Roman Amphora Contents: Reflecting on the Maritime Trade of Foodstuffs in Antiquity (In honour of Miguel Beltrán Lloris): Proceedings of the Roman Amphora Contents International Interactive Conference (RACIIC) (Cadiz, 5-7 October 2015) 1803270624, 9781803270623

Citation preview

Roman Amphora Contents Reflecting on the Maritime Trade of Foodstuffs in Antiquity In honour of Miguel Beltrán Lloris Proceedings of the Roman Amphora Contents International Interactive Conference (RACIIC) (Cadiz, 5-7 October 2015)

Edited by

Darío Bernal-Casasola, Michel Bonifay, Alessandra Pecci and Victoria Leitch

Roman and Late Antique Mediterranean Pottery 17 2021

Roman and Late Antique Mediterranean Pottery Archaeopress Series EDITORIAL BOARD (in alphabetical order)

Series Editors Michel BONIFAY, Centre Camille Jullian, (Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France) Miguel Ángel CAU, Institució Catalana de Recerca i Estudis Avançats (ICREA)/Equip de Recerca Arqueològica i Arqueomètrica, Universitat de Barcelona (ERAAUB) Paul REYNOLDS, Institució Catalana de Recerca i Estudis Avançats (ICREA)/Equip de Recerca Arqueològica i Arqueomètrica, Universitat de Barcelona (ERAAUB) Associate editors Philip KENRICK, Institute of Archaeology, University of Oxford John LUND, The National Museum of Denmark, Denmark Scientific Committee for Pottery Xavier AQUILUÉ, Paul ARTHUR, Cécile BATIGNE, Moncef BEN MOUSSA, Darío BERNAL, Raymond BRULET, Claudio CAPELLI, Armand DESBAT, Nalan FIRAT, Michael G. FULFORD, Ioannis ILIOPOULOS, Sabine LADSTÄTTER, Fanette LAUBENHEIMER, Mark LAWALL, Sévérine LEMAÎTRE, Hassan LIMANE, Daniele MALFITANA, Archer MARTIN, Thierry MARTIN, Simonetta MENCHELLI, Henryk MEYZA, Giuseppe MONTANA, Rui MORAIS, Gloria OLCESE, Carlo PAVOLINI, Theodore PEÑA, Verena PERKO, Platon PETRIDIS, Dominique PIERI, Jeroen POBLOME, Natalia POULOU, Albert RIBERA, Lucien RIVET, Lucia SAGUI, Sara SANTORO†, Anne SCHMITT, Gerwulf SCHNEIDER, Kathleen SLANE, Roberta TOMBER, Inês VAZ PINTO, Caterina VIEGAS, Yona WAKSMAN General advisors Richard HODGES, Richard REECE, Gisela RIPOLL, Bryan WARD-PERKINS, Chris WICKHAM, Enrico ZANINI

Organizers:

With the collaboration of:

Roman Amphora Contents Reflecting on the Maritime Trade of Foodstuffs in Antiquity In honour of Miguel Beltrán Lloris Proceedings of the Roman Amphora Contents International Interactive Conference (RACIIC) (Cadiz, 5-7 October 2015)

Edited by

Darío Bernal-Casasola*, Michel Bonifay**, Alessandra Pecci*** and Victoria Leitch**** * Departamento de Historia, Geografía y Filosofía / Universidad de Cádiz, Cadiz, Spain ** Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France *** ERAAUB /Universitat de Barcelona, Barcelona, Spain

**** Durham University / Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France

Roman and Late Antique Mediterranean Pottery 17

Archaeopress Publishing Ltd Summertown Pavilion 18-24 Middle Way Summertown Oxford OX2 7LG www.archaeopress.com

ISBN 978-1-80327-062-3 (print edition) ISBN 978-1-80327-063-0 (e-pdf edition)

© Authors and Archaeopress 2021 Cover: Amphora from the Garum Shop, Pompeii Back cover: Bou Ferrer wreck amphorae (Left, J. A. Moya); Graeco–Italic amphora, Museo de Zaragoza (Right, J. Garrido Lapeña)

All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of the copyright owners. This book is available direct from Archaeopress or from our website www.archaeopress.com

Contents Preface .....................................................................................................................................................................  1 Ánforas romanas y contenidos. Notas historiográficas ..............................................................................................  7 Miguel Beltrán Lloris

Historical and archaeological indicators  Amphorae: typology and contents ..........................................................................................................................  31 Stefanie Martin‑Kilcher and André Tchernia Shipwrecks, amphorae and contents .......................................................................................................................  41 Franca Cibecchini Amphora contents as commodities: the structure and function oftituli picti in the western Mediterranean in the 1st century AD ..............................................................................................................................................  51 Enrique García Vargas How late antique dipinti contribute to a better knowledge of amphora contents ....................................................  63 Jean-Luc Fournet The reuse of transport amphorae as packaging containersin the Roman world: an overview ..................................  77 J. Theodore Peña

Archaeometric indicators: generalities and case studies  Amphorae and residue analysis: content of amphorae and organic coatings .......................................................... 113 Nicolas Garnier and Alessandra Pecci GC-MS analysis of pitch from Roman amphorae from Cosa in Etruria (Italy) ........................................................... 127 Hitomi Fujii, Carole Mathe, Fabienne Olmer and Cathy Vieillescazes Residue analysis by GC-MS and FT-IR Spectroscopy on Roman amphorae from the archaeological site ‘Nuovo Mercato Testaccio’ (Rome) ..................................................................................................................................... 133 Florinda Notarstefano and Mariateresa Lettieri The contents of ancient Graeco-Italic amphorae.First analyses on the amphorae of the Filicudi F andSecca di Capistello wrecks (Aeolian Islands, Sicily) ............................................................................................................... 141 Nicolas Garnier and Gloria Olcese Inland trade and consumption in context. A case study on the organic residue analysis of transport amphorae from the Balkan Peninsula (Yambol District, South-eastern Bulgaria) ............................................................................. 149 Silvia Polla, Andreas Springer, Birte Gruber, Petra Tušlová and Barbora Weissová The Beirut amphora: residue analysis and contents ............................................................................................... 161 Marshall Woodworth and Paul Reynolds Making Garum. Experimental archaeology methods .............................................................................................. 171 Álvaro Rodríguez-Alcántara, Ana Roldán-Gómez, Enrique García Vargas, Darío Bernal-Casasola and Víctor Manuel Palacios-Macías

Regional and provincial syntheses  West Lusitanian Amphora Contents ................................................................................................................................ 183 Inês Vaz Pinto, Rui Morais, Carlos Fabião, César Oliveira and Sónia Gabriel The Iberian and Roman amphorae of Hispania Citerior .......................................................................................... 205 Ramón Járrega Domínguez and Albert Ribera i Lacomba Amphora contents in Baetica: from thePunic tradition to Late Roman times .......................................................... 215 Darío Bernal-Casasola, Enrique García Vargas, Antonio M. Sáez Romero and Horacio González Cesteros Tituli picti on Spanish amphorae ............................................................................................................................ 241 Piero Berni and Emmanuel Botte The contents of amphorae produced in Gaul in the Imperial period ....................................................................... 249 Fanette Laubenheimer Italian and Sicilian amphorae and their contents: a general overview .................................................................... 259 Simonetta Menchelli The content of amphorae from Adriatic Italy ......................................................................................................... 273 Marie Brigitte Carre and Stefania Pesavento Mattioli African amphora contents: an update .................................................................................................................... 281 Michel Bonifay Amphorae from the Byzacene coast: what did they contain? ................................................................................. 299 Jihen Nacef†

East The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire ..  307 Paul Reynolds Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC‑3rd century AD) ............................................................................................................................. 355 Andrei Opaiţ Ephesus wines ....................................................................................................................................................... 377 Tamás Bezeczky† Egyptian amphorae from the Hellenistic to the Late Roman periods ...................................................................... 389 Delphine Dixneuf

Case studies: From the West to the East  The contents of amphora type T-7.4.3.3 (former type Mañá C2b): ancient problematic and new research perspectives .............................................................................................. 401 Max Luaces The fish-salting production centre of Águilas:late-Roman amphora content analysis .............................................. 409 Alejandro Quevedo, Myriam Sternberg and Juan de Dios Hernández García New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts: salsamenta, garum, lymphatum and other fish products ........................................................................................ 419 Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii ........ 437 Darío Bernal-Casasola, Daniela Cottica, Ricard MarlascaCarmen Gloria Rodríguez Santana and Enrique García Vargas

Amphorae with residues from Southern Sardinia (Cagliari and Nora) ..................................................................... 453 Ignazio Sanna, Laura Soro and Cristina Nervi ‘De profundis’: three amphorae of unorthodox contents retrieved from the Aegean Sea ....................................... 473 George Koutsouflakis Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence from the Punic Amphora Building (Corinth, Greece) ...................................................................................................................... 485 Antonio M. Sáez Romero and Tatiana Theodoropoulou Conclusion ............................................................................................................................................................. 499

Preface It is not possible, from a heuristic perspective, to study Roman amphorae without paying attention to their paleocontents. This is a common metonymic association in archaeology: what archaeologists find is partially the result of time, not the direct trace of what past societies left.

•

Amphoric studies are a mature discipline today, long after Heinrich Dressel reported his exceptional discoveries in Castro Pretorio in the 19th century. Beginning in the 1970s, periodical meetings, conferences and seminars have been held, especially in Europe, to update our knowledge about this primary indicator of the economy and maritime trade in Antiquity. To mention but one classical example, in the introduction of the famous meeting at Siena in 1986 (Amphores romaines et histoire économique. Dix ans de recherche, Collection de l’École française de Rome, 114), the proceedings of which were published three years later, Maurice Lenoir, Daniele Manacorda and Clementina Panella wrote about the importance of ‘considérer l’étude des amphores comme source pour l’histoire économique du monde romain, aussi bien sur le plan de la production que sur celui de la commercialisation. Le renouvellement de nos connaissances permettait, semblait-il, de mettre en parallèle, d’une façon elle aussi renouvelée, documentation archéologique et histoire, économie et société’.

Three institutions with a long tradition in these subjects (Centre Camille Jullian/CNRS and Aix-Marseille Université, ERAAUB – Universitat de Barcelona and Universidad de Cádiz) took the lead organising the RACIIC (Roman Amphora Contents International Interactive Conference. Reflecting on Maritime Trade of food stuffs in Antiquity), held at the Facultad de Filosofía y Letras, University of Cádiz, 5-7 October 2015 (Figure 1).

Thirty years later, the organisers of the event that led to this publication followed very similar premises to the study of amphorae contents (nihil novum sub sole…). The ‘call for papers’, published in early 2015, mentioned that no consensus existed concerning paleocontents with regard to various production areas and typological families, resulting in erroneous characterisations and hampering interpretation. In addition, it was stressed, the two preceding decades had been rich in methodological advances in the field of archaeometry, which can contribute to the identification of organic residues, and many analysis have in fact been carried out on Atlantic and Mediterranean sites, but the coverage afforded by these techniques is still far from systematic. Despite significant advances, information remains dispersed and in need of a holistic perspective. With this in mind, we decided to organise an international conference in Cádiz with the support of the Campus de Excelencia Internacional del Mar – CEIMAR (https://campusdelmar. com). The conference’s main target was threefold: •

•

To reflect on the inherent methodological challenges posed by the products traded in amphorae in Antiquity, and on their identification by historical, archaeological and archaeometric means. To update and compile existing knowledge about paleocontents in Roman amphorae in different regions and provinces.

To generate synergies for the creation of an international and interdisciplinary group for the submission of a research project within the framework of the European Union’s research strategy, following the model of the previous work group CORONAM.

The RACIIC adopted a singular format. Contributions were in English, to stress the markedly international nature of the meeting, and aspired to stimulate debate; hence the reference to ‘interactivity’ in its very title (Figure 2). There were a number of keynote lectures by several prestigious scholars, who presented stateof-the-art syntheses, which were followed by lengthy debates with the participation of researchers working in multiple Atlantic and Mediterranean regions. One of the conference’s main targets was thus to encourage debate and seek points of contact, so the sessions were chaired by international specialists in various disciplines (archaeology, archaeometry, archaeozoology, historians, specialists in chemistry, etc.); methodological and historical issues concerning the content of amphorae were examined; protocols, methodological advances and past errors were discussed, and a programme of research for the future outlined. The agreements reached and the areas in which controversies remained were synthetized at the end of the event. As is customary in archaeological conferences, and in order to encourage participation and debate, poster presentations were scheduled and discussed in one session. Many of the specialists that attended the RACIIC were interviewed, allowing those who could not come to ‘put a face’ to researchers from different countries: these 60 micro-interviews, as well as footage of the conference, what we called ‘sea-pills’, are available in CEIMAR’s Youtube channel (https://www.youtube.com/c/ CampusdelmarEsceimar/videos). Five years later, we have become used to these dynamic multimedia formats, especially as they have greatly contributed to minimise the effects of the COVID‑19 pandemic. At the time, it was regarded as a novel and valuable format by the attendants and the scientific community at large. For logistic reasons, attendance was limited to one hundred participants, from 12 countries and over 30 institutions and businesses (Figure 3): Germany

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 1-5

Preface

Figure 1. RACIIC conference poster, which illustrates the contents of a Late Punic Amphora from the Bottega del Garum, Pompeii, during sampling.

Figure 3.- Participants of RACIIC, in the Aula Magna (A) and courtyard (B) of the Facultad de Filosofía y Letras, University of Cádiz.

Figure 2. Presentations by our late colleagues Tamás Bezeczky (†) and Jihen Nacef (†) during the RACIIC sessions.

(U. Ehmig, University of Frankfurt am Main); Austria (T. Bezecksky, Austrian Academy); Belgium (J. Poblome, Katholieke Universiteit Leuven, P. Monsieur, Ghent University); Spain (P. Berni, R. Járrega, H. GonzálezCesteros, Institut Catalá d’Archeologia Classica; A. Ribera, Ayuntamiento de Valencia; M. Beltrán Lloris, Museo de Zaragoza; A. Pecci, University of Barcelona, M. A. Cau, P. Reynolds, University of Barcelona – ICREA; D. Bernal-Casasola, M. Luaces, V. Palacios, A. Rodríguez, A. M. Roldán, A. M. Sáez, J. Sánchez, University of Cádiz; E. García Vargas, University of Seville); USA (T. Peña, Berkeley University), France (E. Botte, M. Bonifay, M.‑B. Carre, M. Sternberg, A. Tchernia, Centre Camille Jullian – Aix-Marseille Université/CNRS; D. Dixneuf, Laboratoire d’Archéologie Médiévale et Moderne en Méditerranée – Aix-Marseille Université/CNRS; J.‑Y. Empereur, Centre d’Études Alexandrines/CNRS; F. Cibecchini, DRASSM; J.‑L. Fournet, Collège de France – Paris; N. Garnier, Laboratoire Garnier; C. Hänni, École Normale Superieure – Lyon; F. Laubenheimer, University of Paris – Nanterre; S. Lemaître, University of Poitiers); Greece (G. Koutsouflakis, Greek Maritime Euphoria); Italy (S. Menchelli, University of Pisa; G. Olcese, University of Rome – La Sapienza; S. Pesavento Mattioli, University of Padova); Portugal (I. Vaz Pinto, Troia Resort; C. Fabião, University of Lisbon; R. Morais, University of Oporto); United Kingdom (M. Woodworth, Oxford University); Tunisia (J. Nacef, University of Monastir); Turkey

2

Preface (A. K. Senol, Egean University Izmir) and independent scholars (A. Opaiţ, Canada). In the years since the celebration of the event, some of these researchers have changed institution and some, sadly, are no longer with us. We want to use this opportunity to remember Jihen Nacef and Tamas Bezeczky with affection; you will never be forgotten. Some of the specialists who were invited could not attend, as they had previous commitments (J. P. Bernardes, J.-P. Brun, C. Carreras, S. Demeschtika, C. Franco, M. Hansson, M. Kbiri Alaoui, W. Van Neer, A. Wilson). Conversely, other colleagues could join us to the meeting and at the publication stage (S. Bakardzhiev, N. Boichot, S. Bombico, N. Borislavova, R. Bourgoing, R. Brulet, M. P. Buonincontri, J. Clarke, A. Contino, D. Cottica, G. M. Crisci, D. Djaoui, C. De Juan, G. Di Pasquale, H. Fujii, S. Gabriel, S. Grainger, B. Gruber, V. Hasenbach, J. D. Hernández García, V. Leitch, M. Lettieri, M. Loughton, S. Martin Kilcher, R. Marlasca, C. Mathe, D. Miriello, J. Miró, J. Muslin, A. A. Nagy, C. Nervi, F. Notarstefano, C. Oliveira, F. Olmer, G. Piques, S. Polla, A. Quevedo, J. Ream, C. G. Rodríguez Santana, N. Rovira, C. Sánchez, C. M. Sánchez Moral, I. Sanna, L. Soro, A. Springer, T. Theodoropoulou, M. Thomas, M. Tillier, P. Tušlová, I. Van der Graaff, C. Vieillescazes, B. Weisová). The RACIIC was, therefore, attended by a representative section of the field of ‘amphorology’, in both disciplinary and geographical terms.

Figure 4.- Public tribute to Professor Miguel Beltrán Lloris, attended by academic authorities from the University of Cádiz and Campus de Excelencia Global CEIMAR, alongside some of the organisers.

The steering and scientific committees saw it fitting to dedicate the event to Dr Miguel Beltrán Lloris, Director of Museo de Zaragoza, who at the time reached the happy age of retirement and gave the ‘opening lecture’, the only one in this publication to be printed in Spanish, as a deference to him and to give the reader the opportunity to read about the problems related to the paleocontent of amphorae in his own words. He is, no doubt, one of the pioneers of this topic in Spain, with his well-known monograph Las ánforas romanas en España (Zaragoza, 1970), published 50 years ago almost to the day. Following an affectionate speech by Professor André Tchernia, who honoured us with his presence, Miguel Beltrán was given a Punic amphora as a token of our affection and as a reference to the longue durée that presides our subject of study (Figure 4). With his characteristic good humour, he said that all attendees should have been given a ‘full’ amphora, not an empty one, in direct reference to the central topic of the conference. It is worth stressing that Professor Beltrán is often made reference to in our fieldwork, when the excavation of early and mid-imperial Roman contexts yields a ‘Beltrán II A’, a ‘Beltrán II B’, or a ‘Beltrán 68’, a constant reminder of the importance and pioneering nature of his work.

Figure 5.- Opening lecture of the RACIIC conference by Professor André Tchernia, with the intervention of Dr Enrique García Vargas. indicators’) addressed methodological issues, which were divided into four key aspects: ‘typology and content’, superbly presented by A. Tchernia, followed by E. García Vargas’s reply (Figure 5); ‘reutilisation of Roman amphorae’ (presented by T. Peña and M. A. Cau); ‘Amphorae contents and shipwrecks’ (F. Cibecchini and G. Koutsouflakis); and ‘amphora inscriptions and content’ (U. Ehmig – J.-L. Fournet and P. Berni – E. Botte). Sessions 2 and 3 dealt with archaeometry: ‘amphorae and residue analysis’ (N. Garnier – A. Pecci and C. Oliveira – M. Woodworth); ‘DNA and amphorae’ (C. Hänni and N. Garnier); and ‘zooarchaeology and the analysis of amphora contents’ (M. Sternberg – D. Bernal-Casasola). The two following sessions aimed to update the available evidence for the main Atlantic and Mediterranean regions (‘Regional syntheses I – the West’; and ‘Regional syntheses II – Africa and the East’), which, for logistic reasons, consisted in a number of comprehensive lectures: ‘Hispanic and Mauretanian amphorae’ (presented by H. González Cesteros on behalf of D. Bernal-Casasola, M. Kbiri Alaoui, R. Morais, R. Járrega, A. M. Sáez, I. Vaz Pinto; and discussed by C. Fabião, E. García Vargas and A. Ribera); ‘Gallic amphorae and imitations in Central Europe’ (F. Faubenheimer and P. Monsieur); and ‘Italian amphorae’,

The conference’s programme followed a lateral approach. For each topic, a ‘speaker’, generally acting on behalf of several authors, gave a lecture, which was then commented upon by one or several ‘discussants’. In six sessions, the RACIIC tried to synthetize the most pressing issues. Session 1 (‘Historical and archaeological

3

Preface As editors, it is also important to apologise for the delay in bringing these proceedings out. Over five years have lapsed between the celebration of the conference in Cádiz and the publication of this monograph; too long a time, and we are aware of this: the frantic pace of institutional academia, constant fieldwork demands, and the effects of the COVID‑19, which has sent the project back by one whole year, are the main reasons for this delay. Some of the authors submitted their contributions over three years ago. In other cases, text and bibliography have been updated. We decided not to submit the papers to peer review, since all contributions sprang from invited lectures that had already been evaluated by the scientific and steering committees. Naturally, all authors are responsible for their own texts. The reader is kindly asked to take into account that some of the papers are over three years old. This is especially significant concerning archaeometric results (organic residue analysis), which can age very quickly indeed as a result of the rapid advance of techniques, equipment and interpretation.

Figure 6.- Participants debate during the conference. (S, Menchelli and M.-B. Carre - S. Pesavento). The second ‘geographical’ session examined ‘African amphorae’ (M. Bonifay and J. Nacef), ‘Egyptian amphorae’ (D. Dixneuf and J.-Y. Empereur), ‘Eastern amphorae’ (P. Reynolds and T, Bezeczky, S. Lemaître, A. K. Senol) and ‘Black sea amphorae’ (A. Opait and P. Reynolds). The event ended with a ‘conclusions and proposals’ session (Session 6), led by J. Poblome (‘Synthesis of proposals, agreements, disagreements and open topics’).

Finally, we would like to express our deepest gratitude to the institutions and individuals that made the RACIIC possible. First, the Campus de Excelencia Internacional del Mar CEIMAR, and especially its former General Coordinators Carmen Garrido and Fidel Echevarría, for taking this initiative under their wing and support it wholeheartedly. To Casimiro Mantell, Deputy ViceChancellor of Research and Javier Benavente, Director General of Research, University of Cádiz, for their support; and the technical staff at CEIMAR and UCC+i (UCA) for their help, especially in terms of outreach and transference. The members of the Steering Committee also want to thank Max Luaces, at the time a PhD candidate at the universities of Cádiz and Lyon, who was at the head of the conference’s secretary for several intense months, and to several members of research group HUM-440 (Alejandro González Blas, Rosa Arniz Mateos and Lorena Hoyo Gómez) for their logistic support during the event.

The conference prioritised debate, which was ample and illuminating, after both the lectures (Figure 6) and poster presentations. From typology to the semantics of reuse; from the sealing of amphorae to polyvalent shapes; these debates became veritable think tank sessions in which participants raised very important issues. The programme included two complementary activities: a visit to the Phoenician-Punic and Roman collections of the Museo de Cádiz, led by its then director, Dr Juan Alonso de la Sierra Fernández, whom we want to thank for his generosity; and a presentation and tasting connected to the project for the experimental reconstruction of Pompeii garum, ‘Flor de Garum’, coordinated by professors V. Palacios and A. M. Roldán, of the Chemical Engineering and Food Technology Department, University of Cádiz.

We also want to thank the members of the Scientific Committee, especially those who could not participate in the conference (A. Arévalo, L. C. Juan Tovar, M. Kbiri Alaoui, J. L. López Castro, B. Raissouni and A.Wilson).

The contents of the conference have been ‘reformatted’ for this publication, which includes 34 works with the main arguments presented to the RACIIC. The volume follows the same structure as the conference, but it adapts to the needs of the authors and topics, as explained in more detail in the conclusions. It is not always easy for authors to stick to the agreed topics, but virtually all participants of the conference finally feature in the volume, which is a matter of great satisfaction for the editors. We leave it to the reader to judge our results and to ponder about issues in which further work is needed, for instance about the creation of a working group for a European project dealing with amphoric contents, one of the conference’s targets. Unfortunately, to date this has not met with success, since the various ‘calls’ and ‘topics’ of Horizon 2020 did not include the topics at hand or those related to the protection of Historical Heritage, as is well known.

The conference was co-funded by Centre Camille Jullian, Aix-Marseille Université/CNRS, Università della Calabria and ERAAUB, Universitat de Barcelona. Several institutions collaborated by assuming the costs incurred by a number of participants, especially SECAH – Ex Officina hispana, the University of Oxford, the Université de Louvain, the Institut Català d’Arqueologia Classica, the Centre d’Études Alexandrines, CNRS, and Laboratoire Nicolas Garnier. This publication is co-funded by project GARVM III (PID2019108948RB-I00/ AEI / 10.13039/501100011033) Gobierno de España/Feder; Operative Program FEDER 2014-2020 and Consejería de Transformación Económica, Industria, Conocimiento y Universidades, Junta de Andalucía (project: ARQUEOSTRA, FEDER-UCA18-104415); and

4

Preface

Figure 7.- Whole amphora, with operculum, titulus pictus and possible remains of paleocontents (based on weight) from the Late Roman shipwreck of Ses Fontanelles, Palma de Mallorca (© Consell de Mallorca – Universitat de Barcelona - Universidad de Cádiz – Universitat de les Illes Balears, 2019). project ARQUEOFISH (P18-FR-1483), Programa de Ayudas a la I+D+I, Plan Andaluz de Investigación, Desarrollo e Innovación (PAIDI 2020); RACAMed I & II (HAR201784242-P and PID2020-113409 GB-100), funded by Ministerio de Ciencia e Innovación, contract RYC 2013-13369, as part of ERAAUB (2017 SGR 1043).

number of amphorae whose paleocontents are well known is under one hundred, and those in which physical residues and tituli picti are found together are even fewer (Figure 7), veritable ‘treasures’ that need investigating properly. We hope that technical advances will make these protocols part of the standard toolkit of archaeology in the next few decades.

The future will tell what is the real scope of our results. The conference was, at least, valuable in raising awareness among archaeologists about the need to handle amphorae in which remains of paleocontents have survived with care; these are exceptional finds, which must not be just ‘emptied’, but subject to careful protocols of analysis that include sieving, floating, triage, and as many archaeometric techniques as possible, from organic residue analysis to biomolecular analysis. Mentioning garum, with all its types and varieties, and the absence of scientific consensus, largely the result of the paucity of the available evidence, should suffice to stress the complexity of the matter. The

We want to stress that the reader will not find answers to all their questions in these pages. The paleocontents of Roman amphorae is a complex and unresolved issue. Yet, we aim to present an up-to-date picture, organised by topic and geographical region, with the aid of the top specialists in this fascinating aspect of the Roman economy. Cádiz – Aix – Barcelona – Toulon, December 2020 Darío Bernal-Casasola, Michel Bonifay, Alessandra Pecci and Victoria Leitch

5

Ánforas romanas y contenidos. Notas historiográficas Miguel Beltrán Lloris

Director de la cátedra Galiay de la Institución Fernando el Católico, Zaragoza

Abstract: The author would like first to show his appreciation to the organizers of the International Conference for having named the event in honour of Miguel Beltrán Lloris. The author briefly reviews his vocation towards amphorae, which started with an exemplar that belonged to Antonio Beltrán until its donation to the Museum of Zaragoza and on which M. Beltrán based his first article on Roman amphorae in 1965 (2. Introduction. My first amphora). The development of studies that have specifically affected the problem of amphora contents is analyzed, highlighting the most important contributions since the monograph Roman amphorae in Spain, 1970 (3. Historiographic notes for a discussion on the content of the Roman amphorae from Hispania). The author also emphasizes that amphorae are only one part of trade in antiquity (3.2. The amphorae are only part of the trade) and the importance of the epigraphic information to characterize much of the known packaging, aspects that have laid the foundations of traditional knowledge (4. The value of epigraphic information). The analysis of organic residues is highlighted as one of the key elements of knowledge since the 1950s (5. Analysis of organic residues). The author lists the problems arising from the phenomenon of copying amphora containers (6. Copy), raising the question of the identity between the amphorae’s shape and content, which can sometimes be misleading (7. It is always true that shape equals content?), focusing the discussion on some demonstrative examples, such as the Greco-Italic amphorae of wine from Gades or the Baetican Dressel 1 shape and its relationships with saltings, or on the contrary, the Dressel 7-11 shapes and their wine contents, both in Andalusia and in the Tarraconense. The value of other amphorae as multipurpose vessels is highlighted (8. Multipurpose Amphorae: Haltern 70 from Guadalquivir), stressing the importance of reutilisation or alternative contents that help to clear lurking doubts (9. Issues concerning the reutilisation or alternative contents). The author raises then the various dilemmas of the Baetican salted fish and fish sauces regarding its allocation to an amphora model, paying particular attention to the tituli picti (10. Baeticae salsamenta and fish sauces. What amphorae for what type of salting?). The same principle applies for wine amphorae from Tarraconense, on which few painted signs are known (11. Tarraconense wines. What amphorae for that type of wine?). The typological standard derivations (12. Stable types and typological conversions) are analysed, as well as the smaller versions of certain amphorae (13. Miniaturized versions) and the problems of the controversial supply of empty amphorae to distant centers for their packaging far away from their place of origin (14. Medium — and short —distance supply of empty containers). The presentation ends with a succinct summary of the Functional assignments by typology (section 15), by typology and territorial context (section 16) by typology, territorial contexts and waste/organic coatings (section 17), finishing with a reference to amphorae whose content remains unknown (18. Amphorae without content). Key words: Amphora, contents, trade, epigraphic information, organic residues, copy, alternative contents, multipurpose amphorae, salsamenta, fish sauces, wine, typology.

1. Roman Amphora contents in honour of Miguel Beltrán Lloris He tenido la fortuna de trabajar durante más de cuarenta años en la mejor profesión del mundo y además me han pagado por ello. Y, por si fuera poco, ahora, los amigos de las ánforas quieren dedicarme esta International Interactive Conference, con base en mis supuestos méritos en el campo de las ánforas, en el que he tenido más yerros que aciertos, partiendo sobre todo de aquellos pecados de juventud que contiene mi libro sobre Las ánforas romanas en España, que para mi sorpresa tuvo una acogida inesperada en el mundo científico. Nada raro, en aquellas lejanas fechas, teniendo en cuenta la escasísima bibliografía que sobre las ánforas hispanas se había producido en nuestro territorio. La dedicatoria del libro decía entonces: A Pío y Antonio Beltrán, mis maestros. Ahora, con el paso del tiempo, me gustaría ampliarla a Carmenchu, Cecilia y Daniel y, sobre todo, a los muchos amigos nacidos de las ánforas, cuya lista

enuncié en aquel momento y que ahora haría extensiva al centenar de participantes en esta Conferencia, encabezada, sobre todo, por Darío Bernal‑Casasola, Michel Bonifay y Alessandra Pecci, que han querido ligar mi nombre a esta celebración que sin duda alguna será un hito en la investigación anforológica. Me siento un hombre inmensamente afortunado por poder contar con vuestro afecto y amistad, sentimientos que os devuelvo con renovada gratitud. 2. Introducción. Mi primera ánfora Mi mundo de las ánforas, como tantas otras cosas en mi vida, se construyó a partir de la influencia de Antonio Beltrán, mi padre, maestro y amigo. En mi casa paterna, en la plaza de San Francisco de Zaragoza, en su recibidor, se alzaba solitario, un extraño ser antropomorfo, que a mi se me aparecía como una

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 7–27

Miguel Beltrán Lloris

acompañar a mi padre a las excavaciones arqueológicas y a descubrir universos fascinantes de hace miles de años. Enseguida me vi afectado por esa maravillosa enfermedad que se llama vocación y decidí dedicar mi existencia, científica, al mundo arqueológico, bajo la tutela de mis dos maestros, mi abuelo Pío, eminente numismático, y mi padre, cuyos pasos he seguido. Pero esa es otra historia que ahora no interesa. Volvamos al comienzo de mi relato. En el año 1965, para celebrar mi mayoría de edad, preparé, mi primer artículo, que estuvo dedicado, como no podía ser menos, a aquel «antropomorfo cerámico» que conocía desde mi infancia en el recibidor de casa (Beltrán Lloris 1965). En aquel momento, descubrí la existencia de la tabla de Dressel, que me causó cierta frustración al no poder clasificar nuestro ejemplar en su tipología. Al año siguiente, cayó en mis manos un sugestivo artículo del año 1957 de Fernand Benoit, que yo, como gran ignorante en la materia, no conocía: «Typologie et épigraphie amphoriques» (Benoit 1957: 252). Gracias a dicho artículo, descubrí los errores de mis primeras líneas y publiqué más tarde una rectificación completa a las notas iniciales (Beltrán Lloris 1968). Aquel episodio me dejó tan frustrado que me llevó a revisar y leer con fruición todo lo que cayó en mis manos relativo a las ánforas romanas, que comenzaron a formar parte de mis obsesiones más queridas. Así, en junio del año 1969 leí como tesis de licenciatura en la Universidad de Zaragoza, Las ánforas romanas en España, que tomaron forma editorial al año siguiente, en 1970 (Beltrán Lloris 1970), hace 45 años. 3. Notas historiográficas para una discusión sobre los contenidos de las ánforas romanas de Hispania

Figura 1. Ánfora grecoitálica de las Alhambras, Teruel. Museo de Zaragoza, 96.28.1. Donativo de Antonio Beltrán Martínez. Fotografía: Museo de Zaragoza, J. Garrido Lapeña.

No es este el lugar para desarrollar in extenso los problemas en torno al contenido de las ánforas, que la presente reunión ha canalizado en dos discursos lógicos, el de los indicadores históricos y arqueológicos y el de la arqueometría, que sumados a las síntesis regionales presentarán, sin duda, el estado actual de conocimiento. El contenido de las ánforas, encuadrado en la tipología formal, debe ser abordado desde el análisis de los residuos (Biers y Mc Govern 1990), los revestimientos internos con resina o pez, la información contenida en los tituli picti, las fuentes escritas y el contexto de los hallazgos en los centros de producción (Noguera Celdrán y Antolinos Marín 2011-2012).

esbelta joven con gráciles brazos y estilizado cuerpo, que procedía de una localidad turolense de sugestivo nombre: las Alhambras (Figura 1). Aquella «figura» estuvo en el recibidor de mi casa durante muchos años y yo mantuve con ella una especial relación, nacida de la curiosidad que provocaban aquellas formas y su posible utilidad, aunque esta es una cuestión que yo descubrí enseguida. Aquel recipiente, de esbelto cuello, servía para depositar, en el último momento, los papeles perdidos, los chicles, envoltorios de golosinas, billetes caducados y otras menudencias que llevábamos los chiquillos en los bolsillos. Además, claro está, de alojar los paraguas en los días de lluvia. Sin duda aquel extraño ser, muy antiguo, tenía múltiples usos para los distintos habitantes de aquella casa, y para mí se convirtió en un mudo personaje al que saludaba y despedía todos los días.

3.1. Las primeras aproximaciones Haciendo excepción ahora de las obras de síntesis generales, que no enumeraremos en aras de la brevedad (Bernal-Casasola y Ribera i Lacomba 2008, 2012; Étienne y Mayet 2000, 2002, 2004) no han sido excesivos los trabajos que se han dedicado de forma específica a los contenidos de las ánforas, desde los pioneros de Étienne en 1990 sobre las ánforas lusitanas (Étienne 1990) o de Fabião y Guerra (1993)

Este fue mi primer encuentro con las ánforas romanas, después vendría mi vocación por la antigüedad y el mundo de los museos, desde muy temprano, cuando comencé a

8

Ánforas romanas y contenidos. Notas historiográficas

insistiendo en el mismo sentido, hasta las aportaciones de Bernal‑Casasola y Petit años más tarde (Bernal-Casasola y Petit Domínguez 1994‑1995; 1999), las exhaustivas líneas de García Vargas en 1998 sobre los envases gaditanos (García Vargas 1998), o las más recientes de Bernal-Casasola (2000) insistiendo no solo en los productos tradicionales sino en otros tenidos erróneamente como secundarios, además de los vinos béticos (Bernal-Casasola 2009). Finalmente sobre el mundo productivo de las salsas y conservas de pescado, el Congreso Internacional Cetariae 2005 (Lagóstena Barrios, Bernal-Casasola y Arévalo 2007), o el simposio internacional en Homenagem a Françoise Mayet de 2006 (Tavares da Silva y Soares 2006), así como la síntesis meridional de García Vargas (2006), el trabajo de Lagóstena Barrios (2001) y finalmente, el luminoso encuadre de la cuestión de BernalCasasola (2015), sin que las más recientes aportaciones hayan insistido especialmente en el apartado de los contenidos (Morais 2017: 333; Járrega Domínguez 2018).

Collet de Sant Antonio, la Almadrava Nord (Gisbert Santonja 2009: 128), recordando, además, las palabras de Tchernia (1986: 304), a propósito de la importancia del tonel desde el siglo II, como corrobora el comercio de los Uritii (Martin-Kilcher 2002: 350) o la producción vinícola de Barcino en el siglo III d. C. (Beltrán de Heredia 2009; Beltrán de Heredia y Comas i Sola 2009). 4. El valor de la información epigráfica Desde las primeras referencias pompeyanas en el CIL IV (supp. 2, 1871), y los trabajos pioneros de Dressel sobre el Castro Pretorio y el Testaccio de Roma (Dressel 1879, 1899) (Figura 2), glosados por Remark en 1912 (sin tener en cuenta la tipología anfórica), más tarde, en 1965, por Callender (apoyándose excesivamente en el CIL IV) (Callender 1965: 37-39) y en 1966, por Fausto Zevi (1966), se ha venido cimentando la importancia de los tituli picti para el contenido de las ánforas (Marichal 1975).

Desde la monografía de Las ánforas romanas en España, del año 1970, en donde se contenían breves capítulos alusivos a los contenidos según formas y tipos, el panorama ha evolucionado notablemente. Los más recientes trabajos han estabilizado la tríada de contenidos anfóricos (aceite, vino y salsas/salazones) y en el momento presente el proyecto Amphorae ex Hispania: paisajes de producción y de consumo (http://amphorae.icac.cat/), unido ahora al anterior Roman Amphorae: a digital resource (http:// archaeologydataservice.ac.uk), y complementado por la información derivada de las bases de datos del CEIPAC (http://ceipac.ub.edu) sirven de perfecto termómetro para calibrar hasta qué punto se ha avanzado sobre el contenido de las ánforas.

Epigráficamente, vino y salazones de pescado constituyen los contenidos clásicos de las ánforas hispanas, corroborados a partir de la información generada desde los años 70 (Manacorda 1977; Hahn 1989; Martin-Kilcher y Schillinger 1989-1990; Marichal y Liou 1993; MartinKilcher 1994; Ehmig 1996; Liou 1998; Liou y RodríguezAlmeida 2000; Martin-Kilcher 2002), siendo relevante el papel de los tituli béticos salsarios, cuya influencia (en las producciones galas e itálicas) se ha puesto de relieve (Martín-Kilcher 2004). A estos, se ha unido el contenido oleario (oleum dulce) sobre ánfora de forma Dressel 23, encontrada en Aiguablava (Gerona) (Anónimo 2015). Se conocen cerca de 40 menciones de nombres de vinos, con sus calidades (vinum simplex, vetus, mulsum…) y variedad geográfica (gaditanum, falernum…).

Desde un punto de vista historiográfico han primado hasta fechas muy recientes los atributos de la forma, el origen y la cronología, habiendo quedado en un término más discreto el referente a la función/contenido, que es absolutamente necesario para situar las ánforas (y las cerámicas en general) (Notarstefano 2012: 15) en su contexto real.

Se han estabilizado las ánforas Dressel 1-4 como vinarias; las formas Dressel 7-9, 11, 12, 14-16 y Beltrán II A y B como salazoneras; otras se han asociado a las aceitunas y sus derivados, como la Dressel 23 (olivas colombares y olivas salitas CIL XV: 4803 a-b, 4804) y 26 (CIL XV: 4855, 4898), un cuarto grupo por la aparente contrariedad de los contenidos de los tituli, está en terreno indeterminado como la Dressel 6 (vino, aceite, liquamen), y otros recién interpretados han recobrado su fisonomía como la forma Dressel 21-22, cuya epigrafía salazonera (Botte 2009), se ratifica ahora por los paleocontenidos (BernalCasasola 2015). Los tituli hispanos más abundantes aluden a las ánforas de salazones, cuya síntesis y crítica ha hecho Lagóstena Barrios (2004: 206), y siguen siendo minoritarios los relativos al vino (Revilla Calvo 2004: 190).

3.2. Las ánforas son solo una parte del comercio No me corresponde ahora, sino enunciar la historiografía del contenido de las ánforas a través de una serie de propuestas de trabajo, que resumen algunas de las controversias generadas en torno al contenido anfórico. Estos recipientes son solo una parte del comercio de difusión de productos de consumo, a los que habría que sumar los dolia, los envases de vidrio, toneles (Beltrán Lloris, Ortiz Palomar y Paz Peralta 1999: 137) y odres (Lund 2004), además de los pequeños recipientes especializados y costosos, que difundieron producciones selectas (Cappelletto et al. 2013), ciertas jarras que también tuvieron importante difusión, e incluso las lucernas de aceite como testimonios indirectos (Morillo Cerdán 2000: 627‑628). Los dolia documentan la diversificación de la calidad de los productos vinarios transportados, como el ejemplo tarraconense del

La correcta interpretación del amineum vetus en las ánforas galas (Gauloise 4, Dressel 2-3 de Lyon) como una variedad de uva transplantada a las Galias ayuda a interpretar los tituli referentes a la muria hispana y al garum hispanum en los envases Dressel 9 de Frejus o Lión, que además envasaron la muria de Antipolis (Desbat 2003; 2004: 294-295).

9

Miguel Beltrán Lloris

Figura 2. Tituli picti sobre ánforas hispánicas de Castro Pretorio, Roma (Dressel 1879: láminas XIII-XIV). Un caso extraordinario en la información epigráfica lo proporcionan los salazoneros sellos figurados del «envasador gadirita» en las ánforas Ramon T-9111 (Sáez Romero 2016b).

1998; Beck et al. 1989; Bertucchi 1992), sino también a los salazoneros, (ya en proceso de análisis desde los años 70: Formenti 1993; Bernal-Casasola y Petit 1999; Bargagliotti 2000: 1116), y recientemente a las ánforas de aceite (Garnier, Silvino y Bernal-Casasola 2011), eliminando la exclusividad de este revestimiento interno solo para los envases vinarios (Bernal-Casasola 2009) y mostrando la coexistencia entre resinas y aceites (Pecci y Cau Ontiveros 2010), que deja abierta la puerta a los aceites no alimentarios en los que la resina o pez intervendrían desde el punto de vista olfativo (MartínezMaganto, García Jiménez y Bernal-Casasola 1997: 104).

5. Los análisis de residuos orgánicos Los contenidos epigráficos de Dressel tuvieron su comprobación en los primeros análisis en 1955 sobre ánforas vinarias (Jongkees 1955), en 1976 de Condamin y Formenti sobre trazas de aceite y vino en ánforas Dressel 20 y Dressel 2-4 (Condamin y Formenti 1976, 1978), y en 1978 sobre el vino en la forma Lamboglia 2, a los que siguieron otros trabajos en la misma década (Oddy y Bateman 1973; Formenti, Hesnard y Tchernia 1978; Addeo, Piccioli, Altiri 1979) y posteriormente (Formenti 1989: 563) sobre Dressel 26 (aceite) y Dressel 2-4 (vino), sumándose a estas conclusiones la información de los pecios con restos de ictiofauna (estorninos, jureles, sardinas) (Desse-Berset y Desse 2000; Delussu y Wilkens 2000), que coincide con la obtenida en determinadas factorías (Olisipo) (Assis y Amaro 2006).

6. La copia Es evidente el fenómeno de la copia, que en el mundo romano se muestra inicialmente deudor del helenismo, como representa la réplica itálica en el siglo I a. C. de las ánforas de Cos para el transporte del vino, origen de la Dressel 2-4 y su secuela mediterránea y atlántica (tarraconense, bético, gálico, helvético o británico). Este fenómeno obedece a la corriente generada en la formación del imperium romanum y del arte nuevo que se desarrolla desde el prestigio e identificación de las formas griegas. Resulta difícil deslindar la mera copia de un contenedor de productos de prestigio del simple engaño comercial para disfrazar producciones bajo falsas apariencias (Desbat 2004: 291).

Se añadieron, además, otros contenidos, que sumados a los ya conocidos vinarios (Formenti y Duthel 1995) o de salazones (Tchernia 1986), ampliaron el abanico funcional de las ánforas, desde las chuletas de cerdo del pecio Planier 3 (Tchernia, Girard y Poplin 1986: 242), a los dátiles (forma Camulodunum 189) (Peacock y Williams 1986: 109), los restos de frutos (como las nueces, avellanas, melocotones, ciruelas o cerezas) (Bruni 2000), y otros productos minoritarios (Bernal-Casasola 2004: 238).

Se parte de un principio bien definido. La forma de un ánfora garantiza el origen y la calidad del producto que transporta, axioma que se mantiene desde sus primeras definiciones, basado en la identidad entre forma y contenido. La copia busca la reproducción de una forma anfórica con la intención de conseguir un grado máximo de semejanza y asociar el prestigio del producto original

Se ha producido además, la ampliación de los revestimientos de pez, no solo a los envases vinarios (Celades, Pascual Guasch y Villate 1983; Martínez Maganto y Petit Domínguez

10

Ánforas romanas y contenidos. Notas historiográficas

al de la reproducción conseguida. El copista permanece preso de la técnica de ejecución, pero la caracterización arqueométrica permite la distinción entre los modelos originales y sus derivados, hasta el punto de que son ciertamente escasas las copias verdaderas, como ya expuso Tchernia hace años (1992: 213). En ausencia de información sobre el ánfora, ignoramos cómo podía el comprador identificar el producto de diferentes orígenes envasado en recipientes análogos, debiendo suponer la existencia de información en otros materiales perecederos (Eiring y Finkielsztejn 2004: 463).

ánforas Dressel 1C (Fase II, 2º 1/3 a. C.) y Dressel 21‑22 (Lomba de Canho 67) (Fase III) (Bernal-Casasola y Jiménez-Camino 2004: 595) y a través de Op. M. Lucretius, envases Dressel 12, en principio envases vinarios y de salazones dependientes del mismo productor. En esta línea opinaron en su día Domergue (1973: 114), y después Étienne y Mayet (1994: 131), repitiendo el modelo del ager cosanus en donde se identificó por E. L. Will (1987) la posibilidad (¿?) de una producción de la Dressel 1A para vino y 1C para salazones, apoyándose en las ánforas Dressel 1 de Albinia (que transportaron como producto primario el pescado) (Costantini 2007).

7. ¿Siempre se cumple el axioma: forma igual a contenido?

7.3. Las Dressel 7-11 vinarias

Uno de los axiomas más contundentes en el mundo de las ánforas es el de la identidad entre forma y contenido. Las exigencias funcionales conformaron los contenedores: cuellos largos y estrechos para las sustancias líquidas (vino), bocas amplias para los alimentos semisólidos (salazones) y cuellos cortos y estrechos para los envases olearios. Y así parece en la mayoría de las formas conocidas en lo que constituye una interpretatio normalizada (Fabião y Guerra 1993), puesta en duda, sin embargo, desde hace tiempo, con base en contextos mas complejos que aconsejan desconfiar de normas apriorísticas excesivamente rígidas que obligan a reconsiderar muchas conclusiones (Silvino y Poux 2005: 501).

La Bética La experiencia epigráfica y analítica demuestra que el grupo anfórico bético de las formas Dressel 7-11 estuvo destinado al transporte de las salazones. Lo corroboran los restos orgánicos de caballa común (scomber scombrus) en ánforas Dressel 7 y 9 del pecio de Sud Perduto y las Dressel 12 de Cap Bear (scomber japonicus de talla media) (Desse-Berset y Desse 2000: 75). No deja de llamar la atención que para el ánfora de Clase 67 (Lomba de Canho 67) (costera y del Guadalquivir), se ha planteado un uso dual desde las salazones iniciales (Fabião 1989: 68), con base en la indefinición, de contenidos, sus antecedentes formales, y siguiendo el ejemplo de la forma Haltern 70 (Fabião 2000a: 673; Molina Vidal 2000: 638; García Vargas 2012: 200; García Vargas, Conlin y Maestre 2013).

7.1. Grecoitálicas gaditanas y salazones En la correspondencia, forma-contenido, se observan algunas contradicciones, como las grecoitálicas de imitación de la costa gaditana, cuyos análisis de paleocontenidos han proporcionado salsas derivadas del pescado y no productos vinarios —Baelo Claudia— (Bernal-Casasola et al. 2003: 310; Bernal-Casasola, Roldán Gómez y Blánquez Pérez 2004), aunque no se eliminen las posibilidades de localizar envases costeros dedicados al vino (¿?) (García Vargas et al. 2016).

Otras dualidades se han estimado para la Dressel 10, a partir del conocido titulus de Roma (CIL XV: 4533): (vinum) amin(eum), sobre forma similis non magna, que atenuaría la hipótesis dual, y el dudoso CIL XV: 4731 que se lee Has(…) (¿Hastense vinum?).

Esta alternancia de contenidos, es un fenómeno constatado especialmente en los tiempos tardorrepublicanos, como ha comprobado García Vargas (2012: 200) en el Guadalquivir y parece un fenómeno dual intenso en los tiempos preaugusteos, que se irá decantando con el transcurso del tiempo.

En la Dressel 9, con referencias mayoritarias al garum, liquamen, muria y ventresca de atún (abdomina membratim caesa), se menciona la lumpha (vino rebajado en agua) hasta en ocho tituli (Pesavento y Buonopane 2002; Arata 1994; García Vargas 1998: 204; Liou 1993: 140), además de las pepitas de uva en envases del pecio Sud Perduto 2, en las Bocas de Bonifacio y Cádiz (Chic García 1980). La lumpha se constata igualmente en ánforas Dressel 7-11, Beltrán IIB (García Vargas 2004: 511) y Dressel 12 de Pompeya: Lump(hatum)-(CIL IV: 5609), lympha(tum) vetus penuarium?/excelens (Giordiano y Casale 1990: n 374-5).

7.2. Dressel 1 C bética y salazones En lo relativo a la Dressel 1 C bética, en función de los antecedentes grecoitálicos gaditanos, se han postulado dos usos con base en territorios distintos: salazonero en la costa (Pérez Rivera 2000: 236) y vinario en el valle del Guadalquivir por la ausencia de factorías salazoneras (Almeida, García Vargas y González Cesteros 2016). Es evidente que faltan los argumentos definitivos (García Vargas et al. 2016).

En el Guadalquivir, la forma Maña C2b (T-7.4.3.3), cuya morfología indica derivados del pescado, ha proporcionado tituli referidos al vin[um?]/du(…) y restos orgánicos en forma de pepitas de uva e incluso mejillones en vinagre (García Vargas, Almeida y González Cesteros 2011: 199; Sáez Romero et al. 2016: fig. 1). Las presentes constataciones vinarias tienen además otro punto de

Se inscriben en este fenómeno las conocidas producciones algecireñas de S(ocii) C(etarii) G(aditani) (Étienne y Mayet 2002: 117 ss.; Hesnard 1998) (Figura 3) que fabricaron

11

Miguel Beltrán Lloris

Figura 3. Ánforas Dressel 1 C de Baelo Claudia (Etienne y Mayet 2000: figura 27). refrendo que nos parece ciertamente estable, según su presencia en los contextos funerarios de la Belgica oriental, en donde las ánforas Dressel 7-11 béticas, forman «servicio» con la vajilla metálica destinada al consumo del vino, asumiendo el papel que antes ostentaban las ánforas Dressel 1 y Pascual 1 (Silvino y Poux 2005: 503).

76). Este fenómeno supondría una dualidad de contenidos vinarios, o de salazones, para las formas Tarraconense 1 A, C y D de salazones y las variantes 1 B y E, adjudicadas al vino (Miró Canals 2016). Se ha aducido igualmente el factor geográfico, al tratarse de envases producidos lejos de la costa, (a veces no más de 10 km). La situación de los centros productores tierra adentro parece abogar hacia contenidos vinarios, por lo forzado de la producción salazonera en dichos ámbitos (Fernández-Izquierdo 1995: 211) y el coste adicional del transporte de los envases vacíos hasta los hipotéticos centros costeros.

La Tarraconense Se ha planteado que ciertas producciones de la forma Dressel 7-11 tarraconense, anteriores a mediados del siglo I d. C., pudieron haber contenido vino (Silvino y Poux 2005; Beltrán Lloris 2008b: 272; López-Mullor 2009:

12

Ánforas romanas y contenidos. Notas historiográficas

la situación en un envase multiusos, como ya sugirió Martin-Kilcher (1994: 385), según los tituli: defrutum/ sapa/olivae y muria, sin olvidar la posibilidad de que hubiera podido mantener un uso distinto en su primera etapa de producción (República y Augusto), antes de que se definieran los restantes envases béticos (Dressel 7‑11, Dressel 20). Entre los restos orgánicos se evidencian restos de aceitunas en Sud Lavezzi (Liou 1982: 444), Cala Culip IV (Nieto Prieto et al.: 1989) y Saintes-Maries-de-laMer II (Liou 2000).

Es clara la convivencia de distintos modelos de ánforas vinarias en los alfares tarraconenses. Sirve de ejemplo en el Campo de Tarragona y en el Bajo Ebro, con la coetánea producción de ánforas Pascual 1, Oberaden 74 y Dressel  2-3 (Carreras Monfort y González Cesteros 2012: 213), siendo paradigmático el ejemplo de l’Aumedina de Tivisa, que fabricó además Dressel 7/11 (Revilla Calvo 1993: 51), como el centro de Sant Boi de Llobregat (Pascual 1, Dressel 3, Dressel 7-11) (López-Mullor 2013: 67). La base plana de la Oberaden 74 explica su adaptación para su difusión fluvial/terrestre (Ebro, Aude, Garona…), pero no su contenido, y el resto de los envases debería obedecer a distintas calidades del vino.

Otro tanto se ha supuesto para las copias lusitanas, añadiendo además los contenidos piscícolas según el territorio (Mayet y Silva 2002). La forma Verulamium 1908 del Guadalquivir (forma tardía de la Haltern 70) (Carreras Monfort 2012: fig. 1), remite al mismo contenido (olivae ex defructum) (Carreras Monfort y Marimon Ribas 2003a). Las imitaciones de esta en la Galia Narbonense confirman uno de los contenidos principales: oliva alba o nigra, según los tituli (Carreras Monfort y Marimon Ribas 2003b: 41) y se ha optado por la atribución vinaria para imitaciones de la Haltern 70 en Mérida —Mérida I— (Bustamante Álvarez y Cordero Ruiz 2013: 83) (Figura 5).

Este fenómeno se comprueba en otros productores, como Philodamus que envasó en Pascual 1, Dressel 2-3 y Dressel 30 (Pascual Guasch 1992: 174-176); ANTH, que produjo Dressel 8 y Dressel 2-3 (Pascual Guasch 1992: n. 22; Tremoleda i Trilla 2000) y otros que fabricaron formas Dressel 2-3 y Pascual 1 (L.C.P —Pascual Guasch 1992: 119—, PSC —Pascual Guasch 1992: 187—, etc.). TIBISI estampilló Pascual 1, Dressel 2-4 y Dressel 7/11 (Revilla Calvo 1993: 79), y el lusitano Lucius Arvenius Rusticus sobre variantes locales Dressel 7/11 A, B, C, Pascual 1 y Haltern 70 (Caedos, Rodrigues y De Sepúlveda 2006).

Otro tanto se ha planteado, en la década de los ochenta, sobre la forma salazonera (muria) Beltrán IIB, que García Vargas sugirió como la sucesora de la polivalente Haltern 70 (García Vargas 1998: 205) y que ha proporcionado tituli de muria, vino y lympha (Liou 1987; Lagóstena Barrios 2004), además del titulus de Fos con defr(utum) excell(ens) (Liou 1987: n. 35).

Las ampuritanas ánforas de ANTH, que produjo ejemplares típicamente vinarios, como la Dressel 2-4 y otros de tipología salazonera, como la Dressel 8, estas últimas interpretadas como vinarias, pese a su tipología (Figura 4). Sin embargo, los restos de conservas de pescado en el ejemplar de Tolegassos, junto a Ampurias, han dirimido la cuestión a favor de las salazones (Tremoleda i Trilla 2000). Un modelo análogo de comportamiento observamos en los envases de época augustea del liberto Iulius Theophilus, que produjo en los centros del Bajo Llobregat ánforas Pascual 1 vinarias y Dressel 21-22, que asimiladas a sus homólogas itálicas son de salazones, caracterizadas por su amplia boca (Beltrán Lloris 2014).

La falta de información ha llevado a postular para las Ovoides I, según su situación interior o costera, diversos usos: aceite, vino, salazones o derivados del pescado (García Vargas, Almeida y González Cesteros 2011: 214), con atribuciones igualmente plurifuncionales para la Ovoide 2-4, 8, 9 y 10 (¿vino, defrutum, aceite?). 9. Reutilizaciones o contenidos alternativos

La asociación de determinadas ánforas salazoneras a contenedores relacionados con el vino en tumbas de Goeblange, hizo pensar a S. Martin Kilcher (2000: 769), en el probable contenido de hydrogarum o oenogarum para algunos contenedores. Por otra parte, ciertas formas Dressel 7-11 del Guadalquivir, epigráficamente, se asocian al vino aguado (vinum lymphatum) (Pesavento y Buonopane 2002).

El hallazgo en Marsella de un ejemplar de forma Beltrán  IIB, cuyo interior proporcionó restos de conservas de aceitunas con hinojo y hojas de laurel, nos conduce a la reutilización (¿?) del envase para transportar olivas de Provenza (Boyer 1986). Ya desde la década de los 70, fuera de nuestro ámbito, las ánforas de Tomis demostraron la reutilización sucesiva de envases en época tardía (Radulescu 1973: 201), en lo que sin duda se procedió de forma selectiva evitando la contaminación de sabores que podían proporcionar algunos recipientes olearios (Bernal-Casasola 2004: 328). Recuérdese que en el pecio de Grado, todas las ánforas, de diverso origen y tipología (Africana I, Tripolitana I, Dressel 2-4 egeas), se reutilizaron con contenidos piscícolas (Auriemma 2000), o aceite y cereales (Eiring y Finkielsztejn 2004: 464); y el pecio de Sud-Caveaux 1, evidenció la reutilización de ánforas Dressel 10, Oberaden 74, Haltern 70 y Lamboglia 2 para el transporte de pez (Long y Delauze 1996).

8. Ánforas multiusos: Haltern 70 del Guadalquivir Muy controvertidas han sido las teorías sobre el contenido del ánfora bética Haltern 70, históricamente adscrita a diversos usos, desde el garum (Ettlinger 1977: 16), al mundo vinario —Port Vendres II— (Colls et al. 1977), además de los restos de resinas analizados (Morais 2003). Las síntesis de Carreras y Aguilera (Carreras Monfort 2003a; Carreras Monfort y Aguilera Martín 2003; Morais 2005b; Carreras Monfort y Berni Millet 2016), dejan

13

Miguel Beltrán Lloris

Figura 4. Ánforas ampuritanas de ANTH (Tremoleda 2000: figura 101). Ramon ya planteo en la década de los 90 (1995) una serie de contenidos para diversos tipos de ánforas púnicas, de aire salazonero, como T-1.2.13 (restos de atún y palometa), T-1.4.4.1 (carne ovina y bovina en vino), T-2.1.1.2 (huesos de ovicáprido), T-8.1.1.1 y T-8.1.2.1 (huesos de conejo), T-7.5.2.2 (aceitunas y ramas de olivo), T-8.1.3.1 (conserva de cerdo), demostrando la importancia de las conservas cárnicas en salazón. Posteriormente, para BernalCasasola, solo las ánforas Tejarillo 3, pudieron haber estado dedicadas a las salazones cárnicas, por las áreas epónimas interiores donde se han localizado los alfares (Bernal-Casasola 2004: 361).

No parece oportuno alargar estas líneas con otros productos o usos alternativos: nueces, pimienta, habas, lentejas, dátiles, miel, resinas (Callender 1965: 39), resina (Almagro 51 c) (Lequement 1976; Dyczek 2001: 293), o, además, la reutilización de un ánfora Beltrán IIA como vivero (Bernal-Casasola 2011: 147). 10. Las salazones béticas: ¿qué envases para qué salazones? Sigue sin constatarse una especialización formal de las ánforas salazoneras y sus contenidos, cuyos tituli mencionan numerosos aspectos sobre el origen, la materia prima y

14

Ánforas romanas y contenidos. Notas historiográficas

características, así como sobre la vejez o solera del mismo (CIL XV: 4570, Dressel 9: Ti Caesare V Cos/Gaditanum). Las grecoitálicas de Baelo, proporcionaron productos fabricados con más de una especie (atunes, peces pequeños, ovicápridos, suidos y caracoles) (Bernal-Casasola et al. 2003). Los tituli sobre los envases Dressel 7‑11, 12, 14 y Beltrán IIA-B, confirman las conclusiones ya establecidas desde los primeros corpora (Dressel 1891, 1899 y Schoene y Mau 1909). La misma forma contiene productos distintos dentro de las salsamenta y salsas de pescado (garum, liquamen, muria, allex) (García Vargas 2006) y se constata una superposición entre las distintas formas y los contenidos de los tituli (Manacorda 1977; Beltrán Lloris 2000: 444). Tanto el garum como el liquamen (los productos de mayor calidad), están presentes en las formas Dressel 7, 10 y Beltrán IIA y B; y la muria (una salmuera de peor calidad), aparece igualmente sobre las formas Dressel 7, 10 y Beltrán IIB. La forma Dressel 7 también contuvo liquamen (CIL XV: 4731) y cordyla arguta (CIL XV: 4741). Para la Dressel 8 son predominantes los tituli de gari flos (calificado cuatro veces como scombri), aunque también se conocen referencias al liquamen, la muria, y el laccatum (Lagóstena Barrios 2004: 207). La Dressel 9, manifiesta un amplio abanico de contenidos: garum (flos, scombri), liquamen, muria (flos), lymphatum, cordyla y ventresca (abdomina membratim caesa). Sucede otro tanto con el residuo tamizado, el hallec, que solo aparece envasado en las formas Dressel 10 y Beltrán IIA/B, que también transportaron garum y liquamen. La forma IIA contuvo además miscell(um) port(uense) y cordyla arguta y portuensis (Lagóstena Barrios 2004: 215). Predominan las inscripciones con adjetivos de calidad (flos, excellens) y abundan las menciones a la caballa (scomber) (Martínez-Maganto 2000), la especie mas constatada entre los restos de ictiofauna (García Vargas 2006). La Dressel 14, presenta la alternancia entre el liquamen (en sus variantes excellens y flos excellens) y la muria (Étienne 1990; Blánquez Pérez et al. 1998); además de los saxitani vet(eres) de pecio Gandolfo (Étienne 1990).

Figura 5. Ánforas de forma Haltern 70 de producción local emeritense (Bustamante 2011: figura 16). bética o Almagro 51 c (Bernal-Casasola 2000: 275), se han definido según restos orgánicos asociados (salazones, salsas de pescado y moluscos). 11. Los vinos tarraconenses: ¿qué envases para qué vinos? Se ha supuesto que las primeras producciones de ánforas Dressel 1B son un ejemplo de falsificación, ya para exportar vino itálico que llegaba a granel a las costas catalanas (¿?), ya para hacer pasar el vino layetano por un vino italiano, engañando a la clientela (Étienne y Mayet 2000: 255). Estas primeras producciones de la Hispania Citerior, los tipos grecoitálicos, las formas Dressel 1A, B y C y la Lamboglia 2 (López-Mullor y Martín Menéndez 2008a; 2008b: 691) (desarrollados por ceramistas itálicos inmigrantes sobre una base técnica indígena), exportaron producciones locales de nombre desconocido (Beltrán Lloris 2008a). Las imitaciones consecuentes de los prototipos itálicos, Tarraconense 1-3 (segunda mitad del siglo I a. C.), se han adjudicado (desde 1961) al vino, a tenor de sus antecedentes tipológicos y análisis (Vidal y Pascual 1971; Nolla Brufau y Solias 1985: 118).

Hay una cierta especialización de la forma Dressel 12, hacia el garum scomber (la caballa), con más de 12 tituli (González Cesteros 2012), aunque también se referencian el lac(atum) arg(utum) excell(ens) summ(um), la lumpa y la flos muriae. El único titulus sobre Dressel 11, alude a la mur(ia) sec(unda) (CIL XV: 4726). El atún joven, cordula (Apicio 425; Plinio NH: 9, 47; Mart., 3,2,4), fue transportado (ejemplares completos o troceados) de forma mayoritaria por la Beltrán IIA (García Vargas 2006; García Vargas, Martín-Arroyo y Lagóstena Barrios 2016) y con escasa relevancia por la Dressel 7-11 (Liou 1987: 68). La misma alternancia de formas presentan el laccatum (salsa rojiza, por la lacca) y el lymphatum (vino aguado), reducido el segundo exclusivamente a las formas Beltrán II A/B y Dressel 12. La variante Beltrán IIB contuvo flos muriae (Rodríguez Temiño 1990; FernándezChicarro 1953: 438), laccatum y gari scombri flos (MartinKilcher 1994: 406). Otras formas, como la Keay XXII

Solo los escuetos tituli sobre la forma Dressel 2 mencionan al (vinum) laur(entanum) (CIL XV: 4577-4579), atribuido a Lauro en la Layetania (Vallés oriental) comparables, según Plinio, a los de Tarraco y las Baleares (NH: XIV, 71), transportados, entre otras, en ánforas Dressel 2-3 de L.PHAE (CIL XV: 4578, San Boi de Llobregat) (Pascual Guasch 1992: n 173). Para el resto, formas Pascual 1, Dressel 2-4, Oberaden 74/Dressel 28 (Colls et al. 1977: 47) y Gauloise 4 tarraconense, no se conocen tituli, y su contenido vinario deriva de la tipología, restos de pez y geografía productora.

15

Miguel Beltrán Lloris

Vargas, Almeida y González Cesteros 2011: 237-238), también fue copiada fuera de su ámbito, manteniéndose la función principal (Dressel 20 layetana) (Berni Millet 2000; Berni Millet 2011), o con productos asociados semejantes, como el aceite de nuez en las imitaciones Gauloise 13 y 14 (esta con residuos de aceite de semilla de trigo) o con producciones de aceites locales (Laubenheimer 2000: 1126; Formenti 2000; Ehmig 2000: 1146). La sucesora tardía, de este envase, la Dressel 23, es conocida por los tres tituli romanos, referentes a las olivas colombares y olivas sa(li)tas (CIL XV: 4803 a y b, 4804). Del mismo modo para la levantina Oliva 3 se ha supuesto el aceite con base en su parentesco físico con las formas béticas.

Las informaciones literarias hablan del vino de Tarraco, en cuyo territorio se produjeron especialmente ánforas Dressel 2-4 (Járrega i Domínguez 1998) a cuyas producciones del s. I cabría (¿?) atribuir el calificativo de elegantia (Plinio NH: XIV, 71) (aunque los comentarios sean de finales del siglo I d. C.) y calidad semejante a la itálica lacial (Mart XIII: 116; Florus Verg: II) (siempre que no tengamos que clasificar estas atribuciones como un topos literario). Por otra parte, las referencias al vino layetano (Comas i Sola y Carreras Monfort 2006) (territorio del Maresme, Vallés occidental y oriental, Barcelonés y Bajo Llobregat), afectarían a las producciones conocidas: Pascual 1, Dressel 2-4 y similares, a las que cabría atribuir la calidad de vinos rudos y groseros apreciados por su abundancia, como declara Marcial (Epigr: 26, 9; I 49: 21-22; VII: 53, 6), que bien pudieron viajar en dolia, reservándose las ánforas para vinos más selectos (Miró Canals 1988: 108).

13. Las versiones miniaturizadas El mundo de las cerámicas presenta módulos diversos, grandes y pequeños, que se explican por los procesos de almacenamiento para su difusión, o como medidas en sí mismos. En el mundo hispánico, como en el rodio (amphoriskoi) (Aubert 2004: 37) o itálico (Dressel 2-4 grande y pequeña) (Panella 2001: pl. 1, 5-6), hay una serie de formas cuya miniaturización hace pensar en un refinamiento o selección de sus contenidos originales, como las formas Ramon T-9111, 9121, 11210, 12110 (salazones), la Martinhal 4 (afín a Almagro 51 a-b) (Bernardes et al. 2013: 324), o Beltrán II A 1 (Bargagliotti 2000); en otras ocasiones podría tratarse de regulaciones específicas de las exportaciones aceiteras, como en la forma aceitera Tejarillo I (Guadalquivir) (Remesal Rodríguez 1991), o en la Beltrán 72 (Bernal-Casasola 2000: 306), para la que Parker (1977: 39) supuso podría tratarse de una miniatura tardía de la forma Beltrán II B (Bernal-Casasola 2000: fig. 21); o bien la forma Puerto Real 3, respecto de Puerto Real 1 y 2 (Bernal-Casasola y García Vargas 2012: 251), o la Águilas IV (Berrocal Caparrós 2012: 269).

En cuanto a las producciones levantinas, solo pueden aducirse las referencias a los viñedos de Voconius Romanus en Sagunto (Plinio Epist: II, 13) y las de Juvenal (Sat V: 30) y Cornelio Fronto a la mala calidad de dichos caldos, equiparables a los vinos de Creta pero vulgares en su comparación con los de Falerno (Aranegui Gascó 2008). Es dudosa la lectura am(ineum), en el ánfora Dressel 2-4 de M.P.M., de Saguntum (Liou 1998: 94; Aranegui Gascó y Mantilla Collantes 1987). 12. Tipologías estables y reconversiones tipológicas En el desarrollo tipológico y temporal, se observa una sustitución de envases, con distinto ritmo, para los mismos contenidos. Así el ejemplo clásico de la generalización de las más económicas formas Dressel 2-3 en lugar de las Dressel 1, o la secuencia Pascual 1/ Dressel 2-3 o Dressel 20/Dressel 23, etc.

14. Suministro de envases vacíos a corta y media distancia Darío Bernal‑Casasola planteó en 1999, como explicación a la falta de hornos en determinadas áreas productivas de la Tingitana, con gran densidad de cetariae, la hipótesis del suministro de envases gaditanos vacíos (Bernal-Casasola y Pérez Rivera 2000: 861; contra: Lagóstena Barrios 2001: 370), como se comprobaba entre otros ámbitos en Septem Fratres que ha proporcionado ánforas de forma Beltrán IIA y Beltrán IIB, con la estampilla SOCI, cuya arqueometría remite a pastas gaditanas. Dicho comportamiento se repetía en otras factorías, como Rusaddir (Mogador), o Ras Achakar. La explicación radicaba en la economía, del «Círculo del Estrecho» que combinaba todo lo producido en ambas orillas (cetariae, hornos, sal), necesario para el proceso global. Dicho modelo, que parece soportado por los S(ocii) C(etarii) G(aditani) (Mayet 1999; Étienne y Mayet 2002: 112) /S(ocietas) C(etariorum) G(aditanorum) (Lagóstena Barrios 2014), parece también aplicable a Baelo Claudia en época republicana y a Sexi en la etapa bajoimperial (Bernal-Casasola 2006) y puede estar en la base de ciertas anomalías que se han detectado en este terreno (Silvino y Poux 2005).

Llama la atención la producción de ánforas Dressel 14 en el taller gaditano de Villa Victoria, en donde se ha detectado una evolución fabril desde la Dressel 2/4 (vinaria), producida en la fase I del taller (Augusto avanzado) a la Dressel 14, directamente, que sustituye a la anterior en la Fase II (con borde circular y acanaladuras dorsales), en un proceso creativo distinto del lusitano (Bernal-Casasola et al. 2004: 645). Estamos ante la copia de un envase vinario para producir formas destinadas ¿también al vino?, ¿a las salazones?. Las producciones posteriores de la Dressel 14 B de la costa granadina y malagueña, envasaron salazones, como se comprueba por los restos e ictiofauna analizados (Étienne 1990: 17). Son las ánforas Dressel 20 y 23, las que, tradicionalmente en la historiografía, no dejan lugar a dudas. La aceitera forma Dressel 20, y sus eslabones iniciales (Ovoide 6, Oberaden 83 y Haltern 71) (Berni Millet y García Vargas 2016; García

16

Ánforas romanas y contenidos. Notas historiográficas

También se ha planteado como hipótesis, con base en los hallazgos de Castro Marim, con ánforas Ovoides 1 del Guadalquivir, trasladadas a la costa para envasar salazones (García Vargas, Almeida y González Cesteros 2011: 214), en un fenómeno que sugirió Fabião (2000b: 717) a propósito de las salazones del Algarve; o Berrocal (Berrocal Caparrós 2012: 272) para los envases de Mojón, localizados en factorías de salazones (Majuelo, Almuñécar y La Picola, Santa Pola).

la Dressel 12, la Puerto Real 3 (García y Lavado 1996; Bernal-Casasola y García Vargas 2012: 251); la Beltrán IIB, las formas Puerto Real 1 y 2 (García Vargas y Lavado Florido 1996); la Almagro 50, la Lusitana Regional II (Morais 2005b: 133); la Almagro 51 A/B, la forma La Orden (O’Kelly 2012: 281); la Almagro 51 C, la tarraconense Almagro 51 C/Mojón (Berrocal Caparrós 2012: 263), y la Keay XLI la tarraconense Keay XLI/Mojón (Berrocal Caparrós 2012: 264), o la Keay XIX c, la Keay XIX c/ Mojón (Berrocal Caparrós 2012: 267) y la Keay LXXXIX B, la Keay LXXXIX B/Mojón (Berrocal Caparrós 2012: 263); habiéndose tomado el spatheion como referente para las Águilas I-III (Berrocal Caparrós 2012: 268).

15. Atribuciones funcionales por su tipología Junto a los recursos que presta el medio natural, en ausencia de tituli o paleocontenidos, se sitúan los contenidos asignados a determinadas formas con base en las derivaciones tipológicas o con los resultados obtenidos de otras zonas de producción, como los caldos vinarios de la Dressel 28 tarraconense (Tremoleda et al. 2006: 480).

16. Atribuciones funcionales por su tipología y contexto territorial Salazones Bética costera: Dressel 1C (grecoitálicas gaditanas) (Bernal-Casasola y Jiménez-Camino 2004: 589 ss.); Majuelo I (Beltrán 72) (Bernal-Casasola 2000: 300), Venta del Carmen I (Dressel 11, Beltrán II A) (Bernal-Casasola 1998a), Dressel 14 tardía (Dressel 14) (Bernal-Casasola 2000: 278), Puerto Real 1, tardía (Bernal-Casasola 2000: 279), Puerto Real 2 (Keay XC) (Bernal-Casasola 2000: 280), Keay XVI (Bernal-Casasola 2000: 281), Beltrán 72 (BernalCasasola 2000: 286), Dressel 17 (Dressel 12) (BeltránFortes y Loza Azuaga 1997: 109; García Vargas 1998: 104).

Este argumento no está exento de imprecisiones por la cadena de copias que se produce y la solución se complica cuando determinadas formas (sin contenido) son el claro antecedente de otras, posteriores, cuyo contenido si se conoce, como la Ovoide 4 del Guadalquivir (con la que guardan parecido las Tarraconenses 1 C y 1 D), antecedente o primera parte de la multiusos Haltern 70. No parece haber dudas en cuanto a las homónimas de las béticas Dressel 2-4 (García Vargas 2004: 510), Dressel 28 (García Vargas 2004: 511), Dressel 30 (Bernal-Casasola 2009: 39), Gauloise 4 (Bernal-Casasola 1998b: 543), Gauloise 5 (García Vargas, Conlin y Maestre 2013: 288) y la lusitana meridional Dressel 14 (Dressel 14 occ.) (Viegas 2016).

Bética Guadalquivir: Dressel 7-11 (Carreras Monfort 2003b: 647; Carreras 2015). Lusitania occidental: Dressel 14 (Mayet y Tavares da Silva 2002: 225), Almagro 51 A/B (Pinto y Magalhães 2016), Keay XVI (Keay XVI bética) (Almeida y Cordeiro 2016), Almagro 51 c var. A (Mayet y Tavares da Silva 2002: 177), Almagro 51 c var. B (Mayet y Tavares da Silva 2002: 179), Haltern 70 (Simões 2016).

Así, con base en parentescos formales con la forma Dressel 20, se han atribuido al aceite la forma Oliva 3 (Mateo Corredor y Molina Vidal 2016) o las Tejarillo I y II (Bernal-Casasola 2000: 292), o a partir de la Dressel 23, la Keay XV (Keay 1984: 147); desde la forma Apani III, se ha clasificado la Ovoide 6 (García Vargas, Almeida y González Cesteros 2011: 230), siendo mas dudosas las atribuciones de las formas Ovoides 2 y 3 desde la Apani Va (GarcíaVargas, Almeida y González Cesteros 2011: 201).

A partir del contexto territorial, se han atribuido a las salazones la tarraconense Mojón I (Berrocal Caparrós 2012: 255) y las béticas: Eucaliptal 1, Eucaliptal 2 y Terrón 2 (O’Kelly 2012: 282).

En cuanto a las atribuciones vinarias, la Lamboglia 2 ha servido para la Ovoide 10 (García Vargas, Almeida y González Cesteros 2011: 212); la Ovoide 1 ha inspirado la Tarraconense 3 (García Vargas, Almeida y González Cesteros 2011: 262); la Pascual 1, la forma Ramon P 26 (Ramon Torres 2013b); la Dressel 2-3, las formas Ramon P 25 (Ramon Torres 2013a); la Dressel 28, la Bracara Augusta, Local I (Morais 2005b: 564), la Dressel 30, la variante Dressel 30/Mojón (Berrocal Caparrós 2012: 265); la Gauloise 1, la Matagallares I (Bernal-Casasola 2000: 282), la Gauloise 4, la Lusitana Regional I (Morais 2005b: 522) y las Gauloise 5 y 7 han apoyado las referencias al urceus bético (Morais 2005a; 2012: 267).

Vino Tarraconense: Almadrava IV. Por su tipología, base plana y territorio de Dianium (vinario) (Gisbert Santonja 1999: 387). Lo mismo se aplica a las formas Gauloise 4, Segobriga/Oberaden 74 similis ¿para vino mielado? (Oberaden 74) (Almeida y Morín 2012: 242; Almeida et al, 2013: 378) y a la Oberaden 74 similis de la Rioja (Beltrán Lloris 2008b: 298). 17. Atribuciones funcionales por su tipología, contexto territorial y residuos/revestimientos orgánicos Vino Bética costera: Dressel 30 (Bernal-Casasola 2009: 33), Beltrán 68 (resina) (Bernal-Casasola 1996; 2000: 290).

En cuanto a las salazones, la forma Dressel 7-11, ha motivado la Tarraconense P 41 (Ramon 2008: 265);

17

Miguel Beltrán Lloris

Almeida, R. R. de y Cordeiro Raposo, J. M. 2016. Keay XVI (Lusitania occidental). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 . Almeida, R. R. de, García Vargas, E. y González Cesteros, H. 2016. Dressel 1 (Valle del Guadalquivir). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 08 julio, 2016 . Almeida, R. R. de y Morín de Pablos, J. 2012. ¿Ánforas Tipo Segobriga/Oberaden 74 similis? Bases para una producción singular en la Tarraconense interior. En D. Bernal-Casasola y A. Ribera i Lacomba (eds.) Cerámicas hispanorromanas II. Producciones regionales: 231-246. Cádiz, Universidad de Cádiz. Almeida, R. R. de, Morín de Pablos, J., Agustí García, E., Urbina Martínez, D., Urquijo Alvarez de Toledo, C., López Fraile, F., Guerra García, P. y Benito Díaz, L. 2013. Los Vallejos, Casas de Luján II y Rasero de Luján (Saelices, Cuenca). Nuevos datos sobre la producción cerámica en el territorio de Segobriga. En D. Bernal, L. C. Juan, M. Bustamante, J. J. Díaz y A. M. Saez (eds.), Hornos, talleres y focos de producciόn alfarera en Hispania. I Congreso internacional de la SECAH (Cadiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (I): 363-383. Cádiz, Universidad de Cádiz. Amphorae ex Hispania - I. Paisajes de producción y de consumo . Anónimo 2015. Els arqueòlegs troben al fondejador d’Aiguablava la primera àmfora romana d’oli per cuinar que conserva la inscripció. El Digital de Girona, consultado 7 julio, 2015 . Aranegui Gascó, C. 2008. La producción y el comercio de ánforas tarraconenses en el País valenciano. En A. López Mullor y J. Aquilué Abadías (eds.), La producció i el comerç de les àmfores de la provincia Hispania Tarraconensis. Homenatge a Ricard Pascual Guasch. Actes de les Jornades d’Estudi (17-18 de novembre de 2005). Monografies 8: 227-240. Barcelona, Museu d’Arqueologia de Catalunya. Aranegui Gascó, C. y Mantilla Collantes, A. 1987. La producción de ánforas Dr. 2-4 de Sagunto. En El vi a l’antiguitat. Economía, producció i comerç al Mediterrani occidental. Actes del I col·loqui d’arqueologia romana, Badalona, 28, 29, 30 de novembre i 1 de desembre de 1985. Monografies badalonines 9: 100-104. Badalona, Museu de Badalona. Arata, P. 1994. Un relitto da Cala Rossano (Ventotene). En Epigrafia della produzione e della distribuzione. Actes de la VIIe Rencontre franco-italienne sur l’épigraphie du monde romain (Rome, 5-6 juin 1992). Collection de l’École française de Rome 193: 477-496. Roma, École française de Rome. Assis, C. A. y Amaro, C. 2006. Estudo dos restos de peixe de dois sitios fabris de Olisipo. En C. Tavares da Silva y J. Soares (eds.), Simpósio Internacional Produção e comércio de preparados piscícolas durante a proto-história e a época romana no Ocidente da Península Ibérica. Homenagem a Françoise Mayet (Setúbal, 7-9 Maio 2004). Setúbal Arqueológica 13: 123-144.

Salazones Tarraconense: Forma Keay XIX C (Bernal-Casasola 2000: 239). Bética costera: Ramon T-7433 (Sáez Romero et al. 2016), Terrón 1 (Almagro 51 A/B) (O’Kelly 2012: 280); Keay XIX (caballas) (Liou 1973: 586; 1982: 437). Lusitania meridional: Almagro 51 C (caballas) (Liou 1973: 586; Fabião 2004: 379). Lusitania occidental: Almagro 50 sardinas y caballa, (sardina Pilchardus, scomber, restos de pescado) (Chevalier y Santamaria 1971: 32; Almeida y Cordeiro 2010), Almagro 51c (pescado) (Étienne y Mayet 1990: 17), Beltrán 72 (¿garum?) (Bost et al. 1992: 134). 18. Ánforas sin contenido Hay ánforas, para las que no existen todavía propuestas de contenido, en ausencia de modelos de referencia, tituli, geografía productora o análisis de paleocontenidos. Ilustra la situación la forma Pellicer D de la costa, para la que se ha propuesto un contenido multifuncional (vino, aceite, salazones, aceitunas…) (Sáez Romero y Niveau 2016), aunque un ejemplar de esta forma, del valle del Guadalquivir, haya proporcionado grasa vegetal (aceite de oliva) en los análisis (García-Fernández y García Vargas 2010: 115 ss.). Igual ambivalencia ofrece la Ramon T-8211 (Sáez Romero 2016a). Y persisten las dudas para las tarraconenses Keay LXXVIII (Keay 1984: 240), Keay XLI (aceite o vino) (Keay 1984: 252; Bernal-Casasola 2000: 295), Keay XXV (aceite, salazones) (Keay 1984: 252; Lequément 1976: 181), Mojón III (Berrocal Caparrós 2012) y PE-41 (¿salazones o vino?) (Ramon Torres 2008: 265). Y lo mismo para las béticas del Guadalquivir: Ovoide 8 (¿aceite?) (García Vargas, Almeida y González Cesteros 2011: 205); Ovoide 4 («antecedente» de Haltern 70) (García Vargas, Almeida y González Cesteros 2011: 217; García Vargas, Almeida y González Cesteros 2016); Ovoide 9 (García Vargas, Almeida y González Cesteros 2011: 206); Dressel 7-11 (García Vargas, Almeida y González Cesteros 2011: 253); Tejarillo III (¿vino?) (Bernal-Casasola 2000: 292), Keay XIV (¿aceite?) (Bernal-Casasola 2000: 293), Keay XVII (Bernal-Casasola 2000: 294). También para la costera La Orden (¿salazones?) (O’Kelly 2012), etc. Bibliografía Addeo, F., Piccioli, C. y Altiri, F. 1979. Costituendi acidi di una oleoresina di conifere rinvenuta in anfore vinarie durante gli scavi archeologici di Oplontis. Annali della Facoltà di Agraria di Portici 13: 144–172. Almeida, R. R. de y Cordeiro Raposo, J. M. 2010. Almagro 50 (Lusitania occidental). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 8 julio, 2016 .

18

Ánforas romanas y contenidos. Notas historiográficas

Aubert, C. 2004. Le commerce antique en Phénicie d’après les amphores locales et importées de Beyrouth. En J. Eiring y 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: 31-41. Atenas, Danish Institute at Athens. Auriemma, R. 2000. Le anfore del relitto di Grado e il loro contenuto. Mélanges de l’École française de Rome. Antiquité 112-1: 27-51. Bargagliotti, S. 2000. Il relitto dell’Ardenza (Livorno): un carico di anfore betiche naufragato intorno alla meta del I sec. D. C. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 1111-1120. Écija, Gráficas Sol. Beck, C. W., Smart C. J. y Ossenkop D. J. 1989. Residues and linings in ancient Mediterranean transport amphoras. En R. O. Allen (ed.), Archaeological Chemistry IV: 369380. Washington DC, American Chemical Society. Beltrán de Heredia, J. 2009. Premses vineres i instal·lacions vinícoles a Barcino. En C. Carreras Monfort y J. Guitart i Duran (eds.), Barcino I. Marques i terrisseries d’amfores al Pla de Barcelona. Corpus International des timbres amphoriques 15: 119-130. Barcelona, Institut d’Estudis Catalans, Institut Català d’Arqueologia Clàssica. Beltrán de Heredia, J. y Comas i Sola, M. 2009. Instal·lacions vinícoles vinculades a domus. Els exemples de Barcino i Baetulo. En M. Prevosti y A. Martin (eds.), El vi tarraconense i laietà ahir i avui, Documenta 7: 151-165. Tarragona, Institut Català d’Arqueologia Clàssica. Beltrán Fortes, J. y Loza Azuaga, M. L. 1997. Producción anfórica y paisaje costero en el ámbito de la Malaca romana durante el alto Imperio. En Figlinae malacitanae, La producción de cerámica romana en los territorios malacitanos: 107-146. Málaga, Universidad de Málaga. Beltrán Lloris, M. 1965. Sobre un ánfora romana de las Alhambras (Teruel). Caesaraugusta 25-26: 124. Beltrán Lloris, M. 1968. Un ánfora de tradición grecoitálica en Teruel. En Miscelánea ofrecida al Ilmo. Sr. D. José María Lacarra y de Miguel: 85-97. Zaragoza, Librería General. Beltrán Lloris, M. 1970. Las ánforas romanas en España. Monografías Arqueológicas 8. Zaragoza, Institución Fernando el Católico. Beltrán Lloris, M. 2000. Ánforas béticas en la Tarraconense. Bases para una síntesis. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (II): 441-535. Écija, Gráficas Sol. Beltrán Lloris, M. 2008a. La cerámica romana en la Tarraconense: el inicio de las producciones propias. En SFECAG, Actes du Congrès de L’Escala-Empúries:15-32. Marsella, Société Française d’Étude de la Céramique Antique en Gaule.

Beltrán Lloris, M. 2008b. Las ánforas tarraconenses en el valle del Ebro y la parte occidental de la provincia tarraconense. En A. López Mullor y J. Aquilué Abadías (eds.), La producció i el comerç de les àmfores de la provincia Hispania Tarraconensis. Homenatge a Ricard Pascual Guasch. Actes de les Jornades d’Estudi (1718 de novembre de 2005). Monografies 8: 271-317. Barcelona, Museu d’Arqueologia de Catalunya. Beltrán Lloris, M. 2014. Iulius Teophilus. Un nuevo fabricante de salazones tarraconenses en época de Augusto. En Miscelánea de estudios en homenaje a Guillermo Fatás Cabeza: 141-154. Zaragoza, Institución Fernando el Católico. Beltrán Lloris, M., Ortiz Palomar, M. E. y Paz Peralta, J. Á. 1999. La vajilla relacionada con el vino en Hispania. En S. Celestino Pérez (ed.), El vino en la antigüedad romana. Segundo simposio del vino en la antigüedad, Jerez 1996: 129-200. Madrid, Universidad Autónoma de Madrid. Benoit, F. 1957. Typologie et épigraphie amphoriques, Rivista di Studi Liguri 23: 251-254. Bernal-Casasola, D. 1996. Las ánforas del tipo Beltrán 68 en Hispania: problemática y estado de la cuestión. En Actas del XXIII Congreso Nacional de Arqueología, Vol. II: 251-269. Elche, Ajuntament d’Elx. Bernal-Casasola, D. 1998a. Las ánforas de producción local: tipología, caracterización y epigrafía. En D. Bernal-Casasola et al. (ed.) Excavaciones arqueológicas en el alfar romano de la Venta del Carmen. Los Barrios (Cádiz). Una aproximación a la producción de ánforas en la Bahía de Algeciras en época altoimperial: 143-198. Madrid, Universidad Autónoma de Madrid. Bernal-Casasola, D. 1998b. Las ánforas de vino béticas en la Baja Romanidad. Novedades procedentes de recientes excavaciones en centros de producción andaluces. En El vi a l’antiguitat: economia, producció i comerç al Mediterrani occidental. Actes del II Col·loqui Internacional d’Arqueologia Romana (Barcelona 6-9 de maio de 1998). Monografies badalonines 14: 543‑552. Badalona, Museu de Badalona. Bernal-Casasola, D. 2000. La producción de ánforas en la Bética en el s. III y durante el bajo imperio romano. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (I): 239-372, Écija, Gráficas Sol. Bernal-Casasola, D. 2004. Ánforas de transporte y contenidos. A propósito de la problemática de algunos envases de los ss. II y I a. C. En XVI Encuentros de Historia y Arqueología. Las industrias alfareras y conserveras fenicio-púnicas de la Bahía de Cádiz: 321-373. Córdoba, Publicaciones Obra Social y Cultural Caja Sur. Bernal-Casasola, D. 2006. La industria conservera romana en el Círculo del Estrecho. Consideraciones sobre la geografía de la producción. En L’Africa romana XVI. Mobilità delle persone e dei popoli, dinamiche migratorie, emigrazioni ed immigrazioni nelle province occidentali dell’Impero romano: 1351-1394. Roma, Carocci.

19

Miguel Beltrán Lloris

Bernal-Casasola, D. 2009. Ánforas y vino en la Antigüedad Tardía. El ejemplo de la Hispania meridional. En J. Blánquez Pérez y S. Celestino Pérez (eds.), El vino en época tardoantigua y medieval. Serie Varia 8: 33-60. Madrid, Universidad Autónoma de Madrid. Bernal-Casasola, D. 2011. Piscicultura y ostricultura en Baetica. Nuevos tiempos, nuevas costumbres. En D. Bernal-Casasola (ed.) Pescar con arte. Fenicios y romanos en el origen de los aparejos andaluces. Monografías del proyecto Sagena 3: 137-159. Cádiz, Universidad de Cádiz. Bernal-Casasola, D. 2015. What contents do we characterise in Roman amphorae? Methodological and archaeological thoughts on a trending topic. En C. Oliveira, R. Morais y A. Morillo Cerdán (eds.), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 61-83. Esposende, Município de Esposende. Bernal-Casasola, D., Arévalo, A., Lorenzo L. y Aguilera L., 2003. Imitations of italic amphorae for fish sauce in Roman Baetica. New evidence from the salt fish factory at Baelo Claudia. Rei Cretariae Romanae Fautorum Acta 38: 305-313. Bernal-Casasola, D. y García Vargas, E., 2012. Las ánforas del tipo Puerto Real 3. Un nuevo envase de salazones gaditanas en época antonino-severiana. En D. BernalCasasola y A. Ribera i Lacomba (eds.) Cerámicas hispanorromanas. Un estado de la cuestión II. Producciones regionales: 247-254. Cádiz, Universidad de Cádiz. Bernal-Casasola, D. y Jiménez-Camino, R. 2004: El taller de El Rinconcillo en la Bahía de Algeciras. El factor itálico y la economía de exportación (ss. I a.C.-I d.C.). En D. Bernal‑Casasola y L. Lagóstena (eds.), Figlinae Baeticae. Talleres alfareros y producciones cerámicas en la Bética romana (ss. II a.C. - VII d.C.). Actas del Congreso Internacional (Cádiz, 12-14 de noviembre de 2003). British Archaeological Reports. International Series 1266 (II): 589-606. Oxford, John W. Hedges. Bernal-Casasola, D. y Pérez Rivera, J.M. 2000. Las ánforas de Septem Fratres en los ss. II y III d. C. Un modelo de suministro de envases gaditanos a las factorías de salazones de la costa tingitana. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 861-886, Écija, Gráficas Sol. Bernal-Casasola, D. y Petit Domínguez, M. D. 1994-1995. Caracterización de resinas en ánforas romanas por cromatografía de gases. Resultados y aplicaciones en España. Alebus 4/5: 84-98. Bernal-Casasola, D. y Petit Domínguez, M. D. 1999. Análisis químico de residuos en ánforas romanas de vino y salazones de pescado: problemas y resultados. En J. Capel Martínez (ed.) Arqueometría y Arqueología. Serie Monográfica Arte y Arqueología de la Universidad de Granada: 269-294. Granada, Universidad de Granada. Bernal-Casasola, D. y Ribera i Lacomba, A. (eds.) 2008. Cerámicas hispanorromanas. Un estado de la cuestión. Cádiz, Universidad de Cádiz.

Bernal-Casasola, D., y Ribera i Lacomba, A., (eds.) 2012. Cerámicas hispanorromanas. Un estado de la cuestión II. Producciones regionales: 247-254. Cádiz, Universidad de Cádiz. Bernal-Casasola, D., Roldan Gómez, L., Blánquez Pérez, J., Díaz, J. J. y Prados, F. 2004. Las Dr. 2/4 béticas. Primeras evidencias de su manufactura en el conventus gaditanus. En D. Bernal‑Casasola y L. Lagóstena (eds.), Figlinae Baeticae. Talleres alfareros y producciones cerámicas en la Bética romana (ss. II a.C. - VII d.C.). Actas del Congreso Internacional (Cádiz, 12-14 de noviembre de 2003). British Archaeological Reports. International Series 1266 (II): 634-648. Oxford, John W. Hedges. Bernardes, J. P., Morais, R., Pinto, I. V. y Dias, R. 2013. A olaria baixo-imperial do Martinhal, Sagres (Portugal). En D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz y A. M. Saez (eds.), Hornos, talleres y focos de producciόn alfarera en Hispania. I Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (I): 317-329. Cádiz, Universidad de Cádiz. Berni Millet, P. 2000. Amphorae Dressel 20 similes. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 11591167. Écija, Gráficas Sol. Berni Millet, P. 2011. Un ánfora Dressel 20 similis layetana del alfar de Sant Vicenç dels Horts (Baix Llobregat, Barcelona). Boletín Ex Officina Hispana 3: 32-33. Berni Millet, P. y García Vargas, E. 2016. Dressel 20 (Valle del Guadalquivir). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 23 noviembre, 2016 . Berrocal Caparrós, M. C. 2012. Producciones anfóricas en la costa meridional de Carthago-Spartaria. En D. Bernal-Casasola y A. Ribera (eds.), Cerámicas hispanorromanas II. Producciones regionales: 255-277. Cádiz, Universidad de Cádiz. Bertucchi, G. 1992. Les amphores et le vin de Marseille VIe s. av. J.-C.- IIe s. ap. J.-C. Revue archéologique de Narbonnaise Supplément 25. París, CNRS. Biers, W. R. y McGovern, P. E. 1990. Organic contents of Ancient Vessels: Materials Analysis and Archaeological Investigation. MASCA 7. Philadelphia, University of Pennsylvania, Museum of Archaeology and Anthropology. Blánquez Pérez, J., Roldán Gómez, L., Martínez Lillo, S., Martínez Maganto, J., Sáez Lara, F. y Bernal‑Casasola, D. 1998. V. Epigrafía anfórica: los tituli picti de Pecio Gandolfo. En La Carta arqueológica-subacuática de la Costa de Almería (1983-1992). Arqueología Colección: 282-308. Madrid, Universidad Autónoma de Madrid. Bost, J. P., Campo, M., Guerrero, V. y Mayet, F. 1992. L’épave Cabrera III (Majorque). Échanges commerciaux et circuits monétaires au milieu du IIIe siècle après JésusChrist. Publications du centre Pierre Paris 23. París, de Boccard.

20

Ánforas romanas y contenidos. Notas historiográficas

Botte, E. 2009. Salaisons et sauces de poissons en Italie du sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31; Archéologie de l’artisanat antique 1. Nápoles, Centre Jean Bérard. Boyer, R. 1986. Conserve d’olives dans deux amphores trouvées dans les fouilles de la Bourse, à Marseille. Gallia 44-2: 229-233. Bruni, S. (ed.) 2000. Le navi antiche di Pisa: ad un anno dall’inizio delle ricerche. Florencia, Polistampa. Bustamante Álvarez, M. 2011. La cerámica romana en Augusta Emerita en época altoimperial. Entre el consumo y la exportación. Mérida, Ataecina. Bustamante Álvarez M. y Cordero Ruiz, T. 2013. Une exportation viticole à Mérida ? Considération sur la production locale d’amphores de style Haltern 70. En S. Celestino Pérez y J. Blánquez Pérez (eds.), Patrimonio cultural de la vid y del vino / Vine and wine cultural heritage: 81-93. Madrid, Universidad Autónoma de Madrid. Caedos, G., Rodrigues, E. y De Sepúlveda, E. 2006. A olaria romana de Peniche. En C. Tavares da Silva y J. Soares (eds.), Simpósio Internacional Produção e comércio de preparados piscícolas durante a proto-história e a época romana no Ocidente da Península Ibérica. Homenagem a Françoise Mayet (Setúbal, 7-9 Maio 2004). Setúbal Arqueológica 13: 253-278. Callender, M. H. 1965. Roman amphorae with index of stamps. Londres, Oxford University Press for the University of Durham. Cappelletto, E., Bernal-Casasola, D., Cottica D., Bustamante M., Lara M. y Sáez, A. 2013. Urcei per salse da pesce da Pompei-Ercolano: una prima analisi. En D. Bernal, L. C. Juan, M. Bustamante, J. J. Díaz y A. M. Saez (eds.), Hornos, talleres y focos de producción alfarera en Hispania. Congreso internacional de la SECAH (Cadiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (I): 271280. Cádiz, Universidad de Cádiz. Carreras Monfort, C. 2003a. Haltern 70: a review. En Amphorae in Britain and the western Empire. Journal of Roman Pottery Studies 10: 85-91. Carreras Monfort, C. 2003b. Ánforas. En M. T. Amaré Tafalla (ed.), Astorga IV. Lucernas y ánforas. Arqueología leonesa I: 633-673. León, Universidad de León. Carreras Monfort, C. 2012. Verulamium 1908 (Valle del Guadalquivir). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 6 agosto, 2016 . Carreras Monfort, C. 2015. Dressel 7-11 (Valle del Guadalquivir). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 . Carreras Monfort, C. y Aguilera Martín, A. 2003. El producte envasat. En C. Carreras Monfort et al. (eds.), Culip VIII i les àmfores Haltern 70. Monografies del Centre d’Arqueologia Subaquàtica de Catalunya 5:117-132. Girona, Generalitat de Catalunya.

Carreras Monfort, C. y Berni Millet, P. 2016. Haltern 70 (Valle del Guadalquivir). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 08 julio, 2016 . Carreras Monfort, C. y González Cesteros, H. 2012. Ánforas tarraconenses para el limes germano: una nueva visión de las Oberaden 74. En D. Bernal-Casasola y A. Ribera i Lacomba (eds.), Cerámicas hispanorromanas II. Producciones regionales: 207-230. Cádiz, Universidad de Cádiz. Carreras Monfort, C. y Marimón Ribas, P. 2003a. Verulamium 1908 (Haltern 70 tardana). En C. Carreras Monfort et al. (eds.), Culip VIII i les àmfores Haltern 70. Monografies del Centre d’Arqueologia Subaquàtica de Catalunya 5: 32-35. Girona, Generalitat de Catalunya. Carreras Monfort, C. y Marimón Ribas, P. 2003b. Imitacions d’Haltern 70 en la Gàl·lia romana. En C. Carreras Monfort et al. (eds.), Culip VIII i les àmfores Haltern 70. Monografies del Centre d’Arqueologia Subaquàtica de Catalunya 5: 41- 50. Girona, Generalitat de Catalunya. Celades, R., Pascual Guasch, R. y Villate, E. 1983. Estudio analítico de los revestimientos orgánicos aparecidos en el interior de las ánforas. Informació Arqueológica 41: 169-178. Centro para el Estudio de la Interdependencia Provincial en la Antigüedad Clásica (CEIPAC) . Chevalier, Y. y Santamaria, C. 1971. L’épave de l’Anse Gerbal à Port-Vendres, Rivista di Studi Liguri 37: 7-32. Chic García, G. 1980. Acerca de un ánfora con pepitas de uvas encontrada en la Punta de la Nao (Cádiz). Boletín del Museo de Cádiz I: 37-42. Colls, D., Étienne, R., Lequément, R., Liou, B. y Mayet, F. 1977. L’épave Port-Vendres II et le commerce de la Bétique à l’époque de Claude. Archaeonautica 1. Comas i Sola, M. y Carreras Monfort, C. 2006. Les àmfores de la Laietania: estat de la qüestió. En A. López Mullor y J. Aquilué Abadías (eds.), La producció i el comerç de les àmfores de la provincia Hispania Tarraconensis. Homenatge a Ricard Pascual Guasch. Actes de les Jornades d’Estudi (17-18 de novembre de 2005). Monografies 8: 177-188. Barcelona, Museu d’Arqueologia de Catalunya. Condamin, J. y Formenti, F.1976. Recherche de traces d’huile d’olive et de vin dans les amphores antiques. Figlina 1: 143-158. Condamin, J. y Formenti, F. 1978. Détection du contenu d’amphores antiques (huiles, vins). Études méthodologiques. Revue d’Archéométrie 2: 43-58. Costantini, A, 2007. Dressel 1 a Populonia: esportazioni di tonno dalla foce dell’Albegna. En D. Vitali (ed.), Le fornaci e le anfore di Albinia. Primi dati su produzioni e scambi dalla costa tirrenica al mondo gallico. Atti del Seminario Internazionale (Ravenna 2006): 151-156. Bologna, Università di Bologna. Delussu, F. y Wilkens, B. 2002. Le conserve di pesce. Alcuni dati da contesti italiani, Mélanges de l’École française de Rome. Antiquité 112-1: 53-65. Desbat, A. 2003. Amphorae from Lyon and the question of Gaulish imitations of amphorae. En Amphorae in Britain and the western Empire. Journal of Roman Pottery Studies 10: 45-49.

21

Miguel Beltrán Lloris

Desbat, A. 2004. Marques et images de marques. En J.-P. Brun, M. Poux y A. Tchernia (eds.) Le vin. Nectar des dieux. Génie des hommes: 291-301. Estrasburgo, Infolio. Desse-Berset, N. y Desse, J. 2000. Salsamenta, garum et autres préparations de poisson. Ce qu’en disent les os. Mélanges de l’École française de Rome. Antiquité 112-1: 73-97. Domergue, C. 1973: Belo I. La stratigraphie. Publications de la Casa de Velázquez Série Archéologie 1. París, de Boccard. Dressel, H. 1879. Di un grande deposito di anfore rinvenuto nel nuovo quartiere del Castro Pretorio. Bullettino della Commissione Archeologica Comunale di Roma 7: 36-112 (part 1); 143-196 (part 2). Dressel, H. 1891. Corpus Inscriptionum Latinarum XV: Inscriptiones urbis Romae latinae. Instrumentum domesticum, Pars I. Berlín, George Reimer. Dressel, H. 1899. Corpus Inscriptionum Latinarum XV: Inscriptiones urbis Romae Latinae. Instrumentum domesticum, Pars II, fasc. I. Berlina, George Reimer. Dyczek, P. 2001. Roman Amphorae of the 1st-3rd Centuries AD Found on the Lower Danube. Typology. Varsovia, Warsaw University Press. Ehmig, U. 1996. Garum für den Statthalter. Eine Saucenamphore mit Besitzeraufschrift aus Mainz. Mainzer Archäologische Zeitschrift 3: 25-56. Ehmig, U. 2000. Dressel 20/30: ex Baetica originalis - imitatio ex Germania superiore. En Congreso internacional Ex Baetica Amphorae : conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III):1143-1152. Écija, Gráficas Sol. Eiring, J. y Finkielsztenj, G. 2004. Concluding Remarks. En J. Eiring y 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: 459-466. Atenas, Danish Institute at Athens. Étienne, R. 1990. Que transportaient donc les amphores lusitaniennes ? En A. Alarção y F. Mayet (eds.), Ânforas lusitanas. Tipologia, produção, comércio/ Les amphores lusitaniennes. Typologie, production, commerce (actas da mesa-redonda de Conímbriga, 1988): 15-19. Conimbriga París, Museu Monográfico de Conimbriga - de Boccard. Étienne, R. y Mayet, F. 1994. À propos de l’amphore Dressel 1C de Belo (Cadix). Mélanges de la Casa de Velázquez 30‑1: 131-138. Étienne, R. y Mayet, F. 2000. Le vin hispanique. Trois clés de l’économie de l’Hispanie romaine 1. París, de Boccard. Étienne, R. y Mayet, F. 2002. Salaisons et sauces de poisson hispaniques. Trois clés de l’économie de l’Hispanie Romaine 2. París, de Boccard. Étienne, R. y Mayet, F. 2004. L’huile hispanique, Trois clés de l’économie de l’Hispanie Romaine 3. París, de Boccard. Ettlinger, E. 1977. Aspects of amphora-typology seen from the North. En G. Vallet (ed.), Méthodes classiques et méthodes formelles dans l’étude typologique des amphores. Actes du colloque de Rome, 27-29 mai 1974. Collection de l’École française de Rome 32: 9-16. Roma, École française de Rome.

Fabião, C. 1989. Sobre as ânforas do acampamento romano da Lomba do Canho. Cadernos da Uniarq 1. Lisboa, Instituto Nacional de Investigação Científica. Fabião, C. 2000a. Sobre as mais antigas ânforas romanas da Baetica no Ocidente peninsular. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (I): 665-682. Écija, Gráficas Sol. Fabião, C. 2000b. O sul da Lusitania (Algarve português e a Baetica: concurrencia ou complementaridade. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (II): 717-730. Écija, Gráficas Sol. Fabião, C. 2004. Centros oleiros da Lusitânia: balanço dos conhecimentos e perspectivas de investigação. En D. Bernal‑Casasola y L. Lagóstena Barrios (eds.), Figlinae Baeticae: Talleres alfareros y producciones cerámicas en la Bética romana (ss. II a.C.-VII d.C.). Actas del Congreso internacional (Cádiz, 12-14 de noviembre de 2003). British Archaeological Reports International Series 1266: 379-410. Oxford, J. W. Hedges. Fabião, C. 2008. Las ánforas de la Lusitania. En D. BernalCasasola y A. Ribera i Lacomba (eds.) Cerámicas hispanorromanas II. Producciones regionales: 725-745. Cádiz, Universidad de Cádiz. Fabião, C. y Guerra, A. 1993. Sobre os conteúdos das Ânforas Lusitanas. En II Congresso peninsular de História antiga (Coimbra, 18 a 20 de outubro de 1990). Actas: 995‑1016. Coimbra, Universidad de Coimbra. Fernández-Chicarro, C. 1953. Actividades arqueológicas de Andalucía. Archivo Español de Arqueología 26: 435‑443. Fernández Izquierdo, A. 1995. Una producción de ánforas de base plana en los hornos romanos del Mas d’Aragó (Cervera del Maestrat, Castellón). Quaderns de prehistòria i arqueologia de Castelló 16: 211-219. Formenti, F. 1989. Analyse de traces organiques dans les amphores. En Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22‑24 mai 1986). Collection de l’École française de Rome 114: 562-563. Roma, École française de Rome. Formenti, F. 1993. Des protéines dans une amphore de type Pompéi VII. Aquitania XI: 251-253. Formenti, F. 2000. Analyse du contenu d’une amphore Gauloise 13 de Tongres. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (I): 1129-1142. Écija, Gráficas Sol. Formenti, F. y Duthel, J. M. 1995. The analysis of wine and other organics inside amphoras of the Roman period. En P. McGovern, S. Fleming y S. Kats (eds.), The Origins of Ancient History of Wine: 79-85. Nueva York, Gordon and Breach Publishers. Formenti, F., Hesnard, A. y Tchernia, A. 1978. Note sur le contenu d’une amphore Lamboglia 2 de l’épave de la Madrague de Giens. Archeonautica 2: 95-101.

22

Ánforas romanas y contenidos. Notas historiográficas

García Fernández, F. J. y García Vargas, E. 2010. Entre gaditanización y romanización: repertorios cerámicos, alimentación e integración cultural en Turdetania (siglos III-I a.C.). En C. Mata Parreño, G. Pérez Jordà y J. Vives-Ferrándiz Sánchez (eds.), De la cuina a la taula. IV reunió d’ economia en el primer mil·lenni AC. Sagvntvm Extra 9: 115-134. València, Universitat de València. García Vargas, E. 1998. La producción de ánforas en la bahía de Cádiz en época romana: (siglos II A.C-IV D.C.). Écija, Gráficas Sol. García Vargas, E. 2000. La producción de ánforas romanas en el Sur de Hispania. República y Alto Imperio. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (I): 57-174. Écija, Gráficas Sol. García Vargas, E. 2004. Las ánforas del vino bético altoimperial: formas, contenidos y alfares a la luz de algunas novedades arqueológicas. En D. Bernal‑Casasola y L. Lagóstena Barrios (eds.), Figlinae Baeticae: Talleres alfareros y producciones cerámicas en la Bética romana (ss. II a.C.-VII d.C.). Actas del Congreso internacional (Cádiz, 12-14 de noviembre de 2003). British Archaeological Reports International Series 1266: 507-514. Oxford, J. W. Hedges. García Vargas, E. 2006. Garum sociorum: pesca, salazones y comercio en los litorales gaditano y malacitano / época altoimperial romana). En C. Tavares da Silva y J. Soares (eds.), Simpósio Internacional Produção e comércio de preparados piscícolas durante a proto-história e a época romana no Ocidente da Península Ibérica. Homenagem a Françoise Mayet (Setúbal, 7-9 Maio 2004). Setúbal Arqueológica 13: 39-56. García Vargas, E. 2012. Producciones anfóricas tardorrepublicanas y tempranoaugusteas del valle del Guadalquivir. Formas y ritmos de la romanización en Turdetania a través del artesanado cerámico. En D. Bernal-Casasola y A. Ribera i Lacomba (eds.) Cerámicas hispanorromanas II. Producciones regionales: 177-206. Cádiz, Universidad de Cádiz. García Vargas, E., Almeida, R. R. de y González Cesteros, H. 2011. Los tipos anfóricos del Guadalquivir en el marco de los envases hispanos del siglo I A.C. Un universo heterogéneo entre la imitación y la estandarización. SPAL. Revista de Prehistoria y Arqueología de la Universidad de Sevilla 20: 185-283. García Vargas, E., Almeida, R. R. de y González Cesteros, H. 2016. Ovoide 4 (Valle del Guadalquivir). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 . García Vargas, E. y Bernal-Casasola, D. 2008. Ánforas de la Bética. En D. Bernal-Casasola y A. Ribera i Lacomba (eds.) Cerámicas hispanorromanas. Un estado de la cuestión: 661-668. Cádiz, Universidad de Cádiz. García Vargas, E., Bernal-Casasola, D., Sáez Romero, A., Díaz Rodríguez, J. J. 2016. Dressel 1 Baetica Ulterior coast. En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 8 julio, 2016 < http://amphorae. icac.cat/amphora/dressel-1-baetica-ulterior-coast>.

García Vargas, E., Conlin, E. y Maestre, C. 2013. El horno altoimperial del Cortijo del Río, Sevilla. Tipología y producciones cerámicas. En D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz y A. M. Saez (eds.), Hornos, talleres y focos de producción alfarera en Hispania. I Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I: 257-294. Cádiz, Universidad de Cádiz. García Vargas, E. y Lavado Florido, M. L. 1996. Definición de dos nuevos tipos de ánforas gaditanas: Las Puerto Real 1 y 2. SPAL. Revista de Prehistoria y Arqueología de la Universidad de Sevilla 20: 197-207. García Vargas, E., Martín-Arroyo, D. y Lagóstena Barrios, L. 2016. Beltrán IIA (Costa de Baetica). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 8 julio, 2016 . Garnier, N., Silvino, T. y Bernal Casasola, D. 2011. L’identification du contenu des amphores : huile, conserves de poissons et poissage. En SFECAG, Actes du Congrès d’Arles: 397-416. Marsella, Société Française d’Étude de la Céramique Antique en Gaule. Giordiano, C. y Casale, A. 1990. Iscrizioni pompeiane inedite scoperte tra gli anni 1954-1978. Atti della Accademia Pontaniana n. s. 39: 273-378. Gisbert Santonja, J. A. 1999. Àmfores i vi al territorium de Dianium (Dénia): Dades per a la sistematització de la producció amforal al País Valencià. En El vi a l’antiguitat: economia, producció i comerç al Mediterrani occidental. Actes del II Col·loqui Internacional d’Arqueologia Romana (Barcelona 6-9 de mai de 1998). Monografies badalonines 14: 383-417. Badalona, Museu de Badalona. Gisbert Santonja, J. A. 2009. Vi tarraconense al país valencià. Una mirada des dels forns d’amfores. Arqueologia de les vil·les i derelictes de la costa de Dianium (Dénia). En M. Prevosti y A. Martin (eds.), El vi tarraconense i laietà ahir i avui, Documenta 7: 125-150. Tarragona: Institut Català d’Arqueologia Clàssica. González Cesteros, H. 2012. Scomber gaditano en Éfeso. Una Dressel 12 con titulus encontrada en la «Casa aterrazada 2» de Éfeso. Dialogues d’Histoire Ancienne 38-1: 111-124. Hahn, J. 1989. Zwei tituli Picti auf südhispanische Amphoren aus Ladenburg und Heidelberg. Fundberichte aus Baden Württtemberg 13: 267-277. Hesnard, A. 1998. S.C.G. (nº 566) et les Dr. 1C/Dr. 12 de Bétique. En V. Blanc-Bijon, M.-B. Carre, A. Hesnard y A. Tchernia (eds.), Recueil de timbres sur amphores romaines, II. Travaux du Centre Camille Jullian 16: 291-293. Aix-en-Provence, Presses de l’Université de Provence. Járrega Domínguez, R., 1998. La producció amforal romana del camp de Tarragona. Estat de la qüestió. En El vi a l’antiguitat: economia, producció i comerç al Mediterrani occidental. Actes del II Col·loqui Internacional d’Arqueologia Romana (Barcelona 6-9 de mai de 1998). Monografies badalonines 14: 430-437. Badalona, Museu de Badalona.

23

Miguel Beltrán Lloris

Járrega Domínguez, R. (ed.) 2018. Estudios sobre ánforas hispanas. Ex Officina Hispana, Cuadernos de la SECAH 3. Madrid, La Ergástula. Jongkees, J. H. 1955. L’enduit intérieur des amphores romaines. Gallia 13: 256-257. 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, Tempus Reparatum. Lagóstena Barrios, L. 2001. La producción de salsas y conservas de pescado en la Hispania Romana II a.C.-VI d.C. Instrumenta 11. Barcelona, Universitat de Barcelona. Lagóstena Barrios, L. 2004. Las ánforas salsarias de Baetica. Consideraciones sobre sus elementos epigráficos. En J. Remesal Rodríguez (ed.), Epigrafía anfórica. Proyecto Amphorae. Instrumenta 17: 197-219. Barcelona, Universitat de Barcelona. Lagóstena Barrios, L. 2014. Nuevas consideraciones sobre la societas cetariorum gaditanorum. En J. M. Maestre Maestre, J. G. Montes Cala, R. J. Gallé Cejudo, C. Macías Villalobos, V. Pérez Custodio, S. I. Ramos Maldonado y M. Sánchez Ortiz de Landaluce (eds.), Baetica renascens I: 93-108. Cádiz-Málaga, Federación Andaluza de Estudios Clásicos; Alcañiz, Instituto de Estudios Humanísticos. Lagóstena Barrios, L., Bernal-Casasola, D. y Arévalo, A. (eds.) 2007. Cetariae 2005: salsas y salazones de pescado en Occidente durante la Antigüedad. Actas del congreso internacional (Cádiz 7-9 noviembre de 2005). British Archaeological Reports International Series 1686. Oxford, J. and E. Hedges Ltd. - Universidad de Cádiz. Laubenheimer, F. 2000. Imitations d’amphores à huile de Bétique dans l’Est et le Nord des Gaules et en Germanie supérieure. En Congreso internacional Ex Baetica Amphorae : conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 1121-1142. Écija, Gráficas Sol. Lequément, R. 1976. Une épave du Bas-Empire dans la baie de Pampelonne (presqu’île de Saint-Tropez), Revue archéologique de Narbonnaise IX: 177-188. Liou, B. 1973. Direction des recherches archéologiques sous-marines. Gallia 31-2: 571-608. Liou, B. 1982. Direction des recherches archéologiques sous-marines. Gallia 40-2: 437-454. Liou, B. 1987. Inscriptions peintes sur amphores: Fos, Marseille, Toulon, Port-la-Nautique, Arles, Saint-Blaise, Saint Martinde-Crau, Mâcon, Calvi. Archaeonautica 7: 55-140. Liou, B. 1993. Inscriptions peintes sur amphores de Narbonne (Port-la-Nautique). Archaeonautica 11: 131-148. Liou, B. 1998. Inscriptions peintes sur amphores de Narbonne, Port-la-Nautique (Aude), III. Revue archéologique de Narbonnaise 31: 91-102. Liou, B. 2000. Las ánforas béticas en el mar. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (I): 1061-1110. Écija, Gráficas Sol. Liou, B. y Rodríguez-Almeida, E. 2000. Les inscriptions peintes des amphores du pecio Gandolfo (Almería). Mélanges de l’École française de Rome. Antiquité 112-1: 7-25.

Long, L. y Delauze, H.-G. 1996. Bouches du Rhône, au large de Marseille. L’épave Sud-Caveaux 1, une nouvelle expérience en matière d’archéologie sousmarine profonde. En Bilan scientifique du Département des Recherches Archéologiques Subaquatiques et SousMarines: 84-86. Marsella, Ministère de la Culture et de la Communication. López Mullor, A. 2009. Els centres productors d’àmfores de Sant Boi de Llobregat i Darró (Vilanova i la Geltrú). En M. Prevosti y A. Martin (eds.), El vi tarraconense i laietà ahir i avui, Documenta 7: 61-98. Tarragona: Institut Català d’Arqueologia Clàssica. López Mullor, A. 2013. La figlina i les àmfores de l’àrea de les termes romanes de Sant Boi de Llobregat. En C. Carreras Monfort, A. López Mullor y J. Guitart (eds.) Barcino II. Marques i terrisseries d’àmfores al Baix Llobregat. Corpus international des timbres amphoriques 18: 61-81. Barcelona, Institut d’Estudis Catalans i Institut Català d’Arqueologia Clàssica. López Mullor, A. y Martín Menéndez, A. 2008a. Tipologia i datació de les àmfores tarraconenses prodüides a Catalunya. En A. López Mullor y J. Aquilué Abadías (eds.), La producció i el comerç de les àmfores de la “Provincia Hispania Tarraconensis”. Homenatge a Ricard Pascual i Guasch: 33-94. Barcelona, Museu d’Arqueologia de Catalunya. López Mullor, A. y Martín Menéndez, A. 2008b. Las ánforas de la Tarraconense. En D. Bernal-Casasola y A. Ribera i Lacomba (eds.) Cerámicas hispanorromanas. Un estado de la cuestión: 689-724. Cádiz, Universidad de Cádiz. Lund, J. 2004. Oil on the waters? Reflections on the contents of Hellenistic transport amphorae from the Aegean. En J. Eiring y 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: 11-216. Atenas, Danish Institute at Athens. Manacorda, D. 1977. Anfore spagnole a Pompei. En L’Instrumentum domesticum di Ercolano e Pompei nella prima età imperiale. Quaderni di Cultura Materiale 1: 121-133. Roma, «L’Erma» di Bretschneider. Marichal, R. 1975. Paléographie latine et française. En Annuaire 1974/1975. Annuaires de l’École pratique des hautes études, 531-532. Marichal, R. y Liou, B. 1993. Inscriptions peintes sur amphores de Narbonne (Port-la-Nautique). Archaeonautica 11: 131-148. Martin-Kilcher, S. 1994. Die römischen Amphoren aus Augst und Kaiseraugst, 2-3. Die Amphoren für Wein, Fischsauce, südfrüchte (Gruppen 2-24) und Gesamtauswertung. Forschungen in Augst 7/2-3. Augst, Römermuseum Augst. Martin-Kilcher, S. 2000. Amphores à sauces de poisson du sud de la péninsule ibérique dans les provinces septentrionales. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 759-786. Écija, Gráficas Sol.

24

Ánforas romanas y contenidos. Notas historiográficas

Martin-Kilcher, S. 2002. Lucius Urittius Verecundus, négociant à la fin du Ier siècle, et sa marchandise découverte à Mayence. En L. Rivet y M. Sciallano (eds.), Vivre, produire et échanger : reflets méditerranéens, Mélanges offerts à Bernard Liou: 343-353. Archéologie et Histoire romaine 8. Montagnac, Éd. Monique Mergoil. Martin-Kilcher, S. 2004. Fischsaucen: Pinselaufschriften und römische Amphoren. En J. Remesal Rodriguez (ed.), Epigrafia anforica. Instrumenta 17: 245-257. Barcelona, Universitat de Barcelona. Martin-Kilcher, S. y Schillinger-Häfele, U. 1989-1990. Eine südhispanische Amphore für Fischsauce aus MainzWeisenau mit Pinselaufschrift. Mainzer Zeitschrift. Mittelrheinisches Jahrbuch für Archäologie, Kunst und Geschichte 84-85: 274-277. Martínez Maganto, J. 2000. Inscripciones sobre ánforas de salazón: interpretación sobre la estructura y significación comercial de los tituli picti. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (IV): 1207-1219. Écija, Gráficas Sol. Martínez Maganto, J., García Jiménez, R. y BernalCasasola, D. 1997. Ánforas del Museo de Ceuta, Serie Maior, Informes y catálogos 1. Ceuta, Publicaciones del Museo Municipal de Ceuta. Martínez Maganto, J. y Petit Domínguez, D. M. 1998. La pez y la impermeabilización de envases anfóricos romanos. Estudio analítico de una muestra e interpretaciones histórico-económicas. Archivo Español de Arqueología 71: 265-274. Mateo Corredor, D. y Molina Vidal, J. 2016. Oliva 3 (Costa central de Tarraconensis). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 26 octubre, 2016 . Mayet, F. 1999. La production d’amphores Dressel 1C et Dressel 12 dans le détroit de Gibraltar. En Mélanges C. Domergue 2. Pallas 50: 53-61. Mayet, F. y Tavares da Silva, C. 2002. L’atelier d’amphores d’Abul, Portugal. París, de Boccard. Miró Canals, J. 1988. La producción de ánforas romanas en Catalunya. Un estudio sobre el comercio del vino de la Tarraconense (siglos I a. C.-I d. C.). British Archaeological Reports International Series 473. Oxford, British Archaeological Reports Publishing. Miró Canals, J. 2016. Tarraconense 1 (Costa septentrional de Tarraconensis). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 10 julio, 2016 . Molina Vidal, J. 2000. Las primeras exportaciones béticas en el Mediterráneo Occidental. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (II): 637-646. Écija, Gráficas Sol.

Morais, R. 2003. Bracara Augusta: um pequeno Testaccio de ânforas Haltern 70. Considerações e problemáticas de estudo. En D. Bernal‑Casasola y L. Lagóstena (eds.), Figlinae Baeticae. Talleres alfareros y producciones cerámicas en la Bética romana (ss. II a.C. - VII d.C.). Actas del Congreso Internacional (Cádiz, 12-14 de noviembre de 2003). British Archaeological Reports International Series 1266 (II): 546-566. Oxford, John W. Hedges. Morais, R. 2005a. Urceus (Valle del Guadalquivir). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 6 agosto, 2016 . Morais, R. 2005b. Autarcia e Comércio em Bracara Augusta. Bracara Augusta, Escavações Arqueológicas 2. Braga, Unidade de Arqueologia da Universidade do Minho. Morais, R. 2012. Novos dados sobre as ânforas vinarias béticas de tipo urceus. SPAL. Revista de Prehistoria y Arqueología de la Universidad de Sevilla 17: 267-280. Morais, R. 2017. Ánforas de época altoimperial. En Fernández Ochoa, C., Morillo Cerdán, Á. y Zarzalejos Prieto, M., (eds.). Manual de cerámica romana III. Cerámicas romanas de época alto imperial III: cerámica común de mesa, cocina y almacenaje. Imitaciones hispanas de series romanas. Otras producciones: 311-380. Madrid, Comunidad de Madrid. Morillo Cerdán, Á. 2000. Ánforas y envases perecederos: Nuevas aportaciones sobre la comercialización de aceite bético durante la época romana en la región septentrional de la península ibérica. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (II): 621-636. Écija, Gráficas Sol. Nieto Prieto, J., Jover Armengol, A., Izquierdo Tugas, P., Puig Griessenberger, A. M., Alaminos Exposito, A., Martin Menendez, A., Pujol Hamelink, M., Palou Miquel, H. y Colomer Marti, S. 1989. Excavacions arqueològiques subaquàtiques a Cala Culip 1. Centre d’Investigacions Arqueològiques de Girona. Sèrie Monogràfica 9. Girona, Centre d’Investigacions Arqueològiques de Girona. Noguera Celdrán, J. M. y Antolinos Marín, J. A. (eds.) 2011-2012. De vino et oleo Hispaniae. Áreas de producción y procesos tecnológicos del vino y el aceite en la Hispania romana. Anales de prehistoria y arqueología 27-28. Murcia, Universidad de Murcia. Nolla Brufau, J. M. y Solias, J. M. 1985. L’àmfora Tarraconense 1. Característiques procedència, àrees de producció, cronologia. Butlletí Arqueològic De Tarragona V, 6-7:107-144. Notarstefano, F. 2012. Ceramica e alimentazione. L’analisi chimica dei residui organici nelle ceramiche applicata ai contesti archeologici. Beni archeologici conoscenza e tecnologie, Quaderno 10. Bari, Edipuglia. O’Kelly, J. 2012. Las ánforas onubenses de época tardorromana. En D. Bernal-Casasola y A. Ribera Lacomba (eds.). Cerámicas hispanorromanas. Un estado de la cuestión: 279-298. Cádiz, Universidad de Cádiz.

25

Miguel Beltrán Lloris

Oddy, W. A. y Bateman, J. 1973. Analysis of amphora linings. En M. J. Blackman (ed.), Marine Archaeology. Colston Papers 23: 381. Bristol, Archon Books. Panella, C. 2001. Le anfore di età imperiale del Mediterraneo occidentale. En P. Lévêque, J. P. Morel y É. Geny (eds.), Céramiques hellénistiques et romaines III: 177-275. París, Presses Universitaires Franc-Comtoises. Parker, A. 1977. Lusitanian Amphoras. En G. Vallet (ed.), Méthodes classiques et méthodes formelles dans l’étude typologique des amphores. Actes du colloque de Rome, 27-29 mai 1974. Collection de l’École française de Rome 32: 35-46. Roma, École française de Rome. Pascual Guasch, R. 1992. Index d’estampilles sobre àmfores catalanes. Cuadernos de Arqueología 5. Barcelona, Servei del Llibre de L’Estaquirot. Peacock, D. P. S. y Williams, D. F. 1986. Amphorae and the Roman Economy. An Introductory Guide. Londres - Nueva York, Longman. Pecci, A. y Cau Ontiveros, M. Á. 2010. Análisis de residuos orgánicos en ánforas: el problema de la resina y el aceite. En J. M. Blázquez Martínez y J. Remesal Rodríguez (eds.), Estudios sobre el Monte Testaccio (Roma), V. Instrumenta 35: 593-600. Barcelona, Universitat de Barcelona. Pérez Rivera, J. M. 2000. Las imitaciones de ánforas grecoitálicas e itálicas en el Sur de la península ibérica. En Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (I): 227-238. Écija, Gráficas Sol. Pesavento, S. T. y Buonopane, A. 2002. Alcuni tituli picti su anfore di produzione betica rinvenute nel porto di Pisa. En L ‘Africa Romana XIV. Lo spazio marittimo del Mediterraneo occidentale: geografia storica ed economia: 789-800. Roma, Carocci. Pinto, I. V. y Magalhães, A. P. 2016. Almagro 51A-B (Lusitania occidental). Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 08 julio, 2016 . Radulescu, A. V. 1973. Amfore cu inscripţii de la Edificiul roman cu mozaic din Tomis. Pontica 6: 193-207. Ramon Torres, J. 1995. Las ánforas fenicio-púnicas del Mediterráneo central y occidental. Instrumenta 2, Barcelona, Universitat de Barcelona. Ramon Torres, J. 2008. Les àmfores altimperiales d’Ebusus. En A. López Mullor y J. Aquilué Abadías (eds.), La producció i el comerç de les àmfores de la «Provincia Hispania Tarraconensis». Homenatge a Ricard Pascual i Guasch. Actes de les Jornades d’Estudi (17-18 de novembre de 2005). Monografies 8: 241-270. Barcelona, Museu d’Arqueologia de Catalunya. Ramon Torres, J. 2013 a. Ramon PE 25 (Isla de Ibiza). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 8 julio, 2016 . Ramon Torres, J. 2013 b. Ramon PE 26 (Isla de Ibiza). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 .

Remark, P. 1912. De amphorarum inscriptionibus Latinis quaestiones selectae. Dissertatio inauguralis. Tubinga, Laupp. Remesal Rodríguez, J. 1991. El aceite bético durante el Bajo Imperio. Antigüedad y cristianismo: Monografías históricas sobre la Antigüedad tardía 8: 355-362. Revilla Calvo, V. 1993. Producción cerámica y economía rural en el Bajo Ebro en época romana. El alfar de l’Aumedina, Tivissa, Tarragona. Instrumenta 1. Barcelona, Universitat de Barcelona. Revilla Calvo, V. 2004. Ánforas y epigrafía anfórica en Hispania Tarraconensis. En J. Remesal Rodríguez (ed.), Epigrafía anfórica. Proyecto Amphorae. Instrumenta 17: 159-196. Barcelona, Universitat de Barcelona. Rodríguez Temiño, I. 1990. Hallazgos de dos ánforas con tituli picti en Écija (Sevilla). Archivo Español de Arqueología 63: 292-295. Roman Amphorae: a digital resource, University of Southampton, 2005 (updated 2014) . Sáez Romero, A. M. 2016a. Ramon T-8211 (Costa de Ulterior/Baetica). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 . Sáez Romero, A. M. 2016b. Ramon T-9111 (Costa de Ulterior/Baetica). Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 . Sáez Romero, A. M., Bernal‑Casasola, D., García Vargas, E. y Díaz Rodríguez, J. J. 2016. Ramon T-7433 (Costa de Baetica). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 10 julio, 2016 . Sáez Romero, A. M. y Niveau, A. M. 2016. Pellicer D (Costa de Ulterior/Baetica). Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 . Schoene, R. y Mau, A. 1909. Corpus Inscriptionum Latinarum IV, Supplementum, pars II. Inscriptiones parietariae et vasorum fictilium. Berlina, George Reimer. Silvino, T. y Poux, M. avec la collaboration de N. Garnier 2005. Où est passé le vin de Bétique ? Nouvelles données sur le contenu des amphores dites « à sauces de poisson et à saumures » de type Dressel 7/11, Pompei VII, Beltrán II (Ier s. av. J.-C.-IIe s. apr. J.-C.). En SFECAG, Actes du Congrès de Blois: 501-514. Marsella, Société Française d’Étude de la Céramique Antique en Gaule. Simões, V. M. 2016. Haltern 70 (Lusitania occidental). En Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 july 2016 . Tavares da Silva C., y Soares J. 2006. Simpósio Internacional Produção e comércio de preparados piscícolas durante a proto-história e a época romana no Ocidente da Península Ibérica. Homenagem a Françoise Mayet (Setúbal, 7-9 Maio 2004). Setúbal Arqueológica 13.

26

Ánforas romanas y contenidos. Notas historiográficas

Tchernia, A. 1986. Le vin de l’Italie romaine. Essai d’histoire économique d’après les amphores. Bibliothèque des Écoles françaises d’Athènes et de Rome 261. Roma, École française de Rome. Tchernia, A. 1992. Débat. En F. Laubenheimer (ed.), Les Amphores en Gaule: production et circulation. Annales littéraires de l’université de Besançon 213: 207-218. París, Presses Universitaires de France. Tchernia, A., Girard, M. y Poplin, F. 1986. Pollens et ossements animaux de l’épave romaine 3 de Planier (Provence). En L’exploitation de la mer de l’Antiquité à nos jours, 2. La mer comme lieu d’échanges et de communication. VIèmes Rencontres internationales d’archéologie et d’histoire d’Antibes, 24-26 octobre 1985: 231-256, Valbonne, Association pour la promotion et la diffusion des connaissances archéologiques. Tremoleda i Trilla, J. 2000. Industria y artesanado cerámico de época romana en el nordeste de Cataluña (época augustea y altoimperial). British Archaeological Reports International Series 835. Oxford, J. and E. Hedges.

Tremoleda J., Castanyer P., Puigdevall, I. y Dehesa R. 2006. La bòbila romana d’Ermedàs i la seva producció (Cornellà del Terri, Pla de l’Estany, Catalogne). En SFECAG. Actes du Congrès de Pézenas: 477-493. Marsella, Société Française d’Étude de la Céramique Antique en Gaule. Vidal Solá, C. y Pascual Guasch, R. 1971. El pecio de Palamós. En Actas del III Congreso Internacional de Arqueología Submarina (Barcelona, 1961): 117-126. Bordighera, Istituto Internazionale di Studi Liguri. Viegas, C. 2016. Dressel 14 (Lusitania meridional). Amphorae ex Hispania - I. Paisajes de producción y de consumo, consultado 20 julio, 2016 . Will, E.1987. The Roman amphoras. En A. M. McCann (ed.), The Roman Port and Fishery of Cosa: 202-203. Princeton, Princeton University Press. Zevi, F. 1966. Appunti sulle anfore romane, la tavola tipologica del Dressel. Archeologia Classica XVIII: 208-247.

27

Historical and archaeological indicators

Amphorae: typology and contents Stefanie Martin‑Kilcher

Universität Bern Institut für Archäologische Wissenschaften Abt. Archäologie der Römischen Provinzen 3012 Bern, Suisse

André Tchernia

EHESS & Centre Camille Jullian Aix Marseille Univ, CNRS, CCJ, Aix‑en‑Provence, France

Abstract: The main principles upon which the determining of the contents of an amphora type is based are listed and discussed. At the end, their use is exemplified through a study of the contents of the Haltern 70 amphorae: they transported olives preserved in defrutum and, seldom, defrutum alone. Key words: Early and mid-roman amphora types; content; analysis; critique of the sources; interpretation; case study Haltern 70.

1. Introduction Do we all agree that what we want to do is to ascertain the contents of amphora types, not of individual amphorae? We assume that the title given to the opening lecture of this session means that we do. If we keep this goal in mind, individual amphorae should be regarded as hopefully representative samples of a population, which we call an amphora type.1 What, then, is an amphora type? An ideal amphora type is a group of amphorae which share common features in the fields of form and contents as well as generally in chronology and provenance. The form of every amphora cannot be identical. It is more or less narrowly defined, with differences due to the multiplicity of kilns and craftsmanship traditions or to the chronological evolution. But one should be able to isolate the type, through form (and fabric), from the rest of the amphorae. The distinctive features can be very obvious or sometimes very subtle. The larger the series, the clearer the differences. As some forms have been produced in different regions (one thinks of the wine amphorae Dr 1 and Dr 2-4), the cases in which the provenance should be considered as a typological criterium have to be examined and discussed. The same is true for chronology. The long‑lived types undergo an evolution and change in form. And finally, it

 Thanks for information to: Miguel Beltrán, Laurence Benquet, Darío Bernal‑Casasola, Michel Bonifay, César Carreras, Enrique García Vargas, David Djaoui, Fabienne Olmer, Séverine Lemaître, and for the translation Michael Dietler and Andrew Lawrence. Paper submitted in 2017.

is the archaeologists who define a type, thus mirroring the respective state of research. Chronology: The time span of production can be short or very long, but should have a beginning and an end. Production Area: It is ascertained through the distribution of kilns, and more and more through analysis of the fabric. It can be small or large but should be circumscribed. We leave aside for the moment the question of indistinguishable or nearly indistinguishable copies or extents of a production which add patches to the main production area (as e.g. the Pascual 1 in Southern Gaul). The contents are the same. Their form remains identical to its model’s while other kind of imitations, which we would like to call filiations or inspirations, evolve in their own way (e.g. the huge family of the Dr 2-4, but also the amphorae of the middle Rhône valley). Contents: Our basic assumption is that a mass produced amphora type has been conceived or adopted when a region began to commercialize its content. Therefore, we consider that an amphora type is meant to contain a commodity specific to its production area (Figure 1). The content can be more or less homogeneous, but should be definable. We have entered the core of our topic.2

1

 After the conference in Cadiz, the paper by Bernal‑Casasola 2015 was published, which also focussed on this topic. 2

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 31–40

Stefanie Martin-Kilcher and André Tchernia

Figure 1. Ostia. Forum delle corporazioni, statio 51. Although the inscription of the navicularii is not preserved, the ship carrying olive oil amphorae Dr 20 must hail from Baetica (Italica or Gades).

Figure 2. Marbella. Mosaic in the porticus of a villa depicting (amongst other things): kitchen implements, food, ingredients (including an amphora Beltran IIB) and prepared food.

2. Is this statement sensible? It has already been said that ancient Greeks and Romans knew the contents and provenance of an amphora by looking at it. If we may start from the Greek period, let us remember that, when Cassander in 316 BC founded Cassandreia, which was going to act as a port for the neighbouring vineyard of Mende, Lysippus drew an original form of amphora which was meant to be Mende’s own to satisfy Cassander’s desire of fame (φιλοδοξοῦντι καἱ βουλομένῳ ἴδιόν τινα εὑρέσθαι κέραμον; Athénée 784c). Latin authors, not archaeologists, have created the words vinarius, olearius, salsamentarius and applied them to containers (amphora, urna, cadus) (Cato, 10, 2: amphorae oleariae; 135: urnae oleariae, vinariae, which are bought in Capua; Columella, II, 16: lentils should be kept in vasis oleariis salsamentariisque; Pliny XVIII, 307: fabam et legumina in oleariis cadis oblita cinere longo tempore servari, beans can be kept a long time in oleariis vasis topped with ashes; 308 legumina […] in salsamentariis cadis gypso inlinant; XXXII, 89: Sic et ad parotidas utuntur […] quin et testis cadi salsamentarii tusis cum axungia uetere, for salivary glands disease, some people recommend crushed sherds of a cadus salsamentarius mixed up with old grease.

Figure 3. Dougga. Section of a mosaic: A slave is pouring wine from an amphora MR1 into a drinking cup.

A passage from Galen proves that more accurate information could also be retrieved from an amphora form: ‘Think of the Aminean wine of Sicily, the one which is contained in big amphorae. Because the one transported in small lagynoi [that is flat based amphorae] is exactly the contrary, it is harmful to the stomach and provokes headaches’ (Galen, Kühn, X: 834–835).

1. What we know about the production of the commodities in the production area. 2. What we know about the amphorae production in the production area of the commodities. 3. Painted inscriptions on amphorae. 4. Traces of contents.- Macroscopically visible, or invisible and evidenced through chemical analysis. 5. Images.- Notably mosaics like the Marbella one with an amphora Beltrán IIb alongside cooking implements and carefully arranged food (locally sourced fish and meat) (Figure 2), or the Dougga mosaic, with a slave pouring wine from a Mid‑roman 1 amphora (Figure 3). Or a funeral monument on the river Moselle showing a ship with wine‑barrels and imported wine-amphorae from Souterh Gaul (Figure 4). 6. Compatibility contents/characteristics of the amphora.- No large pieces of salted fish in an amphora

The principle of a relationship between type and content may admit some exceptions or individual solutions. It cannot be invalidated. 3. How do we determine the content of an amphora type? Through different kinds of evidence, based on field archaeology and written sources. We have listed eight; we are however well aware that the examples at hand concern the early and middle imperial era.

32

Amphorae: typology and contents

Figure 4. Neumagen. Funeral monument of a wine trader with a wine ship, laden with barrels probably containing wine from the Moselle region and strawwrapped amphorae from southern Gaul. Rheinisches Landesmuseum Trier, Foto Th. Zümer. with a narrow opening; no oil in an amphora with visible — we insist on visible — pitch lining. 7. Global forms.- It is clear that a Dr 2-5, whatever its provenance, was meant to contain wine. The same is true for the flat-based lagynoi of the first century AD and later. On the other hand, things are not so clear for oil amphorae, though, generally speaking, it is admitted that they should be big bellied and short necked. Behind the main forms of the Empire, lie two very different models. On one hand, the ovoid Late Republican amphorae, used as models for the oil amphorae of Italy (e.g. the Brindisi amphorae: Manacorda and Pallecchi 2012) and later in Spain;3 on the other, several late Punic cylindrical amphorae from North Africa (MartinKilcher 2011; Bonifay 2004: 89–98 fig. 46). However, both basic shapes have been, from the first century BC onwards, the basis for a variety of types used for a variety of contents. Among the ovoid, we find for instance the amphorae of the Palamos wreck (Martín Menéndez 2008) (Figure 5): Their insides were treated differently from those of the Brindisi oil amphorae, as they are lined with pitch. These amphorae have been principally produced in Southern Spain (ovoides gaditanas and other types, especially for salted fish products) and a very few in Tarraconensis (stamp L. Volteili): they belong to different types and the contents are different. From the first century BC onwards, a global repertoire of standardized forms, not only as far as wine is concerned but for many other goods was established: in several production areas, different types of containers for long-distance trade were developed for the growing number of different products. A well-known amphora type could be adopted and refined. Or else, it could be reshaped with some different morphological

Figure 5. Palamós (Illes Formigues I). Two of the ovoid amphorae (ovoides gaditanas) from the shipwreck, around 50 BC. Martín Menéndez 2008: Fig. 4.





elements, such as the neck or the handles. Therefore, this argument should be handled with caution. It needs an exact definition of the amphora type (not only by fragments of the rim or neck) and the chronology should be taken into account. In many regards, the first century BC is a period of experimentation, rendering a systematization of types difficult. On these grounds, interesting questions can be raised. For instance some Belo (and other sites in Southern Spain) amphorae (Figure 6), linked to the production of salsamenta, are, on account of their form, supposedly the descendants of the Dr 1C (Domergue 1973; Hesnard 1998; Mayet 1999; Étienne and Mayet 2002), traditionally considered as wine amphorae. However, there could be a chronological gap, since the Italian Dr 1C are dated to around the turn of the second and the early decades of the first century BC. According to the stratigraphically most useful contexts, these types of amphorae seem to be produced in Belo for a relative short period in the second quarter of the first century BC (‘sondage 29, couche Vb’: Domergue 1973: 41).4 Excavation data from related productions seem to point to the earlier first century BC (BernalCasasola et al. 2011), but are based on typological datings from the associated amphorae. Should we nevertheless contemplate the possibility that the contents were the same and consequently the Italian Dr 1C were amphorae salsamentariae?

 The layers correspond with the backfilling of a large basin for the fabrication of fish products. The layer can be dated to the second quarter of the first century BC. 4

 We are omitting other production regions, such as Tripolitania, but also Lusitania (see e.g. Morais 2004). See now García Vargas et al. (eds.) 2019. 3

33

Stefanie Martin-Kilcher and André Tchernia

The Monte Testaccio is another case of significant archaeological context, since it is a dump of oil amphorae. The odds are that an amphora found there has contained oil. However, it is not completely certain. In much lesser quantities, some wine amphorae have been recovered from the Monte Testaccio: Gaulish, Cretan, and even some Dr 2-4 (Carreras Monfort 1999). These are examples of the caution necessary when using some of the indicators of contents we have listed. And that brings us to the fourth point. 4. Occasions of errors Obviously, before reaching a conclusion on contents, the problems, occasions of errors, and discussions are plenty. The snags lie either in establishing the data or in generalizing them. 4.1. Data

Figure 6. Belo. The only completely reconstructed amphora ‘Dr 1C’. Mayet 1999: Fig. 1,1.

For the first section, we have listed 4 main possibilities:

(e.g. Will 1987; Hesnard 1998; Benquet and Mancino 2006). To avoid falling into a circular argument, the zones and period of production of the Italic Dr 1C and their variants must be discussed in more detail (e.g. Benquet and Olmer 2002). If we are actually dealing with vessels for fish products, then production sites for conserved fish on a large scale would have to be located on the Tyrrhenian coast, as well as those linked to the Dr 21/22 (Botte 2009).5 Pottery kilns further inland wouldn’t make sense. On the other hand, biochemical analyses would make sense, as the hypothesis of Dr 1C as a amphora salsamentaria has been repeatedly postulated. 8. The archaeological context of finds.- The particular position within a funerary context is an argument which has been used for the Dr 7-11 amphorae found in aristocratic tombs of the second half of the first century BC and the first half of the first century AD in northern Gaul and Britain. Poux and Silvino (2005) immediately concluded that their position proved they contained wine and therefore that these types were multipurpose. But there is another solution (Martin-Kilcher 2000: 770; 2011): mixed beverages based on garum, but with the addition of wine, as they are described by Pliny (NH 31, 44, 95; 36, 95): ‘Transiit deinde in luxuriam, creveruntque genera ad infinitum, sicuti garum ad colorem mulsi veteris adeoque suavitatem dilutum, ut bibi possit’. Martial (Epigrams VII, 27, 8) describes ‘addet et arcano mixta Falerna garo’. Biochemical analyses of some of these amphorae may yield further information.

1. Painted inscriptions They are not easy to read and their abbreviations are sometimes difficult to understand. A recent example is Dressel’s (1879: 172) reading CER on Dr 21/22, which he interpreted (with caveat) as cerasa, cherries. That has been accepted for a century and a half, until Emmanuel Botte (2009: 128ss) read CET instead of CER, and therefore cer(asa), cherries, became cet(us), tunny. New discoveries can also change long surmised expansions, as David Djaoui (2016a) recently demonstrated with the example of LACCAT( ) when he found a more complete inscription beginning with LACERT(us) mackerel. It occurs not infrequently that the interpretation of a painted inscription remains extremely doubtful. That is why each reading should be checked and more weight should be given to obvious readings and expansions than to disputed readings and far fetched expansions. It is not shameful to delay the solution. The method is to rely on series, not on isolated finds in order to recognise the typical formulas of the inscriptions. 2. Analysis Is a non-scientific person allowed to surmise that errors may occur in interpreting the chemical results, as they occur in interpreting inscriptions or other data? Risks arising from contamination during the subterranean life of the sherds, or from degradation of the contents with the passing of time have recently been emphasized (Garnier 2016: 43). 3. Errors about production areas An old and therefore innocent example is Héron de Villefosse’s 1915 article on Sextus Fadius Secundus Musa

 At Cosa, the installations may concern the fishery (McCann 1987), but definitely not the production of conserved fish. 5

34

Amphorae: typology and contents

and Olitius Apollonius whose names are inscribed in position β on Dr 20 amphorae. Since he had demonstrated they were from Narbonne, he thought that he had also proved that Dr 20 were produced in Gaul and therefore contained wine. He openly contradicted Dressel, and this was unfortunately accepted by several scholars (e.g. Cantarelli 1915; see on this question Thevenot 1959). We could add the example of the LRA 1, considered by Tchalenko 1953-1958 as oil containers, because he thought that their trade was linked to the villages and presses of Northern Syria, and that these had produced olive-oil: a double error. It has now been proved that the LRA 1 principal, although not exclusive, content was wine (Pieri 2005: 69). 4. Errors on typology Errors in the typological assignation of sherds bearing a painted inscription or having benefited from an analysis are of course possible. And so much the more when we consider that many inscriptions belong to, or analysis are made from, comparatively small sherds whose type is not always easy to identify. This is also true when the typological assignation is based on not always reliable, or very small scale, drawings, and the fabric is not, or insufficiently, described.

Figure 7. Amphora Dr 24 and the rim of a Knossos 18. Rizzo 2014: fig. 42a and 43b. wine seems the most probable content. A few amphorae have been recovered from Monte Testaccio. Painted inscriptions on these amphorae are regularly in Greek. Two inscriptions in Latin on amphorae recovered from Monte Testaccio are said to mention oleum (Carreras Monfort 1999: 97–98). The inscriptions are most probably secondary. These amphorae hail from the Greek-speaking East and most of them carry Greek painted inscriptions. It would be wrong and hasty, judging on the two Latin inscriptions, to assume that Dr 24 was a container for olive oil. The evidence points clearly to wine vessels.

But there is a more interesting case: when the study of contents induces one to question the unity of a type and to divide it into several different ones. In 1966, Fausto Zevi still considered the debate on the Dr 6 contents open, because the division between Dr 6A and Dr 6B was not yet clear. Recently Woodworth et al. 2015 proposed that a part of the so called Keay 25 had contained wine, and they split the type into 25, 1, 2 and 3, of which several of the 25, 1 contained wine and the 25, 3 salsamenta.

Before concluding, it is safe to organize the different kinds of data in series, to check their consistency and reliability. Then, the series are combined, thus avoiding giving priority to any type of evidence.

As a rule, if contradictions are revealed in the matters of contents or chronology, the validity of the typology and its definition through form and fabric should be immediately questioned, as the following examples underline.

We will emphasize two problems: 1. Reuse of amphorae Reuse of amphorae is not scarce, and Th. Peña 2007 has thoroughly studied its modes. He will presently tackle the question (see Peña in this volume). In brief, the ultimate content of an amphora found on excavations is not necessarily its original contents. As a consequence, archaeological and epigraphic contexts need to be taken into account when interpreting the inscriptions and analysis. A few ship wrecks, such as Sud-Caveaux 1 and Grado (Long and Delauze 1996: 84–86; Auriemma 2000: 27–51) were principally loaded with reused amphorae. But usually, an amphora found alongside many other identical ones in a wreck is less liable to have been reused than an amphora found in a house.

4.2. Generalizing The most critical phase is when an archaeologist extends to the whole type the comparatively few data he has gathered. Whatever the kind of data he resorts to, the responsibility of the final conclusion is entirely his. At this stage, all the different data are put together. We can see for instance several of our headings at stake in the question of the Dr 24 and Knossos 18, which were regarded as one of the same type and whose contents and production areas were discussed (see a summary in Rizzo 2014: 318–322; Lemaître 2000). In the meantime, they have been revealed to be two chronologically and formally different types (with different sub-types) (Figure 7).

The writing of the inscriptions also gives hints. It is important to check whether the colour, the size and ductus of the letters, the position, composition and phrasing of the inscriptions are consistent and make

Most kilns for Dr 24 have been discovered in cities well known for their wine, such as Clazomenes or Erythraei;

35

Stefanie Martin-Kilcher and André Tchernia

series or not. A number of inscriptions of the same type points to primary contents, while one or two isolated ones do not. 2. Occasional contents The second problem arises from occasional or irregular, in Peña’s words, contents, and from subcategories. In other words, homogeneous product does not always mean identical product. In the written sources, olive-oil is just divided into great production areas, Baetica, Istria, Africa, and some Italian regions. The quality was linked to the provenance. The language of the oil amphora was clear enough. By looking at an oil amphora, the customer got the information he needed (see Figure 1). Conversely, a wine amphora, although containing wine from one specific region, may have contained different types of wine, and different vintages. Hence the usefulness of inscriptions as vetus, mulsum, passum, aminaeum, or Vesuvinum, Surrentinum and so on. These are even more useful in the case of salsamenta. The contents are homogeneous as they are all based on fish, but the treatments create great differences and the fish species are not the same. Hence, the many inscriptions that were necessary to convey this information to the customer.

Figure 8. Ostia. Painted inscription acetum lauronense, vinegar from Lauro on a Dr 2-4 from Tarraconensis. Panella 1970: 131 and Tav. 36, no.561.

The occasional (or irregular) contents are somehow different. Clementina Panella (1970) published a painted inscription acetum lauronense, vinegar from Lauro on a Dr 2-4 from Tarraconensis (Figure 8). That did not upset anyone’s applecart and no one claimed that the territory of Lauro was dedicated to vinegar production; it was immediately understood that vinegar was a by‑product of wine, an occasional, or irregular content, as the inscription indicated.

5.1. Quite a few painted inscriptions A series of undisputable inscriptions (the reference base is overall the Moros and Berni Millett 2004 catalogue):6 Olives preserved in defrutum: Four from three sites (Mainz, 2; Soissons; Vindonissa). Three are identical, olivae nigrae ex defruto (cat. 7, 12, 34); one very probably [olivae] ex defr(uto) excel(lente) penuar(iae) (Mainz, cat. 29).7

An Africana I amphora found in the Plemmirio B wreck contained olive kernels (Parker 1992: 319). But that does not mean that the Africana I are not oil amphorae. They have transported oil and occasionally olives. The contents of the Dr 2-4 and Africana I amphorae can still be considered as preserving their identity as products of the vines or of the olive-trees.

Defrutum by itself: A series of three in the Port-Vendres II shipwreck (cat. 15, 23, 24); two at Pompéi shipped by the same negotiator (cat. 1 et 2): Def(rutum) excel(lens). Sapa: One along the handle, which is unusual, and might denote a secondary inscription (Amiens, cat. 26). The ductus of the letters pleads also for a secondary inscription.

5. Haltern 70: olive containers We would like to finish with a practical exercise: to reconsider, bearing these principles in mind, the case of the Haltern 70, which are the subject, as says Giorgio Rizzo 2014: 202–206, of a ‘conflitto delle interpretazioni’, although in his statistics he considers them as wine amphorae. As is the case for several others, the type is created in the first century BC after the ovoid forms of the Late Republic.

A series of disputed or doubtful inscriptions: Impossible to decide between olives preserved in defrutum or defrutum by itself: Three on three sites (Pompéi, cat. 13: exc(ellens); Augst, cat. 31: [ ]EFR. Anse St-Gervais defr(utum)  After the conference, Djaoui 2016b published several new examples from Arles. 7  Ol(ivae) al(bae) (ex) dulci (Mainz cat. 33) is clearly on a amphora Ha 70 sim. 6

We have three types of data for its contents:

36

Amphorae: typology and contents

excell(ens) on a sherd of uncertain typology (Liou and Marichal 1978: no.35). Doubtful: Olives: two (Pisa, cat. 9: [olivae ] ex defrut(o) vet(ere); Mainz, cat. 41: [olivae ex ex]cel(lente) dul[ci]. Discussed between Mulsum and Muria: Two on two sites (Celsa, cat. 3; Zaragoza, cat. 11). Mulsum should be preferred in both cases. In the first one, the abbreviation muri(a) would be very strange, and the reading of the following word as pen(uaria) is pure fancy; in the second one, because the gap between u and i (?) is too big to be filled only with an r and allow a reading Mu[r]ia.8

1

Disputed Muria: Three on three sites. (MV(?) Pisa, cat. 14); Mur(ia) vet(us) Mainz cat. 32; Mur(ia) arg(uta), Vindonissa) (Figure 9). For the two last ones see below. 5.2. Traces of contents Olive kernels were found in amphorae from the Sud‑Lavezzi II and Culip IV shipwrecks (Liou and Domergue 1990: 29; Nieto Prieto et al. 1989: 76) as well as in an amphora at Fos (Marty et al. 2016). We add the Cevera wreck near Santa Pola (Bernal‑Casasola 2015: 66). An analysis seems to have individuated traces of both olives and defrutum (Juan-Tresserras 1999). The typology of the sherds from Broach of Gurness (van der Werff 2002: 447), which had traces of fish products is unreliable.

2 Figure 9. Vindonissa (1) and Mainz (2). The two massive amphorae with primary painted inscription with large letters show MVR(ia) as their content. The amphorae are probably not Ha 70, but early forms of Dr 14. Photos and Drawings: Vindonissa Museum and S. Martin-Kilcher.

5.3. Production areas The identified kilns are first of all located in the Guadalquivir valley, where their production is associated with Dr 20. The similarity of fabric between Ha 70 and Dr 20, has been noticed in several importation sites and gives a decisive argument for their association.

bear the inscription MVR(ia) written with big letters (Figure 9) are not Ha 70 containing a secondary product, but early forms of the fish-sauce amphora Dr 14, whose exact provenance is as far as now unknown, but pertains certainly to Baetica (Marismas?) (Martin-Kilcher et al. 2009: 356–362; 2011). If the MVR(ia) dipinti can be classed as primary — and everything point in this direction — then we are dealing with two amphora types.

Other points are located on the Marismas zone, in the bays of Cadiz and Algeciras. There, the production of Ha 70 is associated with the production of amphorae for salsamenta (e.g. García Vargas 1998; Bernal‑Casasola and Lagóstena 2004; García Vargas et al. 2011). But an overview on the arborescent amphora production on the south coast of the Iberian Peninsula is actually quite difficult to ascertain on the basis of the published material (mostly rim fragments), and chemical analyses are difficult to interpret (Mateo Corredor 2015).

As for the Ha 70 from the Guadalquivir valley, our interpretation is the following. Olives, when transported, had to be kept in a preserving liquid: the defrutum. This preserving liquid had to be produced. It meant planting vineyards around the olive‑trees. Indeed, olive trees were not an exclusive culture even in the Guadalquivir valley: Remember Pliny XVIII, 93: Bætica quidem uberrimas messes inter oleas metit. If too much defrutum had been made, it could be exported by itself. But it was, as it were, a kind of by‑product of the olives trade. So was the mulsum, which may sometimes have accompanied it as another dulce vinum.

Furthermore, it seems possible that two examples, so far assigned to the Ha 70 type, both massive amphorae from Vindonissa and Mainz (and also recently reported from the Rhône at Arles: Djaoui 2016b: fig. 24) which 8

 See the discussion by J. van der Werff 2002.

37

Stefanie Martin-Kilcher and André Tchernia

To give priority to olives, we rely also on the following argument: the olive trade is mentioned in written sources, while the defrutum trade is not. Defrutum is just grape juice reduced by heating to a third or two thirds of it original volume. The inscriptions on Ha 70 are the only source that bears witness to its overseas trade. We have otherwise no hint that it was a long distance trade commodity. The amphora defriti mentioned in the Tabula Pompeiana Sulpiciorum 80 (Camodeca 1999) was almost certainly not coming from far.

Bibliography Auriemma, R. 2000. Le anfore del relitto di Grado e il loro contenuto. Mélanges de l’École française de Rome. Antiquité 112-1: 27–51. Benquet, L. and Mancino, C. 2006. Les amphores d’Albinia : première classification. In SFECAG, Actes du Congrès de Pézenas: 465–476. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Benquet, L. and Olmer, F. 2002. Les amphores. In J. M. Blázquez Martínez, C. Domergue and P. Sillières (eds), La Loba (Fuenteovejuna, Córdoba). La mine et le village minier antiques: 295–331. Bordeaux, Ausonius Publications. Bernal-Casasola, D. 2015. What contents do we characterise in Roman amphorae? Methodological and archaeological thoughts on a trending topic. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 61–83. Esposende, Município de Esposende. Bernal-Casasola, D. and Lagóstena, L. 2004. Figlinae Baeticae. Talleres alfareros y producciones cerámicas en la Bética romana (ss. II a.C.-VII d.C.). Actas del Congreso Internacional (Cádiz, 12-14 de noviembre de 2003). British Archaeological Reports International Series 1266. Oxford, John W. Hedges. Bernal-Casasola, D., Roldán Gómez, L., Blánquez Pérez, J. and Sáez Romero, A. M. 2011. La producción anfórica de Carteia en época republicana. Primeras evidencias. In J. Abellán Pérez, C. Lazarich González and V. Castañeda Fernández (eds), Homenaje al profesor Antonio Caro Bellido: 63–80. Cadiz, Universidad de Cadiz. Bonifay, M. 2004: Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports International Series 1301. Oxford, Archaeopress. Botte, E. 2009: Salaisons et sauces de poissons en Italie du sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31. Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Camodeca, G. 1999. Tabulae Pompeianae Sulpiciorum (TPSulp.). Edizione critica dell’Archivio Puteolano dei Sulpicii. Vetera 12. Rome, Quasar. Cantarelli, L. 1915. Il Monte Testaccio e i vini della Gallia. Bullettino della Commissione Archeologica Comunale di Roma 43: 41–46. Carreras Monfort, C. 1999. Miscelánea: las otras ánforas del Monte Testaccio (1991-1992). In J. M. Blázquez Martínez and J. Remesal Rodríguez (eds), Estudios sobre el Monte Testaccio I. Instrumenta 6: 91–98. Barcelona, Universitat de Barcelona. Djaoui, D. 2016a. The myth of ‘Laccatum:’ a study starting from a new titulus on a Lusitanian Dressel 14. In I. V. Pinto, R. R. de Almeida and A. Martin (eds), Lusitanian Amphorae: Production and Distribution. Roman and Late Antique Mediterranean Pottery 10: 117–127. Oxford, Archaeopress. Djaoui, D. 2016b. Les amphores de type Haltern 70: olives ou defrutum ? Et que faire des tituli picti concernant la sapa, le mulsum, la muria et l’hallex ? In D. Djaoui (ed.), Histoires matérielles : terre cuite, bois, métal et autres objets. Des pots et des potes: Mélanges offerts à Lucien Rivet. Archéologie et Histoire Romaine 33: 489–512. Autun, Éd. Monique Mergoil.

The matter is different with olives, too often forgotten in the lists of possible amphora-borne commodities. When Pliny XV, 16 says: ‘olives from overseas are considered as better for eating than the Italian ones, but they are inferior for oil-making’9 it makes clear that an export trade of olives towards Italy did exist. The Baetican ones arrived up to the second century AD with the Ha 70 (or their late form respectively). The combined evidence, if one gives preference to well established series, as we advocate, clearly points first to olives preserved in defrutum, then to defrutum or mulsum alone. 6. Brief conclusion No source can be omitted in ascertaining the contents of an amphora type. Each has to be assessed. The best situation is when we can cross‑check data retrieved from archaeology (including painted inscriptions and scientific analysis), with the more global views of written sources, which archaeology often helps us to better understand. New discoveries do not cancel the evidence drawn from other sources. And if new data seem to bring an upheaval, they should be inserted in a thorough survey and re‑examination of all the evidence so far available, of the typological assignments and of the typology itself. Any endeavour aiming to solve pending issues should give priority to the establishment of a clear typology and to the choice of reliable and representative samples. Otherwise, accumulating data might prove disappointing, with the results being questioned and more time wasted than scientific progress being achieved.

9

 Italicis transmarinae praeferuntur in cibis, cum oleo vincantur.

38

Amphorae: typology and contents

Dressel, H. 1879. Di un grande deposito di anfore rinvenuto nel nuovo quartiere del Castro Pretorio. Bullettino della Commissione Archeologica Comunale di Roma 7: 36–112 (part 1); 143–196 (part 2). Domergue, C. 1973. Belo I. La stratigraphie. Publications de la Casa de Velázquez, série archéologie I. Paris, de Boccard. Étienne, R. and Mayet, Fr. 2002. Salaisons et sauces de poisson hispaniques. Trois clés pour l’économie de l’Hispanie romaine, II. Paris, de Boccard. García Vargas, E. 1998. La producción de ánforas en la bahía de Cádiz en época romana: (siglos II A.C-IV D.C.). Ecija, Gráficas Sol. García Vargas, E., Almeida, R. R. de and González Cesteros, H. 2011. Los tipos anfóricos del Guadalquivir en el marco de los envases hispanos del siglo I a.C. Un universo heterogéneo entre la imitación y la estandarización. SPAL, Revista de Prehistoria y Arqueología de la Universidad de Sevilla 20: 185–283. García Vargas, E, de Almeida, R.R., González Cesteros, H., Sáez Romero, A.M. (eds) 2019. The Ovoid Amphorae in the Central and Western Mediterranean. Roman and Late Antique Mediterranean Pottery 13. Oxford, Archaeopress. Garnier, N. 2016. Quel rôle pour les chimistes dans les recherches en archéologie ? In D. Djaoui (ed.), Histoires matérielles : terre cuite, bois, métal et autres objets. Des pots et des potes: Mélanges offerts à Lucien Rivet. Archéologie et Histoire Romaine 33: 31–49. Autun, Éd. Monique Mergoil. Héron de Villefosse, A. 1914. Deux armateurs narbonnais, Sex. Fadius Secundus et P. Olitius Apollonius. Mémoires de la Société des Antiquaires de France 74: 153–180. Hesnard, A. 1998. S.C.G (No 566) et les Dr.1C/Dr.12 de Bétique. In V. Blanc-Bijon, M.-B. Carre, A. Hesnard and A. Tchernia (eds), Recueil de timbres sur amphores romaines 2: 291-293. Aix-en-Provence, Publications de l’Université de Provence. Juan-Tresserras, J. 1999: Estudio analítico de los residuos conservados en el interior de recipientes y asociados a materiales de molienda y trituración. In XXII col·loqui internacional per a l’Estudi del Edat del Ferro. Els productes alimentaris d’origen vegetal a l’edat del Ferro de l’Europa Occidental: de la producció al consum. Sèrie monogràfica del Museu d’Arqueologia de Catalunya, 18: 371–377. Girona, Generalitat de Catalunya. Lemaître, S. 2000. Les importations d’amphores de Méditerranée orientale à Lyon au IIIe siècle ap. J.-C. Rei Cretariae Romanae Fautorum Acta 36: 467–486. Liou, B. and Marichal, R. 1978. Les inscriptions peintes sur amphores de l’anse Saint-Gervais à Fos-sur-mer. Archaeonautica 2: 109–181. Liou, B. and Domergue, C. 1990. Le commerce de la Bétique au Ier siècle de notre ère. L’épave Sud‑Lavezzi 2 (Bonifacio, Corse du Sud). Archaeonautica 10: 11–123. Long, L. and Delauze, H.-G. 1996. Bouches du Rhône, au large de Marseille. L’épave Sud-Caveaux 1, une nouvelle expérience en matière d’archéologie sousmarine profonde. In Bilan scientifique du Département des Recherches Archéologiques Subaquatiques et SousMarines: 84–86. Marseille, Ministère de la Culture et la Communication.

Manacorda, D. and Pallecchi, S. (eds) 2012. Le fornaci romane di Giancola (Brindisi). Bari, Edipuglia. Martin-Kilcher, S. 2000. Amphores à sauces de poisson du sud de la péninsule ibérique dans les provinces septentrionales. In Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 759–786. Ecija, Gráficas Sol. Martin-Kilcher, S. 2011. Formes d’amphores et contenu au Haut Empire, points de repère et questions. In SFECAG, Actes du Congrès d’Arles: 417–426. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Martin-Kilcher, S., Tretola Martínez D. C. and Vogt, R. 2009. Die Amphoren aus dem Grabbezirk von GoeblingenNospelt. In J. Metzler, C. Gaeng, I. Le Goff, S. MartinKilcher, P. Méniel, D. C. Tretola Martínez, R. Vogt, R. Weiller and J.-M. Welter (eds), Goeblange-Nospelt, une nécropole aristocratique trévire. Dossiers d’archéologie du Musée National d’Histoire et d’Art 13: 333–394. Luxembourg, Musée National d’Histoire et d’Art. Martín Menéndez, A. 2008. Àmfores tarraconenses i bètiques en els derelictes de mitjan segle I a.C. a la costa catalana. In SFECAG, Actes du Congrès de l’EscalaEmpúries: 103–127. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Marty, F., Bouby, L., Ivorra, S. and Terral, J.-F. 2016. Conserves d’olives hispaniques en amphores, au premier siècle, sur le site de l’Estagnon (Fos-sur-Mer, Bouches-du-Rhône, Fr). In D. Djaoui (ed.), Histoires matérielles : terre cuite, bois, métal et autres objets. Des pots et des potes: Mélanges offerts à Lucien Rivet. Archéologie et Histoire Romaine 33: 481– 487. Autun, Éd. Monique Mergoil. Mateo Corredor, D. 2015. Producción anfórica de la costa malacitana desde la época púnica hasta el periodo julioclaudio. Lucentum 34: 183–206. Mayet, F. 1999. La production d’amphores Dressel 1C et Dressel 12 dans le détroit de Gibraltar. Pallas 50: 53–61. McCann, A. M. 1987. The Roman Port and Fishery of Cosa. Princeton, Princeton University Press. Morais, R. 2004. Problemàtiques i noves perspectives sobre les àmfores tardo-republicanes. Les àmfores ovoides de producció Lusitana. In C. Carreras Monfort et al. (eds), Culip VIII i les àmfores Haltern 70. Monografies del Centre d’Arqueologia Subaquàtica de Catalunya, 5: 36–40. Girona, Museu d’Arqueologia de Catalunya and Centre d’Arqueologia Subaquàtica de Catalunya. Moros, J. and Berni Millet, P. 2004. Catalogue. Epigrafia sobre Àmfores Haltern 70 Bètiques. In C. Carreras Monfort et al. (eds), Culip VIII i les àmfores Haltern 70. Monografies del Centre d’Arqueologia Subaquàtica de Catalunya, 5: 51–72. Girona, Museu d’Arqueologia de Catalunya and Centre d’Arqueologia Subaquàtica de Catalunya. Nieto Prieto, J., Jover Armengol, A., Izquierdo Tugas, P., Puig Griessenberger, A. M., Alaminos Expósito, A., Martín Menéndez, A., Pujol Hamelink, M., Palou Miquel, H. and Colomer Martí, S. 1989. Excavacions arqueològiques subaquàtiques a Cala Culip 1. Centre d’Investigacions Arqueològiques de Girona. Sèrie Monogràfica 9. Girona, Centre d’Investigacions Arqueològiques de Girona.

39

Stefanie Martin-Kilcher and André Tchernia

Panella, C. 1970. Le anfore. In F. Berti, E. Fabbricotti and A. Carandini (eds), Ostia II. Le terme del Nuotatore. Scavo dell’ambiente V. Studi Miscellanei 16: 102–156. Rome, L’Erma di Bretschneider. Parker, A. J. 1992. Ancient Shipwrecks of the Mediterranean and the Roman Provinces. British Archaeological Reports International Series 580. Oxford, Tempus Reparatum. Peña, J. T. 2007. Roman Pottery in the Archaeological Record. Cambridge, Cambridge University Press. Pieri, D. 2005. Le commerce du vin oriental à l’époque byzantine (Ve-VIIe siècles). Le témoignage des amphores en Gaule. Bibliothèque Archéologique et Historique 174, Beirut, Institut français du Proche-Orient. Rizzo G. 2014. Le anfore, Ostia e i commerci mediterranei. In C. Panella and G. Rizzo (eds), Ostia VI. Le terme del nuotatore. Studi Miscellanei 38: 65–481. Rome, L’Erma di Bretschneider. Silvino, T., Poux, M. and Garnier, N. 2005. Où est passé le vin de Bétique ? Nouvelles données sur le contenu des amphores dites « à sauces de poisson et à saumures » de types Dressel 7/11, Pompéi VII, Beltrán II (Ier s. av. J.-C.-IIe s. apr. J.-C.). SFECAG, Actes du Congrès de Blois: 501–514. Marseille, Société Française pour l’Étude de la Céramique Antique en Gaule.

Tchalenko, G. 1953-1958. Villages antiques de la Syrie du Nord, Le Massif de Bélus à l’époque romaine. Bibliothèque archéologique et historique 50. Paris, Libraire Orientaliste Paul Geuthner. Thevenot, E. 1959. En marge d’une histoire du vin en France. Revue archéologique de l’Est et du Centre-Est 10: 308–345. Van der Werff, J. 2002. Old and new evidence about the contents of Haltern 70 amphoras. In L. Rivet. and M. Sciallano (eds), Vivre, produire et échanger: reflets méditerranéens. Mélanges offerts à Bernard Liou. Archéologie et Histoire Romaine 8: 445–449. Montagnac, Éd. Monique Mergoil. Will, E. L. 1987. The Roman Amphoras. In A. M. McCann (ed.), The Roman Port and Fishery of Cosa: 170–220. Princeton, Princeton University Press. Woodworth, M., Bernal‑Casasola, D., Bonifay, M., Garnier, N., Keay, S., Pecci, A., Poblome, J., Pollard, M., Richez, F. and Wilson, A. 2015. The content of African Keay 25 / Africana 3 amphorae: initial results of the CORONAM project. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 41–57. Esposende, Município de Esposende. Zevi, F. 1966. Appunti sulle anfore romane. Archeologia Classica 18: 208–247.

40

Shipwrecks, amphorae and contents Franca Cibecchini

DRASSM Underwater Archaeology Research Department Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France

Abstract: Without touching on the specifics (covered by some papers in the present volume), the main objective of this paper is to stimulate a constructive debate about various issues and problems around the question ‘Why are shipwrecks so important when talking about contents?’. This question gathers the concept of the contextual capsule, and the notion of the conservation condition of artefacts and materials underwater. A series of examples and topics will be touched upon and highlighted such as: from the pure traces of the content of the amphorae to chemical analyses; from the importance of the discovery’s context to epigraphic analysis (tituli picti), and from the re-examination of old data to the wide frontiers opened by new technologies for future research on deep‑sea wrecks. Key words: Ancient shipwrecks; closed context; amphora; artefacts association; wine; fish sauces; chemical analysis; archaeological fish remains analysis; tituli picti; recycled containers; deep-sea wrecks.

1. Introduction In this paper I highlight a range of issues and problems, hoping to stimulate debate, without going into the specific details of the cases I am discussing. Moreover, many of the themes and subjects that converge in this paper, such as epigraphs or recycling of containers, have been specifically addressed in special sessions of this congress, to which you can refer for further detailed analysis. Why are shipwrecks so important when talking about contents? The question may seem a banal one, or even a little stupid, but I think that it is important to first clarify the strengths, even the exclusiveness, of data collected in this type of archaeological context. In my opinion, there are two key factors: 1. First, the generally excellent conservation conditions of artefacts and materials in underwater contexts. This is true not only for shipwrecks, but for all underwater contexts, especially sites submerged in rivers and lakes (or in the case of ancient underwater sites, such as harbours or anchorages that are now wetlands). 2. Secondly, a shipwreck is a closed context. 2. Conservation of artefacts To highlight the better conservation of artefacts and materials — especially of so-called ‘perishable materials’ — in underwater sites, I will just briefly revisit some main benefits of this type of context.

Dressel 1 amphora from the Madrague de Giens shipwreck, the cover of Andre Tchernia’s monograph on the wine of Roman Italy (Tchernia 1986; Tchernia, Pomey and Hesnard 1978: 13). Remember also the many piles of fish-bones or handfuls of fish-bones, another well-known image from Bernard Liou on of the South Lavezzi 2 wreck (Liou and Domergue 1990). Indeed, it is largely thanks to wrecks that we established the first real evidence about the contents of some forms of amphorae. Likewise, in the history of research on amphorae, the determination of the presence of wine in Lamboglia 2 amphorae, previously believed to contain olive oil, was made thanks again to Madrague de Giens (Formenti, Hesnard and Tchernia 1978). More recently, the image of Lamboglia 2 / Dressel 6 amphorae full of pitch (Figure 1) from Sud-Caveaux 1 (Long 1998), indicates without a shadow of doubt that they were recycled containers. Beyond these well-known cases, I would like to mention a few instances where less ‘obvious’ or unexpected contents were discovered in wrecks or, more generically, in submerged contexts. Unexpected contents such as olive pits or pitch were found in some Etruscan Py 1/2 and Py 3 amphorae, usually containing wine, in the archaic wreck of Giglio Island (Bound 1991: 203–209, fig. 8 and 14).1 Or the volcanic sand and fruits (walnuts, peaches, cherries, olives and plums) found on the San Rossore B wreck in Pisa, contained in Lamboglia 2 wine amphorae (Camilli 2004: 65–66; Camilli and Setari 2005: 57).

2.1. Contents To recall the often excellent preservation of contents in containers recovered from shipwrecks we need only remember the famous image of wine flowing from a

 The bottom of a Py 3 is full of pitch and the site was blanketed with a thick layer of pitch. This layer has enabled the recovery of most of the olive pits, more than 800 in total, and seems to indicate the presence of significant quantities of pitch on board. 1

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 41–49

Franca Cibecchini

Figure 1. Sud-Caveaux 1 shipwreck (France). Sampling of a Lamboglia 2 amphora full of pitch (©F. Bassemayousse). But traces of contents may be less obvious, though still visible, as in the case of the cork of a Greek-Italic amphora soaked in wine residues recovered at Puntone di Scarlino, a site that has been explained as resulting from the clean-up of the hold of a merchant ship (Cibecchini and Bargagliotti 2011: 87–94). Wine residues were invisible on the cork, but left traces of purple on the white cloth in which the cork was wrapped when removed from the water.

determining, on this basis, ‘solid’ versus ‘liquid’ contents and more specifically differentiating between oil and wine. Today, thanks to chemical analysis we know that the practice of coating containers with pitch was a common one, and extended to all kinds of containers, including those for oil (Garnier, Silvino and Bernal‑Casasola 2011). Again, the contribution of Nicolas Garnier and other colleagues address this issue in more detail. For my part, I would simply like to stress that for underwater archaeologists this recent discovery is not unsettling, for indeed the lack of visible traces of pitch in a container from an underwater site is an important clue for the presence of oil. This is due to the fact that the presence of oil causes a chemical reaction that apparently makes pitch disappear.

In recent years we have seen an increase in the accuracy of chemical analyses in determining contents.2 This method has shed light on numerous difficult cases… and created many more! Despite some additional complications, these analyses produce equally good results in the marine environment, but on this issue we refer to the contributions of other colleagues, particularly those of Nicolas Garnier and Alessandra Pecci.

2.3. Tituli picti

Concerning the determination of contents, we must always stress the importance of carefully studying and sieving the contained material, even when this appears to be unnecessary. A good sampling procedure and the correct storage of the object once out of the water are equally important.

Once again, there is an important contribution in this volume dedicated to inscriptions so I will just revisit the fundamental contributions of underwater archaeology to the corpus of tituli picti, by citing a number of new case studies. I cannot stress enough the exceptional conservation of many of these artefacts, especially the painted inscriptions from submerged sites in the Rhône River at Arles (Ehmig 2009; Silvino et al. 2015: 647–650, Djaoui, Piquès and Botte 2014: 182–183). The recent study of the Lusitanian amphorae from the harbour dump site containing the Roman shipwreck Arles-Rhône 33 revealed the painted inscription LAC[--] on the neck of a small

2.2. Coating Let’s recall that on some artefacts from submerged sites we have discovered for the first time the evidence of traces of pitch coating on the interior of different types of containers. The presence or absence of these pitch traces has led archaeologists to speculate about the possibility of  See, for example, the progress made with the determination of protein species, Dallongeville et al. 2011 and 2013. 2

3

42

 Djaoui 2016; Djaoui and Quaresma 2016, especially 360–361.

Shipwrecks, amphorae and contents

A

B

Figure 2. The Roman Bou-Ferrer wreck (Spain). A perfectly preserved amphorae ‘in situ’ during the excavation of the deeper layers in 2013 (2a); one of the amphorae bearing traces of labels for tituli picti (2b) (©J. A. Moya). Dressel 14. The term LAC[--], previously unknown in this type of amphora, has permitted David Djaoui to develop LAC[--] into lac(ertus), a mackerel, referring to the name of the fish from which the liquamen is made.4 This allowed the author to review the entire epigraphic directory of laccatum, now enriched by the new discoveries in the Rhône.

perfectly preserved amphorae collected during the last two campaigns from the deeper layers (Figure 2a). Only the traces of labels for tituli picti have been observed on three specimens from the deeper layers (Figure 2b). As we can see, the question remains unresolved.

The recent Arles-Rhône underwater excavations have shown us that an incredible variety of objects, sometimes surprising, could bear painted inscriptions. The quantity of painted inscriptions recovered on Baetican amphorae for fish products (fish sauces or salted fish) seems to confirm some scholars’ hypothesis concerning the quite systematic presence of these inscriptions, which are fragile and have only been preserved in special cases.5

Regarding this exceptional wreck, the Bou-Ferrer, I would like to offer a methodological remark and relate an anecdote. All of the amphorae from the first layer were filled with large amounts of seashells. These shells were less present in the second layer, and almost nonexistent in the third, and better preserved, layer of cargo. Sometimes the shells were clearly very old and their percentages in the amphorae were highly significant. Nevertheless, the possibility that it was a ‘seafood’ trade of some kind was never considered, although other scholars have suggested this hypothesis during several conferences or seminars.

2.4. Knowledge of the environment

However, the excavation of the Roman Bou-Ferrer wreck (La Villajoyosa, Alicante) seems to contradict or at least mitigate this hypothesis. About 300 Dressel 7-11 amphorae have been recovered from this wreck, which has been under excavation since 2006 (De Juan, Cibecchini and Miralles 2014, De Juan, Cibecchini and Vento 2012). None of them bear painted inscriptions. Not even the numerous

The gradual sifting (3 and 1mm) of a representative sample of amphorae, even from the first layer, and the analysis of the remains made by Gaël Piques (see his contribution in this volume), have clearly shown that the amphorae were filled with a fairly thin fish-sauce made mainly from anchovies (Engraulis encrasicolus), mackerel (Scomber sp.) and horse mackerel (Trachurus sp.). But leaving aside for the moment the results of the content analysis, we were already sure that it was not a seafood

 Djaoui 2016. I would like to thank David for providing data, often unpublished, and for frequent and fruitful scientific discussions. 5  For the epigraphy known on amphorae of Dressel 7-11 family, see an overall picture in Lagóstena 2004, and the observations of Martin‑Kilcher 2011. 4

43

Franca Cibecchini

Figure 4. The Capo Sagro 2 shipwreck (Corsica). The filter jug containing ‘a family’ of at least a dozen crabs of different sizes.

Figure 3. The Cap Bear 3 shipwreck (France). An octopus nesting in an amphora (©P. Foliot, CNRS).

We should always be very attentive to the archaeological and environmental context, and to its documentation. It is vital to bear in mind that the contents of an amphora found in the sea may have been ‘contaminated’, perhaps even in ancient times.

transport. In fact, any underwater archaeologist with solid field experience, at least in the Mediterranean, knows that this phenomenon is fairly common and that it has two rather simple explanations: octopuses and/or the colonisation of ceramic containers by different marine organism (shellfish and bivalves).

It is also essential to look carefully at objects recovered from underwater sites, especially if the context indicates some anomalies. We know that a reused container may be identified through a combination of two to three features, like marks (use-wear and applied marks like graffiti), problems of association among the objects, great heterogeneity and so on.7 And this can be determined even before analysis or the sifting of contents.

