Dilesi is an early example of the well-known Late Roman 2 type. Among the finds from the kiln several upper parts of these early LR 2 amphorae demonstrate different arrangements of the neck, mouth and handles. The discoveries lead to the conclusion that this ceramic workshop, dated in the 4th century AD, produced on the one hand a type of amphora that was exported to the North and on the other hand an early example of the widely known LR2 amphora, at a time when its shape has not yet been standardized. KEYWORDS: LATE ROMAN CERAMICS, CARROT-SHAPED AMPHORA, EARLY LR2 AMPHORA, DILESI CERAMIC WORKSHOP, QUANTITY MEASUREMENT DIPINTO about afterlife, the dogma and the ritual of the church (IG VII 582-584; Duchesne 1879; Platon 1937).

The town of Dilesi on the east coast of Boeotia (Fig. 1) is identified as the ancient Delion originally founded in the Classical era as a branch of the famous sanctuary of Apollo on Delos, and that later became the seaport of Tanagra (Thoukydides IV, 118; Strabo IX, 403; Pausanias IX, 20.1). The rescue excavations at the modern settlement have so far revealed parts of the ancient town such as a stoa-type of building (Pitteros 2000), Roman commercial complexes and workshops where a large number of transport vessels and especially amphorae have been discovered (Charami 1995), a complex of public baths with marble basins that were being kept until recently in the yard of the old Museum at Thebes (Charami 1996a), a well-made rectangular ceramic kiln for bricks with a grate of large dimensions nearby the port, all dated to the Roman period, (Charami 1996b), and a cemetery which was continuously used from the Archaic era until the 7th century AD, located on the nearby low hill of Agrilesa (Chamilaki 2010), as well as sporadic burials on the fringes of the settlement. The development of the town is linked to the prosperity of the fertile area of Tanagra during late Roman period.

A recent rescue excavation that was carried out by the 9th Ephorate of Prehistoric and Classical Antiquities in a plot at a key point of the settlement has revealed a pottery workshop that included five kilns, a well, large depositions of clay, and a rectangular chamber used presumably for storage (Charami forthcoming). This workshop is located just 40m away from the aforementioned rectangular brick kiln. Thus, it is possible that this was a district with workshops for ceramicproduction. Two kilns II and III are dated in the Hellenistic era while the other three, I, ΙV and V are dated in late Roman period. The best-preserved kiln I (5.45 x 5.50m.) has a beautiful arched entrance on the west side as well as a clay floor (Fig. 2). Today it is preserved in the basement of the modern house built on the site. The excavation has revealed a lot of ceramic fragments ranging from large storage vessels, especially amphorae, to triangular clay wedge-supports for vessels in their firing as well as biconical support-rings many of which are inscribed and which were used to assist in the drying or the display of the vessels. All the above constitutes evidence of workshop activity.

According to Synekdemos of Hierokles, in the beginning of the 6th century all the substantial Boeotian cities in antiquity such as Thebes, Chaironia and Tanagra, continued to function as important urban centers after they were christianized (Koilakou 2006). Tanagra was the bishop’s seat since the 4th century: a very important inscription dated to the 4th century has been discovered in the area. It is one of the longest inscriptions of this era, very rich in information about burial customs, beliefs

Up to now 60% of the total of ceramic finds made in the excavation has been sorted. The presentation of this material is today considered preliminary, as a large amount of the total ceramic remains are yet to be studied. A ‘carrot-shaped’ amphora represents an important percentage of the workshop’s products. The best preserved example Α46 (Fig. 3) with a total height of 193

LRCW4

0.58m has the ‘carrot-shaped’ body and hollow toe, two slightly raised strap handles that start and end on the body (rather than ending on the broad and very short neck), and a distinctive wide mouth. The excavation of the site of the kilns has revealed a large number of similar conical bases, about 60. A large amount (60-70) of bodyfragments of similar amphorae has also been found (Fig. 4). By the first assessment, the vessels produced at the Dilesi workshop had approximately the same dimensions: none at least was taller than the A46. Several samples of badly fired upper parts have been identified as discards from the workshop (Fig. 3).

The second type of amphora produced in the kilns of Dilesi is the well-known Late Roman 2 type, which was produced for some three hundred years from the beginning of the 4th to the 7th centuries: it shows a development of form over that long period (Riley 1981, 122; Pieri 2005). There are two well-preserved examples in Dilesi: the first consists of almost the whole body without the base and the second the whole except the neck (Figs 5 & 6). They have a large globular body with straight horizontal combed grooving, the point of the maximum diameter tends to be towards the upper part of the body with two short bowed handles from the shoulder to the neck. The base is rounded and has a small projecting toe or knob. The short conical neck is surmounted by a broad funnelshaped mouth. The lip is rounded and slightly thickened on the inside. This type belongs to the very early examples of LR2. Several body fragments and several upper parts with the funnel-shaped mouth show that this type at Dilesi varies in dimensions, with its height ranging from about 75 to 80cm.

The amphora of Dilesi looks like the ‘carrot-amphora’ type known as Schöne-Mau XV/Camulodunum 189/Peacock and Williams Class 12. This type is commonly associated with Roman military sites in Britain, France, Switzerland and Germany. According to Peacock and Williams ‘a Mediterranean source is likely, but the dearth of finds from this area is curious. It seems that the manufacturing centre directed most of its produce to the northwestern parts of the empire’ (Peacock and Williams 1986, 109). The main difference between the Dilesi ‘carrot-amphora’ and the Camulodunum 189 is the raised handles, while all the published examples from the north-west parts of the Roman Empire bear small thick loop handles. Another difference is the size: the height of the northern finds is between 0.30 and 0.48 m (Reusch 1970).

Certain upper parts of amphorae show different constructions of the neck, mouth and handles (Fig. 7a-f). Some examples do not bear the typical funnel-shaped mouth but instead a cylindrical, collar-shaped or flared mouth. The handles sometimes appear as small and arched, while others are long and angular. According to a preliminary estimation there are various subtypes of LR2 that can be identified among the production of this workshop.

Amphorae termed ‘carrot-shaped’ were also common in the Levant, especially in Lebanon and north-west Palestine. Among the different subtypes, as Reynolds has divided them, the Dilesi ‘carrot-shaped’ amphora has a slight similarity, as far as the shape of the rilled body is concerned, with only two subtypes, the so-called Beirut amphora and the north-Palestinian amphora 14 (Reynolds 2005, 569-570, pl. 10-11, pl.14). Some ‘carrot-shaped’ amphora variants from the area with a similar short ‘collar’ neck and wide mouth have been associated with the cultivation and transport of dates (Reynolds 2005, 571, Fig. 135).

Demesticha (2010) has recently published a preliminary interpretation of the cargo of a shipwreck (shipwreck 7) on the west coast of the south part of the Pagasitic gulf. Although the wreck had been disturbed, it was possible to show that the ship’s cargo consisted of at least three types of amphora which all contained wine and could be dated to the late 4th century. A significant part of the cargo consisted of amphorae (T1) which typologically have many points of similarity to the LR2 type, while the main part of the cargo consisted of typical LR2 amphorae (T2), in one of its early versions. This is the only large group of amphorae of this type known so far. Both types have similarities with the amphorae produced in the workshop of Dilesi. Demesticha has already pointed out that the provenance of the two types of amphorae is probably to be sought in the Peloponnese or nearby areas, and that the similarities between the two types indicate that they probably come from regions located close to each other. It is highly likely that one such region could be Dilesi and its workshop.

Another type of amphora with a similar ‘carrot-shaped’ body but a completely different upper part is the Sinope ‘carrot-amphora’ which is well known all around the Black Sea area, the Lower Danube, and the north and south Black Sea areas. This type can be dated from the second quarter of the 4th century to the end of the 5th century and it is assumed that these products were used as transport vessels for wine (Opaiţ 2010, 113-114). The distinctive features of the Dilesi ‘carrot-shaped’ amphorae are the slightly raised strap handles that start and end on the body rather than on the neck, and the wide mouth. The wide mouth means that these amphorae cannot be associated with liquid commodities but rather with olives or dried fruits, as is also assumed for a comparable amphora found in the Kerameikos as well as for those similar from the north parts of Europe (Böttger 1992, 341-342; Reusch 1970, 61).

Τhe LR 2 amphora is a common type used for a very long period and has been classified in several typologies of excavations from Carthage and Benghazi of North Africa or from several places on the Black Sea shores. Many scholars like Riley or recently Poulou-Papadimitriou and Nodarou (2007) have demonstrated an Aegean origin, as in the Argolid or Cos.

194

GEROUSI

As Riley pointed out already in 1981, although the LR 2 amphora type is widespread throughout the Mediterranean, including south France, and is also found in Britain, it does not predominate at any published site so far. With respect to the Aegean Sea, it has been recorded at a number of sites, covering chronologically a long period from the 4th to the 7th centuries, including Athens (Robinson 1959, M 272, 109, pl. 29 and P 4129, pl.40), Chios (Boardman 1989, 106), Corinth (Adamsheck 1979, 114-115, RC 14 and 16, pl. 26, 27; Slane 2000, 304), Delphi (Pétridis 1995, 281-283; 2003 (2), 534), Samos (Hautumm 1981, 21; Steckner 1989, 5771; Gerousi 1997, 252-254), Thasos (Daux 1965, 947, Fig. 22, 23; Abadie-Reynal and Sodini 1992, 56-57, CC284, CC285, Fig. 24, pl. Va,b,c,e), Crete (Vogt 1991/93, 61-62; 2000, 83-84; Portale-Romeo 2000, 422) and Argos (Aupert 1980, 440, fig 46, no 325a, 326a; Abadie-Reynal 1989 (1), 51-52) to name only a few. But evidence for its quantification is lacking and it remains difficult to establish any statistical framework for the occurrence of the type. The LR2 amphora type was then developed into the so-called Byzantine globular amphora (Hayes 1992, 66-73; Poulou-Papadimitriou 2001, 245247; 2013, 116-118).

refers to the total amount of a cargo and not to the specific amphora. The dipinti on amphorae of LRA 2 type are rare, while the examples of incised graffiti are more numerous. From the published dipinti in amphorae one group concerns written inscriptions that state the capacity or the real weight of the amphorae. A characteristic example is that of around 120 amphorae that were found in Tomi, west of the contemporary port of Constanza, during the excavation of a notable Roman building with mosaics. These amphorae were considered to be from an Aegean site between the 4th and 6th centuries (Rădulescu 1973, 193-207; Scorpan 1976, 178). The amphorae bear both dipinti and graffiti which were written in Greek, usually Ν, ξ and θ and it was suggested that they symbolize the quantity of the product that was carried by these amphorae at some point. Two LRA2 amphorae from Dinogetia of the 6th century bear dipinti in Greek that state the capacity using xestes as a unit of measurement (Barnea 1966, 244-5, Fig. 5.7, 8.7). A large number of fragments of badly fired ceramics that belong to ‘carrot-shaped’ and LR2 amphorae which had been discarded have been also found among other finds during the excavation.

It is generally acknowledged that the LR 2 amphoras are produced in more than one centre. In the region of Greece, very few workshops have been identified until today: just the one in the Argolid which was dated in the end of 6th and the beginning of the 7th centuries (Rudolph 1979, 303-4; Zimmerman Munn 1985, 342-3; Megaw-Jones 1983, 246-247, pl. 24.2, 3). More recently Hayes suggested that an important production center should be located in the Saronic gulf area, possibly in Athens or Corinth (Hayes 2003, 529).

The biconical support rings, which were used as pedestals and show no evidence of burning, are closely linked to the production and have been retrieved in large numbers among the kiln discards (Fig. 9). These rings are required to hold the amphorae upright after firing and especially when they are full with their contents. Some of them bear names, while on others only a few letters are preserved. On one of them the full name “Νικειανού” is preserved who is probably one of the workshop owners. The stands are of different dimensions and we suppose that this fact has to do with the different kinds of amphorae produced in the kilns like the ‘carrot-shaped’ and the LR 2.

The diversity of the upper parts shapes from the Dilesi workshop, mainly in the mouth and handles, and the similarity with the finds from the shipwreck 7 in Pagasitic Gulf indicate an early date in the 4th century for the above mentioned finds from Dilesi, when the shape of the LR 2 has not yet been standardized.

A large number of small wedges was also recovered from the excavation of the kilns (Fig. 10). They were most probably used during the placement of the amphorae in the kiln so that during the firing process they would not be in contact with one another. Moreover a number of lids, all characterized by pastes similar to those of LR2, are also produced on location. The outlines suggest that they would fit vessels with a small diameter like the mouth of the LR2 amphorae.

An interesting dipinto in black is painted in three lines on the shoulder of one LR 2 amphora, which is probably a notation of capacity (Fig. 8). The dipinto, couched in an abbreviated and indecipherable form in three lines, as all similar inscriptions, has been written after the firing of the amphora. Consequently, the dipinto is not directly linked to the amphora’s production but more to its trading and probably to a secondary re-use.

A few basin fragments as well as dishes and bowls were found during the excavation; that leads to the conclusion that they were all produced in the kilns but not in such quantities as is the case with the above mentioned amphorae.

The first upper line is older or contemporary with the two following lines but it is erased and it is rewritten exactly the same in the following second line.

The amphorae produced in this workshop date, as already mentioned, to the 4th century. The finding of a large number of coins dated in the Hellenistic and Roman period, but not after the 4th century AD, is another piece of corroborative evidence, which consolidates the early dating of the amphorae.

Κερ[ά] μια λ The keramion is a unit of measurement and corresponds to the Roman amphora of 8 choe (calculated at a measure of 26.616 lt.) (Lang 1976, 61; Mayerson 2000). The recording of such a big quantity, since the proposed reading is correct, leads to the hypothesis that the dipinto

195

LRCW4

Charami, A. 1996b. Οδός Οδυσσέως. Archaeologikon Deltion 51 (1996), Β1, 279-282.

Thanks to the intense construction activity in recent years the finds from rescue excavations at Dilesi have gradually brought to light a settlement surviving since the Hellenistic era, and which seems to have flourished in late Roman/early Byzantine period (from the 4th until the first half of the 7th centuries). The activity of this ceramic workshop from the Hellenistic era until at least the 4th century - and possibly to a lesser degree in the following centuries - is closely associated with the existence of the port, which allowed for the easy exportation of these products. Further research of the rich material of the kilns (M. Karambinis is studying it thoroughly) and the publication of the abundant ceramic material from the whole area will allow us to define with greater accuracy the range of the production and the exportation of the products from this significant center in the Euboean gulf, located as it is close to important cities of antiquity such as Chalkis, Eretria and Tanagra.

Charami, A. forthcoming. Ανασκαφή οιποπέδου Καββά. Archaeologikon Deltion. Demesticha, S. 2010. Το φορτίο του ναυαγίου 7 στον Παγασητικό κόλπο. Πρώτη ερμηνευτική προσέγγιση. In Papanikola-Bakirtzi and Kousoulakou (eds.), 131142. Daux, G. 1965. Chronique des fouilles et découvertes archéologiques en Grèce en 1964. Bulletin de Correspondance Hellénique 89, 683-1008. Duchesne, L. 1879. Inscription chrétienne de Tanagre. Bulletin de Correspondance Hellénique 3, 144-146. Gerousi, E. 1997. Κεραμικά παλαιοχριστιανικών χρόνων από την περιοχή του Επισκοπείου της Σάμου. Archaeologikon Deltion 47-48 (1992-1993), 251-267. Hayes, J. W. 1992. Excavations at Saraçhane in Istanbul, vol. 2. The Pottery. Princeton N.J., Princeton University Press.

BIBLIOGRAPHY Abadie-Reynal, C. 1989. Les amphores protobyzantines d‘Argos (IVe-VIe siècles). In V. Deroche and J. M. Spieser (eds.), Recherches sur la céramique byzantine, Bulletin de Correspondance Hellénique, Supplément 18, 47-56.

Hayes, J. W. 2003. Amphores. In Ch. Bakirtzis (ed.), Actes du VIIe Congrès International sur la Céramique Médiévale en Méditerranée, Thessaloniki, 11-16 Octobre 1999, 529-534. Athens.

Abadie-Reynal, C. and Sodini, J.-P. 1992. La céramique paléochrétienne de Thasos (Aliki, Delkos, fouilles anciennes). Études Thasiennes XIII. École Française d'Athènes. Paris, de Boccard.

Hautumm, W. 1981. Studien zu Amphoren der spatrömischen und frühbyzantinischen Zeit. Fulda. Koilakou, Ch. 2006. H Βοιωτία κατά την παλαιοχριστιανική περίοδο (4ος-7ος αι.). Μια πρώτη προσέγγιση. Πρακτικά επιστημονικής συνάντησης για το Αρχαιολογικό Έργο Θεσσαλίας και Στερεάς Ελλάδας (27/2-2/3 2003), 1105-1118. Volos.

Adamsheck, B. 1979. Kenchreai IV, The pottery, Leiden. Aupert, P. 1980. Objets de la vie quotidienne à Argos en 585 ap. J.-C. Bulletin de Correspondance Hellénique, Supplément 6, 395-457.

Lang, M., 1976. Graffiti and Dipinti. The Athenian Agora XXI. Princeton.

Barnea, I. 1966. L’incendie de la cité de Dinogetia au VIe siècle. Dacia 10, 237-259.

Mayerson, Ph. 2000. The Monochoron and Dichoron: Standard Measures for Wine based on the Oxyrhynchition. Zeitschrift für Papyrologie und Epigraphik 131, 169-172.

Boardman, J. 1989. The pottery. In M. Ballance, J. Boardman, S. Corbett, S. Hood, Byzantine Emporio. Excavations in Chios 1952-1955. The British School at Athens, Supplement 20, 106-117. Oxford.

Megaw, A. H. S. and Jones, R. E. 1983. Byzantine and allied pottery: A contribution by chemical analysis to problem of origin and distribution. The Annual of the British School at Athens 78, 235-263.

Böttger, B. 1992. Die Kaiserzeitlichen und Spätantiken Amphoren aus dem Kerameikos, Athenische Mitteilungen 107, 315-381.

Opaiţ, A. 2010. Pontic Wine in the Athenian Market. In Papanikola-Bakirtzi and Kousoulakou (eds.), 108130.

Chamilaki, K. 2010. Ταφικά σύνολα υστερορωμαϊκών χρόνων από νεκροταφείο στο Δήλιον Βοιωτίας. Πρώτες παρατηρήσεις. In D. Papanikola-Bakirtzi and N. Kousoulakou (eds.), Κεραμική της Ύστερης Αρχαιότητας από τον ελλαδικό χώρο (3ος-7ος αι. μ.Χ.), Proceedings of a Scientific Meeting, Thessaloniki 12-16 November 2006, Publications of the Archaeological Institute of Macedonian and Thracian Studies 8. 580-609. Thessaloniki.