The octopus loves to nest in amphorae (Figure 3), jars or similar containers, an unfortunate habit for him at times and the origin of a traditional fishing method that is still practiced in Tunisia. This cephalopod also loves to bring home food and ‘souvenirs’, the latter especially for covering the entrance to his ‘home’. This common habit of octopuses not only explains the abundant presence of shells in the amphorae, but may also be the cause of some ‘fairly curious presences’ inside these containers. We can cite, for instance, the presence of small objects or fragments, such as those found in some Dressel 1 amphorae from the wreck of the Madrague de Giens (Tchernia, Pomey and Hesnard 1978: 26), as examples of well-known cases of octopus creations.

2.5. Cold cases A great amount of data comes from early underwater excavations. Even in the 80s and 90s sampling procedures, or even the desalting of ceramic containers were quite rare. Moreover, these containers from old excavations have typically been moved several times, sometimes broken, and in any case they have been abundantly manipulated, which means contaminated. Nevertheless, these containers can still be fruitfully examined, if samples are collected with the suitable precautions, as demonstrated by some recent projects involving the Drassm, CCJ-CNRS, and private researchers.8

For the biological colonisation of ceramic containers, I can cite a clear example from the Capo Sagro 2 shipwreck, lying 500m deep off Corsica.6 According to the number of crab claws found inside an intact filter jug (Figure 4), at least ten specimens of different sizes (claws from 2 millimetres to more than one centimetre) have grown… and died inside this vessel.

We can also cite the analysis of some Baetican amphorae of the Dressel 7-11 family from Sud Lavezzi 2 and Pisa‑San Rossore (Silvino, Poux and Garnier 2005: 508–511) that indicates the presence of wine on amphorae believed

 This shipwreck (−500 meters deep) with a main cargo consisting of tin ingots, for an estimated load of approximately 40 tonnes, was discovered in 2014 and documented in 2015 during a deep-water shipwreck mission in Corsica by the Drassm, Cibecchini, Rico and Poveda 2018. 6

 An overview about the reuse of amphorae in ancient maritime transport in Abdelhamid 2013, with bibliography. 8  Woodworth et al. 2015. The majority of the results are still unpublished. 7

44

Shipwrecks, amphorae and contents

without handles, and one tubular amphora (Figure 5) are visible.10 The production of the tubular amphora is believed to have started no earlier than the middle of the 2nd century BC (Botte 2012; Pascual and Ribera 2014). Its presence in this well dated cargo allows us to raise the start date of tubular amphorae by at least 25-35 years (Cibecchini 2015: 10). 3.2. Association of artefacts I will revisit just one case: the dolia shipwrecks. Since the early studies in the 1970s, it has been assumed that these large jars contained wine, based on the systematic association of dolia with wine amphorae (especially Dr. 2‑4) and the presence of a coating of pitch. This hypothesis, endorsed by most scholars, has been confirmed in recent years by chemical analysis carried out on the pitch of a dolium from the Ouest Giraglia 2 wreck that made it possible to exclude the transport of fish sauces.11

Figure 5. Alistro 1 shipwreck (Corsica). Detail of the tubular amphora. (©DRASSM). to be exclusively used for fish products. A presence seemingly confirmed by the tituli picti in new examples, although the presence of fish derivatives still makes up the majority.9 I bring up these cases only to stress that unfortunately one or two analyses do not allow us to truly answer the fundamental question: is this an exceptional case, more or less isolated or is it a particular form/variant type, produced in a particular period in a given geographical area, to transport unusual contents?

Strictly related to the association of objects on a wreck and to the composition of the cargo, we have the concept of the ‘function of the container’ on a ship.

Without suggesting that all the amphorae of a wreck should be analysed, I believe that only with a proper number of samples, selected in a reasoned way and analysed with established parameters, can we hope to answer all of these questions convincingly.

3.3. Function of the container It is important to determine the function of a container: if it is part of the cargo, whether principal or secondary, or rather an object used by sailors on board, and equally, if this was its first use or if it was being reused. There is great variability in interpretation: we move from small batches to fully charged ships with ‘recycled’ containers, a system involving significant organization. This begs the question of why, and in which cases, were one or more types of containers recycled to carry ‘something different’ from the original contents (Abdelhamid 2013).

3. Shipwrecks as closed contexts 3.1. Dating and composition of cargo Without dwelling too much on the subject, I need to stress the importance of closed-context wrecks and the criticality of correctly identifying them: this has been thoroughly discussed in a famous debate by A. Tchermia in the 1990s (Tchernia 1990). I believe that today the progress made by underwater archaeological studies allows us to be rather ‘for’ than ‘against’ shipwrecks.

I want to stress just one thing. While some studies attest that we can explain the re-use of containers because ‘perhaps there was no specific amphora production’… I think that this assumption is not fully satisfactory.

I would just like to mention a wreck recently discovered and documented in Corsica at more than 400 meters deep, called Alistro 1. This is a cargo of Greco-Italic amphorae, homogeneous and intact, dating from the early 2nd century BC (Cibecchini 2015: 9–10). Among the more than one thousand (at least) Greco-Italic amphorae, two Punic amphorae, probably Ramon T.7‑3.2.1, curiously

Let us look, for example, at the case of the Grado shipwreck, one of the most famous cases of ‘recycled cargo’.12 But with the ‘recycled amphorae’ from around  The type is near of that know in Corinth, Botte 2012: 595, fig. 4, n. 4-5.  Cibecchini, De Juan and Marlier 2017 and Cibecchini 2020. Unpublished analysis by N. Garnier in 2011 revealed no traces attributable to the presence of wine, or even traces of fat or protein. In the absence of traces of the two latter elements, the transport of wine seems more likely, it being the most evanescent element of the three. A similar conclusion led to some paleo-botanic analyses of the pitch of a Diano Marina wreck’s dolium, Arobba, Caramiello and Martino 1997‑1998. For a recent point on chemicals analysis to determine wine content, see Garnier 2015. 12  Auriemma 2000. A cargo of about 600 amphorae, of which more than 200 were African I (usually an oil container), followed by some Tripolitanian 1 (again for oil) and different wine amphorae (150 Knossos A/53 and some Forlimpopoli), all filled with fish products. 10

11

 To my knowledge the cases of the presence of wine, still remain limited to rare examples, whether the results of analyses (two Dressel 8 and one Pompei VII from the Sud Lavezzi 2 wreck and one Dressel 8 and one Dressel 7 from Pisa-San Rossore — Silvino, Poux and Garnier 2005: 510–511 — with an error concerning the citation of the Sud-Perduto 2 wreck) or painted inscriptions (the two Beltran IIB from Arles and from the Saint Gervais 3 wreck — see Ibidem — a Dressel 7-8 specimen from Pisa San Rossore — Barreca 2006) and one or two new pieces from the Rhône River at Arles, still mostly unpublished, Djaoui, Greck and Marlier 2011: 59–60. 9

45

Franca Cibecchini

the Mediterranean there are also small amphorae, ‘Grado 1’, produced in the Adriatic area (Mazzoli, Maritan and Pesavento 2009; Cipriano and Mazzochin 2012: 249) for transporting ‘liquamen’, as clearly indicated by wellpreserved tituli picti. In addition, this same region is well known for the production of amphorae from the Republic to the High-Imperial period, such as the Lamboglia 2 and Dressel 6 types.

studies where the excellent conservation conditions are permitting new chemical research on organic contents. I’m talking about some small pots with two handles called ‘pot du Latium’ in French. The discussion surrounding this particular container also concerns its final destination. The number of specimens found in the Rhône and their frequency in harbour contexts, has led to the hypothesis that they were not part of the commercial cargo but rather, were used on board for the nourishment of the sailors (Djaoui, Piqués and Botte 2014). Some specimens similar to the so-called ‘pot du Latium’ are known in PisaSan Rossore (Figure 6). Like the others from Latium, they clearly contained fish products, as they were discovered full of fish bones, but in this case, the production seems to be local.13

Even more obvious is the case of the Sud-Caveaux 1 wreck mentioned above. The pitch contained in the Lamboglia 2/ Dressel 6 seems to come from an area well known for amphora production, the Tarraconensis region, as shown even by the wine amphorae Tarraconensis 1A and B that complete this cargo (Long 1998). In this case was it a desire to avoid using a type of container too closely associated with the wine transport at a specific time, or perhaps, more simply that old containers, normally not older than a dozen years, were re-used for a ‘secondary’ trade?

Even more interesting, is the discovery of tituli picti on a particular object called an ‘anforisco’. Often believed to be a particular type of amphora stopper or a system or ‘vacuum’ seal amphorae, it is less frequently considered a container for perfumed oils and others curious uses.14 A new titulus pictus from the Arles-Rhône underwater excavation suggests new and unexpected content15 possibilities and seems to confirm that these objects should be interpreted as small containers rather than a very strange kind of amphora stopper. In fact, various pots of this kind have been found on shipwrecks, but no wreckage, to my knowledge, has ever produced amphorae really sealed with an ‘anforisco’.16

There’s another thing that intrigues me. The Sud Caveaux 1 and San Rossore B wrecks are almost contemporary. They are both fully, or almost fully, charged ships with ‘recycled’ containers, and in both wrecks these containers are Lamboglia 2/Dressel 6 amphorae, not really the most common amphorae that could be found between the Tyrrhenian coast and Marseille. This is particularly true for Sud Caveaux 1; the cargo of Pisa-Ship B seems to have been composed in the Campi Flegrei area, where this kind of amphora was more frequent (Pecchioni et al. 2007; Camilli 2004: 65–66).

Curiously, none of the numerous little ‘anforiscos’ retrieved in the exceptional Pisa-San Rossore site have revealed tituli picti (Iardella 2000: 197–209) despite the excellent and exceptional conservation of the artefacts coming from this excavation.

Finally, regarding the reuse of containers, here again, the Rhône River underwater excavations provided us with new data on the reuse of amphorae and the trade of materials hitherto little known. Once more, chemical analysis plays an essential role. These analyses, carried out on four African Ostia LIX amphorae, indicated the disparate presence of ben oil (and sometimes wine and milk, see Djaoui, Garnier and Dodinet 2015). The presence of ben oil is of particular interest, as the natural range of the Moringa sp. oil tree in Africa does not include the area where the amphorae’s production is supposed to be located (Africa Proconsularis). Moreover, the interpretation of painted Pompeii inscriptions (Bonifay et al. 2015), leads us to consider that the primary contents of Ostia LIX may have been olives preserved in brine. Ben oil should be regarded as a secondary contents and its presence enables the reconstruction of a complex redistribution trade circuit, which involved Rome and Egypt (Djaoui, Garnier and Dodinet 2015: 185–186).

All the recent cases mentioned above also show the importance of paying attention to these kinds of studies and analysis, with rigorous methodology and documentation. They also demonstrate the importance of taking the samples properly, without contamination, from materials that are well preserved and in good conservation condition. Only on this basis will it be possible to properly use all the information that the archaeological object, container and contents, can provide to formulate solid hypotheses on the ancient trade and beyond.  For the assignment to the Pisan production, see the article of S. Menchelli in this volume. 14  Beltrán Lloris 1970: 76–82; Rodríguez-Almeida 1974. The same observation about the findings in Pompei, Pavolini 1980: 1010. See Pavolini 1980, especially 1009–1013 for a framework of the possible uses and a good critical approach to many of the assumptions about their function. 15  Djaoui et al. 2020. I would like to thank David Djaoui once again for all the unpublished information provided from the latest research in the Rhône. 16  The same observation about the findings in Pompei, Pavolini 1980: 1010. 13

4. Not only amphorae We have discussed amphorae almost exclusively, but they are not the only containers known from ancient trade. I should mention again the recent Rhône River underwater

46

Shipwrecks, amphorae and contents

traces of resin inside this vessel do not allow us to answer this question with certainty, at least not for the moment. The next step, already being tested, involves the development of robotic excavation techniques, which should be compatible with the scientific principles of archaeology. I want to conclude by highlighting the immense wealth of these wrecks and the incredible mass of data on ancient trade that they can provide, if approached properly. Figure 6. Pisa-San Rossore (Italy). One of the specimens similar to the so-called ‘pot du Latium’, full of fish bones (©F. Cibecchini, DRASSM). Bibliography Abdelhamid, S. 2013. Against the throw-away-mentality: The Reuse of Amphorae in ancient maritime transport. In H. P. Hahn and H. Weis (eds), Mobility, Meaning and transformations of Things: 91–106. Oxford, Oxbow Books. Arobba, D., Caramiello, R. and Martino, G. P. 1997-1998. Analisi paleobotaniche di resine dal relitto navale romano del Golfo Dianese. Rivista di Studi Liguri LXIIILXIV: 339–355. Auriemma, R. 2000. Le anfore del relitto di Grado e il loro contenuto. Mélanges de l’École française de Rome. Antiquité 112-1: 27–51. Barreca, D. 2006. Un titulus pictus dallo scavo delle navi di Pisa San Rossore. Gradus 1-0: 6–7. Beltrán Lloris, M. 1970. Las ánforas romanas en España. Monografías Arqueológicas 8. Zaragoza, Diputación Provincial, Institución Fernando el Católico. Bonifay, M., Botte, E., Capelli, C., Contino, A., Djaoui, D., Panella, C. and Tchernia, A. 2015. Nouvelles hypothèses sur l’origine et le contenu des amphores africaines Ostia LIX et XXIII. Antiquités africaines 51: 189–210. Botte, E. 2012. L’exportation du thon sicilien à l’époque tardo-républicaine. Mélanges de l’École française de Rome. Antiquité 124-2: 577–612. Bound, M. 1991. The pre-classical wreck at Campese bay, Island of Giglio. First season report; Second interim report, 1983 season. Studi e materiali. Soprintendenza ai beni archeologici per la Toscana 6.1: 181–198, 199–244. Camilli, A. 2004. Il cantiere delle navi antiche di Pisa. Note sull’ambiente e sulla periodizzazione del deposito. Archaeologia Maritima Mediterranea 1: 53–75. Camilli, A. and Setari, E. 2005. Le navi antiche di Pisa. Guida archeologica. Milan, Electa. Cibecchini, F. 2015. Les épaves antiques à grande profondeur en Corse. In La Corse et le monde méditerranéen des origines au Moyen-Âge: échanges et circuits commerciaux. Actes du colloque de Bastia - 21‑22 novembre 2013. Bulletin de la Société des Sciences Historiques et Naturelles de la Corse 748-749: 7–23.

Figure 7. Aleria 1 shipwreck (Corsica). A two-handled pot of glazed pottery production (©S. Cavillon, DRASSM). 5. Days of Future Past In recent years, the Drassm’s research projects have focused on key deep wrecks. Our priority was to document those sites (photogrammetry, 3D, etc.). We have raised, for the moment, only a few well-chosen artefacts. For example, from the Aleria 1 wreck (Corsica, more than 300 meters deep) we have raised a pot of a particular production of glazed pottery (Figure 7), probably produced in Rome (Cibecchini 2015: 12 and Cibecchini 2016), for which we would like to know whether they were traded as pottery or containers, without ruling on whether it is an object of life on board. Also in this case, the question is: pottery used on board or a container for the nourishment of the sailors? Unfortunately, the results of the analysis17 of the  The chemical analyses carried out by Nicolas Garnier revealed no trace associated with a content, suggesting that it is rather a cargo vessel. However, this vessel was at the surface of the tumulus after the passage of the trawl nets. It was no longer rooted in its original position and was exposed to any kind of interaction with the external environment. 17

47

Franca Cibecchini

Cibecchini, F. 2016, Et fiat lux … à 20.000 lieux sous la mer ! Nouvelles données sur la cargaison de lampes de l’épave profonde Aléria 1. In D. Djaoui D. (ed.), Histoires matérielles : terre cuite, bois, métal et autres objets. Des pots et des potes: Mélanges offerts à Lucien Rivet. Archéologie et Histoire Romaine 33: 193–201. Autun, Éd. Monique Mergoil. Cibecchini, F. 2020. Pour une nouvelle carte des épaves à dolia : Ouest Giraglia 2, Diano Marina et le commerce du vin en vrac en Méditerranée occidentale. In Carrato C. and Cibecchini F. (eds), Nouvelles recherches sur les dolia. L’exemple de la Méditerranée nord-occidentale à l’époque romaine (Ier s. av. J.-C. - IIIe s. ap. J.-C.). Actes de la tables ronde d’Aspiran, 2013. Revue archéologique de Narbonnaise Supplément 50: 163–196. Montpellier, Éditions de la Revue archéologique de Narbonnaise. Cibecchini, F. and Bargagliotti, S. 2011. Le mobilier portuaire de Portus Scabris (Grosseto-Italie) aux IIe et Ier siècles av. n. è. In SFECAG, Actes du Congrès d’Arles: 83–95. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Cibecchini, F., De Juan, C. and Marlier, S. 2017. The Roman Ouest Giraglia 2 shipwreck (Corsica): architectural study and some elements of reflection about the ship’s cargo. In J. Gawronski, A. Van Holk and J. Schokkenbroek (eds), Ships and Maritime Landscapes. Proceedings of the 13th International Symposium on Boat and Ship Archaeology (Amsterdam 2012): 267– 272. Eelde, Barkhuis publishing. Cibecchini, F., Rico, C. and Poveda, P. 2018. Capo Sagro 2 : une épave romaine à chargement de lingots d’étain à 500 mètres de profondeur. In G. Boetto and E. Rieth (ed.), De re navali. Pérégrinations nautiques entre Méditerranée et océan Indien, Mélanges en l’honneur de Patrice Pomey. Archaeonautica 20: 67–87. Cipriano, S. and Mazzochin, S. 2012. Produzioni anforarie dell’Italia alto e medioadriatica in età romana. In C. S. Fioriello (ed.), Ceramica Romana nella Puglia adriatica: 241–254. Bari, Sedit. Dallongeville, S., Garnier, N., Bernal-Casasola, D., Bonifay, M., Rolando, C. and Tokarski, C. 2011. Dealing with the identification of protein species in ancient amphorae. Analytical and Bioanalytical Chemistry 399 (9): 3053–3063. Dallongeville, S., Garnier, N., Bernal-Casasola, D., Bonifay, M., Rolando, C. and Tokarski C. 2013. Identification of animal species from remains in ancient amphorae by proteomics. In D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz and A. M. Saez (eds), Hornos, talleres y focos de producción alfarera en Hispania. Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (I): 403–417. Cadiz, Universidad de Cádiz. De Juan, C., Cibecchini, F. and Miralles, J. S. 2014. El pecio Bou Ferrer (La Vila Joiosa-Alicante). Nuevos datos sobre su cargamento y primeras evidencias de la arquitectura naval. In J. Nieto Prisco and A. Ramirez Pernia (eds), Arqueología subacuática española, Cartagena, 14, 15 y 16 de marzo de 2013: 133–149. Madrid, Ministerio de Educación, Cultura y Deportes.

De Juan, C., Cibecchini, F., and Vento, E. 2012. El pecio romano Bou-Ferrer, un velero de comercio naufragado en la costa de La Vila Joiosa. In La vila Joiosa. Arqueologia i museu: 178–198. Alicante, Museo Arqueológico Provincial de Alicante. Djaoui, D. 2016. The myth of ‘Laccatum:’ a study starting from a new titulus on a Lusitanian Dressel 14. In I. V. Pinto, R. R. de Almeida and A. Martin (eds), Lusitanian Amphorae: Production and Distribution. Roman and Late Antique Mediterranean Pottery 10: 117–127. Oxford, Archaeopress. Djaoui, D. and Quaresma, J. C. 2016. Lusitanian amphorae from the dump layer above the Arles-Rhône 3 shipwreck. In I. V. Pinto, R. R. de Almeida and A. Martin (eds), Lusitanian Amphorae: Production and Distribution. Roman and Late Antique Mediterranean Pottery 10: 357–367. Oxford, Archaeopress. Djaoui, D., Garnier, N. and Dodinet, E. 2015. De l’huile de ben identifiée dans quatre amphores africaines de type Ostia LIX provenant d’Arles : difficultés d’interprétation, Antiquités africaines 51: 179–187. Djaoui, D, Greck, S. and Marlier, S. 2011. L’épave Arles Rhône 3, Le naufrage d’un chaland antique, enquête pluridisciplinaire. Arles, Actes Sud. Djaoui, D., Piquès, G. and Botte, E. 2014. Nouvelles données sur les pots dits « à garum » du Latium, d’après les découvertes subaquatiques du Rhône (Arles). In E. Botte and V. Leitch (eds), Fish and Ships, Production and commerce of salsamenta during Antiquity. Actes de l’atelier doctoral, Rome 18-22 juin 2012. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17: 175–197. Arles, Errance - Aix-en-Provence, Centre Camille Jullian. Djaoui, D., Augusta-Boularot, S., Garnier, N. and Silvino, T. 2020. Le mythe des amphorisques : tituli picti et analyses chimiques (Arles: le dépotoir portuaire d’AR3 et Lyon : Parking St-Georges et Parc St‑Antoine). In SFECAG, Actes du Congrès de Lyon: 215–228. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Ehmig, U. 2009. Le cas de quelques inscriptions peintes. In L. Long and P. Picard (eds) César, le Rhône pour mémoire. Vingt ans de fouilles dans le fleuve à Arles: 274– 277. Arles, Actes Sud. Formenti, F., Hesnard, A. and Tchernia, A. 1978. Note sur le contenu d’une amphore Lamboglia 2 de l’épave de la Madrague de Giens. Archaeonautica 2: 95–100. Garnier, N. 2015. Identifier les traces de vin archéologique : des structures de production aux vases à boire. Un bilan des méthodologies et des apports de l’analyse chimique organique. In SFECAG, Actes du Congrès de Nyon: 299–313. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Garnier, N., Silvino, T. and Bernal-Casasola, D. 2011. L’identification du contenu des amphores : huile, conserves de poissons et poissage. In SFECAG, Actes du Congrès d’Arles: 397–416. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Iardella, R. 2000. Balsamari ceramici. Vasetti ovoidi e piriformi. In S. Bruni (ed.) Le navi antiche di Pisa. Ad un anno dall’inizio delle ricerche: 197–209. Florence, Polistampa.

48

Shipwrecks, amphorae and contents

Lagóstena Barrios, L. 2004. Las ánforas salsarias de Baetica. Consideraciones sobre sus elementos epigráficos. In J. Remesal Rodríguez (ed.), Epigrafía anfórica. Proyecto Amphorae. Instrumenta 17: 197–219. Barcelona, Universitat de Barcelona. Liou, B. and Domergue, C. 1990. Le commerce de la Bétique au Ier siècle de notre ère. L’épave Sud-Lavezzi 2 (Bonifacio, Corse du Sud). Archaeonautica 10: 11–123. Long, L. 1998. Lucius Volteilius et l’amphore de 4ème type. Découverte d’une amphore atypique dans une épave en baie de Marseille. In El vi a l’antiguitat: economia, producció i comerç al Mediterrani occidental. Actes del II Col loqui Internacional d’Arqueologia Romana (Barcelona 6-9 de maig de 1998). Monografies badalonines 14: 341–349. Badalona, Museu de Badalona. Martin-Kilcher, S. 2011. Formes d’amphores et contenu au Haut-Empire, points de repère et questions. In SFECAG, Actes du Congrès d’Arles: 417–426. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Mazzoli, C., Maritan, L. and Pesavento Mattioli, S. 2009. Anfore da olio e anfore da pesce: le analisi archeometriche. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 239–255. Rome, Quasar. Pascual Berlanga, G. and Ribera y Lacomba, A. 2014. Ánforas tardopúnicas sicilianas en Pompeya. Rei Cretariae Romanae Fautorum Acta 43: 461–466. Pavolini, C. 1980. Appunti sui ‘vasetti ovoidi e piriformi’ di Ostia. Mélanges de l’École française de Rome. Antiquité 92: 930–1020. Pecchioni, E., Cantisani, E., Pallecchi, P., Fratini, F., Buccianti, A., Pandeli, E., Rescic, S. and Conticelli, S. 2007. Characterisation of the amphorae, stone ballast and stowage materials of the ships from the archaeological site of Pisa-San Rossore, Italy: inferences on their provenance and possible trading routes. Archaeometry 49: 1–22.

Rodríguez-Almeida, E. 1974. Sobre el uso del anforisco Cucurbitula. Mélanges de l’École française de Rome. Antiquité 86-2: 813–818. Silvino, T., Djaoui, D., Desbat, A. and Caillaud, C. 2015. Tituli picti et tutti quanti. Nouvelles inscriptions peintes sur amphores dans la vallée du Rhône. In SFECAG, Actes du Congrès de Nyon: 647–658. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Silvino, T. y Poux, M. avec la collaboration de N. Garnier 2005. Où est passé le vin de Bétique ? Nouvelles données sur le contenu des amphores dites « à sauces de poisson et à saumures » de type Dressel 7/11, Pompei VII, Beltrán II (Ier s. av. J.-C.-IIe s. apr. J.-C.). En SFECAG, Actes du Congrès de Blois: 501-514. Marsella, Société Française d’Étude de la Céramique Antique en Gaule. Tchernia, A. 1986. Le vin de l’Italie romaine. Essai d’histoire économique d’après les amphores. Bibliothèque des Écoles françaises d’Athènes et de Rome 261. Rome, École française de Rome. Tchernia, A. 1990. Contre les épaves. In A. Duval, J.-P. Morel and Y. Roman, Gaule interne et Gaule méditerranéenne aux IIe et Ier siècles avant J.C. : confrontations chronologiques. Actes de la table ronde de Valbonne, 11-13 Novembre 1986. Revue archéologique de Narbonnaise Supplément 21: 291–301. Paris, Centre National de la Recherche Scientifique. Tchernia, A., Pomey, P. and Hesnard, A. 1978. L’épave romaine de la Madrague de Giens (Var). Gallia Supplément XXXIV. Paris, Centre National de la Recherche Scientifique. Woodworth, M., Bernal‑Casasola, D., Bonifay, M., Garnier, N., Keay, S., Pecci, A., Poblome, J., Pollard, M., Richez, F. and Wilson, A. 2015. The content of African Keay 25 / Africana 3 amphorae: initial results of the CORONAM project. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 41–57. Esposende, Município de Esposende.

49

Amphora contents as commodities: the structure and function of tituli picti in the western Mediterranean in the 1st century AD Enrique García Vargas Universidad de Sevilla

Abstract: Tituli picti are painted inscriptions in black or red ink on the outer surface of amphorae. Traditionally, they have been considered a form of marketing that fulfil a similar function to modern bottles and labels on cans. However, their highly standardised structure and their widespread use, at least in Roman times, have recently led to the idea of a more specialised function within the commercial circuit of amphorae, namely that of providing information to commercial agents about the basic clauses of maritime insurance contracts presiding over commercial cargos. I think there is no real reason to link tituli picti with insurance terms, and I shall try to demonstrate in this paper that tituli picti or epigraphic labels played a different role. For this, I shall rely on relatively simple examples attested on various types of amphorae filled in different geographical areas. Moreover, I shall argue that their function was commercial, but not related to insurance contracts, but rather to the commercial transactions, that is, to the action of buying and selling the products contained in the amphorae, and especially to the transfer of legal responsibility over the goods from the producer to the trader or from one trader to another. Key words: Roman trade; salted fish products; epigraphy; Roman oil trade; Principate period; defrutum; instrumentum domesticum.

1. Introduction and aims1 The aim of this paper is to provide a general interpretation of tituli picti written on 1st‑century amphorae in the Western Mediterranean, and to compare different epigraphic habits in containers used for different products and in different regions. I begin from the premise that the, sometimes great, similarities presented by the epigraphic formulae found on different amphorae — e.g. the type and order of entries, position on the container — are due to the fact that they played similar roles. Previous suggestions as to what these roles might have been are flawed, and I suggest that the function of tituli picti was directly related in a much more explicit manner to what the amphorae were: containers of goods that were on display for commercial exchange or were in transit in the context of a mercantile operation. In other words, the tituli picti written on commercial amphorae register basic information about the contents of the amphorae, certifying several aspects which would have been key for the commercial transaction to take place. I shall return to this matter shortly, and will use the basic formula found on oil amphorae in order to explain the function of tituli picti in the context of the legal transaction which involved buying and selling the contents of the amphorae.  This work was submitted to the editors in the summer of 2017. Since then, some papers by Emilia Mataix Ferrándiz have been published that study in some depth the structure and function of the inscriptions painted on amphorae (Mataix Ferrándiz 2019) or on the recipients that contained commercial samples (2020) in a similar line to ours, which we believe should be followed in the future as a key for reading this kind of inscriptions. 1

However, before we can proceed, it is useful to briefly examine previous theories on the function of tituli picti. 2. Function of tituli picti on amphorae The tituli picti written on amphorae used to contain salted products have been considered simple marketing devices that aimed to attract customers and promote consumption (Curtis 1984‑1985, 1991; Berdowski 2003: 18–19; Martin‑Kilcher 2002: 344). In my opinion, this interpretation involves an excessively mechanical transposition of the tituli’s text, and it relies too much on our modern notion of a commercial label. Indeed, this interpretation fails to consider that most people, even if they were literate, would not have come into contact with inscriptions on amphorae, and, besides, they would have lacked the knowledge to fully understand the meaning of the abbreviations and other symbols, which were designed by merchants for their own use. As an alternative view, Ulrike Ehmig (2014a and b) argued that the tituli picti written on the amphorae used to contain salted products conveyed key information for the commercial agents involved in the circulation of amphorae. This hypothesis is in many ways similar to mine, but it is still not entirely satisfactory. According to Ehmig’s interpretation, tituli picti would have been something akin to the basic terms of the maritime insurance which covered the goods shipped by a given mercator (name in genitive D). Numeric notations (in C) are interpreted as the value of the freight, and were used by different merchants working with the same shipowner or navicularius if it became necessary to sue for damage that had been inflicted on the cargo

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 51–62

Enrique García Vargas

during the journey. In contrast, the names in genitive (in E) do not, according to this interpretation, refer to the producers of the goods contained within the amphora, as has been traditionally argued (cf. Colls et al. 1977: 91–93), but rather to the merchant’s slaves or employees who accompanied and watched over the cargo until it reached its final destination.

framework will accommodate the changes undergone by commercial bodies in the Early Imperial period, such as the progressive incorporation of oil into official (i.e. promoted by public agents) trade flows. This, along with the growing complexity of markets and commercial practices, brought about increasingly formulistic tituli picti. Although this latter phenomenon also affected other products, the formulae found on amphorae that were not used to contain olive oil did not become completely formulaic until later, during the Flavian period.

This interpretation relies on documents such as Tabula Sulp. 78, dated to AD 38 (Camodeca 1999); this was an insurance policy for 1000 denarii found in the archive of the Sulpicii, in Pompeii. The document was signed by the salted product merchant P. Attius Severus (also known from the tituli picti on Baetican amphorae used to contain salted products: CIL XV, 4748‑4749) and a naukleros named Menelaos; the text also mentions a slave of Attius Severus called Primus. According to the terms of this document, which must be a summary of a longer one, Menelaos pledges to hand Primus 1000 denarii after arriving in harbour, which can only mean that the freighter recognises his responsibility for the safety of the cargo. If the voyage reaches its destination without incident, Menelaos will hand this responsibility back to Primus — that is, he will hand back the insurance premium. Should the cargo be totally lost (and along with it, probably also the lives of Menelaus and Primus), Marcus Barbatos Celer will act as guarantor and will compensate Attius Severus for his losses.

3. Interpreting amphorae contents and the function of tituli picti using oil amphorae With regard to oil amphorae, it is well known that tituli before the Flavian period were simpler that those dated from the Flavian period onwards. The frontal inscriptions painted on the neck and the belly of oil amphorae did not change much over time, and even the earliest labels seem to be fairly standardised. They consist of an entry at the top (α), which indicates the weight of the empty container; another entry at the bottom (γ), which records the weight of the oil within; and finally, another entry in the middle (β), which records a tria nomina in the genitive (mercator). But the element δ, diagonally inscribed under the right handle, is in the earlier examples much less complex than was the case in the late 1st‑century and, especially, 2nd‑century specimens. In general, this is interpreted as the result of an increase in the complexity of tax controls over the production and the distribution of the oil consumed by the state. Until the mid‑1st century, the δ entry remains relatively simple, and is reasonably similar to that found on amphorae used to contain products other than oil (see below).

I find no reason to link tituli picti with the terms of insurance. The amount mentioned in the insurance policy may have matched the total value of the cargo, but this is little more than a legal fiction that stresses the shipowner’s responsibility for the cargo during the trip. Also, the careful annotation of insurance terms on the amphorae would have been rendered useless if the ship sank. Under such circumstances, tituli picti would be of no use whatsoever for the execution of the clauses of the insurance contract, especially since the terms of this contract were recorded in at least two other documents elsewhere.

During the early years of the Imperial period (Augustan and Tiberian periods), the tituli picti on oil amphorae, although in nuce the same throughout the period in which these containers were in circulation, was schematic and its different elements were little developed. In essence, the formula is illustrated by the inscription (Figure 1) found on amphora CIL XV 3657 from the Castra Praetoria. Dressel identifies this specimen with his type 26 similis, but the amphora is an Oliva 3 from the Tarraconensis, as has been recently proven by Berni (in press):

In my opinion, the careful recording of information (e.g. weight of the cargo, its value and the total number of containers in a given shipment) must have played a different role, and in order to clarify this role more attention should be paid to the structure of the information conveyed by these labels.

The interpretation of these entries poses no great difficulties. Entry α refers to the empty weight of the amphora, β to the initials of the tria nomina of the merchant or shipowner that purchased it, and γ to the net weight of the oil. The genitive in δ has traditionally been associated with the producer of the oil, and the numeral that goes with it merely repeats the amount conveyed by γ (although the notations are slightly different).

In the following paragraphs, I shall compare tituli picti on oil amphorae of the Dressel 20 type (or their predecessors) and on amphorae for salted products from different provinces (from Baetica and Italy). The structure and function of tituli picti on oil amphorae will formalise the most comprehensive and abundant type of evidence, and will provide us with an interpretive framework for other kinds of amphorae, for example those used to store and transport salted products in the same period (1st century AD). This interpretive

During the late Julio‑Claudian period, the δ entry on Baetican oil amphorae, which by that time were fully developed Dressel 20 types, became slightly more complicated. From this period onwards, the standard

52

Amphora contents as commodities: the structure and function of tituli picti in the western Mediterranean…

Figure 1. Inscription CIL XV 3657 from the Castra Praetoria (Rome), displaying the typical titulus structure found on early Spanish oil amphorae (in this case on a Tarraconensis Oliva 3 type amphora). entry consisted of a neutral nominative2 or a nomen, cognomen,3 duo nomina4 or tria nomina,5 in the genitive case, followed by the initials a or áá, the word arca and numerals. Although these elements can all appear on the same amphora, the tituli picti found on many specimens are missing one or several of these elements.6

the same personal genitive in δ present identical numerals. If this information is added to that provided by entries α (weight of empty amphora) and γ (net weight of oil, often repeated in δ), the label represents a fully finished transaction (Dig. 18.6.8): quid or what (obviously oil), quale or of what type (from what estate of a producer), quantum sit or how much (γ) and pretium. The buyer is indicated in titulus β and the person who delivers the goods to him (which can be the producer, but does not need to be) is in the genitive or neutral nominative in δ.

A fifth and last element that can appear on the δ entry in the 1st century AD is a seemingly abbreviated name that follows the other four elements of this entry (nominative or genitive names, initials á áá, arca indication and numeral).7 Some authors, inspired by 2nd century δ entries, develop this fifth element as a name in the nominative case.

This led Liou to suggest (Colls et al. 1977: 98) that the entry δ on Baetican oil amphorae aimed to facilitate the tracking of the oil to its source. Later scholars have accepted this suggestion (e.g. Chic García 1988: 146), but whether these notes had a private or an official character is still unclear.

The neutral nominative and the nomina in the genitive are relatively interchangeable, insofar as they both refer to the origin of the oil contained in the amphora, by indicating the estate where it was produced (neutral nominative rooted in a personal name) or by indicating the genitive of the owner (or lessee) of the estate. The indication arca seems to refer to a (state?) fund used to buy the goods, and the signs a and áá + numerals may refer to amounts spent, although this is but a tentative interpretation (cf. Chic García 1981: 251–260; 1981: 257).

The increasing complexity of the δ entry from the late 1st century onwards — especially with regard to the addition of new elements, and the nature of these elements — suggests that, from this period onwards, these control operations were performed by public tax agencies (Chic García 1988). However, for the period immediately prior to the Flavian reforms (which involved the absorption of whole strategic sectors by the tax administration system), and especially before Hadrian’s reform of the Annona (which must have involved indictiones or obligatory sales of oil to the state), it is reasonable to suggest that these annotations were chiefly private in character.

Although the interpretation of the different elements present in the δ entry of oil amphorae in the 1st century AD is not straightforward, it is reasonable to state that these elements refer to the original owner of the oil and to the funds used for its purchase. The numerals represent the amounts spent and other variables that affect the whole cargo to which the amphora belongs — all amphorae with

This does not mean that the oil trade was the exclusive concern of private mercatores but rather that the state purchased olive oil according to the same mercantile rules that applied to everyone else, with the only exception that supply contracts were subject to the rules of public purchase mechanisms such as auctions (Chic García 2002: 338). As such, negotiatores who were successful in a bid had to make their own arrangements to ensure that the goods were purchased, transported and delivered according to the terms of the contract. The notes in the tituli picti

 Montanum: Colls et al. 1977: no.14, c. AD 40.  Calpurni: CIL XV 3647, Castra Praetoria, c. AD 50. 4  L. Vesoni: CIL XV 3688, Castra Praetoria, c. AD 50. 5  Q. Licini Fabiani: Colls et al. 1977: no.12, c. AD 40. 6  Lucreti á cxcii: Colls et al. 1977: no.19, c. AD 40; a arc(a). 7  Luculli arca XIIIS Sec[ ]: CIL XV 3683, c. 50; Iuliani á xvi For [ ]: CIL XV 3643, c. AD 50. 2 3

53

Enrique García Vargas

were written in this context; they probably reproduced in abbreviated form the terms of the sale contract that transferred responsibility for the commodity from sellers to buyers.

Caecilii.8 A final inscription on the vase is the abbreviation of a tria nomina: L.C.S. is superimposed on the black labels and was, therefore, written at a later date. D. Djaoui (2014) has argued that the number indicates the total number of amphorae in the cargo; however, even if the numeral indicated a price, or a volume, this would not change the most likely interpretation of the find, which is that the vase was a sample container for a batch of Baetican oil amphorae. It would have been sold during the Flavian period by DD. Caecilii and purchased by the enigmatic L.C.S.

As previously noted, tituli picti, especially entry δ, seem to record the basic features of the transaction: purchaser, producer‑seller, product, quality (or provenance), quantity and, eventually, price. Under Roman law, meeting the terms of a contract (that is, achieving the effective transferal not only of the goods, but also of the responsibility for their integrity and safety), required not only clearly defined clauses, but a legal act of confirmation (for instance, confirmation following the delivery of the goods and their degustatio, or a physical check of the quality of the products).

Should this interpretation be correct, we would once more be faced with the same epigraphic elements that can be found in the δ entry of 1st‑century Dressel 20 oil amphorae, which can be interpreted as a record of the delivery and degustatio of the goods.

Generally, the degustatio involved providing the customer with smaller sample containers, some of which have survived in the archaeological record. The inscriptions on these sample containers are very similar to those that feature on the earliest Dressel 20 amphorae. In this way, the purchaser could check the characteristics of the cargo that was being delivered by comparing it with the data conveyed by his complementary documentation, which included the sample containers.

The role of the inscriptions written on Dressel 20 amphorae, sample containers, and the rest of the documentation that accompanied these batches of amphorae, was to facilitate commercial transactions. It seems clear that the function of these inscriptions was not to advertise the product; the highly abbreviated nature of the contents and the peculiar calligraphy would have been beyond the comprehension of the general public. We have already rejected the possibility of these inscriptions expressing the terms of maritime insurance. In my opinion, these messages were used by and for oil merchants, as a way to control the commodities that were being transacted — as a sort of summary of the sales contract. As such, the inscription indicated whose responsibility it was to ensure the integrity and safe delivery of the goods (in this case, olive oil) at the end of the transaction.

A small jar for wine from the Rhône found in Arles (Djaoui 2014) has been interpreted as one of these sample containers. The inscription on the jar reads: Alb(anum) Valeri Proculi / Dol(ia) (centum quadraginta) sexsagenaria. That is, the jar contains an Albanian wine, produced by Valerius Proculus, which belonged to a batch that was processed and stored in 140 dolia. The elements present in the inscription are those which would be expected in an act of degustatio: commodity (vinum), quality (Albanum), original owner (Valerius Proculus) and quantity (140 dolia sexagenaria).

4. Tituli picti on amphorae for salted products: structure, function and names

The similarity of this formula to the δ entry found on oil amphorae is revealing. Although these elements do not appear in every case, they can always be related to one or another consideration about the legal transfer of goods: type of oil (or provenance) (Pontianum), owner (Calpurni) and quantity or price (XIIIIS).

Despite the differences in structure and complexity in the tituli picti written on amphorae for salted products and for oil, the content and function of these inscriptions did not always differ. In the first and second thirds of the 1st century AD, olive oil and salted products circulated via the same distribution channels, were transported in the same ships and sold by the same mercatores (Colls et al. 1977). Although the state soon became the main purchaser of olive oil, during this early period the sales followed the same rules that applied to private transactions. The growing ‘official’

Interestingly, the same assemblage from Arles also included a sample container of Baetican oil (pot 2) (Djaoui 2014). The inscription repeats the same formula, with personal names and numerals, which can hardly be a coincidence. The inscription, made with black ink, appears on a Turdetanian‑style vase, which is decorated with the red bands across the shoulder and belly that are characteristic of this tradition. The vase presents a set of numerals which are only partially preserved; it is unclear whether they all represent the same number but if that were the case, the number would be 202. Alongside the numerals, the inscription features several repetitions of the duo nomina in the genitive (not always correctly or fully written) of DD.

 The same nomina and cognomina have been attested on several tituli picti on amphorae for olive oil: only one of these inscriptions can be dated with some certainty (CIL XV 3792A: Horti Torlonia), to AD 149; other DD Caecilii inscriptions are attested in Pompeii (CIL IV 9480), and these must obviously predate August AD 79. Seventy years seems too wide a gap to argue that these inscriptions refer to the same pair of Caecilii. The former inscription is chronologically closer to the Arles example, which is dated, based on archaeological and calligraphic criteria, to the Flavian period (Djaoui 2014). For the QQ. Caecilli on amphorae found in Mantua and Parma see below. 8

54

Amphora contents as commodities: the structure and function of tituli picti in the western Mediterranean…

Figure 2. Inscription structure of Dressel 21-22 Italian amphorae as shown on a specimen from the Garum shop in Pompeii (after García Vargas et al. 2020). interest explains the unique development of tituli picti on Dressel 20 amphorae in the 2nd century, when up to three different lines of text conveyed relevant information for the control of the annona (Chic García 1988).

The care with which these notes were taken emphasises a connection with the commercial transactions. Digesto 18.6.2 points out that wine could be sold in two different ways: by weight or by amphora. In the first system, the buyer would extract, in a fixed day, the wine from the dolia and measure it personally, before bottling it. In the second system, the product was delivered already measured and bottled so the buyer measurement was not necessary. In this case, the buyer only had count the total number of amphorae delivered to him, but there was no occasion for degustatio or the tasting of the product when the buyer picked the wine amphorae. The transfer of responsibility concerning the wine’s quality occurred probably not at that moment, but earlier, during a visit carried out before the product was measured and bottled by the seller (Chic García 2002). Dig. 18.6.4.1 indicates however, that a ‘second’ degustatio was common even after the bottling in amphorae had taken place, when the seller handed over the containers to the buyer. This could explain items such as the wine jar and the pot 22 found in Arles, which has been interpreted as product samples for cargoes in transit. That is, as products that were already bottled by the time they changed hands, and that were produced and bottled hundreds of miles from where they were exchanged, making, rendering a prior degustatio impossible.

During most of the 1st century, however, the tituli on both kinds of amphorae conveyed similar information, referring, as previously noted, to the commercial transaction, the goods that were being traded, the people that were involved in the operation and the responsibilities of each of these participants. This matter of responsibility would have been of great legal importance. Before the contract was brought to completion, any problems or damages suffered by the cargo fell to the seller, whereas once the contract had been ended it was the buyer who bore the damages. The completion of the contract was a formal occasion, which involved quantifying a price and paying a pretium, and it may also have involved a degustatio, the purpose of which was to certify that the quality of the product corresponded to what was agreed in the contract. Only in relation to this commercial transaction, in which the quantity of merchandise and its price were so important from a legal point of view, can we understand the care with which the weight of the goods is annotated on the side of Italian Dressel 21‑22 amphorae (which constitute an interesting parallel for the Baetican Dressel 20). It should be noted that the entries in these Italian records (Figure 2) are arranged differently (García Vargas et al. 2020).

On the other hand, the expression lec( ) or lec( ) fec( ), found on some Dressel 21‑22 amphorae from Rome and Pompeii,9 following the nominative element E, may be interpreted as lectio fecit — that is, this expression indicated that the person whose name appears in the nominative case received the product and, presumably, bottled it. This is the moment when the amphorae can be weighted, first when it was empty (entry C) and later when it was full (entry G). The ductus of the inscription suggests that this operation was carried out in one go,

Similarly, the way the initials of tria nomina of merchants or middlemen (D) — which are normally written by a different hand — are superimposed on the rest of the titulus in Dressel 21‑22 amphorae is reminiscent of the superimposition of tria nomina on the sample container (pot 22) from Arles (the DD. Caecilii); this superimposed notation probably points to a standard commercial transaction that involved the transfer of merchandise.

 MIVR F. (Giordano‑Casale 1990: no.390); SICVLVS LEC F (CIL XV 4793); FAVOR LEC (CIL XV 4786). 9

55

Enrique García Vargas

Figure 3. Epigraphic formula of Baetican salted fish amphorae found north of the Alps and example of a titulus from Mainz (after Martin-Kilcher 1994). and that the genitive with the tria nomina of the merchant (entry D) was added later, sometimes running over the earlier inscriptions.

Entries C and G on Dressel 21‑22 amphorae for salted products from Campania, Apulia, and Sicily, correspond to the empty‑ and full‑weight entry on the Dressel 20 type. As is the case with the oil, this suggests that the product was weighed and bottled in the presence of the seller and the purchaser; the seller features in entry E in the nominative, and the buyer in entry D. The tria nomina were generally in abbreviated form and normally added by a different hand after the measurement and delivery of the merchandise.

In these examples, the protocol is very similar to that followed for Baetican olive oil, when the purchasing and the bottling of the oil occurred simultaneously, and the weight of the container and of the product was measured on the spot. Sometimes, as is the case with pot 22 from Arles, the same cargo was the subject of more than one transaction. These transactions were, in principle, independent from one another. The original purchase had to be accompanied by the degustatio, the measurement and the bottling (diffusio) of the product; the re‑sale of all or part of the product in another harbour, e.g. Arles, involved a new degustatio, but not the measurement of the product, which was already bottled and its weight labelled. In this case, the only matter to be settled was the number of amphorae that were to change hands (202).

With regard to tituli picti on Baetican and Gallic amphorae for salted products, there are substantial differences that are probably related to the various legal requirements. However, the epigraphic formula remains virtually the same as on Dressel 20 and Dressel 21‑22 amphorae, which is unsurprising because it plays the same role: to summarise the conditions of the contract, certify its completion and thus bear witness to the transfer of legal responsibility for the merchandise contained in the amphorae.

In general, tituli picti on amphorae for oil and salted products10 only record the first transaction. The notation, by the same hand, of entries α (weight of empty amphora) and γ (weight of full amphora), indicate that the product was weighed and bottled at the same time, probably in the wharf near the pottery workshop from which the containers had been purchased. The emergence of the earliest tituli picti on Dressel 20 coincides with the creation of the first ‘industrial’ amphora workshops on the shores of the Guadalquivir (La Catria, in Lora del Río) and the Genil (Las Delicias, in Écija). The transaction certified by entry δ, which in our opinion corresponds with the reception of, and payment for, the oil by the mercator mentioned in entry β in the genitive case, also took place at the same time.

Initially, tituli picti on Baetican amphorae for salted products were of great simplicity, in terms of both content and structure. During the Flavian period, these tituli picti became substantially more complex, although they were never subject to the same degree of tax control as the Dressel 20 type, the tituli picti of which became particularly complex from the Flavian period onwards. It is likely that this widespread increase in complexity is indicative of changes in the commercial structures of Baetica in the second half of the 1st century AD (Figure 3), as illustrated by the following formula (Martin‑Kilcher 1994: 405, P‑29, see Figure 3). These changes affect specially to the tituli picti on Dressel 7‑11 and Beltán IIA and B exported to the Rhine through the Rhône Valley. Amphorae found along the course of the Rhine, in Cologne, Mainz and Augst, present an extra inscription, which was added after the basic formula (written during embarkation on the Baetican

 Some Dressel 21‑22 from Rome include an element F, similar to that found on Baetican salted product amphorae. This points to the resale of the containers in the Rhône. 10

56

Amphora contents as commodities: the structure and function of tituli picti in the western Mediterranean…

Find Mainz: Martin-Kilcher 2002, no.2 Mainz: Martin-Kilcher 2002, no.4 Mainz: Martin-Kilcher 2002, no.5 Mainz: Martin-Kilcher 2002, no.6 Mainz: Martin-Kilcher 2002, no.7 Mainz: Martin-Kilcher 2002, no.10 Augst: Martin-Kilcher 1994: 405, P-29 Dr.7-11 Cologne: Ehmig 2007, no.117 BIIA Pompeii : CIL IV 9368, 9369

D PROCVLI ET VRBICI PROCVLI ET VRBI[CI] PROCVLI ET VRBIC[I ] L. [U]RITTI VERECUNDI L. URITTI VE[RECUNDI] L. URITTI VE[RECUNDI] PROCVLI [ET] VRBICI IVNIORVM PROCVLI, L. ET VRBICI

E RVSTICUS DOMESTICUS DOMESTICUS ¿? ¿? DECIMI CITI CASTI ELM ?

F C.I.S. VII L.V.V. XXXV L.V.V LXXV L. V. V. XVI Q. S. L. III M. L. V. VI AFR V VERIDI XXV

Figure 4. Entries D, E and F of tituli picti on Baetican salt-fish amphorae from Mainz (after Martin-Kilcher 2002), Augst (after Martin‑KIlcher 1994), Cologne (after Ehmig 2007) and Rome (after CIL XV). coast). This new entry (F) generally includes tria nomina or nomina in the genitive, along with numerals, the interpretation of which is unclear.

and the person bringing it to the end of the itinerary were seemingly one and the same. It is, however, possible that, even if it were the same person carrying out all these transactions, it was necessary to make the notation for procedural or practical reasons, perhaps because the original seller retained a part of the cargo and sold the rest to other merchants. The possibility of homonymous merchants with matching tria nomina initials cannot, of course, be ruled out.

An assemblage of Beltran IIA amphorae from Mainz (Martin‑Kilcher 2002) have F entries that are filled with abbreviated tria nomina, which sometimes coincide (and sometimes do not) with the unabbreviated tria nomina of the mercatores that appear in entry D. In the following table, we have extracted entries D and F from the data published by Martin‑Kilcher (2002: fig. 5), and we have added similar registers for Augst and Cologne, as well as one from Pompeii which has no element F but which features the same mercatores as Mainz 2‑5 and Augst 29 (Figure 4).

On the other hand, an amphora bearing the name L. Urittius Verecundus (entry D) was found during the excavation of shipwreck Rhône‑Arles 3 (similar to the already known in Martin‑Kilcher 2002). The amphora, which presumably was shipped directly from Baetica, exhibited no trace of entries F and G (to date, these entries feature only on specimens found north of Lyon); we have already noted that the presence of sample containers point towards Arles as an entrepôt which would have been used in the re‑sale and transfer of merchandise.

Generally, names in entry F are interpreted as corresponding to mercatores or their representatives, who would have received the Baetican amphorae of salted products during a stopover in the itinerary of the cargo: for instance, Arles or, further north, at the harbour of Lugdunum (Martin‑Kilcher 2002: 350). The numerals which accompany these names are believed to refer to the number of amphorae transferred in that transaction. The known examples range between III (Martin‑Kilcher 2002: 347, no.7) and LXXV (Martin‑Kilcher 2002: 347, no.5) — a range that is much too wide for the numerals to refer to the weight of the container (which is, in any case, generally recorded in C) or to price (Martin‑Kilcher 2002: 346).

The relative increase in the complexity of the inscriptions found on Baetican amphorae for salted products during the final years of the 1st century AD seems to be a direct result of the growing structuration of provincial trade11 (at least concerning ‘free’ goods, such as salted products and fish sauces). This trade came increasingly under the control of ever‑larger commercial groups which possessed considerable financial resources and commanded wide social and mercantile networks.

Element F, therefore, may be considered an extra ‘control’ element that refers to an additional commercial transaction in transit. The entry records those elements of the contract which make it legally binding by transferring responsibility for the integrity and safety of the merchandise: the name of the purchaser (that of the seller is already in entry D) and the number of amphorae so disposed of (if our interpretation of the numeral in entry F is correct).

From the Flavian period onwards and throughout the 2nd century AD, it is attested that freedmen merchants in general, and especially those involved in the salted product trade, gained substantial social prominence in their communities, as is suggested by their access to urban sevirates; the wealth they gained through their commercial endeavours can only have helped them in this process (García Vargas and Martínez Maganto 2009). This may have been the case for L. Urittius Verecundus, whose name is attested in Arles and Mainz, who belonged to a Gallic gens which specialised in the commercialisation of Baetican salted products and grape by‑products from

Sporadically, the same name appears in both entries D and F: for example, L. Urittius Verecundus (Martin‑Kilcher 2002: 347, no.6). This would appear to be unnecessary repetition, as the person receiving the original cargo, the person selling the amphorae at the mid‑route stopover,

 A similar process, but concerning financial markets, has been analysed recently by F. Lerouxel (2016). 11

57

Enrique García Vargas

Figure 5. Titulus Pictus on a Dressel 8 amphora from Mantua (after Mongardi 2018: 132–133, no.17). the mid‑1st century onwards, if not earlier (cf. amphorae Dr. 20 and Haltern 70 from shipwreck Port Vendres II, which feature Q. Urittius Revocatus in entry D). It is thus possible that Verecundus reached a social position of prominence in his natal or adoptive community, although these merchants were always one social tier below the superior ordines.

primary inscriptions always record, alongside the name of the mercator or the mercatores, the quality of the product, the ageing period and the weight of the amphora — that is, basic information concerning the commercial transaction. However, before we neatly divide the transactions and inscriptions found north (multiple transactions/tituli) and south (simple transactions/tituli) of the Alps into rigid categories, we need to examine in detail the entry H (abbreviated tria nomina and numerals) written on a number of Pompeii VII amphorae found in Pompeii.

Perhaps the increasing complexity of the commercial channels used to distribute Baetican amphorae for salted products, and the related increasing intricacy of the epigraphic formula, were caused by an increase in the demands posed by the annona, and the use of free merchants to meet this demand, who aimed to ensure the smooth flow of supplies, especially towards administrative and military personnel in frontier provinces (Ehmig 1995, 1996; Ehmig et al. 2004). This sort of demand had traditionally enhanced the commercial role of professional collegia, especially in Baetica, but always within the framework of a private commercial environment. The operation of collegia ensured that state agents always had access to the right interlocutor and thus could procure the necessary supplies, whereas private merchants were offered the opportunity to strike lucrative deals with the state (García Vargas and Martínez Maganto 2009: 142–143).

5. A singular case: tituli picti on amphorae ‘of Hispanic types’ found in Modena and Parma An assemblage of amphorae — 15 specimens from Novi Sad Park in Modena, 10 from Via Saffi and 44 from Via Palermo, in Parma — present a similar epigraphic formula to those examples examined thus far, although the arrangement of the elements differs somewhat. These amphorae have been analysed by Mongardi (2014, 2016, 2018) who has suggested the following interpretation (see Figure 5). The formula is very similar to that found in Mainz, but in this case the numeral C is written below the name of the mercator, and not above (as is normally the case), and the tria nomina D are written horizontally, as a final register. Entry E, which often presents a different ductus from the rest of the inscription, has been interpreted (Mongardi 2014: 388) as an indication of price, for it often features fractionary figures.

In any case, inscriptions on most Baetican amphorae for salted products in the first half of the 1st century were rather simpler and merely included entries A, B, C, D and E — that is, the original tituli — which probably means that the cargo had arrived at its destination undivided, as a result of a single commercial transaction. In the late 1st century, secondary inscriptions such as those typically found in Northern Europe, were still rare in areas which had little connection with the supply of military garrisons, such as Pompeii, even though the same mercatores have been attested in both regions: for example, the Proculi et Urbici, who are attested in both Mainz (Martin‑Kilcher 2002: 347, nos2‑5) and Pompeii (CIL IV 9369). In every case,

The reasons behind the different arrangement of tituli in Parma and Modena and in the broader Tyrrhenian and Ligurian regions, are unclear. It is possible that some of the explanation lies in the early date of the specimens (first half of the 1st century AD, provided that the QQ. Caecilii mentioned in the amphora are the same that feature on early 1st‑century Hispanic amphorae), in the prevalence of regional Italian uses (let us remember the nominative E on Dressel 21‑22), or both.

58

Amphora contents as commodities: the structure and function of tituli picti in the western Mediterranean…

Figure 6. Titulus mentioning olivae ex defrutum on a Haltern 70 sherd from the outkirts of Mainz (after Ehmig 2007). The presence of the expression garum Hispanum on many of these amphorae brings to mind the Gallic Dressel 9 similes type, which was used to re‑bottle the product that arrived at Narbonensis in different containers. The typological peculiarities of these amphorae led Mongardi (2014, 2016, 2018) to refer to them as being ‘of Hispanic morphology’, but that does not necessarily mean that they were of Hispanic origin. The clay seems to point us in that direction, but the identification is still unclear, and more analyses will be necessary to confirm this attribution.

in its common association with olive preserve (olivae ex defruto) which was stored in amphorae. On the other hand, we should take into consideration a substantial number of labels which mention other grape by‑products (sapa, mulsum) and even products associated with fish preserves (muria). According to D. Djaoui (2016: 504–506), most of fish sauce inscriptions on H 70, if not all, are the result of the reuse of the amphorae, not of their original content. This brings us back to the idea of the Haltern 70 as a one‑product type, as were most Baetican amphorae during the Imperial period.12 According to Djaoui, this hypothesis is confirmed by the following:

Based on the above‑mentioned evidence, it can be argued, as Mongardi suggests, that in these amphorae the inscriptions indicate product + quality + number of batch + the mercator or mercatores in charge of their distribution + producer +, along the handle, the price of each amphora, written by a different hand. That is, the same commercial variables that we have been pointing out. It remains to be confirmed that E is the price of amphorae, and we should also investigate the reasons behind the palaeographic peculiarities of this entry.

a. The formal homogeneity of the original tituli picti, which only mention two products; these are related to one another and can, to all effects, be regarded as one: the defrutum and the olives preserved within it (olivae ex defruto). b. The total normalisation of the tituli picti, which follow the same structure as those on Baetican and Italian amphorae of salted products and, with some small differences, Baetican oil amphorae from the Guadalquivir Valley. The formula for the Haltern 70 (Figure 6) is as follows (Ehmig 2007: 256–257, 29).

6. Other (apparently) incomprehensible or exceptional examples: Haltern 70 From an epigraphic perspective, Haltern 70 amphorae are a controversial type (Panella and Rizzo 2014: 202–220, Tchernia and Martin‑Kilcher, in this volume). Not one of the few known inscriptions mentions wine, which is the sort of product with which the type is generally associated, following the interpretation of the word defrutum as a type of wine (Colls et al. 1977: 86–89); defrutum is found in the inscriptions written on the amphorae of the shipwreck Port Vendres II (c. AD 40). However, defrutum was not, strictly speaking, a wine, but rather was a sort of grape syrup resulting from the reduction of unfermented grape juice. The non‑alcoholic nature of this substance is clearly stated in the Digesto (33.6.3), which claims that the defrutum is more a preserve than a wine ([…] quod potius conditurae loco fuit). This is also reflected in how often the product is referred to as penuarius (Djaoui 2016) and also

The tituli written on Haltern 70 amphorae reproduce the commercial information that we are already familiar with: the product, its quality of variety, the merchant, the producer, and, when necessary, notes about secondary transactions.13 No quantitative information, such as that provided by entry C on Baetican amphorae of salted products, entries C and H on Italian amphorae of salted products, and entries  Concerning the immediate predecessor of the Haltern 70 type during the Republican period, the Ovoid 4 type, and most Republican Ovoids types in the Guadalquivir, the picture is rather complex: cf. García Vargas, Almeida and González Cesteros 2011; García Vargas et al. 2019). 13  The presence of element F on the Haltern 70 amphorae found at the Rhône, near Arles (Djaoui 2016), again points to this harbour’s role as an important entrepôt and redistribution centre. 12

59

Enrique García Vargas

α and γ on Baetican Dressel 20 oil amphorae, have been identified to date. However, most known inscriptions on Haltern 70 amphorae (Aguilera 2004; Ehmig 2007; Djaoui 2016) are so fragmentary it is impossible to certify with any certainty that this information was, in this case, absent from the record. The number of tituli picti found on Haltern 70 examples is so low, in any case, that arguments can only be tentative. At any rate, I think that Djaoui’s (2016) and Tchernia and Martin‑Kilcher’s (this volume) suggestions that the shape is related to the storage of olives in defrutum (9 mentions in tituli picti) or of defrutum (6 mentions in tituli picti) (Djaoui 2016: 489) are plausible.

(except when the goods were already packaged or bottled) and the degustatio that ensured the products were fit for consumption. Only under these conditions could the transaction be considered complete (perfecta) and the responsibility for the integrity of the cargo passed to the buyer. It is obvious that these transactions must have left documentary traces, but most of these documents are now lost. The main aspects of the transactions, however, were summarily written on the amphorae, which the buyer bought already full or which he filled up himself (diffusio). The information contained in 1st‑century amphorae is remarkably uniform: product, quality and amount/quantity (weight or number of amphorae). The labels also included the name of the buyer and the seller — D and E, for the original transaction, and D and F for re‑sales half‑way through the cargo’s itinerary.

7. Conclusions Although we should emphasise the differences in terms of geographical origin, commercial context and volume of trade, we should also note the large number of amphorae types in the Western Mediterranean which were annotated with remarkably regular labels (tituli picti) written in red or black ink. The information provided by these labels was often the same, and the structure and disposition of the different entries is also very similar.

This information was also expressed on ‘sample’ containers, some examples of which are known to have contained olive oil or wine (supra); these samples were used in the ceremony of the degustatio. The inscription, again, recorded the product (although in the case of pot 22 from Arles, the shape of the jar and its decoration may have been sufficient), the seller, the buyer, and the quantity (in volume or in number of containers). This information must be related to the commercial transaction and the transfer of responsibilities that said transaction involved, rather than the terms of maritime insurance agreed with the shipowner or navicularius. As shown by Tsulp 78, the insurance terms were recorded in a different document, and they were probably summarised on the amphora seals, which would have been the only visible part of the amphora between loading and unloading. Most amphora seals are found in contexts connected with the stopage and transport of goods, such as shipwrecks (cf. Mayer and Olivé 2008; Djaoui 2011).

These labels conveyed, almost always in this order, information on: the product contained in the amphora (A), sometimes accompanied by a general comment on quality (B); the weight (C); the name of the merchant transporting the goods (D); the name of the producer and also the person responsible for loading the product at its point of origin. Certain goods, such as Baetican salted products and olives in grape syrup, in transit along the Rhone‑Rhine ‘fluvial axis’ were labelled with extra inscriptions at entrepôts along the way. There, they were likely reloaded onto different vessels and batches of amphorae were then resold to middlemen mercatores, who assumed the job of taking the now‑divided cargo to its final destination. These secondary labels contained, again, personal names and numerals, which have been interpreted as referring to the commercial agents who now took over responsibility and to the number of amphorae included in each batch re‑sold (F). Finally, in origin, in transit or at the final destination, extra information could be added on the handle (G) or the body (H). These last entries were less ‘formal’ and conveyed information that was not always necessary for the culmination of the contract, such as the final destination or the final consumers of the amphorae.

For oil amphorae, a similar overall picture can be painted, especially before the indictiones system related to the annona, and the attached fiscal structure, came into being. The arrangement of entries differed, but the information provided by the labels is almost the same: buyer, weight, quality (in terms of provenance), seller, and, sometimes, price. The similarities are too great for this to be a coincidence. This basic information continued to appear even after the imposition of obligatory sales with the annona system (in this case, there was an explicit indication of the degustatio: Chic García 2002); from that point onwards, in any case, the formula became much more complex, and much additional information is added, owing to the control exercised over the cargos in each fiscal district involved in the delivery of the product (cf. Chic García 1988). The explicit indication of receptio or gustatio on 2nd‑century Dressel 20 oil amphorae stands as evidence of the fact that, during that period, oil was part of a system of obligatory sales and taxation that was imposed from the Severan period onwards as a result of

The private commercial transfer of goods was dominated by a sense of responsibility for the integrity of said goods — responsibility which was transferred to the buyer along with the ownership of the goods. When the goods in question were wine, oil or preserves, the integrity of the merchandise included its preservation for consumption. The seller was responsible for any deterioration in the quality of the product up to the moment when the transfer of responsibility took place. This transfer was generally preceded by the measurement of the cargo

60

Amphora contents as commodities: the structure and function of tituli picti in the western Mediterranean…

the financial difficulties suffered by the state; traditionally, the state had been a major buyer of olive oil, regardless of the purchasing system in place at any given time.

Curtis, R. I. 1991. Garum and salsamenta. Production and commerce in materia medica. Studies in ancient medicine 3. Leiden ‑ New York ‑ Copenhaguen ‑ Cologne, Brill. Djaoui, D. 2011. Découverte d’un double sceau en bois à date consulaire (épave de Tiboulen de Maïre, Marseille). Étude préliminaire. In SFECAG, Actes du Congrès d’Arles: 625–632. Marseille, Société Française d’Étude le Céramique Antique en Gaule. Djaoui, D. 2014. Découverte d’un pot mentionnant la famille des DD Caecilii dans un contexte portuaire situé entre 50‑140 apr. J.‑C. Découverte subaquatique à Arles, Bouches du Rhône, France. In R. Morais, A. Fernández, and M. J. Sousa (eds), As produções cerâmicas de imitação na Hispania. Actas do II Congresso da Sociedade de Estudos da Cerâmica Antiga da Hispânia (SECAH) (Braga, 4 a 6 de Abril de 2013). Monografias Ex Officina Hispana II (II): 161–178. Porto, Universidade do Porto. Djaoui, D. 2016. Les amphores de type Haltern 70: olives ou defrutum ? Et que faire des tituli picti concernant la sapa, le mulsum, la muria et l’hallex ? In D. Djaoui (ed.), Histoires matérielles : terre cuite, bois, métal et autres objets. Des pots et des potes: Mélanges offerts à Lucien Rivet. Archéologie et Histoire Romaine 33: 489–512. Autun, Éd. Monique Mergoil. Ehmig, U. 1995. Allex oder Anderes. Mainzer Archäologische Zeitschrift 2: 117–130. Ehmig, U. 1996. Garum für den Statthalter. Eine Saucenamphore mit Besitzeraufschrift aus Mainz. Mainzer Archäologische Zeitschrift 3: 25–56. Ehmig, U. 2003. Die römische Amphoren aus Mainz. Frankfurter Archäologische Schriften 4. Möhnesee, Richert Verlag. Ehmig, U. 2007. Die Römischen Amphoren im Umland von Mainz. Frankfurter Archäologische Schriften 5. Wiesbaden, Richert Verlag. Ehmig, U. 2014a. Publicité ou conséquences des risques du transport maritime ? Sur la fonction des tituli picti des amphores romaines dans le contexte des prêts maritimes. Cahiers du Centre Gustave Glotz XXV: 89–106. Ehmig, U. 2014b. Werbung oder Konsequenzen aus den Risiken bei Seetransporten? Zur Funktion von Tituli picti auf römischen Amphoren im Kontext von Seedarlehen. Frankfurter Archäologische Schriften 25: 85–98. Ehmig, U., Liou, B. and Long, L. 2004. Le garum de Caius Saturius Secundus, gouverneur de la province romaine de Rétie. Revue des Études Anciennes 106: 123–132. García Vargas, E. and Martínez Maganto, J. 2009. Fuentes de riqueza y promoción social de los «negotiantes salsarii» béticos durante el Alto Imperio romano: Una aproximación diacrónica. Archivo Español de Arqueología 82: 133–152. García Vargas, E., Almeida, R. R. de and González Cesteros, H. 2011. Los tipos anfóricos del Guadalquivir en el marco de los envases hispanos del siglo I a.C. Un universo heterogéneo entre la imitación y la estandarización. SPAL, Revista de Prehistoria y Arqueología de la Universidad de Sevilla 20: 185–283.

Finally, an interesting avenue of enquiry for the future is the possible connection between changes in the purchasing system, from auction sales to obligatory sales under the annona, and the disappearance of the Haltern 70 type, which would have been filled with grape syrup and olives from the same regions that produced the olive oil. The change in management in Baetican pottery workshops between the late 1st and early 2nd centuries, and the corresponding change in the inscriptions, which now often include the mention of a portus, as well as the growth of the sector in the Guadalquivir Valley, are other strands of the same phenomenon.14

Bibliography Aguilera, A. 2004. Los tituli picti, Epigrafia sobre àmfores Haltern 70 Bètiques. In C. Carreras Monfort et al. (eds), Culip VIII i les àmfores Haltern 7. Monografies del Centre d’Arqueologia Subaquàtica de Catalunya 5: 36–40. Girona, Museu d’Arqueologia de Catalunya and Centre d’Arqueologia Subaquàtica de Catalunya. Berdowski, P. 2003. Tituli picti und die antike Werbesprache für Fischprodukte. Münstersche Beiträge zur Antiken Handelgeschichte 22.2: 18–56. Berni, P. in press. Las otras ánforas del monte Testaccio. In Almazaras y ánforas olearias de la Bética romana. Actas del coloquio de la Casa de Velázquez, Madrid, 29‑30 octubre 2015. Camodeca, G. 1999. Tabulae Pompeianae Sulpiciorum (TPSulp.). Edizione critica dell’archivio puteolano dei Sulpicii. Vetera 12. Rome, Quasar. Chic García, G. 1981. Observaciones sobre las cifras pintadas en las ánforas olearias hispanas. Habis 12: 251–260. Chic García, G. 1988. Epigrafía anfórica de la Bética. Ecija, Gráficas Sol. Chic García, G. 2002. Degustatio o recognitio. In L. Rivet and M. Sciallano (eds), Vivre, produire et échanger: reflets méditerranéens. Mélanges offerts à Bernard Liou. Archéologie et Histoire Romaine 8: 335–342. Montagnac, Éd. Monique Mergoil. Colls, D., Étienne, R., Lequément, R, Liou, B and Mayet, F. 1977. L’épave Port‑Vendres II et le commerce de la Bétique à l’époque de Claude. Archeonautica 1. Curtis, R. I. 1984‑1985. Product identification and advertising on Roman commercial amphorae. Ancient Society 15‑17: 209–228.

 The number of workshops producing Dressel 20 amphorae increased substantially. Approximately 50% of known workshops were active in the mid‑1st century, and 75% by the 2nd century: Chic García 2001: 447. 14

61

Enrique García Vargas

García Vargas, E., Almeida R. R. de, González Cesteros, H. and Sáez Romero, A. (eds) 2019. The Ovoid Amphorae in the Central and Western Mediterranean between the last two centuries of the Republic and the early days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13. Oxford, Archaeopress. García Vargas, E., Bernal‑Casasola, D., Cottica, D. and Toniolo, L. 2020. Epigrafia della pila di anfore della Bottega del garum di Pompei (I 12, 8), In M. Osanna and L. Toniolo (eds), Fecisti Cretaria. Dal frammento ceramico al contesto: studi sul vasellame ceramico del territorio vesuviano. Studi e Ricerche del Parco Archeologico di Pompei 39: 135–142. Rome, L’Erma di Bretschneider. Giordano, C. and Casale, A. 1990. Iscrizioni pompeiane inedite scoperte tra gli anni 1954‑1978. Atti della Accademia Pontaniana 39: 273–378. Lerouxel, F. 2016. Le marché du crédit dans le monde romain (Égypte et Campanie). Bibliothèque des Écoles françaises d’Athènes et de Rome 374. Rome, École Française de Rome. Martin‑Kilcher, S. 1994. Die römischen Amphoren aus Augst und Kaiseraugst, 2‑3. Die Amphoren für Wein, Fischsauce, südfrüchte (Gruppen 2‑24) und Gesamtauswertung. Forschungen in Augst 7/2‑3. Augst, Römermuseum Augst. Martin‑Kilcher, S. 2002. Lucius Urittius Verecundus, négociant à la fin du Ier siècle, et sa marchandise découverte à Mayence. In L. Rivet and M. Sciallano (eds), Vivre, produire et échanger : reflets méditerranéens, Mélanges offerts à Bernard Liou. Archéologie et Histoire romaine 8: 343–353. Montagnac, Éd. Monique Mergoil.

Mataix Ferrándiz, V. 2019. Rethinking the Roman Epigraphy of merchandise: a metapragmatic approach. MAARAV, A Journal for the Study of the Northwest Semitic Languages and Literature 23.1: 177‑205. Mataix Ferrándiz, V. 2020. CIL IV 9591: propuesta reconstructiva de una locatio-conductio para el transporte de mercancías por mar. In V. Revilla Calvo, A. Aguilera Martín, Ll. Pons Pujol and M. García Sánchez (eds), Ex Baetica Romam: homenaje a José Remesal Rodríguez: 787-820. Barcelona, Universitat de Barcelona. Mayer i Olivé, M. 2008. Opercula, los tapones de ánfora: un indicador económico controvertido. In Instrumenta Inscripta Latina, II. Akten des 2. Internationalen Kollokium. Klagenfurt, 5‑8 Mai 2005: 223–239. Klagenfurt, Manfred Hainzmann‑Reinhold Wedenig Hrsg. Mongardi, M. 2014. L’instrumentum fittile inscriptum della colonia romana di Mutina e del suo territorio, Unpublished PhD thesis. Università di Bologna. Mongardi, M. 2016. Anfore di morfologia betica con iscrizioni dipinte dalla regio VIII Aemilia (con un’Appendice a cura di C. Capelli). In R. Járrega and P. Berni (eds), Amphorae ex Hispania: paisajes de producción y consumo. III Congreso Internacional de la Sociedad de Estudios de la Cerámica Antigua (SECAH) ‑ Ex Officina Hispana (Tarragona, 10‑13 de diciembre de 2014). Monografías Ex Officina Hispana III: 557–568. Tarragona, Institut Català d’Arqueologia Clàssica. Mongardi, M. 2018. Firmissima et Splendidissima Populi Romani Colonia. L’epigrafia anforica di Mutina e del suo territorio. Instrumenta 62. Barcelona, Universitat de Barcelona. Panella, C. and Rizzo, G. 2014. Ostia VI. Le terme del nuotatore. Studi miscellanei 38. Rome, L’Erma di Bretschneider.

62

How late antique dipinti contribute to a better knowledge of amphora contents Jean-Luc Fournet

Collège de France

Abstract: I present here a survey of the inscriptions (tituli picti or dipinti) found on Late Roman Amphorae (LRA 1, 2, 4, 5/6, 9, North African amphorae) by describing their typologies and focussing on what we can learn from them about the contents of the various types of amphora. Key words: tituli picti; dipinti; inscriptions; Late Roman Amphorae; Egypt.

1. Introduction While the Latin tituli picti ou dipinti1 on oil or salted fish amphorae from the Roman Period have been studied since the 19th century, particularly by Heinrich Dressel who proposed a masterful interpretation of them in CIL XV,2 the Greek ones have been neglected. It was not until 2008 that the first volume dedicated to Greek dipinti appears, namely those of Tebtynis (Egypt) published by Nikos Litinas.3 But this book deals only with Ptolemaic and Roman material. As far as late antique dipinti (4-7th centuries) are concerned, they remain understudied. This is due in part to a lack of interest in the Early Byzantine era for decades, while Hellenistic amphoric stamps took up all the attention of specialists in amphoric epigraphy (Garlan 2010), but is mostly for reasons related to the late antique amphorae themselves: 1. Their great variety and apparent heterogeneity have long slowed down attempts to develop typologies, and it was not until 1977 that John Riley succeeded in arranging this apparent disorder with a reasoned typology, an essential prerequisite for a proper study of inscriptions. 2. Unlike the dipinti from previous periods, those on late antique amphorae have an extremely deformed writing style that makes them difficult to decipher.

 I use the term dipinto in the conventional sense of an inscription made in ink on an amphora as a short expression for amphoric dipinto. The term, in its narrower sense, has already a certain tradition (e.g., Robinson 1959; Riley 1979; Davies 1984). It is also more convenient than titulus pictus, ‘commercial notation’ (Lang 1976) or ‘vessel’s notation’ (Litinas 2008) that, despite the precision of the last two expressions, have the disadvantage of being in two words and prevent the formation of derivatives as ‘dipintology’ that I proposed in Fournet 2012. 2  Dressel 1899, 560–699 in chapter III (‘Tituli picti in amphoris in monte Testaceo et in emporio repertis’), especially in the introduction (560–565). It is noteworthy that among some Greek dipinti he publishes (no.4860-4897), some are from the Byzantine period and belong to the category I shall discuss below: no.4890–4895 and 4897. 3  Litinas 2008. The reader will find a bibliography concerning dipinti found in Egypt in his introduction, pages 1-8. 1

3. The late antique dipinti are coded notations that can be read only if one decodes the system that informs them, and that depends on how the production and trade that were behind them were organized. After Jean Gascou and Tomasz Derda’s previous research,4 the work carried out in collaboration with Dominique Pieri on the material from Antinoopolis (excavations of the Istituto Papirologico ‘G. Vitelli’) and the study of other Egyptian dossiers I have worked on for the last twenty years5 have led to some advances in our reading and understanding of these inscriptions, which are of the highest interest for the history of ancient trade. I am pleased to present here the results, still provisional, of this long-term study, focusing on the information these dipinti give on amphora contents.6 But since the mention of contents is just a part of a dipinto, we cannot understand it if the whole dipinto is not understood. That is why I shall present here a typology of these dipinti, stressing the issue of contents. I will examine in turn four categories: the LRA 1, the amphorae from the Aegean Sea and mainland Greece (LRA 2), the Palestinian amphorae (LRA 4, LRA 5/6 and 9) and the North African spatheia.

 Gascou 1978, Derda 1992.  Dipinti from the Kellia (IFAO), from Gabbari/Alexandria (Centre d’Études Alexandrines), from Baouît (Louvre-IFAO), from Saqqara (Rijksmuseum van Oudheden), from Oxyrhynchos (Ashmolean Museum). I would like to thank those who have given me these dossiers and authorized me to publish them: Jean-Yves Empereur (Centre d’Études Alexandrines, CNRS), Nicolas Grimal and the successive directors of the IFAO (Bernard Mathieu, Laure Pantalacci and B. MidantReynes), Dominique Benazeth (Musée du Louvre, Paris), Maarten J. Raven (Rijksmuseum van Oudheden, Leyde), Helen Whitehouse (Ashmolean Museum, Oxford), Nick Gonis (University College, London) and Rosario Pintaudi (Università di Messina and Biblioteca Laurenziana, Florence). 6  The present paper will complete and correct what I have published in Fournet-Pieri 2008: 180–210 and Fournet 2012. 4 5

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 63–76

Jean-Luc Fournet

2. LRA 1 This amphora, produced in Cilicia, Isauria, Syria I, Rhodes and Cyprus, has long been considered a wine container (Fournet and Pieri 2008: 210–214, with bibliography). Is it confirmed by the dipinti? These are in fact extremely difficult to read and understand. They are located on several places on the amphora surface (neck, shoulder, under the handles) according to the nature of the data they contain and to the commercialization phases they correspond to. They can be classified into five types (ae). The first three (written in red ink) are inscribed while the wine is being put into the amphorae (as can be seen from the similarity of ink and handwriting), the fourth and fifth ones are written, respectively in black and red ink, at later stages. 2.1. Type a

This is on the shoulder, centred between the handles. It is composed of two, three or four lines (the number of lines increasing over time): • a1: the date according to the indiction system (the word ἰνδικτίων is always omitted). • a2: Christian formulas with, for instance, χμγ (acrostic for ‘Christ born of Maria’) and ϙθ (isopsephism for ‘amen’), the word Θεός ‘God’ or κύριος ‘Lord’ abbreviated, or a religious isopsephism (the most frequent being υπδ = 484 = Θεοῦ ‘from God’; here φπγ = 583 = Θεοῦ ἀμήν ‘from God, amen’?).7 • a3+4: the last line gives a more complex string. In its simplest pattern, attested at the first stages of the development of this amphora (LRA 1A), it is composed of the unit of measure ξέστης ‘sextarius’ abbreviated, followed by the quantity of wine the amphora contains, ranging from 20 to 30 sextarii and almost always expressed with fractions (½ and ¼; rarely ⅓) (= a4).

 This interpretation is due to Grenfell-Hunt 1906-1907: 10–11. There are many other isopsephisms that are still mysterious to me. 7

Type a (old style) (Antinoopolis inv. V110)

64

How late antique dipinti contribute to a better knowledge of amphora contents

Later on, the symbol for sextarius was usually omitted and the quantity was preceded by some letters, very stylized and of enigmatic content (= a3). Some logically thought they represented the product name (Gascou 1978: 26), but I believe they have the same value as in a2, that is to say a religious formula, as can be seen from legible examples.

As a result, the product name was not mentioned in the inscriptions of type a, probably because it was considered obvious and, for that reason, not worth writing down. But the kind of product contained in the amphora can be deduced from one of the other types of inscription (type c).

2.2. Type b

This is located in the area surrounding a, written sometimes upside down or at right angles to it. It only contains a number with fraction(s), ranging from 1 to at least 12, usually around 6. No explanation has been proposed for this. Given the fractions, it can’t be a batch number or anything like that, but it must be a quantity. It is likely to be the tare (weight when empty). Seeing that there was no standardization in late antique amphora production, it was necessary to measure the quantity of wine contained by each amphora in order to write it down on the shoulder (a4). This could not be done with a standard measuring jug: it would have required too much time. Yet the written quantities are very precise, for they have fractions (the smallest one is a quarter). The only

sensible solution was to weigh the amphora first when it was empty and again when filled. Then one subtracted the figure of the tare from the figure of the full amphora and converted the weight into sextarii by using tables of conversion of weight measures into liquid measures — similar to that preserved by the P.Lond. V 1718 (Bell 1917: 154–165) — that allowed the conversion of pounds into sextarii (but in our case, the conversion tables were specifically adapted to the weight of the ceramic and of the wine). But we have not been able to study enough complete specimens to definitively confirm this hypothesis. Nevertheless the tare system is clearly used on the latest form of LRA 1, the Antonov 9 (7th-beg. 8th century AD) and on other Aegean amphorae (see below, § 3.2).

65

Jean-Luc Fournet

2.3. Type c

This inscription is written upside down under the handles. It is very difficult to read because the writer was hampered by the handle (as the modern reader is!) and was obliged to write in small characters (not to mention the numerous phonetisms that distort the words). It is usually composed of 3 lines. Each of the first two lines contains a proper name (in the genitive when not abbreviated); the third one a whole number (up to hundreds). It was suggested that they were the potter’s name and a serial number. They are more likely to give the wine grower’s name (c1) and the batch number (i.e. number of amphorae filled with the same production sold at a certain moment) (c4). As to the second name in line 2 (c2), it could be a patronymic. However no clear anthroponym can be read as in line 1. The words one can decipher are in most cases nonGreek-sounding names. My suggested interpretation is that they were toponyms indicating the place of the wine-producing estate. I have been able to read, on two

fragments, ‘Apamea,’ but in most cases we are dealing with microtoponyms belonging to the native substratum (for instance, Dēbal( ), Garer( ), Katia, Rhana) and for that reason difficult to identify.8 One, Rhadam( ) could be identified on the basis of the Constantinople Synaxarium that cites an estate named Rhandama close to Anarzaba where Saint Marinos of Anarzaba was buried.9 The estate name or more rarely the number is sometimes followed by ὀρινίου (c3), usually abbreviated, which, according to Hesychius, means ‘vine that grows up trees,’ i.e. ‘high-trained vines’ (french ‘conduit en hautain’).10 This adjective, practically unknown to literary sources, designated a climbing vine as opposed to a vine on stakes (Brun 2003: 35–37). This word is so far the only clue given by the dipinti about the content of this amphora: it transported wine.  Fournet and Pieri 2008: 198–199 must be corrected. Angilas, in the published example, is probably a place name: ‘(a place called/field of) Angilas’. But if the place name has primarily a discriminator value, we cannot exclude that, in a number of cases, the family name was used, especially when it was unusual. 9  As a matter of fact, this place name is transmitted in various forms. The Tabula Imperii Byzantini (TIB 5, 1990: 390) gave preference to Rhadamnos after the Acta Sanctorum, but our dipinti show that the other form is the correct one. I owe these references to D. Feissel. 10  Hesychius, ο 1217: ὀρινίαι· ἀναδενδράδες. The adjective ὀρίνιος was a hapax until now. 8

66

How late antique dipinti contribute to a better knowledge of amphora contents

2.4. Type d

This inscription is written in black on the neck. After two or three lines of Christian formulas (d1), we come accross a personal name in the genitive (d2) and then the name of the wine (d3), eventually with an indication of its quality (like πρωτεῖον ‘first quality’) and of its quantity. Given that, in some very clear cases, d was added above a, it is tempting to interpret this inscription as ‘labels’ added by the merchant or transporter advertising his own products. We have been able to collect a large dossier of ‘labels’ of a certain Apollinarios, found in Oxyrhynchos and Antinoopolis.11 You can notice the unusually exuberant first stroke of the χ — which could have been (so to speak) the ‘logo’ of this trader. In one clear case

(Derda 1992: no.II.1), the name is Saint Isidoros, which I’d interpret as a monastery name, that plays here the role of a trading enterprise. The name of the wine is usually ἑμάτι(ο)ν ou ἑμάτινον12 that I interpret as a systematically incorrect13 form of αἱμάτιον ‘blood’ or αἱματίνον ‘blood-coloured’ (Fournet 2012: 257). It could be a metaphorical brand name of a dark-coloured wine (like, mutatis mutandis, the Hungarian Bull’s Blood).14  In Ashmolean inv. GO 464 (my example), it is tempting to read εμωτιν, but parallels suggest the reading of an alpha with a redundant stroke instead of an omega. 13  Same phenomenon with κέρδος often written καίρδος. 14  Not to be confused with the garum called αἱμάτιον (Geoponica, XX 46, 6 ed. Beckh): cf. Botte 2009: 20. 12

 The dossier from Antinoopolis is published in Fournet and Pieri 2008: 203–206. 11

2.5. Type e

This consists of a figure written in large letters on the neck with a brush in red ink. This type is attested only on material found in Antinoopolis, which seems to prove that the dipinto was written in the importing country, probably when the amphorae arrived in Antinoopolis.

in some harbours of the coast under the control of a representative of the business organization who wrote on them the tare (b), then the quantity of wine (a) and the name of the producer and the production zone (c). • d was perhaps written down after a/b/c or instead of a/b/c. The relationship between a/b/c and d is difficult to understand. • e was written down at the arrival of the amphorae.

In conclusion: • a/b/c seem to have been written simultaneously (first b, then a, finally c) at the point of departure, that is to say when filling amphorae: they are always in an identical ink and a-b have in common graphic peculiarities. This suggests that the amphorae were filled and inscribed

Even if not clearly mentioned in the main inscription (a), the content was wine, as can be deduced from c3 and, if I am not mistaken, from the later dipinto written down by the merchant (d).

67

Jean-Luc Fournet

3. Amphorae from the Aegean Sea and mainland Greece The most common type is the LRA 2 (Aegean Sea; cf. Pieri 2005: 85–93). I’ll also take into account the M273 (Samos), which can have dipinti belonging to the same typology. The inscription is located on the lower part of the neck. I have been able to identify at least two types of inscriptions: 3.1. Type 1 This is particularly well represented by a series of inscriptions on amphorae found in Romania and published by Popescu 1976. No pictures are published but only drawings which are hard to interpret. Derda 1992 re-edited them. But the product names are almost unreadable. Type 1 is very similar to the spatheia inscriptions (see § 5.1), but is written in red ink (or sometimes incised).

• a: 2 or 3 lines of Christian formulas, extended by adding Θεοῦ σῴζοντος ‘God saving (us)’ or ‘with God’s salvation’ usually after b. But it can be either before (Derda 1992: no.II.3). • b: personal name or mostly name of a religious institution15. • abis: after the name (but sometimes before, Derda 1992: no.II.3), there is again a Christian formula: Θεοῦ σῴζοντος ‘God saving (us)’ or ‘with God’s salvation’.

• c: product name. It is always illegible except the end -ελέου: it is a compound noun ending in -έλαιον, designating a kind of oil.16 • d: the quantity. It is repeated, probably three times as on other types of amphorae (except LRA 1) even if no example in the corpus gathered above has preserved more than two lines. It seems to be expressed with a measure expressed by a symbol (  judging by Popescu 1976, no.398) which could be for the sextarius or the pound. The second option seems best according to the information given by my type 2.

 The reading ὕρ(χη) γλυκελέου (l. -λαίου) adopted by Derda after Popescu’s bad drawings is highly doubtful: We don’t expect the first to be abbreviated in such a way; moreover, ὕρχη normally refers to a storage container and not a transport container. After Popescu’s drawing (including no.142 and 398), I would be tempted to read γριμελέου (the first three letters seem to suit the traces); but this type of product is attested with certainty only in Palestinian amphorae (see below, § 4.2). So I would suggest not imposing this reading here. 16

 As for the element b of spatheia inscriptions, we should hesitate about the nature of this line: Christian formula to link to a (particularly since it may be followed by another line of Christian formulas, abis) or the producer’s name? By analogy with the spatheia, I favor the second solution (following Derda 1992). 15

68

How late antique dipinti contribute to a better knowledge of amphora contents

3.2. Type 2 I suspect the existence of another type of inscription, coming from a very small number of specimens, which does not really allow us to guarantee its validity. The

particularity of this type lies in the way the amount is expressed and its being written in black ink unlike type 1.

• a: Christian formulas. • c: the quantity. It is given jointly by the tare (κοῦφον, litt. ‘empty (amphora)’17) and by the net weight

(καθαρόν), that is to say the weight of the full amphora minus the tare, repeated three times. Both are calculated in pounds (λίτραι abbreviated).18

17

 The amount is expressed in the same way on the latest LRA 1 (Antonov 9). See above, § 2.2. 18

 Cf. Mayerson 1997.

3.3. Re-uses Here I report a case of re-use: an amphora reused as a honey container.

The decoration around the inscription, unusually neat, indicates that this inscription was written down by the

owner (Lazaros) rather than the producer or merchant, on an amphora which was reused.

69

Jean-Luc Fournet

4. Palestinian amphoras (LRA 4, LRA 5/6 and LRA 9) 4.1. LRA 4 Wine amphora, produced in the Gaza area (4th-7th centuries) (Pieri 2005: 101–114). The LRA 4 dipinti do not have a structure or a writing as homogeneous as those of LRA 1. They are found in lesser quantities. It is therefore difficult to propose fine typologies based on well-documented series. The dipinti (in red or black ink) is located on the belly of the amphora between the two series of ridges. Some are written vertically; other horizontally on the upper part of this zone. They seem to have only one inscription unlike LRA 1. As for their contents, the structure of these inscriptions, when they can be read, is rather close to the LRA 1:

• a1 (= a2 on the LRA 1): Christian formulas. • a2+3 (= a3+4 on the LRA 1): a Christian invocation is followed, directly or separated by a vertical line, by the quantity of wine expressed in sextarii (the measure unit is always omitted) in the form of a figure with fractions (½ and / or ¼).

Because of their large mouths, these amphorae are often the subject of re-uses for various products: figs, olives,19 cheeses (Gabbari 97-14.3, unpublished).  I could identify a set of LRA 4 from the Kellia (Egypt), bearing inscriptions attesting that they were reused as fig and olive containers, in particular within the economical activities of the St. Menas monastery (in the Alexandria area). 19

70

How late antique dipinti contribute to a better knowledge of amphora contents

dipinti force us to look at this amphora type, at least those with dipinti, as an oil amphora, as we will see below.

4.2. Agora M334 (or LRA 9) An amphora dated to the 4th-beginning 8th centuries, considered a wine container (Pieri 2005: 137–138). But

Inscriptions are located on the neck, written in black ink in a single text block.

These dipinti are very similar to those of spatheia:

typical of the Palestine region – I cannot explain the first element of this compound (Fournet and Pieri 2008: 208–210). • c: name of a person or of a religious institution. This is most likely the production domain. • d: the quantity of the product is repeated three times after the symbol of a measure ( or , likely for the pound), often a fourth time on the right of the main inscription.

• a: Christian formulas. • b: product name. The name that comes up most is γριμέλαιον, an otherwise unknown term, which also occurs on another Palestinian amphora, the socalled Bag-shaped type (see § 4.3). Seeing that these amphorae are only of Palestinian origin, it is safe to conclude that this word refers to a kind of oil (ἔλαιον)

71

Jean-Luc Fournet

4.3. Bag-shaped amphora (or LRA 5/6) Amphora regarded as a wine container (Pieri 2005: 114–127). Dipinti are written down in black ink on the shoulder.

The inscription typology and the content are the same as in LRA 9. Contrary to what is generally thought, this

amphora could contain oil as clearly seen from the mention of grimelaion.

5. North African amphorae

5.1. Spatheia

Whatever the type of amphorae, the notations are always written in black ink on the neck in one text block. Their typology is fairly homogeneous.

Garum, wine or oil amphorae. Types cited here after: SPA 1 (4th-mid 5th century) – SPA 2 (5th century) – SPA 3 (6th-7th century) specialized for garum (Bonifay 2004).

72

How late antique dipinti contribute to a better knowledge of amphora contents

• a: Christian formulas • b: personal name [optional]. It is a priori tempting to interpret them as names of producers who may be either laymen (Paulos in my example) or religious institutions (Saint Theodoros, Saint John, Saint Menas).20 But see below. • c: product name [optional, not in my example]. The only product whose reading is clear is garum (Botte 2009: 19–20). It is accompanied by precisions (γάρου πόλεoς, γάρου λευκοῦ πόλεoς, γάρου πόλεος λευκοῦ, κ(ε)υντοπόλεoς) that are problematic.21 • d: the quantity was repeated three times after the measure symbol ( or , sometimes ). The first symbol is likely to be for sextarius, but it results in too low values for this measure. It would be best to interpret it as the symbol for pound (λίτρα).

It is surprising that these inscriptions are written in Greek while Latin was the usual language in Proconsular Africa.22 It could be argued that the use of Greek was justified by exports to the Greek-speaking world (actually, the material is of Egyptian origin), or that Greek was the commercial lingua franca. But what is troubling if one considers that b gives the name of the producer is the mention of the sanctuary of St. Menas, which is located in the Alexandria area. Two explanations are possible: 1. these amphorae have been reused to sell an Egyptian production. 2. these notations are due to the merchants. The name in b would be that of the mercator. If it is correct, this would mean that the monastery of Saint Menas could have acted as a commercial entreprise, which was suspected also in the case of the LRA 1 ‘labels’ (cf. § 2.4). This third explanation is more likely. It confirms that the content was indeed of African origin.

 Derda 1992: 137, ‘A name in the genitive appears, often preceded by the adjective ἅγιος […], presumably referring to the church or the monastery where the amphora was filled with the product mentionned further on in the text.’ He seems to consider that the names without ἅγιος refer also to religious estates. The ellipse of ἅγιος is however impossible in the case of a saint. 21  The interpretation that I proposed in Fournet and Pieri 2008: 207 should be revised in light of some new dipinti recently discovered in Antinoopolis. I will deal with this issue elsewhere. 20

 We have at least one stopper in connection with a spatheion neck (Antinoopolis inv. 2005 [12], unpublished): it is inscribed in Latin. 22

5.2. Other African types I would like to point out a set of African amphorae of medium and large size that contained a previously

unknown product, the aphrospanon. It is evidenced so far by four dipinti (three still unpublished):

73

Jean-Luc Fournet

74

How late antique dipinti contribute to a better knowledge of amphora contents

I suggest this was the trade name of an oil produced in Africa (ἀφρό-) and Spain, i.e. Baetica (-σπανον).23 It is tempting to compare this word with an expression attested in a Roman inscription from Hispalis (CIL II 118024, 161–169), in which Julius Possessor is said to be adiu/tori Ulpii Saturnini praef(ecti) Annon(ae) / ad oleum Afrum et Hispanum recen/sendum. ‘assistant of Ulpius Saturninus, praefectus annonae, for inspecting oil from Africa and Baetica’. Though the expression oleum Afrum et Hispanum refers here to two oil productions, one from Africa, the other from Spain, this parallel shows that these two oils were in some way related to one another. Does the compound noun aphrospanon mean that both oils came from regions administratively connected or that they were regarded as resulting from a similar way of manufacturing practised in Africa and Baetica, so that they could be considered as an identical or similar product and eventually mixed?

have been, a specific identity that made it recognizable by the wine-importing countries. We can say that paleography was in the service of marketing! This homogeneity implies that all the LRA 1 contained the same product. For this reason, one would not feel the need to write down the product name in the a-c dipinti unlike the dipinti of other kinds of amphorae, hence the heterogeneity of their inscriptions. But, despite the differences between LRA 1 dipinti and the others, we are struck by the great unity of formulas evidenced by all the types of dipinti, especially in Christian invocations (acrostics, isopsephisms or others) identical from Carthage to Apamea. 2. These dipinti bring meaningful contributions to several fields: as far as palaeography is concerned, the LRA 1 dipinti are nearly the only testimonies of cursive writing (I mean non epigraphic) in this area of the Byzantine Empire, and indeed they reveal pecularities so far unknown (Fournet 2012). They also give new data about toponymy. But their greatest contribution is to commercial history, by giving names of products which were unknown till now. Thus they help us to fill in some lacunas of literary and papyrological sources and improve our knowledge of commercial Greek.

6. Conclusion I’d like to conclude by stressing two points: 1. The LRA 1 dipinti, displaying a complex system, with several notations located on various places on the amphora, written in red ink in a very stylized scripture, constitute a homogeneous ensemble that differs from the dipinti of the other types of amphorae, which are more easy to read and understand, written usually in black ink in one textual block according to variable formulars. dipinti on LRA 1 in red ink stylised writing at several points of the amphora according to their function homogeneous formular product not mentioned





Bibliography Bell, H. I., 1917. Greek Papyri in the British Museum V. London, London British Museum. Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports International Series 1301, Oxford, Archaeopress. Botte, E. 2009. Salaisons et sauces de poissons en Italie du sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Brun, J.-P. 2003. La vigne et l’huile dans la Méditerranée antique. Viticulture, oléiculture et procédés de fabrication. Paris, Errance. Davies, S. M. 1984. The Dipinti, Stamps and Graffiti. In M. G. Fulford and D. P. S. Peacock (eds), Excavations at Carthage: The British Mission I, 2. The Avenue du President Habib Bourguiba, Salammbo: the pottery and other ceramic objects from the site: 141–153. Sheffield, University of Sheffield. Derda, T. 1992. Inscriptions with the Formula θεοῦ χάρις κέρδος on Late Roman Amphora. Zeitschrift für Papyrologie und Epigraphik 94: 135–152. Dressel, H. 1899. Corpus Inscriptionum Latinarum XV: Inscriptiones urbis Romae Latinae. Instrumentum domesticum, pars II, fasc. I. Berlin, George Reimer. Fournet, J.-L. 2012. La ‘dipintologie’ grecque  : une nouvelle discipline auxiliaire de la papyrologie ? In P. Schubert (ed.), Actes du 26e Congrès international de papyrologie: 249–258. Geneva, Droz. Fournet, J.-L. and Pieri, D. 2008. Les dipinti amphoriques d’Antinoopolis. In R. Pintaudi (ed.), Antinoupolis I. Scavi e materiali 1: 175–216. Florence, Istituto Papirologico ‘G. Vitelli’.

dipinti on other types of amphorae usually in black ink clear writing one text block heterogeneous formular product mentioned

By their very nature, LRA 1 dipinti are evidence of a highly centralized commercial organisation. As a matter of fact, behind the numerous small or medium producers whose name and production quantity is indicated in the type c inscriptions, we can guess that a huge trading organisation covering a large area straddling several administrative provinces (Cilicia, Isauria, Syria I, Rhodes and Cyprus) existed. In the absence of information given by more ‘talkative’ sources, the structures of this organisation remain for us somehow blurred or obscure in detail. The formal homogeneity of the dipinti at the same time testifies the traders’ intention of giving this production, however geographically various it might

 This term may refer to the method of production used for the North African and Betic oils as for σπανέλαιον: cf. Kramer 1990: 261–264. 24  = ILS 1403 = AE 1965, 237 = AE 1971, 171 = AE 1991, 993 = CILA II, 23. 23

75

Jean-Luc Fournet

Garlan, Y. 2010. Progrès récents de l’épigraphie amphorique grecque en Occident. Antienyj Mir i Arkheologija 14: 371–395. Gascou, J. 1978. Amphores byzantines à dipinti grecs de Saqqara. Bulletin de liaison du Groupe international d’étude de la céramique égyptienne 3: 24–27. Grenfell, B. P. and Hunt, A. S. 1906. Excavations at Oxyrhynchus. In E. F. Griffith (ed.), Egypt Exploration Fund, Archaeological Report 1905-1906: 8–16. London, Sage Publications. Grenfell, B. P. and Hunt, A. S. 1907. Excavations at Oxyrhynchus. In E. F. Griffith (ed.), Egypt Exploration Fund, Archaeological Report 1906-1907: 8–11. London, Sage Publications. Harper, R. P. 1997. Upper Zohar. An Early Byzantine Fort in Palestina Tertia. Final Report of Excavations in 1985-1986. Oxford, Oxford University Press. Kirwan, L. P. 1938. Inscriptions. II- The Greek Inscriptions. In W. B. Emery and L. P. Kirwan, The Royal Tomb of Ballana and Qustul, II: 401–405. Cairo, Government Press. Κramer, J. 1990. Die Bedeutung von σπανέλαιον, Zeitschrift für Papyrologie und Epigraphik 81: 261-264 (repr. id., Σπανέλαιον / oleum spanum. Ιn Von der Papyrologie zur Romanistik. Archiv für Papyrusforschung Beiheft 30: 301–306. Berlin - New York 2011). Kuhn, M. 2009. Catalogue of a Collection of Coptic Ostraca in the Graeco-Roman Museum in Alexandria. In J.-Y. Empereur (ed.), Alexandrina 3. Études alexandrines 17: 273–339. Cairo, Institut français d’archéologie orientale.

Lang, M. 1976. The Athenian Agora. Volume XXI. Graffiti and dipinti. Princeton, The American School of Classical Studies at Athens. Litinas, N. 2008. Tebtynis III. Vessels’ Notations from Tebtynis. Fouilles de l’Institut français d’archéologie orientale 68. Cairo, Institut français d’archéologie orientale. Mayerson, P. 1997. A Note on κοῦφα ‘Empties’. Bulletin of The American Society of Papyrologists 34: 47–52. Pieri, D. 2005. Le commerce du vin oriental à l’époque byzantine (Ve-VIIe siècles). Le témoignage des amphores en Gaule. Bibliothèque Archéologique et Historique 174, Beirut, Institut français du Proche-Orient. Popescu, E. 1976. Inscripţiile greceşti şi latine din secolele IV-XIII descoperite în Romaniâ. Bucharest, Academiei Republicii Socialiste Romania. Raven, M. J., Verschoor, V., Vugts, M. and van Walsem, R. 2011. The Memphite tomb of Horemheb, commander-inchief of Tutankhamun, V: the forecourt and the area south of the tomb, with some notes on the tomb of Tia. Papers on Archaeology of the Leiden Museum of Antiquities (PALMA) 6. Turnhout, Brepols. Riley, J. A. 1979. The Coarse Pottery from Berenice. In J. A. Lloyd (ed.), Excavations at Sidi Khrebish Benghazi (Berenice) II. Supplement to Libya Antiqua V (II): 91–467. Tripoli, Society for Libyan Studies. Riley, J. A. 1981. The pottery from the cisterns 1977.1, 1977.2 and 1977.3. In J. Humphrey (ed.). Excavations at Carthage 1977 conducted by the University of Michigan VI: 85–124. Ann Arbor, University of Michigan Press. Robinson, H. S. 1959. The Athenian Agora. Volume V. Pottery of the Roman Period, Chronology. Princeton, The American School of Classical Studies at Athens. Wilfong, T. 2000. Textual Remains. In C. Meyer, L. A. Heidorn, W. E. Kaegi and T. Wilfong (eds.), Bīr Umm Fawakhir Survey Project 1993. A Byzantine Gold-mining Town in Egypt: 25–26. Chicago, The Oriental Institute.

76

The reuse of transport amphorae as packaging containers in the Roman world: an overview J. Theodore Peña

University of California, Berkeley, Department of Ancient Greek and Roman Studies Dwinelle 7233, Berkeley, California, USA

Abstract: The reuse of amphorae as packaging containers in the Roman world raises potentially significant challenges regarding our efforts to employ the amphora record to evaluate the geography, chronology, and intensity of the production, distribution, and consumption of the various foodstuffs and other substances regularly packaged in amphorae. Five different markers indicate or point to the reuse of amphorae as packaging containers: the presence of tituli picti or graffiti indicating multiple episodes of filling; the presence of damage indicating either multiple episodes of filling or protracted use; the presence of content remains indicating multiple episodes of filling; the presence of a label or content remains indicating filling with an irregular substance; and the presence of a non-standard stopper. A modest number of shipwreck sites mostly in the western Mediterranean and ranging in date from the 2nd century BC to the 7th century AD have produced amphora cargoes that bear one or more of these markers and that in some cases display notable levels of heterogeneity within amphora classes in terms of form, size/volume, manufacturing technique, fabric, and/or epigraphy that likewise suggest reuse of these vessels as packaging containers. A small number of facilities for the packaging/distribution of foodstuffs at Pompeii and Korinth have yielded sets of used amphorae apparently collected and being held in storage for reuse as packaging containers. Four papyri of Hellenistic, Roman, and Byzantine date indicate the reuse of amphorae as packaging containers in Egypt, while three inscriptions from Rome may point to the exercise of this practice there during the middle and late imperial periods. The use of an amphora for the packaging of one kind of content did not in many cases render it unsuitable for reuse for the packaging of one or more kinds of content, although in some instances this might have been the case. Given the low cost of new amphorae, it seems unlikely that their reuse as packaging containers reduced costs to any substantial extent, although this practice may have proven convenient in that it reduced the number of empty containers requiring discard or facilitated packaging operations in regions incapable of producing a suitable supply of amphorae or equivalent packaging containers or at times when the supply of such containers was reduced or interrupted. Although the reuse of amphorae as packaging containers may have led to difficulties with the identification of the substances contained in them, it seems likely that tituli picti and/or other labelling devices were employed to avoid this problem. The systematic evaluation of the papyrological corpus for texts relating to amphora reuse and the detailed description of amphora assemblages from shipwreck sites may well expand our evidence for the reuse of amphorae as packaging containers, pointing to regional, chronological, and/or institutional patterns in the practice and/or associations with particular amphora classes and/or kinds of content. The scattered evidence currently available suggests that the reuse of amphorae as packaging containers was not particularly common for transfers by ship, and it may be that amphorae were reused more often in this way for local transfers of foodstuffs. Key words: ancient Rome; trade; amphorae; reuse; foodstuffs; packaging; shipwrecks.

1. Introduction In this contribution I present a general overview of the reuse of amphorae as packaging containers in the Roman world. After discussing various background considerations relating to the reuse of amphorae as packaging containers I review first the archaeological and then the textual evidence for this practice. I then consider three significant questions raised by the reuse of amphorae as packaging containers, indicate research that might be undertaken with a view to improving our understanding of the practice, and conclude by drawing some general conclusions regarding both its incidence and its significance for amphora studies. 2. Background considerations I begin by defining what I mean by the reuse of an amphora as a packaging container. Here what I have in mind is the use of an amphora to hold some substance so

that it could be transferred from that substance’s point of production or collection or from some point further along from this in a distribution chain (in cases in which the substance in question had been packaged in some other container at its point of production/collection) to some other location further along in a distribution chain. I assume that amphorae were produced by their manufacturers and acquired by their initial users (if the two were not the same) principally to serve as packaging containers, although in some instances they would have been manufactured and or acquired for some other purpose. For example, some newly manufactured amphorae (the incidence is impossible to estimate) were presumably employed as storage containers, that is, to hold some substance with the intent that it remain at or very close to the point of filling (e.g., on the same or a neighbouring property). After its initial use as a packaging container (which might have involved the protracted storage of the filled and stoppered vessel at its final destination or at some other point) an amphora

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 77–91

J. Theodore Peña

normally would have been emptied of its content and treated in one of three different ways: it would have been reused — either as a container or for one of a wide variety of other applications (many of which involved some more or less substantial physical modification to the vessel) (Peña 2007a: 61–192); it would have been recycled by being reduced to sherds or smaller-size particles for use as a raw material (Peña 2007a: 250–271); or it would have been discarded (Peña 2007a: 272–318).

If, however, amphorae were regularly and/or intensively reused as packaging containers, this would weaken the validity of the assumptions that undergird studies of this kind, since the containers that are the focus of this work might have held a reuse content different from that class’s principle content and/or that originated outside the locale or region where the class originated. For this reason it would be helpful if we were able to expand our general knowledge of the practice of reusing amphorae as packaging containers, determine the degree to which it was associated with specific amphora classes or kinds of content, and establish whether there were any specific institutional, geographic and/or chronological patterns in its incidence.

One of the ways in which an amphora might have been reused was as a packaging container. This might have involved the vessel’s refilling with a content — here termed a reuse content — identical or similar to its initial content — here termed its prime-use content — or with some significantly different substance. A vessel might have been reused as a packaging container one or more times before being reused for some other purpose, recycled, or discarded. The reuse of amphorae as packaging containers in any locale at any particular time might have been a regular practice and/or an intensive practice, that is, one that involved the ongoing/ continuous reuse of amphorae and/or one that involved the reuse of large numbers of amphorae.

Besides considerations of shape, size, weight, and robustness, the suitability of an amphora for reuse as a packaging container would have been conditioned by the extent to which its opening and emptying in any antecedent use had produced damage that limited its general utility for such a purpose, and/or the presence of a pitch lining or content residues that limited its suitability for the packaging of particular substances. The reuse of amphorae as packaging containers would have involved the acquisition and reconditioning of amphorae that had been emptied of their content. The reconditioning procedure might have involved checking the vessel for overall soundness and specific defects; cleaning to remove precipitates and absorbed residues, followed by drying; reworking the mouth, when this had been subject to damage; pitching (or re-pitching); and, finally, the removal of tituli picti or other labels that had served to identify the vessel’s antecedent content, owner, and/or intended recipient. It was presumably to some extent advantageous if the vessels to be employed belonged to one or at most a limited set of classes that were reused on a regular basis, as this would have simplified reconditioning, filling/stoppering, storage, and transport operations, while also perhaps limiting any confusion that may have arisen regarding the identity of the vessels’ content. Where the reuse of amphorae as packaging containers was a regular and/or intensive practice we may suppose that it was facilitated by brokers who specialized in the collection and perhaps also the reconditioning of used containers and the provision of these to the various sorts of entities that had need of them.

Refilled amphorae might have been employed to transfer their content to locations situated at varying distances from the point where they were filled. In this regard we can recognize three more or less distinct kinds of transfers: local transfers, involving a supply chain that began and ended within the service area of the central place (however defined) where the amphora was filled (and also probably — although by no means necessarily — where its content originated); peri-local transfers, involving a supply chain where the end-point lay in one of the central place service areas that bordered on this territory; and export transfers, involving a supply chain where the end point lay somewhere beyond this cluster of service areas. As will be considered below, the type of transfer in question might have played a significant role in the specific practices involved in the reuse of amphorae as packaging containers. It has been widely assumed that the vessels belonging to a specific amphora class were normally employed in their initial use for the packaging of a single, specific substance that was generally produced or collected in the specific locale or broader region within which that class originated. This content is here termed that class’s principal content, with any other kind of content characterized as an irregular content. Based on this assumption, scholars have evaluated the amphorae from shipwreck sites to gain insights into the mechanisms and geography of the distribution of the substances known or posited to have been the primary content of the classes represented. They have also carried out qualitative and quantitative analyses of amphora assemblages from consumption sites to gain insight into geographic and diachronic patterns in the provision to and consumption of these substances at these locations.

The reuse of amphorae as packaging containers might have occurred in a variety of different economic contexts, including the distribution and sale of substances on the market, the internal transfer of substances in connection with elite self-supply, and the mobilization of substances for a state or ecclesiastical supply initiative of some kind. Export transfers may well have differed from local and peri-local transfers in certain fundamental ways in that they more often involved complex distribution chains that entailed one or more changes in possession and/or

78

The reuse of transport amphorae as packaging containers in the Roman world: an overview

ownership of the filled containers. Among other things, concerns regarding the identity and quality of the content of the amphorae stemming from the fact that they were reused containers may have been more of an issue in export transfers, as potential end-users would have tended to be less familiar with at least some of the individuals and entities that made up the supply chain. 3. Archaeological evidence We can begin our consideration of the archaeological evidence for the reuse of amphorae as packaging containers by considering the markers that either indicate with certainty or raise the possibility that an amphora was reused as a packaging container. We can identify five of these: Marker 1: the presence on an amphora of multiple labels (tituli picti or graffiti) indicating two or more instances of filling. Example: Dressel 2-4 from the Casa del Menandro at Pompeii (Stefani 2003: 211). This container bears three distinct tituli picti (CIL 4.9325). The first two, executed in white paint, appeared one above the other on one side of the vessel’s neck (Figure 1). The first of these is a four-line text that indicates a content consisting of metellianum vinum, a variety of wine from the area of Surrentum that was placed in the container in AD 62, that is, 16-17 years prior to Pompeii’s destruction. The second, limited to a single line, appears immediately below the first in letters of a different size executed in a different hand. It indicates a content of acetum alexandrinum, that is, Alexandrian vinegar. The third text, which consists of a single line executed in red paint, appears on the other side of the amphora’s neck and indicates that the amphora and its contents were the property of a certain Titus Claudius Anthus. The first of the three texts likely relates to the amphora’s prime-use, the second to an episode of reuse, and the third to one or the other of these. Whether the episode of reuse involved the amphora’s use as a packaging container or a storage container is unclear.

Figure 1: Upper part of Dressel 2-4 from the Casa del Menandro at Pompeii bearing three tituli picti (two visible) indicating that it was filled at least twice (Maiuri 1933, Figure 187). Example: Keay 62A from the La Palud 1 shipwreck (Garnier 2007: 27–29). Organic residue analysis carried out on this amphora demonstrated that it had a pine pitch lining and had once contained olive oil. Since olive oil generally has been assumed to be incompatible with the presence of a pitch lining, this raised the possibility that the container had been used to hold some other substance, presumably either wine or a processed fish product of some kind, prior to be being filled with oil. (For the problematic nature of this assumption see Section 5 below.) Marker 4: the presence on an amphora of a label or in an amphora of content remains that indicate that it was filled with an irregular content or with a content that originated at some distance from the container’s point of origin. Example: Keay 35A reused as construction fill in the San Lorenzo Maggiore ecclesiastical complex in Milan (Bocchio 1990: 146; Peña 2007a: 112). While this class of amphora, which originated in Tunisia, is generally regarded as an olive oil container (Bonifay 2004: 135; Bonifay 2007: 14, 24), the example in question bears a titulus pictus in Greek indicating a content of wine, along with a Christian religious slogan. This suggests that the container was reused as a packaging or storage container, quite possibly in connection with an ecclesiastical supply initiative of some kind.

Marker 2: the presence on an amphora of damage (and also possibly repair) that indicates either two or more instances of filling or notably protracted use. Example: Late Roman 1 from the Yassi Ada shipwreck (Van Alfen 1996: 202–203). This container bears three vertical gouges on the inner surface of its rim and the upper part of its neck, evidently produced when its stopper was removed with a sharp instrument. This suggests that the amphora was filled, stoppered, then unstoppered and emptied of its content at some point prior to its filling with the content that it held at the time that it was brought aboard the ship.

Marker 5: the presence in the mouth of an amphora of a stopper produced using a method that differs from the standard method employed for stoppering the class in question or a method that employed materials that originated at some distance from the amphora’s point of origin.

Marker 3: the presence in an amphora of content remains that indicate two or more different contents or of a pitch lining and content remains that are incompatible with a pitch lining.

79

J. Theodore Peña

Figure 2: Neck of Dressel 2-4 of non-Egyptian origin from Quseir al’Qadim with stopper made of plaster and potsherds of Egyptian origin. Left: photo of upper surface of stopper; centre: view of stopper from underside showing sherds; right: profile drawing (Images courtesy of Ross Thomas). Example: Dressel 2-4 from Quseir al-Qadim (ancient Myos Hormos) (Thomas 2011: 15–16, 27, 31). This vessel, the fabric of which indicates that it originated outside Egypt — perhaps in Cilicia or Cyprus — was found with its stopper preserved in situ (Figure 2). This was fabricated by forcing a piece of textile into the upper part of the vessel’s neck and then inserting a plug made up of a mixture of sherds and plaster. The upper surface of the plug was smoothed, impressed with a sealing, and then painted with red pigment. The sealing bears the name of a man who was presumably a merchant who owned the amphora and its content along with an indication that this was wine. The sherds used in the fabrication of the stopper were of Egyptian origin, indicating that it was produced somewhere in Egypt. This suggests that the amphora was imported into Egypt, emptied of its content — almost certainly wine — then filled with a reuse content — certainly wine, presumably of Egyptian origin — and then re-stoppered. The fact that the amphora found its way to Quseir al-Qadim — far from any center of wine production — indicates that this episode of reuse involved the employment of this amphora as a packaging container for an export transfer.

BC (Parker 1992: 211, 291). Among the ship’s cargo were c. 60 examples of the Greco-Italic amphora (a wine container from Tyrrhenian Italy) that contained pitch (Joncheray and Long 2002). These displayed a high degree of heterogeneity in terms of origin and date, and several showed conspicuous damage (e.g., handles and spikes broken away), indicating that they were very probably reused containers.

To date, markers of the kinds just described have been documented for only modest numbers of amphorae, and, by themselves, these vessels provide only limited insights into either the broader patterns or specific practices involved in the reuse of amphorae as packaging containers. In a small number of instances, however, more or less sizable sets of amphorae bearing one or more of these markers have been recovered as cargo at a shipwreck site, and occurrences of this kind have the potential to provide important insights into both broader patterns and specific practices.

Pisa San Rossore B: the remains of a merchantman that sank in the harbor at Pisa (Roman Pisae) in 7 BC or shortly thereafter (Bruni 2000: 42–43). A total of 17 amphorae were recovered from the wreck, including seven Lamboglia 2s, six Dressel 6As, two examples of the Dressel 9 (a fish products container from coastal southern Spain), and two examples of the Haltern 70 (a probable wine container from Spain). The Lamboglia 2s and Dressel 6As contained a variety of irregular substances that are thought to be reuse contents, including fruits (peaches, cherries, plums), nuts (walnuts, chestnuts), and inorganic materials (ocher, arsenic sulfide, volcanic sand) (Pesavento Mattioli et al. 2000). The amphorae were found shimmed in place in the vessel’s hold with chunks of volcanic rock originating in the Bay of Naples, raising the possibility that they were filled and brought aboard the ship somewhere in this region.

Sud-Caveaux 1: the remains of a merchantman that sank off the Mediterranean coast of France c. 30‑20 BC (Cibecchini 2013). This ship was carrying c. 100 amphorae, for the most part examples of the Lamboglia 2 and the Dressel 6A (both wine containers from Adriatic Italy) that were filled with pine pitch. These displayed a high degree of heterogeneity in terms of origin and date, strongly suggesting that they were reused containers. The ship was presumably bound for Marseille with pitch destined for supply to one or more shipyards and/or amphora production workshops. It most likely took on its cargo either at a port in Catalonia or at Narbonne (Roman Narbo Martius).

The list of such sites includes the following: Heliopolis 2 – Nord Levant: the remains of merchantman that sank off the Mediterranean coast of France between the end of the 3rd and the 2nd quarter of the 2nd century

80

The reuse of transport amphorae as packaging containers in the Roman world: an overview

Cala Culip 4/D: the remains of a merchantman that sank off the coast of northeastern Spain in the AD 60s or 70s (Parker 1992: 157). This ship was carrying at least 76 examples of the Dressel 20 (an olive oil container from southern Spain) (Nieto Prieto et al. 1989: 61–74). The majority of these were closed with highly irregular stoppers cut from sherds rather than the purpose‑made stoppers often associated with this class, raising the possibility that they were filled and stoppered on more than one occasion. This ship appears to have been engaged in cabotage trading along the Mediterranean coast of Spain and the adjacent coast of France.

same time, perhaps as a single, unitary cargo. The ship was also carrying a substantial quantity of glass fragments apparently destined for recycling, and was probably also outfitted with a large salt-water tank that served for the transport of live fish. Toniolo (2002-2003: 115–116) reports mean capacities in litres for the sets of containers belonging to each of the four classes of amphorae represented among the ship’s cargo as follows: African 1: 43l; Tripolitanian 1: 85l, Knossos 19: 26l; Grado 1: 17l.1 The Knossos 19s and Grado 1s bore a pitch lining, while the African 1s and Tripolitanian 1s presented no evidence of such a lining. Several of the African 1s, Knossos 19s, and Tripolitanian 1s contained macro‑remains, and the analysis of these revealed that the first two of these classes held salted whole sardines, while the third held Atlantic chub mackerel. The stoppers for all three of these classes involved the use of disks cut from sherds belonging to one or another of these same classes. At least 24 of the Grado 1s preserved traces of a titulus pictus on the neck. While the reading of these presents difficulties, in nine cases it is clear that the text indicates a content of liquamen flos, that is, top-quality liquamen, and in six cases there is a line consisting of the letter A or B followed by the letters XX and a symbol resembling a radiate crown. The letters XX, if a weight figure, likely report the empty weight of the container in pounds.2 As one would expect for amphorae containing liquamen, none of the Grado 1s yielded macro remains.

Grado (‘Julia Felix’): the remains of a merchantman that sank near the head of the Adriatic Sea in the middle decades of the 2nd century AD (Parker 1992: 197). The amphora cargo of this ship is discussed in detail below. Yassi Ada B: the remains of a merchantman that sank off the coast of Anatolia c. AD 625-630 (Parker 1992: 454–455). This ship was carrying c. 900 amphorae, including at least 103 Late Roman 1s and 819 Late Roman 2s (van Alfen 1996; van Alfen 2015; Ward 2015; van Doorninck 2015: 206–208). The excavators believe that the ship may have been operated by an ecclesiastical establishment, and that at the time of its sinking it was engaged in the delivery of foodstuffs under the annona militaris to troops of the Byzantine state campaigning in the Levant, with the some of the containers filled with olive oil and some probably filled with wine. A notably high degree of variability in form and fabric within both groups of amphorae, the presence of two examples of the Late Roman 1 that exhibited damage indicative of an earlier episode of use, and the presence of several examples of the Late Roman 2 bearing graffiti indicating a content of lentils led the excavators to infer that these containers were in significant measure reused amphorae, perhaps assembled in the context of a military emergency.

The contents of all four classes presumably originated somewhere around the head of the Adriatic, an important center for the confection of fish products. The Grado1s were almost certainly the only prime use containers among the ship’s cargo, with the other three classes consisting of reused containers. The distribution of the Grado 1 is limited to northern Italy, both shores of the Adriatic, and the interior of Pannonia and Illyria, suggesting that the ship had sailed from a port somewhere in the upper Adriatic and was probably bound for another port along the shores of this same body of water (Toniolo 2002-2003: 125).

Of these occurrences, the Grado shipwreck warrants detailed discussion, as it provides unusually informative evidence regarding certain aspects of the practices involved in the reuse of amphorae as packaging containers. In this instance, the excavators were able to recover effectively the entire preserved portion of the ship’s cargo. This included at least 566 amphorae belonging to four different classes, among which were 204 examples of the African 1 (an oil container from Tunisia), 23 examples of the Tripolitanian 1 (an oil container from Tripolitania), 154 examples of the Knossos 19/Dressel 5 (a wine container from the Aegean), and 185 examples of the Grado 1 (a fish products container from the upper Adriatic) (Auriemma 2000; Toniolo 2002‑2003; Auriemma and Pesavento Mattioli 2009). These were arranged in the hold by class, with the African 1s placed amidships, the Tripolitanian 1s towards the bow, the Knossos 19s towards the stern, and the Grado 1s arranged in various free spaces that remained near both the bow and the stern. This arrangement suggests that all of the containers were brought aboard the ship at the

Toniolo (2002-2003: 130) considered the practices that lay behind this noteworthy instance of the reuse of amphorae as packaging containers. Observing that the three classes of reused containers display a considerable chronological  Toniolo terms these figures the capacità media for each of the four sets of amphorae. She does not indicate the method employed to calculate vessel capacities, the number of examples of each of the four classes for which a measurement was taken, the capacity measurement for any individual amphora, or the standard deviation figure for any of the four classes. 2 With a median capacity of 17l the examples of this class would have weighed something on the order of 20 Roman pounds and been capable of holding at least c. 40-50 Roman pounds of liquamen. See Bernal‑Casasola et al. 2014: 227 for comparative data for a set of Dressel 21-22s, which, with capacities on the order of 15-18l, had empty weights that ranged from 23 to 32 Roman pounds and were filled with from 53 to 80 Roman pounds of fish sauce. 1

81

J. Theodore Peña

Class Grado 1 Knossos 19 African 1 Tripolitanian 1

Capacity Liters 17 26 43 85

Capacity modules 1 1.5 2.5 5

Capacity sextarii 32 48 80 160

Capacity amphorae 2/3 1 1 2/3 3 1/3

Capacity modii 2 3 5 10

Content liquamen sardines sardines mackerel

Figure 3: Volumetric data and content information for the four sets of amphorae from the Grado shipwreck. The mean capacity figures presented in Column 2 (on which the figures presented in Columns 3-6 are based) are presumably rounded and approximate. spread, she concluded that these vessels were not collected on a casual basis, but rather were deliberately assembled by the owner of the ship with particular aims in mind and curated by him over a substantial period of time, with replacements acquired as individual examples reached the end of their use life. In her view, the characteristic of these containers that made them of interest to the ship owner was not the size of their mouths (which would have rendered them more or less suitable for the introduction and removal of whole salted sardines or mackerel), but rather their overall size and shape, which would have allowed him both to arrange cargo aboard his vessel according to a set of well-established procedures and to sell his merchandise in standardized amounts according to weight. Interesting in this regard is the existence of a set of regular capacity relationships among the four sets of containers. The relevant data, presented in Figure 3, suggest that the mean capacity of the Grado 1s — 17l — corresponds to a module equal to two-thirds of an amphora (that is, the amphora as a standard measure of volume equal to one quadrantal or cubic foot, divisible into 48 sextarii) or, perhaps more conveniently, two modii, a measure equal to one-third of an amphora that was widely employed in the Roman world for dry measures. The capacities of the Knossos 19s, African 1s, and Tripolitanian 1s, in turn, would be equal to three, five, and ten modii, respectively. These four amphora sizes combine to form an additive sequence, with the values of the two smallest containers summing to equal the value of the next largest container, and the values of these three containers summing to equal that of the largest container. The existence of this structured set of relationships, which would not seem to be a random occurrence, re‑enforces Toniolo’s conclusion that the suite of amphorae aboard the Grado ship were carefully selected with specific aims in mind.

among which were 34 Dressel 20s, 16 examples of the Dressel 23 (an olive oil container from southern Spain), 19 examples of the Almagro 50, 16 examples of the Almagro 51C, and 7 examples of the Beltrán 72 (all fish products containers from Portugal) and 32 examples of the African 2 (a fish products container from Tunisia) (Bost et al. 1992: 117–177; Bombico 2015: 28). These were arranged in the hold by class in a way that led the excavators to conclude that all had been loaded aboard the ship at the same time. Several other shipwrecks in the western Mediterranean dating to the second half of the 3rd and the 4th century AD have produced similar suites of African 2s and Portuguese fish products amphorae, and scholars have found it difficult to infer the trade patterns that would have generated cargoes of this kind. Worth considering in this connection are the strongly contrasting levels of homogeneity/heterogeneity displayed by, on the one hand, the set of African 2s and, on the other, two of the three classes of Portuguese fish products amphorae, namely the Almagro 50s and the Beltrán 72s. The relevant data are presented in Figure 4. The African 2s included examples of three different variants — the 2A, 2C, and 2D — that have been thought to represent three discrete moments in the diachronic development of this class, and were manufactured in at least two different fabrics. Thirteen of the 34 vessels bore a stamp, each produced with a different die and representing at least ten and perhaps as many as 13 different stamping entities. All of the vessels bore a pitch lining, suggesting a content of fish products, and whereas two were found to contain remains of fish products, another two contained large numbers of olive pits, pointing to a content of olives. In contrast, 17 of the 19 Almagro 50s bore a maker’s stamp, with these produced using just four different dies representing only two different workshops, while three of the seven Beltrán 72s bore a maker’s mark, with these all produced using the same die as that employed to stamp one of the Almagro 50s.

As already alluded to in relation to the Heliopolis 2, SudCaveaux 1, and Yassi Ada B shipwrecks, in some instances sets of amphorae from a shipwreck display a notably high degree of heterogeneity in terms of form, size/volume, manufacturing technique, fabric, and/or epigraphy that suggests that they represent a group of reused containers. An interesting example of this kind of indicator is presented by the Cabrera 3 shipwreck, the remains of a merchantman that sank off Cabrera, one of the Balearic Islands, in or shortly after AD 267 (Parker 1992: 81). In this case, the excavators were able to document only a moderate portion of the ship’s cargo that had been stowed in the hold amidships. This included 124 amphorae,

From the location of the shipwreck and the likely points of origin of the ship’s cargo the excavators surmised that the vessel likely sank in the course of voyage from Spain to Italy, most likely departing from Cadiz (Roman Gades) and bound for Ostia/Portus. The set of African 2s and combined set of Almagro 50s/Beltrán 72s from Cabrera 3 clearly had substantially different histories before being brought aboard the ship. The former appear to have been

82

The reuse of transport amphorae as packaging containers in the Roman world: an overview

Class African 2 Almagro 50 Beltrán 72

Number documented 32 19 7

Sub-classes

Fabrics

3 1 1

2 1 1

Number stamped 13 17 3

Number entities 10-13 2 1*

Number dies 13 3 1*

Figure 4: Data for sets of African 2, Almagro 50, and Beltrán 72 amphorae from the Cabrera 3 shipwreck. * Same die/stamping entity as one of those attested for Almagro 50s. subject to a complex a set of operations that resulted in the extensive mixing of sets of containers manufactured and presumably initially filled at multiple locations, perhaps at substantially different times, while the latter appear to have been subject to a relatively simple set of operations that saw groups of containers originating and presumably initially filled at the same locus at the same time remaining together. While it cannot be excluded that the African 2s were prime-use containers bound for Italy in the context of a triangular trade that involved shipping cargoes of African fish products and olives to Spain before then sending these on to Italy, a plausible alternative interpretation is that African fish products were shipped from Tunisia to Spain, where they were consumed, with their empty containers reused for the packaging of Spanish fish products and olives destined for export to Italy (Peña 2007a: 76–80).

for use as a facility for the storage and sale of wine and perhaps other foodstuffs (Berry 1997; Fulford 1998; Peña 2007a: 88–94). In the southeast corner of the garden of the Caupona di Amarantus was a group of several (probably 12-16) amphorae set in inverted position, perhaps in two tiers, that presumably had been emptied of their content and set aside in storage for reuse. These included one Gaza amphora (a wine container from Gaza or Lower Egypt), one Cretan wine amphora, and two containers identified as Aegean amphorae (probable wine containers from the Aegean), with most or all of the remaining containers Dressel 2-4s of Campanian origin. The impluvium of the Casa di Mestrius Maximus contained c. 15-20 amphorae lying on their side in a disorderly fashion. These included nine examples of the Cretan 2 (a wine container from Crete), two examples of the Pseudo-Koan amphora (a wine container from the Aegean), and one example of the Dressel 21-22 (fish products amphorae from southern Italy or Sicily). These may have been placed in the impluvium to soak in water as part of the reconditioning process for reuse after having been emptied of their content.

Also informative regarding the practice of reusing amphorae as packaging containers is a small number of packaging/distribution facilities that have yielded sizable sets of amphorae that appear to have been placed in storage for reuse after having been emptied of their content. Four such facilities have been uncovered at Pompeii and its environs and one at Korinth.

Officina del Garum degli Umbrici (1.12.8): a modest atrium house that was converted for use as a facility for the confection and sale of fish products and perhaps also the sale of wine (Curtis 1979; Peña 2007a: 82–88; Bernal‑Casasola et al. 2014). This facility had over 200 amphorae being held in storage in Room 13, a rear courtyard, a large portion of which had likely been emptied of their content and set aside for reuse. These containers, which were segregated into groups by content, included a set of 80 Italian fish products amphorae stacked upside down in the north corner of the courtyard in a four-tiered arrangement, with the mouths of the vessels in tiers two to four set over the spike of a container in the tier below. These included four examples of the Dressel 21/Botte 1, from western Sicily, 69 examples of the Dressel 22/Botte 2, from the Peloritan/Calabrian massif of northeast Sicily and Calabria, and seven examples of the Dressel 22/ Botte 3, from the west coast of Italy between Etruria and Campania. Macro remains of fish products were recovered from the interiors of these containers, the pavement beneath the mouths of the containers in the lowest tier, and the exteriors of the spikes of the containers in the lower three tiers, where these had come into contact with the mouths of the containers in the tier above them. This suggests that these amphorae were stacked in this area immediately after having been emptied of their content. More than a dozen olive oil amphorae of African origin,

The Pompeian sites include the following: Casa del Vinario (9.9.6-7): a modest atrium house that appears to have been converted for use as a facility for the storage and sale of foodstuffs, including wine, olive oil, and fish products (Jashemsky 1967; 1979: 233–242; Peña 2007a: 94–96). One room on the premises (Room O) held 29 amphorae stored in inverted position that presumably had been emptied of their content and set aside for reuse. Fifteen of these bore a titulus pictus, including one example of the Schöne 6 (a small fish products container of local origin), one amphora classified as a Schöne 12 (most likely a Dressel 2-4), and 13 examples classified as a Schöne 11, one of which was probably a Tripolitanian 1, and the remainder most likely examples of the Form 2020 amphora, a class that has been identified as a wine container from eastern Sicily (Peña 2007b) and more recently as an olive container from Tunisia (Bonifay et al. 2015). We have no information regarding the other 14 amphorae from this room. Casa di Q. Mestrius Maximus/Caupona di Amarantus complex (1.9.11-12): a modest atrium house joined with a small structure of unclear type that were converted

83

J. Theodore Peña

including examples of the Tripolitanian 1, Ostia 59, and Mañá C1 and C2, were being stored in the west corner of the courtyard. These were apparently set in inverted position, and had thus presumably been emptied of their content. Approximately 20 wine containers of Campanian origin, including examples of the Dressel 2-4 and the flatbottomed Dressel 2-4, were stored in right-side up position in the east corner of the courtyard. The orientation of these raises the possibility that they were filled at the time of the town’s destruction, although it cannot be excluded that they too had been emptied of their content and set aside for reuse. This array and arrangement of containers may represent the initial step in a set of procedures for the collection of a set of used amphorae of various discreet sizes for use as packaging containers similar to the set of procedures employed to assemble the suite of amphorae aboard the Grado ship.

less complete amphorae were recovered in this room, including eight examples of the Kapitän 2 (a probable wine container from the Aegean), two examples of the Middle Roman 4 (a probable wine container from the south coast of Anatolia and perhaps also Cyprus), two examples of the Rhodian amphora (a wine container probably from Rhodes), three examples of the Cretan 4 (a wine container from Crete), one or two examples of the Forli amphora (a wine container from the Adriatic coast of Italy), and at least two and perhaps as many as five Korinthian amphorae. Slane has conjectured that these were used containers that were assembled on the premises so that they could be employed to package the substance being produced there. 4. Textual Evidence A variety of texts, including papyri, literary works, and inscriptions point in a more or less definitive fashion to the regular and/or intensive reuse of amphorae as packaging and/or storage containers.

Oplontis, Villa B: a warehouse/wine bottling facility located on the shore of the Bay of Naples at Oplontis, c. 3.2km north-northwest of Pompeii (Fergola 2004a; 2004b; Peña and McCallum 2009: 191; Lagi 2015; Thomas 2015; Gazda and Clarke 2016: 62–65, 160–77). At the time of its destruction this facility contained at least 1200 amphorae, c. 90 percent of which were Dressel 2-4s. The groundfloor peristyle contained at least c. 400 amphorae stacked in upside down position in a three-tiered arrangement similar to the one attested at the Officina del Garum degli Umbrici. Also in this space were two stoves, two bronze vessels containing pine pitch residue, and a block of pitch that all might have served for the pitching of amphorae. According to Lagi, the amphorae stored in the peristyle consisted for the most part of locally-manufactured Dressel 2-4s — though also included numbers of used containers that had previously held either Falernum or Surrentinum — again presumably Dressel 2-4s — with this inference presumably based on the presence of tituli picti on these containers. Also present was a small number of Spanish fish products amphorae. While it is not clear what proportion of the locally-manufactured amphorae were used, it would appear that a substantial number of these showed evidence of previous use. In Lagi’s view, the amphorae stored in this part of the facility consisted of used containers collected at Pompeii and its environs that were destined for refilling with locally-produced wine for local distribution, more distant distribution being excluded by the fact that the containers in question had been used. The apparent scale of the facility’s wine bottling operations and its seaside location, however, suggest that a substantial portion of the wine bottled there would have been loaded aboard ships for export transfer.

An extensive corpus of papyri indicates that it was a common practice in Hellenistic, Roman, and Byzantine Egypt to refer to specific types of amphorae by a geographic name that referred to that container’s point of origin (Kruit and Worp 2000; Hickey 2012: 186–197). In many cases it is possible to link these denominations to known amphora classes and to infer the standard content of the container in question. In several instances papyri refer to one or more examples of one of these container types as holding an irregular content, including foodstuffs such as cheese, sweetmeats, legumes, or grain (Mayerson 1992). In many cases — particularly those involving amphorae that originated outside Egypt — these would appear to constitute instances of reuse, although whether this was as a packaging container or a storage container is often unclear. In a limited number of cases papyri either make clear reference to or suggest the intensive reuse of amphorae for the bottling of new wine at locations in Egypt (Peña 2016). From the Hellenistic period we can cite PCairZen. 4.59741 and PSI.8.859, a pair of papyri belonging to the Zenon Archive that probably date to the period 256-248 BC. Together, these two documents constitute an account for the transport of 3200 amphorae from Kerke to the neighboring town of Philadelphia (a distance of c. 10km) by donkey and wagon for pitching (Kruit and Worp 2000: 71; Grace and Empereur 1981: 425). Among these were 1793 six-chous capacity amphorae of what was presumably Egyptian origin and 1407 imported amphorae, termed Kouriaká, Páphia, Pária, and Chía — presumably amphora types associated with the cities of Kourion and Paphos on Cyprus and the islands of Paros, and Chios, respectively. While the specific purpose of this undertaking is not indicated, it presumably involved the preparation of amphorae for the packaging or storage of a large quantity of a liquid or liquids of some

Turning to Korinth, Building 7, a modest commercial structure located immediately to the East of the theater, contained a room (Room 4) that at some point during the second half of the 3rd century AD housed a facility for the packaging and perhaps also processing of some unidentified substance (Slane 2004). Several more or

84

The reuse of transport amphorae as packaging containers in the Roman world: an overview

kind. That this was probably new wine is suggested by the fact that the operation was being carried out during the second half of Payni (i.e., the first half of May), shortly after the vintage. The imported amphorae referred to in the documents were almost certainly used containers, as might have been some or all of the Egyptian amphorae.

2nd century AD (CIL 6.37804) commemorates a certain Gaius Commissius Succesus, whose occupation is given as negotians Porto Vinario lagonaris (jug/small amphora trader at the Portus Vinarius). The most plausible interpretation of this occupational designation is that the man in question worked as an amphora broker in the city’s Portus Vinarius wine trading district, apparently located in Trastevere. His activities might have involved purchasing amphorae (particularly small ones?) that had been emptied of their content and reselling these to wholesalers, retailers, and/or consumers who required empty containers into which they could transfer the content of large-size amphorae, damaged amphorae, large-size casks, skin containers, sacks, basketry containers, and so forth. A second epitaph from Rome, in this case from the Catacombs of Praetextatus, thus dating to the late 2nd century AD or later (CIL 6.9488/ ICUR 5.15389), is for a certain Leontia, whose is termed an ad Porta(m) Trigemina(m) lagunara (jug/small amphora [?] woman at the Porta Trigemina) — here again perhaps a person whose work was in some fashion connected with the collection and/or distribution of used amphorae, in this case based in the vicinity of the Porta Trigemina.

From the Roman period we have PIand.99, dating to the early 3rd century AD (Rathbone 1983: 90–91). This document is concerned with operations relating to the management of the wine being held on an agricultural estate at some unknown location in Egypt, most likely in the Oxyrhynchite nome. It refers first to the tasting of the wine produced on the estate, which it indicates was bottled in xenikokerámia, that is, imported amphorae. As was the case with the documents from the Zenon Archive, these were presumably used containers. The document then addresses the need to replenish the stock of imported wines being held by the estate, indicating that these were bottled in containers termed Hadriánai, Aminaíai, and Tmolítikai. These are presumably amphora types associated with Hadrianum, a wine from the Adriatic coast of Italy, aminaeum, a quality of wine grown in various regions, though particularly associated with Sicily, and Tmolítes, a wine from the Mount Tmolos region of western Anatolia. It seems reasonable to assume that one, two, or all three of these container types were among the xenikokerámia employed for the bottling of the wine produced on the estate.

Finally, a fragmentary monumental inscription of the late 4th century AD found in the area of San Silvestro in Capite in the northeastern Campus Martius (CIL 6.1785=31,931) preserves part of what appears to be an edict issued by the Praefectus Urbi that fixed the amounts of gratuities that landowners were obliged to pay various classes of functionaries upon the consignment of wine being collected by the state as tax in kind under the land tax at the Ciconiae, a landing facility on the left bank of the Tiber River, most likely in the vicinity of the Tor di Nona. While this complex and fragmentary text is open to differing interpretations, one possibility is that the ampullae that it specifies were to be returned to the taxpayer following the completion of the degustatio (tasting) were smallsize amphorae in which the wine in question had been transported to Rome (Peña 1999: 177–178; Vera 2006). If this was the case, it is plausible to infer that the taxpayers transported these containers back to their estates and employed them for the packaging of subsequent consignments of tax wine.

Finally, from the Byzantine period we can cite POxy.55.3804, which can be dated to AD 566. This is the annual account of Theodorus, the estate steward of the Apion family, which had extensive holdings of agricultural land in the Oxyrhynchite nome. Among other things, it records a lot of 1100 containers termed koȗpha palaía (‘old empty jars’ — i.e., used amphorae) that had been purchased and delivered to vine dressers for the bottling of new wine. Turning to literary works, a passage in Pliny the Elder’s Naturalis historia (19.41.142) states that a type of cabbage known as halmyrídia that grows in unspecified sea-side regions remains green even for a lengthy sea voyage if it is packaged in cadi olei (oil amphorae) that have been ‘dried’. If, as appears likely, the drying of these containers refers to the removal of residues of a prior content of oil, this points to the reuse of oil amphorae (rather than the use of new ones) for the packaging of cabbages. Unclear, however, is whether this would have been for the commercial distribution of these cabbages, including in some cases export transfers by sea, and thus perhaps both a regular and intensive practice, or simply for the occasional provisioning of ships’ crews.

5. Questions raised by the reuse of amphorae as packaging containers We may now turn to a consideration of three significant questions raised by the reuse of amphorae as packaging containers. The first of these is the extent to which the presence in an amphora of the remains of its prime-use content or an antecedent reuse content or the presence of a pitch lining would have rendered it unsuitable for reuse for the packaging of certain substances.

In the realm of epigraphical evidence, three inscriptions from Rome may point to the regular and/or intensive reuse of amphorae, presumably as packaging and/ or storage containers. An epitaph on an altar tomb set up outside the Porta Salaria that can be dated to the

85

J. Theodore Peña

As shown above, there is copious evidence that amphorae that had been employed for the packaging of wine (which in many cases would have been provided with a pitch lining) were reused for the packaging and/or storage of wine. This includes the amphora assemblage from Oplontis Villa B and the papyri from Egypt noted above. Further, that amphorae that had held wine could be reused for the packaging of fish products is demonstrated by the Knossos 19s from the Grado shipwreck, that they could be reused for the packing of fruit, nuts, and various inorganic materials is indicated by the amphorae from the San Rossore B shipwreck, and that they could be reused for the packaging of pitch is demonstrated by the amphorae from the Heliopolis 2 and the SudCaveaux 1 shipwrecks. For a long time it was assumed that wine amphorae that had been provided with a pitch lining were probably not as a rule employed to hold oil, the supposition being that, as oil dissolves pitch, any oil placed in a pitched amphora would have suffered some degree of adulteration. Recent research, however, has demonstrated that in many instances amphorae employed for the packaging of oil were, in fact, pitched, requiring the abandonment of this assumption (Garnier et al. 2011: 408–410).

The second question is the extent to which the employment of used amphorae for the packaging of foodstuffs would have represented an advantage of some sort. In terms of cost, it is worth noting that the Edictum de Pretiis, issued in AD 301, sets the price for a lagona with a capacity of one half-amphora — perhaps meaning a small-size amphora — at a figure equal to between 1.67 and 6.25 percent of the price that it sets for one half-amphora of the various grades of wines listed in this same document (Peña 2007a: 28). This suggests that for low-grade wine the cost of a new container might have represented a significant, if still minor fraction of the value of the wine. On the other hand, the Heroninos Archive, a set of papyrus documents pertaining to the management of an estate in the Fayum during the period AD 249-268, shows that new, locally manufactured amphorae wholesaled for between 0.5 and 1.3 percent of the wholesale cost of the wine bottled in them (Rathbone 1991: 167). This substantially lower figure may represent more accurately true wholesale as opposed to retail costs, or it may reflect the situation in Egypt, where the regular and intensive employment of used amphorae for the bottling of new wine as suggested by the papyri noted above worked to hold down the price that could be charged for new amphorae.

There is considerably less evidence for the reuse of amphorae that had held oil as packaging and/or storage containers. The African 1s and Tripolitanian 1s from the Grado shipwreck, however, demonstrate that these could be reused for the packaging of fish products, while the Dressel 20s from the Cala Culip 4 shipwreck may provide evidence that they were reused for the packaging of oil, and the passage from Pliny the Elder noted above may indicate that they were reused for the packaging or storage of cabbages. The oil amphorae from the Officina del Garum degli Umbrici presumably were set aside for refilling with some substance or other, and while we have no evidence regarding what this might have been, processed fish products seems the best conjecture. The Keay 35A reused in the construction of San Lorenzo in Milan noted above suggests that containers that had held oil could also on at least some occasions be employed for the packaging and/or storage of wine. It seems possible that this practice would have produced some adulteration of the wine, however, and it should be noted that this container might well have been reused in the context of an ecclesiastical charitable initiative of some kind, in which issues of quality of content were perhaps less significant than in other circumstances, particularly those associated with market distribution.

With regard to convenience, it is not clear that it would have been more convenient for an establishment that required containers for the packaging of some substance to meet this need by gathering used containers through its own efforts or by purchasing these from a broker than it would have been to obtain new amphorae directly from one or more pottery workshops. Indeed, the acquisition of new amphorae might have offered considerable advantages, since the establishment could have contracted for the delivery of the exact number of containers that it required in specific sizes, rather than having to settle for what happened to be available on the second-hand market. On the other hand, in many places at many times the regular consumption of large quantities of the substances regularly packaged in amphorae would have generated very large supplies of used containers the disposal of which would have represented something of an inconvenience for wholesalers, retailers, and/ or consumers, and it might have been a relatively straightforward process for such establishments to acquire sizable sets of specific kinds of amphorae that suited their purposes in a condition suitable for reuse as packaging containers as these were needed. There would have been a regular incentive for this practice in areas in which for some reason or other the local pottery industry was incapable of manufacturing a sufficient number of suitable containers and where there was no industry capable of providing some alternate type of container, such as casks or skin containers. There may also have been occasional, short-term incentives for this in instances in which there was some interruption of the supply chain for amphorae or in which there was a sudden and/or an unexpected increase in the production of the substances normally packaged in amphorae.

There is also fairly limited evidence for the reuse of amphorae that had held fish products as packaging containers. As with the oil amphorae from this establishment, the Dressel 21s and 22s from the Officina del Garum degli Umbrici had presumably been set aside for refilling with some substance or other, while the African 2s from the Cabrera 3 shipwreck may have been reused for the packaging of fish products and olives.

86

The reuse of transport amphorae as packaging containers in the Roman world: an overview

The third and final question regards the extent to which the reuse of amphorae as packaging containers might have raised problems regarding the identification of the containers’ contents. As noted above, the papyrological evidence makes clear that in Roman Egypt people regularly referred to many different kinds of amphorae by specific names. Interesting in this regard is a passage in Pliny the Elder’s Naturalis historia (14.66) in which he states that, when packaged in lagonae — that is, small amphorae — Tauromenitanum, the wine from Tauromenium, was often passed off as Mamertinum, the wine from Messana, immediately to the north. This suggests that some individuals — perhaps consumers at Rome — tended to associate this form of container with Mamertinum, presumably a wine thought superior to Tauromenitanum, and that unscrupulous sellers — presumably retailers at Rome — took advantage of this by selling inferior wine to incautious buyers at an elevated price. On the basis of evidence like this, scholars have generally inferred that individuals across the empire were not only able to recognize a wide array of amphorae, but also associated some of these with a particular kind of content. The question, then, is whether the reuse of amphorae as packaging containers would have caused confusion on this score, creating problems at the moment of wholesale sale, retail sale, consignment to the state or church in the context of a supply initiative, and/or the levying of customs duties by the state.

packaging container would not have been fundamentally different from that of an amphora in prime use, save that it would have been necessary or desirable to remove a titulus pictus or other type of label referring to its primeuse content in order to avoid confusion. I am not aware of any instance in which there is evidence for the removal of a titulus pictus from an amphora with a new label executed in its place, although an operation of this kind need not have left traces that would be readily apparent. I know of two cases from Rome and four from Pompeii in which an amphora bore two tituli picti identifying different contents, thus instances in which the prime-use label was not removed when a container was refilled, although in none of these is the presence of two labels apt to have caused confusion as to the nature of the container’s reuse content. In cases in which amphorae were not provided with a label indicating their content, merchants involved in the wholesale distribution of these might have relied on accompanying documents that indicated what this was. One example of such a document is TPSulp. 80, a tabella cerata from Pompeii’s Murecine Archive, probably dating to the period c. AD 26-61. This document, which appears to be a shipping manifest, lists sets of amphorae being consigned by a certain Theophilus to a certain Aphrodisias that had been brought to a port in Italy — presumably Puteoli — aboard a ship known as the Octavia, indicating for each set the number and the nature of the containers and their content (Camodeca 1999: 184; Peña 2007b: 251– 252; Purpura 2014: 134; Terpstra 2013: 90–92). As another example of a document of this kind we can cite OBer.1.87, an ostracon from Berenike probably dating to the period c. AD 33-70. This document, which was a customs pass, lists by type a set of amphorae being transported by camel, indicating for each type its name, the number of containers, and their content.

The answer, it seems, is that the reuse of amphorae as packaging containers need not have raised problems of this kind. Whatever people’s ability to recognize different kinds of amphorae and their tendency to associate these with a particular substance, it was a regular practice to provide an amphora with a titulus pictus that identified its content. While it is unclear precisely how widespread this practice was, and in some cases these labels were written in a cryptic fashion that might have rendered them difficult to interpret for persons extraneous to the packaging and distribution process, recent research at Pompeii, where the preservation of tituli picti tends to be exceptionally good, has indicated that a large percentage of amphorae, and, in some instances, perhaps effectively all of those in a particular group were provided with a titulus pictus (Bernal‑Casasola et al. 2014: 224; Timby 2004: 386). This suggests that it was generally thought necessary to provide some indication of its content on the exterior of a sealed amphora.3 If such was the case, the problem of identifying the content of an amphora being reused as a

Frier (1983: 291–294), in tracing the evolution of juridical thought in relation to the problem of vinegar sold as wine, concluded that by the 2nd century AD the wine trade had taken on the forms of mature mercantilism, in which merchants understand that it is to their advantage to comport themselves in such a way as to be perceived as honest partners. It seems fair to assume that the trade in the other foodstuffs commonly packaged in amphorae came to be governed by a similar ethos. If so, during the middle- and perhaps also the late-imperial periods the cases of the deliberate misrepresentation of the content of sealed amphorae probably would have been uncommon at the wholesale level. Merchants regularly involved in the importing and exporting of foodstuffs would also have seen it as being in their interest to establish a reputation for honest dealing with customs officials. We may suspect that deception of this kind was more common in the context of retail sale, particularly in larger towns, and perhaps also in connection with the consignment of foodstuffs levied by the state as tax in kind.

 The content of a sealed amphora may have been attested by some means other than a titulus pictus that is never or only very rarely preserved, such as a lead or wooden tag attached to one of the handles or a stamped, inked or painted indication on the vessel’s stopper. See, for example, Thomas 2011: 23–24, where it is suggested that the layer of red paint regularly applied to the upper surface of many of the amphora stoppers from Quseir al-Qadim may have indicated that the vessel’s content was wine, and that the cobra symbol frequently stamped into this surface along with the name of a merchant may have indicated that the content was a product of the Fayum. 3

87

J. Theodore Peña

disks that were sealed and secured in place by being covered with a substance that appears to be either raw clay or raw clay mixed with plaster.

6. Further research There are various kinds of research that we might pursue in order to improve our understanding of the reuse of amphorae as packaging containers in the Roman world. First, a systematic and comprehensive review of the papyrological evidence would provide us with an enhanced understanding of how and to what degree amphorae were reused for the packaging and storage of foodstuffs and other substances in Roman Egypt. Second, the detailed characterization of amphora cargoes from additional shipwreck sites would perhaps furnish us with additional clear instances of this practice, thereby providing us with more information regarding the circumstances in which it occurred and what this involved. Work of this kind would also provide us with useful information about patterns in the homogeneity/ heterogeneity of amphora cargoes, positioning us to better interpret possible evidence for this practice, such as that from the Cabrera 3 shipwreck discussed above. Finally, programmes of research aimed at identifying the remains of amphora contents and documenting tituli picti will doubtless provide new information regarding practices of and patterns in the reuse of amphorae as packaging containers.

7. Conclusion In the current state of our knowledge it seems fair to say that the evidence for the reuse of amphorae as packaging containers in the Roman world is scattered, uneven, and less than substantial. Specifically, the evidence available to us at present from shipwreck sites does not give us reason to believe that this practice was commonplace for transfers that involved ships. However, more detailed analysis of amphora cargoes focusing on content residues, tituli picti, stoppers, use alterations, and the homogeneity/heterogeneity of groups of containers may change this picture. Where other kinds of evidence are available to us — papyrological evidence from Egypt, sets of amphorae from packaging facilities at Pompeii and its environs — this suggests that used amphorae — both locally manufactured and — in Egypt, at least — those imported from further afield — were employed — perhaps regularly and intensively — for the packaging of wine and perhaps also other substances, including fish products and oil. One possibility is that the incidence of this practice was greater for — perhaps even largely restricted to — substances destined for local or peri-local transfer, where the distance between the bottler and the person who opened and emptied the amphora of its content — perhaps a retailer, perhaps a consumer — was less in both geodesic terms and in terms of number of transactions. Indeed, for transfers of this kind there may have been substantially relaxed expectations regarding the nexus between a type of amphora and its content, with, for example, some understanding that local/regional amphorae and perhaps even certain kinds of imported amphorae also did duty as a sort of general, all-purpose packaging container.

In this connection I would like to note that in June, 2016 (following the initial submission of the manuscript for this contribution) the Pompeii Artifact Life History Project (PALHIP) — a research project that I direct at Pompeii (Peña 2014; Peña and Cheung 2015) — working in collaboration with the Oplontis Villa B initiative of the University of Texas Oplontis Project (Thomas 2015; Gazda and Clarke 2016: 62–65, 160–177; Pecci et al. 2017), carried out the detailed description of a set of 34 of the c. 1200 amphorae recovered at the Oplontis Villa B warehouse/wine bottling facility, which, as noted above, consist in large part of used containers presumably destined for reuse, in many cases probably as packaging containers (Peña 2018: 2, 13–15).4 The amphorae described included 33 examples of the Dressel 2-4 (28 of which were of the locally-produced Panella Type 3, 3 of which might have been examples of this type, and three of which were examples of some other type) and one an example of the Dressel 22/Botte 3. This work recorded interesting evidence for the methods employed at this facility for the stacking of empty Dressel 2-4s destined for reuse, for the emptying of some of these containers by the drilling of one or more tap holes in their wall — in some cases without the removal of the stopper — and for their reconditioning for reuse by the plugging of these tap holes by the insertion of custom-shaped sherd

If the more careful evaluation of sets of amphorae from packaging facilities and individual amphorae and sets of amphorae from consumption sites winds up validating a scenario more or less along these lines, and if the fuller and more careful analysis of amphora cargoes from shipwrecks suggests that the reuse of amphorae as packaging containers for export transfers was substantially more common than can be demonstrated at present, then we will have to rethink our ideas regarding the value of amphorae as evidence for the geography, mechanisms, and intensity of the production, distribution, and consumption of the foodstuffs and other substances commonly packaged in these containers.

 Since this work was undertaken J. Muslin has completed a comprehensive study of the amphorae from Oplontis Villa B; see Muslin 2019. 4

88

The reuse of transport amphorae as packaging containers in the Roman world: an overview

Carlson, D., Leidwanger, J. and Kampbell, S. (eds) 2015. Maritime Studies in the Wake of the Byzantine Shipwreck at Yassiada, Turkey: 17–34. College Station, Texas A&M University. Cibecchini, F. 2013. Sud Caveaux 1 - épave de navire. In Atlas du Patrimoine Archéologique Littoral Méditerranéen, viewed 15 December 2015. . Curtis, R. I. 1979. The garum shop of Pompeii (I.12.8). Cronache Pompeiane 5: 5–23. Fergola, L. 2004a. La villa B e il suo proprietario. In L. Fergola (ed.), Oplontis e le sue ville: 100–11. Pompeii, Flavius. Fergola, L. 2004b. I reperti della Villa B. In L. Fergola (ed.), Oplontis e le sue ville: 114–27. Pompeii, Flavius. Frier, B. W. 1983. Roman law and the wine trade: the problem of “vinegar sold as wine”. Zeitschrift der Savigny-Stiftung für Rechtsgeschichte 100: 258-95. Fulford, M. 1998. The House and Bar of Amarantus (I.9.1112). In J. Berry (ed.), Unpeeling Pompeii: Studies in Region I: 62–8. Milan, Electa. Garnier, N. 2007. Analyse du contenu d’amphores africaines. In E. Papi (ed.), Supplying Rome and the Empire. The Proceedings of an International Seminar Held at Siena-Certosa di Pontignano on May 2-4, 2004 on Rome, the Provinces, Production and Distribution. Journal of Roman Archaeology Supplementary Series 69: 25–9. Portsmouth RI, Journal of Roman Archaeology. Garnier, N., Silvino, T. and Bernal-Casasola, D. 2011. L’identification du contenu des amphores : huile, conserves de poissons et poissage. In SFECAG, Actes du Congrès d’Arles: 397–416. Marseille, Société française d’étude de la céramique antique en Gaule. Gazda, E. and Clarke, J. (eds) 2016. Leisure and Luxury in the Age of Nero: the Villas of Oplontis near Rome. Kelsey Museum publications 14. Ann Arbor, University of Michigan. Grace, V. and Empereur, J.-Y. 1981. Un groupe d’amphores ptolémaïques estampillées. Bulletin de l’Institut français d’archéologie orientale 81-1: 409–26. Hickey, T. 2012. Wine, Wealth, and the State in Late Antique Egypt: the House of Apion at Oxyrhynchus. Ann Arbor, University of Michigan. Jashemski, W. 1967. A Pompeian vinarius. Classical Journal 5, 2: 193–204. Jashemski, W. 1979. The Gardens of Pompeii, Herculaneum and the Villas Destroyed by Vesuvius. New Rochelle, Caratzas Brothers. Joncheray, J.-P. and Long, L. 2002. L’épave profonde Héliopolis 2 - Nord Levant (Var, -80 m). Une fouille d’épave à l’aide de plongeurs à saturation et d’un sous-marin d’observation. Cahiers d’archéologie subaquatique XIV: 131–59. Kruit, N. and Worp, K. 2000. Geographical jar names: Towards a multi-disciplinary approach. Archiv für Papyrusforschung und verwandte Gebiete 46, 1: 65–146. Lagi, A. 2015. Un horreum nel suburbia pompeiano. In C. Parisi Presicce and O. Rossini (eds), Nutrire l’impero. Storie di alimentazione da Roma e Pompei: 96–98. Rome, L’Erma di Bretschneider.

Bibliography Auriemma, R. 2000. Le anfore del relitto di Grado e il loro contenuto. Mélanges de l’École française de Rome. Antiquité 112-1: 27–51. Auriemma, R. and Pesavento Mattioli, S. 2009. I tituli picti delle anfore di Grado. In S. Pesavento Mattioli and M. B. Carre (eds), Olio e pesce in epoca romana. Produzione e commercio nelle regioni dell’alto Adriatico. Atti del Convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 275–80. Rome, Quasar. Bernal‑Casasola, D., Cottica, D., García Vargas, E., Toniolo, L., Rodríguez-Santana, C. G., Acqua, C., Marlasca, R., Sáez, A. M., Vargas, J. M., Scremin, F. and Landi, F. 2014. Un contexto excepcional en Pompeya: la pila de ánforas de la Bottega del Garum (I.12.8). Avance de un estudio interdisciplinar. Rei Cretariae Romanae Fautorum Acta 43: 219-32. Berry, J. 1997. Household artefacts: towards a reinterpretation of Roman domestic space. In R. Laurence and A. Wallace-Hadrill (eds), Domestic Space in the Roman World: Pompeii and Beyond. Journal of Roman Archaeology Supplementary Series 22: 183–95. Portsmouth RI, Journal of Roman Archaeology. Bocchio, S. 1990. Frammenti anforari. In Milano capitale dell’impero romano 286-402 d.C.: 146–8. Milan, Silvana. Bombico, S. 2015. Salted fish industry in Roman Lusitania: trade memories between Oceanus and Mare Nostrum. In F. Themudo Barata and J. Magalhães Rocha (eds), Heritages and Memories from the Sea. 1st International Conference at the UNESCO Chair in Intangible Heritage and Traditional Know-How: Linking Heritage 14-16 January 2015. Évora, Portugal Conference Proceedings: 19–39. Evora, University of Evora. Bonifay, M. 2004: Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports International Series 1301. Oxford, Archaeopress. Bonifay, M. avec une annexe de N. Garnier 2007. Que transportaient donc les amphores africaines  ?. In E. Papi (ed.) Supplying Rome and the Empire. The Proceedings of an International Seminar Held at SienaCertosa di Pontignano on May 2-4, 2004 on Rome, the Provinces, Production and Distribution. Journal of Roman Archaeology Supplementary Series 69: 9–31. Portsmouth RI, Journal of Roman Archaeology. Bonifay, M., Botte, E., Capelli, C., Contino, A., Djaoui, D., Panella, C. and Tchernia, A. 2015. Nouvelles hypothèses sur l’origine et le contenu des amphores africaines Ostia LIX et XXIII. Antiquités africaines 51: 189–210. Bost, J. P., Campo, M., Colls, D., Guerrero, V. and Mayet, F. 1992. L’épave Cabrera III (Majorque): échanges commerciaux et circuits monétaires au milieu du IIIe siècle après Jésus-Christ. Paris, Centre Pierre Paris. Bruni, S. (ed.) 2000. Le navi antiche di Pisa: ad un anno dall’inizio delle ricerche. Florence, Polistampa. Camodeca, G. 1999. Tabulae Pompeianae Sulpiciorum (TPSulp.). Edizione critica dell’archivio puteolano dei Sulpici. Vetera 12. Rome, Quasar.

89

J. Theodore Peña

Maiuri, A. 1933. La Casa del Menandro e il suo tesoro di argenteria. Rome, Libreria dello Stato. Mayerson, P. 1992. The Gaza ‘wine’ jar (Gaziton) and the ‘lost’ Ashkelon jar (Ashkalônion). Israel Exploration Journal 42: 76–80. Muslin, J. 2019. Between Farm and Table: Oplontis B and the Dynamics of Amphora Packaging, Design, and Reuse on the Bay of Naples. Unpublished PhD thesis, University of Texas at Austin. (https://repositories.lib.utexas.edu/ handle/2152/76091). Nieto Prieto, J., Jover Armengol, A., Izquierdo Tugas, P., Puig Griessenberger, A. M., Alaminos Expósito, A., Martín Menéndez, A., Pujol Hamelink, M., Palou Miquel, H. and Colomer Martí, S. 1989. Excavacions arqueològiques subaquàtiques a Cala Culip 1. Centre d’Investigacions Arqueològiques de Girona, Sèrie Monogràfica 9. Girona, Centre d’Investigacions Arqueològiques de Girona. Panella, C. 1974-1975. Per uno studio delle anfore di Pompei: le forme VIII e X della tipologia di R. Schoene. In In memoria di Giovanni Becatti. Studi Miscellanei 22: 151–162. Parker, A. J. 1992. Ancient Shipwrecks of the Mediterranean and the Roman Provinces. British Archaeological Reports International Series 580. Oxford, Tempus Reparatum. Pecci, A., Clarke, J., Thomas, M., Muslin, J., van der Graff, I., Toniolo, L., Miriello, D., Crisci, G. M., Buonincontri, M., and Di Pasquale, G. 2017. Use and reuse of amphorae. Wine residues in Dressel 2-4 amphorae from Oplontis Villa B (Torre Annunziata, Italy). Journal of Archaeological Science Reports 12: 515–21. Peña, J. T. 1999. The Urban Economy during the Early Dominate: Pottery Evidence from the Palatine Hill. British Archaeological Reports International Series 784, Oxford, Archaeopress. Peña, J. T. 2007a. Roman Pottery in the Archaeological Record. Cambridge, Cambridge University Press. Peña, J. T. 2007b. Two groups of tituli picti from Pompeii and environs: Sicilian wine, not flour and hand-picked olives. Journal of Roman Archaeology 20: 233–54. Peña, J. T. 2014.The Pompeii Artifact Life History Project: conceptual background and first season’s results. Rei Cretariae Romanae Fautorum Acta 43: 467–74. Peña, J. T. 2016. Papyrological evidence for the reuse of amphoras as packaging containers in Hellenistic and Byzantine Egypt. Lecture presented at symposium Archaeology and the Hellenistic Near East: A Symposium in Honor of Sharon Herbert. University of Michigan, Ann Arbor, Michigan, May 7, 2016. (http://resromanae. berkeley.edu/node/1105). Peña, J. T. 2018. Pompeii Artifact Life History Project (PALHIP): General report on project operations and results, 2012-2016. Unpublished report. (http://resromanae.berkeley. edu/sites/default/files/POMPEII%20ARTIFACT%20 LIFE%20HISTORY%20PROJECT%202012-2016.pdf). Peña, J. T. and Cheung, C. 2015. The Pompeii Artifact Life History Project: conceptual basis and results of first three seasons. In C. Gambardella (ed.), Heritage and Technology. Mind, Knowledge Experience. Le Vie degli Mercanti XIII Forum Internazionale di Studi. Fabbrica della conoscenza 56: 2115–23. Aversa, Dipartimento di Architettura e Disegno Industriale Luigi Vanvitelli.

Peña, J. T. and McCallum, M. 2009. The production and distribution of pottery at Pompeii: a review of the evidence. Part 2, the material basis for production and distribution. American Journal of Archaeology 113: 165–201. Pesavento Mattioli, S., Mazzochin, S. and Pavoni, M. 2000. Anfore della Nave B. In S. Bruni (ed.), Le navi antiche di Pisa: ad un anno dall’inizio delle ricerche: 131–47. Florence, Polistampa. Purpura, G. 2014. Il χειρέμβολον e il caso di Saufeio: responsibilità e documentazione nel trasporto marittimo romano. Annali del seminario giuridico dell’Università degli Studi di Palermo 57: 129–52. Rathbone, D. 1983. Italian wines in Roman Egypt. Opus 2: 81–98. Rathbone, D. 1991. Economic Rationalism and Rural Society in Third-Century AD Egypt. Cambridge, Cambridge University. Slane, K. 2004. Amphorae — used and reused — at Corinth. 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: 361–69. Aarhus, Aarhus University Press. Stefani, G. 2003. Produzione fittile: attività, consumi, commerci. In G. Stefani (ed.), Menander: La Casa del Menandro di Pompei: 210–33. Milan, Electa. Terpstra, T. 2013. Trading Communities in the Roman World: a Micro-Economic and Institutional Perspective. Columbia Studies in the Classical Tradition 37, Leiden-Boston, Brill. Thomas, R. 2011. Roman vessel stoppers. In D. Peacock and L. Blue (eds), Myos Hormos -Quseir al-Qadim. Roman and Islamic Ports on the Red Sea. 2. Finds from the Excavations 1999-2003. British Archaeological Reports International Series 2286. University of Southampton Series in Archaeology 6: 11–34. Oxford, Archaeopress. Thomas, M. 2015. Oplontis B: a center for the distribution and export of Vesuvian wine. Journal of Roman Archaeology 28: 403–11. Timby, J. 2004. Amphorae from excavations at Pompeii by the University of Reading. 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: 383– 92. Aarhus, Aarhus University Press. Toniolo, A. 2002-2003. Uso e commercio di oggetti da riutilizzare nell’antichità. Un caso nordadriatico. Archeologia veneta 25-26: 109–48. Van Alfen, P. 1996. New light on the 7th-c. Yassi Ada shipwreck: capacities and standard sizes of LRA1 amphorae. Journal of Roman Archaeology 9: 189–213. Van Alfen, P. 2015. The restudy of the LR2 amphorae from the seventh-century Yassiada shipwreck: preliminary evidence for standardization. In D. Carlson, J. Leidwanger and S. Kampbell (eds), Maritime Studies in the Wake of the Byzantine Shipwreck at Yassiada, Turkey: 17–34. College Station, Texas A&M University.

90

The reuse of transport amphorae as packaging containers in the Roman world: an overview

Van Doorninck Jr., F. H. 2015. The seventh-century Byzantine ship at Yassıada and her final voyage: present thoughts. In D. Carlson, J. Leidwanger and S. Kampbell (eds), Maritime Studies in the Wake of the Byzantine Shipwreck at Yassiada, Turkey: 205–16. College Station, Texas A&M University. Vera, D. 2006. Un’iscrizione sulle distribuzioni pubbliche di vino a Roma (CIL VI, 1785 = 31931). In M. Silvestrini, T. Spagnuolo Vigorita and G. Volpe (eds), Studi in onore di Francesco Grelle: 303–17. Bari, Edipuglia.

Ward, C. 2015. Plant remains from the old wine jars on the Byzantine ship at Yassiada. In D. Carlson, J. Leidwanger and S. Kampbell (eds), Maritime Studies in the Wake of the Byzantine Shipwreck at Yassiada, Turkey: 55–62. College Station, Texas A&M University.

91

Archaeometric indicators: generalities and case studies

Amphorae and residue analysis: theorical considerations Nicolas Garnier

SAS Laboratoire Nicolas Garnier – LNG, 32 rue de la Porte Robin, F ‑ 63270 Vic‑le‑Comte associated researcher AOROC ‑ ENS Paris‑Ulm CNRS UMR 8546

Abstract: The very nature of organic materials, i.e. more or less degraded and complex mixtures, and the minute amounts available, have made the identification of organic residues preserved in archaeological context a long‑lasting challenge. Several approaches have been developed over the years so that spot tests and infrared spectrometry have now been abandoned for more secure and informative methods based on chromatography and mass spectrometry. Biomarkers identified through a molecular approach trace back to various sources: the original material, but also its natural or anthropic degradation products, and finally, side contaminations. Today, a rigorous methodology built on artefact selection and sampling choices, extraction and analytical protocols, marker identifications, and the interpretation of the molecular data obtained, enables the unambiguous characterisation of numerous organic archaeological materials. The molecular approach, through the direct characterisation of these materials, appears to be the prevalent choice for the secure identification of amphorae contents, which in turn leads to a better understanding of the goods either transported or stored therein during Antiquity. Key words: adsorbed organic residues; ceramics; amphorae; biomarkers; analytical methods.

1. Introduction In order to gain insight into commerce and trade in Antiquity one needs to address amphorae typology, production sites, commercial routes, and foremost, amphorae contents. But do we identify these contents, mostly organic by nature and thus sensitive to time and prone to natural and microbial degradation, such as a 2 millennia burial leaving only some invisible remains? Past archaeological studies have long relied on amphorae typology, epigraphy and scarce visible residues. If early results in the years 1976‑78 did fall in accordance with archaeological observations (Condamin et al. 1976), organic analysis thereafter rapidly showed discrepancies between content attribution based on typology and data obtained from molecular remains trapped within the porous vessel walls; in the process, cases of wrong assumptions simply based on some visible remains (pitch for example) were revealed. As of 1995, carried by a favorable wind, numerous research teams turned their interests towards archaeological materials, coming up with methodologies which can be regarded as more or less serious and/or well‑grounded. However, technological progress over the past 20 years has reached a point where the chemical characterisation of main recurrent contents (pitch, oil, wine, fish sauce) has become reliable as long as suitable techniques and methodologies are applied. The array of potentially identifiable materials progressively widens, revealing that the links between container typology and content are far more complex than those established in previous models; in addition, amphorae not only support long-distance trade, they also play a crucial role in domestic and funerary contexts and are sometimes reused over several decades. The flow of recently published papers calls for a critical eye regarding technique reliability, adequacy between

archaeological questioning and applied methodologies, quality of reasoning and interpretation, or even veracity of some shady approaches. In this review, the archaeologist, historian, curator and restorer will find the basic facts required for a better understanding of the nature of molecular residues and of the chemical information contained therein, and foremost, he will develop an increased capacity to evaluate the relevance of the methodologies used today and the data thus obtained. Indeed, interpretation, often tainted by a lack of rigor, is a crucial step towards the accurate and secure identification of ancient materials. Unsuitable methods or the overinterpretation of the analytical data can lead to hasty and unfounded results. Awareness of methodological limitations is essential for attaining reliable interpretations and conclusions. Potentials for analysing and identifying ancient organic products will be illustrated by several recent cases. 2. Organic materials in archaeological contexts 2.1. Organic substances and chemical classes The organic world is millions of rich molecules, encompassing a tremendous variety in size and complexity. By definition, all organic molecules contain the basic elements carbon (C) and usually hydrogen (H), but also variable ratios of oxygen (O), nitrogen (N), sulfur (S) and phosphorus (P), these 6 elemental ingredients for life being referred to as the CHNOPS group. Organic molecules each have a given number and a given arrangement of functional groups attached to their carbon backbone, conferring unique chemical properties: hydroxyle on alcohols, carboxyle on acids, carbonyle on ketones, amide on peptides, etc… These molecules are divided into chemical classes including main ones such

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 95–111

Nicolas Garnier

Figure 1. Main chemical families of biomarkers preserved in archaeological samples and their correspondant biological materials. as carbohydrates, lipids, proteins, alkaloids, phenolics (including tannins —  polyphenolic compounds) and nucleic acids (DNA, RNA — the field of paleogenetics) (Figure 1). Lipids, for example, are themselves subdivided into polar lipids (phospholipids, ceramides, glycolipids, sterols, fatty acids) and apolar ones (triglycerides, cerides, sterol esters…), or into saponifiable lipids (fatty acidderived) and unsaponifiables (terpenoids — including sterols and their derivatives).

reason why, in the end, descriptions are generally limited to partial molecular compositions, i.e. a selection of the most relevant markers. As an alternative approach, spectroscopy provides a way of classifying residues based on their spectral fingerprint (global spectra of the material) even if detailed identification of the various compounds is not attainable.

Adding to the level of complexity, organic materials are usually composed of a mix of molecules belonging not only to one but several chemical classes. Describing the chemical composition of these materials (i.e. the molecular structure of their compounds) is quite arduous and falls within dedicated fields of research such as phytochemistry and pharmacognosy. Considering that even such simple substances as oils consist of several coexisting subclasses of chemical compounds, identifying the whole set (defined as biomarkers) within organic materials is a real challenge for chemists. Figure 1 provides a schematic overview of major and minor markers present in the chemical composition of main bioarchaeological materials identified so far. These components each present a large range of molecular weights, chemical structures and concentrations, and thus call on different analytical techniques, even separate extraction and purification protocols; the

The identification of organic materials, i.e. consisting of a complex mix, cannot be based soly on a single compound but requires the association of various molecular markers. While the ‘biomarker’ principle was defined in the early 1990s for the biological, environmental and earth sciences (Evershed 1993), material composition data originate from areas of biochemistry, organic and analytical chemistry. Ancient materials differ in composition from their modern counterparts; factors specific to ancient materials, i.e. aspects related to their evolution, ecology, diagenesis and chemometrics, differentially affect the preservation of the various chemical classes (Poynter and Eglinton 1991). To be considered relevant and selectable, a biomarker should only be present in limited amounts in the biosphere and be quite resistant to degradation and post-depositional processes. In addition, a biomarker may relate to a plant or animal family or genus, or more specifically to a particular species. Defining biomarker

2.2. Biomarker: a key player

96

Amphorae and residue analysis: theorical considerations

associations helps both target the identification process and trace back to the original materials, degradation state permitting.

leaching can therefore lead to a total loss of organic matter in some types of environments such as sites constantly traversed by water (deposits in rivers,1 near sources2…), or to a selective loss of some classes of compounds depending on their polarity, or, on the contrary, to a perfect fossil-like preservation of the organic material such as in quickly silted sites (palafittes – pile dwellings,3 silted ports,4 wells5). Any heating linked or not to human activity, natural or accidental fires leads to the oxidation of the markers. Hence, biomarkers of heated resins, oils, and fats are thermal degradation markers. Very sensitive markers are completely destroyed in the process, such as sugars which are rapidly carbonized into pure carbon residues following an intermediate caramel state. Time also affects the chemical degradation of biomarkers. The markers sensitive to thermal oxidation degrade in the simple presence of oxygen according to slow but continuously operating kinetics. The markers of natural degradation thus formed are oxidized, i.e. they include additional alcohol or acid functional groups that increase the polarity of molecules and hence their solubility in water and consequential loss.

2.3. Preservation and degradation Unlike inorganic materials, organic materials are unstable and rapidly degrade. Reaction rates depend on various factors, both intrinsic (chemical class, nature and number of functional groups, polarity — i.e. hydrophobic vs. hydrophilic molecules, type of reaction and degradation mechanism) and extrinsic (temperature, pH, humidity, redox reaction, microorganisms, metal catalysis…). Therefore, as environmental conditions rapidly and continuously change, it is not possible to predict the degradation pattern of a given molecule over hundreds of years of burial. Aiming to capture the state of decay of an archaeological organic material is thus illusory, even more so, attempts to reconstruct artificial alterations in the lab; these truly correspond to a simplified degradation pattern that will indeed give rise to an altered material but according only to very specific conditions, not necessarily reflecting reality.

Finally, biomarkers which undergo soil microbial activity in the initial stages of diagenesis transform into natural degradation markers. In soil, surrounding organic materials serve as nutrients to bacteria, yeasts, and fungi; this activity not only directly contributes to the loss information but also pollutes the collected samples. The molecular composition of natural products found in archaeological contexts is thus very different from the composition of modern counterparts. The detected markers do not necessarily come from the original content; processes related to the formation of residues (carbonisation, adsorption on porous media, mixing, drying…) or to their degradation (heat, oxidation, dissolution of polar compounds, microbial degradation, contamination) change the chemical composition of the original material. Much information is lost during both the initial preparation and the later burial of the material. Hence, it is highly recommended to entrust the archaeological samples to laboratories specialised in the study of ancient materials which not only follow sampling and sample preparation protocols suitable for these types of artefacts, but also have the know-how to interpret the data according to different contexts and degradation states of the material.

Organic materials may undergo several types of degradation: physical attack related to loss of material by dissolution and diffusion of molecules in the environment, chemical alterations associated with the reactivity of the markers, and biochemical breakdown linked to the activity of soil microorganisms. Briefly, polar molecules (containing acid, alcohol, amine…, functional groups) are partially soluble in water which causes rapid loss by dissolution in runoffs and leaching; sugars rapidly disappear this way. By contrast, apolar molecules (hydrophobic) such as polyaromatic hydrocarbons (tar and bituminous materials), as well as cerides and alkanes of vegetable waxes, are not dissolved and are among the best-preserved compounds. Notably, burial sites are not closed environments and numerous exchanges take place between the deposited material and the surrounding sediment. In terrestrial environments, runoffs favour the migration of molecules from top to bottom; buried ceramic sherds accumulate markers mainly down from the upper layers, including compounds from the fallen vegetation, degraded and broken down into its constituent markers. Among the chemical degradation reactions, hydrolysis and oxidation are the most significant. Through chemical or biochemical processes, saponifiable lipids, such as triglycerides, cerides and esters, are thus hydrolysed in aqueous media into their constituent alcohols and acids; triglycerides are progressively hydrolysed into diglycerides, monoglycerides and finally into fatty acids. If triglycerides, biomarkers of oils and fats, are not initially soluble in water and may well be preserved in arid environments, fatty acids in contrast are soluble (Garnier 2007). The cumulative action of hydrolysis and

 Arles 3 shipwreck and its load, submitted to a constant leaching by the Rhone river (Garnier, Silvino and Bernal‑Casasola 2011). 2  Residues attached to barrel staves preserved in an environment traversed by a reach (Martinez 2014). 3  Archaeological settlements from the Jura (Regert et al. 2008). 4  Port of Pisa San Rossore (Garnier 2003), or strata protected by the silt deposit at Arles (Garnier, Silvino and Bernal‑Casasola 2011). 5  Study of the contents of pottery in a well at Vieille-Toulouse (Vial et al. 2015). 1

97

Nicolas Garnier

restoration of the object. The fluid dialogue among the team archaeologist, restorer and chemist, ensures not only the quality of the study of the designated content but also the preservation of the object and the invisible chemical information it entails.

3. What can be analysed? 3.1. Residues and impregnation If for a long time analyses were performed only on visible macro-remains, it is now possible to analyse invisible molecular traces trapped in mineral substrates. Any porous container, made of ceramic, stone, plaster (yet considered waterproof), is impregnated with the fluid materials it is in direct contact with. The more the content is fluid and the container porous, the longer the contact, the heavier the impregnation will be. Waterproofing with organic (pitch, resin, beeswax, bitumen) or mineral (sulfur) materials, or again using physical processes (glaze), does not prevent the migration of the content’s molecules into the walls, but only slows it down (Romanus et al. 2008). Further, alternating contents may successively impregnate the walls they’re in contact with. The pores of the ceramic gradually take up molecular traces by simple molecular diffusion, an irreversible thermodynamic process that creates particle gradients within the porous substrates. Since no stratigraphic build up takes place, it is impossible to distinguish the sequence of impregnation, all molecular markers gradually mixing by diffusion. Any porous support may be subjected to organic analysis: wine and oil production basins, dolia storage vessels, transport amphorae, serving and consumption dishes. Analyses can also be extended to glass, metal, or stone.

3.3. Sampling for content analysis Distinctive sampling techniques are carried out according to the material of the object (porous or non‑porous), the type of residue (caramel, deposit, invisible impregnation), the shape and the state of preservation of the object. Whenever necessary the analyst must adapt to the object in order to best preserve it. The individualised samples (aggregates, caramels, thick deposits…) are detached with a clean scalpel blade and stored in a glass tube or in a small sheet of aluminum foil. In order to avoid direct contact of the ceramic material with commonly used plastic bags, sherds may be wrapped in aluminum while awaiting sampling. The aluminum foil may be used directly without cleaning; the fine grey aluminum oxide surface covering film, which forms spontaneously on contact with air, does not interfere with any future organic analysis. It is best to let the samples dry before packing to avoid the growth of mould and other living microorganisms. The choice of sample materials is also guided by the physical condition of the objects: for porous objects, basin plasters in agricultural facilities, etc., any part that would have been in contact with the sought-after content may be removed and analysed. Thus, for amphorae, one should remove the bottom or the belly, never the rim. Whole vessels embody the ideal case as they ensure the best preservation and protection of the original content. However, depending on the diameter of the aperture and the shape of the vessel, the preservation requirement of the object may render the sampling difficult or even impossible.

3.2. Chemical analysis: an interdisciplinary approach The successful organic residue analysis of ancient materials relies on the clear definition of the archaeological issues and requires that the analytical methodology be adapted both to the object and the questioning; it also highly depends on the strict interpretation of the chemical data and its archaeological re-contextualisation (Evershed 2008; Garnier 2003; Pecci 2005). The content analysis of an object, a request formulated by the archaeologist, the restorer or conservator, always follows a specific archaeological questioning. It is important the object be well documented. In fact, information relating to the burial and discovery contexts may occasionally prompt the rejection of objects coming from too heavily polluted environments. The sediment of an environment polluted by plastic, by rich humus, or by recent industrial activities, is loaded with plastic markers, plant decomposition markers, etc. Acute environmental contaminations mostly prevent the detection of trace amounts of any remaining organic material, and thus identifying the original content of the vessel under study.

4. How to analyse ancient organic materials? 4.1. Inadequate methods: elemental analysis and spot tests For the past forty years, various methodologies for sample preparation and analysis have been developed with varying degrees of rigor and success. Due to their speed, easy implementation, the non-or minimally invasive factor and the low cost, some have attracted archaeologists, or even careless analysts, at the expense of scientific rigor. Organic residues, consisting of organic molecules built from carbon and mainly hydrogen and oxygen atoms, cannot be identified or even detected by mineral so‑called elemental analysis methods such as X‑ray fluorescence spectrometry (XRF), albeit coupled with electron microscopy (SEM-EDX). In the latter case, the analysis is performed in vacuo and does lead to the detection of light elements, i.e. carbon and oxygen, but as these

These vessels may come from recent or old digs. The sooner the sampling, the more the introduction of postexcavation pollution will be limited. Museum preserved objects, however, may still be analysed even after years. In any case, it is recommended that sampling for organic content analysis be done previous to any planned

98

Amphorae and residue analysis: theorical considerations

elements may come from both mineral (e.g. carbonate) and organic sources, elemental analysis provides no relevant information about the presence or absence of organic matter, even less about its nature.

by Eegriwe in 1933, developed and modified by Feigl in 1966, was still in use until recently to look for traces of tartaric acid (a grape biomarker) in residues from wine storage vessels (Badler et al. 1990). Feigl himself warned against false positives resulting from numerous sugars such as those found in the environment and sediments (Feigl 1966: 470). Hence, this test is not reliable when analysing organic materials in buried objects, and published results (McGovern et al. 1996; Michel et al. 1993) should be taken with extreme caution.

A recent study describes a range of inappropriate approaches to residue analysis applied to the Dr. 1B amphora from the Albenga wreck, supposedly containing wine (Arobba et al. 2014). First, the elemental analysis conducted on the deposits preserved in the amphora show low carbon and nitrogen levels (atoms revealing the presence of organic matter, lipids and proteins), which seems consistent with the high proportions of inorganic compounds found in the archaeological samples and, therefore, the low proportion of organic matter retained; the authors suggest that the organic material could have undergone extensive mineralization. Notably, high observed levels of sodium and chloride are indeed related to exchanges between the content of the sealed amphora and the external marine environment. However, the authors then compare measured δ 18O, δ 2 H and δ 13C isotopic ratios to those found in ethanol in present-day wine; yet, given the high diffusion coefficient of ethanol and the porosity of ceramics, wine ethanol initially contained in the amphora would have rapidly diffused through the walls of the ceramic to then be quickly diluted in the sea water following the sinking. Finally, if the authors argue that the residues found in the amphora contain ‘no measurable trace’ of organic matter, it cannot be linked to the fact that wine markers (tartaric acid, malvidin…) have rarely been detected in ancient shipwrecks, but rather that the methodologies, extraction protocols and analytical techniques previously used, were utterly unsuitable for the characterization of the soughtafter materials.

4.2. Observation methods Archaeological objects and samples are often examined by optical or scanning electron (SEM) microscopy. This initial approach is particularly useful for discerning preserved macro-structures (textile fibers, plant fragments…), but the method rapidly reaches its limits when it comes to characterising amorphous materials (such as waxes, resins, deposits or caramels) other than by their physical appearance (colour, inclusion…). However, some authors are misusing such observations, claiming to identify micro-residues of wax, resin, animal tissues, on South African stone tools (work by Lombard, Wadley et Langejans: Lombard and Wadley 2007; Wadley and Lombard 2007; Wadley, Lombard and Williamson 2004) sterilised and then used for a variety of tasks involving the processing of plants and animal products. Precautions were taken to avoid contaminating the residues. One set of used flakes was stored in sealed plastic bags; the other set was buried in compost for a month and then exposed to open-air conditions for three days. The bagged tools were used for a blind test (Test One, with an interesting and well-argumented criteria: Crowther and Haslam 2007). Others have employed chemical reagents on test strips normally dedicated to medical use, i.e. the detection of hemoglobin in urine (Loy 1983; Loy and Wood 1989), or again, made use of the visual recognition of crystals a priori due to precipitated proteins from a solution in which the tools were soaked (Loy and Hardy 1992; Loy  1983; Loy and Wood 1989), de facto ignoring numerous possible false positives as well as the inability to obtain the same crystallized structures from modern homologous proteins. Such studies applied to amorphous organic materials call for a severe but founded critical review, which confirms their inefficiency (Crowther and Haslam 2007). Additionally, implementing large and expensive techniques such as PIXE, an elemental analysis technique per se inadequate to identify organic materials (Langejans 2007), sometimes gives papers a scientific veneer without providing any proof whatsoever.

Ever since the first tests and physical characterisation of archaeological residues conducted by Chevreul and Berthelot, chemists have sought to identify the chemical elements present in ceramics and clay using dissolutions, tests and solution assays, but without really attempting to isolate organic compounds that may have been contained in the vessels. So-called ‘spot test’ analysis involves reacting a sample with a chemical reagent using either dry (deposits on foil) or wet (in test tubes) methods; chemical functions (alcohol, ketone, sugar…) are revealed by their reactivity. These tests are easy to carry out, quick, extremely cheap, but they do not in any way allow for the identification of molecules, only their chemical functions. Therefore, great care in the application and interpretation of the results must be taken. First developed for pharmacological research and later abandoned in the 1960s, they are still widely used for the study of ancient materials. They can lead to interesting but totally unfounded overinterpretations if incorrectly used: biomarkers cannot be identified soly by their chemical functions, only by their structure. The identification of plant or animal species is even more dubious. The specific so-called ‘Feigl test’, invented

4.3. Spectral methods: Infrared and Raman In the 1970s, infrared spectrometry (IR or FTIR) was applied for the first time to amorphous residues preserved in sealed wine amphorae originating from the Madrague de Giens shipwreck (Condamin et al. 1976).

99

Nicolas Garnier

This technique only allows the identification of chemical functions. F. Formenti was able to associate the observed FTIR absorption bands with acid and alcohol functions present in particular in tartaric acid; she cautiously concludes that it ‘might’ be wine, because the analysis is consistent with this hypothesis (Figure 2). However, since the technique does not enable the full identification of biomarkers, it does not allow to indubitably deduce the nature of the original material, especially mixed residues degraded over 2000 years of burial. Infrared spectrometry, a simple and inexpensive technique to implement, always yields some kind of results. Sample preparation is very fast. However, if the sample is not an aggregate but was scratched off the walls of a ceramic vessel, the obtained spectrum corresponds to the entire collected material, mineral and organic. In this case only silica and clay are detected as shown by the analysis carried out on scraped residues on the walls of vinery jars from Hajji Firuz Tepe (McGovern et al. 1996). It is therefore essential to first extract the organic material in order to exclude the mineral support. The results obtained by FTIR are often overinterpreted. Even though the detection of chemical functions does not allow pinpointing a marker with certainty, some authors go so far as to name the organic materials. For example, in the Malkata Palace pottery jars, based soly on FTIR results, McGovern identifies mixtures of wine and resins, i.e. myrrh in two of the analysed jars and ‘Chios turpentine’ or pistachio resin (Pistacia atlantica Desf. terebinth tree) in the other eight (McGovern et al. 1996). But because of the large number of species listed in the myrrh genus and others associated to it, i.e. bdellium and

balm (over 180 Commiphora sp.), the limited knowledge of their chemistry, the wrong botanical assignment of the Chios resin (Pistacia lentiscus var. chia) on one hand and the difficulty to differentiate the various Pistacia sp. exudates (Assimopoulou and Papageorgiou 2005b; Assimopoulou and Papageorgiou 2005a) on the other hand, it is hard not to be skeptical about the proposed interpretation. FTIR is nowadays considered a pre-analysis method used to assess the presence/absence of mineral and organic matter in a sample, or of specific chemical classes, but by no means does this technique allow us to unequivocally identify organic materials, even more so ancient ones — degraded, mixed, contaminated and preserved only as traces (Boulton and Heron 2000). In order to ensure the secure identification of preserved markers, the structural approach using chromatographic techniques and mass spectrometry should be favoured. Raman spectroscopy, more recently developed, is mainly used for identifying minerals and inorganic pigments (Madariaga 2010). As Raman analysis is performed under experimental conditions which cause organic compounds to fluoresce it rarely gives exploitable spectra. However, when irradiating the sample not in the visible but in the ultraviolet region (FT-Raman variant), organic matter does yield viable spectra, which is the case for resins (Edwards and Falk 1997; Edwards et al. 1996). Given the minute amounts of organic matter preserved in archaeological objects, with the exception of pitch deposits at the bottom of amphorae for example, Raman is rarely applied and a structural organic approach, yielding more accurate results, especially with mixtures of products, is generally preferred. 4.4. Direct methods of global structural analysis

Figure 2. IRTF spectra of the residue preserved in an amphora from the shipwreck la Madrague de Giens, with a residue of actual red wine Medoc (Garnier 2003).

Mass spectrometry analysis, by means of direct injection of the raw sample or of the extract, has been used to obtain more accurate information and is more easily correlatable with the analysed material. The analysis is fast and can be done on the raw sample (

Cathy Vieillescazes

IMBE UMR7263/IRD237, Avignon University/CNRS/IRD/AMU, Restoration Engineering of Natural and Cultural Heritage, F-84000, Avignon, France < [email protected]>

Abstract: The study of archaeological samples is often a challenge for a chemist because of the wide range of compound classes present and of the frequent occurrence of degradation products. This research work is focused on the characterization of resinous material present in the interior surfaces of Roman amphorae from Tuscany (Italy). In 1952, Commander Cousteau and his team carried out research on the Grand Congloué 2 in the bay of Marseille, under the scientific direction of Professor Fernand Benoit, Director of Antiquities of Provence. The wreck of the Grand Congloué 2, dating from the late second century or the beginning of the 1st century BC, contained more than a thousand Roman amphorae type Dressel 1A. Recent studies have definitely shown that these amphorae were from Cosa in Etruria, current Tuscany. The nature of organic residues was analyzed by Fourier-transformed infrared spectroscopy (FT-IR) and gas chromatography-mass spectrometry (GC-MS). The spectroscopic technique allows the identification of the class of substances analyzed based on the transitional vibrations of the functional groups and GC-MS permits a more precise identification of organic compounds present in the samples. The chromatographic analysis carried out on the archaeological samples reveals oxidized and aromatized abietanes. Dehydroabietic acid and retene prove the presence of pine pitch produced by resinous wood of plants from Pinaceae family. The occurrence of methyl diterpenic acids like methyl-dehydroabietate indicates a treatment by destructive distillation (pyrolysis) of softwood. Key words: Roman amphorae; pitch; Infrared spectroscopy (FT-IR); Gas chromatography coupled to mass spectrometry (GC-MS).

1. Introduction In Roman period, trade was thriving in Mediterranean area and amphorae were used to transport and store many different substances. The contents of amphorae were wine, oil, fish, fish sauce (garum), medicines, perfumes, etc. (Colombini, Modugno and Ribechini 2005; Izzo et al. 2013). The ceramic vessels are made from a porous material, so waterproofing materials were used on them. These materials are called pix which is known as organic resinous matter (Font et al. 2007). In 1952, Commander Cousteau and his team carried out research on the Grand Congloué 2 in the bay of Marseille in France, under the scientific direction of Professor Fernand Benoit, Director of Antiquities of Provence. It was the first submarine archaeological project in the world. The wreck ship of Grand Congloué 2 contained more than a thousand of Roman amphorae type Dressel 1A from Cosa in Etruria, now Tuscany in Italy. This ship is dated from the late 2nd century or the beginning of the 1st century BC. The Roman amphorae of Grand Congloué 2 have experienced

many disasters, the main event being a fire in the Château Borely Museum (France), when almost all of the amphorae were broken and then the rest were conserved in Fort St. Jean (Marseille, France) until 2011. Eleven samples were taken at Fort Saint-Jean in 2010. Then in 2015, seven samples were collected at the archaeological repository of the Ministry of Culture (Aix-les-Milles, France), a better place for the conservation of amphorae. All of the eighteen amphorae contained residues. The waterproof treatment which was usually made from natural resins was present inside the amphorae of Grand Congloué 2. The organic coating could contain a wide range of compound classes and their corresponding degradation products. Pine pitch is one of the organic coating that is used in Roman amphorae and many substances are determined during the degradation process (Colombini, Modugno and Ribechini 2005; Duce et al. 2015; Robinson et al. 1987). Retene is a biomarker of pitch, an organic substance obtained by the combustion of resin exudates from pine trees in particular, belonging

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 127–132

Hitomi Fujii, Carole Mathe, Fabienne Olmer and Cathy Vieillescazes

to the Pinaceae family; it is the stable end-product of degradation (Colombini et al. 2003, 2005; Font et al. 2007; Izzo et al. 2013).

Maison Méditerranéenne des Sciences de l’Homme, (Aixen-Provence) in France. All objects have a brown and black visible interior coating.

In the chemical analysis of organic resinous materials, gas chromatography-mass spectrometry (GC-MS) is frequently used to identify organic materials at archaeological sites (Colombini, Modugno and Ribechini 2005; Colombini et al. 2005; Robinson et al. 1987). The other spectroscopic technique is Fouriertransformed infrared spectroscopy (FT-IR) that allows the identification of the class of substance analyzed based on the transitional vibrations of functional groups. This helps us to understand the organic residue and the extent of degradation of archaeological material for further conservation and restoration (Colombini, Modugno and Ribechini 2005; Robinson et al. 1987). The purpose of the present study is the characterization of resinous material present on the interior surfaces of eighteen Roman amphorae from the ship wreck of the Grand Congloué 2 by using FT-IR and gas chromatography coupled to mass spectrometry (GC-MS) techniques. A preliminary study of the organic residues by FT-IR provides a fingerprint and general information about the nature of organic material employed for waterproofing the amphorae and is useful for providing initial hypothesis about the substances present (Font et al. 2007). GC-MS is a very good analytical tool for the analysis of complex mixtures, so gas chromatographic analyses were performed in order to characterize organic compounds present in the waterproof matter.

2.2. Instruments and methods

The identification of the GC peaks was made according to retention time and mass spectrum comparison with pure standards and NIST MS 2.0 database. 2. Experimental

2.2.1. FT-IR spectroscopy The FT-IR analyses were made with a Thermo-Nicolet iZ10 FT-IR spectrometer in transmission mode with OMNIC software. Absorbance was measured in the range of 4004000cm-1 wavenumber recording 32 scans. 2mg of each sample were mixed in a homogeneous way with KBr and pressed under 10T/cm-2 in order to obtain a thick KBr pellet. 2.2.2. Gas chromatography-mass spectrometry GC-MS analyses were performed with a Thermo Scientific device equipped with a GC Focus coupled with an ITQ700 ion trap mass spectrometer. Chromatographic separation was achieved with a Thermo Scientific silica capillary column (30m length, 0.25mm i.d., 0.25µm film thickness). The temperature program of the oven was from 50°C held 2min to 220°C at 8°Cmin-1, from 220°C to 260°C at 2°C, from 260°C to 330°C at 10°Cmin-1 and helium was the carrier gas. Mass spectrometry was performed using the electron ionization mode (EI) at 70eV of ionization energy. Ion source temperature was 250°C. Injections were realized in splitless mode and the data was treated with Xcalibur software. All organic fractions were derivatised by trimethylsilylation. The solutions were evaporated to dryness and added with 100μL of BSTFA at 70°C for 30min. The extracts were dried and dissolved in 1mL of hexane and filtered with PTFE filters. 1μL of each solution was injected in the GC–MS apparatus in triplicate. For each sample, 8mg were grinded into powder homogeneity and extracted before injection by GC-MS. The extracts were obtained following eight different procedures (Fujii 2018):

2.1. Materials and samples 2.1.1 Chemicals Acetone was of analysis grade from Sigma Aldrich (U.K). Chloroform and diethyl ether were of analysis grade, dichloromethane and hexane were of gas chromatography grade, methanol was of spectroscopy grade and all of them purchased from Merck (Germany). KBr of infrared analysis grade was from VWR (USA). N,O,bis (trimethyl) silyltrifluoroacetamide (BSTFA) containing 1% trimethylchlorosilane was purchased from Thermo (USA). PTFE membrane filters (0.45µm of pore size, 25mm diameter) were from VWR International (USA). 2.1.2. Archaeological samples The eighteen samples of Roman amphorae from Grand Congloué 2 are stored in the Centre Camille Jullian,

A. Samples were taken up in 3mL of mixed acetone and dichloromethane (1:2, v/v) in an ultrasound bath for 5min. After filtered with PTFE filters, the extracts were extracted with 1mL of dichloromethane and twice with 1mL of KOH 0.5M. The organic phase was separated (Font et al. 2007). B. Samples were directly derivatised by trimethylsilylation. C. Samples were extracted with 210µL of mixed methanol and chloroform (1:2, v/v) in an ultrasound bath for 15min twice. The extracts were filtered with PTFE filters (Mottram et al. 1999). D. After the extraction, following procedure C, 1.5mL of KOH in methanol (1M) was added for one night. Then the mixed solution with 1.5mL of water was extracted with 1mL of chloroform twice. The phase was separated (Pecci, Cau Ontiveros and Garnier 2013).

128

GC-MS analysis of pitch from Roman amphorae from Cosa in Etruria (Italy)

Characteristic Bands O-H stretching bands C-H vibration bands from methyl and methylene groups C=O stretching bands from carbonyl Deformation vibration bands from CH3 and CH2 COOH vibration bands from carboxyl acid Aromatic group bands

Wavenumbers 3700cm-1 ~3200cm-1 2956cm-1, 2928cm-1, 2868cm-1 1716cm-1, 1696cm-1 1458cm-1, 1383cm-1 1235cm-1, 1174cm-1 1606cm-1, 1498cm-1, 886cm-1, 821cm-1, 755cm-1

Figure 1. Wavenumbers of the characteristic bands of Grand Congloué 2 (sample: GC.2SN.6).

Figure 2. FT-IR spectrum of sample referred GC.2SN.6. E. After the extraction, following procedure C, the extracts were treated with 1.5mL of KOH in water (1M) and were extracted twice with 1mL of chloroform. The organic phase was taken (Pecci, Cau Ontiveros and Garnier 2013). F. The samples were taken up with 3mL of mixed solution (10% methanolic KOH and 10% aqueous KOH (2:3, v/v) heating for 3 hours at 60°C. Neutral organic fractions were extracted with 500μL of hexane three times. HCL (5M) was then added until pH 2 to acidify; the acidic fraction was treated with 500μL of diethylether three times. Both fractions were evaporated (Duce et al. 2015). G. Samples were taken up with 3mL of mixed methanol and chloroform (1:2, v/v) in an ultrasound bath for 15min twice. The phase was evaporated after the centrifugation (3000rpm/20min). 300µL of mixed solution (methanol/chloroform,1:2, v/v) were added and evaporated (Charters et al. 1997). H. Samples were extracted with 3mL of chloroform alone by adapting the same procedure G.

3. Results and discussion 3.1. FT-IR results The FT-IR spectrum of sample GC2.SN.6 (Figures 1 and 2) showed several bands characteristic of organic matter (Colombini et al. 2003, 2005; Font et al. 2007; Izzo et al. 2013). The wide band was observed at 3700cm-1 to 3200cm-1 due to the O-H stretching. This characteristic absorption indicated the presence of carboxylic acid dimer. At 2868cm-1, 2928cm-1 and 2956cm-1 we observed CH3 and CH2 vibration bands corresponding to the methyl and methylene groups that could be caused by the hydrocarbon skeleton from diterpenic structure. The two types of carbonyl group were observed at 1696cm-1 and 1716cm-1. The band of 1696cm-1 related to carboxylic acid and it was possible to correlate the other band with ketone. The observation of the weak band appearing at 1747cm-1 explained the presence of methylic ester. The aromatic ring bands were characteristic at 1606cm-1, 1498cm-1, 886cm-1, 821cm-1 and 755cm-1. In particularly,

129

Hitomi Fujii, Carole Mathe, Fabienne Olmer and Cathy Vieillescazes

Figure 3. GC-MS chromatogram of sample referred GC2SN6 with extraction mode A (after trimethylsilylation). the bands at 1606cm-1 and 1498cm-1 were due to symmetric structure of the ring. The interaction outside of the ring related to 821cm-1 and 755cm-1 was explained by the diterpenic structure hydrogenated in high levels. The bands at 1458cm-1 and 1383cm-1 were attributed to CH2 and CH3 bending. The different vibration bands from COOH at 1235cm-1 and 1174cm-1 were assigned. The FT-IR spectrums of other samples were quite similar and the specific transitional vibrations of the chemical functional groups were observed like sample GC2.SN.6. 3.2. Extraction It was necessary to find the best extraction for the archaeological sample because the collected materials were in very small quantities. So the best extraction method leading to the possibility of the characterization of organic residue material was examined. Moreover, the various extraction types concerning the amphorae of Grand Congloué 2 were assessed. However, the sample GC2.SN.9 was in larger amounts than the others and was extracted using the eight procedures (extractions A-H). The result of extraction through mode A gave the wide range peaks and many peaks were observed with retention times from 20min to 40min. Eighteen substances were identified and most of them had diterpenic markers like norabietatriene, retene and dehydroabietic acid. The oxidation and methylation derivatives of dehydroabiatic acid were equally observed.

Concerning the extraction modes B, C and D, the B one was the simplest method. But peaks were observed with less intensity than with the extraction mode A; although some resin markers were found. In fact, these protocols were insufficient to obtain good results. The substances were extracted in the wide range using the extraction mode E and the resulting chromatogram showed many peaks, similarly to the extraction mode A. However, a methanol solution was used as the extraction solvent and it was possible to have remaining methanol in the sample because its volatility is low. It appeared better to avoid this procedure while methanol could influence the gas chromatography analysis. In the neutral fraction obtained from extraction mode F natural diterpenic substances were present like retene and methyl ester of dehydroabietic acid. In the acidic fraction, the dehydroabietic acid and the oxidation derivatives of dehydroabietic acid were found. The resulting chromatogram showed a good response for main peaks; however, the procedure was not easy because the extraction time was long and the number of detected molecules was not correct in comparison with extraction A. In the case of the extraction protocols G and H, only peaks with small intensity appeared on the chromatogram and the interesting markers such as norabietatriene, simonellite and retene were not detected until 25min.

130

GC-MS analysis of pitch from Roman amphorae from Cosa in Etruria (Italy)

Compounds

GC2. GC2. GC2. GC2. GC2. GC2. GC2. GC2. GC2. GC2. GC2. SN. SN. SN. SN. SN. SN. 24268 Li2.83 24270 24210 24265 24255 24281 SN.1 SN.2 SN.3 SN.4 SN.5 6 7 8 9 10 11

Phenanthrene

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Methylphenantrene

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Dimethylphenantrene

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Tetrahydroretene

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Simonellite

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

9-ethyl-10methylanthracene

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Retene

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

18-Norabieta-8,11,13triene 19-Norabieta-8,11,13triene

8-lsopropyl-1,3dimethylanthrene Methyl ester of dehydroabietic acid Dehydroabieticacid TMS 3-hydroxydehydroabietic acid TMS 7-hydroxydehydroabietic acid TMS Methyl ester of 7-oxodehydroabietic acid 7-oxo-dehydroabietic acid TMS Methyl ester of 15-hydroxy-7-oxodehydroabietic acid 15-hydroxy-7-oxodehydroabietic acid TMS

Figure 4. Identification of all samples by GC-MS with extraction mode A. Therefore, the extraction mode A was chosen for the high quality of the resulting chromatogram and also for the optimal number of compounds detected. Furthermore, it was easy to adapt with samples from Grand Congloué 2, with the best quantities of extracted substances. 3.3. GC-MS results The GC-MS chromatogram of sample GC2.SN.6 (Figure 3) with the extraction mode A allows us to characterize a diterpenic resin. The main peak is retene, dehydroabiatic acid and its methyl ester derivative are also abundant. These compounds are known to be pine pitch markers, a resinous material belonging to the Pinaceae family (Colombini, Modugno and Ribechini 2005). During the procedure of pitch from organic resin, it needs a high temperature condition or distillation with low oxygen condition (Colombini et al. 2005). Therefore, the

abietadienic acids have been transformed to 18-norabietatriene, 19-nor-abietatrien, simonellite and also tetrahydroretene. These substances were identified in sample GC2.SN.6 and thus correspond to the intermediate substances in the thermal degradation of dehydroabiatic acid. The final product in this reaction is retene that is most stable. The high degree of oxidation of diterpenic acid was identified such as 7-hydroxy-dehydroabietic acid, 7-oxo-dehydroabietic acid and 15-hydroxy-7-oxodehydroabietic acid. These oxidation products are from pine pitch because of ageing and through environmental factors (Colombini, Modugno and Ribechini 2005). In addition, the presence of methyl ester of dehydroabietate indicates the presence of wood during the resin heating to prepare pine pitch. Indeed, the methanol released by hard heating of wood reacts easily with diterpenic acids present producing methyl ester

131

Hitomi Fujii, Carole Mathe, Fabienne Olmer and Cathy Vieillescazes

derivatives which are normally absent when the resin is heated alone (Colombini et al. 2005; Izzo et al. 2013). Eighteen organic molecules were identified (Figure 3). All of the samples from the 18 amphorae were studied and 18 compounds were identified in the corresponding GCMS chromatograms (Figure 4). According to the obtained results, a very similar chemical composition is observed in all of these samples, proving a homogeneous preparation of pine pitch for waterproofing the amphorae studied. 4. Conclusion The interior surfaces of eighteen roman amphorae from Etruria (Tuscany, Italy) were studied and many chemical compounds were characterized by GC-MS. The analysis of FT-IR which is a complementary analytical tool also clarified the presence of organic matter. For the GC‑MS analysis of Grand Congloué 2, the method using a mixed solution with acetone and dichloromethane, namely the extraction method A (Font et al. 2007) is highly qualitative and presents a good performance for extracting the organic matter. Therefore, many more peaks in the wide range appeared than with the other extraction methods. The procedure works conveniently with a gas chromatography analysis. It will be very helpful to characterize unknown molecules in future studies. The diterpenoids such as norabietatraene, tetrahydroxyretene, retene, dehydroabietic acid, and the oxidation substances of diterpenoids were identified from all samples of Grand Congloué 2 by the analysis of GC-MS. This proves that the material contained a pitch produced by a destructive distillation of resinous wood from the Pinaceae family. In fact, the technique of pitch production known in the Mediterranean area was used and the pitch applied to the interior of the roman amphorae (Colombini, Modugno and Ribechini 2005). Further, the occurrence of methyl diterpenic acids like methyl dehydroabietate acid indicates a treatment by destructive distillation (pyrolysis) of softwood. This study was the first chemical analysis approach to the roman amphorae of the ship wreck the Grand Congloué 2. The eighteen amphorae contained the same characteristic substances concerning the waterproof coating from pine pitch. Acknowlegments We wish Mrs Nathalie Huet who is responsible for the collection at DRASSM, for sharing samples and for assisting with our understanding of the historical background.

Bibliography Charters, S., Evershed, R. P., Quye, A., Blinkhorn, P. W. and Reeves, V. 1997. Simulation experiments for determining the use of ancient pottery vessels: the behaviour of epicuticular leaf wax during boiling of a leafy vegetable. Journal of Archaeological Science 24: 1–7. Colombini, M. P., Modugno, F. and Ribechini, E., 2005. Direct exposure electron ionization mass spectrometry and gas chromatography/mass spectrometry techniques to study organic coatings on archaeological amphorae. Journal of Mass Spectrometry 40: 675–687. Colombini, M. P., Giachi, G., Modugno, F., Pallecchi, P. and Ribechini, E. 2003. The characterization of paints and waterproofing materials from the shipwrecks found at the archaeological site of the Etruscan and Roman harbour of Pisa (Italy). Archaeometry 45: 659–674. Colombini, M. P., Giachi, G., Modugno, F. and Ribechini, E. 2005. Characterisation of organic residues in pottery vessels of the Roman age from Antinoe (Egypt). Microchemical Journal 79: 83–90. Duce, C., Orsini, S., Spepi, A., Colombini, M. P., Tiné, M. R. and Ribechini, E. 2015. Thermal degradation chemistry of archaeological pine pitch containing beeswax as an additive. Journal of Analytical and Applied Pyrolysis 111: 254–264. Font, J., Salvadó, N., Butí, S. and Enrich, J. 2007. Fourier transform infrared spectroscopy as a suitable technique in the study of the materials used in waterproofing of archaeological amphorae. Analytica Chimica Acta 598: 119–127. Fujii, H. 2018. Étude chromatographique de matériaux organiques contenus dans des amphores romaines. Unpublished PhD thesis, Université d’Avignon and des Pays de Vaucluse. Izzo, F.C., Zendri, E., Bernardi, A., Balliana, E. and Sgobbi, M. 2013. The study of pitch via gas chromatographymass spectrometry and Fourier-transformed infrared spectroscopy: the case of the Roman amphoras from Monte Poro, Calabria (Italy). Journal of Archaeological Science 40: 595–600. Mottram, H. R., Dudd, S. N., Lawrence, G. J., Stott, A. W. and Evershed, R. P. 1999. New chromatographic, mass spectrometric and stable isotope approaches to the classification of degraded animal fats preserved in archaeological pottery. Journal of Chromatography A 833: 209–221. Pecci, A., Cau Ontiveros, M. Á. and Garnier, N. 2013. Identifying wine and oil production: analysis of residues from Roman and Late Antique plastered vats. Journal of Archaeological Science 40 (12): 4491–4498. Robinson, N., Evershed, R. P., Higgs, W. J., Jerman, K. and Eglinton, G. 1987. Proof of a pine wood origin for pitch from Tudor (Mary Rose) and Etruscan shipwrecks: application of analytical organic chemistry in archaeology. Analyst 112: 637–644.

132

Residue analysis by GC-MS and FT-IR Spectroscopy on Roman amphorae from the archaeological site ‘Nuovo Mercato Testaccio’ (Rome) Florinda Notarstefano

Dipartimento di Beni Culturali, Università del Salento

Mariateresa Lettieri

Istituto di Scienze del Patrimonio Culturale, CNR–ISPC, Lecce, Italy CNR–SPIN, via Giovanni Paolo II 132, 84084 Fisciano (Salerno), Italy

Abstract: This paper reports and discusses GC-MS and FT-IR spectroscopy results obtained in the analysis of 19 samples of Early and Middle Imperial Age amphorae from a recently excavated archaeological site in Rome in the area ‘Nuovo Mercato Testaccio’. At this site, a large number of amphorae were recovered, many of which presented residues of the original content and a dark coating on their internal surfaces. The two analytical techniques (i.e., GC-MS and FT-IR) were applied to the characterization of archaeological organic remains, such as amorphous residues, organic material absorbed in the ceramic matrix or as structured remains (fish bone and scales) found in archaeological ceramic vessels. According to the results obtained by gas chromatographic and spectroscopic analysis, it was possible to suggest that the amphorae had transported mainly wine, fish by-products or oil. Key words: Organic residue analysis; GC-MS, FT-IR; amphorae content; Rome.

1. Introduction In recent years, chemical analyses of organic residues in ancient pottery coupled with the traditional archaeological investigations are accepted as an effective approach for determining the function and the content of pottery vessels, but also offer new prospects for identifying economic activities and subsistence practices associated with different cultural and technological traditions (Pollard and Heron 2008; Evershed 2008). Fourier transform infrared spectroscopy (FT-IR) is usually used as a quick and cheap method to screen archaeological samples before subjecting them to more expensive and time-consuming methods (Font et al. 2007; Tarquini, Nunziante Cesaro and Campanella 2014; Lettieri 2015). Chromatographic techniques, especially gas chromatographymass spectroscopy, are widely employed thanks to their sensitivity and reliability of separation and detection of a wide range of organic compounds, in particular lipids, in complex mixtures, after appropriate sample extraction procedures (Dudd, Regert and Evershed 1998; Stern et al. 2000; Evershed et al. 2001; Craig et al. 2007; Gregg and Slater 2010). Within the framework of a scientific collaboration for the study of ceramic objects discovered during the archaeological excavations carried out in the area ‘Nuovo Mercato Testaccio’ in 2005-2009 by the Soprintendenza Speciale per i Beni Archeologici, Roma (Sebastiani and Serlorenzi 2008; Sebastiani and Serlorenzi 2011;

Serlorenzi 2010), the chemical composition of residues found in transport amphorae was examined. The archaeological investigations have brought to light a series of rooms whose walls are made with empty amphorae stacked one upon each other. These structures, dating to the Early Imperial Age (1st century AD-beginning of the 2nd century AD), have been interpreted as a rubbish dump for building material (especially amphorae and bricks). The Middle Imperial Age (second half of the 2nd3rd century AD) is characterized by a structure identified as a warehouse (horreum), consisting of rows of rectangular rooms overlooking a central colonnade square. The space of the square was then filled with waste material, almost exclusively fragmentary amphorae. The great number of amphorae recovered during the excavations represents important evidence for better defining the development of material culture in Rome during the Early and the Middle Imperial Age and, in particular, the history of Mediterranean maritime trade, of which Rome was the arrival point. 2. Materials and methods 2.1. Samples Full descriptions of the analysed samples are given in Figure 1. Nineteen sherds were taken from fragmentary amphorae of different shapes and typologies. The sampling strategy was aimed at the analysis of vessels with amorphous surface residues and those with clear

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 133–140

Florinda Notarstefano and Mariateresa Lettieri

evidence for use, in the form of preserved remains, such as fragments of fish bones and scales. Some amphorae were covered with black carbonised residues on the inner surface, while others still preserved the remains of the original content (fish bones and scales). For these amphorae, samples were collected first by scraping the material adhering to the internal surface of the vessel, then by taking one pottery sherd. All the pottery sherds were water washed and stored in aluminium sheets or paper bags and put in sealable plastic bags, at room temperature, until required for analysis. 2.2. Sample extraction and derivatization for GC-MS analyses Potsherds were crushed to a fine powder using a mortar and pestle. The powdered ceramics were solvent extracted and derivatized using two extraction procedures, as described in Copley et al. 2005a; Charters et al. 1993; Dudd, Evershed and Gibson 1999. Charred surface residues were removed from the pottery using a sterile scalpel and powdered using a pestle and mortar. Extractions then proceeded as for the powdered sherds. The solvents and reagents used were of the highest grade available, in order to avoid possible contamination. Nonadecane (internal standard, IS) and N,Obis(trimethyl)silyltrifluoroacetamide (BSTFA) containing 1% trimethylchlorosilane were purchased from Sigma Aldrich (Milan, Italy). Approximately 1-2g of each powdered ceramic sample was weighed into glass tubes and nonadecane (1µl) was added as internal standard. A chloroform/methanol mixture was used for the extraction (4ml, 2:1 v/v, 2×20min ultrasonication). Following sonication, the test tube was placed in a centrifuge (3000rpm, 15min) to separate the solvent mixture from the inorganic clay particles. Aliquots of the total lipid extract were taken from each sample extract and reduced to a small volume by rotary evaporation, removed to a vial and gently dried under a stream of nitrogen. A portion of the dried extract was saponified by adding 4ml of 0.5M NaOH in MeOH/H2O (9:1 v/v, 10ml) and heating at 70°C for 1 hour in a sealed glass vial in ultrasonic bath. After cooling, the saponified mixture was centrifuged (3000rpm, 20min). The supernatant was acidified though the addition of a small quantity of 1M HCl and tested with pH strips. Lipids were extracted with chloroform (2×5ml). The solvent was then evaporated to virtual dryness under a gentle stream of nitrogen gas. The total lipid extracts were derivatized with few drops of bis(trimethylsilyl)trifluoroacetamide (BSTFA) + 1%

trimethylchlorosilane (TMCS) at 80°C for 30 minutes to produce trimethylsilyl derivatives. The excess solvent was removed under a slow stream of nitrogen and analysed by GC-MS. 2.2.1. Gas chromatography-mass spectrometry Samples were analysed using an Agilent Technologies 6850 II series gas chromatograph (5% phenyl-polymethylsiloxane capillary column, 30m, internal diameter 0.25mm, 0.25µm film thickness), with a split/splitless injection system used in the splitless mode and maintained at 300°C, coupled to an Agilent 5973 Network mass spectrometer operated in the EI mode (70eV). The mass spectrometer was set to scan in the range of m/z 50 to 600 in a total cycle time of 1s. The GC oven temperature was programmed from 100°C to 280°C at 10°C/ min, and held at 280°C for 15min. Helium was used as the carrier gas at a constant flow rate of 1ml/min. Compounds were identified partially by their retention time within the GC, based on comparisons with analysed reference compounds, but mainly by their mass spectra. Mass spectral data were interpreted manually with the aid of the NIST Mass Spectral Library and comparison with published mass spectra and chromatograms. 2.3. FT-IR analyses FT-IR analyses in transmission mode were carried out on the solid residues found inside the sampled vessels. The residues were finely ground with a pestle in an agate mortar, then mixed with KBr (suitable for infrared analysis and provided by Mallinckrodt Baker Chemical Inc.) and compacted in a pellet 13mm in diameter. A FT-IR ThermoNicolet Nexus spectrometer was used to record the FT-IR spectra in transmission mode. All the KBr pellets were analysed in transmission mode, immediately after the preparation. The spectra were acquired in the range of 4000–400cm-1, with a resolution of 4cm-1 and 200 scans per measurement; the background spectrum was collected on a pellet made of KBr only. The ceramic samples were subjected to µ-ATR (microAttenuated Total Reflectance) analyses. The shape and the limited dimension of most of the sherds examined in this study allowed a direct analysis. In this case, sampling or preparation procedures are not required, hence the specimen is not damaged at all and it can be used later for other kinds of analysis. Where a direct analysis cannot be performed, small flakes of the residues on the internal surfaces of the sherds were removed with a scalpel and subsequently analysed in µ-ATR mode. A ThermoNicolet Continuum IR microscope coupled with the spectrometer was used to acquire the FT-IR spectra. A 15X Reflachromat objective with a slideon ATR attachment (Thermo Spectra-Tech), using a Si crystal (refractive index = 3.4; incident angle = 45°;

134

Residue analysis by GC-MS and FT-IR Spectroscopy on Roman amphorae from the archaeological site ‘Nuovo…

Sample

Description

Part sampled

GC-MS: fatty acids yields (µg/g-1)

NMT 3088 3

Lusitanian amphora with charred surface residues

Charred surface residue

621,518

NMT 3088 4

Lusitanian amphora with charred surface residues

Wall sherd

57,245

NMT 3087 5

Lusitanian amphora with charred surface residues and organic remains (fish bones and scales, mollusca shells)

5a. Charred surface residue

82,387

5b. Wall sherd

78,23

6a. Charred surface residue

135,255

6b. Wall sherd

67,836

7a. Charred surface residue

141,188

7b. Bottom sherd

52,889

8a. Organic content

224,26

8b. Bottom sherd

60,787

NMT 3086 6

NMT 3083 7

NMT 3001 8

Lusitanian amphora with organic remains (fish bones and scales, mollusca shells) Tripolitanian amphora with organic remains (fish bones) Lusitanian amphora with organic remains (fish bones)

NMT 2145 9

Cretan amphora with black coating on the inner surface

Wall sherd

65,705

NMT 2145 10

Cretan amphora with brownish coating on the inner surface

Wall sherd

44,233

NMT 2145 11

Cretan amphora with black coating on the inner surface

Wall sherd

50,534

NMT 2145 12

Cretan amphora

Wall sherd

35,815

NMT 2145 13

Cretan amphora

Wall sherd

33,954

NMT 1815 16

Eastern amphora (Agora M54)

Bottom sherd

457,202

NMT 1739 17

Eastern amphora (Agora M54/Schoene XIII)

Wall sherd

43,608

NMT 1397 18

Eastern amphora (Agora M54)

Wall sherd

28,952

NMT 3072 19

Eastern amphora (Riley 203)

Bottom sherd

31,66

NMT 374 20

Eastern amphora (Riley 203)

Wall sherd

22,804

NMT 381 21

Unidentified Eastern Amphora (similar to Riley 203)

Bottom sherd

58,934

NMT 1635 22

Eastern amphora (Dressel 24 similis type)

Bottom sherd

44,305

NMT US 14 23

Eastern amphora (Camulodunum 184) with thick charred black coating on the bottom

Bottom sherd

258,152

Figure 1. Samples analysed. contact area = 50×50mm), was employed to collect the µ-ATR spectra. These spectra were acquired in the range of 4000–650cm-1, with a resolution of 4cm-1 and 200 scans for each measurement; the background spectrum was acquired in air. To evaluate the contribution of the surrounding soil and/ or other contaminants, FT-IR investigations were carried out also on the external side of the sherds (Drieu et al. 2020). All the FT-IR data were processed with the OMNIC 8.1 software (Thermo Fisher Scientific Inc). 3. Results and discussion The gas chromatograms obtained from the lipid extraction and the saponification treatment of the samples,

integrated with the results of FT-IR analyses, show the presence of different classes of compounds, most of them being fatty acids. On the basis of their content, the samples have been divided into three main groups. 3.1. Group 1. Samples 3-8 The amphorae in this group preserved a charred residue on the inner walls together with a solid organic content, mainly fish bones and scales, well visible also to the naked eye. Nevertheless, the ceramic walls of the amphorae were analysed to verify if any other residue related to the content or the coating of the amphorae could be identified. GC–MS analysis showed that the samples mainly contained a lipid component characterized by the presence of a series of linear saturated monocarboxylic

135

Florinda Notarstefano and Mariateresa Lettieri

Figure 2. Chromatogram of the total lipid extract of sample 5a. IS=Internal Standard (nonadecane). fatty acids ranging from 9 up to 24 carbon atoms, with a relative abundance of stearic acid (C18:0), as shown in Figure 2. Cholesterol was also detected. The simultaneous occurrence of fatty acids with linear and odd-number carbon atoms and of cholesterol suggests that the lipid distribution is typical of a degraded lipid material of animal origin (Evershed et al. 2002; Copley et al. 2005b). In addition, a series of long chain fatty acids (C20:0, C22:0, C24:0) and monounsaturated fatty acids (C16:1, C20:1, C22:1) have been detected. These compounds have recently been shown to be present in archaeological pottery and to provide evidence of the processing of marine products (Hansel et al. 2004; Craig et al. 2007; Brown and Heron 2005), while polyunsaturated fatty acids, which characterize marine lipids, are extremely susceptible to peroxidation and are absent in residues from archaeological ceramics (Hansel et al. 2004; Craig et al. 2007). Other fish biomarkers (i.e. isoprenoid fatty acids, phytanic acid) were not identified in our samples probably due to bacterial degradation or to the analytical protocol used. Along with the lipids, dehydroabietic and methyldehydroabietic acids were found in all samples, suggesting that Pinaceae resins were heated directly from the wood to obtain a pitch (Colombini et al. 2003; Colombini, Modugno and Ribechini 2005) probably used as a coating to waterproof the amphora, although no resinous deposits were observed on the interior surface of the sampled vessels. The presence of Pinaceae resins could be related also to an additive to flavour or to preserve the content of the amphora. The FT-IR analyses on the organic content preserved inside the amphorae evidenced that the residues mainly consisted of calcium phosphate (Figure 3a), which can be

Figure 3. FT-IR spectra of sample 5: (a) in transmission mode on solid residue preserved inside the amphora; (b) in µ-ATR mode on the inner wall of the ceramic potsherd. related to fishbone remains (Ribechini et al. 2009). The µ-ATR analysis on the ceramic potsherds detected salts of carboxylic acids and traces of vegetable resins (Figure 3b). In particular, signals around 1580, 1388 and 1254cm-1 account for the presence of tartrate salts (McGovern et al. 1999; Sahaya Shajan and Mahadevan 2005; Guasch-Jané et al. 2006; Qanbari-Taheri et al. 2020; Drieu et al. 2020). In fact, tartrates, usually as calcium or potassium salts, can derive from tartaric acid occurring in large amounts in grapes and have been proposed in many studies as a marker of wine in archaeological potteries (Guasch-Jané et al. 2006; Pecci et al. 2013).

136

Residue analysis by GC-MS and FT-IR Spectroscopy on Roman amphorae from the archaeological site ‘Nuovo…

Figure 4. Chromatogram of the total lipid extract of sample 9. IS=Internal Standard (nonadecane). The FT-IR analyses on the external side of the sherds detected only carbonates and unfired silicates from the soil (Lettieri 2015; Lettieri and Giannotta 2017). According to these results, it appears that the amphorae contained a fish sauce probably mixed with wine or vinegar, used as an additive to preserve or enhance the taste of the fish. A similar fish by-product probably was the oenogarum or the oxygarum described by ancient literary sources (see Grainger 2014). 3.2. Group 2. Samples 9-13, 22-23 The abundant presence of some diterpenoids with an abietane skeleton (dehydroabietic and 7-oxodehydroabietic acids) was detected by GC–MS analysis in all samples, together with diagnostic peaks relative to retene and methyldehydroabietate (Figure 4). The simultaneous presence of abietane acids, retene and methyl-dehydroabietate indicates that the amphorae were coated with a pitch obtained by the heating of wood from plants of the Pinaceae family (Colombini et al. 2003; Colombini, Modugno and Ribechini 2005; Ribechini et al. 2009). The lipid extractions of the samples contained only the saturated fatty acids C16:0 and C18:0 (with C16:0 more abundant) and traces of plant or animal sterols were absent in all samples, proving that the amphorae did not contain diagnostic fatty foodstuffs. The FT-IR spectra acquired on these samples showed the typical peaks of tartrates (around 1575 and 1385cm-1), most likely ascribable to the past presence of wine, or its derivatives, inside the amphorae. Also weak signals at 1707, 1671, 1453 and 1170cm-1, ascribable to vegetable resins

or pitch, were identified. All the detected compounds cannot originate from the surrounding environments since only traces of carbonates and silicates from the soil were found on the external surfaces. According to these results this group of amphorae probably contained wine and an organic coating of pine pitch. The identification of wine biomarkers in archaeological samples is possible also by GC-MS thanks to the recent development of new extraction protocols (Pecci et al. 2013), which have been successfully applied to a study on archaic Greek transport amphorae (Semeraro et al. 2017). At the time of the study presented in this paper our samples have been treated just with the lipid extraction protocol, combined with FT-IR analyses to verify a possible wine content of the amphorae with the aim to confirm these results by further analyses by GC-MS on the same samples. 3.3. Group 3. Samples 16-21 In these samples the characteristic peaks of substances of vegetable origin have been detected by GC-MS. The main component is oleic acid (C18:1) followed by palmitic acid (C16:0), and a minor proportion of stearic acid (C18:0). The predominance of a monounsaturated derivatives suggests a vegetable origin of the fatty acids, probably olive oil (Dudd, Regert and Evershed 1998). This is supported by the presence of the azelaic acid in sample 16 (Figure 5), a bicarboxylic C9:0 acid related to the degradation of oleic acid (Regert et al. 1998), particularly abundant in this sample. These results suggest that the amphorae contained a vegetable oil, but we cannot exclude that other kind of

137

Florinda Notarstefano and Mariateresa Lettieri

Figure 5. Chromatogram of the total lipid extract of sample 16. IS=Internal Standard (nonadecane). foodstuffs of vegetable origin (dried fruits?) had been transported in these containers, considering also the detection of plant sterols in some samples. Furthermore, three samples (18-20) showed low or insufficient levels of fatty acids, therefore further analyses on the same samples, and on other samples will be necessary in order to ascertain the nature of the original content of the amphorae. Some samples contained also traces of dehydroabietic and 7-oxo-dehydroabietic acids. Excluding the possibility that the occurrence of these compounds could be due to an exogenous contamination, it can be argued that the amphorae were coated with a pine resin, with the aim to avoid oil permeation through the ceramic matrix, as has been recently demonstrated by the analyses on samples of oil amphorae associated with oil permeability experiments (Romanus et al. 2009). 4. Conclusions Chemical analyses have identified both the content and the presence of natural coatings on the inner walls of the amphorae, showing the use of pitch and natural resins with waterproofing and/or antibacterial function. On the basis of FT-IR and GC-MS results, some observations relative to the organic composition of samples of the three groups can be drawn. The amphorae in group 1 have been classified as Lusitanian and Tripolitanian types, which are known to have been used for different varieties of fish sauces. Most of them still preserved fish bones and scales. Residue analyses are compatible with the presence of fish by-products,

probably flavoured with other substances. The main organic materials are in fact lipids of animal origin and pine resin, whose amount is lower than the lipid content. Ancient literary sources often describe the production processes of fish foodstuffs: garum, liquamen, and muria were salty liquids, while allec was probably a rather thick salty mush containing scales, bones, and undissolved fish matter (Curtis 2005). Since garum, liquamen, and muria were liquids, skeletal fish remains found in the analysed vessels might represent allec, while the detection of tartrates in one sample could suggest that wine or vinegar was added during the preparation process of fish sauce. Baetican and African garum amphorae, together with Lusitanian containers, are attested in a lower percentage among the amphorae recovered during the excavations. None of these Lusitanian and African amphorae bear tituli picti. The amphorae in group 2 probably transported wine and all the samples were very rich in pine pitch. The amphorae recovered during the excavations of the warehouse were classified mainly as wine containers, coming from different areas of the Eastern Mediterranean, principally Crete (in particular Dressel 43 types) and Asia Minor. The amphorae in group 3 were probably coated with pine resin and some of them could possibly have contained a vegetable oil, whose presence could be related either to olive oil or to other kind of vegetable foodstuffs that only further analyses on the samples could be able to identify. In this regard, it should be considered that the excavation area is close to Monte Testaccio, the rubbish dump organized by the Roman Empire for olive oil

138

Residue analysis by GC-MS and FT-IR Spectroscopy on Roman amphorae from the archaeological site ‘Nuovo…

containers from Baetica and Africa Proconsularis. Most of the vessels recovered at ‘Nuovo Mercato Testaccio’ are instead wine amphorae, especially containers of Eastern Aegean provenance, while oil amphorae (mainly African and Eastern Mediterranean types) are attested in a lower percentage among the pottery finds. Acknowledgments Chemical analyses of samples from the archaeological site ‘Nuovo Mercato di Testaccio’ in Rome have been carried out in the framework of a scientific collaboration between Soprintendenza Speciale per i Beni Archeologici, Roma, CNR-IBAM (Lecce), Università del Salento (Di.S.Te.B.A. and Dipartimento di Beni Culturali). The authors wish to thank Dr Renato Sebastiani, Dr Alessia Contino and Dr Lucilla D’Alessandro of the Soprintendenza Speciale per i Beni Archeologici, Roma, for the opportunity to analyse the amphorae and for their help in understanding the archaeological context and the historical background. The authors also thank Prof. Luigino Troisi (Università del Salento, Lecce) for granting the access to the Di.S.Te.B.A. Organic Chemistry Laboratory and the use of equipment and instrumentation for GC-MS analyses. Author contributions FN conceived this study, performed GC-MS investigation and interpreted the related data. ML performed FT-IR analyses and interpreted the results. Both authors read and approved the final manuscript.

Bibliography Brown, L. D. and Heron, C. 2005. Presence or absence: a preliminary study in the detection of fish oils in ceramics. In J. Mulville and A. Outram (eds), The Zooarchaeology of Milk and Fats. (Proceedings of the 9th ICAZ Conference, Durham August 2002): 67–76. Oxford, Oxbow. Charters, S., Evershed, R. P., Goad, L. J., Blinkhorn, P. W. and Denham, V. 1993. Quantification and distribution of lipid in archaeological ceramics: implications for sampling potsherds for organic residue analysis. Archaeometry 35: 211–223. Colombini, M. P., Modugno, F. and Ribechini, E. 2005. Direct exposure electron ionization mass spectrometry and gas chromatography/mass spectrometry techniques to study organic coatings on archaeological amphorae. Journal of Mass Spectrometry 40: 675–687. Colombini, M. P., Giachi, G., Modugno, F., Pallecchi, P. and Ribechini, E. 2003. The characterization of paints and waterproofing materials from the shipwrecks found at the archaeological site of the Etruscan and Roman harbour of Pisa (Italy). Archaeometry 45: 659–674.

Copley, M. S., Bland, H. A., Rose, P., Horton, M. and Evershed, R. P. 2005a. Gas chromatographic, mass spectrometric and stable carbon isotopic investigations of organic residues of plant oils and animal fats employed as illuminants in archaeological lamps from Egypt. The Analyst 130 (6): 860–871. Copley, M. S., Berstan, R., Mukherjee, A. J., Dudd, S. N., Straker, V., Payne, S. and Evershed, R. P. 2005b. Dairying in antiquity. III. Evidence from absorbed lipid residues dating to the British Neolithic. Journal of Archaeological Science 32: 523–546. Craig, O. E., Forster, M., Andersen, S. H., Koch, E., Crombé, É. P., Milner, N. J., Stern, B., Bailey, G. N. and Heron, C. P. 2007. Molecular and isotopic demonstration of the processing of aquatic products in northern European prehistoric pottery. Archaeometry 49 (1): 135–152. Curtis, R. I. 2005. Sources for production and trade of Greek and Roman processed fish. In T. Bekker-Nielsen (ed.), Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2: 31–46, Aarhus, Aarhus University Press. Drieu, L., Rageot, M., Wales, N., Stern, B., Lundy, J., Zerrer, M., Gaffney, I., Bondetti, M., Spiteri, C., Thomas-Oates, J. and Craig, O. E. 2020. Is it possible to identify ancient wine production using biomolecular approaches? STAR: Science & Technology of Archaeological Research 6 (1): 16–29. . Dudd, S. N., Regert, M. and Evershed, R. P. 1998. Assessing microbial lipid during laboratory degradations of fats and oils and pure triacylglycerols absorbed in ceramic potsherds. Organic Geochemistry 29: 1345–1354. Dudd, S. N., Evershed, R. P. and Gibson, A. M. 1999. Evidence for varying patterns of exploitation of animal products in different prehistoric pottery traditions based on lipids preserved in surface and absorbed residues. Journal of Archaeological Science 26 (12): 1473–1482. Evershed, R. P. 2008. Organic residues in archaeology: The archaeological biomarker revolution. Archaeometry 50 (6): 895–924. Evershed, R. P., Dudd, S. N., Lockheart, M. J. and Jim, S., 2001. Lipids in Archaeology. In D. R. Brothwell and A. M. Pollard (eds), Handbook of Archaeological Sciences: 331–349. London, John Wiley & Sons. Evershed, R. P., Dudd, S. N., Copley, M. S., Berstan, R., Stott, A. W., Mottram, H., Buckley, S. A. and Crossman, Z. 2002. Chemistry of archaeological animal fats. Accounts of Chemical Research 35: 660–668. Font, J., Salvadó, N., Butí, S. and Enrich, J. 2007. Fourier transform infrared spectroscopy as a suitable technique in the study of the materials used in waterproofing of archaeological amphorae. Analytica Chimica Acta 598: 119–127. Grainger, S. 2014. Garum, liquamen and muria: a new approach to the problem of definition. In E. Botte and V. Leitch (eds), Fish and Ships. Production et commerce des salsamenta durant l’Antiquité. Actes de l’atelier doctoral, Rome 18-22 juin 2012: 37–45. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17. Arles, Errance - Aix-en-Provence, Centre Camille Jullian.

139

Florinda Notarstefano and Mariateresa Lettieri

Gregg, M. W. and Slater, G. F., 2010. A new approach for the isolation, concentration, and transesterification of free fatty acids in archaeological ceramics. Archaeometry 52 (5): 833–854. Guasch-Jané, M. R., Andrés-Lacueva, C., Jáuregui, O. and Lamuela-Raventós, R. M. 2006. First evidence of white wine in ancient Egypt from Tutankhamun’s tomb. Journal of Archaeological Science 33: 98–101. Hansel, F. A., Copley, M. S., Madureira, L. A. S. and Evershed, R. P. 2004. Thermally produced ω-(o-alkylphenyl) alkanoic acids provide evidence for the processing of marine products in archaeological pottery vessels. Tetrahedron Letters 45 (14): 2999–3002. Lettieri, M. 2015. Infrared spectroscopic characterization of residues on archaeological pottery through different spectra acquisition modes. Vibrational Spectroscopy 76: 48–54. Lettieri, M. and Giannotta, M. T. 2017. Investigations by FtIr Spectroscopy on Residues in Pottery Cosmetic Vases from Archaeological Sites in the Mediterranean Basin. International Journal of Experimental Spectroscopic Techniques 2: 009. McGovern, P. E., Glusker, D. L., Moreau, R. A., Nunez, A., Beck, C., W., Simpson, E., Butrym, E. D., Exner, L. J. and Stout, E. C. 1999. A funerary feast fit for King Midas. Nature 402: 863–864. Pecci, A., Giorgi, G., Salvini, L. and Cau Ontiveros, M. A. 2013. Identifying wine markers in ceramics and plasters using gas chromatography-mass spectrometry. Experimental and archaeological materials. Journal of Archaeological Science 40 (1): 109–115. Pollard, A. M. and Heron., C. 2008. Archaeological Chemistry. Cambridge, The Royal Society of Chemistry. Qanbari‐Taheri, N., Karimy, A.‐H., Holakooei, P., and Kobarfard, F. 2020. Organic residue analysis of Iron Age ceramics from the archaeological site of Kani‐ zirin, western Iran. Archaeometry 62: 612–625. Regert, M., Bland, H. A., Dudd, S. N., van Bergen, P. F. and Evershed, R. P. 1998. Free and bound fatty acid oxidation products in archaeological ceramic vessels. Proceedings of the Royal Society of London Series B-Biological Sciences 265, 1409: 2027–2032. Ribechini, E., Colombini, M. P., Giachi, G., Modugno, F. and Pallecchi, P. 2009. A multi-analytical approach for the characterization of commodities in a ceramic jar from Antinoe (Egypt). Archaeometry 51 (3): 480–494.

Romanus, K., Baeten, J., Poblome, J., Accardo, S., Degryse, P., Jacobs, P., De Vos, D. and Waelkens M. 2009. Wine and olive oil permeation in pitched and non-pitched ceramics: relation with results from archaeological amphorae from Sagalassos, Turkey. Journal of Archaeological Science 36: 900–909. Sahaya Shajan, X. and Mahadevan, C. 2005. FT-IR spectroscopic and thermal studies on pure and impurity added calcium tartrate tetrahydrate crystals. Crystal Research and Technology 40: 598–602. Sebastiani, R. and Serlorenzi, M. 2008. Il progetto del Nuovo Mercato di Testaccio. Workshop di Archeologia Classica 5: 137–171. Sebastiani, R. and Serlorenzi, M. 2011. Nuove scoperte dall’area di Testaccio (Roma). Tecniche costruttive, riuso e smaltimento dei contenitori anforici pertinenti ad Horrea e strutture utilitarie di età imperiale. In J. Arce and B. Goffaux (eds), Horrea d’Hispanie et de la Méditerranée romaine: 67–95. Collection de la Casa de Velázquez 125. Madrid, Casa de Velázquez. Semeraro, G., Notarstefano, F., Caldarola, R., Quarta, G. and Calcagnile, L. 2017. Investigations on provenance and content of archaic transport amphorae from Castello di Alceste (S.Vito dei Normanni-Br) by chemical analyses through XRF/FP and GC-MS. In Proceedings of 3rd IMEKO International Conference on Metrology for Archaeology and Cultural Heritage (MetroArcheo 2017): 344–349. Budapest, International Measurement Confederation. Serlorenzi, M. 2010. La costruzione di un complesso horreario a Testaccio. Primi indizi per delineare l’organizzazione del cantiere edilizio. In S. Camporeale, H. Dessales and A. Pizzo (eds), Arqueología de la construcción II: 105–126. Madrid-Merida, CSIC, Instituto de Arqueología de Mérida. Stern, B., Heron, C., Serpico, M. and Bourriau, J. 2000. A comparison of methods for establishing fatty acid concentration gradients across potsherds: a case study using late Bronze Age Canaanite amphorae. Archaeometry 42 (2): 399–414. Tarquini, G., Nunziante Cesaro, S. and Campanella, L. 2014. Identification of oil residues in Roman amphorae (Monte Testaccio, Rome): A comparative FTIR spectroscopic study of archeological and artificially aged samples. Talanta 118: 195–200.

140

The contents of ancient Graeco-Italic amphorae. First analyses on the amphorae of the Filicudi F and Secca di Capistello wrecks (Aeolian Islands, Sicily) Nicolas Garnier

SAS Laboratoire Nicolas Garnier

Gloria Olcese

Università degli Studi di Milano

Abstract: The Graeco-Italic amphorae — types IV and V — (Van der Mersch 1994 classification) of Italic origins are considered as wine-amphorae, but no concrete tests for the presence of wine within them have existed until now. The recent re-examination of the cargos of the Filicudi F (300-280/250 BC) and Secca di Capistello (300-280 BC) wrecks (Aeolian Islands), within the Immensa Aequora project (www.immensaaequora.org), in collaboration with Museo Eoliano L. Bernabò Brea di Lipari, has made it possible to study these most-interesting wrecks, enabling for the first time the collection of new information on the area of the amphorae’ origins (Olcese 2010). Different types of samples have been taken in order to identify the content of the amphorae with the gas chromatography-mass spectrometry (GC-MS) method. The amphorae were waterproofed by pitch from coniferous species and contained red wine. Key words: Filicudi F wreck; Capistello wreck; Aeolian wrecks; ancient wine; Graeco-Italic amphorae.

1 The Wrecks and Their Cargos The ancient Graeco-Italic amphorae — types IV and V (Van der Mersch 1994 classification) — of Italic origin are presumed to be wine-amphorae, but no concrete tests for the presence of wine within them have existed until now. For a number of years, the Immensa Aequora project (www.immensaaequora.org) sought to reconstruct the economies and trade in the western Mediterranean basin (Olcese 2013, 2017, 2020 and forthcoming a and b; Olcese et al. 2013)1 using a multidisciplinary approach; Sicily is one of the sample areas for the study of ceramics production and viticulture (Olcese, Razza and Surace 2015, 2017; Brun, Garnier and Olcese 2020; Olcese et al. 2020).2 The recent re-examination of the cargos of the Filicudi F and Secca di Capistello wrecks (Aeolian Islands) (Figure 1), as part of the Immensa Aequora project in collaboration with Museo Eoliano L. Bernabò Brea of Lipari, has made it possible to study these most-interesting wrecks, enabling for the first time the collection of new information on the area of the amphorae’ origins (Olcese 2010: chapter

 Thanks to Museo Eoliano ‘L. Bernabò Brea’ Lipari, Dr F. Pallarés Salvador (director of the excavation with N. Lamboglia), and A. Razza and D. M. Surace for help in revising and editing this article. 2  The latest results from the archaeological and archaeometric studies have been presented as part of the panel ‘A. Making Wine in WesternMediterranean B. Production and the Trade of Amphorae: some new data from Italy’ (organised by the presenters and J.-P. Brun) at the AIAC 2018, XIX International Congress of Classical Archaeology, held at Cologne and Bonn from 22 to 26 May 2018 (Brun, Garnier and Olcese 2020). 1

Figure 1. Location of the Filicudi F and Secca di Capistello wrecks (Map Immensa Aequora project). VII with contributions by I. Iliopoulos and G. Montana; Olcese forthcoming a and b). Furthermore, two series of analysis (GC-MS) have been carried out on the amphorae types IV and V, in order to verify what their contents actually were. 1.1 Filicudi F The wreck, which was partly excavated by N. Lamboglia, is located around the Secca di Capo Graziano, Filicudi Island, Aeolian Islands (Lamboglia and Pallarés 1983; Bernabò Brea and Cavalier 1985; Parker 1992).3 The cargo includes Graeco-Italic amphorae Van der Mersch type  The chronology was proposed on the basis of the black glazed ceramics (Morel 1981: 62) and van der Mersch agrees (1994: 74). 3

Roman Amphora Contents. Reflecting on Maritime Trade of Foodstuffs in Antiquity: 141–148

Nicolas Garnier and Gloria Olcese

IV,4 black gloss pottery (skyphoi and small bowls) and household ware (Figure 2) (Cavalier and Albore Livadie 1985: 89–91; Olcese 2010: 231–239; Olcese 2020; Olcese in preparation). The amphorae that have been stamped present peculiar characteristics: the body is elongated and tapered, the neck cylindrical, the lip section is triangular and elongated. Recently the shipwreck has been reviewed and analyses have been performed on part of the cargo (Olcese 2010; Olcese forthcoming a). Mineralogical analyses have confirmed that all the Graeco-Italic amphorae have a composition compatible with the clays of Ischia and the Gulf of Naples (I. Iliopoulos and G. Montana in Olcese 2010; Olcese forthcoming a and b). The recent discoveries made during excavations for the Neapolitan underground could actually confirm the hypothesis that some of the amphorae from the Filicudi F were produced in the city’s workshops, although for the moment we cannot exclude other centres of production as possible points of origin (Giampaola and Febbraro 2011-2012; Pugliese 2014). Epigraphic data: ΓΑΡ crown M (Olcese 2010: 112–114, 297–298),5 ZΩ (Olcese 2010: 126–129),6 MEΓ (Olcese 2010: 236), ΠΥΘΕΑ (Olcese 2010: 144),7 XAPI (Olcese 2010: 156–161)8 stamped on many Graeco-Italic amphorae. Some of the stamps are attested also in Ischia, in the craft quarter of Santa Restituta, and in Naples (Olcese 2010: chapter V; Olcese 2019). The stamp ΓΑΡ crown Μ, in particular, is present on some amphorae and is sometimes found impressed twice, on the bend of the handle, sometimes even vertically, parallel to the handle (Olcese 2010: 112–114, 297–298). Dating: 300-280 BC. 1.2 Secca di Capistello The wreck, which has been partly excavated, bringing to light part of the wooden hull, is located near the southern end of Lipari island (Aeolian Islands), about 300m from the coastline, near the Secca di Capistello. The cargo, already partially reviewed in publication by H. Blanck and other researchers (Frey 1977; Blanck 1978; Frey, Hentschel and Keith 1978; Albore Livadie, Cavalier and Van der Mersch 1985: 53–64; Parker 1992: 396; Frey 2003; Morel 2004; Olcese 2010: 241–248; Olcese  All the amphorae are Greco-Italic except for one, which is a Punic type.  Based on the examples from Ischia, it was not clear whether the first letter was a Π or a Γ, but the stamps from the shipwreck seem to confirm that it is a Γ and not a Π as previously proposed by van der Mersch (1994: 174) and recently by Pugliese (2014: 91), but the uncertainty remains. 6  It is one of the most commonly attested stamps on Ischia, perhaps as an abbreviation of the name ΖΟΙΛ(ΟΣ). 7  On Ischia, there have been found two stamps reading ΠΥΘΕ; we cannot be certain whether they can be connected to the stamp from the Filicudi. At any rate, the characters and the fabric of the amphorae stamped with ΠΥΘΕ are analogous to those stamps with the crown, and allow us to hypothesize a common origin for the containers. 8  It could have a connection to a group of stamps from Pithecusa that have also been documented at Naples (Pugliese 2014). 4 5

2020), includes Graeco-Italic amphorae Van der Mersch type V, and black gloss pottery (Figure 3). The amphorae are stamped in Greek and present similar, but different morphological and compositional characteristics from the Filicudi F, such as the way they are stamped. C. Van der Mersch, on the basis of the stamps, attributed the amphorae from the wreck to the Greek section of Sicily, though admitted that it was difficult to identify the cities that produced them (Van der Mersch 1985: 64). Recent mineralogical analyses have allowed for the attribution of the amphorae at the Secca di Capistello wreck to Campania and, in particular, to mineralogical group IIa of the pottery recovered in Ischia, to which also some Graeco-Italic amphorae of the Gulf of Naples belong (I. Iliopoulos and G. Montana in Olcese 2010; Olcese forthcoming a). Epigraphic data: BIΩ (Olcese 2010: 109–110), ΔΙΩ (Olcese 2010: 116–119), ΕΥΞΕΝΩ (Olcese 2010: 121–122), ΠΑΡΗ (Olcese 2010: 143), ΠΑΡ (Olcese 2010: 140–142), ΠΙΣΤ,9 ΠOΠ or ΠΟP (?) (Olcese 2010: 144), ΠΤΥ (Olcese 2010: 244),10 ΧΑΡΗΣ between caducei (Olcese 2010: 110, 244), stamped on handles of Graeco-Italic amphorae. Some of these stamps have also been documented on Ischia (in low numbers) and at Naples (Olcese 2010, 2019). The most commonly attested stamp among the amphorae uncovered on the Secca shipwreck is ΠΑΡΗ, having 12 documented examples; it has also been noted on Ischia, among the materials from the Gosetti dump; ΠΑΡ, on the other hand, is present on Ischia, although the matrix of the stamp is different, and is impressed on the shoulders of 4 amphorae from the shipwreck. Dating: 300-280 BC. 2 Amphorae analysed by organic analyses In order to determine the original content and the nature of the foodstuffs being traded, two series of chemical analyses have been conducted, the first one dealing with the sediment directly taken directly from the amphorae, the second one using sherds of the fragmented walls of the amphorae even if no visible residue was present on the inner sides. The second series of analyses looked at five samples presented below (Figure 4), three of which come from amphorae related to the Filicudi F wreck,11 and two from the Secca di Capistello.12  At Ischia, the stamp ΠΙΣ has been documented from the Gosetti dump, cat II B.66 ([ ]IΣ- ?). 10  The stamp appears in publications from the shipwreck, but it has not been found among the Museum’s materials. 11  The first series of analysis included two other samples from Filicudi F wreck which were not adapted to the determination of the original contents of amphorae: Amphora F24 bis (17643): Intact amphora of the Graeco-Italic type IV. Stamp with ΠYΘEA in the rectangular cartouche pressed on the bend of one of the handles. Amphora F63 (17642): Intact amphora of the Graeco-Italic type IV. Stamp with ΓAP/ΠAP crown M in the rectangular cartouche pressed on the bend of one of the handles. Letters A and P are in nexus. 12  The first series of analysis included another sample from Secca di 9

142

The contents of ancient Graeco-Italic amphorae. First analyses on the amphorae of the Filicudi F and Secca…

Figure 2. Graeco-Italic amphorae van der Mersch type IV from Filicudi F wreck in Archaeological Museum of Lipari (Olcese 2010, Immensa Aequora project).

Figure 3. Graeco-Italic amphorae van der Mersch type V from Secca di Capistello wreck in Archaeological Museum of Lipari (Olcese 2010, Immensa Aequora project).

143

Nicolas Garnier and Gloria Olcese

2.1 Amphorae Analysed from the Filicudi F Amphora F20 (17621) Intact amphora of the Graeco-Italic type IV. Two stamps reading ΖΩ in the quadrangular cartouche (2 × 15.5cm and 1.8 × 1.3cm) impressed obliquely on the bend of both handles. Signs of a thumbprint on the base of each handle. Measurements: total height, 74cm; diameter of the outer lip, 18cm; diameter of the inner lip, 12cm; neck, 16.2cm; handle, 18cm; pointed base, 6.8cm. Amphora F53 (17693) Body of a type IV Graeco-Italic amphora with pointed base. Measurements: total height, 56cm; pointed base, 7cm. Amphora F75 (17701/1) Part of the neck without the lip and with the handles, part of the shoulder of a type IV Graeco-Italic amphora. Four stamps reading ΜΕΓ in reverse script across the rectangular cartouche (3 × 1cm), two for each handle, stamped both on the bend of the handle and on the base each handle. Measurements: neck, 16.5cm; handle, 17.5cm. 2.2 Amphorae Analysed from the Secca di Capistello Amphora LIP/SEC 70 (12393) Neck with lip and handles of a type V Graeco-Italic amphora. Stamp with ΧΑΡΗΣ and a caduceus in reverse script on the rectangular cartouche (4.5 × 1cm) pressed obliquely on the bend of one of the handles. Signs of a thumbprint on the base of each handle. Measurements: diameter of the outer lip, 17cm; diameter of the inner lip, 11cm; neck, 12.5cm; handle, 12.5cm. Amphora LIP/SEC 76 (12390) Type V Graeco-Italic amphora. Stamped with ΔΙΩ on the rectangular cartouche (3.5 × 1.5cm) pressed in oblique on the bend of one of the handles. Measurements: diameter of the outer lip, 17cm; diameter of the inner lip, 11.5cm; neck, 13cm; handle, 13.5cm. 3 Method and Results 3.1 Methodology

The study previewed the implementation of the doublestep protocol for the extraction and the analysis of the two successive extracts by GC-MS. Briefly, both types of samples have been extracted according the classical method, using organic solvents under ultrasonication (Charters et al. 1995), then submitted to an acido-catalysed extraction (Garnier and Valamoti 2016). The second step allows the solubilisation of precipitated or insoluble markers such as fruit markers including grape and wine, and the hydrolysis and the extraction of macromolecules like vegetal waxes. This allows us to apprehend the pollution degree of each sample. Each extract is then analysed by gas chromatography-mass spectrometry (GC-MS). This method allows the identification of each individual molecular marker according its mass spectrum, markers being previously separated by chromatography. A second step for interpreting chemical data consists of the identification of materials according the molecular associations previously established. 3.2 First analysis series A first series of three samples of sediment contained in the amphorae and consisting of mainly small seashells, easy to collect, has been analysed. These samples have not been in direct contact with the content because they got inside the amphora after its opening when the stopper had been broken, and the original liquid content disappeared by dilution in the sea. All the material that entered the amphora after it was opened couldn’t have been in contact with and absorbed the original content. Only the lining materials such as pitch, already present on the inner sides, could, by the physical diffusion process, have absorbed the new content. For this reason, only the first extraction-analysis step has been implemented, because it is inappropriate to search for wine in such samples. Therefore, their analysis allows us to evaluate the methodology and especially the choise of samplings. The three samples 12829 (Secca di Capistello), 17642 et 17643 (Filicudi F) reveal a similar chemical composition, with free and methylated diterpenic acids, their oxidation markers (3-, 7-, and 15-hydroxy dehydroabietic acids), and aromatic polyhydrocarbons, all being markers of pitch from coniferous (Figure 5). No markers from wax, animal fat, or vegetal oil were detected.

To determine the content of Graeco-italic amphorae traded in the both shipwrecks of Filicudi F and Secca di Capistello, two series of analyses have been organised according the availability of the material; a first series consisted of the analysis of the sediment contained in the amphorae, the second one dealt with the organic invisible residues absorbed in the inner sides of the amphorae.

The organic analysis of the actual ‘content’ of the amphorae indicates the presence of pitch from coniferous trees, i.e. the waterproofing material using for lining the inner walls of the amphorae in great amounts, but no trace of the real content has been detected using this simple sampling method.

Capistello wreck which was not adapted to the determination of the original contents of amphorae: Amphora 12829: Type V Graeco-Italic amphora. Stamped with ΠΑΡH on the rectangular cartouche pressed on the bend of one of the handles.

To access to the original data of the content, it is essential to take samples from supports which were directly in contact with the original contents, i.e. the internal walls

3.3 Second series of analysis: amphora sherds

144

The contents of ancient Graeco-Italic amphorae. First analyses on the amphorae of the Filicudi F and Secca…

Figure 4. Table of some of the analysed amphorae from the wrecks Filicudi F and Secca di Capistello (photo Immensa Aequora project).

145

Nicolas Garnier and Gloria Olcese

Figure 5. Chromatogram of the first lipid extract obtained from the actual ‘content’ of the amphora no.17693, trimethylsilylated (column ZB5-MSi 20m × 0.18mm, EIMS detection at 70 eV).

Figure 6. Chromatogramms (above) of the first lipid extract from absorbed residues from the inner walls of the amphora F-53, (below) of the second lipid extract and the detail of aldaric acids (column ZB5-MSi 20m × 0.18mm, EIMS detection at 70 eV). of amphorae. In this case, the double-step protocol of extraction and analysis has been implemented. The first lipid extracts of five amphorae (Capistello 12390 and 12393, Filicudi F-20, F-53 and F-75) are very close to those obtained from the first series of samples (Figure 5), consisting of diterpenic acids, widely dominated by the free and methylated dehydroabietic acid. Reten and methyl reten allow us to specify that it is about pitch from a conifer. No trace of wax, animal fat or vegetable oil is detected. Amphorae were thus waterproofed with the pure pitch, mixed with no other natural material. So, the amphorae did not contain oil.

The second stage allows for the access to other compounds presents in small amounts compared with the markers of the pitch, namely the aldaric acids coming from fruits (Figure 6). Diterpenic acids, dehydroabietic and Δ6‑dehydro dehydroabietic, are the most common. The most interesting zone of the chromatogramm is that of aldaric acids (cf insert of the Figure 6). Five amphorae show the presence of tartaric acid, thus grape, as well as syringic acid. This marker is released from the malvidin in the used conditions of extraction. Malvidin is present only in the skin of the black grapes, and in the pulp and the skin of teinturier grapes. It is thus a marker of black grape juice and red wine if there is fermentation. The

146

The contents of ancient Graeco-Italic amphorae. First analyses on the amphorae of the Filicudi F and Secca…

malic acid, in smaller quantities, is present in all the fruits, among those grapes. It is not thus diagnostic of the grape as tartaric acid. Succinic acid, identified in appreciable amounts in the amphora F-53, indicates fermentation. It is thus fermented must, i.e. wine. 4 Conclusion If the first series of samples was not adapted to the determination of the original contents of amphorae, the second series consisting of fragments of walls of the amphorae allowed for the clear identification of pitch from conifers but also black grape juice/red wine. Five amphorae contained grape, and chemistry tells us it was a fermented beverage in one case (F-53). The studied Graeco-italic amphorae were waterproofed using pitch and transported some red wine. From a methodological point of view, this study of two series from amphorae shows the importance of a good choice of samples. Even with successful protocols, a sample unsuitable for the question asked cannot bring any useful information. The samples must have been in direct contact with the contents. Also, an unsuitable protocol (such as a methodology implementing only the first classic extraction method) applied to a correctly chosen sample will never allow for the identification of materials derived from fruits or wine.

Bibliography Albore Livadie, C., Cavalier, M. and Van der Mersch, C. 1985. Relitto della Secca di Capistello. In L. Bernabò Brea and M. Cavalier (eds), Archeologia subacquea nelle Isole Eolie, in Archeologia Subacquea 2. Bollettino d’Arte, Supplemento al n. 29: 53–64. Rome, Istituto poligrafico e Zecca dello Stato. Bernabò Brea, L. and Cavalier, M. (eds) 1985. Archeologia subacquea nelle Isole Eolie. In Archeologia Subacquea 2. Bollettino d’Arte, Supplemento al n. 29: 11–128. Rome, Istituto poligrafico e Zecca dello Stato. Blanck, H. 1978. Der Schiffsfund von der Secca di Capistello bei Lipari. Mitteilungen des Deutschen Archäologischen Instituts. Römische Abteilung 85: 91–111. Brun, J.-P., Garnier, N. and Olcese, G. (eds) 2020. A. Making Wine in Western-Mediterranean. B. Production and the Trade of Amphorae: Some New Data from Italy. Proceedings of the 19th International Congress of Classical Archaeology - Panel 3.5 (Cologne-Bonn 2226 May 2018). Heidelberg, Propylaeum. Cavalier, M., Albore Livadie, C. 1985. Capo Graziano. Relitto F. In L. Bernabò Brea and M. Cavalier (eds), Archeologia subacquea nelle Isole Eolie, in Archeologia Subacquea 2. Bollettino d’Arte, Supplemento al n. 29: 89–91. Rome, Istituto poligrafico e Zecca dello Stato.

Charters, S., Evershed, R. P., Blinkhorn, P. W. and Denham, V. 1995. Evidence for the mixing of fats and waxes in archaeological ceramics. Archaeometry 37: 113–127. Frey, D. 1977. La Secca de Capistello, Lipari. Institute of Nautical Archaeology Newsletter 3-4: 1–4. Frey, D. 2003. La Secca di Capistello, Lipari. Lipari, online paper. . Frey, D., Hentschel, F. and Keith, D. H. 1978. Deepwater archaeology. The Capistello wreck excavation, Lipari, Aeolian Island. International Journal of Nautical Archaeology 7-4: 279–300. Garnier, N. and Valamoti, S. M. 2016. Prehistoric winemaking at Dikili Tash (Northern Greece): Integrating residue analysis and archaeobotany. Journal of Archaeological Science 74: 195–206. Giampaola, D. and Febbraro, S. 2011-2012. Napoli. Piazza Nicola Amore. In G. Olcese (ed.), Atlante dei siti di produzione ceramica (Toscana, Lazio, Campania e Sicilia) con le tabelle dei principali relitti del Mediterraneo occidentale con carichi dall’Italia centro meridionale. Immensa Aequora 2: 356–360. Rome, Quasar. Lamboglia, N. and Pallarés, F.1983. Il relitto F di Filicudi. Forma Maris Antiqui 11-12: 188–199. Morel, J.-P. 1981. Céramique campanienne. Les formes. Bibliothèque des Écoles françaises d’Athènes et de Rome 244. Rome, École française de Rome. Morel, J.-P. 2004. Quatre épaves des îles Éoliennes dans le contexte méditerranéen. In P. Pelegatti and G. Spadea (eds), Dalle Arene Candide a Lipari. Scritti in onore di Luigi Bernabò Brea. Atti del convegno (Genova 3-5 febbraio 2001). Bollettino d’Arte, Volume speciale: 75–102. Rome, Istituto poligrafico e Zecca dello Stato. Olcese, G. 2004. Anfore greco-italiche antiche: alcune osservazioni sull’origine e sulla circolazione alla luce di recenti ricerche archeologiche ed archeometriche. In E. C. De Sena and H. Dessales (eds), Metodi e approcci archeologici: l’industria e il commercio nell’Italia antica. Archaeological Methods and Approaches: Industry and Commerce in Ancient Italy. British Archaeological Reports International Series 1262: 173–192. Oxford, Archaeopress. Olcese, G. 2007. The production and circulation of GraecoItalic amphorae of Campania (Ischia/Bay of Naples). The data of the archaeological and archaeometric research. Skyllis 7: 60–75. Olcese, G. 2010. Le anfore greco italiche di Ischia: archeologia e archeometria. Artigianato ed economia nel Golfo di Napoli. Immensa Aequora 1. Rome, Quasar. Olcese, G. (ed.) 2013. Immensa Aequora Workshop. Ricerche archeologiche, archeometriche e informatiche per la ricostruzione dell’economia e dei commerci nel bacino occidentale del Mediterraneo (metà IV sec.a.C. - I sec. d.C.). Atti del convegno (Roma 24-26 gennaio 2011). Immensa Aequora 3. Rome, Quasar. Olcese, G. 2017. Wine and Amphorae in Campania in the Hellenistic Age: the Case of Ischia. In G. Tol and T. C. A. de Haas (eds), The Economic Integration of Roman Italy. Rural Communities in a Globalising World. Mnemosyne, Supplements 404: 299–321. Leiden, Brill.

147

Nicolas Garnier and Gloria Olcese Olcese, G. 2019. Timbres sur amphores gréco-italiques à Ischia. Archéologie et archéométrie. In N. Badoud and A. Marangou (eds), Analyse et exploitation des timbres amphoriques grecs. Congrès international (Athènes, 5 février 2010): 263–275. Rennes, Presses universitaires de Rennes. Olcese, G. 2020. On land and sea. Production and trade of wine from Campania (3rd BC - 1st AD): Some new archaeological and archeometric data about amphorae. In J.-P. Brun, N. Garnier, and G. Olcese (eds) 2020. A. Making Wine in Western-Mediterranean. B. Production and the Trade of Amphorae: Some New Data from Italy. Proceedings of the 19th International Congress of Classical Archaeology - Panel 3.5 (Cologne-Bonn 2226 May 2018): 105–129. Heidelberg, Propylaeum. Olcese, G. (ed.) forthcoming a. Roman Ceramic Production Sites and Shipwrecks of Tyrrhenian Italy: Fabrics and Mineralogical Analyses. Immensa Aequora Fabrics Atlas. Olcese, G. (ed.) forthcoming b. Per incerta maris. Relitti e commercio romano nel Mediterraneo occidentale in epoca romana. Atti della sessione Relitti e commercio romano del convegno RAC (Roma 17 Marzo 2016). Rome, Quasar. Olcese, G. in preparation (with the contribution of Pallarès, F.). Il relitto di Filicudi F e il suo carico. Rome, Quasar. Olcese, G., Razza, A. and Surace, D. M. 2015. Fare il vino nell’Italia antica: i palmenti rupestri in Sicilia (video documentary). Produced by Class Editori, Milan, visible on the website . Olcese, G., Razza, A. and Surace, D. M. 2017. Vigne, palmenti e produzione vitivinicola: un progetto in corso. In E. F. Castagnino and M. Centanni (eds), Dioniso in Sicilia. Engramma (La tradizione classica nella memoria occidentale) 143, online paper .

Olcese, G., Giunta, S., Iliopoulos, I. and Capelli, C. 2013. Relitti della Sicilia (metà III-I sec. a.C.): primi dati delle analisi archeometriche sulle anfore. In G. Olcese (ed.), Immensa Aequora Workshop. Ricerche archeologiche, archeometriche e informatiche per la ricostruzione dell’economia e dei commerci nel bacino occidentale del Mediterraneo (metà IV sec.a.C. - I sec. d.C.). Atti del convegno (Roma 24-26 gennaio 2011). Immensa Aequora 3: 79–102. Rome, Quasar. Olcese, G., Razza, A., Surace, D. M., and Garnier, N. 2020. Ricerche multidisciplinari sui palmenti rupestri nell’Italia meridionale tirrenica. In J.-P. Brun, N. Garnier, and G. Olcese (eds) 2020. A. Making Wine in Western-Mediterranean. B. Production and the Trade of Amphorae: Some New Data from Italy. Proceedings of the 19th International Congress of Classical Archaeology - Panel 3.5 (Cologne-Bonn 22-26 May 2018): 31–60. Heidelberg, Propylaeum. Parker, A. J. 1992. Ancient Shipwrecks of the Mediterranean and the Roman Provinces. British Archaeological Reports International Series 580. Oxford, Tempus Reparatum. Pugliese, L. 2014. Anfore greco-italiche neapolitane (IV-III secolo a.C.). Naples, Scienze e Lettere. Van der Mersch, C. 1985. Les timbres de la Secca di Capistello. In L. Bernabò Brea and M. Cavalier (eds), Archeologia subacquea nelle Isole Eolie, in Archeologia Subacquea 2. Bollettino d’Arte, Supplemento al n. 29: 61–64. Rome, Istituto poligrafico e Zecca dello Stato. Van der Mersch, C. 1994. Vins et amphores de Grande Grèce et de Sicile, IVe-IIIe s.av. J.-C. Naples, Centre Jean Bérard.

148

Inland trade and consumption in context. A case study on the organic residue analysis of transport amphorae from the Balkan Peninsula (Yambol District, South-eastern Bulgaria) Silvia Polla

Institute of Classical Archaeology, Freie Universität Berlin, Germany

Andreas Springer

Laboratory for Classical and Supramolecular Mass Spectrometry, BioSupraMol, Faculty of Biology, Chemistry and Pharmacy of the Freie Universität Berlin, Germany

Birte Gruber

Faculty of Biology, Chemistry and Pharmacy of the Freie Universität Berlin, Germany

Petra Tušlová

Institute of Classical Archaeology, Charles University, Faculty of Arts, The Czech Republic

Barbora Weissová

Institute of Archaeological Studies, Ruhr University Bochum, Germany

Abstract: The case study presented here concerns the analysis of organic residues in amphorae excavated at two rural Roman and Late Antique inland settlements located along the mid-stream of the Tundza River in the Yambol District (south-eastern Bulgaria). The first group of amphorae finds comes from the Late Antique site of Dodoparon (1.5km north-west of the village of Golyam Manastir). The assemblage incorporates five containers belonging to the type Late Roman Amphorae 2 and two smaller amphorae of type Kuzmanov XIV sub-variant I. The second group of finds comes from the Roman rural settlement known as Yurta (1.5km north-east of the village of Stroyno). During the campaign of 2014, two major types of amphorae were recognized: Kapitän II and Dressel 24 Family. The analysis of nine samples delivered preliminary results on the content of these amphorae types. Site taphonomy and recycling/ reuse practices are also considered in the interpretation of the results. Key words: GC-MS; south-eastern Bulgaria; inland trade; consumption; recycling; LRA 2; Kuzmanov XIV sub-variant I; Kapitän II; Dressel 24 Family.

1. The case study In this paper we present some preliminary results of the first investigation conducted in 2014 and 2015 using gas chromatography – mass spectrometry (GC-MS) for the organic residues analysis1 of a selected set of samples of Roman and Late Antique amphorae belonging to several types circulating in the Eastern Mediterranean region, excavated at two settlements located in Ancient Thrace, nowadays in the territory of south-eastern Bulgaria in Yambol District. Both discussed settlements are located inland, along the mid-stream of the Tundzha River, which was navigable in Antiquity. The Tundzha River, which confluences with the Maritza River at Edirne (ancient Hadrianopolis), created one of the main means  The GC-MS analyses of the amphorae content have been conducted within the framework of the research project ‘Roman Consumption Patterns and the Late Antique Economic Change’ conducted at the Institute of Classical Archaeology of the Freie Universität Berlin. 1

of communication between the inner territory of Ancient Thrace and the Aegean Sea, into which it empties. The connection was reinforced by a Roman road leading south along the river, likely built during the time of Trajan (Madzharov 2009: 237). Another network of Roman roads was connecting the area under discussion with the Black Sea on the east, Thracian inland on the west and the Danubian area on the north, where the province of Moesia Inferior was located during the Roman times (Figure 1). The first of the settlements discussed in this paper, Dodoparon, is placed about 21km west of the river; the second one, Yurta-Stroyno, is located about 12km east of the river; and they are about 40km apart from each other. Both settlements were likely connected with the Tundzha River and the main ‘Trajanic’ road by means of small-scale regional roads which gave them direct connections and trading possibilities with the rest of the Roman Empire.

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 149–160

Silvia Polla, Andreas Springer, Birte Gruber, Petra Tušlová and Barbora Weissová

Figure 1. Map of Bulgaria, Yambol District and Roman Thrace at the end of the 3rd century AD with the sites of Dodoparon and Yurta-Stroyno in the context of the province. Important Roman roads and rivers for our area are marked on the map, as well as selected settlements. By Petra Tušlová, February 2016. Based on the organic residue analysis, we would like in the future to get a better understanding of the inland trade and economy of the Ancient Thrace during the Roman period and Late Antiquity, assessing patterns of distribution, use and reuse/secondary use of transport containers in the context of the individual settlements. The archaeological inland context generally can hint a strong degree of reuse of the transport container on site. For the present preliminary investigation on the content of the transport containers from the two settlements, a total number of nine samples from different types of amphorae was analysed: four samples were taken from LRA 2 and two samples from Kuzmanov XIV sub-variant I amphorae from the site of Dodoparon; two samples were taken from Kapitän II amphorae and one sample from a Dressel 24 family amphora from the site of Yurta-Stroyno. 2. The Dodoparon assemblage The first group of amphorae was uncovered during the 2010 excavation, directed under the cooperation of the Regional Historical Museum of Yambol and Tundzha Regional Archaeological Project (Bakardzhiev 2011; Sobotkova, Longford and Bakardzhiev 2018). The site

under investigation was a fortified hilltop settlement known as Dodoparon (as well as Dadopara), identified on the basis of three inscriptions dated from the 2nd to 3rd centuries AD, found in its vicinity (Velkov 1991: 26; Mihailov 1964: 254). The settlement covers an area of 4.2 hectares stretching over the hilltop in a north-south (585m) and east-west (80–40m) direction. It is built on the highest peak of the Manastirski Highlands (about 600msl), situated 1.5km north-west from the modern village of Golyam Manastir. Three trenches were opened during the excavation, two alongside the settlement fortification wall and one inside the fortified area. Pursuant to the uncovered material, the settlement occupation dates back from the 3rd up to the 6th century AD (Sobotkova, Longford and Bakardzhiev 2018: 207). The central trench revealed foundation walls of a house measuring ca. 8 by 6 meters and stretching in a north-west to south-east direction, following the course of the fortification wall. Within the house, about 57 vessels together with a hoard of nine coins of Justinian I and Justin II were uncovered under a layer of roof tiles and fired mud bricks. The vessels were all broken into small pieces and partly burned as a consequence of the roof collapse caused by a fire (for the detailed excavation report, see: Sobotkova, Longford and Bakardzhiev 2018). The assemblage contained cooking/

150

Inland trade and consumption in context. A case study on the organic residue analysis of transport amphorae…

storage pots (19 vessels); lids (7); jugs, table amphorae and carafes for water/other liquids (12); several cups/ mugs (2); specialized kitchenware (2); dolia (6); transport amphorae (7+1 lid); and one dish of Phocaean Red slip ware (for the amphorae see Tušlová 2017; for the whole assemblage Tušlová 2019). The coin hoard dates the deposition layer after the year AD 572, while the house destruction may relate to the invasion of Avars in the Balkan Peninsula in AD 587, when many cities in inner Thrace were besieged (Velkov 1983: 234). The amphorae are represented by five vessels of a type known as Late Roman Amphorae 2 (Riley 1979 and 1981) and by two vessels Kuzmanov XIV sub-variant I (Kuzmanov 1985: 18) (Figure 2: 1-6). 2.1. Late Roman Amphorae 2 The globular Late Roman Amphorae 2 (LRA 2) are very common in the Aegean Sea area and the Balkan Peninsula — especially along the Danube River and in the area of Dobrudzha, eastern Romania — where they occur in higher quantities than in the rest of the Mediterranean (Karagiorgou 2001: 129). In northern Bulgaria, they are very well represented at Iatrus (Karagiorgou 2001: 134; Typ I, 1 of Böttger 1982), Nicopolis ad Istrum (Falkner 1999: 88 [Ware 94] and 253, Fig. 9.52, 1056–1062), Novae (Dyczek 2007), Trimmamium (Dobreva 2017: 67, Tab. 27) and Dichin (Swan 2004: 373, Fig. 5; Swan 2007: 841, Fig. 3). They are common trade containers of the Late Antiquity from the 4th until the turn of the 6th/7th century AD (Peacock and Williams 1986: 184; Scorpan 1977: 276). According to the closed context, our examples are dated to the last quarter of the 6th century AD, which classes them into the later period — type A2 — distinguished by Scorpan (1976: 160).2 The most common content of the amphorae is considered to be olive oil (Swan 2007: 836; Opaiț 2004a: 297; Karagiorgou 2001: 149; Dyczek 2001: 191; Mitchell 2005: 101–102), although wine3 (Bass 1982: 164–165; Scorpan 1977: 276) and organic remains such as resin (Radulescu 1973: 193–207: at the horrea in Tomis) have been also identified. At the Island of Samos, LRA 2 of the 7th century AD were found in the kitchen of the ecclesiastic complex in association with oil/wine presses, nearby to a grain mill (Steckner 1989: 63–64). A similar discovery was made at the Late Roman camp at Iatrus where the amphorae were found next to a grain mill and a wooden cereal box (Böttger 1982). Consequently, we may assume the LRA 2 might be re-used as cereal containers in both contexts (Steckner 1989: 64, note 15 and 65; see Peña 2007: 64 on cereals as irregular and nonstandard content of amphorae).  Although one of the amphorae (Figure 2: 1) seems to be earlier, dated already to the late 5th century AD, see Tušlová 2019: 96. 3  On the reuse issues and in particular on grape seeds content of Type 2 from the Yassı Ada cargo, see Bass 1982: 158, Fig. 8–4; and on the Yassı Ada amphorae content, see Bryant and Murrey 1982: 328 and van Doorninck 1989: 252. 2

2.2. Kuzmanov XIV sub-variant I (also known as Opaiț B V) Amphorae of the general type Kuzmanov XIV might be commonly associated with the LRA 1 (Peacock and Williams 1986: 185–187; Bjelajac 1996: 72–76), which were very common in the Late Antiquity across the Mediterranean and the Black Sea area (Opaiț 2004b: 8–9). However, the amphorae of the sub-variant I vary considerably from the general type Kuzmanov XIV. Based on their small proportions (rim d. 4‑5cm, body d. 10‑12cm and height 30‑35cm), brown fabric with white and redbrown particles, and white coating covering the outer surface (Kuzmanov 1985: 18–20), they rather correspond to the west Pontic amphorae known from Dobrudzha (Opaiț 2004b: 27).4 Considering these characteristics, the most similar containers to Kuzmanov XIV sub-variant I are the west Pontic amphorae types Opaiţ B V and Opaiț B-Id (Antonova V/Kuzmanov XVI) for which local production in the Roman province of Scythia was suggested. The majority of these containers are dated to the second half of the 6th century AD, with some persisting until the beginning of the 7th century AD. As content, low quality wine transported over a short distance was suggested (Opaiț 2004b: 27–29; Paraschiv 2014: 426). 3. The Yurta-Stroyno assemblage The second group of finds comes from the Roman rural settlement of Yurta-Stroyno. The site, known from the modern literature as ‘Yurta’ (Dimitrova and Popov 1978: 26, no.162), is located in the Elhovo region about 1.5km north-east of the small village of Stroyno. Based on a find of a bronze military diploma of classis Misenensis veteran, dated between AD 152 and 158 (Boyanov 2007: 69–74), the site was classified as a vicus of Roman military veterans (Bakardzhiev 2007). The settlement was excavated in 2006 and 2007 by the Regional Historical Museum of Yambol (Bakardzhiev 2008; Bakardzhiev 2007), and once again from 2014 to 2016 by the Yambol Museum this time in cooperation with the Institute of Classical Archaeology, Charles University, Prague (Tušlová, Weissová and Bakardzhiev 2014; Tušlová et al. 2015; Tušlová, Weissová and Bakardzhiev 2017, 2018). The settlement is dated from the 1st/2nd to the 4th century AD with lower substitution of later material suggesting its continuity to Late Antiquity (5th–6th centuries AD). The amphorae analysed within the framework of the project were found in 2014 inside a five-room house which was recently (ca. in 2004) disturbed by treasure hunters. Consequently, they are all dated based on parallels from Balkan Peninsula / Eastern Aegean and Mediterranean. Two main types of amphorae were recognized among the pottery assemblage: Kapitän II and Dressel 24 Family5 (Figure 2, 7‑9).  For the amphorae description and further discussion, see Tušlová 2017, 2019. 5  Denomination Dressel 24 family relates here to wider group of amphorae including type Dressel 24 and similar amphorae with a funnel/cup-shaped mouth whose typology and terminology is being recently discussed (Opaiţ 2007; Dobreva 2017). 4

151

Silvia Polla, Andreas Springer, Birte Gruber, Petra Tušlová and Barbora Weissová

Figure 2. The amphorae drawings, 1–4 LRA 2 from Dodoparon, 5–6 Kuzmanov XIV sub-variant I from Dodoparon, 7–8 Kapitän II from Yurta-Stroyno, 9 Dressel 24 Family from Yurta Stroyno.

152

Inland trade and consumption in context. A case study on the organic residue analysis of transport amphorae…

3.1. Kapitän II The Kapitän II amphorae are one of the most represented transport containers in the Roman Empire from the end of the 2nd to the beginning of the 5th century AD (Opaiț and Ionescu 2016: 62), with the peak period of their production/ distribution in the 3rd and 4th centuries AD (Dyczek 2001: 143–144; Bezeczky 2013: 149). In Moesia Inferior, these are the predominant Aegean provenance amphorae of the 3rd century AD (Opaiţ and Paraschiv 2013, 322). The production centre is not known, but workshops in the eastern Aegean (Dyczek 2001: 141) and northern Black Sea area, specifically Crimea (Bertoldi 2012: 135), were proposed. This type is well represented in northern Bulgaria (e.g. in Nicopolis ad Istrum, see Falkner 1999: 253, Fig. 9.52, 1050–1053). As a content, wine has been suggested (Carandini and Panella 1981: 500;6 Bjelajac 1996: 41; Bertoldi 2012: 135), and this assumption is being widely recognized by other scholars (e.g. Dyczek 2001: 143; Peacock and Williams 1986: 194; Keay 1984: 137). 3.2. Dressel 24 Family The amphorae of the Dressel 24 family are often seen as predecessor of LRA 2 (Dyczek 2001: 173–199; Opaiț 2007: 627–643) with several different production centres spread over the eastern Aegean (Opaiț and Tsaravopoulos 2011: 275–323 on recently identified centres at Chios and Erythrai) and perhaps also the western Pontus (Bertoldi 2012: 155). They were primarily distributed over western Asia Minor (on Ephesus see Bezeczky 2004), and the Aegean Sea, Black Sea and along the Danube River (for the list of places see Dyczek 2001: 183–184). The first examples known from Pompeii and Ostia date to the 1st century AD (Bruno 2005: 386) while the type is replaced by the LRA 2 within the 4th century AD (Dyczek 2001: 193). In Moesia Inferior (on the coast and in the area of the Middle/Lower Danube) and Thrace (on the coast but also inland, e.g. in Plovdiv), the Dressel 24 Family amphorae are known from the second half of the 1st century AD. Their peak seems to be from the 3rd quarter of the 1st century AD to the middle of the 2nd century AD, with single finds until AD 250 (Dobreva 2017, 224–237). Olive oil is considered to be the most common commodity (Opaiț 2010: 157; Bertoldi 2012: 155; Opaiţ and Tsaravopoulos 2011), although dipinti on some amphorae also relate to garum (as at Novae, Dyczek 2001: 192). 4. The GC-MS analyses of the content 4.1. Sample preparation, extraction, and derivatisation The analyses have been conducted on nine selected samples of amphorae base or inner part of foot. Sherds  Both Kapitän I and II amphorae most likely carrying wine are well represented in Ostia. 6

were first surface-cleaned in order to remove exogenous contamination and a total of about 2g of the powdered samples were extracted using the extraction methods for wine and oil markers analogous to the methodology described by Pecci, Cau Ontiveros and Garnier 2013 and Pecci et al. 2013, including the derivatization with BSTFA before analysis by GC-MS. The main difference to the literature was that we used a rotary evaporator for removal of extraction solvents instead of a flow of dry nitrogen (Pecci, Cau Ontiveros and Garnier 2013; Pecci et al. 2013). The GC-MS method described in the literature was customized for an Agilent G1969A GC‑MS system equipped with a DB-5MS GC column (30m × 0.25mm, Agilent Technologies), including a temperature programme similar to the literature (1min at 50°C, then a ramp of 5°C/min, 10min at 300°C) utilizing added Dotriocontan to mark the end of the chromatogram and as an internal standard. All found compounds were compared to the retention times and mass spectra of standards produced from pure compounds derivatised using BSTFA. GC-MS is a well suited technique to characterize amorphous organic residues preserved in ancient ceramic vessels, allowing first to separate and then to identify organic compounds entrapped in porous ceramic material (Heron and Evershed 1993; Evershed 2008; Regert 2011; Roffet-Salque et al. 2017). 5. Results The content of four LRA 2 from Dodoparon were analysed: in one of the samples, Figure 2, 2 (F41_07), no compounds could be detected. The sample F41_08 (Figure 2, 3 and Figure 3) shows both (plant or animal?) fats (lauric, azelaic, decanoic, palmitic and stearic acids) and other compounds (fumaric, malic and glycolic acids), and seems therefore to corroborate previous evidences and hypotheses (Karagiorgou 2001; Opaiţ 2004a and 2004b) indicating possible interchangeability of contents. Azelaic acid, a saturated dicarboxylic acid, has been detected as an oxidation product of aged olive oil in Egyptian lamps (Regert et al. 1998 and Copley et al. 2005). It is probably a degradation product of oleic acid and it is associated here with other unspecific acidic markers indicating plant or animal fat. Furthermore, no definitive evidence of wine has been detected, as the identified acidic markers of fermentation are common in the environment, highly controversial and recently strongly debated in literature (Drieu et al. 2020). The two remaining samples, F39 (Figure 2, 1 and Figure 4) and F40_16 (Figure 2, 4) show the presence of following unspecific compounds: azelaic, decanoic, dodecanoic, myristic, palmitic, stearic acids. Most of these markers are very common in the environment (i.e. palmitic and stearic acid) and may come from plant oils like e.g. olive oil, a trading good which very likely was transported in LRA 2 (Swan 2007: 836; Opaiț 2004a: 297; Karagiorgou

153

Silvia Polla, Andreas Springer, Birte Gruber, Petra Tušlová and Barbora Weissová

Figure 3. GC-MS chromatograms of the sample belonging to the sample F41_08, LRA 2 from Dodoparon.

Figure 4. GC-MS chromatograms of the sample belonging to the sample F39, LRA 2 from Dodoparon.

154

Inland trade and consumption in context. A case study on the organic residue analysis of transport amphorae…

Figure 5. GC-MS chromatograms of the sample belonging to the sample SY14_069, amphora belonging to the Dressel 24 Family from Yurta-Stroyno. 2001, 148–149 on morphology and metrology of LRA 2 in relation to olive oil as primary content), as well as animal fat (Woodworth et al. 2015: 53–54). Sample F37, a Kuzmanov XIV sub-variant I amphora from Dodoparon (Figure 2, 5), shows the presence of few unspecific fatty acids. The detection of azelaic acid might be considered to the breakdown of olive oil (see also Burton and Levy 2006); lactic acid also suggests degradation processes of the compounds, including microbial degradation. In total, lactic acid was detected in five of the nine analysed samples. Sample F38 (Figure 2, 6) is also characterized by the possible presence of plant oil (?) markers, i.e. azelaic, decanoic, myristic, arachidic, ricinoleic acid associated with glycerol, suggesting the degradation of TAGs, i.e. oxidation of original lipids. Azelaic acid, originated from oxidation of lipids, has been previously detected in relation to olive oil (Copley et al. 2005). Ricinoleic acid, which is abundant in castor oil (Copley et al. 2005 and Pecci et al. 2010), could derive also from lipids metabolites produced by ergot fungi (Lucejko et al. 2012), a common cereal pest. It could therefore be used as an indicator of reuse of the amphorae as cereal container (Peña 2007, 64), although with such a limited amount of markers, all having a limited specificity, castor oil and other sources cannot be ruled out. Both analysed Kapitän II amphorae, SY14_070 (Figure 2, 7) and SY14_122 (Figure 2, 8) from Yurta-Stroyno show an insufficient preservation of compounds. In the first sample, only the highly unspecific markers palmitic

and stearic acid are preserved, while the second sample seems to suggest plant oil as content due to the presence of azelaic (likely to be related to olive oil (?) according to Copley et al. 2005), palmitic and ricinoleic acid (see to this respect the discussion of the markers detected in the Kuzmanov XIV amphora). This would suggest that one of the analysed Kapitän II amphora has been probably used for storing and/or transporting plant oil. Nevertheless, we cannot exclude wine as a content (Carandini and Panella 1981: 500; Bjelajac 1996: 41; Bertoldi 2012: 135), because of the bad preservation of organic compounds in the sample and, in general, of specific wine markers in comparison to lipids, due to their polar nature and high water solubility. Plant fat (oil?) has been identified as the most probable source for the compounds detected in the sample of the amphora of Dressel 24 Family (SY14_069, Figure 2, 9 and Figure 5), i.e. azelaic, ricinoleic, palmitic and stearic acids. Oil is considered to be the most probable commodity carried in this type of amphora (Opaiț 2010). 6. Concluding remarks According to the preliminary character of this study, we can summarize the results as follows. Taphonomy and archaeological context of the sampled amphorae are relevant issues influencing the results and the interpretation of the detected compounds (Garnier 2015: 19, Fig. 1). Regarding our assemblage, the Dodoparon finds are from closed and stratigraphically

155

Silvia Polla, Andreas Springer, Birte Gruber, Petra Tušlová and Barbora Weissová

Location

Dodoparon Dodoparon Dodoparon Dodoparon Dodoparon

2010 Amphorae Types Kuzmanov XIV subvariant I F37 compound, Lactic acid* confirmed by Azelaic acid measurement of Palmitic acid standard (after Stearic acid derivatisation with BSTFA, with GC-MS) Linolenic acid* Year of excavation

Dodoparon

2010 Kuzmanov XIV F38

2010 LRA2 F39

2010 LRA2 F40_16

2010 LRA2 F41_07

2010 LRA 2 F41_08

Glycerol

Lactic acid

Lactic acid

None detected Lactic acid

Yurta Stroyno Yurta Stroyno Yurta Stroyno 2014 Dressel 24 SY14_069

2014 Kapitän 2 SY14_070

Lactic acid*

Palmitic acid Suberic acid

Suberic acid Azelaic acid Azelaic acid

Fumaric acid Azelaic acid

Azelaic acid Decanoic acid Myristic acid

Malic acid

Decanoic acid Dodecanoic acid

Glycolic acid* Stearic acid

Stearic acid

Myristic acid Myristic acid Stearic acid

Lauric acid

Ricinoleic acid*

Palmitic acid Lauric acid

Azelaic acid

Stearic acid Palmitic acid

Decanoic acid

Arachidic acid Stearic acid

Palmitic acid

Ricinoleic acid*

Palmityl alcohol

Palmitic acid

Stearic acid

2014 Kapitän 2 SY14_122

Palmitic acid

Azelaic acid Palmitic acid

Ricinoleic acid*

Stearic acid Stearyl alcohol Probable source of compounds

Oil?

Oil?

Oil?

Oil?

None detected

Oil and wine/fruits?

Oil?

not sufficient

Oil?

Figure 6. Overview of the results of the GC-MS analyses. * Low abundant species, mass spectra close to noise, but RT and significant peaks match well to standard compounds. well-dated contexts, while the amphorae from the rural site of Yurta-Stroyno are from recently disturbed contexts. The well dated assemblage from Dodoparon was however created by the burning down of the house. In this case we have to consider that the glowing timber of the house roof may have destroyed the organic content of amphorae, being too close to the fire, and the chemical noise may have been enhanced by condensation of fume gases, with deleterious effects on the analytical signals. The Yurta-Stroyno settlement, located along the Dereorman River, was illegally excavated recently, which might have caused short-term exposition of the pottery fragments on the surface. In absence of more specific markers, we are not able to relate the acidic markers with certainty to specific commodities, also in the case of plant oils, i.e. distinguishing castor or olive oil drawing on the preserved markers. Copley et al. 2005 detected azelaic acid as an oxidation product in replica lamps filled with olive oil and showed that ricinoleic acid is a strong indicator of castor oil. Reuse practices could also be related to the presence of ricinoleic acid derived from cereals pest (ergot fungi: Lucejko et al. 2018). It is possible that the sampled containers had multifunctional usage, carrying different contents (Figure 3), as suggested for example by the evidence related to the LRA 2. Nevertheless, the low specificity of the detected acidic markers does not allow conclusive results. As this study has still a strong preliminary character, further analyses are needed in order to support the suggested hypotheses according to the detected compounds (Figure 6).

Since the wine markers — all small, highly polar organic acids — are all very soluble in water and might have been washed out due to rain and water logging (Evershed 2008: 910–911), but lipids usually survive interaction with water, this could be the reason for the wine markers not being found (Woodworth et al. 2015). As we do not know the duration of the vessels exposure on the surface, nor the direct effect of the Dereorman River, we cannot rule out the usage of Kapitän II amphorae for carrying wine, even if no definitive markers were found. Archaeological context and taphonomy of samples as well as samples conservation, sampling strategies and adopted analytical methods have a strong impact on the scientific results (Garnier 2015; Dunne et al. 2017). At this point, the analysed set of samples is too small to make profound conclusions about individual types, as well as to try to trace distribution and consumption patterns on the middle stream of the Tundzha River during the Roman times and Late Antiquity. However, the results — matching with or deflecting from the traditional assumptions — give a room for further studies and discussions. Acknowledgments The GC-MS analyses of organic residues, conducted at the Laboratory for Classical and Supramolecular Mass Spectrometry, BioSupraMol (Faculty of Biology, Chemistry and Pharmacy of the Freie Universität Berlin), have been enabled by research funding granted by the Freie Universität Berlin to Silvia Polla. This work was supported by the European Regional Development Fund project ‘Creativity and Adaptability as Conditions of the Success of Europe in an Interrelated World’ (reg. no. CZ.02.1.01/0.0/0.0/16_019/0000734).

156

Inland trade and consumption in context. A case study on the organic residue analysis of transport amphorae…

Bibliography Bakardzhiev, S. 2007. Sondazhni arheologicheski prouchvanya v m. ‘Yurta’ pri s. Stroyno, obshtina Elhovo, oblast Yambol. Arheologicheski otkritya i razkopki prez 2006 g.: 238–241. Bakardzhiev, S. 2008. Sondazhni arheologicheski prouchvanya v m. ‘Yurta’ pri s. Stroyno, obshtina Elhovo. Arheologicheski otkritya i razkopki prez 2007 g.: 471–473. Bakardzhiev, S. 2011. Arheologicheski prouchvaniya na kusnoantichnata krepost v m. ‘Gradishteto’, s. Golyam Manastir, oblast Yambol. Arheologicheski otkritya i razkopki prez 2010 g.: 368–370. Bass, G. F. 1982. The pottery. In G. F. Bass and F. H. Van Doorninck Jr. (eds), Yassi Ada, Volume 1, A SeventhCentury Byzantine Shipwreck: 155–188. College Station, Texas A&M University Press. Bertoldi, T. 2012. Guida alle anfore romane di età imperiale. Forme, impasti e distribuzione. Rome, Espera. 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 Institute at Athens 5: 85–97. Aarhus, Aarhus University Press. Bezeczky, T. 2013. The Amphorae of Roman Ephesus. Forschungen in Ephesus 15/1. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Bjelajac, L. 1996. Amforegornjomezijskog Podunavlja. Posebna Izdanja Monographs 30. Belgrade, Srpska Akademija Nauka i Umetnosti. (summary in English: Amphorae of the Danubian Basin in Upper Moesia). Böttger, B. 1982. Die Gefässkeramik aus dem Kastell Iatrus. In B. Böttger, B. Döhle, G. Gomolka-Fuchs, E. Hajnalová and K. Wachtel (eds), Iatrus-Krivina: Spätantike Befestigung und frühmittelalterliche Siedlung an der unteren Donau. 2, Ergebnisse der Ausgrabungen 1966-1973. Schriften zur Geschichte und Kultur der Antike 17, 2: 33–148. Berlin, Akademie-Verlag. Boyanov, I. 2007. Two fragments from military diploma in Yambol museum. Archaeologia Bulgarica XI-3: 69–74. Bruno, B. 2005. Le anfore da trasporto. In D. Gandolfi (ed.), La ceramica e i materiali di età romana; classi, produzioni, commerci e consumi. Quaderni della scuola interdisciplinare delle metodologie archeologiche 2: 353–394. Bordighera, Istituto Internazionale di Studi Liguri. Bryant Jr., V. M. and Murry Jr., R. E. 1982. Preliminary analysis of amphora contents. In G. F. Bass and F. H. Van Doorninck Jr. (eds), Yassi Ada. Vol. 1, A seventhcentury Byzantine shipwreck: 327–331. Texas, A&M University Press. Burton, M. and Levi, T. E. 2006. Appendix 1 Organic Residue Analysis of selected vessels from Gilat - Gilat Torpedo Jars. In T. Levi (ed.), Archaeology, Anthropology and Cult. The Sanctuary at Gilat, Israel: 849–862. Abingdon, Routledge.

Carandini, A. and Panella, C. 1981. The trading connections of Rome and central Italy in the late second and third centuries: The evidence of the Terme del Nuotatore excavations, Ostia. In A. King (ed.), The Roman West in the Third Century. British Archaeological Reports International Series 109: 487–503. Oxford, Archaeopress. Copley, M. S., Bland, H. A., Rose, P., Horton, M. and Evershed, R. P. 2005. Gas chromatographic, mass spectrometric and stable carbon isotopic investigations of organic residues of plant oils and animal fats employed as illuminants in archaeological lamps from Egypt. The Analyst 130 (6): 860–871. Dimitrova, D. and Popov, Z. 1978. Arheologicheski pametnici v Yambolski okrug. Sofia, Izdatelstvo Septemvri. Dobreva, D. 2017. Tra Oriente e Occidente. Dinamiche commerciali in Moesia Inferior e Thracia in epoca romana i dati delle anfore. Antenor Quaderni 42. Rome, Quasar. Drieu, L., Rageot, M., Wales, N., Stern, B., Lundy, J., Zerrer, M., Gaffney, I., Bondetti, M., Spiteri, C., Thomas-Oates, J. and Craig, O. E. 2020. Is it possible to identify ancient wine production using biomolecular approaches? STAR: Science & Technology of Archaeological Research 6 (1): 16–29. . Dunne, J., et al. 2017. Organic Residue Analysis and Archaeology - Guidance for Good Practice. Swindon, Historic England. Dyczek, P. 2001. Roman Amphorae of the 1st-3rd Centuries AD Found on the Lower Danube. Typology. Warsaw, Warsaw University Press. Dyczek, P. 2007. Late Roman Amphorae from the site of valetudinarium at Novae. 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 (II): 827–834. Oxford, Archaeopress. Evershed, R. P. 2008. Organic residues in archaeology: The archaeological biomarker revolution. Archaeometry 50 (6): 895–924. Falkner, R. K. 1999. The pottery. In A. G. Poulter (ed.), Nicopolis ad Istrum: A Roman to Early Byzantine City. The Pottery and Glass. Reports of the Research Committee of the Society of Antiquaries of London 57: 55–295. London, Leicester University Press. Garnier, N. 2007. Analyse de résidus organiques conservés dans des amphores : un état de la question. 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 (I): 39–58. Oxford, Archaeopress. Garnier, N. 2015. Méthodologies d’analyse chimique organique en archéologie. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 13–39. Esposende, Município de Esposende. Heron, C. and Evershed, R. P. 1993. The analysis of organic residues and the study of pottery use. Archaeological Method and Theory 5: 247–284.

157

Silvia Polla, Andreas Springer, Birte Gruber, Petra Tušlová and Barbora Weissová

Karagiorgou, O. 2001. LRA2: a container from the military annona on the Danubian border? In S. Kingsley and M. Decker (eds), Economy and Exchange in the East Mediterranean during Late Antiquity. Proceedings of a conference at Somerville College, Oxford, 29th May, 1999: 129–166. Oxford, Oxbow Books. Keay, S. J. 1984. Late Roman Amphorae in the Western Mediterranean. A typology and economic study: the Catalan evidence. British Archaeological Reports International Series 196. Oxford, Tempus Reparatum. Kuzmanov, G. 1973. Tipologya i chronologya na rannovizantiyskite amfori (IV-VI v.). Archeologia XV/3: 14–21. Kuzmanov, G. 1985. Rannovizantiyska keramika ot Trakiya i Dakiya (IV-nachaloto na VII v.). Razkopki i prouchvaniya 13. Sofia, Bulgarian Academy of Science. Lucejko, J. J., Del Vais, C., Sanna, I. and Colombini, M. P. 2012. Caratterizzazione di residui organici in ceramiche provenienti dalla laguna di Santa Giusta. Unpublished poster, A.I.Ar. 2012 Modena. VII Congresso Nazionale di Archeometria (Modena, 22‑24 febbraio 2012). Madzharov, M. 2009. Roman Roads in Bulgaria. Contribution to the Development of Roman Road System in the Provinces of Moesia and Thrace. Veliko Turnovo, Izdatelstvo Faber. Malainey, M. 2013. Residue. In J. M. Skibo (ed.), Understanding Pottery Function: 161–189. New York, Springer. Mihailov, G. 1964. Inscriptiones Graecae in Bulgaria Repertae 3.2. Sofia, Academia Litterarum Bulgarica. Mitchell, S. 1983. The Balkans, Asia Minor, and armies across Anatolia. In S. Mitchell (ed.) Armies and Frontiers in Roman and Byzantine Anatolia. British Archaeological Reports International Series 156: 131–152. Oxford, BAR Publishing. Mitchell, S. 2005. Olive cultivation in the economy of Roman Asia Minor. In S. Mitchell and C. Katsari (eds), Patterns in the Economy of Roman Asia Minor: 83–114. Swansea, The Classical Press of Wales. Opaiț, A. 2004a. The Eastern Mediterranean Amphorae in the Province of Scythia. 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: 293–308. Aarhus, Aarhus University Press. Opaiţ, A. 2004b, Local and Imported Ceramics in the Roman Province of Scythia (4th-6th centuries AD). British Archaeological Reports International Series 1274. Oxford, Archaeopress. Opaiţ, A. 2007. From DR 24 to LR 2? 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 (II): 627–643. Oxford, Archaeopress.

Opaiţ, A. 2010. Aspects of the provisioning of the Pontic settlements with olive oil in the Hellenistic and early Roman periods. In D. Kassab Tezgör and N. Inaishvili (eds), PATABS I. Production and Trade of Amphorae in the Black Sea. Actes de la Table Ronde international de Batoumi et Trabzon, 27-29 avril 2006.Varia Anatolica 21: 153–158. Istanbul, Institut français d’Études Anatoliennes-Georges Dumézil. Opaiț, A. and Ionescu, M. 2016. Contributions to the economic life of the city of Callatis in light of new ceramic finds (2nd - 6th centuries AD). Arheologia Moldovei XXXIX: 57–112. Opaiț, A. and Paraschiv, D. 2013: On the Wine, Olive Oil and Fish Supply of the Countryside in Roman Dobrudja (1st-3rd Centuries AD). In L. Buzoianu, P. Dupont, and V. Lungu (eds), PATABS III. Production and Trade of Amphorae in the Black Sea. Production et commerce amphoriques en Mer Noire. Actes de la table ronde internationale de Constanța, 6-10 octobre 2009. Pontica 46 Supplementum II: 317–333. Constant, National History and Archaeology Museum. Opaiţ, A. and Tsaravopoulos, A. 2011. Amphorae of Dressel 24 similis type in the Central Aegean area (ChiosErythrai-Kyme). Annual of the British School at Athens 106: 275–323. Paraschiv, D. 2014. Roman-Byzantine Pontic amphorae discovered at (L)Ibida, in the province of Scythia. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeological Reports, International Series 2616 (I): 425–432. Oxford, Archaeopress. Peacock, D. P. S. and Williams, D. F. 1986. Amphorae and the Roman Economy. An Introductory Guide. London - New York, Longman. Pecci, A., Cau Ontiveros, M. Á. and Garnier, N. 2013. Identifying wine and oil production: analysis of residues from Roman and Late Antique plastered vats. Journal of Archaeological Science 40 (12): 4491–4498. Pecci, A., Salvini, L., Cirelli, E. and Augenti, A. 2010. Castor oil at Classe (Ravenna, Italy): residue analysis of some late Roman amphorae coming from the port. 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): 617–622. Oxford, Archaeopress. Pecci, A., Giorgi, G., Salvini, L. and Cau Ontiveros, M. Á. 2013. Identifying wine markers in ceramics and plasters using gas chromatography-mass spectrometry. Experimental and archaeological materials. Journal of Archaeological Science 40 (1): 109–115. Peña, J. T. 2007. Roman Pottery in the Archaeological Record. Cambridge, Cambridge University Press.

158

Inland trade and consumption in context. A case study on the organic residue analysis of transport amphorae…

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), LRCW 2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. British Archaeological Reports, International Series 1662 (II): 611–625. Oxford, Archaeopress. Radulescu, A. V. 1973. Amfore cu inscripţii de la Edificiul roman cu mozaic din Tomis. Pontica 6: 193–207. Radulescu, A. V. 1976. Amfore romane si romanobizantine din Scythia Minor. Pontica 9: 99–114. Rageot, M., Mötsch, A., Schorer, B., Gutekunst, A., Patrizi, G. et al. 2019. The dynamics of Early Celtic consumption practices: A case study of the pottery from the Heuneburg. PLOS ONE 14 (10): e0222991. . Regert, M. 2011. Analytical strategies for discriminating archaeological fatty substances from animal origin. Mass Spectrometry Review 30(2): 177–220. Regert, M., Bland, H. A., Dudd, S. N., Van Bergen, P. F. and Evershed, R. P. 1998. Free and Bound Fatty Acid Oxidation Products in Archaeological Ceramic Vessels. Proceedings of the Royal Society B: Biological Sciences 265 (1409): 2027–2032. Riley, J. A. 1979. The Coarse Pottery from Berenice. In J. A. Lloyd (ed.), Excavations at Sidi Khrebish Benghazi (Berenice) II. Supplement to Libya Antiqua V (II): 91–467. Tripoli, Society for Libyan Studies. Riley, J. A. 1981. The pottery from the cisterns 1977.1, 1977.2 and 1977.3. In J. Humphrey (ed.). Excavations at Carthage 1977 conducted by the University of Michigan VI: 85–124. Ann Arbor, University of Michigan Press. Roffet-Salque, M., Dunne, J., Altoft, D. T, Casanova E., Cramp, L. J. E., Smyth, J. et al. 2017. From the inside out: Upscaling organic residue analyses of archaeological ceramics. Journal of Archaeological Science, Reports 16: 627–640. Scorpan, C. 1976. Origini şi linii evolutive in ceramic romano-bizantină (sec. IV-VII) din spaţiul Mediteranean şi Pontic. Pontica 9: 155–186. Scorpan, C. 1977. Contribution à la connaissance de certains types céramiques romano-byzantins (IVVII siècles) dans l’espace Istro-Pontique. Dacia 21: 269–297. Sobotkova, A., Longford, C. and Bakardzhiev, S. 2018. Excavations at Dodoparon. In S. Ross, A. Sobotkova, J. Tzvetkova, G. Nehrizov and S. E. Connor (eds), The Tundzha Regional Archaeological Project. Surface Survey, Palaeoecology, and Associated Studies in Central and Southeast Bulgaria (2009-2015). Final Report: 200–210. Oxford, Oxbow Books. Steckner, C. 1989. Les Amphores LR1 and LR2 en relation avec les pressoir du complexe ecclésiastique des Thermes de Samos. In V. Déroche and J. M. Spieser (eds), Recherches sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 57–71. Athens, École française d’Athènes.

Swan, V. G. 2004. Dichin (Bulgaria) and the supply of amphorae to the Lower Danube in the late Romanearly Byzantine period. 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: 371–382. Aarhus, Aarhus University Press. Swan, V. G. 2007. Dichin (Bulgaria): The fifth- and sixthcentury destruction deposit, and their implications for ceramic chronology. 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 (II): 835–844. Oxford, Archaeopress. Tušlová, P. 2017. Late Roman Amphorae from the 6th Century AD House at the Site of Dodoparon in SouthEastern Bulgaria. In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Études alexandrines 43 (II): 671–681. Alexandria, Centre d’Études Alexandrines. Tušlová, P. 2019. Pottery from a Closed Context at the Late Antique Site of Dodoparon, Yambol Region. Archaeologia Bulgarica 3: 71–108. Tušlová, P., Weissová, B. and Bakardzhiev, S. 2014. Stroyno Excavation Project, Introduction to the Site of YurtaStroyno and Report on the 2014 Season. Studia Hercynia XVIII/1-2: 16–24. Tušlová, P., Weissová, B. and Bakardzhiev, S. 2017. YurtaStroyno Archaeological Project. Preliminary Report to the Final Season of 2016. Studia Hercynia XXI/2: 99–112. Tušlová, P., Weissová, B. and Bakardzhiev, S. 2018. Excavation at the Roman Site of Stroyno-Yurta (20142015): An Interim Report. In S. Ross, A. Sobotkova, J. Tzvetkova, G. Nehrizov and S. E. Connor (eds), The Tundzha Regional Archaeological Project. Surface Survey, Palaeoecology, and Associated Studies in Central and Southeast Bulgaria (2009-2015). Final Report: 191–199. Oxford, Oxbow Books. Tušlová, P., Weissová, B., Bakardzhiev, S., Čisťakova V., Frecer, R. and Janouchová, P. 2015. The Stroyno Archaeological Project, Report on the Season 2015. Studia Hercynia XIX/1-2: 243–261. Van Doorninck Jr., F. H. 1989. The cargo amphoras on the 7th century Yassi Ada and 11th century Serçe Limani shipwrecks: two examples of a reuse of byzantine amphoras as transport jars. In V. Déroche and J. M. Spieser (eds), Recherches sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 247–257. Athens, École française d’Athènes. Velkov, V. 1991. Nadpisi ot kabile. In V. Velkov (ed.), Kabile 2: 7–53. Sofia, Izdatelstvo na Bʺlgarskata akademiâ na naukite.

159

Silvia Polla, Andreas Springer, Birte Gruber, Petra Tušlová and Barbora Weissová

Velkov, V. 1983. The Thracian City of Kabyle. In A. G. Poulter (ed.), Ancient Bulgaria: Papers Presented at the International Symposium on the Ancient History and Archaeology of Bulgaria, Nottingham 1981. Monograph Series 1: 233–238. Nottingham, University of Nottingham. Woodworth, M., Bernal-Casasola, D., Bonifay, M., Garnier, N., Keay, S., Pecci, A., Poblome, J., Pollard, M., Richez, F. and Wilson, A. 2015. The content of African Keay 25/ Africana 3 Amphorae: Initial results of the CORONAM Project. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 41–57. Esposende, Município de Esposende.

Zanini, E. 2010. Forma delle anfore e forme del commercio tardoantico: spunti per una riflessione. In G. Guiducci, S. Menchelli and M. Pasquinucci (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): 139–148. Oxford, Archaeopress.

160

The Beirut amphora: residue analysis and contents Marshall Woodworth

Research Laboratory for Archaeology and the History of Art, University of Oxford

Paul Reynolds

Institució Catalana de Recerca i Estudis Avançats (ICREA) ICREA Research Professor and Member of the Equip de Recerca Arqueològica i Arqueomètrica de la Universitat de Barcelona (ERAAUB: Consolidated Group 2017 SGR 1043), Secció de Prehistoria i Arqueologia, Dept. de Història i Arqueologia, Universitat de Barcelona

Abstract: The Beirut amphora type was the principal amphora produced primarily in Beirut, but also in its territory, from the late 2nd century BC to the 7th century AD. Traces of staining inside and running down the outer neck, as well as occasional finds of vessels pierced through the neck, suggested that this type carried wine. Samples of Beirut amphorae and the base of one south Phoenician type (Reynolds AM 14) were analysed in an attempt to identify organic residues and the presence or not of a pitch lining. The results confirm that wine was the principal content of the vessels, all of which were lined with pitch derived from the pine family (Pinaceae). Key words: Beirut amphora type; AM 14; Roman wine production; organic residue analysis.

1. Introduction: The Beirut amphora type It is now many years since a typology for the Beirut amphora type was published (Figure 1).1 The longevity and continuous development of the type through numerous transformations could be recognized through the sequences of contexts excavated in the Anglo‑Lebanese Beirut Souks excavations (Perring 2003; Thorpe 2017‑2018), in much the same way as the same process of continuity can be observed for some other amphora types such as the LRA 3 and its Imperial predecessors, the east Cilician Pompeii  5 type or, relevant here, the south Lebanese Agora M 334 type (Reynolds 2005, 2008). The Beirut amphora type was thus classified into eight forms (Beirut 1‑8) ranging from the end of the 2nd century BC to the 7th century AD. Here, the publication of some complete early forms by ‘Ala Eddine (2005) permitted the pushing back of the introduction of the series to c. 100 BC (or perhaps a little earlier) and their linking with possible Beirut versions of what is likely to be a Sidonian type based on a Koan Hellenistic amphora (imitating its ‘mushroom’ toe) (Figure 1a-b).2  Reynolds 1997-1998 and 2000; with further comments in Reynolds 2005, the type series has been modified with respect to the earliest forms in Reynolds 2008: fig. 1. See also descriptions on the Southampton Amphora website, where some of the forms/stages are presented as separate entries. Although classified and drawn, the pottery from the sequences of contexts from the Beirut Souks excavations (BEY 006 and BEY 007) as well as those of the Imperial Baths (BEY 045) is still unpublished (Reynolds in preparation). For tables and comment on the Beirut assemblage, including the comparative quantities through time of Beirut versus imported amphorae, see Reynolds 2005 and 2010a-b. 2  See Reynolds 2008, for this evaluation of its Sidonian typological links. This Sidonian form, in a sandy yellow or orange-yellow fabric, was also produced to the south in Jiyeh (Wicenciak 2016, Jiyeh Amphora Type 2, one bearig a stamp). Although the typological evolution from the Sidonian/ 1

The Beirut amphora was, however, just one of five amphora types produced in the city of Berytus following the foundation of the Roman colony, the most important of these other forms being the small conical so‑called ‘carrot’ amphora, well-known from the Roman sources (Martial’s ‘twisted cone’), carrying exports of dried fruit (dates, figs and prunes), and a large amphora with large hollow cone toe and grooved handles, perhaps based on the Dressel 14 type, usually associated with exports of fish products (Peacock and Williams Class  12/Camulodunum  189. Reynolds et al. 2008-2009; Dalby 2000: 169). Although now, with the earlier dating of Beirut 1 (projecting pointed rim), the original suggestion that the Beirut amphora series began with the Roman conquest cannot be upheld, and it may be no coincidence that the type seems to have appeared in the years following general liberation of the region from Seleucid control.3 The association of its successor, the Beirut 2 form, with the Augustan colony is however clearer, not only because the date of the associated contexts but also because it was stamped with a reference to the name of the colony COL. BER. There was certainly more than one workshop for the Beirut 2: one was located outside the Hellenistic walls (the aforementioned workshops of Site BEY 015), another was in Khalde, just south of Beirut, and it is now known that the Beirut 2 form was produced Jiyeh form to the Beirut 1 seems a reasonable proposition, it is still unclear if Beirut produced its version of the earlier form. The clay is certainly different, more orange and, in this sense, close to the versions produced in Jiyeh (where we know that the Beirut 2 and 5 were also produced). 3  Tyre was granted independence from the Seleucids in 126/125 BC, Tripoli in 112/111, Sidon in 111 and Seleucia in 109/108 (Butcher 2003: 29). Beirut was under Sidonian control but perhaps managed to gain its freedom from Sidon in the same period?

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 161–169

Marshall Woodworth and Paul Reynolds

in Jiyeh, in a fabric that is macroscopically identical to the Beirut stamped examples (and examples of Beirut 2 sampled and presented in this paper from cistern deposit Context 12237) (Wicenziak 2016, Amphora Type 6). That there were other production sites of Beirut amphorae is indicated by other fabrics encountered, including one group in a kaolinitic fabric. For the Beirut amphora type, it was argued that the dark stains inside the vessel and runs noted outside, especially down the rim and neck, were traces of the original pitch lining (Reynolds 1997-1998 and 2000). This, and the proposed Koan amphora model, suggested that it carried wine, presumably the ‘Berytian’ mentioned by Pliny and considered to second best in the East, on a par with the wines of Sikyon (near Corinth), Cyprus, Lycia, Tripolis (the Syrian city presumably), Tyre and an Egyptian variety (Dalby 2000: 133–134, with reference to Pliny N.H. 14.73 76.). This paper is now able to offer new evidence that the Beirut amphora type indeed carried wine.4 A number of samples were taken from a range of early 1st century BC to late 4th century AD forms of the series. By fortuitous accident, one of the bases sampled (Sample 022) is probably that of the related form classified as AM 14 in the Beirut typology, a common form in Beirut, known to have been produced in southern Phoenicia, associated with wine presses (Reynolds 2005: figs. 95-102). 2. Pre-Sampling and Ceramic Preparation For ceramic samples that were taken from either reconstructed vessels or very large sherds, a portion of the ceramic body (‘pre-sample’) was cut from the vessel using a rotary drill with a diamond-coated cut-off blade. The cut-off blade was cleaned with a solution of dichloromethane/methanol (2:1 v/v) before each presample to prevent cross-contamination. Pre-samples were subsequently wrapped in aluminum foil and stored in labelled paper envelopes. Preparation of powdered samples was conducted in a clean environment in a laboratory setting. Nitrile gloves were worn during all sample preparation steps. The sherds were surface cleaned to approximately 0.25mm using a rotary tool equipped with a tungsten carbide abrasive bit. The abrasive bit was cleaned by washing three times in a solution of chloroform/methanol (2:1 v/v) before surface cleaning and again before sample harvesting. A powdered sample of approximately 1.5g was then taken from the cleaned area of the sherd onto sterile aluminum foil to a depth of approximately 3mm with the rotary tool on a low setting in order to prevent potential thermal degradation of absorbed lipids.  The organic chemical analyses were conducted by Marshall Woodworth as an M.A. thesis at the Department of History and Archaeology of the American University of Beirut (2011). 4

2.1. GC Method Two extraction techniques were used to characterize absorbed organics. Lipid constituents, such as fatty acids and resin acids, were extracted following a modified version of that used by Mottram et al. 1999. A second extraction technique, based upon Pecci et al. 2013, was used to characterize wine biomarkers. All solvents and reagents were obtained from SigmaAldrich and were GC grade. Water was obtained from an in-house Milli-Q system. Lipid Extraction For the lipid extraction, 5µl of tetratricontane (at a concentration of 1mg/ml in hexane) was added (as an internal standard) to 500mg of pulverized ceramic and was extracted 3 times with 3ml of chloroform/methanol (2:1 v/v), vortexed for 2 minutes and sonicated at 50°C for 40 minutes. After sonication, the liquid fraction was centrifuged at 4300RCF. After centrifugation, the liquid fraction was then again separated from any remaining ceramic material and dried using a nitrogen evaporator. Wine Biomarker Extraction For wine biomarker analysis, 5µl of tetratricontane (at a concentration of 1mg/ml in hexane) was added (as an internal standard) to 500mg of pulverized ceramic and was extracted with 3ml of potassium hydroxide in water (1M) and sonicated for 90 minutes at 70°C. After sonication, the liquid fraction was centrifuged at 4300RCF. After centrifugation, the liquid fraction was then separated from the settled ceramic material. Syringe filters were utilized to remove any remaining ceramic material. To prepare for post-centrifugation filtering, the liquid fraction was acidified with hydrochloric acid (16M) to a pH of approximately 4. The liquid fraction was then filtered into a clean vial and acidified again to a pH of approximately 2. Ethyl acetate (3ml) was added and vortexed for 1.5 minutes. The supernatant was removed to a new vial and dried using a nitrogen evaporator. The ethyl acetate extraction was conducted 3 times per sample. Samples were then derivatized by adding 25µl of N,OBis(trimethylsilyl)trifluoroacetamide (BSTFA). After the addition of the derivatizing agent, the sample was vortexed for approximately 30 seconds in order to maximize contact between the reagent and the dried sample extract. After vortexing, the samples were heated at approximately 70°C for 60 minutes. 2.2. GC/MS Conditions After derivatization, 1µl of sample was injected. Analyses were performed on a Thermo Scientific Trace GC Ultra equipped with a DSQ II mass spectrometer at the Kamal

162

The Beirut amphora: residue analysis and contents

Figure 1. Typological evolution of the Beirut amphora.

163

Marshall Woodworth and Paul Reynolds

A. Shair Central Research Science Laboratory at the American University of Beirut, Lebanon. The column used was a Restek TR-5MS (30m length × 0.25mm ID × 0.25μm film thickness). The mass range was scanned from m/z 50-650. Electric ionization was at 70eV. The temperature programme was as follows: initial isothermal hold for 1 minute at 50°C, 50-330°C at 5°C/min, 3 minute hold at 330°C. Several lipid extracts were later analyzed at the Research Laboratory for Archaeology and the History of Art, University of Oxford, United Kingdom. Samples were analyzed using an Agilent 7820A gas chromatograph equipped with a Restek Rxi-5ms column (30m length × 0.25mm ID × 0.25μm film thickness). The mass spectrometer was an Agilent 5975 quadrupole, operated in electron ionization mode (70eV) and the mass scan range was m/z 40-650. The gas chromatograph conditions were as follows: inlet temperature 300°C, flow rate 1.2ml/ min, transfer line temperature 280°C. Helium was used as the carrier gas. The temperature programme for the GC oven was a 50°C hold for 2 minutes, 50-300°C at 10°C/min with a 10 minutes isothermal hold at 300°C. Injections were made by an Agilent 7693A autosampler and sample injection volume was 1μl in splitless mode. 2.3. Analysis In the case of tartaric and syringic acids, standards were prepared from analytical grade tartaric and syringic acids obtained from Sigma Aldrich. Identification of tartaric and syringic acids in archaeological samples was made by comparison of the mass spectra of the archaeological samples against the retention time and mass spectra of authentic standards and NIST mass spectra records. Compounds other than tartaric and syringic acids were identified by the use of the NIST04 (AUB) and NIST11 (Oxford) mass spectral databases. Characterization of the identified compounds was assisted by a spectral database of compounds identified in archaeological samples developed by the Amber Research Laboratory, Vassar College (New York, U.S.), Middleditch (1989) (for identification of contamination constituents) and literature review. 2.4. Samples Nineteen samples of local Beirut amphorae (and one AM 14) were selected for analysis (Figures 2 and 7). The amphorae were recovered from thirteen different contexts dating between the first half of the 1st century BC and the 4th century AD, representing the forms Beirut 1, 1A, 2, 3 and 5 and a single example of Reynolds AM 14, another amphora type related to the Beirut amphora (see above). During sample selection, preference was given to excavated samples that had a base so that samples for analysis could be selected from the bottom of the vessel, an area that would have had extensive exposure to the vessel’s content.

Five of the vessels selected for analysis had visible surface residues on the vessel interiors. On four of the vessels, the residues were macroscopically identified as concretions and/or mineral deposits (from cistern deposit BEY 006.12237: Samples 004, 014; pit fill BEY 006.12233: Sample 011; pottery dump BEY 006.5051: Sample 010, concretions). The fifth sample (AM 14, from BEY.045.1381) was characterized by mortar on the exterior surface of the base (Figure 2, Sample 022). Four of the vessels selected for analysis demonstrated that the end of the base had been modified post-firing and appeared to have been drilled to produce a hole at the lower most part of the base (Figure 3). 3. Results and Discussion In total, twenty samples were analyzed for wine biomarker and lipid constituents. Nineteen of these were from different forms of the Beirut amphora type dating between the 1st and 4th century AD, one of which being the related type ‘AM 14’ (Sample 022). Of the twenty samples, tartaric acid was positively identified in six samples (Figure 8). The Beirut amphora samples corresponded to forms 1, 1A, 2, 3, 4 and 5 of the series. As described in the method, tartaric acid was identified by comparison of mass spectra against a tartaric acid standard and from the NIST11 database (Figures 4a‑c). Tartaric acid is considered to be a specific biomarker for grape and therefore wine in the Mediterranean region during the Roman period. Although tartaric acid is present in many species of New World plants as well as some species from the subcontinent and Southeast Asia, such as tamarind, these plants or their products were not widely traded commodities during the Roman Period (Pecci et al. 2013). Other compounds found in grapes (although not exclusively) co-occurred in samples in which tartaric acid was detected, specifically malonic, fumaric, succinic and malic acids (Figure 5). These other organic compounds also appeared in many samples in which tartaric acid was not detected. Syringic acid was not detected in any of the samples. Of the fourteen samples in which tartaric acid was not detected, malonic, fumaric, succinic and malic acids were identified in seven. In four of the remaining samples, malic acid as well as malonic, fumaric or succinic acids were detected. Malic, fumaric, succinic and malonic acids are compounds that are associated with grapes and their products but not exclusively (Pecci et al. 2013). It is also possible that the significant presence of resin acids associated with the pitch lining of the amphorae may have prevented identification of tartaric acid in some of the samples (Alessandra Pecci, pers. comm.). As these compounds are associated with grapes and their products, although not definitive themselves, they can only provide ancillary support that these other samples may have contained wine or its derivatives or other fruit derivatives. The three remaining samples contained only succinic acid (Sample 019, BEY 006.5051.463), fumaric and succinic acids (Sample 008, BEY 006.5051.549) or malonic acid and succinic acids (Sample 023, BEY 006.8333.21). As

164

The Beirut amphora: residue analysis and contents

Figure 2. Photographs of the Beirut amphorae sampled for analysis.

165

Marshall Woodworth and Paul Reynolds

Figure 3. Sample 019 (BEY 045.1242.x1) illustrating the base’s modification with a hole.

a

described below, these samples contained generally low levels of extractable organics demonstrated in the lipid analyses as well. Lipid analyses identified C16:0 and C18:0 fatty acids in all samples (Figure 6). In addition to the fatty acids, resin acids were also observed. Dehydroabietic acid was identified in all samples. 7-oxodehydroabietic acid and 15-hydroxy-7-oxodehydroabietic acids were also observed in all samples except for Sample 010 (BEY 006.5051.463) and Sample 023 (BEY 006.8333.21). These samples had low levels of dehydroabietic acid relative to other analyzed samples and were also notably depleted in the aforementioned compounds associated with grapes as determined in the wine biomarker analyses in which only succinic acid was detected in Sample 010, whereas malonic and succinic acids were detected in Sample 023. The sum of these analyses indicate a low level of extractable organics. This suggests that an aspect of their depositional environment, such as extensive exposure to water, may have had a negative effect on the preservation of absorbed organic compounds. The identification of tartaric acid in six of the samples corresponding to local Beirut amphorae manufactured suggests the 1st century BC and 4th century AD establishes that these vessels were used as transport vessels for wine. Multiple extractions of the other samples might have increased the number of samples in which tartaric acid could be identified. Notably absent were any detectable fatty acids that would be indicative of a content of a lipid-rich commodity such as vegetable oil (e.g. olive oil) or a fish product (salsamentum). Lipid analyses also identified resin acids consistent with an internal vessel lining with pitch derived from a member of the pine family (Pinaceae). So far it cannot be determined why syringic acid was not detected in any samples. This compound is considered as a biomarker for red wine although, in general, it is less specific than tartaric acid as a biomarker of wine itself without more advanced

b

c

Figure 4a-c. Mass spectra of the peak identified as tartaric acid (as its trimethylsiyl ester) in Sample 004 (BEY006.12237.x2) (top) compared against the mass spectra of tartaric acid (as its trimethylsilyl ester) from the prepared analytical sample (middle) and from the NIST11 database (bottom). extraction techniques (Barnard et al. 2011; Garnier and Valamoti 2016). It is both possible that the samples may not have originally contained syringic acid (i.e. the vessels contained white wine) or that other processes, such as low residual extractable organics due to leeching or diagenetic effects, resulted in insufficiently extant amounts for detection.

166

The Beirut amphora: residue analysis and contents

Figure 6. Partial total ion chromatograph of the total lipid extraction (TLE) of Sample 009 (BEY 006.12389.79).

Figure 5. Partial total ion chromatogram of the wine biomarker extraction of Sample 026 (BEY006.9429.201). Sample Excavation no. Area

Cat. Entry no. 11603 Levelling 171 Context

Date of Context

Form

AD 125-150

Fabric Part of Vessel description (in Sampled Dbase) Heavy local Base

Visible Residue

Vessel Drilled

-White mineral deposits

--

--

--

--

--

003

006

004

006

12237 Cistern fill

x2

Mid 1st c. AD

Beirut 2

Local reduced

006

006

5051 Dump

458

3rd c. AD

Beirut 4

Local

007

006

12237 Cistern fill

xl

Mid 1st c. AD

Beirut 2

Local

008

006

5051 Dump

549

Beirut 3

Local

Base

--

--

009 010

006 006

12389 5051 Dump

79 463

Early 2nd c. AD 100-50 BC 3rd c. AD

Beirut 1A

Local Local

Base Base

-Concretions

---

011

006

12233 Pit fill

139

75-50 BC

Beirut 1

Heavy local

Base

Concretion and mineral deposits

--

012 013

006 006

138 80

75-50 BC 100-50 BC

Beirut 1 Beirut 1A

Heavy local Local

Base Base

014

006

12233 Pit fill 12389 12237 Cistern fill

x3

Mid 1st c. AD

Beirut 2

Local reduced

Base

022

045

1381

19

Early 3rd c. AD

AM 14

As reduced local

Base

016

006

11629 LevellingConstruction

93

AD 125-150

Beirut 3

Local reduced

017

006

12539 Backfill of oven

1

018

006

15

019

045

12239 Drain fill (Secondary)

Late 1st c. AD

Beirut 3

---

Base

--

--

Rough local

Lower wall/ base (of almost complete vessel)

--

--

Local

Base

--

--

Local

Base

--

Base drilled

--

Base drilled

--

Base drilled --

006

1242

11603

xl xl

023

006

8333 Accumulation/ disuse

21

Early 1st c. AD Beirut 1

Local

024 026

006 006

10082 9429 Pit fill

6 201

Late 1st c. AD Late 4th c.

Local Local

Beirut 3 Beirut 5

Lower wall/ base Lower wall/ base

--

--Mineral deposits Mortar on exterior surface

Late 1st c. BC 1st c. AD Beirut 2 (AD 25-50) Beirut 3 3rd c. AD or 4 AD 125-150 Beirut 3

020

Base

Local

Base Base (of almost complete vessel) Base Base

--

---

Base drilled

Figure 7. List of Samples. Excavation area number BEY 006 indicates the excavations in the Beirut Souks, BEY 045 indicates the excavations of the Roman Imperial Baths.

167

Marshall Woodworth and Paul Reynolds

Sample no. 003 004 006 007 008 009 010 011 012 013 014 022 016 017 018 019 020 022 AM 14 023 024 026

Context Malonic acid 11603 12237 5051 12237 5051 12389 5051 12233 12233 12389 12237 1381 11629 12539 12239 1242 11603 1381 8333 10082 9429

X X X X X X X X X X X X X X X X X

Fumaric acid X X X X X X X X X X X X X X X

Succinic acid X X X X X X X X X X X X X X X X X X X X X

Malic acid Tartaric acid Syringic acid Dehydroabietic acid X X X X

X

X

X

X X X X X X X X X X X

X X

X X

X

X

X X X X X X X X X X X X X X X

X X

Figure 8. Principal compounds observed in the wine biomarker extractions.

Bibliography ’Ala Eddine, A. 2005. The development of Beirut amphorae. A general approach. Bulletin d’Archéologie et d’Architecture Libanaises, Hors Série II: 185–205. Amber Research Laboratory Mass Spectral Data Base, Vassar College, viewed 20 February 2010. . Barnard, H., Dooley, A. N., Areshian, G., Gasparyan, B., and Faull, K. F. 2011. Chemical evidence for wine production around 4000 BCE in the Late Chalcolithic Near Eastern highlands. Journal of Archaeological Science 38 (5): 977–984. Butcher, K. 2003. Roman Syria and the Near East. Los Angeles, Getty Publications. Dalby, A. 2000. Empire of Pleasures. Luxury and Indulgence in the Roman World. London and New York, Routledge. Garnier, N. and Valamoti, S. M. 2016. Prehistoric winemaking at Dikili Tash (Northern Greece): Integrating residue analysis and archaeobotany. Journal of Archaeological Science 74: 195–206. Getzov, N., Avshalom-Gorni, D., Gorin-Rosen, Y., Stern, E. J., Syon, D. and Tatcher, A. 2009. Ḥorbat ’Uẓa 2. The 1991 Excavations. Volume II: The Late Periods. Israel Antiquities Authority Reports 42. Jerusalem, Israel Antiquity Authority. Middleditch, B. S. 1989. Analytical Artifacts: GC, MS, HPLC, TLC and PC. Oxford, Elsevier. Mottram, H. R., Dudd, S. N., Lawrence, G. J., Stott, A. W. and Evershed, R. P. 1999. New chromatographic, mass spectrometric and stable isotope approaches to the classification of degraded animal fats preserved in archaeological pottery. Journal of Chromatography A 833: 209–221.

Peacock, D. P. S. and Williams, D. F. 1986. Amphorae and the Roman Economy. An Introductory Guide. London - New York, Longman. Pecci, A., Giorgi, G., Salvini, L. and Cau Ontiveros, M. Á. 2013. Identifying wine markers in ceramics and plasters using gas chromatography-mass spectrometry. Experimental and archaeological materials. Journal of Archaeological Science 40 (1): 109–115. Perring, D. 2003. The archaeology of Beirut: A report on work in the insula of the House of the Fountains. The Antiquaries Journal 83: 195–229. Reynolds, P. 1997-1998. Pottery production and economic exchange in 2nd century Berytus: some preliminary observations of ceramic trends from quantified ceramic deposits from the Anglo-Lebanese excavations in Beirut. Berytus 43: 35–110. Reynolds, P. 2000. The Beirut amphora type, 1st century BC-7th century AD: an outline of its formal development and some preliminary observations of regional economic trends. Rei Cretariae Romanae Fautorum Acta 36: 387–395. 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 Esparraguerra, J. Buxeda i Garrigós and M. A. Cau Ontiveros (eds), LRCW I. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports International Series 1340: 563– 611. Oxford, Archaeopress. Reynolds, P. 2008. Linear typologies and ceramic evolution. FACTA (A Journal of Roman Material Culture Studies) 2: 61–87. Reynolds, P. 2010a. Hispania and the Roman Mediterranean, AD 100-700: Ceramics and Trade. London, Duckworth.

168

The Beirut amphora: residue analysis and contents

Reynolds, P. 2010b. Trade networks of the East, 3rd to 7th centuries: the view from Beirut (Lebanon) and Butrint (Albania) (fine wares, amphorae and kitchen wares). In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 89–114. Oxford, Archaeopress. Reynolds P., Waksman, S. Y., Lemaître, S., Curvers, H., Roumié M. and Nsouli, B. 2008-2009. An early Imperial Roman pottery production site in Beirut (BEY 015): chemical analyses and a ceramic typology. Berytus 51-52: 71–115.

Thorpe, R. 2017-2018. The Insula of the House of Fountains in Beirut, Lebanon. In Archaeology of the Beirut Souks 3. Berytus 57-58 [2019]. Beirut. Wicenciak, U. 2016. Porphyreon. Hellenistic and Roman pottery production in the Sidon hinterland. Warsaw Monograph Series 7. Warsaw, Polish Centre of Mediterranean Archaeology, University of Warsaw.

169

Making Garum. Experimental archaeology methods Álvaro Rodríguez-Alcántara University of Cádiz

Ana Roldán-Gómez University of Cádiz

Enrique García Vargas University of Sevilla

Darío Bernal-Casasola University of Cádiz

Víctor Manuel Palacios-Macías University of Cádiz

Abstract: Recovering garum sauce has been a multidisciplinary project. Many researchers from different disciplines have been involved in this work, such as biologists, archaeologists, chemists, zoologists and many others. Garum sauce was a very popular and valuable food in the Roman era. However, its production process is still unknown, and few remains have been found in the ancient world. The aim of this study was to implement a technical approach to determine the preparation methods for ancient garum. The study starts from remnants of the last garum from Pompeii in AD 79. The archaeometric studies, combined with archaeological research and experimental archaeology methods have increased our knowledge of the contents and allowed us to develop a production method for this type of fish sauce. Key words: garum; experimental-archaeology; production methods; Pompeii; dolia.

1. Introduction Garum was a liquid produced by the hydrolysis of the flesh and internal parts of fishes mixed with salt, aromatic herbs and spices, through natural fermentation over several months (Curtis 1991, 2009). The importance of these fish sauces in daily Roman life, as well as their high impact in Antiquity, have been treated in detail by many authors (Étienne and Mayet 2002; Botte 2009; BekkerNielsen 2005). Literary sources provide few details about the preparation of the sauce, although it is known that it was produced on an artisanal scale using large storage jars — dolia — or other ceramic containers with different capacities, and on an industrial scale in fish-salting factories all around the Mediterranean Sea and the Atlantic Ocean (Ponsich and Tarradell 1965; Desse-Berset and Desse 2000; Étienne and Mayet 2002; Reese 2002). Garum production was cited in some classical texts. The closest to the years of the high Roman Empire is attributed to Gargilio Martial, and belongs to the Latin tradition. Gargilio Martial’s text was extracted from the work Medicine ex oleribus et pomis (3rd Century AD) and located in the St. Gallen monastery, with the title Confectio

liquaminis quod omogarum vocant. This text can be taken as a reference for production. The hypothetical garum production process could be similar to modern fish sauce production processes carried out in Asia (Palacios et al. 2016) and Italy (Carannante et al. 2011). Currently, Asian sauces such as nouc-mam — Vietnam — or nam-pla — Thailand — are made with similar ingredients, and some authors have remarked that modern Asian fish sauces and ancient Roman preparations likely used the same technological resources for manufacture (Curtis 2009; Grimal and Monod 1952). It is possible to use the extensive studies on Asian fish sauces to reconstruct the process of garum sauce production providing contemporary reference materials for comparative analyses. The objective of this study was to implement a technological approach to the production methods and to look at possible ingredients used in ancient garum based on the analysis of the proximal composition, fatty acids and minerals from the remains of solid garum found in six dolia of the Bottega del Garum (Pompeii, Italy), as well as the by-products obtained in the garum reconstruction

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 171–177

Álvaro Rodríguez-Alcántara et al.

process, using as reference a classical recipe (Martial), the primary components identified in these dolia, and the Asian sauce production process. 2. Materials and Methods 2.1. Samples of the archaeological remains Samples of the archaeological remains of garum — allec — were extracted from a depth of 10cm from six dolia — d1‑d6 — preserved in the so called ‘Ambiente 9’ in the ‘Garum Shop’ of Pompeii — 1.12.8 — Samples were found in an extraordinary state of preservation (Figure 1), the six vessel appeared without lapilli or volcanic ashes (Curtis 1979; Bernal‑Casasola et al. 2013). Previous zoological analysis of remnants of these dolia showed anchovy (Engraulis encrasicolus) and pollen from culinary spices (Curtis 1979; Bernal‑Casasola et al. 2013). 2.2. Fresh anchovy samples The anchovy samples were acquired at a local market in Cádiz. Five kilograms of fresh fish were deep-frozen in one piece without beheading or eviscerating, and the rest of fish was used for garum production. All samples were deep-freezen (−80°C) until the time of analysis. Prior to analysis, all the samples were lyophilized and homogenized using a mill and an agate mortar. 2.3. Total N contents The total N contents were determined using the Kjeldahl method (AOAC 2000). Samples — 1g — were digested in an automated digester, model dk6 — Velp Scientific, Italy —, and distilled by steam distillation equipment, model udk127 — Velp Scientific, Italy. Nitrogen measurements were used to calculate the organic matter present in samples and as indicative of degradation levels due to time. However the Kjeldahl method only allows us to extract the total soluble nitrogen from all the sources — aminoacids, nucleics acids, ammonia compounds and others.

2.4. Fat and fatty acid content Fat content was determined by the Soxhlet method (AOAC 2000) using an automated Solvent Extractor ser148 — Velp Scientifica, Italy — and n-hexane as the extraction solvent. Total fat content was calculated as the difference between the dry sample weight — 1g — before and after the fat extraction. The hexane-fat mix was evaporated using a rotary evaporator Hei-vap Value g1 — Heidolph, Germany  — at 60°C until total hexane evaporation. The fat collected was stored at 4°C for methylation and later fatty acid determination. Fatty acids were determined by gas chromatography after derivatization to methyl esters (FAMES) according to the IUPAC standard method. Analysis of the fames (Dieffenbacher and Pocklington 1992) was performed on an Agilent 7890-a series gas chromatography system – Agilent Technologies, Germany — coupled to a triplequadrupole cut-off high resolution mass spectrometer, model Synapt g2-s — Waters, USA —, equipped with a db-5 capillary column (30m × 0.25μm × DFID 0.25mm). 2.5. Ash and mineral content Ash content was determined by the incineration of 1g of dried sample in a muffle oven at 500°C for 2 hours. Ashes, containing the inorganic fraction, were dissolved in nitric acid (AFNOR 1996) for mineral extraction: Phosphorous (P), Calcium (Ca), Sodium (Na), Magnesium (Mg), Potassium (K), Iron (Fe), Copper (Cu) and Zinc (Zn). All the minerals were determined by inductively coupled plasma atomic emission spectroscopy — Iris Intrepid ICPAES, Thermo Scientific, USA. 2.6. Statistical analysis All results are expressed as the mean values ± standard deviation — sd. The data were subjected to one-way analysis of variance — ANOVA. Differences were accepted as statistically significant at a p-value of p < 0.05.

Figure 1. Pompeian samples.

172

Making Garum. Experimental archaeology methods

DI

D2

D3

D4

D5

D6

Allec

Fresh anchovy

Humidity (%)

0.2 ± 0.1

0.3 ± 0.1

0.2 ± 0.1

0.1 ± 0.1

0.2 ± 0.1

0.3 ± 0.1

56.2 ± 1.3

72.3 ± 4.1

Ashes (% dry matter)

92.1 ± 1.3

93.3 ± 1.2

96.9 ± 1.5

98.6 ± 2.3

98.2 ± 1.2

97.7 ± 0.9

50.7 ± 1.8

32.2 ± 2.0

Fat (% dry matter)

2.33 ± 0.11

2.09 ± 0.08

1.14 ± 0.12

0.50 ± 0.04

0.79 ± 0.07

0.99 ± 0.06

10.4 ± 1.3

7.62 ± 1.12

Total Soluble Nitrogen (% dry matter)

0.61 ± 0.06

0.45 ± 0.07

0.26 ± 0.03

0.08 ± 0.01*c

0.17 ± 0.02

0.19 ± 0.03

3.36 ± 0.53

7.84 ± 1.22

Figure 2. Chemical composition. 3. Results and Discussion 3.1. Amphora contents analysis 3.1.1. Chemical composition The results corresponding to the chemical composition of the archaeological remains — garum — with the reference of allec and fresh anchovy samples (Figure 2). The garum remains showed a high degree of dehydration, with a moisture value < 0.3%. The ash levels fluctuated between 92-98%, resulting in a calculated content of organic matter between 2 and 8%. These results suggest a good state of sample conservation considering that they are 2000 years old. Some authors (Luongo et al. 2003; Gurioli et al. 2005) reported that during the AD 79 Mount Vesuvius eruption, the dolia were covered and buried under 1-3m of volcanic ashes. This circumstance, and the dolia placement between

the garum shop walls, could have preserved these samples from the pyroclastic current. Therefore, the temperature inside the dolia could have been maintained under 100°C, limiting the combustion of organic matter. Significant percentages of fat were observed, ranging between 0.5 and 2.33% of the dry weight. These results corroborate the good state of conservation for the archaeological samples and the great physico-chemical and microbiological stability of this group of substances (Evershed 1993; Evershed et al. 2008). As expected, the nitrogen percentages in the samples were relatively low — 0.08-0.61% dry weight — and showed a linear correlation with the fat contents — r2 = 0.94. The linear correlations of total N and fat with ash are r2 =  0.94 and −0.95, respectively. So, these results showed that nitrogen present in the archaeological remains was organic nitrogen.

DI

D2

D3

D4

D5

D6

Allec

Fresh anchovy

C13:0

0.20 ± 0.02

0.04 ± 0.01

0.02 ± 0.01

n/d

0.01 ± 0.00

0.01 ± 0.00

0.01 ± 0.00

0.01 ± 0.00

C14:0

3.01 ± 0.13* 2.03 ± 0.23* 0.57 ± 0.08* 0.11 ± 0.02* 0.15 ± 0.03* 0.20 ± 0.02*

5.18 ± 1.10

1.23 ± 0.04*

C15:0

0.25 ± 0.01* 0.14 ± 0.01* 0.10 ± 0.01* 0.02 ± 0.00* 0.10 ± 0.01* 0.05 ± 0.01*

0.66 ± 0.02

0.43 ± 0.01

C16:0

5.98 ± 0.21* 7.26 ± 0.36* 4.39 ± 0.45* 0.07 ± 0.02* 0.06 ± 0.01*

2.17±0.10*

20.20 ± 1.22

9.30 ± 0.82*

C18:0

n/d

n/d

2.00 ± 0.07*

n/d

n/d

0.72 ± 0.04*

4.17 ± 0.29

1.78 ± 0.06*

C20:0

n/d

n/d

n/d

n/d

n/d

n/d

0.16 ± 0.01

0.10 ± 0.01

0.42 ± 0.03

n/d

0.35 ± 0.02

C22:0

0.02 ± 0.00* 0.04 ± 0.01* 0.09 ± 0.01*

0.35 ± 0.03

∑AGSs

9.85 ± 0.40*a 9.47 ± 0.61*a 7.43 ± 0.64*b 0.22 ± 0.04*c 0.37 ± 0.06*c 3.24 ± 0.18*b 30.74 ± 2.67d

C16:1

0.64 ± 0.12**

C18:1 n-9 C20:1 n-9 C24:1 n-9 ∑AGMs C18:2 n-6 n

0.20 ± 0.01 13.06 ± 0.95e

n/d

0.38 ± 0.04* 0.09 ± 0.01* 0.11 ± 0.02* 0.27 ± 0.05*

2.06 ± 0.23

0.02 ± 0.00*

n/d

0.29 ± 0.05*

n/d

0.08 ± 0.01* 0.21 ± 0.02*

5.04 ± 0.32

1.66±0.17*

2.15 ± 0.21

3.00 ± 0.19

n/d

n/d

0.02 ± 0.00* 0.05 ± 0.01*

1.62 ± 0.09

0.49 ± 0.08*

n/d

0.01 ± 0.00*

0.64 ± 0.03

1.32 ± 0.05* 1.36 ± 0.08* 0.01 ± 0.00*

n/d

4.12 ± 0.38*a 4,36 ± 0,27*a 0.68 ± 0.09*b 0,09 ± 0,01*c 0.22 ± 0.03d 0.544 ± 0.08b 9.37 ± 0.67e n/d

n/d

0.01 ± 0.00*

n/d

0.09 ± 0.01* 0.09 ± 0.02*

0.02 ± 0.01*

n/d

0.75 ± 0.16

1.08 ± 0.18*

0.44 ± 0.08* 3.67 ± 0.51f 1.07 ± 0.09

C18:3 n-3

n/d

n/d

0.09 ± 0.02* 0.09 ± 0.01*

0.75 ± 0.18

0.99 ± 0.03

C20:4 n-6

0.43 ± 0.05

0.13 ± 0.02

0.02 ± 0.01* 0.01 ± 0.00* 0.01 ± 0.00* 0.02 ± 0.01*

0.34 ± 0.05*

0.92 ± 0.07*

C20:5 n3

2.21 ± 0.13

2.93 ± 0.10

0.96 ± 0.07* 0.06 ± 0.02* 0.07 ± 0.01* 0.83 ± 0.06*

2.73 ± 0.09

4.73 ± 0.15*

C22:5

0.38 ± 0.05

0.41 ± 0.03

0.03 ± 0.01* 0.01 ± 0.00* 0.01 ± 0.00* 0.02 ± 0.00*

0.51 ± 0.08

0.26 ± 0.06*

C22:6 n-3

3.54 ± 0.12*

5.59 ± 0.21

1.59 ± 0.07* 0.08 ± 0.02* 0.11 ± 0.02* 0.98 ± 0.05*

6.98* ± 0.28

12.31 ± 0.83*

6.56 ± 0.35*a 9.05 ± 0.36*a 2.63 ± 0.17*b 0.16 ± 0.04*c 0.38 ± 0.06*c 2.01 ± 0.15*b 12.05 ± 0.84d

20.10 ± 1.23e

∑AGPs (∑AGSs + ∑AGMs)/∑AGPs)

2.12

1.52

3.08

1.93

1.55

1.88

3.32

Total Fatty acids (mg/100g dry 20.53 ± 1.131 22.89 ± 1.241 10.74 ± 0.90b 0.48 ± 0.09c 0.96 ± 0.15c 5.79 ± 0.41d 52.16 ± 4.18e matter

Figure 3. Fatty Acids composition.

173

0.83 36.83 ± 2.69f

Álvaro Rodríguez-Alcántara et al.

64 and 31%, respectively, of the total mineral profile in all the archaeological samples, with the exception of d5. The same result is observed in the fresh anchovy; however, the allec reference shows Ca and P contents that are lower due to the predominance of sodium, one of the ingredients of garum preparation. The Na content in all the remains should be high, but the results show otherwise. Smriga et al. (2010) reported similar low Na content in D6 and explained this observation by the effect of rain on the exposed dolia during the post-excavation phase

3.1.2. Fatty acid profile Figure 3 shows the fatty acid content in the garum archaeological remnants, allec reference and fresh anchovy. First, an increase in total fatty acid content — 41% — in the allec reference can be observed with respect to fresh anchovy. The production process of garum involves the addition of herbs and spices to the high fatty acid content. However, these ingredients do not contribute qualitatively to the new fatty acids in the fresh anchovy. Therefore, the content of these compounds in allec is high, although part of them pass to the liquid phase after garum separation. All of the remnants have significant levels of fatty acids, and clusters of dolia are also observed. In this way, d1 and d2 show high similar total fatty acid levels of approximately 20-23mg/100g dry weight (Figure 3). These results for the archaeological samples are 50% of the total fatty acids of the allec reference — 52.16mg/100g dry weight. d3 and d6 show lower total fatty acids levels than d1 and d2 but are similar to each other, at 10.74 and 5.79mg/100g dry weight, respectively. Finally, the fatty acid contents detected in d4 and d5 are the lowest — 0.48-0.96mg/100g dry weight —, representing only 1-2% of the total fatty acids present in the allec reference.

3.2. Production method proposal The experimental recovery of garum sauce has been developed for several years from archaeological remains, archeometric, historiographic and experimental archaeology studies, involving multiple fields of knowledge. As a final result, a general method of making garum sauce has been proposed. Based on the contextualized recipe for making garum in Pompeii and all the historical information related to the products generated it is possible to describe a plausible method with intermediate stages and products.

Significant differences are observed in the fatty acid profile between the allec reference and archaeological remains. With respect to saturated fatty acids (SFA), the c14:0, c16:0 and c18:0 are the majority in the allec reference, while in the remains, c18:0 is not detected except for in d3 and d6. Moreover, the monounsaturated fatty acids (MUFAS) c16:1, c18:1 n-9, c20:1 n-9 and c24:1 n-9 are principally present in the allec reference, while some are not detected in the remains, with the exception of c18:1 n-6 in d3 and d6, and c20:1 and c24:1 in d1 and d2. In these latest dolia, the c20:1 and c24:1 content is higher than in the allec reference, which is remarkable given the age of the samples. This result could indicate that the garum — liquamen — and allec was not filtered.

3.2.1. Raw materials

3.1.3. Mineral composition

According to recipes, the spices used for made garum were very diverse. The ancients producers left us information about raw materials for the elaboration of garum, and spices were included, some exotic ones such as mustard or pepper, in addition to other Mediterranean herbs and

Thanks to classical bibliographic sources, a great variety of raw materials (García-Vargas et al. 2014), both aromatic herbs and spices, types of fish and other additional ingredients, have been determined for making garum. The most commonly named fish species were anchovy, sardine, mackerel, tuna, eel and red mullet. Regarding their size, it is highly variable, the archaeological evidence has shown the use of both young and adult fish (Bernal-Casasola et al. 2015; Bernal-Casasola et al. 2008). For the Pompeii garum fresh anchovies (E. encrasicholus) were used (Figure 5).

Figure 4 shows the mineral composition in the archaeological remnants, allec reference and fresh anchovy. As can be observed, Ca and P were the main minerals and represented D1

D2

D3

D4

D5

D6

Allec

Fresh anchovy

5.26 ± 2.32*b

53.64 ± 3.24*d

P

107.24 ± 12.32*a 155.40 ± 9.28*a 161.10 ± 7.45*a 152.16 ± 4.31*a

142.13 ± 10.92*a

14.28 ± 1.76e

Ca

232.59 ± 25.13*a 312.67 ± 21.43*a 313.46 ± 20.70*a 288.41 ± 22.64*a 238.35 ± 19.87*a 291.52 ± 21.12*a

28.77 ± 1.89b

Na

6.41 ± 0.87*

5.36 ± 1.16*

3.91 ± 0.48

3.22 ± 0.11*

144.67 ± 11.77

5.43 ± 1.41*b

Mg

6.65 ± 2.09*a

7.73 ± 1.88*a

1.87 ± 0.23b

3.18 ± 0.89*c

2.17± 0.73c

4.16 ± 1.10*c

2.16 ± 0.68d

2.08 ± 0.33d

0.99 ± 0.02*

1.14 ± 0.07*

a

K

3.26 ± 1.12*

0.95 ± 0.08*

5.67 ± 1.02

0.72 ± 0.04*

6.44 ± 1.17

13.64 ± 2.09*d

4.23 ± 1.69*

3.88 ± 1.77*

4.89 ± 1.38*

7.42 ± 1.92*

3.83 ± 1.07*

c

Fe

4.55 ± 1.72*

0.09 ± 0.01

0.11 ± 0.07c

Cu

0.06 ± 0.01a

0.06 ± 0.02a

0.05 ± 0.01a

0.07 ± 0.02a

0.05 ± 0.01a

0.04 ± 0.01a

0.60 ± 0.12b

0.02 ± 0.01a

Zn

0.48 ± 0.09

1.07 ± 0.19*

0.95 ± 0.07

0.80 ± 0.05

0.06 ± 0.01

0.79 ± 0.07

0.14 ± 0.03

0.14 ± 0.06c

a a

a

a

b a

b

b

b a

b

4.60 ± 0.34

96.86 ± 3.95*c

1.66 ± 0.17*

a

b

b a

b

c

b

Figure 4. Mineral profile.

174

a

b

b a

b

d

a

c

Making Garum. Experimental archaeology methods

Figure 5. Fresh anchovy (a) and spices (b). spices (LAMIACEAE, POACEAE and LAURACEAE). Some authors wrote that spices used were the typical ones of the area where the sauce was made (García-Vargas et al. 2014). Garum was made using thyme, oregano, savoury, coriander, celery, sage, poppy seeds, fennel, mint and rosemary (Figure 5). 3.2.2. Stacking. Static maceration The arrangement of the ingredients, selected based on archaeological evidence, was carried out, as described in the classic texts: disposing different layers of spices, fish and salt, until the container was filled and covered with a top layer of salt. The amount of fish used determines the amount of salt that must be sufficient to avoid a microbial contamination. It has been established that 5‑7% of the weight of the fish in spices is sufficient to form the spice layer and 15‑20% of the weight of the fish in salt, fulfills the function of salting the fish muscle and maintaining microbiological stability (Figure 6). Garum batches were prepared using anchovy (80%) salt (15%) and Mediterranean spices (5%), placed in layers — spices-fish-salt — inside a 5L thermostated glass vessel. The temperature was fixed at 50 ± 5°C to maximize the fish protease activity, and the vessels were sealed and kept in the dark to simulate an amphorae exposed to the sun. At the moment the layers are arranged, a massive migration of water begins from the body of the fish to the outside and the autolithyc process starts. The autolithyic process generates free amino acid from the fish proteins. 3.2.3. Stirring. Dynamic maceration After a few days, some changes in the organoleptic characteristics of the preparation was observed (Figure 7). Stirring the mix seek aiding the contact of fish protein with the enzyme-rich medium and controlling bacterial growth. At this moment, there was an identical amount of salt in the entire container. Salt controls microbial growth and allows maceration to continue for weeks or months until sauce was finished.

175

Figure 6. Raw materials (a) and fermentation (b) during static maceration.

Álvaro Rodríguez-Alcántara et al.

Figure 7. Final stage of static maceration (a) and first day of dynamic maceration phase (b). 3.2.4. Filtered in linen Once the mixture reaches the optimum point of flavor, it is filtered to obtain the fish sauce and the solid allec. The process is carried out using linen cloths (Figure 8), placed on funnels that allow the process to occur for as long as necessary. The filtering occurs through gravity, without excessive pressure. Remains of truncated conical vessels from the Phoenician-Punic and Roman times have been found and reconstructed for the modern production process. 3.2.5. The products: Garum and allec As a result of the filtering process, the unfiltered remains are obtained, the immiscible parts of the fish and spices. The fish sauce from the preparation is obtained after filtering (Figure 8). It can be treated as a finished product or undergo slight or significant modifications before consumption. Bibliography

Figure 8. Final steps (a), filter and garum samples (b).

AOAC 2000. Official Methods of Analysis. Arlington: AOAC international (17th ed.). Bekker-Nielsen, T. (ed.) 2005. Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2. Aarhus, Aarhus University Press. Bernal-Casasola, D., Díaz, J. J., Acqua, C., Arévalo, A., Bermejo, J., Bustamante, M., Cappelletto, E., JimenezCamino, R., Lagóstena, J., Lara, M., Lorenzo, L., Sáez, A. M., Vargas, J. and Villada, F. 2008. Memoria definitiva de la actividad arqueológica sistemática en Pompeya (regio V, ínsula 4, Cívico 3, Saggio 3). Año 2008. Unpublished report deposited at the Instituto del patrimonio cultural de España del Ministerio de Cultura and at the Universities of Cadiz and Venice. Bernal-Casasola, D., Cottica, D., Bustamante, M., Díaz, J. J., Expósito, J. A., García, E., Gómez, A., Landi, S., Lara, M., Lorenzo, L., Marlasca, R., Riquelme, J. A., Rodríguez, C. G., Sáez, A. M., Vargas, J. and Verdugo, J. 2013. Pesca y garum en Pompeya y Herculano. Cuarta campaña arqueológica. Informes y Trabajos 9, Excavaciones en el exterior: 322–339.

176

Making Garum. Experimental archaeology methods

Bernal-Casasola, D., Marlasca, R., Rodríguez-Santana, C.G., Ruiz-Zapata, B., Gil, M. J., García Vargas, E. and Alba, M. 2015. Garum de sardinas en Augusta Emerita. Caracterización arqueológica, epigráfica, ictiológica y palinológica del contenido de un ánfora Beltrán IIB. Rei Cretariae Romanae Fautorum Acta 44: 737–749. Botte, E. 2009: Salaisons et sauces de poissons en Italie du sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Carannante, A., Giardino, C. and Savarese, U. 2011. In search of garum. The ‘Colatura d’alici’ from the amalfitan coast (Campania, Italy): an heir of the ancient Mediterranean fish sauces. In F. Lugli, A. A.Stoppiello and S. Biagetti (eds). Atti del 4° Convegno Nazionale di Etnoarcheologia, Roma, 17-19 maggio 2006 / Proceedings of the 4th Italian Congress of Ethnoarchaeology, Rome, 17-19 May 2006. British Archaeological Reports International Series 2235: 69–79. Oxford, Archaeopress. Curtis, R. I. 1979. The garum shop of Pompeii (I.12.8). Cronache Pompeiane 5: 5–23. Curtis, R. I. 1991. Garum and salsamenta. Production and commerce in materia medica. Studies in ancient medicine 3. Leiden - New York - Copenhaguen - Cologne, Brill. Curtis, R. I. 2009. Umami and the foods of classical antiquity. American Journal of Clinical Nutrition 90: 712–718. Desse-Berset, N. and Desse, J. 2000. Salsamenta, garum et autres préparations de poisson. Ce qu’en disent les os. Mélanges de l’École française de Rome. Antiquité 112-1: 73–97. Dieffenbacher, A. and Pockilngton, W. D. 1992. International Union of Pure and Applied Chemistry (iupac), Applied Chemistry Division. Commission on Oils, Fats and Derivatives. Standard methods for the analysis of oils, fats and derivatives. 1st suppl. to the 7th edition. London, Blackwell Scientific Publications. Étienne, R. and Mayet, Fr. 2002. Salaisons et sauces de poisson hispaniques. Trois clés pour l’économie de l’Hispanie romaine, II. Paris, de Boccard. Evershed, R. P. 1993. Biomolecular archaeology and lipids. World Archaeology 25 (1): 74–93.

Evershed, R. P., Copley, M. S., Dickson, L. and Hansel, F.A 2008. Experimental evidence for the processing of marine animal products and other commodities containing polyunsaturated fatty acids in pottery vessels. Archaeometry 50: 101–113. García Vargas, E., Bernal-Casasola, D., Palacios Macías, V., Roldán Gómez, A. M., Rodríguez Alcántara, A. and Sánchez García, J. 2014. Confectio Gari Pompeiani. Procedimiento experimental para la elaboración de salsas de pescado romanas. SPAL, Revista de Prehistoria y Arqueología de la Universidad de Sevilla 23: 65–82. Grimal, P. and Monod T. 1952. Sur la véritable nature du garum, Revue des études anciennes LIV: 27–38. Gurioli, L., Houghton, B. F., Cashman, K. V. and Cioni, R. 2005. Complex changes in eruption dynamics during the 79 AD eruption of Vesuvius. Bulletin of Volcanology 67 (2): 144–159. Luongo, G., Perrotta, A., Scarpati, C., De Carolis, E., Patricelli, G. and Ciarallo, A. 2003. Impact of the AD 79 explosive eruption on Pompeii, II. Causes of death of the inhabitants inferred by stratigraphic analysis and areal distribution of the human casualties. Journal of Volcanology and Geothermal Research 126 (3-4): 169–200. Palacios Macías, V., García Vargas, E., Bernal-Casasola, D., Roldán Gomez, A. M., Rodríguez Alcantara A. and Sánchez García J. 2016. Conservas antiguas y gastronomía contemporánea. In D. Bernal-Casasola, J. Á. Esposito Álvarez, L. Medina Grande and J. S. Vicente-Franqueira García (eds), Un Estrecho de Conservas. Del Garum de Baelo Claudia a la melva de Tarifa: 89–105. Cadiz, Universidad de Cádiz. Ponsich, M. and Tarradell, M. 1965. Garum et industries antiques de salaison dans la Méditerranée occidentale. Bibliothèque de l’École des Hautes Études Hispaniques 36. Paris, Presses Universitaires de France. Reese, D. S. 2002. Fish: evidence of specimens, mosaics, wall painting, and Roman authors. In W. F. Jashemski and F. G. Meyer (eds), The Natural History of Pompeii: 274–291. Cambridge, Cambridge University Press. Smriga, M., Mizukoshi, T., Iwahata, D., Eto, S., Miyano, H., Kimura, T. and Curtis, R. I. 2010. Amino acids and minerals in ancient remnants of fish sauce (garum) sampled in the ‘Garum Shop’ of Pompeii, Italy. Journal of food composition and analysis 23 (5): 442–446.