Peacock, D. P. S. and Williams, D. F. 1986. Amphorae and the Roman economy. New York. Pétridis, P. 1995. La céramique paléochrétienne de Delphes. Thèse de Nouveau Doctorat. Paris. Pétridis, P. 2003. Delphes paléochrétienne. In Ch. Bakirtzis (ed.), Actes du VIIe Congrès International sur la Céramique Médiévale en Méditerranée, Thessaloniki, 11-16 Octobre 1999, 534-536. Athens.

Charami, A. 1995. Οικόπεδο ΟΤΕ. Archaeologikon Deltion 50 (1995), Β1, 305. Charami, A. 1996a. Οικόπεδο πρώην ιδιοκτ. Μπίμπιζα. Archaeologikon Deltion 51 (1996), Β1, 282-284.

Pieri, D. 2005. Le commerce du vin oriental à l’époque byzantine (Ve-VIIe siècles). Le témoignage des 196

GEROUSI

amphores en Gaule. Bibliothèque archéologique et historique 174, 86-88. Beyrouth, Institut Français du Proche-Orient.

Cooking Wares and Amphorae in the Mediterranean, and Archaeometry. British Archaeology Archaeological Reports, International Series 1340, 563-609. Oxford, BAR Publishing.

Pitteros, Ch. 2000. Ανασκαφή στωικού οικοδομήματος στο Δήλεσι και το Ιερό του Δηλίου Απόλλωνα. Επετηρίς Εταιρείας Βοιωτικών Μελετών Γ.α, 592-623.

Riley, J. A. 1981. The Pottery from the Cisterns 1977.1, 1977.2 and 1977.3. In J. H. Humphrey (ed.), Excavations at Carthage 1977 conducted by the University of Michigan 6, 85-124, Ann Arbor.

Platon, N. 1937. Χριστιανική επιγραφή εκ Τανάγρας. Archaeologiki Ephimeris 1937, 655-667.

Robinson, H. S. 1959. The Athenian Agora V. Pottery of the Roman Period. Chronology. Princeton, N.J., The American School of Classical Studies at Athens.

Portale, E. C., and Romeo, I. 2000. Le anfore locali di Gortina ellenistica e romana. Rei Cretarie Romanae Fautorum Acta 36, 417-426.

Rudolf, W. W. 1979. Excavations at Porto Cheli and vicinity. Preliminary report V: The early Byzantine remains. Hesperia 48, 294-324.

Poulou-Papadimitriou, N. 2001. Βυζαντινή κεραμική από τον ελληνικό νησιωτικό χώρο και την Πελοπόννησο (7ος-9ος αι.): μία πρώτη προσέγγιση. In E. KountouraGalaki (ed.), Οι σκοτεινοί αιώνες του Βυζαντίου, ΕΙΕ/ΙΒΕ, Αθήνα, Μάιος 1999, 231-266. Athens.

Scorpan, C.1976. Origini şi linii evolutive in ceramica romano-bizantină (sec. IV-VII) din spaţiul Mediteranean şi Pontic. Pontica 9, 155-185.

Poulou-Papadimitriou, N. 2013. Kύθηρα. Η βυζαντινή και η πρώιμη ενετική περίοδος. In Y. Sakellarakis (ed.), Κύθηρα. Το Μινωικό ιερό κορυφής στον Άγιο Γεώργιο στο Βουνό, 3: Τα ευρήματα. Βιβλιοθήκη της Εν Αθήναις Αρχαιολογικής Εταιρείας 282, 25-266. Athens.

Scorpan, C. 1977. Contribution à la connaissance de certains types céramiques romano-byzantines (IVeVIIe siècles) dans l’espace Istro-Pontique. Dacia 21, 269-297. Slane, K. W. 2000. East-west trade in fine ware and commodities: The view of Corinth. Rei Cretarie Romanae Fautorum Acta 36, 299-312.

Poulou-Papadimitriou, N. and Nodarou, E. 2007. La céramique protobyzantine de Pseira: la production locale et les importations, étude typologique et pétrographique. In. M. Bonifay and J.-Ch. Tréglia (eds.), LRCW2. Late Roman Coarse Wares,Cooking Wares and Amphorae in the Mediterranean, Archaeology and Archaeometry.British Archaeological Report, Int Series 1662, 755-76 766 Oxford, BAR Publishing. Radulescu, A. 1973. Amfore cu inscripţii de la edificiul roman cu mosaic din Tomis. Pontica 6, 193-207.

Steckner, C. 1989. Les amphores LR 1 et LR 2 en relations avec le pressoir du complexe ecclésiastique des thermes de Samos. In V. Deroche and J.-M. Spieser (eds.), Recherches sur la céramique byzantine, Bulletin de Correspondance Hellénique Supplément 18, 57-71. Vogt, C. 1991/93. Πρωτοβυζαντινή κεραμική από την Αγία Γαλήνη. Kritiki Estia 4, 39-80. Vogt, C. 2000.The Early Byzantine pottery. In P. Themelis (ed.), Πρωτοβυζαντινή Ελεύθερνα, Τομέας Ι, v. 2, 37-199, Rethimno.

Reusch, W. 1970. Kleine, spitzkonische Amphoren. Saalburg Jahrbuch XXVII, 54-62.

Zimmerman Munn M. L. 1985. A Late Roman Kiln site in the Hermionid, Greece. American Journal of Archaeology 89, 342-343.

Reynolds, P. 2005. Levantine amphoras from Cilicia to Gaza: a typology and analysis of regional production trends from the 1st to 7th centuries. In M. Gurt i Esparraguerra, J. Buxeda i Garrigós, and M. A. Cau Ontiveros (eds.), LRCW1. Late Roman Coarse Wares,

197

LRCW4

Fig. 1. Map of the region of Boeotia, Greece.

198

GEROUSI

Fig. 2. The best preserved kiln I with the arched entrance on the west side.

Fig. 3. The best preserved example of the carrot-shaped amphorae, A46 of Dilesi workshop.

199

LRCW4

Fig. 4. Various body parts of carrot-shaped amphorae

Fig. 5. Well-preserved example of early LR 2 amphora without the base

Fig. 6. Well-preserved example of early LR 2 amphora without the neck.

200

GEROUSI

Fig. 7 a-f. Upper parts of LR2 amphorae from the Dilesi workshop.

Fig. 8. Dipinto in black on the shoulder of one of the early LR 2 amphorae

201

LRCW4

Fig. 9. Various parts of biconical support-rings. Some of them bear inscriptions with names.

Fig. 10. Small wedges used probably during the firing process.

202

EMPOLI, A LATE ANTIQUE POTTERY PRODUCTION CENTRE IN THE ARNO VALLEY (FLORENCE, TUSCANY, ITALY) FEDERICO CANTINI1, GIOVANNI BOSCHIAN2, MARZIA GABRIELE3 1

Department of Civilisations and Forms of Knowledge, University of Pisa; [email protected] 2

Department of Biology, University of Pisa; [email protected]

3

Department of Civilisations and Forms of Knowledge, University of Pisa and CEPAM – UMR 7624 CNRS/University of Nice Sophia Antipolis; [email protected]

The main aim of this paper is to characterise the pottery production of Empoli (Florence, Tuscany, Italy) in terms of ceramic classes, types, fabrics (by thin section) and chronology. In Roman times, this site held a strategic location along the road that connected the cities of Pisa and Florence, close to the Arno river and its tributary the Elsa. Pottery – including amphorae, red slip, red painted and fine wares – and waste from pottery production have been found in many rescue excavations, inside and west of the town. The production dates to between the 4th and the 5th/early 6th centuries AD. KEYWORDS: 4TH-5TH/EARLY 6TH CENTURIES, AMPHORAE, RED SLIP, RED PAINTED WARE PRODUCTION SITE, TABLE WARE, THIN SECTIONS. The results will be published in the Proceedings of the International Conference, L’anfora “di Empoli”: produzioni e diffusione in età romana (Empoli, 14-16 October 2010).

Introduction (F. C.) In this paper we present the results of the study of the pottery found in the old town of Empoli, in the Middle Arno valley, near Florence (Fig. 1, 1).

Outside the region, this type of amphora has been found in Lazio (Ostia, Rome), Liguria (Ventimiglia), Sardinia

Excavations in the centre of the town demonstrate that the site has a long history, beginning from the Hellenistic period. The settlement was at a strategic point of the region, along the road connecting the cities of Pisa and Florence (the via Quinctia) and close to the Arno river and its tributary the Elsa (AA.VV. 1984; Ferretti et al. 1995; Maiuri 2006).

(Porto Torres), Campania (Miseno), Corsica (Mariana), France (Marseilles), Spain (Tarragona) and Africa (Carthage) (Cantini 2011, 163). The analysis of the finds from the town of Empoli allowed us to define the chronology and typology of the Empoli amphora, and to identify other classes of pottery produced here: fine, red slip and red painted wares.

In Roman times it was probably a vicus or a mansio, but the data currently available do not allow for definite conclusions.

This paper presents only the results of the red slip ware thin sections, since at the visual and stereomicroscopic scale the red painted and fine wares have an identical texture and mineral composition. We are still awaiting the results of the analysis of the amphora thin sections, to be published in the Proceedings of the International Conference at Empoli.

The town is mainly known in the archaeological literature for amphora production waste, found in Empoli and Avane (west of the city). This type of amphora was first identified in the Ostia excavations, and named “Ostia IV, 260, no. 35” by Prof. Daniele Manacorda (1984). In 1986 Cambi published the production waste from Empoli. Thin section analysis revealed a fabric with microfossils (radiolarians) (Cambi 1989), and the analysis of organic residues confirmed that it was a wine amphora (Pecci et al. 2010).

This work is part of a project funded by the University of Pisa and by the Italian Ministry of Education, Universities and Research, aimed at studying the transformation of the Arno valley between late Antiquity and the Middle Ages (PRIN 2008).

Other production centres for very similar amphorae, with a composition including gabbro, have been identified along the Tyrrhenian coast, near Vada Volaterrana (Livorno) (Pasquinucci et al. 1998; 1999).

Archaeological contexts (F.C.) The finds come from eight excavations carried out by the Soprintendenza per i Beni Archeologici della Toscana in collaboration with the Middle Valdarno Archaeological Association.

The presence of numerous kilns and different variants of this type of amphora led the Soprintendenza per i Beni Archeologici della Toscana to carry out new thin section analyses of many samples from different Tuscan sites. 203

LRCW4

Five shallow bowls have a large diameter and an everted, indistinct or enlarged rim (Fig. 2, IR.7): for IR.7.1: Cantini 2009, 61, no. 3, second half of the 4th-early 7th centuries; for IR.7.4: Cantini 2009, 60, no. 18, 4th-6th centuries. Type IR.7.2 imitates the African red slip ware Hayes 83 (second half of the 5th century).

Seven of these are inside the urban area: Del Vivo (SDV), Piazza Farinata (SPF) (Rustici 1998), Piazza San Giovanni (RSG), Banca Nazionale del Lavoro (BNL), Montefiori (RM) (Filippi 2007, 17), Pratesi (PR) (Alfani 1996/7; Maiuri 2006, 28-30; Papanti 2006), Municipio (RCE) (Ferretti et al. 1995). Only one (Avane-SA) is situated west of Empoli, near the Arno river (Alderighi et al. 2010) (Fig. 1, 2). Unfortunately these were not stratigraphical but rescue excavations. Conversely, the only stratigraphical investigation in Empoli, i.e. Piazza della Propositura (Rastrelli 2002), is being studied by the Soprintendenza: we were not granted permission to analyse it.

By contrast, we have only one flanged bowl decorated with an incised wave, probably a mortar (Fig. 1, IR.6.1). In Tuscany, red slip ware dates from the second half of the 3rd century, and maybe as early as the late 1st-2nd centuries, to the 7th century (Cantini 2009; 2010; 2011, Fumo 2010). Parallels for the Empoli pottery belong to a chronological period between the 4th and 5th/early 6th centuries, with few types dating to before the 4th or after the first half of the 6th centuries.

The pottery productions (F. C.)

Thanks to kiln waste, the red painted bowls, dishes and cups with thickened and inverted rims, and the small amphorae can also be attributed to the Empoli workshops (Fig. 2, DR) (for DR.1.1: Cantini 2007, 279, no. 14.4.1, second half of the 4th century; 2009, 69, no. 19).

Empoli produced tableware, storage ware and amphorae, as shown by the production waste (Fig. 1, 3-7). The fabrics of the vessels have a 5YR reddish- yellow 6/6 Munsell colour. As tableware, the red slip ware includes jugs, small amphorae, bowls, dishes, cups and mortars. The jugs can be divided into three groups:

The decorative motifs used are broken lines, triangles, arches, circles and points, radial tabs, waves, festoons with dots and dashes, comma motifs, crosses formed by tangent circles intersected by a line, square metopes and fishes.

-large wide-mouthed jugs, with a swollen and everted rim (Fig. 1, IR.1). For IR.1.2.1: Cantini 2009, 64, no. 34, second half of the 3rd-second half of the 4th centuries;

This production is common in Northern Tuscany between the first half of the 4th and the 5th/early 6th centuries (Cantini 2009; Filippi 2007).

-pear-shaped jugs with an everted rim, short neck and a handle just below the rim (Fig. 1, IR.2). For IR.2.1.1: Cantini 2009, 63, no. 8, 64, no. 1, 4th-7th centuries. For IR.2.1.2: Cantini 2009, 63, no. 17, 64, no. 11, and Cantini 2007, 277, no. 11.4.1, 4th-6th centuries;

Fine ware was also produced at Empoli (Fig. 3, AD): large conical bowls with everted and thickened rims, dating from the 3rd-4th centuries to the 6th century (Menchelli et al. 2007, 326, no. 23; Cantini 2007, 264, no. 1.1.4), jugs similar to the red slip jugs, and flasks.

-small jugs, often with ribbed walls (Fig. 1, IR.3). For IR.3.1: Alderighi, Cantini (eds.) 2011, 67, Fig. 26, no. 8, 4th-5th centuries. For IR.3.3: Cantini 2009, 64, no. 14, late 5th-mid 6th centuries.

Finally, several fragments of Empoli amphorae were found. They are often melted into a vitrified condition, and date to a period between the 4th and the early 6th centuries (Cantini, in press). Though the results of the thin section analysis are not yet available, we observed other samples under a stereomicroscope and identified a fabric with some microfossils, together with iron oxides, quartz, mica and small unidentified grey and white particles.

The open shapes include: -cups with an everted rim (Fig. 1, IR.4). For IR.4.1.2: Alderighi, Cantini (eds.) 2011, 67, Fig. 26, 5, 4th-5th centuries; Cantini 2009, 61, no. 26, 4th century; 2007, 275, no. 11.1.9, 6th century. For IR.4.2: Cantini 2007, 275, no. 11.1.1, 6th century. For IR.4.3: Cantini 2009, 61, no. 20, 5th century; Ciampoltrini et al. 2010, 324, Fig. 3, no. 3, AD 370-380;

We identified five different amphora types (Fig. 3, ANFE.1-5) on the basis of the shape of the rim:

-deep bowls with a thickened, inverted or vertical rim (Fig. 1, IR.5). For IR.5.2.1: Cantini 2007, 275, no. 11.1.2, 4th-5th centuries. For IR.5.3: Cantini 2009, 61, no. 13, late 4th-early 5th centuries. For IR.5.4: Cantini 2009, 61, no. 15, second half of the 3rd-second half of the 4th centuries. One type of bowl has handles (Fig. 1, IR.5.5: Cantini 2009, 61, no. 6, first half of the 7th century).

-thickened and squared (ANFE.1.1-2). For ANFE.1.1 (4th-first half of the 7th centuries): Bien 2005, 291, Fig. 2, no. 14; Cantini 2007, 280, Table XVII, no. 16.15; 2010, 358, Fig. 3, nos 18 and 21; Ciampoltrini, Rendini 1989, 518-519, Fig. 4, no. 5; Ciampoltrini et al. 2010, 325, Fig. 4, no. 4; Faggella 1990, 401, nos 22, 24, 25, 28; Menchelli et al. 2007, 325, nos 2-3; Vaccaro 2010, 141143, plate VIII, no. 1, plate LXXVIII, no. 1, plate LXXIX, no. 1, plate LXXXIII, no. 1. Tape-shaped (ANFE.1.3): Faggella 1990, 402, no. 38, late 4th-early 5th centuries;

We also identified one type of dish (Fig. 2, IR.8.1) and two large bowls (Fig. 2, IR.9.1-2). For type IR.9.2: Cantini 2009, 62, no. 8, mid 6th-first half of the 7th centuries; 2007, 276, no. 11.2.7, second half of the 7th century.

204

CANTINI- BOSCHIAN- GABRIELE

were located along a street, probably the Florentia-Pisae, which ran near the Arno river.

-internally thickened (ANFE.2): Vaccaro 2010, 82-83, plate XXIV, no. 7, 141-143, plate LXXVIII, nos 2-3, plate LXXXI, no. 2, plate LXXXII, no. 2, 4th-mid 5th centuries;

We can observe various associations of production waste in different locations: red painted ware in BNL; all the classes in RM; amphorae, fine and red painted wares in PR; amphorae only in RCE and SDV; amphorae and red slip ware in RSG; only fine ware in SPF.

-sharp and everted (ANF.3.1-2). For ANFE.3.1 (4th-first half of the 7th centuries): Alderighi, Cantini (eds.) 2011, 51, Fig. 6, 1; Bien 2005, 291, Fig. 2, no. 13; Faggella 1990, 402, no. 30; Milanese et al. 1997, 60; Pasquinucci et al. 1998, 363, no. 20; Menchelli et al. 2007, 325, no. 4; Macias et al. 2005, 131, nos 1-2 and 4; Vaccaro 2010, 8283, 141-143, plate XXIV, no. 6, plate LXXIX, nos 2, 3, plate LXXX, no. 3. For ANFE.3.2 (4th-5th/6th centuries): Faggella 1990, 402, nos 36 and 37; Pasquinucci et al. 1998, 363, no. 32; Vaccaro 2010, 141-143: plate LXXXIV, no. 1;

The simultaneous production of different classes of pottery at the same workshop is also confirmed by the discovery of two amphora fragments with traces of red slip. This evidence confirms a specialisation in pottery production at Empoli, dating back at least from the middle of the 2nd to the 3rd centuries, when the town produced “ceramica a pareti sottili” (Rustici 1998, 25).

-internally swollen with a flattened top (ANF.4, late 4thearly 5th centuries): Macias et al. 2005, 131, no. 3; -everted and expanded (ANF.5) (mid 3rd-5th/6th centuries). For ANFE.5.1: Faggella 1990, 401, no. 26; Genovesi 2010, 343, Fig. 4, no. 6; Pasquinucci et al. 1998, 363, no. 23; Vaccaro 2010, 141-143, plate LXXXIII, no. 3. For ANFE 5.2: Cantini 2005, 202, Table 56, no. 11.14; Cantini 2007, 280, Table XVII, no. 16.14; Cantini 2010, 358, Fig. 3, nos 19 and 20 (similis); Cyegelman, Millemeci 2007, 374, Fig. 13, A; Gambaro 1997, 70; Menchelli et al. 2007, Fig. 2, n. 4; Pasquinucci et al. 1998, 363, no. 18; Vaccaro 2010, 141-143, plate LXXXI, no. 2, plate LXXXV, nos 1 and 3; Villedieu 1984, 196 and 330, no. 288.