177

Regional and provincial syntheses

West

Lusitanian Amphora Contents Inês Vaz Pinto

Troia Resort / CEAACP – Centro de Estudos de Arqueologia, Artes e Ciências do Património da Universidade de Coimbra

Rui Morais

FLUP/UI&D CECH / Faculdade de Letras da Universidade do Porto, Portugal

Carlos Fabião

UNIARQ - Centro de Arqueologia da Universidade de Lisboa. Faculdade de Letras. Universidade de Lisboa. Portugal

César Oliveira

Laboratório HERCULES, Évora, Portugal LAB2PT, Braga, Portugal Departamento de Ciências e Técnicas do Património da Universidade do Porto, Portugal

Sónia Gabriel

Laboratório de Arqueociências, Direcção Geral do Património Cultural (DGPC) CIBIO - InBIO, Laboratório Associado Centro de Investigação em Biodiversidade e Recursos Genéticos UNIARQ - Centro de Arqueologia da Universidade de Lisboa

Abstract: Roman amphorae produced in Lusitania were usually considered as fish product containers since amphora kilns were mainly known on the coast where they co-existed with fish-salting workshops, and the rare tituli picti known confirmed this assumption. Over the past 25 years the progress in research has brought up new issues: the contents of two flat-base types was questioned and wine was the new content proposed for them; a number of amphora types from the second half of the 1st century BC – early 1st century AD, earlier than the known fish-salting installations, were identified; and different sets of ceramic material from inland sites, some very far from the coast, revealed amphorae in local fabrics that cannot be for fish products. Recent technical advances on content analysis, such as gas chromatography, are an opportunity to clarify contents, and have begun to be applied to Lusitanian amphorae, and the published results are presented. The products likely to be carried in amphorae will also be considered. The contents of Lusitanian amphorae will be discussed relying on existing evidence such as typology, provenance, epigraphy, fish remains and chemical analysis when possible. Key words: Amphora contents; fish product amphorae; wine amphorae; fish remains; tituli picti; garum; liquamen; muria; hallec.

1. Introduction Lusitanian amphorae are certainly, as are amphorae in general, the best indicator of the circulation and trade of foodstuffs in Lusitania, and from this province to the capital and other provinces of the Roman Empire. But did they carry just the fish products from the coastal location of their kilns, as suggested by their Baetican counterparts? Are there arguments to consider some wine amphorae? And what kind of fish products might have been carried in these amphorae? Classical authors are not of great help in this matter since they rarely refer to Lusitanian products and exports. Tituli picti have given us precious information for at least a century, but only a few have survived and been identified. Some decades ago, shipwrecks came to be a valuable capsule of information, often with well‑preserved

amphorae, sometimes with organic remains, but only a few with confirmed Lusitanian amphorae have been consistently studied and their organic remains analysed. More recently chemical analysis of amphora residues came to determine the nature of the content carried in them, but only a few Lusitanian amphorae have been subjected to this kind of study. Progress has been very slow but the last few decades have produced new data, not only new amphorae with contents, but also the products likely to be carried in amphorae, which can be inferred from the fish remains in the fish-salting units. In this study the available evidence on Lusitanian amphora contents will be considered and discussed, and a reassessment of the principal content of the main types of amphorae will be attempted.

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 183–204

Inês Vaz Pinto, Rui Morais, Carlos Fabião, César Oliveira and Sónia Gabriel

2. Earlier approaches to the contents of Lusitanian amphorae The amphorae produced in the Roman province of Lusitania were traditionally considered as fish product containers since early modern scientific approaches did not offer any clues about what the amphora could transport. In fact, the classical and most abundant types such as the Dressel 14 for the Early Empire, and the Almagro 51c for Late Antiquity were mass-produced in the same pottery production centres located near fish-salting production centres in Western Lusitania (Figure 1). This is the case with Sado Valley’s amphora workshops, such as Pinheiro (Mayet and Silva 1998) and Abul (Mayet and Silva 2002) and there are at least seven more (Fabião 2004; Almeida et al. 2014a), located on the right bank of the river and also along its estuary, a short boat trip from Troia, which is a large fish-salting production centre with 27 workshops identified (two of them still unpublished) (Pinto, Magalhães and Brum 2011; Pinto, Magalhães and Brum 2014; Pinto, Magalhães and Brum 2017). Other fish-salting installations in the region, such as those in Setúbal (Travessa de Frei Gaspar) (Silva, Coelho-Soares and Soares 1986) and Praça do Bocage (Silva and Soares 1980-81), Comenda (Silva and Soares 1986: 194), Creiro (Silva and Soares 1987; Silva and Coelho-Soares 2016) and Sesimbra (Pereira 2014) revealed that the low Sado area and the surrounding region encompassed a very large fish product production area and it seems likely that the amphorae from the regional potteries would have been used to export those fish products. The Tagus Valley, nowadays more urbanized and difficult for research, revealed the potteries of Garrocheira, Porto dos Cacos and Quinta do Rouxinol (Fabião 2008), on the south bank of the river and known since the 1980s. These production sites present the same pattern of amphora production (with the same forms) as those from the Sado Valley workshops, considered as containers for fish products. The fish-salting installations recorded in Lisbon (Bugalhão 2001), in Belém (Filipe and Fabião 2007) but also in Cascais (Cardoso 2006) and on the left bank of the river, in Cacilhas (Barros and Amaro 1984-85) and Porto Brandão (Santos, Sabrosa and Gouveia 1996), suggest that this may have been a very active production area for fish products. In the Algarve, in Southern Lusitania, the evidence of amphora and fish-salting production is rather scarce in the 1st and 2nd centuries, with only two amphora production sites known, Castro Marim and Manta Rota. These activities developed from the 3rd to the 5th century, apparently never reaching the importance they had in Western Lusitania, but the amphora kilns and the fishsalting factories were widely distributed along the coast in a clear association, and the same generic amphora types were present, the Almagro 51c, Almagro 50 and a

variant of the Almagro 51a-b, the Algarve 1, being the most significant and also considered as fish product amphorae (Bernardes and Viegas 2016). The area of Peniche, also on the coast, was the northernmost in Lusitania with the most recent finds of amphora kilns — in Morraçal da Ajuda. The site is Augustan and produced vessels mostly inspired by Baetican fish amphorae and therefore considered as such (Cardoso et al. 2016). The rare tituli picti and fish remains related to these amphorae solidify this assumption and in the first meeting dedicated to Lusitanian amphorae held in Conimbriga in 1988 (Alarcão and Mayet 1990), R. Étienne (1990) pointed out that the main three types of Lusitanian amphorae, Dressel 14, Almagro 50 and Almagro 51c, were used to carry fish products: the Dressel 14 amphora carried fish sauces (proved by the inscriptions liquamen and muria found by H. Dressel at Castro Pretorio in Rome) and salsamenta (as suggested by fish cut into pieces in an amphora retrieved from the Gandolfo wreck); the Almagro 50 also carried both salsamenta (Port-Vendres I wreck) and sauces (Planier VII wreck); the fish contents of the Almagro 51c relied only on some bones in a Lusitanian amphora from the Cabrera III wreck (Figure 2). This evidence was complemented by the Almagro 50 amphorae of the Randello shipwreck, (Sicily) full of whole sardine, and of the Sud-Lavezzi 1, probably containing mackerel, mentioned in the first article about Lusitanian amphora contents published in 1993 (Fabião and Guerra 1993). A new summary on the subject presented in 2002 refers to the Cala Reale A shipwreck (Sardinia) with Almagro 51ab and Beltrán 72 with whole sardine (Étienne and Mayet 2002), these new data reinforcing the fish product character of Lusitanian amphorae. At the Conimbriga meeting, in 1988, Dias Diogo suggested a wine content for one specific type of Lusitanian amphora, the so-called Lusitana 3 (Alarcão and Mayet 1990: 189). The typology of Lusitanian amphorae this author published in 1991 (with the publication date of 1987) included the Lusitana 3 type, a small flat-based amphora copying the Gauloise 4 type, presented as a wine amphora. Although a controversial subject, and not recognized as such in the Sado Valley workshops, where this type is considered as the precursor to the fish-product amphora Almagro 51c – Almagro 51c, Variant A (Mayet and Silva 1998), its wine contents is starting to be accepted, and awaits scientific analysis for a categorical identification. The importance of the fish product amphorae and the possibility of a wine amphora (Lusitana 3) is reinforced in the cited article on amphora contents published a few years later by C. Fabião and A. Guerra (1993). In 1998, C. Fabião suggests the Lusitana 3 is a wine amphora and considers

184

Lusitanian Amphora Contents

Figure 1. Distribution of amphora kilns and cetariae. 1: Morraçal da Ajuda, Peniche; 2: Cascais; 3: Lisboa (several); 4: Garrocheira; 5: Porto dos Cacos; 6: Cacilhas and Porto Brandão; 7: Quinta do Rouxinol; 8: Sesimbra; 9: Creiro; 10: Rasca and Comenda; 11: Setúbal (several); 12: Zambujalinho; 13: Quinta da Alegria; 14: Tróia; 15: Pinheiro; 16: Abul; 17: Enchurrasqueira; 18: Bugio; 19: Parvoíce, Alcácer do Sal; 20: Barrosinha; 21: Sines; 22: Ilha do Pessegueiro; 23: Beliche; 24: Martinhal; 25: Salema; 26: Boca do Rio; 27: Burgau; 28: Praia da Luz; 29: Lagos (several); 30: Monte Molião; 31: Praia do Vau; 32: Portimões and Portimão; 33: Praia de Aveiros; 34: Praia de Santa Eulália; 35: Cerro da Vila; 36: Loulé Velho; 37: Salgados; 38: Quinta do Lago; 39: São João da Venda; 40: Faro; 41: Olhão; 42: Quinta do Marim; 43: Torre de Ares, Quinta das Antas and Pedras del Rei; 44: Quinta do Muro; 45: Manta Rota; 46: São Bartolomeu de Castro Marim.

185

Inês Vaz Pinto, Rui Morais, Carlos Fabião, César Oliveira and Sónia Gabriel

Figure 2. Shipwrecks with Lusitanian amphorae with content remains (based on Bombico 2017: Anexo II): 1: San Antonio Abad / Grum de Sal (Ibiza); 2: Cabrera III; 3: Port-Vendres I; 4: Dump above Arles-Rhône 3; 5: Anse SaintGervais (Fos-sur-Mer); 6: Punta Vecchia I; 7: Sud-Lavezzi 1; 8: Cala Reale A; 9: Randello. Lusitana 9, a small barrel-shaped amphora (Diogo 1987), as another wine amphora, this one from the 3rd-5th century, which would have replaced the first one, produced from the late 1st until the early 3th century. They were both believed to be for local or regional distribution in the province of Lusitania, but the progress of research has shown that Lusitana 3 reached Baetica in significant quantities and dominated the imports in the lower Guadalquivir Valley between the late 2nd and the mid 3th century AD (García Vargas 2016) and even the small Lusitana 9 has been identified in Baetica (Bernal‑Casasola 2016a: 304) and also in other inland places, such as the Lusitanian town of Ammaia (Venditti 2016) or the villae of São Cucufate (Vidigueira) (Pinto and Lopes 2006). C. Fabião (1998: 187) also drew attention to a Lusitanian version of the wine amphora Dressel 28, possibly produced in the Tagus Valley (Raposo 1990: 127) and produced in the workshop of Pinheiro (Fabião 2008: 740), but it is so rare in production and consumption contexts that it is perhaps a rather insignificant type (Almeida et al. 2014b: 657). In recent years, other amphora types of small size identified in the cities of Mérida and Conimbriga, located in inland regions with an agricultural vocation, were assumed to be wine amphorae. For instance the Haltern 70

produced in Augusta Emerita (Bustamante and Cordero 2013; Bustamante and Heras 2016) and the Class 45‑46 found in consumption contexts in Conimbriga (Correia et al. 2015a and b). This new amphora from Conimbriga was subjected to gas chromatography analysis whose results proved the assumption to be correct. Once again, these small amphora types were probably chiefly used for local/regional distribution and consumption. But Pliny the Elder mentioned a much appreciated type of olive from Augusta Emerita (NH 15.17). One can assume that those olives where known and sought by Roman elites, who travelled abroad to other regions of the Roman Empire. So the use of a small amphora with the shape of a Haltern 70 amphora type, known to transport preserved olives, could give us a suggestion about the product that could have been transported in these amphorae. Obviously, this is just a hypothesis that needs to be confirmed by further investigations. The small amphorae from Conimbriga could suggest that other small containers could be used for local/regional distribution of foodstuffs. Over the past 15 years, early types of amphorae from the second half of the 1st century BC‑early 1st century AD imitating Baetican models and of the so-called ‘ovoid type,’ with fabrics typical of the Sado and Tagus valleys,

186

Lusitanian Amphora Contents

have been progressively identified in the territory of Lusitania in Augustan times (Morais 2004; Morais and Fabião 2007). Due to the very probable areas of production near marine resources, they are assumed to be for fish products even though no contemporary fish-salting installations have been identified so far. Most likely based on Italian and Baetican forms, these amphorae are varied, but have similar chronologies of production. Generally, they are characterized by having short necks and handles, with an identical oval or more or less cylindrical body, similar to the known Baetican amphorae Haltern 70 of earlier production, including the so-called ‘small variant’ or Ovoid type 4 from the Guadalquivir region (Almeida 2008; García Vargas, Almeida and González Cesteros 2011). Some fragments show formal characteristics that seem to fit in the so-called Lusitana 12 type from the typology of Dias Diogo (1987: fig. 7) and are also very similar to the first productions of Dressel 14, variant A of F. Mayet and C. Tavares da Silva (2002). According to the fabrics, these earlier productions must come from the pottery workshops located along the Tagus and Sado Rivers, with a probable prevalence of productions from the Sado Valley (Mayet and Silva 2016; Pimenta, Ferreira and Cabrita 2016). They also come from the Augustan pottery workshop in Morraçal da Ajuda (Peniche) (Cardoso and Rodrigues 2005; Cardoso, Rodrigues and Sepúlveda 2006; Cardoso et al. 2016) that produced amphorae similar to the Dressel 7-11. We should underline here that we are only concerned with Roman amphora production: however, it is interesting to note that there are Iron Age westernmost amphora productions, with so-called pre-Roman shapes, which have been identified in several pre-Roman archaeological contexts, chiefly in the Tagus estuary (Sousa and Pimenta 2014). The local ‘Tagan’ production was identified by its fabric, similar to the Roman local amphorae, but actually, no production centres were yet identified. But if the westernmost area of the Iberian Peninsula, the area that was later the Provincia Hispania Ulterior and then Hispania Ulterior Lusitania, has its local tradition of amphorae production since the Iron Age, one could expect a specific local dynamic not so different from those known in other areas of the Iberian Peninsula with a similar history. In other words, we don’t yet know about the contribution of local craftsmen to the Lusitanian Roman amphora production nor if there is some continuity from pre-Roman to Roman production and its chronology. The results of the analyses on the organic contents found in the studied amphorae, including those earlier ones of ovoid typology that we have been highlighting, revealed fish content (Oliveira, Morais and Araújo 2015: 197). These results are consistent with the data known through archaeology about these amphorae in which there is a correlation between their places of production and the fish-salting factories.

Due to the significance of fish contents in Lusitanian amphorae, and the interest in interpreting their archaeological remains, it is important to mention the different possible fish products carried in amphorae. This subject that has been broadly studied in the past decades (Curtis 1991; Étienne and Mayet 2002; Botte and Leitch 2014 among many others) but due to the scarce, unclear and sometimes contradictory literary sources, and possibly a slight change of the meaning of the terms over time, remains a challenging study. To offer a brief summary of a very complex matter, classical literature and archaeological evidence distinguish salted fish (salsamenta) from fish sauces. Salted fish would have been processed whole or in chunks, with or without bones depending on fish size. In the case of small fish, like sardines or anchovy, they would be salted whole, including flesh and bones. Sauce production is more complex. Literary sources and painted inscriptions on amphorae (tituli picti) mention four main types: garum, liquamen, muria and hallec. According to the classical authors, garum and liquamen were filtered sauces from the maceration of blood, viscera and/or small fish in salt (Geoponica 20.46 apud Étienne and Mayet 2002: 44), therefore liquid and without residue (or very little residue). Muria would also be a liquid sauce, apparently a sort of fish brine, probably without previous maceration of the fish components (Étienne and Mayet 2002: 47–48). Hallec could be either the solid residue of garum and liquamen (Pliny, NH 31.95 and Geoponica 20.46 apud Étienne and Mayet 2002: 44) or a sort of paste made from very small fish, becoming a refined product with a number of varieties according to Pliny (NH 31.95). The fish sauces are very disparate both in composition and consistency and so it is not easy to tell from the fish remains within the vats or inside amphorae, what sort of sauce is made or transported. In amphorae, scattered bones of very small fish could be remains of hallec, and bones of whole large fish should be an indication of salsamenta while garum, liquamen and muria should not leave residues, or only very few from a defective filtering. But is is often not possible to infer the type of product from the zooarchaeological data. 3. Evidence of Lusitanian amphora contents The evidence of amphora contents will be analysed in detail, namely the information that can be obtained from tituli picti, macro-remains and chemical analysis of residue. This rather scarce information, in the case of fish content, will be compared to the remains of fish products collected in Lusitanian fish-salting factories. 3.1. Tituli picti Tituli picti with a reference to contents are rather rare on Lusitania amphorae, and they all appear on Dressel 14 amphorae (Figure 3).

187

Inês Vaz Pinto, Rui Morais, Carlos Fabião, César Oliveira and Sónia Gabriel

Reference to content LIQUA/MEN LIQ(uamen) LIQ(uamen)

Amphora type Dressel 14 Dressel 14 Dressel 14 parva

Provenience Rome, Castro Pretorio Rome, Castro Pretorio Dump above Arles-Rhône 3, no.6 Dump above Arles-Rhône 3, LIQ(uamen) Dressel 14 parva (Tagus/Sado) no.28 LIQ(uamen) EXC(ellens) Dressel 14 Anse Saint-Gervais (Fos-sur-Mer) LIQUAMEN/FLOS/EXC(ellens) Dressel 14, Variant A (Tagus/Sado) Dump above Arles-Rhône 3, no.19 LIQ(uaminis) FL(os) Dressel 14 Rome, Domus Aurea EXCELLENS MURI(A) Dressel 14 Rome LAC(ertus) Dressel 14 parva (Tagus/Sado) Dump above Arles-Rhône 3, no.27

Bibliographical Reference CIL XV: 4716 CIL XV: 4714 Djaoui and Quaresma 2016: 361 Quillon 2011: 109; Djaoui and Quaresma 2016: 361 Liou and Marichal 1978 Djaoui and Quaresma 2016: 361. CIL XV: 4719 CIL XV: 479 Djaoui and Quaresma 2016: 361

Figure 3. Tituli picti with a reference to content in Lusitanian amphorae. Some were found in Rome and were published by H. Dressel, one comes from Anse Saint-Gervais (Fos-surmer) (Figure 4) and the others come from the river site Arles-Rhône 3. Most of them refer to liquamen, either fully written or abbreviated. In two of them liquamen is qualified as excellens, and in two others it is specified as liquamen flos excellens. According to Pliny (NH 31.93–95), gari flos is a garum without condiments, a virgin garum, and the same must be true of liquamen. Just one example from Rome presents a painted inscription referring to muria, and recently the titulus pictus LAC appeared on the neck of a Lusitanian Dressel 14 parva from the river site Arles-Rhône 3 (Djaoui 2016) (Figure 5). The study of this titulus pictus by D. Djaoui concluded that LAC must be developed as LAC(ertus) and lacertus is a small mackerel, solving the question of the enigmatic laccatum sauce (Étienne and Mayet 2002: 52–53). Moreover, the position of LAC in the second line of the titulus pictus suggests it characterizes the type of product unreadable in the first line, most probably liquamen.

Figure 5. Neck of a Dressel 14 parva amphora with the titulus pictus LAC[--] (photograph L. Roux/ CG13; drawing A. Veleva; vectorized by D. Djaoui) (Djaoui 2016: fig. 1). This amphora thus carried a fish product like small mackerel, most probably liquamen, the most common sauce mentioned in tituli picti on Dressel 14.

Figure 4. Titulus pictus on a Dressel 14 amphora from Anse de Saint-Gervais: LIQ(uamen)/EXC(ellens)/SABINI ET AVITI (Liou and Marichal 1978: fig.16, n. 32 apud Bombico 2017: fig. 75).

No titulus pictus refers to garum, the most well-known sauce. Was liquamen an alternative designation of garum? In the 1st century AD garum and liquamen appeared side by side, for example in the amphorae of the Castro Pretorio, in Rome, and in the house of Aulus Umbricius Scaurus, in Pompeii, meaning they were perceived as different products in that period. In the early 4th century, in Diocletian’s price edict, the Latin text uses the term liquamen for fish sauce, translated as garos in the Greek text (Grainger 2014: 40), meaning the difference between the two had vanished at that time. Recently S. Grainger

188

Lusitanian Amphora Contents

Type

Dressel 14

Fish remains Sardina pilchardus (sardine: 17-18cm) Whole fish Sardina pilchardus (sardine 16-21cm) Whole fish Sardina pilchardus (sardine: 20-26cm) -Whole fish Fish bones

Almagro 51c

Almagro 50

Almagro 50 or Sado 1/ Keay 78 Almagro 51a-b Beltrán 72 similis

Fish bones (undet.) Sardina pilchardus (sardine: 11-18cm) Whole fish

Provenience

Date

Bibliographical Reference

Setúbal (R. Joaquim António Granjo, 19)

second half 1st c.

Gabriel and Silva 2016

Setúbal (R. Francisco Augusto Flamengo, 10-12)

end 1st c.-2nd c.

Gabriel and Silva 2016

Tróia (Workshop 2, vat 9) second half 2nd c. San Antonio Abad / Grum de Sal shipwreck (Ibiza) Cabrera III (Mallorca) shipwreck

second half 1st c. - mid 2nd c.

Vilar-Sancho and Mañá 1964 apud Bombico 2017

AD 257

Étienne 1990: 17

Randello (Sicily) shipwreck

3rd-4th c.

Sardina pilchardus (sardine: 22-25cm) -Whole fish

Port-Vendres I (Anse Gerbal, France) shipwreck

end 4th c.-early 5th c.

Fish bones, possibly mackerel Sardina pilchardus -Whole with scales Amphora 894: a few bones of sardine, labridae, sparidae (?) Sardina pilchardus -Whole with scales

Sud-Lavezzi I (Bonifacio, end 4th c.-mid 5th c. Corsica) shipwreck Cala Reale A (Sardinia) shipwreck

Gabriel 2012

end 4th c.-mid 5th c.

Wheeler and Locker 1985; Parker 1989; Desse and Desse 2000: 93 Chevalier and Santamaria 1971: 32 apud Desse and Desse 2000: 92 Liou 1982: 438-440 Spanu 1997; Gasperetti 2012; Delussu and Wilkens 2000: 60-61, 65

Figure 6. Fish remains from Lusitanian amphorae. (2014) proposed that the difference between the two was that garum was the liquid sauce made of blood and viscera, a luxury table item, while liquamen was the liquid sauce made of small fish and used for cooking, which would explain the constant references in the cookbook of Apicius. What matters for Lusitanian amphora contents is that the few known tituli picti refer to the 1st-2nd centuries, and liquamen may have been used as a general term for a filtered sauce or, if S. Grainger is right, it referred to the liquid sauce made out of small sardines, whose leftovers are so common in the Lusitanian fish-salting cetariae. Tituli picti only provide information on the Dressel 14 amphora content and only two types of sauce are mentioned, liquamen and muria, and only one type of fish, small mackerel (LAC(ertus)), probably referring to a type of sauce. Sardines might have been the common base of liquamen but it was not necessary to specify this. 3.2. Macro remains Not many Lusitanian amphorae have revealed macro remains in their contents. Fish bones are the only type of macro remains identified in Dressel 14, Almagro 50 (and possibly Sado 1/Keay 78), Almagro 51c, Almagro 51a-b and Beltrán 72 similis coming both from inland contexts and from shipwrecks (Figure 6).

Three samples of fish remains from Dressel 14 amphorae found in inland sites on the riverbanks of the Sado estuary were studied. The first one, the earliest found in Lusitania, comes from the lower part of a Dressel 14 of undetermined variant found in Setúbal (Rua António Joaquim Granjo, 19), in a context dating from the second half of the 1st century AD (Gabriel and Silva 2016). It was composed of sardines (Sardina pilchardus) 17-18cm long, with the presence of all bones but scattered and in low frequency, with sand being dominant in the sample (Figure 7). It is not possible to know if these remains result from salsamenta or if they are the residue of a fish sauce, but they are the earliest example of sardine inside a Lusitanian amphora. Another sample from Setúbal, found in 10-12, Rua Francisco Augusto Flamengo, inside a Dressel 14, Variant C, comes from a rubbish pit that accumulated from the end of the 1st century to the 2nd century. It was composed of massive numbers of superimposed bones in articulation (mostly cranial bones and vertebrae) of sardines 16-21cm long (Figure 7), suggesting they derive from salsamenta manufactured using whole fish (Gabriel and Silva 2016: 115). The third sample is from the second half of the 2nd century AD and was collected in Tróia. Even though it was found inside a vat (vat 9 of workshop 2), it is not a residue of

189

Inês Vaz Pinto, Rui Morais, Carlos Fabião, César Oliveira and Sónia Gabriel

Figure 7. Physical appearance of the fish bone samples recovered at Rua Francisco Augusto Flamengo (RFAF10‑12) and Rua António Joaquim Granjo (RAJG-19), Setúbal (Gabriel and Tavares da Silva 2016: fig. 3). Photographic credits: José Paulo Ruas – DGPC. fish-salting production, but a concentration of fish bones attached to a large fragment of Dressel 14 in the context of a rubbish pit (Pinto, Magalhães and Brum 2010) and it was considered as the contents of the broken amphora. The remains consisted of large numbers of vertebral centra and head bones, thus indicating that whole fish had been contained in the amphora. Based on the minimum number of individuals (MNI) sardines represent 91% (MNI= 29) of the fish present in the sample that also includes jack mackerel (Trachurus trachurus, MNI= 1), seabream (Diplodus sp., MNI= 1) and mackerel (Scomber sp., MNI= 1). The sardines were 20-26cm long, the longest so far found in a Lusitanian amphora, suggesting salted sardine (salsamenta). Only one shipwreck has revealed Lusitanian Dressel 14 with contents. It is the San Antonio Abad/Grum de Sal shipwreck, from the second half of the 1st century to the mid 2th century, in which the majority of the 67 amphorae recovered were Dressel 14 reported as containing fish bones (Vilar-Sancho and Mañá 1964 apud Bombico 2017: ficha 10). R. Étienne stated that the Dressel 14 amphorae from the Gandolfo shipwreck contained fish cut up into pieces, but those are Baetican amphorae, one of them with a titulus pictus from Sexi (Almuñecar) (Pascual Guash 1968; Liou and Rodríguez Almeida 2000; Bombico 2017: ficha 2).

Contents in Almagro 50 amphorae were found in shipwrecks such as Randello, Port-Vendres I and SudLavezzi 1. The Randello (Ragusa, south of Camarina, Sicily) shipwreck, dating to the 3rd-4rd century, transported Lusitanian Almagro 50 amphorae (Parker 1989; Bombico 2017: ficha 57) carrying Sardina pilchardus c. 11‑18cm long studied by A. Wheeler and A. Locker (1985). Does the presence of whole individuals indicate salted sardines or could it be the residue of a fish paste like hallec? The Port-Vendres I (anse Gerbal) shipwreck, from the end of the 4th century - beginning of the 5th (Liou 1974: 427–428) revealed one Almagro 50 amphora, very likely Lusitanian (Bombico 2017: ficha 19), containing Sardina pilchardus, mostly 22-25cm long and interpreted as salsamenta (Desse-Berset and Desse 2000). In the Sud-Lavezzi I shipwreck (Bonifacio, Southern Corse), amphora no.8, representative of a group of three amphorae, and possibly not an Almagro 50 but a Keay 78/ Sado 1 amphora (due to the body and base shape, but the drawing is not conclusive), was full of fish bones, possibly mackerel (Liou 1982: 438–440).

Therefore, according to the finds in Setúbal and Tróia, the Dressel 14 amphora, with several tituli picti indicating liquamen and muria, also carried whole sardines (salsamenta), pointing to the versatility of this amphora type.

R. Étienne (1990) also suggested that the Almagro 50 amphorae from the Planier VII (anse Gerbal) shipwreck contained pectunculus pilosus shells interpreted as an indication of garum. This shipwreck carried other Lusitanian amphorae, but the Almagro 50 were considered as amphores à saumure de Bétique in the original publication, a generally accepted fact (Bombico 2017: ficha 23).

The Almagro 51c, the most abundant Lusitanian amphora from the 3rd to the 5th century AD, only has one instance of bones inside, in an amphora from the Cabrera III shipwreck, according to oral information from R. Étienne (1990: 17).

The whole sardines carried by these Lusitanian Almagro 50, and possibly Sado 1/Keay 78 in one case, indicate these amphorae did not carry a liquid fish sauce like garum, liquamen or muria, but they had a solid content,

190

Lusitanian Amphora Contents

Figure 8. Lusitanian amphorae studied by chemical analysis: a) ovoid amphora from Castro de Vigo, Pontevedra, Spain; b) amphora from Peniche, Peniche type 1 (Braga, Cavalariças, no. 1997-1031); c) Dressel 14 parva amphora from Albergue Distrital at Braga (no.1997-0451); d) Dressel 14 amphora (or ovoid type) from Albergue Distrital at Braga (no.1994-0315); e) class Conimbriga 45-46 - no.1 – 71.CRP.N.(cano) and f) class Conimbriga 45-46 - no.11 – 72.B.F6 (10). probably salted fish (salsamenta) in the case of the PortVendres 1 and Sud-Lavezzi I amphorae, and salsamenta or hallec in the case of the Randello amphorae with medium sized sardines. The Cala Reale A shipwreck is of great interest since it carried a large cargo (with 1866 amphora bases counted) of Lusitanian amphorae of four types: Sado 3 (956 rims), Almagro 51a-b (625 rims), Beltrán 72 similis (156 rims) and Almagro 51c (64 rims) (Gasperetti 2012; Bombico 2017: ficha 42). The Sado 3 and Almagro 51c did not reveal any content while six whole amphorae of Almagro 51a‑b and Beltrán 72 type contained whole sardines with scales. One Almagro 51a-b with a narrower neck only revealed a few vertebrae from sardines, a bone of labridae, and one of labridae or sparidae: the archaeoichtyologists (Delussu and Wilkens 2000) suggest an intrusion or a different product explain this. Possibly two amphorae of the same type but with different proportions may have carried different products. The wreck was dated from the end of the 4th to the mid 5th century (Spanu 1997), and in our opinion the reduced number of Almagro 51c and the large number of Sado 3 and Almagro 51a-b, variant C point to the latest moment of that time span.

These are the first Lusitanian Almagro 51a-b and Beltrán 72 similis with content, a solid content, except for one individual of Almagro 51a-b, but unfortunately the sardine size is not estimated and therefore it is not possible to hypothesize what kind of solid product they carried or compare the contents of the two amphora types. 3.3. Chemical analysis Chemical analyses are not yet a common method to determine content in Lusitanian amphorae. The few results come only from analyses performed at the Chemistry Centre of Minho University (Braga, Portugal) and from a short reference to the amphorae from the Punta Vecchia I (Cape Corse, Corsica) that we could not clarify. Several different types of amphorae were subjected to chemical analysis by Gas Chromatography, coupled with Mass Spectrometry (GC/MS) to determine their content (Figure 8). Six different amphorae were studied at the Chemistry Centre of Minho University (Braga, Portugal): an ovoid amphora (similar to Lusitanian 12), an amphora from Peniche, Peniche type 1, collected at Braga (Cavalariças, no.1997-1031), two Dressel 14 amphorae (one parva) from Braga (Albergue Distrital, no.1997-0451

191

Inês Vaz Pinto, Rui Morais, Carlos Fabião, César Oliveira and Sónia Gabriel

Amphora type Lusitanian Ovoid Amphorae (similar to Lusitana 12) Lusitanian amphorae from Peniche (similar to Dressel 7-11) Dressel 14

Provenience

Chromatography Fish products sweetened Western Lusitania: Sado-Tagus probably with honey or sweet Valleys wort Fish products sweetened Western Lusitania: Peniche probably with honey or sweet wort Fish products sweetened Western Lusitania: Sado-Tagus probably with honey or sweet Valleys wort

Class Conimbriga 45-46 - no.1 – Central Lusitania: Conimbriga 71.CRP.N. (cano) Class Conimbriga 45-46 - no.11 Central Lusitania: Conimbriga – 72.B.F6 (10) Almagro 51c

Punta Vecchia 1 (Corsica) shipwreck

Bibliographical Reference Morais, Oliveira and Araújo 2016 Morais, Oliveira and Araújo 2016 Morais, Oliveira and Araújo 2016

Wine

Correia et al. 2015a: 169-184

Wine

Correia et al. 2015a: 169-184

Fish - probably whole fish

Leroy de La Brière and Meysen 2007: 88-89 apud Bombico 2017: ficha 32

Figure 9. Results of chemical analyses on Lusitanian amphorae performed at the Chemistry Centre of Minho University (Braga, Portugal). and 1994-0315) two Conimbriga class 45‑46 amphorae (no.1-71.CRIP-NORTE.Cano and no.11-72.B.F6). The ovoid amphora (similar to Lusitanian 12), the amphora from Peniche (Peniche type 1) and the two Dressel 14 amphorae revealed the presence of fish derivatives (Oliveira, Morais and Araújo 2015; Morais, Oliveira and Araújo 2016). The high amounts of carbohydrates detected suggested the addition of sweetened products like honey or sweet wort. Several vegetal biomass-burning tracers were detected, indicating that processed fish products were in contact with smoke, being either cooked or smoked. The presence of several vegetable oil tracer compounds indicates the addition of either plants or spicy herbs to the processed fish, or that the fish products were in a close proximity treated with plants or herbs exudates (Morais, Oliveira and Araújo 2016: 205) (Figure 9). The data from the analysis seems to correspond with some passages from ancient authors such as Pliny (NH 31.95) when he refers to Melligarum/Mellogarum and, indirectly, Apicius (De re coq. 1.8) when he mentions that fish-based products of inferior quality had a nauseating odour, contrary to good quality products made with mackerel blood. For that reason, it would be necessary to cover up the odour using a recipe based on fumigation with laurel or cypress, with some honey or fresh must. The Punta Vecchia 1 (Cap Corse, Corsica) shipwreck, from the end of the 3rd century, beginning of the 4th, revealed 45 broken Lusitanian Almagro 51c amphorae of different sizes that were subjected to chemical analyses revealing vestiges of fat amino acids in the resin residues, suggesting they carried fish and not garum (Leroy de La Brière and Meysen 2007: 88–89 apud Bombico 2017: ficha 32). The chemical results obtained by the Minho University group were complemented by DNA-based analyses of fish remains still preserved inside a doliola from the end of the 4th century / beginning of the 5th, found in Boca do Rio (Budens, Algarve) (Figure 10).

A combination of a DNA barcode approach with HTS (High Throughput Sequencing) technology allowed the identification of six species of bony fish and one large sized shark used in fish sauce preparation (Landi et al. 2015: 237–245). However, these data should be treated with caution as some replicate previous information and additional analyses are being currently carried out to verify and deepen these findings. Those preliminary results confirmed the data from archaeology and chemical analyses of contents, revealing the economic importance of this activity in all the coastal area of Lusitania. New types of local or regional Lusitanian amphorae are still to be defined, which should be considered as minority productions in terms of trade. A good example is the above referred new type recently identified in the North of Lusitania, in the Roman city of Conimbriga, the ‘Class Conimbriga 45-46’ of small dimension, about 46.5cm long and 25cm wide, datable from the 1st to the 3rd century AD. Two specimens were analysed by gas-chromatography with mass detection and the results (Figure 9) revealed that they were wine containers (Correia et al. 2015a: 169– 184). Plants and/or seed oil residues were also detected, as well as several biomass-burning tracers, particularly in sample 72.B.F6 which presented the higher intensities of these compounds. Pinaceae residues were present on both amphorae, indicating that their content was contaminated with biomass burning emissions, so it should have been heated or boiled, probably with the referred plants and seeds (Correia et al. 2015a: 169–184). Another possible explanation is the use of a fumigation procedure to clean the winery, wine storage vessels and other paraphernalia for winemaking (Geoponica 6.11 apud White 1970). This revision on traces of contents shows the complementarity of the different analyses. The painted inscriptions indicated liquid sauces like liquamen or muria, the study of macro-remains indicated fish species

192

Lusitanian Amphora Contents

Figure 10. Doliola from Boca do Rio (Budens) whose content was subject to DNA-based analyses. and the significance of sardine, fish size and its possible correspondence to salsamenta or hallec, and the chemical analyses allowed the identification of wine residue in a small local amphora from Conimbriga and fish derivatives in early Lusitanian amphorae, as well as the complexity of the recipes to prepare the products. 4. Products likely to be transported in amphorae According to the evidence just reviewed, fish products and wine were the two products carried by Lusitanian amphorae, fish products being by far the most significant in the long term. What do we know of these products in their areas of production? How do they compare to the remains found in amphorae? 4.1 Fish products Considering that amphorae were the main transport containers of Lusitanian fish products, Lusitanian amphorae must have been filled with the products of the fish-salting factories in that territory. To date, little zooarchaeological evidence has been reported from Lusitanian fish-salting factories but the outcomes are of great interest in view of the paucity of fish samples collected in amphorae and the chance to compare them with the painted inscriptions (Figure 11).

A number of fish-salting factories are known in Lusitania and a survey in the digital platform RAMPPA — AtlanticMediterranean Excellence Network on Ancient Fishing Heritage documents — 92 cetariae (production units). Even though the numbers do not reflect the production capacity, since Western Lusitania had larger and longerlasting factories than the Algarve or the southwestern coast, it is worth noting that 20 of those are in the Tagus Valley, 31 in the Sado Valley, 8 in the southwestern coast and 33 in Algarve, often several production units belonging to the same production centre (Figure 1). Among these, three production centres stand out, Troia, with the highest capacity production, Lisbon, with many remains under the modern city, and Lagos, in the Algarve, also with many cetariae in an urban setting. Very often, the fish-salting vats of these cetariae preserve the remains of the last production, usually a light sediment, either loose or dense, with a significant concentration of fish bones. The samples of this residue collected in vats that have been analysed are presented in Figure 11 in a chronological sequence. The studied samples cover a time period from the late 2nd century until the first half of the 5th century, and they are mainly from Western Lusitania, Tagus Valley (Lisbon and Belém) and Sado Valley (Setúbal and Tróia),

193

Inês Vaz Pinto, Rui Morais, Carlos Fabião, César Oliveira and Sónia Gabriel

Date late 2nd c. 3rd c. 3rd/ 4th c. 5th c. 5th c. 5th c. 5th c.

5th c.

Production Site

Main species Scombridae, cf. Scomber Tróia (workshop 23, vat 7) scombrus (mackerel: >30cm) — (Gabriel – unpublished data) Heads only Sardina pilchardus (sardine: Quinta do Marim, Olhão 6-14cm, max. 19cm) — Whole (Desse-Berset and Desse 2000) fish Tr. Frei Gaspar, Setúbal (Desse- Sardina pilchardus (sardine: Berset and Desse 2000) 6-14cm, max. 19cm) — Whole fish Rua dos Correeiros, Lisbon Sardina pilchardus (sardine: (Assis and Amaro 2006) 8-18cm) — Whole fish Mandarim Chinês, Lisbon Sardina pilchardus (sardine: (Assis and Amaro 2006) 8-18cm) — Whole fish Sardina pilchardus (sardine: Casa do Governador, Belém representing average 67 to (Gabriel -unpublished data) 100% in 13 samples, 9-13cm, máx.19cm) — Whole fish Tróia (workshop 1/2) (DesseSardina pilchardus (sardine: Berset and Desse 2000) 8-14cm, max. 19cm) — Whole fish

Tróia (worshop 2, vat 1) (Gabriel 2012)

Secondary species

Minor species Sardina pilchardus (sardine: PAVLI MACEDONIS’, also on the top edge of a Flavian amphora, only has the nominal element (Bogaers and Haalebos 1989: 56, fig. 7.2; Berni 2017: 302, no.282). Even so it is of great scientific interest, as it helps us to better understand, with a particular example, the system of distribution of food within the Roman legions stationed in Germania, with Macedo being the final recipient of the oil to be distributed among the soldiers of the centuria of Paulus. The last example belongs to an amphora of the third quarter of the 2nd century AD (Berni 2017: 313, no.337), has no nominal element, and follows the more usual pattern for this kind of record. The greater unit of VII modii appears on the lip and the fractional number of XIIII sextarii in the curvature of the handle, to make a

242

Tituli picti on Spanish amphorae

Figure 3. Post cocturam graffiti from military context from Nijmegen. total of 68.936 liters. Taking into account that most of the Antonine Baetican amphorae follow a true standard of 216 Roman librae (Rodríguez Almeida 1981: 133–151), which is equivalent to the total capacity of 9 modii or 78.768 litres, 87.5% of the oil packaged in Baetica would have been subtracted at Nijmegen in the first instance.

ideas that have been debated by researchers in recent years, although with different points of view. The analysis of certain characteristics inherent in these inscriptions take into consideration the autopsy of the epigraphic document itself, which has not always been carefully examined in reference to certain hypotheses.

The official inscriptions of the Hispanic oil amphorae are always written in indelible black ink, which in the Roman world took the generic name atramentum for its colour (Aguilera and Berni 1998: 264–266). These inscriptions play a role similar to the adhesive paper labels on our plastic or glass bottles, so they were also placed during the canning process of the product being packed. The epigraphic system of the Dressel  20 amphora is well known from Monte Testaccio thanks to the work of Dressel in CIL XV and later by Rodríguez Almeida (Rodríguez Almeida 1984: 175ff). As Dressel noticed, the inscriptions painted on Dressel 20 are organized with five elements which, unlike other amphorae, share two important features: they are always written in a fixed position and their meaning is constant (Figure 2). On the neck the element α gives the tare of the amphora (c. 30kg). In the upper part of the belly the Greek letter γ gives the net weight of the olive oil (c. 70kg). The sum of these two figures gave the total weight of the oil amphora (c. 100kg), a value obtained from a simple calculation that was used to organize the transport of goods by water and land routes. In the upper bell is β with the name of the merchant, or even and often with a society of olearii, related or not, in possession of the merchandise. The only cursive inscription is δ, written with calamus transversely under the right handle, with useful data on the bottling and strict fiscal control. This is the only fixed inscription that will evolve over time, becoming increasingly long and detailed in a number of elements. At the base of the handle, next to the previous or sometimes almost overlapping element, a fifth inscription called ε appears frequently, a numerical figure of unknown significance that Dressel considered as an anomaly of these oil amphorae.

One of the most peculiar features of the Dressel 20 amphora tituli picti system is that its five elements are written in fixed positions and their meaning remained constant. This can be explained by the fact that the Baetican olive oil was a product of strategic consumption for the Roman State, as demonstrated through the archaeological record, or on the famous and monumental Monte Testaccio of Rome, the garbage dump of the Annona, or through the hundreds of thousands of fragments of these amphorae in German and British military sites (Remesal 1986; Carreras and Funari 1998). In parallel, there was a large and high private trade that allowed for the arrival of Baetican oil in any corner of the Roman Empire, as we also demonstrated with the collection of this material in the territory of presentday Catalonia (Berni 1998). Roman State consumption captured the largest volume of the market, but left the private sector enriched by trade, to the extent that this group became organized as a legal entity under the name ‘corpus oleariorum’. (Tabales Rodríguez and Jiménez Sancho 2001). Some of the merchants, who carried out commercial activity with the annona of Rome, became publicly known from the Antonine dynasty, through characters who are also registered in the tituli β of Testaccio, with the title or honor of ‘Diffusor olei ad annonam urbis’ (Chic et al. 2001). We believe that this peculiar and orderly epigraphic system was the product of the need to establish rules of control on the labelling of oil that would be understandably useful for the collective olearii and the administration of the Roman State. This is an epigraphic system of consensus between public and private bodies as two different legal entities that were joined in a common commercial enterprise.

At this point, it is useful to make a series of observations on specific important aspects of the function and meaning of this epigraphy, to clarify or specifying some

The following unique feature of the Dressel 20 amphora epigraphic system is known, although with a greater incidence, in other Hispanic amphorae productions,

243

Piero Berni and Emmanuel Botte

Figure 4. Cursive titulus δ from Testaccio with action from the ponderatores. [1] ‘pensit Fe[---]’ (Rodríguez Almeida 1994: 62, no.63). [2] ‘p(ondo) CCXV’ (CIL XV 4392). such as the tituli picti of the fish-sauce amphorae that have different structures and more complex and irregular records. We refer to the fast writing system that conditioned the epigraphic formulas and the technique of calligraphy for inscriptions. This characteristic is clearly expressed in the epigraphic composition of the Dressel 20 amphora. The enormous volume of the production and trade of Baetican olive oil required a practical and quick system of bottling, for the distribution every year of thousands of globular amphorae during the season of the diffusio olearia that was carried out between spring and summer. The tituli picti were written in positions indicated on the amphora (Figure 2) with the aid of a writing box (pictaccium) that was impregnated with a liquid which helped to smooth the surface and to fix the ink in any position on the surface of the object. First the empty amphora was weighed, then a brush was used to mark in α the tare in Roman librae. Then, with the amphora suspended on the Roman steelyard it was filled with oil to obtain the net weight of the product packed in γ. Then δ was written with a short calamus and with cursive writing under the right handle of the amphora. Then, with a small brush, the name of the owner of the merchandise was registered frontally in the upper part of the globular body. The bottling process ended with the closing of the Dressel 20 amphora mouth, using a round ceramic lid (operculum) with a thick layer of lime mortar that shaped the stopper. The cap in its fresh state was marked with a security stamp bearing the name of the character of β, in order to security seal the product and ensure that the amphora were not open without the permission of their owner (Berni and Gorostidi 2013). The busy activity of the diffusio olearia forced the development of a fast and efficient bottling system to speed up waiting times. Weighing the amphorae saved a great deal of time and effort, since filling them with a measure of capacity would have required the use of calibrated containers to carry out this calculation, which would have slowed down the work. Note that many tituli δ carry the abbreviation p(ondo) followed by the numeral γ (Figure 4, no.2). In some cursive inscriptions from the 3rd century AD the single letter p appears in front of servile names written in the Latin nominative case, or even with the total development of pensit into the verb pendo, going against the general rule (Figure 4, no.1). The reason proposed by Rodríguez Almeida is that the verb pendo was used in a transitive sense applied to the character linked to

the action of weighting, whereas the verb pondero followed by the numeral γ was intransitively applied to the weight of the amphora (Rodríguez Almeida 1994: 62, no.63). In the paleography of the tituli α and γ, the particularity of writing the weights quickly, with frequently repetitive values, drawn by means of a rectangular section brush, results in signs that resemble musical notes by the simple fact of combining thick horizontal strokes with thin verticals (Dressel 1878: 163). Dressel called them ‘Hispanic numerals’, although in reality we now know that they only occurred in Baetica (Aguilera and Berni 1998). To carry out this writing the values 1, 5, 10, 50 and 100 were used combined. From the beginning of the 2nd century AD the ductus was continued, linking the letters to each other, which made it possible to write the numbers more quickly, trying to raise the hand as few times as possible. Another feature, a little strange but effective, was to put small units in front of large ones to subtract quantities. Thus the construction of the final number was shorter in signs: CXCVIII instead of CLXXXXVIII (CIL XV 4491), XCI instead of LXXXXI (CIL XV 4491). In addition, it also helped to correct the final calculation without having to repeat the whole operation by rewriting the result: ICI to return to C (CIL XV 4044), IXCI to subtract the unit from further obtaining XC (CIL XV 4491). Of all the inscriptions on the Dressel 20 amphora the cursive titulus δ was the one that evolved the most over the years. During the 1st century AD they are simple, with a few elements written in a single line (a name in the genitive, another in the nominative, the abbreviation À, the value of γ). During Hadrian’s reign new elements appear (R barred, a city name, a neutral name, other numerals, the abbreviation acc() ). Some years later with Antoninus Pius consular dating is introduced. With the reform of Septimius Severus new elements appear and others disappear. The longest forms occupy three lines of writing. The way of making the cursive inscription was also thought to save time when written with a synthetic language based on keywords, mostly abbreviated. The elements followed a logical order, proper names were governed by the grammatical case, and to avoid confusion about the context of the names abbreviated attributes were added in front. In this way the cursive titulus δ did not lose or alter their meaning, making it possible to reconstruct the original full document knowing the writing system codex.

244

Tituli picti on Spanish amphorae

Figure 5. Two cursive tituli picti from Testaccio with the indication of the bottling location in the valley of the Guadalquivir (CIL XV 4350 and 4371). The cursive δ was an official fiscal registration of the bottling process (Dressel 1878: 169), as a result of the sale-purchase of the oil between two physical entities, the producers or hoarders and the agents of the merchant. Dressel 20 amphorae were bottled in facilities specially prepared to carry out this busy activity of the diffusio olearia in the mare apertum months. We know from the inscriptions of Testaccio that some Stationes fisci were annexed to the figlinae or within the fundi figlinarum on the banks of the rivers Guadalquivir and Genil, if it was possible to have the appropriate port infrastructures to withstand the high volume of merchandise traffic. Rodríguez Almeida published, some years ago, an illustrated scheme that is very useful for understanding the dynamics of this economic activity in Baetica (Rodríguez Almeida 1993: 101, fig. 2), connecting with arrows the fundi olearii of the inland province with the bottling officinae next to the riverside potteries. In the following illustration (Figure 5) we present two examples from Testaccio with the explicit mention of the place where the olive oil was packaged. In the titulus on the left it reads ‘R fig( ) Scimnia( ) As[tig](is) / Aurelio Cae[s]are III C[ommodo II co(n) s(ulibus)] / arcle(n)se Soranae’ with consular date of the year AD 161 (CIL XV 4350). The one on the right is dated 179 AD and reads ‘R at port(um) XIII [---] / fulvianum Camp[--- ------] / Imp(eratore) Commodo II et V[ero II co(n)s(ulibus)]’. It should be noted first that the interpretation of these synthetic records will depend on the theoretical model used by each researcher. To date, there is still no consensus on the meaning and formulation of the different elements that shape these cursive inscriptions. There are many scientific theories, sometimes conflicting (Rodríguez Almeida 1984: 175ff ; Chic 1986; Remesal 1986: 21ff; Liou and Gassend 1990: 208; Liou and Tchernia 1994). We cannot discuss all these issues here, but will focus on certain aspects linked to our own historical questions. From the titulus δ of CIL XV 4350 we know that the oil packaging process was carried out in the Astigitanian figlina of the Scimniani family, located in Las Delicias, next to the right bank of the Genil River and near Écija (the Roman city of Astigi) (Berni 2008: 426). The consular dating, which was first introduced in the time of Antoninus Pius, controlled the date on which the tax was paid in the year the oil was bottled. The letter R barred, which Dressel believed at first to be the abbreviation Rat(ionis), appears on most of the

inscriptions from the middle of the 2nd century AD almost always accompanying a city name from the Guadalquivir valley (Astigis, Hispalis, Corduba), the Baetica coast (Lacca, Malaca, Sel), or the border of Tarraconensis (Castulo). Dressel changed his mind (CIL XV: 562), influenced by Mommsen and suggested reading R(ecognitum) when he saw an official character in the state controlled operation of the oil trade. Rodríguez Almeida suggested reading R(ecensitum) (Rodríguez Almeida 1972: 126–128), which in the Roman administrative language has the value of enumerating, counting and census, after taking the phrase ‘ad oleum afrum et hispanum recensendum’ of the hispalense epigraph with the cursus honorum of the praefectus annonae Ulpius Saturninus (CIL II 1180). In Remesal’s opinion (Remesal 1986: 21–22) the R would read as recognitum, and next to the name of the cities, which he interprets as the fiscal district from which the amphora was issued, the form translated into modern terminology could have the meaning: ‘Checked in the district of Astigi …’. This last hypothesis conflicts with the documentation of Testaccio, as we know several examples where the location of the oil packaging does not correspond territorially with the administrative name of the city where the fiscal check was supposed to be performed. For example, the Astigitanus check CIL XV 4294 is associated with the stamp SNR of La Catria in the conventus Hispalensis. Or the most striking case, because of the geographic distance, that we find in CIL XV 4031, is in Corduba on an amphora from the territory of Hispalis with the stamp DATSCOL of Azanaque-Castillejo figlina. Remesal also proposed that the name of the city be the capital of the administrative convent where the fiscal check was carried out, but we have among the names of coastal cities Lacca, Malaca, Sel (CIL XV 3926, 4203; Testaccio VI nos463 and 479) which was part of the conventus Gaditanus. In addition, there is a cursive mark of Castulo (CIL XV 4137), belonging to the conventus Carthaginensis, for a Tarraconensis oil that had to be transported for packaging, after covering a great distance, at some navigable point of the Guadalquivir river below the colonia of Corduba. Therefore, we believe that these city names would not allude to the administrative capital of the place where the fiscal check was carried out, but to the fiscal districts of the oil to be bottled. We believe it has to do with an attempt to eradicate fiscal fraud, which

245

Piero Berni and Emmanuel Botte

would have been generated, for example, by packaging oil from Corduba at a fiscal statio in the territory of Hispalis. Perhaps that is the reason why Dressel’s initial proposal on R barred as an abbreviation of Rat(io) can be rescued, a term that is neither alien nor incompatible with oil trade (The tituli β of 3rd century AD of the ‘Fisci rationis patrimoni provinciae Baeticae’: CIL XV 4124, or ‘Fisci rationis patrimoni provinciae Tarraconesis’: CIL XV 4134-4137), if it is understood that the tax generated by the sale-purchase of the oil was to be entered in the Fiscus, the personal chest of the emperors of Rome, with the place of origin of the packaged product. There are other tituli picti β of the Ratio Fisci, dated around AD 220-224, that include at the end of the inscription names of stationes, such as Sel, Sex and Po[---], which correspond to port towns on the coast of conventus Gaditanus (CIL XV 4142; Testaccio II no.125, 120–127; Martínez et al. 2017).

and trade fish-salted products and sauces. Within these two categories are all kinds of fish, from tuna to sardines, and for the latter four varieties of products, known as garum, liquamen, muria, and allec.

Returning to the Astigi inscription of CIL XV 4350 (Figure 5.1), in the last line there is a neuter arcle(n)se in front of a female name Sorana. The neuter names of the tituli picti δ have endings -um and -e(n)se and usually accompany a person’s name in the genitive. We, following Dressel, think these names are place names. When they end in -um as in fulvianum (Figure 5.2) they allude to the oil producer, and if they end in -e(n)se the fundus olearius. Remesal (1979) thinks that it is the place where the fiscal check was carried out, but in the amphora of Las Delicias we already have that indication with the name of the figlina Scimniana. On the other hand, if we develop the text of the third line as ‘(oleum) arcle(n)se Soranae’, the genitive name could fit semantically as the owner of the product to be packaged. Remesal (1979; 1986: 22) believes that the characters in the genitive are the publicani in charge of the fiscal control. In this example it could be a woman who may have stocked oil from different fundi olearii.

These letters each correspond to specific information:

As regards the inscription CIL XV 4371 (Figure 5.2) a similar formula to the previous one is observed, except that instead of the name of city we have the word ‘ad port(um)’ that Dressel believed to refer to the portus of Gades (Dressel 1878: 161). For us this indication associated with the neutral fulvianum should be related to the bottling facilities of the Arva river port (Berni 2008: 165), which were annexed to an important sector of the craft industry that has left its print on the amphorae with the PORT·ARVA stamps (Berni 2008: 287, table 45). We know that the Fulvii of Arva practiced different economic activities linked to agricultural production (CIL II 1064, Suppl.: 837), such as the manufacture of amphorae (Berni 2008: 194–195), and the oil trade (CIL XV 3876). 3. Amphorae for fish products Here we gather the main amphorae for fish-products from the Iberian peninsula, namely Dr. 7-11, Dr. 12, Beltrán II, Dr. 14, Schoene VII (also called Pompeii VII). From the end of the Republican period and throughout the early Roman period, these amphorae were used to transport

The majority of scholars now follow the pattern of organization of tituli picti proposed by S. Martin-Kilcher (Martin-Kilcher 2002, 2004 and 2011), but there are also other systems, notably that used by R. Étienne and Fr. Mayet (Étienne and Mayet 2002: 211–214), who apply the codification of H. Dressel to fish amphorae for reading the inscriptions on Dressel 20 oil amphorae. The diagram proposed by S. Martin Kilcher corresponds to Figure 6, organized from letters A to H, with primary inscriptions from the letters A to E and secondary inscriptions from F to H.

We will find in the first line the name of the product, often abbreviated, and this is one of the major differences with wine and oil amphorae, for which the nature of the content is never indicated. This implies of course that there was little doubt that they were transporting this type of foodstuff. For amphorae for fish products, this is also due to the fact that a wide variety of preparations could be placed in these containers: garum, liquamen, muria, and other products whose meaning is not yet fully understood: lymp, laccatum… We do not have enough space here to dwell on the meaning of all the inscriptions found on line A, and for which there is an abundant bibliography, including recently (for example papers of U. Ehmig, more precisely Djaoui 2016; Bernal‑Casasola and Domínguez Bella 2011; Quillon 2016). The mention of the product is sometimes accompanied by a product-specific qualifier, such as vet for vet(us) or arg for arg(utus). On the lines of group B are first qualifiers on the quality of the product: excellens or flos being the most common, and one also finds penuar or summaur. The number of years of aging can be found, though not systematically, for example in the AIIA formula that indicates two years. Line C corresponds to a figure, which almost certainly corresponds to a unit of measure of the amphora content, and it is assumed that it must be a unit of measure in pounds, although sometimes figures have very high values (Martin-Kilcher 1994: 420–421; Botte 2009: 162). The inscription of line D will give us the name of the merchant, the negotiator, sometimes in a simple form but more frequently in the form of duo or tria nomina. Finally, under the handle, perpendicular to the group of inscriptions C to D, there is the inscription E, composed of a number and / or a name whose meaning is not obvious

246

Tituli picti on Spanish amphorae

Figure 6. Organization of the tituli picti on fish amphorae (Martin-Kilcher 2002). because it is often written with small cursive letters. It is generally considered to be a control mark written by a negotiator’s representative. Finally there are the secondary inscriptions F to H. For F and G, S. Martin‑Kilcher suggests the signature of a merchant and the number of amphorae that he transported between the port of arrival of the containers and the consumption sites (Martin-Kilcher 2002: 345–346). Others suggest the name of the producer and the number of amphorae leaving the fish-salting factory (Étienne and Mayet 2002: 213–214). Finally the last inscription H refers to the owner of the amphora once it arrived at its destination. It may have been painted under the line D, sometimes even over the group of primary inscriptions A to D, and it is clear that it is a different hand, sometimes even in a different colour. Bibliography Aguilera Martín, A. and Berni Millet, P. 1998. Las cifras hispánicas. In J. Mateu Ibars (ed.), Calligraphia et tipographia. Arithmetica et numerica. Chronologia: 257–282. Barcelona, Universitat de Barcelona. Bernal-Casasola, D. and Domínguez Bella, S. 2011. Colorantes y pigmentos en las pesquerías hispanorromanas. CuPAUAM. Cuadernos de Prehistoria y Arqueología de la Universidad Autónoma de Madrid 37‑38: 671–685. Berni Millet, P. 1998. Las ánforas de aceite de la Bética y su presencia en la Cataluña romana. Instrumenta 4. Barcelona, Universitat de Barcelona.

Berni Millet, P. 2008. Epigrafía anfórica de la Bética. Nuevas formas de análisis. Instrumenta 29. Barcelona, Universitat de Barcelona. Berni Millet, P. 2017. Chapter 8. Amphorae-Epigraphy: Stamps, Graffiti and Tituli Picti from Roman Nijmegen. In C. Carreras and J. van der Berg (eds), Amphorae from the Kops Plateau (Nijmegen): trade and supply to the Lower-Rhineland from the Augustan period to AD 69/70. Archaeopress Roman Archaeology 20: 185–282, 289– 343. Oxford, Archaeopress. Berni Millet, P. 2018. Calendar graffiti on Dressel 20 amphoras. Asiaticus: another paradigmatic case with a new find from Brijuni (Croatia). in T. Bezeczky (ed.), Amphora research in Castrum Villa on Brijuni Island. Archäologische Forschungen 28: 143–172.Vienna, Austrian Academy of Sciences Press. Berni Millet, P. and Gorostidi Pi, D. 2013. C. Iulius Valerianus et C. Iulius Iulianus: mercatores del aceite bético en un signaculum de plomo para ánforas Dressel 20. Journal of Roman Archaeology 26: 167–189. Bogaers, J. E. and Haalebos, J. K. 1989. Opgravingen op het terrein van het voormalige Canisiuscollege, 19871989. Overdruk uit Nuaga 36: 49–60. Botte, É. 2009. Salaisons et sauces de poissons en Italie du Sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31. Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Callender, M. H. 1965. Roman Amphorae, with index of Stamps. Oxford, University Press. Carreras Monfort, C. and Funari, P. P. A. 1998. Britannia y el Mediterráneo. Instrumenta 5. Barcelona, Universitat de Barcelona. Chic García, G. 1986. El comercio del aceite de la Astigi romana. Habis 17: 243–264. Chic García, G., García, E., Romo, A. and Tabares, M. A. 2001. Una nueva inscripción annonaria de Sevilla: M. Iulius Hermesianus, diffusor olei ad annonam urbis. Habis 32: 353–374. CIL II: Corpus Inscriptionum Latinarum, vol. II. Inscriptiones Hispaniae Latinae, A. Hübner, A. (ed.), Berlin, 1869; Inscriptionum Hispaniae Latinarum Supplementum, Hübner A. (ed.), Berlin, 1892. CIL XV: Corpus Inscriptionum Latinarum, vol. XV. Inscriptiones urbis Romae Latinae. Instrumentum domesticum, Pars II, fasc. 1, H. Dressel (ed.), Berlin, 1899. Colls, D., Étienne, R., Lequément, R., Liou, B. and Mayet, F. 1977. L’épave Port-Vendres II et le commerce de la Bétique à l’époque de Claude. Archaeonautica 1. Djaoui, D. 2016. The myth of ‘Laccatum’: a study starting from a new titulus on a Lusitanian Dressel 14. In I. V. Pinto, R. R. de Almeida and A. Martin (eds), Lusitanian Amphorae: Production and Distribution. Roman and Late Antique Mediterranean Pottery 10: 117–127. Oxford, Archaeopress. Dressel, H. 1878. Ricerche sul Monte Testaccio. Annali dell’Istituto di Corrispondenza Archeologica 50: 118–192. Étienne, R. and Mayet, F. 2002. Salaisons et sauces de poisson hispaniques. Trois clés pour l’économie de l’Hispanie romaine, II. Paris, de Boccard.

247

Piero Berni and Emmanuel Botte

Grenier, A. 1934. Archéologie gallo-romaine (Manuel d’archéologie préhistorique, celtique et gallo-romaine de J. Déchelette, VI). Deuxième partie : L’archéologie du sol : 601642. Paris, A. Picard. Liou, B. and Gassend, J.-M. 1990. L’épave Saint-Gervais 3 à Fos-sur-Mer (milieu du IIer siècle ap. J.-C.). Inscriptions peintes sur amphores de Bétique. Vestiges de la coque. Archaeonautica 10: 157–264. Liou, B. and Tchernia, A. 1994. L’interprétation des inscriptions sur les amphores Dressel 20. In Epigrafia della produzione e della distribuzione. Actes de la VIIe Rencontre franco-italienne sur l’épigraphie du monde romain (Rome, 5-6 juin 1992). Collection de l’École française de Rome 193: 133–156. Rome, École française de Rome. Martin-Kilcher, S. 1994. Die römischen Amphoren aus Augst und Kaiseraugst, 2-3. Die Amphoren für Wein, Fischsauce, südfrüchte (Gruppen 2-24) und Gesamtauswertung. Forschungen in Augst 7/2-3. Augst, Römermuseum Augst. Martin-Kilcher, S. 2002. Lucius Urittius Verecundus, négociant à la fin du Ier siècle, et sa marchandise découverte à Mayence. In L. Rivet and M. Sciallano (eds.), Vivre, produire et échanger : reflets méditerranéens, Mélanges offerts à Bernard Liou. Archéologie et Histoire romaine 8: 343–353. Montagnac, Éd. Monique Mergoil. Martin-Kilcher, S. 2004. Fischsaucen: Pinselaufschriften und römische Amphoren. In J. Remesal Rodríguez (ed.), Epigrafía anfórica. Instrumenta 17: 245–257. Barcelona, Universitat de Barcelona. Martin-Kilcher, S. 2011. Formes d’amphores et contenus au Haut-Empire, points de repère et questions. In SFECAG, Actes du Congrès d’Arles: 417–426. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Martínez Rodríguez, F., Márquez Cruz, M., Sarompas Cazorla, C. E. 2017. Tituli picti en ánforas olearias béticas alusivos a la statio sel y a la statio sex, hallados en el Monte Testaccio de Roma, y su contexto geo‑histórico. Cuadernos de prehistoria y arqueología de la Universidad de Granada 27: 373–397.

Mayet, F. 1986. Les figlinae dans les marques d’amphores Dressel 20 de Bétique. In Hommage à Robert Étienne. Publications du Centre Pierre Paris 17. Revue des études anciennes 88: 285–305. Quillon, K. 2016. Les amphores à salaisons et sauces de poissons de Bétique et de Tarraconaise : typologie et contenu (fin de la République - Haut-Empire). Unpublished PhD Thesis, Aix-Marseille Université. Remesal Rodríguez, J. 1977-78. La economía oleícola bética: nuevas formas de análisis. Archivo Español de Arqueología 50-51: 87–142. Remesal Rodríguez, J. 1979. Recensión a Archaeonautica I (1977). Archeologia classica 31: 379–389. Remesal Rodríguez, J. 1986. La annona militaris y la exportación de aceite bético a Germania. Madrid, Universidad Complutense. Rodríguez Almeida, E. 1972. Novedades de epigrafía anforaria del Monte Testaccio. In Recherches sur les amphores romaines. Actes du Colloque de Rome (4 mars 1971) Collection de l’École française de Rome 10: 107–241. Rome, École française de Rome. Rodríguez Almeida, E. 1981. Varia de Monte Testaceo. Cuadernos de Trabajos de la Escuela Española de Historia y Arqueología en Roma 15: 105–164. Rodríguez Almeida, E. 1984. Il Monte Testaccio: ambiente, storia, materiali. Rome, Quasar. Rodríguez Almeida, E. 1993. Graffiti e produzione anforaria della Betica. In W. V. Harris (ed.), The inscribed economy: production and distribution in the Roman empire in the light of instrumentum domesticum: the proceedings of a conference held at the American Academy in Rome on 10-11 January, 1992. Journal of Roman Archaeology Supplementary Series 6: 95–107. Ann Arbor, University of Michigan. Rodríguez Almeida, E. 1994. Los tituli picti. In J. M. Blázquez Martínez (ed.), Excavaciones arqueológicas en el Monte Testaccio (Roma): 36–129. Madrid, Instituto de Conservación y Restauración de Bienes Culturales. Tabales Rodríguez, M. A. and Jimenez Sancho, A. 2001. Hallazgo de una nueva inscripción referente al cuerpo de olearios en el Alcazar de Sevilla. Habis 32: 375–385. Thévenot, E. 1951. Una familia de negociantes de aceite establecida en la Bética en el siglo II: los Aeli Optati. Archivo Español de Arqueología 25: 225–231.

248

The contents of amphorae produced in Gaul in the Imperial period Fanette Laubenheimer

CNRS, UMR 7041, Paris Ouest – Nanterre

Abstract: This paper is a survey of Gaulish amphorae contents in the Imperial period: wine, fish sauces, mead (?) and oil. Imitation of foreign amphora types and typical Gaulish production are all investigated. Tituli picti show that different qualities of wine from sometimes different provenance were filled in the same types of amphorae, for instance Gauloise 4 or 5 from Narbonnensis. Tituli picti also demonstrate that the Lyon 3 and 4 amphorae types from the Lyon area, usually considered as fish sauce amphorae, were in fact used for a range of contents: several fish sauces, defrutum or mead. So, we need to be aware of the fact that sometimes, one amphora type is not associated with a single content. Key words: Gaulish amphorae; contents; wine; fish sauces; mead; olives; oil.

1. Introduction We are interested in the study of amphorae not for themselves but for what they represent: commercial containers for the transport of commodities. It is indeed their economical significance that is at the heart of the matter and that can help us understand the issues of production and exchange in the Imperial period. Yet the arguments at our disposal in order to certify or deduce the contents of an amphora are not always evident. Clear, unambiguous and well-preserved inscriptions on amphorae are rare. Other parameters are taken into consideration such as the shape of the amphora, the residues it contains, the presence of resin on the inside, the context of the discovery, the localisation of production, iconographical documents, ancient texts, etc. However, in many cases when we only have probabilities to establish the content of an amphora, we often without hesitation deduce important orientations of Roman economy. What is the situation in Gaul? Can we be careful and make progress at the same time? What should we call into question? There are several scenarios to be considered. On the one hand amphorae of which the shape is of foreign origin and copied in Gaul. On the other hand those of which the shape is typical of Gaulish creation. We shall question the relation between content and container: is there a single commodity, always the same one, from the same origin? Or are there different qualities of goods, from various origins? We shall also study the amount of production and its distribution; we shall compare mass-produced amphorae and those produced in small quantity for local distribution. Their impact on economy is quite different.

We shall approach the subject by examining amphorae according to the type of product they contain: wine, fish sauce, mead (?), olives and oil.1 2. Wine In Gaul, wine was the ingredient most frequently transported in amphorae, even though barrels were also in use from the 1st century A.D. 2.1. In Gallia Narbonensis Imported models of amphorae From the Augustan period, models of wine amphorae from Italy were copied. Dressel 1 Apart from the imitation of a few Greco-Italic amphorae, like in Marseilles, Dressel 1 were copied in Gaul; many painted inscriptions show that they were used in their country of origin for the transport of various qualities of wine. Workshops where they were manufactured have been located in Marseilles (Bertucchi 1992b), in or around the Rhone Valley, in Mazan in the Vaucluse (Laubenheimer and Schmitt 2009), in Saint-Just in the Ardèche (Laubenheimer, Odiot and Leclère 1989) or in Lyon (Desbat and Dangréaux 1997). We have no painted indication to show the kind of wine they contained. The quantity of amphorae produced appears to be limited and their distribution little known. However we do know that Italian craftsmen probably ran the workshops where they were produced, particularly in Mazan and Saint Further information on painted inscriptions on Gaulish wine and fish amphorae in Gallia Narbonensis can be found in Fabrice Bigot’s thesis Nouvelles données sur la production et la diffusion des amphores gauloises à partir de l’étude de contextes portuaires et littoraux de Gaule Narbonnaise (Ier s. av. – IVe s. ap. J.-C.) University of Montpellier 2, December 2017. 1

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 249–257

Fanette Laubenheimer

Just, where the production of plates and architectural pottery were characteristic of Italian savoir-faire. Dressel 2/4 During the same period and shortly after, in some of these workshops and in a large number of others, appear productions of Dressel 2/4 amphorae, another Italian model imported from Greece. The oldest and the most numerous workshops are situated in Gallia Narbonensis (Mauné 2013) but the model will soon spread throughout Gaul during the Empire, as in Sens in the Yonne (Laubenheimer 2003), Crouzille-Mougon in the Indreet-Loire (Laubenheimer 1989) or Mazières-en-Mauge in the Maine-et-Loire (Laubenheimer 2000b). For the Narbonensis, two painted inscriptions discovered in Castro Pretorio in Rome indicate the wine’s origin: Béziers. The inscription is in white instead of black like on other Gaulish amphorae. The first one (CIL XV, 4542) shows an apparently original formula,2 SVM VET V, ‘Best quality (wine) aged five years’.3 This is not an ordinary old wine (usually a year old) but clearly an aged wine, which gives it a special quality. On the second amphora (CIL XV, 4543) the wine from Béziers is called white, a unique designation referring probably to the great Roman wines that were white. Thus we know that Gaulish Dressel 2/4 amphorae could contain white wine but probably not always: on our first amphora the colour of the wine is not specified and that might be because it was implicitly inferred that it contained red wine. What about the contents of other Dressel 2/4 amphorae in Gaul? We do not know for certain but we presume that it was wine. However we must point out the fact that Dressel 2/4 amphorae made in workshops throughout Gaul are always produced in parallel with other types. There are no workshops specialised in Dressel 2/4 models, as for instance for the Gauloise 4 amphorae. There appears to be no massive and standardised production, the shape varies from one workshop to another, we have occasionally discovered several contemporary variants within the same workshop, for example at Castellet in the Var (Laubenheimer 1989), Crouzilles-Mougon or Mazières-en-Mauge. Are they diversified containers for different wines? Or just proof of the potter’s skill to create different standards? Exportation is not massive and is not well known. Pascual 1 In the Augustan period and in the 1st century, several workshops in Gallia Narbonensis start producing Pascual 1 amphorae, an imitation of a Tarraconensis model (Mauné 2013). Like Dressel 2/4 they are not standardised productions and most of the time they are related to other types of amphorae. Workshops have also been localised in the West of Gaul, those in the Centre export to Brittany (Laubenheimer,  The word SVM must be read SVM(MVM), of superior quality, like on others tituli picti, and not “I am” like it was before. 3  D. Coutourier, F. Laubenheimer, A. Schmitt, forthcoming study. 2

Cantin and Schmitt 2005). The clay used in these regions is orange and grey in the centre and the potters covered the amphorae with a white coat, probably to make them look like the Tarraconensis P1 model made from white clay. Thus the appearance of an amphora, its very colour, is significant. Is it related to the content? Did these amphorae transport wine and vinegar like their models? We have no trace of a painted inscription to say so. Gaulish models Gauloise 2 Also from the Augustan period, Marseilles began to produce flat-based amphorae of the Gauloise 2 type (Bertucchi 1992b). Other workshops in Gallia Narbonensis did the same (Mauné 2013). The quantities involved were not large. The amphorae from Marseilles in characteristic micaceous clay, are clearly recognisable and we can follow the trace of their exportation northwards through the Rhône Valley to many Gaulish markets but always in small quantities. For example in the large NorthWest region only seventeen specimens have been found (Laubenheimer and Marlière 2010: 39). So far we have not discovered any painted inscription indicating that they contained wine. Gauloise 1 Similar to the Greek wine amphorae from Marseilles but flat-based, the Gauloise 1 model is distinctive. It is the unique production of some workshops in the Gard that used sandy clay during the Augustan period and the 1st century. They will later be produced on a small scale in certain large Gauloise 4 workshops (Mauné 2013). For a long time we believed that they contained wine. This has been confirmed by the recent discovery of a specimen with a painted inscription in the Rhone in Arles (communication D. Djaoui, F. Bigot; Bigot 2020). The production of these amphorae was localised in the South of Gallia Narbonensis and as far as Castres in the Tarn near Toulouse (Laubenheimer and Schmitt 2009: 133). Their exportation was mainly regional, rarely beyond, for example to Lyon (Laubenheimer and Bonnet 1998). They contained Narbonensis wine, which was never much favoured by the large markets of the Empire. Gauloise 5 This typically Gaulish model is characteristic of Provence (Laubenheimer 1985). Oddly we have found no trace of their production in the Languedoc (Mauné 2013). An imitation was made in the Bordeaux region (Berthault 1992a). Because of its shape and the Provençal context we believe that it was a wine amphora, which is confirmed by painted inscriptions. Gauloise 5 amphorae are a typical and clearly recognisable production with limited exportations, specially to Rome. They contained two different qualities of wine; one amphora from Fossur-Mer contained picatum excellens and others contained wine described as vetus and Massicum (Laubenheimer 2004). I shall come back to the term.

250

The contents of amphorae produced in Gaul in the Imperial period

Gauloise 3 This flat-based type less well defined because of many variants is also a wine amphora. Two specimens from Fos-sur-Mer show that it contained amineum (Laubenheimer 2004). Produced together with other models in some workshops of Gallia Narbonensis (Mauné 2013), they were imitated elsewhere in Gaul. Gauloise 7 These are rare and limited productions related to certain large workshops (Mauné 2013). Only one specimen found in Fréjus mentions the transport of amineum (Laubenheimer 2004). Gauloise 4 It is the standardised Gaulish amphora par excellence, destined for large-scale trade (Laubenheimer 2001). Mass produced from the second half of the 1st century to the 3rd century in Gallia Narbonensis in dozens of workshops (Mauné 2013) in connection with vineyards, wine-growing villae and ports, they were exported to the large markets of the Roman Empire and beyond (Laubenheimer 2015: fig. 59). The Gauloise 4 amphorae contained various types of wine as indicated by over thirty painted inscriptions (Laubenheimer 2004): mainly amineum vetus but also picatum vetus and maybe even mulsum and depletum (racked wine). Sometimes the wine was simply described as vetus. Thus various qualities of wine were related to the same container. Some non-standardised imitations of Gauloise 4 were produced outside Gaul but in small quantities and there are no inscriptions to indicate their role. We shall not consider here the types of flat-based Gaulish amphorae, Gauloise 8 and 9. They are minor productions; their type is not clearly defined for we do not have enough whole specimens and no painted inscription. We can safely say that the great majority, if not all of the flat-based Gaulish amphorae from Gallia Narbonensis, were wine amphorae. But what about the wine they transported, did it all come from that region? It is questionable when we see the word Massicum appear on a dozen Gauloise 4 or Gauloise 5 amphorae (Laubenheimer 2004). The word relates to the Massic, a region that produced excellent wine in Campania. Does that mean that the famous Italian wine was delivered in bulk and reconditioned in Gaulish amphorae in a port of arrival such as Marseilles? It is unlikely that a good wine, also called amineum or vetus in Gaul, would have been transported in this way. Could it not rather refer to a Gaulish wine made in the Campanian tradition? Or else could sometimes the abbreviation Mass simply allude to wine from Marseilles? We know that different kinds of wine can be transported in one type of amphora: for example amineum or picatum in Gauloise 4. Therefore in Gaul the shape of the amphora is not associated to a specific wine, the container can be used for several kinds of quality or variety.

2.2. In the Three Gauls The Dressel  2/4 and Pascual  1 Italian models have sometimes been imitated in the Three Gauls, for example in workshops in the West of the country such as CrouzillesMougon and Mazières-en-Mauge or in the Seine basin at Sens (Yonne). There are no painted inscriptions to confirm that they were used for wine as suggested by their shape. They are regional productions, rarely exported, which can be related to the development of vineyards in the North of Gaul from the 1st century onwards. Furthermore models of Narbonensis wine amphorae were copied without strictly conforming to standardised characteristics. For example Gauloise 5 in Bordeaux or Gauloise 4 at Noyon in the Oise (Laubenheimer and Marlière 2010: 494–495), Gueugnon in the Saône-etLoire (Laubenheimer 1989 and 2003) or Sens (Yonne). Were they really wine amphorae? There is no painted inscription to confirm the fact. Other original models of flat-based amphorae were created outside Gallia Narbonensis. There is for example the small Lyon 4 amphora from the region of Lyon with an inscription on a specimen from Cologne showing that it contained defrutum hispanum, a ‘Spanish’ vin cuit (Ehmig 2007: no.168). Whereas other amphorae of the same type contained fish sauce. We shall return to the subject later. There is also the Gauloise 12 model from some workshops in and around Normandy that produced and exported several modules of a specific type (Laubenheimer and Marlière 2010: 544–545). The nature of their content is thought to be wine because of the flat base and the presence of resin on the inside but it remains hypothetical. There are still a number of variants of flat-based amphorae listed in the Three Gauls that have not been clearly associated with local wine. We only have the relation between ancient texts on wine growing, traces of wine storehouses or vineyards and amphorae workshops to work on. One example among others is the Thésée-Pouillé workshop in the Loir-et-Cher (Laubenheimer 1989) where we have found several original types but the quantity we have is too small and the fragments too rare for us to be able to establish a true typology. The relation between these amphorae and their contents remains an enigma. Most often we tend to evoke wine even though we have no definitive proof. If we have no doubt about the contents of several types of flat-based wine amphorae in Gallia Narbonensis thanks to their painted inscriptions and to the geography of the workshops in relation to the vineyards, we must admit that concerning many sorts of amphorae in the Three Gauls, we have not yet been able to clearly determine their contents.

251

Fanette Laubenheimer

3. Fish sauces Many fewer than wine amphorae, Gaulish fish amphorae are nevertheless present. Indeed traces of fisheries along the French Mediterranean coast are not rare. Punctually we have found fragments of Dr. 7/11 amphorae, imitations of an Iberian type, produced in very small quantities in Narbonensis workshops dedicated mainly to wine amphorae: one specimen in Sallèles d’Aude (Aude), a few in Velaux (Bouches du Rhône), or in Marseilles, for example (Laubenheimer 1985), or in the region of Lyon. We must also mention a small production of Dressel 14, imitations of Lusitanian amphorae in Fréjus in the Var (Laubenheimer 1991). However, several types of Gaulish shaped amphorae were specialised in the transport of fish sauces, as is indicated by their painted inscriptions. First of all there are the Dressel 16 of which six specimens exported garum, muria or liquamen from Antibes to London, Rome, Ostia, Augst or Narbonne (Laubenheimer 1992, and 2004). Once again the same type of amphora can contain different products as we have seen with wine (Figure 1). The term Antipolitana, from Antibes, which is associated with all three kinds of fish sauce, is apparently not only connected to the town but also to the area or to the recipe because one of the amphorae (the one from London) was made in Fréjus, not Antibes (Picon in Laubenheimer 1991) and two others found in Augst are also attributed by analysis to the area of Fréjus (Desbat, Picon and Schmitt 1994: 652). Furthermore a Dressel 16 exported to Cologne contained muria hispana excellens (Ehmig 2007: no.160). Augst 33 amphorae (Figure 1), of Provençal origin and more specifically from Fréjus (Desbat, Picon and Schmitt 1994: 652) also contained muria flos Antipolitana as shown by two beautiful specimens in Cologne (Ehmig 2007: no.151 and no.152). The case of fish amphorae from Lyon is more complex. Three types of Gaulish shaped amphorae are concerned: Lyon 3A, from the 1st century to the days of ClaudiusNero, Lyon 3B, more numerous, in the Flavian period and small-sized Lyon 4 in the 1st century (Figure 2). In any event, the first ones, Lyon 3A, contained garum flos or scombri according to about ten inscriptions that have reached us. The second ones, Lyon 3B, contained essentially muria (excellens or not) and less frequently garum according to twenty-four inscriptions (Laubenheimer 2004). Finally Lyon 4, of small capacity, contained garum or liquamen or even dufrutum (vin cuit), which has nothing to do with fish (Ehmig 2007: no.168). Once more, concerning Lyon 3B and Lyon 4, different kinds of content may correspond to the same type of amphora. Liquamen or muria in the case of Lyon 3B have

been described as Antipolitana or Hispana. There are also two qualifications for Lyon 4: garum pompeianum primum and muria antipolitana (Ehmig 2007: no.164, 165, 166). To sum up, in the present state of our knowledge, muria antipolitana was transported in Provençal Dressel 16 and Augst 33 as well as in Lyon 3 and 4. Muria hispana has been found in Dressel 16 and in Lyon 3, produced in two different regions. So has liquamen antipolitanum. On the other hand garum antipolitanum is only associated with Dressel 16 and garum hispanum with Lyon 3. Dressel 16 and Lyon 3 transport the largest variety of products: four, the same, but garum is antipolitanum in Provence and hispanum in the region of Lyon. As for garum pompeianum, it has only been found in Lyon 4. Does the geographical name of the various sauces correspond to the origin of the products or to the recipes (‘in the manner of ’)? We notice that concerning the fish amphorae from Baetica, the general term Hispana indicating the country, is never used to qualify the origin of the sauces when they are designated by the name of a town such as Gades or Malaga. We have seen above that the term Antipolitanum probably referred more to a region or a recipe than to the town of Antibes. In that case, since the Lyon 3 and Lyon 4 are produced in the Lyon region, where do the products they transport come from? We have often referred to the bulk importation of fish sauces (henceforth Hispana and Antipolitana), which would then have been transferred into amphorae in the Lyon region in order to be commercialised. This scenario appears difficult to apply to Provencal Dressel 16 amphorae. The question is open. We may also wonder why large sized containers like Lyon 3 and small modules like Lyon 4 (from 44 liters to 9 liters) were used for identical products. In any event, it appears that the amphorae from Lyon were used indifferently for products of different nature, quality or even origin. We must point out that their zone of distribution is situated northwards (Desbat, Dangréaux 1997: fig. 33; Monsieur, De Paepe and Braet 2007) rather than towards the Mediterranean regions, which may be one explanation. 4. Mead We are not used to hearing of mead amphorae. Nevertheless a Lyon 3B type of amphora (Figure 3) found in the legionnaire camp of Bonn shows an explicit painted inscription published as follows by Ulrike Ehmig (Ehmig 2008):

252

The contents of amphorae produced in Gaul in the Imperial period

Dressel 16

Augst 33

ANTIPOLITANA garum muria liquamen

HISPANA

ANTIPOLITANA

muria

muria

Lyon 3

Lyon 4

ANTIPOLITANA muria liquamen

ANTIPOLITANA POMPEIANA garum muria

HISPANA garum muria

Figure 1. Contents of Gaulish fish amphorae: Dr 16, Augst 33, Lyon 3, Lyon 4.

253

Fanette Laubenheimer

Lyon 3

FISH ANTIPOLITANA

muria liquamen

MEAD? HISPANA

aqua mulsa

muria garum

Lyon 4

FISH ANTIPOLITANA

muria

WINE Defrutum

POMPEIANA

garum

Figure 2. Different contents of Lyon 3 and Lyon 4 amphorae. (Aqua) mul(sa) Stillic(idium) Exc(ellens)

Mur(ia) Stillic(idii) Excellens

Excellent mead made with rainwater. Mead is one of the oldest alcoholic beverages in Gaul (Laubenheimer 2015). For the first time there seems to be proof of its commercialisation in amphorae.

On the inscription from Besançon, there is no doubt about the R of Muria but the inscription from Bonn is less obvious and the L may well be a R. In which case the two inscriptions would be similar and would indicate Muria, instead of Aqua Mulsa, made Stilicidii, a characteristic drip method.

Now the recent discovery of a very clear new painted inscription on this type of amphora calls into question Ulrike Ehmig’s interpretation and thereby the presence of mead in amphorae.

It clearly appears that Lyon 3 containers transported very different kinds of products: sauces named from Antibes or Spain of different quality and production. Just like the Lyon 4 could carry defrutum as well as fish sauces.

The titulus picti of a new Lyon 3 amphora, with no rim, from the Musée de Beaux Arts et d’Archéologie of Besançon, reads as follows (Figure 4):

Thus we can deduce that the Lyon 3 and Lyon 4 types from the region of Lyon were indifferently intended to transport various kinds of content; this brings into

254

The contents of amphorae produced in Gaul in the Imperial period

Figure 3. Mead Lyon 3 amphora discovered in Bonn (Ehmig 2008).

Figure 4. Painted inscription indicating Muria Stillicidi on the neck of a Lyon 3 amphora from the Musée des Beaux Arts et d’Archéologie of Besançon (picture J.-L. Daisson). question one of the great principles of amphorology: that the shape of an amphora is characteristic of its’ content. 5. Olives Augst 21 Gaulish amphorae made in Lyon and the Rhône Valley during the Flavian era and the 2nd century are considered to be an adaptation of the older Iberic model: Haltern 70. Twenty-one painted inscriptions testify that they transported, like their model, olives that we know were green: alba or viridis mixed with sweet wine, ex dulcis sometimes described as excellens (Laubenheimer 2004; Ehmig 2007). We note, like for Lyon 3 amphorae, a difference between the areas of production of the content, in this case olives, and the place of production of the amphorae, the area around Lyon does not grow olive-trees. There are never large quantities of Augst 21 amphorae on the markets but they are well distributed in the provinces of North Germany, Brittany and Gallia Belgica.

Figure 5. Gauloise 16 amphora from Meaux, detail of the foot (picture Véronique Brunet). We must mention a new type of Gaulish amphora discovered in Meaux (Seine et Marne) where it was produced, according to physic-chemical analyses. We shall call it Gauloise 16 (Figure 5). The rim shows a double external inflexion, the neck is truncated, the body ovoid and the foot pointed. The handles have a central groove. The general aspect is that of a Lyon 3A except that the foot is different. Instead of being hollow it is full, filled with a roll of clay that is rounded off more or less coarsely rather like the foot of the Haltern 70. Gauloise 16 amphorae appear in the Claudian levels and are still present in the Flavian levels. So far we have no proof of exportation. The variety of influences in the shape of Gauloise 16 brings into question their content: were they fish or mead (?) amphorae like Lyon 3 or rather imported sweet wine and olive amphorae like Haltern 70 or Augst 21? Or were they something quite different? The question remains open.

255

Fanette Laubenheimer

6. Oil? So far we have not found oil amphorae in the South of Narbonensis, land of olive trees. Gaulish olive oil must have been transported in other containers. The surprise came from the North and the East of Gaul and from upper Germania where they produced amphorae, which were meant to be replicas of Baetica oil amphorae (Laubenheimer 2000a; Ehmig 2000). The most characteristic imitations are Gauloise 14 produced in several Eastern workshops with the same profile and capacity as Dressel 20 but with differences in details of Gaulish creation. They were produced in the second half of the 2nd century and the beginning of the 3rd century. Exportations were neither numerous nor distant: limes camps in the Rhine, civilian agglomerations, vici, villae within the proximity of the workshops. The content was not olive oil, which came from Baetica in Dressel 20 amphorae. The hypothesis of walnut oil, put forward a few years ago, remains the most likely; it waits to be confirmed by the analysis of content. 7. Conclusion The question of the content of amphorae, as one can see, is essential to understand the commerce of commodities, the economy of exchange and alimentary habits. Since Dressel’s remarkable study, the corpus of painted inscriptions has considerably increased but our way of thinking has not changed. We tend to associate a shape of amphora to a specific content, in particular when the amphora has no tituli picti left but has a shape that we connect with a certain use. The method seems abusive. Take for example a Lyon 3 amphora that does not contain fish but defrutum, we must realise that we may be mistaken. Fortunately our diagnostics are not all false and research does move forward. Nevertheless a recent development in our knowledge of the content of Gaulish amphorae creates a number of interrogations, and indeed may call into question our previous approach. In Gaul, we have seen the dichotomy between imitation and invention of shape and the diversity of function for one type of amphora. For wine, the copy of well-established existing models gave way to a creative impulse for new shapes. Why so many, often for the same products? Conversely, why was the same container used for so many different qualities?

Why are we so embarrassed by the difference between the zones of production of the amphorae and those of their contents? I refer to the Narbonensis amphorae that contained wine from Béziers as well as Massicum. I also refer to the production of amphorae in the Lyon region or the Rhone Valley, far from the fishing zones or olive groves. I furthermore refer to the Provencal amphorae that contained fish sauces called Hispana. Why do we find so many different recipes of sauces, from Antibes or from Spain, in the same containers? The use of amphorae from the Lyon region for the transport of different ingredients is now evident. It is probably related to the fact that their zone of distribution was directed exclusively towards Northern countries, which differentiates them from Mediterranean amphorae. Thus it appears that the study of the contents of amphorae must be closely connected to that of their distribution. Once again, why is there such a variation between the shape of an amphora and its content? Remember the strange Gauloise 16 from Meaux or the Gauloise 14, copies of Dressel 20 from the Northeast. Finally there are all the minor productions of amphorae for which we know nothing of the content but that we try to imagine. What is the connection between the vineyards of Normandy and Gauloise 12 or between the vineyards of the three Gauls and the workshops that produced amphorae, which we presume were for wine? Admittedly we still have a number of doubts but we are progressing by extending our way of thinking beyond the principle that prevailed for such a long time: that one type of amphora corresponded to one content. Progress in the analysis of content will certainly improve our research work because we will have good samples at our disposal and sufficient means to analyse them. The recent enlargement of the corpus of painted inscriptions on Gaulish amphorae has shaken our certitudes. Is the case of Gaulish amphorae a peculiarity? Acknowledgment: I am very grateful to Amanda Paquin for the translation of this article.

256

The contents of amphorae produced in Gaul in the Imperial period

Bibliography Berthault, F. 1992a. Production d’amphores dans la région bordelaise. In F. Laubenheimer (ed.), Les amphores en Gaule - I. Production et circulation. Annales Littéraires de l’Université de Besançon 474, série Amphores: 93– 100. Paris, Les Belles Lettres. Bertucchi, G. 1992b. Les amphores et le vin de Marseille VIe s. av. J.-C.-IIe s. ap. J.-C. Revue archéologique de Narbonnaise Supplément 25. Paris, CNRS. Bigot F., 2020. Production et commerce des amphores gauloises sur le littoral de la Narbonnaise, Toulouse, éd. Mergoil. Desbat, A. and Dangréaux, B. 1997. La production d’amphores à Lyon. Gallia 54: 73–104. Desbat, A., Picon M. and Schmitt, A. 1994. Sur l’origine de quelques amphores d’Augst. In S. MartinKilcher (ed.), Die römischen Amphoren aus Augst und Kaiseraugst. Forschungen in Augst 7, 3: 649–656. Augst, Römermuseum Augst. Ehmig, U. 2000. Dressel 20/30 : Ex Baetica originalis imitatio ex Germania superiore. In Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Eć ija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 1143-1152. Ecija, Gráficas Sol. Ehmig, U. 2007. Tituli picti auf Amphoren in Köln. Kölner Jahrbuch 40: 215–322. Ehmig, U. 2008. Aqua mulsa, Ein exzellenter Topfen Met. Bonner Jahrbücher 208: 62–72. Laubenheimer, F. 1985. La production des amphores en Gaule Narbonnaise sous le Haut-Empire. Annales Littéraires de l’Université Besançon 327. Paris, Les Belles Lettres. Laubenheimer, F. 1989. Les amphores gauloises sous l’Empire : recherches nouvelles sur leur production et leur chronologie. In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22-24 mai 1986). Collection de l’École française de Rome 114: 105–138. Rome, École française de Rome. Laubenheimer, F. 1991. Les vides sanitaires et les amphores de la Porte d’Orée à Fréjus (Var). Gallia 48: 229–265. Laubenheimer, F. 1992. Production d’amphores à Fréjus. In F. Laubenheimer (ed.), Les amphores en Gaule. Production et circulation. Annales Littéraires de l’Université de Besançon 474: 15–24. Besançon, Université de Besançon. Laubenheimer, F. 2000a. Imitations d’amphores à huile de Bétique dans l’Est et le Nord des Gaules et en Germanie supérieure. In Congreso internacional Ex Baetica Amphorae : conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 1121–1142. Ecija, Gráficas Sol. Laubenheimer, F. 2000b. Les amphores. In G. Berthaud (ed.), Mazières-en-Mauges gallo-romain (Maine-et-Loire), un quartier à vocation artisanale et domestique: 60–63. Angers, ARDA.

Laubenheimer, F. 2001. La standardisation des amphores Gauloise 4, des ateliers de Narbonnaise à la production de Denia (Espagne). In F. Laubenheimer (ed.), 20 ans de recherches à Sallèles d’Aude: 33–50. Besançon, Presses Universitaires Franc-Comtoises. Laubenheimer, F. 2003. Amphorae and vineyards from Burgundy to the Seine. In J. Plouviez (ed.), Amphorae in Britain and the western Empire. Journal of Roman Pottery Studies 10: 32–44. Laubenheimer, F. 2004. Inscriptions peintes sur les amphores gauloises. In M. Feugère and P.-Y. Lambert (eds), Dossier, L’écriture dans la société gallo-romaine. Gallia 61: 153–192. Laubenheimer, F. 2015. Boire en Gaule. Paris, CNRS. Laubenheimer, F. and Bonnet, C. 1998. Assainissements place des Célestins à Lyon. In F. Laubenheimer (ed.), Les amphores en Gaule II. Production et circulation : 205–235. Besançon, Presses Universitaires FrancComtoises. Laubenheimer, F. and Marlière, E. 2010. Échanges et vie économique dans le Nord-Ouest des Gaules. Besançon, Presses universitaires de Franche-Comté. Laubenheimer. F. and Humbert, S. forthcoming, Échanges et vie économique en Franche-Comté dans l’Antiquité. Le témoignage des amphores chez les Séquanes. Besançon, Presses universitaires de Franche-Comté. Laubenheimer, F. and Schmitt, A. 2009. Amphores vinaires de Narbonnaise, Production et Grand Commerce, Création d’une base de données géochimique des ateliers. Travaux de la Maison de l’Orient et de la Méditerranée 51. Lyon, Maison de l’Orient et de la Méditerranée. Laubenheimer, F., Cantin, N. and Schmitt, A. 2005. Vin et vignoble du Centre-Ouest de la Gaule au Ier s. Le témoignage des amphores Pascual 1 et Dressel 2/4 blanches à pâte orangée à cœur gris. In SFECAG, Actes du Congrès de Blois: 377–382. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Laubenheimer, F., Odiot T. and Leclère H. 1989. Sous Auguste, un atelier de potiers italianisant à SaintJust (Ardèche). In M. M. Mactoux and E. Geny (eds), Mélanges P. Lévêque, 2: 295–329. Paris - Besançon, Université de Besançon, Centre de Recherches d’Histoire Ancienne. Mauné, S. 2013. La géographie des productions des ateliers d’amphores en Gaule Narbonnaise pendant le Haut-Empire. Nouvelles données et perspectives. Revue archéologique de Narbonnaise 46: 335–374. Monsieur, P., De Paepe, P. and Braet, C. 2007. Lyon Amphorae in the North: Studies in Distribution, Chronology, Typology and Petrology. In S. Y. Waksman (ed.), Archaeometric and Archaeological Approaches to Ceramics. British Archaeological Reports International Series 1691: 103–111. Oxford, Archaeopress.

257

Italian and Sicilian amphorae and their contents: a general overview Simonetta Menchelli University of Pisa

Abstract: This paper deals with Italian and Sicilian amphorae and their contents. The methodology takes into account some general issues: episodic/systematic recycling; planned use of certain amphorae forms for foodstuffs different from those they would usually contain; widespread use of small and non-ceramic containers. These issues need to be considered for fish products and olive oil in particular, as the Italian amphorae containing these foodstuffs were chronologically and quantitatively limited according to the archaeological evidence, while literary works and the resources of the Italian sea and countryside, from the ancient times up to present day have given a different picture. On the contrary, as is known, the production of Italian wine is archaeologically well documented and recent research has increased the number of known workshops of labelled and non-labelled wine amphorae. The importance of archaeometric studies applied to the contents of amphorae has been highlighted in the hope that this type of scientific analysis will become standard procedure in future for archaeological research. Key words: Italian and Sicilian Amphorae; fish products; olive oil; wine.

The topic I have to deal with — the production and trade of foodstuffs contained in Italian and Sicilian amphorae — is very complex. In order to discuss this subject I think first of all it would be appropriate to distinguish between: A. The main contents: what has been established in studies as the foodstuff usually associated with a given amphora form (regarding the multifaced relationships between amphorae and their contents see in particular Manacorda 2019: 39–41). B. The alternative goods: substances different from those usually contained in that amphora form. In the latter case it is very difficult to distinguish between the following instances: B1. episodic recycling, not necessarily connected with a local production; B2. more frequent/systematic recycling, most probably evidence of manufacturing activities, which at the moment is faint but which could be identified by further research; B3. planned use of specific types (in terms of dimensions or other peculiarities) in a given amphora form. Obviously the episodic recycling is very difficult to interpret, we label it ‘episodic’ because we have no tools for including it in a general model or in a productive system. It is certain that vessels, and in particular amphorae might have had a longer life beyond the obvious phases (productiondistribution-consumption-discarding) as pointed out in many studies (Peña 2007; Peña 2014 and in this book; Hahn and Weiss 2013; Abdelhamid 2013; Giannichedda 2014). Therefore we should take this possibility into consideration when we try to reconstruct the volume of business in the Roman economy.

Regarding the most frequent/systematic recycling, and particularly in the case of fish products, we have a lot of evidence both along the Tyrrhenian and Adriatic coasts. In this case the most difficult task is determining whether the amphorae were frequently/continuously reused or intentionally produced for a content different from the usual one. In these instances, archaeometric analyses of the contents provide the decisive answer, and therefore it is to be hoped that they will become part of standard procedure for archaeological research. As is well known, the Italian fish amphorae were: • the Tubular amphorae produced in Sicily in the Late Republican period in which tuna was exported, mostly to the Eastern part of the Mediterranean (Botte 2012) (Figure 1.1). • the Dressel 21-22 amphorae, recently classified in the following types: Botte 1 (manufactured in Sicily, early-late 1st century AD), Botte 2 (produced in the Calabrian area, early-late 1st century AD), Botte 3 (manufactured in Campania, mid-late 1st century BC.late 1st century AD); (Figure 1, respectively 2-4). • the small Adriatic amphorae type Grado (2nd century AD-early 3rd century) (Figure 1.5) and similar forms (anfore con orlo a fascia (1st century-3rd century AD) (Figure 1.6) (Carre, Pesavento and Belotti 2009; and below Carre and Pesavento). Literary sources cite fishing and fish-processing in numerous sites along the Italian and Sicilian coasts (see Curtis 1991: 85–101; Botte 2009b: 42–48; Marzano 2007; Marzano 2013), consistent with current fish-related activities and gastronomic traditions. On the contrary, the circulation of these amphorae was chronologically

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 259–272

Simonetta Menchelli

Figure 1. (1) Tubular amphora (Botte 2012: fig. 4.2); (2) Botte 1a amphora (Botte 2009a: fig. 4, 43167); (3) Botte 2 amphora (Botte 2009a: fig. 7, 41418); (4) Botte 3 amphora (Botte 2009a: fig. 8 , 25430); (5) Grado type amphora (Carre, Pesavento and Belotti 2009: fig.5); (6) Orlo a fascia amphora (Carre, Pesavento and Belotti 2009: fig. 9C).

260

Italian and Sicilian amphorae and their contents: a general overview

Figure 2. Funerary stele of argentarius L. Calpurnius Daphnus (Marzano 2013: fig. 42). (as a whole mid-2nd century BC- early 3rd century AD), geographically and quantitatively very limited. It is evident that, apart from the above-mentioned forms, the Italian fish products were contained and sometimes traded in other types of amphorae, and that urcei, pots (see for example Djaoui, Piquès and Botte 2014) and non-ceramic containers (goat-skins, wooden barrels) were also utilized. For example the consumption of fresh fish was a status marker in the Roman world and in particular in Rome where literary and epigraphic sources document astronomical prices for large and highly sought-after fish, coming from the nearby coasts, which were sold at auctions in which argentarii were available to advance the considerable cash needed for the purchase (Figure 2; Marzano 2007: 307–310; Marzano 2013: 283–286). The way in which some of this precious fish was transported can be documented by the funerary stele of argentarius L. Calpurnius Daphnus (1st century AD), (Figure 2) who is depicted flanked by two men carrying — on their shoulders — heavy wooden baskets, whose contents are explained by the fish held by the figure on the left and by Daphnus himself (Marzano 2013: 286). Starting from the Tyrrhenian Sea, and particularly from Etruria and the ager Cosanus, Strabo referred to a tunny watch (thynnoskopeion: Strabo, V, 2, 8) and for some time now fisheries have been identified in the area. The interpretation of the one in the Cosa Harbour (McCann 1987) is not universally accepted (Botte 2009b: 43–44), while the fish-processing workshops in the Feniglia tombolo between Cosa and Monte Argentario (Cavallo, Ciampoltrini and Shepherd 1992) and in the islands of Giglio and Giannutri (Rendini 2003) are widely acknowledged.

As regards the containers of the Cosan fish-products, F. Laubenheimer (2007: 75–78) identified, among the Albinia amphorae, a production of the Dressel 1C which were characterized by 3-letter stamps on the lip and differently distributed from the wine Dressel 1 (Figure 3). She hypothesized that these amphorae might have transported fish products, also on the basis of the comparison with Baetican productions, for example the Dressel 1C bearing the well-known stamp SCG (Societas Cetariorum Gaditanorum) found both at Baelo, famous for its fisheries, and at El Rinconcillo near Algesiras, together with the fish amphorae Dressel 7/11, Dressel 12 and Beltrán II a. (Laubenheimer 2007: 77; Étienne and Mayet 1994: 131–138; García Vargas and Bernal‑Casasola 2008: 668). The Laubenheimer hypothesis could be borne out by the Albinian Dressel 1 amphora which contained the remains of a single species of tuna, Thunnus thynnus (Costantini 2007): at least 11 young tunas had been carefully slaughtered, as documented by the many small underthroat pieces (De Grossi Mazzorin 2006). The analysis of the contents of the amphora body carried out by A. Pecci (2006) has identified large quantities of cholesterol, the marker of animal fat, confirming the presence of tunas which contain copious amounts of this steroid. The analyses seem to exclude the presence of olive oil or traces of tartaric acid, a wine marker. Therefore this amphora never contained wine, it was not recycled but intentionally used for a foodstuff different from the one usually associated with the Dressel 1 amphora. The fact that this container, which originally must have been in a storage room, has been found in a Medieval layer (10th‑12th centuries: Costantini 2006: 169) does not seem to affect its documentary value.

261

Simonetta Menchelli

Figure 3. Dressel 1 C amphorae from the Albinia Workshop (Laubenheimer 2007: 77). There remains the problem of whether the tunas in this amphora, whose fabric would appear to be Albinian, came from the Cosa area or whether the amphora was recycled in Populonia, to contain local tunas. In fact tuna fishing and processing in this area are also documented by archaeological data (tunny fishing works near Punta delle Tonnarelle: Shepherd 2003; Shepherd and Dallai 2003; a cetaria in Poggio al Mulino: Genovesi and Megale 2016) and literary sources (Strabo — V, 2, 6 — referred to a thynnoskopeion; Rutilius Namatianus — 379-380 —mentioned vivaria at Falesia, directly East-South-East of Populonia). The use of the Dressel 1 amphorae for fish products was probably quite widespread in Northern coastal Etruria. For example in the Portus Pisanus area in the Vallinbuio workshop, which produced Dressel 1 amphorae (Cherubini, Del Río and Menchelli 2006), one of these was found to have been filled with whole or sliced fish of the Centrocanthus cirrus and Spicara smaris species, of the Centracanthidae family (Bulzomì 2013). These species

were very common in the Italian seas and generally in the Mediterranean, for example they were also abundantly documented in the Bottega del garum in Pompeii (I,12, 8) (Carannante 2008-2009). Along the Northern Etruscan coasts the pots were also used to contain fish-derived products: the item displayed in Pisa, in the Cantiere della Navi Antiche (see Cibecchini in this volume, fig. 6) presents morphologic and technical aspects, in particular the carbonatic fabric-peculiarities typical of the local coated wares (see Menchelli and Pasquinucci 2012; about this topic Picchi and Menchelli 2018: fig. 3, 9-10). The context where it was found, in Pisa’s urban port of call, suggests a similar use to that of the Latial pots studied by Djaoui, Piquès and Botte (2014). These vessels, (forms Ostia II, 401 and similar jugs), evidence of the fish-processing along the Latial coasts, are dated from the Augustan to Antonine times, with the greatest presence in the Flavian period. Due to their

262

Italian and Sicilian amphorae and their contents: a general overview

Figure 4. Campanian urcei and pots. (1-4) urcei (Cappelletto et al. 2013: fig.10); (5) : Schöne Mau 1 pot (De Caro 1987: fig.48, 72); (6-9) Gasperetti 1996 1213a-d pots (Gasperetti 1996: fig. 2, 15-18). limited capacity (from 0.30-0.45 litres to a maximum of 1.70-2.25 litres) they have been considered unsuitable for large-scale export activities and more appropriate for a redistribution trade of fresh small fish intended for the sailors working in the route from Ostia to Arelate. Baetican pots for fish-derived products (allec or salsamenta) have also been identified (Piquès et al. in this volume): evidently this must have been a widespread practice along the Mediterranean coasts, at least along the Western ones. Up to now we have not found archaeological data for Sardinian fish pots and amphorae but the evidence of salt production and alieutic activities (salteries, tunny fishing works, tanks for fish-processing; toponomastic information) is substantial (collected by C. Nervi 2014) and therefore we can presume that the local gastronomic products, e.g. the bottarga, derived from ancient traditions. Regarding the production and trade of garum in Pompeii and generally in Campania is difficult to add anything to

Botte’s studies and very recent ones by D. Bernal‑Casasola et al. (2014). The pile of amphorae (Dressel 21‑22/Botte 1, Botte 2 and Botte 3 types) preserved in the Bottega del Garum is an extraordinary finding. As is well known these amphorae were upside-down, empty and ready to be filled with fish-products. The Calabrian Botte 2 amphorae were the most numerous (69 items out of 81). Moreover, some small fish salting workshops, built in the 2nd century BC. — interpreted as a ‘cottage production’ — have recently been found in Pompeii, near Porta Stabia, and others have been identified near Porta Ercolano (Ellis 2011: 59–88). Concerning Umbricius Cordo, who most probably carried out various activities including importing Spanish garum, and the Campanian urcei, (Figure 4.1-4) we know that these vessels, which mostly contained about 3 litres, appear to have had a distribution restricted to the Vesuvian area and, as stressed by D. Bernal‑Casasola, they were not suitable for long-distance trade (Cappelletto et al. 2013). Furthermore, epigraphic and archaeological evidence documents the generalized use of pots (particularly the Schöne Mau 1 — Figure 4.5 — and Gasperetti 1213a-d forms — Figure 4.6-9 — for the

263

Simonetta Menchelli

local-regional and perhaps wider transport of more solid fish products (Picchi, Menchelli 2018: 404–406). As regards Bruttium, Pliny extolled, among the places in the Empire whose garum production was famous, Thurii, the only town in Italy involved in alieutic activities apart from Pompeii (Pliny, NH, 31, 94). In the early Imperial Age, the Dressel 21-22/Botte 2 forms were the regional fish-amphorae produced along both the Tyrrhenian and the Ionian coasts (Sangineto 2006: pl. 81, no.1-6, 314; Luppino and Sangineto 1992: pl. 38, 20; Botte 2009a: 158; Rizzo 2014: 136) and widely exported to Pompeii, as mentioned above. In Bruttium fish-processing also continued in Late Antiquity, as documented by the Monasterium vivariense founded by Cassiodorus, south of Scolacium in the mid‑6th AD, whose vivaria guaranteed the intensive farming and sale of fish (Cass., Inst., 29, see Ianelli and Cuteri 2007). Therefore, as suggested by D. Bernal‑Casasola (2010: 19–20), most probably, at least a part of the Calabrian Keay 52 amphorae might have been used for transporting fish. In Sicily there is a lot of evidence indicating the presence of many fish-processing workshops and we have already referred to the Sicilian Dressel 21/Botte 1 amphorae, of which some workshops have been identified at Alcamo Marina and Solunto (Botte 2009b: 121–131; Rizzo 2014: 147–148).

Turning to the oil amphorae, as is well known Pliny said that Italian olive oil was the best (Pliny, NH, XV, III, 2); olive trees have always been one of the most typical elements in Italian landscapes and Roman oil presses have very frequently been found in the whole Peninsula, mostly connected with villas (Fabiani 2019; Van Limbergen 2016). In spite of all this, the association of amphora forms and olive oil is only documented in the Adriatic area of the Peninsula with the Brindisian ovoid (Figure 6.1) , Dressel6B (Figure 6.2) amphorae and those with funnel-shaped necks (Figure 6.3) Obviously I will not be talking about these amphorae which have been the subject of important studies (respectively Manacorda 2012 and 2019; Cipriano 2009; Mazzocchin 2009; Carre and Pesavento in this volume; García Vargas et al. 2019). I will just stress some aspects: There were no specific olive-oil amphorae in Tyrrhenian Italy but, between the 1st cent. BC and the Augustan Age, the ‘Brindisian’ ovoid amphorae were produced in some western districts. As is well known, within their complex morphological group, Ovoid Amphorae, at least those having a short neck and a large rim, could have been used as olive oil containers (see the remarks of Manacorda 2019, 40). These Tyrrhenian workshops have been identified in the following sites: Etruria:

As regards the Adriatic area, side by side with the production of specific fish amphorae, the reutilization of the Dressel 6a wine amphorae and the Dressel 6b oliveoil amphorae is well documented especially in Cisalpina (Carre, Pesavento and Belotti 2009; Carre and Pesavento in this book).

Albinia Some amphorae, 2 of which were whole, found in the Albinia workshop, are considered locally produced and dated between 125-75 BC (Benquet and Mancino 2007: 55; Benquet and Capelli 2019: 275–280).

In conclusion, regarding the relationship between fish products and amphorae, in Italy we can confirm both the episodic/systematic recycling of other forms of amphorae, and perhaps the planned use of specific types (in terms of dimensions or other peculiarities) in a given form, such as the Albinian Dressel 1C and the small Dressel 6b produced in Loron, as suggested by Y. Marion (2009: 286) (Figure 5). Moreover the use of pots, goat-skins and wooden containers must have been part of daily life, as documented by literary, epigraphic and iconographic sources.

Le Grottacce Site (the area between Anzio and Torre Astura): ovoid amphorae have been found in an amphorae workshop connected to a large villa (De Haas, Attema and Pape 2008: 547–548)

In any case Italian alieutic production did not have a strong in terms of exportation on the Mediterranean economy, judging by the restricted quantity and circulation areas of the Italian fish amphorae. Moreover, the low level of Italian productivity is reflected in the limited infrastructure capacity of the workshops known at the moment, much lower than that of Spain, Tunisia and Brittany, listed by E. Wilson (2007).

Latium-Campania:

Torre Astura: a workshop which produced, side-by-side, Graeco-Italic and ovoid amphorae has been discovered through surveys (Empereur and Hesnard 1987: 35). Fondi: ovoid amphorae have been collected in the locality of Canneto, stamped by P. Veveius Papus who also stamped Brindisian ovoid forms (Empereur and Hesnard 1987: 35; Manacorda 1994, 37-38). Liri Valley: Giancola 6 amphorae stamped by M. Tuccio Galeo identified as the M.Tuccio mentioned in a letter from M. Caelius Rufus to Cicero in 51 BC (Ad Fam.VIII, 8, 1) and as the Galeo who died in 47 BC and had named Cicero as his heir (Ad Att. XI, 12, 4) (Cipriano and Carre 1989: 74–77, fig.7)

264

Italian and Sicilian amphorae and their contents: a general overview

Figure 5. Small Dressel 6b amphorae from Loron (Marion 2009: fig.2). Most probably these amphorae were of Brindisian/ Salentine origin as maintained by D. Manacorda (1994: 32–37; 2012: 161–162), but epigraphic data found along the Tyrrhenian coast (evidence of the gens Tuccia and references, in the stamps, to the tribus Tromentina which included the inhabitants of Fabrateria Nova) and petrographic results suggest that these amphorae were mainly manufactured along the strip of land between Southern Latium and Northern Campania (see Rizzo 2014: 107–108 and the cited bibliography). Alliphae A group of ovoid amphorae, most probably locally produced, have been identified in Alife (Caserta) (Di Mauro 2011).

Bruttium: Blanda Iulia, here the ovoid amphora production appears to be linked with the business of Postumus Curtius whose workshops were responsible for a very variegated manufacturing activity (Dressel 2-4, Tyrrhenian Lamboglia 2, Dressel 21-22 and ovoid forms: Panella 2010: 48). It is evident that the Roman ruling class was deeply involved in these productions, both the senatores and equites (Panella 2010: 48) who, like M.Postumus Curtius, M. Tuccio Galeo, and P. Veveius Papus, aside from the praedia in Apulia-Calabria, might have owned others in various Italian regions and therefore carried out their activities with the same systems and tools.

265

Simonetta Menchelli

Figure 6. (1) Ovoid amphora from the Giancola Workshop (Manacorda 2019: fig.1, 3); (2) Dressel 6b amphora (Panella 2001: 17); (3). Funnel-shaped neck amphora (Panella 2001: 19); (4) Flat-bottomed amphora, Type 3,1 from the Vingone Workshop (Martelli 2008: fig.116); (5) Flat-bottomed amphora from the Naxos Workshop (Ollà 2001: 54 no.2). These ovoid amphorae were no longer produced after the Augustan Age, so there remains the problem of the lack of archaeological identification of the olive-oil containers in vast sectors of Italy, for many centuries. Moreover, also with regard to Histria, the richest region in terms of documentation, since the Dressel 6b amphorae were not produced after the 4th century AD, we should assume that there were other kinds of olive-oil containers in the subsequent centuries, at least up to the mid-6th century when Cassiodorus described the region as still being very fertile and productive (Cassiodorus, Variae, 12, 22, 4: gravidam vini olei vel tritici).

In the last few years the archaeometric analyses carried out on the olive-oil amphorae contents have provided a lot of new data, for example another kind of vegetable oil, castor oil, was identified in some African spatheia and in cylindrical amphorae from the Port of Classe (Pecci et al. 2010; Garnier in this book). Furthermore, a very important datum is that resin was also used for coating oil amphorae (Pecci et al. 2010; Pecci, Salvini and Cantini 2010). This result is becoming increasingly common and it will change the system of classifying amphorae normally used, especially if

266

Italian and Sicilian amphorae and their contents: a general overview

fragmented, since until now it was considered impossible that resin-lined amphorae could have been utilized to transport olive oil. In any case, this new information must be used with caution: the remains of pitch are hardly ever visible in containers which probably transported oil, perhaps because the decomposition of the pitch in the oil was absorbed in the pores of the pottery. Therefore, the amphorae presenting clearly visible remains of thick pitch lining should have contained goods other than olive oil (for this topic see Bonifay 2018: 332). Regarding the wine amphorae produced in Italy, the bibliography, even the recent one, is enormous and I will just mention the works by G. Olcese (2012, 2013) and G. Rizzo (2014), thanks to whom we can see a great increase in the number of workshops in the whole Peninsula and Sicily, both in the most famous wine districts, referred to by Ancient authors, and in the areas with non-labelled wines. Moreover, C. Panella (2010) has some years ago written a paper of fundamental importance about the productive boom of Roman Italy and the role of wine and wine amphorae in this new economy. This paper confirms that amphorae models were first produced in the Campanian area and subsequently imitated in the whole of Italy and the Provinces and I wonder whether, besides Dressel 1 and Dressel 2-4 amphorae, the flat-based small amphorae model might not also have had a Campanian origin. As is well known, it is an old hypothesis of P. Arthur (1991) who maintained that in the Augustan Age in the Garigliano basin, in the Masseria S. Donato workshop (located near Suessa Aurunca), Dressel 2-4 bearing the stamp MAESCELS and Dressel 28/Oberaden 74 amphorae were produced side-by-side (Arthur 1991: 74–76; see also Tchernia 1996: 208). According to the author, these flat-bottomed forms or similar ones, were manufactured in this area as early as the beginning of the 1st century BC as they were found in the Spargi shipwreck. Now we know that the dating of these amphorae in the early 1st century BC can be corroborated by other finds (Albenga and Sant’Andrea shipwrecks; Albintimilium, Portus Cosanus, Cosa and ager Cosanus: see Menchelli and Picchi 2016: 229–230). A. Tchernia considered this hypothesis ‘envisageable’ but lacking in corroborative evidence (Tchernia 1996: 208). In 2001 C. Panella wrote that the flat-bottomed model derived from Late-Republican prototypes, citing the chronological evidence provided by N. Lamboglia (1952; 1955) but, analyzing Paul Arthur‘s arguments, she came to the conclusion that it was not possible to determine a chronological priority among the productions of these amphorae in Campania, Narbonensis and Tarraconensis (Panella 2001: 184, 214, note 30). But now, we can try to support P. Arthur’s idea, thanks to amphorae workshops recently excavated in Italy which appear to have produced the flat-based amphorae in the last decades of the 1st century BC. They are:

1. Vingone near Florentia (Florence-Tuscany: Martelli 2008; Shepherd 2015) (Figure 6. 4); 2. Ca lo Spelli in the ager Pisanus (Livorno, Tuscany: Picchi et al. 2010); 3. Various centres in the coastal ager Volaterranus (Livorno, Tuscany: Cherubini and Del Río 1997); 4. Albinia near Cosa (Tuscany: Benquet, Vitali and Laubenheimer 2013); 5. Naxos near Tauromenion (Sicily: Ollà 2001) (Figure 6.5). The Gaulish amphorae do not seem to have been produced before the early Augustan period, and both the Tarraconensis Oberaden 74, recently analyzed by Carreras and González (2012), and the Baetican flat-based amphorae designated as urcei by R. Morais (2007) can be dated to the same period. Therefore, it can be thought that this model derived from an experimental phase in the Central Tyrrhenian area of Italy, where the flat-bottomed amphorae appear to have been manufactured from the beginning of the 1st century BC. This basic model spread increasingly and became standardized, with the obvious regional variants, in Italy and in the Western Provinces, in the late-Republicanearly Imperial period, linked with Augustan colonization and his policy of conquest, particularly because of the requirements of the soldiers and veterans. In the following decades and centuries many other amphorae characterized by flatted-bottoms or small spikes were manufactured in Italy (Spello, — Figure 7.1 — Forlimpopoli — Figure 7.2 —, Empoli — Figure 7.3 —, Keay 52 — Figure 7.4 — forms) and in the Provinces, due to different regional and — in general — historical and economic factors, and all of them kept wine as the first food-stuff up to the latest Calabrian and Sicilian productions (Casalini 2014; Franco and Capelli 2014a and 2014b). There is not enough space to talk about the foodstuffs exported to Italy from the Provinces and in any case this would be too strenuous a task. At any rate, from the methodological point of view, I think we should take into careful consideration the regional aspects and possible regional differences in the consumption of imported wares. We should try to get statistically reliable data about the highest possible number of districts, applying a bottom-up approach, without uncritically accepting general models (for example the whole of Italy’s general dependence on provincial goods). In fact, the comparative analysis of the results of two survey projects — one dealing with the ager Pisanus and the other with the ager Firmanus — has led to very different results, even though both these districts were deeply involved in the Mediterranean trade-routes as documented by the imported cooking and table wares. Concerning the 1st-5th centuries AD period, in the Pisan area the imported amphorae from other Italian regions and provinces were very numerous (86 out of a total of 246

267

Simonetta Menchelli

Figure 7. (1) Spello type amphora (Panella 2001: 12); (2) Forlimpopoli type amphora (Panella 2001: 21); (3) Empoli type amphora (Panella 2001: 22); (4) Keay 52 amphora (Keay 1984: forma LII). items = 35%), while in the ager Firmanus they only amount to 15% (145 out of a total of 991 items) (Menchelli and Picchi 2014; Menchelli 2014). In the case of the latter, as in many other Italian districts, especially in Late Antiquity, no amphorae models were produced specifically for fish or olive oil, therefore, given such scant imports, we must suppose a local/sub-regional supply of foodstuffs in non-ceramic containers and/or in pots and recycled amphorae of other forms. In conclusion, in my opinion, we have a much more complex framework than we could possibly have foreseen. The development of research, the interdisciplinary

approach, and the internet communication system have provided us with an exceptional amount of data, and at the same time have complicated or destroyed many ideas well established : for example we now know that fish amphorae were not necessarily resin coated, while the olive oil amphorae might have been coated. The imitation of foreign amphorae forms appears to have been widespread throughout the whole of the Roman world. A strict correspondence between forms and contents is not always certain but we know that within large families of amphorae (such as the Ovoid ones) typological details could be markers of different transported goods (see Bernal‑Casasola, Pecci and Sáez Romero 2019: 400).

268

Italian and Sicilian amphorae and their contents: a general overview

In any case, the original content could have been mixed, for example olives and defrutum, as in the case of the Haltern 70 forms, and the amphorae might be reused innumerable times, for different foodstuffs in different areas of the Roman world. Moreover, part of Roman trade (on how big a scale and which products?) was carried out using small and/or non ceramic containers which therefore elude our research. Obviously we cannot expect to verify all these processes, but we should bear them all in mind when we use amphorae to try to reconstruct historical frameworks, in order to define, as reliably as possible, the relevant economic and social trends. Bibliography Abdelhamid, S. 2013. Against the throw-away-mentality: The Reuse of Amphorae in ancient maritime transport. In H. P. Hahn and H. Weis (eds), Mobility, Meaning and transformations of Things: 91–106.Oxford, Oxbow Books. Arthur, P. 1991. Romans in Northern Campania. Archaeological Monographs of the British School at Rome 1. London, British School at Rome. Benquet, L. and Capelli C. 2019. The Italic ovoid amphorae in the Toulouse area at the end of the Iron Age (HauteGaronne, France). In E. García Vargas, R. R. de Almeida, H. González Cesteros and A. M. Sáez Romero 2019 (eds), The Ovoid Amphorae in the Central and Western Mediterranean: Between the Last Two Centuries of the Republic and the Early Days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13: 274–297. Oxford, Archaeopress. Benquet, L. and Mancino, C. 2007. Le anfore di Albinia: primo saggio di classificazione. In D. Vitali (ed.), Le fornaci e le anfore di Albinia. Primi dati su produzioni e scambi dalla costa tirrenica al mondo gallico. Atti del Seminario Internazionale (Ravenna 2006): 55–66. Bologna, Università di Bologna. Benquet, L., Vitali, D. and Laubenheimer, F. 2013. Nouvelles données sur l’atelier d’amphores d’Albinia (Orbetello, Italie): campagnes de fouille 2003-2006. In F. Olmer (ed.), Itinéraires des vins romains en Gaule, IIIe-Ier siècles avant J.-C., confrontation de faciès. Actes du colloque européen organisé par l’UMR 5140 du CNRS, Lattes, 30 janvier - 2 février 2007: 513–529. Lattes, Centre National de la Recherche Scientifique. Bernal-Casasola, D. 2010. Iglesia, producción y comercio en el Mediterráneo tardoantiguo. De las ánforas a los talleres eclesiásticos. 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): 19–31. Oxford, Archaeopress.

Bernal-Casasola, D., Pecci, A. and Sáez Romero, A. M., 2019. Preliminary organic residue analysis of few Ovoid 1 and Ovoid 5 amphorae from the Guadalquivir valley. In E. García Vargas, R. R. de Almeida, H. González Cesteros and A. M. Sáez Romero 2019 (eds), The Ovoid Amphorae in the Central and Western Mediterranean: Between the Last Two Centuries of the Republic and the Early Days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13: 391–402. Oxford, Archaeopress. Bernal-Casasola, D., Cottica, D., García Vargas, E., Toniolo, L., Rodríguez Santana, C., Acqua, C., Marlasca, R., Sáez, A. M., Vargas, J. M., Scremkin, F. and Landi, S. 2014. Un contexto excepcional en Pompeya: la pila de ánforas de la Bottega del Garum (I, XII, 8). Rei Cretariae Romanae Fautorum Acta 43: 219–232. Bonifay, M. 2018. The distribution of African pottery under the Roman Empire: evidence versus interpretation. In A. Wilson and A. Bowman (eds), Trade, Commerce and the State in the Roman World: 327–351. Oxford Studies on the Roman Economy. Oxford, Oxford University Press. Botte, E. 2009a. Le Dressel 21-22: anfore da pesce tirreniche dell’alto impero. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 149–171. Rome, Quasar. Botte, E, 2009b. Salaisons et sauces de poissons en Italie du Sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard, 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Botte, E. 2012. L’exportation du thon sicilien à l’époque tardo-républicaine. Mélanges de l’École française de Rome. Antiquité 124-2: 577–612. Bulzomì, F. 2013. Vallin Buio. In M. Pasquinucci (ed.), Guida archeologica delle coste livornesi. Porti antichi, vita quotidiana, rotte mediterranee: 101. Florence, Nardini. Cappelletto, E., Bernal-Casasola, D., Cottica, D., Bustamante, M., Lara Medina, M. and Sáez, A. M., 2013. Urcei per salse di pesce da Pompeii-Ercolano: una prima analisi. In D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz and A. M. Sáez (eds), Hornos, talleres y focos de producción alfarera en Hispania. Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (I): 271– 280. Cadiz, Universidad de Cádiz. Carannante, A. 2008-2009. L’ultimo garum di Pompeii. Analisi archeozoologiche sui resti di pesce dalla cosiddetta ‘Officina del garum’. AUTOMATA, Journal of Nature, Science and Technics in the Ancient World 3-4, 1: 43–53. Carre, M.-B., Pesavento, S. and Belotti, C. 2009. Le anfore da pesce adriatiche. In S. Pesavento Mattioli and M.B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 215–238. Rome, Quasar.

269

Simonetta Menchelli

Carreras, C. and Gonzalez, H., 2012. Ánforas tarraconenses para el limes germano: una nueva visión de las Oberaden 74. In D. Bernal-Casasola and A. Ribera i Lacomba (eds), Cerámicas Hispanorromanas II. Producciones Regionales: 255–277. Cadiz, Universidad de Cádiz. Casalini, M. 2014. Anfore di piccole dimensioni a fondo piatto dell’Italia meridionale e della Sicilia. Alcune riflessioni a partire dalla documentazione romana. Rei Cretariæ Romanæ Fautorum Acta 43: 271–278. Cavallo, D., Ciampoltrini, G. and Shepherd, E. J. (eds) 1992. La pesca nell’agro di Cosa in età romana: prospettive di ricerca e nuove acquisizioni. In V Rassegna di Archeologia Subacquea. Atti. Giardini Naxos 19-21 ottobre 1990: 103–114. Messina, Edizioni P&M Associati. Cherubini, L. and Del Rio, A. 1997. Officine ceramiche di età romana nell’Etruria settentrionale costiera: impianti, produzioni, attrezzature. Rei Cretariae Romanae Fautorum Acta 35: 133–141. Cherubini, L., Del Rio, A. and Menchelli, S. 2006. Paesaggi della produzione: attività agricole e manifatturiere nel territorio pisano-volterrano in età romana. In S. Menchelli and M. Pasquinucci (eds), Territorio e produzioni ceramiche. Paesaggi, economia e società in età romana: 69–76. Pisa, Università di Pisa. Cipriano, S. 2009. Le anfore olearie Dressel 6B. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 173–189. Rome, Quasar. Cipriano, M. T. and Carre, M.-B. 1989. Production et typologie des amphores sur la côte adriatique de l’Italie. In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22-24 mai 1986). Collection de l’École française de Rome 114: 67–104. Rome, École française de Rome. Costantini, A. 2006. Le anfore del saggio III. In M. Aprosio and C. Mascione (eds), Materiali per Populonia 5: 169–173. Pisa, Edizioni ETS. Costantini, A, 2007. Dressel 1 a Populonia: esportazioni di tonno dalla foce dell’Albegna. In D. Vitali (ed.), Le fornaci e le anfore di Albinia. Primi dati su produzioni e scambi dalla costa tirrenica al mondo gallico. Atti del Seminario Internazionale (Ravenna 2006): 151–156. Bologna, Università di Bologna. Curtis, R. I. 1991. Garum and salsamenta. Production and commerce in materia medica. Studies in ancient medicine 3. Leiden - New York - Copenhaguen - Cologne, Brill. De Caro, S. 1987. Villa rustica in località Petraro (Stabiae). Rivista dell’Istituto Nazionale di Archeologia, 10/3: 5–89. De Grossi Mazzorin, J. 2006. Testimonianze di lavorazione del tonno a Populonia? In M. Aprosio and C. Mascione (eds), Materiali per Populonia 5: 263–272. Pisa, Edizioni ETS. De Haas, T., Attema, P. and Pape, H., 2008. Amphorae from the coastal zone between Anzio and Torre Astura (Pontine region, Central Italy): the GIA excavations at Le Grottacce, a local amphora collection and material from surveys in the Nettuno area, 2008. Palaeohistoria 49-50: 517–615.

Di Mauro, S. 2011. Un gruppo di anfore a fondo piatto di Alife. Oebalus 5: 289–299. Djaoui, D., Piquès, G. and Botte, E. 2014. Nouvelles données sur les pots dits « à garum » du Latium, d’après les découvertes subaquatiques du Rhône (Arles). In E. Botte and V. Leitch (eds), Fish and Ships. Production and commerce of salsamenta during Antiquity. Actes de l’atelier doctoral, Rome 18-22 juin 2012. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17: 175–196. Arles, Errance - Aix-en-Provence, Centre Camille Jullian. Ellis, S. J. R, 2011. The rise and the re-organization of the Pompeian salted fish industry. In S. J. R. Ellis (ed.), The Making of Pompeii: Studies in History and Urban development of an ancient town. Journal of Roman Archaeology Supplementary Series 85: 59–88. Portsmouth RI, Journal of Roman Archaeology. Empereur, J.-Y. and Hesnard, A. 1987. Les amphores hellénistiques. In P. Lévêque and J. P. Morel (eds), Céramiques hellénistiques et romaines II: 9–71. Paris, Les Belles Lettres. Étienne, R. and Mayet, F. 1994. À propos de l’amphore Dressel 1C de Belo (Cadix). Mélanges de la Casa de Velázquez 30-31: 131–138. Fabiani, F., 2019. Tecniche e impianti per la produzione dell’olio in epoca romana. Esempi in Toscana e Liguria. In S. Segenni (ed.), L’agricoltura in età romana: 103–126. Milan, Ledizioni. Franco, C. and Capelli, C., 2014a. New archaeological and archaeometric data on Sicilian wine amphorae in Roman period (1st to 6th century AD). Typology, origin and distribution in selected western Mediterranean contexts. Rei Cretariae Romanae Fautorum Acta 43: 547–555. Franco, C. and Capelli, C., 2014b. Sicilian flat-bottomed amphorae (1st-5th century AD). New data on typochronology and distribution from an integrated petrographic and archaeological study. In D. Malfitana and G. Cacciaguerra (eds), Archeologia classica in Sicilia e nel Mediterraneo. Didattica e ricerca nell’esperienza mista CNR e Università. Il contributo delle giovani generazioni. Un triennio di ricerche e di tesi universitarie. Ricerche di archeologia classica e post-classica II: 341–362. Catania, Consiglio Nazionale delle Ricerche. García Vargas, E. and Bernal-Casasola, D., 2008. Ánforas de la Betica. In D. Bernal-Casasola and A. Ribera i Lacomba (eds), Cerámicas hispanorromanas. Un estado de la cuestión: 661–687. Cadiz, Universidad de Cádiz. García Vargas, E., Almeida R. R. de, González Cesteros, H. and Sáez Romero, A. (eds) 2019. The Ovoid Amphorae in the Central and Western Mediterranean between the last two centuries of the Republic and the early days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13. Oxford, Archaeopress. Gasperetti, G. 1996. Produzione e consumo della ceramica da mensa e dispensa nella Campania romana. In M. Bats (ed.), Les céramiques communes de Campanie et de Narbonnaise (Ier s. av. J.-C. - IIe s. ap. J.-C.), La vaisselle de cuisine et de table. Actes des journées d’étude (Naples, 1994). Collection du Centre Jean Bérard 14: 19–63. Naples, Centre Jean Bérard.

270

Italian and Sicilian amphorae and their contents: a general overview

Genovesi, S. and Megale, C. 2016. The Roman Settlement of Poggio del Molino: the Late Republican Fort and the Early Imperial Farm. In Fasti OnLine Documents & Research 347: 1-15. . Giannichedda, E. 2014. Chi ha paura dei manufatti? Gli archeologi hanno paura dei manufatti? Archeologia Medievale 41: 79–93. Hahn, H. P. and Weiss, H. (eds) 2013. Mobility, Meaning and Transformations of Things. Oxford, Oxbow Books. Ianelli, M. T. and Cuteri, F. A. 2007. Il commercio e la lavorazione del pesce nella Calabria antica e medievale con particolare riferimento alla costa tirrenica. In Lagóstena, L., Bernal-Casasola, D. and Arévalo, A. (eds), Cetariae 2005. Salsas y Salazones de Pescado en Occidente durante la antigüedad. Actas del Congreso Internacional (Cádiz 7-9 de noviembre de 2005). British Archaeological Reports International Series 1686: 285–300. Oxford, Archaeopress. 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, Tempus Reparatum. Lamboglia, N. 1952. La nave romana di Albenga. Rivista di Studi Liguri 18: 131–203. Lamboglia, N. 1955. Sulla cronologia delle anfore romane di età repubblicana (II-I sec. a.C.). Rivista di Studi Liguri 21: 241–270. Laubenheimer, F. 2007. A propos de timbres d’amphores de l’atelier d’Albinia (Prov: De Grosseto, Italie). Vin et Poisson. In D. Vitali (ed.), Le fornaci e le anfore di Albinia. Primi dati su produzioni e scambi dalla costa tirrenica al mondo gallico. Atti del Seminario Internazionale (Ravenna 2006): 67–80. Bologna, Università di Bologna. Luppino, S. and Sangineto, A. B. 1992. Appendice. Il deposito di anfore di Trebisacce ed un recipiente per la pix Bruttia. In F. Costabile (ed.), Polis e Olympieion a Locri Epizefiri: 174–191. Soveria Mannelli, Rubbettino Editore. Manacorda, D. 1994. Produzione agricola, produzione ceramica e proprietà nella Calabria romana tra Repubblica ed Impero. In Epigrafia della produzione e della distribuzione. Actes de la VIIe Rencontre francoitalienne sur l’épigraphie du monde romain (Rome, 5-6 juin 1992). Collection de l’École française de Rome 193: 3–59. Rome, École française de Rome. Manacorda, D. 2012. Le anfore commerciali di Brindisi: la produzione di Giancola. In D. Manacorda and S. Pallecchi (eds), Le fornaci romane di Giancola (Brindisi): 141–169. Bari, Edipuglia. Manacorda, D. 2019. Produzioni di anfore ovoidi di area brindisina. In E. García Vargas, R. R. de Almeida, H. González Cesteros and A. M. Sáez Romero 2019 (eds), The Ovoid Amphorae in the Central and Western Mediterranean: Between the Last Two Centuries of the Republic and the Early Days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13: 35–41. Oxford, Archaeopress.

Marion, Y. 2009. Le Dressel 6B de petites dimensions de Loron. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 281–288. Rome, Quasar. Martelli, E. 2008. Le anfore. In E. J. Shepherd, G. Capecchi, G. Demarinis, F. Mosca, and A. Patera (eds), Le fornaci del Vingone a Scandicci. Un impianto produttivo di età romana nella valle dell’Arno. Rassegna Archeologia Classica e Postclassica 22 B: 137–158. Florence, All’Insegna del Giglio. Marzano, A., 2007. Fish Salting versus Fish Breeding: the case of Roman Italy. In L. Lagóstena, D. Bernal‑Casasola, and A. Arévalo (eds), Cetariae 2005. Salsas y Salazones de Pescado en Occidente durante la antigüedad. Actas del Congreso Internacional (Cádiz 7-9 de noviembre de 2005). British Archaeological Reports International Series 1686: 301–313. Oxford, Archaeopress. Marzano, A. 2013. Harvesting the Sea: The Exploitation of Marine Resources in the Roman Mediterranean. Oxford, Oxford University Press. Mazzocchin, S. 2009. Le anfore con collo ad imbuto: nuovi dati e prospettive di ricerca. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 191–213. Rome, Quasar. McCann, M. 1987. The Roman port and the fishery of Cosa. Princeton, Princeton University Press. Menchelli, S. 2014. Survey in Toscana e nelle Marche: la metodologia ed alcuni risultati a confronto. In G. Baldelli and F. Lo Schiavo (eds), Amore per l’antico. Dal Tirreno all’Adriatico, dalla Preistoria al Medioevo e oltre. Studi di antichità in ricordo di Giuliano de Marinis: 531– 541. Rome, Scienze e Lettere. Menchelli, S. and Pasquinucci, M. 2012. Ceramiche con rivestimento rosso nella Tuscia settentrionale. Rei Cretariae Romanae Fautorum Acta 42: 229–237. Menchelli, S. and Picchi, G. 2014. Distorsioni interpretative e concretezza epistemologica nello studio delle anfore romane: l’esempio dell’ager Firmanus (Marche meridionali, Italia). In Fasti OnLine Documents & Research 304: 1–26. . Menchelli, S. and Picchi, G. 2016. Late Republican-early Imperial flat-bottomed amphorae: some remarks about their origins and widespread success. Rei Cretariae Romanae Fautorum Acta 44: 229–238. Morais, R. 2007. Ánforas tipo urceus. In L. Lagostena, D. Bernal-Casasola, and A. Arévalo (eds), Cetariae 2005. Salsas y Salazones de Pescado en Occidente durante la antigüedad. Actas del Congreso Internacional (Cádiz 7-9 de noviembre de 2005). British Archaeological Reports International Series 1686: 401–405. Oxford, Archaeopress. Nervi, C. 2014. La Sardinia tra Penisola Iberica e Africa immersa in un mare di «sale» I d.C.-VII d.C. In E. Botte and V. Leitch, (eds), Fish and Ships, Production and commerce of salsamenta during Antiquity. Actes de l’atelier doctoral, Rome 18-22 juin 2012: 199–212. Arles, Errance - Aix-en-Provence, Centre Camille Jullian.

271

Simonetta Menchelli

Olcese, G. (ed.) 2012. Atlante dei siti di produzione ceramica (Toscana, Lazio, Campania e Sicilia): con le tabelle dei principali relitti del Mediterraneo occidentale con carichi dall’Italia centro meridionale: IV secolo a.C. - I secolo. Immensa Aequora 2. Rome, Quasar. Olcese, G. (ed.) 2013. Immensa Aequora Workshop. Ricerche archeologiche, archeometriche e informatiche per la ricostruzione dell’economia e dei commerci nel bacino occidentale del Mediterraneo (metà IV sec.a.C. - I sec. d.C.). Atti del convegno (Roma 24-26 gennaio 2011). Immensa Aequora 3. Rome, Quasar. Ollà, A. 2001. La produzione di anfore vinarie a Naxos (III a.C. - V d.C.). In M. C. Lentini (ed.), Naxos di Sicilia in età romana e Bizantina ed evidenze dai Peloritani. Catalogo Mostra Archeologica Museo di Naxos: 47–60. Bari, Edipuglia. Panella, C. 2001. Le anfore di età imperiale del Mediterraneo occidentale. In P. Lévêque, J. P. Morel and É. Geny (eds), Céramiques hellénistiques et romaines III: 177–275. Paris, Presses Universitaires FrancComtoises. Panella, C. 2010. Roma, il suburbio e l’Italia in età medio e tardo-repubblicana: cultura materiale, territori, economie. FACTA (A Journal of Roman Material Culture Studies) 4 : 11–122. Pecci, A., 2006. Analisi organiche del contenuto di un’anfora romana rinvenuta a Populonia. In M. Aprosio and C. Mascione (eds), Materiali per Populonia 5: 273–282. Pisa, Edizioni ETS. Pecci, A., Salvini, L. and Cantini, F. 2010. Residue analysis of some late Roman amphora coming from the excavations of the historical center of Florence. 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): 363–367. Oxford, Archaeopress. Pecci, A., Salvini, L., Cirelli, E. and Augenti, A. 2010. Castor oil at Classe (Ravenna, Italy): residue analysis of some late Roman amphorae coming from the port. 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): 617–622. Oxford, Archaeopress. Peña, J. T. 2007. Roman Pottery in the Archaeological Record. Cambridge, Cambridge University Press.

Peña, J. T. 2014. The Pompeii Artifact Life History Project: conceptual background and first season’s results. Rei Cretariae Romanae Fautorum Acta 43: 467–474. Picchi, G. and Menchelli, S. 2018, Pots for food. Some regional instances in Italy. Rei Cretariae Romanae Fautorum Acta 45 : 229–238. Picchi, G., Cabella, R., Capelli, C., Ducci S., Menchelli S., Pasquinucci M. and Piazza, M. 2010. Attività manifatturiere nel retroterra di Portus Pisanus. Rei Cretariae Romanae Fautorum Acta 41: 291–302. Rendini, P. 2003. Impianti per la lavorazione di pesce conservato al Giglio e Giannutri. In M. Giacobelli and A. Benini (eds), Atti del II Convegno Nazionale di Archeologia Subacquea, Castiglioncello, 7-9 settembre 2001: 175–188. Bari, Edipuglia. Rizzo, G. 2014. Le anfore, Ostia e i commerci mediterranei. In C. Panella and G. Rizzo (eds), Ostia VI. Le Terme del Nuotatore. Studi Miscellanei 38: 65–481. Rome, L’Erma di Bretschneider. Sangineto, A. B. 2006. Anfore. In G. F. La Torre and F. Mollo (eds), Blanda Iulia sul Palecastro di Tortora. Scavi e ricerche (1990-2005): 310–335. Messina, Dipartimento di Civiltà Antiche e Moderne, Università di Messina. Shepherd, E. J. 2003. La tonnara di Baratti. In C. Mascione and A. Patera (eds), Materiali per Populonia 2: 271–280. Florence, All’Insegna del Giglio. Shepherd, E. J. 2015. Produrre per Florentia. In V. D’Aquino, G. Guarducci, Nencetti, S. and S. Valentini (eds), Archeologia a Firenze: città e territorio. Atti del Workshop Firenze, 12-13 Aprile 2013. Archaeopress Archaeology: 183–192. Oxford, Archaeopress. Shepherd, E. J. and Dallai, L. 2003. Attività di pesca al promontorio di Piombino (I sec.a.C. - XI sec. d.C.). In M. Giacobelli and A. Benini (eds), Atti del II Convegno Nazionale di Archeologia Subacquea, Castiglioncello, 7-9 settembre 2001: 189–207. Bari, Edipuglia. Tchernia, A. 1996. Maesianus Celsus et Caedicia Victrix sur des amphores de Campanie. In Les élites municipales de l’Italie péninsulaire des Gracques à Néron. Actes de la table ronde de Clermont-Ferrand (28-30 novembre 1991). Collection de l’École française de Rome 215. Rome, École française de Rome. Van Limbergen, D. 2016. A Note on olives and olive oil from Picenum (Marche, northern Abruzzo): an obscured food product within the economy of central Adriatic Italy in Roman times ? Picus 36: 301–312. Vitali, D. 2007 (ed.). Le fornaci e le anfore di Albinia. Primi dati su produzioni e scambi dalla costa tirrenica al mondo gallico. Atti del Seminario Internazionale (Ravenna 2006). Bologna, Università di Bologna. Wilson, A. 2007. Quantification of fish salting infrastructure capacity in the Roman world. Oxford Review of Economic Policy Working paper 21: 1–11.

272

The content of amphorae from Adriatic Italy Marie Brigitte Carre

Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France Maison Méditerranéenne des Sciences de l’Homme, 5 rue du Château de l’Horloge, CS 90412, 13097, Aix-en-Provence (France)

Stefania Pesavento Mattioli

Università degli Studi di Padova, Dipartimento dei Beni Culturali Piazza Capitaniato 7, 35139 Padova (Italia)

Abstract: This contribution focuses in particular on the contents of amphorae from Adriatic Italy. Data collected from historical, archaeological and archaeometric sources are presented, followed by an investigation into various research issues with some suggested directions for future multidisciplinary research initiatives. Amphorae for the transport of wine, oil, olives, sauces and preserved fish were produced in Adriatic Italy, from Istria to Apulia. The shape of these Adriatic amphorae always indicates a specific content and any anomalous content preserved demonstrates its reuse or a primary but unusual use of the containers to carry different types of food. Tituli picti which mention fish on Adriatic wine and oil amphorae are in effect ‘labelling’ a content which is different from the primary one, while tituli picti on small amphorae for fish and olives provide information on the specific quality of these products. Sporadic examples of reuse, whose importance is often exaggerated, are also attested: for example pine resin, augite sand or pozzolan, found in Lamboglia 2 wine amphorae. Key words: Adriatic Italy amphora; reuse; wine; sauces and fish preserve; oil.

1. General overview (Figure 1) Amphorae for the transport of wine, oil, olives, sauces and fish preserves were produced in Adriatic Italy, from Istria to Apulia. They are reasonably well understood and it is not necessary to re-examine the subject of their production centres, which cannot always be clearly identified, but have been investigated in a number of studies (Carre 1985; Cipriano and Carre 1989; Carre and Pesavento Mattioli 2003a; Carre and Pesavento Mattioli 2003b; Carre, Monsieur and Pesavento Mattioli 2014; Cipriano and Mazzocchin 2017; Carre and Pesavento Mattioli 2018). Furthermore, the volume on materials found in the north-eastern area of Terme del Nuotatore in Ostia contains a broad and in-depth analysis of the history and geography of productions from Venetia and Histria, Aemilia, Umbria, Picenum, Apulia and Calabria (Rizzo 2014: 117–130, with a very rich bibliography for more information on the morphology, chronology and distribution of the containers). New finds, such as those at the Nuovo Mercato Testaccio in Rome (Cafini and D’Alessandro 2010; D’Alessandro 2011; D’Alessandro 2013), or the revision of the complex of amphorae with stamps in the Modena area (Mongardi 2018) have enriched the picture of the Adriatic amphorae in various ways. In particular, the study and publication of a substantial set of materials from Aquileia, among which a large quantity of amphorae, were important: an excavation carried out in 2004‑2005 investigated a section of Canale Anfora, one of the waterways which constituted the harbour system of the emporium, slowly infilled between the last years of the 1st and the 3rd

centuries, to then be totally filled within the 4th (Maggi et al. eds 2017). Further perspectives on the production of amphorae along the Adriatic Sea are now opened by investigation on the east coast, in Liburnia, where Crikvenica furnaces have been identified and extensively studied (Lipovak Vrkljan 2011; Lipovac Vrkljan and Konestra eds 2018). However, there is still room for an investigation into the content of western adriatic amphorae, starting from a very short summary of the data acquired on the basis of historical, archaeological and archaeometric criteria, followed by a discussion of various issues concerning the evidence and suggestions for possible future new avenues for multidisciplinary research. Greco-Italic, Lamboglia 2 and Dressel 6A amphorae have a long production history, between the 3rd century BC and the first half of the 1st century AD, as confirmed by the existence of transitional shapes and the fact that they were often produced in the same kilns (Rizzo 2014 and Panella 2010). They were used for the commerce of wine. We disagree with the hypothesis (Lindhagen 2009; Lindhagen 2013) of the pioneering and primary role of central Dalmatia in the production of these amphorae, although there may have been kilns in that region — an idea that has yet to be confirmed (Carre, Monsieur and Pesavento Mattioli 2014). Aside from Dressel 6A, Dressel 2-4 amphorae are produced, albeit in smaller numbers. As of the mid-1st century AD (or perhaps earlier), flat-bottomed amphorae started to be produced:

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 273–279

Marie Brigitte Carre and Stefania Pesavento Mattioli

kilns are attested mainly in Emilia (Aldini 1978 and Stoppioni 1993), but also in Abruzzo (Staffa 2003), Istria (Maggi and Marion 2011) and on the eastern shore of Liburnia (Lipovac Vrkljan 2011). The deposition conditions of Canale Anfora allowed the preservation of several tituli picti, now identified for the first time on these containers: mostly, they refer to numeric series interpreted as ponderal indications. Most amphorae are connected to kiln sites in Romagna, but many remain without attribution, expanding the picture of known productions (Gaddi and Maggi 2017: 263–284). In the Republican period Adriatic oil was traded in ovoid amphorae from the territory of Brindisi (Manacorda and Pallecchi (eds) 2012; Palazzo 2013) and from the central Adriatic. These are followed, as of the mid-1st century BC and up to the 4th century AD, by the Dressel 6B series in its various phases, produced in territories pertaining to the central Adriatic and eastern Cisalpine, where plenty of kilns are attested in the Istrian peninsula (for an overview: Carre and Pesavento Mattioli 2003b, Cipriano 2009; for Istrian productions in Fasana and Loron: Bezeczky 1998 and Maggi and Marion 2011). A significant context is that of Canale Anfora in Aquileia: its chronology justifies the low quantities of Dressel 6B of the first three phases, while very consistent (more than 300 vessels, corresponding to the 65% of the total Dressel 6B) are the quantities of those of the fourth phase, datable within the 2nd and the 4th centuries AD. Therefore, it was possible to dedicate a thorough analysis of them, allowing to distinguish Loron and Fažana products and the different variants, to add to the group some small scale examples similar to the small fish amphorae (cf. infra) but lacking the inner resin lining which was otherwise preserved within this context, and to enlarge the panoramic of the stamps (Gaddi and Maggi 2017: 286–308). Another group of containers, which certainly had an Adriatic origin, identified as ‘anfore con collo a imbuto’ (amphorae with a funnel-like neck, also named Portorecanati after the first find) thanks to their peculiar form, were probably for oil. These include those based on the numerous exemplars found mainly in northern Italy, but also in north-eastern provinces and central‑Adriatic Italy. There are two types and they date from the mid-1st to the mid-3rd century AD (Mazzocchin 2009; Gaddi and Maggi 2017: 308–313). Finally, other foodstuffs were carried in amphorae from Adriatic Italy. The Schörgendorfer 558‑type amphorae (also known as ‘anfore troncoconiche da olive’), probably from the Picenum area, contained olives, and are attested from the 1st to the second half of the 2nd century AD, mainly in northern Italy and in the Danubian provinces (Pesavento Mattioli 2008; Pesavento Mattioli 2011). Small amphorae of the Grado type, Aquincum 78 and similar have been associated with products related to fish processing; among these, documented between the end of the 1st and the end of the 3rd century AD,

specific production areas have yet to be identified (Carre, Pesavento Mattioli and Belotti 2009). Many new finds or newly recognised specimens were to follow, for example in the territory of Modena (Mongardi and Rigato 2013) and in particular within the Canale Anfora assemblage, where 136 individuals have been recognised, forming a significant sample from the morphological standpoint as well, and confirming the validity of the proposed classification while enlarging it with another variant. Within all Aquileian small amphorae, resin lining has been preserved, which distinguishes them from the small Dressel 6B (Gaddi and Maggi 2017: 314–327). 2. Main contents, uncertain contents and reuses The contents of these amphorae is known thanks to extensive data, which do not need to be repeated, such as sources on the use of the different territories, kiln waste, tituli picti, remains of the contents in specific contexts, the presence or lack of pitch coating, etc. But one aspect needs to be remembered: most of the containers found and used for our typo-chronological analyses had been reused in reclamation and drainage projects, which were quite numerous in northern Italy (cf. e.g. Pesavento Mattioli ed. 1998). In these cases no content was ever found, although the amphorae were intact, and the action of water erased any trace of pitch coating inside them. The only exception, as we have said, are the materials of Canale Anfora in Aquileia: the characteristics of the deposit have in fact conserved the internal pitch and many tituli picti. It should also be added that, despite the numerous wrecks reported or investigated in the Adriatic sea, especially along the eastern coast (cf. Parker 1992 and Jurišić 2000), very few have been published systematically and much of the data needs to be analysed. We are mainly lacking valuable information deriving from the analyses of the content of the amphorae included in the cargoes, which do not seem to have been carried out yet. Based on our experience and the most recent studies, it therefore seemed advisable to suggest some cases where, according to us, it is possible to distinguish primary contents and secondary contents/reuse of the amphorae and others that are still complicated and deserve more in-depth analyses. 2.1. Lamboglia 2: anomalous contents Contrary to what has been written (Loughton 2014: 77), Lamboglia 2 amphorae were not multi-purpose containers. The first studies on Lamboglia 2 from the Madrague de Giens wreck have proved that their use as oil containers initially suggested by Lamboglia was wrong (Formenti, Hesnard and Tchernia 1978). Some sporadic examples of reuse, often overrated, are attested: pine resin in about a hundred exemplars from the SudCaveaux 1 wreck, where it is clear that the amphorae, once

274

The content of amphorae from Adriatic Italy

Figure 1. Main types of Italic Adriatic Amphoras (sc. approx 1:20). Wine amphoras: a. Greco-Italic; b. Lamboglia 2; c. Dressel 6A (Picenum); d. Dressel 6A (Cisalpina); e. Flat-bottomed amphora (San’Arcangelo di Romagna); f. Flatbottomed amphora (Forlimpopoli). Oil and olive amphoras: g. Brindisi (Apani IIA); h. Brindisi (Apani III), i. Adriatic Ovoid; j. Dressel 6B (Cisalpina); k. Dressel 6B (Istria); l. Funnel-shape amphora; m. Schörgendorfer 558 A; n. Schörgendorfer 558 B; o. Small fish amphora.

275

Marie Brigitte Carre and Stefania Pesavento Mattioli

in the Iberian peninsula and filled with a new content, were then carried to Marseille to meet the needs of an amphora or dolium workshop or of a shipyard (Long 1998); sand with granular elements of pyroxene, sanidine and biotite from a volcanic area, such as Campania or Latium, perhaps to be used as a temper for the preparation of ceramics, in Adriatic amphorae from Ship B in Pisa (Pesavento Mattioli and Mazzocchin 2002: 786; Giachi and Pallecchi 2000: 350–351); pozzolan in a Lamboglia 2 from the wreck Chrétienne A, maybe used to seal the stoppers of the Dressel 1 amphorae loaded on the same ship (Hesnard and Gianfrotta 1989: 398). These cases must not force us to doubt that their primary content was wine: the wide diffusion of Lamboglia 2 amphorae around the Mediterranean (cf. Loughton 2014: 359–366) made a great number of them available to be reused to carry different materials and foodstuffs, after being emptied of their original content in the place where it was consumed.

Figure 2. Amphora Dressel 6A from Salzburg with titulus pictus FLOS/VET/X (Heger 1986: 133–137).

2.2. Fish content in Dressel 6A and Dressel 6B amphorae Remains of fish or tituli picti mentioning garum and liquamen were found in some Dressel 6A (in Salzburg — Figure 2 —, Verona, Milan and Forum Sempronii) and in a few Dressel 6B (in Milan and Magdalensberg). We have come to the conclusion that these are not containers specifically produced to transport something different from wine and oil, but represent the reuse of amphorae which were widely available, where the ‘anomaly’ was indicated by the ‘label’ of the titulus itself. These vessels could contain a sauce imported probably from the Iberian peninsula in casks or barrels or occasionally produced in some locality along the Adriatic coast (Carre, Pesavento Mattioli and Belotti 2009: 215–220). The latter option has been confirmed by the content of a Dressel 6A found in Vicenza (Figure 3), consisting of a mix of saltwater and freshwater fish, which probably comes from the nearby Venetian Lagoon; although the amphora was found in an embankment area where the other amphorae recovered (over 300) were empty, it was probably buried with part of its content still on the bottom (Mazzocchin and Wilkens 2013; Mazzocchin 2013: 90–92). The absence of a titulus pictus showing a content different from wine does not surprise us, since very few painted inscriptions have ever survived in similar contexts. 2.3. Content of ‘small amphorae’ of the Grado type (‘anforette adriatiche’) The extensive analysis carried out during the 2007 conference on the production and trade of oil and fish in the Adriatic area (Pesavento Mattioli and Carre eds 2009) led to the proposal of a typological classification of small fish amphorae, which might have to be rethought and expanded in the light of continuous new findings. No problems arise as to the content of those that bear clear inscriptions mentioning liquamen, garum, muria, often

Figure 3. Amphora Dressel 6A from Vicenza with contents (Mazzocchin and Wilkens 2013: 106). first quality (flos) and/or that are abundantly covered in pitch, but we have obviously wondered whether this might be another case of occasional reuse of containers which were originally meant for oil (for their morphological similarity to Dressel 6B, in spite of their smaller dimensions). This hypothesis is contradicted by the presence of an abundant pitch coating even when there are no tituli picti and the large quantity of widespread exemplars, in contrast to the few Dressel 6A and 6B with an ‘anomalous’ content. Mineralogical and chemical analyses of the ceramic body of some exemplars

276

The content of amphorae from Adriatic Italy

have also shown the non-compatibility with any group of Dressel 6B, confirming various productions, which have yet to be localised, but clearly focused on specific contents (Carre, Pesavento Mattioli and Belotti 2009: 221–234 and then Schimmer 2012: 57–58 and table 61; Mongardi and Rigato 2013; Mongardi 2018: 135–138). According to us, the presence of almost 200 small fish amphorae in the ship in Grado, which also carried approximately 400 other amphorae of different origins (Africa, Tripolitania, western Mediterranean) originally produced for wine and oil but full of fish remains, can be explained with an extraordinary availability of fish in some parts of the Adriatic; small amphorae were used for liquamen, which probably had a tradition of being locally-produced and did not need big salting plants, whereas larger amphorae were reused for other types of preserves. No problems are raised by those small amphorae which, despite having quite similar morphological features, do not have any pitch coating and have no tituli picti: this is the case of the small amphorae produced in Loron, Istria (Marion 2009; Maggi and Marion 2011: 177–178), where the absence of the above-mentioned elements has been explained with the finding of kiln waste, suggesting that this was a production site, and since the amphorae had not yet been filled we cannot say whether or not pitch and tituli picti would have been added at a later stage. However, we must remember that among the materials found in the fill of Canale Anfora in Aquileia, where, as previously mentioned, the original pitch coating is still present, there are some bases identical to the ones of the small amphorae but without a pitch coating: only targeted analyses could help in understanding if they contained oil and must be therefore identified as very small Dressel 6B exemplars. It should be remembered that the tituli picti in two of the small amphorae from Canale Anfora (Figure 4) explicitly mention liquamen Aquileiense, allowing the identification of a local origin for the fish sauce (Maggi 2016: 426–427).

Figure 4. Small fish amphora from Aquileia – Canale Anfora with titulus pictus LIQ/AQUIL/XVIII (Gaddi and Maggi 2017: 323).

Figure 5. Funnel-shape amphora from Aquileia – Canale anfora with stamp C.IVLI MARCELLI (Gaddi and Maggi 2017: 312.) also because only one exemplar bearing the stamp C. IVLI/MARCELLI was pitch coated (Figure 5). This stamp, like IVLI palm/PAVLINI, also appears on flatbottomed amphorae that were certainly used for wine (Mazzocchin 2009: 206–207): we can therefore imagine either a different type of content, oil versus wine, or an identical content of a different quality (Gaddi and Maggi 2017: 311–313). 3. Conclusion

It is likely that these amphorae contained oil, because of the absence of pitch coating and the correspondence of the diffusion patterns with those of Dressel 6B, which in some ways they also bear some resemblance to, but no content remains have ever been identified (please note that there is evidence from wrecks or contexts where it could have survived) and there are no tituli picti mentioning it.

In conclusion, at least as far as Adriatic Italy is concerned, we firmly believe that the shape of the amphorae always indicates a specific content and that any anomalous content preserved shows a reuse or a primary but fortuitous use of containers produced to carry different types of food. Tituli picti which mention fish on Adriatic wine and oil amphorae ‘label’ a content which is different from the primary one, while tituli picti on small amphorae and olive amphorae provide information on a specific quality of the content: liquamen, garum or muria and other indications that remain unclear on small amphorae, oliva alba or nigra ex dulci on olive amphorae (perhaps ex muria on the ones with no titulus).

Among the exemplars from Canale Anfora in Aquileia, the constant lack of traces of pitch coating on the inner surfaces might suggest that the containers were actually used for the commerce of oil, confirming the most plausible hypothesis, but the problem remains,

In view of future multidisciplinary studies, useful results to better define the content of Adriatic amphorae can be obtained with analyses on amphorae with funnel-like necks and small amphorae with no pitch coating; samples are available for this purpose.

2.4. Content of amphorae with funnel-like necks (‘anfore con collo a imbuto’)

277

Marie Brigitte Carre and Stefania Pesavento Mattioli

Bibliography Aldini, T. 1978. Anfore foropopiliensi. Archeologia classica 30: 236–245. Bezeczky, T. 1998. The Laecanius amphora stamps and the villas of Brijuni. Vienna, Österreischen Akademie der Wissenschaften. Cafini, M. L. and D’Alessandro, L. 2010. Anfore adriatiche a Roma. Rinvenimenti dall’area del Nuovo Mercato Testaccio. Rei Cretariae Romanae Fauctorum Acta 41: 93–100. Carre, M.-B. 1985. Les amphores de la Cisalpine et de l’Adriatique au début de l’Empire. Mélanges de l’École Française de Rome. Antiquité 97: 207–245. Carre, M.-B. and Pesavento Mattioli, S. 2003a. Anfore e commerci nell’Adriatico. In F. Lenzi (ed.), L’Archeologia dell’Adriatico dalla Preistoria al Medioevo. Atti del convegno internazionale. Rimini 2001: 268–285. Florence, All’Insegna del Giglio. Carre, M.-B. and Pesavento Mattioli, S. 2003b. Tentativo di classificazione delle anfore olearie adriatiche. Aquileia Nostra 74: 453–475. Carre, M.-B. and Pesavento Mattioli, S. 2018. The amphorae of the western Adriatic: an update. In G. Lipovac Vrkljan and A. Konestra (eds), Pottery Production, Landscape and Economy of Roman Dalmatia. Interdisciplinary approaches. Archaeopress Roman Archaeology 47: 7–13. Oxford, Archaeopress. Carre, M.-B., Monsieur, P. and Pesavento Mattioli, S. 2014. Transport amphorae Lamboglia 2 and Dressel 6A: Italy and/or Dalmatia? Some clarifications. Journal of Roman Archaeology 27: 417–428. Carre, M. -B., Pesavento Mattioli, S. and Belotti, C. 2009. Le anfore da pesce adriatiche. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 215–238. Rome, Quasar. Cipriano, M. T. and Carre, M.-B. 1989. Production et typologie des amphores sur la côte adriatique de l’Italie. In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (2224 mai 1986). Collection de l’École française de Rome 114: 67–104. Rome, École française de Rome. Cipriano, S. 2009. Le anfore olearie Dressel 6B. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 173–189. Rome, Quasar. Cipriano, S. and Mazzocchin, S. 2017. Western Adriatic amphorae production: the research status, in G. Lipovac Vrkljan, I. Radić Rossi and A. Konestra (eds), AdriAmphorae. Proceedings of the workshop (Zagreb 21st April 2016): 33–47. Zagreb, Institut za arheologiju. D’Alessandro, L. 2011. Anfore adriatiche a Roma: attestazioni del bollo THB dal Nuovo Mercato Testaccio. In SFECAG, Actes du Congrès d’Arles: 607–617. Marseille, Société Française d’Étude de la Céramique Antique en Gaule.

D’Alessandro, L. 2013. Anfore adriatiche a Roma: dati epigrafici dal Nuovo Mercato Testaccio. In D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz and A. M. Saez (eds), Hornos, talleres y focos de producción alfarera en Hispania. Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (I): 351–364. Cadiz, Universidad de Cádiz. Formenti, F., Hesnard, A. and Tchernia, A. 1978. Note sur le contenu d’une amphore Lamboglia 2 de l’épave de la Madrague de Giens. Archaeonautica 2: 95–100. Gaddi, D. and Maggi, P. 2017. Anfore italiche. In P. Maggi, F. Maselli Scotti, S. Pesavento Mattioli and E. Zulini (eds), Materiali per Aquileia. Lo scavo di Canale Anfora (2004-2005): 263-328. Trieste, Editreg. Giachi, G. and Pallecchi, P. 2000. Analisi preliminari sui materiali/Analysis of the materials preliminary results. In S. Bruni (ed.), Le navi antiche di Pisa. Ad un anno dall’inizio delle ricerche: 348–351. Florence, Edizioni Polistampa. Heger, N. 1986. Frührömische Amphoren aus der Stadt Salzburg (Mozartplatz 4). Bayerische Vorgeschichtsblätter 51: 131–161. Hesnard, A. and Gianfrotta, P. A. 1989. Les bouchons d’amphore en pouzzolane. In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22-24 mai 1986). Collection de l’École française de Rome 114: 393–441. Rome, École française de Rome. Jurišić, M. 2000. Ancient shipwrecks of the Adriatic maritime transport during the first and second centuries AD. British Archaeological Reports International Series 828. Oxford, Archaeopress. Lindhagen, A. 2009. The transport amphoras Lamboglia 2 and Dressel 6A: a central Dalmatian origin? Journal of Roman Archaeology 22: 83–108. Lindhagen, A. 2013. The freedmen milieus at Delos and Narona. New perspectives on the Lamboglia 2 wine trade. In A. L. Schallin (ed.), Perspectives on ancient Greece Papers in celebration of the 60th anniversary of the Swedish Institute at Athens. Acta Instituti Atheniensis Regni Sueciae 22: 231–250. Stockholm, Paul Astroms. Lipovac Vrkljan, G. 2011. Lokalna keramičarska radionica Seksta Metilija Maksima u Crikvenici- Crikveničke amfore rannog dna (Local potteryworkshop of Sextus Metilius Maximus in Crikvenica-Crikvenica flatbottomed amphorae). In G. Lipovac Vrkljan, I. Radić Rossi and B. Šiljeg (eds), Rimske Keramičarske i staklarske radionice. Proizvodnja i trgovina na Jadranskom prostoru: 3–18. Crikvenika, Institut za arheologiju Zagreb. Lipovac Vrkljan, G. and Konestra, A. (eds) 2018. Pottery Production, Landscape and Economy of Roman Dalmatia. Interdisciplinary approaches. Archaeopress Roman Archaeology 47. Oxford, Archaeopress. Long, L. 1998. Lucius Volteilius et l’amphore de 4ème type. Découverte d’une amphore atypique dans une épave en baie de Marseille. In El vi a l’antiguitat: economia, producció i comerç al Mediterrani occidental. Actes del II Col loqui Internacional d’Arqueologia Romana (Barcelona 6-9 de maig de 1998). Monografies badalonines 14: 341–349. Badalona, Museu de Badalona.

278

The content of amphorae from Adriatic Italy

Loughton, M. 2014. The Arverni and Roman Wine. Roman Amphorae from Late Iron Age sites in the Auvergne (Central France): Chronology, fabrics and stamps. Archaeopress Roman Archaeology 2. Oxford, Archaeopress. Maggi, P. 2016. Tituli picti su anfore di produzione adriatica dallo scavo di Canale Anfora ad Aquileia. In F. Mainardis (ed.), ‘Voce concordi’. Scritti per Claudio Zaccaria: 423–438. Trieste, Editreg. Maggi, P. and Marion, Y. 2011. Le produzioni di anfore e di terra sigillata a Loron e la loro diffusione. In G. Lipovac Vrkljan, I. Radić Rossi and B. Šiljeg (eds), Rimske Keramičarske i staklarske radionice. Proizvodnja i trgovina na Jadranskom prostoru: 176-187. Crikvenika, Institut za arheologiju Zagreb. Maggi, P., Maselli Scotti, F., Pesavento Mattioli, S. and Zulini, E. (eds) 2017. Materiali per Aquileia. Lo scavo di Canale Anfora (2004-2005). Trieste, Editreg. Manacorda, D. and Pallecchi, S. (eds) 2012. Le fornaci romane di Giancola (Brindisi). Bari, Edipuglia. Marion, Y. 2009. Les Dressel 6B de petites dimensions de Loron. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 281–288. Rome, Quasar. Mazzocchin, S. 2009. Le anfore con collo ad imbuto: nuovi dati e prospettive di ricerca. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 191–213. Rome, Quasar. Mazzocchin, S. 2013. Vicenza. Traffici commerciali in epoca romana: i dati delle anfore. Trieste, Editreg. Mazzocchin, S. and Wilkens, B. 2013. Fish and Crustaceans from a Roman Amphora in Northern Italy. Archaeofauna 22: 105–111. Mongardi, M. 2018. Firmissima et splendidissima populi romani colonia. L’epigrafia anforica di Mutina e del suo territorio. Instrumenta 62. Barcelona, Universitat de Barcelona. Mongardi, M. and Rigato, D. 2013. Tituli picti su anforette adriatiche da pesce: Modena e il suo territorio (Italia). In SFECAG, Actes du Congrès d’Amiens: 615–620. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Palazzo, P. 2013. Le anfore di Apani (Brindisi). Rome, Scienze e Lettere.

Panella, C. 2010. Roma, il suburbio e l’Italia in età medioe tardo-repubblicana: cultura materiale, territori, economie. FACTA (A Journal of Roman Material Culture Studies) 4: 11–123. Parker, A. J. 1992. Ancient Shipwrecks of the Mediterranean and the Roman Provinces. British Archaeological Reports International Series 580. Oxford, Tempus Reparatum. Pesavento Mattioli, S. (ed.). 1998. Bonifiche e drenaggi con anfore in epoca romana: aspetti tecnici e topografici. Atti del seminario di studi (Padova 1995). Modena, Franco Cosimo Panini. Pesavento Mattioli, S. 2008. Le anfore troncoconiche da olive: spunti di riflessione. In P. Basso, A. Buonopane, A. Cavarzere and S. Pesavento Mattioli (eds), Est enim ille flos Italiae. Vita economica e sociale nella Cisalpina romana. Atti delle giornate di studio in onore di Ezio Buchi (Verona 2006): 335–348. Verona, Qui Edit. Pesavento Mattioli, S. 2011. Le anfore Schörgendorfer 558 e il commercio delle olive adriatiche. In G. Lipovac Vrkljan, I. Radić Rossi and B. Šiljeg (eds), Rimske Keramičarske i staklarske radionice. Proizvodnja i trgovina na Jadranskom prostoru: 165–173. Crikvenika, Institut za arheologiju Zagreb. Pesavento Mattioli, S. and Carre, M. B. (eds) 2009. Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15. Rome, Quasar. Pesavento Mattioli, S. and Mazzocchin, S. 2002. La nave B del porto di Pisa: ipotesi su una rotta commerciale di età augustea. In L’Africa romana XIV. Lo spazio marittimo del Mediterraneo occidentale: geografia storica ed economia: 779–787. Rome, Carocci. Rizzo G. 2014. Le anfore, Ostia e i commerci mediterranei. In C. Panella and G. Rizzo (eds), Ostia VI. Le terme del nuotatore. Studi Miscellanei 38: 65–481. Rome, L’Erma di Bretschneider. Schimmer, F. 2012. Amphoren aus Cambodunum/Kempten. Ein Beitrag zur Handelgeschichte der römischen Provinz Raetia. Wiesbaden, Reichert Verlag. Staffa, A. R. 2003. Impianti produttivi d’età romana nel territorio di Pescara: le fornaci. Rivista di Topografia Antica 13: 117–154. Stoppioni, M. L. 1993. Le anfore. In M. L. Stoppioni (ed.), Con la terra e con il fuoco. Fornaci romane del Riminese: 145–154. Rimini, Guaraldi.

279

African amphora contents: an update Michel Bonifay

Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France

Abstract: This papers aims to update the previous syntheses made by the author in 2004 and 2007 about the contents of African amphorae. Firstly, there is a revision of the different arguments providing information on these contents (typology, epigraphy, distribution, stoppers, macro-remains, archaeometry), and second, an updating of the different hypotheses relating to contents (oil, fish products and wine) by types. The conclusion is that very little progress has been made for 15 years, and the hypotheses about contents can still be divided into three categories: certain (oil for types Africana I and Tripolitana I/III; salsamenta for type Africana II C; wine for imitations of types Dressel 2/4, Gauloise 4 and LRA 1); controversial (amphorae of Punic tradition, early Roman types, Africana II and III); and completely conjectural (5th to 7th century types). The author calls for the launch of an ambitious programme of chemical analyses in order to reach a better understanding of the foodstuffs transported in the African amphorae from the mid 2nd century BC to the end of the 7th century-beginning 8th century AD. Key words: Africa; amphorae; oil; wine; salsamenta; trade.

1. Introduction While the African amphora typology and chronology are becoming better understood, the same cannot be said of their contents. Africa is perhaps the region of the Roman world where amphora contents remains a major problem. During the fifteen years after the work carried out on African amphorae by Fausto Zevi and André Tchernia in 1969, there was little doubt about their content. Their large distribution overseas was supposed to provide evidence of the development of olive cultivation in Africa under the Antonine dynasty (Carandini 1970). Therefore olive oil was considered to be the content of most of these containers, first at Ostia for the 2nd‑4th century types (Panella 1973; Manacorda 1977), and then in Catalonia for the latest types of the 5th to 7th century (Keay 1984). Nevertheless, the idea of a single content was not entirely accepted among the specialists. In 1969 it was clear to Fausto Zevi and André Tchernia that two different amphora types produced in the same workshops could have meant two different contents, the largest amphorae (Africana II) being probably devoted to olive oil, and the smaller ones (Africana I) to garum or salted fish (Zevi and Tchernia 1969: 185–187). On the other hand, Robert Lequément’s work during the 70s had drawn attention to the pitch lining of some African amphorae which was from a general point of view totally incompatible with an oleaginous content. As a matter of fact, such a pitch lining was present inside numerous Africana II and III amphorae (Lequément 1975 and 1976). Lastly, a chemical analysis programme undertaken in 1980 by André Tchernia and chemist Françoise Formenti made it possible to reverse the initial hypothesis, and to determine that Africana I was intended for the shipping of olive oil (due to traces of fat acid), while Africana II (which did not show fat acid traces, but traces of pitch) probably transported another foodstuff, perhaps salted fish products.

During the 80s and 90s, different research schools have developed, some specialists remaining faithful to olive oil as the content of most African amphorae (Keay 1984; Peacock and Williams 1986), others leaning more towards a strict separation between the contents according to typology (Sciallano and Sibella 1991), yet others adopting a prudent approach suggesting a possible interchangeability of the contents (Panella 1982; Laubenheimer 1990). In the last 25 years I have tried several times to address this issue, taking into account all the available arguments, archaeological and archaeometric (Ben Lazreg et al. 1995; Bonifay 2004, 2007 and 2016). Recent data, as well as a new scientific approach led by a young generation of researchers, might justify this attempt at a provisional updating (Figure 1). 2. Current debates 2.1. Amphora shape Is the typology indicative of the content? I would be inclined to follow Bernard Liou when he asserts that ‘il faut évidemment croire dur comme fer aux principes de base: qu’il y a des amphores à vin, des amphores à huile et des amphores à saumure que distingue leur forme’ (Liou 1988: 172). If the African potters imitated the Dressel 2/4, Gauloise 4 or MR 1, that is probably because the African producers wished to package their own production of wine in such containers (Figure 4.61‑63). This assertion, verifiable for the African imitations of non-African types, is perhaps also applicable to most African amphorae of Punic tradition and Roman-African typology. This is not very far from our current notion of intelligent packaging. What our logistics specialists propose to do with nanotechnologies, the ancients tried to do through the amphora shape, the stamps and the dipinti…

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 281–297

Michel Bonifay

1 2 3

Van der Werff 1 Van der Werff 2 Van der Werff 3

Traces of pitch ■ ■ ■

4

Schöne-Mau XL

?

5

Hammamet 1

□

6 7 8

Hammamet 1E Hammamet 2 Hammamet 3

□ □ □

9

Tripolitaine II

□

10

Benghazi LR 7

?

11

Leptiminus II

■

12 13

Ancient African Dressel 26 Carthage EA IV

□ □ ■

14

Leptiminus I

?

15

Uzita Pl. 52, 10

□

16

Ostia LIX

□

17

Ostia XXIII

□

18

Africaine II pseudotripolitanian

?

19

Bonifay 16

□

20

Bonifay 17 Station 48 piazzale delle Corporazioni

?

22

Africaine I

□

23

Africaine II A

■

24

Africaine II B

■

25

Africaine II C

■

26 27 28

Africaine II D Tripolitaine I Tripolitaine III

■ □ □

29

Africaine III A (= Keay 25.1)

■

30

Africaine III B (= Keay 25.3)

■

31

Keay 25 V

?

32

Africaine III C (= Keay 25.2)

■

33

spatheion 1

□

34-37 38 39 40

spatheion 3 Keay 27 Keay 36 Keay 59/8B

■ □ □ □

Fig.

21

Type

Archaeological arguments

not attested at the Monte Testaccio; long hollow spike not attested at the Monte Testaccio attested at the Monte Testaccio

Archaeometry traces of wine traces of wine

salsamenta? wine? salsamenta? wine? salsamenta? wine? salsamenta? salsamenta? wine? oil? ? ?

not attested at the Monte Testaccio; coastline and inland workshops not attested at the Monte Testaccio; long hollow spike

Possible content

salsamenta? wine? ? salsamenta? traces of oil?

long hollow spike fish macro-remains; long hollow spike not attested at the Monte Testaccio; long hollow spike

oil? interchangeable? ? (oil ?) salsamenta? salsamenta? salsamenta? olives? salsamenta? wine?

cork stopper attested at the Monte Testaccio; ceramic stopper; shape close to type Africana I inland workshops; attested at the Monte Testaccio? attested at the Monte Testaccio; shape close to type Africana I

oil? oil? oil? ? salsamenta? wine?

■ attested at the Monte Testaccio; traces of oil ceramic stopper rare at the Monte Testaccio; cork no traces of oil stopper anchovy DNA rare at the Monte Testaccio coastline workshops; lead labels with trident; fish macroremains? no traces of oil attested at the Monte Testaccio attested at the Monte Testaccio no traces of oil, chemical marker of wine no traces of oil, no traces of wine, fatty acids?

no traces of pitch (ex.: Pisa); olive pits (ex.: Dramont E wreck); fish macro-remains (ex.: Tarragona) dipinti inland workshops? inland workshops? inland workshops

282

no traces of oil, chemical marker of wine?

oil salsamenta? wine? salsamenta? wine? salsamenta? salsamenta? wine? oil oil wine? salsamenta? ? (oil ?) salsamenta? wine? interchangeable? interchangeable? oil, olives, salsamenta, wine? garum? oil? oil? wine? oil?

African amphora contents: an update

41

Keay 35A

Traces of pitch □

42

Keay 35B

■

43

Keay 39

□

44-46 Keay 55-56-57

□

Fig.

47 48 49 50 51

Type

? □ ? ? ?

53

Keay 64 Keay 3/5 Keay 40 Keay 41 Keay 62R Keay 62Q / Albenga 11/12 Keay 62A

54

Bonifay 47

■

52

55 56 57 58 59 60 61 62 63 64 -

?