Red slip ware fabric and mineralogical analysis (G. B., M. G.) A mineralogical and petrographic study was carried out on the red slip ware, to define its composition, characterise the Empoli production and check it against the geological characteristics of the nearby outcrops of parent material. The whole pottery assemblage (27 samples) was observed at microscopic scale on freshly broken surfaces under a Leica MZ125 stereomicroscope; this assemblage was divided into fabric groups based on mineral composition (as ascertainable by this method), and texture. Four samples (PR35, PR43, RM45, RSG8) were chosen for thin section study, and cut into standard petrographic (27x42mm) thin sections.

The base is a small empty spike (ANF.6, late 4th-first half of the 6th centuries): Cantini 2007, 281, Table XVIII, nos 16.21-23; Faggella 1990, 403, nos 40-43.

These thin sections were observed under a Leica DM/LP standard polarising microscope, and described following the standard formalised by Stoops (2003) for soils, as no generally accepted standard for pottery description is available at present.

Chronology (F. C.) In the absence of stratigraphical reference points, we tried to define the period of activity of the Empoli workshops by plotting a diagram of the number of ceramic types against their date, as defined in the literature. Time intervals are plotted on the x-axis, while the number of types whose chronology lies within each interval is indicated on the y-axis. We obtained a curve representing the production trend, with a peak between AD 300 and 525 (Fig. 3, no. 1) that indicates the period of production at Empoli; the low part of the curve on the more recent side may indicate the continuing use of old vessels. Conversely, the roughly specular part on the less recent side of the curve is populated by types belonging to contexts dated to broad (i.e. poorly dated) time spans.

The groups identified at the visual and microscopic scale differ slightly in terms of grain-size and mineral composition, and can be grouped within one single class, with moderate variability. The clay fabric is usually rather fine, seldom somewhat coarser, rich in Fe-oxides, and includes fine silt-size quartz angular grains and mica (mostly muscovite) flakes, the latter being sand-sized in some cases (Fig. 4; a, b). The temper is very fine to medium sand-sized, and includes angular to subangular quartz, a few feldspar grains of the same size and shape, sandstone fragments, flint and sparse calcareous bioclasts.

We obtained similar results using the values for each class of pottery (Fig. 3, no. 2).

Some diagnostic elements, like sandstone fragments (Fig. 4; c, d), microfossils, minute echinid and coral parts (Fig. 4; e, f), suggest that the parent material may derive from fine detritic formations outcropping to the North and South of the town of Empoli, i.e. yellow sands, sandstones, clay, sandy clay and conglomerates (Fig. 5, A), or clays and sandy clays with occasional coarser and finer horizons (Fig. 5, B).

Organization of production (F. C.) Excavations in the centre of Empoli and its suburbs have indicated a significant presence of tubuli for pottery kilns and of production waste, indicating that the workshops 205

LRCW4

Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry, Comparison between western and eastern Mediterranean, British Archaeological Reports, International Series 2185, 353-362. Oxford, BAR Publishing.

Compared to the assemblage of Aiano-Torraccia di Chiusi (Fumo 2010; Cantisani, Cavalieri, Lofrumento, Pecchioni, Ricci 2011), this composition includes microfossils, while the plagioclase grains common at Aiano-Torraccia do not occur. Bibliography

Cantini, F. 2011. Dall’economia complessa al complesso di economie. Tuscia (V-X secolo). Post Classical Archaeologies 1, 159-194.

AA.VV. 1984. Città di Empoli. Mostra archeologica del territorio di Empoli. Catalogo della mostra. Empoli, Comune di Empoli.

Cantini, F., in press. Produzioni e indicatori di produzione ceramica a Empoli tra IV e VI secolo. In L. Alderighi, L. Terreni (eds.), L’anfora “di Empoli”: produzioni e diffusione in età romana (Empoli, 14-16 October 2010).

Alderighi, L., Filippi, M., Maiuri, W. and Terreni, L. G. 2010. Empoli: scavo in località Avane, lato sud del tratto finale di via Magolo con angolo di via del Pozzo. Notiziario della Soprintendenza per i Beni Archeologici della Toscana 5, 2009, 23-33.

Cantisani, E., Cavalieri, M., Lofrumento, C., Pecchioni, E. and Ricci M. 2011. Ceramic findings from the archaeological site at Aiano-Torraccia di Chiusi (Siena. Italy): a multi-analytical approach. Archaeological and Anthropological Sciences, 10.1007/s12520-011-0000-1, 1-18.

Alderighi, L., Cantini, F. (eds.), 2011. Capraia e Limite. La villa dei Vetti. Nuove e vecchie indagini archeologiche, Notiziario della Soprintendenza per i Beni Archeologici della Toscana 6, 2010, 47-81. Alfani, I. 1996/7. Anfore romane ad Empoli: il materiale del recupero Pratesi. Unpublished thesis, University of Florence.

Ciampoltrini, G., Andreotti, A., Notini, P., Rendini, P. and Spataro, C. 2010. Traffici e consumi ceramici nella valle del Serchio in età teodosiana. In LRCW3, 319-328.

Bien, S. 2005. Des niveaux du VIIe siècle sous le music-hall de l’Alcazar à Marseille. In J. Ma. Gurt i Esparraguera, J. Buxeda i Garrigós, M. A. Cau Ontiveros (eds.), LRCW1. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry, British Archaeological Reports, Int. Series 1340, 285-298. Oxford, BAR Publishing. Cambi, F. 1989. L'anfora di Empoli. In Amphores

Ciampoltrini, G. and Rendini, P. 1989. Un insediamento tardo antico nella Valle dell’Osa (Orbetello, Gr), Indagini di superficie. Archeologia Medievale 15, 513-522. Cyegelman, M. and Millemeci, G. 2007. Vetulonia, via Garibaldi (Castiglione della Pescaia, Gr): scavi 20032006. In L. Botarelli, M. Coccoluto, M. C. Mileti (eds.), Materiali per Populonia 6, 345-386. Pisa, ETS.

romaines et histoire économique, Actes du colloque de Sienne (22-24 mai 1986), Collection de l'École Française de Rome 114, 564-567. Rome.

Faggella, F. 1990. Contenitori da trasporto. In G. De Marinis (ed.), Archeologia urbana a Fiesole. Lo scavo di Via Marini-Via Portigiani, 249-269. Firenze, Giunti.

Cantini, F. 2005. Archeologia urbana a Siena. L’area dell’ospedale di Santa Maria della Scala prima dell’ospedale, Altomedioevo. Firenze, All’Insegna del Giglio.

Ferretti, E., Macii, R. and Terreni, L. 1995. Ritrovamenti archeologici nel territorio di Empoli. Fucecchio, Edizioni dell’Erba.

Cantini, F. 2007. Ceramica e pietra ollare. In F. Cantini, C. Cianferoni, R. Francovich and E. Scampoli (eds.), Firenze prima degli Uffizi. Lo scavo di via de’ Castellani. Contributi per un’archeologia urbana fra tardo antico ed età moderna, 183-286. Firenze, All’Insegna del Giglio.

Filippi, M. 2007. La ceramica dipinta tarda dal centro storico di Empoli: i rinvenimenti "Montefiori" e Piazza della Propositura. Milliarium VII, MMVII, 1627.

Cantini, F. 2009. Produzione, circolazione e consumo del vasellame decorato con ingobbio rosso in Toscana tra I-II e XIII secolo. In E. De Minicis (ed.), Le ceramiche di Roma e del Lazio in età medievale e moderna, Atti del VI Convegno di Studi (Segni, 6-7 maggio 2004), La ceramica dipinta in rosso. I contesti laziali a confronto con altre realtà italiane, 59-79. Roma, Edizioni Kappa.

Fumo, A. 2010. Le ceramiche dipinte di rosso della villa di Aiano-Torraccia di Chiusi (San Gimignano, Siena): uno studio archeologico e archeometrico. www.fastionline.org/docs/FOLDER-it-2010-178.pdf. Gambaro, L. 1997. Economie ed insediamenti nella montagna pesciatina in età romana. In Atti del Convegno sul l’Archeologia in Valdinievole, 51-79. Buggiano.

Cantini, F. 2010. Circolazione e consumo di vasellame ceramico e anfore nel Medio Valdarno tra IV e VII secolo: nuovi dati da San Genesio (San Miniato, Pisa) e Firenze. In S. Menchelli, S. Santoro, M. Pasquinucci, G. Guiducci (eds.), LRCW3. Late Roman

Genovesi, S. 2010. Materiali tardoantichi dal Portus Pisanus (S. Stefano ai Lup, Livornoi). In LRCW3, 337-344. 206

CANTINI- BOSCHIAN- GABRIELE

Macias Solé, J. M. and Remolà Vallverdú J-A. 2005. La cultura material de Tarraco-Tarracona (Hispania Tarraconensis-regnum Visigotorum): cerámica común y ánforas. In LRCW1, 125-135.

Pasquinucci, M., Del Rio, A. and Menchelli, S. 1999, Contenitori da trasporto e da magazzino nella fascia costiera alto-tirrenica dal tardoantico al medioevo. In Contenitori da trasporto e da magazzino tra Tardo Antico e Basso Medioevo, Atti del XXX Convegno Internazionale della Ceramica, 59-65. Firenze, All’Insegna del Giglio.

Maiuri, W. 2006. La città, il territorio, il porto: Empoli in età romana. Milliarium VI, MMVI, 28-39. Manacorda, D. 1984. L’identificazione dell’anfora di Empoli. In Città di Empoli. Mostra archeologica del territorio comunale. Catalogo della mostra, 23-28. Empoli, Comune di Empoli.

Pecci, A., Salvini L. and Cantini, F. 2010. Residue analysis of some late roman amphora coming from the excavation of the historical center of Florence. In LRCW3, 363-367.

Menchelli, S., Capelli, C., Pasquinucci, M. and Picchi, G. 2007. Corsica tardo-antica: anfore italiche e ceramica comune da Mariana. In M. Bonifay and J.-C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry, British Archaeological Reports, International Series 1662, 313-328. Oxford, BAR Publishing. Milanese, M., Patera, A. and Pieri E. (eds.), 1997.

Rastrelli, A. 2002. Lo scavo nella Piazza della Propositura di Empoli. Milliarium 1, IV, 2-7. Rustici, S. 1998. Una domus romana ad Empoli. Milliarium I, I, 16-25. Stoops, G. 2003. Guidelines for Analysis and Description of Soil and Regolith Thin Sections. Madison, Soil Science Society of America. Vaccaro, E. 2010. Sites and Pots: Settlement and Economy in Southern Tuscany (AD 300-900). British Archaeological Reports, International Series 2191. Oxford, BAR Publishing.

Larciano. Museo e territorio. Roma, «L’Erma» di Bretschneider. Papanti, A. 2006. Uno studio sulla monetazione dello scavo Pratesi a Empoli. Milliarium VI, MMVI, 40-43.

Villedieu, F. 1984. Turris Libisonis. Fouille d’un site romain tardif à Porto Torres, Sardaigne. British Archaeological Reports, International Series 224. Oxford, BAR Publishing.

Pasquinucci, P., Del Rio, A. and Menchelli S. 1998. Produzioni di vino nell’Etruria settentrionale costiera. In El Vi a l’antiguitat. Economía, produccíó i commerçal Medíterraní occídental, II Colloqui Internacional d’arquelogia romana, 357-363. Badalona.

207

LRCW4

Fig. 1. 1, position of Empoli in Italy; 2, the excavations (1-BNL, 2-RM, 3-PR, 4-RCE, 5-RSG, 6-SPF, 7-SDV; 8-Piazza della Propositura; 9-SA); 3-7 production waste; IR1.1-IR.6, red slip ware types (4th-5/early 6th century).

208

CANTINI- BOSCHIAN- GABRIELE

Fig. 2. IR.7.1-IR.9.2, red slip ware types (4th-5/early 6th century); DR.1.1-DR.4.2, red painted ware types (4th-5th/ early 6th century).

209

LRCW4

Fig. 3. AD.1.1-AD.4.1, fine ware types (4th-5/early 6th century); amphorae types (4th-5/early 6th century); graph 1; graph 2. 210

CANTINI- BOSCHIAN- GABRIELE

Fig. 4. Thin sections: a, c, e (PPL); b, d, f (XPL); a, b, c, d-sample PR43; e, f- sample RSG8.

211

LRCW4

Fig. 5. Source formations (Carta geologica d’Italia 1:100000).

212

ANFORE DI TIPO SICILIANO DAL TERRITORIO DI AGRIGENTO MARIA SERENA RIZZO1, LUCA ZAMBITO2, FRANCESCO GIANNICI3, RENATO GIARRUSSO4-, ANGELO MULONE4 1

Parco Archeologico Valle dei Templi Agrigento; [email protected] 2

3

Dottore di Ricerca Università di Messina; [email protected]

Dipartimento di Fisica e Chimica, Università di Palermo; [email protected] 4

Geolab s.r.l. Palermo; [email protected]; [email protected]

During recent archaeological excavations and surveys in rural settlements of late Roman date in the hinterland of Agrigento many amphorae, probably made in Sicily, have been found: they are small, with ovoidal body and flat bottom. Amphorae with similar characteristics were found in different areas of Sicily, particularly along the northern coast and in the eastern part of the island. In the hinterland of Agrigento were identified three sites where they were produced. In this paper we present some samples and the archaeometric analysis of amphorae found in Cignana, Canalicchio near Calamonaci and Verdura near Sciacca, in the land of Agrigento. KEYWORDS: AGRIGENTO, SICILIAN AMPHORAE, ARCHAEOMETRIC ANALYSIS Ci sembra che l’espressione generica “anfore di tipo siciliano” possa servire da una parte a dare conto della molteplicità dei centri produttori e della loro diffusione

Con l’espressione “anfore di tipo siciliano” ci riferiamo ad un tipo di contenitore di piccole dimensioni, con corpo ovoidale allungato segnato da cordonature, base convessa talvolta con una sorta di rozzo piede ad anello, caratteristiche piccole anse a maniglia impostate sulla spalla e sotto l’orlo, spesso con evidenti segni di ditate in corrispondenza dell’attaccatura, collo più o meno allungato e semplice orlo, del quale sono state riconosciute diverse varianti; ci riferiamo, insomma, al tipo che viene spesso indicato come Termini Imerese 151/354, dopo che esso fu identificato come probabile produzione locale tra i materiali provenienti dagli scavi urbani condotti in questo centro della Sicilia settentrionale (Rizzo 1993, 223-225). Ricerche successive hanno dimostrato la diffusione di anfore simili anche nella Sicilia nord-orientale, a Caronia Marina (Bonanno e Sudano 2007, 356; Bonanno 2009, 37-40) e a Terme Vigliatore-S. Biagio (Borrello and Lionetti 2008, 77) e individuato alcuni siti di produzione: producono infatti anche anforette di questo tipo le fornaci di Capo d’Orlando (Spigo et al. 2006) e di Caronia, contrada Chiappe (Wilson 1990, 263; Bonanno e Sudano 2007), mentre scarti di fornace sono stati rinvenuti a Furnari, in località Tonnarella (Bonanno e Sudano 2007, 356; 2009, 39). Ci sembra probabile, inoltre, che siano attribuibili a questo gruppo almeno alcune delle anfore prodotte dalle fornaci di Santa Venera al Pozzo (Amari 2006; in questo volume; Malfitana 2008, 135-137). Nella Sicilia meridionale un sito di produzione, individuato grazie alla presenza di scarti di fornace, è noto a Campanaio, presso Montallegro (Wilson 2000), circa 20km ad ovest di Agrigento.

sul territorio insulare, dall’altro a raggruppare nell’ambito di un tipo ben definito che, pur con varianti, mantiene comunque una propria identità dal punto di vista morfologico, produzioni che sono state spesso classificate in modi diversi e che rientrano invece, secondo noi, nell’ambito di un gruppo omogeneo e dotato di una propria identità. Ciò, tuttavia, non esclude che rimangano problematiche e da definire le relazioni di questo gruppo con altre produzioni siciliane, ed in primo luogo con le Keay LII (per una probabile produzione di anfore di questo tipo a Caronia, Bonanno 2009, 36-37), con le quali i contenitori di “tipo siciliano” hanno indubbiamente somiglianze morfologiche, ma che sono chiaramente distinguibili per diversi particolari della forma e per gli impasti e la cui produzione è localizzata nell’area dello stretto (Pacetti 1998, 190-192). Ugualmente, va distinto dai nostri contenitori il tipo cd. Crypta Balbi 2 (Saguì 1998, 321; Panella et al. 2010), la cui produzione, localizzata anch’essa nella Sicilia nord-orientale, sembra copra un arco cronologico compreso tra la seconda metà del V secolo e tutto il VII. Potrebbero invece essere pertinenti al gruppo di anforette di cui ci occupiamo almeno una parte dei contenitori Ostia I, 455-456; Ostia IV, 166, nel cui ambito vennero d’altronde classificate le anfore di tipo siciliano rinvenute nell’area della necropoli paleocristiana di Agrigento (Lima 1995, 281-282). Questi contenitori rappresentano con ogni probabilità i successori tardo antichi delle anfore note come MidRoman 1/M254 (Riley 1979, 177-180), prodotte, almeno in parte, in Sicilia (Bonifay 2004, 148; Capelli and Bonifay 2007, 554), in atéliers diffusi, stando a quanto 213

LRCW4

in superficie, si distribuiscono su una superficie superiore ai tre ettari, dato che lascia supporre che alla villa tardoantica sarebbe seguito un villaggio, vissuto, fino al VII secolo. Fra le anfore raccolte si segnalano: Keay II A, Keay 62= Bonifay 46 Fig. 74.7, Keay 35 A e Keay 35 B/85 = Bonifay tipo 41.6, Keay 35 = Bonifay 44.1, Keay 61 D = Bonifay Fig. 75, Keay 61 C = Bonifay 48.1, Sidi Jdidi 1 = Bonifay 54.4, Spatheion tipo 3 A di Bonifay. Si raccolgono alcuni frammenti attribuiti ad anforette di tipo siciliano, una delle quali, con evidenti inclusi di tipo vulcanico, appartiene però ad una produzione non agrigentina (Tav. I, 2).

suggeriscono le prime preliminari indicazioni delle indagini archeometriche, in aree diverse dell’isola (analisi archeometriche, che attestano l’origine siciliana di campioni provenienti dal territorio di Enna in Bonanno et al. 2010, 265). Per quanto riguarda il territorio agrigentino, sono state classificate come Mid-Roman 1 (Capelli and Bonifay 2007, 554) alcune delle anfore prodotte dalle fornaci di Campanaio (Wilson 2000, Fig. 20); un esame diretto dei materiali ha rilevato che tra gli scarti sono certamente presenti anche anfore di tipo “siciliano”; ciò potrebbe provare la provenienza da una stessa officina dei due tipi di anfora. Frammenti certamente pertinenti a questo tipo, comunque, e con argille analoghe a quelle della anforette di tipo siciliano, si rinvengono in superficie sia a Campanaio, sia a Giammaritaro. Analisi archeometriche inedite, inoltre, condotte su campioni rinvenuti nell’entroterra agrigentino, sembrano attestare la compatibilità di almeno un fabric di Mid-Roman 1 con le argille locali.

3 C.da Cignana (Palma di Montechiaro, Ag) Il sito di c.da Cignana si trova in una conca formata dai rilievi di Monte Narbone e della Piana di Cignana che chiudono a Nord, est ed ovest il vallone di Cignana (Burgio e Belvedere 2010). La natura alluvionale dei suoli e le numerose fonti di acqua hanno da sempre favorito l’insediamento anche in relazione alla prossimità dello sbocco a mare e della via di comunicazione paralitoranea. Ad una villa mosaicata con un impianto termale si sovrappone, a partire dalla metà del V secolo d.C., un villaggio in cui i vari lotti abitativi si distribuiscono in maniera disorganica lungo il pendio (per le anfore Rizzo and Zambito 2010, 293-295). Le anforette di tipo siciliano (Tav. I, 6 e II, 9) rappresentano circa il 14% delle attestazioni anforiche complessive del sito.

M.S.R. Le “anfore di tipo siciliano” nel territorio di Agrigento Scavi e ricerche di superficie condotti in anni recenti in diversi siti dell’agrigentino hanno messo in evidenza l’ampia diffusione di questo tipo di anforette in questa parte dell’isola (Lima 1995, 281-282; Parello et al. 2010; Rizzo and Zambito 2010, 294-295). Gli esemplari presi in esame (Tav. I, 1) hanno, nei casi in cui essa può essere determinata, un’altezza media di 40/43cm ca., diametro dell’orlo tra gli 8 e i 9cm, diam. massimo compreso tra i 23 e i 26cm e capacità intorno agli 8 litri. Un certo numero di varianti si riscontra nella forma dell’orlo, che può essere ripiegato a becco, estroflesso a sezione triangolare, indistinto con leggero inspessimento all’esterno, indistinto e sottolineato da una scanalatura alla base. Gli impasti, ad un esame macroscopico, sono piuttosto diversificati da sito a sito.

4 C.da Margio Canneddaro (Naro, Ag) Il sito, individuato con la ricognizione, probabilmente un ampio villaggio, si trova in prossimità del corso del Burraitotto alla confluenza con il Naro e, soprattutto, in prossimità sia della costa agrigentina meridionale sia di un varco fluviale che mette in collegamento, ancora oggi, l’entroterra con il mare. Nell’area, oggetto di una prospezione, si rinvengono, oltre a numerose anfore africane (Keay XXVI, LV-LVII, LXII, LXI), ed orientali (LR 1 e 2), frammenti fittili pertinenti alle forme in TSA D Hayes 76, 99, 104, 105). Si raccolgono inoltre alcuni esemplari di anfore siciliane.

1 Saraceno (Favara, Ag) Il villaggio tardoantico di C.da Saraceno (CastellanaMcConnel 1990) amplia un impianto precedente riferibile ad una villa della prima età imperiale. Le ultime indagini di scavo hanno consentito di isolare due fasi di vita tarde, la prima cessa attorno alla metà del V, mentre la seconda attorno alla metà del VII dopo una ripresa nel VI d.C. Sono presenti anfore cilindriche di grandi dimensioni e spatheia africani, anfore di importazione orientale. Le anforette di tipo siciliano sono attestate in entrambe le fasi cronologiche (Tav. I, 3-4).

5 C.da Cangiana (Naro, Ag) Il vicus tardo antico e bizantino di c.da Cangiana si estende su oltre 6 ettari in una vallata fluviale e su un precedente impianto ellenistico e romano. Il sito, noto da ricognizione, può forse essere messo in relazione con la Massa Cinciana di cui parla Gregorio Magno in una sua lettera. Si segnalano, oltre ad anforette di tipo siciliano (Tav. II, 10), spatheia type 1-3, dolia, anfore LR 1 e 2, Keay XXV, XXXV, LXII, LXII R, LXI, XXXIV, 3B “similis”, Hayes 91 D, 93, 76, 67, 88, 99, 105.

2 C.da Stefano (Favara, Ag) Il sito di contrada Stefano, interessato da saggi di scavo negli anni ’80, (Castellana 1984-5) si estende a breve distanza dal sito di Saraceno, cui è collegato da una direttrice di fondovalle. I frammenti tardo antichi, raccolti 214

RIZZO-ZAMBITO-GIANNICI-GIARRUSSO-MULONE

6. S. Leone (Ag)

10 Calamonaci

S. Leone, a meno di tre chilometri a sud di Agrigento è il porto, connesso con la foce fluviale, dell’antico centro di Agrigentum e risente delle dinamiche insediative della città. Agli inizi del ‘900 è stato indagato un vasto complesso forse riferibile a magazzini. Saggi recenti hanno portato al rinvenimento di numerosi esemplari di anforette siciliane.

In contrada Canalicchio di Calamonaci si estende un piccolo agglomerato rurale (Parello and Amico in questo volume), del quale sono state recentemente scavate due distinte unità edilizie; il quadro generale delle attestazioni anforiche è ancora in corso di studio, ma è già evidente la netta prevalenza delle importazioni africane. Le anfore di tipo siciliano (Tav. I, 5) sono in piccola quantità e risultano, dalle indagini petrografiche, di produzione locale.

7. C.da Menta (Racalmuto, Ag) In c.da Menta (Racalmuto) sorge un grosso agglomerato conosciuto soltanto da indagini di superficie. Si osservano piatti Hayes 76, 104 e coppe Hayes 99. Anfore Keay XXV, XXVI a e b, Keay XXXV. Si è rinvenuta un’anforetta (Tav. II, 8) con impasto arancio e grossi inclusi calcarei. Superficie esterna con evidenti segni del tornio.

11 Vito Soldano (Canicattì, Ag) In contrada Vito Soldano, si estende un ampio insediamento, caratterizzato da una lunga continuità di vita (La Lomia 1961; Bonacasa Carra 1987). Occupato già nella prima età imperiale da un abitato, è interessato, tra la fine del III e gli inizi del IV secolo, dalla costruzione di un edificio termale, che rappresenta oggi l’emergenza monumentale più evidente. La costruzione della terma si inserisce forse in un programma di monumentalizzazione dell’abitato, in connessione con un rilancio del suo ruolo, anche nell’ambito della ristrutturazione del cursus publicus in età costantiniana. In letteratura si tende infatti a riconoscere nel sito di Vito Soldano la mansio di Cosconiana, una delle mansiones nunc institutae menzionate dall’Itinerarium Antonini lungo la via Agrigento-Catania.

8. C.da Ramocoburro (Naro, Ag) Il sito di c.da Ramocoburro (Naro, Ag) si sviluppa attorno a una monumentale struttura idraulica composta da quattro vasche rivestite in cocciopesto e poste su due diversi livelli e si pone in stretta relazione con la via di comunicazione est ovest. 9. Verdura, Carabollace e Cozzo Ferraro (Sciacca, Ag)

12 C.da Firrìo (Racalmuto, Ag)

Nel territorio di Sciacca, alla foce di due torrenti, il Verdura e il Carabollace, sorgono in età tardo antica due insediamenti. Il primo vive per un periodo molto ristretto fra la seconda metà del IV e la metà del V d.C. (Parello et al. 2010); il secondo insediamento, invece, ha una ripresa fino alla fine del VI d.C. dopo una cesura a metà V (Caminneci et al. 2010, Caminneci 2011). Entrambi sono connessi ad una direttrice di scambi commerciali verso l’entroterra e sull’asse costiero E-W.

Anche il sito di c.da Firrìo ubicato in posizione strategica in contatto con l’entroterra agrigentino e con la vallata fluviale del Platani, si connota come un grande vicus in età tardo antica su preesistenze tardo repubblicane e imperiali; la presenza di reperti peculiari indica che la vita nel sito in epoca imperiale era connessa all’estrazione e alla raffinazione dello zolfo.

A Verdura (TAV. I, 1.17) sono attestate, oltre alle anforette (Tav. II, 1-6), anfore Keay XXV e Spatheia Bonifay Type 1, Keay XXXV A e B, Keay LV-LVII, LVII (D’Angelo 2010, 285). Analogo è il quadro delle presenze anforiche africane di Carabollace così come uguale è la scarsa presenza di materiali orientali (Franco 2011, 274-275). Si sottolinea l’assenza a Carabollace delle anforette di cui ci stiamo occupando (è stato trovato, infatti solo un esemplare nel terreno superficiale (Caminneci 2010).

Centri produttivi agrigentini Ad oggi sono tre i siti agrigentini in cui certamente erano attive fornaci che producevano le nostre anforette. Solo in uno di essi, quello di c.da Campanaio, sono state condotte negli anni passati indagini stratigrafiche (Wilson 2000) mentre gli altri due sono individuati solo attraverso indagini di superficie. Ciascuna area produttiva presenta litotipi abbastanza riconoscibili anche al solo esame macroscopico; in futuro indagini archeometriche potranno servire a caratterizzare meglio ciascuna associazione di minerali e dunque consentire di poter risalire agli atéliers di provenienza. La conoscenza di queste tre figlinae assieme ad analisi condotte su alcuni campioni ha potuto mettere in evidenza come accanto alla stragrande maggioranza di anforette sicuramente pertinenti al territorio siano presenti due esemplari alloctoni. Del primo si è già discusso (vedi supra) e potrebbe provenire da area etnea.

Un altro sito saccense sorge nell’entroterra, in c.da Cozzo Ferraro (TAV. I, 1.16) ed è caratterizzato dalla presenza di due spuntoni di calcare evaporitico che chiudono una sella in cui si è depositato uno spesso strato di humus. A valle di Cozzo Ferraro è documentata una piccola area di lavorazione della ceramica. Nella parte alta, invece, si è raccolto materiale che da età ellenistica arriva ad epoca alto imperiale.

215

LRCW4

Abbondanti inclusi bianchi di medie dimensioni, rari piccoli inclusi neri e raro quarzo.

Del secondo, le indagini archeometriche hanno indicato come probabile area di provenienza il centro Italia.

L.Z. 13 Agrigento Agrigento in età tardo antica pare risentire di un momento di crisi a metà del V d.C. (presenza di ampi strati di distruzione), ma presto si riorganizza, forse contraendo l’area dell’abitato. L’area produttiva sorge a nord e ovest rispetto al cosiddetto quartiere ellenistico-romano. Sono note alcune sezioni stratigrafiche dopo interventi di emergenza almeno dagli inizi degli anni ’40 del secolo scorso mentre ricognizioni di superficie hanno consentito di recuperare, oltre a una gran quantità di scarti di laterizi, anche alcuni orli di anfore tipo Mid-Roman 1(uno forse mal cotto) e due scarti di fornace delle nostre anforette (Tav. I, 7).

Problemi di cronologia I dati offerti dai siti agrigentini contribuiscono a meglio precisare l’arco cronologico entro cui sembra si sviluppi la produzione di questo tipo di anfora. Gli esemplari più antichi potrebbero essere quelli di Campanaio, datati da Roger Wilson nella seconda metà del IV secolo (Wilson 2000); entro la seconda metà del V si datano invece i contesti di Verdura. A Saraceno, i cui scavi recenti sono ancora inediti, si rileva comunque la presenza di anforette siciliane sia negli strati tardo antichi, che sembrano cessare intorno alla metà del V secolo, sia in quelli della prima età bizantina, databili entro la fine del VI secolo. Gli esemplari da Canalicchio provengono invece da uno strato di abbandono contenente materiale di cronologia compresa tra la fine del V e i primi decenni del VII secolo.

14 Giammaritaro (S. Elisabetta, Ag) Il sito si estende su un dolce pendio collinare e nel pianoro alla base (dove è rintracciabile traccia dell’area abitativa) in prossimità di vasti giacimenti argilliferi, di numerosi corsi d’acqua e sorgenti e soprattutto di un’arteria di comunicazione con la direttrice che conduce attualmente da Agrigento a Palermo sfruttando l’alveo fluviale del Platani. Oltre a un consistente nucleo di materiali tardo-repubblicani e alto imperiali sono stati raccolti orli di anfore Keay XXV, LXII, LV-LVII, e rari frammenti di sigillate africane, nonché la valva di una matrice per la produzione di lucerne assimilabili alla forma VIII (Parello et al. 2010). Si raccolgono inoltre numerosi scarti di fornace di ceramica comune ed uno riferibile ad un’anforetta di tipo siciliano. I prodotti di Giammaritaro hanno impasto depurato di colore arancio scuro, in frattura numerosi micro vacuoli; numerosissimi inclusi bianchi di piccolissime dimensioni. Rari inclusi di medie dimensioni.

Per definire le fasi finali della produzione, i dati più interessanti provengono dai contesti di Cignana: in questo sito, infatti, le anfore di tipo siciliano sono presenti, pur in piccole quantità, sia negli strati di uso e di crollo del vano 2/2006, che in quelli dell’ambiente 3/90, contesti più o meno coevi, databili tra la seconda metà del VI e la prima metà del VII secolo; certo, i contesti di Cignana contengono tutti una certa quantità di materiale residuale, tuttavia dal vano 3/90, tra i più affidabili stratigraficamente del sito, proviene uno degli esemplari più integri, conservato in buona parte. Le attestazioni di Cignana potrebbero essere dunque le più tarde finora note. Rientrano tutti infatti nell’ambito del V secolo i contesti di Termini Imerese da cui provengono le anfore di tipo 151/354, così come si chiudono entro la metà del V gli strati della necropoli paleocristiana di Agrigento nei quali sono state rinvenute le anforette. E’ datata invece tra la metà del V e la metà del VI secolo la seconda fase di utilizzo dell’edificio termale di Bagnoli S. Gregorio, cui va riferito il riempimento dell’ambiente 4, che ha restituito una grande quantità di questi contenitori (Spigo et al. 2006, 453-459).

15 Campanaio (Montallegro, AG) Il centro produttivo di C.da Campanaio sorge nell’immediata periferia W dell’attuale centro di Montallegro (Wilson 2000), sul versante collinare di un rilievo gessoso e a breve distanza dalla costa, in un’area ricca di sorgenti e in cui doveva essere facile approvvigionarsi di combustibile. Sono stati scavati ampi settori del sito e, accanto ad impianti per la trasformazione di prodotti agricoli e, forse, ad una piccola fornace per la produzione di calce, sono stati individuati più forni per produzione di ceramica.

La frammentarietà degli esemplari di Cignana non consente di verificare se le produzioni più tarde presentino esattamente le stesse caratteristiche morfologiche degli esemplari più antichi: in generale, non sembra, per il momento, di poter individuare indizi di una evoluzione morfologica del tipo. Per gli esemplari di Termini Imerese si è creduto di poter distinguere tra varianti più antiche con orlo estroflesso “a becco” ed esemplari con orlo indistinto, che potrebbero provenire da contesti leggermente più tardi; analogo significato cronologico è stato attribuito alle due varianti riconosciute tra i materiali di Capo d’Orlando. Per quanto riguarda i materiali agrigentini, al momento non sembra che si possa attribuire valore cronologico a queste piccole variazioni nella forma dell’orlo.

Scavi sistematici hanno documentato nell’area una lunga tradizione produttiva che arriva agli inizi del V secolo d.C., epoca a cui risale l’abbandono della fornace (Wilson 2000); vi sono sicuramente prodotte le anforette di cui ci stiamo occupando (Tav. II, 7). Esse hanno impasti macroscopicamente differenti, ma ne prevale un tipo ben depurato, rosso mattone, con numerosi vacuoli.

216

RIZZO-ZAMBITO-GIANNICI-GIARRUSSO-MULONE

caratteristiche che ne denuncino chiaramente la destinazione, anche se la mancanza della vaschetta di decantazione, che caratterizza di solito gli impianti di produzione dell’olio, potrebbe spingere a ritenere più probabile che esso servisse a produrre vino, raccolto direttamente in dolia nei quali avveniva la fermentazione. Dolia, in questa fase e nella successiva della prima età bizantina, si rinvengono in diversi punti del sito, ma si concentrano in particolare in un vano nel quale il più antico pavimento in opus spicatum è stato perforato per inserirvi diversi grandi contenitori. A Campanaio, invece, è stata individuata una vasca per la separazione dell’olio dall’acqua, mentre dolia, destinati evidentemente a contenere prodotti agricoli, furono rinvenuti nei pressi della vasca stessa (Wilson 1982). Questo dato è particolarmente interessante, considerato che nel sito è certamente attestata anche la produzione di anforette, che è difficile non mettere in relazione con le derrate che vi venivano certamente prodotte. Benché questa relazione non sia dimostrabile in modo certo, l’evidenza di Campanaio invita quanto meno a non dare per scontata la funzione di contenitori esclusivamente vinari delle nostre anforette, che vanno forse considerati piuttosto polifunzionali, come d’altronde sembra fossero anche gran parte delle anfore africane (Bonifay 2004, 463-475).

Alcuni dati sui siti di produzione Come si è visto, sono stati finora identificati tre siti nei quali sembra venissero prodotte anforette di tipo siciliano. Essi sono installati sia in siti rurali, sia nei pressi dell’area urbana e, secondo quanto documentano i casi di Campanaio e Giammaritaro, non specializzati unicamente nella produzione di anfore, come attestato d’altronde in altre parti dell’isola, per esempio a Bagnoli, dove veniva fabbricata anche ceramica da cucina (Spigo et al. 2006) e Santa Venera (Amari 2008; in questo volume), le cui fornaci producevano anche tegole e ceramica comune. Tra gli esami archeometrici che vengono qui presentati, il campione VE4 di Verdura ha caratteristiche che potrebbero essere compatibili con argille locali. Nella stessa direzione vanno anche i risultati, inediti, dell’esame di alcuni campioni provenienti da siti dell’entroterra agrigentino, che risultano compatibili con le formazioni litologiche della Sicilia centrale e occidentale. Tra i campioni presentati in questa sede, invece, diversi presentano caratteristiche che, se pur compatibili con una provenienza siciliana, non consentono di escluderne l’origine africana. In questo caso ci sembra che i dati archeologici, che attestano una ridottissima presenza del tipo in nord Africa e la sua assenza tra le produzioni degli atéliers africani, ormai in gran parte ben conosciuti, inducano a ritenere assai improbabile una loro fabbricazione in quest’area. Quarzo arrotondato è stato d’altra parte già identificato in produzioni insulari ed è presente anche in uno scarto di ceramica comune di Giammaritaro.