Archaeological arguments

Possible content oil?

coastline workshops; fish macroremains; attested in fish-salting factories

salsamenta? no traces of oil in Keay 55

inland workshops? inland workshops? ceramic stopper

? (oil?) ? (salsamenta? wine?) ? ? (oil?) ? (oil? wine?) ? (oil? wine?) ? ? (oil?)

traces of oil

□

Keay 61C, Keay 61 A/D Keay 8A Keay 34 Keay 50 ‘orlo a fascia’ Schöne-Mau XXXV pseudo-Dressel 2/4 Benghazi MR 1 Dressel 30 Keay 1B Carthage LRA 1 Globular 3 (‘Castrum 65 Perti’) 66 Bonifay 57 67 Bonifay 58 68 Augst 73 ■: traces of pitch; □: no traces of pitch.

Archaeometry

? (oil?) ? (salsamenta? wine?) ? (oil?) ? ? ? ?

□ ? ■ ? ? ? ■ ? ? ■ ■ ?

imitation of wine amphora imitation of wine amphora imitation of wine amphora imitation of wine amphora imitation of wine amphora? imitation of wine amphora

wine wine wine wine wine? salsamenta? wine

?

imitation of wine amphora?

wine?

? ? ?

garum? garum? ointment?

Figure 1. African amphora contents. Summary of hypotheses. 2.2. Stamps

2.3. Dipinti

(Figure 3.27) from Pompeii, which bear the wording shmn — ‘oil’ in Punic (Panella 1977: 145). Normally, the dipinti on the Tripolitanian amphorae from the Monte Testaccio (Aguilera 2007) consist of tria nomina abbreviations, sometimes followed by two letters mentioning the social status of the person (the negotiator?), and below by a number indicating the quantity of oil transported. Capacity is also specified on graffiti that sometimes replace the dipinti (Liou and Silvino 2005: 727, PSG 11). In two cases, the warehouse at the port of departure — Hadrumetum — is also mentioned on Africana II amphorae (Torres 2007). For that reason, one may ask whether the mention of a specific content or a specific origin is intended to warn the consumer that this content (or origin?) is not the usual one according to the amphora shape. This is the hypothesis put forward by André Tchernia.

Dipinti which are rare on African amphorae, rarely provide information on content (this is true of other types of amphorae too: Martin‑Kilcher 2011: 417). One exception is represented by two Tripolitana I amphorae

Following this assumption, a possible example of unusual content is given by the wording olei on an amphora Africana II from Ostia, which led us to believe, erroneously, that this type was the normal container for olive oil from

Stamps rarely provide information on the content. Those present on African amphorae generally mention tria nomina; sometimes in Tripolitania the social status of the individual or the imperial status of the workshop; and sometimes in Mauretania Caesariensis, Zeugitana and Byzacena, the name of the city where a workshop is located. Nevertheless, one case was recently evidenced at Leptiminus of a stamp likely to inform us about the content: A series of stamps mentioning the city name LEP(timinus), are decorated with vine scrolls (Mattingly et al. 2011: 217 and fig. 6.11). D. Mattingly and D. Stone interpret this stamp as a possible evidence for a wine content in amphorae Africana II D and Africana III (Figure 3.26 and 29), on which this stamp is printed.

283

Michel Bonifay

Roman Africa (Zevi and Tchernia 1969: 187). An example of unusual origin is perhaps illustrated by the dipinti on spatheia type 3 (Figure 3.34‑37) discovered in Egypt, where the content — garum — is associated with names of monasteries. Among these latter, at least one — Saint Menas — can only be located in Egypt. We don’t know if these African containers were reused for the packaging of Egyptian fish sauce, or if the African garum was distributed through Egyptian monasteries (Fournet and Pieri 2008; see also J.‑L. Fournet in this volume). More difficult to solve is the case of the dipinti on amphorae Ostia LIX in Pompeii (Figure 2.16), in particular the wording TAVR previously translated (uinum)Taur(omenitanum) — wine from Taormina (see most recently Peña 2007). This hypothesis has to be rejected, because the type Ostia LIX, also named ‘Type 2020’ (Scotti 1984: Pl. 160), is definitely an African amphora, not a Sicilian one. Instead of Tauromenium, the name of Tauraca, an alteration of Thabraca/Tabarka was recently proposed, as a working hypothesis. As the other abbreviations attested on the same type in Pompeii, MOL, VIR et STR, seem to refer to different kinds of olives, (oliuae) MOL(lles), VIR(ides), and STR(ictiuae), the wording TAVR could have a geographic significance, in relation to the preparation of the olives — (olivae) TAVR(acenses) (Bonifay et al. 2015). Even if this set of assumptions should prove to be correct, we don’t know whether olives were the usual content of Ostia LIX, or if it was a particular content signalled by the dipinto, or even the reuse of the amphorae. An amphora of the same type discovered in the House of Menander in Pompeii carries a curious dipinto N?AP/OLEVM. In order to explain the traces of moringa oil evidenced by chemical analyses in four amphorae Ostia LIX found in Arles (see below), it has been recently proposed to develop NAP as Napata, a city of Upper Egypt where moringa growing is attested from the Roman period onwards (Djaoui, Garnier and Dodinet 2015). In this case the dipinto could obviously signal the reuse of the container. Finally, it is more common to find dipinti with content information on amphorae of Punic tradition. Several amphorae from Pompeii, Herculanum and Rome, derivatives of type Van der Werff 1, have dipinti (Martin‑Kilcher 1999: fig. 10) mentioning fish products (hallec, muria) that are for the moment difficult to define more precisely (Martin-Kilcher 2011: 417).

2.4. Lead labels Lead labels seem to be characteristic of African amphorae and have been discovered wrapped around handles of Africana II C amphorae (Figure 3.25) in the Annaba wreck. Robert Lequément demonstrated that the officina(e) mentioned on these lead labels might be interpreted as fish‑salting factories, a hypothesis also reinforced by the representation of a trident on one of these labels (Lequément 1975). Other lead labels have been found on some other wrecks, always associated with a cargo of Africana II C (e.g. on the wreck of Pampelonne: Lequément 1976). 2.5. Workshop location Using workshop location as a clue to content is difficult in Africa proconsularis because the workshops are not located close to foodstuff production centres, but close to the ports in charge of the distribution of these foodstuffs. Thus, at Salakta, Leptiminus and Nabeul amphorae for the distribution of oil, salsamenta, and wine were produced, products that were not necessarily produced on the territory of these cities (for Salakta and Nabeul: Ben Moussa, Bonifay and Nacef forthcoming; for Leptiminus: Mattingly et al. 2011). Some exceptions are reported in Byzacena, for example the workshops of the South side of the sebkhra Sidi el-Hani, which produced the so-called ‘pseudo‑Tripolitanian’ Africana II amphorae (Figure 2.18), the content of which was certainly not fish products. The situation is different in Tripolitania, at least in the hinterland of Oea and Lepcis Magna, where the workshops seem to be linked with agricultural properties, as shown by the Tripolitana III amphorae (Figure 3.28) produced in the Tarhuna plateau region (Ahmed 2019). One remaining problem is that amphorae Tripolitana II (Figure 2.9) were produced not only on the coastline, West of Lepcis Magna (Capelli and Leitch 2011) and at Oea (Panella 1973: 564), but also in the Tarhuna plateau region (Ahmed 2019). Should we think about two different contents (salsamenta and wine) for the same amphora type? Workshop location was also used in order to assign a wine content to the amphorae Van der Werff 2‑3 (Figure 2.2‑3) produced in the centre of Jerba Island (Fentress 2001). 2.6. Distribution There are very few new elements in this field. We can just confirm the primacy of Africana I and Tripolitana I/III

Figure 2. African amphora contents. 1: Van der Werff 1 (Ramon 1995: T7421); 2: Van der Werff 2 (Ramon 1995: T7422); 3: Van der Werff 3 (Ramon 1995: T7411); 4: Schöne-Mau 40 (Panella 2001; Pompeii); 5: Hammamet 1 (Bonifay 2004; Pupput); 6: Hammamet 1D (Bonifay 2004; Pupput); 7: Hammamet 2 (Bonifay 2004; Pupput); 8: Hammamet 3 (Caprino 1999; Tivoli); 9: Tripolitaine II (Bonifay 2004; Pupput); 10: Benghazi LR7 (Bonifay, Capelli, Muçaj 2010; Lepcis Magna); 11: Leptiminus II (Bonifay 2004; Marseille); 12: Ancient African (Sciallano and Sibella 1991; Camara 2 wreck); 13: Carthage EA IV (Martin-Kilcher 1999; Carthage); 14: Leptiminus I (Opaiţ 2000; Leptiminus); 15: Uzita Pl. 52, 10 (Bonifay 2004; Rougga); 16: Ostia LIX (Bonifay et al. 2015; Pompeii) 17: Ostia XXIII (Excoffon and Pellegrino 2015; Fréjus); 18: Africana II ‘pseudo-Tripolitanian’ (Bonifay 2004; El Jem); 19: Bonifay 16 (Bonifay 2004; Pupput); 20: Bonifay 17 (Bonifay 2004; Pupput); 21: Station 48 Piazzale delle Corporazioni (Bonifay 2004; Pupput).

284

African amphora contents: an update

?

?

?

?

?

? ?

?

?

?

0

10 cm

1

2

3

4

5

6

oil

wine

salsamenta

7

olives

interchangeable

unknown

LEP_A08

? ?

?

?

?

? ?

11

0

1

10 cm

8

?

9

?

10

?

12

?

13

?

14

?

?

?

?

0

10 cm

15

16

17

18

19

285

20

21

Michel Bonifay

amphorae (Figure 3.22 and 27‑28) at Monte Testaccio, which was the major dump for oil amphorae in Rome at least until the mid 3rd century AD. Along with these three dominant types, some other African amphorae are also present. Some of them can be considered as exceptions, either in regard to their content (following the example of the Africana II amphora from Ostia with a dipinto olei), or in regard to the place where they were discarded, if we accept that amphorae for fish products or wine could have been exceptionally dumped at the Monte Testaccio. Among these exceptions, I would place the types Africana II A, II B, II C and II D (Figure 3.23‑26), as well as the type Carthage EA IV (Figure 2.13), generally quite rare. Nevertheless, one might note that a greater proportion of Africana II A has been recently evidenced in contexts of the first half of the 3rd century (Revilla 2010: 399–400). On the other hand, the presence at Monte Testaccio of some minority types could imply that they were devoted to the transportation of olive oil. This could be the case for the Africana II ‘pseudo-Tripolitanian’ (Figure 2.18) (Revilla 2007: fig. 15, no.1‑3, rather than Africana II D), and Hammamet 1E (Figure 2.6) (Revilla 2007: fig. 3, no.16‑17), which differs from the other variants of the same type by its greater size (Bonifay 2004: 96 and fig. 49). Lastly, it is particularly interesting to note that among the two types attributed to the early production of North-Western Zeugitana, Ostia LIX and Ostia XXIII (Figure 2.16‑17), only the second one is attested in contexts of the 2nd century at Monte Testaccio (Revilla 2007: fig. 3.9‑14, and fig. 14.2‑3, rather Ostia XXIII than Ostia LIX). This clue for a difference of content for the amphorae Ostia LIX and XXIII is supported by observations made on the cargo of the wreck ‘of the Columns’ at Camarina, where the amphorae Ostia LIX were sealed with a cork stopper, while the containers Ostia XXIII without traces of pitch (see below) were sealed with a ceramic stopper (Bonifay, Franco and Cacciaguerra 2016: 391–392), an observation that gives to the latter type another characteristic of a container for oil, and justifies its presence at Monte Testaccio. 2.7. Modes of sealing and opening the amphorae The distinction between ceramic and cork stoppers was investigated between 2004‑7. The former are mainly associated with Africana I (Bonifay 2004: 467 and fig. 263, no.1), and some late types, such as amphora type

Albenga 11/12 (Figure 3.52) from the wreck of La Palud (Long and Volpe 1998: fig. 297, no.80). The latter are rather associated with Africana II and III (wrecks of Giglio Porto and Cap Blanc, see Parker 1992: 193 and 99), spatheion 1, and Keay 35B (Figure 3.33 and 42) (wreck of Dramont E, see Santamaria 1995: 125). In spite of some contradictory data, the previous example of the Camarina wreck shows that the mode of sealing the amphorae could help distinguish a possible difference of content in two similar types of amphorae. On the other hand, the criterion of the mode of opening the amphorae (Bonifay 2004: 467–470) needs further discussion and analysis. 2.8. Macro-remains The most frequently reported macro-remains in African amphorae coming from underwater excavations are shells and crustaceans, but it is not possible to know whether they are evidence of the content or traces of colonization of the amphora by these animals after the sinking of the ship (see also Koutsouflakis in this volume). Fish remains, generally stuck to the pitch, are a more reliable testimony of the original content. Less frequently mentioned in the literature, these remains are more specifically observed in Africana II amphorae, in most cases without specification of the variant (Bonifay 2007: fig. 3a, table drawn up after Parker 1992). Olive pits are also reported in amphorae Africana II, III C and spatheia. Lastly, a neck of type Africana II D, coming from the wreck of Pointe de la Luque B, with grape pips glued to the pitch, could suggest a wine content (Bonifay 2004: fig. 262.4); nevertheless, we must also take account, following Stefanie Martin‑Kilcher, of more complex products mixing fish sauce and wine (Martin‑Kilcher 2011: 420). 2.9. Pitch New data on this issue are abundant, and suggest, on first reading, to a complete reversal of previous hypotheses (see also Garnier and Pecci in this volume). Pitch was regarded ten years ago as the cornerstone of the argumentation on (African) amphora contents. As a matter of fact, if we examine amphorae coming from underwater excavations or port silting — it is preferable to deal with anaerobic contexts to get the pitch conserved — we can see with the naked eye that

Figure 3. African amphora contents. 22: Africaine I (Bonifay 2004; Pupput); 23: Africaine II A (Bonifay 2004; Pupput); 24: Africaine II B (Bonifay 2004; Pupput); 25: Africaine II C (Panella 1973; Ostia); 26: Africaine II D (Bonifay 2004; Planier VII wreck). 27: Tripolitana I (Panella 1973; Lepcis Magna); 28: Tripolitana III (Panella 1973; Lepcis Magna); 29: Africana III A (= Keay 25.1) (Panella 1973; Thaenae); 30: Africana III B (= Keay 25.3) (Bonifay 2004; Antibes); 31: Keay 25 V (Keay 1984; Tarragona); 32: Africana III C (= Keay 25.2) (Joncheray 1975; wreck Dramont E); 33: spatheion type 1 (Santamaria 1995; wreck Dramont E); 34: spatheion 3A (Villa 1998; Udine); 35: spatheion 3B (Bass et al. 1982; Yassi Ada wreck); 36: spatheion 3C (Gagnière 1965; Saint-Etienne de Candau) 37: spatheion 3D (Biffino et al. 1998; San Giusto); 38: Keay 27 (Keay 1984; Tarragona); 39: Keay 36 (Keay 1984; Can Samarruga); 40: Keay 59/8B (Bonifay, Capelli and Moliner 2011; Marseille); 41: Keay 35A (Opaiţ 1997-98; Tomis); 42: Keay 35B (Bonifay 2004; Dramont E wreck); 43: Keay 39 (Bonifay, Capelli and Moliner 2011; Marseille); 44: Keay 57 (Sciallano and Sibella 1991; Fos-surMer); 45: Keay 56 (Palol 1987; Fornels); 46: Keay 55 (Opaiţ 1997-98; Tomis).

286

African amphora contents: an update

?

?

?

?

?

?

23

24

0

10 cm

22

?

?

25

?

?

26

?

?

34

? 35

36

37

27

28

?

? ?

?

31

? 0

10 cm 0

33

10 cm

29

30

32 oil

?

wine

?

?

41

42

38 salsamenta

olives

interchangeable

39

unknown

0

10 cm

40

43

287

44

45

46

Michel Bonifay

some of them bear traces of pitch, while others do not. Many similar observations were made during field work (Bonifay 2004: 464, fig. 260; Bonifay 2007: 12, fig. 3, both tables drawn up after Parker 1992), the best example being that of the Dramont E wreck, where amphorae type Keay 35B (Figure 3.42) show traces of pitch, while amphorae type Keay 35A (Figure 3.41) do not. Another survey was carried out a posteriori, in 1993, in the storerooms of the French Department of maritime and underwater excavations (DRASSM), with the observation of more than 200 complete or fragmentary African amphorae, stating that no Africana I amphorae had traces of pitch, while most of the Africana II, III and late types (Keay 35B, 62) did (Bonifay 2004: 465–467, fig. 261; for a complete inventory, see Ben Lazreg et al. 1995: 133–135). Following this survey, I proposed to assign content other than olive oil to all the African containers showing traces of pitch — that is to say the majority.

see. Nicolas Garnier perfectly summed up the situation after having analysed Spanish amphorae Dressel 20 together with amphorae Africana I, all of them coming from underwater excavations in the river Rhone. He could demonstrate that all these amphorae contained traces of olive oil, and almost all of them traces of pitch as well. But, even if these amphorae were coming from an anaerobic context, no traces of pitch were visible to a naked eye. He explained this phenomenon by the dissolving action of oil, which was deeply impregnated in the ceramic material, and the disappearance of the pitch from the internal surface of the container (Garnier et al. 2011: 411: ‘comme la poix est poussée à l’intérieur des parois, il est normal de n’en retrouver aucune trace visible, ou exceptionnellement’). The macroscopic observation of a lack of pitch, even if proved to be false from a chemical point of view, can help to identify the content of an amphora — in this case possibly oil.

Archaeometry (see below) has provided the evidence to contest this model. Indeed, the analysis by Nicolas Garnier of a series of samples in 2004 (Garnier, Annexe, Analyse du contenu d’amphores africaines, in Bonifay 2007: 25–29) showed traces of pitch together with traces of oil in one African amphora Keay 62A (Figure 4.53) coming from La Palud wreck, even though no traces of pitch were visible to the naked eye. This observation, considered as an exception at the time, was explained by the nature of oil transported in this amphora (fuel for lamps) and/or a possible reuse of the amphora (Garnier in Bonifay 2007: 29). This first reassessment of the principle governing the characterization of the amphorae devoted to the transportation of olive oil was followed by many others over the last few years. The analysis by Kerlyne Romanus of samples collected on Eastern amphorae from Sagalassos, carried out in parallel with a series of ex novo experiments, showed that pitch is fully compatible with oil (Romanus et al. 2009). The same results were obtained by Alessandra Pecci on amphorae from the port of Classe, as well as on Medieval ceramics (Pecci et al. 2010; Pecci 2009). These results challenge the argument that traces of pitch exclude oil as the content of an amphora (Pecci and Cau 2010; see also Garnier and Pecci in this volume).

2.10. Archaeometry

Nevertheless, faced with this new data, we must keep a ‘cool head’ (Bonifay and Tréglia 2010: 1037). The chemists’ viewpoint is not the same as the archaeologists’, or more exactly chemists can see things that archaeologists cannot

The development of archaeometry also helped with the determination of the contents themselves (see Garnier and Pecci in this volume). A series of surveys highlighted some unsuspected contents. We have already discussed the moringa oil detected in the amphorae Ostia LIX (Figure 2.16) from Arles (see above). But this is also the case for the marker of ricinoleic acid found in the African amphorae from storeroom 17 of the port of Classe, burnt at the end of the 5th century, composed of late variants of spatheia type 1 and large-sized amphorae of unknown typology (Classe/Magazzino 17, type 1, probably not Keay 57, 62Q and 61A, see Pecci et al. 2010: table 2). This marker was interpreted as the evidence for the transportation of castor oil in all these amphorae (Pecci et al. 2010). This kind of oil is known for having been used as fuel for lamps in Egypt, and as body ointment, but traces of production in North Africa are lacking. Considering that the same marker was found in amphorae of different typologies and origins (for example in a LRA 2 from Lesbos, where it was interpreted as a reuse of the container), I wonder whether ricinoleic acid provides evidence for the content or if castor oil was used during the preparation of the container, as an antiseptic (in the same way as the sulphur that was detected in some Africana I amphorae, see Garnier et al. 2011) or a component of the pitch lining. Indeed, by examining the analyses of the amphorae

Figure 4. African amphora contents. 47: Keay 64 (Wilson 1996; Campanaio); 48: Keay 3/5 (Bonifay 2016; Rome); 49: Keay 40 (Bonifay 2016; Cartagena); 50: Keay 41 (Remolà 2000; Tarragona); 51: Keay 62R (Freed 1995; Carthage); 52: Albenga 11/12 (Pallarès 1987; Albenga); 53: Keay 62A (Albore Livadie 1984; Filicudi Porto); 54: Bonifay 47 (Bonifay 2004; Marseille); 55: Keay 61C (Bonifay 2004; Arles); 56: Keay 61A/D (Liou 1975; Pomègues); 57: Keay 8A (Murialdo 1988; Pieve del Finale); 58: Keay 34 (Bonifay 2004; Rougga); 59: Keay 50 (Radic 1993; Lastovo wreck); 60: ‘con orlo a fascia’ (Cunja 1996; Koper); 61: Schöne-Mau XXXV (Panella 1973; Pompeii); 62: pseudo-Dressel 2/4 (Bonifay 2004; Zarzis); 63: Dressel 30 (Panella 1973); 64: Keay 1B (Long 2002; Camargue); 65: Globular Type 3/Castrum Perti (Saguí 1998; Rome). Small containers: 66: Bonifay 57 (Bonifay 2016; Lambaesis); 67: Bonifay 58 (Bonifay 2004; Nabeul). 68: Augst 73 (Bien, Bonifay and Moliner 2007; Marseille).

288

African amphora contents: an update

?

?

52

53

?

0

10 cm

47

48

49

50

51

?

0

10 cm

54

55 oil

56 wine

57

salsamenta

olives

interchangeable

59

58

60

unknown

0

5 cm

? ?

?

?

?

66

67

0

10 cm

61

62

63

64

65

289

68

Michel Bonifay

from Classe, it is troubling to note that the marker of acid ricinoleic is always associated with the marker of pitch, and in about half of the cases with markers of other contents — other vegetal oils and even fermented products (Pecci et al. 2010: Table 2). Archaeometry was also used in order to better characterize the two other main products transported by amphorae beside oil, namely fish products and wine. Concerning the salsamenta, very innovative methods were used: paleogenetic on one side, and proteomic on the other side. The first ones allowed for the analysis of the ancient DNA present in residues conserved in the bottom of amphorae type Africana II from the Parc Saint-Georges excavations in Lyon. In one of them, probably originating from Salakta according to the petrography, the analyses revealed the presence of fish‑products composed of anchovies (Silvino et al. 2007). The second method is still at a development stage, but results are encouraging on samples of present‑day ceramics soaked in a fish mixture (Dallongeville et al. 2010). Lastly, the more traditional method by lipidic extraction led to the development of a protocol for the research of fish‑based contents, notably in comparison with samples collected in salting vats of the fish-salting factory of Nabeul (mentioned in Garnier et al. 2011: 405; about the fish‑salting factories of Nabeul, see Slim et al. 2007). Concerning wine, the DNA analyses provided evidence for such content in an African amphora of unknown typology from Lyon (Silvino et al. 2007: fig. 3, no.4). On the other hand, the protocols for traditional chemical analyses are now well developed, and have already provided results, mainly for Africana III (see below). 3. Updating of the content distribution by types The quantity of available data has significantly increased over the last ten years, but their fragmentary and casual nature have not notably changed. It is certainly possible to update the hypotheses made in 2007, but we must bear in mind that they remain fragile. 3.1. Oil We know very little about the content of the earliest RomanAfrican types. Traces of oil seem to have been detected in the so-called ‘Ancient African’ type (Figure 2.12) (mid 2nd century BC‑early 1st century AD), but this should be checked in future (Contino and Capelli 2019: 58). According to the ovoid shape of this container, a fish-based product would also be plausible (Martin‑Kilcher 2011: fig. 3), and wine cannot be excluded by comparison with some other ovoid amphorae produced in Spain (García Vargas et al. 2019: 410). The same issue remains pending for the successor of this type in the 1st century AD, the Dressel 26 (Contino et al. 2016: 155). Only one type of Punic tradition from the Hammamet gulf (type 1E) (Figure 2.6), attested at the Monte Testaccio, could have transported olive oil, but this amphora is poorly distributed.

All the arguments mentioned above demonstrate that the African amphorae of the 2nd‑3rd century devoted par excellence to the transportation of oil are Africana I (Figure 3.22), and Tripolitana I and III (Figure 3.27‑28). Among these arguments, their presence at the Monte Testaccio is not the least, and allows us to include the type Ostia XXIII (Figure 2.17), which has a slightly early onset. On the other hand, oil is probably not the usual content of type Ostia LIX or Africana II, except perhaps the so-called ‘pseudo-Tripolitanian’ variant of the latter. The fact that it is more difficult to follow the evolution of these oil containers from the 4th century onwards has led us to consider a crisis in African olive growing during this period (Panella 1982: 178). Nevertheless, it is virtually certain that Africana I survives as a variant of Africana III (the same evolution is known for the amphorae Africana II C). It remains to be discussed which variant(s) of type Africana III was/were devoted to the transportation of oil. A good candidate could be the variant Keay 25 V (Figure 3.31), due to the similarity of rim with the latest examples of the Africana I. Tripolitana III is produced throughout the 4th century, even if less distributed than during the 3rd century (Panella 2001: 211). In the 5th century, other types seem to take over, but there is insufficient information on their content: Keay 27 (?) in the region of Carthage (Figure 3.38), Keay 35A in Nabeul (Figure 3.41), and Keay 59/8B in Southern Byzacena (Figure 3.40). The situation is even more uncertain in the Sahel region and in Tripolitania, where data on amphora production during the 5th century is almost completely lacking. In the 6th century, types Keay 62Q/Albenga 11/12 and mainly Keay 62A (Figure 4.52‑53) are possible candidates for the transportation of olive oil, while the situation becomes totally unclear in the 7th century (types Keay 61 and 8A) (Figure 4.56‑57). 3.2. Fish products Republican African amphorae of Punic tradition (type Van der Werff 1‑3) (Figure 2.1‑3) and their ‘Romanized’ successors (type Carthage EA IV) (Figure 2.13) are traditionally considered to have transported fish products (Martin‑Kilcher 1999 and 2011; Botte 2009: 115). Nevertheless, wine has been also suggested for Van der Werff 2‑3 on the basis of the archaeological survey of Jerba (Fentress 2001), and recent chemical analyses (Pecci and Giorgi 2019). The most difficult problem to solve is the content of Imperial types Tripolitana II, Dressel 18 and Hammamet 1‑3 (Figure 2.9, 2.5, 2.7‑8): as for their predecessors, we might be tempted to expect a fishbased main content, but the location of their production, sometimes far from the coast, weakens this exclusive hypothesis: wine is also a possible solution. Lastly, on a morphological bases — in particular the long hollow spike often considered as typical of the amphorae for salsamenta, a fish‑based content is suggested for the early Imperial types Leptiminus I and Uzita Pl. 52.10 (Figure 2.14‑15).

290

African amphora contents: an update

The container which could better correspond to the abundant production of salsamenta in Africa (Ben Lazreg et al. 1995) during the mid Roman period is the Africana II. There is a high probability for the variants A and C, which are linked from a typological point of view, and provide a series of arguments, including archaeometric ones (pitch traces, macro-remains, lead labels, DNA analyses). Nevertheless, it is not certain that all variants of Africana II had the same content (Panella 1982: 176). In particular, there is a doubt for variant D (Figure 3.26), which could equally contain wine, as suggested by D. Mattingly and D. Stone (Mattingly et al. 2011: 217). Africana II C (Figure 3.25) survives in the first half of the 4th century through the adoption of the slim shape of Africana III (Bonifay 2004: types Africana II C2 and II C3). On the other hand, we don’t know if other medium-sized cylindrical amphorae classified under type Africana III had the same content. It could be the case of variant III B (= Keay 25.3) (Figure 3.30) of the second half of the 4th century, the rim shape of which seems to announce the development in the 5th century of the large‑sized amphorae of type Keay 35B (Figure 3.42). The latter type, produced in Nabeul, with different sizes (Bonifay, Capelli and Moliner 2011: 240), brings together several arguments for a fish-based content (traces of pitch, presence on fishsalting fabrics, macro-remains in one case). Nothing secure can be said about types Keay 57‑56‑55 (Figure 3.44‑46), which seem to supersede Keay 35 at the end of the 5th century and the 6th century, unless they have traces of pitch and we know that Keay 55, at least, didn’t contain oil (Garnier in Bonifay 2007: 26‑27). The latest evidence of transportation of African fish sauces is given by the dipinti of difficult interpretation on spatheia type 3 found in Egypt. Besides these series of amphorae, even small ones, there are some very small containers devoted to the transportation of very small quantities of products (see below). Should we not consider that some of them (types Bonifay 57 and 58) (Figure 4.66‑67) could have transported fish sauces (Bonifay 2004: 288)? 3.3. Wine The reluctance expressed in the past about the production and distribution of African wine has faded over the last years. But the support of the scientific community for a wine content in several amphora types of the 4th perhaps went too far. Surely, there is no doubt that wine was the main content of African amphorae whose form imitates the typology of wine amphorae from Italy or Gaul (Schöne‑Mau XXXV, pseudo-Dressel 2/4, Dressel 30) (Figure 4.61‑63), even if some cases are dubious (e.g. the content of type Keay 1B, see Reynolds 2010: note 193) (Figure 4.64). The same hypothesis for Africana III is more hazardous. Robert Lequément was the first to suggest that the medium-sized cylindrical amphorae transported wine

(Lequément 1980: 191). He was relying on the great number of exemplars presenting traces of pitch, which has been confirmed by the systematic examination of the amphorae stored in the DRASSM storerooms. In addition, the wide distribution of these amphorae better corresponded to the distribution of wine than to the distribution of oil or fish products. The fact that these amphorae appear at the end of the 3rd century or more probably at the beginning of the 4th century, at the same time as the canon uinarius is established, is troubling. Lastly, the help of archaeometry was decisive, with the discovery of wine traces in four exemplars of Africana III A from three different underwater sites (Woodworth et al. 2015: wrecks of Héliopolis 1, Pointe de la Luque B, and excavations in the river Rhone). This hypothesis of wine for the amphorae Africana III, put forward with great caution, concurrently with the hypothesis of salsamenta that is more generally accepted (Reynolds 2010: note 331), was reproduced as a certainty in many publications dealing with this amphora type. Nevertheless, the arguments remain contradictory. Firstly, all variants of Africana III, which are not always recognizable with a simple rim sherd, did not transport the same content (see above for variants with a possible content of oil or salsamenta, even if I agree with Paul Reynolds that ‘[Keay 25] did not normally carry oil’, see Reynolds 2010: note 331). On the other hand, the resemblance of Africana III to Africana II D is striking, and could the key to the problem, if we knew the content of the latter type. But for this type also, we hesitate between salsamenta and wine. The geographical proximity between the amphora workshops and the fish-salting factories could be an argument against the wine hypothesis, but in two workshops (Oued el‑Akarit and Nabeul) the production of others types corresponding to other contents (late variant of type Africana I) is also attested. Lastly, the solution can only come from archaeometry. Research protocols for identifying a wine content are beginning to show results, and it seems that traces of wine are frequently highlighted in amphorae Africana III (Alessandra Pecci, personal communication), at least for variant A (Figure 3.29). 3.4. Other contents and reuse of amphorae Besides the three main amphora contents, oil, fish products and wine, other minority contents should be considered. For example, preserved olives are possibly mentioned in dipinti at Pompeii on type Ostia LIX (see above) and at Augst: oli(vae) on an amphora Africana III (Martin‑Kilcher 1994: no.248), and attested by olive pits in amphorae Africana III C and spatheia type 1 (Figure 3.32‑33) in the Dramont E wreck (Santamaria 1995). In the same wreck, traces of preserved meat have been also reported in fragmentary African amphorae of unknown typology (Santamaria 1995: 123). Nevertheless, in this last example as well as in others mentioning the discovering of fig pips or date stones,

291

Michel Bonifay

we must always bear in mind that the containers could have been reused. This could be a possible explanation for the presence of specific vegetal oils — castor oil or moringa oil — in some African amphorae (see above). This phenomenon of reusing the containers, which perhaps does not have the same frequency according to the region or chronology, is difficult to assess, but it need not stop us trying to determine the primary content of each amphora form. 3.5. Interchangeable contents A specific problem is raised by spatheia type 1 (Figure 3.33). These amphorae were sometimes interpreted as small variants of Africana III C, with a similar content. Nevertheless, the examination of several exemplars coming from underwater or port excavations progressively convinced me that many different contents were possibly attested in these small-sized amphorae. Indeed, the exemplars from the Dramont E wreck obviously transported preserved olives; another one discovered at Tarragona contained macro‑remains of fishes; lastly, some others, although complete, didn’t show any traces of pitch. This inconsistency in the data led me to consider that the spatheia were invented in order to serve as a space filler in cargoes composed of large‑sized cylindrical amphorae. Indeed, the distribution map of amphorae in the Dramont E wreck show that the spatheia were inserted, even in a disorderly manner (Poveda 2012), in the interstices left with the stacking of Keay 35A and B in the ship’s hold (Santamaria 1995: fig. 9). According to this hypothesis, the spatheion type 1 would not be a specific amphora type with a specific content, but an instrument of optimisation of the cargo, created at the moment when the distribution mode of the African foodstuffs was changing, and the traders were perhaps inclined to make the cargoes as full as possible. The lifetime of this small container seems to be short, between the beginning of the 5th century and the beginning of the 6th century, but the small containers spatheia type 3 of the 7th century don’t necessarily have filiations with those of the 5th century. 3.6. Other containers First of all, it must be reminded that in Africa amphorae were used not only for transportation, but also for the storage of foodstuffs, in the absence of dolia (Brun 2003), at least in Africa proconsularis (not in Mauretania caesariensis, see Amraoui 2017: 282–285). These African storage amphorae, which cannot be classified within the classical transport amphora classifications, adopt different morphologies, some of Punic tradition, others of GrecoRoman type, and all of them reach impressive capacities (up to 140 litres). Exceptionally, these amphorae could have travelled. Some examples are known on wrecks, but in this case they don’t necessarily belong to the cargo (for example at Pakostane: Huguet 2012), and even on some terrestrial sites (for example, at Rome: Contino and D’Alessandro 2014).

On the other hand, as in other regions of the Empire (especially in the Latium), small containers, which we cannot call amphorae, even if they have two handles (see type Bonifay 57), are suspected to have transported small quantities of products. We already mentioned the small pots that could have contained fish sauces (types Bonifay 57 and 58, see above), but a lot of other forms could have a similar purpose, notably a series of small jugs of the late Antique period (especially types Bonifay 61, 62.6, etc.). The content of these small containers is unknown, except the content of type Augst 73 (Figure 4.68), which could be an ointment (Marty and Capelli 2016). Finally, the storage and transport of foodstuffs in skins and barrels in Roman Africa is now being more recognised, as attested by a good deal of representation on sculptures, mosaics and ceramics (Marlière and Torres Costa 2007). It is certain, as mentioned by the ostraca from Carthage (Peña 1998), that oil produced in the inland regions of Africa was transported in skins to the port warehouses, where it was poured into amphorae, and then traded overseas. It is also probable that most of the intraprovincial terrestrial trade of wine, oil and salsamenta, was made by means of skins and barrels. Lastly, it is not completely impossible that part of the maritime trade of Africa used barrels. This issue remains to be confirmed, in particular for the fish products (Marlière and Torres Costa 2007: 106 and note 97), and the later period (Horden and Purcell 2000: 372, ‘the post‑amphora Mediterranean’). Nevertheless direct evidence is lacking in order to support this hypothesis. 4. Conclusion So, fifty years after the first synthesis on Roman Africa amphorae (Zevi and Tchernia 1969), we can say that their content is still poorly documented. To sum up, the presumed contents of African amphorae can be divided into three categories: 1. those which are unanimously accepted; 2. those which are matter of controversy; 3. and those which remain completely conjectural. Within the first category, we can be sure that oil was the usual content of Africana I, and Tripolitana I and III; that fish products were the content of Africana II C; and that wine was the content of African imitations of Dressel 2/4 (= Schöne-Mau XXXV and local variants), Gauloise 4 (= Dressel 30) and LRA 1 (= Henchir Chekaf IV). In the second category, we can classify: a. the amphorae of Punic tradition, for which we still hesitate between salsamenta and wine; b. the early Roman-African types (Ancient African, Dressel 26, Ostia LIX) for which different contents have been suggested (oil, preserved olives, salsamenta, wine);

292

African amphora contents: an update

c. the Africana II, mainly its variant D, the content of which could be salsamenta or wine; d. Africana III, though it is not completely clear which variants were carrying wine (variant A?), salsamenta (variant B?), and eventually oil. Lastly, the third category — the ‘grey area’ — gathers almost all the large‑sized amphorae of the late Antique period, for which we can only put forward some very fragile hypotheses, for example: Keay 27 (oil?), Keay 35A (oil?), Keay 35B (salsamenta?), Keay 39 (oil?), Keay 59/8B (oil?), Keay 62 (oil?), etc. (Figure 1). As in other regions of the Empire (e.g. in Spain), we face the same difficulties of content identification during a period of experimentation in amphora production, which seems to last perhaps longer in Africa, from the fall of Carthage to the end of the 1st century‑beginning 2nd century AD. The contents are clearer from the mid 2nd century to the end of the 4th century, the more standardised African amphorae becoming at this stage ‘intelligent packaging’, as in the other regions of the Empire. Lastly, as everywhere, the most difficult problems are posed by the late Antique amphorae, from the 5th century onwards. Fifty years ago, at the beginning of the research on African amphorae, it was unanimously stated that olive oil was the main content of these containers. Then, this hypothesis was contested, and more recently African amphorae have been less and less considered as having carried olive oil. In fact, the point is not to deny the possibility that African amphorae transported oil, but to accept the idea that the contents of these amphorae were more diverse than previously assumed. When dealing with African amphorae on consumption sites, we have to bear in mind not only that the contents are not always the same in the same period (e.g. Africana I and II), but also that the main products exported overseas may have changed over time: oil dominating between mid‑2nd and mid‑3rd century, then possibly salsamenta during the 3rd century‑beginning 4th century, and perhaps wine during the 4th century. It is distressing to see that the record on African amphora contents is perhaps one of the least advanced in the Roman Mediterranean, while these containers are among the most important in terms of production and distribution. More than ever is crucial to launch a programme of systematic chemical analyses on the contents of the African amphorae.

Bibliography Aguilera Martín, A. 2007. Les tituli picti des amphores oléaires tripolitaines et tunisiennes. In A. Mrabet and J. Remesal Rodríguez (eds), Africa et in Hispania: Études sur l’huile africaine: 257–268. Instrumenta 25. Barcelona, Universitat de Barcelona. Ahmed, M. A. M. 2019. Rural Settlement and economic Activity: Olive Oil and Amphorae Production on the Tarhuna Plateau during the Roman Period. Society for Libyan Studies Open Access Monograph 1. London, The Society for Libyan Studies. Albore Livadie, C. 1984. Relitto Porto A di eta tardoimperiale. In Archeologia subacquea 2. Isole Eolie. Bolletino d’Arte, Supplemento 29: 95–97. Rome, Istituto poligrafico e zecca dello Stato. Amraoui, T. 2017. L’artisanat dans les cités antiques de l’Algérie (Ier siècle avant notre ère - VIIe siècle après notre ère). Archaeopress Roman Archaeology 26. Oxford, Archaeopress. Bass, G. F. and F. H. Van Doorninck Jr. (eds) 1982. Yassi Ada, Volume I, A Seventh-Century Byzantine Shipwreck. College Station, Texas A&M University Press. Ben Lazreg, N., Bonifay, M., Drine, A. and Trousset, P. 1995. Production et commercialisation des salsamenta de l’Afrique ancienne. In L’Afrique du Nord antique et médiévale. Production et exportations africaines. Actualités archéologiques. Actes du VIe colloque d’Histoire et d’Archéologie de l’Afrique (Pau, 25-29 octobre 1993): 103– 142. Paris, CTHS. Ben Moussa, M., Bonifay, M. and Nacef, J.† forthcoming. Sullecthum et Neapolis : comparaisons entre deux villes portuaires de l’Afrique romaine. In Porti e retroterra produttivi. In Actes du colloque international (Livourne, 26-28 mars 2009). Biffino, A., De Felice, G., Fiorello, C. S., Lapadulla, E. and Pietropaolo, L. 1998. I materiali ceramici e metallici. In G. Volpe (ed.), San Giusto. La villa, le ecclesiae. Primi risultati dagli scavi nel sito rurale di San Giusto (Lucera) 1995-1997: 263–276. Bari, Edipuglia. Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique, British Archaeological Reports International Series 1301. Oxford, Archaeopress. Bonifay, M. avec une annexe de N. Garnier 2007. Que transportaient donc les amphores africaines ?. In E. Papi (ed.) Supplying Rome and the Empire. The Proceedings of an International Seminar Held at SienaCertosa di Pontignano on May 2-4, 2004 on Rome, the Provinces, Production and Distribution. Journal of Roman Archaeology Supplementary Series 69: 9–31. Portsmouth RI, Journal of Roman Archaeology. Bonifay, M. 2016. Annexe 1. Elements de typologie des céramiques de l’Afrique romaine. In D. Malfitana, and M. Bonifay (eds), La ceramica africana nella Sicilia romana - La céramique africaine dans la Sicile romaine. Monografie dell’Istituto per i Beni Archeologici e Monumentali, C.N.R. 12: 507–574. Catania, IBAM/CCJ.

293

Michel Bonifay

Bonifay, M. and Tréglia, J.-C. 2010. De Vigo a Voitenki, en passant par Pise et Parme. In S. Menchelli, S. Santoro, M. Pasquinucci, and G. Guiducci (eds), LRCW 3. 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): 1033–1039. Oxford, Archaeopress. Bonifay, M., Capelli, C. and Moliner, M. 2011. Les amphores africaines de la basilique de la rue Malaval à Marseille (Ve siècle). In SFECAG, Actes du Congrès d’Arles: 235–254. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Bonifay, M., Capelli, C. and Muçaj, S. 2010. Amphores tardives de tradition punique : observations sur le type Benghazi LR Amphora 7. In C. Ebnöther and R. Schatzmann (eds), Oleum non perdidit. Festschrift für Stefanie Martin-Kilcher zu ihrem 65. Geburtstag. Antiqua 47: 151–159. Bern, Archäologie der Schweiz. Bonifay, M., Franco, C. and Cacciaguerra, G. 2016. Analyse micro-régionale de la diffusion des céramiques africaines en Sicile. In D. Malfitana and M. Bonifay (eds), La ceramica africana nella Sicilia romana - La céramique africaine dans la Sicile romaine. Monografie dell’Istituto per i Beni Archeologici e Monumentali, C.N.R. 12: 353–401 and 735–839. Catania, IBAM/CCJ. Bonifay, M., Botte, E., Capelli, C., Contino, A., Djaoui, D., Panella, C. and Tchernia, A. 2015. Nouvelles hypothèses sur l’origine et le contenu des amphores africaines Ostia LIX et XXIII. Antiquités africaines 51: 189–210. Botte, E. 2009. Salaisons et sauces de poissons en Italie du Sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard, 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Brun, J.-P. 2003. Les pressoirs à vin d’Afrique et de Maurétanie à l’époque romaine. Africa, n. s., Scéances Scientifiques 1: 7–30. Capelli, C. and Leitch, V. 2011. A Roman amphora production site near Lepcis Magna: petrographic analyses of the fabrics. Libyan Studies 42: 69–72. Caprino, C. 1999. Rinvenimenti a Villa Adriana (Tivoli). Accademia Nazionale dei Lincei, Monumenti Antichi, Studi Miscellanei VI-1. Rome, Bretschneider. Carandini, A. 1970. Produzione agricola e produzione ceramica nell’Africa di età imperiale. In Omaggio a R. Bianchi Bandinelli. Studi Miscellanei 15: 97–119. Rome, De Luca. Contino, A. and Capelli, C. 2019. Late Republican and Early Imperial ovoid amphorae: the African production. In E. García Vargas, R. R. de Almeida, H. González Cesteros and A. Sáez Romero (eds), The Ovoid Amphorae in the Central and Western Mediterranean between the last two centuries of the Republic and the early days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13: 42–61. Oxford, Archaeopress. Contino, A. and D’Alessandro, L. 2014. Materiali ceramici dagli scavi della porticus Aemilia (Testaccio, Roma). Campagna di scavo 2011-2012. Rei Cretariae Romanae Fautorum Acta 43: 323–334.

Contino, A., Capelli, C., Milella, M., Pacetti, F., Ungaro, L. and Bonifay, M. 2016. L’anfora «Dressel 26» del Castro Pretorio. Antiquités africaines 52: 145–156. Cunja, R. 1996. Capodistria Tardoromana e Altomedievale. Lo scavo archeologico nell’ex orto dei cappuccini negli anni 1986-1987 alla luce dei reperti dal V al IX secolo D.C. Koper, Società storica del Litorale. Dallongeville, S., Garnier, N., Bernal-Casasola, D., Bonifay, M., Rolando, C. and Tokarski, C. 2011. Dealing with the identification of protein species in ancient amphorae. Analytical and Bioanalytical Chemistry 399 (9): 3053– 3063. Djaoui, D., Garnier, N. and Dodinet, E. 2015. De l’huile de ben identifiée dans quatre amphores africaines de type Ostia LIX provenant d’Arles : difficultés d’interprétation. Antiquités Africaines 51: 179–187. Excoffon, P. and Pellegrino E., with an appendix by C. Capelli 2015. Amphores africaines du type Ostia XXIII à Fréjus. Antiquités africaines 51: 155–165. Fentress, E. 2001. Villas, wine and kilns: the landscape of Jerba in the late Hellenistic period. Journal of Roman Archaeology 14: 249–268. Fournet, J.-L. and Pieri, D. 2008. Les dipinti amphoriques d’Antinoopolis. In R. Pintaudi (ed.), Antinoupolis I. Scavi e materiali 1: 175–216. Florence, Istituto Papirologico ‘G. Vitelli’. Freed, J. 1995. The late series of Tunisian cylindrical amphoras at Carthage. Journal of Roman Archaeology 8: 155–191. Gagnière, S. 1965. Les sépultures à inhumation du IIIe au XIIIe s. de notre ère dans la Basse Vallée du Rhône. Essai de chronologie typologique. Cahiers Rhodaniens XII: 53–110. García Vargas, E., Almeida R. R. de, González Cesteros, H. and Sáez Romero, A. (eds) 2019. The Ovoid Amphorae in the Central and Western Mediterranean between the last two centuries of the Republic and the early days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13. Oxford, Archaeopress. Garnier, N., Silvino, T. and Bernal-Casasola, D. 2011. L’identification du contenu des amphores : huile, conserves de poissons et poissage. In SFECAG, Actes du Congrès d’Arles: 397–416. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Horden, P. and Purcell N. 2000. The Corrupting Sea. A Study of Mediterranean History. Oxford, Blackwell. Huguet, C. 2012. La céramique et les matériaux de construction en terre cuite. In G. Boetto, I. Radić Rossi, S. Marlier and Z. Brusić (eds), L’épave de Pakostane, Croatie (fin IVe- début Ve s. apr. J.-C.). Résultats d’un projet de recherche Franco-Croate. Archaeonautica 17: 128–132. Joncheray, J.-P. 1975. L’épave E du cap Dramont, sigillée claire D et amphores rescapées du pillage. Cahiers d’archéologie subaquatique IV: 141–146. Keay, S. J. 1984. Late Roman amphorae in the Western Mediterranean, A typology and economic study: the Catalan evidence. British Archaeological Reports International Series 196. Oxford, Tempus Reparatum.

294

African amphora contents: an update

Laubenheimer, F. 1990. Le temps des amphores en Gaule. Paris, Errance. Lequément, R. 1975. Étiquettes de plomb sur des amphores d’Afrique. Mélanges de l’École française de Rome. Antiquité 87: 667–680. Lequément, R 1976. Une épave du Bas-Empire dans la Baie de Pampelonne (presqu’île de Saint Tropez). Revue archéologique de Narbonnaise IX: 177–188. Lequément, R. 1980. Le vin africain à l’époque impériale. Antiquités africaines 16: 185–193. Liou, B. 1975. Informations archéologiques. Gallia 33-2: 572–605. Liou, B. 1988. Le contenu des amphores, typologie et épigraphie : quelques cas aberrants ou embarrassants. In SFECAG, Actes du Congrès d’Orange: 171–178. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Liou, B. and Silvino T. 2005. Inscriptions sur les amphores du parc Saint-Georges à Lyon. In SFECAG, Actes du Congrès de Blois: 721–727. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Long, L. 2002. Carte archéologique. Au large de la Camargue. In Bilan scientifique du Département des Recherches Archéologiques Subaquatiques et SousMarines, 2000: 57–61. Paris, Ministère de la Culture et de la Communication. Long, L. and 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 and Y. Rigoir (eds), Fouilles à Marseille. Les mobiliers (Ier-VIIe s.). Etudes Massaliètes 5: 317–342. Paris, Errance. Manacorda, D. 1977. Le anfore. In Carandini, A. and Panella C. (eds), Ostia IV. Le Terme del Nuotatore. Scavo dell’ambiente XVI e dell’area XXV. Studi Miscellanei 23: 117-285. Rome, De Luca. Marlière, E. and Torres Costa, J. 2007. Transport et stockage des denrées dans l’Afrique romaine : le rôle de l’outre et du tonneau. In A. Mrabet and J. Remesal Rodríguez (eds) 2007. In Africa et in Hispania: Études sur l’huile africaine. Instrumenta 25: 85–106. Barcelone, Universitat de Barcelona. Martin-Kilcher, S. 1994. Die römischen Amphoren aus Augst und Kaiseraugst, 2-3. Die Amphoren für Wein, Fischsauce, südfrüchte (Gruppen 2-24) und Gesamtauswertung. Forschungen in Augst 7/2-3. Augst, Römermuseum Augst. Martin-Kilcher, S. 1999. Karthago 1993. Die Füllung eines frühkaiserzeitlichen Pozzo. In F. Rakob (ed.), Karthago III. Die deutschen Ausgrabungen in Karthago: 403–434. Mainz, Philipp von Zabern. Martin-Kilcher, S. 2011. Formes d’amphores et contenus au Haut-Empire, points de repère et questions. In SFECAG, Actes du Congrès d’Arles: 417–426. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Marty, P. and Capelli, C. 2016. Quels sont ces petits vases qui gisent dans nos tombes ? Le balsamaire Augts 73, un vase spécifiquement funéraire ? Un état de la question dans l’Empire. In SFECAG, Actes du Congrès d’Autun: 203–226. Marseille, Société Française d’Étude de la Céramique Antique en Gaule.

Mattingly, D. J., Stone, D. L., Stirling, L. M., Moore, J. P., Wilson, A. I., Dore, J. N. and Ben Lazreg, N. 2011. Economy. In D. Stone, D. J. Mattingly and N. Ben Lazreg (eds), Leptiminus (Lamta): Report no. 3, The Urban Survey. Journal of Roman Archaeology Supplementary Series 87: 205–272. Portsmouth RI, Journal of Roman Archaeology. Moliner, M., Bien, S. and Bonifay, M. 2007. La céramique de la basilique paléochrétienne de la rue Malaval à Marseille : approche préliminaire. 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 (I): 275–285. Oxford, Archaeopress. Mrabet A. and Remesal Rodríguez, J. (eds) 2007. In Africa et in Hispania: Études sur l’huile africaine. Instrumenta 25, Barcelona Universitat de Barcelona. Murialdo, G. 1988. Necropoli e sepolture tardo-antiche del Finale. Rivista di Studi Liguri LIV: 221-242. Opaiţ, A. 1997-1998. North African and Spanish Amphorae in Scythia Minor. In Il Mar Nero. Annali di archeologia e storia III: 47–95. Opaiţ, A. 2000. Early Roman Amphorae from Leptiminus. Rei Cretariae Romanae Fautorum Acta 36: 439–442. Pallarès, F. 1987. Alcune considerazioni sulle anfore del Battistero di Albenga. Rivista di Studi Liguri LIII: 269– 306. Palol, P. de (ed.) 1987. La basílica des Cap des Port, de Fornells, Menorca. In II Reunió d’Arqueologia Paleocristiana Hispànica (Montserrat, 2-5 novembre 1978). Publicacions Eventuals 31: 353–454. Barcelona, Institut d’Arqueologia i Prehistòria. Panella, C. 1973. Le anfore. 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 Miscellanei 21: 463–633. Rome, De Luca. Panella, C. 1977. Anfore tripolitane a Pompei. In M. Annecchino, A. M. Bisi Ingrassia and A. Carandini (eds), L’intrumentum domesticum di Ercolano e Pompei nella prima età imperiale: 135–149. Rome, L’Erma di Bretschneider. Panella, C. 1982. Le anfore africane della prima, media e tarda età imperiale, tipologia e problemi. In Actes du colloque sur la céramique antique (Carthage, 23-24 juin 1980). CEDAC Dossiers 1: 171–186. Carthage, Centre d’Études et de Documentation Archéologique de la Conservation. Panella, C. 2001. Le anfore di età imperiale nel Mediterraneo occidentale. In P. Lévêque, J. P. Morel and É. Geny (eds), Céramiques hellénistiques et romaines III: 177–275. Paris, Presses Universitaires FrancComtoises. Parker, A. J. 1992. Ancient Shipwrecks of the Mediterranean and the Roman Provinces. British Archaeological Reports International Series 580. Oxford, Tempus Reparatum. Peacock, D. P. S. and Williams, D. F. 1986. Amphorae and the Roman Economy. An Introductory Guide. London - New York, Longman.

295

Michel Bonifay

Pecci, A. 2009. Analisi funzionale della ceramica e alimentazione medievale. Archeologia Medievale XXXVI: 21–42. Pecci, A. and Cau Ontiveros, M. Á. 2010. Análisis de residuos orgánicos en ánforas: el problema de la resina y el aceite. In J. M. Blázquez Martínez and J. Remesal Rodríguez (eds), Estudios sobre el Monte Testaccio (Roma) V. Instrumenta 35: 593–600. Barcelona, Universitat de Barcelona. Pecci, A. and Giorgi, G. 2019. Le analisi dei residui organici e la determinazione del contenuto di alcune anfore del progetto Impianto Elettrico. In D. Bernal‑Casasola and D. Cottica (eds), Scambi e commerci in area vesuviana. I dati delle anfore dai saggi stratigrafici I.E. (Impianto Elettrico) 1980-81 nel Foro di Pompei. Roman and Late Antique Mediterranean Pottery 14: 157–165. Oxford, Archaeopress. Pecci, A., Salvini, L., Cirelli, E. and Augenti, A. 2010. Castor oil at Classe (Ravenna-Italy): residue analysis of some late Roman amphorae coming from the port. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 617–622. Oxford, Archaeopress. Peña, J. T. 1998. The mobilization of state olive oil in roman Africa: the evidence of late 4th-c. ostraca from Carthage. In J. T. Peña, J. J. Rossiter, A. I. Wilson, C. Wells, M. Carroll, J. Freed and D. Godden (eds), 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: 117– 238. Portsmouth RI, Journalof Roman Archaeology. Peña, J. T. 2007. Two groups of tituli picti from Pompeii and environs: Sicilian wine, not flour and hand-picked olives. Journal of Roman Archaeology 20: 233–254. Poveda, P. 2012. Hypothetical reconstruction of the Dramont E Shipwreck. In Günsenin N. (ed.), Between continents: proceedings of the XIIth International symposium on boat and ship archaeology (Istanbul, 2009): 331–336. Istanbul, Ege Yayınları. Radic, I. 1993. Adriatic in Mediterranean communications and trade in Roman times according to underwater finds. In J. Pavúk (ed.), Actes du XIIe Congrès International des Sciences Préhistoriques et Protohistoriques (Bratislava, 1-7 septembre 1991): 337–344. Bratislava, Union Internationale des Sciences Préhistoriques et Protohistoriques. Ramon Torres, J. 1995. Las ánforas fenicio-púnicas del Mediterráneo central y occidental. Instrumenta 2, Barcelona, Universitat de Barcelona. Remolà i Vallverdù, J. A. 2000. Las ánforas tardo-antiguas en Tarraco (Hispania tarraconensis). Siglos IV-VII d. C. Instrumenta 7. Barcelona, Universitat de Barcelona. Revilla Calvo, V. 2007. Les amphores africaines du IIe et IIIe siècles du Monte Testaccio (Rome). In A. Mrabet and J. Remesal Rodríguez (eds) 2007. In Africa et in Hispania: Études sur l’huile africaine. Instrumenta 25: 269–298. Barcelona, Universitat de Barcelona.

Revilla Calvo, V. 2010. Las ánforas tunecinas y tripolitanas de mediados del siglo III d.C. (campañas 1995-1997). In J. M. Blázquez Martínez and J. Remesal Rodríguez (eds), Estudios sobre el Monte Testaccio (Roma) V. Instrumenta 35: 317–343. Barcelona, Universitat de Barcelona. Reynolds, P. 2010. Hispania and the Roman Mediterranean, AD 100-700. Ceramics and Trade. London, Duckworth. Romanus, K., Baeten, J., Poblome, J., Accardo, S., Degryse, P., Jacobs, P., De Vos, D. and Waelkens, M. 2009. Wine and olive oil permeation in pitched and non-pitched ceramics: relation with results from archaeological amphorae from Sagalassos, Turkey. Journal of Archaeological Science 36: 900–909. Saguí, 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 in onore di John W. Hayes (Rome 1995): 305–330. Florence, All’Insegna del Giglio. Santamaria, C. 1995. L’épave Dramont E à Saint-Raphaël (Ve s. ap. J.-C.). Archaeonautica 13. Sciallano, M. and Sibella, P. 1991. Amphores : comment les identifier ? Aix-en-Provence, Edisud. Scotti, C. 1984. Anfore. In M. Bonghi Jovino (ed.), Ricerche a Pompei. L’insula 5 della Regio VI dalle origini al 79 d. C. I (campagna di scavo 1976-1979): 270–316. Rome, L’Erma di Bretschneider. Silvino, T., Maudet, C., Duffraisse, M. and Hänni, C. 2007. Lyon, Saint-Georges : identification et origine de résidus organiques contenus dans des céramiques antiques. 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 (I): 305–311. Oxford, Archaeopress. Slim, L., Bonifay, M., Piton, J. and Sternberg, M. 2007. An example of fish salteries in Africa Proconsularis: the officinae of Neapolis (Nabeul, Tunisia). In L. Lagóstena, D. Bernal‑Casasola, and A. Arévalo (eds), Cetariae 2005. Salsas y Salazones de Pescado en Occidente durante la Antigüedad. Actes du congrès international (Cadix, 7-9 novembre 2005). British Archaeological Reports International Series 1686: 21–44. Oxford, J. and E. Hedges Ltd. - Universidad de Cádiz. Stefani, G. 2003. Menander. La casa del Menandro di Pompei. Catalogo della mostra (Antiquarium di Boscoreale, 2003). Milan, Electa. Torres Costa, J. 2007. Ex Hor(Reis) Had(rumetinis). A propos d’un titulus pictus mentionnant les entrepôts d’Hadrumetum au IIIe s. ap. J.-C. In . A. Mrabet and J. Remesal Rodríguez (eds) 2007. In Africa et in Hispania: Études sur l’huile africaine. Instrumenta 25: 99-313. Barcelona, Universitat de Barcelona. Villa, L. 1998. Alcuni aspetti della circolazione di prodotti di importazione in Friuli tra VI e VII secolo. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del Convegno in onore di John W. Hayes (Rome 1995): 275–288. Florence, All’Insegna del Giglio.

296

African amphora contents: an update

Wilson, R. J. A. 1996. Rural life in Roman Sicily: excavations at Castagna and Campanaio. In R. J. A.Wilson (ed.), From River Trent to Raqqa. Nottingham University archaeological fieldwork in Britain, Europe and Middle East, 1991-1995: 24–41. Nottingham, University of Nottingham.

Woodworth, M., Bernal‑Casasola, D., Bonifay, M., Garnier, N., Keay, S., Pecci, A., Poblome, J., Pollard, M., Richez, F. and Wilson, A. 2015. The content of African Keay 25 / Africana 3 amphorae: initial results of the CORONAM project. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics, Chromatography and DNA Analysis in Archaeology: 41–57. Esposende, Município de Esposende. Zevi, F. and Tchernia, A. 1969. Amphores de Byzacène au Bas-Empire. Antiquités africaines 3: 173–214.

297

Amphorae from the Byzacene coast: what did they contain? Jihen Nacef†

(1973-2018) Formerly Lecturer at the Institut Supérieur des Études Appliquées en Humanités de Mahdia (University of Monastir, Tunisia)

Abstract: This paper discusses the content of the African amphorae produced in the Sahel region of Tunisia from the Republican period to the end of the Byzantine period. After a first phase of production of amphorae of Punic tradition, both for wine and fish sauces contents, the main phase of production from the 2nd to the beginning of the 5th century is characterized by very standardised amphora types devoted to specific contents, as for example oil for type Africana I, and salsamenta and wine for type Africana II. After a phase of disruption of the production during the 5th century, the reorganisation of the workshop gives birth to new types, the content of which remains largely unknown. The last phase of production during the 7th century is dominated by globular amphorae of Byzantine tradition, imitating wine containers from the Eastern Mediterranean. Key words: Byzacena; amphorae; workshops; typology; contents.

1. Introduction

The production of the first city, Hadrumetum (Zevi and Tchernia 1969: 181; Bonifay 2004: 13) has been recognized through stamps, but to this day there are no traces of potter’s workshops. However, Leptiminus and Sullecthum have produced the remains of pottery kilns and stamps (Peacock et al. 1989; Bonifay 2004: 31–35; Leptiminus 3; Nacef 2015a). Others workshops are known in Thapsus.

workshops remains uncertain. However, there is evidence that Punic tombs contain Greco‑Italic amphorae and an amphora of these two Punic types, strongly suggestive of wine (Bisi 1989: 594–596, see examples of funeral furniture in the region around Salakta, Ben Younes 1986 and 1988). If not maybe they contained oil or salsamenta. The coastal location of the workshop at Gaala might suggest that salsamenta was the content for these amphora, and the same could be said for El Maklouba, located a few kilometers from the sea (they may have contained Gummi for example). However, in the absence of archaeometric analyzes, the three classic contents (wine, oil and salsamenta) all remain possible. Indeed, the cultivation of vines, even if not on a large scale, leaves open the possibility of a wine content for this geographical area. Additional suggestions come from the Punic necropolis of Smirat (in Byzacena) which is furnished with bowls and jugs of patterned pottery, decorated with the impression of grape seeds grouped in twos or threes (Gobert and Cintas 1941: 99).

2. Pre-Roman tradition

3. Punic to Roman transition

Let us start with the amphorae of the late Punic period. Caesar, according to the account of his war in Africa: Chap. [67] (2) sought food and stopped at Aggar, probably at the site of El Maklouba, where he ‘returns to camp with lots of barley, oil, wine, figs, little wheat’. So we can deduce from the food brought from the farms that the two contents from Punic type amphorae were oil and wine. These amphorae were connected to the standard Van der Werff 3 and Van der Werff 2 types (Figure 2.1‑2) highlighted on the site of Uzita in Byzacena which were produced in the workshop of El Maklouba (Peacock et al. 1989: 195–196; Nacef 2015a), as for other sites of Byzacena between the 3rd and the 1st centuries BC. The data for these amphora is sparse compared to the amphorae of the Roman period. These amphorae, based on the archaeological surveys of Jerba’s workshops, carried wine (Fentress 2001: 263), while the contents of those coming from the workshops of the Byzacena coastal

Before looking at the Roman amphorae, I will mention a particular amphora production placed on the outskirts of the city of Thapsus, which are of Punic tradition, that continue into the Roman period (Nacef 2015b). They are small amphorae, derived from the Punic type Uzita 2 (Figure 3.3). Their diffusion is poorly understood and was probably small-scale and their content not previously discussed. The two classic goods, wine and salsamenta are possible candidates, if we accept the idea that they are derived from the Punic model. Since we know that vine growing was possible in the region, the production of wine is certainly possible. Another argument supports this hypothesis: the Thapsus amphorae are a maximum of 70/80cm high. Compare this with the small African/ African I (oil amphora) which can reach 1m in height, have a wider rim diameter and an average capacity of 40 litres. Moreover, it is not only Thapsus that had a production

In this paper, I will discuss examples of coastal Byzacene amphorae illustrating two sets of productions: classic African and late African. The area of Byzacene is known as the Sahel in Tunisia. This region has abundant remains of workshops and production facilities of ancient ceramics, in some cases only dealing with amphorae production. This craft is located in four major cities-ports that are respectively north to south: Hadrumetum, Leptiminus, Thapsus and Sullectum. The latter has a navicular office at the Piazzale delle Corporazioni at Ostia (Figure 1).

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 299–303

Jihen Nacef†

Carthage

N

Nabeul/Neapolis

Sousse/Hadrumetum Lamta/Leptiminus Moknine Thapsus Salakta/Sullecthum

0

10 cm

El Jem/Thysdrus

Bonifay, 2004

3

0

50

100Km 0

5 cm

Figure 1. Location map of the sites. 4

5

Figure 3. 3: No.4: Early Roman derivative from type Van der Werff 2 (Thapsus workshop). Nos5-6: types Sullecthum 1 and 8 (Sullecthum workshops: Nacef 2015a). 4. Roman and Late Roman amphorae The case of ancient workshops of Sullecthum is remarkable insofar as they experienced a boom at the end of the 2nd to 3rd centuries. The containers are of different dimensions, as already mentioned in the first study by Zevi and Tchernia (Zevi and Tchernia 1969). The two large and small models from these workshops suggest two different contents according to these authors. Indeed, the Africana type I, devoted to the transport of olive oil, is the main type at Sullecthum (Figure 4.6), associated with the production of the large Africanas, especially Africana IIA and IID (Figure 4.7‑8), for other contents.

Figure 2. Nos1-2: Types Van der Werff 3 and Van der Werff 2 (El Jem museum: Nacef 2001). of amphorae of Punic tradition: at Leptiminus and Sullecthum we see similar forms. Both types Leptiminus 2/ Sullecthum 1 and Leptiminus 1/Sullecthum 8; Figure 3.4‑5) (Nacef 2015a: 36 and 44–46) do not seem to be oil amphorae (Bonifay et al. 2002‑2003: 175). They are distinguished from classical Africana amphorae by their form (absence of neck on some, handles on the body etc.) I don’t think this is accidental, but connected to a use for wine.

With the naked eye, the Africana  I examples in the storerooms at the museum of Salakta do not bear traces of pitch, even if we know that they were pitched, like most of the oil amphorae (Garnier, Silvino and Bernal‑Casasola 2011). A question then arises: at which time/stage were the amphorae pitched knowing that they could have come from workshops installed in the vicinity of the port and their warehouses or from workshops located in the hinterlands of a city (e.g. Ghar Draj is 6/6.5km from Thapsus and El Maklouba about 10km from Salakta). More precisely, is the amphora pitched at the same time that it is filled? Sometimes, these oil amphorae were reused to transport fish content. The case of 204 Africana I amphorae from the Grado ship wreck, originating from Salakta’s workshops is enlightening. They were used for the transport of sardines (Auriemma 2000).

300

Amphorae from the Byzacene coast: what did they contain?

7 5 cm

0

Musée Salakta

9

10

8

0

12 6

0

10 cm

5 cm

11

Figure 4. No.6: type Africana I (Sullecthum museum: Nacef 2015a). Nos7-8: types Africana 2A and 2D (Sullecthum workshops: Nacef 2015a). The Sullecthum productions are also found in urban contexts in Lyon. They help demonstrate evidence of residue content and allow us to track changes further downstream in the contents of this type of amphora, and thus show ‘l’importance des arrivages de Salakta jusqu’à la fin du IIIe s.’ (Lemaitre et al, 2011: 214). Indeed, the type Sullecthum 6 (Figure 5.9) is evidenced by a single pitched example, as well as Africana II. As from these workshops have grape seeds inside them, relating to either a fish content or wine (Lemaitre et al. 2011: 206). In the second half of the 3rd century, there is clear evidence of the same type, from the same source, whereas the African type I from Salakta, intended for the transport of oil, is only modestly attested and disappeared in the middle of the 4th century (Lemaitre et al. 2011: 210). So at this point the Africana IID and then the Africana III containing fish derivatives and wine, dominate this market. So far the amphora stamps of ancient Sullecthum bear no relationship to the contents. On the other hand, the example of the LEP stamp decorated with a grape vine on African III from Leptiminus workshops is interpreted as a production linked to Liber Pater and reveals a wine content (Leptiminus 3: 218). It is possible that this decoration is associated with the symbol of a corporation or association that deals with wine trade. Note that in Byzacena, the Salakta example has stamps with a palm in the centre, composed either of a single line (Sullecthum 2 Uzita Pl. 52, 10; Figure 5.10) or of two lines on later examples such as the African I or II, separated by a palm

Figure 5. No.9: types Sullecthum 6. No.10: Sullecthum 2/Uzita Pl. 52, 10. No.11: Sullecthum 11/Keay 25. No.12: Sullecthum 12/Dressel 30 (Sullecthum workshops: Nacef 2015a). (Nacef 2015a: table no.3, stamps nos12, 26, 51). There is scarcely any mention of the cultivation of palm trees or the trade of dates. Instead, we could perhaps consider the symbol of an association or a brotherhood which manages and deals with commerce or simply a decoration or mark of the potter. The classic Africana amphorae are the only African production with stamps. Rare cases of early examples of Late Roman amphorae bear stamps, such as on Africana III/Keay 25 (Figure 5.11). An amphora of this type with a stamp inserted in a text box on the neck is displayed at the Mahdia Museum. While the late Africana amphorae generally bear no inscriptions or symbols stamped or painted. Let us look now at the concept of ‘intelligent’ packaging based both on form/typology and on writing. Some forms imitate those known for their wine content, such as the Dressel 30 (Figure 5.12), which were produced on a small scale (these were identified only at the Catacombs workshop/Sullecthum  12 type). It is very probable that wine was produced in small quantities and it was circulated in atypical packaging for this region. The second set corresponds to Late Roman amphorae, including a first generation amphora Africana III, type Bonifay/Keay 25, produced at Salakta and Leptiminus for wine, as evidenced by the analyzes of amphorae from the wreck Heliopolis 1 whose cargo is linked to the Sullecthum workshops (Ben Lazreg et al. 1995).

301

Jihen Nacef†

13

16

17

14

19 18

15

0

5 cm

Figure 6. Nos13-15: types Keay 62A, 61C, Bonifay 47 (Henchir Chekaf workshop: Nacef 2015a). No.16: type spatheion 3D (Moknine 2 workshop: Nacef 2017). No.17: type Henchir Chekaf IV (Henchir Chekaf workshop: Nacef 2015a). No.18: type Castrum Perti (Moknine 2 workshop: Nacef 2017). No.19: type Keay 8A (Dkhila workshop: Nacef† and Capelli 2019). Late Antiquity experienced a recovery after a vacuum of ceramic production or artisanal activity which started around the 5th century at coastal Byzacena, including the regions of Salakta and Ksour Essef, placed north between the territory of Thapsus and Leptiminus (Nacef 2015a: 101). Survey between the territory of Salakta and Thapsus did not reveal traces of workshops or production of the 5th century. The same phenomenon is observed at Leptiminus and its territory (see Leptiminus 3). The transformations affecting the Mediterranean world and essentially Rome affected the cession of ceramic production in the Tunisian Sahel. These late amphorae come from these workshops and are particularly dense in the region of Ksour Essef, Bir El Hammam and Moknine. They are Ech Chekaf I/Keay 62, Ech Chekaf II/Keay 61C and Ech Chekaf III/Bonifay type 47 (Figure 6, 13‑15) (Nacef 2007). These workshops are linked to a production system different from the previous period and are no longer installed, for the most part, near the ports. They were partly identified by the Tunisian‑English team, (Peacock, Bejaoui and Belazreg 1989), including Ras Aïd, Chtiouine, Henchir Ech Chekaf, not forgetting the new set of kilns identified in Chrahel, located between the territory of the city Sullecthum and the city of Leptiminus. What we can say from this evidence is that there was a recovery in production after the fifth century break, and these typically large containers date to the sixth and the seventh centuries and represent the second ‘boom’ (Bonifay and Raynaud  2007). Again, the situation is increasingly controversial with regard to content, but oil remains the most likely commodity. From the second half to the end of the 6th century, the repertoire changes with two smaller types: Bonifay’s spatheia 3D type (Figure 6.16), a miniaturised amphora with a small capacity (height

c. 40cm and a diameter of c. 8 to 9cm) (Bonifay 2004: 127), and the globular amphorae of Byzantine tradition. Perhaps this change relates to a variation of contents in the Byzantine era? Or a more precious content, probably not be olive oil, in a very small container which was easy to carry? Note also that most of the known examples are pitched. 5. Byzantine amphorae I conclude this discussion with an examination of the globular amphorae of Byzantine tradition. Two production sites for this type have been identified at Henchir Ech Chekaf (Nacef 2007: Type Ech Chekaf IV/ imitation of LRA1) in the territory of Salakta and at Moknine which was probably in the territory of Leptiminus and seem to signal a new generation of Byzantine tradition which is similar to Arabic productions in their first phase (Nacef 2017). The continuity of these amphorae is remarkable, which is most probably observed up to the Aghlabid period in the 10th century (see Gragueb et al. 2011: fig. 4). Both types of globular amphora (Ech Chekaf IV and globular amphora Moknine, Figure 6.17‑18) are imitations of oriental forms (LRA 1 and 2). These productions, as well as the latest cylindrical type Keay 8A (Figure 6.19) are distributed across Southern Gaul until the 8th century (Bonifay and Raynaud  2007:  102), in Rome and at continental sites in Europe such as San’Antonino, which was under Byzantine authority. We see thus a continuity in trade relations with the Arabs of Ifriqiya despite the diplomatic rupture.

302

Amphorae from the Byzacene coast: what did they contain?

Bibliography Auriemma, R. 2000. Le anfore del relitto di Grado e il loro contenuto. Mélanges de l’École française de Rome. Antiquité 112-1: 27–51. Ben Lazreg, N., Bonifay, M., Drine, A. and Trousset, P. 1995. Production et commercialisation des salsamenta de l’Afrique ancienne. In L’Afrique du Nord antique et médiévale. Production et exportations africaines. Actualités archéologiques. Actes du VIe colloque d’Histoire et d’Archéologie de l’Afrique (Pau, 25-29 octobre 1993): 103– 142. Paris, CTHS. Ben Younes, H. 1986. La nécropole punique d’El Hkayma Mai 1984. Reppal II: 31–172. Ben Younes, H. 1988. La nécropole punique d’El Hkayma seconde campagne Septembre 1985. Reppal IV: 49–159. Bisi, A. M. 1989. Associazioni di anfore puniche Mañá C1=Uzita 3 e di «greco-italiche», in contesti punici della Sicilia e del Nordafrica. In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22-24 mai 1986). Collection de l’École française de Rome 114: 594–596. Rome, École française de Rome. Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports International Series 1301. Oxford, Archaeopress. Bonifay, M. and Raynaud, C. 2007. Les échanges et la consommation. In M. Heijmans and J. Guyon (eds), Antiquité tardive, haut Moyen Age et premiers temps chrétiens en Gaule méridionale, 2. Gallia 64: 93–161. Bonifay, M., Capelli, Cl., Martin, T., Picon, M. and Vallauri, L. 2002-2003. Le littoral de la Tunisie, étude géoarchéologique et historique (1987-1997). La céramique. Antiquités africaines 38-39: 125–202. Fentress, E. 2001: Villas, Wine and Kilns: the landscape of Jerba in the late Hellenistic period. Journal of Roman Archaeology 14: 249–268. Garnier, N., Silvino, T. and Bernal-Casasola, D. 2011. L’identification du contenu des amphores : huile, conserves de poissons et poissage. In SFECAG, Actes du Congrès d’Arles: 397–416. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Gobert, E. G. and Cintas, P. 1941. Smirat. Revue tunisienne: 83–121. Gragueb, S., Treglia, J.-C., Capelli, C. and Waksman, Y. 2011. Jarres et amphores de Sabra Al Mansuriya (Kairouan, Tunisie). In P. Cressier and E. Fentress (eds), La céramique maghrébine du Haut moyen âge (VIIIeXe siècle), état des recherches, problèmes et perspectives. Collection de l’École française de Rome 446: 197–220. Rome, École française de Rome.

Lemaître, S., Duperron, G., Silvino, T., Bonnet, C., Bonifay, M. and Capelli, C. 2011. Les amphores africaines à Lyon du règne d’Auguste au VIIe siècle : réflexions à propos de la circulation des marchandises sur l’axe rhodanien. In SFECAG, Actes du Congrès d’Arles: 203–222. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Leptiminus 3 = Stone, D. L., Mattingly, D. J. and Ben Lazreg, N. (dirs) 2011. Leptiminus (Lamta). Report no. 3. The field survey. Journal of Roman Archaeology Supplementary Series 87. Portsmouth RI, Journal of Roman Archaeology. Nacef, J. 2001. Étude de la collection d’amphores de la réserve du Musée d’El Jem. Unpublished thesis (Diplôme d’Études Approfondies en Patrimoine et Archéologie, spécialité Archéologie), Université de Tunis I. Nacef, J. 2007. Nouvelles données sur l’atelier de Henchir Ech Chekaf (Ksour Essef-Tunisie). 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 (II): 581–591. Oxford, Archaeopress. Nacef, J. 2014. Nouveaux témoignages sur la production de la céramique antique du sahel tunisien. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeological Reports International Series 2616 (I): 103–111. Oxford, Archaeopress. Nacef, J. 2015a. La production de la céramique antique dans la région de Salakta et Ksour Essef (Tunisie). Roman and Late Antique Mediterranean Pottery 8. Oxford, Archaeopress. Nacef, J. 2015b. Un atelier de potier à la périphérie de la ville Thapsus aux premiers siècles de l’Empire. Antiquités africaines 51: 17–44. Nacef, J. 2017. Moknine 2. Nouvelles données sur un atelier de potier d’époque tardive en Byzacène. In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. Études alexandrines 43 (I): 491–515. Alexandria, Centre d’Études Alexandrines. Nacef, J.† and Capelli, C. 2019. L’atelier céramique de Dkhila (Tunisie). Antiquités africaines 54: 151–156. Peacock, D. P. S., Bejaoui, F. and Belazreg, N. 1989. Roman amphora production in the Sahel region of Tunisia. In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22-24 mai 1986). Collection de l’École française de Rome 114: 179–222. Rome, École française de Rome. Zevi, F. and Tchernia, A. 1969. Amphores de Byzacène au Bas-Empire. Antiquités africaines 3: 173–214.

303

East

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire Paul Reynolds

Institució Catalana de Recerca i Estudis Avançats (ICREA) ICREA Research Professor and Member of the Equip de Recerca Arqueològica i Arqueomètrica de la Universitat de Barcelona (ERAAUB: Consolidated Group 2017 SGR 1043), Secció de Prehistoria i Arqueologia, Dept. de Història i Arqueologia, Universitat de Barcelona

Abstract: This paper offers an overview of the supply of olive oil to the provinces of the Roman and early Byzantine East (1st to 7th centuries) and the possible western and eastern amphorae and other containers (e.g. dolia, other large jars, glass vessels, barrels) that may have carried this prime commodity. Whereas mainland Greece and the Aegean Islands were prominent producers and suppliers in the early and late Empire for local consumption and export in specific amphorae (Dressel 24 and related forms, Dressel 25, Agora M 235, LRA 2 and successors), other regions produced oil for local needs (Egypt, Lebanon, NW Syria), whereas other regions (the Lower Danube, Crimea) were unable to grow olives for climatic reasons. The role and mechanisms of the State, the Army and the annona in the supply of oil here and in other regions (e.g. Egypt, Syria) are discussed. In the course of this typological revision, other amphorae and possible contents (e.g. wine, fish products, dried fruit) are also discussed and illustrated. New data on amphora contents based on organic residues analyses are noted. Key words: Olive oil; Wine; Fish products; Dried fruit; Amphorae; Jars; Barrels; Carboys; Annona; Peloponnese; Aegean; Organic residues analyses.

1. Introduction1 Whereas the long distance distribution of olive oil in the Roman West can be associated with specific amphora types (e.g. Dressel 20, Dr 6B, the Tripolitanian series and a range of Tunisian forms: Bonifay 2004, 2016), the identification of amphora types in the Roman East that may have served a similar purpose is more problematical. Apart from the more well-known amphorae of the East (e.g. in Robinson 1959; Riley 1981; Hayes 1983; Keay 1984; Peacock and Williams 1986; Panella 1986; Hayes 1992; Marangou-Lerat 1995; Pieri 2005a; Reynolds 2005; Bertoldi 2012), the number of newly discovered amphora types across the region, not to mention the scores of new forms I have personally documented during my work in Lebanon, Syria, Albania and Greece, demonstrate the actual typological complexity of amphora production, as well as our lack of knowledge in unchartered regions.2  An earlier version of this paper (‘The oil supply in the Roman East: western imports versus local modes of production and redistribution’) was originally presented in December 2009 at a conference organised by Séverine Lemaitre at the University of Poitiers (L’économie de l’huile en Méditerranée grecque (de l’époque classique à la période romaine tardive), finally updated and prepared for publication in 2015, but sadly never reaching the happy conclusion of its publication. I would like here to express my thanks to Séverine for suggesting to me this topic, inviting me to that excellent conference and for her input in the evolution of this paper. I would also like to thank Andrei Opaiţ for his most useful comments on this paper, which, as he pointed out, goes beyond the principal topic of the supply of oil, including possible eastern oil amphorae, in the East and includes eastern amphorae that carried (or may have carried) other goods (wine, dried fruit, salsamenta). I have taken this liberty to provide discussion and illustration of other eastern forms in circulation, some of which are in some cases problematical with regard to their contents, and therefore may be of interest for readers of this volume. 2  Well-known to western archaeologists, that is. What was the wellknown repertoire of the Black Sea for Russian, Bulgarian or Rumanian 1

As in the West (e.g. Dressel 2-4, Keay 19, Keay 25), whereas some forms were specific to certain cities and their territories (the Beirut amphora, the Agora M 334 of Akko Ptolemais, Knidian amphorae, Rhodian amphorae, Sinopean ‘carrot’ amphorae), other eastern types provided models that were widely imitated, only to be distinguished by their specific fabrics, whether across the Aegean and Asia Minor (e.g. the Agora M 273 type, LRA 3 and its imperial predecessors, the pinched handle amphora Agora G 199, LRA 4 and LRA 5, from southern Phoenicia to northern Egypt). However, with the exception of Crete, southern Phoenicia, Egypt, and the Black Sea to some extent, and some isolated cases, the vast majority of eastern amphorae encountered, including some of the principal exported types, are still unsourced, with their workshops-production sites being still to be found, even though broad regional sources for them (sometimes quite disparate) have been suggested.3 If this were not problematical enough, the identification of the goods that were carried in eastern amphorae is still primarily guess-work based on their theoretical regional archaeologists (e.g. Zeest 1960), has now reached a wider audience through the ongoing work of, for example, Andrei Opaiţ (2004a and his many articles), Sergei Vnukov (e.g. 2004, with bibliography), Andrei Sazanov (e.g. 1997, 2017), with their bibliographies. Robinson’s 1959 publication of amphorae from the Athenian Agora is an example of the complex range of forms in Athens and even then, there are many other unpublished forms in the stores (Reynolds in preparation b). 3  Here the new Web Atlas of Ceramic kilns in Greece, the initiative of Eleni Hasaki (University of Arizona) is a welcome tool for research. Nikoula Kougia (PhD candidate, University of Patras) has already made a major contribution through her ongoing systematic documentation of over 50 kilns and ceramics of the early Roman period in Patras. Nefeli Eleni Pirée Iliou (PhD candidate, University of Oxford) has also been busy documenting rural sites-villas, presses and kilns across Epirus.

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 307–354

Paul Reynolds

Figure 1a. Location of principal sites mentioned in the text. sources and assumptions as to function according to typological traits. The key, clearly, is the analysis of the traces of organic residues, the subject of this conference, as well as observations at oil and wine as well as other industrial sites. Advances with respect to the contents carried by some of the major 3rd and 5th century Black Sea forms can now be reported (see below). Though, as predicted, analyses prove that the Beirut amphora (one of 5 forms produced in the city in the early Imperial period) carried wine (Woodworth and Reynolds, this volume), and we know that the small Beirut ‘carrot’ amphora/ Schöne-Mau XV/Peacock and Williams Class 12 carried dried fruit (Figure 4a-c), the purpose of the other forms produced in Beirut and elsewhere in, I suggest, northern Lebanon needs urgent attention (see below, Figure 16). None of the new forms discovered in Epirus for which we now have a partial typology have been analysed for their contents (Reynolds 2020: Appendix A, 1st to 3rd century forms, ERAM 1-13).

Figure 1b. Location of principal Levantine sites.

The lack of archaeological evidence for fish processing factories and salsamenta in the Eastern Empire outside the Black Sea, despite textual references to them in Asia Minor and the Aegean islands, Lycia, Cilicia, Cyprus (Curtis 1991: 106–109, 112–147, fig. 8: see especially p. 130, note 93), is another surprising phenomenon noted by Curtis that needs explanation and investigation, especially because some of the sites are those we will discuss with respect to oil and wine production and their possible amphora containers (e.g. Cyme — Lydia — referred to by the 3rd century AD writer Julius Africanus, Kεστοι 7.19;

308

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Figure 2. Regions producing olive oil in 1936 (from Bull 1936: fig. 5). Clazomenae: praised by Pliny, HN 31.94; Megisti, on the island of Castellorizo — Lycia — where fish salting vats have been found; Cyprus: production noted by Pliny HN 37.66) or for which we have possible amphora candidates but no archaeological evidence or contents analysis (see below, Agora M 54 and G 198, San Lorenzo V, Ostia I 568, the ‘Chalk’ amphora, Corinthian ‘fruit’ amphorae). Others in Curtis’ list are more likely locations where salsamenta were traded (Miletus; Antioch; Phaselis — Cilicia — whose merchants were involved in the Black Sea fish trade in the 4th century BC; Anazarbus: here referring to the tariff inscription). Others are known to have been involved in tunny fishing in pre-Roman times from installations (Halicarnassus-Bodrum), tunny watches (Lesbos) and the frequent depictions of tunny on their coinage (Methymna — Lesbos — Chios, Samos, Clazomenae, Teos, Tarsus) (Curtis 1991: 129–131). As for olive oil from the Western provinces, this was rarely exported to the East. Bearing in mind the problems and limitations just outlined, this paper offers an overview of the mechanisms of the oil supply in the eastern Roman provinces in an attempt to identify how and in what containers oil was supplied to both military and civilian markets. It appears that state- and military-organised supply of amphora-borne oil to both Rome and the army in the West does not find an exact parallel in the East, except in the case of the army on the Danube and perhaps also the Roman army in Syria, if it was oil that was carried in the locally produced pale Syrian amphora class. 2. Oil in the Roman West For those studying the economy of the Roman West, oil production has come to be associated with the annona and its two primary markets: the state-organised supply of oil to both the city of Rome and to the Roman armies

of Britannia and the Rhine-Upper Danube frontier. The primary regions involved in this oil production were Baetica, Africa-Tunisia and Tripolitania (Reynolds 2010a: Chapter 1; Keay 1984; Mattingly 1988; Brun 2004: Chapter 7). Some points need to be underlined, however. As finds of Dressel 20s attest, annona oil surpluses were sold in the open market, north-east Tarraconensis and Narbonensis being the principal secondary markets, probably in the hands of the same shippers and merchants that supplied Rome. Here, Baetican oil, whether surplus annona oil or not, as well as grain, was also bought on the open market. Other cities did the same (Reynolds 2010a: 29–30, note 33). Many regions did not have access to annona oil and supplied their own needs, if the climate was favourable for the production of olives (see Figure 2, for a map of oil producing regions in 1936: Central-northern Spain, Galicia and Cantabria, as well as much of France, and of course Germany and Britain were excluded). Otherwise, they used non-olive oil alternatives such as nut oils or animal fat-lard (Reynolds 2010a: 25–26, notes 78 and 85). From the early Severan period onwards, Baetican oil exports for the annona dropped in quantity, due to several factors, Tunisian competition and a drop in demand on the Rhine frontier being two of the principal likely causes (Reynolds 2010a: 24–27). The shift to the provision of lard in the soldiers’ rations under Severus is surely also relevant (Carreras Monfort 2002: 87). From the 4th century it was primarily Tunisian, not Baetican, oil which was supplied to Rome. From the late 4th century, outside Hispania, Baetican oil (Guadalquivir) and BaeticanLusitanian fish products (primarily Cadiz, Malaga, Sado) targeted major Gallic centres such as Arles, Lyon and Trier, as well as Rome (Reynolds 2010a: Tables 17a and 18; Rizzo 2014: 393–440). The northern troops used locally produced nut oil packaged in imitation Dressel 20s

309

Paul Reynolds

(Reynolds 2010a: 25 and 33–39). A significant point, and in contrary to what was generally proposed (Keay 1984), the majority of Tunisian amphorae of the 2nd to 5th centuries found on sites around the Mediterranean were not for oil. They carried fish, and, in the late 3rd-4th century, wine.4 3. Oil in the Roman East Western exports of oil to the Roman East during the 1st to 3rd centuries, whether from Baetica, Africa or Tripolitania, were rare indeed. Only certain major port hubs were supplied, and in small quantities, with Baetican Dressel 20s and even fewer Dressel 23s, Ephesus and Alexandria being two of these, the latter perhaps only as a consequence of ease of supply through redistribution from Rome on returning Egyptian grain ships.5 The legionary fortress at Zeugma, on the Euphrates, is a case where, in the mid-3rd century, the combination of Dressel 20 oil, and Campanian and Gallic wine amphorae, though all in very small quantities, may reflect a more state-directed supply to the army (Reynolds 2013; Reynolds 2010a: 28). In marked contrast to the substantial supply of Lusitanian, Baetican and Tunisian fish products to Beirut in the 3rd to early 5th centuries, very little Campanian wine and no Baetican oil and Gallic wine were supplied to Beirut in the 3rd century, so the focus on Zeugma, hence also Antioch, seems significant.6 However, as in the case of Alexandria, the supply may have resulted from an ad hoc redistribution from Rome, in this case accompanied by African Red Slip Ware? (Bonifay and Tchernia 2012).7 Here, it would be interesting to see if the legionary camp at Raphanea, near Emesa, was similarly supplied. If the Roman eastern provinces did not import western olive oil, how was the supply of olive oil or alternatives managed? Is there any evidence for the supply of oil to civilian and, in parallel with the western model,  Bonifay and Pieri 1995; Bonifay 2004: 463–475; Reynolds 2010a: note 331 and Index, Oil, Tunisian. Residue analyses now indicate that Keay 25 was used not only for fish but for wine during the late 3rd-4th centuries, perhaps, like Mauretanian Keay 1, in response to Aurelian’s demand for cheap wine for Rome: Bonifay 2007b; Woodworth et al. 2015. 5  Reynolds 2010a: 70–71, note 88, note 272; Will 1983; for rare finds of Dressel 20, a rim and 22 sherds, at the imperial quarries of Mons Claudianus, see Tomber 2006: 159. Elizabeth Will’s manuscript on early Imperial stamped amphorae in the Athenian Agora remains sadly unpublished. Dressel 20s (and Gallic amphorae of the second half of the 1st century) are attested in Corinth, but seem to be scarce, in quantities that suggest their indirect supply (Slane 1989). For Dressel 20s in Ephesus, where early Imperial Baetican fish amphorae are probably as common as they are in Beirut, see Bezeczky 2013: 139–142. For rare finds of Dressel 20 in Dacia, as well as an increase in Spanish oil imports (Dressel 23) in the 4th and 5th centuries (Tomis, Ulmetum, Dinogetia, Topraichioi, Murighiol and Ulmetum, here dated to AD 470-480), see Opaiţ forthcoming. 6  There were 3 handle fragments of Campanian amphorae and two walls (out of a total of 45 Campanian amphora RBHS) that could be dated to between the Antonine period and the early 3rd century. Most are Augustan to mid-1st century examples (in the same contexts as Baetican fish amphorae and Haltern 70: for some of these, see Reynolds 2000). 7  In Crete, the varied supply of Imperial western amphorae at Knossos, including Gallic, Campanian and Iberian (fish) amphorae, together with the supply of Italian sigillata, may be noted (Hayes 1983). 4

military markets? Where did the major cities, including Constantinople and Antioch, obtain their oil? Here, it is important to bear in mind that the some regions, including the lower Danube and Crimea, may not have had the necessary climate to sustain the production of olive trees (Figure 2), a possibility that would explain the regular importation of olive oil from the Peloponnese in the 5th and 6th centuries (see below). The archaeological evidence for the local production and distribution of oil, olive or other classes, in the Roman East is more elusive than at first might seem. Archaeological surveys provide evidence for regions engaged in oil production, primarily through the presence of presses: in Egypt, Palestine, Syria, southern Anatolia, and the Peloponnese, for example (Brun 2004: passim). But, with the exception of Dressel 24 and LRA 2 (both types previously classified as wine amphorae), the identification of associated amphorae or other containers is not at all straightforward. Apart from modern IsraelPalestine, few press sites have been found associated with the production of amphorae. Are we looking for the wrong containers? In some oil producing regions, were amphorae used at all? The majority of amphorae in the Roman East have been classified as either wine or, more rarely, fish amphorae, such as the massive wide-necked amphorae of the Crimean Straits of Kerch for example (see below, Figure 13d). Any amphora with a tall narrow neck and Koan-style doublerod handles, or one derived from it, ought to have carried wine: e.g. the good Koan imitations of eastern Cilicia and Ras al Basit on the north Syrian coast (Reynolds 2005: fig. 12; Mills and Reynolds 2014: fig. 3.6); the large ‘Amrit’ amphora of the 2nd to 4th centuries AD (Reynolds 2005: figs. 46–53); the huge dolia from Ras al Basit that reached Egypt (Mills and Reynolds 2014: fig. 5); the 3rd-4th century Ras al Basit and perhaps Amrit globular amphorae with a small ring base (Reynolds 2005: figs. 39-43; Mills and Reynolds 2014: fig. 4); the east Cilician Pompeii V which has thick, pseudoKoan handles and its successor, the early 3rd century late Pompeii V (with a bulbous rim and wide groove-moulded handles clearly related to those of the proto-LRA 1 of the mid-3rd century — Reynolds 2005: figs. 19-22; Reynolds 2008; for a complete example of late Pompeii V, with a large pear-shaped body and tiny small flat toe, bearing the typical dipintos of proto LRA 1, see the Cesnola Collection, Metropolitan Museum of Modern Art, 74.51.2434 —). The so-called pseudo-Koan amphorae of eastern Cilicia (Agora G 198 and Agora M 54) are notably wide-necked versions of the same Koan shape, both with horned double-rod handles: the possibility that these carried fruit in sweet passum wine, the famous Cilician apples perhaps (Dalby 2000: 168) (or some mixture of garum and wine?) is being investigated (Figure 3a-b).8 There is  Reynolds 2005: figs. 6-11, with references; Bertoldi 2012, for more examples; for a shipwreck off Knidos with Agora M54, Cilician 8

310

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

a

c

b Pompeii Agora G 198 (Panella 1986: fig. 16)

Alexandria Agora M 54 (Empereur and Picon 1989: fig. 5)

d

Beirut Dressel 21-22 BEY 006 6721.2 Sicilian?

Beirut Dressel 21-22 BEY 006 6721.1 Campanian

Figure 3. Wide-neck amphorae. a. Agora G 198, b. Agora M 54 and c-d. Dressel 21-22. an overall resemblance between these (especially Agora G 198) and Dressel 21-22, such as those from the so-called Garum shop at Pompeii, all of which carried fish products (Bernal‑Casasola et al. 2014),9 but Dressel 21-22 does not have a Koan-type double rod handle, rather one with a central convex rib or band (Figure 3c-d). In addition to the figs, dates and prunes exported to the Danube, Pompeii, Rome, southern Gaul, Britain and Germany in the Schöne-Mau XV (Figure 4a-c), the Colchester 105, quite a large, free-standing amphora with a ring base, vertical convex rim (for a lid) and Beirut/Akko-type handles with a central band, also carried dried fruit (dates) as well as olives (in two examples from Avenches) (Figure 4d). Several related smaller Levantine amphorae, including Augst 46, and Célestins 2A (found in Lyon and Beirut), which may have carried dried fruit (or possibly olives), are illustrated (Figure 4e-i).10 Dressel 2-4 and Agora G 199, inter alia, as well as for comments on the wines of Cilicia, see Opaiţ, Davis and Brennan 2018: my thanks to Andrei Opaiţ for this reference. For the Spanish Ministry I+D projects RACAMed (HAR2017-84242-P), directed by Alessandra Pecci and myself, and GARVM III (PID2019-108948RB-I00), ten Agora G 198 amphorae in Pompeii have been sampled. In the stores by the Forum, this is by far the most common Levantine import. There is, for example, only one Pompeii V, the Tyrian amphora of the British excavations (Timby 2004: fig. 7), and a small group of carrot amphorae from Beirut. 9  These, 81 of them, were piled upside down in the northern corner of Room 13, and were from three Italian sources: Alcamo Maria-Trapani (north-western Sicily) (Botte Type 1), Messina-Monte PeloritaniCalabria (Botte Type 2) and somewhere between southern Etruria and northern Campania (Botte Type 3). 10  Colchester 105/Peacock and Williams Class 65/Lyons-Célestins 1A (Lemaître 2000; Reynolds et al. 2008-2009: 72, 85, fig. 11a-b). The typological similarity of Colchester 105 to that of a small module amphora with a ring base (Figure 4e), both forms from Gallic contexts,

Amphorae with a tall, sealable narrow neck but lacking double rod handles are generally classified as wine amphorae on the basis of form and often their derivation from Hellenistic models, as is the case of early Imperial Chian amphorae, Imperial Roman ‘Pamphylian’ amphorae identified by Virginia Grace (exported to Beirut, Athens and Arles) (Figure 5a-e),11 Rhodian amphorae with ‘horn’ handles (we now know that these have also been found carrying figs: Dugonjić 2015: 249) or the Knidian series. Cretan amphorae of the early Imperial to late Roman periods are all classified as wine amphorae. Some indeed bear references to sweet wine made from sundried grapes (passum), also associated with west Cilician Agora 199 (‘pinched’ handle) and Samian amphorae.12 The led to their fabric comparison through petrology by Claudio Capelli and the conclusion that they share a similar provenance (Lebanon-Phoenicia, but not Beirut) (Gohier and Capelli 2013; see also Long and Duperron 2013: 139–140, fig. 23). The small Athenian Agora amphora P 11638 (Figure 4f), which in turn has a similar rim to the latter, and the same body and button base of the contemporary type Augst 46 (and Augst 47) (Figure 4g-i), which has a different, band rim and different fabric, is surely not coincidental. I would suggest that all carried dried fruit. 11  Reynolds 2010a: Table 4b, BEY 045 early 3rd century piscina contexts and mid-3rd century BEY 5051, Fabric FAM 169. For an unclassified example from the 3rd-century Arles-Rhône wreck, see Long and Duperron 2013: fig. 24 (here reproduced as Figure 5e). Claudio Capelli’s description and analysis of the fabric (which corresponds well with my FAM 169) is also provided. It is unfortunate that Capelli was unaware of the Pamphilian attribution of this form. 12  Rauh and Slane 2000, and Rauh 2004, for Agora G 199; MarangouLerat 1995: 5–29, for literary references to Cretan wine for the table and medicinal purposes, with ample references to the use of sun-dried grapes (raisins), i.e. vinum passum; 128–152, for dipintos referring to Cretan wine from the Athenian Agora (e.g. 129, P1: Πασσον) and, especially, Pompeii: e.g. 131, P6, on a Crétoise AC 2 dated to AD 74: Cret(icum vinum) Exc(ellens)…V(inum). Vet(us) (vintage wine) Mulsum

311

Paul Reynolds

a d b Beirut BEY 015 workshop (Reynolds et al. 2008-2009: fig. 7.1)

e

f Arles-JBM (Gohier and Capelli 2013: fig. 1.1)

Carrot amphorae/ Schöne-Mau XV/ Naples (Vipard 1995: Peacock and Williams Class 12 fig. 3.8)

c

g

Athenian Agora P 11638 Augst 46-47/ Célestins 1A 'hybrid'

h

Lyon. Célestins 1A / Colchester 105/Peacock and Williams Class 65 c, AD 190-250 (from Lemaître 2000, Fig. 9.6 and 8)

Lyon. Célestins 2A (Lemaître et al. 2005: fig. 11.1). AD 200-250 Augst. Augst 46 (Martin-Kilcher 1994: fig. 196.4)

Pompeii. Exhibition display (January 2020) Peach, figs, walnuts and a Beirut ‘carrot’ amphora (Photo: P. Reynolds)

i Beirut. Célestins 2A. BEY 006 5051.198 and 207. Mid- 3rd c.

Figure 4. Amphorae for dried fruit from Roman Phoenicia. a-c. ‘Carrot’ amphorae, d-f. Célestins 1A/ Colchester 105 and related small containers, g. Augst 46, h-i) Célestins 2A. early Imperial wines of Chios, Lesbos, Mitylene, Eresus, Methyma, none mixed (Koan-style) with sea-water, and the sweet wines of Mysia, Smyrna (‘black Perperene’), Nicomedia, ‘black Aegeate, Aphrodisias, Sardis (dry and sweet ‘Tmolite’) and Thera are mentioned by Pliny and Galen (Dalby 2000: 133–138). But smaller closed form containers in Crete, such as the Knossos 41 (Figure 5f), as well as in the Greek Islands and Asia Minor, could have carried other liquid delicacies (e.g. the famous quince-honey — melomeli — and grape and quince wine of the port of Cydonia-Khania in western Crete — kydonites — mentioned by Dioscorides) (Dalby (with honey); 134, P1, on a Crétoise AC 2: Γλυκυ (sweet wine); 149, P101, on a Crétoise AC 3, from Narbonne: Mel(is) Flos (with best quality honey); 134, P19 μυρτιτησ (with myrtle). For rare references to Samian wine in the early Byzantine period, see Dalby 2000: 151 and Pieri 2005a: 136. Pieri here notes that all the (many) examples from Marseille, PortVendre and Fos were pitched.

2000: 152). I see no reason, however, why Crete, like mainland Greece, could not have also been engaged in oil production and exports, as was the case in modern times and is documented for the 2nd century AD (Figure 2, again; Hadjisavvas and Chaniotis 2012). Perhaps the 6th century (TRC 2) and later imitations of LRA 2C (‘LRA 13’) served that purpose. Generally classified as wine amphorae are the 4th to first half of 6th century Agora M 273-Pieri LRA 8 type and related forms (Figure 6), with a large troncoconical body widening towards the base, whether the micaceous versions of Samos and cities on the mainland opposite (Ephesus, Miletus?, Clazomenae?), or other regional Aegean and Asia minor products following this model (large buff versions, typical imports in Beirut in the 4th and 5th centuries: Figure 6d; several different unsourced, volcanic-fabric large versions common in 6th-century contexts contexts in Butrint and Durres:

312

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

c

a

b

Athenian Agora P 3467. AD 0-50 (Grace 1973: fig. 8, no. 15)

b-c) Athenian Agora P 21176. Early 3rd c. (Grace 1973: fig. 8, no. 17)

(Photo: P. Reynolds)

d f

e

Athenian Agora P 25616. AD 267 (Grace 1973: fig. 8, no. 18)

Arles-Rhone 14, wreck ‘Pamphilian’ amphora (Long and Duperron 2013: fig. 24)

Knossos Type 41 (Hayes 1983: fig. 26, 92)

Figure 5. a-e. ‘Pamphylian’ amphorae, finds in Athens and Arles, f. Knossos Type 41.

313

Paul Reynolds

e

a

b

c

But 5381.13 Early 6th c. Micaceous pale salmon

f

Yassi Ada II. Late 4th c. (Bass and Van Doorninck 1971: figs. 8-9)

Pagasitikos Gulf Wreck 7 Agora M 273. Late 4th c.? (Demesticha and Spondylis 2011)

r

Calcareous Buff-tan Typical M 273 type in mid4th-5th c. in Beirut

d Butrint But 3257.12 Early 5th c.

s

Beirut BEY 006 13017.256 c. AD 410

Flanged rim Lid seat

t But 5058.4 Mid 6th c.

But 5058.13 Mid 6th c.

‘Samian’ (rough, abrasive micaceous red-brown)

h

5139.1 But 5139.1. 6th c. Coarse, lime-rich with mica flakes: Phocean?

g j

m k

Samos Cistern type

But 1899.4 Late 5th-550

o

v

w Nicopolis Early 7th c. (Reynolds and Pavlidis 2014)

But 5175.43 j-n) Fab 10 Black surface red fabric volcanic glass

'Ikarian' RC 22

q

p

BHX 3.102 Nicopolis Early 7th c. (Reynolds and Pavlidis 2014) cf. MykaleGlauke Limen

But 1676.111

But 1790.3

Bezeczky Type 56 Soapy Ephesus Meander Valley

But 1609.7 Mid-late 6th c. Coarse Samian with clear volcanic glass?

n

l

But 5013.3 Samos Cistern Mid 6th c. flanged rim

But 5013.5/ But 5005.11

i

But 1152.138-139 c. 525-550

u u

But 4023.1 Early 6th c. Micaceous Soapy yellow Related to Meander Valley

o-p) Nicopolis Early 7th c. (Reynolds and Pavlidis 2014)

‘Ikarian’ RC 22

Butrint But 3156.25 Early-mid 5th c.

Figure 6. a-w. Amphorae of Agora M 273 type and derivatives, 4th to 7th century.

314

Beirut AD 551 Probably Fab 10 (Hayes 2000: fig. 30.3). If so, the body shape differs from Agora M 273 et al.

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Figure 6h-m and possibly the complete profile Figure 6n;13 the smaller so-called Ikarian amphora/Kenchriae Adamscheck RC 22/Pieri LRA 10, with a corresponding lid: Figure 6q; small c. 25-29cm Koan versions at the workshop of Mastichari)14. From the late 6th to early 7th century the end of the ‘Samian’ series takes the form of the so-called ‘Samos Cistern-type’ (two quite different sources for these regularly occur in Butrint and Nicopolis, in Epirus), of smaller capacity than the earlier Agora M 273, with a taller, narrower, more cylindrical, but still tronco-conical body (Figure 6v-w). Production of one of the Samos Cistern-type variants on the mainland, on the coast immediately opposite Samos, is now known.15 Some ‘Samian’ forms in this series, in early-mid 5th as well as mid- 6th century contexts in Butrint, the latter also within the Samos Cistern type, have a flanged inner rim for a lid (Figure 6s-u). Bezeczky Type 56 (appearing in the late 5th or early 6th century?) was an Ephesian-Meander Valley version of the Agora M 273, size as RC 22 (Figure 6o-p). The continued contemporary production of LRA 3 into the early 7th century is attested by the presence of the form on the c. AD 62 Yassi Ada shipwreck, and perhaps even later by the find of one example in the Cripta Balbi, as late as the end of the 7th century.16 In the southern sector of Roman Phoenicia, some tall carrot-bodied amphorae with raised long handles (Agora M 334; my AM 14) are associated with wine, not oil presses (Reynolds 2005: 570–572, with references). A hole pierced  Quite early imports of Agora M 273 are found in Aquileia, probably the same date as the Africana III amphorae in the same well deposit, late 3rd or 4th century (Ventura and Braidotti 2015). For examples in Beirut and Butrint, see Reynolds 2010b; Fantuzzi 2020, for petrological analyses of Butrint examples of Agora M 273 variants. John Hayes (2000: fig. 30.3) classified Figure 6n as a product of the same source as Beth-Sh’an LRA 6, because of its grooving and black surface. However, there is a strong resemblance with Fab 10 amphorae, just as hard-fired and of similar colour and even profile, e.g. the rim Figure 6k. 14  Didoumi 2014: figs. 14, 16a-c. These have a knobbed toe, atypical of Agora M 273, etc. and recall small late Knidian amphorae which have the same body shape (e.g. complete examples in the Athenian Agora stores: Robinson 1959: M 305 and M 306; Grace 1979: fig. 66, right; Opaiţ 2014b: figs. 3-15, for drawings). Note, however, that Fab 10 amphorae sometimes also have a knobbed toe (see Figure 6l). 15  Lohmann, Kalaitzoglou and Lüdorf 2017: 271–287, Taf. 18-21. My thanks to Horacio González for this information. A pottery workshop dump with wasters of Samos Cistern type amphorae was found at the far western end of the Mykale Mountain range, ‘in the southern of the two bays of Dip Burnu, the Glauke Limen harbour’. These are possibly the same products as two examples in an early 7th century deposit in Nicopolis: Reynolds and Pavlidis (2014: fig. 9.2-3), relatively well-made, with even rim tops (HX 3.98 and 102) (one is reproduced here, Figure 6v). Fig. 10.7 (BHX 3.107) is a smaller module version. The LRC ‘3A’ (3C?) illustrated in the Mykale publication with the amphorae is surely residual (Lohmann, Kalaiitzoglou and Lüdorf 2017: Taf. 18, no.183). 16  Horacio González informs me that both LRA 3 and Type 56 occur together in Ephesus. In the early 7th century context of Nicopolis, Type 56 occurred alone, with 7 almost complete and 18 diagnostic fragments of Samos Cistern amphorae (Reynolds and Pavlidis 2014). However, what must be a Type 56 was found in a late 5th- early 6th century deposit in the Butrint Triconch Palace (But 5381). I am grateful to Andrei Opaiţ for reminding me of the presence of a complete example of LRA 3 on the Yassi Ada wreck (Bass and Van Doorninck 1982: fig. 8-19, P 74, a lower wall-base with an open foot, which normally would have been dated to the 5th century) and the example present in the Cripta Balbi late 7th century assemblage, which, being complete, was thought not to be residual (Saguì 1998: 318). 13

through the neck, a feature of these as well as some Beirut amphorae that are clearly typologically related in the case of the body, foot and handles, is fortunately a good indication that the amphorae carried wine, permitting the escape of gases during the fermentation of wine sealed within.17 The presence of a pitch lining seems to be associated with wine as well as fish products (the Garum shop Dressel 21-22 were pitched). Oil, not for consumption, but for lighting for example, could also be carried in a pitched amphora, as seems to have been the case with some Tunisian forms (Bonifay 2004: 463-475).18 4. Antioch and the limestone massif villages The supply of Antioch and its relation to the production of both oil and wine for export in the villages to the East in the Limestone Massif has long been problematical (Tchalenko 1953-1958; Callot 1984; Tate 1992). Many oil presses have now been re-classified by Jean-Pierre Brun as wine presses (Brun 2004: 105–123). There are nevertheless a fair number of oil presses, the majority of which date to the later Byzantine period. It has long been argued that both wine and oil from these villages were carried in LRA 1, the most common amphora in Mediterranean assemblages of the 5th to 7th centuries, produced in both Cilicia and Cyprus (Decker 2000; Kingsley and Decker 2000; Casana 2004: 113–115). It is significant, however, that workshops of LRA 1 are known solely at coastal locations, for example at Sebaste, Soli, Aegiae, in eastern Cilicia, and at Paphos, Zygi and probably Salamis, in Cyprus (Empereur and Picon 1989; Ferrazzoli and Ricci 2007; Reynolds 2005: 565–567). Not only is LRA 1 likely to have been for wine not oil but the coastal locations of workshops suggest that this wine was carried down from estates in skins, as was the case for some regions sending oil to Carthage for the annona (Fournet and Pieri 2008; Peña 1998). It is possible that there are no LRA 1 workshops in Roman Syria itself. Indeed, visiting two of the Syrian villages, Beyho and Qalb Loz (Figure 1b), I discovered that LRA 1 was absent, whereas pale, sometimes painted, calcareous Syrian amphorae were the only amphorae to be seen (Reynolds 2005: 265–266). Furthermore, a study of the amphorae from the Amuq Valley survey has not yielded any evidence of the production or special presence of LRA 1 (Mills forthcoming a: with thanks to Phil Mills for sharing this prior to publication). These 5th to 7th century western Syrian amphorae (plain or minimal painting) and the related Euphrates fabric painted series (complex painting) are ubiquitous  Jean-Yves Empereur (in this Amphora Contents congress at Cádiz), given that such holes are not found on all examples of Egyptian amphorae, suggested that they were drilled when the temperature was particularly hot. In contrast to what is being suggested here, in Egypt the wine was fermented in dolia (as also in Ras al Basit) and then transferred to amphorae. 18  There is growing evidence from organic residues contents analysis, nevertheless, that many amphorae bore pitch, even though this is not now visible. 17

315

Paul Reynolds

Zeugma AM 16

Zeugma AM 15 a

b

7064.1

7061.1

Zeugma AM 17

c

d 7004.2

7065.1

Figure 7. a-d. Zeugma. Syrian painted amphora forms AM 15, 16 and 17 (from Reynolds 2013).

316

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Figure 8a-b. Roman relief of two ships carrying separate cargoes of barrels and amphorae (Allard Pierson Museum, Amsterdam). in Homs, at Dehes and on military sites on the Euphrates from Zeugma to Resafa, and are a clear indication of the self-sufficient, provincial nature of the supply of goods carried in them to military, as well as urban sites (Figure 7) (Reynolds 2010a: note 275; Sodini et al. 1980; Reynolds 2013 and 2014; Vokaer 2013; Apamea: Pieri 2005b, Vokaer 2017).19 These may have carried both oil and wine, given their association with press sites of either nature (contents analysis is clearly needed to clarify their role). Perhaps the enigmatic painted decoration on these amphorae was an indication of the contents.20 That the supply to military sites, and certainly to the army when on campaign, was state/army — driven is implicit in the references to the annona of Oriens in the Theodosian Code (Reynolds 2010a: note 272). LRA 1 is found on the same Syrian sites, but is always well outnumbered by the central Syrian amphorae. In his speech In Praise of Antioch (20), written in AD 356, Libanius states not only that Antioch was supplied from its vast territory, but that local goods, wine, and especially oil, were also shipped abroad from its port at Seleucia (Reynolds 2010a: 269; Casana 2004: 113–115). At this precise time LRA 1, if they did serve as containers for Antioch’s oil surpluses, were only rarely exported,21 the main period of exports of LRA 1 beginning in the late 4th century. More relevant here perhaps, exports of the Syrian painted series and their early Byzantine predecessors are so far unattested for any period. We are left with a problem, then. What vessels carried these Antiochean oil exports in the mid-4th century? How  There are two main groups of the ‘globular’ type, one in a granular calcareous fabric, with minimal or no painting (e.g. AM 16) and those with complex painting in Euphrates a hard, fine, ‘plastic’ clay fabric (with multi-coloured inclusions) (e.g. AM 17). These are all called ‘North Syrian Amphora 1’ by Pieri, but the typology is far more complex than this implies: see Reynolds 2013. A free-standing type with a narrower body type found in Apamea is Pieri ‘North Syrian Amphora 2’ (Vokaer 2017: 785–786, fig. 7.1-6; apparently one of the Homs pieces belongs to that typology: Reynolds 2014: fig. 8, Type 7). 20  The decoration is usually referred to as ‘scrolls’. Curly vine tendrils? Oil palm leaves? As Agnes Vokaer has noted, these Syrian amphorae (both NSA 1 and NSA 2, given finds at Zeugma and Apamea) are to be associated with removable lids, though they may have served additional purposes (‘saucer’ form with a central handle) (Vokaer 2017: fig. 7.7‑8), perhaps, following Opaiţ’s comment on LRA 2 lids (forthcoming), an indication that it was oil that was carried? 21  Beirut: mid- 4th century examples; Athens?; those of a decade later found in situ at Kourion following the Cretan earthquake of AD 365: Williams 1987. 19

a b Figure 9. a. Kition. Marquié 2003: Type III, Pl. 28.1. b. An example of Hayes 1991, Paphos Type II (Cesnola Collection, Semitic Museum, Harvard University, inv. No.1995.10.95). was this Syrian oil exported? A relief of early Imperial date, thought to have originated in the eastern Mediterranean, now in the Allard Pierson Museum (Amsterdam), depicts two large ships, one carrying amphorae, the other a cargo of barrels (Figure 8a-b). The amphorae, with their markedly carinated shoulder and long cone-like body are almost certainly a representation of a distinctive 2nd to early 3rd century form found at Paphos, Kition, Amathonte (Hayes 1991: Type II; Marquié 2003: 4–5, Plate 28.1, Type III; Amathonte: Empereur 1987: Plate 26.4b) and perhaps also in Beirut (Figure 9). The ships may have been sailing from Paphos or perhaps from a port on the mainland such as Laodicea or Tripoli. The fabric of an example of this type found at Kition may indicate a coastal (north) Levantine, rather than (eastern) Cypriot source, as the fabric appeared to contain many foraminifera (like the products of Amrit).22 Given that wine from the region (East and West Cilicia, Cyprus, Ras al Basit, Amrit) was regularly packaged in amphorae in this period, did the barrels contain oil, rather  My observation of an example in the museum of Kition, courtesy of Sandrine Marquié. 22

317

Paul Reynolds

territories that were the primary markets for Levantine amphorae (including other common Levantine forms such as south Phoenician-Palestinian LRA 5 and Agora M 334, and Gazan LRA 4), well before exports reached significant levels elsewhere.24 It was the LRA 2 and its successors, primarily from the Peloponnese, and then from the Aegean, not the LRA 1, that carried oil to Constantinople (see below). Unfortunately, until excavations in the metropolis ensue, we are unable to document trends in imports for the 4th and 5th centuries, prior to the mid-5th to 9th century evidence presented by Hayes from Saraçhane (1992).

Figure 10. 16th century engraving by J. Della Strada depicting activities in an oil press (from Brun 2003: 149). than wine, the commodity usually associated with them (Brun 2003; Marlière and Costa 2007). A modern parallel may be relevant. In the 19th and early 20th centuries, oil exported from the Peloponnese was shipped in barrels (Blitzer 1990: 705; 2004). A 16th century Italian engraving depicting the stages of olive oil production includes its transport in barrels carried by a donkey (Figure 10). It would seem that barrels containing oil do not leak, or seep. Once the wood reaches a certain saturation point, it remains stable and the oil penetrates no further.23 Another possibility is that oil was transported in large globular glass vessels, similar to those of early modern times. These, known as carboys, formed a major component of two related Beirut AD 551 debris assemblages (BEY 006 8751 and 8752: c. 23kg of glass in the latter), a minimum of 13 vessels, from a possible storeroom in the ‘House of the Fountains’ (Jennings 2004‑2005: 182–185). Some points ought to be made with regard to the oil supply of Constantinople. Though grain was distributed free to a section of the population, following the model established in Rome, this was not the case for oil (as it was in Rome from the reign of Severus onwards) (Reynolds 2010a: note 275; Sirks 1991: Chapter 6). In the case of LRA 1, the assertion, often repeated by scholars, that the primary catalyst for the production of the type was Constantinople (Decker 2000; Kingsley and Decker 2000) does not reflect the appearance and distribution of either the mid-3rd century predecessors (‘proto LRA 1’, found in Beirut, Caesarea, the Sinai or, in quantity, at the port of Taposiris Magna, just to the south of Alexandria, for example: Le Bomin, Marchand and Reynolds forthcoming) or the mid-4th century exports of LRA 1 already mentioned. Proto-LRA 1s, in contrast, are conspicuously absent in the mid-3rd century Sassanian sack contexts in the Athenian Agora. Fourth- century LRA 1 exports, based on observations of trends in Beirut, should be seen in the context of the contemporary economic renaissance of Antioch and other Levantine cities and their  My thanks to Isabel Bonora Andújar for this information (at the Poitiers conference). 23

The presence of oil, as well as predominantly wine dealers recorded on tombstones in the port of Corycus (Rough Cilicia) has been noted and there is other evidence for both oil and wine production in late Antiquity in some coastal regions of Anatolia, including Corycus, Myra and Miletus and Sinope (as shown also in modern times on the map, Figure 2) (Iacomi 2010: 22–23; Izdebski 2013). Regarding oil production in southern Anatolia, Andrei Opaiţ in a recent paper (forthcoming) has proposed that the large amphora known as San Lorenzo 7 carried oil (Figure 11a‑d). This form, whose source has generated much debate, given the general observation of volcanic inclusions in the fabric, he suggests was produced in eastern Cilicia, on the basis of an example probably of this type classified as local in the wide range of finds of the workshop excavated at Sebaste.25 Though a source somewhere close is possible, but still needs better definition, Opaiţ gives no reason as to his identification of the type as an oil carrier, apart for the fact that it is large, appears not to be pitched and has a lid.26 Here, the typological similarity of San Lorenzo 7  Reynolds 2005 and Reynolds 2010a: Chapter 3.1-2. Trends in Antioch itself in the 3rd and 4th centuries, the capital of the diocese of Oriens, or the port of Laodicea, to the south, from which Cilician goods were exported to Egypt in the Imperial period (Tomber 1998), seem for ever to be a total unknown. East Cilician roof tiles in the yellow fabric of LRA 1 (produced at the same coastal locations as large basins and clay coffins: Aegeae?), not transport amphorae, were the primary east Cilician exports of the 1st to 3rd centuries in Roman Phoenicia and northern Palestine (Mills 2013; Mills forthcoming b). 25  For the typology of San Lorenzo 7, see Opaiţ forthcoming, with distribution and bibliography; also Bertoldi 2012: 137; Rizzo 2014: 333–334, who also comments on the similarity with Almagro 50/Keay 16, noting its presence in the phase of AD 190-200/225; Iulia Concordia: Belotti 2004: 82–83, 105, fig. 18; Turin: Quiri 2015: 167, fig. 7.1; Benghazi: Riley 1979: 208, D298 and D299; Elaiussa Sebaste: Ferrazzoli and Ricci 2010: 817, fig. 4.18. My own observations of the complete Agora P 12298 San Lorenzo 7 type piece (fired pale green) noted similarities with the mid-3rd century proto-LRA 1 (common in Beirut), as well as differences with the only example of the San Lorenzo 7 type in Beirut, though the latter also had east Cilician traits (BEY 045 1510.46, from the Imperial Baths early 3rd century piscina fill deposits: Coarse, lumpy with quartz, argillaceous and lime. Some possible sandstone. Two fragments of volcanic glass). Originally classified as Keay 16 (Reynolds 2000), the fabric suggests otherwise. 26  A good point, as this is the case for (Argolid) LRA 2s, which have lids in the same (Argolid) fabric. Here I would add Agora M 235 also, given the fabric of some lids found in the same deposits as these amphorae in Butrint (Figure 22m). He argues that amphorae carrying quality wines needed proper stoppers, not lids that could be easily removed and replaced during use. There are plenty of lids in Butrint contexts in the distinctive fabric of the so-called ‘Ikarian’/RC 22 amphora that follows the typological model of the Agora M 273 type and is more likely to have carried wine, not oil. The rim of the illustrated example has a lid seat (Figure 6q) but this is not a common variant, most rims of RC 22 being plain on the inside, like Ephesus Type 56. 24

318

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

a

Peacock and Williams Class 50 ‘Chalk’ amphora

Ostia I 568/ Alba 41

San Lorenzo 7 e

i

j

Arras Concordia Sagittaria (Venice) (Rizzo 2014: fig. 45, from Belotti 2004: fig. 18, 96)

Athens Agora P 8164 AD 140-170 (Opaiţ 2014a: fig. 26)

Vienne (Godard 1995 fig. 34.117 and 119)

f

b

Ostia IV Tav LV.445

g

c

Ostia (Ostia I Tav XLIII.568)

Beirut BEY 045 1510.46

h d

Beirut BEY 006 9923.15

Quseir al-Qadim (Whitcomb 1982: Pl. 16a, from Opaiţ 2014a: fig. 24)

h Beirut BEY 045 1510.46

Figure 11. a-d. San Lorenzo 7, e-h. Ostia I 568/Alba 41 and related forms, i. Peacock and Williams Class 50 ‘Chalk’ amphora.

319

Paul Reynolds

a

d c

Chersonesos Zeest 72 (Klenina 2015: fig. 5.1)

b

Chersonesos Zeest 73 (Klenina 2015: fig. 5.6)

Chersonesos Zeest 75 (Klenina 2015: fig. 6.1)

e f

Beirut Zeest 72 BEY 006 5051.553 Mid 3rd century Ostia Zeest 80 var. Late 2nd c. (Panella 1986: fig. 26) Athenian Agora Zeest 80. AD 257 (Photo: P. Reynolds)

Figure 12. Black Sea amphora types: a-b. Zeest 72, c. Zeest 73, d. Zeest 75, e-f. Zeest 80.

320

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

with Keay 16, the rim and handles at least, which actually often leads to their confusion, may be noted. On this basis, San Lorenzo 7 could possibly have carried fish products. This is also the favoured content suggested for the relatively large number of finds in the Paphos Theatre Excavations (Waddington 2003: 13 examples). Note here the aforementioned 1st century AD production and salting vats on the island of Castellorizo, off the coast of Lycia (Pirazzoli 1987) and the fish tanks discovered at Lapithos, in northern Cyprus (Nicolaou and Flinder 1976). Another amphora of interest with a similar distribution and date, is Ostia I 568/Alba 41, with deeply grooved handles and a distinctive fabric bearing angular pinkish inclusions (Figure 11e-h).27 Complete examples of the type (those illustrated) as well as the vessel from Qasrawet D-50 (Sinai), indicate that it has a wide, fairly cylindrical body ending in a solid spike with a mushroom toe like that of Agora G 199.28 It is possible, therefore, given its distribution, that its source may lie somewhere in Pamphylia or Rough Cilicia, but west of the production zone of Agora G 199 (as that has quite a different fabric) (Rauh 2004, for colour photographs). There are no indications as to what it carried, but some surely carried oil because they are found at the Monte Testaccio, dated to AD 252, 254 and 247-258, but classified erroneously as Gauloise 3 (Marimon Ribas and Puig Palerm 2007: 348‑349, notes 26 and 27, fig. 81, nos2-4). The unusual 3rd to 4th century Peacock and Williams Type 50/‘Chalk’ amphora, with a wide neck, rounded oval handles, long carrot-shaped body and flat hollow toe, relatively well- known in Gaul and Britain, has a fabric that suggests a source in eastern Cilicia (cf. Pompeii V amphora fabrics), Cyprus or perhaps northern Lebanon (Figure 11i).29 The dipinto (Figure 11j), which may be from  For Ostia I 568/Alba Pompeia 41, see Bertoldi 2012: 138, who illustrates 4 examples from Ostia as well as a complete example from Brescia, with a truly cylindrical, much longer body like contemporary African forms (Bruno 2002: 305, fig. 72; for a better drawing, see Opaiţ 2014a, fig. 27); Turin: Quiri 2015: 170–171, fig. 4, 21 examples (more common than San Lorenzo 7, 7 examples), fig. 7.2, with reference to Italian finds at Alba Pompeia (Piedmont), Novara, Aquileia, Venetia, Brescia, Milan, Angera; Athens: Opaiţ 2014a: 50–52, Agora P 8164, with the typical ‘mushroom’end spike; Paphos: Hayes 1991: 205, fig. LXX, no.7, later 2nd century: Hayes compares it to his Paphos Type IV, but that is a Beirut amphora; the Sinai: Arthur and Oren 1998: 201, Qasrawet 2498, fig. 5.8, complete profile; Mons Claudianus: Tomber 2006: 171, Type 59, fig. 1.66, top five examples, dated rather earlier than usual: Trajanic, Trajanic+, Antonine and even mid-1st century, with reference to the complete example from Quseir al-Qadim (Figure 11f) and others in Qana’ (Yemen), the island of Socotra off the south coast of Arabia, Alexandria, Mons Porphyrites and Berenike. A petrological analysis of the type is also presented; Beirut (e.g. Figure 11h, from one of the Imperial Baths early 3rd- century piscina fill deposits). For Opaiţ (2014a, 50-52), the form carried wine. 28  Agora G 199 (Robinson 1959). Bertoldi 2012: 141. Reynolds 2005: 564, with references. Production sites of the main exported series (in a micaceous fabric) have been located in western (Rough) Cilicia (Rauh and Slane 2000), whereas a calcareous buff version is thought to have been produced opposite, in northern Cyprus. 29  Peacock and Williams 1986: 200–201, Class 50, with a fabric description: ‘Scatter of subangular quartz and potash feldspar grains up to about 0.75mm across, together with crystalline limestone, foraminifera and occasional lava fragments. The inclusions are set in an optically anisotropic matrix, with a sparse scatter of quartz grains up to 0.05mm across.’ The (greenish) cream outer surface also recalls 27

a vessel of this type, is written in large letters typical of 3rd century proto-LRA 1. If so, and given the existence of other contemporary wine amphorae in the region, this type probably carried something else, to be determined (Fish products? Olives? Dried fruit?). The second amphora type Opaiţ (forthcoming) has added to the eastern repertoire of oil transport vessels is the large, rather globular 3rd century Zeest 80/Riley MRA 5/Hayes Knossos 39 and its later versions, for which he suggests an origin in the Sea of Marmara, the Dardanelles or Thessaly, where it is relatively common. The type is common in the Herulian sack contexts of the Athenian Agora (Figure 12f), with sporadic finds in the lower Danube and occasional finds elsewhere including Knossos, Corinth, Butrint (a complete example in the museum), Benghazi, Brindisi (where it occurs with 3rd century Sinopean/Knossos 26-27 amphorae), Ostia (an earlier version of the late 2nd century: Figure 12e) and Rome.30 Whereas handles of some examples are plain (e.g. Palaczyk 2015: fig. 2.10), others are scored with distinctive grooving which could point to typological links with the contemporary Crimean Zeest 72-73 (Figure 12a-c). Opaiţ (forthcoming) links Zeest 72 with wine production, as he does also the large Sinopean series of the 2nd to 3rd century (Knossos 26-27), and large 4th century and smaller 5th century ‘carrot-bodied’ amphorae and narrownecked Sinopean fine pale green fabric (‘argile claire’) successor of the first half of the 5th century (Figure 13a-d).31 Pompeii V. Fabric, perhaps related to Fulford ‘casserole’ 37 (actually a deep cooking pot), a form which is particularly common in Tarsus (Fulford and Peacock 1984: 26, 189, fabric 3.10; Evrin 2005). I would like to thank Catherine Godard for sending me samples of the type many years ago and bring it to my attention. I have seen only one example, in Butrint, which again immediately suggested a Cilician-north Levantine origin (For 430.1: coarse fabric with quartz, lime and red inclusions). The rounded oval handles and the calcareous fabric recall ‘north Lebanese’ forms (notably my AM 52, but the limestone is more typical of my AM 202). In France, an example from the villa of Anthée (Cité de Tongres) has a red brown fabric and light beige coat (the thinsection description is similar to that of David Williams: sub-angular and sub-rounded quartz, feldspar, white micas, pyroxenes, rock fragments, foraminifera and iron oxide) (Vilvorder, Symonds and Rekk 2000: 484). They also note that of the many examples found at Arras in a context of 370-380, two wall fragments bore dipintos in Greek (probably one from Arras that I have reproduced here, Figure 11j). However, not one recognisable example occurred in my Beirut assemblage. 30  See Opaiţ forthcoming for a summary of the distribution of Zeest 80. Bertoldi 2012: 156; Knossos: Hayes 1983: Type 39; Benghazi: Riley 1979: 188–189; Brindisi and the northern Adriatic: Auriemma and Quiri 2004, Auriemma, Degrassi and Quiri 2015; Ostia: Panella 1986: 628, fig. 26; Rome: Rizzo 2014: 351–352 (only 1 example). 31  Opaiţ 2017: note 91, refers to north Pontic amphorae and wine. For examples of these forms, see Opait 2004b (and Bertoldi 2012: 157). For Sinopean amphorae, see Kassab Tezgör and Tatlican 1998; Kassab Tezgör 2010 and 2011; Opaiţ 2010b, with reference to Sinopean ‘carrot’ amphorae and wine; Bertoldi 2012: 151–153. (and Bertoldi 2012: 151–153). For additional examples of most, at Chersonesos, see Klenina 2015 (Zeest 72, 75, 80). For Crimean and Sinopean amphorae in Beirut, see Reynolds 2010b: figs. 2 and 4. I found examples of ‘argile claire’ amphorae (which often have a hard, fine fabric with erupting lime, similar to some LRA 1s) in the likely warehouse located in the port area of Seleucia, the port of Antioch, together with Sinopean carrot amphorae and LRC (see Reynolds 2005). I believe it was these, not LRA 1, that were sampled by Empereur and Picon (1989), which led to their

321

Paul Reynolds

a

b

c

d

Beirut Sinope amphora Knossos 26-27 BEY 006 8652.1 Early or mid 3rd c.

Demirci (Sinope) (Kassab-Tezgör 2010: tav. 19, no. 1) 4th century.

Beirut Sinope ‘carrot’ BEY 006 13357.10 c. AD 410.

Late Sinopean (argile claire) AUB Museum

Figure 13. a-d. Sinopean forms, 3rd to 7th century. Whereas the later (AD 450-7th century) production of a small narrow-bodied amphora in a new pale green fabric at Demirci (Sinope) could indicate wine production and seems to be confirmed by analysis of the contents of Beirut examples (Woodworth 2017) (Figure 13d),32 the classification of the Zeest 72 (Figure 12a‑b) as wine amphorae is surprising, but probably based on the existence of finds associated with wine presses as well the massive fish factories of Panticapeion-Myritake. Recent analyses by Marshall Woodworth of Beirut examples, however, suggests that they were pitched and contained salsamenta. Though I have long suggested that the 2nd century to AD 450 Sinopean series, were fish amphorae (thinking of the Greek coinage bearing tunny fish), it seems that the 2nd and early 3rd century (mostly yellow fabric) large Sinopean amphorae (Figure 13a) may have carried attribution of Sinopean carrot amphorae to Syria and not Sinope. Only the handles tend to bear more visible tempering with black volcanics in order to strengthen them. 32  This may imitate LRA 1 (rather too narrow bodied, though, for the second half of the century) or, perhaps a late Cretan TRC 4: this is usually later, i.e. (late) 6th-7th century, but apparently appears in the 5th century at Gortina (Portale and Romeo 2000b: 422, fig. 5; Sazanov 2000, 2007; Smokotina forthcoming). The late Roman successors of the small early Roman amphorae of Heracleia Pontica (on the coast, to the west of Sinope) also bear a resemblance to LRA 1s, indicating a possible similar change, exports of wine for the annona? (these were presented by Vnukov at the Per Terram, Per Mare amphora conference in Cyprus: unfortunately not included in the publication). Unlike the Sinopean argile claire late series, these are not attested in Beirut.

olive oil (reaffirming Strabo’s description of the region as being full of olive groves), whereas the 4th-5th century red fabric Sinopean ‘carrot’ amphorae (Figure 13c) are all pitched and seem to have carried wine, three of them being possibly re-used to carry oil.33 The amphora that Opaiţ (2007b) associates with Crimean fish exports, with a wide distribution in the Black Sea, is the large, wide-necked Zeest 75 (Figure 12d). This, we may note, has handles with a deep concave moulding and central groove that could indicate also a product in wine (like the Agora M 54). 5. Egypt Tripolitania’s 1st to 3rd century connections with Egypt are emerging from recent work on sites along the Nile and its tributaries, and in the less accessible oases, an indication of cross-desert inter-regional traffic (Bonifay 2007a; Mazou 2015). However, the distribution of Tripolitanian oil was minimal by the 4th century and there is a marked rise in Tunisian non oil-bearing amphorae from the mid5th century (cf. Serapeum context in Alexandria: Bonifay and Leffy 2002).

 I am extremely grateful to Marshall Woodworth for allowing me to make these preliminary comments prior to the publication of this important work. See Opaiţ (2010b) for his definition of the Sinopean ‘carrot’ series as wine amphorae. 33

322

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire a

Egypt, otherwise, in the mid-Roman and early Byzantine period derived oil from its own resources, primarily from olive oil produced in the Fayum. Whereas the 5th to 7th century monasteries in northern Egypt were engaged in wine production, those in the Fayum, at Karanis and Abu Jeremias in Memphis, and that of St Symeon at Aswan produced olive oil (Brun 2004: Chapter 6, especially 169–183). Other oil, whether for lighting, food preparation or the perfume industry — an important consumer of specially prepared oils — was derived from non-olive sources: linseed, linum and sesame oil (Pecci et al. 2010) for cooking and lighting; poppy seed oil for cooking (Díaz-Iglesias and Galán 2007).

a

Homs Survey Baalbek amphorae (Reynolds 2014: fig. 2h)

In Egypt, as in the rest of Oriens, soldiers were stationed in the cities, as well as in the forts of the eastern limes. That soldiers in Egypt received oil in kind is clear. As in Syria, the supply of this oil was from local resources, managed in the later Empire through the annona system, controlled by the military, supplied through taxation or imperial rents in kind gathered and transported by local civilian authorities (Reynolds 2010a: note 272; notes 275, 276; Gascou 1989; Isaac 1990; Brun 2004).

b

b

Homs Survey Roman storage jars (Baalbek products) (Reynolds 2014: fig. 2i)

As elsewhere, Egyptian amphora forms, whether those of the Nile Valley or the ‘bag-shaped’ LRA 5 amphorae of Palestinian tradition produced in the 5th century and later monasteries or hermitages such as the Kellia, are generally thought to have carried wine (the classic tall, bi-conical amphorae of the Imperial period and late Roman types such as Egloff 172 and LRA 7 are often associated with wine presses and dolia) (Dixneuf 2011). The use of animal skins for the local transport of oil is, however, mentioned in late 1st and early 2nd century AD texts (Brun 2004: 169, 172; Brun 2003: 164–165, for other examples and the listing of the uter olearius in Diocletian’s Price Edict, 10.14 and 15.98).

 Reynolds 2010a: note 189 : my thanks to Zana Gabriel for showing me the pottery she has collected from the Mtein press sites; for the Roman winery at Kamed al Loz, see the web site of University of Freiburg, : my thanks to Hanna Hamel for this reference. 34

Homs Survey Roman storage jar rims and handles (Baalbek products) (Reynolds 2014: fig. 2i)

c

6. Lebanon-Phoenicia Roman Lebanon, like its modern successor, was a major producer not only of wine but also of olive oil. Surveys in the Mtein region and excavations at Kamed al Loz, in the Beqa Valley, bear witness to early Imperial Roman wine production, both lying within the modern wine production zone (Map, Figure 1b).34 Baalbek’s 3rd century small table amphorae have the grooved handles of Koan amphorae that surely identified them as wine vessels (Figure 14a; Hamel 2008; Reynolds 2014). Other amphorae of local type (unfortunately too fragmentary even to draw: they are not Beirut or the usual north Lebanese amphora forms) are occasional finds on Roman wineries in Mtein, alongside ubiquitous fragments of large

Homs survey Baalbek amphorae (Reynolds 2014: fig. 2h)

Beirut. BEY 006 12271.54 Late 1st century AD small amphora (oil?) North Lebanese fabric (lime-rich FAM 43C)

Figure 14. a. Homs survey Baalbek amphorae, b. Homs Survey Roman storage jars, c. Beirut. Late 1st century small amphora (for oil?), North Lebanese fabric. storage jars. Similar jars, for wine and/or oil, in typical Baalbek fabric and others are common on sites surveyed to the east of Baalbek and in the Homs region (FischerGenz and Ehrig 2005; Fischer-Genz 2014; Reynolds 2014) (Figure 14b: note the grooved handles, similar to those of Baalbek amphorae). As in modern Lebanon, perhaps, it is possible that olive oil production was even more widespread (Figure 15a). The late 1st to 2nd century expansion of settlement

323

Paul Reynolds

b

Mtein

a

Beit Shebab jar (Mtein) (Hankey 1968: fig. 2)

Figure 15. a. Olive oil production in Modern Lebanon. Map of modern production area, taken from www.lebaneseoliveoil.com, b. Beit Shebab jar (Mtein). After Hankey 1968: fig. 2: the scale is 10cm. into the uplands of northern Lebanon can be associated with a range of north Lebanese amphorae, such as those found at Yanouh (Figure 16a).35 Beirut was certainly a market for the goods carried in them in the 1st to mid3rd centuries, but no 4th century or later north Lebanese amphorae reached Beirut, if they existed (Figures 16b-d, f-j, for the most common forms; for a small jar, which may have carried smaller, more portable quantities of north Lebanese oil, see Figure 14c: Reynolds et al. 2008-2009: fig. 15.6). My Type 52, which is possibly north Lebanese (Tripoli?), is a common import in early 3rd century Beirut and is perhaps the only 3rd century (Lebanese) type to regularly reach the West, or at least the ports of southern Gaul (Arles, Toulon) (Figure 16e), though what this large amphora carried needs to be investigated (Oil? Wine?). The presence of a variant of AM 52 at Yanouh is encouraging with regard to a north Lebanese provenance (Gatier et al. 2001: fig. 1.5). Though Yanouh was certainly producing olive oil in the 6th and 7th centuries (JeanSylvain Cailloux, in the Poitiers conference), I know of  Gatier et al. 2001: Pl. 1.1: one of three solid toe bases illustrated, but fig. 1.2 may be the base of an Amrit amphora; see Butcher (2003) and Reynolds (2003a, 2005, 2014), for the political development and Flavian date of ceramics and settlements in inland Lebanon, following the end of the Herodian dynasty, who uptil then held much of this territory, including the Beqaa Valley. Baalbek, included within the territory of the Augustan colony of Beirut, was granted ius italicum and its own lands by Septimius Severus and it may be that the production of wine carried in their own black-fired amphora with a grooved handle began then, possibly linked to that of the local jars in the same fabric previously mentioned (Figure 14a-b). My sincere thanks to Michel Pasqualini and Jean-Pierre Brun for allowing me to publish the drawing of the (so far unique) complete example of AM 52 from Toulon in advance of their publication of the excavations in the port (Figure 16e). The fabric of this example differs from typical examples of AM 52 in Beirut (for which I suggest a source in Tripoli). It is closer to a fine version of the fabric of AM 202. 35

no contemporary local amphorae that could have carried these goods and it is possible that amphorae were not used for that purpose. Two other oil production sites should be mentioned, at Khalde and Chhim (see Map, Figure 1b). At the coastal mansio at Khalde, 30km south of Beirut, in the 5th and 6th centuries a basilica, perhaps with a martyrium, formed the nucleus of a settlement in which oil was pressed by almost every household (Brun 2004: 102, 104; Duval and Caillet 1982). Like the oil production at Chhim, further to the south, there is a strong possibility that the oil production at Khalde was in the hands of the Church. In the mountains at Chhim, east of the port of Jiyé/Porphyreon, a 5th or 6th century press room is possibly to be associated with the finds there of a large local amphora type, though this could equally date to the early Imperial period (Figure 17) (Kowarska and Lenarczyk 2014; Reynolds 2004b; Reynolds 2005: 570, figs. 92-94). Also typical at Chhim are large storage jars with a pale, kaolinitic fabric, probably imported from the Hula Valley in northern Galilee or the Golan (Figure 18) (Hartal 2002: fig. 24). Jars like these, with a collar neck and ring handles on the shoulder, are found in Mtein and are typical finds on rural sites around Baalbek and Homs, as we have seen (Figure 14b). These should be seen as the principal containers for either wine or oil during the early Imperial period. The specialised modern production and trade of similar storage jars (muzannar) at Beit Shebab (Mtein) in the mountains to the north of Beirut should here be mentioned (Figure 15b) (Hankey 1968). The late Ottoman production of oil jars, in various sizes, down to c. 30cm (cf. the Roman vessel Figure 14c), is well attested at the southern Lebanese site of Rashaiya al Fouhar (personal observation in 2000).

324

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

c

AM 202

a BEY 045 1766.2 Mid-2nd c.?

b

d BEY 006 10128.94 c. 100-150

BEY 045 1228.9 Early 3rd c.

‘North Lebanese’ Lime-rich, quartz, iron oxide fabrics FAM 43

Yanouh (Gatier et al. 2001: Pl. 1.1)

Beirut workshop BEY 015 Form 3

m

Late 1st-Early 2nd c.

f BEY 006 11630.39 FAM 43C c. 125-150 (c. 150?)

AM 52 Unsourced Similar to Beirut Tripoli?

e

g

BEY 045 1242.41 Early 3rd c.

k

BEY 045 1242.732 Early 3rd c.

h

l BEY 045 1381.3 Early 3rd c.

BEY 045 1242.137 Early 3rd c.

j i

BEY 006 13053.13 FAM 43F? Early 5th c. BEY 006 2318.56 Late 4th/5th c.

AM 52 Toulon. US 303 (Brun and Pasqualini forthcoming)

Figure 16. North Lebanese amphorae. a. Yanouh (Gatier et al. 2001: Pl. 1.1), b-d. Beirut, Reynolds AM 202, e. Port of Toulon, US 303, Reynolds AM 52 (Brun and Pasqualini forthcoming), f-l. Beirut, AM 52, 2nd to 5th century examples.

325

Paul Reynolds

Figure 17. Chhim oil amphora (from Reynolds 2004b).

Figure 18. Khirbet Zemel. Roman Golan storage jars (from Hartal 2002: fig. 24). The scale is 10cm.

Figure 19. Koroneika storage jars (from Blitzer 1990: 687, fig. 2). In late Ottoman Greece, the potters of Koroni on the south-western coast of the Peloponnese (Messenia) also produced a range of sizes of storage jars that were traded all over the Mediterranean. These were used to store oil in each household, or surplus stocks of oil, in the case of the oil merchants and are common finds in Provence, perhaps as a result of their export to Marseille for the

local soap industry (Figure 19) (Blitzer 1990). The north Tunisian jars of Nabeul and Jerba, probably containing oil, were also well traded in the same period (Blitzer 1990; Michaelides and Bakirtzis 2003; observation of extant examples from Nabeul at Carthage-Dermech and in the Durres Archaeological Museum). The Koroni jars recall the large Roman dolia of the Byzantine period, such as those

326

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

produced at Ras al Basit and Amrit, that were still in situ in the Beirut houses destroyed in the AD 551 earthquake (Thorpe 2017-2018; Reynolds in preparation a). The Koroneika were primarily for oil because it was said that wine did not keep well — their wide necks would certainly have been a problem given that the contents needed to be air tight. Where barrels could not be afforded, the Koroni jars could be filled with wine so long as the tops were sealed and wine was accessed through a hole drilled through the wall. Given the availability of amphorae that carried both local, as well as a wide range of imported wines, was it common practice in Beirut to store local olive oil in those late Roman dolia? Alternatively, these may have served simply as containers for water if there were no convenient cisterns.36

at the Monte Testaccio in Rome bears a dipinto dating to AD 254, possibly referring to this city, would suggest that the vessel carried oil.38 The contemporary 3rd century form Knossos 18, which also has a cupped rim, if reconstructed correctly from examples in the Punta Mazza wreck (Milazzo, Sicily), is quite a different form, with a much longer, more cylindrical body than Dressel 24, akin to Tunisian amphorae, perhaps imitating Keay 3/Africaine I, which probably carried oil (Figure 20g-i).39 A small module, free-standing amphora in Dressel 24 fabric with the same handles and vertical convex rim and a domed ring base is very common in the Athenian Agora (M 177) and presumably carried either smaller, more manageable quantities of oil or something else very specific (Figure 20f).

7. Greece and the Aegean: Dressel 24, Agora M 235 and LRA 2

Dressel 24 was previously generally assumed to have carried wine (Chian wine, for example; and Chios was just as famous for its exports of multi-purpose mastic), but it (and derivatives) is now universally accepted to have transported olive oil on the basis of the identification of the type at the Testaccio in layers dating to between AD 247 and 254 (Marimon Ribas and Puig Palerm 2007: 351– 353, figs. 79-80). Note, however, that of the nine examples, whereas some are probably Knossos 15 (cf. Figure 20j) (Marimon Ribas and Puig Palerm 2007: fig. 80, nos.4 and 6, similar to the published vessels from Trieste, and possibly nos.2, 3 and 5), three are clearly early versions of LRA 2 (one illustrated, Figure 23d), contemporary with those in the Athenian Agora (see below, Figure 23a, P 25618), or other early (Peloponnesian?) variants with a straighter rim close to those found in Dobrogea and Butrint (Figure 23e-i; Reynolds 2020). This is good at least for evidence for early exports and the oil carried by these possible Peloponnesian precursors of LRA 2. The unusual abundance of Knossos 15 at Trieste in the northern Adriatic in a mid-3rd century deposit may, theoretically, be connected with the contemporary drop in the Baetican oil supply for the military annona and its replacement by Asia Minor sources (Auriemma and Quiri 2004: 49–50).40

The classic Dressel 24 type is found in contexts of the late 1st to 3rd century AD (Figure 20d, from Lyon, having an identical to Dressel’s type piece). The number of regional 1st to 3rd century variants related to Dressel 24, including presumably direct successors of the 1st century Dressel 24, is becoming complex (Figure 20: Opaiţ 2007a, 2010a; see Rizzo 2014: 319, fig. 42, for a summary of Opaiţ’s Dressel 24 similis A-D forms). Hellenistic amphorae, possible predecessors of Dressel 24, as well as the later large-bodied Zeest 90/MRA 18 were produced at Erythrai -Ildiri (on the mainland, opposite Chios) (Figure 20b-c).37 The latter was also produced in Chios (Figure 20a) and possibly at Kyme: a shoulder fragment classified as Dressel 24 similis  One such Beirut house dolium was sampled for its contents but proved not to bear traces of wine or oil and so may have been used to store water. I would especially like to thank Alessandra Pecci for these analyses. 37  A workshop for the Hellenistic predecessor of Dressel 24 (with a band rim, lid seat and Hellenistic Chian-type toe, hollow underneath) has been found at Erythrai-Ildiri, opposite Chios (Özyigit 1990, with a photograph of rims, fig. 5: I am grateful to Sevérine Lemaître for sending me a copy of this article; Opaiţ and Tsaravopoulos 2011: 327–404, including fabric descriptions and petrology, pp. 314–315). Amphorae with a tall rim and very tall neck, some stamped, produced in Chios are apparently 1st century BC predecessors of Zeest 90 (see here Figure 20a; Opaiţ 2007a; for full details, also Opaiţ and Tsaravopoulos 2011). If I understand correctly, later versions of Zeest 90/Riley MRA 18, bearing Greek and latinised Greek stamps, such as the many examples found in the late 2nd-3rd century barracks at Novae, were also produced at Erythrai, one bearing a coin of the city impressed on a wall fragment classified as being of this type (Dyczek 2008: fig. 6). Both Andrei Opaiţ (2007a) and Piotr Dyczek consider Zeest 90 to be the predecessor of LRA 2. Early ‘classic’ Dressel 24 examples and Dressel 24 similis of the second half of the 1st century AD occur in Pompeii (Panella 1986: 624–625, fig. 22, with Dressel 24 similis, fig. 23) and Lyon (Figure 20d, Dressel 24, with a tall neck and less convex rim; Bertrand 1992; Lemaître 1995). The form Knossos 15, of the mid-2nd to mid-3rd century, in a ‘smooth, clean brick red’ fabric, with a relatively small, more ‘articulated’ cupped rim and tronco-conical neck, should or may belong in the series (Figure 20j) (Hayes 1983: 141, 147, A 46; Bertoldi 2012: 155). For other examples from Trieste, where the form is very common (52% of the eastern amphora imports), see Auriemma and Quiri (2004: 49–50, fig. 10A). Related to that Adriatic supply perhaps, Dressel 24 appears in Dürres in the 2nd century (Reynolds 2010a: 203, Table 3a) and there are two wrecks carrying (Chian) Zeest 90/Dressel 24 similis sunk in the southern Euboean Gulf (Vidličková 2015). A total of 7 Dressel 24 diagnostics (5 rims and one base) were found in the Butrint Forum assemblage, most appearing in the mid-3rd century (Contexts 26 and 98), with only one occurring on the Vrina Plain (VP 5017.9: Reynolds 2020, 145, fig. 8.3.3). 36

 KΥΜΔΙΩΝ. Though the Spanish authors interpreted the reading differently, as the name of the producer: Κυ(ρίω) μ(οι) Διων(υςίω) (Remesal Rodrίguez and Garcίa Sánchez 2007: 176–177, no.519), Opaiţ and Tsaravopoulos (2011: 314–317) present good grounds for considering this a reference to Kyme (in Lydia), a major oil producer from the 4th century BC onwards. Opaiţ classified the shoulder as that of a Dressel 24 similis. 39  The few examples of Knossos 18 I have seen in Beirut and in the Athenian Agora have a fabric unlike that of Dressel 24 (at least the hard, fine micaceous Dressel 24s I have seen, see below). My macroscopic description of P 32199 (Figure 20i): Greenish cream outer coat, smooth surfaces. Very pale orange cream inner surface. Very pale orange yellow fabric. Well fired, fairly granular. Abundant fine to 0.5mm gold mica. Common 0.5mm lime. The base is of Dressel 24 type. My macroscopic description of the small Dressel 24 similis, P 12467 (Figure 20e): Hardfired, micaceous fabric. Yellowy-orange fabric. The break is quite fine and rather uneven. Abundant mica dust glitter and moderate 0.5-1 mm gold mica flakes on the surface, which bears a cream yellow buff coat (firing). This is what I consider normal for Dressel 24. 40  Dyczek (2008: 517) also refers to an example of Zeest 90 from Romula (Dacia) bearing a dipinto oleum in Latin, whereas one from the military hospital at Novae carried oysters according to a dipinto (ΑΛΜΟΣΤΡΑ) and a second refers perhaps to ‘Pontic nuts’ (ΠΟΝΤΙΚΟ [TO] ΚΑΡΥΟΝ), with a 38

327

Paul Reynolds

b

e

d

a

Athenian Agora Agora P 12467 Dressel 24 similis Small module AD 50-100

Chios workshop (Opaiţ and Tsaravapoulos 2011: fig. 10a-b)

f

Chersonesos Dressel 24 sim./ (Zeest 90) (Erythrai product) (from Opaiţ and Tsaravapoulos 2011: fig. 53b)

Lyon (St Juste) Dressel 24 AD 50-100 (from Bertrand 1992: fig. 8.4)

Athenian Agora Agora P 12314 cf. Agora M 177 Small module with Dr 24 top (oil?)

g

c

h

Knossos Knossos Type 18 (Hayes 1983: fig. 22, A 49)

i

Athenian Agora Agora P 32199 Knossos 18 type AD 267

Vimniacum ‘Dressel 24 similis D’/(Zeest 90) AD 150-200 (Opaiţ 2007a: fig. 8.42)

j Punta Mazza Wreck Knossos 18 type 3rd century (Ollà 1997: Tav VII.16)

Knossos Knossos Type 15 (Hayes 1983: fig. 22, A 46)

Figure 20. a-j. Dressel 24 and related forms (Dressel 24 similis, Agora M 177, Knossos 18, Knossos 15).

328

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

In the case of the East, Dressel 24 and forms related to it were the most common amphora imports on Danube military sites (Opaiţ 1996, 2004a-b; Bjelajac 1996; Dyczek 2001, 2008; Benea 2000, with reference to its abundance at Histria and military vicus of Tibiscum). The type (in classic hard, fine micaceous Dressel 24 fabric: handles, bases, but rarely rims) appears in Beirut in the first half of the 2nd century, was relatively common in contexts of the late 2nd and early 3rd century, but was practically absent in mid-3rd century contexts (a single example). Turning to the Greek mainland and to the Peloponnese in particular, Aigion and Sikyon, packaged and exported oil from the late 2nd century BC to early 1st century AD in amphorae related to Dressel 25, with its distinctive flange at the base of the rim, bowed round-sectioned handles and ovoid body (Figure 21b-d, from Ephesus).41 A late 2nd century BC-first half of 1st century BC version with a taller bell-shape rim was its predecessor, Opaiţ’s protoDressel 25 (Figure 21a, from the 1st century BC Styra A shipwreck).42 A similar but much later variant dated AD 20-40+ is also illustrated (Figure 21e).43 The similarity of these early Roman vessels to late Republican-Augustan Brindisi oil amphorae and Spanish derivatives (e.g. Baetican ‘Oberaden 83’, the predecessor of Dressel 20) has long been suggested and it is assumed that they carried olive oil.44 Another, smaller type found in the second half of the 1st century AD in Athens (Figure 21f, P 12371), with a shorter neck, and round section handles, in a dark brown (Corinthian?) fabric has appeared in fair numbers in Aquileia, sites in the Veneto and Padua in 1st century contexts (where it has also been classified as ‘Dressel 25’).45 find of amphorae of this type in the River Danube at Batin, near Novae, still containing walnuts. Oysters were another well-known product of the Gulf of Actium (Dalby 2000: 143), with good numbers of them appearing in early Roman contexts in the Augustan Actium Victory Monument at Nicopolis (Reynolds forthcoming). These, I would have thought, need not have been carried in amphorae. 41  For Aigion and the production of three amphora forms, types II (with a double flat band, stepped moulding on the neck: end 2nd century BC to early 1st century AD) and III (with a double ridge moulding: 50-0 BC) being related to Dressel 25 (Sikyon and Apani forms), see Filis 2019; For imports at Ephesus, see Bezeczky 2013: 91–93, Type 22, nos.172‑185, contexts dating from 30/27 BC to 23/45 AD. No.180, with a much shorter neck is possibly the same version as Agora P 12371, here Figure 21f. In general, see Opaiţ 2010a:155–156. 42  Cf. Opaiţ 2010a: PL 87.1, Athenian Agora P 6795, which is close to the Styra Wreck vessel illustrated here. For another from the same context, see Grace 1979: fig. 38, P 6796. According to Filis (2019: 17) this is a product of Aigion. 43  Agora P 33075. Very pale salmon orange smooth surface coat with fine hard pale pink orange fabric. Mica dust and some fine gold mica. Common 0.5-1mm yellow lime. Moderate 0.5-1mm red pellets (iron oxide or fine mudstone). Well made, thin round section handles. 44  For Brindisi, see Manacorda and Pallecchi 2012 and Manacorda 2019. Production of the form at Corinth, following the Roman sack in 146 BC, would presumably not have begun until after the foundation of the Augustan colony. 45  Mazzocchin and Gualtieri 2004, for details of its distribution in Italy. The examples from Padua occurred in a drainage deposit dating to the second half of the 1st century AD, like the Agora piece. Thin-section analysis of one of the 47 examples in this deposit determined that the inclusions were primarily lime-limestone with additional rocks, calcite and chert. The absence of mudstone, but presence of chert and calcite is more typical of Epirote amphora fabrics, but these are not usually fired brown in this period. The complete example from the Agora (see here

Indeed, this form, with its rounded rim and well-curved handles, is closer to the amphora illustrated by Dressel than the wide-necked P 33075 and parallels presented as Dressel 25 by Opaiţ (2010a). The Lechaion rim (Figure 21g), of early-mid 1st century date, was a rare diagnostic related to this type (a few round sectioned handles and base fragments were also found in the same foundation contexts of the basilica excavated in 2018). The late Roman type Agora M 235/Remolà Tipo Tardío A type has generally been classified as a wine amphora (Figure 22d-j, for classic 4th-5th century, early Byzantine versions). Opaiţ (2014a: fig. 1) has presented a summary of the possible development of the series, beginning as early as the early 1st century (Figure 22a), though this needs confirmation, the forms gradually reducing in size and becoming more oval and squat from the second half of the 5th century (P 12707) to the mid- 6th century (Figure 22l, P 13468). This general sequence can be observed in the display of the amphorae in the Agora collection in the basement of the Stoa of Attalos (Reynolds 2010b: fig. A.13f). The latter two Agora pieces, however, with their short tronco-conical necks, seem to bear a typological relation to Cretan TRC 2, which begins in the early 6th century (cf. Figure 22o). I argued for a Cretan origin for Agora M 235 on the basis of its fabric, general typological details and characteristic (string?) cut-grooving on the body, because the latter grooving and fabric are identical to one of several versions of the 6th century late Cretan form TRC 2 present in Butrint (some are certainly Cretan), with a more squat, globular but still tronco-conical body than earlier Cretan amphorae (Figure 22o).46 However, Agora M 235 proper is not common in Crete and there are good grounds to consider these string?-cut products to be south Peloponnesian, and of the city and port of Messene or its territory.47 Indeed, there are a fair number in Sparta (Pickersgill and Roberts 2003), but its major market was clearly coastal, with finds in Argos, Corinth, including Lechaion (Figure 22i), Athens, Butrint, and Figure 21f, P 12371), with dark brown chert or mudstone inclusions and lime, struck me as being possibly from Corinth (the base, not visible in this photograph, is a small, flattish knob type). The most common local amphorae in Augustan and mid-1st century contexts in Lechaion, the western port of Corinth, correspond to other amphorae related to the Corinth 243 type (Slane 1990: 114, fig. 28; Reynolds 2020: Appendix A, Fig. A.10, ERAM 6) and not Dressel 25, though some diagnostics were certainly present (e.g. Figure 21g). 46  Reynolds 2010b: 96, fig. 5c. For these late Cretan amphorae, see Portale and Romeo 2000a-b and Yangaki 2007. In the Agora store these M 235s and TRC 2s were presented as a continuous series in the stacks, precisely because they share the same fabric and body treatment (groove and flat band ribbing), which could well be a Peloponnesian trait. Messene in the 6th century appears to have adopted the late Cretan shape, which has direct predecessors in Crete (Cretan MCR 2), here breaking with its long tradition in the production of M 235. For similar pieces in Corinth, see Slane and Sanders (2005: 278, fig. 11, 4-26), and in Argos, see Aupert (1980: 440, fig. 46, no.326a), with the same (string-) cut grooving. 47  See Yangaki 2005 and 2014, for the evidence that Agora M 235 was one of the products of Messene, found at a pottery workshop and in typical local fabric.

329

Paul Reynolds

a b

Ephesus. Bezeczky 2013: pl. 15, no. 172

c

Styra A shipwreck Opaiţ ‘proto-Dressel 25’ First half 1st c. BC? (Vidličková 2019: fig. 3.2)

d

Ephesus. Bezeczky 2013: pl. 15, no. 176

Ephesus. Bezeczky 2013: pl. 16, no. 179. Aigion product?

f

e

Athens Agora P 33075 Context R 10: 1 Dressel 25 AD 20-40+ (Reynolds in preparation b)

g

Athens Agora P 12371 Context N 20: 2 AD 50-100 Dr 25 (Corinthian?) (Photo: P. Reynolds)

Lechaion (Corinth) LHSLP 18Cb 09 002 Context 3729.4

Figure 21. a-f. Dressel 25 and related forms.

330

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

b

aa

Athens. Agora P 4130 Early Ist c. AD Precursor M 235? (Opaiţ 2014a: fig. 3) Early Dressel 24?

c

Argos. 2nd or 3rd c. (Abadie-Reynal 2007: pl. 71, 444.2)

ig

f

k

Lechaion. Agora M 235 Context 18Ba 01 028. 3473. Late 4th c.

Leptis Magna. c. 350-360 (Bonifay and Capelli 2013: fig. 14)

j

h

l

Gortina Agora M 235 Probably 4th c. (Portale and Romeo 2000a: 351-2, pl. LXXa)

Athens Agora P 16074 ‘AD 300-350’ (Opaiţ 2014a: fig. 7)

Athens Agora P 25617 (Mid-) 3rd century (Opaiţ 2014a: fig. 5)

g

Argos. ‘4th c.’ (Abadie-Reynal 2007: pl. 70, 444.1)

e

dd

Leptis Magna. c. 400 (Bonifay and Capelli 2013: fig. 19)

m) Butrint Lid. But 1300.48 c. AD 530. Fabric as Agora M 235

m o

q

p

n

n) Butrint But 5014.55 Precursor of TRC 2 type? Early 6th c.

Butrint But 3257.23 c. 430-450

r Athens. Agora P 13468 (from Opaiţ 2014a: fig. 9). Mid- 6th c.

Athens Agora P 4108 (from Opaiţ 2014a: fig. 8). AD 400-450

Tarragona (ANG-12) Mid-5th century (From Remolà i Vallverdú 2000: fig. 87.5)

o) Butrint But 5383.2 6th c. TRC 2

Butrint But 1422.70 Late 5th-e. 6th c. Corinth (from Slane and Sanders 2005: fig. 5, 2-18). AD 450-500

Figure 22. a-k, p-r. The development of the Agora M 235 type, early 1st century AD to c. AD 500, l. Agora P 12468, related to TRC 2, m-o. 6th century TRC 2 and lid in likely Peloponnesian fabrics.

331

Paul Reynolds

further afield in surprising numbers, at Leptis Magna (Bonifay and Capelli 2013: Figure 22g-h) and Tarragona (Figure 22j; Remolà i Vallverdú 1993, 2000), where it was first properly brought to our attention.48 Perhaps its distribution was connected with the distribution of green porphyry marble from the quarries of Croceae (in the Taygetus mountains), which was listed in the Price Edict of AD 301 as one of the two most expensive marbles.49 The classic, traded, Agora M 235 is found usually in the late 4th and 5th century, but with definite earlier examples (with more vertical handles) appearing in Athens and in a context of c. 350-360 at Leptis Magna (Figure 22g-h). Other major type pieces (those from Argos and Gortina illustrated here: Figure 22b, e-f) are unfortunately poorly dated (Abadie-Reynal 2007: 245, Pls. 70-71, 444.1 and 444.2, with only one other example noted; Portale and Romeo 2000a: 341, fig. 166, 351-2, Tav. LXXa: of a total of only 4 examples; Reynolds 2010b: fig. 5b, P 12713, a complete example with a dipinto, from the Athenian Agora). The 4th-5th century form is often pitched (noted by D. Pieri for the Gallic examples, and by A. Opaiţ for those of the Agora) and so the type is thought to have carried wine. However, the possible early 1st century version of the type — if indeed from the same source — has a notable resemblance to Dressel 24 (Figure 22a, P 4130). The similarity may be noted between what Opaiţ now (2014a: fig. 1 and 5, P 25617) sees as the mid-3rd century form of the Agora M 235 series (with the same grooving and a rather cylindrical neck) (Figure 22c, Figure 23c) and a very similar piece, but with a more globular body and different fabric in the same AD 267 deposit (Figure 23ab, P 25618), perhaps an Argolid early LRA 2.50 Given that oil presses are to be found in the territory of Messene  For other finds, as well as suggestions as to its development, not without its problems, see Opaiţ (2014a: 43–48), who suggests that the characteristic grooving is due to the use of string to hold the vessel together during the manufacturing process, having observed this technique in Tunisia. The Agora M 235 type was classified by Pieri as his LRA 12 (Pieri 2005a: 139, Pl. 60) and by Remolà i Vallverdú as Tarragona Vila-roma 8.198/Tipo tardío A (1993 and 2000: 234, figs. 87-88). For finds in Leptis Magna in contexts of c. 350-360 (Period 3) and c. 400 (Period 4) (illustrated here, Figure 22g-h), and possibly also earlier, in Period 2 (c. 290‑310), see Bonifay and Capelli 2013. 49  Dalby 2000: 142 and Karagiorgou 2001: 186; see also Sodini 2002-2007, for the major role of mainland Greece and the Aegean islands in the export of marble in Late Antiquity; for the many marble quarries in Thessaly, notably those of verde antico, the third most expensive marble in the Price Edict, that became very sought-after in the 6th century, see also Karagiorgou 2001: 171–187. 50  Opaiţ (2007a: 632, fig. 9.50-51) previously classified both these vessels, which occur together in Deposit B: 5 over a floor and associated with the Herulian sack of 267, as early versions of LRA 2. Whereas P 25617 has occasional lime (and I classified it as possibly the same source as Agora M 235), P 25618 has a lot of lime, some large and exploding (like a normal Argolid fabric, but without the gold mica flakes). P 25618 has flat(tish) band ribbing whereas P 25617 has separate ‘cut’ ribbing typical of M 235. Though the rim and handles are identical, a key difference is the more tronco-conical neck (akin to LRA 2) and body of P 25618. That both occur together cannot be a coincidence. They are presumably two contemporary workshops, relatively close, trying to produce the same type, with different results. 48

(Karagiorgou 2000; 2001; Yangaki 2005) and this was also the source of the koroneika in Ottoman times, it is possible that the Agora M 235 type also carried oil. In this respect, one of the Argos examples, probably 4th century, given its tall neck, bears a dipinto referring to oil, though it has often been commented that such dipintos indicate the exception to the rule (Figure 22f). In Butrint, mid- to late 5th century tall-necked amphorae with two handles (Figure 22p-q) that are in a hard, micaceous fabric very similar to that of both Agora M 235 (e.g. the aforementioned complete example P 12713) and the ‘Lakonian’ one-handled Agora M 315 (Figure 25d), do not ‘fit’ Opaiţ’s proposed development of Agora M 235, though they share the same rim and handles. Of the same date (second half of 5th century) is a more complete top from Corinth (Figure 22r), with a similar tall neck to Fig. 22p and the typical (string-) cut grooves of Agora M 235, and is indeed classified as an Agora M 327 (i.e. M 235) (Slane and Sanders 2005: 262–263, fig. 5, Assemblage 2, 2-18). The Butrint piece Figure 22q is even later in date and it is possible that these vessels are from another south Peloponnesian source, perhaps not so distant or the same as that of Agora M 315. Also with likely Peloponnesian origins and with long distance exports in the mid-3rd century,51 as we have seen (Figure 23), the oil amphora par excellence on sites in Byzantine Greece and the Aegean, the Lower Danube and Crimea in the 5th and 6th centuries was the LRA 2A-B and in the 7th century, its developed version, Pieri LRA 2C (‘LRA 13’, ‘LRA 2/13’) (Figure 24a-c, possibly d, for Argolid products).52 Though LRA 2 is usually labelled an ‘Aegean’ amphora, the classic early Byzantine LRA 2A, of the late 4th-5th centuries, is more likely to have been a product of the Greek mainland, in the Peloponnese and Boeotia especially, but not the Aegean, until the 6th century. The demands of the state for alternative sources and stricter control of the annona for the Danubian troops and Gothic foederati following the serious devastation of the late 5th century (Poulter 2007a-b; Liebeschuetz 2007: 107; Heather 2007: 173–188) led to the reorganisation by Justinian of the supply, and, in theory, also the expansion of production, of olive oil and other annona goods, including wine in LRA 1, through the creation of the new office of the quaestura exercitus (Justinian, Nov. 41:  Opaiţ 2007a; Pieri 2005a: 85, for the 3rd century form (fig. 42, right, a complete example from Dobrougea; fig. 43, middle, the complete Athenian Agora piece illustrated by Grace [1979: fig. 37], from the AD 267 Herulian sack material, though, like the Tunisian amphora and small Kapitän II beside it, it may date c. 300? (there are a number of groups of redeposited sack material with later pieces in the Agora). 52  The 7th century evolution of LRA 2 is commonly known as LRA 13, as well as LRA 2/12, though Riley’s type comprises three quite different (1979: 231–232, fig. 93, no.373, a ‘classic’ version, fig. 94, no.374, with a ribbed neck, fig. 94, no.375, probably not in this series, with an ovoid body and arched handles, a Cretan or early Medieval amphora?). It is difficult to know what to call the wide range of variants of the period. Following Frederick Van Doorninck, it is perhaps more correct to refer to Pieri’s sequence of forms, hence name them variants of LRA 2C than refer to LRA 13 (Pieri 2005a: 86–93). 51

332

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

a

b

c

a-b) Athens. Agora P 25618, views of both sides (straight walled; more globular). Context B15: 5. Post-AD 267 sack debris (Photo: P. Reynolds)

c) Athens. Agora P 25617 Context B15: 5. Post-AD 267 sack debris (Photo: P. Reynolds)

d f

d) Rome-Testaccio AD 254. Proto LRA 2 (Marimon Ribas and Puig Palerm 2007: fig. 79.1)

e

f) Pagasitikos Gulf Wreck 7 Early LRA 2. Late 4th c.? (Demesticha and Spondylis 2011)

h

g

VP 7133.1

e) Dobrugea Early LRA 2. 3rd c. (Pieri 2005a: fig. 42, after Opaiţ 1980: pl 4)

i

VP 7224.1

VP 2092.2/ VP 7026.1

g-i) Vrina Plain (Butrint) 4th c. precursors of LRA 2. South Peloponnesian? (Reynolds 2020 figs. 3.16.1, 3.12.8 and 3.15.14)

Figure 23. a-i. Early versions of LRA 2, 3rd to 4th century.

333

Paul Reynolds

Argolid

Argolid

b

a

Source?

c d

Angular grey and red flint a) Butrint-Triconch But 5014.59 Early 6th century

e

i

Local chert

Corfiot? Epirote?

Organics: Aegean?

g

f

Epirote?

j

Epirote

h

b-h) Nicopolis Early 7th c. versions of LRA 2 (Reynolds and Pavlidis 2014)

Cypriot?

l k

i) Butrint-Triconch LRA 2.6 (cf. Agora M 235). But 5201.3 Mid/late 6th c.

n

k) Beirut. LRA 2C BEY 045 507.1/ 503.76-77. Early 8th c. (Reynolds 2018: fig. 50)

j) Amathus LRA 2 (from Demesticha 2003: fig. 4)

m

Halisarna workshop (Kos) l) LRA 2A and m) LRA 2C (from Diamanti 2010: figs. 84 and 99)

o

n-o) Yassi Ada LRA 2C. c. AD 625 (Bass and Van Doorninck 1982: fig. 8-5, CA 13 and 20

Figure 24. a-o. LRA 2 and later versions (Argolid and other sources).

334

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

de quaestoribus id est praefectis insularum). Based from 536 at Odessus/Varna, at the mouth of the Danube, the command included the supervision of the provinces of Caria, Cyprus, the Cyclades, Moesia and Scythia (Reynolds 2010a: 266–267, 271, for general comments; for lead seals associated with the quaestura, see Curta 2002, with useful bibliography). We know from the sequence of statue bases of the 3rd to early 5th century erected in the principia of Novae (Moesia II) and Oescus (Dacia Ripensis) that previously, up to the 430s (the last three at Novae date to 430, 431 and 432), the supply to the Danubian army was organised by primipili, with interesting references to the specific regions of supply they were involved with (Ephesus-Asia, the Islands, Syria Palaestina, i.e. probably Caesarea) (Rizos 2015).53 This would seem to provide a context for the Abydos tariff inscription, some 50 years earlier than the quaestura (pre-Anastasian reform, c. 470), which refers to the special tax benefits of the Cilician merchants transporting wine to Constantinople, acting on behalf of the state (Guillou and Durliat 1984; Iacomi 2010). Justinianic period lead seals (82 specimens) have been found specifically in Scythia Minor and Moesia Inferior (with most at Tomis/Constanţa) but rarely elsewhere on the Danube (and not in Chersonesos), over half bearing the names of individuals, probably merchants, with 13 others being seals of official documents, 9 of these being of functionaries (including eparchs, consuls, a secretis, chartularii), working under Justinian. The last mention of the quaestura in the law codes was in 575 (Justin II, nov.162) (Curta 2002; see also Chiriac and Munteanu 2014, for 59 4th century seals from the Lower Danube, linked to many cities of Asia Minor — e.g. Smyrna, Ephesos, Hypaipa, Koloe — and Pamphylia, which they date to the campaigns of Valens against the Goths in 366–369). Given the long evolution of the supply of state-army annona goods, clearer chronologies and more secure identification of sources for the wide range of LRA 2s of  Oescus: [Lib]ero patri | conservatori | dd(ominorum) nn(ostrorum) Augg (ustorum) | Fl(avius) Zosimus p(rimi)p(ilarius) || ex provincia Asi | a civitate Efisia | norum votum | posuit | [p] os(t) pastum militum. (To Liber Pater, protector of our lords the Augusti. Flavius Zosimus, primipilarius from the province of Asia from the city of the Ephesians, dedicated this offering after fulfilling the service of provisioning the soldiers). Oescus: [---d(ominorum) nn(ostrorum)] | [Au]gg(ustorum) Fl(avius) | Tatianus | [pr]imip(ilarius) ex pro|(vi)ncia Suria [Pal]estina || [po]st pastum | militum | votum po|suit (branch with five leaves) | Ἀντωνίνου. ([…] of our lords the Augusti. Flavius Tatianus, primipilarius from the province of Syria Palaestina, set up this offering after fulfilling the service of provisioning the soldiers. (By/Of) Antoninos). Novae: FF(lavii) Agapitus | et Diogenis pp(rimipilarii) ex pro[v(incia)] | insulanea pasto mili|te statuam eregeru||nt in vulto Dionisisi [sic] ind(ictionis) | xiiii in conss(ulatibus) Anthioci [sic] | et Bassi vv(irorum) cc(larissimorum) (The two Flavii, Agapitus and Diogenis, primipilarii from the province of Insulae, after provisioning the soldiers, set up a statue in the presence of Dionysius, in the fourteenth indiction year, during the consulship of the viri clarissimi Antiochus and Bassus [=AD 431]). As Rizos notes (Rizos 2015), though Oescus lay beyond the command of the Praetorian Prefect of the East, its location was difficult to access from Illyricum (by the Prefect of Illyricum) and was easier to supply from the Black Sea-Lower Danube. 53

the early, mid- and late 6th century are essential for a better understanding of their actual development across Greece and the Aegean. Though there is some confusion over the sources of the earliest 4th-5th century phase, the existence by the late 6th or early 7th century of completely new workshops of LRA 1 and LRA 2C in the Aegean is irrefutable and an important development that would seem to reflect the organisation of the quaestura exercitus: these new amphorae began production after that final legal reference to the quaestura. The in situ finds at Dichin and Capidava (see below) perfectly illustrate the military destination of annona goods. For the first phase, relatively few production sites of LRA 2 are known, including a 4th century workshop at Delion/Dilesi (Boeotia) (Gerousi 2014), and several in the Argolid, one of the earliest being at the port of Halieis/ Porto Cheli, in the Argolic Gulf. Trends indicated by Abadie-Reynal at Argos are nevertheless surprising: LRA 2 ‘appears timidly in the 5th century’, to become during the 6th century the most common type on the site (28% of the all 6th century amphorae) (AbadieReynal 1989: 51–52), a point reaffirmed by Sarri (2014: 66–68, figs. 6-10) in her publication of a major group of LRA 2s from the destruction of a bath building in Argos. One would have expected a stronger presence from the early 5th century, if not before. LRA 2 (and Agora M 235) production needs to be carefully plotted with respect to their proposed production regions, their quantification and patterns in the date and distribution of oil presses (in the same way this has been done for Tunisian amphorae). Olga Karagiorgou led the way with her pioneer work on the distribution of LRA 2 but similar documentation is needed across Greece. There is as yet, as far as I know, no published evidence for workshops on Chios, though this is the source attributed by Paul Arthur (1998: 168–169) to one of the most common late fabrics of LRA 2 and LRA 2C (with exploding lime, as ‘Argolid’ products, but with additional gold mica flakes): common in mid- to late 6th century Butrint, less common in 6th- century Lechaion and 6th-early 7thcentury Nicopolis, but apparently that of the majority of LRA 2C in the mid-7th century destruction deposits of the Byzantine fortress of Emporio on Chios (Boardman 1989; according to Arthur 1998: 168–169).54 LRA 2s in  Note that the drawing of the Emporio LRA 2C no.236 has the correct height, but far too wide a rim and neck in comparison to the photograph on Plate 24, where the amphora has more typical proportions (I have therefore not reproduced this drawing, which is the only one in the publication). As yet, no clear production sites of LRA 2 have been found in Chios, though, as Dominique Pieri underlines (Pieri 2005a: 90–92), this has often been assumed (by myself included) because it was thought that the workshops discovered by Tsaravopoulos (1986) were for LRA 2, whereas they are actually for the much earlier form Zeest 90/Dressel 24 similis (Opaiţ and Tsaravopoulos 2011, who curiously make no reference to this problem). This said, two newly discovered workshops near the harbour of Limia at Volissos (NW Chios) and at Komi (SE Chios) have been signalled by Theodoulou et al. (2015: 51, with reference to information from D. Tsardhaka of the Chios Ephorate of Antiquities). 54

335

Paul Reynolds

5th-6th century contexts in Lechaion, the western port of Corinth, are, in contrast, almost entirely in a dense fabric like that of the Argolid, but lacking the large lime lumps and surface ‘explosions’ of the latter. The first fabric may correspond to Paul Arthur’s observations of the Porto Cheli fabric; the second, to ‘Argolid’ products in Slane and Sanders 2005, so one wonders where exactly in the Argolid these, the most common exports, were produced, and we return to the problem of Chios. The fabric of the illustrated example of the early 6th century LRA 2A from Butrint (Figure 24a) had a pale red core, yellow ochre surfaces, with lime but less evident mica. It may be no coincidence that forms related to Dressel 24 and Zeest 90, with their cup-shaped rims, are typologically close to LRA 2, whose role as an oil container is suggested from dipinti and other evidence, such as presence on oil press sites or likely oil-filled examples that exploded in the fortress fire at Dichin (Swan 2004, 2007). We may note that in Lechaion, where 1000s of LRA 2 fragments were found in the relatively small Late Roman areas excavated, the building of Area A produced organic evidence for the industrial processing of olives in one of the rooms in the 5th century, perhaps associated with numerous funnels (in LRA 2 fabric), a re-used amphora base and a plasterlined vat.55 Archaeo-botanical evidence, plant remains as well as pollen, can offer us a most valuable resource for our understanding of the local environment and the foods consumed as well as grown locally. The early Byzantine coastal villas of Opountian Locris on the mainland fronting the island of Euboea, provide evidence of oil presses and silos (Zachos 2018: 46–47). Fourth to 6th century LRA 2s are also predominant on rural and urban sites in Chalcidice (e.g. Vasiliou and Tsigarida 2014) and Boeotia (Dafi 2014), where production of what could be considered the combed-bodied Medieval descendants of the LRA 2, Günsenin 2 and its successor Günsenin 3, has at last recently been identified at Thebes and the port of Chalcis, this region being known for exports of wine and oil in the Medieval period (Waksman et al. 2016).

I hope that we soon learn more about these key sites and their ouput. Paul Arthur’s description of the ‘Chian LRA 2’ fabric (presumably of these Zeest 90s: Arthur 1989: 82, note 2), described it as the ‘standard’ LRA 2 fabric, with a little mica and lime eruptions on the surface. His description of the fabric of the LRA 2 Kounoupi workshop material at Porto Cheli says that ‘Some of the fabrics appear macroscopically to be rather similar to the Chian fabrics, though do not appear to have the same amount of mica or the lime eruptions characteristic of the latter’. Colour photographs of samples of the Chian Zeest 90s, showing the section break (but unfortunately not the surface) appear in Opaiţ and Tsaravopoulos (2011: 302, with petrology by David Williams, p. 293). 55  The 2018-2019 excavations of the American School-California State University at Lechaion, directed by Paul Scotton. The archaeobotanical remains, over a hundred olive pits, were studied by Chantel White (University of Pennsylvania). The ARS fine wares include many examples of ‘late Hayes 61B’ of the mid-third quarter of the 5th century. In the building to the north-east, Area B, a room with a large number of redeposited LRA 2s was found (no rims). For a recent overview of the archaeo-faunal record from a selection of urban and rural sites across the Mediterranean, including Corinth, see MacKinnon (2019).