Le anforette di tipo siciliano ed il commercio Il tentativo di delineare una carta di distribuzione delle anforette di tipo siciliano è reso estremamente difficile dalle difficoltà di identificazione cui si accennava all’inizio: esse infatti sono state finora raramente riconosciute e, nelle pubblicazioni di scavi, sono state spesso classificate nell’ambito di tipi meglio noti. Tutto ciò impedisce oggi di avere un quadro certo, o anche soltanto probabile, della loro diffusione e, a maggior ragione, delle quantità, in termini assoluti e percentuali, presenti nei diversi contesti. Allo stato attuale si può soltanto constatare che una parte di questi contenitori veniva esportata, come attesta il relitto delle Egadi che viene qui presentato (Oliveri, in questo volume) e che essi, come già era stato rilevato dagli studiosi (Rizzo 1993, 224; Bonanno e Sudano 2007, 355; Panella et al. 2010, 66) sono presenti in alcuni contesti romani. Sembrerebbe, tuttavia, che in termini quantitativi si tratti di una presenza decisamente minoritaria, rispetto ad altre produzioni siciliane molto meglio attestate, prima tra tutte quella delle Keay LII. Ciò è evidente ad esempio nel contesto della seconda metà del V secolo della Magna Mater, dove le anfore Termini Imerese 151/354 sono state riconosciute con certezza (Panella et al. 2010, 66). Nei depositi di V secolo della Basilica Hilariana sul Celio, il complesso degli esemplari riferiti al tipo Ostia IV, 166 e dei frammenti di difficile identificazione di probabile provenienza italica, nel cui ambito potrebbe aver trovato posto anche qualche anfora di “tipo siciliano”, rappresentano una percentuale assolutamente minoritaria degli stessi contenitori di provenienza italica, laddove le Keay LII contano da sole per il 22% del totale delle attestazioni anforiche (Bertoldi e Pacetti 2010, 434). In qualche caso rimane il dubbio sulla possibilità di

Le anforette e la produzione agricola: primi dati dal territorio agrigentino E’ ovvio che uno dei problemi più interessanti riguarda il contenuto delle nostre anforette, che in genere, per le ridotte dimensioni, vengono ritenute vinarie. I rinvenimenti agrigentini, pur non apportando dati certi su questo argomento, offrono però almeno la possibilità di mettere in relazione i contenitori con i siti nei quali sono stati rinvenuti, e con le attività che vi si svolgevano. I siti di rinvenimento, come si è visto, sono sia settori dell’antico centro urbano di Agrigento, sia, e sono la maggior parte, insediamenti rurali chiaramente finalizzati allo sfruttamento agricolo del territorio, frutto di quella riorganizzazione dell’insediamento rurale, ben evidente nelle campagne siciliane, e agrigentine in particolar modo, tra IV e V secolo. In due casi si tratta certamente di insediamenti nei quali si svolgevano attività di trasformazione dei prodotti agricoli. A Saraceno, infatti, è stato messo in luce da una recente campagna di scavo un piccolo impianto produttivo costituito da una base circolare, probabilmente per un torchio, connessa con un canale di adduzione, ottenuto con uno spatheion, che sfociava in una vaschetta rettangolare, rivestita con pietre e tegole ma non impermeabilizzata e destinata dunque forse ad alloggiare i dolia nei quali il liquido prodotto veniva riversato. L’impianto è pertinente all’insediamento rurale di V secolo e non sembra avere 217

LRCW4

forse in modo non continuo e costante, come avveniva invece per i prodotti di altre aree dell’isola, poteva venire immesso nei canali dell’esportazione transmarina.

attribuire al nostro tipo anfore classificate, ad esempio, come “evoluzioni tarde” delle Keay LII, come un esemplare proveniente da Vigna Barberini (Rizzo et al. 2004, tav. II, 11) o altri, rinvenuti al Celio e nell’area a SE della Crypta Balbi, già segnalati da Carmela Bonanno (Bonanno 2007, 356). In generale, Roma, stando a quanto oggi sappiamo, sembra essere la principale destinazione di quest’anfora al di fuori dell’isola; se ne può forse inoltre sospettare una debole circolazione in ambito tirrenico, se, ad esempio, sono attribuibili a questo gruppo alcuni esemplari da Marsiglia (Bonifay and Villedieu 1989, Fig. 14, 18; Bonifay et al. 1998, Fig. 94, 144; Fig.136, 243; Fig. 201, 316; Fig. 247, 96); tra le anforette di probabile origine siciliana rinvenute a Mariana, in Corsica (Pasquinucci et al. 2007), tuttavia, provenienti essenzialmente dall’area dello Stretto, non sembra ve ne siano attribuibili al nostro gruppo; è possibile invece che siano contenitori vicini al nostro gruppo due anfore rinvenute a Nicotera (Cuteri et al. 2007, Fig. 8, 21-22).

Non ci risulta che siano state condotte, ad oggi, indagini archeometriche su anfore di questo tipo rinvenute fuori dalla Sicilia: sarebbe certamente interessante, per la comprensione delle dinamiche economiche del territorio agrigentino, sapere se esso abbia partecipato al rifornimento in vino o olio della capitale o se l’attività di esportazione di questa parte dell’isola riguardasse esclusivamente i cereali, rimanendo confinate le altre produzioni agricole all’autoconsumo. Analogo interrogativo si pone per i predecessori delle nostre anforette, le Mid-Roman 1, prodotte, a quanto sembra, anche nell’agrigentino: ci auguriamo che in futuro si possa giungere a capire se all’ampia diffusione nel Mediterraneo di questi contenitori abbiano partecipato tutte le aree dell’isola o se essa sia stata alimentata essenzialmente da determinati distretti produttivi i cui prodotti agricoli specializzati erano particolarmente vocati all’esportazione: quegli stessi, possiamo immaginare, che avrebbero continuato anche successivamente ad esportare su larga scala il proprio vino, prevalentemente conservato entro anfore Keay LII, prodotte anche nella Calabria meridionale, ereditandone sostanzialmente i circuiti commerciali.

Qualche esemplare, infine, sembra aver raggiunto Cartagine e Atene (Robinson 1959, Agorà M 323 e 324 e, in livelli di III-IV, L 31 e 32). Si tratta comunque di presenze quantitativamente modeste, per quanto questa valutazione possa essere viziata dal possibile, e in alcuni casi probabile, non riconoscimento del tipo.

In ogni caso, l’identificazione di produzioni anforiche nel territorio agrigentino contribuisce a delineare il quadro di un’economia rurale vivace che, alle produzioni con ogni probabilità dominanti e destinate prevalentemente all’esportazione, come quella cerealicola, affianca attività agricole finalizzate forse prevalentemente all’autoconsumo, tali da rendere il territorio non completamente dipendente dai prodotti d’oltremare, pur presenti in abbondanza. Questi ultimi, quantitativamente assai rilevanti in tutti i siti rurali finora indagati nell’agrigentino, potrebbero denunciare una produzione di vino ed olio insufficiente al consumo interno, tanto da rendere necessario un forte contributo delle importazioni, sia dall’Africa, sia dal Mediterraneo orientale. Possono però testimoniare anche, proprio perché si affiancano alle produzioni locali, l’esistenza di un’ampia classe in grado di alimentare, anche nei villaggi rurali e fino all’avanzato VII secolo, la domanda di beni di maggior pregio.

In Sicilia, come si è visto, le anforette sono ampiamente diffuse; nella maggior parte dei casi, però, in ciascun sito si rinvengono in prevalenza esemplari prodotti localmente o comunque nell’area circostante il sito di rinvenimento: sono questi in genere i dati che si ricavano dagli esami archeometrici degli impasti, che, ad esempio, hanno rilevato la provenienza da officine localizzate nella Sicilia nord-orientale per le anfore di Caronia (Bonanno e Sudano 2007, 356) e la compatibilità con argille locali per gli esemplari di Termini Imerese; sembrano nel complesso prodotti prevalentemente nella Sicilia centrale ed occidentale gli esemplari rinvenuti nel territorio agrigentino. Almeno una parte della produzione di queste anforette sembrerebbe dunque destinata in primo luogo all’autoconsumo e ad una circolazione in un ristretto ambito sub-regionale, come, evidentemente, i prodotti che contenevano: d’altronde, se l’insediamento di Campanaio può costituire un modello, almeno una parte dei siti di produzione dei nostri contenitori dovevano essere rappresentati da villaggi rurali, nei quali la fabbricazione di anfore si affiancava a quella di tegole e ceramica comune (come confermano tutti gli esempi finora noti, anche quello di Giammaritaro) e ad altre attività artigianali (la fabbricazione della calce, attestata anche a Saraceno) (Wilson 2000), così come la trasformazione dei prodotti agricoli e la produzione delle derrate che dovevano essere contenute nelle anfore doveva affiancarsi a quella dei cereali, da considerare forse primaria, se dobbiamo dare per scontata la vocazione essenzialmente cerealicola delle campagne agrigentine, e attestata, nello stesso sito di Campanaio, dal rinvenimento di semi di grano (Wilson 1990, 189). Ciò non esclude la produzione di un modesto surplus che,

M.S.R. Le analisi petrografiche I reperti ceramici oggetto di studio sono stati analizzati al microscopio polarizzatore, su sezione sottile, al fine individuare gruppi omogenei riconducibili a differenti tipologie di impasto e riconoscere la produzione locale dalle importazioni. I gruppi sono stati definiti in base ai caratteri mineralogico – petrografici dello scheletro sabbioso e altri parametri tessiturali quali l’addensamento, la granulometria, la forma dei clasti, la morfologia e la dimensione dei pori, il colore e l’aspetto microstrutturale della pasta ceramica.

218

RIZZO-ZAMBITO-GIANNICI-GIARRUSSO-MULONE

I risultati esposti di seguito, che si riferiscono alle descrizioni dei caratteri mineralogici e tessiturali dei singoli campioni, costituiscono la base di partenza per l’analisi dei raggruppamenti e le relative ipotesi di provenienza.

un’argilla calcarea ricca in ossidi di ferro e contiene numerosi pori da decomposizione termica di microfossili e bioclasti. Gruppo 2 (scheletro quarzoso a granulometria da siltosa a sabbiosa molto fine). I caratteri mineralogici e tessiturali sono piuttosto comuni e compatibili con le materie prime argillose locali.

Le indagini, condotte mediante osservazione di sezioni sottili al microscopio polarizzatore, hanno consentito di suddividere i reperti ceramici indagati nelle seguenti categorie:

Campione VE – A4 Anfora Impasto caratterizzato da scheletro smagrante mediamente abbondante (addensamento circa 15%) a granulometria da siltosa a sabbioso molto fine, costituito da clasti sub-angolosi di quarzo e, subordinatamente, feldspati. Nell’impasto inoltre sono presenti sporadici bioclasti calcarei delle stesse dimensioni dei clasti silicatici e rari calcinelli di dimensioni grossolane (1mm). La matrice ceramica ha l’aspetto tessiturale di un’argilla calcarea a basso grado di cottura.

Gruppo 1 (scheletro quarzoso bimodale con elementi ben arrotondati, subordinati microfossili e bioclasti decomposti per cottura, rari clasti di quarzo policristallino, selce, feldspati. Le caratteristiche composizionali e tessiturali dell’impasto sono compatibili con una produzione siciliana ma non si può escludere la provenienza africana. E’ rappresentato dai seguenti campioni: Campione VE – A2 Anfora

Gruppo 3 (scheletro quarzoso abbondante a granulometria grossolana con elementi ben arrotondati, subordinati bioclasti e litoclasti calcarei). I caratteri mineralogici e tessiturali sono compatibili con una provenienza siciliana o dal nord-Africa.

Impasto caratterizzato da scheletro smagrante poco abbondante, mal distribuito con addensamento medio pari a circa 10%, a granulometria bimodale con prevalenza delle frazioni siltosa e sabbiosa media (diametro massimo: 0.3mm). La frazione media è rappresentata prevalentemente da granuli sub-arrotondati e arrotondati di quarzo, sia monocristallino che policristallino, e subordinatamente da feldspati (ortoclasio, microclino e plagioclasio) e raramente da granuli di selce. La pasta ceramica ha l’aspetto tessiturale di una argilla calcarea ricca in ossidi di ferro, ad elevato grado di cottura. In essa si osservano numerosi pori da decomposizione termica di originari microfossili calcarei contornati da aloni di schiarimento.

Campione CA10-A15 Impasto caratterizzato da scheletro abbondante (addensamento circa 25%) a granulometria eterogenea, tendenzialmente bimodale con prevalenza delle frazioni della sabbia molto fine e della sabbia media (dimensioni massime: 0.7mm). La frazione fine è costituita da quarzo angoloso; quella grossolana da granuli sub-arrotondati e arrotondati di quarzo, bioclasti (piastrine monocristalline di echinodermi), e di calcare bioclastico a cemento da microsparitico a sparitico. Raramente si osservano clasti di feldspato (ortoclasio). La pasta di fondo ha l’aspetto tessiturale di un’argilla calcarea sottoposta a medio grado di cottura. Essa presenta numerosi relitti di microfossili calcinati e frequenti vacuoli di forma allungata.

Campione VE – A3 Anfora Impasto caratterizzato da scheletro smagrante mediamente abbondante, mal distribuito con addensamento medio pari a circa 20%, a granulometria bimodale con prevalenza delle frazioni siltosa e sabbiosa media (diametro massimo: 0.3mm). La frazione media è rappresentata prevalentemente da granuli sub-arrotondati e arrotondati di quarzo, sia monocristallino che policristallino, e subordinatamente da feldspati (ortoclasio, microclino, plagioclasio). La pasta ceramica ha l’aspetto tessiturale di una argilla calcarea ricca in ossidi di ferro, a medio grado di cottura. In essa si osservano numerosi pori da decomposizione termica di originari microfossili e rari calcinelli.

F.G., R.G., A.M. Bibliografia Amari, S. 2006. I materiali in esposizione nell’Antiquarium. Sale I-II-III. In M. G. Branciforti (ed.), L’area archeologica di Santa Venera al Pozzo Acium. Antiquarium, 105-183. Catania.

Campione CGN3456

Amari, S. 2008. A Late Roman Amphorae Production in Eastern Sicily. In O. Menozzi, M. L. Di Marzio and D. Fossataro (eds.), SOMA ’05, Symposium of Mediterranean Archaeology, Università degli Studi di Chieti 24-26 febbraio 2005, British Archaeological Reports, International Series 1739, 473-479. Oxford, BAR Publishing.

Impasto povero di scheletro (addensamento: 15%) a granulometria bimodale con prevalenza delle frazioni siltosa media con diametro massimo pari a 0,3 mm. La frazione siltosa è costituita da clasti angolosi di quarzo; quella sabbiosa da clasti sub arrotondati e arrotondati di quarzo e raramente di feldspato (ortoclasio e plagioclasio). La pasta ceramica ha l’aspetto tessiturale di 219

LRCW4

Bertoldi, T., Pacetti, F. 2010. Materiali di V secolo dalla Basilica Hilariana sul Celio: analisi tipologica delle ceramiche comuni. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Comparison between western and eastern Mediterranean, British Archaeological Reports, International Series 2185, 433-445. Oxford, BAR Publishing.

Castellana, G., McConnell, B. 1990. A rural Settlement of imperial and byzantine date in Contrada Saraceno near Agrigento, Sicily. American Journal of Archaeology 94, 1, 25-44. Cuteri, F. A., Corrado, M., Iannelli, M. T., Paoletti, M., Salamida, P., Sangineto, G. B. 2007. La Calabria fra tarda antichità e alto medioevo: ceramiche, commerci, strutture. In LRCW2, 461-476. D’Angelo, F. 2010. Le anfore. In Parello et al., 285.

Bonacasa Carra, R.M. 1987. Agrigento paleocristiana. Zona archeologica e antiquarium. Palermo.

La Lomia, M. R. 1961. Ricerche archeologiche nel territorio di Canicattì. Vito Soldano. Kokalos VII, 1961, 405-408.

Bonanno, C., Sudano F. 2007. L’insediamento in località Pantano di Caronia Marina (Messina): contesti tardo antichi e bizantini. In M. Bonifay and J.-C. Treglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry, British Archaeological Reports, International Series 1662, 353-363. Oxford, BAR Publishing. Bonanno, C. 2009. Kalè akté: scavi in contrada Pantano di Caronia Marina, Messina, 2003-2005. Roma.

Lima, M. A. 1995. Le anfore. In R. M. Bonacasa Carra (ed.), Agrigento. La necropoli paleocristiana sub divo, 237-291. Roma. Malfitana, D. 2008. Roman Sicily Project («Rsp»): Ceramics and Trade, a multidisciplinary approach to the study of material culture assemblages. First overview: the transport amphorae evidence. In D. Malfitana, J. Poblome, and J. Lund (eds.), Facta. A Journal of Roman Material Culture Studies 2, 135137.

Bonanno, C., Cabella, R., Capelli, C., Piazza, M. 2010. Nuove esplorazioni in località Gerace (Enna-Sicilia). In LRCW3, 261-272.

Pasquinucci, M., Menchelli, S., Capelli, C., Picchi, G. 2007. Corsica tardo antica: anfore italiche e ceramica comune da Mariana. In LRCW2, 313-328.

Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports, International Series 1301. Oxford. BAR Publishing.

Pacetti, F. 1998. La questione delle Keay LII nell’ambito della produzione anforica in Italia. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo, Atti del Convegno di studi in onore di John W. Hayes, 185-208. Firenze, All’Insegna del Giglio.

Bonifay M., Carre, M.-B., Rigoir, Y. 1998. Fouilles à Marseille. Les mobiliers (Ier-VIIe siécles ap. J.-C.). Marseille.

Panella, C., Saguì, L., Coletti, F. 2010. Contesti tardo antichi di Roma. Una rilettura alla luce dei nuvi dati. In LRCW3, 57-78.

Bonifay, M., Villedieu, F. 1989. Importation d’amphores orientales en Gaule (V-VII siècles). In Deroche, V., Spieser, J. M. (eds.), Recherches sur la céramique byzantine, Bulletin Correspondance Hellenique, Supplément 18, 17-46. Atene.

Parello, M. C., Amico, A., D’Angelo, F. 2010. L’insediamento alla foce del Verdura in territorio di Sciacca (Agrigento, Sicilia, Italia). I materiali ceramici. In LRCW3, 283-291.

Borrello, L., Lionetti, A. 2008. La ceramica. In G. Tigano (ed.), Terme Vigliatore-S. Biagio. Nuove ricerche nella villa romana (2003-2005), 65-79.

Riley, J. 1979. The Coarse Pottery from Berenice. In J. A. Lloyd (ed.), Excavations at Sidi Khrebish Benghazi (Berenice) II, Libya Antiqua, Supplement 5, 91-466. Tripoli.

Burgio, A., Belvedere, O. 2010. Sulla viabilità nel territorio di Palma di Montechiaro. Brevi note in margine al “Progetto Cignana”. In C. Marangio and G. Laudizi (eds.), Palaià Filia. Studi in onore di G. Uggeri. Galatina.

Rizzo, G., Capone, M., Costantini, C., Gafa, R., Pentiricci, M., Munzi, M. 2004. Vigna Barberini, settore D, Periodo IV: 540/550-580/590 d.C. In L. Paroli e L. Vendittelli (eds.), Roma dall’antichità al medioevo II. Contesti tardoantichi e altomedievali, 72-90. Milano.

Caminneci, V. 2010. Tra il mare e il fiume. Dinamiche insediative nella Sicilia occidentale in età tardo antica: il villaggio in contrada Carabollace (Sciacca, Agrigento, Sicilia, Italia). http://www.fastionline.org

Rizzo, M. S. 1993. Anfore. In O. Belvedere, A. Burgio, R. Macaluso and M. S. Rizzo (eds.), Termini Imerese. Ricerche di topografia e di archeologia urbana, 217225. Palermo.

Capelli, C., Bonifay, M. 2007. Archéometrie et archéologie des céramiques africaines: une approche pluridisciplinare. In LRCW2, 551-567. Castellana, G. 1984-1985. Ricerche nel territorio di Favara e Palma di Montechiaro. Kokalos 30-31, 521527.

Rizzo, M. S. and Zambito, L. 2010. Ceramiche comuni ed anfore dal villaggio tardo antico di Cignana (NaroAgrigento, Sicilia, Italia). In LRCW3, 293-300.

220

RIZZO-ZAMBITO-GIANNICI-GIARRUSSO-MULONE

and J. Lund (eds.), Old Pottery Studies. Atti del Convegno Internazionale di Studi, 451-464. Catania.

Robinson, H. S. 1959. The Athenian Agora V, Pottery of the Roman Period, Chronology. Princeton, N.J., The American School of Classical Studies at Athens.

Wilson, R. J. A. 1982, Una villa romana a Montallegro (Agrigento). Sicilia Archeologica XV, 48, 7-20.

Saguì, L. 1998. Il deposito della Crypta Balbi: una testimonianza imprevedibile sulla Roma del VII secolo? In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo, Atti del Convegno di studi in onore di John W. Hayes, 305-330. Firenze.

Wilson, R. J. A. 1990. Sicily under the Roman Empire. Warminster. Wilson, R. J. A. 2000. Rural Settlement in Hellenistic and Roman Sicily: excavations at Campanaio (Ag) 199498, Papers of the British School at Rome LXVIII, 337-369.

Spigo, U., Ollà, A., and Capelli, C. 2006. La ceramica di produzione locale dalle terme di Bagnoli-S. Gregorio a Capo d’Orlando (Me). In D. Malfitana, J. Poblome,

221

LRCW4

Tavola I: 1. Distribuzione delle anforette nel territorio agrigentino (11-13 centri produttivi); 2. Anforetta da c.da Stefano; 3-4. Anforette da c.da Saraceno; 5. Fondo di anforetta da Calamonaci; 6. Anforetta da Cignana; 7. Scarti di anforette da Agrigento (c.da Bonamorone).

222

RIZZO-ZAMBITO-GIANNICI-GIARRUSSO-MULONE

Tavola II: Anforette da c.da Verdura, Racalmuto, Cignana, Campanaio: scala 1:4 Foto delle sezioni sottili

223

LRCW4

224

OBSERVATIONS ON THE LATE ROMAN COARSE WARES PRODUCTION FROM THE FACTORY OF ACIUM IN SICILY (ITALY) SUSANNA AMARI University of Catania (Italy); [email protected] In the archaeological area of Santa Venera al Pozzo (Acicatena-Catania) is situated the mansio Acium, placed at the ‘cross-roads’ of four principal ways of the Roman road network on the island and located along the ancient river Akis at 4km from the estuary in the port of Capo Mulini. At Acium was found a late Roman factory with three different types of common wares, amphorae and kilns for building materials as well as many rooms related to pottery manufacture and a warehouse for storing the products of the kilns. The factory is dated from the 4th until the middle of the 5th centuries AD. This paper focuses on the study of late Roman coarse wares produced at the factory, all in a red-orange clay and some decorated with impressed wave lines. The aim of this research is to present Acium as a rural settlement with an important role in the economical life of Sicily participating in the distribution of foodstuffs and wares and in the trade carried out in the Mediterranean. KEYWORDS: ACIUM (CATANIA-ITALY), PRODUCTION SITE, LATE ROMAN, AMPHORAE, CERAMIC BUILDING MATERIAL, COARSE WARE POTTERY, LONG-DISTANCE DIFFUSION the legend of Akis and Galatea, transformed into two rivers in order to stay together forever. Modern local sources mention that S. Venera al Pozzo was still frequented in the medieval ages because of a church, the free-market, sources of healing water and the so-called ‘Spedale di S. Venera’ a reconstruction of the Roman spa for modern use of the therapeutic baths. In the modern site exists also a church in front of the well where St. Venera was submitted to her martyrdom. In 1873, the two warm sulphurous water springs (22°C) were piped to a modern spa building (at 2kms), in Acireale. The main source is under the well and the secondary is at few metres from the caldaria. The archaeological area of S. Venera al Pozzo and the Antiquarium are open to the public (Branciforti 2006). It is currently the headquarters of the Archaeological Park Service and Landscaping of the Valley of the Aci.

I hereby would like to present some results of a research at Santa Venera al Pozzo in Sicily concerning a production site of amphorae, ceramic wares and building materials that played an active role in the late Roman society and the Sicilian economy. The mansio Acium is situated in proximity to two sites of great movement of people, goods and ideas, i.e. the road Pompeia and the port of Capomulini. Considering the logistic advantages of trade via the river Akis it is self evident that Acium had an important economic role in the production of raw materials (e.g. wood), agricultural products (fruit, vegetables, cereals, wine, oil, etc.) and ceramic wares used in harvest, sale, storage and consumption and also in and their distribution by land-, river- and sea-routes. The archaeological area

Recent archaeological excavations, in which I participated as a scientific collaborator, have revealed that the site, rich in natural water springs has been frequented since the prehistoric period; in the Greek age it was a rural centre connected with the cult of the waters and of the chthonian deities Demetra and Kore. Furthermore, we found remains of Roman occupation which is explained by the fact that it is a site located in a naturally fertile area. The island in the Roman period was characterized by intensive occupation and extensive rural landscapes related to the expansion of agricultural production, regional scale craft and industrial production of pottery, amphorae, and ceramic building materials.

Research-programmes funded by the EU and carried out by the Superintendence of Catania in the years 1999-2001 and 2004-2008 at S. Venera al Pozzo around the Roman baths, the water springs and the temple, led to the identification of ancient Acium, listed in the Antonine Itinerary along the road between Messina and Catania. The Roman thermal building was known from the 18th century. Some sculptures, mosaics, coins, wares, and the podium of the Roman temple were discovered in the following centuries. The possible religious function of the Roman settlements is supported by the discovery of a temple and of a marble head of Minerva – it seems that the Romans were using the sulphurous water springs at Bath. Regarding the Roman cult, the tradition here would be the martyrdom of St. Venera, decapitated in AD 143 and her head thrown into a well that made the water miraculous. Concerning the presence of water in the area during the Greek period, the classical sources hand down

In the 1st century AD a rural villa was built at S. Venera al Pozzo on the ruins of the Greek Akis, 160m north of the Roman baths. The building and the adjoining agricultural land extend at 1700m2 around an inner court. The villa is located near a water spring, from where water 225

LRCW4

was piped into a well built in opus caementitium at the end of the 1st century. The villa was in use until the end of the 3rd century AD.

entitled “Sicilian Amphorae (AD I-VI): an economic analysis of production and distribution”, at the University of Oxford (supervisor: Prof Andrew Wilson). All the samples collected will be studied by Claudio Capelli (University of Genova).

Products in the factory From the 2nd decennium of the 4th until the middle of the 5th century AD - when all activity came to an end possibly because of Vandal attacks – the Roman building was transformed into an active factory with kilns producing three types of common wares, amphorae and ceramic building materials. They were all active in the mid-4th century. Many rooms of the villa were utilized for the manufacture and storage of ceramic wares. The well of the villa continued to be used and inside the building was built a ceramic drain. It is clear from the evidence already presented above that there were raw materials (water, clay and wood) around the site. Normally, the figlinae have a supply area near the workshop. A ceramic workshop is still active on the slopes of the clay hill, just a few hundred meters from S. Venera al Pozzo.

Amphorae: new data In the late Roman factory of Acium at least three shapes of amphorae were produced and they were identified by their wasters; they were all used for transportation, sale, and storage of wine and maybe oil. The first form is the Agora M254/Ostia I, 453-454/Ostia III, 46/Benghazi Mid Roman 1a/Keay LXXXI/Peacock and Williams class 40; the second shape is the Agora M234/Keay LII and the third is the LRA 1 similis (Amari 2006; 2008). They were all produced in a characteristic red-orange clay, with mica, and contain calcareous and black volcanic rock inclusions. They have a light cream-pink or pale brown color on the exterior and the upper part of the interior. A few are beige on the external surface. The ceramic body is fine, hard, compact, and the section is clean. In some cases the core of the ceramic body is darker than the surfaces as a result of the firing process. The discovery of the factory is of great importance for the production of forms and their variants, especially for the MR1 and Keay LII. The MR1 is known from many sites of consumption in the Mediterranean area. According to petrographic analysis it seems to represent production predominantly of North-Eastern Sicily. Similar forms might have been produced in some factories in Africa (Capelli and Bonifay 2010, 554, fig. 10a), but not vice versa. Acium and Naxos are two production sites of MR1 and Keay LII and stationes listed in the Antonine Itinerary. Naxos has a long tradition of producing amphorae. To today we only know that this workshop of Acium began its production around 320 AD, when the shape of the best known amphora that began to be produced for the first time is Keay LII, with its many types and variations. Now at S. Venera al Pozzo there are no traces of the previous productions or other workshops. The Acium shapes are characterized by numerous variants in the rims and bases. Some examples with swollen or split rim can be distinguished from shape Keay LII, characterized by a cauldron-shaped body and a flat base. Some variants of the rim (fragments without neck and shoulder) are difficult to classify between the two forms MRA1 and Keay LII, on the basis of the current published parallels. Here are presented only some types of these three forms. A detailed typological study of the amphorae is in preparation.

The production of Acium consists of vessels useful in everyday activities within the workshop (including lamps, sieves, tanks, basins, potter’s wheels, support disks, spacers of kilns), and products targeted especially for the market. The manufacturing chronology and the life-span of the workshop are confirmed by the recovery of good luck coins found under the floor of the kiln’s combustion chamber (Amari 2006, 174-7) and several forms of African red-slip ware - Hayes forms 32/58, 50A/B, 61B, 67, 78, (Amari 2006, 165-7) -, as well as African amphorae and lamps dated from the 4th until the first half of the 5th centuries AD (Amari 2006, 160-1). Only some of the rooms of the Roman factory were fully excavated in 2006 but we have chosen to highlight the whole perimeter of the factory. After some preliminary publications on the amphorae (Amari 2008) and the building materials (Amari 2007), the objects in the Antiquarium (Amari 2006) and Acium’s activities (Amari 2011), this paper focuses on the production of late Roman coarse wares in the factory. Many samples bear impressed wavy lines (amphorae and coarse ware) or streaked finger marks (ceramic building materials). The fabric is in all specimens in a red-orange colour (Munsell 5YR, 7/8 and 7.5YR, 7/4-7/6) with black rock volcanic inclusions (more abundant in brick and tiles) and mica. There are a few calcareous inclusions. The ceramic body is hard, compact, and the section is clean. In a few specimens the core of the ceramic body is darker than the surfaces as a result of the firing process. Most samples have an external pinkish surface, others are pale brown and beige. On about twenty samples archaeometrical nondestructive analysis was carried out with a portable LNS PIXEalpha system, in the LANDIS laboratories of the National Institute of Nuclear Physics at the University of Catania. PIXE analysis applied on the surface of the samples detected elements as light as K, Ca, Fe, Rb, Sr, Zr (Amari 2007, 122). Selected local late Roman samples have been submitted for thin-section analysis. The results will be discussed by Carmela Franco in her D.Phil. thesis,

The Acium Mid Roman 1a (Fig. 1.1-3) has a rim with rib more or less pronounced (diameter range 5.5-6.8cm), a narrow and cylindrical neck, rounded or grooved handles (few to ‘fiorellino’), a carinated shoulder, a swollen ovoid body and wide base with ring foot (diameter range 8.29.5cm). Many samples have a beige (Fig. 1.2) or pale brown (Fig. 1.1) external surface. Often, fingerprints to attach the loops are seen on the top of the handles.

226

AMARI

The flat-bottomed Acium Keay LII production (Fig. 1.48) has a rim of triangular section, swollen, split or mouldmade (diameter range 6.0-7.6cm), a tall cylindrical neck, a curved shoulder indistinguishable from the grooved ovoid body tapering at the bottom, rounded or slightly grooved handles and flat base with narrow ring (diameter range 6.4-7.5 to 7.7-8.5cm) and a concave bottom. Many samples have wavy lines impressed around the neck. Often, fingerprints to attach the loops are seen on the top of the handles. The peculiar rim-band type (moulded) finds parallels among the published examples from Rome (Pacetti 1998, Celio Figs. 1.2, 3.1-2, 13.2) which have also a triangular section, swollen and split rims (Pacetti 1998, Palatino Figs. 4.1, 13.1 and Celio Fig. 6.1).

productions of the Italian peninsula. The recent discovery of some integral imbrices, 58cm in length, allowed us to distinguish with certainty the use of the two types of tegulae produced. The tegulae of 58x48cm were used for roofing (Fig. 2.1), while those of 72x49-52cm for paving. In the workshop of Acium, the latter were used to cover the pavement of the east portico and the duct through the building. Both types are present in Pompei and Ostia. In the factory we found many caps (diameter range c. 8.814cm, thickness 2.3-2.8cm) obtained from rectangular flat-tegulae (Fig. 1.12). The pastes of the bricks and the tiles found at Acium are reddish-orange with many sizeable inclusions, presumably volcanic rock inclusions, ranging between 2mm. The majority of the production has a beige or pale pink colour surface, rich in calcium (CaCO 3 ). There are no traces of sodium (Na) or chlorine (Cl) as the PIXE analysis has shown (Amari 2007). The lower part of the majority of the bricks bears finger marks (single groove, double lines or double curves, wavy lines, single or triple X) or incised motifs (stars and herringbones) formed on the wet clay. These were expedient methods for setting the lime mortar better and distinguishing the wares of the specific workshop. The curving tiles, last production of the factory in the first half of the 5th century AD, are decorated with streakedfinger lines on the upper surface and/or combed lines (Fig. 2.3). The Acium type keeps the exclusive way of the factory of marking the bricks and tiles with wide finger marks. The workshop, strongly rooted in its tradition, did not use only the new combed lines. This could be an experimental production of the factory. The production of Acium attested by a few samples but also from the furnace wasters, is of particular importance. It seems that before the mid-5th century AD, two systems of laying roof tiles co-existed: one with flat tegulae being covered by curved imbrices (Fig. 2.1); the second system which soon substituted the first consisted of rows of curved pan tiles laid in rows and cover tiles were set on top of the joints (Fig. 2.2). This system is still in use on roofs of the 'Sicilian' style. Furthermore, the curving tiles mark the introduction of a new type of tile in the first half of the 5th century AD which, although modelled with the usual hard and dense ceramic body, preludes, together with the introduction of decoration comb, a new production that will characterize the combed tiles of the Byzantine era. Tiles have also been found in the sea, near the Catanian coast; they are in a red fabric which does not seem to be of Acium production. The presence of tile shipwrecks in the Mediterranean Sea confirms the existence of a seatrade of bricks and tiles and the hypothesis that the Acium production of building materials was intended also for marketing.

Acium LR1 similis (Fig. 1.9-11) are containers with a swollen and slightly ribbed body, generally with a high ribbed rim (diameter range 9.2-10.7cm), with rounded, everted or oblique edge and rounded or slightly grooved handles. Most specimens have wavy lines impressed before firing around the cylindrical neck. Fingerprints to attach the loops are seen inside the neck and on the top of the handles. They seem to have been used as containers for wine, olives, olive-oil and other staples for the local or regional markets. These amphorae are unknown elsewhere. There is no evidence for their distribution in the vicinity or elsewhere (Amari 2008). We know more about the distribution of the MRA 1a and Keay LII amphorae (Roma, Napoli, Corsica, France, Spain, Aegean, etc.; cf. Bonifay and Pieri 1995, 114-5, fig. 12, 76-77; Panella 2001, 196; Capelli and Bonifay 2007; Menchelli et al. 2007). The excavation data suggests a contemporaneous production of the three forms, but we still do not know at what percentage. The first and second shape, although in lower percentages than the MRA1, had contemporary distribution at Rome (Pacetti 1998, 189, tables 1-2) until the first half of the 5th century AD, when the activities of the Acium-factory ceased. In the 4th century shipwreck at Yassi Ada one MR1 and one Keay LII amphora were found (Bass and van Doorninck 1971, 34-6, figs. 26, 28). Samples probably similar to the Acium Keay LII have been found in the excavations in Rome on the slopes of the Palatine (Casalini and Crespi 2010, 101-2, 106, figs. 1, 1.8, 4.a Fabric 1. Keay LII, gruppo a; pers. comm.). Ceramic building materials: classical and experimental good quality types. Metrological update. The production consists of many bricks and tiles in many shapes and of good quality, used for building and tiling of walls, columns, arches, floors, roofs, concamerationes, suspaensurae with ring-bricks. The tools and brick types used for daily tasks in the factory are not presented here. Different types are present in the workshop: many shapes of bricks, the rectangular flat-tegulae, the semicircular imbrices and the curving tiles (Amari 2006, 149-158, 169-171; 2007). In the 4th and 5th centuries AD regional types of bricks and tiles were produced in the Roman Empire. Regional production occurred also at Acium, although some types correspond in size to other

Coarse ware: shapes, function and contents. The factory of S. Venera al Pozzo was specialized in the production of coarse wares. The pottery made in the factory was not valuable, but of good quality and part of the production was useful in the daily agricultural and domestic activities, such as harvest, storage, and consumption of food and drink. These containers were 227

LRCW4

produced in a red-orange clay, with mica and black rock inclusions. Most have a light cream-pink coloured exterior surface, some have wavy lines impressed before firing. The ceramic body is fine, hard, compact, and the section is clean. Often the core of the ceramic body is darker than the surfaces as a result of the firing process. There are open and closed shapes found at various quantities in the factory and some in the neighbouring area. Many samples and wasters are exhibited in the Antiquarium (Amari 2006).