Shipwreck 7 of the Pagasitikos Gulf, thought to date to the late 4th century, carried a cargo of an unclassified widebodied amphora with a wide tronco-conical neck (fruit? olives?) (Figure 25g) and early LRA 2 (Figure 23f), with rarer variants of Agora M 273 (Figure 6c) and a Palestinian LRA 5 (Demesticha and Spondylis 2011).56 This shipment may have been heading north, to Thessalonika or beyond and could be seen within the context of annona cargoes of the province of Achaea which, as an inscription of AD 401/402 found at Megara indicates, were gathered at the two central horrea bases, at the ports of Kenchreai and Skarpheia in Epicnemidian Locris, serving the Peloponnese and eastern mainland Greece respectively (see Map, Figure 26a).57 In the Lower Danube oil for the troops had to be imported because the climate was not favourable for olive oil production (Figure 2). At the fort of Dichin in the 480s a warehouse caught fire and amphorae stored on the first floor came crashing down. The two (Argolid) LRA 2s, presumably carrying oil, exploded, whereas the LRA 1s did not, presumably because they carried wine.58 There were also Danubian versions of LRA 2 in the warehouse, possibly carrying some form of local oil alternative.59 Examples of LRA 2 of similar date and into the first half of the 6th century have been published from Novae (Biernacki and Klenina 2015: 105, figs. 6-7). A second phase of expansion in production can be identified in the appearance of the ‘Chian’ (lime eruptions and mica flakes) fabric and other new products in Butrint in contexts of the second quarter of the 6th century and the third quarter of the 6th century (related to the quaestura exercitus).60 Other, even more common products in Butrint in these contexts (Figure 24i) are small amphorae with a cupped rim (classified as ‘LRA 2.6’ in Butrint), with a non I would like here to thank Stella Demesticha for sending me a good quality pdf of this paper. The scale of the published drawings was at 1:8. 57  Trombley 1989: 217, who also comments on the western Boeotian port of Vathy and its direct access to the Corinthian Gulf in late Antiquity (pp. 219–220); Rizos 2015: 296–298. An inscription of c. AD 402 (by the augusti Arcadius, Honorius and Theodosius) , in several cities of the province of Achaea, surviving at Corinth and Megara, provided schedules for goods to be sent to the praepositura horreorum, with Boeotian, Euboean and Aetolian cities delivering at Skarpheia, while the Peloponnesian ones at Corinth were collected at Kenchriae, and surely also at its western port at Lechaion, where major harbour facilities, including a lighthouse and warehouses are coming to light through recent surveys on land and sea. Another state warehouse further up the Corinthian Gulf has recently been found at Zahloritika, south of Aegion, and has been dated to the Justinianic period (Didioumi and Sdrolia — forthcoming —, with reference to Koumousi and Theodoropoulou (Theodoropoulou et al. 2015: 115–117). There was another base at Maroneia (Hagios Charambolos), in Thrace. We should not forget here the key role of the imperial capital and papal vicariate of Thessalonika in these distribution networks (Jones 1966: 254–255). As an indication of their importance, both Achaea and Asia were, like Africa, the only provinces governed by a proconsul (in the Notitia Dignitatum) (see Figures 26a-b). 58  Swan 2004 and 2007. For the final publication, see Swan 2020. 59  Swan 2004: figs. 6, 7, 9 and 10. 60  See Reynolds 2004a: 229–234, Appendix 1, contexts 1152 (c. 525-550) and 1676 (550-580: no ARS 105s), for fabrics and quantities, though these will need to be revised for the final publication: under Fab 15, with the exception of Fab 15H, are the ‘Epirote’ products, with LRA ‘Chian’ (lime and mica: could also be Argolid) and a LRA 2 fabric with large lime (sometimes sunken into the surface), separately quantified; Fab 15H is thin-walled, rather dense, with poor, smudged grooves and is more common in But 1676 than But 1152; see also Reynolds 2020). 56

336

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

a a

Corinthian? S Epirote? (Nicopolis?)

Corinthian? S Epirote? (Nicopolis?)

b

‘Lakonian’

d

c

Peloponnesian

ee f

Nicopolis Agora M 325 Early 5th c. (Reynolds and Pavlidis 2017: fig. 8.3)

Argos. Agora M 325 Early 5th c. (Ivantchik 2002: Type 8, fig. 18.129)

Corinth (Slane and Sanders 2005: fig. 8, 3-24) End 6th century

Argos. Agora M 325 Early 5th c. (Ivantchik 2002: Type 8, fig. 18.130)

Corinthian

k

g

Pagasitikos Gulf Wreck 7 Type 1. Fruit? Olives? (Demesticha and Spondylis 2011). Height: 70 cm

h

Argolid

Lechaion. ‘Fruit amphora’. Context 18Ba 02 004.3420 Mid-6th century

Corinth (Slane and Sanders 2005: fig. 11, 4-14) Mid-7th c. or 650-675

Athens. Agora P 13172 Agora M 315 type. 5th c.

Corinthian

i

Corinthian

j

Corinth. ‘Fruit amphora’ ‘Gritty Corinthian fabric’ Height: 80 cm (Slane and Sanders 2005: fig. 13, 4-16) Mid- 7th c. or 650-675

Corinth. ‘Fruit amphora’ (Slane and Sanders 2005: fig. 11, 4-17) Mid-7th c. or 650-675

Corinth. ‘Fruit amphora’ (Slane and Sanders 2005: fig. 8, 3-27) End of 6th century

Figure 25. Various Peloponnesian types, 5th to 7th century, a-c. Agora M 325, d. ‘Lakonian’ Agora 315, e-f. Corinth, Slane and Sanders 2005: fig. 8, 3-24 and fig. 11, 4-14, g. Pagasitikos Gulf Wreck, Form 1, h-k. Corinthian-Argolid ‘fruit’ amphorae (wine?).

337

Paul Reynolds

globular, rather ovoid body, ‘string’ grooving and small button base that recall the shape and manufacturing techniques of Peloponnesian Agora M 235 (in fact, have I been wrong all along, and this was the model, not LRA 2?: see for example Figure 22e f or k). A vessel of this type was found in situ in the horreum of the fortress of Capidava, along with two LRA 2s, two LRA 1s, a LRA 3, and many Gazan and Pontic amphorae (small Opaiţ B, especially), in destruction material associated with a ‘hoard’ (a purse probably) of folles suggesting a date in the 580s, this event preceding the final phase of building on the site (Opriş and Raţiu 2016; Gândilă 2009). It would seem that the quaestura exercitus was still functioning. The Butrint ‘LRA 2.6’, however, was established earlier, in the late 5th century (larger examples). Indeed, the large numbers and calcareous fabric with very fine chert suggest a source in Epirus, relatively close to Butrint (for their petrology and comparison with rather similar 3rdcentury Zemer 57 — a Cretan shape — also extremely common and with a macroscopically similar fabric, see Fantuzzi 2020). In theory, these Epirote amphorae, LRA 2 or Agora M 235 imitations, carried local oil. Indeed, this region, as well as Corfu and the Ionian Islands to the south, stand out on the 1936 map of oil producing regions: Figure 2). Perhaps the 6th century large, globular-bodied amphorae in chert fabric found in Butrint and the villa of Diaporit, possibly to be associated with flanged lids in the same fabric, were for oil (Reynolds 2020: fig. A.16, f-h), as also a common early 5th century form with cupped rim in buff fabric at Nicopolis, possibly derived from Agora M 235 (Reynolds 2020: fig. A.16m-n). In Corinth, another series with a similar top to the ‘Epirote’ LRA 2.6 appears in the late 6th (Figure 25e: close in shape to LRA 2.6) and in the 7th century (Figure 25f), in the latter case with a quite distinctive tall, narrow body and domed base, perhaps following Lakonian models, i.e. like those of the 5th century one-handled amphora typical of Sparta, Agora M 315, M 336 and P 13172 (Figure 25d), a type that it extremely common in the Athenian Agora and in Argos.61 A content other than oil is likely in this case (wine?). A third phase of late 6th-7th century production associated with LRA 2C and imitations of LRA 1 has now be identified on Kos (Figure 24m) and Paros, with imperial control stamps (the bust of the emperor and the monogram of an official eparch or kommerkiarios) on some examples of variants of LRA 2C: 20 on Kos itself, one on the c. 625 Yassi Ada ship, one in the Imperial palace of Constantinople, 5 in the Athenian Agora and single examples on the island of Geronisos, in a monastery in the Kellia, and in Tocra (Diamanti 2010, 2012; Poulou Slane and Sanders 2005: 271, 276, 287, Assemblage 3, fig. 8, 3-23 (greybrown ware) and 3-24, Assemblage 4, fig. 11, 4-13 and 4-14. The authors refer to parallels at the site of Berbati (Argolid). For examples of Agora M 315, in a number of variants, see Ivantchik 2002: 392–393, figs. 23-25, nos144-153. 61

Papadimitriou and Didioumi 2010; Diamanti, Kouzeli and Petridis 2014; Didioumi 2014; Opaiţ and Diamanti 2014). We may note that a single example of LRA 2A in the same fabric as the LRA 2C also appeared at the Koan workshop (at Halisarna), evidence for earlier production (nearby?) (Figure 24l). Charikleia Diamanti has also begun to demonstrate how not just Constantinople, but also the troops and sailors on the Cyclades received annona goodsrations (including oil in LRA 2C and wine in local LRA 1s) carried in these Aegean amphorae (e.g. Despotikó, with Parian amphorae of the late 6th century to 700+; Kythnos, with Koan LRA 2C; Naxos, with Parian and Argolid imports; Thera, with Parian LRA 1s).62 The Crimea in the late 6th and 7th centuries was another target for these and other supplies (especially wine in LRA 1 — originals and Aegean imitations —, and extraordinary numbers of Gazan amphorae and the narrow-bodied Cretan TRC 4) (Sazanov 2000, 2007; Smokotina forthcoming). With the loss of much of the lower Danube by the end of the 6th century, and its remaining coastal ports by 615-617, as well as much of Greece to the Slavs and other incomers, during the reign of Heraclius (Liebeschuetz 2007), and then Antioch and the Levantine provinces to the Arabs (634-637), as well as Egypt and Cyrenaica (640s) and the rest of North Africa by the 660s, alternative sources of supply for Byzantium were critical. This went parallel with a shift from tied navicularii to private, free naukleroi in the 7th century, with the application of new sea laws (the Nomos Naukleros) probably at the same time (see Lopez 1959). Cyprus now turned to the production of LRA 2C, where the form was quite alien to the island. These have characteristic separate bands of combed decoration (Figure 24j), several examples appearing in a much later, early 8th century context in what was Umayyad Beirut (Figure 24k: Reynolds 2003b).63 Though we usually associate Cyprus with wine, the island had had a long history of olive oil production from the Bronze Age onwards, exports of the 5th to 3rd centuries BC being carried in the so-called ‘Cypriot basket handle amphora’.64  A paper given at the British School at Athens, January 2020, in the symposium organised by Maria Duggan and Evangelia Kiriatzi: Distant Seas, Connected Worlds: Tintagel, Britain and Greece in late Antiquity. 63  See Demesticha (2003), for kiln sites in Amathus, Zygi and Paphos. These, I understand, are not independently dated. For John Hayes’ comments on Cypriot LRA 1s and LRA 2Cs, see Hayes 2007: 437–438; For dated examples, see Hayes (2003): a single example from mid7th/AD 654? Deposit 11, not a piece Hayes describes as local, just in a ‘variant’ light orange fabric; and several, well preserved, more likely Cypriot examples from late/end of 7th century Deposit 14. For similar possible Cypriot LRA 13s and Egyptian versions (Egloff 167) in 8th century Umayyad Beirut, see Reynolds (2003b; 2018). The fabric of the illustrated example from the Imperial baths still functioning in the Umayyad period (Figure 24k) is close to that of FAM 169, that of Virginia Grace’s supposed ‘Pamphilian’ amphorae in 3rd century Beirut contexts (see Figure 5a-e) amphorae: Gritty fabric, with moderate 0.5 mm rounded dark red brown material and laminar dark grey-black inclusions (pyroxenes?). Fired yellow buff, with a very pale red core, inner surface more yellow ochre. Surfaces smoothed over but rough and gritty where disturbed by combing. 64  Hadjisavvas 1992. Oil-pressing installations are well attested all over Cyprus. These indicate a long history of oil production from the Hellenistic to Byzantine periods but the sites remain for the most part 62

338

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

aa

b

Figure 26. a. Map of the Dioceses of the East, c. AD 400 (from Wikimedia: Cplakidas 2007), b. Map of the provinces and Praetorian Prefectures of the East, c. 400 (from Wikimedia).

339

Paul Reynolds

A proposed production on Samos needs also to be verified.65 However, the composition of the cargo of the early 7th century Yassi Ada ship (Leidwanger 2014), comprising LRA 2Cs primarily in micaceous Samian or Meander Valley-Ephesian fabrics and LRA 1s, some probably still Cilician, but with micaceous Samian or Meander Valley-Ephesian examples also,66 as well as the aforementioned stamped Koan vessel, tells us of the shift in production by c. 625 and the likely key role of Ephesus, seat of the governor of Asia and capital of the Diocese of Asiana, as well as the ecclesiastical see of Asia, in the consumption as well as redistribution of goods (Maps, Figure 26a-b).67 This was a restricted, Aegean market. These micaceous LRA 2Cs and LRA 1s have not been found in Butrint or Nicopolis (and are also absent in Corinth?). In the case of Nicopolis (the capital of the province of Epìrus Vetus), an early 7th century deposit in one of the most important ecclesiastical residences of the city (the ‘House of the ekdikos Georgios’) comprised a large number of amphorae from the ‘orbit’ of Ephesus and Samos (see above, Figure 6o, p, v-w), together with a more classic version of LRA 2C in the same fabric and grooving as the Butrint ‘Epirote LRA 2.6’ (Figure 24g), and standard ‘Argolid’ fabric (Figure 24b-c; d?), porous whitish-buff (Corfiote?) (Figure 24f) and local-Epirote coarse chert (Figure 24h) versions, and a possible Aegean piece tempered with organics (Figure 24e), demonstrating the wide range of Epirote-close regional as well as imported oil in this period (Reynolds and Pavlidis 2014). Like 6th century military sites on the Danube, these were accompanied by LRA 1s in almost equal numbers, together with Tunisian amphorae (Reynolds and Pavlidis 2014: figs. 5-6 and Table, fig. 13). In Butrint, LRA 1s and LRA 2s also dominate assemblages of the 5th to early 7th rather imprecisely dated. The Cypriot ‘basket handle’ amphora was certainly for carrying oil: it has been found in situ by Hellenistic presses (Hadjisavvas 1992). From observation of stratified examples in Beirut, it is clear that the form was just as prevalent in fabrics that indicate its production in Ras al Basit, Amrit and the north Lebanese coast (probably Byblos). Note that the major ecclesiastical oil processing installations excavated at Salamis date to the 7th century (Argoud et al. 1980; Diederichs 1980; Hadjisavvas 1992: 45). 65  There appears to have been a production site at Perre (southern coast of Samos) (Pieri 2005a: 91, note 132). The twenty LRA 2s found in the 7th century phase of the basilica complex installed in the baths of Samos (Steckner 1989), like the 120 ‘LRA 1’s in the adjacent room (note that the single example drawn, with its raised handles, looks more like a Cretan amphora than LRA 1), associated with a press and dolium, were not stated to have been produced there (for a summary of these and other finds, see also Karagiorgou 2001: 11–12). 66  Leidwanger (2014) here presents petrological work on a number of the LRA 1s, with some comments on the LRA 2s. I also had the opportunity to examine the amphorae and was struck by the predominant number of unusual LRA 1s (in terms of fabrics and morphology), micaceous LRA 1s and LRA 2s (from the same source: X8, with spaced narrow bands of combed grooves; UN8, very micaceous; UN9: W8: wide grooved bands). I noted one Cypriot example, like the products of Amathus (Y200). One perhaps Koan (W4: fine red-brown ware, with fine micas on the surface and occasional lime). No obvious Argolid examples? 67  Horacio González presented a major context of these micaceous LRA 2Cs in Ephesus at the International Conference held in Berlin (October 2017): The 8th century: Patterns of transition in economy and trade throughout the Late Antique, Early Medieval and Islamicate Mediterranean. For him, these are Ephesian, not Samian amphorae. It is curious that contemporary Samos cistern amphorae were not a part of the Yassi Ada cargo.

centuries, with a host of Aegean Agora M 273s and Samos Cistern amphorae, and lesser Tunisian and Palestinian amphorae (Reynolds 2004a, 2010b). Spectacular in situ finds of the first half of the 8th century versions of ‘globular LRA 2/13’ at Gortina, possibly due to an earthquake, have been published by Costa (2015: 714–716, fig. 3), all bearing spaced grooved bands (like the earlier examples on the Yassi Ada wreck and Cypriot LRA 13s of the same 8th century date in Beirut). Here an Aegean or Cretan origin for these amphorae is assumed. As usual, the role of the Church in the production and movement of goods for its own purposes (and they could have in addition sold surpluses to the state) is always a question to reflect on (Bernal‑Casasola 2010). The storerooms of the ecclesiastical complex in the baths at Samos (built c. 565-578) (see note 65, below), in the rooms around the portico of the domus of the ekdikos at Nicopolis, the dolia (?) in the episcopal quarter at Byllis (Beaudry et al. 2003: fig. 9, for a plan), the 7th century oil press room in Salamine (Cyprus), perhaps the Yassi Ada ship, are ecclesiastical and the Church had access to the same type of goods in amphorae that were supplied to the army. We must not forget that the huge numbers of LRA 2s on sites across Greece, e.g. those in Butrint or the industrial buildings at Lechaion, are not annona goods, just (oil) surpluses that would have been bought on the open market (as too were the Dressel 20s found in NE Spain or Narbonne). It is important to note, however, that the Levantine ports, such as Beirut and Caesarea, were largely off the circuit of LRA 2A-C.68 Their absence in Syria, well exemplified by the supply to Zeugma, Homs and Apamea, is also pretty striking. This can be contrasted with finds of not insignificant numbers of LRA 2 (70 MNV) with LRA 1 (356 MNV), Agora M 273 (113 RBH), Gazan amphorae (38 RBH) and LRA 5 (39 RBH) in a 6th- century levelling layer by the horreum of Andriake, the port of Myra in Lycia (Lemaître and Vener-Marksteiner 2018), and is a good indication that the port served as a transhipment point for traffic heading for the Aegean or the western Mediterranean, sailing out of Antioch? The scarcity of Sinopean and African amphorae here is notable. For the Aegean, the newly documented shipwrecks off Chios are illuminating: two with LRA 2 cargoes, two with LRA 13/LRA 2C cargoes and one with a LRA 1 cargo (and additional ‘Samos cistern’ and late Sinopean amphorae) (Theodolou et al. 2015). Quantification (which certainly needs to be updated) for the regular but small scale supply of LRA 2 to ports in the West such as Ravenna, Carthage, Naples, Marseilles, Tarragona, the mid- 6th century emporium at Benalúa (Alicante) (Reynolds 1995: Appendix C.1; 2010a, 2015) or  In the AUB Anglo-Lebanese excavations, there were only 15 diagnosticsvessels (according to RBH) of LRA 2 and an additional 45 wall fragments, so a maximum of 60 vessels if wall finds are included (contemporary and residual finds in Medieval contexts). This, in contrast to 5732 LRA 1 RBHS (4736 maximum vessels). For quantification from contexts of AD 551 and the late 6th century, see Reynolds 2010a: Table 24. 68

340

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Byzantine Cartagena (Reynolds 2011), indicates regular but low numbers, more indicative of returning western (Tunisian or Italian) ships or additional goods loaded on outward eastern ships at transhipment points such as Andriake (Reynolds 1995; 2010a: Tables 20, 22 and 23). The supply of non-Epirote LRA 2s at Butrint (c. 326 RBH, 498 RBHS) and Nicopolis can be taken as normal for Greece (or can serve as a model for comparison) that may be contrasted with Ravenna. Figures at Egnatia (Apulia) (1265 RBHS) seem to mirror those in Epirus, across the water, and the unusual numbers in Cyrenaica could well reflect shipping coming down from Crete. The presence of c. 22 LRA 2 amphorae and 80 LRA 2 lids on the Marzamemi ‘church’ wreck, off Sicily, is a good example of how these marble products accompanied other quite distinct primary cargoes, in this case a shipment probably heading for Justinianic Carthage (Leidwanger 2018). The supply of LRA 2 to Tintagel, in Cornwall, and other sites in SW Britain over the period c. 450-550 (an ever increasing number of LRA 2s from the recorded 61 examples in Duggan 2018), marked by accompanying, but actually fewer LRA 1s, is a major contrast with respect to trends in western Mediterranean ports and even with the emporium at Vigo where Phocean LRC, Aegean and Limyran cooking pots and Lycian unguentaria and low numbers of LRA 2, and raw glass, inter alia, follow Benalúa-Alicante, the initial landing point of eastern ships prior to their onward journey to Vigo and Britain (Fernández 2014). Tintagel was clearly the desired final destination of these (Argolid) LRA 2s and the oil they carried, these (in my opinion eastern ships) returning with much needed tin as the primary cargo. Finally, with respect to the amphorae in the Peloponnese, two other important types should be mentioned. Agora M 325 is a 5th century amphora with ovoid body and small indented-domed base that could be southern Epirote (Nicopolis?) (or perhaps Corinthian), given its coarse, reduced-fired, chert-rich fabric, similar to that of Epirote cooking wares (which share the same indented base in this period). It is well represented in Athens it seems (Robinson 1959: 115, Plate 32, with several more on display), and is certainly so in Corinth (Slane and Sanders 2005: fig. 3.127, AD 400-450; fig. 5, 2-31, AD 450-500), Argos (Figure 25ac: Ivantchik 2002: 383, 388, Type 8, nos.129-131), Lechaion and Nicopolis, in this case dated to the early 5th century (Figure 25c) (Reynolds and Pavlidis 2017: fig. 8.3, one of three in the deposit). The contents need to be determined, but this may be wine, given the shape which appears to imitate Cretan forms (cf. MRC 3 and TRC 4). A second group of amphorae of notable large size and with wide necks (more than one size is registered) is particular to Corinth and the Argolid, appearing in both general fabrics (Figure 25 h-k). They have been classified as fruit amphorae. This is of course an attractive proposition, but if fruit, surely it would only be logical to package them in amphorae, especially huge ones, if they were preserved in a sweet liquid of some kind (e.g. cherries).

Typologically, the possibility they carried fish products (cf. Dressel 21‑22) needs to be considered and tested through organic residues analyses. However, definitive evidence for their use has been found in excavations inside a tower of a farmstead at Pyrgouthi, in the Berbati Valley, in the Argolid, destroyed in the 6th century. Here one of several of this type on the site was found in situ in the tower, with a treading floor in one corner and next to what would have been a small wine press, these vessels presumably being used to store wine during the fermentation process (Hjolman 2005: especially p. 161, vessel nos.33-37). In this respect, given the large body size and wide neck, one is reminded of the trade in Roman and Ottoman ‘jars’ in northern and southern Lebanon. 8. Tunisian oil and other exports in the East Finally, extremely rare in the East in the 1st and 2nd centuries, Tunisian amphorae (oil, fish) appear in the early 3rd century (e.g. Beirut: Reynolds 2010a: Table 4b), to perhaps drop in the mid-3rd (Beirut: Reynolds 2010a: Table 4b; I saw none in the Athenian Herulian sack material), then appear regularly from c. 300, testimony here being the various shipwrecks recently discovered in the waters of Euboea that would have been heading north for Thessalonika or Constantinopole (Koutsouflakis and Argiris 2015). Fourth to late 6th century Tunisian amphorae and African Red Slip Ware are regular finds on sites in the East, perhaps more than we originally thought (see Bonifay 2005, for a first overview of the distribution of the amphorae, with references to sporadic finds; Opaiţ 1997‑1998 and 2004a, with Tomis documenting the largest African amphora assemblage of the Black Sea, particularly for the second half of the 5th to early 6th century). In Alexandria, following the high numbers of spatheia Keay 25 (wine and fish) during the 4th-mid 5th century, when Egypt was still engaged in the supply of grain to Rome, Vandal and Byzantine Tunisian imports were scarce indeed (Bonifay, Capelli and Şenol 2015: preVandal Keay 27 and Vandal forms Keay 36 and 41, all probably oil amphorae). A very similar trend in Tunisian amphora imports is found in Beirut with few imports from c. 450 to the early 7th (Reynolds 2010a: Tables 9b, 15 and 24). This may be contrasted with the definite stronger Vandal-Byzantine supply to the Aegean, lower Danube and Crimea (Reynolds 2010a: 855). Some would have accompanied the shipments of Tunisian grain to Constantinople, other administrative hubs of the annona network in the East (Athens-Piraeus, Antioch, Caesarea Maritima, Alexandria, Ephesus, Thessalonika), the latter at the end of the fortified Via Egnatia running across northern Greece from the Adriatic port of Dyrrachium/ Dürres, with its own defensive castra,69 with opportunity  For finds of the second half of the 6th century at the coastal castrum of Velika (Thessaly), identified as the Justinian and earlier fort of Kentavropolis, with abundant LRA 1s and LRA 2, as well as African amphorae, see Didioumi and Sdrolia (forthcoming). 69

341

Paul Reynolds

for redistribution to the Byzantine army, navy and administration, in Danubian fortresses and forts and in mainland and island ports. We know from Procopius that African navicularii bought and sold merchandise in the port of Constantinople (H.A. 25: 7–10). However, the rarity of Tunisian amphorae in contrast to the major quantities of African Red Slip Ware in the Saraçhane deposits of Constantinople published by Hayes (1992) is a notable fact that suggests the primary Tunisian export to the Byzantine capital from c. 450 to late 7th century was grain (Panella 1993: 639: Reynolds 1995: 129; Bonifay 2005: 576). This is a pattern of supply echoed in Beirut and other eastern ports: e.g. Tocra, Alexandria, Corinth, Lechaion (recent excavations: Tunisian amphorae are notably rare in comparison to Vandal to mid- 6th century ARS), Argos, Kourion, Ramat Hanadiv (with large numbers of Vandal ARS and late Sinopean amphorae, but no Tunisian amphorae, which should reflect the supply to Caesarea otherwise still poorly documented) (Hayes 1972; Reynolds 1995 and 2010a; Calderon 2000; Slane and Sanders 2005; Hayes 2007). Comparative trends in Ephesus when published will here be interesting. For the Aegean, the finds of the AD 668 (?) destruction deposits of the fort of Emporio (Chios), built possibly as late as Constans II (641-668), are well known, if poorly recorded: these include complete examples of oil amphorae Keay 8A and Keay 62Q, perhaps again rarer than the contemporary ARS (Ballance et al. 1989: 6–7, 106, Plate 24, nos.225 and 231). For the lower Danube, the presence of a few Vandal Tunisian amphorae (Keay 62Q: oil) in the c. 480 destruction contexts of the horreum at the fort of Dichin (Bulgaria) as well as a good number of small late spatheia in the final destruction contexts of c. 580 has been documented (Swan 2007, 2020). A typical example of the range of Tunisian amphora imports, with some basic quantification, has been published from Novae (Biernacki and Klenina 2015: 113–114, fig. 14): large spatheia Keay 25/ Bonifay Types 29 (no % given) and Keay 32A (1% of the 5th-6th century amphorae, total not given) (for fish and wine); of special interest to us here are amphorae probably carrying oil, represented by Keay 59 (late 4th century to c. 450: rare), Keay 8B (the successor of Keay 59), produced at Majoura (the suggested production at Iunca is less certain, Michel Bonifay, pers. comm.), usually found in contexts of c. 450 to the early 6th century (9% of 5th-6th century amphorae, total not given), Keay 62Q/Bonifay Type 45 (late 5th-first half of 6th century?), a few pieces here in the destruction levels of 470-480, and Keay 61A/Bonifay Type 49 of the late 6th to second half of the 7th century, in this case found in (late) 6th century contexts. At the port of Tomis-Constanţa (Rumania), the number of Vandal to early Byzantine period Keay 8B, found in mausolea T104 and T105, are impressive, alongside examples of Keay 35, 62, 55‑57 (all carrying olive oil) (Opaiţ 1997‑1998). This pattern is equally evident in the Crimea, with Keay 55 (Nabeul), 56 (Nabeul), 62Q, 62R and many examples of Keay 8B (Smokotina 2014: fig. 6). After c. 550 the supply to the Black Sea is quite different. With the exception of small

spatheia, whose contents are still unclear (salsamenta?; olives?), other Tunisian amphorae of the period, such as Keay 61, and ARS appear to be scarce indeed in the lower Danube (Opaiţ 1997‑1998, 2004a; Bonifay 2005; Swan 2007) and in late 6th-7th century contexts in Chersonesos (Sazanov 2000, 2017) and are absent in a context of the late 6th century in Pantikapeion-Kerch (Fedoseev et al. 2010). This may be contrasted with the major supply of Byzantine Tunisian oil and other foodstuffs to military sites and certain major ports in the West, such as Rome, Naples, Marseille, Tarragona, Cartagena and the emporium of Vigo in north-western Spain (Bonifay, Carre and Rigoir 1998; Saguì 1998; Mannoni and Murialdo 2001; Bien 2007; Reynolds 2010a: Chapters 4-5, with references; Reynolds 2011; Fernández 2014; Rodríguez Martorell and Macias Solé, forthcoming). Butrint, technically in the Byzantine East, receives equal large numbers of Tunisian amphorae and fine wares, as well as LRC and Aegean amphorae throughout the 5th to mid-7th centuries (related to Tunisian and Eastern shipping heading for Ravenna and the military sites of Illyricum, perhaps) (Reynolds 2004a, 2020), whereas Nicopolis has a similar Tunisian and Aegean supply but, surprisingly, no LRC in the 6th and 7th century deposits I have studied (Reynolds and Pavlidis 2014, 2017; some still unpublished). These are in their different ways related to the complex inter-relationship and circular maritime traffic of Byzantine Africa, Ravenna, forts and ports in the Aegean and Constantinople. 9. Conclusions Since the presentation of my original paper in 2009 in Poitiers, there has been a marked increase in the publication of amphora production sites and amphorae in the Aegean and Greek mainland which, coupled with ongoing work in Epirus, is focusing attention on the key role of these regions as producers of surpluses, including oil for export and their local-regional markets. The possible role of the Church in the production and distribution of some of these latter long distance exports (LRA 2 in the Aegean; Egyptian Egloff 167 copies of LRA 2C), as well as wine (e.g. LRA 1, LRA 5, Agora M 273), clear in the case of Egyptian LRA 5 derivatives in the monasteries of the Nile Delta, is still a major question. Some of the problems regarding the identification of the sources of LRA 2, and its typological-source relationships to the Dressel 24 and related forms, as well as Agora M 235, have also been noted and really need to be addressed. The key evidence from press sites, many identified through archaeological surveys (especially in Greece and Turkey), needs also to be considered together with the presence or absence of amphorae associated with them. It may be no coincidence that a recurring pattern is the production of amphorae right on the coast, for these to be filled with liquid goods transported there in animal skins from inland estates.

342

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

A persistent and frustrating problem is the identification of what amphorae produced in the East actually carried. We are still behind here with respect to the major advances achieved for Tunisian amphorae. The recent analyses of Pontic amphorae appear to contradict what has been proposed by some scholars in the past. In many cases, quite different views on contents have been put forward for the same amphora types. We can theoretically suggest that the vast majority carried wine (as well as products in wine?), fish or dried fruit in a number of cases, and that a fairly small number of forms (Dressel 25, Dressel 24 and Zeest 90, LRA 2 and its later derivatives, Agora M 235) carried surplus oil for long distance export as well as for local-close regional markets, with Syria supplying its troops with local oil carried in western Syrian painted amphora forms. But there is general agreement that we cannot continue in this fashion and progress ‘scientifically’ without the, fortunately now possible, but still costly, chemical, including perhaps DNA, analysis of organic residues in eastern amphorae, especially those in shipwrecks which have been proven to be actually some of the best preserved cases, despite their watery environment.70 Such analyses are essential for a correct, or more correct, reconstruction of economic trends and history. The problem of the organic residues analyses of what may be re-used amphorae is very pertinent, and can only be resolved by repeated analyses of the same problematic forms.71 It will be interesting to read a paper on the theme of amphora production and contents in the East in ten years’ time. In the meantime, some of what we know and the problems that have to be tackled have here, I hope, been identified. 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. Paris, de Boccard. Argoud, G., Callot, O. and Helly, B. 1980. Salamine de Chypre XI. Une résidence byzantine. “L’Huilerie”. Paris, de Boccard. Arthur, P. 1989. Aspects of the Byzantine economy: an evaluation of amphora evidence from Italy. In V. Déroche and J.-M. Spieser (eds), Recherches sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 79–93. Athens, École française d’Athènes. Arthur, P. 1998. Eastern Mediterranean amphorae between 500 and 700: a view from Italy. In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del Convegno in onore di John W. Hayes (Rome 1995): 157–183. Florence, All’Insegna del Giglio. 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.  Woodworth et al. 2015; Papers presented at this conference.  A point highlighted by the aforementioned Spanish funded project on amphora contents (note 8). 70 71

Aupert, P. 1980. Objets de la vie quotidienne à Argos en 585 ap. J.-C. In Études argiennes. Bulletin de Correspondance Hellénique Supplément 6: 395–457. Athens, École française d’Athènes. Auriemma, R. and Quiri, E. 2004. Importazioni di anfore orientali nell’Adriatico tra primo e medio impero. 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 2629, 2002. Monographs of the Danish Institute at Athens 5: 43–55. Aarhus, Aarhus University Press. Auriemma, R., Degrassi, V. and Quiri, E. 2015. Eastern amphora imports in the Adriatic Sea: evidence from terrestrial and underwater contexts of the Roman Imperial age. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 139– 160. Uppsala, Åströms förlag. Ballance, M., Boardman, J., Corbett, S. and Hood, S. 1989. Excavations in Chios, 1952-1955. Byzantine Emporio. Annual of the British School at Athen, Supplementary volume 20. Athens, The British School of Archaeology at Athens - London, Thames and Hudson. Bass, G. F. and F. H. Van Doorninck 1971. A Fourth-Century Shipwreck at Yassi Ada, American Journal of Archaeology 75: 27–37, Plates 1–4. Bass, G. F. and F. H. Van Doorninck Jr. (eds) 1982. Yassi Ada, Volume I, A Seventh-Century Byzantine Shipwreck. College Station, Texas A&M University Press. Beaudry, N., Blanc, P., Cerova, Y., Chevalier, P., Haxhimihali, M., Hobdari, E., Islami, A., Kozelj, T., Muçaj, S., Nallbani, E., Sodini, J.-P., Tassignon, I., Vanderheyde, C. and Wurch-Kozelj, M. 2003. Byllis (Albanie), Bulletin de Correspondance Hellénique 127-2: 622–639. Belotti, C. 2004. Ritrovamenti di anfore romane a Ivlia Concordia. Aspetti topografici ed economici. L’Album 10. Venice, Fondazione Antonio Colluto. Benea, D. 2000. Les amphores de Tibiscum. Les relations commerciales entre la Dacie et les territoires de la Méditerranée orientale. Rei Cretariae Romanae Fautorum Acta 36: 435–438. Bernal-Casasola, D. 2010. Iglesia, producción y comercio en el mediterráneo tardoantiguo. De las ánforas a los talleres eclesiásticos. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 19–31. Oxford, Archaeopress. Bernal-Casasola, D., Cottica, D., García Vargas, E., Toniolo, L., Rodríguez-Santana, C. G., Acqua, C., Marlasca, R., Sáez, A. M., Vargas, J. M., Scremin, F. and Landi, F. 2014. Un contexto excepcional en Pompeya: la pila de ánforas de la Bottega del Garum (I.12.8). Avance de un estudio interdisciplinar. Rei Cretariae Romanae Fautorum Acta 43: 219–232.

343

Paul Reynolds

Bertoldi, T. 2012. Guida alle anfore romane di età imperiale. Forme, impasti e distribuzione. Rome, Espera. Bertrand, E. 1992. Les amphores d’un vide sanitaire du 1er siècle à Lyon (Saint-Just). SFECAG, Actes du Congrès de Tournai: 265–277. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Bezeczky, T. 2013. The Amphorae of Roman Ephesus. Forschungen in Ephesus 15/1. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Bien, S. 2007. La vaisselle et les amphores en usage à Marseille au VIIe siècle et au début du VIIIe siècle: première ébauche de typologie évolutive. 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 (I): 263–274. Oxford, Archaeopress. Biernacki, A. B. and Klenina, E. Y. 2015. Amphorae of the 4th-6th centuries AD from Novae (Moesia Secunda): typology and chronology. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 99–120. Uppsala, Åströms förlag. Bjelajac, L. 1996. Amforegornjomezijskog Podunavlja. Posebna Izdanja Monographs 30. Belgrade, Srpska Akademija Nauka i Umetnosti. (summary in English: Amphorae of the Danubian Basin in Upper Moesia). Blitzer, H. 1990. Κορώνείκά storage-jar production and trade in the traditional Aegean. Hesperia 59: 675–711, Pl. 99–112. Blitzer, H. 2004. Chapter 6. Agriculture and subsistence in the Late Ottoman and post-Ottoman Mesara. In The Plain of Phaistos. Cycles of Social Complexity in the Mesara Region of Crete. Monumenta Archaeologica 23: 111–217. Los Angeles, University of California, Cotsen Institute of Archaeology. Boardman, J. 1989. Chapter 3. The Finds. In M. Ballance, J. Boardman, S. Corbett, and S. Hood, Excavations in Chios, 1952-1955. Byzantine Emporio. Annual of the British School at Athens Supplementary volume 20: 86–142. Athens-London, The British School of Archaeology at Athens-Thames and Hudson. Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports International Series 1301. Oxford, Archaeopress. Bonifay, M. 2005. Observations sur la diffusion des céramiques africaines en Méditerranée orientale durant l’Antiquité Tardive. In Mélanges Jean-Pierre Sodini. Travaux et Mémoires 15: 565–581. Paris, Collège de France. Bonifay, M. 2007a. Observations préliminaires sur les amphores africaines de l’oasis de Bahariya. In S. Marchand and A. Marangou (eds), Amphores d’Égypte de la Basse Époque à l’époque arabe. Cahiers de la Céramique égyptienne 8 (I): 451–462.

Bonifay, M. (with appendix by N. Garnier) 2007b. Que transportaient donc les amphores africaines ?. In E. Papi (ed.) Supplying Rome and the Empire. The Proceedings of an International Seminar Held at SienaCertosa di Pontignano on May 2-4, 2004 on Rome, the Provinces, Production and Distribution. Journal of Roman Archaeology Supplementary Series 69: 9–31. Portsmouth RI, Journal of Roman Archaeology. Bonifay, M. 2016. Amphores de l’Afrique romaine : nouvelles avancées sur la production, la typochronologie et le contenu. In R. Járrega and P. Berni (eds), Amphorae ex Hispania: paisajes de producción y consumo. III Congreso Internacional de la Sociedad de Estudios de la Cerámica Antigua (SECAH) - Ex Officina Hispana (Tarragona, 10-13 de diciembre de 2014). Monografías Ex Officina Hispana III: 595–611. Tarragona, Institut Català d’Arqueologia Clàssica. Bonifay, M. and Capelli, C. 2013. Les thermes du Levant à Leptis Magna : quatre contextes céramiques des IIIe et IVe siècles. Antiquités africaines 49: 67–150. Bonifay, M. and Leffy, R. 2002. Les céramiques du remplissage de la citerne du Sérapéum à Alexandrie. In Alexandrina 2. Études Alexandrines 6: 39–84. Cairo, Institut français d’archéologie orientale. 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. Bonifay, M. and Tchernia, A. 2012. Les réseaux de la céramique africaine (Ier-Ve siècles). In S. Keay (ed.), Rome, Portus and the Mediterranean. Archaeological Monographs of the British School at Rome 21: 315– 333. London, The British School at Rome. Bonifay, M., Capelli, C. and Şenol, A. K. 2015. Amphores africaines tardives à Alexandrie. Archéologie et archéometrie. In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Études alexandrines 43 (II): 845–857. Alexandria, Centre d’Études Alexandrines. Bonifay, M., Carre, M.-B., and Rigoir, Y. (eds) 1998. Fouilles à Marseille. Les mobiliers (1er-VIIe siècles ap. J.-C.). Études Massaliètes 5. Paris, Errance. Brun, J.-P. 2003. Le vin et l’huile dans la Méditerranée antique. Viticulture, oléiculture et procédés de fabrication. Paris, Errance. Brun, J.-P. 2004. Archéologie du vin et de l’huile dans l’Empire romain, Paris, Errance. Brun, J.-P. and Pasqualini, M. forthcoming. Fouilles dans le port antique de Toulon (Var), in press. Bruno, B. 2002. Importazioni e consumo di derrate nel tempo: l’evidenza delle anfore. In F. Rossi (ed.), Nuove ricerche sul Capitolium di Brescia. Scavi, studi e restauri: 277–307. Milan, Et. Bull, W. E. 1936. The olive industry of Spain. Economic Geography 12.2: 136–154. Butcher, K. 2003. Roman Syria and the Near East. London, The British Museum Press. Calderon, R. 2000. Roman and Byzantine pottery. In Y. Hirschfeld (ed.), Ramat Hanadiv Excavations, Final Report of the 1984-1998 Seasons: 91–115. Jerusalem, Israel Exploration Society.

344

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Callot, O. 1984. Huileries antiques de la Syrie du nord. Bibliothèque archéologique et historique 118. Paris, Librairie orientaliste P. Geuthner. Carreras Monfort, C. 2002. The Roman military supply during the Principate. Transportation and staples. In P. M. Erdkamp (ed.), The Roman army and the economy: 70–89. Leiden, Brill. Casana, J. 2004. The archaeological landscape of late Roman Antioch. In I. Sandwell and J. Huskinson (eds), Culture and society in later Roman Antioch. Papers from a colloquium, London 15th December 2001: 102–125. Oxford, Oxbow Books. Chiriac, C. and Munteanu, L. 2014. Trade Connections between Asia Minor and the Western Pontic Area in the 4th Century CE. Some Sphragistic Considerations. In V. Cojocaru, A. Coşkun and M. Dana (eds), Interconnectivity in the Mediterranean and Pontic World. Pontica et Mediterranea III: 299–330. Cluj-Napoca, Mega. Costa, S. 2015. An archaeology of domestic life in Early Byzantine Gortyna. Stratigraphy, pots and contexts. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 711–721. Uppsala, Åströms förlag. Curta, F. 2002. Quaestura exercitus: The evidence of lead seals. Acta Byzantina Fennica n. s. 1: 9–26. Curtis, R. I. 1991. Garum and salsamenta. Production and commerce in materia medica. Studies in ancient medicine 3. Leiden - New York - Copenhagen - Cologne, Brill. Dafi, E. 2014. Amphorae and cooking wares from the coastal site of Antikyra in Boeotia. In N. PoulouPapadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. The Mediterranean: a Market without Frontiers. British Archaeology Reports International Series 2616 (II): 737–748. Oxford, Archaeopress. Dalby, A. 2000. Empire of Pleasures. Luxury and Indulgence in the Roman World. London, Routledge. Decker, M. 2000. Food for an Empire: Wine and oil production in North Syria. In S. Kingsley and M. Decker (eds), Economy and Exchange in the East Mediterranean during Late Antiquity. Proceedings of a conference at Somerville College, Oxford, 29th May, 1999: 69–86. Oxford, Oxbow Books. Demesticha, S. 2003. Amphora production in Cyprus during the Late Roman Period. In C. Bakirtzis (ed.), VIIe Congrès International sur la Céramique Médiévale en Méditerranée (Thessalonique, 11-16 octobre 1999): 472–474. Athens, Ministère de la Culture. Demesticha, S. and Spondylis, E. 2011. Late Roman and Byzantine trade in the Aegean. Evidence from the HIMA Survey Project at Pagasitikos Gulf, Greece. Skyllis 11.1: 34–40. Diamanti, C. 2010. Local production and import of amphoras at Halasarna of Kos island (5th-7th c.). S. Saripolos Library 115. Athens, National and Kapodistrian University of Athens.

Diamanti, C. 2012. Byzantine Emperors on stamped Late Roman/Early Byzantine amphoras. Rei Cretariae Romanae Fautorum Acta 42: 1–5. Diamanti, C., Kouzeli, K. and Petridis, P. 2014. Archaeology and Archaeometry in Late Roman Greece: the case of mainland and insular settlements, workshops and imports. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. The Mediterranean: a Market without Frontiers. British Archaeology Reports International Series 2616 (I): 181–192. Oxford, Archaeopress. Díaz-Iglesias, L. and Galán. J. M. 2007. El aceite en el Antiguo Egipto. In J. M. Blázquez Martínez and J. Remesal Rodríguez (eds), Estudios sobre el Monte Testaccio (Roma) IV. Instrumenta 24: 553–587. Barcelona, Universitat de Barcelona. Didioumi, S. 2014. Local pottery in the island of Cos, Greece from the early Byzantine period. A preliminary report. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeological Reports International Series 2616 (I): 169–180. Oxford, Archaeopress. Didioumi, S. and Sdrolia, S. forthcoming. Late Roman amphorae from the castrum of Velika (Thessaly, Greece). A preliminary report. In LRCW 6. Diederichs, C. 1980. Salamine de Chypre, IX. Céramiques Hellénistiques, Romaines et Byzantines. Salamine de Chypre 9. Paris, de Boccard. Dixneuf, D. 2011. Amphores égyptiennes. Production, typologie, contenu et diffusion (IIIe siècle avant J.-C.-IXe siècle après J.-C.). Études alexandrines 22. Alexandria, Centre d’Études Alexandrines. Duggan, M. 2018. Links to Late Antiquity. Ceramic exchange and contacts on the Atlantic Seaboard in the 5th to 7th centuries AD. British Archaeological Reports British Series 639. Oxford, BAR publishing. Dugonjić, P. 2015. Rhodian amphorae in the Adriatic: a preliminary report on the evidence from Croatia. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 245–256. Uppsala, Åströms förlag. Duval, N. and Caillet, J. 1982. Khan Khalde (ou Khalde III). Les fouilles de Roger Saidah dans les églises, mises en œuvre d’après les documents de l’auteur. In Archéologie au Levant, Recueil à la mémoire de R. Saidah. Collection de la Maison de l’Orient méditerranéen 12, Série archéologique 9: 311–394 Lyon, Maison de l’Orient. Dyczek, P. 2001. Roman Amphorae of the 1st - 3rd centuries AD found on the Lower Danube, Typology. Warsaw, Warsaw University Press. Dyczek, P. 2008. On the origins of amphora Zeest 90/ Dyczek 25 - ultimate solution? Rei Cretariae Romanae Fautorum Acta 40: 515–521.

345

Paul Reynolds

Empereur, J.-Y. 1987. Les amphores. In V. Karageorghis et al. (eds), La Nécropole d’Amathonte, Tombes 113‑367. II, Céramiques non chypriotes. Études chypriotes 8: 44. Nicosia, Department of Antiquities - Athens, École française d’Athènes. 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 recherche. Actes du colloque de Sienne (22‑24 mai 1986). Collection de l’École française de Rome 114: 223–248. Rome, École française de Rome. Evrin, T. Ç. 2005. The cooking wares of the Romans discovered during the excavations at the Republic Square in Tarsus, Cilicia. In J. M. Gurt i Esparraguerra, J. Buxeda i Garrigós, and M. A. Cau Ontiveros (eds), LRCW I. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports International Series 1340: 681–689. Oxford, Archaeopress. Fantuzzi, L. 2020. Appendix C. Petrographic analysis of pottery from Butrint and environs. In P. Reynolds, Butrint 6. Excavations on the Vrina Plain. Volume 3. The Roman and Late Antique Pottery from the Vrina Plain Excavations. Butrint Archaeological Monograph Series 6, 3: 269–300. Oxford, Oxbow Books. Fedoseev, N. F., Domżalski, K., Opaiţ, A. and Kulikov, A. V. 2010. Post-Justinian pottery deposit from Pantikapeion-Bosporos: Rescue excavations at 12, Teatral’Naja St. in Kerch, 2006, Archeologia 61: 63–94. Fernández, A. 2014. El comercio tardoantiguo (ss. IV-VII) en el Noroeste Peninsular a través del registro cerámico de la Ría de Vigo. Roman and Late Antique Mediterranean Pottery 5. Oxford, Archaeopress. Ferrazzoli, A. F. and Ricci, M. 2007. Elaiussa Sebaste: produzioni e consumi di una città della Cilicia tra V e VII secolo. 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 (II): 671–688. Oxford, Archaeopress. Ferrazzoli, A. F. and Ricci, M. 2010. Un centro di produzione delle anfore LR 1: Elaiussa Sebaste in Cilicia. Gli impianti, le anfore. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 815–826. Oxford, Archaeopress. Filis, K. 2019. The ovoid amphorae from Aigion, in the north-west Peloponnese. The connections with Corinth and the Brindisi area. In E. García Vargas, R. R. de Almeida, H. González Cesteros and A. M. Sáez Romero 2019 (eds), The Ovoid Amphorae in the Central and Western Mediterranean: Between the Last Two Centuries of the Republic and the Early Days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13: 3–34. Oxford, Archaeopress.

Fischer-Genz, B. 2014. Local pottery traditions in the rural settlements around Baalbek/Heliopolis, Lebanon. In B. Fischer-Genz, Y. Gerber and H. Hamel (eds), Roman Pottery in the Levant. Local Production and Regional Trade. Proceedings of the round table held in Berlin, 19-20 February 2010. Roman and Late Antique Mediterranean Pottery 3: 79–87. Oxford, Archaeopress. Fischer-Genz, B. and Ehrig, H. 2005. First results of the archaeological survey project in the territory of ancient Heliopolis-Baalbek. In M. Van Ess and K. Rheidt, Archaeological research in Baalbek. A preliminary report on the 2004 and 2005 seasons. Bulletin d’Archéologie et d’Architecture Libanaises 9: 135–138. Fournet, J.-L. and Pieri, D. 2008. Les dipinti amphoriques d’Antinoopolis. In R. Pintaudi (ed.), Antinoupolis I. Scavi e materiali 1: 175–216. Florence, Istituto Papirologico “G. Vitelli”. Fulford, M. G. and Peacock, D. P. S. 1984. Excavations at Carthage: The British Mission, Vol. I (ii). The Avenue du President Habib Bourguiba, Salammbo: the pottery and other ceramic objects from the site. Sheffield, University of Sheffield. Gândilă, A. 2009. Un tezaur de monede bizantine timpurii descoperit la Capidava. Cercetări numismatice 15: 87–105. Gascou, J. 1989. La table budgétaire d’Antaeopolis (P. Freer 08.45 c-d). In J. Lefort and C. Morrisson (eds), Hommes et richesses dans l’ Empire byzantin. I, IVe-VIIe siècle. Réalités byzantines 1: 279–313. Paris, Lethielleux. Gatier, P.-L., Charpentier, G., Harfouche, R., Maqdissi, M., Mercier, F., Nordiguian, L., Pieri, D., Poupet, P. and Rousset, M. O. 2001. Mission de Yanouh et de la haute vallée du Nahr Ibrahim. Rapport Préliminaire 19992001. Bulletin d’Archéologie et d’Architecture Libanaises 5: 93–152. Gerousi, E. 2014. A late Roman workshop at Dilesi in Boeotia. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeology Reports International Series 2616 (I): 193–202. Oxford, Archaeopress. Godard, C. 1995. Quatre niveaux d’abandon de la ville de Vienne (Isère): éléments pour la chronologie des céramiques de la fin du IIe siècle et du IIIe siècle après J.-C. In SFECAG, Actes du Congrès de Rouen: 285–322. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Gohier, P. and Capelli, C. 2013. Les amphorettes levantines du dépotoir du site « 5 Place Jean-Baptiste Masillon » à Arles. Un module miniature des amphores Célestins 1A? Approches typologiques et pétrographiques. In SFECAG, Actes du Congrès d’Amiens: 549–554. Marseille, Société Française d’Étude de la Céramique Antique en Gaule.

346

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Grace, V. 1973. Imports from Pamphylia. In Études Déliennes. Bulletin de Correspondance Hellénique Supplément 1: 183–208. Athens, École française d’Athènes. Grace, V. R. 1979. Amphoras and the ancient wine trade. Agora Picture Book 6. Princeton, American School of Classical Studies at Athens. Guillou, A. and Durliat, J. 1984. Le tarif d’Abydos (vers 492). Bulletin de Correspondance Hellénique 108-1: 581–598. Hadjisavvas, S. 1992. Olive oil processing in Cyprus from the Bronze Age to the Byzantine period. Studies in Mediterranean Archaeology 99. Göteborg, Åströms förlag. Hadjisavvas, S. and Chaniotis, A. 2012. Wine and olive oil in Cyprus and Crete. Socio-economic aspects. In Parallel Lives: Ancient Island Societies in Crete and Cyprus. British School at Athens Studies 20: 157–173. Athens, The British School at Athens. Hamel, H. 2008. Die Reste der Feste. Datierung und Funktion eines spätantiken Fundkomplexes aus Baalbek (Lebanon). Unpublished M.A. thesis, Institut für Klassische Archäologie, Freien Universität Berlin. Hankey, V. 1968. Pottery-making at Beit Shebab, Lebanon. Palestinian Exploration Quarterly 100: 27–32. Hartal, M. 2002. Excavations at Khirbet Zemel, Northern Golan: An Ituraean settlement site. In Z. Gal (ed.). Eretz Zafon: Studies in Galilean Archaeology: 75–117. Jerusalem, Israel Antiquities Authority. Hayes, J. W. 1972. Late Roman Pottery. London, British School at Rome. Hayes, J. W. 1983. The Villa Dionysos excavations, Knossos: The pottery. Annual of the British School at Athens 78: 97–169. Hayes, J. W. 1991. Paphos Volume III. The Hellenistic and Roman Pottery. Nicosia, Department of Antiquities, Cyprus. Hayes, J. W. 1992. Excavations at Saraçhane in Istanbul. Volume 2. The Pottery. Princeton, Princeton University Press. Hayes. J. W. 2000. From Rome to Beirut and beyond: Asia Minor and eastern Mediterranean trade connections. Rei Cretariae Romanae Fautorum Acta 36: 285–297. Hayes, J. W. 2003. Hellenistic and Roman pottery deposits from the ‘Saranda Kolones’ castle site at Paphos. Annual of the British School at Athens 98: 447–516. Hayes, J. W. 2007. Pottery. In A. H. S. Megaw, Kourion. Excavations in the Episcopal Precinct. Dumbarton Oaks Studies 38: 435–475. Washington DC, Dumbarton Oaks Research Library and Collection. Heather, P. 2007. Goths in the Roman Balkans c. 350-500. In A. G. Poulter (ed.), The transition to Late Antiquity on the Danube and beyond. Proceedings of the British Academy 141: 163–190. Oxford, Oxford University Press. Hjohlman, J. 2005. Pyrgouthi in Late Antiquity. In J. Hjohlman, A. Penttinen and B. Wells (eds), 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: 127–266. Stockholm, Åströms förlag.

Iacomi, V. 2010. Some notes on Late-Antique Oil and Wine Production in Rough Cilicia (Isauria) in the Light of Epigraphic Sources: Funerary Inscriptions From Korykos, LR 1 Amphorae Production in Elaiussa Sebaste and the Abydos Tariff. In Ü. Aydinoğlu and A. K. Şenol (eds), Olive Oil and Wine Production in Anatolia during Antiquity. Symposium Proceedings, MersinTurkey (November 2008): 19–32. Istanbul, Kilikya Arkeolojisini Arastirma Merkezi. Isaac, B. 1990. The Limits of Empire. The Roman Army in the East. Oxford, Clarendon Press. Ivantchik, A.I. 2002. Un puits d’époque paléochrétienne sur l’agora d’Argos. Bulletin de Correspondance Hellénique 126-1: 331–404. Izdebski, A. 2013. The economic expansion of the Anatolian countryside in Late Antiquity: the coast versus inland regions. In L. Lavan (ed.), Local Economies? Production and Exchange of Inland Regions in Late Antiquity. Late Antique Archaeology 10: 343–376. Leiden, Brill. Jennings, S. 2004-2005. Vessel Glass from Beirut. BEY 006, 007 and 045. In Archaeology of the Beirut Souks 2. Berytus 48–49. Jones, A. H. M. 1966. The Decline of the Ancient World. London, Routledge. Karagiorgou, O. 2000. LRA2: a container from the military annona on the Danubian border? In S. Kingsley and M. Decker (eds), Economy and Exchange in the East Mediterranean during Late Antiquity. Proceedings of a conference at Somerville College, Oxford, 29th May, 1999: 129–166. Oxford, Oxbow Books. Karagiorgou, O. 2001. Urbanism and economy in late antique Thessaly (3rd to 7th century A.D.): the archaeological evidence (vols I-III). Unpublished PhD thesis, University of Oxford. . Kassab Tezgör, D. 2010. Le réseau commercial des amphores sinopéennes entre les IIe-IIIe s. et le Vie s. de notre ère. In PATABS I. Production and Trade of Amphorae in the Black Sea. Actes de la Table Ronde international de Batoumi et Trabzon, 27-29 avril 2006. Varia Anatolica 21: 167–173. Istanbul, Institut français d’Études Anatoliennes-Georges Dumézil. Kassab Tezgör, D. 2011. Types amphoriques romains tardifs produits simultanément à Demirci près de Sinope et dans d’autres centres. In Ch. Tzochev, Ch., T. Stoyanov and A. Bozkova (eds), PATABS II, Production and Trade of Amphorae in the Black Sea. Acts of the International Round Table held in Kiten, Nessebar and Sredetz, September 26-30, 2007: 259–264. Sofia, Bulgarian Academy of Sciences/St. Kliment Ohridski University of Sofia. Kassab Tezgör, D. and Tatlican, I. 1998. Fouilles des ateliers d’amphores à Demirçi, près de Sinope en 1996 et 1997. Anatolia Antiqua 6: 423–442. Keay, S. J. 1984. Late Roman Amphorae in the Western Mediterranean. A Typology and Economic Study: the Catalan Evidence. British Archaeological Reports International Series 196. Oxford, Tempus Reparatum.

347

Paul Reynolds

Kingsley, S. and Decker, M. 2000. New Rome, new theories on inter-regional exchange. An introduction to the East Mediterranean economy in Late Antiquity. In S. Kingsley and M. Decker (eds), Economy and exchange in the East Mediterranean during Late Antiquity. Proceedings of a conference at Somerville College, Oxford, 29th May, 1999: 1–27. Oxford, Oxbow Books. Klenina, E. 2010. Trade relations between the Mediterranean and the Black Sea region in the 3rd6th centuries A.D. in the light of ceramic artifacts. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 1005–1014. Oxford, Archaeopress. Klenina, E. 2015. Amphorae of the 1st century BC-4th century AD from Chersonesos Taurica: the issue of trade relations. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 79– 97. Uppsala, Åströms förlag. Koumousi, A. and M. Theodoropoulou 2015. Ζαχλωρίτικα Αιγιαλείας: συμβολή στην ιστορική τοπογραφία και αρχαιολογία της περιοχής κατά την Πρωτοβυζαντινή περίοδο. Patras. Koutsouflakis, G. V. and Argiris, X. 2015. Roman North African amphorae in the Aegean: the evidence of shipwrecks. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 3–22. Uppsala, Åströms förlag. Kowarska, Z. and Lenarczyk, S. 2014. Pithos-type vessels from the excavations in Chhîm, Lebanon. In B. Fischer-Genz, Y. Gerber and H. Hamel (eds), Roman Pottery in the Levant. Local Production and Regional Trade. Proceedings of the round table held in Berlin, 19-20 February 2010. Roman and Late Antique Mediterranean Pottery 3: 125–134. Oxford, Archaeopress. Le Bomin, J., Marchand, D., Reynolds, P. forthcoming. Evidence of a Long-Lasting Trade: Proto-LR 1 Amphorae found at Taposiris Magna (Egypt). Bulletin de Liaison de la Céramique Égyptienne 31, 2022, in press. Leidwanger, J. 2014. A preliminary archaeometric analysis of the Late Roman 1 amphoras from the cargo of the seventh century Yassiada shipwreck, Turkey. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeology Reports International Series 2616 (II): 897–906. Oxford, Archaeopress. Leidwanger, J. 2018. New investigations of the 6th-c-AD “church wreck” at Marzamemi, Sicily. Journal of Roman Archaeology 31: 339–356.

Lemaître, S. 1995. Les importations d’amphores orientales à Lyon de l’époque d’Auguste au début du IIIe siècle après J.-C.. Étude préliminaire. In SFECAG, Actes du Congrès de Rouen: 195–205. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Lemaître, S. 2000. Les importations d’amphores de Méditerranée orientale à Lyon au IIIe siècle ap. J.-C. Rei Cretariae Romanae Fautorum Acta 36: 467–476. Lemaître, S. and Vener-Marksteiner, B. 2018. Faciès céramiques de l’antiquité tardive dans la ville portuaire d’Andriakè en Lycia (Turquie). Résultats d’une prospection près des horrea. In P. Ballet, S. Lemaître and I. Bertraud (eds), De la Gaule à l’orient méditerranéen. Fonctions et statuts des mobiliers archéologiques dans leur contexte. Collection Archéologie et Culture: 259–269. Rennes, Presses Universitaires de Rennes - Cairo, Institut français d’archéologie orientale. Lemaître, S., Waksman, S. Y., Reynolds, P., Roumié, M. and Nsouli, B. 2005. À propos de l’origine levantine de plusieurs types d’amphores importés en Gaule à l’époque impériale. In SFECAG, Actes du Congrès de Blois: 515–528. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Liebeschuetz, J. H. W. G. 2007. The Lower Danube Region under Pressure: from Valens to Heraclius. In A. G. Poulter (ed.), The transition to Late Antiquity on the Danube and beyond. Proceedings of the British Academy 141: 101–134. Oxford, Oxford University Press. Lohmann, H., Kalaitzoglou, G. K. and Lüdorf, G. 2017. Forschungen in der Mykale I, 1. Survey I der Mykale (Dilek Daglan / Aydin) 2001-2009. Landeskunde eines westkleinasiatischen Gebirgszuges vom Chalkolithikum bis in spätosmanische Zeit. Asia Minor Studien 77. Bonn, R. Habelt. Long, L. and Duperron, G. 2013. Navigation et commerce dans le delta du Rhône: l’épave Arles-Rhône 14 (IIIe s. ap. J.-C.). In S. Mauné and G. Duperron (eds), Du Rhône aux Pyrénées. Aspects de la vie matérielle en Gaule Narbonnaise II. Archéologie et Histoire Romaine 25: 125–167. Montagnac, Éd. Monique Mergoil. Lopez, R. S. 1959. The role of trade in the economic readjustment of Byzantium in the seventh-century. Dumbarton Oaks Papers 13: 67–85. MacKinnon, M. 2019. Constistency and change: zooarchaeological investigation of Late Antique diets and husbandry techniques in the Mediterranean region. In F. Dugast and C. Michel D’Annoville (eds), L’alimentation dans l’Antiquité tardive (IVe-VIIIe siècles). Antiquité Tardive 27: 135–148. Manacorda, D. 2019. Produzioni di anfore ovoidi di área brindisina. In E. García Vargas, R. R. de Almeida, H. González Cesteros and A. M. Sáez Romero (eds), The Ovoid Amphorae in the Central and Western Mediterranean: Between the Last Two Centuries of the Republic and the Early Days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13: 35–41. Oxford, Archaeopress. Manacorda, D. and Pallecchi, S. (eds) 2012. Le fornaci romane di Giancola (Brindisi). Bari, Edipuglia.

348

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Mannoni, T. and Murialdo, G. (eds) 2001. S. Antonino: Un insediamento fortificato nella Liguria bizantina. Collezione di Monografie Préistoriche ed archeologiche XI. Bordighera, Istituto Internazionale di Studi Liguri. Marangou-Lerat, A. 1995. Le vin et les amphores de Crète de l’époque classique à l’époque impériale. Études Crétoises 30. Athens, École française d’Athènes. Marimon Ribas, P. and Puig Palerm, A. 2007. Miscelánea: Las otras ánforas del Monte Testaccio. In J. M. Blázquez Martínez and J. Remesal Rodríguez (eds), Estudios sobre el Monte Testaccio (Roma) IV. Instrumenta 24: 345–357. Barcelona, Universitat de Barcelona. Marlière, E. and Costa, J. T. 2007. Transport et stockage des denrées dans l’Afrique romaine : le rôle de l’outre et du tonneau. In A. Mrabet and J. Remesal Rodríguez (eds), In Africa et in Hispania: Études sur l’huile africaine. Instrumenta 25: 85–106. Barcelona, Universitat de Barcelona. Marquié, S. 2003. Appendix VIII. Un lot de céramiques de la première moitié du IIIe siècle de notre ère dans le Secteur I. In V. Karageorghis (ed.), Excavations at Kition. VI. The Phoenician and later levels, Part II: 387–404. Nicosia, Department of Antiquities. Martin-Kilcher, S. 1994. Die römischen Amphoren aus Augst und Kaiseraugst. Ein Beitrag zur römischen Handels- und Kulturgeschichte. 2, Die Amphoren für Wein, Fischsauce, Südfrüchte (Gruppen 2-24) und Gesamtauswertung. Forschungen in Augst, 7/2.3. Augst, Römermuseum. 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. Mazou, L. 2015. Nouvelles données sur les amphores d’Afrique vers la Cyrénaïque et l’Égypte. De la fin de l’époque romaine aux premiers temps de la conquête arabe. In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. Études alexandrines 43 (II): 881–900. Alexandria, Centre d’Études Alexandrines. Mazzocchin, S. and Gualtieri, S. 2004. Le anfore Dressel 25 rinvenute a Padova. Un caso di studio archeologico e archeometrico. In F. Berti et al. (eds), Metodologia di ricerca e obiettivi degli studi : lo stato dell’arte. Atti della 6a Giornata di archeometria della ceramica, Ferrara, 9 aprile 2002: 73–77. Bologna. Michaelides, D. and Bakirtzis C. 2003. “Aράβίκοί”. Aμφορεισ στην Κύπρο. In C. Bakirtzis (ed.), VIIe Congrès International sur la Céramique Médiévale en Méditerranée (Thessalonique, 11-16 octobre 1999): 125–136. Athens, Ministère de la Culture. Mills, P. 2013. The Ancient Mediterranean Trade in Ceramic Building Materials. A Case Study in Carthage and Beirut. Roman and Late Antique Mediterranean Pottery 2. Oxford, Archaeopress. Mills, P. forthcoming a. The Hellenistic and Roman pottery and CBM from the Amuq Valley Survey 200911. In L. Swartz Dodd (ed.), Field walking at the Amuq Valley Survey 2009-11, Turkey. University of Chicago.

Mills, P. forthcoming b. The economics of ceramic building material in the late Roman Mediterranean. In LRCW 6. Mills, P. and Reynolds, P. 2014. Amphorae and specialized coarsewares of Ras al Bassit, Syria: local products and exports. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. The Mediterranean: a Market without Frontiers. British Archaeology Reports International Series 2616 (I): 133–142. Oxford, Archaeopress. Nicolaou, K. and Flinder, A. 1976. Ancient Fish Tanks at Lapithos, Cyprus. International Journal of Nautical Archaeology 5: 133–141. Ollà, A. 1997. Osservazioni preliminari sul carico del relitto romano-imperiale nelle acque di Punta Mazza. In Ritrovimenti Subacquei a Milazzo e il relitto di Punta Mazza: 65–98. Azienda Autonoma di Soggiorno e Turismo, Milazzo - Sopritendenza per i Beni Culturali ed Ambientali, Sezioni Archeologica, Messina. Opaiţ, A. 1996. Aspecte ale vieţii economice din Provincia Scythia (secolele IV-VI p. Ch). Producţia ceramicii locale şi de import. Bibliotheca Tracologica 16. Bucarest, Institutul Român de Tracologie. Opaiţ, A. 1997-1998. North African and Spanish Amphorae in Scythia Minor. In Il Mar Nero. Annali di archeologia e storia III: 47–95. Opaiţ, A. 2004a. Local and Imported Ceramics in the Roman Province of Scythia (4th-6th Centuries AD). British Archaeological Reports International Series 1274. Oxford, Archaeopress. Opaiţ, A. 2004b. The Eastern Mediterranean amphorae in the province of Scythia. 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: 293–308. Aarhus, Aarhus University Press. Opaiţ, A. 2007a. From DR 24 to LR 2? 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 (II): 627–643. Oxford, Archaeopress. Opaiţ, A. 2007b. A Weighty Matter: Pontic Fish Amphoras. In V. Gabrielsen and J. Lund (eds), The Black Sea in Antiquity, Regional and Interregional Economic Exchanges. Black Sea Studies 6: 101–121. Aarhus, Aarhus University Press. Opaiţ, A. 2010a. Aspects of the provisioning of the Pontic settlements with olive oil in the Hellenistic and early Roman periods. In D. Kassab Tezgör and N. Inaishvili (eds), PATABS I. Production and Trade of Amphorae in the Black Sea. Actes de la Table Ronde international de Batoumi et Trabzon, 27-29 avril 2006.Varia Anatolica 21: 153–158. Istanbul, Institut français d’Études AnatoliennesGeorges Dumézil.

349

Paul Reynolds

Opaiţ, A. 2010b. Sinopean, Heraklean and Chersonesan “Carrot” Amphorae. Ancient Civilisations from Scythia to Siberia 16: 371–401, 522–556. Opaiț, A. 2014a. Defining more Roman amphora types from the Athenian Agora: too much history, too little typology (I). Rei Cretariæ Romanæ Fautorum Acta 43: 43–54. Opaiţ, A. 2014b. The baggy amphora shape: a new fashion? In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and archaeometry. The Mediterranean: a Market without Frontiers. British Archaeology Reports International Series 2616 (I): 441–450. Oxford, Archaeopress. Opaiţ, A. 2017. On the local production and imports of wine in the Pontic and Lower Danube regions (1st century BC to 7th century AD). In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. Études alexandrines 43 (I): 579–612. Alexandria, Centre d’Études Alexandrines. Opaiţ, A. forthcoming. Supplying olive oil to the Lower Danube border region (2nd-6th century AD). In LRCW 6. Opaiţ, A. and Diamanti, C. 2014. Imperial stamps on early Byzantine amphoras. The Athenian Agora examples. Rei Cretariae Romanae Fautorum Acta 43: 55–61. Opaiţ, A. and Tsaravopoulos, A. 2011. Amphorae of Dressel 24 similis type in the central Aegean area (Chios-ErythraiKyme). Annual of the British School at Athens 106: 275–323. Opaiţ, A., Davis, D. and Brennan, M. L. 2018. Roman commerce and elite markets in the eastern Mediterranean: a case study of three 2nd-century shipwrecks off Knidos. In Koinè et Mobilité Artisanale entre la Méditerranée et la Mer Noire dans l’Antiquité. Pontica LI Supplementum V: 299–334. Constanţa, Musée d’Histoire Nationale et d’Archéologie. Opriş, I. C. and Raţiu, A. 2016. An early Byzantine amphorae deposit from Capidava. Journal of Ancient History and Archaeology (Bucharest) 3.1: 24–42. Özyigit, Ö 1990. 1988 yili Erythrai sondaj çalismalari. Kazi Sonuçlari Toplantisi 11: 125–150. Palaczyk, M. 2015. Quite new in Eretria - not very new in the Aegean? Roman amphorae from the Swiss excavations: preliminary observations. In S. Demesticha (ed.), PER TERRAM - PER MARE, Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 31–40. Uppsala, Åströms förlag. Panella, C. 1986. Oriente e Occidente: considerazioni su alcune anfore egee di età imperiale a Ostia. In J.-Y. Empereur and Y. Garlan (eds), Recherches sur les amphores grecques. Bulletin de Correspondance Hellénique Supplément 13: 609–636. Athens, École Française d’Athènes. Panella, C. 1993. Merci e scambi nel Mediterraneo tardoantico. In Storia di Roma III, 2: 613–641. Turin, Einaudi. Peacock, D. P. S. and Williams, D. F. 1986. Amphorae and the Roman Economy. An Introductory Guide. London - New York, Longman.

Pecci, A., Salvini, L., Cirelli, E. and Augenti, A. 2010. Castor oil at Classe (Ravenna-Italy): residue analysis of some late Roman amphorae coming from the port. In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 617–622. Oxford, Archaeopress. Peña, J. T. 1998. The mobilization of state olive oil in Roman Africa: the evidence of late 4th-c. ostraca from Carthage. In J. T. Peña, J. J. Rossiter, A. I. Wilson, C. Wells, M. Carroll, J. Freed and D. Godden (eds), 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: 117– 238. Portsmouth RI, Journal of Roman Archaeology. Pickersgill C. and Roberts, P. 2003. New light on Roman Sparta: Roman pottery from the Sparta theatre and stoa. Annual of the British School at Athens 98: 549–87. Pieri, D. 2005a. Le commerce du vin oriental à l’époque byzantine (Ve-VIIe siècles). Le témoignage des amphores en Gaule. Bibliothèque archéologique et historique 174. Beirut, Institut français du Proche-Orient. Pieri, D. 2005b. Nouvelles productions d’amphores de Syrie du Nord aux époques protobyzantine et omeyyade. In Mélanges Jean-Pierre Sodini. Travaux et Mémoires 15: 583–596. Paris, Collège de France. Pirazzoli, P. A. 1987. Submerged Remains of Ancient Megisti in Castellorizo Island (Greece): A Preliminary Survey. International Journal of Nautical Archaeology 16.1: 59. Portale, E. C. and Romeo, I. 2000a. Contenitori da trasporto. In Gortina V.3. Lo Scavo del Pretorio (1989-1995) - Vol. V.3 - I Materiali - Tavole. Monografie della Scuola Archeologica Italiana di Atene e delle Missioni Italiani in Oriente XII: 260–410. Padua, Bottega d’Erasmo. Portale, E. C. and Romeo, I. 2000b. Le anfore locali di Gortina ellenistica e romana. Rei Cretariae Romanae Fautorum Acta 36: 417–426. 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), LRCW 3. 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, Archaeopress. Poulter, A. G. 2007a. The transition to Late Antiquity. In A. G. Poulter (ed.), The transition to Late Antiquity on the Danube and beyond. Proceedings of the British Academy 141: 1–50. Oxford, Oxford University Press. Poulter, A. G. 2007b. The transition to Late Antiquity on the Lower Danube: the city, a fort and the countryside. In A. G. Poulter (ed.), The transition to Late Antiquity on the Danube and beyond. Proceedings of the British Academy 141: 51–97. Oxford, Oxford University Press.

350

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Quiri, E. 2015. Imports of eastern transport amphorae to Turin (Italy). In S. Demesticha (ed.), PER TERRAM - PER MARE, Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 161– 180. Uppsala, Åströms förlag. Rauh, N. K. 2004. Pirated Knock-offs: Cilician Imitations of Internationally Traded Amphoras. 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: 329–336. Aarhus, Aarhus University Press. Rauh, N. K. and Slane, K. W. 2000. Possible amphora kiln sites in W Rough Cilicia. Journal of Roman Archaeology 13: 319–330. Raven, S. 1993. Rome in Africa. New York (3rd edition, 1st edition, London 1969). Remesal Rodrίguez, J. and Garcίa Sánchez, M. 2007. Los tituli picti sobre ánforas olearias orientales. In J. M. Blázquez Martínez and J. Remesal Rodríguez (eds), Estudios sobre el Monte Testaccio (Roma) IV. Instrumenta 24: 173–182. Barcelona, Universitat de Barcelona. Remolà i Vallverdú, J. A. 1993. Un tipo de ánfora tardoromana poco conocido (VLR 8.198). Archivo Español de Arqueología 66: 300–310. Remolà i Vallverdù, J. A. 2000. Las ánforas tardo-antiguas en Tarraco (Hispania Tarraconensis). Siglos IV-VII d. C. Instrumenta 7. Barcelona, Universitat de Barcelona. Reynolds, P. 1995. Trade in the Western Mediterranean, AD 400700. The ceramic evidence. British Archaeological Reports International Series 604. Oxford, Tempus Reparatum. Reynolds, P. 2000. Baetican, Lusitanian and Tarraconensian amphorae in classical Beirut: some preliminary observations of trends in amphora imports from the western Mediterranean in the Anglo-Lebanese excavations in Beirut (BEY 006, 007 and 045). In Congreso internacional Ex Baetica Amphorae: conservas, aceite y vino de la Bética en el Imperio Romano (Écija y Sevilla, 17 al 20 de diciembre de 1998). Actas (III): 1035–1065. Ecija, Gráficas Sol. Reynolds, P. 2003a. Amphorae in Roman Lebanon: 50 BC to AD 250. Archaeology and History in Lebanon 17: 120–130. Reynolds, P. 2003b. Pottery and the economy in 8th century Beirut: An Umayyad pottery assemblage from the Roman Imperial baths (BEY 045). In C. Bakirtzis (ed.), VIIe Congrès International sur la Céramique Médiévale en Méditerranée (Thessalonique, 11-16 octobre 1999): 725–734. Athens, Ministère de la Culture. Reynolds, P. 2004a. The Roman pottery from the Triconch Palace, The Medieval amphorae and Appendix 1. 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 Surveys 1994-99: 224–277, 327–395. Oxford, Oxbow Books. Reynolds, P. 2004b. The Pottery. In R. Ortali and B. Stuart, Two rock-cut Roman tombs in Chhîm. Bulletin d’Archéologie et d’Architecture Libanaises 6 (2002): 126–133.

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 Esparraguerra, J. Buxeda i Garrigós and M. A. Cau Ontiveros (eds), LRCW I. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports International Series 1340: 563– 611. Oxford, Archaeopress. Reynolds, P. 2008. Linear typologies and ceramic evolution. FACTA (A Journal of Roman Material Culture Studies) 2: 61–87. Reynolds, P. 2010a. Hispania and the Roman Mediterranean, AD 100-700. Ceramics and Trade. London, Duckworth. Reynolds, P. 2010b. Trade networks of the East, 3rd to 7th centuries: the view from Beirut (Lebanon) and Butrint (Albania) (fine wares, amphorae and kitchen wares). In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 89–114. Oxford, Archaeopress. Reynolds, P. 2011. A 7th century pottery deposit from Byzantine Carthago Spartaria (Cartagena, Spain). In M. À. Cau, P. Reynolds and M. Bonifay (eds), LRFW 1. Late Roman Fine Wares: Solving problems of typology and chronology. A review of the evidence, debate and new contexts. Roman and Late Antique Mediterranean Pottery 1: 99–127. Oxford, Archaeopress. Reynolds, P. 2013. Transport amphorae of the 1st to 7th centuries: Early Roman to Byzantine periods. In W. Aylward (ed.), Excavations at Zeugma, conducted by Oxford University, II: 92–161. Los Altos CA, The Packard Humanities Institute. Reynolds, P. 2014. The Homs Survey (Syria): Contrasting Levantine trends in the regional supply of fine wares, amphorae and kitchen wares (Hellenistic to early Arab periods). In B. Fischer-Genz, Y. Gerber and H. Hamel (eds), Roman Pottery in the Levant. Local Production and Regional Trade. Proceedings of the round table held in Berlin, 19-20 February 2010. Roman and Late Antique Mediterranean Pottery 3: 53–65. Oxford, Archaeopress. Reynolds, P. 2015. Material culture and the economy in the age of Saint Isidore of Seville (6th and 7th centuries).. In J.-P. Caillet, I. Velázquez and G. Ripoll (eds), Isidore de Séville et son temps. Antiquité Tardive 23: 163–210. Reynolds, P. 2018. Amphorae in Beirut from the Umayyads to the Crusaders: A guide to trends in local and imported products. Archeologia Medievale 45: 91–110. Reynolds, P. 2020. Butrint 6. Excavations on the Vrina Plain. Volume 3. The Roman and Late Antique Pottery from the Vrina Plain Excavations. Butrint Archaeological Monograph Series 6, 3. Oxford, Oxbow Books. Reynolds, P. forthcoming. The Pottery (with Appendices A-C). In K. Zachos, The Trophaeum Actium Victory Monument of Augustus, Nicopolis. American Journal of Archaeology Supplements.

351

Paul Reynolds

Reynolds, P. in preparation a. The Classical Pottery. The Archaeology of the Beirut Souks series, Berytus. Reynolds, P. in preparation b. Amphorae of the Imperial Roman period: late 1st century BC to 3rd century AD. American Excavations in the Athenian Agora. Reynolds, P. and Pavlidis, E. 2014. Nikopolis (Epirus Vetus): an early 7th century pottery assemblage from the ‘Bishop’s house’ (Greece). In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeology Reports International Series 2616 (I): 451–467. Oxford, Archaeopress. Reynolds, P. and Pavlidis, E. 2017. An early 5th century pottery deposit from the cloaca by Basilica A, Nicopolis (Epirus, Greece). In Σπείρα. Επιστημονική συνάντηση προς τιμήν της Αγγέλικας Ντούζουγλη και του Κωνσταντίνου Ζάχου. Πρακτικά [Conference in honour of Angelika Douzougli and Konstantinos Zachos] (Ioannina, 1st-3rd November 2012): 651–666. Athens. Reynolds P., Waksman, S. Y., Lemaître, S., Curvers, H., Roumié M. and Nsouli, B. 2008-2009. An early Imperial Roman pottery production site in Beirut (BEY 015): chemical analyses and a ceramic typology. Berytus 51‑52: 71–115. Riley, J. A. 1979. The Coarse Pottery from Berenice. In J. A. Lloyd (ed.), Excavations at Sidi Khrebish Benghazi (Berenice) II. Supplement to Libya Antiqua V (II): 91– 467. Tripoli, Society for Libyan Studies. Riley, J. A. 1981. The pottery from the cisterns 1977.1, 1977.2 and 1977.3. In J. Humphrey (ed.). Excavations at Carthage 1977 conducted by the University of Michigan VI: 85–124. Ann Arbor, University of Michigan Press. Rizos, E. 2015. Remarks on the logistics and infrastructure of the annona militaris in the eastern Mediterranean and the Aegean. Antiquité Tardive 23: 287–302. Rizzo, G. 2014. Le anfore, Ostia e i commerci mediterranei. In C. Panella and G. Rizzo (eds), Ostia VI. Le Terme del Nuotatore. Studi Miscellanei 38: 65–481. Rome, L’Erma di Bretschneider. Robinson, H. S. 1959. The Athenian Agora. Volume V. Pottery of the Roman Period, Chronology. Princeton, The American School of Classical Studies at Athens. Rodríguez Martorell, F. and Macias Solé, J. M. forthcoming. Tarracona: Pottery and trade in the 7th and 8th centuries. In LRCW 6. Saguì, L. 1998. Il deposito della Crypta Balbi: una testimonianza imprevidibile sulla Roma del VII secolo? In L. Saguì (ed.), Ceramica in Italia: VI-VII secolo. Atti del Convegno in onore di John W. Hayes (Rome 1995): 305–330. Florence, All’Insegna del Giglio. Sarri, E. 2014. Aspects of a Roman bath at Argos. Rei Cretariae Romanae Fautorum Acta 43: 63–70. Sazanov, A. 1997. Les amphores de l’Antiquité tardive et du Haut Moyen-Age: continuité ou rupture? Le cas de la Mer Noire. In G. Démians d’Archimbaud (ed.), La céramique médiévale en Méditerranée. Actes du VIe Congrès de l’AIECM2, Aix-en-Provence (13-18 novembre 1995): 87–102. Aix-en-Provence, Narration.

Sazanov, A. 2000. Les ensembles clos de Chersonese de la fin du VIe-troisième quart du 7e siècles: la chronologie des céramiques. In M. Kasanski and V. Soupault (eds), Les sites archéologiques en Crimée et au Caucase durant l’Antiquité Tardive et le Haut Moyen-Age. Colloquia Pontica 5: 123–149. Leiden-Boston-Kőln, Brill. Sazanov, A. 2007. Les amphores d’époque protobyzantine au nord de la Mer Noire: chronologie et typologie. 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 (II): 803–815. Oxford, Archaeopress. Sazanov, A. 2017. Un ensemble de la fin du VIe siècle. Secteur nord de Chersonèse (Crimée). In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. Études alexandrines 43 (I): 651–669. Alexandria, Centre d’Études Alexandrines. Sirks, B. 1991. Food for Rome: the legal structure of the transportation and processing of supplies for the Imperial distributions in Rome and Constantinople. Amsterdam, J. C. Gieben. Slane, K. W. 1989. Corinthian ceramic imports: the changing pattern of provincial trade in the first and second centuries AD. In The Greek Renaissance in the Roman Empire: Papers from the Tenth British Museum Classical Colloquium. Bulletin of the Institute of Classical Studies Supplement 5: 219–225. London, University of London, Institute of Classical Studies. Slane, K. W. 1990. Corinth XVIII.2. The Sanctuary of Demeter and Kore on Acrocorinth: The Roman Pottery and Lamps. Princeton, The American School of Classical Studies at Athens. Slane, K. W. and Sanders, G. D. R. 2005. Corinth: Late Roman horizons. Hesperia 74-2: 243–297. Smokotina, A. 2014. The North African red slip ware and amphorae imported into Early Byzantine Bosporus. Rei Cretariae Romanae Fautorum Acta 43: 71–80. Smokotina, A. forthcoming. Cretan amphora imports into Bosphorus in the 6th century. In LRCW 6. Sodini, J.-P. 2002-2007. Marble and Stoneworking in Byzantium, Seventh-Fifteenth Centuries. In A. E. Laiou (ed.), The Economic History of Byzantium From the Seventh through the Fifteenth Century: 129–146. Washington DC, Dumbarton Oaks. Sodini, J.-P., Tate, G., Bavant, B., Bavant, S., Biscop, J.L., Orssaud, D., Morrisson, C. and Polplin, F. 1980. Déhès (Syrie du Nord), Campagnes I-III (1976-1978): recherches sur l’habitat rural. Syria 57-1: 11–302. Steckner, C. 1989. Les amphores LR 1 et LR 2 en relation avec le pressoir du complexe ecclésiastique des thermes de Samos. In V. Déroche and J. -M. Spieser (eds), Recherche sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 57–71. Athens, École française d’Athènes.

352

The oil supply in the Roman East: identifying modes of production, containers and contents in the eastern Empire

Swan, V. G. 2004. Dichin (Bulgaria) and the supply of amphorae to the Lower Danube in the late Romanearly Byzantine period. 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: 371–382. Aarhus, Aarhus University Press. Swan, V. G. 2007. Dichin (Bulgaria) : The fifth- and sixthcentury destruction deposits, and their implications for ceramic chronology. 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 (II): 835–844. Oxford, Archaeopress. Swan, V. G. 2020. Amphorae and Allied Transit Vessels. In Late Roman and early Byzantine pottery, in A. G. Poulter (ed.), The Transition to Late Antiquity. Excavations and survey at Dichin, a Late Roman to Early Byzantine fort and a Roman Aqueduct: 523–556. Oxford, Oxbow Books. Tate, G. 1992. Les campagnes de la Syrie du Nord du IIe au VIIe siècle. Un exemple d’expansion démographique et économique à la fin de l’Antiquité. Bibliothèque archéologique et historique 133. Paris, Institut français d’archéologie du Proche-Orient. Tchalenko, G. 1953-1958. Villages antiques de la Syrie du Nord, Le Massif de Bélus à l’époque romaine. Bibliothèque archéologique et historique 50. Paris, Libraire Orientaliste Paul Geuthner. Theodoulou, T., Foley, B., Kourkoumelis, D. and PrekaAlexandri, K. 2015. Roman amphora cargoes in the sea of Chios - the 2008 mission. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 1–54. Uppsala, Åströms förlag. Thorpe, R. 2017-2018. The Insula of the House of Fountains in Beirut, Lebanon. In Archaeology of the Beirut Souks 3. Berytus: 57–58. Timby, J. 2004. Amphorae from Excavations at Pompeii by the University of Reading. 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: 383–392. Aarhus, Aarhus University Press. Tomber, R. 1998. ‘Laodicean’ wine containers in Roman Egypt. In O. E. Kaper (ed.), Life on the Fringe. Living in the Southern Egyptian Deserts during the Roman and earlyByzantine Periods. CNWS Publications 71: 213–220. Leiden, Research School CNWS. Tomber, R. 2006. The pottery. In V. A. Maxfield and D. P. S. Peacock (eds), Survey and Excavations. Mons Claudianus 1987-1993, Vol III, Ceramic Vessels and Related Objects from Mons Claudianus. Fouilles de l‘ Institut français d’archéologie orientale 54: 1–236. Cairo, Institut français d’archéologie orientale.

Trombley, F. R. 1989. Boeotia in Late Antiquity: epigraphic evidence on society, economy and Christianization. In Boiotika: Vorträge Vom. 5. Internationalen Böotien - Kolloquium. Institut für Alte Geschichte LudwigMaximilians Universität München 13.-17. Juni 1986: 215–228. Munich, Editio Maris. Tsaravopoulos, A. 1986. Η αρχαία πόλη της Χίου. Στοιχεία για την τοπογραwία της από σωστικές ανασκαwές. Horos 4: 124–144. Vasiliou, S. and Tsigarida, E. B. 2014. Late Roman pottery from recent excavations at Cassandra in Chalcidice. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeology Reports International Series 2616 (II): 723–735. Oxford, Archaeopress. Ventura, P. and Braidotti, E. 2015. Le anfore tardoantiche dal pozzo di via dei Patriarchi. In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. Études alexandrines 43 (I): 313–330. Alexandria, Centre d’Études Alexandrines. Vidličková, L. S. 2015. Roman amphorae from the Portolafia and Tourkolimano shipwrecks. In S. Demesticha (ed.), PER TERRAM, PER MARE. Seaborne Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 23–30. Uppsala, Åströms förlag. Vidličková, L. S. 2019. Northern Peloponnesian amphorae with convex-concave rims from the Styra A shipwreck. In E. García Vargas, R. R. de Almeida, H. González Cesteros and A. M. Sáez Romero 2019 (eds), The Ovoid Amphorae in the Central and Western Mediterranean: Between the Last Two Centuries of the Republic and the Early Days of the Roman Empire. Roman and Late Antique Mediterranean Pottery 13: 228–236. Oxford, Archaeopress. Vilvorder, F., Symonds, R. P. and Rekk, S. 2000. Les amphores orientales en Gaule septentrionale et au sud-est de la Grande Bretagne. Rei Cretariae Romanae Fautorum Acta 36: 477–486. Vipard, P. 1995. Les amphores ‘carottes’ (Forme SchöneMau XV). État de la question. In SFECAG, Actes du Congrès de Rouen: 51–77. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Vnukov, S. Y. 2004. Pan-Roman Amphora Types Produced in the Black Sea Region. 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: 407– 415. Aarhus, Aarhus University Press. Vokaer, A. 2013. Pottery production and exchange in Late Antique Syria (fourth-eighth century A.D.). A study of some imported and local wares. In L. Lavan (ed.), Local Economies? Production and Exchange of Inland Regions in Late Antiquity. Late Antique Archaeology 10: 567–606. Leiden, Brill.

353

Paul Reynolds

Vokaer, A. 2017. Late Roman amphorae from Apamea, Syria. In D. Dixneuf (ed.), LRCW 5. Late Roman Coarse Wares, Cooking wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. Études alexandrines 43 (II): 779–803. Alexandria, Centre d’Études Alexandrines. Waddington, C. J. 2003. The Paphos Theatre Excavations and the Late Roman Pottery of Cyprus. Unpublished PhD thesis, The University of Sydney. Waksman, S. Y., Skartsis, S. S., Kontogiannis, N. D., Todorova, E. P. and Vaxevanis, G. 2016. Investigating the origins of two main types of Middle and Late Byzantine amphorae. Journal of Archaeological Science, Reports 21: 1111–1121. Whitcomb, D. S. 1982. Roman Ceramics. In D. S. Whitcomb and J. H. Johnson, Quseir al-Qadim 1980: Preliminary Report. American Research Center in Egypt Reports 7. Malibu, Undena Publication. Will, E. L. 1983. Exportation of olive oil from Baetica to the eastern Mediterranean. In J. M. Blázquez and J. Remesal Rodríguez (eds), Producción y comercio del Aceite en la Antigüedad. Segundo Congreso Internacional (Sevilla, 24-28 febrero 1982): 391–440. Madrid, Universidad Complutense. Williams, D. F. 1987. Roman amphorae from Kourion, Cyprus. Report of the Department of Antiquities of Cyprus 32: 235–238. Woodworth, M. 2017. Absorbed organic residue analysis of amphorae from the Black Sea region (3rd to 6th c. AD): analyses and methodological considerations. Unpublished PhD Thesis, St Catherine’s College, Oxford. Woodworth, M., Bernal‑Casasola, D., Bonifay, M., Garnier, N., Keay, S., Pecci, A., Poblome, J., Pollard, M., Richez, F. and Wilson, A. 2015. The content of African Keay 25 / Africana 3 amphorae: initial results of the CORONAM project. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 41–57. Esposende, Município de Esposende.

Yangaki, A. G. 2005. La céramique des IVe - VIIIe siècles ap. J.-C. d’Eleutherna : sa place en Crète et dans le bassin égéen. Athens, Scripta. Yangaki, A. G. 2007. Amphores crétoises de forme globulaire : remarques préliminaires. 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 (II): 767–774. Oxford, Archaeopress. Yangaki, A. G. 2014. Late Roman pottery from Funerary Monument 18 and the adjacent monuments of the Arcadian Gate (Ancient Messene, Peloponnese). In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeological Reports International Series 2616 (II): 769–776. Oxford, Archaeopress. Zachos, G. A. 2018. The City of Opous and its Effect on the Settlement Pattern of Opountian Locris in the Roman Period. In V. Di Napoli, F. Camia, V. Evangelidis, D. Grigoropoulos, D. Rogers and S. Vlizos (eds), What’s New in Roman Greece? Recent Work on the Greek Mainland and the Islands in the Roman Period. Proceedings of a Conference held in Athens, 8-10 October 2015. MΕΛΕΤΗΜΑΤΑ 80: 39–54. Athens, National Hellenic Research Foundation/Institute of Historical Research. Zeest, I. B. 1960. Keramičeskaja tara Bospora. Materialy i issledovanija po archeologii SSSR 83. Moscow, Institut Arkheologii (in Russian).

354

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC‑3rd century AD) Andrei Opaiţ

Independent researcher

Abstract: This paper presents an overview of the production of wine and fish products in the Pontic and Lower Danube areas and their presence in the provincial, regional and international markets. In addition to discussing amphora typologies, the author points out the need for a better understanding of the characteristics of these commodities and the features that imposed certain design limits on them. The discussion establishes a foundation for improved understanding of their production and the dynamics of circulation in certain regions, as well as the nature of that trade. In the second part, the author presents the wine and fish product containers made around the Pontic coastal areas. He concludes that there was an intensive provincial and regional trade between these Pontic regions, one able to support not only the garrisons of the Danubian border, but also the Pontic cities and their hinterlands. However, these Pontic products are found less outside the Black Sea area since they were perhaps unable to compete with similar southern foodstuffs. Key words: Black Sea; amphorae; wine; fish products; Sinope; Heraclea; Amastris; Chersonesus.

1. Introduction Writing about purely local production of wine and fish products from a certain area is very challenging as it cannot be compared with similar products that were imported and arrived in the consuming market in specific amphorae, as a result of long distance trade, and thus it is more difficult to see the commercial dynamics. Agricultural and piscatorial production which targeted only the local or provincial market in many cases used perishable containers made of skin or wood that left no archaeological traces. The only exceptions are the table and pitcher amphorae that were quite abundantly made and reused for many years. Another trade was the mostly intra-regional trade, carried in transport amphorae made in the southern, eastern and northern parts of the Black Sea, areas that continued their specialization in vintage wine achieved during the Hellenistic period. Occasionally these amphorae were exported to the Mediterranean. However, local production, in aggregate, was bigger and satisfied a larger sector of the population than the imported vintage wine or fish products. There was undoubtedly a certain hierarchy of these foodstuffs but this hierarchy is not easily discernible if we study only imported amphorae. There were different layers of social strata that were connected to different networks of supply, networks, which, most of the time, acted on multifarious channels of trade, but, in the end, all of them sustained the local society. What is important to observe, in the end, is the balance between the local and imported products. Therefore, to restrict our study only to containers, mostly imported vessels, will shed a warped light on the economic realities of a site or province. I suggest that it will be more useful to start our investigation paying attention first to contents and after that to containers. I feel that this approach is necessary

because there is a fine line between rule and exception from the rule. We should not forget that we are dealing with the results of human activities and it is very difficult to find universally applicable patterns. Therefore, before discussing the local production of these agricultural and piscatorial goods, it is useful to tackle some theoretical issues regarding the amphora and its use. In defining the concept of type, I agree with those scholars who see a type only as an abstract, mental representation of characteristics which form a certain vessel (Hautumm 1981: 89; Adams and Adams 1991: 30). However, when we are dealing with amphorae, some distinctions must be made, and I want to draw attention to them, as quite often the approaches of some scholars, in discussing amphora manufacture, are influenced by fine ware products, considering amphorae as universal containers made for carrying all kinds of liquid, semiliquid and solid products. Basically, the distinction between prime-use and reuse of an amphora is nullified (Peña 2007: 62–70). This distinction is clearly attested as the Egyptian papyri unmistakably distinguished the ‘maker of wine amphora’ from the ‘maker of fine pottery’ (Cockle 1981: 88). Unlike fine wares, which were made in very restricted and specialized workshops by highly specialized potters, using selected clays, specific technologies, and responding to a certain demand of the market for table vessels, amphorae were made in numerous workshops set up for the export of a surplus of agricultural products. Therefore, its necessary basis was a certain agricultural/piscatorial surplus product, the discovery of a large market in need of that yield, addressing means of transport, and the setting up of a reliable trade network, and only after that, specific containers were manufactured. Also, the producer/seller had a constant desire to establish standing relations with

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 355–375

Andrei Opaiţ

the buyers, which will bring repeated orders from his customers and rapid turnover. In an establish market, based on ‘mature mercantilism’, the relationships sellerbuyers were based on goodwill and honesty not on narrow practices (Frier 1983: 291). Another mistake is to put wine, olive oil, and fish products into one category only because they were all carried in amphorae, containers which are not always easily distinguishable. We need to differentiate which amphorae were used for wine, olive oil, garum and salsamenta, as they were not alike. They had distinctly physical properties that differentiated them from many points of view, as we should expect to have different shapes for containers carrying liquids (wine), low viscosity (olive oil), high viscosity (garum), and solid (salted fish) products. However, I agree that in some areas such as North Africa, some cultural and manufactural traditions prevailed. Olive oil, garum, and wine were carried in cylindrical amphorae from Punic to early Byzantine times, the only compromise, made during the Roman time, being the adoption of vertical handles. Though, the other parts of the Mediterranean and in the Pontic area were less conservative and adopted specific forms for specific contents. All these differences have their impact on the amphora design (Opaiţ 2007). Whereas, for example, a small empty space left between the stopper and the wine create some problems (Frier 1983: 259), sometimes it is required (Pliny NH 14.135), in a wine amphora; I think that the kind of wine stored in a container might be important in choosing a certain type of amphora stoppage. However, the reduction of oxygen in an olive oil container is a key factor in preserving the quality of olive oil. Thus, when olive oil is stored in large vessels, these are best kept rather full (Piscopo and Poiana 2012: 208). The problem becomes more complicated when these amphorae were reused. For example, the dissolution of some substances from reused tinplate containers decreases the quality of oils, whereas new containers did not alter olive oil during a oneyear storage period (Piscopo and Poiana 2012: 211). In my opinion, if this is true for virgin oil, both in antiquity as well as modern times, the ancients were perhaps less worried when they transported third-rate olive oil for lighting. Recently, there has been a lively debate regarding the applied sealant coating on the amphora interior, suggesting that pitch was either impossible (Heron and Pollard 1988) or possible to be used also for olive oil amphorae (Romanus et al. 2009; Foley et al. 2009; Oliveira, Morais, and Morillo 2015). The problem is even more complex as it seems that the ancients used to dissolve resin in a small quantity of olive or linseed oil or fat with lime. The addition of lime to resin-oil mixture will react creating calcic soap that produces a good and lasting sealing. This mixture of oil-lime was used in the Roman time for making a damp-proof mortar and sealing the buildings (Dorrego, Carrera and Luxán 2004). We have to make sure that some kinds of pitch were compatible with olive oil while others were not. Approaching amphorae from an emic point of view, i.e. to separating the amphora types ‘in more or less the

same way as the makers and users would presumably have done’ (Adams and Adams 1991: 223), resulted in their separation in groups of containers with a precise purpose such as wine amphorae, olive oil amphorae and fish products amphorae. From this point of view, it is revealing how a modern potter can determine the function of a vessel just by looking at a picture of an item and examining its height, neck, size and the position of the handles (Ionas 2000: 50, note 3). Only after clarifying the first use of an amphora can we tackle its many reuses. I consider that, for understanding amphora production, we need to take into account at least three elements that were considered by a landlord who exported these agricultural products: predictability, perishability and profitability. The first element, predictability, was decisive in establishing the quantities of an amphora manufactured at a certain estate. A clear exemplification is provided by three Egyptian papyri. In Roman Egypt, the leases of the pottery workshops begin with the new year, after 1 Thoth (29 August), and the potter will deliver ‘winter manufacture, well fired and coated with pitch’, therefore these amphorae were ready to receive wine in spring when the wine was transferred from dolia into amphorae, ready to be conserved or shipped (Cockle 1981: 90). By the time of hiring the potter, the landlord knew how much wine would be made during that season. This is the reason that the quantities of amphorae differed in three different leases (Cockle 1981: 91) and perhaps the reason that the landlord preferred to hire itinerant potters on short‑term contracts, rather than to have potters hired on permanent contracts. The latter type of contract was probably used for potters who worked for a city or for an imperial domain, or for the state, in the outskirts of a city, in a special ‘potters’ quarter, to meet the needs of a constant and large agricultural output. We should not forget that the state was supposed to provide amphorae for the wines exacted as taxes from the wine growers (Grace and Empereur 1981: 421–422, note 12). Large olive oil amphora workshops such as those from North Africa, Spain or western Asia Minor knew how much oil was needing to be bottled for the tax payment and how much, more or less, was available for free trade. We should also not forget the technological advance that came about in the processing of wine and olive oil by using screw presses perhaps from the end of the Hellenistic period (Lewit 2012). From the physical differences above mentioned, resulted the second variable, the perishability of these products, which also was a decisive element in the use and the timing of export of these agricultural and piscatorial products. Vintage wine was more profitable if it was preserved for many years until it reached a certain maturity and this made it more valuable, being preserved, sometimes for many years, in amphorae or dolia, either in the cellar of the producer, or of the buyer cellar, or in some entrepôts.

356

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

In addition, as most of the traded wine was of low quality, quite often this wine was subject to organic deterioration and a class of aerobic bacteria oxidized the alcohol of the wine and the wine quality perished and changed into vinegar. However, even in such cases vinegar had a certain value and it was a marketable product (Frier 1983: 277). Roman jurists paid considerable attention to the wine trade trying to secure legal protection for buyers and control the powerful network of negotiatores vinarii (Frier 1983). If wine dolia were partially buried, to keep the wine cool, olive oil dolia sat on the ground so the oil cannot become turbid (Curtis 2015: 176). Olive oil can preserve its qualities for a maximum two years. According to Piscopo and Poiana ‘The olive oil storage could be as short as possible: its consumption could not exceed the production year: the oil gains in excellence about 3-8 months after production, after that its quality starts decreasing […]. The packaging in the selling container is the last unit of operation: it is very important for product stability, maintaining high quality level and also value added’ (2012: 207). Therefore, it must be exported as soon as possible, in available high-quality containers. Thus, for example, of over 100 Panathenaic amphorae only two or three were pitched inside (Hautumm 1981: note 172). If we have in mind the tons of olive oil that were produced annualy in specialized regions, olive oil was impossible to be locally stored. Perhaps we should view in this light the imperial edict (CTh 13.5.33, 409 E) where shippers were prohibited to deviate from the course set and the violation of this ruling carried the death penalty. Fish must be processed immediately, closer to its catching points, where we should also find amphora workshops for bottling the garum. If the catch was unexpectedly very abundant, the best method of preservation was that part of it should be transformed into garum and liquamen, and bottled in special amphorae while another part should be salted and exported either in special, sturdier, amphorae for salsamenta or, in their absence, in reused amphorae. An excellent example is the Di Grado shipwreck where garum was exported in specially designed amphorae while salted fish was in reused amphorae (Auriemma 2000); another example is the export of garum of the Officina del Garum degli Umbrici in small urcei, while other parts of fish, which cannot be used for garum, were salted and exported in reused amphorae (Peña 2007: 82–87). Also, olive oil amphorae manufactured in huge workshops, e.g. Leptiminus, could be used, in extremis, to carry fish products. And here we come to another important variable that needs to be pointed out: profitability. Many imported amphorae, due to their physical qualities, were desirable containers for many forms of reuse. Indeed, according to epigraphic information, there was a certain hierarchy even among these reused containers as some amphorae retained a higher value even after they were empty. Acquiring amphorae that had been emptied of their content at auctions was a common practice that began at least in the fifth century BC and continued until the late

Roman and early Byzantine periods. This is the case with the Chian and Erythraean amphorae of the 5th century BC that retained a higher value than other amphora types (Amyx 1959: 175–178; Pliny NH 35.46). For small businesses, such as those revealed by Peña (2007: 82–98), it was necessary to have such containers, acquired at a lesser price than that of new amphorae, which were needed for less expensive products, such as salted fish, table wine, vinegar, meat etc. Vinegar, an important culinary ingredient and preservative (Dalby 2003: 343), was also a significant portion of this trade and we can see that sometimes it was exported in even larger quantities than wine (for example, a tablet from Puteoli mentions the ship Octa which carried six medium‑size amphorae of wine and seventy‑seven of vinegar cf. Camodeca 1999: 184; for vinegar carried in skins see Pliny NH XXIII, 56 and the discussion in Brun 2003: 92). It was a usual product of everyday life, stronger than wine and was used mixed with water, satisfying the thirst and without the drunkenness effects of the wine. In the 4th century the daily ration of vinegar was of 0.15 liters (Le Roux 1994: 404). In addition, as Pliny (NH 14, 128) advised, it was better to transfer the wine in amphorae that had contained vinegar. In economic terms, it would be unjustified to add the high price of a newly made amphora to a cheap product that was traded mostly on the local or provincial market. The best form of pitch was applied to an amphora when it was still hot after removing it from the kiln (Brun 2003: 69); at this moment, the pitch would close the pores (Rice 1987: 163; Arnold 1985: 140). Later repitching would not have the same properties as the first pitching, regardless of how well the cleaning of the amphora interior was done and thus it would be risky to fill that vessel with a vintage wine or virgin olive oil. Only rarely were these products considered valuable enough to be the subject of a long‑distance trade, and the second-hand amphorae filled with salted fish found in the Di Grado shipwreck might represent an exception. It is clear that we cannot establish an exact pattern of the reuse of these containers. However, these reused amphorae had a limited economic value that cannot impede our understanding of the economy of a certain site or province. What really mattered in economic and quantitative terms were the millions of amphorae of prime use manufactured in specialized workshops, which were designed to carry a specific product, vintage wines, olive oils or garum, to large consuming markets either through the annona or private systems. 2. Wine production (1st century BC–mid-1st century AD) 2.1 Lower Danube/sub-Carpathian wine amphorae To the end of the Hellenistic period, this area experienced the same phenomenon, already attested since the Classical period, of acclimatization of new, stronger vines, able to produce a vintage wine that lasted for

357

Andrei Opaiţ

Figure 1. Map-Romanian vineyards areas. many years, getting more valuable, and becoming an important complement to the prestige of the local elites. During the Hellenistic period, this phenomenon became widespread. It was not only due to the sickness of old vines (Billard 1913 (1998): 376–392), but also the desire to achieve a greater productivity (Sealey 1985). It is attested in Italy by adopting the vine of Thasos (Thasiavitis) (Dalby 2000: 149), while in Maronea the Phoenician vine was naturalized, transplanted from Maronea in Egypt (Salviat 1990: 462–464). The latter region also attracted other Greek vines such as the Knidian and the Chian (Rostovtzeff 1922: 95; Forbes 1965: 119); the Chian wine was also transplanted in Galilee (Shipley 2000: 345). Some archaeological discoveries add weight to this phenomenon, such as those fragments of vines (with earth adhering) found on the El Sec shipwreck (Parker 1992: 393). A similar process of adaptation of new Aegean vines took place in the Lower Danube/sub-Carpathian area. Inside this area, we witness the occurrence of Rhodian and Coan amphora imitations. This, in our opinion, attests to a successful adaptation of these vines to the local soil. Our situation is similar to that of Gaul, where the local population learned how to create new vines by cutting imported vines (Pliny, NH XVII, 116), or to orientate the new vineyards (Pliny, NH XVII, 2, 21). In addition, an

impressive technical progress was the local production of amphorae, which involved choosing a certain clay and processing it, effective kilns, controlling the kiln’s temperature, knowledge that could come only through a profound Greek influence. We witness, therefore, an abandonment of the subsistence economy, creating a superior output ready for export, and a network for distributing this vintage wine on the provincial market (Glodariu 1976: pl. 2). Based on the discoveries of these amphorae I can suggest that this process started at the end of the 2nd or the beginning of the 1st century BC and continued until the middle of the 1st century AD, including the sub-Carpathian area from Oltenia to the northern Moldavia, as well as north eastern Bulgaria (Varbanov 2011. Basically, it is the same area that is renowned even nowadays for its Romanian vintages — Figure 1 —). So far, the Rhodian vessel imitation is indicated by some fragmentary amphorae with handles that are rounded in cross section. They have rectangular or rounded stamps with anepigraphic or different geometric representation (Figure 2.1a‑c) (Glodariu 1976: pl. 18‑20). The shapes of Coan amphorae were also imitated, having either bifid or pseudo bifid handles. Two such upper part amphorae, discovered at Răcătău, were covered by a dark red slip, perhaps for a better isolation of the walls.

358

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

this time the shape of Colchian amphorae is conical with a pronounced ‘waist’ at the middle of the body (Figure 3.9) (Tsetskhladze and Vnukov 1992; Vnukov 2003: 160–194, fig. 66; 2011 and 2017; Kovalevskaja and Sarnowski 2003: fig. 3.2). Although I have minor disagreements regarding the dating of some variants, Vnukov’s typology remains the most reliable and helpful for understanding the amphora production of this area at the beginning of early Roman times. These transport amphorae were used only for regional trade, their main market being not only the northern and western Pontic shores but also the northern steppes and sub-Carpathian region. Worth noting is the presence of the local table wine on the western and northern Pontic coasts, represented by table amphorae and pitchers (Figure 3.10) (Strjeletskiiy et al. 2005: pl. 33.374). 3. Wine production (mid-1st–3rd century AD)

Figure 2. (1) Stamped Getic amphorae, after Vulpe 1965: fig. 3. In parallel with this production for the high end of the market, cheap, table wine was also massively produced as the literary (Strabo, Geogr. VII. 3,11) and archaeological sources (Opaiţ 2013: 38–45, figs 5‑6) attest. 2.2 South and East Pontic wine A similar phenomenon of Aegean amphorae imitations is recorded not only on the southern Pontic coast but also far inland of the Anatolian plateau at Pompeiopolis. The recent petrographic and morphologic studies undertaken especially by S. Vnukov shed a new light on the south and east Pontic amphorae. He identified some Heracleian (Figure 3.2‑3) (Vnukov 2003: 28–102, types S I and S II, figs 10, 34) and Sinopeian (Figure 3.4) (Vnukov 2003: 133–141, type Sin II, III, V) imitations of Coan and Rhodian amphorae. However, in parallel with these imitations, these cities manufactured amphorae with their own tradition (Figure 3.5‑6) (Vnukov 2003: 102–117, type S III, fig. 38; 130–133, type Sin I, fig. 51/1‑4). Both cities also exported table amphorae (Vnukov 2003: 118–128; 147–156; S IVA1 and Sin IV) (Figure 3.7‑8). The spread of the Coan vine and the imitation of Coan amphorae seem to be wider than we previously thought as the recent excavations from Pompeiopolis revealed the presence of this Pseudo‑Coan amphora locally made at that site. The Sinopean and Heracleian wine seems to have been well-appreciated, reaching the Crimea (Vnukov 2003: 147–156), the east Carpathian area, and Dobrudja (Opaiţ and Paraschiv 2012: 116–117, fig. 7; Histria, personal observation in the museum’s store room). The evolution of Colchian amphorae from Hellenistic to early Roman times is also presented by Vnukov. During

The Roman position was well-consolidated at the Lower Danube beginning with the reign of Claudius, and after the conquest of Dacia Roman power was certainly more secure in this area. The presence of over 100000 Roman soldiers (Protase 2010: 114; Matei-Popescu 2010: 275; 2015) to which we should add numerous veterans, administrative officials and colonists that arrived here ex toto orbe Romano, made the demands of this new market very impressive. These demands were impossible to satisfy only through long‑distance trade, either subsidized or free. Basic staples such as grains, table wine and fish products were provided by local production, leaving the burden for providing the other basic foodstuff, the olive oil, to the state, as this was impossible to procure from this area, the Sinopean production being enough only for a very limited, local area. Therefore, the main forms of trade in wine and fish products were provincial and intra-regional. 3.1 South and East Pontic wine Between the mid-1st century and the end of the 3rd century AD Heracleia developed a series of amphorae that evolved from amphorae of Pseudo‑Coan into type Shelov A/Vnukov type Sin IV A1 into Shelov B/Vnukov type Sin IV A2, Shelov C, and D. It becomes smaller and the capacity is reduced from 6‑7 litres between the end of the 1st century AD and mid‑3rd century to almost 3 litres in the second half of the 3rd century AD (Figure 4.11‑13). What is remarkable with this amphora is its narrow, long neck, a morphologic characteristic that might be due to a specific quality of the Heracleian wine. Quite frequently these amphorae bear stamps or dipinti on the neck, perhaps the officinator or owner names (Il’yashenko 2013: fig. 3). The Sinopean amphorae of type Vnukov Sin II changed the pear-shaped body into an ovoid shape of type Snp B I and Snp B II, according to Kassab Tezgör’s typology that discovered this amphora type in the workshop of Demirci (Figure 4.14) (Kassab Tezgör 2010: 126, pl. 15. 1‑2).

359

Andrei Opaiţ

Figure 3. (2) Pseudo Coan-Heraclea. (3) Pseudo Rhodian-Heraclea. (4) Pseudo Coan-Sinope. (5) Sinopean amphora. (6) Sinopean amphora. (7) Table amphora-Heraclea. (8) Table amphora-Sinope. (9) Colchian amphora-after Kovalevskaja and Sarnowski 2003: fig.3.2. (10) Table amphora-after Strjeletskiiy et al. 2005: pl. 33.374.

360

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

Figure 4. (11) Heraclean amphora-Noviodunum, courtesy of Archaeological museum, Tulcea. (12) Heraclean amphora-Noviodunum, courtesy of Archaeological museum, Tulcea. (13) Heraclean amphora-Gura Portiţa, courtesy of Archaeological museum, Tulcea. (14) Sinopean amphora, after Strjeletskiiy et al. 2005: pl. 10.1. (15) Amastris (?), after Arsen’eva and Naumenko 1992: fig. 26.1. (16) Colchian amphora, after Vnukov 2003: fig. 66.347. (17) Colchian amphora, after Vnukov 2003: fig. 66.342.

361

Andrei Opaiţ

Figure 5. (15) Amastris (?), courtesy of T. M. Arsen’eva and S. A. Naumenko. An amphora type showing a strong Sinopean influence, with a very similar body, was manufactured somewhere on the southern Pontic coast (Amastris?) and it is known in the literature as type Zeest 76/77 (Zeest 1960: 113‑114, pl. 32; Arsen’eva and Naumenko 1992: 146, 159, figs 26, 40; Naumenko 2012: 70, figs 11.1; 15.12). Many examples have been found at Tanais, and Gorgipija (Alekseeva 1997: plates 95.1,3; 112.13; 134.6; 166.12; 222.6). Remarkable is the engraved stamp of Faust(i)nos (Φαυστεῖνος/ Φαυστείνου) that occurs on at least two examples (Figure 4.15 and Figure 5) (Kruglikova 1966: 208–209; Kropotkin 1970: 13; Shelov 1972: 124; Krapivina 1993: 99) — I am deeply grateful to T. M. Arsen’eva and S. A. Naumenko for allowing me to publish the picture of this amphora and stamp). Colchian amphorae had a more ovoid body perhaps due to a Sinopean influence, but two shapes of the rims, one in a cup-shape and another in a chalice-shape, and different fabrics suggest the existence of different workshops (Figure 4.16‑17) (Vnukov 2003: fig. 66.3; 2011; 2017; Opaiţ 2015). During the early Roman period, the wine production of these areas recorded a real boom and it supplied a large area. The Heracleian and Sinopean wines were constantly

present on the northern and western Pontic shores and penetrated deep in the Lower Danube reaching Pannonia (Sinope: Nagy 2015: 186; Heracleia: Harshegyi 2008: 176, fig. 10), Dacia (Ardeţ 2006: 103–104, pl. XIV.122; Bondoc 2014: 109, fig. 9‑10), Moesia inferior (Baumann 1995: 91, no.25, pl. 67.2), and also the north of the Danube in Moldavia (Croitoru 2011, plates 43‑45). In addition to this intra-regional trade some Sinopean, Heracleian and Colcheian amphorae have been found not only in the eastern Mediterranean in Crete (Sinope: Hayes 1983: 151, types 26, 27, fig. 24.A67, A68, A70; Heracleia: Hayes 1983: 147, type 14, fig. 21.A32), Athens (Heracleia: Opaiţ 2010: 110–113), Corinth (Heracleia: Slane 1986: 298, no.124, fig. 18), Beirut (Sinope: Reynolds 1997‑1998: fig. 202), Caesarea (Johnson 2008: 103, nos1236, 1237), but also in Italy at Rome (Heracleia: Rizzo 2014a: 563, fig. 6.2), Napoli (Sinope and Colchis: Miniero 2008: 171), and Iulia Concordia (Sinope: Belotti 2004: 80–82, fig. 17). 3.2 North Pontic wine The early Roman period experienced an economic boom, seen in the north Pontic wine industry. This activity is attested by many wine presses and platforms for trampling grapes discovered at Chersonesus (Carter and Mack 2003: 84–85, fig. 8.33, 36), and Panticapaeum

362

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

(Vinokurov 1999). The presence of many local transport amphora types is also remarkable. Although some potters tried to reinvigorate the local Hellenistic shape (Figure 6.18) (Strjeletskiiy et al. 2005: 69, pl. VII.1), or to imitate the Sinopean type Snp A I (amphorae à col renflé) in a local fabric (Figure 6.19) (Strjeletskiiy et al. 2005: 67, urn 4, pl. 37), the most distinctive Crimean amphorae of this period are types Zeest 72 and 73 (Figure 6.20‑21) which are basically variants of the same type (Zeest 1960: 111–112; Strjeletskiiy et al. 2005: pl. V.2, urn 96; Klenina 2004: fig. 5. 39‑47). These latter amphorae were manufactured also in fractionary sizes (Figure 6.22). Thus, these volumes varied between 7l, 19l, 63.5l, and 86l. The amphora types enjoyed not only a regional distribution being traded in the Roman provinces of the Lower Danube (Opaiţ 2004: 27–28), or Pompeiopolis in Paphlagonia (Personal observation in the Pompeiopolis excavation store room), but also have been found in the eastern Mediterranean in Crete (Hayes 1983: 153–156, types 34, 35, 43; Sackett 1992: 255, plates 198.133-136; 199, 145), Caesarea (Johnson 2008: 112, no.1350-1351; 113, no.1363); and Beirut (Reynolds 2010: 93, fig. 2.c-d). In addition to this vintage wine that travelled intra and inter-regions there were table wines as the table amphorae (Sedikova 1996: fig. 2.1) and table pitchers fully attest to (Figure 6.23-25). 3.3 West Pontic/Lower Danube wine These areas did not develop a vintage wine as no transport amphora workshop has been identified so far. However, the significant presence of table amphorae and table pitchers attest to a considerable boom of cheap wine, enough to satisfy the huge number of soldiers garrisoned in this area. A strong competition seems to develop between them during the early Roman period, but in the end the former container was the winner, and it continued to be the typical container for local wine between the 4th and the 6th centuries AD (Opaiţ 2003). Usually table amphorae have a globular body and a ring base with two sturdy handles while the table pitchers have a cylindrical neck, a strap like handle, and a conical body that ends in a trapezoidal, tubular base (Figure 7.26‑28) The table amphorae have capacities varying between 18 and 25 litres, while the table pitchers have volumes varying between 10 and 15 litres. Both vessel types are present in the urban as well the rural milieu of the Lower Danube (Opaiţ and Paraschiv 2012; Opaiţ 1980: 291–294, plates I‑II; Bondoc 2014: 111, fig. 15; Nuţu et al. 2014: 58–61, plates 12‑15; Bjelajac 1996: 99–100, pl. XXXV; Kabakchieva 1986: 21, plates 23‑27). The richness in wine of this area is attested also by many other archaeological discoveries (Pilipović 2013). It is worth mentioning the occurrence of a table wine amphora that seems to imitate the Forlimpopoli amphorae (Bjelajac 1996: 22; Ardeţ 2006: 71) and which experienced inter-provincial trade both in Moesia

Superior, Pannonia, and Dacia Inferior. I would not exclude, therefore, a transplantation of vines from northern Italy into this mid-Danube area. It has been found at Viminacium, Singidunum (Bjelajac 1996: 22‑25, type III, fig. IV. 31‑33; Brukner 1981: plates 165‑166), Aquincum (Harshegyi 2008: 174, fig. 7), Pojejena (Ardeţ 2006: 71–72, pl. VI.61) (Figure 7.29), and Tibiscum (Figure 7.30) (Ardeţ 2006: pl. VII.65). It is also worth pointing out the appearance of table pitchers in a very restricted area north of the Danube (Croitoru 2011: map 46). However, in the northern Danubian area, a local table amphora, which perhaps has its origins in a Hellenistic container, dominates (Figure 7.31) (Bichir 1973: 80–81; plates 112.4; 114‑119; Croitoru 2011: 133, no.1810, fig. 69.1810). Additionally to these ceramic containers, the local population and the soldiers used barrels. Although archaeological excavations cannot document the barrels, the famous Column of Trajan (Lepper and Frere 1988: 49, pl. V; 104, pl. XLII‑XLIII; Marlière 2002: 150–151, R59‑R62) and some bas-reliefs depicting barrels offer enough support for the presence of these containers. One of them was has been found in Dobrudja (Dragomir 1962; Bărbulescu 2001: 206; Covacef 2002: 129–130; Bordenache 1969: no.115, pl. LI) while from Bulgaria (Novae) comes a funerary stela dedicated to Iulius Iero, who was a negotiator vinarius (Kolendo 1965: 132–138). Another three funerary bas‑reliefs, also discovered in Bulgaria, have representations of barrels carried by carts (Kolendo 1965). The presence of a intra-provincial wine trade between Thracia and Moesia Inferior with Dacia is also epigraphically attested (Ardevan 1988; Bounegru 2002: 94; Matei-Popescu 2012: 88–89). 4. Processed fish production (1st–3rd century AD) This important economic activity that took place in the Black Sea area is one of the best known, perhaps less due to literary and epigraphic information (Curtis 1991; 2001; 2005; Wilkins 2005; Lund and Gabrielsen 2005) but more due to archaeological discoveries such as salting factories and amphorae for fish products (Højte 2005). As my opinion on Pontic amphorae for fish products has been recently published I will not elaborate on this topic (Opaiţ 2007). In comparison with agricultural products such as wine and olive oil, piscatorial activity was more dependent on environmental factors. In some areas fish was caught on a regular basis, especially around the straits or the big rivers such as the Danube, the Dniester, the Dnieper, the Bug, the Don and the Kuban, in others not. Large fish that can reach 2‑3m, such as sturgeon, pike, and catfish were a regular catch in the slow currents of these rivers. Perhaps this was the reason that fish product amphora workshops were not developed everywhere. Ultimately there were the environmental factors and the cultural traditions that played a decisive role in determining a certain shape of amphora.

363

Andrei Opaiţ

Figure 6. (18) Pontic amphora, after Strjeletskiiy et al. 2005: 69, pl.VII.1. (19) Pontic amphora, after Strjeletskiiy et al. 2005: pl.37.urn 4. (20) Pontic amphora-Zeest 72, after Strjeletskiiy et al. 2005: pl. V.2, urn 96. (21) Pontic amphora-Zeest 73, courtesy of Chersonesus archaeological museum. (22) Pontic amphora-Zeest 72-small size, after Strjeletskiiy et al. 2005: pl. 42, urn 75. (23) Table amphora, after Sedikova 1996: fig. 2.1. (24) Table pitcher, after Strjeletskiiy et al. 2005: pl. 42, urn 81. (25) Table pitcher, after Strjeletskiiy et al. 2005: pl. 43, urn 101.

364

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

4.1 South Pontic amphorae for fish products: Sinope Sinope, besides manufacturing wine amphorae, which followed a Hellenistic tradition, developed an amphora type that, due to its morphologic characteristics, carried a processed fish product. This amphora was named ‘Type B Snp III’ by Kassab Tezgör in her typology (2010: 126, pl. 15.3-4). The differences between this vessel and the wine vessels (Type B Snp I and II) are striking. The neck is troncoconic, with no separation from eggshaped body that ends in a conic base, which is a typical base for the Sinopean amphorae. In this way it ensured an excellent fluidity for the pouring of presumably high viscosity contents (Figure 8.32). The height is of c. 90 cm, the maximum diameter of 47cm, and the capacity is of c. 67 litres. This amphora is frequently found in Dacia, especially in military milieu, at Drobeta, Tibiscum, Ulpia Traiana Augusta Dacica Sarmizegetusa, Romula, Micăsasa, Dierna (Ardeţ 2006: 108–110, pl. XIX, XX.166-173), Napoca (personal communication Viorica Rusu-Bolindeţ) Cioroiul Nou (Bondoc 2014: 109, fig. 11), in Pannonia Inferior at Aquincum (Hárshegyi 2016: 628, fig. 2.12), while in Moesia Inferior it occurs at Troesmis (personal observation in the Museum of History Tulcea), and Durostorum (personal communication St. Honcu). In the Black Sea, besides Sinope, it occurs at Samsun (personal observation in the Samsun museum), Sebastopolis (Sukhum-Abkhazia) (personal communication S. Vnukov), and perhaps Gorgippia (Alekseeva 1997: pl. 118.A; 119.4). It is worth noting that this amphora has been found in the eastern and central

Figure 7. (26) Table amphora-Troesmis, after Opaiţ 1980: pl. 1.4. (27) Table amphora-Callatis, after Opaiţ and Ionescu 2016: pl. VIII.46. (28) Table pitcherCallatis, after Opaiţ and Ionescu 2016: pl. IX.48. (29) Table amphora-Mid Danube-Pojejena, after Ardeţ 2006: pl. VI. 61. (30) Table amphora-Mid Danube, after Ardeţ 2006: pl. VII. 65. (31) Table amphora-courtesy of Bacău archaeological museum.

Figure 8. (32) Sinopean fish amphora, after Ardeţ 2006: pl. XXIX.163.

365

Andrei Opaiţ

Figure 9. (33) Zeest 75-Greci, after Opaiţ 1980: pl. X.3. (34) Sovhoz 10, after Strjeletskiiy et al. 2005: pl. VII.2. (35) Miniature Zeest 75-courtesy of Chişinău archaeological museum. Mediterranean at Corinth (Slane 2000: fig. 14.g), Crete (Hayes 1983: 151, fig. 24.69; 155, type 36, fig. 25.80, could be also a south Pontic fish product amphora; Sackett 1992: 255, pl. 199.147), Caesarea (Johnson 2008: 103, no.1234), Milano (Bruno 2003: 88, fig. 3.10), and Rome (Rizzo 2014a: 349–350, fig. 52; Rizzo 2014b: 559, fig. 4.1; 563, fig. 6.4). 4.2 North Pontic amphorae for fish products As I showed in a recent paper, Pontic fish amphorae have the same structure as Mediterranean fish product amphorae: a wide mouth, a large tronconic neck and an ovoid body ending in a massive spike (Opaiţ 2007). Although only one, Panticapaean amphora workshop has been identified, the pattern of their distribution and some local fabric characteristics, gives us some indications. I will give only a brief presentation of some, very common, amphora types. 4.2.1. Type Zeest 75 Is one of the well-known types. The body is large, ovoid, ending in a massive spike; the rim is set on a large, tronconic neck and the handles are massive with a deep

groove on the external and internal sides. The body and the neck are covered by shallow grooves. The shape of the rim suggests the existence of at least four contemporary variants: A. It has a thick rim, triangular in its upper part, well pointed towards exterior and separated by an offset at its lower part. In addition to the fullsized ovoid body discovered in Dobrudja at Greci (Figure 9.33) (Opaiţ 1980: 306–308, type XI, pl. X.3; XII.4), Slava Cercheză (personal observation in the Archaeological Museum of Tulcea), and Olbia (Zeest 1960: 113, pl. 31. 75a), there is a fractionary size having a smaller, conical body discovered in the cemetery Sovhoz no.10 in the Chersonesan territory (Figure 9.34) (Strjeletskiiy et al. 2005: urn 97, pl. VII.2), Balaklava (Klenina 2000: 123–134, fig. 25.3), Bliznetsy (Klenina 2004: 24, type 4, fig. 7.70), the territory of Olbia (Burakov 1976: 71, pl. II.16; Malyukevich 1991: 73, fig. 2.14),Tanais (Arsen’eva and Naumenko 1992: 40–41, fig. 18), and Gorgipija (Alekseeva 1997: pl. 127.2). If the volume of the full-size subtype is c. 194l, the small size capacity is only one third of it, c. 66l. It is worth noting

366

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

Figure 10. (36) Zeest 75-Tomis, courtesy of Muzeul de Istorie Naţională Constanţa. (37) Zeest 75-Kalos Limen, courtesy of Evpatoria archaeological museum. (38) Zeest 75-Sovhoz 10, after Strjeletskiiy et al. 2005: pl. V.6. the occurrence of a miniature copy at Chişinău. Most likely it was a sample vessel for degustatio (Figure 9.35). B. It has a massive, rolled rim which has small facets and it occurs at Tomi (Figure 10.36) Chersonesos (Strjeletskiiy et al. 2005: grave 127, pl. VI.1), Balaklava (Klenina 2000: 123–134, fig. 25.2,5), Olbia (Krapivina 1993: 99, figs 30.26; 72.15), Bliznetsy (Klenina 2004: 24, type 4, fig. 7.68, 69), Tanais (Arsen’eva and Naumenko 1992: 144–145, fig. 24.2; 29.3), and Gorgipija (Alexeeva 1997: pl. 90.6). C. This variant has a trapezoidal rim with its top rounded and a marked offset at its lower, external side. A large example has been discovered at Kalos Limen (personal communication V. Stolba) (Figure 10.37). A small-size variant has been discovered in the necropolis of Sovhoz no.10 near Chersonesos (Strjeletskiiy et al. 2005: 68, urn 105) (Figure 10.38). D. The last variant has an almost rectangular massive rim, with its top bevelled towards the exterior; it occurs at Histria (Figure 11.39) (Zeest 1960: 113, pl. 31.75б), Olbia (Krapivina 1993: fig. 72.16), and Gorgipija (Alekseeva 1997: pl. 106.11; 124.12). A small-sized variant has been discovered in the necropolis of Chornorichens’kiiy (Babenchikov 1963: 119, pl. XVI.10). 4.2.2. Type Zeest 76 similis /Kelemen 20 This is also a massive and heavy amphora with many variants. The rim is massive and rolled, but sometimes it is bevelled at the exterior and hanging at its lower edge. The handles are also massive, ovoid in section, but

367

Figure 11. (39) Zeest 75-Histria, after Zeest 1960: pl. XXXI. 75б.

Andrei Opaiţ

with a deep cut on the internal side (Figure 12.40). Many variants occur in the Chersonesan territory (Strjeletskiiy et al. 2005: 70, urn 237, pl. VII.4), Mirmekion (personal observations on the site), and Tanais (Shelov 1965: 68, fig. 11 top; Arsen’eva and Naumenko 1992: 144–145, fig. 20). Perhaps variants of this type have been discovered at Ostia (Panella 1986: fig. 26; Rizzo 2014b: 563, figs 6.7, 6.8), Caesarea (Johnson 2008: 113, no.1364), and Knossos (Hayes 1983: 155, type 39, fig. 25.91). An interesting discovery comes from Aquincum. This amphora has a large dipinti: LEG, which suggests a special delivery, perhaps of a fish product, to the legion based at Aquincum (Kelemen 1990: 172–174, fig. 5.2‑4). 4.2.3. Type Zeest 83 This is an amphora made in the Bosporan kingdom (Figure 12.41). The handles are round in cross section but the rim is mostly flattened at exterior, slightly thickened and separated from the neck only by a small offset. One of the workshops that manufactured this amphora has been recently discovered at Panticapaeum (Yermolin and Fedoseev 2013). One example occurs at Callatis (Opaiţ and Ionescu 2016: 69, pl. XV.93). 4.2.4. Type VI-Sovhoz This is a less known amphora type and only a few examples have been discovered so far. It was also manufactured in at least two sizes. The rim is bevelled towards exterior; the neck is tall, broad sloping; the massive ovoid handles with a central rib are arched reaching almost the rim level; the conical body ends in a massive spike. Two amphorae discovered at Sovhoz no.10 illustrate these two variants (Figure 12.42‑43) (Strjeletskiiy et al. 2005: 70, grave 184, plates VII.3, 24.184, urn 237, pl. 49.237). Similar amphorae were discovered at Balaklava (Klenina 2000: fig. 25.1), Kozyrskoe (Burakov 1976: 75, pl. IV.1-3), and Tanais (Alexeeva 1997: 142, fig. 21). It has also been found on the Lower Danube at Viminacium in Moesia Superior (Bjelajac 1996: 64, fig. XX.114), and in Dobrudja (Tulcea museum, inv. nos 38580 and 46255). 4.2.5. Table amphorae The occurrence of this type of amphora is very interesting as it seems to mirror the wine table amphorae. However, their morphological features, which are very similar to the fish product amphorae, make me consider them as part of the fish product containers. They also might contain samples for tasting, the seller in this way avoiding having to open a large amphora before selling it. Among them I distinguish at least three variants. A. This variant is characterized by its sloppy, hanging, thin rim; a large mouth, a tall and large neck; a handle ovoid in cross section with a central sharp groove; short, sloppy shoulders; an ovoid body with flaring ring base. Its height is of c. 50cm. Its capacity is of c. 7.7l (Figure 13.44). It has been found in the necropolis Sovhoz no.10, near Chersonesos (Strjeletskiiy 1959:

139–145; 142; Strjeletskiiy et al. 2005: 71, type X, pl. VI.6), Bezymyannaya (Chersonesan chora) — personal observation in the Chersonesos museum —, Tanais (Arsen’eva and Naumenko 1992: 157, fig. 38.1), Gorgipija (Alexeeva 1997: pl. 131.2), and Mirmekion (Gajdukevich 1987: 171 fig. 191). The fabric of the jars from Sovhoz no.10 and Bezymyannaya indicates a local production probably around Balaklava. Interesting is the occurrence of a rim fragment at Pompeiopolis but in a different fabric and that suggests the existence of a multitude of workshops. B1. This variant has been discovered at Olbia and has smaller dimensions, i.e. a height of 31cm and a maximum diameter of the body of 18cm (Figure 13.45). A whitish wash covers the exterior of the amphora. It has a rim diameter of 10cm, height of 31 and maximum diameter of 18 (Krapivina 1993: 98, type 27, fig. 30.25). A variant B2 with a similar beaded rim but a wider, tronconic, ribbed neck, conical body and ring base has been found at Pompeiopolis (Figure 13.46). The rim diameter is of 9cm, the height is of c. 25cm, and maximum diameter is of 19cm. The fabric is red with abundant brownish inclusions (iron oxides?) and few white, large (calcareous?) inclusions. It might be a south Pontic production. C. The last variant comes from Tanais and it really looks like a miniature amphora. The preserved height of 17cm, RD of 4cm, and MD 10cm (Figure 13.47). The fabric is light with black inclusions and mica (Arsen’eva and Naumenko 1992: 56, fig. 73.6), suggesting perhaps an imported product. Although the authors do not date this jar, the bifid handles hint to an earlier date of the end of the Hellenistic or beginning of the Roman times. 4.3. West Pontic fish products amphorae In spite of the richness in fish of this area of the Black Sea we do not have too much archaeological information regarding the presence of the fish processing activities. This is not due to a lack of archaeological activities, but perhaps to some environmental factors such as the absence of a stony coast similar to the Crimean coast. Also, the continuous advance of the Danube Delta perhaps eventually buried the ancient fish processing installations. However, the written information and some archaeological discoveries suggest that piscatorial activities were well-developed in this region. One of the most important pieces of information regarding the fish processing that took place in the region of the Danube Delta is provided by a famous inscription, the Horothesia of Laberius Maximus, the legate of Lower Moesia in AD 100. This inscription of the second half of the 2nd century or the beginning of the 3rd century, reiterates a series of previous decisions taken by Roman legates of the 1st century AD with regard to the people of Histria because ‘the town is so great a distance from the mouths of the river […]’ and ‘the revenue there from

368

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

Figure 12. (40) Zeest 76 & similis-Sovhoz 10, after Strjeletskiiy et al. 2005: pl. VII.4. (41) Zeest 83-Gorgipija, after Alexeeva 997: pl.149.2. (42) Sovhoz 10, after Strjeletskiiy et al. 2005: pl.VII.3; 24.184; urn 237, pl. 49.237. (43) Sovhoz 10, after Strjeletskiiy et al. 2005: pl. 49.237.

369

Andrei Opaiţ

Figure 13. (44) Table amphora-Sovhoz 10, after Strjeletskiiy et al. 2005: pl. VI.6. (45) Table amphora-Olbia, after Krapivina 1993: fig. 30. 25. (46) Table amphora-Pompeiopolis. (47) Table amphora-Tanais, after Arsen’eva and Naumenko 1992: fig. 73.6.

Figure 14. (48) Histrian amphora stamp-Noviodunum, courtesy of Tulcea archaeological museum. pickled fish (ταρειχευομένου ίχθύος) is just about the city’s only revenue […]’ the governor ‘decided that the same freedom to fish according to your custom at the Peuce mouth and to transport the pine wood without tax for the use of each individual should remain to you’ (Oliver 1965: 155). A most likely Hellenistic amphora handle, discovered at Noviodunum, bears two parallel stamps, with the name of an official ΘΕΟΓΝΗТΟΥ while the second has the name ΙΣΤΡΙΗ (Figure 14.48) (Opaiţ and Bauman 2006: 396). As Histria does not have good enough conditions to develop

370

Figure 15. (49) Pruteni amphora-after Vornic et al. 2007: pl. 33-36.

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

vineyards, even less for vintage wines, the most plausible product for export must be pickled fish. The frequent occurrence, in archaeological excavations in Dobrudja and eastern Romania, of a large amphora with morphological characteristics similar to the previous discussed type Zeest 76 similis, might suggest that this was the container that was used to export the Danubian pickled fish (Opaiţ 1987: 247, fig. 2.2; Opaiţ and Bauman 2006). Another archaeological discovery was an amphora workshop on the Prut river at Pruteni (Vornic et al. 2007: 91–101, plates 33‑36). The design of this amphora was most likely influenced by a north Pontic amphora, perhaps of type VI‑Sovhoz, previously discussed. The large dimensions of the mouth and the neck, and the egg-shaped body suggest a processed fish product (Figure 15.49). Also, the location of this workshop, north along the Prut, at only 3km from the river, suggests a fish related activity. Most likely this point was a good one for catching large catfish that swam up the river for reproduction. 5. Concluding remarks Beginning with the second half of the 1st century BC, profound political and economic changes took place in the Lower Danube and Pontic region: creation of new veteran colonies on the southern shore of the Black Sea, the colonization with north Danubian Getic population ad praestanda tributa (CIL XIV 3608 = ILS 986 = IIT IV 1, 125 = IDRE I 112; Pippidi 1962: 106–132; 1967), followed by the inclusion of the Lower Danube in the Roman empire, a tight Roman control of the northern and eastern coasts of the Black Sea. The transformation of the Black Sea into a Roman lake had visible economic consequences over the whole this region. It is worth noting that this area was already well-integrated into the Hellenistic world prior to the Roman arrival. An excellent proof is the adaptation of new Aegean vines not only in the Greek area of southern Black Sea but also between the Carpathians and the Danube. However, the Roman domination accelerated and increased the integration of the Lower DanubePontic basin with the Mediterranean world. Not only did this region become more integrated into the Roman economy, receiving huge quantities of olive oil, and vintage wines, either using controlled mechanisms or free trade, but also exploiting better its own agricultural and piscatorial resources. Regarding the export of Sinopean olive oil, I have serious doubts that this city, in spite of the literary testimonia (Xenophon, Anabasis 6. 4. 6; Strabo 2. 1. 5; 12. 3. 12) on the presence of oliveyards, was a large exporter of such products. Most likely this olive oil was enough to support the local necessities. Both the Hellenistic and the Roman amphora designs support only wine and fish product exports. If the Sinopean

wine is well-known on the northern and western Pontic coasts during the Hellenistic and Roman times, the occurrence of amphorae of type B Snp III, which have the morphologic characteristics of a fish product amphora, shed new light on the Sinopean economy. The processing of fish at Sinope is also documented by literary sources (Strabo 7.6. 2; 12. 2. 11-12; 12. 3.12; Pliny NH 9.18; for their discussion see Doonan 2002: 187–192; De Boer 2013: 112). The wide distribution of these amphorae not only in important cities and fortifications of the Lower Danube but also in the Mediterranean area seems to confirm the high prestige this Sinopean product enjoyed (Diodorus Siculus 37. 3. 5). Of course, this amphora type is missing on the northern Pontic shore. Interesting is also the proliferation of these small containers with a ring base, easy to move and transport. This could also be proof of its valuable content. There was certainly an important grain production in the Lower Danube that was able to supply the local garrisons as part of the annona, while the north Pontic and north Danubian grain was also exported especially in periods of crisis (Durliat 1990: 207, notes 107 and 108; Pliny N.H. 18.66). However, as the local wine production attests, the rhythm of this economic development was different inside this large region. Those areas that had a longer tradition in wine making and processing fish, such as the Pontic Greek cities, had the necessary economic background to continue and amplify it on a scale unprecedented in the Hellenistic period. Other areas, such as the Lower Danube provinces, perhaps due to the responsibility of supplying the garrisons with staples and many other fiscal obligations, managed only to develop a flourishing local wine industry as the many table amphorae and pitchers, or bas-reliefs depicting barrels attest. The vicus system created in the Lower Danube, such as the 46 vici epigraphically attested in the province of Moesia Inferior, with a population resulting mainly from colonization with veterans, at least in its first phase (Bărbulescu 2001: 282–283), was the backbone of support for the military garrisons. Basically, we witness a massive mobilization of the production forces to support a huge army with basic foodstuffs, while the other Pontic centres were able to supply luxury goods, such as vintage wine and processed fish products, to a superior echelon of this Danubian-Pontic society. There was evidently an excellent balance between different products attained by different levels of the local society. The only missing variant of this equation was the olive oil that arrived into this area either through a controlled mechanism, especially in the garrisons of the border, or by free trade in the Pontic basin. An excellent proof of this situation is the massive presence of Dr 24 and Dr 24 similis amphorae in this entire region. The huge numbers of these vessels are the best evidence of the integration of this area into the complex economic mechanism created by Roman power.

371

Andrei Opaiţ

Bibliography Adams, W. I. and Adams, E. W. 1991. Archaeological Typology and Practical Reality. A Dialectical Approach to Artifact Classification and Sorting. Cambridge, Cambridge University Press. Alekseeva, E. M. 1997. Antichnyiiy Gorod Gorgippij. Moscow, Editorial URSS (in Russian). Amyx, D. 1959. The Attic stellai: part III. Hesperia 27: 163–310. Ardeţ, A. 2006. Amforele din Dacia Romană. Timisoara, Mirton. Ardevan, R. 1988. Porolissum si Augusta Traiana. Observaţii asupra inscriptiei IGB, III/2, 1590. Acta Musei Porolissensis 12: 291–295. Arnold, D. E. 1985. Ceramic theory and cultural process. Cambridge, Cambridge University Press. Arsen’eva, T. M. and Naumenko, S. A. 1992. Usad’by Tanaisa (‘Estates of Tanais’). Moscow, RAN In-t Arkheologii (in Russian). Auriemma, R. 2000. Le anfore del relitto di Grado e il loro contenuto. Mélanges de l’École française de Rome. Antiquité 112-1: 27–51. Babenchikov, V. P. 1963. Chornoriches’kiiymogilnik. Arkheologichni Pam’yatki URSR 13: 90–122 (in Russian). Baumann, V. H. 1995. Asezări rurale antice în zona Gurilor Dunării. Tulcea, Institutul de Cercetǎri Eco-Muzeale. Bărbulescu, M. 2001. Viața rurală în Dobrogea Romană (sec. I - III p. Chr.). Bibliotheca Tomitana III. Constant, Muzeul de Istorie Naţională şi Arheologie. Bekker-Nielsen, T. (ed.) 2005. Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2. Aarhus, Aarhus University Press. Belotti, C. 2004. Ritrovamenti di anfore romane a Ivlia Concordia. Aspetti topografici ed economici. L’Album 10. Venice, Fondazione Antonio Colluto. Bichir, Gh. 1973. Cultura carpicǎ. Bucharest, Academiei Republicii Socialiste România. Billard, R. 1988. La vigne dans l’antiquité. Lyon, Lardanchet [1913]. Marseille, Laffitte Reprints. Bjelajac, L. 1996. Amforegornjomezijskog Podunavlja. Posebna Izdanja Monographs 30. Belgrade, Srpska Akademija Nauka i Umetnosti. (summary in English: Amphorae of the Danubian Basin in Upper Moesia). Bondoc, D. 2014. Roman amphorae from Cioroiul Nou (Romania). Rei Cretariae Romanae Fautorum Acta 43: 105–112. Bordenache, G. 1969. Sculture greche e romane del Museo nazionale di antichità di Bucarest. Bucharest, Academia Bucarest. Bounegru, O. 2002. Comerţ şi navigatori la Pontul Stîng şi Dunărea de Jos (sec. I-III p. Chr.), Iaşi. Brukner, O. 1981. Rimska keramika u Jugoslovenskom delu provincije Donje Panonije. Belgrade, Savez arheoloskih drustava Jugoslavije. Brun, J.-P. 2003. Le vin et l’huile dans la Méditerranée antique. Viticulture, oléiculture et procédés de fabrication. Paris, Errance.

Bruno, B. 2003. Le anfore della cava di UC VII. Considerazioni sulle anfore nei contesti databili tra la tarda età antonina e la prima età severiana. In S. Lusuardi Siena and M. P. Rossignani (eds), Dall’Antichità al Medioevo. Aspetti insediativi e manufatti. Ricerche archeologiche nei cortili dell’Università cattolica. Atti delle giornate di studio, Milano 24 gennaio 2000, Milano 24 gennaio 2001: 85–97. Milan, Vita e Pensiero. Burakov, A. V. 1976. Kozyrskoegorodishcherubezha i pervyhstoletiiynasheyery. Kiev, Akademiianauk Ukrainskoi SSR, Institut arkheologii (in Russian). Camodeca, G. 1999. Tabulae Pompeianae Sulpiciorum (TPSulp.). Edizione critica dell’archivio puteolano dei Sulpicii. Vetera 12. Rome, Quasar. Carter, J. C. and Mack, G. 2003. Crimean Chersonesos. City, Chora, Museum, and Environs. Austin, Institute of Classical Archaeology. Cockle, H. 1981. Pottery Manufacture in Roman Egypt. A New Papyrus. Journal of Roman Studies 71: 87–97. Covacef, Z. 2002. Arta Sculpturală în Dobrogea Romană, secolele I-III. Cluj-Napoca, Nereamia Napocae. Croitoru, C. 2011. Roman discoveries in the east Carpathian barbaricum (1st century B.C. - 5th century A.D.). Braila, Brăila Museum . Curtis, R. I. 1991. Garum and salsamenta. Production and commerce in materia medica. Studies in ancient medicine 3. Leiden - New York - Copenhaguen - Cologne, Brill. Curtis, R. I. 2001. Ancient food technology. Technology and change in history 5. Leiden - Boston, Brill. Curtis, R. I. 2005. Sources for production and trade of Greek and Roman processed fish. In T. Bekker-Nielsen (ed.), Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2: 31–46. Aarhus, Aarhus University Press. Curtis, R. I. 2015. Storage and transport. In J. Wilkins and R. Nadeau (eds), A Companion to Food in the Ancient World: 173–182. Oxford, Wiley Blackwell. Dalby, A. 2000. Empire of Pleasures. Luxury and Indulgence in the Roman World. London and New York, Routledge. Dalby, A. 2003. Food in the Ancient World from A to Z. London, Routledge. De Boer, J. 2013. Stamped amphorae from the Greek Black sea colony of Sinope in the Mediterranean during the Hellenistic period. In G. R. Tsetskhladze, S. Atasoy, A. Avram, Ş. Dönmez, and J. Hargrave (eds), The Bosporus: Gateway between the Ancient West and East (1st Millennium BC-5th Century AD). Proceedings of the Fourth International Congress on Black Sea Antiquities. Istanbul, 14th-18th September 2009: 109–114, British Archaeological Reports International Series 2517. Oxford, Archaeopress. Doonan, O. P. 2002. Production in a Pontic landscape: the hinterland of Greek and Roman Sinope. In Faudot, M., Fraysee, A. and Geny, É. (eds), Pont Euxin et Commerce: la genèse de la « route de la soie»: 185–198. Besançon, Presses universitaires de Franche-Comté. Dorrego, F., Carrera, F. and Luxán, M. P. 2004. Investigations on Roman amphorae sealing systems. Materials and Structures/Matériaux et Constructions 37: 369–374.

372

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

Dragomir, I. T. 1962. Doua reliefuri Dionsiace descoperite la Făgărașul Nou (r. Hîrșova, reg. Dobrogea). Studii şi Cercetări de Istorie Veche şi Arheologie 13: 421–429. Durliat, J. 1990. De la ville antique à la ville Byzantine. Le problème des subsistances. Collection de l’École française de Rome 136. Rome, École française de Rome. Foley, B. P., Dellaporta, K., Sakellariou, D., Bingham, S. B., Camilli, R., Ryan M., Evagelistis, E. D., Ferrini, V. L., Katsaros, K., Kourkoumelis, D., Mallios, A., Micha, P., Mindell, D. A., Roman, C., Singh, H., Switzer, D. S. and Theodoulo, T. 2009. The 2005 Chios ancient shipwreck survey. New methods for underwater archaeology. Hesperia 78: 269–305. Forbes, R. J. 1965. Studies in Ancient Technology. Vol. III. Leiden, Brill (2nd ed.). Frier, B. W. 1983. Roman law and the wine trade: the problem of ‘vinegar sold as wine’. Zeitschrift der Savigny-Stiftung für Rechtsgeschichte 100: 258–295. Gadukevich, V. F. 1987. Antichnye goroda Bospora Mirmekiiy. Leningrad, Nauka (in Russian). Glodariu, I. 1976. Dacian Trade with the Hellenistic and Roman World. British Archaeological Reports International Series 8. Oxford, BAR publishing. Grace, V. and Empereur, J.-Y. 1981. Un groupe d’amphores ptolémaïques estampillées. Bulletin de l’Institut français d’archéologie orientale 81-1: 409–426. Harshegyi, P. 2008. Roman amphorae from the East along the ripa Pannonica. Rei Cretariae Romanae Fautorum Acta 40: 173–178. Hárshegyi, P. 2016. Supplying the Roman army on the Pannonian Limes. Amphorae on the territory of Budapest, Hungary (Aquincvum and Albertfalva). Rei Cretariae Romanae Fautorum Acta 44, 619-32. Hautumm, W. 1981. Studien zu Amphoren der spätrömischen und frühbyzantinischen Zeit. Fulda. Hayes, J. W. 1983. The Villa Dionysos excavations, Knossos: The pottery. Annual of the British School at Athens 78: 97–169. Heron, C. and Pollard, A. M. 1988. The analysis of natural resinous materials from Roman amphorae. In E. A. Slater and J. O. Tate (eds), Science and Archaeology, Glasgow 1987. British Archaeological Reports British Series 196: 429–448. Oxford, BAR publishing. Højte, J. M. 2005. The Archaeological Evidence for Fish Processing in the Black Sea Region. In T. BekkerNielsen (ed.), Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2: 133–160. Aarhus, Aarhus University Press. Il’yashenko, S. M. 2013. Standartnye dipinti na uzkogorlyh svetloglinyanyh amforah Tanaisai ego okrugi III-IV vv. n. e. Bosporskie Issledovaniya 29: 3–285 (in Russian). Ionas, I. 2000. Traditional Pottery and Potters in Cyprus. The Disappearance of an Ancient Craft Industry in the 19th and 20th centuries. Aldershot, Ashgate and University of Birmingham.

Johnson, B. L. 2008. The Pottery. In Patrich, J. (ed.), Archaeological Excavations at Caesarea Maritima. Areas CC, KK, and NN. Final Reports: 66–119. Jerusalem, Israel Exploration Society. Kabakchieva, G. 1986. Keramika ot vilata pri Ivaiylovgrad. Razkopkiiprouchvaniya XV. Sofia. Kassab Tezgör, D. 2010. Les fouilles et le matériel de l’atelier amphorique de Demirci près de Sinope. Varia Anatolica 22. Paris, de Boccard. Kelemen, M. H. 1990. Roman amphorae in Pannonia III. Acta Archaeologica Academiae Scientiarum Hungaricae 42: 147–193. Klenina, E. Yu. 2000. Kuhnayakeramika/Küchengefässe. In T. Sarnowski and. O. Savelja (eds), Römische Militärstation und Heiligtum des Iupiter Dolichenus: 134–140. Warsaw, Institute of Archaeology, Warsaw University. Klenina, E. Yu. 2004. Keramicheskie sosudy II-III v. n. e. iz usad’by „Bliznetsy” (Hora Hersonesa Tavricheskogo). Poznań, Instytut Historii Uniwersytetu. Kolendo, J. 1965. Études sur les inscriptions de Novae. Archeologia 16: 124–148. Kovalevskaja, L. and Sarnowski, T. 2003. La vaisselle des habitants d’une maison rurale de l’époque romaine dans la chôra de Chersonèse Taurique. Rei Cretariae Romanae Fautorum Acta 38: 229–235. Krapivina, V. V. 1993. Ol’biya. Material’naya kul’tura I-IV vv. n.e. Kiev. Kropotkin, V. V. 1970. Rimskie izdeliya v Vostochnoiy Evrope (II v. do n.e. - V v. n.e.). Moscow (in Russian). Kruglikova, I. T. 1966. Bospor i pozdneantichnoe vremya: ocherk ekonomicheskoiy istorii. Moscow (in Russian), Lepper, F. and Frere S. 1988. Trajan’s Column. A New Edition of the Cichorius Plates. Gloucester, Alan Sutton. Le Roux, P. 1994. Le ravitaillement des armées romaines sous l’Empire. In R. Étienne (ed.), Du latifundium au latifundo. Publications du Centre Pierre Paris 25: 394– 416. Paris, de Boccard. Lewit, T. 2012. Oil and wine press technology in its economic context. Screw presses, the rural economy and trade in late antiquity. Antiquité tardive 20: 137–149. Lund, J. and Gabrielsen, V. 2005. A fishy business: transport amphorae of the Black Sea region as a source for the trade in fish and fish products in the Classical and Hellenistic periods. In T. Bekker-Nielsen (ed.), Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2: 161–169. Aarhus, Aarhus University Press. Malyukevich, A. E. 1991. O torgovyh kontaktah poseleniiy nijnego podnestrov’ya v pervyevekan.e. In V.P. Vanchugov (ed.), Severo-Zapadnoe Prichernomor’e: kontaktnaia zona drevnikh kul’tur: 71–82. Kiev, Nauk Dumka. Marlière, E. 2002. L’outre et le tonneau dans l’Occident romain. Monographies Instrumentum 22. Montagnac, Éd. Monique Mergoil. Matei-Popescu, F. 2010, The Roman Army in Moesia Inferior. Bucharest, Conphys.

373

Andrei Opaiţ

Matei-Popescu, F. 2012. The origin of the tradesmen in Dacia. In D. Boteva-Boyanova, L. Mihailescu-Bîrliba and O. Bounegru (eds), Pax Romana: Kulturaustausch und Wirtschaftsbeziehungen in den Donauprovinzen des römischen Kaiserreichs. Akten der Tagung in Varna und Tulcea 1.-7. September 2008: 85–98. Kaiserslautern, Parthenon-Verlag. Matei-Popescu, F. 2015. The Auxiliary units in Moesia Superior and Dacia. A review and an update. In L. Vagalinski and N. Sharankov (eds), Limes XXII. Proceedings of the 22nd International Congress of Roman Frontier Studies Ruse, Bulgaria, September 2012. Bulletin of the National Archaeological Institute 42: 407–418. Sofia, National Archaeological Institute with Museum (NAIM-BAS). Miniero, P. 2008. I reperti archeologici del riempimento del Ninfeo. In P. Miniero and F. Zevi (eds), Museo archeologico dei Campi Flegrei: catalogo generale.3. Liternum, Baia, Miseno: 165–171. Naples, Electa. Nagy, A. A. 2015. La circulation des amphores dans la colonie civile de Brigetio. Raport préliminaire. In L. Borhy (ed.), Studia Arheologica Nicolae Szabó LXXV annos nato Dedicata: 183–190. Budapest, L’Harmattan Kiadó. Naumenko, S. A. 2012. Sostav amfor v zakrytyh komplesah Tanaisa knotsa I - serediny III v.n.e. Vestnik Tanaisa 3: 63–88 (in Russian). Nuţu, G., Stanc, S. M. and Paraschiv, D. 2014. Niculiţel. A Roman Rural Settlement in North-east Moesia Inferior. Archaeological and Archaeozoological Research. Kaiserslautern, Parthenon-Verlag. Oliveira, C., Morais, R. and Morillo Cerdán, A. 2015. ArchaeoAnalytics. Chromatography and DNA Analysis in Archaeology. Esposende, Município de Esposende. Oliver, J. H. 1965. Texts A and B of the Horothesia dossier at Istros. Greek, Roman, Byzantine Studies 6: 143–156. Opaiţ, A. 1980. Consideraţii preliminare asupra amforelor romane şi romano-bizantine din Dobrogea. Peuce 8: 291–327. Opaiţ, A. 1987. Amfore romane de mare capacitate. Consideraţii tipologice. SCIV 38: 245–258. Opaiţ, A. 2003. ‘Table’ Amphora Versus ‘Table’ Pitcher in Roman Dobrudja. Rei Cretariae Romanae Fautorum Acta 38: 215–218. Opaiţ, A. 2004. Local and Imported Ceramics in the Roman Province of Scythia (4th-6th Centuries AD). British Archaeological Reports International Series 1274. Oxford, Archaeopress. Opaiţ, A. 2007. A Weighty Matter: Pontic Fish Amphoras. In V. Gabrielsen and J. Lund (eds), The Black Sea in Antiquity, Regional and Interregional Economic Exchanges. Black Sea Studies 6: 101–121. Aarhus, Aarhus University Press. Opaiţ, A. 2010. Pontic wine in the Athenian market. In D. Papanikola-Bakirtzi and N. Kousoulakou (eds), Κεραμικη Τησ υστερησ αρχαιοτητασ απο τον ελλαδικο χωρο (3ος -7ος αι. μ.Χ.), Επιστημονίκή Συνάντηση (Θεσσαλονίκη, 12-16 Νοεμβρίου 2006): 108–130. Thessaloniki, Archaeological Institute of Macedonian and Thracian Studies-Aristote University Thessaloniki.

Opaiţ, A. 2013. Producția şi consumul de vin în ținuturile dintre Carpați şi MareaNeagră (sec. II a. Chr. - III p. Chr.): unele considerații. Studii şi Cercetări de Istorie Veche şi Arheologie 64: 21–65. Opait, A. 2015. Some east Pontic amphorae of Roman and early Byzantine times. In G. R. Tsetskhladze, A. Avram and J. Hargrave (eds), The Danubian Lands between the Black, Aegean and Adriatic Seas (7th century BC - 10th century AD). Proceedings of the fifth international congress on Black Sea antiquities (Belgrade- 17-21 September 2013). Archaeopress Archaeology: 283– 291. Oxford, Archaeopress. Opaiţ A. and Baumann, H. 2006. Consideraţii privind posibilitatea industrializării peştelui în Dobrogea antică. In M. Mănucu Adameşteanu (ed.), Orgame/ Argamum, Suppl. 1: A la recherche d’une colonie. Actes du Colloque International 40 ans de recherche archéologique a Orgame/Argamum (Bucarest-Tulcea-Jurilovca, 3-5 octobre 2005): 397–405. Bucharest, Asociatia Generala a inginerilor din Romania. Opaiţ, A. and Ionescu, M. 2016. Contributions to the economic life of the city of Callatis in light of new ceramic finds (2nd-6th centuries AD). Arheologia Moldovei XXXIX, 57–112. Opaiţ, A. and Paraschiv, D. 2012. Rare amphora finds in the city and territory of (L?)ibida (1st-6th centuries AD). Rei Cretariae Romanae Fautorum Acta 42: 113–124. Panella, C. 1986. Oriente e Occidente: considerazioni su alcune anfore egee di età imperiale a Ostia. In J.-Y. Empereur and Y. Garlan (eds), Recherches sur les amphores grecques. Bulletin de Correspondance Hellénique Supplément 13: 609–636. Athens, École Française d’Athènes. Parker, A. J. 1992. Ancient Shipwrecks of the Mediterranean and the Roman Provinces. British Archaeological Reports Series 580, Oxford, Tempus Reparatum. Peña, J. T. 2007. Roman Pottery in the Archaeological Record. Cambridge, Cambridge University Press. Pilipović, S. 2013. Wine and the Vine in Upper Moesia. Archaeological and Epigraphic Evidence. Balcanica 44: 21–34. Pippidi, D. M. 1962. Epigraphische Beiträge zur Geschichte Histrias in hellenistischer und römischer Zeit. Berlin, Berlin Akademie-Verlag. Pippidi, D. 1967. Hotarnicia consularului Laberius Maximus. In Contribuţii la istoria veche a României: 349– 385. Bucharest, Editura Stiintifica. Piscopo, A. and Poiana, M. 2012. Packing and storage of olive oil. In I. Muzzalupo (ed.), Olive Germplasm, The Olive Cultivation, Table Olive and Olive Oil Industry in Italy: 201–222. London, IntechOpen. Popescu, M.-C. 2013. Hellenistic and Roman Pottery in PreRoman Dacia. Bucharest, Arhitectură. Restaurare. Arheologie. Protase, D. 2010. Organizarea militară. In D. Protase, A. Suceveanu (eds), Istoria Românilor. Vol.2: 99–136. Bucharest, Editura Enciclopedică (2nd ed.).

374

Feeding the Lower Danube and Pontic areas with local wine and fish products (1st century BC-3rd century AD)

Reynolds, P. 1997-1998. Pottery production and economic exchange in 2nd century Berytus: some preliminary observations of ceramic trends from quantified ceramic deposits from the Anglo-Lebanese excavations in Beirut. Berytus 43: 35–110. Reynolds, P. 2010. Trade networks of the East, 3rd to 7th centuries: the view from Beirut (Lebanon) and Butrint (Albania) (fine wares, amphorae and kitchen wares). In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 89–114. Oxford, Archaeopress. Rice, P. M. 1987. Pottery Analysis. A Sourcebook. Chicago, University of Chicago Press. Rizzo, G. 2014a. Le anfore, Ostia e i commerci mediterranei. In C. Panella and G. Rizzo (eds), Ostia VI. Le terme del nuotatore. Studi Miscellanei 38: 65–481. Rome, L’Erma di Bretschneider. Rizzo, G. 2014b. Pontus and Rome: Trade in the Imperial Period. In V. Cojocaru, A. Coşkun, and M. Dana (eds), Interconnectivity in the Mediterranean and Pontic World during the Hellenistic and Roman Periods: 555–577. ClujNapoca, Mega. Romanus, K., Baeten, J., Poblome, J., Accardo, S., Degryse, P., Jacobs, P., De Vos, D. and Waelkens, M. 2009. Wine and oil permeation in pitch and non-piched ceramics: relations with results from archaeological amphorae from Sagalassos, Turkey. Journal of Archaeological Science 36: 900–909. Rostovtzeff, M. 1922. A large estate in Egypt in the Third century B.C. A Study in Economic History. Madison, University of Wisconsin Studies. Sackett, L. H. 1992. Knossos, From Greek City to Roman Colony: Excavations at the Unexplored Mansion II. London, The British School of Archaeology at Athens. Santoro, S., Pasquinucci, M. and Guiducci, G. (eds) 2010. LRCW 3. Late Roman Coarse Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. Comparison between Western and Eastern Mediterranean. British Archaeological Reports International Series 2185. Oxford, Archaeopress. Salviat, F. 1990. Vignes et vins anciens de Maronée à Mende. In T. Petridis (ed.), Μνήμη Δ. Λαζαρίδη. Πόλις και χώρα στην αρχαία Μακεδονία και Θράκη. Πρακτικά Αρχαιολογικού Συνεδρίου, Καβάλα 9-11 Μαΐου 1986. Ελληνογαλλικές Έρευνες Ι: 457–476. Thessaloniki, Ypourgeio Politismou, Archaiologiko Mouseio Kavalas - Athens, École française d’Athènes. Sealey, P. R. 1985. Amphoras from the 1970 excavations at Colchester Sheepen. British Archaeological Reports British Series 142. Oxford, BAR publishing. Sedikova, L. V. 1996. Kompleks nahogok pozdne antichnogo vremeni iz tsisterny v VI kvartale Hersonesa. Khersoneskiy Sbornik 7: 175–181 (in Russian).

Shelov, D. B. 1965. Raskopki severo-vastochnogo uchastka Tanaisa. In Drevnosti Nizhnego Dona: 56–129. Moscow, Institut arkheologii Akademiia nauk SSSR (in Russian). Shelov, D. B. 1972. Tanais i nijniiy Don v pervyeveka nasheiy ery. Moscow (in Russian). Shipley, G. 2000. The Greek World After Alexander 323-30 BC. London, Routledge. Slane, K. W. 1986. Two Deposits from the Early Roman Cellar Building, Corinth. Hesperia 55: 271–318. Slane, K. W. 2000. East-West trade in fine wares and commodities: the view from Corinth. Rei Cretariae Romanae Fautorum Acta 36: 299–312. Strjeletskiiy, S. F. 1959. Pozdneantichn’iymogil’nik v Inkermanskoiydoline. Kratkie Soobshcheniya Instituta Arkheologii Akademii Nauk SSSR 8: 139–145 (in Russian). Strjeletskiiy, S. F., Vysotskaya, T. N., Ryzhova, L. A. and Jestkova, G. I. 2005. Naselenie okrugi hersonesa v pervoiy polovine I tysyacheletiya novoiy ery (po materialam nekropolya ‘Sovhoz 10). Stratum (20032004) 4: 27–277 (in Russian). Tsetskhladze, G. R. and Vnukov, S. Yu. 1992. Colchian Amphorae: Typology, Chronology, and Aspects of Production. Annual of the British School at Athens 87: 357–386. Varbanov, V. 2011. Late Hellenistic anepigraphic amphora stamps from northeastern Bulgaria. Archaeologia Bulgarica XV, 1: 53–59. Vinokurov, N. I. 1999. Vinodelie antichnogo Bospora (‘Winemaking of Ancient Bosporus’). Moscow (in Russian). Vnukov, S. Yu. 2003. Prichernomorskie amfory I v. do n.e.-II v. n.e. (morphologiya) I (‘Black Sea amphorae of the 1st c. B.C.-2nd c. A.D., morphology’). Moscow, Institut Arkheologii (in Russian). Vnukov, S. Yu. 2011. ‘Colchean’ amphorae from Abkhazia. In C. Tzochev, T. Stoyanov and A. Bozkova (eds), PATABS II. Production and Trade of Amphorae in the Black Sea. Acts of the international round table held in Kiten, Nessebar and Sredetz. September 26-30, 2007: 271–278. Sofia, National Archaeological Institute with Museum by the Bulgarian Academy of Sciences - St. Kliment Ohridski University of Sofia. Vnukov, S. Yu. 2017. Overseas trade in the Black Sea region and the formation of the Pontic market from the first century BCE to the third century Ce. In V. Kozlovskaya (ed.), The Northern Black Sea in Antiquity: Networks, Connectivity, and Cultural Interactions: 100– 138. Cambridge, Cambridge University Press. Vornic, V., Telnov, N., Bubulici, V. and Ciobanu, L. 2007. Pruteni.Un centru de olărie dacic din epoca romană. Chisinau, Academia de Ştiinţe a Moldovei, Institutul Patrimoniului Cultural, Centrul de Arheologie. Vulpe, A. 1965. Reprezentări umane pe cupele getice de la Popeşti [Antropomorphic representations on Getic bowls from Popeşti]. Studii şi cercetări de istorie veche 16/II: 341-349. Wilkins, J. 2005. Fish as a source of food in Antiquity. In T. Bekker-Nielsen (ed.), Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2: 21–30. Aarhus, Aarhus University Press.

375

Andrei Opaiţ

Yermolin, A. L. and Fedoseev, N. F. 2013. A ceramic amphorae production workshop of the first half of the 3rd century AD in Pantikapaeu. In L. Buzoianu, P. Dupont, and V. Lungu (eds), PATABS III. Production and Trade of Amphorae in the Black Sea. Production et commerce amphoriques en Mer Noire. Actes de la table ronde internationale de Constanța, 6-10 octobre 2009. Pontica 46 Supplementum II: 185–197. Constant, National History and Archaeology Museum.

Zeest, I. B. 1960. Keramičeskaja tara Bospora. Materialy i issledovanija po archeologii SSSR 83. Moscow, Institut arheologii (in Russian).

376

Ephesus wines Tamás Bezeczky†

(1949-2018) Formerly Institut für Kulturgeschichte der Antike der Österreichischen Akademie der Wissenschaften

Abstract: Amphorae produced in the region of Ephesus, and their imitations in Pergamon, Chos, Aphrodisias and Sardis were devoted to the transport of wine and distributed throughout the Mediterranean and even in India from the late 1st century BC to the 7th century AD. This paper provides a comprehensive summary of the epigraphy, typology, production areas, and contents of these containers. Key words: Ephesus; amphorae; epigraphy; production; wine.

1. Introduction Ephesus is located in Asia Minor at the mouth of the River Cayster (Küçükmenderes = ‘Small Meander’), which flows between the Tmolos and Messogis Mountains. The amphorae used in Roman commerce have been found in significant numbers in Ephesus, some of which may have been produced in and around Ephesus. However, no villas or amphora workshops have yet been uncovered in the vicinity of Ephesus. A comprehensive geological survey was undertaken in the region during the 1990s. In addition to the samples from the local clay deposits, different types of ceramics and amphorae were also collected at Ephesus, and from various other sites off the coast of Asia Minor.1 (Figure 1) The geological surveys have identified several types of clay in the Cayster Valley and in the region of Kuşadasi. After comparison with petrological analyses of the amphorae produced from Hellenistic times to the Late Roman/Byzantine period, three main groups of raw materials can be identified.2 In a recently completed  This research was sponsored by Ulrike Outschar´s FWF-Project (No. 9280). The results of this survey are documented in an unpublished manuscript (Sauer 1995) and have also been used as reference material. Some of the data, however, has already been mentioned in other publications: Outschar 1993: 52; Bezeczky 2005: 205; Ladstätter 2008: 180; Bezeczky 2013: 28-31. 2  The petrological analysis provides a detailed description of the fabrics. Sauer 2013: 28–31 and 200-201. There are three major types. Fabric A: micaceous groundmass, which is rich in garnet (which points to the vicinity of Kuşadasi and the valley of Meander). Fabric B: micaceous groundmass, which is rich in epidote / clynozoisite, which is characteristic of the vicinity of Ephesus and the valley of Cayster. Fabric C: micaceous groundmass, which is rich in carbonate and has very frequent sheet silicates (muscovite, oxidized sheet silicates, biotite/phlogopite). Such carbonate‑rich clays (marls) also occur frequently in the surroundings of Ephesus. The amphorae types suggest that fabrics A and C were used in Hellenistic and Early Roman times and fabric B was used in the mid and late Roman period. In the Athenian one‑handle jars Lang (1955) differentiates three types of fabric: a dark buff non‑micaceous clay, a micaceous buff clay, and a highly micaceous red brown clay which may be fired black, brown or red. Concerning the origins of the amphorae, J. W. Hayes (1976: 117; 1992: 63) suggests the Hermos or the Meander valley basin, while J. Riley (1979: 184) proposes Asia Minor. I. B. Zeest (1960: 118–119) suggests a Samian origin. 1

research programme, I wanted to determine whether Ephesian amphorae could be found in some large centres of the Roman Empire.3 Relying on both the formal/ typological characteristics of the amphorae and the results of earlier and current petrological analyses (Sauer 1995; Sauer 2013: 197–211), it proved possible to differentiate the highly micaceous vessels produced in Ephesus (Figure 2) from those produced in other workshops. In a few cases it was also possible to identify other production centres; the remaining unresolved questions may become an object of investigation in the future. I hope that my results will contribute to a more complete understanding of Ephesian economic history and draw attention to other centres of production — onehandled jars of identical shape and amphorae of a similar type were produced in a number of places.4 The first part of this paper tentatively identifies the contents of the amphorae produced in the Ephesus region, based on the ancient written sources and the inscriptions on the amphorae. I shall also use the results of petrological analyses. The second part of this article will survey the various Ephesan vessel shapes. Ancient written sources mention Ephesian wine: 1. Strabo (XIV 1, 15) ‘[…]  The Ephesian and the Metropolites are good wines, but the Mesogis, the Tmolus, the Catacecaumene, Cnidos, Smyrna, and other more obscure places, are distinguished for the excellence of their wines, whether for gratification or dietetic purposes […]’.

V. Grace (1979) proposes an Egyptian origin, but this can be ruled out entirely based on the petrological analysis. D. Peacock and D. Williams were the first to carry out a petrological analysis of these amphorae. Based on the heavy mineral analysis, D. Williams (1982: 104) considers Asia Minor to be the most likely place of origin, while Peacock (1984: 22) favors Sardis. 3  Austrian Science Fund, FWF project number P20776. 4  There was a lecture on this topic at the Per Terram, Per Mare, Production and Transport of Roman Amphorae in the Eastern Mediterranean Conference in Nicosia, Cyprus on April 12-15.

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 377–387

Tamás Bezeczky†

2.1. kond(…) In the Athenian Agora there is a fragment of a micaceous jar with a black dipinti: kon(…) and κορι/Θ.5 The documentation available on the internet (the catalogue card) shows — as was mentioned by Mabel Land — that the dipinti were not written by the same hand. The letters kon( ) are written with Latin, while the second line is written with Greek characters. The letters kon(ditum) may refer to the name of the wine. The second line may contain the traders name of the wine (like Corintian) (Figure 3). In the Vedius Gymnasium (Ephesus) there was excavated a Late Roman 3 amphora with a post cocturam graffito kond(…). The graffito may refer to the type of the konditum wine6 (Figure 3). Figure 1. Geological survey in Asia Minor (after Sauer 1995).

2.

3. 4. 5.

Pliny (NH XIV, 75) ’[…] as for the vintage of Mesogis, it has been found to cause headache, and that of Ephesus has also proved to be unwholesome, because sea–water and boiled must are employed to season it […]’. Dioscorides (5, 10) mentions Ephesium. It was produced in Pygela near Ephesus. The other wine mentioned, Messogites, comes from the region Asia and the Tmolus mountain. Athenaeus (I, 30-31) ‘there is a mountain village near Ephesus, formerly called Leto’s village, but now Latorea from an Amazon of that name; in this pramnianwine was produced’. Pliny (NH XIV, 54) the pramnian wine […] is grown in the territory of Smyrna (Izmir) […]. The Expositio totius mundi et gentium (cap. 47) describes the Ephesus region as rich in various wines, olive oil, grain, good purple dye, and spelt. One of the stone inscriptions at the Temple of Artemis mentions that the Temple had estates in the Cayster Valley. The inscription speaks of a wine co‑operative (συνεργασία), and of the way wine was sold in the city centre and in the small shops in the street between the stadium and the theatre (Knibbe 1985: 71–77; Engelmann 1986: 107).

The inscription concerning taxation in Ephesus tells us that the proportion of wine as a means of settling the taxes was higher than that of cash or olive oil (Schulten 1906: 40-70). 2. The inscriptions The number of the dipinti and graffiti is fairly small. There are few stamps on the amphorae produced in the region of Ephesus. First, I shall take a look at the amphorae with inscriptions then I shall mention the published stamped jars.

Conditum was a well known spiced honey wine, as mentioned by Apicius (I, 1‑2). There was two different types: conditum paradoxum —  spiced wine surprise (honey-wine) and conditum melizomum viatorium — spiced honey‑wine for travellers. This is the type of wine in Diocletian’s edict II.1. (κονδείτιον = conditi spiced wine 1 Italian pint den. 24). 2.2. προ There is a graffiti (Figure 3) in the Ephesus Terrace House 2 that mentions προ = προπεν (aperitif, cocktail).7 Martial mentions προπεν twice (12, 82, 11; 5.78.3). It is also mentioned by Petronius.8 There are (highly micaceous) vessels probably produced in Ephesus in the Athenian Agora. Black dipinti contain the abbreviation προ (Robinson 1959: M 256-P 9800, M 258-P 22009, M 259-P 11583, M 278-P 11582) (Figure 3). Mabel Lang read and Henry Robinson published the abbreviation as πρότροπος. It was defined as sweet Mytilenean wine.9 This type of wine is mentioned by Athenaios: ‘[…] The Mitylenaeans call the sweet wine of their country prodromus; others say protropus […]’.10 Plinius also described the liquor protropum.11 The must obtained before pressing was placed in a special vessel  Lang 1976: 81, He 44, Pl. 47 (P22833). She published: ‘Roman κον = kon. The first line might be the producer’s name, but since it is in a different hand might be the contents of second use: e.g., κονδειτιον. The second must give contents, whether something flavoured with coriander or a trade name like Corinthian κορι( ) ξ(έσται) Θ´’. 6  Ladstätter 2008: 181; At the conference, Jean-Luc Fournet noted that it was carved on the vessel in the course of its later use. This vessel also contained wine. 7  Taeuber 2005: GR 51; Liddell-Scott Greek-English Lexicon, 1494, drink before or first; take a snack before dinner. 8  Petronius 28, […] Trimalchio hoc suum propin esse dicebat […]. 9  Robinson 1959: 107; Lang 1976: 81, He 44, Pl. 47; Liddell‑Scott 1968: 1537, πρότροπος a sweet Mytilene wine, that flowed from the grape before the treading. 10  Athenaios I.30b and 2.45e. 11  Pliny XV. 85, ‘[…] but among these varieties ought also to be placed the liquor called in Greek protropum, the name given by some people to 5

378

Ephesus wines

Figure 2. Amphora types produced in the Ephesus region (Bezeczky 2013). (lagoena). Dioscorides also described protropum among the sweet wines(creticum and pramnium).12 Like Pliny, he does not mention that this wine could only be made on Lesbos. The προπεν reading of the abbreviation προ found in Ephesus, as well as protropum mentioned by Pliny and Dioscorides, makes it possible to assume that the vessels contained wine. Whichever reading is accepted, it is obvious that the dipinti occur on vessel from Ephesus. These belong to [the] Ephesus 4 type (Agora M 240, M 255‑259, M 277-278) (Figure 2). The inscriptions on the late Roman jars and amphorae make it safe to assume that they contained wine from Ephesus. Since their shape change only slightly, the earlier one handle jars may also have contained wine. It seems likely that the vessels from Ephesus contained good quality wine. The vessels can be found from Britain to India. must that flows down of its own accord before the grapes are trodden. This, as soon as it flows, is put into special flagons (lagoen) and allowed to ferment, and afterwards to dry for forty days of summer that follows, just at the rise of the Dog-star […]’ (Tr: H. Rakham, LCL 243). 12  Dioscorides 5.9. Dioscorides emliti ‘[…] Passum made from sun-dried grapes or those dried on the branches and pressed is called creticum, protropum, or pramnium […]’.

The wines also mentioned by ancient sources (phygelites/ ephesian and pramnian) have not been found on amphorae from Ephesus. Pramnian is known in a number of regions (on the islands of Lesbos and Icaria and the region of Smyrna). It is no surprise that the pramnian dipinto is known on vessels that do not come from Ephesus (Lang 1976: 74–75, Hd 5, Hd 17, 79, He 26). There are two more dipinti on one handle jars in the Athenian Agora. Their reading is not quite certain (Figure 3). 2.3. μη Robinson found a dipinto (μη) in black paint on a M 241 fragment. He published it as μηλιτης (apple or quince wine). The reading was suggested by Mabel Lang (Robinson 1959: 106–107). Later, in Lang’s book published in 1976, the same dipinto is read as μηνιαϊος (monthly) (Lang 1976: 93, L 43). 2.4. παλ The black dipinti on two (also micaceous) vessels (M 242 and M 257) can be read as παλ which could signify παλαιός (Robinson 1959: 106 and 109, M 242-P 11590 and M 257-P 22008; Lang 1976: 93, L 43e-g).

379

Tamás Bezeczky†

2.5. άκινο

3.1. Ephesus production

The interpretation of the inscription on the M 307 amphora changed between 1959 and 1976. Robinson wrote: between handles, a graffito άκινο (the jar might have held wine flavoured with άκινος, wild basil) (Robinson 1959: 117). Lang refers to this as Άκιν ( ) and mentions a graffito on the bottom of a vessel (shallow bowl with wide ring foot). The amphora can be dated to the 5th century and the other vessel is dated to the first half of the 1st century. This is probably a name and not a wine (Lang 1976: 50, F 320-P 11569 and F 254).

3.1.1. Early Roman period

2.6. Stamps On the one handle jars we have three stamps in Ephesus. Two of them are from the Agora, and the third [one is at] the Basilica Stoa. Two have initials of names Єλ(…), Mάσ(…) and a gem imprint showing a young head in profile (Bezeczky 2013: 66). Ελουιος and Mάσκουλος names are known from inscriptions in Ephesus. I have recently found a micaceous handle fragment with a CE(…) stamp among the finds of the Athenian Agora13 (Figure 3). 2.7. Other graffiti It should also be mentioned that there are a number of post firing (post cocturam) graffiti on the one handled jars of Ephesus of the early and middle Roman periods. Usually they are beside the handle. Lang found that they refer to the dates of the vessels. ‘Since the range of graffito numbers in such that they give dates on the Actian era which compare favorably with the context dates of the jars, we are perhaps justified in exploring this interpretation in detail’ (Lang 1955: 279). 3. Amphora production area and fabric The one-handled jars formerly called ‘micaceous jars’ form a group of their own. They were produced from the Augustan to the Late Roman period. During this time the shape of the vessel underwent many changes. In the Late Roman era, the two-handled versions were the most widespread of the series. The investigation of the Roman vessels showed that the available typologies are insufficient. Although this article presents categories using the analyzed vessels, this does not represent a new typology, as it consistently refers to earlier publications.

 Inv. SS 2776, I am grateful to John Camp for the permission to study the amphorae. The handle fragment excavated in the late Roman context but the handle belong to an early Ephesian type (F 65 or F 66). 13

Ephesus type 1 One-handled jar (Agora F 65, F 66, Robinson 1959: 17). Date: From the last quarter of the 1st century BC to the 1st century AD (Robinson 1959: 10 and 17; Bezeczky 2013: 66). Fabric: The Ephesian jars and the jar from Magdalensberg were produced in Ephesus (Bezeczky 2013: 69, nos80, 81, 87, 89, 91; Magdalensberg MB 75 NG/45). One of the jars from Pergamon Type F 66 possibly has an Ephesus Fabric (Japp 2014: 148, fig. 6). Distribution: According to the publications, this type can be found in the following places: France in the Rhône Valley, Germany: Haltern, Noricum: Magdalensberg, Italy: Comacchio, Pompeii, Greece: Athens and Turkey: Pergamon, Sardis (Loechscke 1909: 212; CIL IV. 5911, Robinson 1959: Pl. 2; Berti 1986: 189; Lemaître 1997: 312– 313; Bezeczky 2013: 66). 3.1.2. Mid-Roman period Ephesus type 2 One-handled jar (Agora M 45, J 46-47, M 125, Robinson 1959: 55, 88 and 95). Date: This group has been dated from the early 2nd  century to the late 3rd century AD (Robinson 1959: 55, 88 and 95; Bezeczky 2013: 70). The Ephesian pieces were found in mixed layers. The vessels found at the excavation in Ostia belong to four chronological groups: AD 120‑140, 160‑180, 180‑210 and 230‑250 (Bezeczky 2014: 390, T 72, courtesy of C. Panella and G. Rizzo). Fabric: According to Sauer’s analysis, the vessels from Ephesus, Sagalassos and Ostia were produced in the Ephesus region (Bezeczky 2013: 71, no.104, 594‑600, 712‑713 and 99‑258, 99‑317, 88‑93, 3009; Sagalassos 39, SA 2009 JP 95; Ostia Bezeczky 2014: 388, T 72, no.13). Some of the vessels from Ostia and Pergamon require further analysis, as their origins could not be determined with certainty (Ostia: Bezeczky 2014: 388, T 72, nos14 and 15; Japp 2014: fig. 7 = 131 and fig. 8 = 128). I have also found Ephesus 2 type vessels in Aquincum, Arles, Corsica, Ostia, Pergamon and Zadar. The petrological analyses showed that the vessels from Aquincum, Arles, Corsica, and some of the vessels in Ostia were not made in Ephesus (Aquincum 62.13.215 and 86.8.409; Arles NH, RHO 07.00.1488; Corsica US 45, 116 and US 15/9 159; Ostia: Bezeczky 2014: 388, T 72, no.17 and 19). The thin section analysis revealed volcanic inclusions, but the place of production could not be identified. However, one of the vessels was made in the Pergamon region (Pergamon 125 RNH PE 70 G2-3 Kan So 94). Distribution: According to the publications, this type can be found in the following places: Greece: Athens, Corinth, Olympia, Knossos; Lybia: Benghazi–Berenike; Egypt: Berenike, Mons Claudianus; Cyprus: Paphos; Turkey: Tarsus, Zeugma; Black Sea: Tanais; France: Lyon, Marseille, Vienne, Frejus; Italy: Aquileia, Ostia,

380

Ephesus wines

Figure 3. Dipinti, graffiti and stamps of the Ephesian jars and amphorae.

381

Tamás Bezeczky†

Brescia, Pisa; Dalmatian coast: Istria-Brijuni, Zadar; Raetia: Augst; Noricum: Virunum; Pannonia: Aquincum (Robinson 1959; Panella 1973: Ostia III 368; Riley 1979: 183–185; Bezeczky 2013: 70). Ephesus type 3 One-handled jar, similar Agora M 279, (Robinson 1959: 110). Date: The Athenian jar is dated to the late 4th century, while the Ephesian is dated to the third century. Two vessels from Villa dels Munts (Altafulla, near Tarragona) date from the second part of the third century (Robinson 1959: 110; see the forms Williams 2005: LR 3 the one handled jars in the pictures collection). Fabric: We did not carry out a petrological analysis of the jars from Ephesus and Tarragona. The thin section analysis of the jars from Pergamon and Ospring (courtesy of D. Peacock) show that were probably produced in Ephesus. Possible Ephesus Fabric: Pergamon and Ospring. Distribution: Turkey: Pergamon; Spain: Tarragona; Britannia: Ospring (Japp 2014: 144, fig. 11; Peacock and Williams 1986: 188, fig. 107 B; Williams 2005: LR 3). 3.1.3. Late Roman period Ephesus type 4 One-handled jar, Agora M 240, M 255 - 259, M 277 - 278, (Robinson 1959: 106–109). Date: This type is dated to the 4th century AD at the Athenian Agora (Robinson 1959: 106, 108–109. The Ephesian pieces of this type were found in mixed layers). Fabric: The vessels found in Terrace House 2, the Ephesus Harbour and Poreč are characterized by typical Ephesus fabric (Bezeczky 2013: 163, no.601 and 602; 84‑88, 3012; Poreč). The vessel found in the harbour is black in colour. The vessels from Aquileia and Aquincum were probably made in the Ephesus region (Aquileia 372338‑1; Aquincum 86.8.409). The vessel found in Ovilava in Noricum may also have been produced in Ephesus, but further petrological analyses are required to determine its exact origins. A vessel from Marseille was produced in a volcanic area according to the thin section (Marseille, Harbour 1884, 412601). Distribution: According to the publications, this type can be found in the following places: Greece: Athens, Argos; Black Sea: Tanais; Israel: Caesarea Maritima; Egypt: Mons Claudianus; Italy: Luni; Slovenia: Ajdovščina; Spain: Hispalis; Tunisia: Carthage (Robinson 1959: 106–107, 108, 110; Abadie 1989: 48 fig. 2; Lusuardi Siena 1977: 223; Tomber 1988: 500, nos374‑379 fig. 19; Arsen’eva and Naumenko 1992: 149, fig. 29, 1; 158 fig. 38, 3‑4; Oleson 1994: 16, A 18 fig. 4; Pieri 2005: 191–195; Vidrih Perko–Žbona Trkman 2005: 282, fig. 9, nos9, 12; Tomber 2006: 168, Type 54; Amores Carredano, García Vargas and González Acuña 2007: 136, fig. 4, no.30). Ephesus type 5 One-handled jar, Agora M 275 – 276, (Robinson 1959: 110; Pieri 2005: 98, LR 3 B1). Date: This type is dated to the 4th century AD at the Athenian Agora and Gaulis necopolis. 3rd  century

in Marseille and late 4th and early 5th century in Caesarea (Robinson 1959: 110; Feugère et al. 1987: 85–86; Pieri 2005: 98; Tomber 1999: 301). Fabric: The M 275 vessels found in the Ephesus harbour and at the Athenian Agora showed the characteristics of Ephesus fabric when examined with a hand-held magnifying glass. The fabric of the M 276 vessel from Athens is also micaceous, but it contains larger (0.5‑1mm) brown components that are not characteristic of vessels produced in Ephesus. Distribution: Ephesus harbour; According to the publications, this type can be found in the following places: Istria: Poreč, Dalmatia: Zadar and Aquileia (all unpublished), Marseille and Caesarea (Pieri 2005: 98 and footnote 163; Tomber 1999: 314). Ephesus type 6 Amphora, Agora M 307, Amphora (Robinson 1959: 112; Pieri 2005: 98, LR 3 B3). Date: This type is dated to the 5th century AD at the Athenian Agora (Robinson 1959: 112). Fabric: Typical Ephesus fabric. The vessels found in the harbour are gray and black in colour. Ephesus Fabric: Ephesus and Sagalassos (Ephesus 81‑631, 3008, 84-3012, 89-18 3023, 88-3025; Sagalassos SA 2004 NEG 66). Distribution: Apart from Ephesus, they can be found in the following locations: Ampurias, Arles, Marseille, Port-Vendres, Ravenna, Split, Athens, Argos, Samos, Thasos, Jerusalem, Ballana, Jalame, Cyprus and the Black Sea region: Istanbul, Ulcicà, Tomis, Ravenna and Brijuni unpublished (Pieri 2005: 99; Ravenna: Maioli and Stoppioni 1989: 571). Ephesus type 7 Late Roman Amphora 3 (Agora M 373, M 335, L 50-51, Robinson 1959: 119; Annis 1976: 31; Riley 1979: 229–230, LR 10 Berenike; Carthage LR 3; Peacock and Williams 1986: 188–189; Pieri 2005: LR 3 A3, 95-6; Zeest 95; British Biv). Date: This type can be dated from the end of the 4th century AD to the end of the sixth century, or perhaps even later. In Rome the amphorae from the Palatine Hill were found in layers dated to 470-480 (Robinson 1959: 79, 115, 119, 4th to 6th century). Fabric: The LR 3 amphorae analysed earlier were found at Ephesus (Bezeczky 2013: nos381, 852-853, 856, 859‑860). They all belong to Ephesus Fabric group B. The new petrological analyses confirm the vessels’ Ephesian origin in most cases in Bracara, Barcelona, Carthage, Tarragona, Marseille, Rome, Brijuni, Poreč, Rhodes (Ephesus: Bezeczky 2013: 166, no.381; 167, nos852, 853, 859; 97-012, 99-058-608; 88-27-3032, 89-32-H7-3028, 8932-H3-3031; Bracara S67, S51; Barcelona 79-03-33063, 1101-14510; Carthage Peacock 1984: 121, no.3, fig. 34, 4-5; Tarragona Remolà 2000: 211, fig. 4; Marseille P 219a, P 220, 2709, 2097-10, 2095-4; Rome Palatino 286, B31 and 2240; Brijuni 43; Poreč 107; Rhodes ΠX 1184, ΠX 1185); in a few cases Arles, Marseille, Nice, Carthage, Rome, Brijuni, Kos Mastichari it is likely, but not certain (Arles AT 22.007, 18;

382

Ephesus wines

Marseille – 29; Nice, Château 1964, 59-54, 169; Carthage F13 III, 298; Rome, Palatinus PNE 04, II 2240 B1, Crypta Balbi II, 713; Brijuni Castrum yard sl. III; Kos, Mastichari 119). One LR 3 amphora was made of a micaceous fabric, but also contains volcanic inclusions (Kos AE 268, 269). This suggests that LR 3 amphorae were also produced elsewhere. Distribution: Widespread across the Roman Empire from Britain to the eastern Alps region, from Spain to Asia Minor and the Black Sea region, and even India where was just a body sherd (Williams 2005; Pieri 2005; Tomber 2008: 166; Bezeczky 2013: 165; Cirelli 2014: 542–543). Ephesus type 8 Amphora, Ephesus 56 (Peacock 1984: 123; PacettiPaganelli 2001: 214–215; Bezeczky 2013: 167–170). Date: The dating is the same as that of the LR 3 amphorae from the end of the 4th century AD to the end of the 6th/ beginning of the 7th century AD (Bezeczky 2013: 167). Fabric: The analysed amphorae came from the Agora and from Arap-Dere in Ephesus two pieces in Marseille and the one in Zadar were also made in Ephesus. They all belong to Ephesus Fabric group B (Ephesus, Bezeczky 2013: 170, nos397; 861, 865, 869; Harbour 88-44, 3021; Marseille 2570.12, 6007,4; Zadar 24). In the case of the vessels in Nice, Carthage and Brijuni, the petrological analyses suggests that they were probably produced in Ephesus (Nice Château 19-2-60; Carthage Peacock 1984: 123, no.8, fig. 36 no.16; Brijuni 209, 150482). Distribution: Carthage, Rome, Bodrum, Marseille, Brijuni, Zadar and Samos (Carthage: Peacock 1984: 123, nos89, fig. 36; Tomber 1988: 525, no.532; Pacetti-Paganelli 2001: 214–215 Marseille, Bourse Inv.-nos6007-4, 257012; Brijuni 25937; Zadar 352, Necr.; Hautumm 1981: 224 no.187, fig. 332). The form was copied on Mastichari, on the island of Kos (Didioumi 2014: 171, figs. 14 and 16). Ephesus type 9 LRA 3 B2, small amphora (Pieri 2005: 98). Date: There was no opportunity to date the pieces from Ephesus during the excavation of the harbour. Fabric: Ephesus fabric, examined only with a magnifying glass. Distribution: Many specimens were found during the excavation of the Ephesus harbour, Ravenna, Mljet and Marseille (Ephesus: Zabehlicky 1999: no.1, T. 106; Ravenna: Maioli and Stoppioni 1989: 573; Mljet: Brusić 1988: fig. 4; Marseille: Pieri 2005: 98). 3.2. Other production sites (Figure 4) 3.2.1. Pergamon There are micaceous one-handled jars in the Pergamon collection (for permission to study these, thanks to Sarah Japp and Felix Pirson). The petrological analyses shows that some of the vessels were unquestionably produced in Ephesus, while further vessels are also likely to have been made there.

Three further vessel types of local manufacture can also be identified. Among these, Sauer distinguishes between jars made in Pergamon (Type F 66) (Pergamon TR 27.2880 N Zi and PE 75 Gsd/NSC 80) and those made in the area around Pergamon Type M 125 (Pergamon PE 70 G2-3 KanSo 94); Type J 46 (Japp 2014: 149, fig. 9); Pergamon Type 1 (Pergamon PE.80, F 10a Zi 9 and PE 86 23 r II, 13); Pergamon Type 2 (Japp 2014: 148, fig. 3. The thin section analyses show that the fabric of two types are identical with those of the locally produced vessels: (micaceous fabric with volcanic inclusions). The fabric of the Type Pergamon 3 (Pe 126) vessel also contained volcanic inclusions, but it is different from the Pergamon fabric and its origins are unknown. Distribution: Type Pergamon 1 of the vessels made in Pergamon can be found in Athens, Knossos and the Pula region (Lang 1955: P 14082, Pl. 79(p); Hayes 1983: 158, no.105). 3.2.2. Kos An amphora production workshop was identified on the island of Kos (Didioumi 2014: 169–180). A ceramic workshop also produced amphorae in a north-western part of the island called Mastichari (Courtesy S. Didioumi, A. Giannokouri and 4th Ephorate of Byzantine Antiquities. The petrological characteristics (thin section and heavy mineral) of the rocks and soil collected in the region are the same as those of the amphorae. The fabric is basically micaceous, with chert and volcanic fragments. The heavy mineral assemblage consists of hornblende, clinopiroxen, garnet, chromium spinel epidot, diallage, rutil, brown hornblende (Type Pieri B4 AE 26: Similis Ephesus 56 AE 268-269 = Didioumi 2014: fig. 14). The workshop produced amphorae from the 6th century to the beginning of the 7th century. One of the types imitates the LR 3 form with some differences, like Pieri LR 3 B4 form. The other type is modelled on the Ephesus 56 amphora. Parts of both types resemble the Ephesian archetype quite closely. One LR 3 amphora was also found that may have been produced in Ephesus (Mastichari 119). Distribution: A vessel found in Caesarea Maritima and one in Saraçhane (Pieri 2005: 98; Hayes 1992: fig. 48). There is probably another vessel among the Nice amphorae, although we did not carry out a petrological analysis of that fragment. 3.2.3. Aphrodisias Three Aphrodisias amphora types have been published. They were found in the course of a survey (De Staebler 2012: 63, fig. 4). These amphorae have a flat bottom made of a micaceous fabric. They have not been found at any other site yet. There is a Late Roman 3 amphora in Ravenna which has an ‘Aphrodisios’ (άφoρδίσιος) post cocturam graffito. Aphrodosios is interpreted as a personal name (Fiaccadori 1983: 239, 23.1a-b). It is presumed to be a name — perhaps that of the producer or the merchant. 3.2.4. Sardis There was another amphora workshop (perhaps near Sardis) that produced micaceous one-handled jars. The shape of the vessels published by M. L. Rautman

383

Tamás Bezeczky†

Figure 4. Amphorae from other production sites (Bezeczky 2013).

384

Ephesus wines

(1995: 64, figs. 19 and 21) is very different from the series in Athens. I have not found Sardis amphorae at the sites that I examined or among the published finds. Fabric: ‘[…] that is moderately coarse grained and highly micaceous. Small particles of lime occur together with abundant micain fine particles and large flecks […]’ (Rautman 1995: 49). 3.3. Unknown production centre in a volcanic area 3.3.1. Agora M 126, one-handled jar (Robinson 1959: 95‑96) However, one-handled jars type Agora M 126 are formally and petrologically different from the amphorae produced in Ephesus (Robinson 1959: 95–96 notes: Layer IV — fragments of two other jars with tubular foot: one similar to M 126 in shape but of micaceous fabric; the other made of fabric similar to that of M 126 but comes from a jar with much shorter toe. Arles RHO 97 Z 62.935; Ostia Bezeczky 2014: 388, T 72, 17; Brijuni Verige Underwater exc. C6/C7 97.9.17; Zadar, Zaton 2006 Qu A 01 l-4). They have volcanic inclusions in their fabric. This form occurred during the second and third centuries in the western Mediterranean (Rhône Valley, Corsica, Aquileia), Northern provinces and the Black Sea region. The site of production is not known. They contained different volcanic materials from that mentioned previously. Volcanic Fabric (Augite, volcanic piroxene, volcanic plagioclase). 3.3.2. Other types of vessels were also made in volcanic regions. The place of production are currently unknown. Type F 66: Pergamon, Type J 46: Pergamon, Type M 240: Marseille 1884; Wreck Heliopolis (Type F 66: Pergamon Tr 2728-80 N Zi; Type J 46: Pergamon 121PE 80 10a Zi; Type M 240: Marseille 1884; Heliopolis DRASM 19117-16). 4. Conclusion The one-handled jars and amphorae collected at different sites show that the wine produced in the Ephesus region was consumed across the Mediterranean, and even in India from the Hellenistic to the Late Roman period. The petrological analysis of the amphorae produced elsewhere shows particles of volcanic origin or other types of fabric which are markedly different from those produced in Ephesus. It is clear that the forms presented here were also produced at different locations, and we will need to continue to search for these production centres. Acknowledgements While I was collecting the amphorae, the archaeologists of many museums and institutes generously helped me: Philip Bes, Michel Bonifay, Zdenko Brusić, John Camp, Darío Bernal‑Casasola, Sophia Didioumi, David Djaoui, Skerlou Elpida, Melina Filimonos-Tsopotou, Angeliki

Giannikouri, Smiljan Gluščević, Horacio González Cesteros, Alain Grandieux, Vladimir Kovačić, Georg Ladstätter, Luciana Mandruzzato, Franca Maselli-Scotti, Clementina Panella, Mira Pavletić, Andrei Opaiţ, Jeroen Poblome, Florence Richez, Giorgio Rizzo, Christa Schauer, Eleni Schindler-Kaudelka, Fani Seroglou, Roberta Tomber, Verena Vidrih Perko and 4th Ephorate of Byzantine Antiquities in Rhodes and Kos. Special thanks are due to my brother Gábor Bezeczky who translated the manuscript and Helka Németh who made the digital drawings for press. I am grateful to Stephen Greep who checked the language of the English version.

Bibliography Abadie, C. 1989. Les amphores protobyzantines d’Argos (IVe-VIe siècles). In V. Déroche and J.- M. Spieser (eds), Recherche sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 47–56. Athens, École française d’Athènes. Amores Carredano, F., García Vargas, E. and González Acuña, D. 2007. Ánforas tardoantiguas en Hispalis (Sevilla, España) y el comercio mediterráneo. 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 (I): 133–146. Oxford, Archaeopress. Annis, B. 1976. Amphora Sixti. In Festoen opgedragen aan A.N. Zadoks and Josephus Jitta bij haar zeventigste verjaardag. Scripta Archaeologica Groningana 6: 29–40. Groningen-Bussum, Wolters-Noordhoff. Arsen’eva, T. M. and Naumenko, S. A. 1992. Usad’by Tanaisa (‘Estates of Tanais’). Moscow, RAN In-t Arkheologii (in Russian). Berti, F. 1986. Le anfore di età romana del Ferrarese. Primo contributo ad una loro classificazione. In La civiltà comacchiese e pomposiana dalle origini preistoriche al tardo medioevo. Atti del Convegno nazionale di studi storici (Comacchio 17-19 maggio 1984): 183–209. Bologna, Nuova Alfa. Bezeczky, T. 2005. Late Roman Amphorae from the Ephesian Agora. In F. Krinzinger (ed.), Spätantike und mittelalterliche Keramik aus Ephesus. Archäologische Forschungen 13: 203–223. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Bezeczky, T. 2013. The Amphorae of Roman Ephesus. Forschungen in Ephesus 15/1. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Bezeczky, T. 2014. Contenitori monoansati delle Terme del Nuotatore di Ostia. In C. Panella and G. Rizzo (eds), Ostia VI, Le Terme del Nuotatore. Studi Miscellanei 38: 387–390. Rome, L’Erma di Bretschneider. Brusić, Z. 1988. Antička luka u Polačama na otoku Mljetu (Der antike Hafen in Polače auf der Insel Mljet). Izdanja Hrvatskog Arheološkog Društva 12: 139–150.

385

Tamás Bezeczky†

Cirelli, E. 2014. Typology and diffusion of amphorae in Ravenna and Classe between the 5th and 8th centuries AD. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeological Reports International Series 2616 (I): 541–552. Oxford, Archaeopress. De Staebler, D. P. 2012. Roman Pottery. In Ch. Ratté and P. D. De Staebler (eds), The Aphrodisias Regional Survey. The Aphrodisias Papers 5: 59–86. Mainz, Verlag Philipp von Zabern. Didioumi, S. 2014. Local pottery in the island of Cos, Greece from the early Byzantine period. A preliminary report. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeological Reports International Series 2616 (I): 169–180. Oxford, Archaeopress. Engelmann, H. 1986. Degustation von Götterwein (I.v.E. 2076). Zeitschrift für Papyrologie und Epigraphik 63: 107–108. Engelmann, H. 2011. Inschriften und Heiligtum. In U. Muss (ed.), Der Kosmos der Artemis von Ephesos. Sonderschriften des Österreichischen Archäologischen Instituts 37: 33– 44. Vienna, Verlag des Österreichischen Archäologischen Institutes. Feugère, M., Joly, J.-P., Pellecuer, C. and Peyre, A. 1987. La nécropole gallo-romaine tardive de La Brèche (Laudun, Gard). Premières données. Archéologie en Languedoc 4: 81–89. Fiaccadori, G. 1983. I frammenti iscritti. In G. Bermond Montanari (ed.), Ravenna e il porto di Classe, venti anni di ricerche archeologiche tra Ravenna e Classe. Fonti e Studi 7: 238–241. Bologna, University Press. Grace, V. R. 1979. Amphoras and the ancient wine trade. Agora Picture Book 6. Princeton, American School of Classical Studies at Athens. Hautumm, W. 1981. Studien zu Amphoren der spätrömischen und frühbyzantinischen Zeit. Fulda. Hayes, J. W. 1976. Pottery: stratified Groups and Typology. In J. H. Humphrey (ed.), Excavations at Carthage 1975, conducted by the University of Michigan: 47–123. Tunis, Institut National d’Archéologie et d’Art - American Schools of Oriental Research. Hayes, J. W. 1983. The Villa Dionysos excavations, Knossos: The pottery. Annual of the British School at Athens 78: 97–169. Hayes, J. W. 1992. Excavations at Saraçhane in Istanbul. Volume 2. The Pottery. Princeton, Princeton University Press. Japp, S. 2014. Micaceous waterjars in Pergamon: a specific variant of the Late Roman 3 amphora. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. The Mediterranean: A Market without Frontiers. British Archaeological Reports International Series 2616 (I): 143–149. Oxford, Archaeopress.

Knibbe, D. 1985. Der Asiarch M. Fulvius Publicianus Nicephorus, die Ephesischen. Handwerkerzünfte und die Stoa des Servilius. Jahreshefte des Österreichischen Archäologischen Instituts in Wien 56: 71–77. Ladstätter, S. 2008. Funde. In M. Steskal and M. La Torre (eds), Das Vediusgymnasium in Ephesos. Archäologie und Baubefund. Forschungen in Ephesos 14-1: 97–18. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Lang, M. 1955. Dated jars of Early Imperial times. Hesperia 24-4: 277–285. Lang, M. 1976. The Athenian Agora. Volume XXI. Graffiti and dipinti. Princeton, The American School of Classical Studies at Athens. Lemaître, S. 1997. L’amphore de type Agora F65/66, dite « monoansée ». Essai de synthèse à partir d’exemplaires lyonnais. In SFECAG, Actes du Congrès du Mans: 311–319. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Liddell, H. G. and Scott, R. 1968. Greek-English Lexicon. Oxford, Clarendon Press. Loeschcke, S. 1909. Ausgrabungen bei Haltern. Die keramischen Funde. Ein Beitrag zur Geschichte der augusteischen Kultur in Deutschland. Mitteilungen der Altertums-Kommission für Westfalen 5: 101–322. Lusuardi Sienna, S. 1977. Anfore. In A. Frova (ed.), Scavi di Luni, II. Relazione delle campagne di scavo 1972-1973-1974: 218–270. Rome, Bretschneider. Maioli, M. G. and Stoppioni, M. 1989. Classe, Podere Chiavichetta anfore di produzione locale. In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22-24 mai 1986). Collection de l’École française de Rome 114: 572–573. Rome, École française de Rome. Oleson, J. P. 1994. Amphoras. In J. P. Oleson (ed.), The Harbours of Caesarea Maritima. Result of the Caesarea Ancient Harbour Excavation Project 1980-1985. Vol II, The Finds and the Ship. British Archaeological Reports International Series 594: 3–24. Oxford, BAR Publishing. Outschar, U. 1993. Produkte aus Ephesos in alle Welt? Berichte und Materialien des Österreichischen Archäologischen Instituts 5-2: 46–52. Pacetti, P. and Paganelli, M. 2001. Anfore. In M. S. Arena, P. Delogu, L. Paroli, M. Ricci, L. Saguì and L. Vendittelli (eds), Roma dall’antichità al medioevo. Archeologia e storia: 211–219. Milan, Electa. Panella, C. 1973. Le anfore. 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 Miscellanei 21: 460–633. Rome, De Luca. Peacock, D. P. S. 1984. Petrology and origins. The amphorae: typology and chronology. In M. G. Fulford and D. P. S. Peacock (eds), Excavations at Carthage: The British Mission I, 2. The Avenue du President Habib Bourguiba, Salammbo : the pottery and other ceramic objects from the site: 6–28, 116–140. Sheffield, University of Sheffield. Peacock, D. P. S. and Williams, D. F. 1986. Amphorae and the Roman Economy. An Introductory Guide. London - New York, Longman.

386

Ephesus wines

Pieri, D. 2005. Le commerce du vin oriental à l’époque byzantine (Ve-VIIe siècles). Le témoignage des amphores en Gaule. Bibliothèque archéologique et historique 174. Beirut, Institut français du Proche-Orient. Rautman, M. L. 1995. Two Late Roman Wells at Sardis. Annual of the American Schools of Oriental Research 53: 37–84. Remolà i Vallverdù, J. A. 2000. Las ànforas tardo-antiguas en Tarraco (Hispania tarraconensis). Siglos IV-VII d. C. Instrumenta 7, Barcelona, Universitat de Barcelona. Riley, J. A. 1979. The Coarse Pottery from Berenice. In J. A. Lloyd (ed.), Excavations at Sidi Khrebish Benghazi (Berenice) II. Supplement to Libya Antiqua V (II): 91–467. Tripoli, Society for Libyan Studies. Robinson, H. S. 1959. The Athenian Agora. Volume V. Pottery of the Roman Period, Chronology. Princeton, The American School of Classical Studies at Athens. Sauer, R. 1995. Bericht über die archäometrischen Untersuchungen für das FWF-Projekt Nr. 9280. Produktionszentren späthellenistischer und römischer Keramik an der W-Küste Kleinasiens (mit einem Anhang von U. Outschar). Unpublished Manuscript. Sauer, R. 2013. Petrology. In T. Bezeczky (ed.), The Amphorae of Roman Ephesus: 199–211. Forschungen in Ephesus 15/1. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Schulten, A. 1906. Zwei Erlasse des Kaisers Valens über die Provinz Asia. Jahreshefte des Österreichischen Archäologischen Instituts in Wien 9: 40–70. Taeuber, H. 2005. Graffiti und Dipinti. In H. Thür (ed.), Das Hanghaus 2 in Ephesos. Die Wohneinheit 4. Baubefund, Ausstattung, Funde. Forschungen in Ephesos 8-6: 132–143. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Tomber, R. 1988. Pottery from the 1982-83 excavations. In J. H. Humphrey (ed.), The Circus and a Byzantine Cemetery at Carthage, vol. 1: 437–528. Ann Arbor, The University of Michigan Press.

Tomber, R. 1999. Pottery from the sediments of the Inner Harbour (area I 14). In K. G. Holum, A. Raban and J. Patrich (eds), Caesarea Papers 2. Herod’s temple, the provincial governor’s praetorium and granaries, the later harbour, a gold coin hoard, and other studies. Journal of Roman Archaeology Supplementary Series 35: 295–322. Portsmouth RI, Journal of Roman Archaeology. Tomber, R. 2006. The pottery. In V. A. Maxfield and D. P. S. Peacock (eds), Survey and Excavations. Mons Claudianus 1987-1993, Vol III, Ceramic Vessels and Related Objects from Mons Claudianus. Fouilles de l’Institut français d’archéologie orientale 54: 1–236. Cairo, Institut français d’archéologie orientale. Tomber, R. 2008. Indo-Roman Trade. From pots to pepper. London, Duckworth. Vidrih Perko, V. and Žbona Trkman, B. 2005. Ceramic finds from Ajdovščina - Fluvio Frigido. An early Roman road station and late Roman fortress Castra. Rei Cretariae Romanae Fautorum Acta 39: 277–286. Williams, D. F. 1982. The petrology of certain Byzantine amphorae. Some suggestions as to origins. In Colloque sur la céramique antique (Carthage 1980). CEDAC Dossier 1: 99–110. Carthage, Centre d’Études et de Documentation Archéologique de la Conservation. Williams, D. F. 2005. Agora F 65-66. Late Roman Amphora 3. In Roman amphorae. A digital resource, University of Southampton, viewed 2017. . Zabehlicky, H. 1999. Die Grabungen im Hafen von Ephesos 1987-1989. In H. Friesinger and F. Krinzinger (eds), 100 Jahre österreichische Forschungen in Ephesos. Akten des Symposions (Wien 1995). Archäologische Forschungen 1/ Denkschriften der philosophischhistorischen Klasse 260: 479–484. Vienna, Verlag der Österreichischen Akademie der Wissenschaften. Zeest, I. B. 1960. Keramičeskaja tara Bospora. Materialy i issledovanija po archeologii SSSR 83. Moscow, Institut arheologii (in Russian).

387

Egyptian amphorae from the Hellenistic to the Late Roman periods Delphine Dixneuf

Aix Marseille Univ, CNRS, LA3M, Aix-en-Provence, France

Abstract: Archaeological research undertaken in Egypt over the past 25 years has led to the discovery of large assemblages of amphorae that have allowed for the establishment of a new typological and chronological classification of Egyptian amphorae, from the beginning of the Hellenistic period to beginning of the medieval era. Four archaeological sites have provided the basis for this classification: Pelusium (Sinai Peninsula), Alexandria, Buto (Western Delta) and Bawit (Middle Egypt). Key words: Egypt; amphorae; Pelusium; Alexandria; Buto; Bawit.

1. Introduction Archaeological research undertaken in Egypt over the past 25 years has led to the discovery of large assemblages of amphorae that have allowed for the establishment of a new typological and chronological classification of Egyptian amphorae, from the beginning of the Hellenistic period to beginning of the medieval era (Dixneuf 2011). Four archaeological sites have provided the basis for this classification: Pelusium (Sinai Peninsula), Alexandria, Buto (Western Delta) and Bawit (Middle Egypt). Two large families of clay have been used for the production of Egyptian amphorae. The first is that of the Mareotid clays; after firing they present a light brown to buff colour, sometimes red. The texture is medium fine and generally sandy, with white and grey inclusions, sometimes fragments of seashells. The second family includes the alluvial clays created by Nile silt. Attested all along the Nile valley, in the Fayyum and Delta, these siliceous clays are brown to red in colour, sometimes with a black core. They contain grains of quartz, mica and vegetal particles, all of varying quantity.

for storing wine and other foodstuffs are also affected by this development, the beginnings of which can be seen from the start of the 4th century BC (Defernez and Marchand 2006: 73). Potters were using local clay to create new amphora models, largely inspired by Aegean containers that circulated in Egypt. During the Hellenistic period, these amphorae, known today as AE 1 and AE 2 (Amphores Égyptiennes 1 and 2; Figure 1), were principally produced, as far as we know, on the north coast to the west of Alexandria (Majcherek and Shennawi 1992: 133), in the Mareotid (Empereur and Picon 1998: 77, fig. 2-3), in the Delta (Coulson, Wilkie and Rehard 1986:

It seems entirely justified to suppose that each form and model of amphora should correspond to a specific product; the general shape might advertise the content. Nevertheless, it is now well established that it is not the case for many types of amphorae and the situation is not that simple. We will see what questions arise for Egyptian productions, but one should already note that the determination of contents started very late in Egypt compared to research in this field carried out in Western Europe.

b

0

10 cm

2. Hellenistic amphorae From the beginning of the Hellenistic period, the repertory of Egyptian ceramics expanded with new shapes introduced from the Greek world. One can observe this change in the different categories of receptacles, in particular table and cooking ware. Containers intended

a

Figure 1. Hellenistic amphorae AE 1 and AE 2 (Empereur and Picon 1998: 77, fig. 2-3).

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 389–398

Delphine Dixneuf

543–550; Kenawi 2012: 312 and 315–316, fig. 6-7) and in the Fayyum (Marchand, Chang and Nannucci 2018). While essentially traded within Egypt, these containers have also been attested, in limited quantities, on a few Eastern Mediterranean sites. 3. Roman amphorae From the end of the Hellenistic and beginning of the Roman period, Egyptian amphora types gradually differ from Aegean and Greek models. One can observe the emergence of a new form, commonly designated by the term ‘bitroncoconic’. The morphological characteristics of this new family are the following: an elongated cylindrical neck, smooth or fluted handles affixed to the upper part of the neck, below or touching the rim, a cylindrical or bulbous body emphasized by a central depression. This form, that was present in Egypt from the beginning of the Roman period until the first half of the 8th century AD, displayed various regional variants of which only the principal types will be presented here. The workshops on the southern shores of Lake Maryut produced essentially three forms of AE 3 amphorae: two bitronconical forms produced between the end of the 1st and the 3rd century (AE 3-1.4 and AE 3-1.5; Figure 2, a-b) — these two forms differ one from the other by the greater height of the first and the section of the rim, which is either rolled or triangular on the second — and a very particular cylindrical shape (AE 3-1.6; Figure 2, c). In fact, the amphorae of this last group have a long cylindrical body and neck; the junction of the neck and the body is marked by a rounded but not very pronounced shoulder. The rim, roughly shaped, forms a high, thick band with a more or less protruding internal flange most probably designed to fix a stopper. The badly managed firing — one observes numerous variations of colour in the breaks and on the surfaces — the rough manufacture and the highly stereotyped shape, are evidence of mass production, quasi-industrial, of these amphorae between the 2nd and 3rd centuries. It seems highly probable that the AE 3 amphorae of the Mareotid were destined for the storage and trade of locally produced wines. The study of amphora material discovered on the site of Buto in the Delta has allowed us to identify a series of containers with an extremely bulbous body, marked with a central depression, and a high cylindrical neck (AE 3-2; Figure 2, d). Most probably the product of local workshops not far from Buto in the western Delta, these containers seem to have been produced between the end of the 1st and the 2nd century. They are of alluvial clay, slightly sandy and red to brown in colour. Sometimes a white slip has been applied to the upper part of the amphora. Their contents are not known; however, several examples display a perforation on the neck that might suggest that they held wine. The sites of the Fayyum have provided several amphorae whose shape is characterized by a spherical body that

gradually thins towards a conical bottom (AE 3-3.1; Figure 2, e). The neck is cylindrical, often ringed, and finishes with a small rim, triangular in section (Bailey 2007: 235, fig. 1.7). The dating of this amphora on the site of Tebtynis runs from the 1st to the beginning of the 2nd century (Marangou and Marchand 2007: 291–292, fig. 145-150 and 292, fig. 153-154; Ballet and Południkiewicz 2012: 182, 327, fig. 791), and even into the 3rd century on the site of Medinet Madi where some of these amphorae were lined with resin (Silvano et al. 2007: 1707). Amphorae produced in Middle Egypt are of two types. The first has been identified on the site of El-Ashmunein, and would seem to be, in accordance with its discovery, a local product (Bailey 1998: 125–129, pl. 78, U1). This amphora has a high, regularly fluted, cylindrical neck and a small rim, triangular in section, the upper edge of which is concave (AE 3-3.2; Figure 2, f). The handles are affixed to the upper part of the neck, touching the rim. The body has carinated shoulders and a central depression, and stretches down to a cylindrical base. The second type corresponds to an amphora that one finds principally on the sites of Egypt’s eastern desert and one supposes that the production workshops were located in Middle Egypt (Bailey 1998: pl. 72-73; Brun 2007: 508–509, fig. 3.4 and 6). This amphora is characterized by a high, regularly fluted, cylindrical neck and a triangular sectioned rim with a protruding exterior (AE 3-3.3; Figure 2, g). The shoulder is pronouncedly carinated and the body thins gradually downwards to a pointed conical base. Dated principally from the mid-1st and the 2nd century on the site of the eastern desert (Tomber 2006: 148, type 12; Brun 2007: 505–509), this amphora appears to have been used primarily for storing wine. Upper Egypt also played a role in the creation and evolution of the morphological repertory of Egyptian amphorae. Indeed, the potters of the area between Quft (Coptos) and Luxor produced a specific series of containers that fit quite clearly into the tradition of the AE 2 amphorae of Graeco-Roman period produced in that area (Figure 2, h-i). However, the trade in these amphorae is rather more directed towards the commerce of Egypt’s eastern desert and the ports of the Red Sea (especially Brun 2007: 502–523; Tomber 2007: 525–536). The AE 4 amphora produce in Egypt and, more precisely, on the banks of Lake Maryut not far from Alexandria (Empereur 1986: 605–606), are attached to the large amphora family of Dressel 2/4 produced on the island of Kos in the Aegean Sea and widely imitated in the Mediterranean and even beyond, all the way to Great Britain (Tomber 1992: 140). The general morphology is characterized by significant height (up to 148cm). They display an elongated neck in the form of a truncated cone with the middle section ringed (Figure 3). The lip is rolled; the bifid handles are affixed to the rim and above the junction, marked by a protrusion, of the neck and the shoulder. The body is cylindrical and extends downwards to a knob-shaped base. The dating covers the period from

390

Egyptian amphorae from the Hellenistic to the Late Roman periods

20 cm

0

a

b

f

g

c

h

d

i

e

j

Figure 2. Roman amphorae AE 3 a. Dixneuf 2011: 322, fig. 87 no.P11538; b. Empereur and Picon 1998: fig. 4; c. Dixneuf and Lecuyot 2007: 141, fig. 13; d. Dixneuf 2011: 332, fig. 98 no.169; e. Bailey 2007: 235, fig. 1.7; f. Spencer and Bailey 1982: fig. 35a; g. Brun 2007: p. 508, fig. 3.4; h. Mysliwiec 1987: no.947; i. Tomber 2006: 144, fig. 1.55; j. Brun 2007: 508, fig. 2 no.2.

391

Delphine Dixneuf

0

10 cm

Figure 3. Roman amphorae AE 4 (Empereur 1986: fig. 2-5). the second quarter of the 1st to the middle of the 3rd century (Empereur 1986: 608). As with AE 3 amphorae produced in the Mareotid, these containers were also designed to hold locally made wine. 4. Late antiquity-beginning of the Arab period During late antiquity and at the beginning of the Arab period, four great families of containers were produced in Egypt. The late AE 3 bitronconic amphorae fit into the typological and chronological continuity of AE 3 amphorae produced during the Roman period and, more precisely, of the containers manufactured from alluvial clay and from production sites located in the Delta, the Fayyum and along the Nile valley in Middle Egypt. The principal morphological characteristics of these amphorae (AE 3T, ‘Amphore Égyptienne 3 Tardive’) are the following: a body with two bulges separated by a central depression, a high regularly ringed, cylindrical neck and a small conical bottom with a ring-shaped bulge (Figure 4). The handles

are affixed to the upper part of the neck, touching the rim, which is either rolled or as a band with a concave exterior, rounded lip and a flange at its base. During the 7th and beginning of the 8th century, the shape changed and became noticeably thinner. The general designation AE 5/6 refers to spherical or egg-shaped containers, known in the Levant as LRA 5/6 or bag-shaped amphora. In Egypt, such amphorae are clearly imitations of Levantine products since we observe no original prototypes. Two types of clay were used in their manufacture: a calcareous clay characteristic of the Abu Mina (Engemann 1992: 153–159) and Mareotid, and an alluvial clay that might correspond to the production of Kom Abu Billu/Terenuthis in the Delta (Ballet 1994: 355–363); this is a hard clay with a red and grey-blue core with blackened vegetal particles. On the other hand, the macroscopic analysis of a number of AE 5/6 amphorae discovered on the site of Bawit in Middle Egypt has led to the possibility of there being another production centre, not yet found, in this region.

392

Egyptian amphorae from the Hellenistic to the Late Roman periods

c

a

b

0

10 cm

d

Figure 4. Late Roman amphorae AE 3T. a. Dixneuf and Lecuyot 2007: 139, fig. 7; b. Dixneuf 2011: 345, fig. 121 no.221; c. Dixneuf 2011: 351, fig. 127 no.252; d. Dixneuf 2011: 352, fig. 128 no.254. Two principal forms have been recognized: the first has a spherical body and a rounded base with no foot. Small handles are affixed to the upper part of the body. The second is a modest-sized amphora with a pear-shaped or cylindrical body and rounded base. These two forms were manufactured in both calcareous (Figure 5, a-e) and alluvial clay (Figure 5, f-k) (AE 5/6-1 and 2). On the site of Kellia, the calcareous clay containers appear towards the first quarter of the 7th century, while those of alluvial clay are not attested before the middle of the 7th century (Bonnet 1983: 442–443; Ballet 2003: 146). The AE 7 amphorae, generally termed LRA 7, are the most widely spread and distributed wine containers throughout Egyptian territory during late antiquity and the first centuries of the Arab period. The AE 7 type is a slender amphora of modest capacity — 6 to 7 litres — made of a brown, micaceous alluvial clay. At the same time, this family includes many variants (Figure 6), a fact that can be explained by the large number of workshops that must have produced these amphorae and the long time period of their use, from the second half of the 4th century until the Fatimid era (Gayraud 2003: 558). The workshops that manufactured this amphora were principally located in Middle Egypt, in the heart of the great wine-growing region, between Qarara and El-Ashmunein, as well as further south, between Esna and Edfu (for a bibliography regarding AE 7 production workshops, see Dixneuf 2011: 157–163).

The last group, known as AE 8, includes two families of containers that clearly imitate amphorae imported into Egypt. The first group is of imitations of LRA 2/ LRA 13 amphorae. The amphorae have a tronconical or cylindrical neck and rather large, spherical or eggshaped body (Figure 7, a). One can note several networks of fine stripes on the shoulder and base. According to ceramological investigations undertaken at Kellia, AE 8-1 containers have only been identified in the most recent occupation phases, that is from the mid-7th century and the first half of the 8th century. There appearance would then be contemporary with that of AE 5/6 alluvial clay amphorae (Bonnet 1994: 363; Ballet 2003: 149), a dating that is confirmed by G. Majcherek, according to studies on the amphorae of Kom el-Dikka in Alexandria (Majcherek 2004: 235). The second group is that of imitations of LRA 1 amphorae (Figure 7, b-c). This is an amphora with a cylindrical body marked by a slight central depression, and a convex umbilicated base. The neck is cylindrical and ends with a small rolled rim with a projected ridge. This form is attested from the mid-7th century at Bawit as well as Kellia (Ballet 2003: 153). Acknowledgment Thanks to Colin Clement for the English translation.

393

Delphine Dixneuf

a

d

b

e

c

f

g

i

k

h

j 0

10 cm

Figure 5. Late Roman amphorae AE 5/6. a. Egloff 1977,: 117, no.186; b-e. Engemann 1992: 157, fig. 1, 10a, 11a, 12a; f. Dixneuf 2011: 358, fig. 136 no.279; g. Bailey 1998: pl. 85, forme W7; h. Dixneuf 2011: 360, fig. 140 no.292; i. Dixneuf 2011: 360, fig. 141 no.295; j. Dixneuf 2011: 361, fig. 142 no.299; k. Dixneuf 2011: 361, fig. 143 no.301.

394

i

a

b

j

k

c

d

l

e

m

f

n

g

0

h

10 cm

Egyptian amphorae from the Hellenistic to the Late Roman periods

Figure 6. Late Roman amphorae AE 7. a. Şenol 2001: 394, fig. 10 no.28; b-d. Dixneuf 2011: 367-368, fig. 155-157 nos317, 320, 322; e. Lecuyot and Pierrat 1992: 175, fig. 1; f-n. Dixneuf 2011: 370-373, 376-377, fig. 159-162, 164, 168-171 nos326, 330, 334, 335, 339, 351, 353, 354, 356.

395

Delphine Dixneuf

a

c

0

5 cm

b Figure 7. Late Roman amphorae AE 8. a. Bonnet 1994: fig. 225; b-c. Dixneuf 2011: 385, fig. 179 nos382-383.

396

Egyptian amphorae from the Hellenistic to the Late Roman periods

Bibliography Bailey, D. M., 1998. Excavations at el-Ashmunein 5. Pottery, Lamps and Glass of the Late Roman and Early Arab Periods. British Museum Expedition to Middle Egypt 5. London, British Museum. Bailey, D. M., 2007. A form of Amphores Égyptiennes 3 from the South-West Fayum. In S. Marchand and A. Marangou (eds), Amphores d’Égypte de la Basse Époque à l’époque arabe. Cahiers de la Céramique égyptienne 8 (I): 227-236. Ballet, P., 1994. Un atelier d’amphores Late Roman Amphora 5/6 à Kôm Abou Billou (Egypte). Chronique d’Égypte 69/138: 353–365. Ballet, P., 2003. Classification de la céramique. In P. Ballet, N. Bosson and M. Rassart-Debergh, Kellia 2. L’ermitage copte QR 195. Céramique, inscriptions, décors. Fouilles de l’Institut français d’archéologie orientale 49: 69–207. Cairo, Institut français d’archéologie orientale. Ballet, P. and Południkiewicz, A., 2012. Tebtynis V. La céramique des époques hellénistique et impériale (campagnes 1988-1993) : production, consommation et réception dans le Fayoum méridional. Fouilles de l’Institut français d’archéologie orientale 68. Cairo, Institut français d’archéologie orientale. Bonnet, F. 1983. Poteries, verres, monnaies, décors et inscriptions. In R. Kasser (ed.), Survey archéologique des Kellia (Basse-Egypte). Rapport de la campagne 1981. Mission suisse d’archéologie copte de l’Université de Genève, EK 8184/1: 423–480. Leuven, Peeters. Bonnet, F., 1994. Le matériel archéologique récolté en 1977, 1982 et 1983 aux Qouçoûr er-Roubâ’îyât. In R. Kasser (ed.), Explorations aux Qouçoûr er-Roubâ’îyât’. Rapport des campagnes 1982 et 1983. Mission suisse d’archéologie copte de l’Université de Genève, EK 8184/2: 349–406. Leuven, Peeters. Brun, J.-P., 2007. Amphores égyptiennes et importées dans les praesidia romains des routes de Myos Hormos et de Bérénice. In S. Marchand and A. Marangou (eds), Amphores d’Égypte de la Basse Époque à l’époque arabe. Cahiers de la Céramique égyptienne 8 (II): 505–523. 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. Athens, École française d’Athènes. Defernez, C. and Marchand, S. 2006. Imitations égyptiennes de conteneurs d’origine égéenne et levantine (VIe s.-IIe s. av. J.-C.). In B. Mathieu, D. Meeks and M. Wissa (eds), L’apport de l’Égypte à l’histoire des techniques. Méthodes, chronologie et comparaisons. Bibliothèque d’Étude 142: 63–99. Cairo, Institut français d’archéologie orientale. Dixneuf, D. 2011. Amphores égyptiennes. Production, typologie, contenu et diffusion (IIIe siècle avant J.-C.-IXe siècle après J.-C.). Études alexandrines 22. Alexandria, Centre d’Études Alexandrines.

Dixneuf, D. and Lecuyot, G. 2007. Note préliminaire sur les amphores découvertes par la mission « Recherches sur les ateliers hellénistiques et romains de Bouto » (2002-2003). In S. Marchand and A. Marangou (eds), Amphores d’Égypte de la Basse Époque à l’époque arabe. Cahiers de la Céramique égyptienne 8 (I): 135–141. Egloff, M. 1977. Kellia. La poterie copte. Quatre siècles d’artisanat et d’échanges en Basse-Égypte. Recherches Suisses d’Archéologie Copte III. Genève, Georg. Empereur, J.-Y., 1986. Un atelier de Dressel 2-4 en Égypte 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. Athens, École française d’Athènes. 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. Athens, École française d’Athènes. Engemann, J. 1992. A propos des amphores d’Abou Mina. Cahiers de la céramique égyptienne 3: 153–159. Gayraud, R.-P., 2003, La transition céramique en Égypte, VIIe-IXe siècles. In M. Bonifay (ed.), Table-ronde De Rome à Byzance ; de Fostat à Cordoue : évolution des faciès céramiques en Méditerranée (Ve - IXe s.), in C. Bakirtzis (ed), VIIe Congrès International sur la Céramique Médiévale en Méditerranée (Thessaloniki, 11-16 Octobre 1999) : 558–562. Athens, Ministère de la Culture. Kenawi, M. 2012. Beheira Survey: Roman pottery from the Western Delta of Egypt. Surface pottery analysis Kilns. Rei Cretariae Romanae Fautorum Acta 42: 309–317. Lecuyot, G. and Pierrat, G. 1992. À propos des lieux de production de quelques céramiques trouvées à Tôd et dans la vallée des Reines. Cahiers de la céramique égyptienne 3: 173–180. 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. Aarhus, Aarhus University Press. Majcherek, G. and Shennawi, A. 1992. Research on amphorae production on the Northwestern Coast of Egypt. Cahiers de la céramique égyptienne 3: 129–136. Marangou, A. and Marchand, S. 2007. Conteneurs importés et égyptiens de Tebtynis (Fayoum) de la deuxième moitié du IVe siècle av. J.-C. au Xe siècle apr. J.-C. (1994-2002). In S. Marchand and A. Marangou (eds), Amphores d’Égypte de la Basse Époque à l’époque arabe. Cahiers de la céramique égyptienne 8 (I): 239–294. Marchand, S., Chang, R. L. and Nannucci, S. 2018. Philadelphie 2018. Amphores égyptiennes locales AE 1 en pâte calcaire. Époque ptolémaïque, seconde moitié du IIIe s. av. J.-C. Bulletin de liaison de la Céramique Égyptienne 28: 125–154. Mysliwiec, K. 1987. Keramik und Kleinfunde aus der Grabung im Tempel Sethos ‘I. in Gurna. Archäologische Veröffentlichungen 57. Mainz, von Zabern.

397

Delphine Dixneuf

Şenol, A. K. 2001. Amphoras from the Necropolis of Gabbari. In J.-Y. Empereur and M.-D. Nenna (eds), Nécropolis 1. Études alexandrines 5: 369–396. Alexandria, Centre d’Études Alexandrines. Silvano, F., Colombini, M. P., Modugno, F. and Ribechini, E. 2007. Roman amphorae from Fayum Oasis (Medinet Madi). In J.-C. Goyon and C. Cardin (eds), Proceedings of the Ninth International Congress of Egyptologists, Vol. 2. Orientalia lovaniensia analecta 150: 1708–1711. Leuven, Peeters. Spencer, A. J. and Bailey, D. M. 1982. British Museum Expedition to Middle Egypt, Ashmunein (1981). London, British Museum.

Tomber, R. 1992. Early Roman pottery from Mons Claudianus. Cahiers de la céramique égyptienne 3: 137–142. Tomber, R. 2006. The pottery. In V. A. Maxfield and D. P. S. Peacock (eds), Survey and Excavations. Mons Claudianus 1987-1993, Vol III, Ceramic Vessels and Related Objects from Mons Claudianus. Fouilles de l’Institut français d’archéologie orientale 54: 1–236. Cairo, Institut français d’archéologie orientale. Tomber, R., 2007, Early Roman Egyptian amphorae from the Eastern Desert of Egypt: A chronological sequence. In S. Marchand and A. Marangou (eds), Amphores d’Égypte de la Basse Époque à l’époque arabe. Cahiers de la Céramique égyptienne 8 (II): 525–536.

398

Case studies: From the West to the East

The contents of amphora type T-7.4.3.3 (former type Mañá C2b): ancient problematic and new research perspectives Max Luaces

Scientific and technical staff attached to the GREPURE (Grecia Púnica Redescubierta) research project, HUM 152 Research Group, University of Seville

Abstract: Although its morphology has been well known for many decades, defining the contents of the T-7.4.3.3 amphora has been a challenge, as this Punic type was previously clustered with other amphoric forms. Furthermore, contradictory evidence has been discovered since the 1990s as to whether the T-7.4.3.3 amphora contained garum, fish or wine. In this paper, we will try to clarify some of the existing confusion by specifying the typological and contextual features of this amphora from the area of the Strait of Gibraltar. We will then review several sources of documentation related to the identification of its contents, taking into account archaeological and historical data. In the end, even if we have proof of multiple products for this ancient container, fish products seem to be the main content that this amphora transported during the Late Republican era. Key words: Late Republican economy; Punic amphorae; fish product; Circle of the Strait; Maritime archaeology.

1. Introduction The identification of the possible contents for the amphora type known today under the appellation of T-7.4.3.3 may not seem to be problematic. This type was one of the various forms of amphora that made up the former group of the Mañá C2b type, essentially in relation to the SG7.4.3.0 of Ramon Torres (Figure 1). It is a particular class of container, dated from the Late Republican era, which was associated with the economic activities of the Circle of the Strait.1 Until now, the T-7.4.3.3 amphora was generally considered to have been used for the maritime transport of garum and fish products.2 It is through its packaging of these goods that it is considered in the typology of J. Ramon Torres (Ramon Torres 1995: 264–265). If the whole matter seemed settled, several elements have recently suggested that we should question the traditional analysis of the contents of this type of amphora. First of all, the discovery of an epigraphic inscription indicating a content related to wine products, on the neck of a T-7.4.3.3, led to nuance the first interpretations on the subject. Recent archaeological data concerning the distribution of this type, in connection with other discoveries in the sector of the Strait of Gibraltar, also brought new information. Finally, the existence of old typo-morphological confusions, between the diverse forms associated with the T-7.4.3.3 container, seem to have added new ambiguities. In these conditions, an update of the documentation relating to the contents of this amphora seems appropriate. Firstly, an inventory of all the data on the subject needs to be made, confronting  This paradigm defines a cultural area that combined both shores of the Strait of Gibraltar, associating several city-states related to the Phoenician colonization of the beginning of the Iron Age (Callegarin 2016; Bernal-Casasola 2016). 2  Recent studies carried out by the University of Cadiz and the University of Seville allowed for a better understanding of the ancient production patterns of garum and its by-products from the sector of the Strait of Gibraltar (García Vargas et al. 2014). 1

the multiple indications and cultural and chronological peculiarities of the T-7.4.3.3 type. However, it would be necessary before anything else to attempt to clear up the numerous confusions that characterize this specific container, from the point of view of its typology and of its identification, a subject that turns out to be much more complex than previously thought. 2. The typo-morphological confusions associated with the ‘SG-7.4.3.0’ as a first source of controversy To examine the contents of the T-7.4.3.3 type, the researcher needs to look at the preliminary identifications associated with it. Indeed, if this class and all of its subgroup relatives from the SG-7.4.3.3 (T-7.4.3.1, T-7.4.3.2 and T-7.4.3.3) (Figure 1) are all considered as garum containers, it is largely because of their assimilation with the Dressel 18 type (Botte 2009: 31; Guerrero Ayuso 1986: 175; Van Der Werff 1978: 182; Ramon Torres 1995: 265). One titulus pictus on this latter type, read as Hal(lex) Coc(tiva) or Hal(lex) Soc(iorum), has been interpreted for a long time as evidence for a fish content for this type (Dressel 1899: 681, no.4730). The fact that the amphora used for the traditional interpretation of the contents of the Dressel 18 has been untraceable for a long time, and that its technical drawing showed morphological similarities with the form SG-7.4.3.0, led to the assimilation of these diverse productions as one and the same type (Boube 1975: 170–171). Consequently, the contents of the related classes of the Ramon Torres typology were interpreted from the information of the Dressel 18 type. Nevertheless, as we shall have the opportunity to see further on, its connection to the T-7.4.3.3 type turns out to be inappropriate (Capelli et al. 2017: 28–30). Besides this link with the Dressel 18 type, the T-7.4.3.3 amphora possesses numerous typological names. This plurality of appellation is currently undermining the

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 401–407

Max Luaces

Figure 1. General overview of the T-7.4.3.3 amphora and its typological relatives, formerly regrouped under the Mañá C2 type (A: T-7.4.3.1; B: T-7.4.3.2; C: the T-7.4.3.3 type itself). analysis of the T-7.4.3.3 type and that of its contents. The preliminary characterization of this same amphora is known through the studies of Cintas, under the form names Cintas 312 and Cintas 313 (Cintas 1950: 233). The classification made by J. M. Mañá presented the opportunity to propose another characterization, under group C of his typology, a group that was divided between the C1 and C2 forms (Mañá de Angulo 1951: 40). The field work of Vuillemot in Algeria led him to regularly encounter these same morphologies, identified according to the names, AN2, AN6 and T7 (Vuillemot 1965: 302). The study by J.-P. Morel on the Moroccan site of Thamusida offered another characterization based on various forms of rim, in correspondence with the groups D1, D3 and D4 (Morel 1965: XLIX). The publication of the archaeological material of Uzita in Tunisia provided the opportunity for new typological proposals, the same morphologies being identified as Van der Werff 1 type (Van der Werff 1978: 177–181).

The scholar Ramon Torres suggested refining the classification of Mañá (Ramon Torres 1981), by distinguishing for the first time the Iberian origin of some of these morphologies, isolated by the Mañá C2b type. In 1986, V. M. Guerrero Ayuso made a first report on the profound confusion around the various forms of containers related to the Mañá C group (Guerrero Ayuso 1986: 149–163). He proposed additional precision, with the constitution of two sub-types (Mañá C1a and C1b). The same year, the study of C. Florido Navarro on the amphorae of Tartessos brought this archaeologist to observe similar amphorae, but he decided to distinguish this type by the name type XII. Finally, a complete redefinition of these various classes and morphologies was made by J. Ramon Torres in his recent typology (Ramon Torres 1995). The main merit of this classification is in the characterization of multiple types for each of the previous groups, according to the differences of

402

The contents of amphora type T-7.4.3.3 (former type Mañá C2b): ancient problematic…

rim, neck and bodies. Some of these distinctions were related to the specific origin of some of the distinct morphologies. The T-7.4.3.3 type was established from the numerous discoveries of this amphora in Spain and in Morocco. This form of container was then considered as the evolution of another Iberian amphora, the T-7.4.3.2 type, which seems to have appeared among the Punic communities of the Strait of Gibraltar between 150 and 140 BC. These two types (T7.4.3.2 and T-7.4.3.3) were more particularly adaptations of a first amphora, which was previously produced in the territories of modern Tunisia, named T-7.4.3.1 (Ramon Torres 1995: 210-213). To summarize, the type that we are analyzing here under the name of T-7.4.3.3 is regularly likened to other types: Dressel 18, Cintas 312/Cintas 313, Mañá C2, the AN2/AN6/T7, D1/ D3, Van der Werff 1, Mañá C2b, Type XII. However, in spite of their typo-morphological similarities, it turns out that these various namings refer in reality to several different amphoric productions. Each of them belonged to different chronologies and areas of production, and need thus to be clearly distinguished to enable us to better understand the Late Republican and Early Empire economy. In this particular case, the T-7.4.3.3 type is a specific shape of amphora that was mainly, if not exclusively, produced in the extended area of the Strait of Gibraltar, and that was only produced after the fall of Carthage. This production pattern is not the same as that of the Van der Werff 1 type, for example, or of all the container related with the Mañá C2 type. These precedents explain why Ramon Torres had divided his subgroup SG‑7.4.3.0 into three forms, even if they show very similar morphologies and are following the same typological and progressive line. Each of these three types refer to different contexts of production and chronologies, so it is not sensible to mix them up. It is probably because these various types were not well enough distinguished from each other, as a result of their similar forms, that we sometimes observe contradictory comments between the chronologies of production and their contents. In this instance, if we wish to analyze the specific contents of the T-7.4.3.3 type, it is imperative not to take into account the archaeological data of Tunisia, which should better be associated with the T-7.4.3.1 type. Furthermore, we should focus mainly on the data from the Late-Republican era, a period that corresponds to the era in the course of which our T-7.4.3.3 type appeared and was mainly produced. Although they need to be used cautiously, the data associated with the Early Empire should not be completely dismissed, as it seems that some later version of the T-7.4.3.3 type was manufactured until AD 50, a period that marked the final passage of ancient Morocco under Roman rule.3

2.1. Dressel 18 and T-7.4.3.3, a false typo-morphological assimilation The majority of comments related to the contents of the amphora T-7.4.3.3 were made with regard to its connection with the form number 18 of Heinrich Dressel’s table. As previously indicated, this latter class of amphora had previously been closely associated with the morphology which corresponds to the T-7.4.3.3 type. In the light of several recent discoveries, the connection between these two types is nonetheless clearly erroneous, the form 18 of the table of Heinrich Dressel being in reality a very specific production that has to be distinguished from the type T-7.4.3.3 (Figure 2). In this particular case, the Dressel 18 presents distinct characteristics from our T-7.4.3.3 type (Capelli et al. 2017: 29–30). This first amphora was an evolution of the Tunisian productions associated with the type van der Werff 1, which was manufactured during the 1st century AD. Various archaeometric analysis, carried out on Dressel 18 types discovered in Ostia and Pompeii, suggest that it was most probably a production from the eastern coastline of modern Tunisia.

 Note that the chronological situation is also complex for SG-7.4.3.3, because of a productive continuity in Tunisia of the Punic forms of the T-7.4.3.1 type till the end of the Republic and the Early Empire (Ben Jerbania 2013). I limit myself geographically, to try not to complicate further an already rather vague situation. 3

403

Figure 2. Representation of a Dressel 18 amphora (A) (from Capelli et al. 2017) and a T-7.4.3.3 amphora (B) Beyond their distinct production chronologies and area of provenance, these two types are in fact quite different from a typo-morphological point of view, as we can see.

Max Luaces

we should also look more closely at the archaeological documentation from several sets of T-7.4.3.3 amphorae from the western Mediterranean. 3.1. The documentation from the region of the Strait of Gibraltar

Figure 3. Reproduction of the titulus pictus on the Dressel 18 isolated by H. Dressel (from Dressel 1899: 681). Its reading by H. Dressel was already hypothetic and should be considered with caution. It is also important to note that the titulus pictus which testified that garum was contained in Dressel 18 is doubtful. The reading of Hal(lex) Coc(tiva) proposed by Heinrich Dressel was never questioned, but an examination of the transcription published in the corpus of Latin inscriptions leave numerous doubts (Figure 3). If we accept the relationship between the Dressel 18 and the Van der Werff 1 types, we should note that Van der Werff had considered the existence of a plurality of contents for his type, because of the diversity of the production area that he had noticed. In order to get a better understanding of these problematic, it seemed to him that: ‘the safest way is to see what were, by region, the products which entered the most into play for the export’ (Van der Werff 1978: 182). To conclude, it is advisable, on one hand not to assimilate the T-7.4.3.3 type with that of the Dressel 18 amphora, and therefore not to consider automatically T-7.4.3.3 amphora as garum containers. On the other hand, it is necessary to underline that the relationship between the Dressel 18, as a production specific to the Africa Vetus and the Early Empire era, and a garum container are not completely congruent, or at least that they must be qualified. However, the questions related to Dressel 18 amphora exceed the framework of this paper, and our interest focuses on the T-7.4.3.3 type. 3. Archaeological discoveries and historical data for a new characterization of the contents Having highlighted the clear distinction between the Dressel 18 type and the amphora T-7.4.3.3 we will now examine the contents of the latter. There are several opinions on the matter. The first one concerns the contextual information regarding the T-7.4.3.3 type in the area of the Strait of Gibraltar. In fact, looking at the economic and cultural situation in this region could give us some insights into the possible contents of this amphora, by examining both the historical data and the archaeological information. On the other hand,

The ancient sources offer us numerous information on the economic activities of the area of the Strait of Gibraltar. The Ancients insisted in particular on the importance of fishing activities. Several references refer to the export of salsamenta and garum by the main cities of the region, such as Gadir and Malaka (García Vargas and Ferrer Albelda 2001: 549–555; Lagóstena Barios 2001: 203–220). In addition, these same communities have brought us numerous proofs of a local production of T-7.4.3.3 amphorae, in connection with the export of processed fish products (Lagóstena Barrios 1996). Some production contexts of the T-7.4.3.3 amphora discovered in Cadiz have proved to be associated with fish factories. The site of Gregorio Marañón, situated in the eponymous street of modern Cadiz, is an illustration of such a connection (Blanco Jiménez 1991; Perdigones Moreno and Muñoz Vicente 1990). The different levels of this site offered various discoveries of wasters of the T-7.4.3.3 type, associated with the remains of fish tanks. Beyond this information on the contexts of the Strait of Gibraltar, we should take into account the sociocultural environment associated with the T-7.4.3.3 type. As mentioned before, this amphora was produced mainly during the Late Republican era. The various cities of the Circle of the Strait connected with its production also present material with Punic culture characteristics during this period. This can be observed in the fine wares and architectonic material, and even in the economic patterns (Luaces 2015), in the course of the Late Republican period. This evidence speaks of the continuity of the Phoenician and Punic culture, until the area of the Strait of Gibraltar had a more Roman character, during the Early Empire (Luaces 2017: 363–672). To envisage the production of garum, a typically Roman product that was connected to the food practices from the end of the Republic, in a region that was more Punic than Roman does not necessarily seem congruent. It is, at least, the case for the very specific case of the T-7.4.3.3 amphora, this type being deeply connected to the Punic tradition. However, such an argument is difficult to defend, as information on the sociocultural situation of the Strait of Gibraltar during the Late Republican period is particularly scare. To get a clearer picture of the issues, it would be necessary to consider other sources of information. The numerous archaeological excavations carried out in the region of the Strait of Gibraltar allowed us to obtain a whole set of empirical data on the T-7.4.3.3 amphora. Several recent discoveries have helped studies on the contents of this type. The first discovery concerns the excavation of ichthyological remains inside a T-7.4.3.2

404

The contents of amphora type T-7.4.3.3 (former type Mañá C2b): ancient problematic…

amphora of Gaditan origin. This container was discovered in some of the most ancient strata of the Roman town of Baelo Claudia (province of Cadiz, Spain), by the team of Darío Bernal‑Casasola. The study of these remains has allowed us to identify the remains of the anterodorsal trunk of a fish related to the tuna species (Bernal‑Casasola et al. 2007: 355–374). From a zoological point of view, these remains could correspond to some of the fish that can be found in the area of the Strait. Moreover, the identified remains seem to be more specifically related to a salsamenta kind of contents, a chosen set of salted fish parts, and not a sauce or a completely processed fish product. Although this discovery was made in an amphora T-7.4.3.2 and not a T-7.4.3.3, note that these two types are closely connected, as both are representative of the amphoric production of the area of the Strait of Gibraltar. The T-7.4.3.3 amphora is more exactly a direct evolution of the T-7.4.3.2, as we have shown above (Figure 1). Two examples of T-7.4.3.3 type found in the ancient city of Lixus also brought other empirical data. A first amphora was discovered in the archaeological contexts of a warehouse. Sediments inside have proved to be filled with grape seeds (Bonet Rosado et al. 2005: 131). The second container was discovered in another context of Lixus with remains of mollusks inside, mainly Mytiloida (mussels) (Aranegui Gascó 2004: 181). Although outside contamination is possible, these two discoveries open up the possibility of multiple contents for the amphora T-7.4.3.3. But first, we should consider the chronological features of these two objects, as both seem to be dated from the end of the Republic, or the Early Empire. We are confronted thus with maritime packaging used for commercial activities more than a century after the end of manufacturing of this same type in southern Spain. In addition, the malacofauna was a particularly difficult product to preserve during antiquity. Thus, these mollusks were probably not intended for long distance shipment. This amphora may instead have been intended for a local or regional market, such as eastern Mauritania. Let us now look at the discovery of various fragments of T-7.4.3.3 in productive contexts that were not directly connected to the coast. There is evidence for the production of this amphora in the Gaditan country, several kilometres from the coast (Sáez Romero 2008: 500–503). Furthermore, the discovery of a titulus pictus interpreted as VINUM/D, adds more evidence to the question about the interpretation this amphorae for fish only (García Vargas 1998: 65–68). The reading of this painted inscription opens up the possibility for wine contents for our Late Republican amphora. The indications of Strabo, concerning wine production around the Guadalquivir valley and the region of the Strait, illustrate the existence of this activity at least from the Republican period (Strabo III.2.6). We are once again confronted with a level of confusion, between data clarifying the presence of a fish content

and other evidence for different sea products, and even wine… In view of these diverse indications, the question of a multiple contents for the T-7.4.3.3 amphora needs to be considered. Nevertheless, it is necessary to mention that the data from Lixus concerns very specific forms of T-7.4.3.3, dated from the Early Empire, and the data for wine content is associated with artifacts that could have been storage vessels and related to domestic activities rather than commercial packaging. In the end, the archaeological documentation of the region of the Strait of Gibraltar raises many more questions than answers. However, the documentation of another Mediterranean region, more precisely Southern Gaul, offers further useful data. 3.2. The archaeological documentation from Southern Gaul Until not so long ago, the T-7.4.3.3 amphorae was rather rare in Southern Gaul. Some of our studies changed this fact thanks to the identification of a large number of these amphorae in this region (Luaces 2017: 383–454). I will focus here on a context clearly connected with commercial traffic, which could conveniently be compared with the data presented for the area of the Strait of Gibraltar. Besides some isolated amphorae, numerous T-7.4.3.3 were discovered in various underwater deposits in Southern Gaul. The ‘Chrétienne M2’ shipwreck, in particular, provided a large number of these containers. This underwater site is situated off the island of the Veilles, between the natural harbour of Agay and the current port of the French city of Fréjus. Several shipwrecks were in fact discovered at the foot of the La Chrétienne reef. The wreck number M2 is the one that interests us here (Joncheray and Joncheray 2002: 57–62). This shipwreck offered an assemblage made up of several dozens of amphorae type Dressel 1, of which the majority, if not the whole, were of Iberian origin, and T-7.4.3.3 amphorae from the Strait of Gibraltar. This type latter container is less numerous than the Dressel 1 amphorae, but still forms a significant part of this ancient cargo. Besides the amphorae, it is necessary to indicate the presence of numerous epigraphic marks, written in three different languages: Latin, Greek and neo-Punic (Luaces 2017: 403–405). The archaeological material of the ‘Chrétienne M2’ shipwreck is particularly rich, but it is the various examples of T-7.4.3.3 that interest us here. It turns out that the majority of these artifacts were marked by large traces of pitch on the internal surface (Figure 4). I will not discuss here the problems associated with the pitch coating on the interior of amphorae, but will follow the school of thought here that there was a relationship between pitch and fish contents, and also wine products (Garnier, Silvino and Bernal‑Casasola 2011). Pitch, on the other hand, was not generally used

405

Max Luaces

Figure 4. Pictures of two fragments from two different T-7.4.3.3 amphorae, with the interior coated with pitch, discovered among the cargo of the Chrétienne M2 shipwreck. for amphorae transporting oil products. The almost systematic presence of pitch for the T-7.4.3.3 amphora of this shipwreck strengthens the hypothesis that they were used for fish products, and perhaps also wine. Nonetheless, we need to take into account the fact that the Dressel 1A amphorae in this cargo, a well-known type for wine products, which in this case originated from the same region as the T-7.4.3.3 amphora, were not coated with pitch. This difference in the preparation of the amphorae suggests the presence of different contents, and different amphora types. On the other hand, if this ‘Chrétienne M2’ shipwreck transported wine from the Ulterior province in these Dressel 1, it is likely that the amphorae T-7.4.3.3 was used for another contents. As the pitch coating of ancient maritime containers excluded oil products, it is all the more tempting to suggest that fish products were the contents. 4. Conclusion The documentation that was reviewed here could be seen as rather contradictory, as illustrated by the diverse sources around the study of the contents of the T-7.4.3.3 amphora. However, we have observed several documents that could on the contrary lift the veil on some of the ambiguities which still mark these containers. If we traditionally envisaged the T-7.4.3.3 class as a container used mainly for garum, the question of the contents now needs to be reconsidered. On one hand, the clear distinction between this type from the Circle of the Strait and the Dressel 18 type suggests that we must differentiate their contents: the use of T-7.4.3.3 exclusively for the transportation of garum should now be questioned. On the other hand, there were several variants of the T-7.4.3.3 amphora, which were made in multiple autonomous production centres, over a very long period. This may have led to a multitude of contents. Nevertheless, it would seem that the main commodity transported by the T-7.4.3.2 type — the parent of the T-7.4.3.3 — was fish products. This interpretation does not exclude the transport of garum, but it would then be a question of associating this content with the versions of the T-7.4.3.3 amphorae dated to the Early Empire. For Late Republican era, the evidence from the Roman city of

Baelo Claudia suggests that this container was used for salsamenta. Such content turns out to be more consistent with the Phoenician and Punic sociocultural environment of the Circle of the Strait, than garum. This report does not fully prevent either the use of these amphorae as a wine container, or even their use for winemaking. But such content should rather be seen as a secondary use, in connection with domestic activities or with that of a specific local market. The different archaeological data also suggest the existence of a relationship between the content and the commercial market for which these amphorae were intended. The possible transportation of malacofauna in these amphorae, for example, would only have been for a restricted distribution. Other data from Southern Gaul consolidates an interpretation which would see the preferential use of the T-7.4.3.3 type as a container for fish products, at least within the framework of the great trade routes of the Mediterranean. We should nevertheless take into account the diversity and long lasting use of our amphora. In the end, more than a binary association between the shape of this container and a specific content, we should add a further level of differentiation, in connection with the different periods during which the T-7.4.3.3 was active (three phases, between 150 BC and AD 50, in connection with three kind of different contents?), and take into account the deep sociocultural and economic transformations in the area of the Circle of the Strait in this timeline, between the end of the second Punic War and Claudius’s reign. Bibliography Aranegui Gascó, C. 2004. Lixus (Larache, Marruecos). Bienes culturales: revista del Instituto del Patrimonio Histórico Español 3: 171–182. Ben Jerbania, I. 2013. Observations sur les amphores de tradition punique d’après une nouvelle découverte près de Tunis. Antiquités africaines 49: 179–192. Bernal-Casasola, D. 2016. Le Cercle du Détroit, une région géohistorique sur la longue durée. Karthago 2: 7–50. Bernal-Casasola, D., Arévalo González, A., Morales Muñiz, A. and Roselló Izquierdo, E. 2007. Un ejemplo de conservas de pescado baelonenses en el siglo II a. C. In A. Arévalo González and D. Bernal-Casasola (eds), Las cetariae de Baelo Claudia: avance de las investigaciones arqueológicas en el barrio meridional (2000-2004): 355–374. Cadiz, Universidad de Cádiz. Blanco Jiménez, F. J. 1991. Excavaciones de urgencia en un solar de la calle Gregorio Marañón, Cádiz. Anuario arqueológico de Andalucía 1989-III, Actividades ele Urgencia. Informes y Memorias: 78–81. Bonet Rosado, H., Fumadó Ortega, I., Aranegui Gascó, C., Hassini, H. and Vives-Ferrándiz Sánchez, J. 2005. La ocupación mauritana. In C. Aranegui Gascó, (ed.), Lixus-2 Ladera Sur. Excavaciones arqueológicas marocoespañolas en la colonia fenicia, Campañas 2000-2003. Sagvntvm Extra 6: 87–140. Valence, Universitat de Valencia - Institut National des Sciences de l’Archéologie et du Patrimoine.

406

The contents of amphora type T-7.4.3.3 (former type Mañá C2b): ancient problematic…

Botte, E. 2009. Salaisons et sauces de poissons en Italie du sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Boube, J. 1975. Marques d’amphores découvertes à Sala, Volubilis et Banasa. Bulletin d’Archéologie Marocaine 9: 169–236. Callegarin, L. 2016. L’efficience d’un paradigme d’antiquistes. Karthago 29: 51–72. Capelli, C., Contino, A., Djaoui, D. and Rizzo, G. 2017. Anfore neo-puniche del I secolo d.C. di Arles, Ostia e Roma: classificazione tipo-petrografica, origine e diffusione. The Journal of Fasti Online. . Cintas, P. 1950. Céramique punique. Publications de l’Institut des Hautes Études de Tunis III. Paris, C. Klincksieck. Dressel, H. 1899. Corpus Inscriptionum Latinarum XV: Inscriptiones urbis Romae Latinae. Instrumentum domesticum. Berlin, George Reimer. García Vargas, E. 1998. La producción de ánforas en la bahía de Cádiz en época romana: (siglos II A.C-IV D.C.). Ecija, Gráficas Sol. García Vargas, E. and Ferrer Albelda, E. 2001. Producción y comercio de salazones y salsas saladas de pescado de la costa malagueña en épocas púnicas y romana republicana. In F. Wulff Alonso, G. Cruz Andreotti and C. Martínez Maza (eds), Comercio y comerciantes en la historia antigua de Málaga, siglo VIII a.C. - año 711 d.C. Actas del II Congreso de Historia Antigua de Málaga 1998: 547–572. Malaga, Centro de Ediciones de la Diputación de Málaga. García Vargas, E., Bernal-Casasola, D., Palacios Macías, V., Roldán Gómez, A. M., Rodríguez Alcántara, A. and Sánchez García, J. 2014. Confectio Gari Pompeiani. Procedimiento experimental para la elaboración de salsas de pescado romanas. SPAL, Revista de Prehistoria y Arqueología de la Universidad de Sevilla 23: 65–82. Garnier, N., Silvino, T. and Bernal-Casasola, D. 2011. L’identification du contenu des amphores : huile, conserves de poissons et poissage. In SFECAG, Actes du Congrès d’Arles: 397–416. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Guerrero Ayuso, V. M. 1986. Una aportación al estudio de las ánforas púnicas Mañá C. Archaeonautica 6: 147–186. Joncheray, A. and Joncheray, J.-P. 2002. Chrétienne M, trois épaves distinctes, entre le cinquième siècle avant et le premier siècle après Jésus-Christ. Cahiers d’archéologie subaquatique XIV: 57–130.

Lagóstena Barrios, L. 1996. Explotación del salazón en la Bahía de Cádiz en la Antigüedad. Aportación al conocimiento de su evolución a través de la producción de las ánforas Mañá C. Florentia Iliberritana 7: 141–169. Lagóstena Barrios, L. 2001. La producción de salsas y conservas de pescado en la Hispania Romana II a.C.-VI d.C. Instrumenta 11. Barcelona, Universitat de Barcelona. Luaces, M. 2015. La relation entre le temps et la rationalité économique dans les contextes archéologiques de Gadir/Gadès (VIe-Ier s. av. J.-C.). Pallas 99: 245–265. Luaces, M. 2017. Production et diffusion des amphores tardo-puniques en Méditerranée occidentale. L’apport des contextes de la Gaule méridionale. Unpublished PhD thesis, University of Cadiz and University of Lyon 2. Mañá de Angulo, J. M. 1951. Sobre tipología de las ánforas púnicas. In M. Beltrán Lloris (ed.) Crónica del VI Congreso Arqueológico del Sudeste Español (Alcoy, 1950): 203–210. Cartagena, Papelería Española. Morel, J. P. 1965. Les niveaux préromains. In J. P. Callu, J. P. Morel, R. Rebuffat and G. Hallier (eds), Thamusida 1. Fouilles du service des antiquités du Maroc. Mélanges d’archéologie et d’histoire, Suppléments 2: 61–112. Rome, École française de Rome. Perdigones Moreno, L. and Muñoz Vicente, A. 1990. Excavaciones arqueológicas de urgencia en los hornos de Torre Alta. San Fernando, Cádiz. Anuario arqueológico de Andalucía 1988.III, Actividades de Urgencia. Informes y Memorias: 106–112. Ramon Torres, J. 1981. Ibiza y la circulación de ánforas fenicias y púnicas en el Mediterráneo occidental. Trabajos del Museo Arqueológico de Ibiza 5. Ibiza, Museo Arqueológico de Ibiza. Ramon Torres, J. 1995. Las ánforas fenicio-púnicas del mediterráneo central y occidental. Instrumenta 2. Barcelona, Universitat de Barcelona. Sáez Romero, A. M. 2008. La producción cerámica en Gadir en época tardopúnica (siglos -III/-I). British Archaeological Reports International Series 1812. Oxford, Archaeopress. Van der Werff, J. H. 1978. Amphores de tradition punique à Uzita. Bulletin Antieke Beschaving (Babesch) 52-53: 171–200. Vuillemot, G. 1965. Reconnaissances aux échelles puniques d’Oranie. Autun, Musée Rolin.

407

The fish-salting production centre of Águilas: late-Roman amphora content analysis Alejandro Quevedo

University of Murcia, associate collaborator of the Centre Camille Jullian, Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France

Myriam Sternberg

Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France

Juan de Dios Hernández García Águilas Municipal Archaeological Museum

Abstract: Águilas (Hispania Tarraconensis) was a major salted fish production centre in Late antiquity. Recent archaeological excavations have allowed two production phases to be distinguished: a first phase covering a small area and dated to between the second half of the 3rd century AD and the first half of the 4th century AD, and a second phase that spread over a wider area between the second half of the 4th century and the 5th century AD. Numerous samples have been collected in the archaeological excavations carried out since the 1980s. Their study provides us with an understanding of the different forms of production and their evolution at Águilas. Two content analyses are discussed: one from an Africana II D amphora considered to have been reused from the first phase and one from a local spatheion dating from the second phase. Key words: Late roman pottery; Amphora workshops; spatheia; salsamenta; trade; North Africa.

1. Introduction The coastal town of Águilas (Murcia) is in the south-east of the Iberian Peninsula, near where the provinces of Baetica and Hispania Tarraconensis meet (Figure 1.1). To date we have not been able to ascertain its vicus or its name, although some scholars have linked it to Urci (Hernández García 2010: 268). Late antiquity traces of salted fish production have been found in the area, including salting tanks, amphora workshops and fill layers mixed with rubbish and fish bones (Figure 1.2). This phenomenon can be found in other places on the eastern coast of the provincia Karthaginenis, where cetariae and figlinae linked to fish salting and packing are documented (Lagóstena 2001: 169–175, Quevedo 2021). Of particular note is the Mazarrón inlet, where there are various factories, including the Insignia building, El Alamillo and La Azohía, as well as the large El Mojón pottery workshop (Ramallo 2006: 51–53, 134–149). To this list we can probably also add Carthago Nova itself, although archaeological finds there are very few and far between (Quevedo 2015: 49–50). Although traditionally the amphorae manufactured on the region’s coast are associated with pisciculture products (Berrocal 2012), until now no content analyses had been carried out on them. Only the remains of Atlantic saury (Scomberesox saurus) had been identified in a tank in Puerto de Mazarrón (Ramallo 1985: 441). 2. The production of fish preserves and sauces in Águilas during late antiquity The development of Águilas as an urban enclave began towards the end of the Republican period with the

exploitation of the mines in the area. From the Augustan period, the locality underwent a notable expansion, eventually covering between 7 and 8 hectares. During the 1st century AD, various domus were built, as well as artisanal areas and open spaces, possibly of a public nature. During the Flavian era, the so-called Western Baths were built, followed in the second third of the 2nd century AD by the Eastern Baths. The port was especially active in trade with North Africa and covered a wide territory that extended as far as the town of Eliocroca (Lorca). However, in the first half of the 3rd century AD, the urban and commercial dynamic underwent a profound change and from then on it was re-orientated towards the exploitation of the resources of the sea. The fish-salting industry in Águilas and its hinterland developed between the 3rd and the 5th centuries AD in two well-differentiated phases (Hernández García 2010: 274–276, with bibliography). The first stage, which began in the second half of the 3rd century AD, bears the features of small-scale production. The structures at this time were not very large and there was only one basin. Good examples of this are the finds at 4 Conde de Aranda Street and 1 San Juan Street (Figure 2.1-2), where numerous fish remains were also found, as well as items related to fishing. We do not know of any local amphorae from this period and the evidence from finds in both areas suggests that imported vessels may have been reused (Hernández García 2002: 353–354). This is the case of different Africana II type amphorae

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 409–417

Alejandro Quevedo, Myriam Sternberg and Juan de Dios Hernández García

Figure 1. 1. The location of Águilas in the south-east of the Iberian Peninsula and the main enclaves cited in the text [A. Quevedo, own preparation]; 2. Installations linked to the production of salted fish in Águilas during late antiquity [background map courtesy of J. D. Hernández García]. manufactured in what is today Tunisia, that would have been filled with local products (Figure 2.3). The second stage took place between the 4th and 5th centuries AD, a time at which the production of salted fish reached its floruit, becoming the main economic driving force of Águilas. Archaeology has documented numerous dumps with the remains of ichthyofauna from this time in the urban area, which in some cases appear to have been in use until the 6th century AD (Figure 1.2). In contrast to the previous phase, the new factories were larger and made up of different tanks, as illustrated by the Cassola Street-Paseo de la Constitución archaeological site (Figure 2.4). The development of the salting industry extended to the hinterland of Águilas, with the best example being the vast complex on the nearby Isla del Fraile, as confirmed by an excavation carried out in September 2020 (Quevedo and Hernández García 2020).

In this second phase, next to the cetariae there are numerous kilns for manufacturing amphorae. A characteristic feature of the archaeology of Águilas is that some of these pottery workshops were set up in the former bath complexes, when they lost their original function (Hernández García and Pujante 2006; Ramallo 1983-84: 105–115). Kilns were built in both the Western Baths (Figure 2.5-6) and the Eastern Baths (Figure 2.7-8) and some of the structures were reused, including a natatio in the latter as a kind of pond with an adjoining basin to decant the clay. Excavations carried out in recent years have revealed further kilns that illustrate the expansion of the late-period pottery workshops in Águilas, such as on the Renfe plot (Hernández García 2004) or those at 8 Isabel la Católica Street and 5 Muñoz Calero Street (unpublished). Five main types of local amphora were produced between the 4th and the 5th centuries AD: Águilas Types I, II, III,

410

The fish-salting production centre of Águilas: late-Roman amphora content analysis

Figure 2. 1-4. The cetariae of Águilas [figures courtesy of J. D. Hernández García]: 1. Conde Aranda Street, 4 (3rd-4th century); 2. San Juan Street, 1 (3rd-4th centuries); 3. Africana II D amphora with fish remains, Conde Aranda Street; 4. Calle Cassola-Paseo de la Constitución (4th-5th centuries). 5-8. Figlinae of Águilas: 5-6. Western Baths, different phases of the excavation of the kiln with spatheia [figures courtesy of S. F. Ramallo Asensio]; 7-8. Eastern Baths, remains of amphorae produced in situ and of resin for waterproofing them [figures courtesy of J. D. Hernández García].

411

Alejandro Quevedo, Myriam Sternberg and Juan de Dios Hernández García

IV and V (Hernández García and Pujante 2006: 401–404, fig. 10-14). Águilas Type I is the most common: it is a fine spatheion that is found repeatedly both in the spoil heaps and the kilns of the baths, several of which collapsed with their final loads inside (Figure 2.5-8). Some of the types described have an African influence, including the Águilas IIA (similar to a Keay XXV) and Águilas V types (Keay LXXIXB). To these vessels we can add others that complete the repertory. They are clearly imitations of African containers, such as types Keay IA, IB, XXV and the Africana II C shapes from the Renfe site pottery workshop, as well as an amphora found on the Isla del Fraile (and perhaps made there? Quevedo and Hernández García 2020).

(Figure 3.2-7). These three gregarious species live near the coast and are quite numerous in the seas around Águilas. Fish with an average size of 12cm were piled up without being scaled or boned, although they may have been eviscerated. The good preservation of otoliths (Figure 3.3) suggests that the fish underwent little corrosive treatment by drying and/or salting, which could mean that the product was salted fish (salsamenta), rather than sauce (garum). These species are not a very popular for eating today, as was already the case with the blotched picarel in antiquity.

The excavations carried out since the 1980s have allowed the recovery of fish remains from the interiors of numerous amphorae. In this paper we analyse for the first time the content of two of them, an Africana II D amphora and a local spatheion belonging to each of the identified production phases. The objective is to define their content precisely and understand the evolution of the salted fish production system in Águilas between the 3rd and the 5th centuries AD.

The second amphora analysed was a spatheion of the type manufactured in abundance in Águilas during late antiquity. It is an Águilas I Type, a small-capacity vessel with an elongated cylindrical body up to 70cm high (Figure 4). The neck is long without a differentiated rim and it has a quadrangular-section handle and a solid apex (Ramallo 1985: 436, fig. 2).

3. Africana II D Amphora. Content Analysis The first container studied comes from the excavation at 4 Conde de Aranda Street, a modest salted fish production facility built over an early-Roman house (Figure 2.1). The excavation of the cetaria revealed numerous dumps of fish remains, items related to fishing, such as hooks and net weights, and various amphorae with the remains of their contents. The complex was developed between the mid-3rd century AD — as attested by the abundant Africana C terra sigillata and a sesterce of Maximinus Thrax (AD 235‑238) — and the mid-4th century AD (Hernández García 1999: 263–264). The analysed amphora is of the large Africana type, a group redefined by Clementina Panella (1973), whose classification is still valid (Bonifay 2004: 115–117). It is an Africana II D shape or perhaps an early variation of that type1 (Figure 2.3, 3.1). It was manufactured in the province of Byzacena, in potteries of the Tunisian Sahel such as Leptiminus, Salakta and Thaenae (Bonifay 2016: 600, fig. 5.18). The amphora is characterised by a seal SAT impressed below the rim, probably with a metal punch, and the red titulus N. The seal has been documented in two examples of Africana II A amphoras: one identical example found at Ostia (Palma and Panella 1968: pl. XLVI, 591) and another from the Nuovo Mercato Testaccio in Rome (Coletti 2013: 313, fig. 16). The content of the amphora consisted largely of small sparidae: bogue (Boops boops) and salema (Sarpa salpa) mixed with blotched picarel (Spicara maena) (1/8e),  Although the distinct carination of the body meant that it was initially published as an Africana II B. Our sincere gratitude to Michel Bonifay and Víctor Revilla for their comments on the typology of the amphora. 1

4. Local Águilas I Type spatheion. Content analysis

The archaeometric analysis carried out on samples of the Águilas I type from the Eastern Baths revealed the use of clay from two different outcrops (Arana 1985). The fabric of the spatheia is either porous and a creamy-yellowywhite colour (Figure 4.1) or reddish and more compact (Figure 4.2). It is sometimes possible to see the line of the seam between the upper and lower parts, which would have been thrown separately (Figure 4.2). The examples with a clear fabric were fired at temperatures of between 700 and 800°C, while those with the darker fabric were fired at between 800 and 900°C. In general, these vessels have a rough, porous appearance, and on their surface it is possible to see schist and quartz particles, as well as other inclusions. The example analysed here also comes from the Eastern Baths (Figure 5.1). The content mainly consisted of pilchards (Sardina pilchardus), mixed with blotched picarel (Spicara maena) and/or bogue (Boops boops) (1/6e) (Figure 5.2-7). The pilchards, which were probably caught in the vicinity, had an average size of 15cm. Easily preserved, pilchards were often used for by-products and in the spatheion they were stored whole, having been neither scaled nor eviscerated, like the mendoles and/or bogues. This suggests that the product in the spatheion could also be salted fish (salsamenta), as in the first amphora. 5. Discussion The amphorae used for products deriving from fishing could have very varied contents: small and medium-sized salted fish, sauces — garum, hallec, liquamen or muria — or unfinished versions of the aforementioned products, such as unfiltered garum (Bernal‑Casasola 2015: 64–66). The analyses show that the two receptacles from Águilas were probably transporting salted fish (salsamenta) and not

412

The fish-salting production centre of Águilas: late-Roman amphora content analysis

Figure 3. 1. Africana II D amphora (or a possible variation) found in the Conde Aranda Street factory [drawing by A. Quevedo]; 2. Fish species that made up the content of the amphora; 3-7. Bone remains identified in its interior [figures courtesy of C. Durand].

413

Alejandro Quevedo, Myriam Sternberg and Juan de Dios Hernández García

Figure 4. Local Águilas I spatheia and detail of the two possible tones of the fabric: yellowish or reddish-orange, Eastern Baths [drawing by A. Quevedo, photos courtesy of the Águilas Municipal Archaeological Museum]. garum. This is especially interesting for the Africana II D amphora, a type whose content — salsamenta or wine — is still the subject of speculation (Bonifay 2016: 596, fig. 2). On a macroscopic level, neither the Africana II D amphora nor the Águilas I spatheion contained any remnants of waterproofing resin, although research has shown that this could be present, even if it cannot be detected at first (Garnier, Silvino and Bernal-Casasola 2011).

The containers were not used to store exactly the same species. The content of the African amphora consisted exclusively of small coastal fish, while the spatheion was mainly filled with small pelagic migrant fish. Were these differences due to the availability of the resources, to the specialisation of containers or to a change in consumer tastes over time? In the case of the spatheion, the narrow mouth leads us to believe that the product was in a semi-liquid state.

414

The fish-salting production centre of Águilas: late-Roman amphora content analysis

Figure 5. 1. Local Águilas I spatheion, Eastern Baths [drawing by A. Quevedo, photo courtesy of J. D. Hernández García]; 2. Fish species that made up the content of the amphora; 3-7. Bone remains identified in its interior [figures courtesy of C. Durand].

415

Alejandro Quevedo, Myriam Sternberg and Juan de Dios Hernández García

It is even more complicated to work out whether the amphora was reused (Bernal‑Casasola, 2015: 76–78), as has been suggested for the Africana II D and other African amphorae found in the Cassola Street cetaria (Hernández García 2002: 354). Was it a wine amphora reused for local salted fish? The design of the archetype suggests that the morphology of the container was especially suitable for a particular product. In that case, is it possible to ascertain whether an amphora with African salsamenta had been refilled with Hispanic salsamenta? Beyond the study of new samples, it appears that the most pertinent method for getting to the bottom of this question is DNA analysis, especially of the non-migratory fish species that live near the coast. However, this line of study faces some challenges. On the one hand, the different species do not always have subspecies in which the markers of different populations can be distinguished. On the other, the fact that they are fish with little current commercial value, such as the bogue (Boops boops), the salema (Sarpa salpa) and the blotched picarel (Spicara maena) analysed here, means that the research is less developed. It is somewhat less surprising to hypothesise that Águilas, which is in an area known for its rich fishing and fish-related products, could have imported African amphorae of salted fish in an initial production phase. However, despite Pliny, Strabo and other ancient writers reporting on the quality of the garum from Carthago Nova and its surrounding area, we know very little about the cetariae and the local amphorae from the early-Roman period2 (Quevedo 2021). This tendency is reversed in the late period, with the figlinae and cetariae in the Mazarrón and Águilas inlets. Until now, researchers had paid more attention to the vessels from the Mazarrón area, while the Águilas amphoras were defined as ‘a production of a reduced typology’ (Berrocal 2012: 268). However, the latest finds oblige us to put this perception into context and lead us to reconsider the large variety of containers for local transport. Although archaeometric differences between the fabrics from each locality have been established (with that of Águilas generally the coarsest, Arana 1985: 447), there is an imperative need to carry out new analyses. The fact that Águilas and Mazarrón are in the same geological environment means that it is not always possible to differentiate the fabrics on a macroscopic level and that they may contain common

 Their comprehensive study is currently underway as part of the ‘New methodologies for the study of the economy of the south-east in the Roman period. Amphoras: typological-archaeometric characterisation and organic content analysis’ project directed by Alejandro Quevedo and financed by the Séneca Foundation of the Science and Technology Agency of the Region of Murcia. 2

elements such as silver mica. Moreover, the pottery workshops in both zones were operating at the same time, often producing identical types. A final question refers to the volume of production and its scope. The spatheia had a smaller capacity. Were they designed for a more select product or as space-fillers to place in the gaps in the ships’ cargo holds? However, we have to take into account the manufacture of types with larger volumes, the abundance of kilns and, above all, the considerable size of some of the late-period cetariae (such as that of the Insignia building in Puerto de Mazarrón). All this leads us to believe that the distribution of the salted fish produced on the eastern coast of the Karthaginenis, which until now had been documented on a regional level (Berrocal 2012: 272–273), would have had a much greater commercial impact, a hypothesis on which we are currently working. 6. Conclusion Together with Mazarrón and probably Carthago Nova, Águilas was one of the most important salted fish production centres in south-eastern Hispania. The analyses undertaken have allowed us to characterise, for the first time, the content of some of the amphorae found in the cetariae of this port enclave. The study of the ichthyofauna has demonstrated that they produced and packed salsamenta, although the different species — bogue, salema, blotched picarel — and the proportions in which they were used varied over time. In a first phase (3rd-4th centuries AD) it is possible that the product was packed in reused imported amphorae, such as the Africana II. We are keeping this working hypothesis open, but it is surprising that, at a time of small-scale production, they were able to fill amphorae as large as those from Africa. In the later period (4th-5th centuries AD), the Águilas I spatheia — among other types produced locally — were clearly destined for trade in a paste or semi-liquid product made largely of sardines. Many questions remain concerning the volume of production, exports and the evolution of the halieutic industry of Águilas and the eastern coast of Cartagena during late antiquity. The issues raised by these preliminary analyses should begin to find answers in future analyses of new samples.3

 This study will be conducted within the framework of a collaboration between the University of Murcia, the Águilas Archaeological Museum and the UMR 7299 of the National Centre for Scientific Research (CNRS, France). A contribution made within the framework of Alejandro Quevedo’s project: ‘Hispania / Africa: modelos urbanos y relaciones de interdependencia entre las dos orillas del Mediterráneo occidental (s. I-V d.C.)’, financed by the University of Murcia. 3

416

The fish-salting production centre of Águilas: late-Roman amphora content analysis

Bibliography Arana Castillo, R. 1985. Envases para salazón en el bajo Imperio (II): estudio mineralógico de las cerámicas romanas de Águilas y Mazarrón (Murcia). In VI Congreso Internacional de Arqueología Submarina, Cartagena 1982: 443–450. Madrid, Ministerio de Cultura. Bernal-Casasola, D. 2015. What contents do we characterise in Roman amphorae? Methodological and archaeological thoughts on a trending topic. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 61–83. Esposende, Município de Esposende. Berrocal Caparrós, M. C. 2012. Producciones anfóricas en la costa meridional de Carthago-Spartaria. In D. Bernal-Casasola and A. Ribera (eds), Cerámicas hispanorromanas II. Producciones regionales: 255–277. Cadiz, Universidad de Cadíz. Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports International Series 1301. Oxford, Archaeopress. Bonifay, M. 2016. Amphores de l’Afrique romaine : nouvelles avancées sur la production, la typochronologie et le contenu. In R. Járrega and P. Berni (eds), Amphorae ex Hispania: paisajes de producción y consumo. III Congreso Internacional de la Sociedad de Estudios de la Cerámica Antigua (SECAH) - Ex Officina Hispana (Tarragona, 10-13 de diciembre de 2014). Monografías Ex Officina Hispana III: 595–611. Tarragona, Institut Català d’Arqueologia Clàssica. Coletti, F. 2013. Nuove acquisizioni sull’epigrafia anforaria africana. Contesti romani a confonto di età medio e tardo imperiale, In D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz and A. M. Sáez (eds), Hornos, talleres y focos de producciόn alfarera en Hispania. I Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I: 219–316. Cadiz, Universidad de Cádiz. Garnier, N., Silvino, T. and Bernal-Casasola, D. 2011. L’identification du contenu des amphores : huile, conserves de poissons et poissage. In SFECAG, Actes du Congrès d’Arles: 397–416. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Hernández García, J. D. 1999. Actuaciones arqueológicas de urgencia en Águilas. Memorias de Arqueología de la Región de Murcia 8 (1993): 257–290. Hernández García, J. D. 2002. La factoría de salazones de la calle Cassola - Paseo de la Constitución (Águilas). Memorias de Arqueología de la Región de Murcia 11 (1996): 339–358.

Hernández García, J. D. 2004. Intervención arqueológica en la P4 de la Unidad de Ejecución E7, Terrenos de Renfe (Águilas). Alfar tardorromano. In XV Jornadas de Patrimonio Histórico y Arqueología de la Región de Murcia (Murcia del 24 de noviembre al 2 de diciembre de 2004): 67–69. Murcia, Dirección General de Cultura, Servicio de Patrimonio Histórico. Hernández García, J. D. 2010. Evolución del asentamiento romano de Águilas y la relación con su hinterland. In J. M. Noguera (ed.), Poblamiento rural romano en el Sureste de Hispania. 15 Años después: 255–283. Murcia, Universidad de Murcia. Hernández García, J. D. and Pujante Martínez, A. 2006. Termas orientales altoimperiales y centro alfarero tardorromano. Excavación en Calle Juan Pablo I esquina con Calle Castelar (Águilas), Memorias de Arqueología de la Región de Murcia 14 (1999): 387–408. Lagóstena Barrios, L. 2001. La producción de salsas y conservas de pescado en la Hispania Romana II a.C.-VI d.C. Instrumenta 11. Barcelona, Universitat de Barcelona. Palma B. and Panella C. 1968. Anfore. In Carandini, A. (ed.), Ostia I. Le terme del Nuotatore. Scavo dell’ambiente IV. Studi Miscelanei 13: 97–116. Rome, De Luca. Panella, C. 1973. Le anfore. 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 Miscellanei 21: 463–633 Rome, De Luca. Quevedo, A. 2015. Contextos cerámicos y transformaciones urbanas en Carthago Nova (s. II-III d.C.). Roman and Late Antique Mediterranean Pottery 7. Oxford, Archaeopress. Quevedo, A. 2021. La producción anfórica de Carthago Nova y la costa del Sureste de la península Ibérica: estado de la cuestión. SPAL, Revista de Prehistoria y Arqueología de la Universidad de Sevilla 30.1: 196–221. Quevedo, A. and Hernández García, J. D. 2020. Arqueología de la Hispania tardoantigua: un nuevo proyecto de investigación en la isla del Fraile (Águilas). Sagvntvm 52: 135–152. Ramallo Asensio, S. F. 1983-1984. Algunas consideraciones sobre el Bajo Imperio en el litoral murciano: los hallazgos romanos de Águilas. Anales de la Universidad de Murcia, Letras 42 (3-4): 97–124. Ramallo Asensio, S. F. 1985. Envases para salazón en el bajo Imperio (I). In VI Congreso Internacional de Arqueología Submarina, Cartagena 1982: 435–442. Madrid, Ministerio de Cultura. Ramallo Asensio, S. F. 2006. Mazarrón en el contexto de la romanización del Sureste de la península Ibérica. In S. F. Ramallo Asensio, S. Agüera Martínez and J. García Sandoval (eds), Carlantum. III Jornadas de estudio sobre Mazarrón: 11–164. Mazarron, Universidad Popular de Mazarrón.

417

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts: salsamenta, garum, lymphatum and other fish products Gaël Piquès

ASM - Archéologie des Sociétés Méditerranéennes, UMR 5140, Univ Paul-Valéry, CNRS, Ministry of Culture, INRAP and Labex Archimède (IA-ANR-11-LABX-0032-01), Montpellier, France

Núria Rovira

ASM - Archéologie des Sociétés Méditerranéennes, UMR 5140, Univ Paul-Valéry, CNRS, Ministry of Culture, INRAP and Labex Archimède (IA-ANR-11-LABX-0032-01), Montpellier, France

Margaux Tillier

ASM - Archéologie des Sociétés Méditerranéennes, UMR 5140, Univ Paul-Valéry, CNRS, Ministry of Culture, INRAP and Labex Archimède (IA-ANR-11-LABX-0032-01) Montpellier, France ISEM - Institut des Sciences de l’Evolution Montpellier, UMR 5554, Univ Montpellier, CNRS, IRD, EPHE, CIRAD, INRAP, Montpellier, France

Franca Cibecchini

DRASSM Underwater Archaeology Research Department and Centre Camille Jullian, Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France

David Djaoui

Musée départemental Arles antique and Centre Camille Jullian, Aix Marseille Univ, CNRS, CCJ, Aix-en-Provence, France

Carlos De Juan

Universitat de València

Abstract: Salted-fish and fish sauces, as well as wine and oil, are a key research subject because of their significant place not only in the diet but also in the many economic and socio-cultural spheres of Roman society. For a long time, identification has been based on painted inscriptions (tituli picti) and their association with the specific form of the containers, but new multidisciplinary approaches have been developing over the last twenty years. Following this trend, this contribution aims to show, from a methodological point of view, the potential and limits of archaeozoological and archaeobotanical analyses, applied to underwater contexts, in the study of paleo-contents concerning both fish and plant remains. In addition, this article presents new data from two archaeological sites (Arles-Rhône 3 and Bou Ferrer) that enable us to reveal the recurrent presence of these two types of remains together, which could indicate some salted-fish products mentioned in ancient literary sources, in particular related to garum or perhaps lymphatum. Key words: paleo-contents; salted-fish and sauces; wine; underwater excavations; archaeozoologie; archaeobotany; Early Roman Empire.

1. Introduction Since the articles of P. Grimal and T. Monod (1952) on ‘La véritable nature du garum’, C. Jardin (1961) on ‘Garum et sauces dans l’antiquité’ and M. Ponsich and M. Tarradell (1965) on ‘Garum et industries antiques de salaison dans la Méditerrannée Occidentale’, historians and archaeologists have continued to study the nature of salted-fish (salsamenta) and other fish products mentioned in ancient texts such as garum, liquamen, allec, muria, as well as other more enigmatic products revealed by the

tituli picti such as ‘laccatum’1 and lymphatum (Curtis 1978 and 1991; Lagóstena Barrios 2001; Étienne and Mayet 2002; Trakadas 2005; Lagóstena Barrios, Bernal-Casasola and Arévalo 2007; Botte 2009; Martin-Kilcher 2011; Grainger 2014; García Vargas et al. 2018).

 In the case of ‘laccatum’, considered until now as an enigmatic fish sauce, the resumption of the epigraphic studies has proved that it was in fact fish-salting of juvenile mackerels: LACCATV(m) has to be developed into LAC(ertus) CAT(ulus) V(etus) (Djaoui 2016). 1

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 419–436

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

Roman Empire which was conducted as part of a project, funded by the LabEx Archimède (ASM UMR5140), that seeks to contribute to our knowledge of salted-fish and sauces from Roman times and their containers. The material studied, consisting of fish bones and plant macro-remains, was recovered from ceramics found in two different underwater contexts, which allowed for the preservation of waterlogged organic remains. Both of these contexts dated to the Flavian period. The first example comes from ceramics found at a freshwater context at the site of a garbage dump associated with the port of Arles (France) on the Rhône River, which was excavated as part of the Arles-Rhône 3 archaeological project (Djaoui, Greck and Marlier 2011) (Figure 1). The second example involves Dressel 7/11 amphorae from the salt-water wreck of Bou Ferrer (La Vila Joiosa, Spain), which had set out from Cadiz (Cibecchini et al. 2014).

Figure 1. Location of the archaeological sites: the shipwreck and garbage dump of Arles-Rhône 3 (Arles, France) and the wreck of Bou Ferrer, situated off La Vila Joiosa (Alicante, Spain) (Map: H. Bohbot, CNRS). Our understanding of the amphorae used for the transport of these products is largely based on painted inscriptions (tituli picti) and their association with the specific form of the container (Martin-Kilcher 2011). The inscriptions show that fish products were quite numerous, that they could have been of different degrees of quality, could have been refined for several years, were sometimes flavoured with spices, and were often made from tuna, young tuna and mackerel. Archaeozoological studies conducted on fish remains contained in amphorae, in particular Dressel 7/11, and on remains from fish-salting factories, which have developed over the past twenty years, provide us with information on the nature of these products as well as on the relationship between the container and its contents (Desse-Berset and Desse 2000; García Vargas et al. 2018). They are also a means of analysing, as a complement to literary sources, what the specialties of the different production areas could have been. However, studies on the paleo-contents of amphorae are not (or should not be) limited to fish remains. Indeed, salted-fish products could have also involved aromatics, as well as other foodstuffs of plant and/or animal origin. Hence the interest, when possible, in multiproxy approaches involving in particular archaeobotanical and chemical analyses, the latter in an attempt to detect components that do not leave macroscopic traces (Garnier 2014; Bernal-Casasola 2015; Pecci et al. 2018). The aim of this contribution is therefore to show, from a methodological point of view, the potential and limits of archaeozoological and archaeobotanical analyses, applied to underwater contexts, in the study of paleocontents concerning both fish and plant remains. In addition, this article presents new data that enable us to reveal the recurrent presence of these two types of remains together, which could indicate some salted-fish products mentioned in ancient literary sources. We thus present the results of a preliminary study on several contents of jars and amphorae from the Early

2. The underwater excavations of the Rhone at Arles (France) 2.1. Materials and methods The paleo-contents from Arles come mainly from the excavation of a vast port dump that covered the wreck of the Roman barge designated as Arles-Rhône 3 (Djaoui, Greck and Marlier 2011). This dumping ground was formed between the sinking of the ship around AD 50‑60 and a terminus ante quem around AD 130‑140. Taphonomic conditions are exceptional here in this anaerobic environment, allowing for the preservation of organic and perishable materials, including numerous seeds and other plant remains such as leaves. During the Arles-Rhône 3 excavation, the contents of the jars and amphorae were, as far as possible, systematically removed and placed in the refrigerator for analysis. A portion was sifted to a minimum mesh of 1.1mm by the excavators and the rest in the laboratory through a sieve column of 4mm, 2mm, 1mm, 0.5mm and 0.250mm meshes. The sorting of the different samples first focused on the plant remains and secondly on the fish remains. For the latter, only those of the jars have so far been examined and studied, while archaeobotanical analyses were conducted on more varied containers (Tillier 2019). In the majority of the jars, the fish remains were distributed between 2mm and 0.5mm and between 1mm and 0.5mm for preparations based on fry. Below 0.5mm, there are only fragments of fin spines and some barely formed vertebrae. Unlike the ichthyofauna, the use of a 0.5mm sieve is insufficient to collect many of the seeds of the different varieties of aromatic plants. This is particularly the case for those of the Lamiaceae family (oregano, marjoram, thyme, rosemary…), despite the fact that the leaves and/or floral tops are the anatomical parts used as aromatics. Therefore, when working in wet contexts, it is necessary to use a 0.250mm or at least a 0.315mm mesh in order to ensure that the very small seeds, likely to be used in food preparations, are also recovered.

420

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts…

The advantage of freshwater contexts is that all marine fish remains can be associated with the primary content of the vases. On the other hand, a risk of intrusion is possible for the plant remains in unopened jars and amphorae. Note that waterlogged contexts have the advantage of allowing for the preservation of plant macroremains (seeds and fruits, but also sometimes flowers and stems), in particular fruit trees, vegetables, herbs, spices, oilseeds and many wild plants which are often less abundant and diversified in carbonized or mineralised contexts. On the other hand, in some of these contexts, such as the peri-urban dumping ground of Arles-Rhône 3, it is often difficult to find or identify ‘closed assemblages’ that would allow us to clearly circumscribe in space and time plants assemblages belonging to palaeobiocoenoses (Marinval 1988; Willerding 1991), in particular when the containers are not closed. The methods of transportation for the plant assemblages must also be taken into account, since the Rhône is a powerful watercourse which can carry plants from quite a distance. The question arises, therefore, whether the archaeobotanical remains found in the jars and other containers from ArlesRhône 3 were part of the primary content (ingredients of food preparations) or whether they happened to be there as part of the waste generated by other activities of the city’s population or as part of the natural vegetation growing around the river. On the one hand, we should remember that the majority of aromatics (celery, dill, fennel, caraway…) grow wild in this Mediterranean region (Rameau et al. 2008), as well as in Italy (Pignatti 1982), and that they can also be cultivated there. On the other hand, regarding grape pips or olive stones, which were also used in certain food preparations, they could also come from consumption or production activities (wine making, olive oil processing…) corresponding rather in this case to waste thrown into the river dumps. Thus, from a methodological point of view, in order to know if the archaeobotanical remains are intrusions linked to taphonomic processes, the taxonomic composition of the jars’ contents must be compared to that of control samples taken from the surrounding sediments. In the case of Arles-Rhône 3, this could not be done, but the archaeobotanical results could be compared with those obtained in three samples taken from under the wreck (Tillier et al. 2016; Tillier 2019). 2.2. The contents of the Latium jars Among the ceramics that produced fish remains, there is a category of jars (Ostia II 401) produced in the Lazio region called ‘garum jars’ or ‘Latium jars’ (Djaoui, Piquès and Botte 2014). This generic name derives originally from two isolated finds, one in Pompeii (Gasperetti 1996: 32) and the other in Port-La-Nautique, near the city of Narbonne (Sanchez 2003: 117/118 fig. 53 no.8). In parallel, the excavations of the Rhone at Arles also revealed the inscription on a jar sardi (RHO.05.Z99.349), interpreted as sardina or sardine (Long et al. 2009: 588). Currently, most of the discoveries have been made in

this last region, in particular during the Arles-Rhône 3 project. The analyses on 18 of these jars has revealed that 16 of them still contained fish remains trapped in the pitch, which confirms their function as a container for fish products. These jars are attested in Gaul only in ports: Arles, Port-la-Nautique (Narbonne), Antibes, Fréjus, Marseille (Djaoui, Piquès and Botte 2014) and recently Lattara2. Given this observation, as well as the few finds made only on boats, we associated their contents with food on board, intended for the sailors. The jars were carried away from the ports of Rome and thrown in the port of arrival once empty or almost empty (Djaoui, Piquès and Botte 2014). Several of these jars, however, still contained enough remains to identify the product (Figure 2). The first three jars presented here illustrate three different preparations made from mackerel: salsamentum, allec and perhaps garum (Figure 3). The first jar (AR3.2031.9), although incomplete, still contained around a hundred fish remains, including a rachis and connected bones of the opercular series, adhered to the wall. The bones are mainly from Atlantic mackerel (Scomber scombrus) and represent at least three individuals. In addition, we found 5 vertebrae of large (around 20cm) sardine or sardinella (Sardina pilchardus or Sardinella aurita). Mackerels are roughly equivalent in size, between 20 and 25cm. All the anatomical parts are represented, with the exception of the anterior part of the skull, in front of the eye. This observation, as well as the presence of a cut quadrate, indicates a mode of cutting attested on mackerels from the Dressel 9 amphorae of the wreck Sud Perduto 2 (Desse-Berset and Desse 2000: 75–79). The head of the fish was cut off on a transverse axis, slightly at an angle, passing through the eye. According to the authors, this cut would empty the mackerels of their blood, which would prevent the rapid decomposition of the flesh and improve conservation. In addition, the blood and entrails of the fish were allegedly used to make garum (Desse-Berset and Desse 2000: 77). It would then appear that the mackerels from this first jar from the Rhone excavation were probably gutted. This hypothesis is based on the absence of basipterygium, but also of all the parts of the branchial arches and the hyoid arch, which tends to indicate that the pelvic fins have been cut and the gills torn off. This is a process that can only go together with the evisceration of the fish. Note here that the cutting of the head at eye level could also be performed in order to save space in the jar, as in our current tinned sardines. By reproducing the gesture, we can observe that this method of cutting also makes it possible to tear out the gills and the intestines more quickly, all at once, which would be advantageous when processing on a large scale.  One Latium jar was found during the last archaeological campaign in 2019 (G. Piquès dir.) and the study is still in progress. 2

421

Figure 2. Arles-Rhône 3 – Inventory, preliminary results (fish and plant remains) and interpretation of the Latium and two other jars’ paleocontents.

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

422

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts…

Figure 3. Arles-Rhône 3 – Preparations made from mackerel identified by the fish remains recovered in three Latium jars (© G. Piquès, CNRS; ceramics drawing: A. Véléva).

423

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

accompanied by the guts, which do not leave any bone trace after decomposition with the exception of possible stomach contents (Van Neer and Parker 2008). Among the fish remains a hundred bones of at least 6 sardines measuring between 12 and 15cm long were also found. We have not observed any indication, such as deformities, suggesting that they are stomach contents, but the hypothesis cannot be excluded. It is also difficult to characterize this product, which could make one think of garum haimation (made up of gills, blood and guts) with the sardines in excess. It could also be allec, a processing residue of garum (Geoponics XX, 46). Once again, this food preparation made of gills reveals another by-product of salted mackerel, as well as the probable processing in the same fish-salting factories of different food preparations directly in these jars.

Figure 4. Arles-Rhône 3 – Fish remains from the Latium jar AR3.2001.51 (© G. Piquès). This preparation of beheaded and eviscerated mackerels would correspond rather to a salsamentum of salted and dried fish, or cured by brining. In the next two pots, we will see that the cut heads and guts of the salted mackerels were by-products used in other preparations. The AR3.2001.51 jar still contained its original contents because it had been stoppered with a cloth impregnated with pitch that was found still in place (Djaoui, Piquès and Botte 2014: 185). It was half-filled with a mass of fish bones caught in a muddy gangue. The sifting of the contents revealed nearly 8000 remains (Figure 4). After a long work of identification and counting we arrived at the following observation: the bones contained in this jar correspond above all to 30 young shad (Alosa fallax), 9 sprats (Sprattus sprattus), 3 anchovies (Engraulis encrasicolus) and a small sole (Soleidae). These fish were on average less than 12cm long and whole (all parts of the skeleton are represented). Added to this, bones of Atlantic mackerel were also found, surprisingly overrepresented for the front part of the skull compared to the rest of the body. The counting of the bones, as well as the presence of certain sliced ​cranial pieces (frontal, quadrate), made it possible to deduce the presence of 6 small whole mackerels (between 12 and 20cm), to which are added 8 heads of other larger mackerels (between 22 and 25cm). In other words, for the latter, the by-product of the preparation previously described above. This preparation formed from a mixture of small fish and mackerel processing waste would have consisted of a kind of spread that could be eaten with bread. Small bones, which are softened by salt, could easily be ingested. The ingredients and method of preparation could correspond to the definition of allec given by Pliny the Elder. What is interesting here is that garum processing waste would have become a product by itself made from a multitude of species and detritus. Finally, in a third jar (X-16373), empty at first sight, 74 elements of branchial arches corresponding to the gills of two mackerels were recovered. The gills were probably

Still in the same category of jars, another one (AR3.1012.8) produced more than 2000 fish remains, almost exclusively sardines less than 8cm long. All the bones of the skeleton are represented. Counting a selection of multiple head bones provides a minimum of 55 individuals. In addition, 2 anchovies and a small pandora (Pagellus sp.), less than 10cm in size, were caught with the sardines which form the basis of this preparation. It is unclear whether these remains are the residue of a poorly filtered sauce made from fry (apua garum?), similar to the one identified in Lattara — Lattes, France — (Piquès 2005), or whether the fish originally filled the entire jar. In the latter case, it would rather be a kind of fish mash (allec). Three other jars (AR3.3007.65, AR3.3013.1 and AR3.3020.1) provided only scales which, for two of them, appear to belong to fish of the Clupeidae family (sardine, sardinella or shad). For the third jar (AR3.3013.1), the scales belong to a small tuna or bonito, perhaps a piece of meat saltpreserved. Analyses on the contents of two of these jars also revealed the presence of seeds, in particular from aromatic plants used for seasoning. Note here the identification in the jar AR3.3020.1 of dill (Anethum graveolens), black mustard (Brassica nigra), caraway (Carum carvi) and fennel (Foeniculum vulgare). In the jar AR3.3007.65 coriander (Coriandrum sativum) was present with dill and fennel, along with other seeds also found (Figure 5). These seeds consisted of the pips or pits of three fruits (figs — Ficus carica —, olives — Olea europaea — and grapes — Vitis vinifera —), along with four other seeds of wild or undetermined taxa. In regard to fruits, in particular olives, we cannot exclude the possibility that they could also have been part of the food preparation (Tillier 2019), since the association of olives and garum is attested by tituli picti such as the one found on a Roman context at Lattara, in France (Bats 2010). But the question arises concerning the presence of the other seeds which seem to be intrusions. In this regard, it should also be noted that both the fruits and several of these aromatics (celery and fennel), as well as others (oregano and rosemary), are also attested in the samples taken under the wreck (Tillier et al. 2016). As mentioned previously, many of these herbs grew wild or were cultivated around Arles. However, the

424

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts…

Figure 5. Arles-Rhône 3 – Contents of the Latium jars AR3.3020.1 and AR3.3007.65 composed of Clupeidae scales (Sardina sp./Sardinella aurita) and aromatics (© G. Piquès, CNRS; ceramics drawing: A. Véléva; seeds pictures: © M. Tillier). hypothesis that they were used as condiments remains plausible, because of, on the one hand, the consistency of the results obtained on jar AR3.3020.1 and, on the other hand, because of the testimonies given by Latin authors (André 2009: 201–202). Other ingredients that could be incorporated into these food preparations were also added to the herbs. They are indicated by two inscriptions painted on two of the Rhone jars, one mentioning olives and the other onions (Allium cepa). The jar (AR3.3025.5) with the inscription CEPA, the onion, which delivered some unfortunately unidentified fish remains, could have contained allec in which onions were immersed. In this sense, we should remember Martial’s statement (Epigrams 3, 77, 5) ‘Capparin and putri cepas allece natantis’ evoking capers and onions swimming in putrid allec. Regarding the jar (AR3.2029.6) with the inscription OLIV (-), it was already mentioned that olives are known to be present in certain preparations based on fish. 2.3. Jars from Campania and Baetica The Latium jars are not the only containers from the Rhone trash dump which have produced fish remains. This is also the case for a jar originally from Campania, as well as a jar from Cadiz (Baetica). The first jar (AR3.3001.173) has two handles and belongs to a series of ten also found during the rescue excavation of the Roman barge Arles-Rhône 3. Regarding the origin of these ceramics many hypotheses have been formulated, such as Etruria, Africa or even southern Gaul, although a recent study has suggested Campania as the probable production area (Djaoui and Capelli 2017). The sifting of the sediment using a 0.315mm mesh allowed for the recovery of an assemblage of about 5000 fish

remains perfectly preserved. Eight taxa were identified. They are all small juvenile or adult specimens, the size of which does not exceed 12cm. All the anatomical parts are represented and no trace of cutting was observed. The proportion of vertebrae compared to the head bones corresponds to complete skeletons. The bones of the branchial arches are present, which indicates that the fish still had their gills and that they were probably not eviscerated. According to the count of a selection of head bones, this jar contained the remains of a minimal number of 88 individuals: 29 individuals from the Gobiidae family (33%), 26 sardines (29.4%) and 24 anchovies (27.3%), to which are added 3 red mullets — Mullus sp. — (3.4%), 2 horse mackerels — Trachurus sp. — (2.3%), 2 sand smelts — Atherina sp. —, 1 pandora and a small garfish — Belone belone — (1.1%). Among them is also a shrimp rostrum (Palaemon sp.). The mixture, in certain preparations, of fish and crustaceans is attested in a terrestrial context on a Dressel 6A amphora found in the north of Italy at Vicenza (Mazzocchin and Wilkens 2013). This amphora contained bones of picarel (Spicara smaris), one fish from the Cyprinidae family, shrimp (Palaemon adspertus) and crabs. Regarding the Campania jar found in Arles-Rhône 3, the presence of the shrimp seems involuntary. All the species and age classes represented likely derive from the fishing of an estuary or lagoon. This preparation, a kind of fish mash, could also be allec (Figure 6). The following jar (AR3.2007.126) belongs to a series of about twenty individuals with an umbilical bottom, a thick wall (almost 1cm) and bands of ochre colour deriving from Cadiz or Punic territory (Djaoui 2014: 163). A milling sample of the internal wall was taken in five of them in order to identify the nature of the organic impregnations and deduce their content. Chemical analyses revealed certain markers common to all vases, including compounds from resinous species (Pinus sp.),

425

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

and others specific to certain jars only. All the vases were waterproofed with pitch. Analysis of fatty acids, sterols and terpenes provides additional information on the content. Three of the vases, including the jar AR3.2007.126, contained fatty material from nonruminant animals. Chemical markers indicated for the jar AR3.2007.126 a large quantity of coprosterol, whose relative proportions were interpreted as horse dung (Djaoui 2014: 165). Later sieving of the sediment from this jar provided further information. This jar contained more than a thousand remains of small fish of only two species. According to the count of a selection of several head bones, this assembly consists of 78% sprat — Sprattus sp. — (MNI: 65) and 22% anchovies (MNI: 18). They are all small sized (less than 14cm long) and a priori were whole (the branchial arches are present). Once again, we have a product based on small fish, which could also meet the definition of allec (Figure 7).

Figure 6. Arles-Rhône 3 – Content of the Campanian jar AR3.3001.173 composed of small whole fish (© G. Piquès, CNRS; ceramics drawing: A. Véléva).

Thus, the study of these paleo-contents adds new elements to our knowledge of the products contained in these jars, used equally as a commercial sample of Baetican oil (Djaoui 2014), as potted food for sailors (Djaoui and Capelli 2017), or in Cadiz as funeral urns. 3. The underwater excavations at La Vila Joiosa: the wreck of Bou Ferrer The Bou Ferrer wreck is situated at a depth of approximately 25m along the coast of La Vila Joiosa (Alicante, Spain). It is a large ship, exceptional for the Early Roman period, with a width of about 12m and a length of about 32m, with a wellrounded profile and a shallow draught (De Juan, Cibecchini and Vento 2008; Cibecchini et al. 2014; De Juan, Cibecchini and Miralles 2014). It was carrying a homogenous cargo of Dressel 7/11 amphorae, the identification of which is currently under discussion even if four main subtypes (BF14) have been identified to date, stowed in three layers with a very likely fourth located midship. The majority seems to consist of Dressel 11-type amphorae (BF1-2), with parallels from shipwrecks at Sud-Lavezzi 2 and Tour Sainte Marie 1. Few amphorae of type BF4 (likely Dressel 9) find parallels again from Sud-Lavezzi 2. A dozen BF3 amphorae, almost proto-Beltran IIB, find the best parallels in workshops such as Villanueva de Puerto Real in Cadiz (De Juan, Cibecchini and Vento 2008: 188–189). The discovery of two coins during the 2017 campaign gives a terminus post quem for the wreck of AD 66, while the typology of the amphorae, the ingot epigraphy and the ceramics of the ship’s crew, allow us to assign a date range to the shipwreck between AD 66 and 68. 3.1. Materials and methods

Figure 7. Arles-Rhône 3 – Content of the Baetican jar AR3.2007.126 composed of small whole sprat and anchovies (© G. Piquès, CNRS; ceramics drawing: A. Véléva).

In 2007, a first examination of the content of 13 amphorae sieved in the port of La Vila Joiosa and sorted during the archaeological campaign was carried out. This sorting, although rapid and not exhaustive, made it

426

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts…

possible to give a first overview of the different remains present in the amphorae and to make some hypotheses concerning the product they contained (Cibecchini et al. 2014: 310–311). Based on these initial observations, other amphorae have since been sifted to verify that all the contents were fish-based and to determine if the ship was carrying other types of products in the Dressel 7/11 amphorae. Since 2007, the contents of 33 amphorae have been sieved and sorted. The amphorae were pitched, half or almost completely filled with a sandy-muddy sediment, and the stopper was usually found inside. The sieving was carried out by the staff of the museum of La Vila Joiosa using a sieve column of 4mm and 1.4mm meshes. The volume of sediment that was sorted was on average 7 litres per amphora and varied from 2.5 to 17 litres.

Figure 8. Bou Ferrer – Shellfish, crustacean fragments and other intrusive remains recovered in the 4-mm fraction of the amphora 550-2013-24 (© N. Rovira, University Paul Valéry-Montpellier 3).

Overall, the amphorae contained numerous shells and other marine remains, such as fragments of crustaceans and sea urchins, which appeared to correspond to intrusions, at least regarding the shells (Figure 8). The sorting of the debris from the screens also allowed for the discovery of a small number of fish remains. In general, we found between 50 and 400 bones per amphora, which is very little compared to the number of fish remains found in jars from Arles. This material was also very fragmented, in particular the bones of the 4mm fraction, and some of them were worn (Figure 9). The rate of undetermined remains (52.7%) was therefore relatively high. The most significant feature was the over-representation of otoliths (sagitta) compared to other elements of the skeleton (Figure 10). For instance, if we compare this with the number of vertebrae, a fish skeleton has on average around 40 vertebrae for two otoliths. In each of our samples, the number of otoliths was almost always greater than that of the vertebrae, representing sometimes more than double (Figure 11). In contrast with the contexts of Arles-Rhône 3, the question arises here of determining whether these remains of marine origin were part of the primary content of the amphorae, or whether they were intrusions representing thanatocoenoses (dead fish, shells and/or crustaceans brought into the amphora by octopus or introduced with the sediment). With regard to the plant remains, underwater contexts, in particular wrecks, generally present fewer risks of intrusion, even if they are not foolproof. Note in particular the possibility of finding plant products alongside amphorae or other ceramic vases coming from the cargo, such as cereals, fruits or vegetables, stored in bulk or in containers made of perishable materials (bags, goatskins, baskets…). This could also result from food stored for the crew’s consumption, or even food waste. In addition, intrusions are also possible depending on the proximity of the wreck to the coast or the depth at which it is located, related both to the existence of habitats or nearby areas of activity, or even

Figure 9. Bou Ferrer – Example of fish remains recovered in the 1.4mm fraction (© G. Piquès, CNRS).

Figure 10. Bou Ferrer – Otoliths recovered in the 4mm fraction consisting in the main fish remains found in many amphorae contents (© G. Piquès, CNRS).

427

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

2012 - 2 2012 - 4 2012 - 5 2012 - 6 2012 - 7 2012 - 9 2012 -10 2012 - 12 2012 - 13 2012 - 14 2012 - 15 2012 - 17 2012 - 18 2012 - 19 2012 - 20 2012 - 22 2013 - 10 2013 - 18 551 561 - 2013 - 22 580 - 2013 - 16

Otoliths 12 25 7 153 82 59 132 61 51 8 61 11 32 45 10 207 40 144 16 93 22 1271

Vertebrae 14 84 18 44 80 26 60 30 55 55 82 30 22 53 42 104 30 24 31 70 58 1012

Head bones 6 6 6 6 8

Teeth 3 9 2 9 14 2 30 3 2 4 4 3 4

5 3 5 2 8 4 2 7 4 8 5 5 4 10 7 111

31 11 11 4 10 2 158

Scales 19 20 20 49 152 13 48 37 29 47 9 3 9 23 87 15 17 4 20 5 626

Fin spines 3 4 2 9 10 2 3 2 17 2 14 1 4 9 6 2 3 5 5 4 9 116

Total 57 148 55 270 346 1020 278 136 159 118 178 49 67 123 85 439 104 206 64 207 103 3294

Figure 11. Bou Ferrer – Inventory of the different types of fish remains identified for a sample of 21 amphorae. To date, the content of only one of the sealed amphorae has been examined in detail. The amphora (2014-35) was first x-rayed after the desalting stage. The imagery showed a solid mass covered with a liquid, but apparently no skeletons or large fish bones. It was decided then to open it up with a grinder to observe its content (Figure 12). The solid mass consisted of very fine sediments, which filled almost a third of the amphora (12.5 litres). The liquid mass was sea water. We could therefore asses that the ceramic stopper, not completely hermetic, did not allow any elements larger than 4mm Ø to pass through into the amphora. No fish remains were observed by the naked eye, nor in the sediment, nor stuck to the walls of the amphora. The filling sediment was then divided into 6 samples taken every 10cm, from top to bottom, and sieved using screens of 2mm, 1mm and 0.5mm meshes. Figure 12. Bou Ferrer – Sawing of the sealed amphora 2014-35 (Dr. 7/11 type, BF-1) in order to access its content. It was filled (1/3 of the volume) with very fine sediments containing no elements larger than 4mm Ø (© G. Piquès, CNRS).

3.2. The contents of the amphorae

a watercourse. Archaeobotanical macro-remains of crops and wild plants, including spices and herbs, can thus be found in the sediments covering the wreck. It is therefore undeniable that meticulous protocols must be applied in particular when the containers are not closed, which in turn necessitates a systematic sampling of external sediment in order to compare the results. In the specific case of Bou Ferrer, this was not done, but the presence of sealed amphorae provided an additional element of control, which was considered sufficient enough to assess the primary contents.

A total of 24 taxa were identified. The Sparidae family was the most represented with at least six species, occasionally identified when the state of conservation allowed for it: Diplodus vulgaris, Pagellus sp. Boops boops, Sarpa salpa, Oblada melanura and Dentex sp. The elements found are mainly otoliths, some fragmented vertebrae and dental, as well as numerous isolated teeth. Among other common species we can include hake (Merluccius merluccius), present in 7 out of 11 amphorae with 54 remains in total (Figure 13). Once again, otoliths are the most abundant element for this species. The same occurs for several other taxa: Phycis sp. and Sciaenidae

A detailed study of fish remains has been done so far for the contents of only 11 amphorae, along with a rapid spot-checking of 22 (2012 and 2013 campaigns) and 13 (2007) other amphorae.

428

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts… Amphorae Dressel 7/11 - Bou Ferrer 2012 -6 2012 -7 2012 -9 2012 -12 2012 -17 2012 -19 2012 -20 2012 -22 2013 -14 Sparidae Scaena umbra/Umbrina sp. Carangidae Sardina sp./Sardinella aurita Merluccius merluccius Conger conger Pleuronectiforme Phycis sp. Labridae Serranidae Trisopterus sp. Trachinidae Scorpaena sp. Muraenidae Engraulis encrasicolus Gobiidae Elasmobranch (ray) Elasmobranch (shark) NRD (determined remains) Indeterminate TOTAL of remains Nombre de taxons

37 4 6 4 4 2 1 2 2

1 63 77 140 10

23 2 5 11 4 1 2

18 2 1 1 1

39 1

1

1 2

5

3 4

1

2

9

25

1

1 2 5 2

1 2 2 1 3

24 61 85 6

51 39 90 7

1 6

1

56 113 169 10

13

15 26 41 5

2

1

3

40 61 101 8

25 23 48 5

146 3 11 9 25 3 3 15 9 4 5 2

42 7 5 3 8 1 8 1

2 2

1

1

2 238 227 465 14

561 -2013 580 -2013 -22 -16 36 15 1 1 1 11 8 4 31 2 1 4 1 3 1 1

75 56 131 8

74 83 157 13

3 55 32 87 7

TOTAL

Ubiquity

403 21 31 42 54 43 15 33 20 14 11 6 6 3 4 5 4 1 716 798 1514

11 8 8 7 7 7 7 6 6 6 4 4 3 3 2 2 2 1

Figure 13. Bou Ferrer – Inventory of the fish remains per taxa identified in the contents of 11 amphorae. (Sciaena umbra and/or Umbra sp.). Finally, other fish regularly found in these amphorae were Clupeidae (Sardina sp. or Sardinella aurita) and Carangidae (cf. Trachurus sp.), whose vertebrae, unlike the previous taxa, are better represented. For Carangidae the dimensions of the vertebrae, about 15 to 18cm long, correspond to that of horse mackerel (Trachurus trachurus), and for Clupeidae (between 10 and 15cm long) to sardines (Sardina pilchardus). Clupeidae and anchovy, which were also present in some of the amphorae, are among the fish known for their use in the making of salting-fish products, as shown by data from Arles and other studies (Wheeler and Locker 1985; Sternberg 2000; Piquès 2005; García Vargas et al. 2018). This is also the case for mackerel (Scomber cf. scombrus) which appeared together with the two other taxa in two amphorae BF1-type from the 2007 campaign (Cibecchini et al. 2014). All these species are pelagic fish, suggesting that it is unlikely that they are intrusions. However, this question arises for the other taxa since they can live in environments like the one where the wreck was found. In some amphorae, intrusive remains were detected. This is the case, for example, of 31 well-preserved vertebrae and head bones of a small conger (Conger conger) from the amphora 580-2013-16. Several remains of Muraenidae and small fish abundant near the wrecks, such as Labridae, Sparidae, Serranidae, Gobiidae and lionfish (Scorpaena sp.) could also be intrusions. We can also add isolated teeth from sharks, as well as otoliths of Gadidae (Trispoterus sp. and Phycis sp.) and Sciaenidae (Sciaena umbra and Umbra sp.), because they are worn (polished) by water and because they are almost the only evidence of these species. Otoliths and teeth, more robust than bones and overrepresented, could be part of thanatocoenoses around the wreck, introduced with the sediment or by animals into the amphorae. The role of octopuses in this intrusion, on the other hand, seems minimal because these animals feed mainly on crustaceans, cephalopod and bivalve molluscs, and more rarely on fish (Ajana, Techetach and Saoud 2018; see also F. Cibecchini’s contribution on this volume).

Despite what has just been said, all the identified fish species can potentially be used for the production of salted-fish products. Some of them have never been found in Baetican amphorae contents, but they are documented in fish-salting factories. This is the case, for instance, of the forkbeard (Phycis sp.) found in Seville (Plaza de la Encarnación) and associated, among others, with sardines, in a fish-salting vat dated to AD 20/120 (Amores et al. 2007: 337–338; García Vargas et al. 2018: 298). Finally, note also that almost all taxa of Bou Ferrer’s amphorae correspond to those identified in another fish-salting vat in Malaga dated to the 4th-6th centuries (LozanoFrancisco 2017). What do we know about the production of salted-fish and sauces from the Cadiz Bay, the region from which the amphorae found at Bou Ferrer, currently under analysis, derive?3 We should remember that the ship, probably charted by Roman negociatores to buy foodstuffs in this area to supply Rome, was in use during the 1st century AD. Some data are provided for this period by the excavations of ‘Olivillo’ and ‘Colegio Mayor Universitario de Cadiz’, a large trash dump belonging to a fishsalting factory (Bernal‑Casasola, Vargas Girón and Lara Medina 2019). The study of the fish remains provides information concerning the nature of the products and the species used in their preparation, in particular the production of salted tuna, a Phoenician-Punic tradition well attested in the region, as shown through the cutting and preparation techniques (Bernal‑Casasola et al. 2019b and 2019c). Atlantic chub mackerel (Scomber japonicus) was also present, in particular as waste of haimation garum comprising blood, gills and guts (Bernal‑Casasola et al. 2019d). Sardines’ bones were particularly abundant in the dump and corresponded to small specimens (less than 15cm long). They were often associated with anchovy and small Carangidae, and sometimes with  Thin-sections analyses on the amphorae fabrics are underway (C. Capelli and F. Cibecchini). 3

429

Dr 7/11 Dr 7/11

Dr 7/11

Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11

Dr 7/11

Dr 7/11 Dr 7/11 Dr 7/11

Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11

Dr 7/11

Dr 7/11

Dr 7/11

Dr 7/11

2012 - 12 2012 - 13

2012 - 14

2012 - 15 2012 - 16 2012 - 17 2012 - 18 2012 - 19

2012 - 20

2012 - 21 2012 - 22 557 - 2012 - 23

430

2013 - 7 2013 - 10 2013 - 13 2013 - 14 2013 - 18

551

554 - 2013

555 - 2013 - 8

561 - 2013 - 22

1,4mm 1,4mm 1,4mm 1,4mm 1,4mm

34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93

1,4mm 1,4mm 1,4mm 1,4mm 1,4mm

42,93

0.5mm

Gades 34,28-42,93 1,4mm

Gades 34,28-42,93 1,4mm

Gades 34,28-42,93 1,4mm

Gades 34,28-42,93 1,4mm

Gades 34,28-42,93 1,4mm

Gades Gades Gades Gades Gades

Gades 34,28-42,93 1,4mm

Gades 34,28-42,93 4mm Gades 34,28-42,93 1,4mm Gades 34,28-42,93 1,4mm

Dr 7/11 BF-1 Gades

Dr 7/11

Dr 7/11

34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93

Gades 34,28-42,93 1,4mm

Gades Gades Gades Gades Gades

Gades 34,28-42,93 1,4mm

Gades 34,28-42,93 1,4mm Gades 34,28-42,93 1,4mm

12.5 (before sieving)

8,2

2,85

8,2

3,8 8,35 6,7

2,6

4,25

6,22 1,27

5,2

6

10,85

2,9 17 4,45 7,9

4,7

55

74

75

238

25

40

15

51

63 56 24

NRd

>16

103

207

64

206

104

249

439

85

178 59 49 67 123

118

136 159

278

148 55 270 346 102

57

520

NR

Sardina sp./Sardinella aurita (NR: >4)

Sardina sp./Sardinella aurita (NR: 11) Engraulis encrasicolus (NR: 1)

Sardina sp./Sardinella aurita (NR: 1)

No studied

Sardina sp./Sardinella aurita (NR: 6) Engraulis encrasicolus (NR: 2)

No studied Sardina sp./Sardinella aurita (NR: 3) Sardina sp./Sardinella aurita (NR: 5)

Sardina sp./Sardinella aurita (NR: 1) Sardina sp./Sardinella aurita (NR: 9) No studied Sardina sp./Sardinella aurita (NR: 8) Engraulis encrasicolus (NR: 2) No studied

Sardina sp./Sardinella aurita (NR: 2) Engraulis encrasicolus (NR: 3)

Sardina sp./Sardinella aurita (NR: 3) No studied

Sardina sp./Sardinella aurita (NR: 3) Engraulis encrasicolus (NR: 1)

Sardina sp./Sardinella aurita (NR: 3) Sardina sp./Sardinella aurita (NR: 2)

Sardina sp./Sardinella aurita (NR: 5) Engraulis encrasicolus (NR: 2)

Sardina sp./Sardinella aurita (NR: 1) Sardina sp./Sardinella aurita (NR: 4) Sardina sp./Sardinella aurita (NR: 11) Sardina sp./Sardinella aurita (NR: 1)

Sardina sp./Sardinella aurita (NR: 3)

Fish remains Composition (Clupeidae & Engraulidae) Sardina sp./Sardinella aurita (NR: 28) Engraulis encrasicolus (NR: 11)

Vitis vinifera (>2 pips+2 pedicels)

Vitis vinifera (1 pip)

Vitis vinifera (1 pip)

Vitis vinifera (1 pip)

Vitis vinifera (1 pip)

Vitis vinifera (2 pips)

Vitis vinifera (1 pip)

Vitis vinifera (1 pip)

Vitis vinifera (2 pips)

Vitis vinifera (1 pip)

14 fragments stems/roots+4 seeds

1 seed

Seed and fruit remains Herbs, spices Wild/Others

Vitis vinifera (1 pip)

Fruits

Tituli picti

fish sauce… oenogarum? oxygarum? lymphatum?

fish sauce…

oenogarum? oxygarum? lymphatum?

fish sauce… fish sauce… oenogarum? oxygarum? lymphatum?

fish sauce…

fish sauce…

fish sauce… fish sauce… fish sauce… oenogarum? oxygarum? lymphatum? fish sauce… fish sauce…

fish sauce… fish sauce… fish sauce… fish sauce… fish sauce… fish sauce… fish sauce… oenogarum? oxygarum? lymphatum? fish sauce… fish sauce… oenogarum? oxygarum? lymphatum? fish sauce…

fish sauce…

Product interpretation

Figure 14. Bou Ferrer – Inventory, preliminary results (fish and plant remains) and interpretation of 33 Dressel 7/11 amphorae paleocontents. Only the presence of ‘sardines’ (Sardina sp./Sardinella aurita) and anchovies is indicated in order to show the contents probably corresponding to fish sauces and to relate them to Vitis vinifera finds.

2014 - 35

580 - 2013 - 16

2013 - 1

1,4mm 4mm 1,4mm 1,4mm 1,4mm 1,4mm 1,4mm

Gades 34,28-42,93 1,4mm

34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93 34,28-42,93

Dr 7/11

Gades Gades Gades Gades Gades Gades Gades

Samples Mesh Volume (L)

Gades 34,28-42,93 1,4mm

2012 - 10

Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11 Dr 7/11

Dr 7/11

2012 - 1

2012 - 2 2012 - 3 2012 - 4 2012 - 5 2012 - 6 2012 - 7 2012 - 9

Amphorae Type Origin Volume (L)

Registration no.

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts…

small Sparidae. The association of the first three taxa is particularly found in many Gaditan Dressel 7/11 amphorae supposed to contain fish sauces such as sardines’ garum (Bernal‑Casasola et al. 2019a and 2019e). Regarding the contents of Bou Ferrer’s amphorae, the first observation we can make is that the 46 examined amphorae do not contain mackerel salsamenta, in particular Atlantic chub mackerel (Scomber japonicus) as those identified in the Dressel 7 and 9 amphorae from the wreck of Sud-Perduto 2, or in the Dressel 12 amphorae from Cap Bear 3 (Desse-Berset and Desse 2000). In our case, the products, according to the scarcity of fish bones, correspond to fairly well filtered sauces (liquamen, garum). The few remains recovered confirm that these products were made at least from Clupeidae, probably sardines (as in Olivillo), but also anchovies and small Carangidae. Note that the remains of Clupeidae, although they are present in small quantities (between 1 and 28 vertebrae per amphora), they are attested in 20 amphorae out of 28 (Figure 14). Concerning the other taxa, the question then arises: are they intrusions? An answer is provided by the sealed amphora 2014-35. Although the stopper was not completely hermetic, few remains had entered inside the amphora. Thus, we determined that the bioarchaeological macro-remains were not intrusions. This amphora, partially sorted (1/10), has revealed for the moment 4 vertebrae of Clupeidae (Sardina sp./Sardinella aurita), less than 12cm long, as well as 7 teeth of Sparidae, 2 otoliths and 1 vertebra of Carangidae, 1 otolith of hake (Merluccius merluccius) and an otolith of a flat fish (Pleuronectiformes). Therefore, the presence of Clupeidae and anchovies in many of the amphorae from Bou Ferrer is for us firm evidence that they contained fish products. To support this hypothesis, another significant fact can be added: the discovery of archaeobotanical remains in several of the amphorae containing Clupeidae bones, including the sealed one. The archaeobotanical remains found in amphora 2014‑35 are currently limited to 2 grape pips and 2 pedicels (Vitis vinifera), as well as to 14 fragments of stems/roots and 4 indeterminate seeds (Figure 15). In the 32 other amphorae entirely sorted, only 9 delivered waterlogged seeds, in a very small number, and almost exclusively limited to grape pips. No herbs or spices have been identified to date. The grape remains do not seem to be intrusive in the open amphorae, despite the use of vine shoots to wedge the amphorae on the boat. We can state this due to the absence of grape bunches among the shoots recovered and the concordance of the results provided by the sealed amphora. These rare Vitis vinifera remains are thus indicative of the presence of a product derived from grapes, probably a liquid, but which one? Ancient winemaking is a fairly well-known process thanks to literary, iconographic and archaeological

Figure 15. Bou Ferrer – Grape pips found in the sealed amphora 2014-35 (2 on the left) and in the amphora 2012-14 (1 on the right) accompanying in particular sardine or sardinella remains (© N. Rovira, University Paul Valéry-Montpellier 3). sources (Tchernia and Brun 1999; Brun 2004; André 2009). In regard to archaeobotany, the work carried out by E. Margaritis and M. Jones (2006) using ethnobotany and experimentation in order to characterize the products and by-products of each stage of the winemaking process, as well as the types of wines processed, should be noted. During the crushing and/or pressing of the grape bunches’ stages, there is always a significant number of by-products (especially seeds and pedicels, but often also fragments of stalks and skins) which remain in the grape must that is then fermented. Filtering occurs more or less rapidly depending of the desired kind of wine (white, rose or red, among others), in part before conditioning the must in the containers for fermentation (dolia, wooden tubs…) or when the wine is poured into the amphorae. The filtering of the wine lees can be more or less accurate and let pass, either voluntarily or involuntarily, a certain number of by-products, in particular pips but also some pedicels and small fragments of stalks and skins. Taking these practices into account, the low presence of grape pips, sometimes accompanied by pedicels, in some of the amphorae from Bou Ferrer indicates the use of wine, or even vinegar (acetum), fairly well filtered or in small quantities, in the making of the product contained in the amphora. The question regarding the quantity used of this product is interesting. We know that in these amphorae the final product also contained fish. However, is it a mixture of equal parts of wine (or vinegar) and a fish-based sauce (garum, liquamen)? Are one of the two ingredients predominant? Spices seem to be absent here, but the wine (if it is really wine…) could have been sweetened (mulsum?) or even mulled (carenum, defrutum, sapa?), as described in the Geoponics and other literary sources (Étienne and Mayet 2002; André 2009). This mixture of a fish sauce and a liquid derived from grapes seems to indicate an oenogarum or an oxygarum, but it could also be something else (perhaps lymphatum)? Chemical analyses could probably help identify the nature of the grape product used here by distinguishing at least wine from vinegar. Another interesting fact is that so far, all the examined amphorae seem to contain an identical product. This leads to the question of whether the other Dressel 7/11

431

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

amphorae on the ship, the number of which is estimated at nearly 3000, contained other products such as salted mackerel or tuna (salsamenta). A cross-check between the different types of amphorae (BF1-4) and their content should be carried out, to find out in particular whether or not grape remains appear in specific ones. Thus, the next step would be to select different contents to be studied according to their spatial distribution in the cargo in order to identify all the products transported by the boat. 4. Fish and grape products: new data, new questions The analyses of the contents of several types of ceramic containers from two underwater sites, Arles-Rhône 3 and Bou Ferrer, have highlighted several types of fish-based food preparations in which plant products were also used. This work, currently in progress, has raised several questions, both methodological and interpretative, and thus fuel discussions not only on the variety of products produced in Roman fish-salting factories but also on the versatility of containers (Silvino and Poux 2005; MartinKilcher 2011). The first set data presented here on the paleo-contents of Bou Ferrer seem to indicate the processing of products involving the mixture of several substances. The majority of the Dressel 7/11 amphorae analysed (22 out of 33) contain fish bones from species (mainly sardines and anchovies) known in the making of fish sauces, to which are added in 11 containers grape pips, and pedicels in the sealed amphora. The number of residues in both cases is very low, which suggests a rather liquid or very well filtered product. The possibility that this sauce accompanied pieces of meat from large fish (tuna? bonito?), boneless, which would have disappeared over time without leaving traces, cannot be excluded. We have set forth the hypothesis that the sauce could be therefore a mixture of garum (or liquamen?) cut with wine (or defrutum or vinegar), without being able to assume which of these products would have been predominant: fish sauce cut in wine/vinegar? Or wine cut with a fish sauce? In all cases it is conceivable that we are dealing with a kind of oenogarum or oxygarum, or perhaps liquamen, mixed with a derivative of grapes, as described by ancient literary sources and many researchers (e.g. Grimal and Monod 1952; Curtis 1991; Étienne and Mayet 2002; Botte 2009; Martin-Kilcher 2011; Grainger 2014). Recently, such garum products have been identified in fish-salting factories on the Atlantic coast of Gaul, thanks to a successful implementation of a multidisciplinary analytical approach (Driard, Dréano and Garnier 2017). Bou Ferrer is not the first case where grape remains are attested in fish-salting amphorae. Many other discoveries of wrecks (Cala Rossano, Sud-Perduto II, Grand Congloué…) and in other contexts (Punta de la Nao in Cadiz) (Arata 1994; Silvino and Poux 2005: 509–510) have led researchers to evoke other types of food preparations, such as lymphatum, known mainly by tituli picti and by some rare literary

mentions (Arata, 1992; Silvino and Poux 2005; Lagóstena Barrios 2007; Martin-Kilcher 2011). F. P. Arata (1994) suggested, based on H. Dressel’s assumptions and on the discovery of a stem and numerous grape pips associated with Dressel 9 amphorae from the wreck of Cala Rossano, that lymphatum was a product based on garum mixed with another substance, for him raisins (in bunch?) or grape must, and intended for ageing. From an archaeobotanical point of view, filling an amphora of raisins and/or grape must is likely to provide a fairly high number of remains, especially in the case of the former. If raisins were mixed with garum we could expect to find an assembly composed of several anatomical parts such as pips and skins (perhaps fragments of exocarp) and also, depending on the method of processing (destemmed berries or whole bunches), stalks (pedicels) and stems. In the case of grape must, the quantity of residue corresponding to grape pomace and their composition may vary depending on the quality of the filtering and the intention or not to make the product ferment in the amphora, but usually we would find mainly seeds, stalks and small fragments of skins. The fact that in the case of the amphorae from Bou Ferrer there are so few, does not seem to correspond to what one would expect if grape must would have been used, or even mulled wine (defrutum, sapa), but this is only an assumption. To our knowledge, no experimental data are available to know what by-products we might expect in these cases. In the course of the discussion and given the new data obtained thanks to the development of archaeozoological, archaeobotanical and chemical analyses in the study of the paleo-contents of ceramic containers and fish-salting factories, even if there is still progress to be made, it would be interesting to explore the question about the versatility of Baetican fish amphorae Dressel 7/11 and about wine production in the Cadiz Bay and neighbouring regions (Silvino and Poux 2005). The analyses of the contents of the Latium and other jars from Arles are also an example of the implementation of a multidisciplinary analytical approach. They concern not only the study of ichthyofauna but also plant macroremains, which highlights the use of herbs and spices in at least two of these preparations. The question has been raised of the reliability of these interpretations given the risk of intrusion into the open jars. If caution is required, flavouring salted-fish and fish sauces is a highly plausible practice since it is well described in ancient literary sources. We did not find in the bibliography consulted other references reporting similar archaeobotanical identifications, but other analyses, in particular palynological, have also been carried out on the contents of fish-salting amphorae, for example at Augusta Emerita (Bernal‑Casasola et al. 2015). Concerning potential condiments or plants used in fish preparations, the results of these analyses showed in particular pollen from Asteraceae, Chenopodiaceae and Brassicaceae, plants that despite their non-belonging to the main families of aromatics (Apiaceae and Lamiaceae) present certain

432

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts…

species which could have this function. The authors also suggested in the case of Chenopodiaceae the use of vegetables macerated with fish. The implementation of chemical analyses in the study on the paleo-contents should also be noted in the case of the jars from Arles (Djaoui, Piquès and Botte 2014), as in so many others (Silvino and Poux 2005), which, despite the occasional impossibility of precisely identifying certain products including fish, have become essential for this type of research. A final question that we wish to highlight in this paper concerns methodological aspects. In anaerobic environments, particularly underwater contexts, ceramic containers are often likely to preserve all or part of their content. Thus, the recovery of archaeobotanical macroremains, in particular fragile and varied anatomical parts (often absent in aerobic context), should not be a problem if certain precautions are taken. The contents of the vases, once withdrawn, must be always stored waterlogged and in a cool place (preferably in a fridge). In regard to sifting, a minimum mesh of at least 0.5mm but preferably 0.315mm or 0.250mm is necessary to recover small seeds, in particular certain aromatics. For ichthyofauna, the question of the conservation of the bones in wet environments does not arise, but that of the minimum mesh size used for sieving the samples is very important. We have seen through the case of the paleo-contents from Bou Ferrer and Arles that the number of faunal remains may not be high and that often small fish, such as sardines and anchovies or juveniles, are present. It is therefore necessary to also use fine meshes for sieving, following the same protocols as for archaeobotanical macro-remains. In conclusion, the implementation of meticulous protocols for sampling and sieving adapted to the recovery and identification of plant and animal macroremains conserved in wet environments during the two archaeological operations presented made it possible to bring to light evidence of food preparations today little known from an archaeological point of view and certainly undervalued. If the tituli picti on Dressel 7/11 amphorae are relatively numerous and give information about certain fish products and other substances, often associated with grape derivatives, such as lymp(hatum), other products, perhaps hidden under the common name of garum, are not always attested through these painted marks. Texts, including gastronomic ones, also indicate many practices involving animal and plant products that we can also try to trace by using multidisciplinary approaches. Funding This work was funded by LabEx ARCHIMEDE (‘Investissement d’avenir’ ANR-11-LABX-0032-01 program), university Paul Valéry-Montpellier 3, Archéologie des Sociétés Méditerranéennes (CNRS, UPVM, MCC, Inrap UMR5140).

Acknowledgments We would like to thank A. Espinosa Ruiz and the VilaMuseu (La Vila Joiosa, Alicante) team for their warm welcome and help for everything related to the Bou Ferrer project, especially during the excavation of the sealed amphora in 2017. We also thank D. Bernal‑Casasola for gently sending his latest work on ‘Olivillo’ (Cadiz). Finally, we are extremely grateful to B. L. Luley (Gettysburg College, USA) for his thorough reading of this article and his comments to improve it. Bibliography Ajana, R., Techetach, M. and Saoud Y. 2018. Diet of Octopus vulgaris from the Moroccan Mediterranean Coast. Thalassas: An International Journal of Marine Science 34: 415–420. Amores, F., García Vargas, E., González, D. and Lozano, M. C. 2007. Una factoría altoimperial de salazones en Hispalis (Sevilla, España). In L. Lagóstena, D. Bernal‑Casasola and A. Arévalo (eds), Cetariae 2005: salsas y salazones de pescado en Occidente durante la Antigüedad. Actas del congreso internacional (Cádiz, 7-9 noviembre de 2005). British Archaeological Reports International Series 1686: 335–339. Oxford, J. and E. Hedges Ltd. - Universidad de Cádiz. André, J. 2009. L’alimentation et la cuisine à Rome. Paris, Les Belles Lettres, (2nd ed.). Arata, F. P. 1994. Un relitto da Cala Rossano (Ventotene). Tituli picti su anfore e bollo su lingotti di stagno. In Epigrafia della produzione e della distribuzione. Actes de la VIIe Rencontre franco-italienne sur l’épigraphie du monde romain (Rome, 5-6 juin 1992). Collection de l’École française de Rome 193: 477–496. Rome, École française de Rome. Bats, M. 2010. Une lettre sur plomb à Lattes. In T. Janin (ed.), Premières données sur le cinquième siècle avant notre ère dans la ville de Lattara. Lattara 21: 749–756. Lattes, Association pour le Développement de l’Archéologie en Languedoc-Roussillon. Bernal-Casasola, D. 2015. What contents do we characterize in Roman amphorae? Methodological and archaeological thoughts on a ‘trending topic’. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 61–83. Esposende, Município de Esposende. Bernal-Casasola, D., Marlasca, R., Rodríguez Santana, C. G., Ruiz Zapata, B., Gil, M. J., García Vargas, E. and Alba, M. 2015. Garum de sardinas en Augusta Emerita. Caracterización arqueológica, epigráfica, ictiológica y palinológica del contenido de un ánfora Beltrán IIB. Rei Cretariae Romanae Fautorum Acta 44: 737–749. Bernal-Casasola, D., Vargas Girón, J. M. and Lara Medina, M. (eds) 2019. 7 metros de la historia de Cádiz… Arqueología en El Olivillo y en el Colegio Mayor Universitario. Cadiz, Universidad de Cádiz.

433

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

Bernal-Casasola, D., Marlasca Martín, R., Vargas Girón, J. M and Pascual Sanchez, M. Á. 2019a. Paleocontenidos ícticos en ánforas Dressel 7/11. In D. Bernal-Casasola, J. M. Vargas Girón and M. Lara Medina 2019 (eds), 7 metros de la historia de Cádiz… Arqueología en El Olivillo y en el Colegio Mayor Universitario: 509–516. Cadiz, Universidad de Cádiz. Bernal-Casasola, D., Marlasca Martín, R., and Vargas Girón, J. M. 2019b. Atunes en salazón y en conserva en las chancas gaditanas: perspectivas desde El Olivillo. In D. Bernal-Casasola, J. M. Vargas Girón and M. Lara Medina 2019 (eds), 7 metros de la historia de Cádiz… Arqueología en El Olivillo y en el Colegio Mayor Universitario: 517–534. Cadiz, Universidad de Cádiz. Bernal-Casasola, D., Marlasca Martín, R., and Vargas Girón, J. M. 2019c. Ronqueo de atunes en El Olivillo: evidencias osteológicas. In D. Bernal-Casasola, J. M. Vargas Girón and M. Lara Medina 2019 (eds), 7 metros de la historia de Cádiz… Arqueología en El Olivillo y en el Colegio Mayor Universitario: 535–541. Cadiz, Universidad de Cádiz. Bernal-Casasola, D., Marlasca Martín, R., and Vargas Girón, J. M. 2019d. Caballas desangradas en El Olivillo: evidencias del garum haimation y de los estorninos gaditanos en salazón. In D. Bernal-Casasola, J. M. Vargas Girón and M. Lara Medina 2019 (eds), 7 metros de la historia de Cádiz… Arqueología en El Olivillo y en el Colegio Mayor Universitario: 548–552. Cadiz, Universidad de Cádiz. Bernal-Casasola, D., Marlasca Martín, R., and Vargas Girón, J. M. 2019e. De la ictiofauna de El Olivillo: sardinas y otros taxones en época julio-claudia. In D. BernalCasasola, J. M. Vargas Girón and M. Lara Medina 2019 (eds), 7 metros de la historia de Cádiz… Arqueología en El Olivillo y en el Colegio Mayor Universitario: 553–558. Cadiz, Universidad de Cádiz. Botte, E. 2009: Salaisons et sauces de poissons en Italie du sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Brun, J.-P. 2004. Archéologie du vin et de l’huile dans l’Empire romain. Paris, Errance. Cibecchini, F., De Juan, C., Vento E. and Piquès, G. 2014. Il Bou Ferrer (Villajoyosa, Alicante, Espagna). Risultati delle campagne di scavo 2006-2007. In D. Leone, G. Volpe and M.Turchiano (eds), Atti del III Convegno di Archeologia Subacquea, Manfredonia (Italia), 4-6 ottobre 2007. Insulae Diomedeae 24: 301–312. Bari, Edipuglia. Curtis, R. I. 1978. The production and commerce of fish sauce in the western Roman Empire: a social and economic study. Unpublished Ph. D., University of Maryland. Curtis, R. I. 1991. Garum and salsamenta. Production and commerce in materia medica. Studies in ancient medicine 3. Leiden - New York - Copenhaguen - Cologne, Brill. Desse-Berset, N. and Desse, J. 2000. Salsamenta, garum et autres préparations de poisson. Ce qu’en disent les os. Mélanges de l’École française de Rome. Antiquité 112-1: 73–97.

De Juan, C., Cibecchini, F. and Miralles, J. S. 2014. El pecio Bou Ferrer (La Vila Joiosa-Alicante). Nuevos datos sobre su cargamento y primeras evidencias de la arquitectura naval. In J. Nieto Prieto and M. Bethencourt Núñez (eds), Arqueología subacuática española. Actas del I Congreso de Arqueología Naútica y Subacuática Española, Cartagena, 14, 15 y 16 de marzo de 2013: 113–124. Cadiz, Universidad de Cádiz. De Juan C., Cibecchini F. and Vento E. 2008. Intervención arqueológica subacuática en el pecio Bou Ferrer (Alicante, España): resultados preliminares de la campaña 2006. In J. Pérez Ballester and G. Pascual Berlanga (eds), Comercio, redistribución y fondeaderos. La navegación a vela en el Mediterráneo. V Jornadas Internacionales de Arqueología Subacuática (Gandia, 2006): 269–277. Valencia, Universitat de València. Djaoui, D. 2014. Découverte d’un pot mentionnant la famille des DD Caecilii dans un contexte portuaire situé entre 50-140 apr. J.-C. Découverte subaquatique à Arles, Bouches du Rhône, France. In R. Morais, A. Fernández, and M. J. Sousa (eds), As produções cerâmicas de imitação na Hispania. Actas do II Congresso da Sociedade de Estudos da Cerâmica Antiga da Hispânia (SECAH) (Braga, 4 a 6 de Abril de 2013). Monografías Ex Officina Hispana II (II): 161–178. Porto, Universidade do Porto. Djaoui, D. 2016. The myth of ‘Laccatum:’ a study starting from a new titulus on a Lusitanian Dressel 14. In I. V. Pinto, R. R. de Almeida and A. Martin (eds), Lusitanian Amphorae: Production and Distribution. Roman and Late Antique Mediterranean Pottery 10: 117–127. Oxford, Archaeopress. Djaoui, D. and Capelli, C. 2017. Objets d’importation ou objets personnels ? La dotation de bord des marins au regard du grand commerce, l’exemple du dépotoir portuaire d’Arles-Rhône 3. In SFECAG, Actes du Congrès de Narbonne: 115–132. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Djaoui, D, Greck, S. and Marlier, S. 2011. L’épave Arles Rhône 3, Le naufrage d’un chaland antique, enquête pluridisciplinaire. Arles, Actes Sud. Djaoui, D., Piquès, G. and Botte, E. 2014. Nouvelles données sur les pots dits « à garum » du Latium, d’après les découvertes subaquatiques du Rhône (Arles). In E. Botte and V. Leitch (eds), Fish and Ships. Production and commerce of salsamenta during Antiquity. Actes de l’atelier doctoral, Rome 18-22 juin 2012. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17: 175– 196. Arles, Errance - Aix-en-Provence, Centre Camille Jullian. Driard, C., Dréano, Y. and Garnier, N. 2017. Les sauces de poisson produites sur la côte atlantique des Gaules : sources archéologiques et productions diversifiées des ateliers de salaison. Gallia 74-2: 183–205. Étienne, R. and Mayet, Fr. 2002. Salaisons et sauces de poisson hispaniques. Trois clés pour l’économie de l’Hispanie romaine, II. Paris, de Boccard.

434

New data, new questions on the paleo-contents studies of Roman jars and amphorae in underwater contexts…

García Vargas, E., Roselló Izquierdo, E., Bernal-Casasola D., and Morales Muñiz, A. 2018. Salazones y salsas de pescado en la Antigüedad. Un primer acercamiento a las evidencias de paleocontenidos y depósitos primarios en el ámbito euro-mediterráneo. In D. Bernal-Casasola and R. Jiménez-Camino Álvarez (eds), Las cetariae de Ivlia Traducta. Resultados de las excavaciones arqueológicas en la calle San Nicolás de Algeciras (2001-2006): 287–312. Cadiz, Universidad de Cádiz. Garnier, N. 2014. Analyse chimique des sauces et des conserves de poissons : un état de la question. In E. Botte and V. Leitch (eds), Fish and Ships. Production et commerce des salsamenta durant l’Antiquité. Actes de l’atelier doctoral, Rome 18-22 juin 2012: 17–36. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17. Arles, Errance - Aix-en-Provence, Centre Camille Jullian. Gasparetti, G. 1996. Produzione e consumo della ceramica da mensa e dispensa nella Campania romana. In M. Bats (ed.), Les céramiques communes de Campanie et de Narbonnaise (Ier s. av. J.-C. - IIe s. ap. J.-C.), La vaisselle de cuisine et de table. Actes des journées d’étude (Naples, 1994). Collection du Centre Jean Bérard 14: 19–63. Naples, Centre Jean Bérard. Grainger, S. 2014. Garum, liquamen and muria: a new approach to the problem of definition. In E. Botte and V. Leitch (eds), Fish and Ships. Production et commerce des salsamenta durant l’Antiquité. Actes de l’atelier doctoral, Rome 18-22 juin 2012: 37–45. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17. Arles, Errance - Aix-en-Provence, Centre Camille Jullian. Grimal, P. and Monod T. 1952. Sur la véritable nature du garum, Revue des études anciennes LIV: 27–38. Jardin, C. 1961. Garum et sauces de poisson de l’Antiquité. Rivista di Studi Liguri XXVII: 70–96. Lagóstena Barrios, L. 2001. La producción de salsas y conservas de pescado en la Hispania Romana II a.C.-VI d.C. Instrumenta 11. Barcelona, Universitat de Barcelona. Lagóstena Barrios, L. 2007. Sobre la elaboración del garum y otros productos piscícolas en las costas béticas. Mainake XXIX: 273–289. Lagóstena Barrios, L., Bernal-Casasola, D. and Arévalo, A. (eds) 2007. Cetariae 2005: salsas y salazones de pescado en Occidente durante la Antigüedad. Actas del congreso internacional (Cádiz, 7-9 noviembre de 2005). British Archaeological Reports International Series 1686. Oxford, J. and E. Hedges Ltd. - Universidad de Cádiz. Long, L., Piton, J. and Djaoui, D. 2006. Les céramiques communes des gisements du Rhône à Arles, le faciès portuaire d’époque impériale. In M. Pasqualini (ed.), Les céramiques communes d’Italie et de Narbonnaise, Structures de production, typologie et contextes inédits (IIe s. av. J.-C. - IIIe s. ap. J.-C.). Actes de la table ronde de Naples (2-3 novembre 2006). Collection du Centre Jean Bérard 30: 201–215. Naples, Centre Jean Bérard. Lozano-Francisco, M. C. 2017. Estudio ictiológico de los residuos encontrados en las cetariae malacitanas. A propósito de las factorías del teatro romano de Málaga. In M. Corrales Aguilar (ed.), Aportaciones a la Malaca tardorromana y bizantina. Excavaciones en la factoría de salazones del teatro romano de Málaga (siglos IV-VI d C.): 143–164. Sevilla, Junta de Andalucía.

Margaritis, E. and Jones, M. 2006. Beyond cereals: crop processing and Vitis vinifera L. Ethnography, experiment and charred grape remains from Hellenistic Greece. Journal of Archaeological Science 33: 784–805. Marinval, P. 1988. Cueillette, agriculture et alimentation végétale de l’Épipaléolithique jusqu’au 2ème âge du Fer en France méridionale. Apports palethnographiques de la carpologie. Unpublished PhD thesis, EHESS-Paris. Martin-Kilcher, S. 2011. Formes d’amphores et contenu au Haut-Empire, points de repère et questions. In SFECAG, Actes du Congrès d’Arles: 417–426. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Mazzocchin, S. and Wilkens, B. 2013. Fish and Crustaceans from a Roman Amphora in Northern Italy, Archaeofauna 22: 105–111. Pecci A., Domínguez-Bella, S., Buonincontri, M., Miriello D., De Luca, R., Di Pasquale, G., Cottica, D. and BernalCasasola, D. 2018. Combining residue analysis of floors and ceramics for the study of activity areas at the Garum Shop at Pompeii. Archaeological and Anthropological Sciences 10: 485–502. Pignatti, S. 1982. Flora d’Italia. 3 vol. Bologna, Edagricole. Piquès, G. 2005. Les déchets d’une fabrication de sauce de poisson dans le comblement d’un puits gallo-romain et la question du sel à Lattes. In G. Piquès and R. Buxó (eds), Onze puits gallo-romains de Lattara (Ier s. av. n. è.IIe s. de n. è.). Fouilles programmées 1986-2000. Lattara 18: 293–305. Lattes, Association pour le Développement de l’Archéologie en Languedoc-Roussillon. Ponsich, M. and Tarradell, M. 1965. Garum et industries antiques de salaison dans la Méditerranée occidentale. Bibliothèque de l’École des Hautes Études Hispaniques 36. Paris, Presses Universitaires de France. Rameau, J.-C., Mansion, D., Dumé, G., and Gauberville, C., 2008. Flore forestière française 3. Région méditerranéenne. Paris, Institut pour le Développement ForestierCNPPF. Sanchez, C. 2003. Au carrefour des influences méditerranéennes et continentales. Le rôle de Narbonne dans le commerce antique. In E. Dellong (ed.), Carte archéologique de la Gaule, 11/1. Narbonne et le Narbonnais: 127–123. Paris, Académie des Inscriptions et Belles-Lettres. Silvino, T. and Poux, M. with the collaboration of N. Garnier 2005. Où est passé le vin de Bétique ? Nouvelles données sur le contenu des amphores dites « à sauces de poisson et à saumures » de type Dressel 7/11, Pompei VII, Beltrán II (Ier s. av. J.-C.-IIe s. apr. J.-C.). In SFECAG, Actes du Congrès de Blois: 501–514. Marseille, Société Française d’Étude de la Céramique Antique en Gaule. Sternberg, M. 2000. Données sur les produits fabriqués dans une officine de Neapolis (Nabeul, Tunisie). Mélanges de l’école française de Rome. Antiquité 112-1: 135–153. Tchernia, A. and Brun, J.-P. 1999. Le vin romain antique. Paris, Glénat.

435

Gaël Piquès, Núria Rovira, Margaux Tillier, Franca Cibecchini, David Djaoui and Carlos De Juan

Tillier, M. 2019. Économie végétale et échanges en Méditerranée romaine (1er s. av. n. è.- 5ème s. de n. è.). Étude carpologique de contextes portuaires. Unpublished PhD thesis, University Paul Valéry-Montpellier 3. Tillier, M., Bouby, L., Rovira, N. and Lefèvre, D. 2016. Carpologie en contexte portuaire romain: économie végétale et environnement des sites de Caska (île de Pag, Croatie), du Castélou-Mandirac (Narbonne) et d’Arles-Rhône 3. In C. Sanchez and M. P. Jézégou (eds), Les ports dans l’espace méditerranéen Antique. Narbonne et les systèmes portuaires fluvio-lagunaires. Actes du colloque international tenu à Montpellier du 22 au 24 Mai 2014. Revue Archéologique de Narbonnaise, Supplément 44: 381–395. Montpellier-Lattes, Éditions de l’Association de la Revue archéologique de Narbonnaise.

Trakadas, A. 2005. The archaeological evidence for fish processing in the Western Mediterranean. In T. Bekker-Nielsen (ed.), Ancient Fishing and Fish Processing in the Black Sea Region. Black Sea Studies 2: 47–82. Aarhus, Aarhus University Press. Van Neer, W. and Parker S. 2008. First archaeozoological evidence for haimation, the invisible garum. Journal of Archaeological Science 35: 1821–1827. Willerding, U. 1991. Präsenz, Erhaltung und Repräsentanz von Pflanzenresten in archäologischem Fundgut. In W. Van Zeist., K. Wasylikowa and K.E. Behre (eds), Progress in Old World Palaeoethnobotany: 25–51. Rotterdam, A.A. Balkema. Wheeler, A. and Locker, A. 1985. The estimation of size in sardines (Sardina pilchardus) from amphorae in a wreck at Randello, Sicily. Journal of Archaeological Science 12: 97–100.

436

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii Darío Bernal-Casasola Universidad de Cádiz

Daniela Cottica Università Ca’ Foscari

Ricard Marlasca

Posidonia S.L.

Carmen Gloria Rodríguez Santana Museo y Parque Arqueológico Cueva Pintada

Enrique García Vargas Universidad de Sevilla

Abstract: Between 2008 and 2012, a joint project called ‘Fishing and Fish-processing at Pompeii and Herculaneum’ explored the exploitation of marine resources in the region around mount Vesuvius in the Roman period. All available evidence for the marine species fished and consumed in the area (from iconography to archaeozoological remains), the fishing tackle, the areas used for processing and preparing fish, and the local/regional ceramic vessels (mainly amphorae and urcei) used for trading with fish marine foodstuffs were collected and analyzed. Field work was conducted in the so-called Garum Shop or Bottega del Garum (1, 12, 8), the only place clearly being used at the time of the Plinian eruption in AD 79 for the preparation and sale of fish preserves. An exceptional deposit of around one hundred complete amphorae, stored in the second courtyard of the Garum shop (room 13), known as the ‘pila d’anfore’, was analyzed; most of these containers belonged to the Dr. 21-22 family. The deposit was dug, analyzed in detail and partially published. The importance of this deposit is that for the first time we were able to demonstrate Botte’s hypothesis, that these italic amphorae were used for the bottling of fish products and not dried fruit. Most of the Dr. 21-22 contained archaeozoological remains, as they had been emptied just before the eruption (scales, vertebrae and other fish bones attached to their walls). The archaeozoological study carried out determined the existence of different fish-families (Clupeidae, Engraulidae, Carangidae, Scombridae and Sparidae); especially interesting is the connection of these archaeozoological remains with the tituli picti that refer to the palaeocontent (mainly MAL, SP and COP but also AB, CE, COP AB and VR are known) of the amphorae. Based on this evidence, we can argue that these are not abbreviations of fish names as previously assumed, but products manufactured with the said species, as in many cases the same fish species are associated with amphorae bearing different inscriptions. In this paper, the ichthyological palaeocontent of 8 amphorae from the ‘Pila d’Anfore’ is presented in detail, which confirms the use primarily of two taxa as the main ingredients (picarel — Spicara smaris and anchovy — Engraulis encrasicolus). These data verify the relationship of these well-known Italic amphorae types with Italian fish-processing plants from the 1st century BC to the 2nd century AD. Key words: Italic Amphorae; fish; garum; Pompeii; Garum Shop; tituli picti; archaeozoology.

1. Introduction Until recently, Italian Dr. 21/22 amphorae were generally believed to have been used for the storage and transport of fruit preserves: this idea was put forward by H. Dressel based on the tituli picti that some of them displayed, for instance Mal(a) Cum(ana) — apples from Cumae — or CE(rasa) — cherries  — (Dressel 1879: 167-172). This hypothesis was accepted by most researchers, from Callender and Zevi to the main reference works about Roman amphorae (Peacock and Williams 1986: 96–97; Sciallano and Sibella 1991). This widespread belief was challenged by E. Botte’s analysis of the Dr. 21/22 amphorae from Pompeii and

other areas of Thyrrhenian Italy and Sicily, following his detailed examination of the chronological and typological evolution of the family and, especially, a rereading of the painted inscriptions (Botte 2007; 2009a; and especially 2009b: 120–161). In 2008, in order to investigate the halieutic cycle in Pompeii, works resumed in the Bottega del Garum (1, 12, 8), and the question was re-opened owing to the large number of Dr. 21/22 found in this oficina salsamentaria, which specialised in producing and selling garum at the time of burial by Vesuvius’ eruption in AD 79 (a synthesis of the project can be found in Bernal-Casasola

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 437–452

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas

A

B

C

D

Figure 1. Dr. 21/22 amphorae from the Bottega del Garum, Pompeii, with inscriptions that allude to the contents, CE (A) and MAL (B); ichthyologic remains adhered to the inner walls, mixed with resin (C); fish remains resting on the paving, discovered when one of the amphorae in the bottom layer was removed (D). and Cottica 2013). This deposit is exceptional on three counts: the number of whole amphorae belonging to the same type found together (over a hundred); the presence of inscriptions in nearly all of them (Figure 1 A, B) (and, when they are missing, it is because of preservation issues); and, the abundant presence of fish residues inside the amphorae (Figure 1 C) or on the floor when the amphorae were lying on the ground (Figure 1 D). That is, the context presented a rare opportunity to analyse the content of Italian Dr. 21/22 amphorae, because the presence of tituli picti, macroscopic fish remains and well-preserved amphorae from which precise typological inferences can be made in the same archaeological context is truly exceptional. Some of our results have already been presented, for instance concerning area 13

of the Bottega del Garum, which was presented in the Fautores Conference at Catania, to which the reader is directed for contextual and functional issues (BernalCasasola et al. 2014); and also about the inscriptions and the interpretation of the deposit, which were presented at the conference Fecisti Cretaria (García-Vargas et al. 2020; Bernal-Casasola et al. 2020). These results confirmed for the first time Botte’s insightful hypothesis, by linking these amphorae with fish contents. Now it is widely accepted that Italian Dr. 21/22 were used for the storage and transport of fish products (Bertoldi 2012: 104; BernalCasasola and Cottica 2019; Menchelli in this volume). The aim of this paper is to present the first detailed analytical data, so that precise links can be drawn between

438

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii

typological variants and contents and to encourage the discussion among specialists. As such, in what follows we present the study of the fish palaeocontents of eight amphorae found in the 2009 excavation season; future works will present the results of the analyses carried out on all the amphorae identified in the Bottega del Garum, which are currently being undertaken by Universidad de Cádiz and Università Ca’ Foscari Venezia. 2. Characteristics of the sample and methodology The amphorae under study were those found to contain visible remains of palaeocontents during the initial stages of the excavation of the Bottega del Garum. The sample comprises eight specimens of Dr. 21/22 amphorae from the ‘pila d’anfore’, found in situ (A5, A8, A9, A11, A14, A36), and two more that were stored in the ‘Granai del Foro’ since they were dug out by Maiuri in the 1960s (no.43108 and 43129). As illustrated by figures 2 and 3, all of them belong to type Botte 2, from the so called Calabrian-Peloritan region, except for no.43129 which is a Botte 1 Sicilian amphora.1 To date, the study of the paleocontents of these eight amphorae2 remained unpublished (Rodríguez Santana and Marlasca, 2011), and had only been referred to in very general terms, in publications dealing with the excavation of the Bottega del Garum (Bernal-Casasola et al. 2014). The characterisation of the samples began with the description of the macroscopic remains and the verification of the excavators’ preliminary observations (Figure 4): A05 (Sample 1 = S1): to the naked eye, the residue collected from this amphora did not contain anything of note, and its detailed examination through a binocular lens confirmed the presence of no ichthyofauna (Figure 5 A); A08: the examination of the residue found inside the area of the belly revealed the presence of a large number of fish scales (Figure 5 B-D); A09: the residue found in the inner walls of the rim and neck contained a small quantity of scales and small bones; this amphora was illustrated elsewhere, as it is one of the very few in the Bottega del Garum which, in addition to paleocontents, presents a titulus — SP — and a stamp which alludes to the workshop in which it was manufactured (Bernal-Casasola et al. 2014: 226, fig. 5); the sediment was treated with water, which allowed for the identification of some of the bones (Figure 5 E-F); A11: the residues were found to contain only a few fish scales (Figure 5 G); A14: the inside of these amphorae was found to contain an abundance  Two more amphorae with fish remains from the Bottega del Garum and stored in the Granai del Foro were analysed, but are not presented here because they belonged to different types. One is a late Punic Serie 7/ Maña C2b — Sample 7, no.43102 — and the other an Italian Dr. 2/4 — Sample 10, no.43133. Numbers A(mphora) + no. were assigned during excavation, and the other two ones just with numbers correspond to those in the Soprintendenza (currently Parco Archeologico) catalogue. 2  Undertaken within the framework of a cooperation agreement between Universidad de Cádiz and Cabildo de Gran Canaria - Museo y Parque Arqueológico Cueva Pintada (Gáldar, Gran Canaria). 1

of bones and vertebrae (Figure 5 H, I); A36: was found to contain only a small number of scales in the interior of the rim and neck; the residue was not sieved owing to the small amount of faunal remains identified (Figure 5 J, K); 43108: the interior of this amphora yielded an earthy residue which, after being sieved through a 0.5mm mesh was found to contain only a few unidentifiable remains (Figure 5 L, M, N); 43129: this Sicilian amphora yielded multiple tiny bones within an earthy sediment matrix (Figure 5 O, P); only 2mm and 1mm sieves were used, owing to the small amount of sediment present, which made the use of the 0.5mm mesh redundant. Following this, the remains found were examined in the laboratory with the assistance of reference collections. The state of preservation of the material found in the amphorae was generally good, and in some cases the amount of residue present was sufficient for sieving and selection. In some instances, only very fragmentary anatomic elements could be identified (e.g. acantotrichia, lepidotrichia, costae or scutae). The different mesh sizes used for sieving heavily conditioned the type of sediment and bone remains found (Figure 4): • Sieve 1 (2mm): fragments of neurocraneum, shoulder and abdominal girdle, some hyomadibulare, and vertebrae. In addition to this, this mesh size yielded plant remains, insects, stones and other elements. • Sieve 2 (1mm): neurocraneum and viscerocraneum, as well as shoulder and abdominal girdle and axial skeleton; large number of vertebrae; little sediment. • Sieve 3 (0.5mm): large number of vertebrae, small fragments of viscerocraneum (but well preserved, especially concerning the most robust joints), abundant fragments of neurocraneum and axial skeleton; little sediment. • Sieve 4 (0.25mm): very fragmentary bone remains, and some tiny bones found whole; multiple remains almost reduced to dust. • Residue: ash, sediment and bone dust. 3. Anatomical characterisation and quantification of ichthyofaunal remains The archaeozoological remains were compared with specimens in reference collections. Since the samples comprise small species, the specific determination focused on diagnostic vertebral and cranial remains. In order to determine a Minimum Number of Individuals (MNI), the laterality of paired bones (sinistrum and dextrum) was examined whenever the state of preservation of the bones allowed. Since the bones were found in sealed contexts (amphorae) MNI was calculated separately for each amphora. The total number of remains examined in the 2009 season was 2356, leading to the specific characterisation of 1763 (Rodríguez and Marlasca 2011). Excluded from these figures are serial elements, such as vertebrae (both centra and processus spinosi); skeleton pinnarum (fins); acantotrichia; lepidotrichia; pterygoforia;

439

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas

20 cm

20 cm

A5

0

A8

0

20 cm 20 cm

0

A9

A11

0

Figure 2. Drawing and photograph of Dr. 21/22 amphorae from the Bottega del Garum being studied (A5, A8, A9, A11).