In the factory we have found several types of jugs. The type with circular mouth has a narrow neck, grooves around the shoulder, ribbed tape handle, and a discoid high foot (Fig. 4.7). The jugs with trilobate mouth have globular or ovoid ribbed-body, grooves around the shoulder, ribbed tape handle, and flat base (Fig. 4.8-9). An original production of Acium is reflected in the small three-handled jugs, with a more or less swollen and ribbed body and a high foot (Fig. 4.10). One sample has a mark under the foot (base 4.5cm), incised before firing it is an X inside a circle. This shape is attested in some sites (Pistunina and Cefalù) of the north coast of Sicily (Amari 2006, 146; Dr Paola Puppo, pers. comm.). Some single- and three-handled jugs of common ceramics are painted with drippings of black paint on the body. Other closed shapes were also found, namely round-mouthed jugs and jars decorated with impressed circles and medallions with figures (Amari 2006, 141-142). They seem to have been manufactured in this workshop but in the absence of wasters one cannot be certain; hopefully there will be sufficient evidence in the future in order to confirm their production at Acium.

In the factory we found situlae (Fig. 3.1) and basins with oblique walls, and everted, flat or horizontal rim (Fig. 3.2-5). There are samples undecorated or decorated with one or two wavy lines around the body, sometimes cut by a horizontal groove (Fig. 4.3). Some basins with everted lip have inside a grooved rim; other samples with a flat lip have horizontal handles. Fingerprints to attach the loops are seen inside and on top of the handles. The third type of basin with horizontal rim (more or less pronounced) presents undecorated and decorated samples, the latter with wavy lines and grooves impressed with the fingers on the wet clay of the rim. The bowls have curved or flared wall, with a low flat base (diameter range 4-6cm). The first type (Fig. 3.6) has rounded or slightly curved lip on the inside. The second type has straight walls often with rounded lip (diameter range 8.8-16cms and height 4.4-8.2cm) and sometimes a flat lip with a groove at the base to accommodate a lid, and a flat base (Fig. 3.7).

Conclusions In the factory the three circular kilns of vertical type were in use at the same time from the middle of the 4th century AD. This implies an activity of intensive production and also strong demand for these materials by the markets. Little is known about the economic and social context of Acium as a production site and its exchange relationships with the nearby settlements involving ceramic and agricultural products. We think that at Acium there was a market for oil, olives, grain, fruit and other food-stuffs of local provenance and from the countryside in the Acireale territory where a granary was uncovered at Casalotto and a storage room at Aciplatani. The new discoveries are very important regarding the function and role of Acium, a production site located at 14km from Catania, at the ‘cross-roads’ of four of the principal ways of the Roman road network on the island and placed along the ancient river Akis at 4km from the estuary in the important port of Capo Mulini frequented since the prehistoric times, as well as, the S. Venera al Pozzo area. The river route until the port was indicated especially for bricks and tiles, which were very heavy wares and hence hard to transport by land. This analysis focused on the features of the materials, and their distribution and commercialization in the surrounding area, in the short- and long-distance markets. Some shapes are rare, such as the three-handled jugs, and it is easier to follow their distribution, while others compare with regional (Agrigento, Caltanissetta, Capo D’Orlando, Naxos, etc.) or African productions and require more detailed petrographic analysis. It seems that in the 4th century the annexation of Sicily and Bruttium to the Vicariate of Italy Suburbicarian with the annonaria imposition of the canon vinarius contributed to the development and marketing of the amphorae Keay LII (Panella 2001, 180; Panella et al. 2010, 58-9). A further consequence must have been the establishment of the factory in Acium in the second decade of the 4th century

The lids have a flat knob and a flat lip curved slightly upwards, or a rounded lip, or a swollen rim with rib (Fig. 3.8-10). Large dish-lids have a hollow knob, a ribbed body, a flat rim and are made in a gritty clay (Fig. 3.11). In the factory were produced two types of funnels that were for the sale, one with a wide mouth (diameter range 27-15.5cm), and one with a more or less elongate cylindrical canal (Fig. 3.12-13). A few samples have waves impressed around the canal or the body. These objects were useful in processing, storing and marketing agricultural products. Similar artifacts were found on ships (e.g., Bass and van Doorninck 1971, 34, 36, fig. 30). We found many jars with ovoid body; some samples are partly ribbed, they have a high neck, one vertical handle, narrow foot with concave bottom, and they are undecorated or have one or two wavy lines impressed around the shoulder (Fig. 4.1-3). There are fragmentary samples of a second type of jars, neckless, with a swollen body with groove under the two horizontal handles, undecorated shoulder, or with one/two wide wavy lines around the groove, sometimes cut by a horizontal groove. Fingerprints to hold the loops are seen inside and on top of the handles. A small jar has swollen body, high ribbed neck, wide flat base, and decorated shoulder with one impressed wavy line (Fig. 4.4). The cups are rounded with everted lip, ribbed tape handle, and discoid foot (Fig. 4.5-6).

228

AMARI

and its prosperity during the next three decades, also useful for the sale of vina fiscalia. The workshop which was very close to the sea, ceased its activity due to the attacks by the Vandals in the first half of the 5th century. We do not know if production continued in another factory nearby. We are waiting for new financing that will allow to continue the research with archaeological excavations and acquire new data regarding local production of common wares, ceramic building materials and amphorae, the organisation of production and distribution of the products, a better understanding of local society and economy (with emphasis on the social condition of craftsmen in the factory and the organization of the work). The study of the late Roman coarse wares from S. Venera excavations needs to focus on the study of production sites within their immediate wider context in order to see how they were organised. The location of the rural settlement of Acium, a production site with a local market in a key place in the agricultural landscape, is strategic for the distribution and marketing of goods. This allows to formulate the hypothesis that Acium had an important role in the economical life of Sicily during the late Roman period with an organized trade in the broader area, in Rome and elsewhere wherever there were free markets. Important information derives also from the coins discovered in the figlina of Acium – along with the fact that they help dating the activities of the workshop. It is interesting to note that there are coins minted in Rome, Aquileia, the Balkans and the East Mediterannean (Siscia, Sirmium, Thessaloniki, Cyzicus, Bithynia [Heraclea, Nicomedia], Antioch and Alexandria). The coins and the African red slip ware are useful indicators of trade around the Mediterranean and at present the coin evidence suggests limited contact with the north-western part of the Mediterannean (e.g. at Acium there are only few Spanish amphorae). It seems certain that the amphora Keay LII, although present in many areas of the Mediterranean from the west to the east, was a suitable container for the Roman market. We know that from the 4th to the first half of the 5th centuries, about one-fifth of the supplies of wine came within Keay LII amphorae (Pacetti 1998, 191-2), small containers easily portable from the port to the final destination, that played an important role in the distribution of wine in small quantities. From the percentages of the recoveries in Rome, however, it emerges that the 'wine' contained in the MR1 was less 'required'. Perhaps it was a 'different' type of wine intended rather for free markets. The question remains open and a line of research is aimed at verifying the possible presence of amphorae and other products of Acium in sites located in the Mediterranean. Petrographic analysis in progress, comparison of data, new researches and the publication of the ceramic wares could be useful in identifying production and distribution of Acium foodstuffs and wares in the western and eastern Mediterranean.

current director of the Greek-Roman Archaeological Park in Catania, and the architect Carmelo Distefano, director of the Archaeological Park and Landscaping of the Valley of the Aci (where is S. Venera al Pozzo), for giving me permission to publish this paper with the drawings and photos. Giovanni La Scala produced the drawings and Giuseppe Barbagiovanni the photographs of the material. I would like to thank them. Bibliography Amari, S. 2006. I materiali in esposizione nell’Antiquarium. Sale I-II-III. In M. G. Branciforti (ed.), L’area archeologica di Santa Venera al Pozzo - Acium. Antiquarium, 105-183. Palermo. Amari, S. 2007. A late Roman pottery and brick factory in Sicily (Santa Venera al Pozzo). In S. Y. Waksman (ed.), Archaeometric and Archaeological Approaches to Ceramics. Papers presented at EMAC '05, 8th European Meeting on Ancient Ceramics, Lyon 2005. British Archaeological Reports, International Series 1691, 121-128. Oxford, BAR Publishing. Amari, S. 2008. A Late Roman Amphorae Production in Eastern Sicily. In O. Menozzi, M. L. Di Marzio and D. Fossataro (eds.), SOMA 2005. Proceedings of the IX Symposium on Mediterranean Archaeology, Chieti (Italy), 24-26 February 2005. British Archaeological Reports, International Series 1739, 473-479. Oxford, BAR Publishing. Amari, S. 2011. Acium. Le attività. In M. G. Branciforti (ed.), Natura e archeologia a Santa Venera al Pozzo, Progetto Scuola-Museo. Regione Siciliana, 44-65. Assessorato dei Beni Culturali e dell’Identità Siciliana, Palermo. Bass, G. F. and van Doorninck, Jr. F. H. 1971. A FourthCentury Shipwreck at Yassi Ada. American Journal of Archaeology, 75.1, 27-37. Bonifay, M. and Pieri, D. 1995. Amphores du Ve au VIIe s. à Marseille: nouvelles données sur la typologie et le contenu. Journal of Roman Archaeology 8, 94-120. Branciforti, M. G. (ed.) 2006. L’area archeologica di Santa Venera al Pozzo - Acium. Antiquarium. Palermo. Branciforti, M. G. and Amari, S. 2009. L’area archeologica di Santa Venera al Pozzo, brochure. Palermo. Capelli, C. and Bonifay, M. 2007, Archéométrie et archéologie des céramiques africaines: une approche pluridisciplinaire. In M. Bonifay and J. C. Tréglia (eds.), LRCW2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports, International Series 1662 (II), 551-68. Oxford, BAR Publishing.

Acknowledgements

Casalini, M and Crespi, M. 2010. Anfore tardoantiche di piccole dimensioni a fondo piatto dalle pendici nordorientali del Palatino. Nuovi dati alla luce di un riesame

I thank Maria Grazia Branciforti, archaeologist director of the archaeological researches at S. Venera al Pozzo, 229

LRCW4

tipologico e petrografico. Rei Cretariae Romanae Fautorum Acta 41, 101-111.

Panella, C., Saguì, L., Casalini, M., Coletti, F. 2010. Contesti tardoantichi di Roma: una rilettura alla luce di nuovi dati. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds.), LRCW3. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean Archaeology and Archaeometry. Comparison between western and eastern Mediterranean. British Archaeological Reports, International Series 2185 (I), 57-78. Oxford, BAR Publishing.

Menchelli, S., Capelli, C., Pasquinucci, M. and Picchi, G. 2007. Corsica tardo-antica: anfore italiche e ceramica comune da Mariana. In LRCW2, 313-28. Pacetti, F. 1998. La questione delle Keay LII nell'ambito della produzione anforica in Italia. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del convegno in onore di John W. Hayes, 185-208. Firenze, All’insegna del Giglio. Panella, C. 2001. Le anfore di età imperiale del Mediterraneo occidentale. In É. Gény (ed.), Céramiques hellénistiques et romaines III, 177-275. Paris.

230

AMARI

Class Amphora

Shape MR1a Keay LII

LRA 1 similis Ceramic Building Material

Curving tile

Imbrex Tegula Coarse Ware

Cap Situla Basin

Bowl-lid Lid

Large dish-lid Funnel Jar Small jar Mug Jug circular mouth Jug trilobate mouth Jug three-handles

No SV 1091 SV 1318 SV 1317 SV 927 SV1270 SV 1295 SV 677 SV 359 SV 349 SV 252 SV 701 SV 721

Dimensions in cms High 16.5. Lip 6.5 High 9.1. Lip 6.1/6.8 High max 8.6. Diameter 8.3 High 7. Lip 7.5/8.3. Hyperfired High 9.1. Lip 7 High 9.2. Lip 6.4/8.4 High 11. Lip 6.3/7.6 High 17 Diameter 8 High 35. Lip 10.7 High 18.5. Lip 7.2/9.2 High 12.5. Lip 9.2 High 7. Width 22. Length 32 max

Fig. 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.11 1.10 1.9 2.2

SV 949 SV1293 SV 732 In situ SV 1225 SV 58 SV 152

Side 9.2x6.4. Thick 1.9/2. Grooves distance 0.25 High 7. Width 17. Length 58 High 3/5. Width 48. Length 58 High 3/5.5. Width 48. Length 72 High 2.4. Diameter 8.8 High 31.6. Rim 30.4/26.6. Base 17.4 High 17.5. Rim 35.5. Base 17

2.3 2.1 2.1

SV 634 SV 1223 SV 697 SV 1124 SV 53 SV 1150 SV 1184

High 16.7. Rim 36. Base 19.5 High 12. Lip 28.5 High 9.7. Rim +40 High 6.5. Lip 14.8. Base 4.3 High 3.5. Lip 9.7. Foot 5.2 High 7.5 Lip. 17.6. Knob 4.1 High 7.3. Lip 12. Knob 4. Body bell, lip curved upward High 6.5. Lip 16.5. Knob 4.5 High 6.5. Rim 17.8. Knob 5.5. Rounded lip High 6.5. Rim 17. Knob 4 High 11.5. Lip 33.8. Knob 7.1 High. 20.5. Lip 20.5. Canal 4 High. 10.5. Lip 27. Canal 3.5 High 35.5. Lip 17.5. Foot 8.5 High 12. Lip 16.7 High 16.5. Rim 10.7. Base 11.9 High 12. Lip 6.4. Foot 4.2 High 9.4. Lip 6. Foot 3.8 High 19.5. Lip 4. Foot 4.3 High 21. Lip 7.1-7.6. Base 8.5 High 19. Lip 6.8-7.3. Base 8.4 High 14. Rim 8.8. Foot 3.3

SV 255 SV 105 SV 78 SV 1154 SV 1161 SV 1187 SV 254 SV 1157 SV 39 SV 1144 SV 1063 SV 1143 SV 1122 SV 1125 SV 107

Table of data

231

1.12 3.1 3.2 3.3 3.5 3.4 3.7 3.6 3.8 3.9 3.10 3.11 3.12 3.13 4.1 4.2 4.4 4.5 4.6 4.7 4.8 4.9 4.10

LRCW4

232

AMARI

233

LRCW4

234

DISTRIBUTION AND CONSUMPTION

ARCHÉOMÉTRIE ET ARCHÉOLOGIE DES CÉRAMIQUES AFRICAINES: UNE APPROCHE PLURIDISCIPLINAIRE, 2. NOUVELLES DONNEES SUR LA CERAMIQUE CULINAIRE ET LES AMPHORES. CLAUDIO CAPELLI1, MICHEL BONIFAY2 1

Università degli Studi di Genova, Dipartimento di Scienze della Terra, dell’Ambiente e della Vita, Corso Europa 26, 16132 Genova, Italia ; chercheur associé au Centre Camille Jullian; [email protected]

2

CNRS, Centre Camille Jullian, Maison Méditerranéenne des Sciences de l’Homme, 5 rue du Château de l’Horloge, BP 647, 13094 Aix-en-Provence, France; [email protected]

The poor lithological and sedimentological variability of the North African regions is responsible for the well-known difficulty in differentiating the African fabrics. However, a strictly integrated archaeological and petrographic approach undertaken in 1998 made it possible to get more precision in terms of classification, provenance and diffusion of some Late Roman African wares. In particular, in the last five years our collaboration has provided new data on the origin of the main classes of African pottery. First, different surveys have allowed a better characterisation of the different amphora workshops in the Sahel and Nabeul regions along the Tunisian coastline, pointing out the distinguishing petrographic and/or typological markers. Secondly, new productions of cooking ware have been identified in poorly explored regions, such as eastern Tripolitania and central Algeria. Lastly, more amphora types have been characterised from an integrated point of view, which allowed the formulation of new hypotheses about their origin. Further results are now expected from the development both of a detailed geological cartography and the field survey of pottery workshops in Africa. KEYWORDS: NORTH AFRICA, ARCHAEOMETRY, THIN-SECTION ANALYSIS, TYPOLOGICAL STUDIES, WORKSHOPS, PROVENANCE AND DISTRIBUTION, COOKING WARE, AMPHORAE au dehors de la Tunisie côtière et en conduisant des recherches sur plusieurs sites de consommation africains et méditerranéens.

1. Introduction: nouvelles études, nouveaux territoires Dans une première contribution, publiée dans les actes du colloque LRCW2 (Capelli et Bonifay 2007), on a pu montrer qu’il était possible de distinguer différentes productions d’Afrique du Nord en associant étroitement la démarche archéologique aux investigations archéométriques, notamment pétrographiques dans notre cas. Certes, cette démarche est rendue extrêmement difficile du fait de la généricité de la plupart des pâtes et des sédiments africains, notamment tunisiens, composés principalement de grains de quartz (les plus gros, dont les dimensions atteignent généralement quelques dixièmes de mm, arrondis et aux surfaces opaques, montrent des caractéristiques typiquement éoliennes) et de composantes accessoires génériques (feldspaths, microfossiles, calcaires, grès, argilites). Cependant, c’est la possibilité de pouvoir prospecter quelques ateliers tunisiens et caractériser leurs productions d’un point de vue intégré, typologique et pétrographique, qui a permis de mettre en évidence les éléments discriminants de ces ateliers et aussi de comprendre, en général, que l’on peut arriver à différencier les productions africaines en prenant en compte un ensemble de caractéristiques texturales et compositionnelles des pâtes, techniques et typologiques.

Une conséquence de nos études est la réalisation d’une importante base de données de céramiques africaines d’époque romaine, qui est composée à présent d’environ 3,000 lames minces et beaucoup plus d’échantillons macroscopiques. On présentera synthétiquement ici quelques résultats de nos recherches récentes, menées souvent en collaboration avec des archéologues des deux rives de la Méditerranée, dans le cadre de missions archéologiques de fouilles ou de prospections, de projets de recherche collectifs ou de l’encadrement de thèses de doctorats (cf. infra, remerciements). 2. Géologie et typologie africaines et non-africaines Si la définition géopolitique de l’Afrique est assez claire, tout au moins pour l’époque romaine tardive (Fig. 1a), cantonnée à l’Ouest par la frontière entre la Maurétanie tingitane (rattachée depuis le début du IVe s. au diocèse d’Espagne), et à l’Est par les autels des Philènes, dans le golfe de Syrte, ligne de partage traditionnelle entre la pars occidentalis et la pars orientis de l’Empire, il n’en est pas de même de sa définition géologique.

Dans les dernières années, on a pu valider cette méthode et améliorer les connaissances en augmentant le nombre d’ateliers caractérisés, en étendant les aires d’investigation à d’autres régions productrices africaines 235

LRCW4

du type précoce Dressel 26: Sicile ou Afrique?). Toutefois, d’autres productions siciliennes peuvent être bien distinguées des productions africaines en raison de la présence d’inclusions diagnostiques. C’est le cas, déjà évoqué précédemment (Capelli et Bonifay 2007, 554) des amphores de type Benghazi Mid Roman 1, riches en inclusions volcaniques, métamorphiques et sédimentaires, produites dans le nord-est de la Sicile, même si des imitations sont connues en Afrique (Salakta et Tripoli: Capelli et Bonifay 2007, 554, pâtes riches en quartz) et en Cyrénaïque (Mazou et Capelli 2011: pâtes riches en éléments calcaires).

De ce point de vue (Fig. 1c), l'Afrique est divisée en deux domaines majeurs séparés par la ligne sud-atlasique (South Atlassic Front): 1) le domaine des chaînes, portant l'empreinte de la tectonique alpine, et 2) le domaine de la plate-forme saharienne, avec une empreinte tectonique moins prononcée. Le domaine des chaînes, constitué de reliefs modelés au cours du Cénozoïque, est composée par les ensembles tectono-sédimentaires suivants, du nord au sud: l'Atlas tellien, les Hauts plateaux (High plateaux) et l'Atlas (Tunisian Atlas, Saharan Atlas, High Atlas). La plate-forme saharienne (Saharan Platform), qui appartient au Craton Nord Africain, est constituée d’un socle précambrien sur lequel repose une puissante couverture sédimentaire phanérozoïque (Guiraud et al. 2005 et bibliographie citée). Toutefois, ce domaine rentre seulement partiellement dans l'Afrique romaine, principalement dans sa partie sud-orientale.

Parfois, c’est la pétrographie qui prime même lorsque la typologie est typiquement africaine. C’est le cas des nombreuses imitations, parfois très fidèles, de céramiques culinaires africaines en Espagne (Aquilué 2008) et en Gaule du Sud (Pellecuer et Pomarèdes 1991; Mauné 1996: «Brune-Orangée Biterroise»), qu’il est aisé de différencier de leurs modèles africains par un simple examen macroscopique de la pâte, comportant des inclusions diverses peu compatibles avec les pâtes africaines (par exemple du mica abondant dans la production gauloise).

Même si les roches sédimentaires sont dominantes (notamment sur la côte tunisienne), on trouve également des secteurs caractérisés par la présence de roches métamorphiques (par exemple l'Algérie du Nord) et, plus rarement, volcaniques (par exemple le Massif du Gharian, en Libye). Il faut aussi souligner que l'on peut trouver des roches et sédiments similaires à ceux de l’Afrique du Nord dans d'autres secteurs méditerranéens qui ont subi la même histoire géologique. Par exemple, l'Algérie du Nord, la Tunisie du Nord et la Sicile centro-septentrionale font partie de la chaîne Maghrébide, qui est liée aux chaînes Cénozoïques de la Méditerranée occidentale; les plateaux de l’Algérie du Sud, de la Tunisie orientale et de la Sicile du Sud font partie de l’avant-pays alpin; l’Algérie méridionale et la Tunisie centrale et méridionale sont constituées par les chaînes atlasiques (Guiraud et al. 2005 et bibliographie citée).

Parfois, enfin, les timbres certifient l’origine africaine d’une production que la pâte ne permet pas de reconnaître avec assurance comme telle. C’est le cas des amphores de Tubusuctu, bien identifiées par l’épigraphie (Laporte 2010) mais dont la pâte est caractérisée non pas par du dégraissant quartzeux éolien mais par des inclusions sédimentaires génériques (Capelli et Bonifay 2007, 555). La difficulté subsiste pour d’autres types d’amphores qui ne portent pas de timbre, comme celles du type Keay 1B, dont la pâte est riche en éléments volcaniques et métamorphiques qui auraient pu tout aussi bien indiquer une provenance de Sicile nord-orientale. Dans ce cas, c’est principalement la vraisemblance archéologique qui nous fait préférer l’hypothèse d’une provenance algérienne. Enfin, l’hypothèse d’une localisation à Cherchell (Ben Abed et al. 1999) de la production des amphores du type «Station 48 du Piazzale delle Corporazioni», avec le timbre M-palmier-C et une pâte riche en quartz éolien compatible avec les pâtes tunisiennes et très différente des précédentes, n’a toujours pas reçu de confirmation.

On voit que des sédiments et des roches - et par conséquent des pâtes - similaires se trouvent dans des régions africaines et non-africaines très distantes les unes des autres. Il est donc difficile, même dans le cas où l’on trouve des inclusions appartenant à des roches bien caractérisables du point de vue pétrographique, de reconnaître la provenance des céramiques sans l’apport des données typologiques. Parfois, la typologie nous aide à distinguer des productions non-africaines parmi des céramiques dotées d’une pâte typiquement africaine (éolienne). C’est le cas des amphores de type hispanique Dressel 7/11 produites dans quelques ateliers de la région de Ceuta (Capelli et al. 2013) ou encore des amphores de type Dressel 21/22 produites en Sicile nord-occidentale, notamment sur l’atelier d’Alcamo Marina (Capelli et Piazza 2006; Botte 2009).

3. Définition des ateliers d’amphores: plus de données, plus de complexité La poursuite des caractérisations d’ateliers au gré de leur découverte, lors de visites de sites ou de programmes de recherches en Tunisie, dessine un panorama plus précis mais également plus complexe de la production amphorique africaine. En effet, la poursuite des analyses confirme que chaque zone de production peut être distinguée des autres par ses caractéristiques pétrographiques et typologiques. En outre, dans quelques zones où les prospections ont été plus poussées, une certaine variabilité des pâtes et de la typologie permet même, dans quelques cas, d’arriver à l’identification précise des ateliers. Nous prendrons trois exemples.

Le second exemple est particulièrement important dans la mesure où plusieurs parties de la Sicile centroseptentrionale possèdent des affleurements de roches et de sédiments éoliens avec les mêmes caractéristiques que ceux du nord de la Tunisie (notamment le Flysch Numidien; Thomas et al. 2010). Ainsi, dans l’attente de nouvelles analyses, on ne peut exclure une provenance sicilienne pour un certain nombre de types d’amphores et de céramiques communes habituellement attribués à l’Afrique (on citera en particulier le problème de l’origine 236

CAPELLI-BONIFAY

similaire, même si normalement les pourcentages d’inclusions sont très différents. On pourrait justifier ces ressemblances par l’utilisation des mêmes types de matières premières régionales (argiles pliocéniques), ce qui peut expliquer que, parfois, il est difficile de distinguer ces deux ateliers par le simple examen des pâtes; la distinction typologique reste dans ce cas déterminante. Enfin, pour les mêmes raisons, la zone d’atelier de Teboulba (Nacef dans le présent volume) pourrait correspondre à certaines pâtes partiellement similaires à celles de Moknine (notamment par les caractéristiques des microfossiles; dans ce cas, les deux centres sont peu distants) mises en évidence au sein des amphores Keay 8A de S. Antonino (Gandolfi et al. 2010, 38). Ici, la typologie sera d’un secours moins évident car le répertoire de formes (comprenant également des Keay 61) est partiellement comparable à celui de Moknine. De même, il convient de rappeler que la forme Keay 8A n’est pas seulement produite à Teboulba mais également dans les ateliers suburbains de Leptiminus (Dore 2001).

3.1. Les ateliers du Sahel tunisien A la différence de ce qu’on a pu penser précédemment, on sait maintenant que le Sahel n’est pas caractérisé par un seul type de pâte homogène mais par plusieurs types de pâtes qui correspondent à différentes zones de productions. En conséquence, afin de confirmer l’origine sahélienne d’une amphore trouvée sur un site de consommation, il est nécessaire d’identifier précisément le centre producteur. Quatre sites producteurs sont désormais bien caractérisés par la typologie et la pétrographie (Fig. 2). Les ateliers littoraux situés dans la périphérie des grandes villes portuaires de Byzacène, ne posent plus, et depuis longtemps (Capelli 2002-2003; Capelli et al. 2006), de problème d’identification: c’est le cas des ateliers de Salakta et de Leptiminus, actifs principalement jusque dans le courant de la première moitié du Ve siècle. Rappelons que les productions de Salakta sont caractérisées par une pâte rouge orangée ou plus souvent bicolore rouge et grise, riche en inclusions blanches ; au microscope, ce dégraissant est constitué par un sable bien classé composé de calcaire, de microfossiles, de quartz et d’une quantité accessoire de feldspath, de pyroxène et de fragments de roches basaltiques. Celles de Leptiminus se distinguent par une matrice orange clair et des inclusions abondantes composées principalement de quartz bien arrondi et bien ou moyennement classé.

Pour ce qui concerne l’aspect technique, il faut souligner que la texture des pâtes des deux ateliers côtiers de Leptiminus (Fig. 2b, n° 4) et Salakta (Fig. 2b, n° 1), par ailleurs plus anciens, suggère l’utilisation d’un dégraissant sableux (probablement prélevé dans de sédiments quaternaires d’origine marine) sélectionné, tandis que les pâtes des autres deux ateliers, plus à intérieur des terres et tardifs, ne montrent pas d’évidences sûres d’un mélange de composantes argileuses et sableuses différentes.

Les ateliers tardifs (Ve-VIIe s.) caractérisés plus récemment, comme ceux de Henchir Chekaf (Nacef 2007, annexe C. Capelli) et de Moknine (Gandolfi et al. 2010), ou encore celui récemment mis en évidence à Teboulba par J. Nacef (voir sa contribution dans le présent volume) sont parfois moins faciles à distinguer les uns des autres. L’atelier de Henchir Chekaf a produit plusieurs variantes d’amphores Keay 61/62 (Nacef 2007: types I, II et III) ainsi que des imitations d’amphores orientales LRA 1 (Nacef 2007: type IV). Si on s’en tient aux exemplaires les plus caractéristiques, la pâte de cet atelier (Fig. 2b, n° 2) est tout à fait différente de celles des ateliers de Salakta (Fig. 2b, n° 1), pourtant peu éloignés: la matrice est riche en fer, macroscopiquement rouge à nuances violacées, et les inclusions, peu classées et jamais très abondantes, sont composées principalement de quartz rarement arrondi, associé à des microfossiles fréquemment dissociés par la cuisson et peu visibles (parmi lesquels des radioles d’échinoïdes) et à de rares fragments de calcaire. La zone d’ateliers de Moknine, fouillée dans les années 90 par N. Ben Lazreg, est à l’origine d’une grande part des amphores Keay 61A/D importées en Gaule du Sud et en Ligurie (Gandolfi et al. 2010). La pâte typique de ces ateliers (Fig. 2b, n° 3) montre une matrice partiellement calcaire (macroscopiquement orange), des inclusions abondantes de quartz et de microfossiles particuliers (foraminifères, radioles d’echinoïdes, fragments de coquilles de bivalves) et souvent de gros nodules (pouvant atteindre quelques mm) bien évidents à l’œil nu, composés de calcaires fossilifères. Bien que ces deux ateliers soient très éloignés l’un de l’autre, la composante fossilifère de quelques pâtes d’Henchir Chekaf et de Moknine est partiellement

3.2. Les ateliers de Nabeul L’avancement des connaissances sur les ateliers de Nabeul font de cette zone productive l’une des mieux documentées de Tunisie. On connaissait la présence d’ateliers d’amphores à Nabeul depuis la proposition faite par Cl. Panella de développer le timbre C.I.N, attesté sur les amphores Africaines II C, en C(olonia) I(ulia) N(eapolis). Mais les premières découvertes d’ateliers datent de la fin des années 1990, avec la mise en évidence de celui de Sidi Zahruni, à l’origine d’une production de masse aux Ve-VIIe s. des types Keay 25.2, spatheion 1, Keay 35, 55, 56, 57, 62A, etc., et de celui de Sidi Aoun, plus ancien, avec une production d’amphores Africaines I, II (notamment II C) et III (Keay 25.1, 2 et 3), du IIe-au début du Ve s. (Ghalia et al. 2005). Stimulée par cette découverte, une équipe tunisienne dirigée par A. Mrabet et M. Ben Moussa a multiplié, dans les années 2000, le nombre d’ateliers connus dans une aire de 75km2 et permis de retrouver le timbre C.I.N, notamment sur l’atelier de Sidi Aoun (Mrabet et Ben Moussa 2007). Enfin, une quatrième phase de recherches a permis, en 2008, d’opérer la caractérisation pétrographique de neuf des dix ateliers précédemment reconnus (Bonifay et al. 2010a). Alors même que les recherches continuent, menées par M. Ben Moussa, l’archéométrie fait clairement apparaître trois zones géologiques/géographiques et l’archéologie

237

LRCW4

ateliers à caractéristiques très différentes, comme Salakta et Henchir Chekaf, se trouvent implantées sur la même formation géologique du Pliocène, et qu’à l’inverse d'autres ateliers à caractéristiques similaires, comme certains de Nabeul, sont localisés dans deux formations différentes, du Pliocène et du Miocène-Pliocène.

deux zones chronologiques/ géographiques (IIe-Ve s. et Ve s.-VIIe s.) Ffig. 3a). Tout d’abord, les nouvelles analyses pétrographiques confirment que les productions de Nabeul ont des caractéristiques communes (typologiques et pétrographiques, macroscopiquement une couleur généralement orange vif) qui permettent de les différencier des autres principales productions étudiées (ex.: Salakta, Leptiminus). Elles peuvent toutefois être divisées, principalement sur la base des différences de texture, en trois ensembles correspondant à trois zones géographiques. Les pâtes des deux zones principales A et B (Fig. 3b, n° 1 et 2), sont bien caractérisées par de nombreuses inclusions fines de quartz et de microfossiles, avec des grains de quartz éolien majeurs et des fragments de grès à quartz accessoires. Toutefois, les pâtes des deux zones et des divers ateliers sont distinguées entre eux par des différences secondaires dans les caractéristiques texturales et les rapports relatifs entre les différentes composantes. Plus différente est la pâte de l’atelier de la zone C (Fig. 3b, n° 3), qui est caractérisée par une matrice plus pure et des inclusions de quartz bien classées. Selon la Carte Géologique de Tunisie, les trois zones A-B-C correspondent, mais pas exactement, à trois formations géologiques différentes, d’âge respectivement pliocènique, mio-pliocènique et quaternaire.

De toute évidence, l’argile utilisée par les potiers n’est pas toujours celle qui affleure (ou telle qu’elle est cartographiée en surface sur les cartes). Ainsi sur l’île de Jerba, où des affleurements d’argiles mio-pliocèniques sont cartographiés à l’emplacement même des ateliers actuels de Gallala, on sait que les potiers traditionnels creusent de profondes galeries sous les niveaux quaternaires afin d’atteindre les couches les mieux adaptées à leur travail (Combès et Louis 1967). 4. Classification et typologie: avancées sur les céramiques culinaires Pour la Tunisie, le modèle présenté précédemment (Capelli et Bonifay 2007, 553), distinguant d’une part les productions de la zone de Carthage (culinaire A et C), d’autre part celles de Byzacène (culinaire B) a été confirmé dans ses grandes lignes par la poursuite des analyses, notamment sur les sites consommateurs de Méditerranée occidentale. Toutefois, le choix d’un plus grand nombre d’échantillons sur certains ateliers déjà connus, comme celui de Leptiminus, a permis de progresser dans la caractérisation de leurs différentes productions. De petites disparités apparaissent en fonction de la typologie ou de la sous-catégorie (culinaires engobées et non-engobées) (Leitch 2010, annexe C. Capelli) mais l’ensemble demeure assez homogène pour être distingué des autres zones de production (par exemple Sidi Khalifa et Thaenae). Des productions régionales moins diffusées, comme celle de la région de Sidi Jdidi, ont également été caractérisées, une partie pouvant être rattachée à la production tardive de l’atelier de Sidi Khalifa ou d’ateliers voisins (Mukai 2010, annexe C. Capelli). Mais les résultats les plus novateurs de ces dernières années concernent des productions non-tunisiennes: les céramiques culinaires de Lepcis Magna (Libye), d’une part, et celles de quelques sites de la partie occidentale de l’Algérie, d’autre part.

D’autre part, la typologie des productions permet de regrouper tous les ateliers de la zone A et ceux de la zone B – sauf Sidi Zahruni – dans une même phase chronologique comprise entre le IIe s. et la première moitié du Ve s. (types Africaine I, II, III, spatheion 1, Dressel 30), tandis que l’atelier de Sidi Zahruni et l’atelier de la zone C (Labayed) sont datables des VeVIIe s. (types Africaine III C, III, spatheion 1, Keay 35, 55-57, 62, etc.). Le fait que les deux zonages (géologique et chronologique) ne se superposent pas permet d’arriver, dans certains cas, à la précision de l’atelier (Exemple: Barnoussa et Sidi Zahruni: pâte semblable mais typologie différente). On se trouve donc ici approximativement dans la même situation que dans le Sahel, avec un déplacement (et un regroupement?) des ateliers dans le courant de la première moitié du Ve s., impliquant des changements dans les lieux d’approvisionnement des matières premières. Ces modifications sont majeures dans le cas du Sahel (autre formation géologique, autre technique), mineures dans le cas de Nabeul. Un troisième changement se produira à Nabeul au courant du VIIe s. avec un nouveau déplacement (partiel?) des ateliers, cette fois vers l’intérieur du périmètre urbain (Slim et al. 2002), perceptible dans le changement radical des argiles qui deviennent calcaires (pâte claire), tout en restant pliocèniques, et non plus ferriques (pâtes rouges). Ces ateliers ont notamment produit les spatheia miniatures de type 3C (Bonifay 2004a, 39 et fig. 19).

4.1. La céramique culinaire de Lepcis Magna L’étude du mobilier issu des fouilles de la Mission Française en Libye dans les Thermes du Levant à Lepcis Magna (dirigées jusqu’en 2010 par A. Laronde†) a permis d’identifier quatre grandes catégories de céramiques culinaires régionales: deux sont proches morphologiquement des productions non engobées (à surface extérieure grise) de Byzacène méridionale (forme Hayes 183), tandis que la troisième possède des formes plus originales et une pâte présentant des similitudes avec certaines amphores locales (Tripolitaines II et III) (Bonifay et Capelli 2013). Mais nous souhaiterions ici

Enfin, pour ces deux zones d’ateliers (Sahel et Nabeul), il convient de signaler les difficultés d’utilisation des cartes géologiques régionales à petite échelle (notamment la Carte géologique de la Tunisie, éch. 1:500.000). Sans vérification sur le terrain, on risquerait de déduire que des 238

CAPELLI-BONIFAY

mettre l’accent sur la quatrième production, dénommée provisoirement «culinaire B/T» car elle s’apparente à la fois à la culinaire B et aux productions tripolitaines que l’on vient d’énumérer.

la région de Zana/Diana Veteranorum, celle des Hautes Plaines entre Sétif et Batna (Filah 1986), permettent de mettre en évidence certaines particularités des productions de l’Est algérien.

La céramique culinaire B/T de Lepcis Magna livre un répertoire de formes composé presque uniquement de faitouts (proches du type Hayes 183) et de couvercles à lèvre en demi-lune (proches du type Sabratha 104) (Fig. 4b, n° 1-2). Elle apparaît dans les niveaux de la fin du IIIe s. ou du début du IVe s. et reste présente jusqu’au début du Ve s. au moins. Il s’agit d’une production non engobée mais très soignée, cuite à température moyenne ou élevée, avec une pâte orange homogène et une surface de même couleur soigneusement lissée qui la rapproche des culinaires engobées de Byzacène. L’analyse en lame mince (Fig. 4d) montre une matrice argileuse pure riche en fer et des inclusions composées de rares grains de quartz éolien plus grossier et d’abondants grains fins (