440

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii

20 cm

20 cm

0

0

43108

43129

20 cm

20 cm

0

0

A14

A36

Figure 3. Drawing and photograph of Dr. 21/22 amphorae from the Bottega del Garum being studied (A14, A36, no.43108, no.43129).

441

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas

Sample

Amphora

Residue

Weight (g)

>2

>1

>0.5

>0.25

Faunal remains

Tituli picti

S1 S2 S3 S4 S5 S6 S8 S9

A05 A08 A09 A11 A14 A36 no.43108 no.43129

Body Body Rim/Neck Neck Neck Rim/Neck Inner sediment Inner sediment

0.05 7 0.5 0.3 0.4 0.2 36.5 5.2

2.7 6.3 1.2

1.3 3 1

0.9 -

0.6 -

Scales and other remains Different remains water sieved Small scales Different remains Scales Different remains Minimal remains

MAL _ SP CE _ _ _ _

Figure 4. Fish samples from the Bottega del Garum in Pompeii, with indications of the location of remains and their composition (total weight and weight after sieving - S2 also includes abundant residues), and tituli picti referring to content.

Figure 5. Sediment in the amphorae: minimal remains inside A05 (A – S1); sediment (B), indeterminate fish remains (C), including scales (D) in A08 (S2); sediment after watersieving (E) and indeterminate elements (F) in A 09 (S3); small quantity of sediment from A11 (G), including a few scales (S4); sediment following sieving and identifiable fish remains (H, I; S5, A14); small quantity of sediment (J) and indeterminate elements (K) in A36 (S6); sediment from amphora 43108 (L, M), with detail of the process of bone selection (N) and anatomic and specific identification (S8); sediment from amphora 43129 (O), and selection of remains (P) from sample S9.

442

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii

Cranium (NR) Viscerocranium Even elements Maxillare Praemaxillare Dentale Articulare Praeoperculare Columna vertebralis (NR) Odd elements Urostylus Serial elements Vertebrae

Amphora 09 (S3) Spicara smaris (Linnaeus, 1758)

7 s 1 1

d

1 14

1 2 1

1 13

TOTAL

21

Figure 6. Relative frequency of the anatomical parts preserved in sample S3 from amphora A09 (the number columns refer to number of remains or NR). radialia; basalia; costae; branchiostegalia; teeth (molariform, canines and incisors); scutae (scales); and parts of the archus branchialis.3 This category also includes small flakes of bones from the viscerocranium and the neurocranium which, owing to their extreme state of fragmentation, risked double-counting and altering the final taxonomic characterisation. The following figures refer to taxonomic groups (Families and Gender) and not species because the identification of the remains to the species level often proved impossible. For each taxon, the anatomical elements identified are listed (in the case of paired bones, dextrum and sinistrum bones are also specified), and total numbers given. The figures also present total numbers for two anatomical groups: cranium (neurocranium, viscerocranium and zonoskeleton anterius or shoulder girdle), which include the head bones; and columna vertebralis, which includes the rest of the body of the fish. Photographs of relevant bone specimens are also included.4  The number of these serial elements were counted in some samples in order to estimate their representativeness, but with these estimates were not taken into consideration in terms of the interpretation of the sample. 4  The images were taken with a binocular magnifying lens and processed by Jacob Morales Mateos, Universidad de Las Palmas de Gran Canaria, whom we want to thank for his work. 3

4. Anatomical and specific determination of fish remains. In what follows, we summarise the results pertaining to the four amphorae that have yielded significant archaezoological remains, and in the following section we shall analyse the implications of these results. Amphora 09 (S3) A small quantity of ichthyological remains was found in the sediment adhered to the interior wall of the neck and rim (Figure 6). Despite the paucity of these remains, it could be established that the amphora contained a single species, the picarel (Spicara smaris). Unsurprisingly, vertebrae (which are more robust) clearly predominate over the rest of the bones. Comparison with reference specimens (the size and weight of which are known) suggests the use of very small specimens, one year old, around 60-70mm in length; a weight of 2-2.5g also suggests fishes of around one living year. Amphora 14 (S5) A small quantity of residue was collected from the interior wall of the neck. Despite the paucity of these remains, it was attested that the amphora contained a single species (Figure 7), the sprat (Engraulis encrasicolus) with around 60-70mm of whole length. Elements from

443

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas

Amphora A14 (S5) Engraulis encrasicolus (Linnaeus, 1758)

Cranium (NR) Neurocranium Undetermined Viscerocranium Even elements Articulare Hyomandibulare Columna vertebralis (NR) Vertebrae

9 5 s 43 43

TOTAL

d 2 2

52

Figure 7. Quantification and graphic representation of the relative frequency of anatomical parts identified in sample S5 in amphora A14 (the number columns refer to number of remains or NR). A

C

B

Figure 8. Vertebrae (A), articulari dextra — norma lateralis — (B) y hyomandibulare — norma medialis — (C) of anchovies (Engraulis encrasicolus) in amphora A14 (S5).

444

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii

Amphora 43108 (S8) Spicara smaris (Linnaeus, 1758)

Cranium (NR) Neurocranium Even elements Posttemporale Otolitus Viscerocranium Odd elements Urohyale Even elements Palatinum Maxillare Praemaxillare Dentale Dentale Articulare Operculare Hyomandibulare Epi-keratohyale Shoulder girdle Even elements Scapula Supracleithrale Columna vertebralis (NR) Odd elements Urostylus Serial elements Vertebrae Vertebrae praecaudalis Vertebrae caudalis TOTAL

70 s

s 3 2 2 5 2 6

5 5 1 5

1

4

Cranium (NR) Viscerocranium Even elements Quadratum TOTAL

d

Amphora 43108 (S8) Spicara spp.

12 s 5

Amphora 43108 (S8) Undetermined

Cranium (NR) Neurocranium Undetermined Even elements Frontale Posttemporale Viscerocranium Odd elements Urohyale Even elements Palatinum Maxillare Dentale Operculare Shoulder girdle Even elements Cleithrum Scapula Columna vertebralis (NR) Serial elements Vertebrae praecaudalis Vertebrae caudalis TOTAL

d 3 2 4 7 6 2

4 1 226 17 44 51 114 296

12

d 7

26 s 4

s

2 1 1 1

d 4

d 1 1 1

6 6 2 1 1 28

Figure 9. Anatomical parts of fish remains in amphora 43108 (S8), with NR (column numbers). the cranium and the rachis were identified, which is unsurprising, considering that these small fish were introduced whole in the container (Figure 8). In addition to this, multiple serial bones (acantotrichia, lepidotrichia, costae, procesus spinosus…) were also identified, but not quantified, because this information is irrelevant in terms of body parts and specific determination. At any

rate, these remains appear to match the diagnostic elements. The state of preservation of these remains was excellent (Figure 8). As usual, vertebrae clearly predominate over cranial remains, owing to the comparatively robust nature of vertebrae in anchovies (Engraulis encrasicolus).

445

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas

A

E

B

F

uncertain (Figure 9). Bones that could not be identified include at least one additional fish species, perhaps a small member of the Sparidae family, larger than the picarel. Picarel remains include 8 quadrata (6 dextra and 2 sinistrum); based on this and on the size of these remains, the MNI in the sample is 8 (Figure 10). These fished with have a whole length around 60-70mm, and their weight around 2-2.5g suggests again around one living year. Amphorae no.43129 (S9)

C

The extraction of a sample of earthy residue yielded a large number of tiny bones. The sample is overwhelmingly dominated by picarels — Spicara smaris — (Figure 11). Although the sample collected is not large, it is clear that small specimens were used, like in amphora A09 (S3) — 60-70mm of length; 2-2. g of weight, around one year old. A praemaxillare dextrum found within the sample belongs to another species that could not be established. It may be an accidental intrusion, because it seems clear that the producers’ intention was to make a picarel-only product (Figure 12).

G

D

Results from the four amphorae in which the specific identification of fish bones was possible have allowed for a general characterisation of contents. It must be taken into account, however, that in many of our amphorae specific identification was not possible, at least in the small samples collected (from the neck and rim areas). At any rate, the fact that they include fish remains (serial elements such as scales) is not open to doubt,5 and this is in itself highly significant, as this seems to confirm the use to which these amphorae were put.

H I

The first thing worth mentioning is the overwhelming presence of two species in the record: anchovy (Engraulis encrasicolus. Linnaeus, 1758) and picarel (Spicara smaris. Linnaeus, 1758), but always separately. There could be several reasons behind this separation, such as different catches at sea or culinary recipes. The fact is that anchovies and picarels are never mixed in the same container.

Figure 10. Illustration of fish remains found in amphora 43108 (S8): detail of remains found in the sieve (A); maxillare sinistrum — norma medialis — (B); articulare sinistrum — norma lateralis y medialis — (C); Dentalia dextrum (D) and sinistrum (E) — norma lateralis —; operculum dextrum — norma lateralis y medialis — (F); Otoliti — norma medialis and lateralis — (G); vertebra praecaudalis — norma lateralis — (H); and vertebra caudalis — norma lateralis — (I). Except for A, all remains correspond to Spicara smaris.

Second, the presence of other species (small Sparidae and other species that could not be defined) is likely accidental. The idea that other species could be introduced in small quantities to alter the flavour is suggestive, but the amounts in which they appear seems to rule it out. It is more plausible to think about the accidental presence of a member of another species in a catch predominantly made up of sprats or picarels.

Amphora no.43108 (S8) The earthy residue was partially extracted and examined, leading to the recovery of a large number of bones. The bones overwhelmingly belong to picarels; in fact, the remaining Sparidae bones could also belong to this species, but their state of fragmentation makes this

In fact, a recent study carried out on the coasts of the island of Eivissa, shows how fishing carried out with traditional gear, aimed at fishing for picarel, would catch 90% picarels (so called ‘gerret’ on the island), and the remaining 10% would be distributed among different species of diverse  In those cases in which only scales have been attested (A08, A11, A36) a precise identification will need to rely on biomolecular techniques. 5

446

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii

Cranium (NR) Neurocranium Undetermined Even elements Frontale Posttemporale Viscerocranium Odd elements Urohyale Even elements Quadratum Dentale Articulare Operculare Operculare Suboperculare Epi-keratohyale Shoulder girdle Even elements Scapula Columna vertebralis (NR) Serial elements Vertebrae TOTAL

Cranium (NR) Neurocranium Undetermined Viscerocranium Even elements Praemaxillare TOTAL

Amphora 43129 (S9) Spicara smaris (Linnaeus, 1758)

28 s

1 3 2

d

1 s 4 2

1 1 2 2

d 5 2 1

1 58 58 86 Amphora 43129 (S9) Undetermined

2 1 s 2

d 1

Figure 11. Anatomical fish remains identified in amphora 43129 (S9), with NR. families, among which there would be, for example, sparids, carangids or wrasses (Montero et al. 2018), which seems to be perfectly reflected in the samples studied.

appear in ranges of 8-9cm in length, being the majority or exclusive in ranges of 17 to 20cm. Therefore, we believe that they would have been captured in the spring.

The dimensions of the identified specimens allow them to be defined as juvenile, just over one year old, so most, if not all, would be female, since, following a study carried out on the Croatian coast (Dulcic et al. 2003), male individuals would

Third, the type of container seems to be irrelevant for the species contained within them (Figure 13). At any rate, picarels seem to be the most abundant species, regardless of the type of amphora under consideration.

447

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas A

D

Our first conclusion is that most samples analysed for this paper yielded fish remains (seven out of eight), which confirms that the Dr. 21/22 amphorae in the ‘Pila d’Anfore’ in the Bottega del Garum were used to store fish products. Concerning the species represented, a large proportion corresponds to members of the Sparidae family, overwhelmingly picarel (Spicara smaris), perhaps with the addition of blotched picarels (Spicara maena) and some members of the Sparidae family.

B

C

We want to highlight, as has also been done recently (Carannante 2019: 380), that when the analyses presented here were carried out in Pompeii in 2012, the results were recorded in a preliminary study (BernalCasasola et al. 2014), and at that point the picarel (Spicara smaris) and blotched picarels (Spicara maena) species were integrated into the Centracanthidae family; but a study published that same year 2014 (Santini, Carnevale and Sorenson 2014) integrated these species within the Sparidae family, as done for this study.

E

Figure 12. Illustration of remains of Spicara smaris found in amphora 43129 (S9): Articularia (A); Articulare dextrum — norma medialis — (B); Keratohyale and Epihyale anatomically connected — Norma lateralis — C); Operculare — norma lateralis — (D); and vertebrae (E). As such, it can be concluded that amphorae Dr. 21-22 were predominantly used to store a product based on picarels, although they were also used to contain anchovy-based products. It is worth stressing that anchovies are widely represented in the Bottega del Garum, notably in the content of the dolia found in room 9 (Bernal-Casasola and Cottica 2013: 47–49; Rodríguez-Alcántara et al. 2020). It must also be highlighted that both Calabrian and Sicilian amphorae were used to store the same products. 5. Conclusions and perspectives. Dr. 21/22 and fish. But, what products? First, it is worth emphasising that these eight Dr. 21/22 amphorae from the Bottega del Garum in Pompeii are the first to have their fish palaeocontents published in detail. Earlier publications included data concerning fish families and species in 28 of the Dr. 21/22 amphorae found in this context, but only in very general terms (Bernal-Casasola et al. 2014: 229, fig. 7) and without addressing the potential relationship between fish content and type of amphora; only the contents of the two amphorae preserved in the Granai del Foro (no.43108 and no.43129) had been published in some detail (Bernal-Casasola and Cottica 2019: 127, fig. 5).

The picarel (Spicara smaris) clearly dominates three of the contexts under analysis (A09, 43108, 43129). On two occasions they appear mixed with other species, which appear in very small proportions and were likely included in the batch accidentally, a situation that has numerous parallels (Bernal-Casasola et al. 2016). This species had already been associated with Dr. 21-22 amphorae, when it was found inside one of the containers from the Bottega del Garum currently on display in the Antiquarium of Boscoreale (Stefani et al. 2015: 87, no.99; Bernal-Casasola and Cottica 2019: 124–128, fig. 6). Recently, the contents of an amphora found in the Bottega del Garum during Maiuri’s excavations have been analysed, although unfortunately the container is lacking in context information and its typology is uncertain: the results of the analysis, however, determined that the amphora contained picarels, mostly female and approximately one year old (they were between 10-13cm in length) (Carannante 2019); this confirms the importance of this species for trade in Pompeii. The picarel was not highly commended in the classical sources, but must have been frequently used to make garum and other fish products in the Mediterranean, judging by its common presence in archaeological contexts. For instance, a Dr. 6 amphora, the first in which the species contained could be identified (Lepiksaar 1986), was found to include up to 24 different small fish species, mostly sardine but also picarel. This species was also identified in an amphora, probably produced in Africa, found in Olbia, although in this case the specimens used were larger, approaching 20 cm in length (Brusci and Wilkens 1996; Delusu and Wilkens 2000: 57). The species has also been identified in production contexts, for instance in the salting vats of the cetariae of Neapolis (Nabeul, Tunisia), where it was found in small quantities in

448

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii

Sample 1 2

Reference Amphora 5 Amphora 8

Type Dr. 21/22, Calabrian fabric Dr. 21/22, Calabrian fabric

Predominant species -

3

Amphora 9

Dr. 21/22, Calabrian fabric

Spicara smaris

4

Amphora 11

Dr. 21/22, Calabrian fabric

-

5

Amphora 14

Dr. 21/22, Calabrian fabric

Engraulis encrasicolus

6

Amphora 36

Dr. 21/22, Calabrian fabric

-

8

Amphora 43108

Dr. 21/22, Calabrian fabric

Spicara smaris *

9

Amphora 43129

Dr. 21/22, Sicilian

Spicara smaris *

Figure 13. Fish species identified in the amphorae (the asterisks denote intrusions, or species other than the predominant; fish drawings by A.M. Arias in Ictionimia andaluza, 2019, Madrid). products dominated by anchovies (Engraulis encrasicolus), round sardinellas (Sardinella aurita) and sardines (Sardina pilchardus) (Sternberg 2000), as well as in a number of Spanish production contexts such as Santa Pola and Cerro del Mar (García-Vargas et al. 2018: no.20, 43). For this reason, it can be argued that the picarel was a key ingredient in the preparation of fish sauces in Pompeii and its region, as the presence of remains of both Spicara maena and Spicara sp. in the well-known sewer under Cardo V, in nearby Herculaneum, seems to confirm (Rowan 2014: 67). Also significant is the identification of anchovies (Engraulis encrasicolus), unmixed with other species, in another of the amphorae (A14). The amphora was found whole and in situ in so-called Level 2, so nothing suggests that it might have been reused.6 This species has been found in other containers in the ‘Pila d’Anfore’, either on its own (A17, A49, A59, A76, A78, A83, A84) or mixed with other species (A17, A47, A48, A53, A55, A68, A74, A79, A80, A81), and the taxon is clearly associated with this shape (Bernal-Casasola et al. 2014: 229, fig. 7). The zooarchaeologial studies undertaken in the Bottega del Garum clearly indicate that garum was being fermented (‘brewed’) in the dolia found in room 9 when the eruption hit, and that this garum was based on anchovies (Curtis 1979; Rodríguez Alcántara et al. 2020), so it is a possibility that these had just come out of the Dr. 21-22 in the ‘Pila d’Anfore’. It is not easy, however, to draw a direct link, such as the one outlined above, between all the elements found in this archaeological context, as picarels are not found in the dolia, as would have been the case if the amphorae were emptied into the dolia. In any case, the evidence available suggests that the Dr. 21/22 in the ‘Pila’ had just been emptied and stored for sale/reuse/ recycling; no obvious sign of reuse exists, such as double tituli picti. At any rate, all the amphorae stored in room 13 were fish amphorae, while other types, such as African oil  In that case, the anchovies would have been mixed with the original paleocontent. 6

amphorae and Dr. 2/4 wine amphorae, were stored in a different area of the second courtyard of the Bottega, and were in all cases free from fish remains (Bernal-Casasola et al. 2020: 213, fig. 1).7 The product contained in the Dr. 21/22 amphorae is not easy to elucidate. The contents (small fish species) and the small size of the samples suggest garum-like sauces made with fermented fish, poorly filtered, which would explain the presence of some bones. However, two things make us think that the amphorae were full of a semi-solid paste with solid remains: first, the amphorae in the pile were nearly empty and upside down, so the animal remains found in the sample are likely nonrepresentative of the original content; second, we also know the contents of the bottom half of the Dr. 21/22 from the ‘Pila d’anfore’ on display in Boscoreale, in which at least one fourth or one fifth of the total capacity of the amphora was filled with remains of Sparidae (Spicara sp.) and other species (Rodríguez-Santana and Marlasca 2011; Bernal-Casasola and Cottica 2019: 128, fig. 6); prior to breaking and its contents dehydrating, this amphora would have been nearly full of fish bones. Since these did not belong to species of blue fish rich in fats, the most highly appreciated for the production of salted products (salsamenta), the logical conclusion is that they contained some sort of allec, the semi-solid residue that results from the filtering of garum/liquamen. Unfortunately, the tituli picti do not clarify the issue. Of the three present in the amphorae here considered (Figure 14 — MAL in A05; CE in A11; and SP in A09 —), only one can be clearly associated with picarel (A09), while the meaning of the other two is uncertain. It is, however, tempting to read SP as S(a)P(erda) or SP(arus), as has been proposed (Botte 2009b: 138–140), which would match our Spicara smaris. The association of picarels (or rather, Sparidae, as not all remains can be identified at the species level) with other  Only one specimen of Italian Dr. 2/4 (no.43133), can be interpreted as having been reused in this fish establishment. 7

449

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas

and not deliberate: it is likely that these compositions were determined by the catches at sea and that quality was given priority over quantity. Concerning species, it is important to stress the total absence of tuna remains in the Dr. 21/22 amphorae, or in any of the other archaeological contexts associated with the Bottega del Garum (Rodríguez Santana and Marlasca 2011). Recent finds in the Hellenistic and Roman cetariae of Portopalo di Capo Passero, in southeast Sicily, have for the first time linked tuna and Dr. 21/22 amphorae, both directly (bone remains have been found in some containers) and indirectly (through the large number of Dr. 21/22 amphorae, in some archaeological contexts directly in association with tuna remains, in a fish preserve factory that specialised in the fishing and processing of tuna); it is plausible to think, therefore, that, in addition to small fish species, these containers were also used for the commercialisation of red tuna (Bernal-Casasola et al. 2021). The confirmation of this point should be one of our future research priorities, and involves the characterisation of scales through biomolecular techniques (ancient DNA), since they often cannot be identified by anatomical comparison. The other important question that remains to be answered is why this cetaria in Pompeii dealt with fish so far away from the Vesuvian area, as 85.2% of the Dr. 21/22 amphorae were from Calabria and 4.9% from Sicily, with only 8.5% corresponding to the type Botte 3, from the central Tyrrhenian area. Perhaps the content of these amphorae was so-called ‘raw garum’ used as a condiment in different recipes. This could potentially explain the prevalence of fermented fish from Calabria and Sicily.8

Figure 14.- Readable Tituli picti on the amphorae under analysis: MAL (A05), CE (A11) and SP (A09). tituli (COP — A54, A79, A80 — or CE — A64 —, Bernal-Casasola et al. 2014: 229, fig. 7) indicates that these are not references to other fish species, as suggested by other authors, but to products created with similar ingredients. So, at present we can neither develop the seven abbreviations concerning products in the A record corresponding to this shape — COP, MAL, SP and to a lesser extent AB, CE, COP AB and VR — (Botte 2009b; Bernal-Casasola et al. 2014; García Vargas et al. 2020), nor determine production processes, which should be one of our priorities in future. What we do know is that many of these products were made with these sort of species and that they were similar to allec, with a semi-solid texture and many bone residues, hardly micronized. The results also suggest the preparation of mixed sauces, with several species but based on anchovies and picarel, the other species being added only in small quantities. It seems obvious, however, that when other species appear in such small percentages, as in our samples, this is merely accidental

Finally, it is worth pointing out that we already have the first organic residue analysis results yielded by five Dr. 21/22 amphorae from project ‘Impianto Elettrico’ in Pompeii. Although these results are still preliminary, they reveal the presence of Pinaceae-based products, related to the resin/pitch used to seal the walls of the amphorae. The analyses have also detected several acids that could also point to vegetal resins and fatty acids present in fish, alongside residues related to post-depositional contamination, vegetal oils and succinic acid present in fermentation. These deposits do not relate to wine, as there are no traces of tartaric acid, so the Dr. 21/22 in Pompeii present fish rather than wine-related markers, although further confirmation is needed; the presence of traces of vegetal oils is interpreted as evidence for the reuse of these containers (Pecci and Giorgi 2019).

 It is for this reason that we are not sure whether the interesting paleoenvironmental results coming from the study of the samples preserved in the so called ‘Laboratorio di Ricerche Applicate’ in Pompeii reflect the conditions of the picarels fished in Calabria and Sicily; or the ones of the local catches in the Bay of Naples, as has recently been suggested (Carannante 2019). 8

450

Dressel 21-22 Italic amphorae for fish: the archaeozoological confirmation from the garum shop at Pompeii

These results demonstrate that the Dr. 21/22 found in the Bottega del Garum were used to contain fish. The abundance of ichthyologic evidence found in this exceptional archaeological context illustrates the complexity of fish-processing palaeocontents associated with these amphorae. The evidence suggests that the samples analysed represent some sort of semi-solid product similar to allec, based on small fish species such as anchovy (Engraulis encrasicolus) and picarel (Spicara smaris). Bibliography Bernal-Casasola, D. and Cottica, D. 2013. Il progetto Dalla pesca al garum: lo sfruttamento delle risorse del mare nell’area vesuviana, 2008-2012. Una collaborazione italo-spagnola. In A. Arévalo González, D. BernalCasasola and D. Cottica (eds), Ebusus y Pompeya, Ciudades Marítimas. Testimonios monetales de una relación. Ebusus e Pompei, Città Marittime. Testimonianze monetali di una relazione: 29–59. Cadiz, Universidad de Cádiz. Bernal-Casasola, D. and Cottica, D. 2019. Pescado itálico en el Impianto Elettrico. Reflexiones sobre la filiación de las ánforas Dressel 21-22. In D. Bernal-Casasola and D. Cottica (eds), Scambi e commerce in area vesuviana. I dati delle anfore dai saggi stratigrafici I.E. (Impianto Elettrico) 1980-81 nel Foro di Pompei. Roman and Late Antique Mediterranean Pottery 14: 117–143. Oxford, Archaeopress. Bernal-Casasola, D., Cottica, D, García Vargas, E., Toniolo, L., Rodríguez Santana, C., Acqua, C., Marlasca, R., Sáez, A. M., Vargas, J. M., Scremin, F. and Landi, S. 2014. Un contexto excepcional en Pompeya: la pila de ánforas de la Bottega del Garum (I, XII, 8). Rei Cretariae Romanae Fautorum Acta 43: 219–232. Bernal-Casasola, D., Marlasca, R., Rodríguez Santana, C. G., Ruiz Zapata, B., Gil, M. J. and Alba, M. 2016. Garum de sardinas en Augusta Emerita. Caracterización arqueológica, epigráfica, ictiológica y palinológica del contenido de un ánfora Beltrán IIB. Rei Cretariae Romanae Fautorum Acta 44: 737–749. Bernal-Casasola, D. Cottica, D., García Vargas, E., Toniolo, L., Sáez, A. M., Bustamante, M., Cappelletto, E., Pecci, A., Expósito, J. A., Lara, M., Díaz, J. J., Vargas, J. M., Marlasca, R, Rodríguez, C. G. 2020. Ánforas, dolios y cerámica de la Bottega del Garum (1, 12, 8): reflexiones funcionales y socio-económicas. In M. Osanna and L. Toniolo (eds), Fecisti Cretaria. Dal frammento al contesto: studi sul vasellame ceramico del territorio vesuviano. Studi e Ricerche del Parco Archeologico di Pompeii 40: 211–225. Rome, L’Erma di Bretschneider. Bernal-Casasola, D., Malfitana, D., Díaz, J. J. and Mazzaglia, A. 2021. Portopalo, Vendicari y los orígenes del atún siracusano: novedades y desafíos. In Le cetariae ellenistiche e romane di Portopalo (Sicilia). Primi risultati da ricerche interdisciplinari. HEROM. Journal on Hellenistic and Roman material culture, Supplement 1: 461–544. Rome, CNR. Bertoldi, T. 2012. Guida alle anfore romane di età imperiale. Forme, impasti e distribuzione. Rome, Espera.

Botte, E. 2007. Les amphores Dressel 21-22 de Pompei. Quaderni di Studi Pompeiani 1: 169-186. Botte, E. 2009a. Le Dressel 21-22: anfore da pesce tirreniche dell’alto impero. In S. Pesavento Mattioli and M.-B. Carre (eds), Olio e pesce in epoca romana: produzione e commercio nelle regioni dell’Alto Adriatico. Atti del convegno (Padova, 16 febbraio 2007). Antenor Quaderni 15: 149-171. Rome, Quasar. Botte, E, 2009b. Salaisons et sauces de poissons en Italie du Sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard, 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Brusci, T. and Wilkens, B. 1996. Conserves de poisson à partir de quatre amphores romaines. Archaeofauna 5: 165–169. Carannante, A. 2019. The last garum of Pompeii: Archaeozoological analyses on fish remains from the ‘garum shop’ and related ecological inferences. International Journal of Osteoarchaeology 29: 377–386. Curtis, R. I. 1979. The garum shop of Pompeii (I.12.8). Cronache Pompeiane 5: 5–23. Delusu, F. andt Wilkens, B. 2000. Le conserve di pesce. Alcuni dati da contesti italiani. Mélanges de l’École française de Rome. Antiquité 112-1: 53–65. Dressel, H. 1879. Di un grande deposito di anfore rinvenuto nel nuovo quartiere del Castro Pretorio. Bullettino della Commissione Archeologica Comunale di Roma 7: 36–112 (part 1); 143-196 (part 2). Dulcic, J., Pallaoro, A., Cetinic, P., Kraljevic, M., Soldo, A. and Jardas, I. 2003. Age, growth and mortality of picarel, Spicara smaris L. (Pisces: Centracanthidae), from the Eastern Adriatic (Croatian coast). Journal of Applied Ichthyology 19: 10–14. García-Vargas, E., Bernal-Casasola, D., Cottica, D., and Toniolo, L. 2020. Epigrafia della pila di anfore della Bottega del Garum di Pompeii (1, 12, 8). In M. Osanna and L. Toniolo (eds), Fecisti Cretaria. Dal frammento ceramico al contesto: studi sul vasellame ceramico del territorio vesuviano. Studi e Ricerche del Parco Archeologico di Pompei 39: 135–142. Rome, L’Erma di Bretschneider. García Vargas, E., Roselló Izquierdo, E., Bernal-Casasola, D. and Morales Muñiz, A. 2018. Salazones y salsas de pescado en la Antigüedad. Un primer acercamiento a las evidencias de paleocontenidos y depósitos primarios en el ámbito euro‑mediterráneo. In D. Bernal-Casasola and R. Jiménez-Camino Álvarez (eds), Las cetariae de Ivlia Tradvcta. Resultados de las excavaciones arqueológicas en la calle San Nicolás de Algeciras (2001-2006): 287–312. Cadiz, Universidad de Cádiz. Lepiksaar, J. 1986. Tiereste in einer römischen Amphora aus Salzburg (Mozartplatz, 4). Bayerische Vorgeschichtsblätter 51: 163–185. Montero, I., Pantoja, J., Castaño, P., Marí, V., Barceló, C., García, M., Sureda, A. y Box, A. 2018: Pesca tradicional de Spicara smaris (Linnaeus, 1958) en la isla de Ibiza. In VII Jornades de Medi Ambient de les Illes Balears, Ponències i resums: 287–290.

451

Darío Bernal-Casasola, Daniela Cottica, Ricard Marlasca, Carmen G. Rodríguez Santana and Enrique García Vargas

Peacock, D. P. S. and Williams, D. F. 1986. Amphorae and the Roman Economy. An Introductory Guide. London - New York, Longman. Pecci, A. and Giorgi, G. 2019. Le analisi dei residui organici e la determinazione del contenuto di alcune anfore del progetto Impianto Elettrico. In D. BernalCasasola and D. Cottica (eds), Scambi e commerci in area vesuviana. I dati delle anfore dai saggi stratigrafici I.E. (Impianto Elettrico) 1980-81 nel Foro di Pompei. Roman and Late Antique Mediterranean Pottery 14: 157–165. Oxford, Archaeopress. Rodríguez-Alcántara, A., Roldan-Gómez, L., BernalCasasola, D., García-Vargas, E. and Palacios-Macías, V. 2020. Analysis of the productive capacity of the Dolia found in the Garum Shop (I.12, 8) using experimental archaeological methods. In M. Osanna and L. Toniolo (eds), Fecisti Cretaria. Dal frammento ceramico al contesto: studi sul vasellame ceramico del territorio vesuviano. Studi e Ricerche del Parco Archeologico di Pompei 40: 227–230. Rome, L’Erma di Bretschneider. Rodríguez Santana, C. G. and Marlasca, R. 2011. Análisis y estudio de las ictiofaunas arqueológicas de Pompeya y Herculano (Nápoles, Italia). Galdar, unpublished report.

Rowan, E. 2014. The fish remains from the Cardo V sewer. New insights into consumption and the fishing economy of Herculaneum. In Botte, E. and Leitch, V. 2014. Fish and Ships. Production et commerce des salsamenta durant l’Antiquité. Actes de l’atelier doctoral, Rome 18-22 juin 2012. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17: 61-73. Arles, Errance - Aix-en-Provence, Centre Camille Jullian. Santini, F., Carnevale, G. and Sorenson, L. 2014. First multi‐locus timetree of seabreams and porgies (Percomorpha: Sparidae). The Italian Journal of Zoology, 81: 55–71. . Sciallano, M. and Sibella, P. 1991. Amphores : comment les identifier ? Aix-en-Provence, Edisud. Stefani, G., Borgongino, M., Fergola, L. and Sodo, A. M. 2005. Cibus. L’alimentazione degli antichi romani. Schede. In Cibi e sapori a Pompeii e dintorni: 83–94. Castellammare di Stabia, Flavius. Sternberg, M. 2000. Données sur les produits fabriqués dans une officine de Neapolis (Nabeul, Tunisie). Mélanges de l’école française de Rome. Antiquité 112-1: 135–153.

452

Amphorae with residues from Southern Sardinia (Cagliari and Nora) Ignazio Sanna

Funzionario per l’archeologia subacquea, Soprintendenza Archeologia Belle Arti e Paesaggio per la città metropolitana di Cagliari e delle province di Oristano e Sud Sardegna

Laura Soro

PhD, ISEM CNR Research Grant

Cristina Nervi

CPIA, Ministero dell’Istruzione, dell’Università e della Ricerca

Abstract: Sardinia is a turning point along the shipping routes and the goods exchanges. Over the past years, the Soprintendenza ABAP in collaboration with the University of Cagliari and with different specialists, discovered various underwater findings and wrecks. These remains are related to a wide chronology and were found along the mid Southern Sardinian coast, originating from all the Mediterranean area and from the Lusitania. Two main towns were included within routes of stuff transport and perhaps of transhipment: Cagliari and Nora. Cagliari is the main port of the South coast of Sardinia. The underwater research in the port of Cagliari discovered a Late Republican wreck with a mixed cargo containing olive oil amphorae from Brindisium, Lamboglia 2 wine amphorae from Apulia and Dr. 1A, B and C amphorae with a rich epigraphic set and opercula with cork and pozzolana were also recovered. In the harbour a group of imperial wine pitched amphorae was also recovered: Dr. 9 fish amphorae, Pascual 1 and Dr. 2-4 forms, from Catalan and CampanianVesuvian area (from the second half of the 1st century AD); it was also found a Late Rhodian amphora. Moreover, it was found an AE3 amphora (2nd century AD), absolutely rare in the western Mediterranean, which comes from Egypt. Other significant findings of the 2nd century AD resulting from the harbour excavation context are some Gauloise 1, 4 and 5, Tripolitana amphorae, Dr. 20 oil amphorae and a Beltrán IIB pitched amphora from Baetica. This is an aspect which does not exclude the transport of salsamenta and processed seafood, as well as wines. Some fragments of Kapitän II and a very rare amphora (tyrian coast), of the 3rd century AD, coming from the Eastern Mediterranean, were also found. The transport of foodstuffs by sea is witnessed by other recoveries in other parts of the Gulf of Cagliari near the coast: the sea of Capo Sant’Elia coast returned a rare ‘Pseudo-Cos’ en cloche (1st to 2nd century AD) containing internal pitch remains, engraved with a ‘M’ letter at the base of the handle, and a cargo of North African and Gallic amphorae datable to mid-3rd century, all pitched inside. In the Late Roman period the transport of wine is also documented thanks to some African resined amphorae, Keay 25, and to some forms presumably made in relevant workshops in Sicily: Catanian MR 1 variants amphorae (MR1a and flat-bottomed types) and a Keay 52 (Termini Imerese variant) amphora. Samples of the 6th to 7th centuries from Eastern areas are LRA 1, LRA 2 and a globular amphora with a greek inscription engraved on the shoulder. Nora was a Phoenician, Punic and Roman city located on a promontory faced on the Gulf of Cagliari. It was undoubtedly an important reference in the field of trade and maritime transport of foodstuffs. The large amount of discovered materials is providing numerous data related to the treatment and content residues. Key words: Southern Sardinia; underwater archaeology; amphorae contents; maritime trade; foodstuffs.

1. Introduction The centrality of Sardinia in the Western Mediterranean and its resources undoubtedly had an influence, directly and indirectly, on the involvement of the island in ancient marine traffic and commercial exchanges. The most ancient archaeological evidence regarding the island’s imports/exports by sea, and the circulation of goods/exchange that must have characterised the commercial traffic in the Mediterranean, dates back to the 6th millennium BC (Freund and Batist 2014) and continued over the subsequent millennia with an increase in evidence. During the Iron Age, products such as meat and wine conveyed through the Phoenicians’ colonial activities, later further developed by the Punics, are not only attested in the Mediterranean Basin, but travelled

as far as the Atlantic Coast of Spain. An example of these are the small askoid Nuragic pitchers recovered in Huelva and Cadiz together with the Phoenician-Nuragic amphorae, of the Sant’Imbenia type, already attested in Carthage, which could have contained a particular exported Sardinian wine (Botto  2011).1 The most substantial underwater archaeological documentation concerns the entire Roman phase and the spread-out nature of the numerous sites with finds is indicative of the routes travelled, which joined the various centres of the Mediterranean and the Atlantic together, including  An already advanced level of cultivation and selection of vines for wine-making has been found in various Nuragic contexts on the island relating to the Middle-Recent and Final Bronze Ages (1350‑1150 BC) (Ucchesu et al. 2015). 1

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 453–472

Ignazio Sanna, Laura Soro and Cristina Nervi

Sardinia. These involved not only travel through, but also stops in, the more important coastal locations, such as Cagliari and Nora in the south of the island, to restock and exchange commodities and foodstuffs for other local products. On the other hand, navigation in Sardinian waters could have been a predetermined and usual route, but it could also have been an obligatory option, dictated by seasonal weather and sea conditions.

amphorae recovered, type Gr.-Ita. V,5 were also produced in Campania (Ischia-Gulf of Naples) (Figure 1.3). As already attested in Spain (Asensio 2010), this data item indicates the establishment of a commercial route from the Italic centres, focusing predominantly on the production of wine products and the exportation of the relative production surplus, above all from the former Greek colony of Ischia and the Campania area.

The materials and contexts that are presented below, which are pertinent to the Roman phase, datable between the 4th century BC and the 7th-8th centuries AD, are the result of underwater research conducted over the course of the last twenty years on the seabeds, ports and lagoons of the central-southern area of the island. The data acquired significantly increase the underwater archaeological database, in which the presence of Central Southern Sardinia was clearly very limited until the early 1990s. The interventions involved various marine locations spread around in the proximity of the coasts, at depths of no more than -50/-60m, including the Gulf of Cagliari (Soro and Sanna 2020) and the marine area of Nora (Sanna 2016), but the investigations in progress inside the current port area of Cagliari are offering a noteworthy contribution.2

In the proximity of Cagliari, outside of the waters of the current port, in the locality of Sant’Elia (SEL 2), another Magna Graecian context has been identified, also including, in addition to the amphorae, a clay loutèrion, similar to that of Sabaudo 2, with the bowl decorated with herringbone motifs (Figure 1.14). The rest of the cargo is contained inside the mattes out of which the material indicated above came. The small quantity of fragments recovered does not enable us to identify the chronology of the amphorae with any precision, however the triangular profiles of the rims and the inclination of the upper side, between 45° and 50°, could refer to Gr.-Ita. V-VI forms from the second half of the 3rd century BC (Figure 1.4-6).6 The context of Cagliari-Sant’Elia may be associated with the materials found in the cargo of wreck C of Gonnesa (Carbonia-Iglesias), in the south-western waters of Sardinia, where numerous amphorae, type Gr.-Ita. V-VI,7 internally resined and associated with black painted pottery produced in Campania (Salvi and Sanna 2000), have been identified and recovered. The transportation of wine from Campania between the 4th  and 3rd century BC is also attested in the southernmost sector of the Sardinian coast, in the locality of Capo Malfatano, where in the same context various samples of Greek Italic amphorae, type IV-V, datable in the mid3rd century BC, have been recovered, associated, as in the case of Gonnesa, with black painted pottery produced in Campania. A particular case concerns Nora (Cagliari), since during recent investigations in the western bay, some location contexts have been found where types Gr.-Ita. IV‑V are associated with torpedo-shaped Punic amphorae with umbonate bottoms (types T-4.2.1.2.), datable in the 4th-early 3rd century BC (Ramon Torres 1995), together with other amphorae, type Corinth B and Ionian, all extensively resined and presumably used for transporting wine. The case of Nora is comparable with the finds on land in the excavations in Carthage and provides further evidence of the circulation of wines of various origin by sea, probably in this case taken by the Punics to the territories belonging to them, such as Sardinia, after being brought together in the Punic ports of North Africa (Bechtold and Docter 2010: 115‑116).

2. The Gulf of Cagliari 2.1. Hellenistic/Magna Graecian Age At the end of the 4th-first quarter of the 3rd century BC, we see the first evidence regarding landings in Cagliari of ships coming from locations in the Tyrrhenian Sea during the period of the total control of southern Sardinia by Carthage (Sanna 2019). In the northern sector of the port of Cagliari a context has been discovered, referred to as Sabaudo 2, consisting primarily of wine amphorae, type Gr.-Ita. IV and V (Figure 1.1‑3), in addition to ceramic materials and an important clay loutèrion3 (Figure 1.13). From the site, which is still surrounded by sediments, some almost fully intact amphorae have been recovered, others fragmented, which still conserve abundant internal resin sealing and external dripping on edges and necks. An amphora, type Gr.-Ita. IV,4 conserves the stamp TINƟ in Greek characters on the handle, inserted in a rectangular cartouche (Figure 1.2). This rare stamp can be traced back to the wine production of the island of Ischia (Naples); the few examples known to date are only attested in Ischia, in the Campania area and in Sicily (Olcese 2010; Cibecchini and Capelli 2013: 435). The other  Since 2005 systematic surveys have been conducted by the Archaeological Heritage Service of Cagliari, including probes and trial excavations. 3  It has a large moulded rim delimiting a bowl with a roller decorated interior surface; the fabric is grey, rich in polished black volcanic inclusions referable to production from the Campania-Vesuvius area (4th and early 3rd century BC) (Russo 2011: 25–28). 4  The upper third is conserved, the neck is almost tronco-conical and the edge almost prickly; it has a pale-whitish slip and grey-hazel fabric. 2

 The body is ‘top-shaped’, the neck is short and tronco-conical, the rim is triangular with the upper part slightly tilted. 6  The characteristics of the fabrics would also indicate amphorae produced in Campania in this case. 7  A pink fabric, rich in volcanic black inclusions, comparable to productions from Campania (Cibecchini and Capelli 2013). 5

454

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

2.2. Republican Age The cargo of the wreck Sabaudo 1 belongs to the Late Republican Period, datable to the second quarter of the 1st century BC, and was recovered in 2006 during the first underwater archaeological intervention in the port of Cagliari, in an extensive area saved from dredging (Sanna et al. 2010). The cargo was composed of a small number of oil amphorae and a large quantity of wine amphorae, among which the form Dressel 1 from the Tyrrhenian Sea is not prevalent, as you might expect, but is almost equivalent in numbers to the Lamboglia 2 amphorae. So far in total 82 Dressel 1s have been recovered, in the variants A-B-C and 74 Lamboglia 2 amphorae (Figure 1.712),8 all internally resined, in addition to two ovoid Brindisine, not resined, very similar to type III from Apani (Palazzo 1989: 550, fig.1, n.3). Among the various stamps recovered, one printed on the rim of Dr. 1C is not inscribed, but has the representation of an anchor, others on Lamboglia 2 have inscriptions, including the following: ARCHEL(A) (Buora et al. 2007), PACONI,9 and one on two lines ARTHEMO (first line) MALLEOLI ∆ L•S•H (second line) (Figure 1.12). In the same context, other different samples of amphorae have also been recovered: an unresined Africana Antica (Capelli and Contino 2013)10 and two Hispanic resined ovoid amphorae: one Lomba do Canho and an Ovoid 4 (García Vargas, de Almeida and González Cesteros 2011). The cargo of the Sabaudo 1 also included black painted pottery from Campania A, B and Bi-oid, with thin ceramic walls, for the storeroom and the kitchen, and for use on board ship, there was Greek pottery and common ware from the Celtic tradition, oil lamps and a small round millstone. There were resinous products inside the containers discovered in the cargo of the wreck Sabaudo 1 (thin-walled bottle, lagynoi, ollae warehouse). In some cases the resin covered the walls with a waterproofing function, in other cases it was the product contained. Even though analyses are still underway to determine its exact nature, the amount and disposition of the resinous material discovered, as well as its smell, suggest it could have been used as a perfume or a perfumed substance, similar to the ones found at Nora.11 However, these resinous products could likewise have had flavouring and preservative functions for foods such as wines or meats, such as the Punic amphorae samples  For the conspicuous number of Lamboglia 2s, around 50 out of 500/700 amphorae, the wreck at Punta de Algas (Cartagena) (Pérez Ballester and Pascual Berlanga 2004) must be cited, as well as the wrecks at Sa Nau Perduda, El Golfet (Girona) and Colònia de Sant Jordi-Mallorca (Asensio i Vilaró 2010). 9  From Carthage a L•PACONI on Lamboglia 2 (Pérez Ballester and Pascual Berlanga 2004). 10  The association between the Lamboglia 2 amphorae and the Africana Antica type has also been found in the cargo of a wreck recently identified in deep water 150km south of Cagliari and 60km from Tunisia; it is currently being investigated by the authors here. 11  A recent discovery in Nora (South Sardinia), in the underwater context NRS 881, is a Roman amphora, type Dr. 2-4 (1st‑2nd century BC), which contained resinous products and aromatic woods (Sanna et al. forthcoming). 8

of the underwater archaeological context of the Santa Giusta lagoon (Del Vais and Sanna 2012; Lucejko et al. 2012). The cargo, prevalently wine, was accompanied by other foodstuffs, such as grapes, sloes, walnuts, hazelnuts, almonds, olives and pine cones, well preserved by the muddy anaerobic sediment (Figure 1.15).12 In addition to the undoubted need for these on board ship, we also cannot rule out the trading of these foodstuffs, as with the other major commodities transported. A second Late Republican context, from the third quarter of the 1st century BC, was found beneath a sandy seabed during the summer of 2015, around 150m from the coast, in the waters of Villasimius (Cagliari), in the far south-east of Sardinia. In the same location the presence was identified of a second cargo on top of the first, datable in the late 2nd or early 3rd century AD and composed prevalently of North African amphorae (Tripolitana I and III, of which some were resined), and around a thousand oil lamps of African production, together with G4 Gallic amphorae. The remains of the Late Republican cargo that have been brought to light and partially recovered include Dressel 1B and 1A wine amphorae, inside which, in addition to abundant resin, the cork stoppers are still preserved, sealed by the resin, and also partly the pozzolana stoppers. Some amphorae have small stamps on the rims or on the body, inscribed on rectangular cartouches, others without inscriptions show plant motifs. Adriatic amphorae have not been recovered in the context, a fact that may justify on the one hand a slightly later dating than Sabaudo 1, plausibly in the Pre-Augustan decade, but it could also represent a different origin for the ship and probably also an alternative market of exchange. 2.3. Imperial Age A cargo clearly specialising in the transportation of wine and fish by-products, referred to as Sabaudo 3 (Figure 1.1), datable in the first half of the 1st‑2nd century AD, has been identified in the central-northern sector of the port of Cagliari, north of the Ichnusa Dock. The heterogeneous lot of 32 amphorae recovered includes types Pascual 1B (Figure 2.9-10),13 Dressel 2-4 (Figure 2.1-8) and 7-11 (Figure 2.11-14). The production of the Pascual 1Bs for the external market arrived shortly after the mid‑1st century AD, and for some subsequent decades its diffusion was only local. In the period of large-scale production of the Pascual and wine from Tarragona the main exportation is attested in Gaul, with little in Italy, with routes passing through Bonifacio, as indicated by the wrecks of Sud-Lavezzi 2 and 3; the finding in Cagliari is therefore new data (López Mullor and Martín Menéndez 2008).  The drained sediments are filtered in the excavation phase or floated in the laboratory. 13  Pseudo bifid handles, high tronco-conical rim, with bulging and protruding upper edge, short neck. The Fabrics are beige-pink and beige-yellow with fine texture, referable to Laietana workshops such as Can Tintorer-El Papiol (Pascual Guasch 1977). 12

455

Ignazio Sanna, Laura Soro and Cristina Nervi

Figure 1. 1. Port of Cagliari: sectors of underwater research and contexts discovered (draw by I. Sanna and S. Fanni); 2-4. Greek-Italic amphorae from Sabaudo 2 shipwreck (4th-3rd century BC); 5-6. Amphorae from Sant’Elia (gulf of Cagliari)(draw by R. Arcaini); 7-12. Amphorae from Sabaudo shipwreck (first half of the 1st century BC) (draw by R. Arcaini and S. Fanni); 13. Loutèrion Sabaudo 2; 14. Loutèrion SEL 2; 15. Fruits and seeds (olea stones, grape pips, hazelnuts, walnuts) contained in the Sabaudo 1 (photo I. Sanna).

456

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

The group of 15 amphorae recovered, type Dr. 2-4, include Italic samples from the Campania area, along with Tarraconensis, Laietan and Gallic ones from Frejus. The various productions are distinguished by the types of fabric and form, with variants also within the same area of origin.14 From morphological examination and from the fabrics we find a prevalence of Catalan productions, followed by Italic, while the Gallic ones are in a minority. It is worth highlighting two samples of non-western wine amphorae belonging to the cargo: an amphora from Rhodes and an Egyptian one, type AE 3, both incomplete and without bases, but still in good condition and well resined internally. The one from Rhodes is datable between the 1st century BC and the 2nd AD (Dangreaux and Desbat 1987),15 while the Egyptian amphora, which is rarer than the one from Rhodes (Figure 2.16), was produced in various workshops located in the area of the mouth of the Nile (Şenol 2007; Empereur and Picon 1989).16 An important detail, which connects well with themes of the sea transportation of foodstuffs, is the presence, in addition to the obvious drips of resin on the outside of the rim, of a small hole intentionally made at the base of the neck, near to the lower attachment of the handle, used to favour the fermentation of the wine contained inside, which has already been found in the containers (Figure 2.18).17 As regards the 12 Dr. 9 amphorae, we have no complete samples; the upper thirds of the containers have been recovered, with the interior surfaces very resined and with obvious external drippings. All the specimens have mixed morphological characteristics, referable to the intermediate productions, not the late ones, datable between the first and second quarter of the 1st century AD (Figure 2. 12-14).18 On one Dr. 9 the tituli picti are conserved on the neck, arranged on at least three lines. The Dressel 7-9 type form is referable to Catalan productions imitating those of Baetica, very similar to the upper part of the PE41 from Ibiza, such as the one recovered in Nora (Cagliari) (Figure 2.11). The reference to the Catalan-Balearic area is also confirmed by the characteristics of the grey-hazel coloured fabric, which is fine and purified, with rare pale

 The fabrics of the Tarraconensis and Laietana amphorae reveal differences in the matrix and the tempers, as well as in the colour, corresponding to geological deposits with different and non-homogeneous characteristics (López Mullor and Martín Menéndez 2008). 15  It has a clear slip, the fabric is beige-pink, the handles have a circular section and the elbow is tall and acute. 16  The fabric is quite hard, the colour is red-orange, the matrix is sandy, with little temper; it often has one white-rosy slip, applied with obvious residues in the proximity of the handles and in the body. 17  Available from The Alexandrian Centre for Amphora Studies, http://www. amphoralex.org/amphores/AE/AmphoresAE3.php [25-09-2017]. 18  They have flared rims curved towards the outside, more or less thickened at the top, and wide and short necks; the handles have two or three mouldings, the only toe is short, tronco-conical and hollow. The pale yellow external engobe is still present, the fabrics vary from pinkorange to beige-grey. 14

inclusions.19 From the marine sector, dominated by the characteristic Capo S. Elia, we can report a rare specimen of wine amphora, in a good state of conservation, with a ‘Pseudo-Cos’ en cloche form (=Agorà M54) (Figure 3.15).20 Its incorporation inside the matte of Posidonia indicates the high potential of the location context. Extensively resined inside with external dripping, it has a letter M engraved below the lower position of the handle. Dated between the 1st and the 2nd century AD and coming from the Aegean area (Cyprus, Cilicia), it was made as an imitation of the well-known amphora of Cos and the equally famous and prized wine transported in it (Empereur and Picon 1989: 231–232). Their main circulation was in the Near East (Reynolds 2005: 564), while it is rare in the West: among the few examples, we find this amphora in Provence (Sciallano and Sibella 1991), Leptis Magna, Lusitania and Tarragona. The recoveries of dolia, such as the two important specimens found complete and recovered in the waters of Capoterra in the locality of Frutti d’Oro (Cagliari) (Figure 2.17), in the proximity of the coast of the Gulf of Cagliari, fall within the dynamics of wine trade by sea. The interiors of the first two containers are completely resined, as are the exteriors around the bottom and 60cm of the wall, probably in order to insulate and protect the wine contained from the absorption of salty bilge water inside the hold. The data available are increased by the recent recoveries of other large fragments of rims and walls relating to five different containers found in the industrial port of Cagliari and a further fragment of rim similar to the previous ones recovered inside the port, not far from the site of the context of Dr. 2-4 and 7‑9 already described. The discovery in the sea of dolia in the Southern Mediterranean is a new case for Sardinia, compared to the other known cases of Corsica and the coasts of Latium and Tuscany, Liguria and Provence. The data in this case also reveal a southern alternative of the circuit of wine transported in bulk by sea, until now concentrated primarily in the central-northern sector of the Mediterranean, on the basis of the various wrecks. I.S. 2.4. The Late Empire and the Late Antique period The attestations relating to the Late Antique phase that emerged during the underwater research projects were concentrated in the central-eastern sector of the port of Cagliari (Figure 1.1) and in the stretch of sea to the south outside the port, as far as Capo Sant’Elia and the more distant south-eastern tip of the Gulf, in the locality of Villasimius. These were cargos that did not make their destination, moving through the difficult waters approaching the city of Cagliari at the mercy of the Sirocco, and in some cases  Its production seems to be concentrated in the southern region of the Catalonia, no later than AD 80 (López Mullor and Menéndez 2008). 20  Recovery carried out by underwater fishermen. 19

457

Ignazio Sanna, Laura Soro and Cristina Nervi

Figure 2. Amphorae from Sabaudo 3 shipwreck (1st-2nd century AD). 1-8. Dressel 2-4; 9-10. Pascual 1 b; 11. Dressel 7-9; 12-14. Dressel 9; 15, 18. Egyptian amphora type AE3, with drips of resin and hole in the neck; 16. Rhodian amphora (draw by R. Arcaini and S. Fanni); 17. Dolia recovered in the locality of Frutti d’OroCapoterra (Cagliari)(photo L. Soro).

458

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

complicated by the presence of semi-submerged rock formations, and at Cagliari hampered by a long sandbank running almost parallel and close to the coast (Soro 2019). The materials and contexts identified are significant markers of the highly structured and complex panorama of exchanges and supplies of food commodities, not only from the centres of the Western Mediterranean, but also from the more distant and renowned locations on the Near Eastern horizon, which also involved South Sardinia and highlight the important role of the city of Carales.21 In the marine zone of S. Elia, approximately 900m from the coast, a mixed cargo of amphorae has been identified, datable in the first half the 3rd century AD, included among which are the Africana IIA type, with abundant resinous residues inside, Africana IIC, unresined and, from Baetica, Beltrán IIB late amphorae, variant b (García Vargas and Bernal-Casasola 2008), and with them also some examples of Gauloise 4 from the last phases of production (Dangreaux and Desbat 1987: 128–129), as is know, carrying the fine Gallic wine. An amphora of the MRA 1b variant type was recovered in the 3rd-century layers in Leptis Magna (Bonifay and Capelli 2013) and Ostia (Panella and Rizzo 2014: 140–141). Imports of wine products are documented by some sicilian samples. A MR 1a amphora (Catanian MR 1a type, form 1), (Figure 3.7-8), datable between the end of the 2nd century and the second half of the 3rd century AD,22 and the underwater find at the port of Cagliari of a second similar specimen (Catanian flat-bottomed type, form 3 from inside the port basin of Cagliari), allow us to extend the panorama of the Sardinian evidence, which until now has been limited and indicated the importation of established Sicilian wines. The relationship with the production zones of North Africa, particularly Tripoli, is attested by some examples referable to the form Tripolitana III, from the latest phase of production (mid-3rd century AD) (Bonifay 2004: 104–105; Revilla 2010: 398–399; Panella and Rizzo 2014: 285–286); on the inner walls, in the proximity of the location of the handles, are the characteristic deep impressions left by the artisan in the phase of workmanship (Revilla 2007: 275) (Figure 3.1-2). None of the samples being examined have resinous residues, a factor that is rather usual for the Tripolitanas, considered to be oil amphorae par excellence (Mattingly 1985; Panella 2001: 211; Bonifay 2004). But the excavations conducted  All the materials presented below are still in the midst of the study phase and form have been the subject of the doctoral thesis Commercial traffic and port landings in Southern Sardinia, through the study of amphora containers coming from underwater archaeological research (3rd‑7th century) (University of Cagliari, Department of History, Cultural Heritage and Territory, tutor Prof. Rossana Martorelli, 2017-2018). 22  The new classification is proposed by C. Franco (Franco 2014: 250 fig. 4.38, n. 2). Catanian MR 1a type is attested in the 3rd-century layers in Leptis Magna (Bonifay and Capelli 2013: fig. 25, 1.7). 21

recently in Villasimius23 have unearthed an important cargo of hundreds of oil lamps — which are currently in the study phase — associated with amphorae type Africana IA and Tripolitana III. Some of the latter show clear and abundant traces of resining internally and on the rim sull’orlo (Figure 3.16). As regards the Tripolitana amphorae, the data are significant, because they indicate that the product contained not oil but perhaps wine. This hypothesis requires confirmation using appropriate chemical analyses, which are already in progress, bearing in mind results that have emerged in recent years from the chemical analyses conducted on the samples of resined amphorae, indicating a possible use (or reuse?) of the resined container for transporting oil (Garnier et al. 2013; Woodworth et al. 2015). The identification of resin on the Tripolitanas therefore opens up a new scenario, which leads us to the formulation of two hypotheses: first, it is possible that initially the resined Tripolitanas transported wines and/or fish derivatives. Another key to interpretation, (which, furthermore, does not exclude the former) a more plausible one, considering the cargo of oil lamps, would indicate a reuse of the amphorae, following an initial employment for transporting wine/fish, as the identification of the resin does not in itself constitute a sufficient element to rule out the presence of olive oil (Woodworth et al. 2015): the ‘incompatibility’ between resin and an oil-like substance could also be overcome by hypothesising a use of the resined containers for transporting an oil (with olive oil also not excluded) of a lower quality; the association with the large cargo of oil lamps would therefore make it reasonable to hypothesise that the resined amphorae could contain oil to be used for combustion and therefore for use with the oil lamps themselves.24 The circulation of commodities coming from the MiddleEastern area towards the West is a phenomenon that has already been identified during the peak of the Imperial period inside the Gulf of Cagliari, registering a certain continuity over the centuries.25 Kapitän II type amphorae have been recovered26 in various sectors of the current port of Cagliari, with traces of internal resining, originating from the Aegean Sea or, as has recently been proposed, from the area of the Black Sea (Swan 2004: 379–380; Reynolds 2010: 90), which made its appearance at the end of the 2nd century but continued to be present at Leptis Magna in contexts of the 3rd‑4th century27 (Bonifay and Capelli 2013: 72, 117) (Figure 3.10). The detail of Eastern importing is confirmed by two small associated fragments of rim, belonging to an amphora probably produced in Tyre, (Figure 3.11) referred to  Cf. supra.  The fuel used for the oil lamps could have been pure olive oil or mixed with tallow, but also vegetal and/or animal fat (Mastino 1995: 68–70). 25  Cf. supra, amphora of the type AE3 recovered in the Sabaudo Dock. 26  A further two examples have been recovered in the harbour of Nora and in the bay of Capo Malfatano (Sanna and Soro 2013: 775–777). 27  In the castrum of Novae, in the Danubian sector, Kapitän II are documented for the early 4th century (Dyczek 2007: 828). 23 24

459

Ignazio Sanna, Laura Soro and Cristina Nervi

as ‘Phoenician Levantine amphora’ by Paul Reynolds, the result of the long typological-formal evolution of Phoenician amphorae produced over a limited period of time of about a hundred years, until the first half of the 3rd century AD. The evidence is restricted almost exclusively to the eastern area (Regev 2004: 345–346; Reynolds 2005: 570, 599, fig. 89; 2010: 90–91).28 The sample from Cagliari does not present traces of resin, so we may suppose the transportation of other products, such as fruit, cannot be ruled out, either fresh or preserved, as attested in other Levantine amphorae. The finding of a spatheion in the Eastern sector of the port of Cagliari, probably of Hispanic production, features in the same location context as a specimen form Keay 25.3 (Figure 3.9). The A spatheion29 (Figure 3.9) from the Eastern sector of the port of Cagliari, probably of Hispanic production, is comparable with the productions attested on the Iberian peninsula, in the area of Mazarrón, particularly in El Mojón and Águilas (Padilla Monge 2001: 400; Bernal Casasola 2008: 32), where some kilns have been identified in recent decades (Berrocal Caparrós 2012). The production, dated in the 4th-5th century30, is characterised by the typical deep incision on the neck and shoulder and by a profile with a rounded edge. The presence of the spatheion identified, corresponding to the type El Mojón I (Hernández García and Pujante Martínez 1999), appears significant also in the light of the small quantity of finds reported, not only in Sardinia, but also outside of the area of production.31 Regarding the contents of these spatheia, as well as wine products, the transportation of salsamenta cannot be ruled out, considering the proximity of the kilns to the centres where the products were made.32 Equally rare,  Considering the rarity of the specimen, which is in the study phase, we present here a preliminary macroscopic description of the fabric, which has a matrix with a large quantity of black and dark grey sharpedged microlithics, quartz and white and gilded mica. There are nodules of calcite and carbonate skeleton microfossils present and impression pores due to the thermal decomposition of the microfossils during baking (Ardizzone 2001). The temper consists of inclusions of polished black volcanic origin and abundant translucent quartz (cf. with petrographic analyses conducted on some samples of Lebanese amphorae in D. Williams (Opaiţ and Paraschiv 2012: 123; Freed 1994; Bonifay 2004: 458–459). 29  The dimensions are very reduced (little more than 40cm of residual height; diameter 8.1cm; wall thickness approx. 7mm), containing traces of resin inside very much altered by the location; the fabric is hard and compact, characterised by the presence of pale opaque quartz, calcite and a fair quantity of gilded mica of small dimensions. 30  The spatheia recovered in Hispalis, Seville, have been identified in contexts dated in the first quarter of the 6th century (Amores Carredano et al. 2007: fig. 5, 32). 31  There are extensive examples in the port area of Mazarrón (Berrocal Caparrós 2012: 272), while on a regional scale their diffusion seems to fall. Furthermore few specimens have been reported coming from the Central Mediterranean Basin: one from the deposit in Pupput, in the Gulf of Hammamet, in a context dated in the 5th century, and two from the marine area overlooking the site of Skerki Bank, between Sicily and Tunisia (Freed 1994; Bonifay 2004: 458–459). 32  Regarding the analyses conducted on organic residues extracted from some amphorae from the production centre of Águilas, see Quevedo Sternberg and Hernández García in this volume. 28

both underwater and on land, is the recovery from the same sector where the Hispanic spatheion originates of a perfectly intact Keay 52, with abundant resinous residues inside, characterised by a rim with a slightly flared lip and a circular section (Figure 7.6). Starting from mid-4th century, wines produced in the CalabrianSicilian countryside are widely distributed, destined for trade on a vast scale and to be placed alongside African commodities, by now insufficient to meet the new demographic needs of the Pars Orientis following the foundation of Constantinople (Romei 2004: 279–282; Malfitana et al. 2008: 147; Panella et al. 2010: 66). Thus we see the significant marketing of these containers in Roman and North-African contexts and in the western provinces (Bonifay and Pieri 1995: 115–116; Pacetti 1998: 187, 203; García Vargas and Vázquez Paz 2006; Bonifay and Capelli 2013; Casalini 2014: 273). Thanks to recent archaeometric data, the amphorae being examined, datable to around the end of the 5th century through comparisons with some examples identified in Termini Imerese, show macroscopic33 and morphological characteristics referable to the areas of production located along the north-eastern coasts of Sicily (Casalini and Crespi 2010; Casalini 2014: 275, 277; Franco and Capelli 2014a: 349; 2014b: 550). Starting from the 4th century, in parallel with the increase in wine production in Southern Italy and Sicily (Malfitana et al. 2008), in the eastern provinces a system of agricultural production was developed to cater for the massive demographic increase that followed the foundation of Constantinople. There thus occurred, de facto, a production of surpluses to be used for exports to the West (Reynolds 1995; Panella 2001: 180; García Vargas 2011), alongside the still widespread African goods,34 as documented by the wrecks of Yassi Ada 1 (Bodrum) and La Palud (Port-Cros), possessing very heterogeneous cargos with amphorae and other African products of the 7th century together with the eastern productions (Long and Volpe 1998; Van Doorninck 2002). The dynamism of the economic framework described above does not seem to suffer any lulls with the conquest of North Africa by the Vandals (Panella 2001; Palmieri 2008: 1082). In the case of Sardinia, also under Vandal rule for around eighty years (456-533) (Pani Ermini 1987), the socio-political factor ended up causing a decrease in economic functionality as regards Rome’s  A coarse and not very hard fabric, rough surfaced, externally of a non uniform colour ranging from dark to light brown, due to the conditions of the location, brick-orange colour in section. The porosity is widespread due to the presence of vacuoles deriving from the manufacturing process or caused by the presence of calcite granules, which disappear in baking (Ardizzone 2001). Abundant temper, with a prevalence of translucent and opaque quartz, with sub-rounded edges and conchoid fractures, gilded mica and muscovite, not widespread and of small dimensions. 34  We remind readers of the presence of early forms of LRA1, 2 and 4 in contexts 3 and 4 of Leptis Magna, recovered together with abundant attestations of African amphorae (Bonifay and Capelli 2013). 33

460

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

Figure 3. The Late Empire and the Late Ancient Period amphorae from central-eastern sector of the port of Cagliari. 1-2. Tripolitana III; 3. Africana IIA; 4-5. Keay 25; 6. Keay 52; 7-8. Catanian MRA 1 amphora; 9. Hispanic spatheion; 10. Kapitän II; 11. Tyrian amphora; 12. LRA 1; 13. LRA 2; 14. LRA 2/13 (draw by L. Soro; 10-11: S. Fanni); 15. Pseudo-Cos ‘en cloche’ amphora recovered in the locality of Sant’Elia (Cagliari) (photo L. Soro); 16. Tripolitana III amphora with residues of resin in the rim and neck; 17-19. Some amphorae in their underwater context; 20. Keay 62 from Cape Malfatano (Teulada), Southwestern Sardinia, amphora form (Photo L. Soro).

461

Ignazio Sanna, Laura Soro and Cristina Nervi

restocking of supplies;35 in contrast, for the decades under Vandal rule, an increase is recorded in the importance of the island in the supply of wheat, minerals, metals and animals destined for Carthage (Cosentino 2002: 1). With the Justinian renovatio, Sardinia fell politically within the Eastern Byzantine orbit but the fact that it was located geographically at the westernmost extremity of the Empire could have determined a particular development in the dynamics of the circulation of goods. As confirmation of the vitality of the island,36 the underwater investigations inside the port of Cagliari have also provided some fragmentary samples of amphorae of Eastern production: an LRA1b, from the 6th-7th century (Figure 3.12) one of the most attested forms in Mediterranean contexts starting from the 5th and until the 7th century37 (Reynolds 2005; Long and Volpe 1998; Bonifay and Pieri 1995: 107–108; Pieri 2007a; García Vargas 2011), the origin and production of which is documented along the coast of Syria and Cilicia, as well as in Caria and in some islands of the Aegean (Demesticha 2003; Opaiț 2004: 294–295; Pieri 2007b: 614–616; Reynolds 2005: 565–566). We must also highlight an LRA 2 characterised by handles pushed in towards the shoulder with an elliptical profile; observations under the microscope have enabled us to note residues of resining among the signs of internal turning (Figure 3.13).38 Among the amphorae of Eastern origin documented, only one possesses epigraphic apparatus, graffiti, namely globular amphora LRA 13 (Figure 3.14).39 Along the shoulder there appears a scratched inscription in Greek, which is incomplete due to the fracture: + (cross) A — perhaps with a connection that indicates the letter Γ (?) — Λ, Ο, Ι [ …]. Other signs are also visible, very abraded and difficult to understand, scratched along the upper portion of the shoulder and another cross in the proximity of the attachment of a handle. On the interpretation of the inscription,40 which is in the study phase, and of which we are providing preliminary information, there are various hypotheses that could be put forward; of the many, perhaps the most convincing could be associated with the content of the amphora, perhaps of a fatty and resinous nature — ἈΛΟΙ[ΦΉ]. Continued studies will confirm or deny this hypothesis.  Attestations of the corporationes navicularii Turritani and of the navicularii et negotiantes Karalitanis, who handled the loading and transporting of commodities, appear on the mosaics of the port of Ostia (CIL XIV 4549, 19 and 21; Meloni 1987: 252). 36  For an in-depth analysis of influences, imports and artistic-Byzantine evidence, see the contributions by the late Prof. Roberto Coroneo (2011). 37  In Beirut it is already known from the 4th century, in its small forms (Reynolds 2010: 92–93). 38  From a morphological perspective, we propose reference to an amphora-like container found in Athens Agora Amphora P13468 (Opaiț 2014: 46) and an LRA 2 recovered in Argos (Ivantchik 2002: fig. 17, 122). 39  Presence of chamotte in the fabric. Some morphological details relate back to globular amphorae, classified as LRA 13 (Riley 1979: 231), the result of the evolution of LRA2s (Diamanti, Kouzeliç and Petridis 2014; Reynolds 2016). 40  There are obvious comparisons with Byzantine productions (Van Doorninck 1989; Reynolds and Pavlidis 2014). 35

The bridge between Southern Sardinia and the East could be identified in the African coastal sites, Carthage first and foremost, from where cargos of goods brought there from the East set off, together with the products of local workshops, all destined for the markets of the Provençale and Tarraconensis coasts.41 To confirm a maritime relationship between the East and Sardinia,42 fragments of LRA 4 and LRA13 at Capo Malfatano are significant, in the south western sector of Sardinia, recovered together with a Keay 62A (Sanna and Soro 2013) (Figure 3.20), an LRA1343 and an LRA 5 in Nora (Solinas and Sanna 2005) (Figure 4.13). L.S. 3. Nora Nora, as a seaside town, was certainly integrated into the movements of goods and exchanges in the Mediterranean, which reached Southern Sardinia and probably revolved around the port system of the nearby Carales, of which it was a trade and economic partner. The settlement system at Nora had an urban centre and agricultural area, which extended inland from the town, where various sites are located dedicated to the exploitation of the territory’s multiple resources: the abundance of water, favourable to agriculture and the breeding primarily of sheep and goats, then later pigs and cattle; the presence of protrusions of rock, which enabled intense mining activities and the evidence of minerals, associated with the useful and fruitful working of metals. Such a synergy allowed the development of the town over the centuries, forming the raison d’être of the urban centre, but there followed a sometimes autonomous development, not always governed by the requirements of the inhabited centre on the coast to which it was related. For the reasons briefly listed above, the waters surrounding the port have provided data and materials relating to a very extensive period of time, from the 8th century BC to the contemporary age. The finds are referable to wrecks, probable wrecks, to materials associated with the manoeuvres for loading and unloading goods, which were undoubtedly carried on small boats, or were taken to the town using mobile — or fixed — gangways to reach the ships (Sanna 2016). There is a completely different marine context for the zones neighbouring the current area of settlement, where the data gathered in the current sea zones suggest buildings or contexts that at one time  Ballet, Bonifay and Marchand 2012. On this subject, is the recovery, in the same area of the port of Cagliari from which LRAs 1 and 2 originate, of a Sigillata Africana D dish, almost complete, form Hayes 104A, with a painted diamond-shaped cross at the centre, on the lower sides and on the top portion of which three doves are shown (Carandini and Tortorella 1981: 94–95, tav. XLI, n. 9). 42  For evidence of Eastern amphorae on land in Sardinia, see Cisci 2006: 134–136 with the relative bibliography, in the area adjacent to the cemetery of Bonaria (Cagliari), with scratched globular amphorae from the 8th century (Mureddu 2002: 237–241). Recent finds of globular amphorae from the 8th-9th century originate from the excavations under the Bastion of Santa Caterina (Cagliari) Cisci et al. 2013). 43  Cf. infra. 41

462

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

were not submerged, belonging to the urban fabric of Nora. Interesting studies have been conducted by the Underwater Division of the Archaeological Heritage Service of Sardinia, in association with the University of Padua, and will be published in the near future (Bonetto and Sanna forthcoming). 3.1. The Underwater Sites (Figure 4.1) Underwater investigations, under way since 1996, have produced an impressive quantity of data. If we follow the chronological order, the most ancient data — 8th century BC — have been collected in the sea space south of the promontory of Coltellazzo (NRS31). From the Hellenistic Period — and particularly the 3rd century BC — are the Gr.-Ita. V wine amphorae recovered in the Western Bay (NRS787). We then move on to the Late Republican Age, in this case in the Eastern Gulf there are materials found from the 1st century BC (NRS841). We find Dressel 1A and 1B wine amphorae, in association with Lamboglia 2, and these could relate to the site already identified by Michel Cassien between the late 1970s and the first half of the 1980s, the data on which has recently been revisited (Bertelli 2014). The evidence identified in the arm of the sea south of the forum date back to the same century (NRS343). Slightly further south is a context from the 1st century AD (NRS844), and specifically from the second half of the century, associated with Iberian containers that transported salsamenta (Dressel 7-11) and oil (Dressel 20) with Haltern 70 of type C (Figure 4.2-3), dating back to the age of Tiberius or Claudius, when the market of the exports from Baetica had stabilised. The data seem to favour a wreck with a mixed cargo coming from the Atlantic zone of the Iberian Peninsula. Perhaps a similar wreck is that of Port Vendres II, the heterogeneous goods are similar to those from Nora (Colls et al. 1977). The site (NRS872) is from between the 1st and 2nd century AD, associated with Dressel 20 C oil amphorae from Baetica. Most of the sea contexts identified date back to the 2nd century; the data are not surprising if they are compared with the development of the occupation of the territory of Nora. In the agricultural area linked to the town, it is precisely the second half of the 2nd century AD that marked the peak in the number of farms and settlements – based on the data coming from the various ceramic classes, not only from the amphorae (Nervi 2016). The zone opposite the most ancient recoveries from the Phoenician Age (NRS31) is from the 2nd century AD, where, in addition to Dressel 9 and 20, an almost complete PE‑41 has also been found, produced in Ibiza, indicating a cargo that, though originating from Baetica, had passed through the Balearic Islands. The context of NRS864 includes Dressel 20D oil amphorae, associated with one of the same type, but with a smaller module — Dressel 20 parva (Figure 4.4). The latter is not very widespread, but in

Sardinia it is also attested among the deposits of the port of Olbia, along the eastern coast; also in the context are Beltrán IIA and IIB, variant B — together with Gauloise 4 and Dressel 2-4 wine amphorae from the Tarraconensis area. So here we find ourselves — presumably — dealing with a mixed cargo with South and North-Hispanic — Tarraconensis — and Gallic materials: the ship had travelled the route north, originating in the Southern Iberian Peninsula. The sector of NRS887 has provided the largest quantity of materials indicating two separate phases, one from the late 2nd‑early 3rd century AD and the other from the late 3rd‑early 4th  century  NRS887, situated — once again — in the western bay. The site reveals a very heterogeneous series of containers: Ibizan wine amphorae (PE 25) (Figure 4.6‑7), Baetica (Beltrán IIB‑B), for the context of 2nd-early 3rd century AD, while those of the 3rd‑4th century AD are the Lusitanian (Almagro 50 and Almagro 51C), Hispanic (Beltrán 72 — Figure 4.8 —, Dressel 20C), North African (Tripolitana III, African IIC — Figure 4.10 —, IID — Figure 4.11-12 —, Africana IIIA/ Keay XXV.1 — Figure 4.9 — and Dressel 30 — Figure 4.5 —), Gallic (Gauloise 4), associated with forms of African Cooking Ware (saucepans Hayes  181/Bonifay  3B) and African common ware. The data could indicate two heterogeneous cargos; in particular the later one would seem to consist of products coming from Baetica, Gallia, the Balearics and North Africa. For the discoveries relating to the Late Antique period, we find — at the same site as materials from the 1st century AD — in the eastern bay of Nora, a wine container of the type Pieri LRC2C (Pieri 2005: 88–89), recently referred to (e.g. by Paul Reynolds in a conference paper) as LR2/13. An LRA 5, the upper third of the vessel (Solinas and Sanna 2005: fig. 5b) (Figure 4.13), could come from the Palestinian-Egyptian area; it was recovered in the stretch of southern sea NRS31 already indicated for PE‑41 and the Phoenician material from the 8th century BC. 3.2. Containers with Organic Residues After this rapid overview, which introduces the data that will be discussed below, we move on to analyse those amphorae that have shown traces of tarring, or organic residues in the content. PE 25 Balearics — Claudian-Neronian Age — wine. Of the four samples recovered: two of which fragmentary — only the upper part has been conserved (Figure 4.6‑7) — and the other two conserve rim, neck and handles. Their resining enables us to associate them with the transportation of wine, but what is much more significant is their presence in the waters of Nora, which links them to a fragment discovered in the urban excavation (Piccardi 2003), and this type of amphora has rarely been identified outside of the Balearic Archipelago. A fragment of a later type — late 2nd-early 3rd century AD — was recovered during the survey on the territory of the town of Nora (Nervi and Piccardi 2013; Ramon Torres 2008: 264).

463

Ignazio Sanna, Laura Soro and Cristina Nervi

PE 41 Balearics — Claudian-Neronian Age — wine. The wine amphora PE 41 (site NRS844) is a rare sample in a marine context; its pale fabric, pink in colour, presents traces of a rope etched onto the external surface, when the body was not yet fully solidified (Ramon Torres 1991: 125–126): a decoration typical of Ibizan productions. Dressel 2-4 Tarraconensis — late 2nd century AD — wine. The amphora type is referable to the same site as Gauloise 4 (wreck NRS864); this wine amphora presents traces of tar and, considering the morphology of the handle and the characteristics of the fabric, could be from the zone of Tarragona. Gauloise 4 — 2nd century AD — wine. The fragments of this type (wreck NRS864) are incomplete and resined internally; they present very homogeneous and thin fabrics, with an ochre-chamois colour, similar to the productions of the Rhone Valley, datable in the 2nd century AD. The amphora type is only rarely present in the town of Nora (Piccardi 2003: 215) and finds in the territory associated with it (Nervi 2016). Dressel 30/Keay 1B — late 2nd-3rd century AD — wine. This amphora — deriving in its forms from Gauloise 4 — presents traces of resin in the fragment conserved at the site NRS 887. The container from Nora could have a proconsular origin, in view of the very grainy and porous grey-pink fabric. Africana IIIA/Keay XXV.1 — last quarter of the 2nd-4th century AD — wine. This type, present in the context NR887 (Figure 4.1), has numerous attestations (7 samples) of productions from Salakta (Nacef 2015: 48–50), or Nabeul: all resined. The analyses conducted by Marshall Woodworth (Woodworth et al. 2015), presented during the Congress, have enabled Keay XXV to be related to wine transportation. The type is also documented in the urban excavation in Nora (Piccardi 2003: 221) and is among the finds of the survey. Beltrán IIA – variants A and B — late 1st century AD — salsamenta. The amphorae present in the context NRS887 are resined and have fabrics attributable to the area of Cadiz, or Granada (García Vargas, de Almeida and González Cesteros 2011). Variant II A-A is also attested among the finds of the survey, as is the later type IIB-B. Africana IID — late 3rd-4th century AD — salsamenta. The type is attested in the site NRS887, untarred, with a fabric from the area of Sullecthum (Nacef 2015: 46–48) (Figure 4.11) and resined with an fabric attributable to the zone of Nabeul (Figure 4.12). We therefore set about analysing the — dubious — issue of the certain attribution of the amphora type to a sole content, or whether — in relation to different commodities — the same type of amphora could be used, or reused, for a number of products. There may not have been a direct correlation between product and form, but rather an adaptation of the form to suit different commodities containable in it.

Haltern 70C — Claudian-Neronian Age — wine, defrutum, olives. The type recovered in the context NRS844 is attributable to the variant Haltern 70C: an amphora which probably had a varied contents (Figure 4.2-3) (Carreras Monfort and Berni Millet 2012). This amphora was traded particularly in the Claudian-Neronian Period, as attested by the recoveries from Nora (Piccardi 2003: 224) and of its territory. During the survey containers were identified as type Haltern 70C, as well as — due to the fragmentary nature of the finds — others that were more generically like the amphora Haltern 70, without the possibility of further precision. Imports to the town of Nora -— generally — covered the entire Mediterranean basin. The Phoenicians were the first to make this site one of the points along the routes they took across the Mare Nostrum, importing and exporting all the types of merchandise then in fashion: wine and meat were two key elements of the imports-exports of the Colonial period. It could have been the Punics who conveyed the goods in the 4th-3rd centuries BC, as in the transitional phase of Romanization they continued to organise the routes and exchanges. So — perhaps — the wine amphorae Gr.‑Ita. V present among the underwater and land-based finds and the Gr.-Ita. VI found during the course of the survey could have arrived there, triggering a synergy between Μεγάλη Ἑλλάς (Megálē Hellàs) and Nora, a link that was to continue in a less pronounced way with imports of the Adriatic type Lamboglia 2, rarely attested in all contexts. The dawn of the Imperial Age marked an initial moment of floruit for Nora and so it was that the town and its agricultural area witnessed the marked presence of Dressel 1 wine amphorae, — in the variants A, B and C: a type that promises to be well documented among the finds of Nora and its territory. Tyrrhenian wine gained its place among the products imported by Nora between the Late Republic and the Empire, but Dressel 1, also South Hispanic/Baetica products were also present in the territory, alongside products from the Italic Peninsula. During the 1st century AD, among the Hispanic amphorae we must mention the Dressel 7-11, used for transporting products deriving from the processing of fish — present in Nora-town (Piccardi 2003: 216–217) and its territory —, as well as Haltern 70, which fitted well within the market on account of their multi-purpose use, as containers for wine, olives or defrutum, reaching the town and its territory. It is interesting to note the presence of Balearic amphorae — PE 25 and PE 41 — both in urban contexts and from the survey, confirming the data emerging from their presence in underwater contexts. Starting from the 2nd century AD, imports of materials coming from the most disparate areas increased. The allochthonous imported products reached Nora and its territory inside amphora types coming from the Iberian Peninsula — Dressel 14 for oil, Lusitania; Dressel 2-4 for wine,

464

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

Figura 4. Nora (Pula-Cagliari). 1. The underwater contexts discovered and areas of research (draw M. Piras and I. Sanna); Some amphorae recovered. 2-3. Haltern 70; 4. Dressel 20 parva; 5. Dressel 30; 6-7. PE-25; 8. Beltrán 72; 9. Keay 25; 10. Africana IIC; 11. Africana IID (with stamp); 12. Africana IID; 13. LRA 5 (draw. R. Arcaini; LRA 5: L. Ruvioli)

465

Ignazio Sanna, Laura Soro and Cristina Nervi

both Tarraconensis and South-Hispanic — and in a very wide range of containers coming from North Africa, covering a range that is both wide and interesting. It follows that the second half of the 2nd and the 3rd century AD mark the floruit of Nora, and in that context imports multiplied, opening it up to the dynamics of exchange; these exchanges can only be partially reconstructed, and deserve more research in future. The types attested in the survey suggest a high volume of commerce; here we will mention those attested from the 3rd4th century AD, which also recur in the underwater projects: Tripolitana III, Africana IIIA, IIC, IID, Dressel 30, comparable to the type Gauloise 4, present among the urban finds and those from the territory; such amphorae also bring with them all the controversy concerning types and their ‘copies/ imitations’, in other areas different from that of their origin. This is too extensive a discussion to be resolved here, so we will merely hint at it briefly. The Late Antique period marked an important increase in salted products coming from the Lusitanian area: Almagro 50 and Almagro 51C are present in urban contexts, from the survey and the sea, and demonstrate the exchanges of products from the processing of fish from the Atlantic to the centre of the Western Mediterranean, with Nora a part of this exchange process; beside these are the African containers, which — based on the data for Nora — never succeed in equalling the Portuguese, at least for the second half of the 4th century AD. Late Antiquity saw the increased arrival of goods from the eastern basin of the Mediterranean, present in Nora town and in its surrounding territory — LR 1, LR 5, LR 2/13 —; these are also found in underwater contexts, except till now the LR 1. The products coming from the Levantine areas, but also from the Aegean Islands and Argolis, reached Nora, perhaps stopping in an important storage port along the navigation routes. The end of the 7th century AD marked the definitive decline of the territory of Nora and the town itself, perhaps to the advantage of nearly Cagliari, the Episcopalian centre. The decline appears gradual, we could even say ‘physiological’. Starting from the 6th century AD we witness a gradual, slow abandonment of the settlements outside the town, while Nora town was already transformed, if we do not wish to use the debasing term ‘ruralised’. C.N. 4. Conclusions Although the data presented are preliminary and currently being studied, they are already valid starting points for reflections on the role played by Central-Southern Sardinia within ancient mercantile circuits (Sanna 2019). Through the finds and the relative contexts identified it is possible to reconstruct an evolution of economiccommercial strategies over time as a consequence of political-administrative and social choices in the period, which influenced the circulation of foodstuffs.

For the Hellenistic phase the data that have emerged indicate the passage into South Sardinia of wines from Magna Graecia and Campania-Latium, transported by their own carriers or by the Punics, with an initial flow that was not yet intense, often also combined with the circulation of Greek wines (Olcese 2010: 43). From the chronological perspective, the association between Gr.‑Ita. IV and V amphorae has already been confirmed in other contexts (Olcese 2010), so the Cagliari layer would represent a new and interesting example of the importation of wine from Ischia and Campania to Sardinia — perhaps not heading exclusively for Cagliari, but also for other western centres on the island, to then continue towards Corsica, the Balearics and Catalonia (Asensio i Vilaró 2010), and not excluding North Africa, particularly Carthage (Bechtold and Docter 2010). Transportation of wine by sea changed significantly in the Republican Period, with a further development in the late phase, in concomitance with a ‘non limitation’ of food cargos and the Romanization of the new provinces (Sanna 2019). The contexts of the port of Cagliari (Sabaudo 1) and Villasimius belong to the crescendo of traffic and exchanges between the east and west of the Mediterranean, with cargos associated with other important commodities and coinciding with the socio-political changes and the different Apulian-Adriatic and Tyrrhenian production standards of the Italic peninsula (Manacorda, Olcese and Patterson 1995). From the analysis of the batch of amphorae from the Sabaudo 3 context (1st-2nd century AD), it is interesting to note the heterogeneity of the commodities transported and the containers, with fish derivatives from Baetica (in the types Dr. 7-11) and Tarraconensis, Italic (in Dr. 2-4), Aegean (Rodie) and Eastern European (AE3) wines. Different from the various Provençale and Corsican wrecks (Sciallano and Liou 1985),44 the proportional relationship between the containers used for fish byproducts and those for wine is almost equal. On the subject of the possible locations chosen for the trading of commodities, the attestations of the port of Cagliari and its gulf, which also include other rare yet significant Eastern types (‘Pseudo Cos’ en cloche/Agorà M54) could indicate a flow of western amphorae and the relative contents towards the markets of the Near East and in return the transportation, albeit minimal, of Eastern products (Soro and Sanna 2020).45 The city of Carthage could presumably have been a point of collection on these routes or a stage in the itineraries towards the east. The character of heterogeneity also emerges from the analysis of the later contexts, where we find possible associations, verifiable through future investigations, of amphorae of wine and oil/salsamenta  For Pascual 1 amphorae, together with Dr. 2-4, see also the wreck SudLavezzi 3 (Sciallano and Liou 1985). 45  The presence of Dr. 2-4 amphorae and the Hispanic ones Dr. 7-11 and 9 is attested in Egypt, even if type AE3 and the consumption of local wine is predominant (Şenol 2007). 44

466

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

produced in North Africa, with G4, Dr. 20 from Baetica and, also in this case, together with Eastern specimens (Kapitän II and an amphora produced in Tyre) that are very rare in the South-Western Mediterranean. As regards 1st-3rd century, Gallic wine, with the G4s, was distributed in the entire central-southern sector of the island, in association with other types of contemporary non-Gallic containers. For the Late Antique Period, the distribution of materials requires further stratigraphic investigation in order to define the associations of materials with greater precision. From the preliminary data it is possible that various cargos were present, certainly with a heterogeneous character — evidence of specific commercial strategies that, in the current state of the research, are also well documented in other excavations on land and undersea, such as that of Cabrera III (Balearics), where a we find a cargo of amphorae coming from Baetica, Lusitania and Africa (Bost et al. 1992). The hypothesis put forward in the case of the Balearic wreck is that of the loading of the ship on a single occasion, probably in a major port centre (not Cabrera, considered a secondary landing point for navigation along the major routes), perhaps Cadiz, where it was plausible that various goods and products converged to then travel together (Bost et al. 1992: 200–201; Bonifay 2007: 254). In the case of Cagliari, material datable in the 3rd century shows a significant similarity with the finds that formed the cargo of a wreck of Ognina Sud 1 (Sicily) (first half of the 3rd century), which transported Dressel 20 amphorae, together with numerous samples of Africana I, Tripolitana I‑III and amphorae of type Kapitän I and II (La Fauci 2002). For the later phase, the types of amphorae produced in the Calabrian-Sicel area makes it possible to examine a possible stopping point for a cargo coming from the Near East, passing through Sicily and heading east. Again, LRA 1, 2 and 13, datable to the 6th7th century, recovered, together with a Hayes 104A plate, confirm the centrality of Cagliari in the first centuries of the Byzantine dominion. In the case of Nora, as for Cagliari, even though there is still much undiscovered archaeological potential, the results obtained to date highlight a chrono-typological heterogeneity of materials, correlated with products transported and zones of origin. From the data there emerges a marked prevalence of the Hispanic provinces, shared evenly between Atlantic and Mediterranean. Into this commercial flow, which seems to have proceeded for many centuries, the Balearic containers are included as a true innovation. The statistical value increases because amphorae PE-41 and PE-25 have been identified in various areas of the sea around Nora, linking them to different episodes and times, associated in all cases with other containers and finds of Hispanic origin. The most recent investigations are unearthing heterogeneous contexts, between the late 2nd and early 4th century AD, representing not only direct routes towards Rome but also towards the western sectors of the Mediterranean: the Hispanic and Gallic coasts.

In order to reconstruct the itineraries of the most widespread commodities, such as wine, oil, fish derivatives, fruit, etc., it is useful to also consider other factors which may affect the economy of a society and, opportunistically or in a predetermined way, the choices as regards the route and the possible stages along it. Commodities of various kinds, for example, undoubtedly influenced the choice of stages, as well as the various types of vessels used for transportation. Consider the products that are difficult to recover in underwater contexts, but are attested in the ancient sources, such as hides and wool, or other types of food products such as walnuts, hazelnuts, grape pips, olea stones, as in the case of the Late Republican wreck Sabaudo 1 (Figure 1.15). Another type of material certainly influenced the system of importation and circulation of the most significant foodstuffs: metals (Bigagli 2002). The exchange of wines, first from Apulia, then Tyrrhenian, using Lamboglia 2 and Dressel 1, with the metal ingots produced in Carthago Nova (Pérez Balester and Pascual Berlanga 2004), also involved Sardinia, the land of mineral resources, which since the Bronze Age exchanged copper ingots, tin and lead, by sea in the East and West Mediterranean. In most of Central-Southern Sardinia, the development of research and studies on these themes can be further advanced through more underwater archaeological investigations or deeper studies. Almost all the contexts presented are the result of preliminary investigations or are works in progress. This potential is still present in the shallow seabeds in the proximity of the Sardinian coasts and in the large internal lagoons, beneath mud and sand, discoverable due to accidental natural circumstances, due to human intervention or through targeted and organised investigations, such as those presented here.

Bibliography Amores Carredano, F., García Vargas, E. and González Acuña, D. 2007. Ánforas tardoantiguas en Hispalis (Sevilla, España) y el comercio mediterráneo. 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 (I): 133–146. Oxford, Archaeopress. Ardizzone, F. 2001. Rapporti commerciali tra la Sicilia occidentale ed il Tirreno centromeridionale nell’VIII secolo alla luce del rinvenimento di alcuni contenitori da trasporto. In G. P. Brogiolo (ed.), Atti del II congresso nazionale della S.A.M.I. (Brescia, 28 settembre-1 ottobre 2000): 402–407. Florence, All’Insegna del Giglio. Asensio i Vilaró, D. 2010. El comercio de ánforas itálicas en la Península Ibérica entre los siglos IV y I a.C. y la problemática en torno a las modalidades de producción y distribución. Bollettino di Archeologia on line I, viewed 31 January 2016. .

467

Ignazio Sanna, Laura Soro and Cristina Nervi

Ballet, P., Bonifay, M. and Marchand, S. 2012. Africa vs Aegyptus: routes, rythmes et adaptations de la céramique africaine en Égypte. In S. Guédon (ed.), Entre Afrique et Égypte: relations et échanges entre les espaces au sud de la Méditerranée à l’époque romaine. Scripta Antiqua 49: 87–117. Paris, de Boccard. Bechtold, B. and Docter, R. 2010. Transport amphorae from Punic Carthage: an overview. In L. Nigro (ed.), Motya and the Phoenician ceramic repertoire between the Levant and the West, 9th-6th century BC. Proceedings of the International Conference held in Rome, 26th February 2010. Quaderni di archeologia fenicio-punica 5: 85–116. Rome, Sapienza Università di Roma. Bernal-Casasola, D. 2008. El final de la industria pesqueroconservera en Hispania (ss. V-VII d.C.). Entre obispos, Bizancio y la evidencia arqueológica. In J. Napoli (ed.), Ressources et activités maritimes des peuples de l’Antiquité. Actes du colloque international de Boulogne-surMer (12, 13 et 14 mai 2005). Les Cahiers du Littoral 2 (6): 31–57. Boulogne-sur-Mer, Centre de Recherche en Histoire Atlantique et Littorale, Université du Littoral Côte d’Opale. Berrocal Caparrós, M. C. 2012. Producciones anfóricas en la costa meridional de Carthago-Spartaria. In D. Bernal-Casasola and A. Ribera (eds), Cerámicas hispanorromanas II. Producciones regionales: 255–277. Cadiz, Universidad de Cádiz. Bertelli, A. 2014. Catalogo dei rinvenimenti. In J. Bonetto (ed.), Nora e il mare. I. Le ricerche di Michel Cassien (1978‑1984). Scavi di Nora IV: 473–514. Padova, Padova University Press. Bigagli, C. 2002. Il commercio del piombo ispanico lungo le rotte attestate nel bacino occidentale del Mediterraneo. Empùries 53: 155–194. Bonetto, J. and Sanna, I. (eds) forthcoming. Nora e il mare. II. Padova, Padova University Press. Bonifay, M. 2004. Études sur la céramique romaine tardive d’Afrique. British Archaeological Reports International Series 1301. Oxford, Archaeopress. Bonifay, M. 2007. Cargaisons africaines : reflet des entrepôts? Antiquités africaines 43: 253–260. Bonifay, M. and Capelli, C. 2013. Les thermes du Levant à Leptis Magna : quatre contextes céramiques des IIIe et IVe siècles. Antiquités africaines 49: 67–150. Bonifay, M. and Pieri D. 1995. Amphores du Ve au VIIe s. à Marseille : nouvelles données sur la typologie et le contenu. Journal Roman Archaeology 8: 94–120. Botto, M. 2011. Interscambi e interazioni culturali fra Sardegna e Penisola Iberica durante i secoli iniziali del I millennio a.C. In M. Álvarez Martí-Aguilar (ed.), Fenicios en Tartesos: nuevas perspectivas. British Archaeological Reports International Series 2245: 33– 67. Oxford, Archaeopress. Buora, M., Carre, M.-B., Tiussi, C. and Ventura, P. 2007. Bolli su anfore Lamboglia 2 o simili dall’area aquileiese. Rei Cretariæ Romanæ Fautorum Acta 40: 285–304. Capelli, C. and Contino, A. 2013. Amphores tripolitaines anciennes ou amphores africaines anciennes  ? Antiquités africaines 49: 199–208.

Carreras Monfort, C. and Berni Millet, P. 2012. Haltern 70 (Valle del Guadalquivir). In Amphorae ex Hispania I, Paisajes de producción y consumo, viewed 31 January 2016. . Casalini, M. 2014. Anfore di piccole dimensioni a fondo piatto dell’Italia meridionale e della Sicilia. Alcune riflessioni a partire dalla documentazione romana. Rei Cretariæ Romanæ Fautorum Acta 43: 271–278. Casalini, M. and Crespi, M. 2010. Anfore tardoantiche di piccole dimensioni a fondo piatto dalle pendici nord-orientali del Palatino. Nuovi dati alla luce di un riesame tipologico e petrografico. Rei Cretariæ Romanæ Fautorum Acta 41: 1–11. Cibecchini, F. and Capelli, C. 2013. Nuovi dati archeologici e archeometrici sulle anfore grecoitaliche: i relitti di III secolo del Mediterraneo occidentale e la possibilità di una nuova classificazione. In F. Olmer (ed.), Itinéraires des vins romains en Gaule, IIIe-Ier siècles avant J.-C., confrontation de faciès. Actes du colloque européen organisé par l’UMR 5140 du CNRS, Lattes, 30 janvier - 2 février 2007: 423-452. Lattes, CNRS. Cisci, S. 2006. Contenitori per la conservazione ed il trasporto (VI a.C. - VIII d.C.). In R. Martorelli, R. and D. Mureddu (eds), Archeologia urbana a Cagliari. Scavi in Vico III Lanusei (1996-1997): 112–136. Cagliari, Scuola Sarda Editrice. Cisci, S., Messina, M. G., Mureddu, D. and Tatti, M. 2013. Cagliari. Indagini archeologiche presso il Bastione di Santa Caterina. Campagna 2012-2013. In R. Martorelli (ed.), Settecento-Millecento: storia, archeologia e arte nei ‘secoli bui’ del Mediterraneo: dalle fonti scritte, archeologiche ed artistiche alla ricostruzione della vicenda storica: la Sardegna laboratorio di esperienze culturali. Convegno di studi (Cagliari, 17-19 ottobre 2012): 235– 348. Cagliari, Scuola Sarda Editrice. Colls, D., Étienne, R., Lequément, R., Liou, B. and Mayet, F. 1977. L’épave Port-Vendres II et le commerce de la Bétique à l’époque de Claude. Archaeonautica 1. Coroneo, R. 2011. Arte in Sardegna dal IV alla metà dell’XI secolo. Cagliari, AV. Cosentino, S. 2002. Potere e istituzioni nella Sardegna bizantina. In P. Corrias and S. Cosentino (eds), Ai confini dell’impero. Storia, arte e archeologia della Sardegna bizantina: 1–13. Cagliari, M&T. Dangreaux, B. and Desbat A., 1987. Les amphores du dépotoir flavien du Bas-de-Loyasse à Lyon. Gallia 45: 115–153. Del Vais, C. and Sanna, I. 2012. Nuove ricerche subacquee nella laguna di santa Giusta (OR) (campagna del 2009-2010). In M.G. Arru, S. Campus, R. Cicilloni and P. Ladogana (eds), Ricerca e confronti 2010. Giornate di studio di archeologia e storia dell’arte a 20 anni dall’istituzione del Dipartimento di Scienze Archeologiche e Storico-artistiche dell’Università degli Studi di Cagliari (Cagliari, 1-5 marzo 2010). ArcheoArte 1, suppl.: 201-233. Demesticha, S. 2003. Amphora production in Cyprus during the Late Roman Period. In C. Bakirtzis (ed.), VIIe Congrès International sur la Céramique Médiévale en Méditerranée (Thessalonique, 11-16 octobre 1999): 472– 474. Athènes, Ministère de la Culture.

468

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

Diamanti, C. Kouzeli, K. and Petridis, P. 2014. Archaeology and archaeometry in Late Roman Greece: the case of mainland and insular Settlements, workshop and imports. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. The Mediterranean: a Market without Frontiers. British Archaeological Reports International Series 2616 (I): 181–192. Oxford, Archaeopress. Dyczek, P. 2007. Late Roman Amphorae from the site of valetudinarium at Novae. 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 (I): 827–834. Oxford, Archaeopress. 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 recherche. Actes du colloque de Sienne (2224 mai 1986). Collection de l’École française de Rome 114: 223–248. Rome, École française de Rome. Franco, C. 2014. Sicilian Amphorae (1st-6th centuries AD): typology, production and trade. Unpublished PhD thesis, Oxford, Trinity Term. Franco, C. and Capelli, C. 2014. Sicilian flat-bottomed amphorae (1st-5th century AD). New data on typochronology and distribution and from an integrated petrographic and archaeological study. In D. Malfitana and G. Cacciaguerra (eds), Archeologia classica in Sicilia e nel Mediterraneo. Didattica e ricerca nell’esperienza mista CNR e Università. Ricerche di archeologia classica e postclassica II: 341–362. Catania, Consiglio Nazionale delle Ricerche. Freed, J. 1994. The Pottery from the Late-Roman Shipwreck. In A. M. Mc Cann and J. Freed (eds), Deep Water Archaeology: a Late Roman Ship from Carthage and an ancient Trade Route near Skerki Bank of Northwest Sicily. Journal of Roman Archaeology Supplementary Series 13: 21–48. Ann Arbor, Journal of Roman Archaeology. Freund, K. P. and Batist, Z. 2014. Sardinian Obsidian Circulation and Early Maritime Navigation in the Neolithic as Shown Through Social Network Analysis. Journal of Island & Coastal Archaeology 9: 364–380. García Vargas, E. 2011. Oriental Trade in the Iberian Peninsula during Late Antiquity (4th-7th Centuries AD): An archaeological perspective. In D. Hernández de la Fuente (ed.), New Perspectives on Late Antiquity: 76–118. Newcastle upon Tyne, CSP. García Vargas, E. and Bernal-Casasola, D. 2008. Ánforas de la Bética. In D. Bernal-Casasola and A. Ribera i Lacomba (eds), Cerámicas hispanorromanas. Un estado de la cuestión: 661–687. Cadiz, Universidad de Cádiz. García Vargas, E. and Vázquez Paz, J. 2006. Sevilla y el comercio transmarino en el Bajo Imperio y en la Antigüedad Tardía: el testimonio de la arqueología. In S. Ventura (ed.), La catedral en la ciudad (III). De Isidoro a Ad bar Rahman: 44–99. Sevilla, Aula Hernán Ruiz, Catedral de Sevilla.

García Vargas, E., de Almeida, R. R. and González Cesteros, H. 2011. Los tipos anfóricos del Guadalquivir en el marco de los envases hispanos del siglo I a.C. Un universo heterogéneo entre la imitación y la estandarización. SPAL, Revista de Prehistoria y Arqueología de la Universidad de Sevilla 20: 185–283. Garnier, N., Silvino, T., Bernal-Casasola, D., Tokarski, C. and Rolando, C. 2013. Comment identifier des traces d’huile d’olive dans des céramiques archéologiques ? In D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz and A. M. Saez (eds), Hornos, talleres y focos de producciόn alfarera en Hispania. Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (II): 487–497. Cadiz, Universidad de Cádiz. Hernández García, J. D. and Pujante Martínez, A. 1999. Excavación en C/ Juan Pablo I, esquina con C/ Castelar: termas orientales Águilas (Murcia). Anales de prehistoria y arqueología 15: 179–191. Ivantchik, A. I. 2002. Un puits d’époque paléochrétienne sur l’agora d’Argos. Bulletin de Correspondance Hellénique 126-1: 331–404. La Fauci, F. 2002. Nuove osservazioni sui relitti di Capo Ognina (Siracusa). In Archeologia subacquea. Studi, ricerche e documenti III: 335–349. Rome, Istituto Poligrafico dello Stato. Long, L. and Volpe, G. 1998. Origini e declino del grande commercio nel Mediterraneo tra età arcaica e tarda antichità. I relitti della Palud (Port-Cros, Francia). In L’Africa romana XI: 1235–1284. Sassari, Il Torchietto. López Mullor, A. and Martín Menéndez, A. 2008. Tipologia i datació de les àmfores tarraconenses produïdes a Catalunya. In A. López Mullor and J. Aquilué Abadías (eds), La producció i el comerç de les àmfores de la «Provincia Hispania Tarraconensis». Homenatge a Ricard Pascual i Guasch. Actes de les Jornades d’Estudi (17-18 de novembre de 2005). Monografies 8: 33–94. Barcelona, Museu d’Arqueologia de Catalunya. Lucejko, J. J., Del Vais, C., Sanna, I. and Colombini, M. P. 2012. Caratterizzazione di residui organici in ceramiche provenienti dalla laguna di Santa Giusta. Unpublished poster, A.I.Ar. 2012 Modena. VII Congresso Nazionale di Archeometria (Modena, 22‑24 febbraio 2012). Malfitana, D., Botte, E., Franco, C., Morgano, M. G. and Palazzo, A. L. 2008. Roman Sicily Project (‘RSP’). Ceramics and trade. A multidisciplinary approach to the study of material culture assemblages. FACTA. A Journal of Roman Material Culture Studies 2: 127–192. Manacorda, D., Olcese, G. and Patterson, H., 1995. Le anfore di Giancola (BR): archeologia, archeometria, storia. In G. Olcese (ed.), Ceramica romana e archeometria: lo stato degli studi: 277–284. Florence, All’insegna del Giglio. Mastino, A. 1995. La produzione ed il commercio dell’olio nella Sardegna antica. In M. Atzori and A. Vodret (eds), Olio sacro e profano: tradizioni olearie in Sardegna e Corsica: 60–76. Sassari, EDES Editrice Democratica. Mattingly, D.J. 1985. Olive oil production in Roman Tripolitania. In D. J. Buck, and D. J. Mattingly (eds), Town and Country in Roman Tripolitania. British Archaeological Reports International Series 274: 27– 46. Oxford, BAR Publishing.

469

Ignazio Sanna, Laura Soro and Cristina Nervi

Meloni, P. 1987. L’età imperiale. In S. F. Bondì, M. Brigaglia, A. Guillou, G. Lilliu, P. Meloni and L. Pani Ermini (eds), Storia dei Sardi e della Sardegna. Dalle origini alla fine dell’età bizantina, I: 235–262. Milan, Jaca Book. Mureddu, D. 2002. Cagliari, area adiacente il cimitero di Bonaria: un butto altomedievale con anfore a corpo globulare. In P. Corrias and S. Cosentino (eds), Ai confini dell’impero. Storia, arte e archeologia della Sardegna bizantina: 237–241. Cagliari, M&T. Nacef, J. 2015. La production de la céramique antique dans la région de Salakta et Ksour Essef (Tunisie). Roman and Late Antique Mediterranean Pottery 8. Oxford, Archaeopress. Nervi, C. 2016. Il paesaggio di Nora. British Archaeological Reports International Series 2833. Oxford, BAR publishing. Nervi, C. and Piccardi, E. 2013. Continuità e trasformazione di flussi commerciali tra Isole Baleari, Sardegna, Corsica ed Alto Tirreno in epoca romana. In L’Africa Romana XX. Momenti di continuità e rottura: bilancio di trent’anni di convegni: 2462–2474. Rome, Carocci. Olcese, G. (ed.) 2010. Le anfore greco italiche di Ischia: archeologia e archeometria. Artigianato ed economia nel Golfo di Napoli. Immensa Aequora 1. Roma, Quasar. Opaiț, A. 2004. The Eastern Mediterranean Amphorae in the Province of Scythia. 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: 293–308. Aarhus, Aarhus University Press. Opaiț, A. 2014. Defining more Roman amphora types from the Athenian Agora: too much history, too little typology (I). Rei Cretariæ Romanæ Fautorum Acta 43: 43–54. Opaiţ, A and Paraschiv, D. 2012. Rare amphora finds in the city and territory of (L)IBIDA (1st-6st centuries AD). Rei Cretariæ Romanæ Fautorum Acta 42: 113–124. 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 (Rome 1995): 185–208. Florence, All’Insegna del Giglio. Padilla Monge, A. 2001. Comercio y comerciantes en el mundo tardorromano en Málaga. In F. Wulff Alonso, G. Cruz Andreotti and C. Martínez Maza (eds), Comercio y comerciantes en la Historia Antigua de Málaga (siglo VIII a.C.- año 711 d.C.). Actas del II Congreso de Historia Antigua de Málaga: 385–417. Malaga, Centro de Ediciones de la Diputación de Málaga. Palazzo, P. 1989. Le anfore di Apani (Brindisi). In Amphores romaines et histoire économique. Dix ans de recherche. Actes du colloque de Sienne (22-24 mai 1986). Collection de l’École française de Rome 114: 548–553. Rome, École française de Rome. Palmieri, L. 2008. I Vandali e l’olio. Produzione e commerci nell’Africa del V secolo d.C. In L’Africa Romana XVII. Le ricchezze dell’Africa. Risorse, produzioni, scambi: 1081– 1090. Roma, Carocci.

Panella, C. 2001. Le anfore di età imperiale nel Mediterraneo occidentale. In P. Lévêque, J. P. Morel and É. Geny (eds), Céramiques hellénistiques et romaines III: 177–275. Paris, Presses Universitaires FrancComtoises. Panella, C. and Rizzo, G. 2014. Ostia VI. Le terme del nuotatore. Studi miscellanei 38. Roma, L’Erma di Bretschneider. Panella, C., Saguì, L., Casalini, M. and Coletti, F. 2010. Contesti tardoantichi di Roma: una rilettura alla luce di nuovi dati. In G. Guiducci, S. Menchelli and M. Pasquinucci (eds), LRCW 3. 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, Archaeopress. Pani Ermini, L. 1987. La Sardegna nel periodo vandalico. In S. F. Bondì.M. Brigaglia, A. Guillou, G. Lilliu, P. Meloni and L. Pani Ermini (eds), Storia dei Sardi e della Sardegna. Dalle origini alla fine dell’età bizantina, I: 297–327. Milan, Jaca Book. Pascual Guasch, R. 1977. Las ánforas de la Layetania, In G. Vallet (ed.), Méthodes classiques et méthodes formelles dans l’étude typologique des amphores. Actes du colloque de Rome, 27-29 mai 1974. Collection de l’École française de Rome 32: 47–96. Rome, École française de Rome. Pérez Ballester, J. and Pascual Berlanga, G. 2004. The adriatic amphora type L.2 recovered from the environment of Cartagena (Murcia, Spain). In M. Pasquinucci and T. Weski (eds), Close Encounters: Sea-and Riverborne Trade Ports and Hinterlands, Ship Construction and Navigation in Antiquity, the Middle Ages and in Modern Time. British Archaeological Reports International Series 1283: 27–37. Oxford, Archaeopress. Piccardi, E. 2003. Anfore. In B. M. Giannattasio (ed.), Nora. Area C, scavi 1996-1999: 209–236. Genova, Brigati. Pieri, D. 2005. Le commerce du vin oriental à l’époque byzantine (Ve-VIIe siècles). Le témoignage des amphores en Gaule. Bibliothèque d’Archéologie et d’Histoire 174. Beirut, Institut français du Proche-Orient. Pieri, D. 2007a. Béryte dans le grand commerce méditerranéen. Production et importation d’amphores dans le Levant protobyzantin (Ve-VIIe s. ap. J.-C.). In M. Sartre (ed.), Productions et échanges dans la Syrie gréco-romaine. Actes du 2e colloque international sur la Syrie antique (Tours, 12-13 juin 2003. Topoï, Supplément 8: 297–327. Paris, de Boccard. Pieri, D. 2007b. Les centres de production d’amphores en Méditerranée orientale durant l’Antiquité tardive: quelques remarques. 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 (I): 611–625. Oxford, Archaeopress. Ramon Torres, J. 1991. Las ánforas púnicas de Ibiza. Trabajos del Museo Arquelógico de Ibiza 23: 125–126.

470

Amphorae with residues from Southern Sardinia (Cagliari and Nora)

Ramon Torres, J. 1995. Las ánforas fenicio-púnicas del Mediterráneo central y occidental. Instrumenta 2. Barcelona, Universitat de Barcelona. Ramon Torres, J. 2008. Les àmfores altimperiales d’Ebusus. In A. López Mullor and J. Aquilué Abadías (eds), La producció i el comerç de les àmfores de la «Provincia Hispania Tarraconensis». Homenatge a Ricard Pascual i Guasch. Actes de les Jornades d’Estudi (17-18 de novembre de 2005). Monografies 8: 241–270. Barcelona, Museu d’Arqueologia de Catalunya. Regev, D. 2004. The phoenician Transport Amphora. 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: 337–352. Aarhus, Aarhus University Press. Revilla, V. 2007. Les amphores africaines du IIème et IIIème siècles du Monte Testaccio (Rome). In A. Mrabet and J. Remesal Rodriguez (ed.), In Africa et in Hispania. Études sur l’huile africaine. Instrumenta 25: 269–299. Barcelona, Universitat de Barcelona. Revilla, V. 2010. Ánforas africanas del primer cuarto del siglo III d.C. del Monte Testaccio (campañas 19981999). In . M. Blázquez Martínez and J. Remesal Rodríguez (eds), Estudios sobre el Monte Testaccio (Roma) V. Instrumenta 35: 397–413. Barcelona, Universitat de Barcelona . Reynolds, P. 1995. Trade in the Western Mediterranean, AD 400-700. The ceramic evidence. British Archaeological Reports International Series 604. Oxford, Tempus Reparatum. 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 Esparraguerra, J. Buxeda i Garrigós, and M. A. Cau Ontiveros (eds), LRCW I. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and Archaeometry. British Archaeological Reports International Series 1340 (II): 563–611. Oxford, Archaeopress. Reynolds, P. 2010. Trade networks of the East, 3rd to 7th centuries: the view from Beirut (Lebanon) and Butrint (Albania) (fine wares, amphorae and kitchen wares). In S. Menchelli, S. Santoro, M. Pasquinucci and G. Guiducci (eds), LRCW 3. 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): 89–114. Oxford, Archaeopress. Reynolds, P. 2016. From Vandal Africa to Arab Ifrīqiya. Tracing Ceramic and Economic Trends through the Fifth to the Eleventh Centuries. In S. T. Stevens and J. P. Conant (eds), North Africa under Byzantium and Early Islam. Dumbarton Oaks Byzantine symposia and colloquia: 129–172. Washington DC, Dumbarton Oaks Research Library and Collection.

Reynolds, P. and Pavlidis, E. 2014. Nicopolis (Epirus Vetus): an early 7th century pottery assemblage from the ‘Bishop’s house’ (Greece). In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and archaeometry. The Mediterranean: a Market without Frontiers. British Archaeology Reports International Series 2616 (I): 451–467. Oxford, Archaeopress. Riley, J. A. 1979. The Coarse Pottery from Berenice. In J. A. Lloyd (ed.), Excavations at Sidi Khrebish Benghazi (Berenice) II. Supplement to Libya Antiqua V (II): 91–467. Tripoli, Society for Libyan Studies. Romei, D. 2004. Produzione e circolazione dei manufatti ceramici a Roma nell’alto Medioevo. In L. Paroli and L. Vendittelli (eds), Roma dall’antichità al Medioevo II, contesti tardoantichi e altomedievali: 278–311. Milan, Electa. Russo, A. 2011. Loutéria, Thymiateria e bruciaprufumi. In M. Grimaldi et al., La casa di Marco Fabio Rufo. Lo scavo dei giardini e i materiali. In Fasti OnLine Documents & Research 117: 25–28. . Salvi, D. and Sanna, I. 2000. L’Acqua e il Tempo. Prospezioni di archeologia subacquea nelle acque di Gonnesa. Cagliari, GIA. Sanna, I. 2016. La marina di Nora in età romana: i reperti subacquei quali indicatori di contatti e scambi commerciali. In S. Angiolillo, M. Giuman, R. Carboni and E. Cruccas (eds), Nora Antiqua. Atti del Convegno di Studi (Cagliari, 3-4 ottobre 2014). Scavi di Nora V: 3–14. Perugia, Morlacchi University Press. Sanna, I. 2019. Approdi e traffici transmarini nella Cagliari punica. I dati della ricerca archeologica subacquea. In R. Martorelli (ed.), Know the sea to live the sea. Conoscere il mare per vivere il mare. Atti del Convegno di Studi (Cagliari, 7-9 marzo 2019): 273–294. Perugia, Morlacchi University Press. Sanna, I. and Soro, L. 2013. Nel mare della Sardegna centro-meridionale tra 700 e 110 d.C. Un contributo dalla ricerca archeologica subacquea. In R. Martorelli (ed.), Settecento-Millecento. Storia, archeologia e arte nei ‘Secoli Bui’ del Mediterraneo (Cagliari, 17-19 ottobre 2012): 761–807. Cagliari, Scuola Sarda Editrice. Sanna, I., Arcaini, R. and Fanni, S. forthcoming. Rapporti commerciali tra penisola italica e iberica attraverso i contesti subacquei repubblicani nella Sardegna centro meridionale. In H. Uroz Rodríguez and A. Ribera i Lacomba (eds), Cultura material romana en la Hispania republicana. Atti Congreso Internacional de Arqueología. Lezuza (22-24 Abril 2016). Sanna, I., Le Bourdonnec, F. X., Poupeau, G. and Lugliè, C. 2010. Ossidiane non sarde in Sardegna. Analisi di un rinvenimento subacqueo nel Porto di Cagliari. In C. Lugliè (ed.), L’ossidiana del Monte Arci nel Mediterraneo. Atti del V Convegno internazionale (Pau, 27-29 Giugno 2008): 99–119. Ales, NUR. Sciallano, M. and Liou, B. 1985. Les épaves de Tarraconaise à chargement d’amphores Dressel 2-4. Archaeonautica 5: 5–178.

471

Ignazio Sanna, Laura Soro and Cristina Nervi

Sciallano, M. and Sibella, P. 1991. Amphores : comment les identifier ? Aix-en-Provence, Edisud. Şenol, A. K. 2007. A Statistical Essay on the Distribution of Imported Amphorae Finds of the CEAlex Salvage Excavations. In S. Marchand and A. Marangou (eds), Amphores d’Égypte de la Basse Époque à l’époque arabe. Cahiers de la Céramique égyptienne 8 (I): 57–75. Solinas, E. and Sanna, I. 2005. Nora: documenta submersa. In B. M. Giannattasio, C. Canepa, L. Grasso and E. Piccardi (eds), Aequora, póntos, jam, mare… Mare, uomini e merci nel mediterraneo antico. Atti del Convegno internazionale (Genova, 9-10 dicembre 2004): 253– 257. Florence, All’Insegna del Giglio. Soro, L. 2019. L’approdo portuale di Cagliari in età tardoantica e bizantina: traffici commerciali e relazioni di scambio. In R. Martorelli (ed.), Know the sea to live the sea. Conoscere il mare per vivere il mare. Atti del Convegno di Studi (Cagliari, 7-9 marzo 2019): 273– 294. Perugia, Morlacchi University Press. Soro, L. and Sanna, I. 2020. Merci e approdi nella marina di Cagliari: il quadro archeologico subacqueo. In R. Martorelli and D. Mureddu (eds), Archeologia urbana a Cagliari. Scavi nella chiesa di Sant’Eulalia alla Marina. Il quartiere dalle origini ai giorni nostri: status quaestionis all’inizio della ricerca: 177–194. Perugia, Morlacchi University Press.

Swan, V. G. 2004. Dichin (Bulgaria) and the supply of amphorae to the Lower Danube in the late Romanearly Byzantine period. 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: 371–382. Aarhus, Aarhus University Press. Ucchesu, M., Orrù, M., Grillo, O., Venora, G., Usai, A., Serreli, P. F. and Bacchetta, G. 2015. Earliest evidence of a primitive cultivar of Vitis vinifera L. during the Bronze Age in Sardinia (Italy). Vegetation History and Archaeobotany 24: 587–600. Van Doorninck Jr., F. H. 1989. The cargo amphoras on the 7th century Yassi Ada and 11th century Serçe Limani shipwrecks: two examples of a reuse of byzantine amphoras as transport jars. In V. Déroche and J. M. Spieser (eds), Recherches sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 247–257. Athens, École française d’Athènes. Van Doorninck Jr., F. H. 2002. Byzantine shipwrecks. In A. E. Laiou (ed.), The Economic History of Byzantium: from the Seventh through the Fifteenth Century. Dumbarton Oaks studies 39: 899–905. Washington DC, Dumbarton Oaks. Woodworth, M., Bernal, D., Bonifay, M., Garnier, N., Keay, S., Pecci, A., Poblome, J., Pollard, M., Richez, F. and Wilson, A. 2015. The content of African Keay 25/Africana 3 Amphorae: Initial results of the CORONAM Project. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 41–57. Esposende, Município de Esposende.

472

‘De profundis’: three amphorae of unorthodox contents retrieved from the Aegean Sea George Koutsouflakis

Ephorate of Underwater Antiquities Hellenic Ministry of Culture and Sports

Abstract: The question of the contents of Roman amphorae has apparently never been in the forefront of Aegean amphora studies, neither has the investigation of Roman shipwrecks. In contrast to what happened in the central and western Mediterranean, underwater excavations in the Aegean have been mostly motivated by assemblages of the previous Greek or the following Byzantine periods. While a great number of Roman amphora wrecks have been spotted and preliminarily documented during surveys conducted over the last fifteen years, the study of amphorae has been limited exclusively to typological or petrological issues in the very best cases. Any speculations put forward on the question of amphora contents, chiefly depend on observations and conclusions based on similar or related assemblages found in the rest of the Mediterranean. However, despite the relative absence of generalized data, several isolated finds from the sea provide ample information about some less common, or even exceptional commodities, shipped and transported into new or reused amphorae. The aim of this paper is to present a short overview of the present state of knowledge on the contents of amphorae raised from the Aegean archipelagos and to focus on three exceptional cases which unveil some less ordinary contents. Key words: amphora contents; amphora studies; Aegean; Mediterranean trade; murex; Galathea stringosa; Phournoi, Makronisos.

1. Introduction1 As noted elsewhere in the past, shipwrecks provide evidence for the advance of archaeological studies that no other terrestrial or underwater site could possibly present.2 They offer a unique opportunity to trace the nexuses between trade, trading routes, commercial strategies, shipping and ship‑management. The potentiality of studies focused on associated groups of artifacts within a wider ‘closed find’ reveals another substantial aspect of their value. Yet, no doubt, the most widely pronounced quality of shipwrecks — if not the most important one — is the large range and quality of intact artifacts they provide, especially clay containers. After the investigation of the first shipwrecks, the study of amphorae was catapulted in less than a decade from a self‑limited pastime to a new discipline. The intactness and ample number of amphorae located in underwater cargoes, allowed the advance of specialized studies, focused on the particulars of production. Typologies of amphorae along with the study of stamps, inscriptions, stoppers, coatings and contents were seriously influenced and enriched by the input of shipwrecks.  During the writing of this paper I received valuable help from a number of colleagues and specialists. I have benefited immensely from conversations with A. Sáez Romero, D. Mylona, A. Opait, Th. Webb and N. Boichot to whom I’m grateful. I am grateful as well to Chr. Avronidaki for proofreading patiently my manuscript at different stages of preparation. 2  Parker 1992: 2–4; Adams 2001; Gibbins and Adams 2001. For the outstanding contribution of shipwrecks in amphora studies see more specifically Fr. Cibecchini in the present volume. 1

During the early decades of systematic archaeological research on ancient shipwrecks, defining the content of an amphora was not considered as a substantial issue. On the contrary, contents were considered either as self‑evident from the shape of the vessel itself, the staple products of the region from where an amphora was manufactured or were somewhat arbitrarily concluded from excerpts of literary sources or pictorial art.3 A more systematic approach was adopted from the 1980s onwards, when the study of amphorae also extended to the description and analysis of residues found in their interiors, as remains indicating contents, coating or preservatives. The introduction of better sampling methods by G. Bass (1982: 163–165) resulted in a series of new data, regarding mostly the presence of botanical remains. As a result, archaeologists became less careless when examining the debris found inside intact amphorae and started to adopt more strict protocols. The advance of sampling methods and the introduction of multidisciplinary studies applied into this new field, provided some answers but most importantly created a new scepticism towards long standing common fallacies and clichés, at the time raising new questions. The most intriguing, dubbed as ‘the million‑dollar question’ (Bernal‑Casasola 2015: 76), concerns the reuse  The notion that amphorae a) were specifically designed according to criteria of density, weight, physical and/or chemical properties and domestic use of the commodity they were intended to accommodate and b) were thus recognized in markets as typical containers of that particular commodity, linked with specific geographic areas, dominated amphora‑studies from their very first steps, still retains great value and is fervently supported by a number of researchers (Opait 2007: 101–122; 2010: 154–155). 3

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 473–483

George Koutsouflakis

of amphorae and seems to haunt to the present day any attempt towards definitivity: Were amphorae containers of one commercial use? Were they primarily designed for a specific commodity? Were they refilled with different products — and if yes to what extent per period? Were they traded empty as single containers? Yet, even today, more than four decades later, amphorae recovered from ancient shipwrecks, do not easily reveal their original content. Shipwrecks are by definition unexpected events, deposited in an environment that, if protected, suffers the minimum of secondary human activity; however the amphorae comprising their cargos can hardly be considered ‘undisturbed’. Excluding some exceptional cases, the vast majority of amphorae recovered from ancient shipwrecks lack their stoppers and are filled with solid and liquid debris as a result of post‑depositional contamination. Liquid contents, no doubt the most common in antiquity, escape the vessels without leaving any visible trace. Semi‑liquids, according to their solubility in sea water, may as well disappear or solidify, depending on their nature. Solids may disintegrate, leaving tiny remains, easily removable by marine action. The content of any amphora recovered from a wreck‑site represents mainly the environment of the surrounding sea‑floor. Remains of foodstuffs like pits, seeds, shells, animal or fish‑bones found in amphora interiors may well not belong to their original content. Amphoras deposited underwater tend to get colonized by shells, cephalopods fish and crustaceans, which leave their own remains or collect inside objects from the vicinity of their dome.4 This dynamic in‑and‑out exchange of material in conjunction with — the less debated today — reuse that amphorae may undergo, obscures and perplexes the theoretically clear canvas on which we tend to delineate the potentiality of intact containers found underwater. Grape seeds for example are among the botanical remains, most commonly reported inside intact amphorae.5 Their solid structure enables them to survive while their tiny dimensions make them a less possible candidate to be played with by cephalopods. However, the information they provide is not necessarily univocal.  Shells, mainly of the Gastropoda and Bivalvia families, constitute one of the most usual items found in the interior of amphorae. The most common species of gastropods reported are Bolma rugosa, Cerithium vulgatum, Conomurex persicus, Haliotis myconosensis and Semicassis granulada, while the list of bivalve species is far to long to be mentioned here. Taking into consideration that most gastropods are herbivores, predatory carnivores or scavengers, the interior of an amphora can hardly be an ideal feeding ground to them. Their presence inside amphorae is mainly due to the collecting habits of Octοpus vulgaris. However they might as well have entered the vessels as living organisms. Several examples of complete amphorae in the storerooms of the Hellenic Ephorate of Underwater Antiquities in Athens demonstrate that shells may survive and thrive over long periods of time inside amphorae, long enough to grow larger and not be able to exit the vessel. 5  The examples from the Aegean are numerous, although mostly understudied or unpublished. For sone published some examples see Carlson 2003: 583; Pulak and Townsend 1987: 91; Greene et al. 2008: 704; Hocker 2005: 104; Günsenin 2005: 120. 4

Grape seeds have been found together with olive pits and pine seed hulls in the same amphorae of the Yassi‑Ada 7th century AD shipwreck (Bryant and Murry 1982: 328– 329; Van Doorninck 1989: 252–253). Although grape seeds outnumbered the other botanical remains and were considered as evidence for the transportation of a kind of cheap wine whose impurities had not been carefully removed, the question of an earlier use of these vessels remained open. The same amphorae might have carried olive oil, olives or olives preserved in wine during an earlier trip, remnants of which were the olive pits stuck on the coating of the amphora (Van Doorninck 2005: 96). Moreover, grape seeds stuck into the coating pitch might as well be remnants of an earlier use, not necessarily corresponding with the final use of the vessels (Ward 2015: 60). Clearly in many instances, botanical remains fail to give definite answers to long standing questions (though sometimes uttered in an unsuspected or even naïve manner), like ‘wine or oil?’. The introduction of DNA analysis performed on the inner walls of amphorae in order to trace micro‑residues imbued in the permeable pores of the clay, showed surprising and baffling results. New commodities, made their first appearance as ingredients, creating an ever‑growing puzzle. Archaeologists and historians, for example have assumed for almost a century that Chian piked‑shaped amphorae of the 4th century BC were wine‑carriers, since the Greek island of Chios was renowned in antiquity for its fine and distinctive vintages. Fifth and fourth centuries BC Chian coins depict grapes dangling above amphorae of similar shape. However, DNA analysis on a sample recovered from a shipwreck in the east side of the island, revealed on the contrary that the main commodity was olive oil flavored with oregano and other herbs, while a second type of amphora from the same wreck showed traces similar to mastic (pistacia lentiscus) or pistacia terebinthus (Hansson and Foley 2008: 1173–1174).6 Further results from DNA analysis in a larger number of Corkyrean and Mendean amphorae of the Classical and early Hellenistic periods, revealed mixed traces of wine and oil, suggesting either the extensive reuse of amphorae or the transport of non single‑species commodities (Foley et al. 2012: 397). It therefore seems that some of the most celebrated ‘wine amphorae’ of the Classical period were not exclusively used as wine jars and this fact begs the question of the content of any other type of amphora characterized as such. Commercial amphorae may have been intended to be, already from the time of their manufacture, more

 It is not clear if these traces correspond to the residues of the content or the lining. A mastic gum lining is reported in 4th century A.D. vessels from the Athenian Agora (Robinson 1959: 8, 93). Terebinth resin was the content of nearly 150 two‑handled Canaanite jars recovered from the cargo of the Uluburun shipwreck (Mills and White 1989). 6

474

‘De profundis’: three amphorae of unorthodox contents retrieved from the Aegean Sea

multifunctional than previously thought:7 while Mende was one of the most praised wine producing areas in the Greek world (Papadopoulos and Paspalas 1999), nine of the ten Mendean amphorae recovered from the Tektaş Burnu shipwreck contained black pitch (Carlson 2003: 588–589), a commodity that could be used both in shipbuilding, for waterproofing wood, or in amphora lining, to prevent oozing and losses (Koehler 1986: 50– 51). Certainly this was not exceptional.8 An excerpt from Polybios (5.89.6) mentions the donation of Macedonian king Antigonos Doson to the Rhodians of ‘a thousand of amphorae of raw pitch’, after the destructive earthquake of 227/6. Although pitch was in great demand in antiquity (Meiggs 1982: 467–471), there is no such thing so far as special ‘pitch amphorae’ and as in the case of the Tektaş Burnu shipwreck, Antigonos Doson would probably have had to rely on any available batch of amphorae. As seen above intact amphorae retrieved from the sea, if treated according to specific, strict protocols, may enrich our knowledge on the issue of amphora content or may instigate studies into more composite approaches, opening a continuous multidisciplinary dialogue. Examples of undisputed content are still limited and mainly confined in cases of solids or semi‑solids which leave more firm residues. The three following cases are included in this category of solid contents. 2. A Lamboglia 2 amphora from the Gerolimnionas‑bay shipwreck, off the island of Makronisos (South Euboean Gulf S/W no.26) The Gerolimnionas‑bay shipwreck at Makronisos island, along the east coast of Attica, was discovered at the entrance of the bay, during an extended underwater survey conducted in 2013 by the Hellenic Institute of Underwater Antiquities in cooperation with the Ephorate of Underwater Antiquities.9 The wreck‑site spread over an area of more than 1000 square meters along the coast in a depth ranging from 25 to 48m. Although deposited in the shores of a deserted and uninhabited island which suffered the minimum of human action, the wreck had been seriously disturbed by the installations of a fish farm that had operated in a nearby spot for several years  The suggestion that amphorae were designed as general‑purpose containers, has been put forward mainly for the Late Roman and Byzantine production, when reuse and re‑circulation of amphoras seems to have become a generalized phenomenon with less evident affinities between form and content (Van Halfen 1996: 209; Steckner 1989: 65). Y. Garlan (2000: 90–91) and P. Dupont (2001: 455) challenged some years ago the hitherto common assumption of a primary content also for the amphoras of the earlier, Greek periods. The implications of a ‘primary context approach’ and a ‘multipurpose approach’ are commented on by Lawall (2011: 23‑24, 30–33). 8  Amphorae filled with pitch or resin are reported from shipwrecks of all ancient and mediaeval periods, usually in limited numbers compared to the rest of the consignment (see Günsenin 2005: 120; Parker 1992: 168, 284, 301, 372–373; Frost 1969: 22). 9  For the results of the survey see Koutsouflakis 2013. For a preliminary report on the finds of the shipwreck no.26 in Gerolimnionas bay see Koutsouflakis and Argiris 2015: 5–6 and Koutsouflakis 2017: 59–62. 7

Figure 1. Lamboglia 2 amphora from Makronisos (BE 2013/21-4). during the 1990s. The cargo was homogenous consisting of Tripolitanian I amphorae, most of them found in fragmentary condition, which should be dated to the 1st or 2nd century AD. A Lamboglia 2 amphora (BE 2013/21‑4, Figure 1) was found amid the main concentration of the cargo, side by side with some of the very few intact Tripolitanian I amphorae. Together with a Benghazi MR 2 intact amphora, located somewhat further away, they comprise the only amphorae that seem to differ from the rest of the consignment. There is no certain way to prove that both vessels belonged originally to the ship. The Benghazi MR 2 amphora could be a later intrusion to the site while the Lamboglia 2 sample might well have pre‑existed in the same location. Generally, the Lamboglia 2 type precedes the Tripolitanian I type for at least one and a half centuries with the two types co‑existing for a short period in the markets along the later 1st century BC. If they were indeed in the same ship, the Lamboglia 2 amphora might well have been an old reused vessel incorporated into the ship’s equipment, for storing supplies, liquids and foodstuffs for the crew.10 The heavy construction of  There is ample evidence for the reuse of transport amphorae as purely storage vessels in households excavated in terrestrial sites throughout the Aegean. In some cases these amphorae, stored in cellars, are remarkably old and found upside down, in wait for any use (for a characteristic case in Thasos see Empereur and Garlan 1992: 212–213; Grandjean 1992: 578). In this sense the survival of some older amphorae within the cargo of a shipwreck should not surprise us — although the unstable environment of any ship by definition does not facilitate a long‑term resistibility against breakage. 10

475

George Koutsouflakis

Makronisos island and was simply dumped into the sea because its content had gone bad.12

Figure 2. Bones of capra aegagrus hircus found inside amphora BE 2013/21-4. any Lamboglia 2 amphora would perfectly fit to that task. Furthermore, the Lamboglia 2 sample exhibited clear signs of a secondary structural interference i.e. the solid spike had been cut short and a handmade vertical hole had been drilled through the remaining spike into the interior of the vessel. A total of four almost complete amphorae were raised from the cargo. The content of the two Tripolitanian I and the one Benghazi MR 2 samples was thoroughly investigated through a sieve providing nothing but the ordinary: sand, pebbles, crustacean exoskeletons, bivalve and univalve seashells, pieces of calcium carbonate, fragments of calcareous algae and tiny pieces of broken ceramics. The same content was also testified in the Lamboglia 2 sample, with the addition of a good number of animal bones (Figure 2). The bones consisted of pieces of the lower legs, jaws, ribs and vertebrae from at least two different animals belonging to the same capra aegagrus hircus species, the common Greek goat breed, found in domestic, semi‑domestic or wild state in mainland Greece, the islands and Asia Minor.11 Some of the bones bore clear cut‑marks of the butchering implement. The animals’ flesh, perhaps dried, salted or smoked, if not processed in a different manner, had been cut into big chunks of a size suitable for passing through the vessel’s neck and had been stored inside the amphora. The vertical hole on the spike might have served to drain moisture or the residues of the meat’s melting fats. The amphora, clearly undergoing re‑use — a kind of reuse not initially intended for it — should be connected to food‑supplies stored on board, regardless of whether linked to the Tripolitanian I cargo or not. If it pre‑existed on the site of the Tripolitanian I cargo, it could have been part of the supplies of any other ship that passed the east coast of  The identification of the bones was carried out by the specialist Th. Webb, whom I warmly thank here for his contribution. 11

A few other cases of meat stored in amphorae for transit are known: a Mendean and a Pseudo‑Samian amphora containing butchered beef bones were raised from the Tektaş Burnu Classical wreck (Carlson 2003: 589) and are the only other examples to be found so far in the Aegean. Butchered meat, stored in Punic amphorae as a trading commodity or as ship’s supplies is also known from the wrecks Coltellazzo A (Parker 1992: no.329), Olbia AB (Parker 1992: nos758‑9) and Santa Gilla (Solinas 1993: 32). The presence of scattered animal bones is also reported from a number of shipwreck sites without it being always clear if they represent food eaten on board or if they are the result of post depositional contamination.13 It seems that packing salted or processed meat in amphorae and consuming on board, was not as exceptional as previously thought. Ancient sources are however less eloquent in revealing the minutiae of traded, processed meat. The 5th century BC comic poet Hermippus in his lost work Phormoiphoroi (the basket‑carriers) lists beef chops from Thessaly and pork from Sicily among the imports available in the Athenian market (Athenaeus 1.27.e‑f). The title of the play does not necessarily imply that all commodities listed were transported in phormoi (baskets) and probably phormoiphoroi stands for the more generic meaning of porters or carriers. A papyrus from the Zenon archive (P. Cairo Zen. 59012, col. II, lines 38‑48), dated to 259 BC, refers to the valuation of goods imported to Egypt on two ships, recording among others (dried fish, salted fish, etc.) ‘two earthenware jars of wild boar‑meat’, using the term keramion (Austin 1981: no.237). Other papyroi of the same vintage refer to keramia filled with calf and deer meat (Panagou 2016: 327), while pickled calf and pork meat are also mentioned in papyroi as contents of amphorae from the prefecture of Gaza (Kruit and Worp 2000: 101– 103, SB XX 14574.2, 14212.8). The term keramion (literally earthenware jar) stands almost certainly for amphora. 3. A ‘Proto‑Dressel 25’ amphora from the Gerolimnionas bay off the island of Makronisos Not far from the spot of wreck no.26 described above, in the opposite (southern) side of the entrance of Gerolimnionas bay, an isolated amphora was located during the 2013 season, half‑buried in the sand at a depth of about 55m. The vessel was set free from the sediments and raised  As stated, the finds pot of wreck no.26 in Gerolimnionas bay was along the shoreline of an island that was isolated and uninhabited both in ancient (Strabo 10.5.3) and modern times. The Gerolimnionas bay was used in antiquity as a temporary anchorage; there are however no traces of any kind of installation or settlement on land in the area surrounding the bay to which the amphora BE 2013/21‑4 could be possibly related. This leaves no doubt as to its connection with ship supplies. 13  Scattered bones of cattle, pig, sheep and goat are reported in Serçe Limani (Bass 2005: 114), Vulpiglia (Parker 1992: no.1230), Porticello (Parker 1992: no.879), Culip D (Parker 1992: no.347), Mallieha (Frost 1969: 27), Titan (Parker 1992: no.1149). 12

476

‘De profundis’: three amphorae of unorthodox contents retrieved from the Aegean Sea

up (BE 2013/21‑6, Figure 3). In the vicinity of the find spot, several other amphora sherds were documented without providing any reliable context: amphora‑sherds dated from the Classical to the late Byzantine period were scattered throughout the whole area of the bay’s entrance, no doubt as a result of having been jettisoned from ships passing by or entering the anchorage. The amphora’s shape displays close affinities to the still not thoroughly studied family of the presumable ancestors of Dressel  25 type, occasionally labeled Proto‑Dressel 25 and linked to the production of the Northern Peloponnese, after the destruction of the town of Corinth by the Romans in 146 BC (Opait 2010: 155–156; Grace 1979: 19).14 Its shape and profile are close to that of P 6795 (Opait 2010: pl. 90.1) and P 6796 (Grace 1979: fig. 38, third to the right) found in the Athenian Agora, dated to the second half of the 2nd century BC.15

Figure 3. Proto-Dressel 25 amphora from Makronisos (BE 2013/21-6).

The content of amphora BE 2013/21‑6 amphora was a concreted mass of heavily crushed shells of the Muricidae shellfish family (Figure 4),16 well known as the source for the production of purple dye that was used to colour textiles and fibres in the Aegean already from the Middle Bronze Age. Installations and workshops for the production of purple dye are known in several littoral and hinterland sites, in the outskirts of towns and settlements and even in isolated, uninhabited islets in the Aegean.17 Apart from dye production, murex was considered suitable for human consumption and could be used as well as fishing bait, while murex shells could be used as chalk, ornaments or as a source of lime (Alfaro and Mylona 2014: 160; Bélis 1999). For the crushed shells, the inevitable by‑product of any purple‑dye workshop, there was apparently a further use, i.e. as a  The form is also known under the terms ‘Will 11c’ (Will 1987: 205), ‘Greek Brindisian’ (designation by V. R. Grace) or simply ‘Northern Peloponnesian’ based on their origin (Lawall 2010: 395). 15  There is little doubt that the post‑destructional cluster of amphorae developed in the workshops of Northern Peloponnese has close similarities to (if they are not directly evolved from) the late versions of Corinthian A’ amphorae. The connection however to the later Dressel 24 and 25 families needs yet to be verified. The amphora presented here has the ovoid body and handles of the Agora examples, but exhibits a rather cup‑shaped mouth that misses the typical pronounced plastic bands on the rim and the level of the upper joint of the handles to the neck. 16  ‘Purple shellfish’ or ‘murex’ are generic terms to describe a large number of members of the Muricidae family. The Mediterranean species exploited for the production of purple dye are Heraplex trunculus L., Purpura haemostoma L. and Bolinus brandaris L. The condensed masses chiseled out of the amphora also contained discernible pieces of common bivalve shells. The underwater environment from which the amphora was selected did not exhibit marine sediments of this kind. 17  Production sites are usually indicated by spreads or small heaps of crushed shells, evidently the waste from purple dye production. They are much more numerous in the Aegean than those listed by Reese (2000), and can be found in the most unexpected places: such scatters were observed by the author on the islets of Phtena, two tiny twin‑islets located 2 nautical miles south of the island of Anaphi in Southern Cyclades, together with the foundations of a large building. The restricted inland area of these two islets would have made it impossible for them to have been populated. 14

Figure 4. Crushed murex shells from the interior of amphora BE 2013/21-6. supplementary component of certain kinds of cement. This use is attested in the western Mediterranean during the Roman times including the shells’ smashing into a white silt perfect for plasters, but also their addition in the mix of opus signinum, used for fish‑vats, cisterns etc.18 In the Aegean, they are reported as an ingredient of the plaster underlying mosaics of the Hellenistic period in Delos (Bruneau 1969: 766–767) and Pergamon (Kawerau and Wiegant 1930: 54). In the case of Delos, constructors made probably use of the disposable wasters of workshops operating in the industrial quarter (Bruneau 1969: 765–779; 1978: 110–114). At Pergamon the raw material had to be carried from the shore inland at a distance of nearly 30km.  I would like to thank A. Sáez Romero for his valuable advice on the subject and for sharing with me unpublished details from the sites of Camposoto and Cerro de la Batería in the southern Cadiz island, where the use of shells in the opus signinum is testified. I would like to add here some interesting information, of a more recent exploitation of murex shell heaps left from ancient ateliers: they were used during the 19th and 20th century in Spain as extra calcium for feeding chickens and other cattle (A. Sáez Romero, per vocem). 18

477

George Koutsouflakis

Shells, as a source of calcite material could have been transported over short or longer distances by land or sea, and we may even suspect that there was some kind of economic interest in the exploitation of existing resources. The notion however, of transporting raw materials (like plaster, alum, sand, minerals, pozzolana etc., Panagou 2016: 329) in amphorae designed originally to carry edible commodities, cannot possibly be considered as an extended and generalized practice. Certainly there were other more practical and less expensive ways available to fulfill this task. But surprisingly enough, the Makronisos Northern Peloponnesian amphora is not the only example reported in the wider area of the Aegean. An amphora of about the same date, full of crushed murex shells, was documented in the district of the House of the Comedians in Delos and should be connected with the Delian workshops operating on the island (Bruneau 1969: 764–765, fig. 2). A second unidentified and unpublished example of unknown context is reported from the city of Chania in Crete (M. Vlazaki per vocem in Alfaro and Mylona 2014: 160). Apart from these few examples found on dry land, we hear nothing else about this practice in the Aegean. Sea‑shells deliberately packed in amphorae are mentioned also in wrecks, presumably as ingredients of a garum made out from mollusks. Pectunculus pilosus shells were found inside some of the Almagro 50 Lusitanian amphorae from shipwreck Planier G, dated to the first half of the 4th century AD (Benoit 1962: 159). While the concoction of a special garum including shells may well be the case in shipwrecks like Planier G or Titan (Parker 1992: nos830 and 1147), the given evidence is too poor in wrecks like the Giglio Porto (Parker no.453), where both the existence of mortar or a shell‑fish edible product has been insinuated. In any case, the nature and density of fragmented shells inside the Makronisos amphora certainly exclude any possibility of an edible commodity. The original source of the content of amphora BE 2013/21‑6 remains unknown. Makronisos island has not yet provided any traces of murex‑processing sites; neither has the eastern coast of Attica. The island of Agh, Georgios, some miles further to the south has nevertheless provided ample evidence of crushed murex shells and the whole area of the Evripos channel, further to the north, was regarded as a rich hunting ground for murex. The amphora was probably thrown overboard as an act of jettison of a ship in danger, or was disposed of because its content, presumably in the form of wet mortar, was concreted and the vessel consequently disused. 4. A Zeest 72 amphora from a shipwreck off the island of Fournoi (Fournoi S/W no.15) The wreck was located in 2015, during a survey conducted by the Hellenic Ephorate of Underwater Antiquities in cooperation with the R.P.M. Nautical Foundation, at

Fournoi, a cluster of islands slightly south to the islands of Samos and Icaria, in the Eastern Aegean. Although the wreck‑site was known to the community of Kalymnian sponge divers for many years and might have suffered some extent of looting, the main concentration of the cargo still remained in relatively good condition, half‑buried in a sloping sandy seafloor, in a depth of about 45‑52m. The shipwreck can be considered as relatively coherent. The cargo consisted mainly of amphoras, although several concreted columns of tableware were also traced between the heavier containers. The inspection of the wreck site revealed an heterogeneous amphora consignment consisting of Zeest 7219 and Shelov E/Zeest 104 amphora types as main cargo, the first such cargo to be found in the Aegean (Figure 5). There were at least two different variations of the Zeest 72 type: the vast majority of amphorae displayed an oblong form, which corresponded to the Zeest 72 type already known (Krapivina 2010: 69–70, pl. 39:1). However a sub‑variation, less tall, with wider shoulders, was also noticed. A typical example of the Kapitän II and some small sized Shelov D/ Zeest 91b amphorae were also recognized and sampled inside the consignment. The Zeest 72 type is almost exclusively known from sites of the northern and northwestern Black Sea littoral (Pantikapaion, Tanais, Chersonesos, Mirmeki. Iluraton, Olbia, Scythian Neapolis, Histria, Kavarna, Sozopol, Romula). Distribution maps display their highest density in the area of Crimea and the Azov Sea, while the chemical composition of the clay also seems to point towards that area (Dyczek 2001: 229, 232). Pantikapaion and Mirmeki, ancient towns in the area of Crimea were proposed as the manufacturing centers (Zeest 1960: 111).20 This amphora is hardly known outside the Black Sea. According to the evidence of terrestrial sites, Zeest 72 type amphorae made their first appearance in the late 2nd century AD, with production and circulation culminating during the 3rd century. The type went out of use somewhere in the first half of the 4th century AD. The content of Zeest 72 type was assumed to be salsamenta on the basis of amphorae found in cisterns used for salting fish, excavated in Mirmeki (Gajdukievič 1952: 206–207, figs. 123‑124). Two intact amphorae of the Zeest 72 type were raised from the Fournoi wreck, both lacking stoppers. The one  Zeest 1960: 111‑112, pl. XXX. The type is also known in eastern classifications as Abramov 6.11‑6.13 (Abramov 1993: pl. 54) or Krapivina 19 (Krapivina 1993: fig. 30, 1‑3). The paucity and limited distribution of this amphora outside the area of the Black Sea has resulted in its long absence from any major Mediterranean typology or classification of amphorae. Four decades after its original description and classification, the type was incorporated into P. Dyczek’s typology as Type 31 (2001: 228–233). 20  Strangely enough, while commenting on the similarity of the chemical composition of the clay to sources around the area of Crimea and displaying distribution patterns that demonstrate the same origin, Dyczek (2001: 231) proposed an origin in the Aegean or in Ionia, based on ‘convincing chemical and physical examinations’. 19

478

‘De profundis’: three amphorae of unorthodox contents retrieved from the Aegean Sea

Zeest 72 amphora (BE 2015/5‑45, Figure 6) was found in a semi‑standing position, with the top facing the surface (Figure 5, right). It is not known if this position was due to secondary removal made by previous visitors on the site or if it was originally deposited in this way. Certainly an upstanding position would contribute a great deal to the preservation of traces of the original content. Investigation of the Zeest 72 sample through a sieve revealed some unexpected finds. While almost the whole upper half of the vessel was filled with sand, disintegrated marine flora, fragments of shells and pieces of crustaceans, the density of crustaceans became enormously larger towards the lower part of the vessel. Having removed the upper two thirds of its content, the remaining third constituted of a dense layer of concreted shells from crustaceans that had to be chiseled out in large chunks (Figures 7‑8). These chunks consisted mainly from intact pieces of cephalothorax and claws attributed to a kind of crustacean unknown to the local community of fishermen at Fournoi. The exact species of those crustaceans has not been identified yet with absolute certainty. There are strong indications however that it corresponds to the Galathea stringosa (Linnaeus 1761), a squat lobster of the Galatheide family wide spread from the norteast Atlantic Ocean to the Mediterranean and the Red Sea.21 Ancient Greeks used two terms, more or less generic, to denote the two main families of the dekapoda order, astakos and karavos; both terms are loosely used, however, in ancient sources.22 The inconsistency as well in the  Ι am indepted to Dr Dimitra Mylona for the identification of the species (made by images sent) which I fully trust is correct. Galathea stringosa is described as an edible member of the crustacea subphylum with small — if any — commercial value at present. It can be found on any depth from the sublittoral zone and down to 600 meters, on gravely and rocky seafloors. 22  The terms have survived in modern Greek (astakos, karavida) but the exact meaning might well have changed after two millennia. Today astakos signifies mainly the large‑sized Greek kind of lobster (Palinurus vulgaris) but it is also used occasionally for the less common European lobster (Homarus gammarus). The term karavida, on the contrary, is applied to any kind of small‑sized fresh or saltwater crayfish that carries claws. Aristotle provides the earliest extant literary discussion in his treatise Historiae Animalium which remains the best evidence towards understanding the classification of crustaceans in the ancient Greek world. According to him, the two species can be distinguished by the presence or absence of large claws. Accordingly he called astakos the smaller crustaceans with claws and karavos the crustaceans without claws (HA, 525 a 30–33). If so, the ancient meaning of the terms contradicts their modern use, with the term karavos, clearly a maritime creature, applied to Palinurus vulgaris (Palinurus elephas) and the term astakos to any other kind of clawed, fresh‑water (Astacus‑astacus) or maritime crayfish (Nephrops norvigicus, Homarus gammarus, Galathea stringosa, etc). But even Aristotle was not strictly consistent in using the terms. Moreover, Aristotle often had to Adopt periphrastic names when reporting creatures that were not given any specific name up to his time (Voultsiadou and Vafidis 2007: 105, 108). Among the unclassified malakostraca cited by him, a special mention should be made of a ‘genus small like crabs but in appearance similar to lobsters’ (HA 525b10). This description fits more or less well to the squat lobsters of the Galatheidae family. 21

Figure 5. Fournoi S/W nr15. Part of the exposed cargo. usage of these terms in modern language often generates confusion,23 since the jargon used by fishermen and markets may well differ from the existing scientific nomenclature. The second Zeest 72 amphora sampled from the wreck, was found in a sidelong, slightly inverted position with nothing of remains of Galathea stringosa inside and the same goes also for the remaining types of amphorae raised from the site. The Zeest 72 in question, may have carried supplies for the ship’s crew or some limited amount of delicacies aimed at a very restricted market. Crayfish (and lobster), praised in the works of ancient comedy as a much sought‑after delicacy (Athenaeus 3.104.d‑e), were no doubt available in local markets, in presumably small amounts. Despite the ubiquity of representations on mosaics in Hellenistic and Roman dining halls, we miss evidence for a more intensive exploitation of local resources. Both of them are certainly not listed among the staple products of the Northern Black Sea region and ancient writers, contrary to their attitude towards other maritime species, do not reveal any famed place of origin for them. Some interesting connections might be sought in the iconography of Pontic and Bosporian coinage (Koutrakis et al. 2009: 1538–1540). The depiction of a staple product as a main or secondary motif on ancient coins is well attested in antiquity. The crayfish is depicted on the reverse of coins of two former Milesian colonies in Mysia and Thracia. The most striking iconographic example comes from the city of Priapos (modern Karabiga, Turkey) in the southwestern shores of the Marmara Sea. Coins minted during the 1st century BC illustrate a head of Apollo on the one side and a full‑framed, clawed crayfish to the left on  For treatises on identification and nomenclature of both species in ancient literature and modern terminology see Voultsiadou and Vafidis 2007: 107–109 and Koutrakis et al. 2009. It should be also stressed that the exact meaning of the specific terms might have undergone changes during later antiquity. Athenaeus, a writer of the 2nd‑3rd century AD, refers to genuine karavos (6.244.b) a fact that demonstrates the ambiguity of ancient terminology. 23

479

George Koutsouflakis

from the Classical to the early Hellenistic age (Hind 1985: 103–104). The depiction of the obverse may range from a head of Apollo, to a head of Medusa or a naked Apollo seated on the omphalos. The reverse consistently exhibits in full frame an archaic‑styled anchor sided by a tiny crayfish in the field to the left or to the right of the anchor’s shank. The main side is clearly dedicated to the eponym god while the other hints at the city’s safe harbour dominating over a long harbourless coast to the south, and should be related to the city’s longstanding interconnections to the maritime environment. It is not clear if the crayfish is to be related to a local commodity that presumably played some role in the sustenance or the economy of the city. Populations of Astacus leptodactylus, a narrow clawed crayfish have been attested in areas surrounding Apollonia Pontica as well as in several coastal areas in the Black Sea (Koutrakis et al. 2009: 1539).25 Studies however demonstrating the extent of Galathea stringosa do not favour a wide outspread in the Black Sea, although populations of the species are attested in the Sea of Marmara (Müller 1986: 121). In any case, there is no substantial evidence of a systematic exploitation or even trading of local resources.26 Figure 6. Zeest 72 amphora from Fournoi (BE 2015/5-45).

Figure 7. Mass of concreted crayfish exoskeletons from the interior of amphora BE 2015/5-45. the other.24 The potential interconnection however with the exploitation of the city’s maritime resources remains doubtful, in the absence of more concrete evidence. A crayfish‑like creature is also present in a full series of silver tetradrachms, drachms and diobols from Apollonia Pontica (modern Sozopol, Bulgaria) dated  The town is mentioned by Strabo (13, 587) who comments that the area produced fine wine and that the God Priapus (son of Dionysus and Aphrodite) gave the town its ancient name. If wine was indeed the main product of the city, the presence of crayfish on the coins might have had other symbolic interconnections. The city of Priapos had strong relations to the sea. Thucydides in the 5th century BC refers to the town as an important naval station, while God Priapus was a patron god of merchant sailing as well (Neilson 2002). 24

The evidence from the so far unexcavated Fournoi shipwreck no.15 seems to denote the presence of a rare delicacy within the cargo, stored probably in one of the available amphorae on board. It has to be noted though that even in cases like that, where evidence appears to be as solid as it can get, conclusions have to be articulated with extreme caution: regardless of the fact that we are yet unaware if amphora BE 2015/5‑45 can speak for the rest of the cargo (probably not), the list of speculations can grow much larger: could it be that the raised amphora was in fact colonized by populations of Galatheiae stringosae right after the ship sunk and that the exoskeletons of generations of crustaceans were accumulated in the lower part of the vessel because of its half‑standing position, before the amphora was finally silted? We do not know much about the habits of Galatheia stringosa, but as a nocturnal and photo‑phobic invertebrate it tends indeed to populate rock interstices or generally any closed and dark environment. Supporting testimony from other shipwrecks in the Mediterranean is ill‑defined and rather obscure. Langoustine shells are reported inside amphorae raised from La Luque B and Pampelone shipwrecks in southern  This species is found in both fresh and brackish waters, e.g. lagoons, channels, estuaries, as well as in slow running freshwater rivers in the Ponto‑Caspian Basin. It is mentioned as a highly Adaptable crustacean that expanded also to vast tracts of waterland throughout Central and Western Europe. 26  On the contrary Hind (1985: 90–92) suggested that the depiction of crayfish (ancient astakos) on coins of Apollonia Pontica was a hint to the population of the hinterland called Astai and the whole area Astike Thrake, creating a symbolic pun between the two words. 25

480

‘De profundis’: three amphorae of unorthodox contents retrieved from the Aegean Sea

category, demonstrating a rare kind of reuse the justification of which remains on practical terms, enigmatic and ill‑understood. With the exception of pitch, amphorae are less often reported to contain non‑edible commodities and whenever this happens they are utilized purely as storage jars.27

Figure 8. Cephalothorax and claws from crayfish, selected from the amphora BE 2015/5-45. France dated in the early 4th century AD (Parker 1992: nos611 and 783), without elaborating much further on the identification of the exact species and the density of remains. 5. Epilogue All three examples presented here highlight different stages in the life of an amphora (Peña 2007: 8–9), with one common denominator: apart from being loaded on a ship and ending at the bottom of the sea, they all contained commodities they were not originally destined to carry. The timeworn and deliberately truncated Lamboglia 2 amphora stands as a characteristic example of a transportation container that, having fulfilled its main commercial task, was removed from circulation but remained in use for a specific, secondary, storage purpose, a purpose it was not originally designed for, demanding thus some needful adaptations. This phenomenon is frequent in ancient ships; almost every excavated shipwreck produced some amphorae that do not tally, either typologically or chronologically, with the main cargo and are regarded as vessels left aboard from previous consignments and put inside the ship’s cupboard. Amphorae are highly durable and adaptable containers and reuse of this kind is frequently attested in domestic or nautical contexts. The basic concept of a container essentially made to transport (and subsequently store) drinks and foodstuffs is not violated yet. Ancient Greeks, Phoenicians and Romans were however less sensible and far more practical when utilizing food containers. After completing an initial task amphorae could lose or not lose their commercial affiliations. An undefined quota of them was sent (back?) to production centers for refilling, while others were disposed of or were channeled in urban and extra muros sites for secondary uses. And no doubt, a good number of them remained disposable in markets, workshops, ships, ports and households in order to undergo a second, more multi‑functional use, in some cases diverging from carriage or storage of foodstuffs. The ‘proto‑Dressel 25’ amphora from Makronisos is comprised in this latter

The latest example, the Zeest 72 amphora from Fournoi, raises a lot of questions which we cannot fully answer yet. The nature of the cargo that the shipwreck Fournoi no.15 displayed, seems to point towards a well concerted, targeted consignment, showing no obvious traces of reshipped, refilled or reused amphorae whatsoever. The Zeest 72 amphora type itself remains to a great extent an understudied type that apparently appeared little in the amphora‑currency of the 3rd‑4th century AD Aegean and did not travel far out from the Black Sea. Suggestions on the content until now were based on mere speculations. The general characteristics of the shape, the large mouth, large capacity and the fact that it circulated (and was probably produced) in an area well‑known for the processing and trade of salt‑cured products may insinuate salsamenta as the primary content. On the other hand, lobster‑like crustaceans are never mentioned between the staple‑products of the northern and northwestern Black Sea and exploitation of local resources, if existed, presumably never grew to the extent of demanding a specific type of amphora for its shipment and trading: we would expect to find more evidence about that, historical, epigraphical or archaeological. Galatheia stringosa is not referred to as an endemic species in the Black Sea littoral. In every likehood the carriage of lobster was not a ‘random, unorthodox primary use’ (Lawall 2011: 30) but rather a commodity loaded on available empty amphorae somewhere on the way to the Aegean or on the way back to the Black Sea. The Zeest 72 type, could be a multi‑purpose kind of amphora or an amphora that accomplished a primal use and was travelling on board, refilled with a product not available in the homeland. Bibliography Abramov, A. P. 1993. Antichnye amfory. Periodizatsija i khronologija, Bosporskiy Sbornik 3: 4–135. Adams, J. 2001. Ships and boats as archaeological source material. World Archaeology 32: 292–310.

 Pigments and iron nails are reported to have been found inside Late Roman amphorae in the vaulted rooms of the mosaic‑floored Edifice of the ancient port of Tomis (modern Constanţa, Radulescu 1973: 197–198). Pigments in reused amphorae also came to light in the Athenian Agora (Lawall and Jawando 2004). Petroleum (naphtha) is reported as the content of several Shelov D/Zeest 91b amphorae discovered in the city of Tanais (Kozlovskaya and Ilyashenko 2013: 91–92). What is less known, is the tendency of ancient sailors to store small‑sized pots inside amphorae. Intact lamps and cups are often retrieved from the interior of amphorae (e.g. Carlson 2003: 107–108). Although octopi are most usually blamed for this apparent ‘displacement’, the instances are far too many to be imputed only to the collecting habits of cephalopods. 27

481

George Koutsouflakis

Alfaro, C. and Mylona, D. 2014. Fishing for purple shellfish (Muricidae) in Ancient Greece: Acquisition Technology and First Steps in Purple Dye Production. In C. Alfaro, M. Tellenbach, and J. Ortiz (eds), Production and Trade of Textiles and Dyes in the Roman Empire and Neighbouring Regions. Purpureae Vestes IV: 149–165. Valencia, Universitat de València. Austin, M. M. 1981. The Hellenistic World from Alexander to the Roman Conquest. A Selection of Ancient Sources in Translation. Cambridge, Cambridge University Press. Bass, G. F. 1982. The pottery. In G. F. Bass and F. H. Van Doorninck Jr. (eds), Yassi Ada, Volume I, A Seventh‑Century Byzantine Shipwreck: 155–188. College Station, Texas A&M University Press. Bass, G. F. (ed.) 2005. Beneath the Seven Seas: Adventures with the Institute of Nautical Archaeology. New York, Thames and Hudson. Bass, G. F. and F. H. Van Doorninck Jr. (eds) 1982. Yassi Ada, Volume I, A Seventh‑Century Byzantine Shipwreck. College Station, Texas A&M University Press. Bélis, M. 1999. The use of purple in cooking, medicine and magic: An example of interference by the imaginary in rational discourse. In R. Buxton (ed.), From Myth to Reason? Studies in the Development of Greek Thought: 295–316. Oxford, Oxford University Press. Benoit, F. 1962. Nouvelles épaves de Provence (III). Gallia 20: 147‑176. Bernal‑Casasola, D. 2015. What contents do we characterize in Roman amphorae? Methodological and archaeological thoughts on a ‘Trending Topic’. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 61–83, Esposende, Município de Esposende. Bryant, V. M. J. and Murry, J. R. M. 1982. Preliminary analysis of amphora contents. In G. F. Bass and F. H. Van Doorninck Jr. (eds), Yassi Ada, A seventh century byzantine shipwreck: 327–331. College Station, Texas A&M University Press. Bruneau, P. 1969. Documents sur l’industrie délienne de la pourpre. Bulletin de Correspondance Hellénique 93, 2: 759–791. Carlson, D. N. 2003. The Classical Greek shipwreck at Tektaş Burnu, Turkey. American Journal of Archaeology 107: 581–600. Dupont, P. 2001. Trafics méditerranéens archaïques : quelques aspects. In R. Eichman, R. and H. Parzinger (eds), Migration und Kulturtransfer. Die Wandel vorderer‑ und zentralasiatischen Kulturen in Umbruch vom 2. zum 1. vorchristlichen Jahrtausend. Akten der Internationalen Kolloquiums Berlin (23‑26 nov.1999): 445–460. Bonn, Rudolf Habelt. Dyczek, P. 2001. Roman Amphorae of the 1st ‑ 3rd centuries AD found on the Lower Danube, Typology. Warsaw, Warsaw University Press. Empereur, J.‑Y. and Garlan, Y. 1992. Bulletin archéologique : amphores et timbres amphoriques. Revue des Études Grecques 105: 176–220.

Foley, B. P., Hansson, M. C., Kourkoumelis, D. P. and Theodoulou, T. A. 2012. Aspects of ancient Greek trade re‑evaluated with amphora DNA evidence. Journal of Archaeological Science 39: 389‑398. Frost, H. 1969. The Mortar Wreck in Mellieha Bay. London, Appetron Press. Garlan, Y. 2000. Amphores et timbres amphoriques grecs entre érudition et idéologie. Mémoires de l’Académie des Inscriptions et Belles Lettres, Nouvelle série 21. Paris, de Boccard. Gajdukievič, V. F. 1952. Razkopki Mirmekija v 1935‑1940. Materialy i issledovanija po archeologii SSSR 25: 135–222 (in Russian). Gibbins, D. and Adams, J. 2001. Shipwrecks and maritime archaeology. World Archaeology 32: 279–291. Grandjean, Y. 1992. Contribution à l’établissement d’une typologie des amphores thasiennes. Bulletin de Correspondance Hellénique 116: 541–584. Greene, E. 2005. An Archaic ship finally reaches port: Pabuç Burnu, Turkey. In G. F. Bass (ed.), Beneath the Seven Seas: Adventures with the Institute of Nautical Archaeology: 59–63. New York, Thames & Hudson. Günsenin, N. 2005. A 13th‑century wine carrier: Çamalti Burnu, Turkey. In G. F., Bass (ed.), Beneath the Seven Seas: Adventures with the Institute of Nautical Archaeology: 118–123. New York, Thames & Hudson. Hansson, M. C. and Foley, B. P. 2008. Ancient DNA fragments inside Classical Greek amphoras reveal cargo of 2400‑year‑old shipwreck. Journal of Archaeological Science 35: 1169–1176. Hocker, F. 2005. Sampling a Byzantine vintage: Bozburun, Turkey. In G. F. Bass (ed.), Beneath the Seven Seas: Adventures with the Institute of Nautical Archaeology: 100–105. New York, Thames & Hudson. Kawerau, G. and Wiegand, T. 1930. Altertümer von Pergamon, V.1, Die Paläste der Hochburg. Berlin, de Gruyter. Koehler, K. G. 1986. Handling of Greek transport amphoras. In J.‑Y. Empereur and Y. Garlan (eds), Recherches sur les amphores grecques. Bulletin de Correspondance Hellénique Supplément XIII: 49–67. Athens, École française d’Athènes. Koutrakis, E., Machino, Y., Mylona, D. and Perdikaris, C. 2009. Crayfish terminology in Ancient Greek, Latin and other European languages. Crustaceana 82: 1535–1546. Koutsouflakis, G. V. 2013. Navigation and Commercial Transportation in Southern Euboean Gulf (6th cent. BC ‑ 14th cent. AD). Unpublished Phd Thesis, University of Athens. Koutsouflakis, G. V. 2017. Underwater archaeological survey in the Southern Euboean Gulf (2010‑2016). Part A: Petalioi ‑ Makronesos. Enalia XII, 32–75. Koutsouflakis, G. V. and Argiris, X. 2015. Roman North African amphorae in the Aegean: the evidence of shipwrecks. In S. Demesticha (ed.), PER TERRAM ‑ PER MARE, Seaborn Trade and the Distribution of Roman Amphorae in the Mediterranean. Studies in Mediterranean Archaeology and Literature 180: 1–20. Uppsala, Åströms förlag.

482

‘De profundis’: three amphorae of unorthodox contents retrieved from the Aegean Sea

Kozlovskaya, V. and Ilyashenko S. M. 2013. The Lower City of Tanais. In M. Manoledakis (ed.), Exploring the Hospitable Sea. Proceedings of the International Workshop on the Black Sea in Antiquity held in Thessaloniki, 21‑23 September 2012. British Archaeological Reports International Series 2498: 83– 94, Oxford, Archaeopress. Krapivina, V. V. 1993. Ol’biya. Material’naya kul’tura I‑IV vv. n.e. Kiev. Kruit, N. and Worp, K. 2000. Geographical jar names: Towards a multi‑disciplinary approach. Archiv für Papyrusforschung und verwandte Gebiete 46, 1: 65–146. Lawall, M. L. 2010. Transport amphoras. In N. A. Lejpunskaja, P. Guldager Bilde, J. Munk Hojte, V. V. Krapivina and S. D. Kryziskij (eds), The Lower City of Olbia (Sector NGS) in the 6th Century BC to the 4th Century AD: 355–405. Aarhus, Aarhus University Press. Lawall, M. L. 2011. Socio‑economic conditions and the contents of amphorae. In Ch. Tzochev, Ch., T. Stoyanov and A. Bozkova (eds), PATABS II, Production and Trade of Amphorae in the Black Sea. Acts of the International Round Table held in Kiten, Nessebar and Sredetz, September 26‑30, 2007: 23–33. Sofia, Bulgarian Academy of Sciences/St. Kliment Ohridski University of Sofia. Lawall, M. L. and Jawando, A. 2004. Amphoras as paint pots? Hesperia 71: 416–419. Mills, J. S. and White, R. 1989. The identity of the resins from the Late Bronze Age shipwreck at Ulu Burun (Kaş). Archaeometry 31: 37–44. Müller, G. J. 1986. Review of the hitherto recorded species of Crustacea Decapoda from the Bosphorus, the Sea of Marmara and the Dardanelles. Cercetari Marine 19: 109–130. Neilson, H. R. 2002. A terracotta phallus from Pisa Ship E: more evidence for the Priapus deity as protector of Greek and Roman navigators. International Journal of Nautical Archaeology 31, 2: 248–253. Opaiţ, A. 2007. A Weighty Matter: Pontic fish amphoras. In V. Gabrielsen and J. Lund (eds), The Black Sea in Antiquity, Regional and Interregional Economic Exchanges. Black Sea Studies 6: 101–121. Aarhus, Aarhus Universitetsforlag. Opaiţ, A. 2010. Aspects of the provisioning of the Pontic settlements with olive oil in the Hellenistic and early Roman periods. In D. Kassab Tezgör and N. Inaishvili (eds), PATABS I. Production and Trade of Amphorae in the Black Sea. Actes de la Table Ronde international de Batoumi et Trabzon, 27‑29 avril 2006.Varia Anatolica 21: 153–158. Istanbul, Institut français d’Études Anatoliennes‑Georges Dumézil. Panagou, T. 2016. Transport amphoras and their contents. In M. Giannopoulou and C. Kallini (eds), ἠχάδιν ‑ Honorary Volume for Stella Drougou, Vol. II: 312–334, Athens, T.A.P.A.

Papadopoulos, J. K. and Paspalas, S. A. 1999. Mendaian as Chalkidian wine. Hesperia 68: 161–188. Parker, A. J. 1992. Ancient Shipwrecks of the Mediterranean and the Roman Provinces. British Archaeological Reports International Series 580, Oxford, Tempus Reparatum. Peña, J. T. 2007. Roman Pottery in the Archaeological Record. Cambridge, Cambridge University Press. Pulak, C. R. and Townsend, R. F. 1987. The Hellenic shipwreck at Serşe Limani, Turkey: Preliminary report. American Journal of Archaeology 91: 31–57. Radulescu, A. V. 1973. Amfore cu inscriptii de la edificiul roman cu mozaic. Pontica 6: 193–207. Reese, D. S. 2000. Iron Age shell purple‑dye production in the Aegean. In J. W. Shaw and M. C. Shaw (eds), Kommos IV. The Greek Sanctuary: 643–646. Princeton, Princeton University Press. Robinson, H. S. 1959. The Athenian Agora. Volume V. Pottery of the Roman Period, Chronology. Princeton, The American School of Classical Studies at Athens. Solinas, E. 1993. Uno scavo nell’ anfora: Stratificazione millenarie nei contenitori. Archeologia Viva 37: 32. Steckner, C. 1989. Les amphores LR 1 et LR 2 en relation avec le pressoir du complexe ecclésastique des thermes de Samos. In V. Déroche and J. M. Spieser (eds), Recherches sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 57–71. Athens, École française d’Athènes. Van Alfen, P. G. 1996. New light on the 7th‑c. Yassi Ada shipwreck: Capacities and standard sizes of LRA1 amphoras. Journal of Roman Archaeology 9: 189–213. Van Doorninck Jr., F. H. 1989. The cargo amphoras on the 7th century Yassi Ada and 11th century Serce Limani shipwrecks: Two examples of a reuse of Byzantine amphoras as transport jars. In V. Déroche and J. M. Spieser (eds), Recherches sur la céramique byzantine. Bulletin de Correspondance Hellénique Supplément 18: 247–257. Athens, École française d’Athènes. Van Doorninck Jr., F. 2005. The ship of Georgios, priest and sea captain: Yassiada, Turkey. In G. F, Bass (ed.), Beneath the Seven Seas: Adventures with the Institute of Nautical Archaeology: 92–97. New York, Thames & Hudson. Voultsiadou, E. and Vafidis, D. 2007. Marine invertebrate diversity in Aristotle’s zoology. Contributions to Zoology 76, 2: 103–120. Ward, C. A. 2015. Plant remains from the old jars on the Byzantine ship at Yassiada. In N. D. Carlson, J. Leidwanger and S. M. Kampbell (eds), Maritime Studies in the Wake of the Byzantine Shipwreck at Yasiada, Turkey: 55–62. College Station, Texas A&M University Press. Will, E. L. 1987. The Roman amphoras. In A. M. Mc Cann (ed.), The Roman Port and Fishery of Cosa. A Center of Trade: 170–220, Princeton, Princeton University Press. Zeest, I. B. 1960. Keramičeskaja tara Bospora. Materialy i issledovanija po archeologii SSSR 83. Moscow, Institut arheologii.

483

Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence from the Punic Amphora Building (Corinth, Greece) Antonio M. Sáez Romero

Departamento de Prehistoria y Arqueología, Universidad de Sevilla Facultad de Geografía e Historia, Calle Doña María de Padilla s/n, 41004, Sevilla

Tatiana Theodoropoulou

Equipe GReNES, CEPAM-CNRS (UMR7264) Université Côte d’Azur, Pôle Universitaire Saint Jean d’Angély-SJA 3, 24 avenue des Diables Bleus, 06300 Nice

Abstract: Although archaeoichthyological studies from the Aegean Sea are increasing, the discovery of zooarchaeological remains linked to processed fish from the Classical and Roman Antiquity remains scarce. The faunal material found in the so-called Punic Amphora Building at Corinth (dating mid-5th century BC), together with hundreds of Punic transport amphorae, has been interpreted as the evidence of considerable trade of salted bluefin tuna (tárichos) developed between the Strait of Gibraltar region and the Eastern Greek cities as early as the 5th century BC. More western Punic amphorae found elsewhere in the Aegean, Sicily and Magna Graecia, as well as some quotes in the Greek classical sources of the 5th-4th centuries BC confirm the magnitude of that commercial and gastronomic connection. A recent review of both the amphorae and the faunal material from the tavern allow for a detailed account of the fish remains and also an experimental replication of the salting process. In addition, the analysis of the amphorae from the Corinthian building and other sites in the Central and Eastern Mediterranean supports the hypothesis of a commercial struggle between the western Punic cities and the Carthaginian fish by-products at least from the late 5th century BC. Key words: Ancient Corinth; Gadir; Maritime Archaeology; Tuna; Carthage.

1. Introduction The so-called Punic Amphora Building (PAB) at Corinth holds an important place in the history of Mediterranean trade in Antiquity, and especially the long-distance trade of salted fish. The building features among the earliest evidence found in Classical Greece of trade and consumption of the renowned processed tuna produced by the western Punic cities of the Strait of Gibraltar region. Even though it can be considered as a key context for the study of the rise of the international production and consumption of salted fish during the Classical period and for the analysis of the relationships between the Greeks and the Punic Far West. Only brief accounts of the amphora types and fishbones were included in the preliminary reports of the excavations (Williams 1978, 1979 and 1980). The amphorae were classified in two main typological groups, and the first archaeometric analysis on the western Punic vessels (Maniatis et al. 1984) indicated two different groups of fabrics. Fish remains, essentially consisting of packs of scales, scarce vertebrae and cranial bones, were attributed to bluefin tuna and gilthead sea bream (Zimmerman-Munn 2003). These preliminary studies immediately turned the PAB into an international reference in relation to the study of tuna fishing and its consumption in the form of preserved and salted fish, and in general for the study of longdistance trade in Classical times. The interpretation of

the building as a small tavern specialized in the retailing of quality Greek wines, mainly from Chios, and salted fish, makes this the only establishment of these characteristics known from Classical Greece to this day (Kelly-Blazeby 2006; Glazebrook and Tsakirgis 2016). The material remains linked to the trade and consumption of salted fish products offer an exceptional archaeological case that confirms the relationship between the Phoenician and Greek worlds and traces down the beginning of the development of a ‘globalized market’ and millenarian Mediterranean cooking patterns. Between 2014-2021 a new research program has been conducted on the material remains from the PAB contexts, including both fishbones and the rest of the finds documented inside the building. In collaboration with the staff of the American School of Classical Studies at Athens (ASCSA), a first short visit to the site during the summer of 2014 aimed at evaluating the possibility of resuming the study of the unpublished remains of the Punic Amphora Building. During that launch phase, a direct examination of part of the archaeological material documented in the excavations of 1977-1979 was carried out, focusing on the western Punic amphorae and the fish remains. The importance of the observations that stemmed from the study of that selected set of amphorae and associated fishbone assemblage led to implementing an integrated project that, in the medium term, would result in the publication of the final excavation report.

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 485–497

Antonio M. Sáez Romero and Tatiana Theodoropoulou

Following that first sample study, four subsequent research seasons were carried out (2015-2016 and 2018‑2019) with the aim of including all the pottery finds and faunal remains of the 1977-1979 excavations. Due to limited funding, the short 2015 summer season mostly focused on the examination of the potential of the deposits in relation to the study of the Punic amphorae that support the existence of fluid commercial networks in classical times between Greece and the western Punic cities. Hundreds of fragments were processed and dozens of individuals classified as T-11213 amphorae. The longer study seasons carried out in 2016 and 2018‑2019 allowed us to process the pottery of the PAB contexts, namely Punic and Greek amphorae, and to study other associated evidence (metal and lithic finds, coins, etc.). A specific part of the work carried both at Corinth, and in the ASCSA headquarters in Athens was dedicated to the collection and examination of the excavation archives, including photographs and excavation journals, numismatic archives, blueprints and plans. During the 2016 and 2018-2019 seasons, the information was digitally processed, which allowed producing a preliminary version of the 3D reconstruction of the building (Figure 1), as well as a photogrammetric documentation of selected pottery material. During the four study seasons completed up to date, no less than five thousand objects have been drawn, classified, inventoried, and placed within their stratigraphic contexts. Thanks to the recording of high-resolution information about the building and its stratigraphy, it is henceforth possible to identify the original position of many of the items recovered in the interior and exterior spaces of the PAB. The main goal of this first presentation of the results of the research program described above, will focus on the connection between the Punic amphorae and the fish remains, as well as on the exceptional evidence for determining the nature of the fish products transported by the Punic vessels. The paper provides a summary of previous research, some preliminary observations drawn from the study of the Punic transport vessels, a detailed account of the fish remains and some discussion about the results of the digital modelling and experimental archaeology conducted during the research project. Other major aspects, such as the architectural features of the building, the distribution of the finds, the function of the rooms as well as the discussion of the chronological framework, will be addressed in forthcoming articles (Sáez, Theodoropoulou and Belizón 2020). The pottery and archaeoichthyological results are discussed and further contextualized through literary and historical evidence. The analysis of the amphorae finds from the building and other related locations within the ancient city of Corinth supports the hypothesis of a commercial clash between the tárichos (τάριχος) produced in the western Punic cities and the Carthaginian salt-fish products. Thanks to the

increasing datasets of amphorae typologies and fabrics, archaeoichthyological studies, and recent discoveries of several fish-salting facilities and pottery workshops in the western and Sicilian Punic cities, it is henceforth possible to re-evaluate earlier interpretations of the finds from the PAB and to explore new paths in the analysis of ceramic vessels and their contents in ancient Mediterranean. 2. A renowned stranger. Historiographical preface Although archaeoichthyological studies from the Aegean Sea have increased over the last decades, evidence of processed fish from the Classical and Roman Antiquity remains scarce (for a recent overview of research and specific studies in the ancient Mediterranean, see the special issue of the Journal of Mediterranean Archaeology, Mylona and Nicholson 2018; on the potential and methodological limits of identifying processing practices on fish remains, Theodoropoulou 2018). The earliest faunal material hitherto found in Greece comes from the so-called Punic Amphora Building at Corinth, and is dated to the mid-5th century BC. Corinth was one of the most important city-states in classical Greece and a wealthy trading and artisanal centre as early as the Archaic period (7th-6th centuries BC). Corinth’s location in the centre of the Greek peninsula is crucial in explaining its importance as a commercial hub. The construction of a paved trackway for the boats in the 6th century BC further facilitated the passage of ships and goods from the Adriatic towards the Aegean Sea through the Corinthian Gulf. The flourishing Corinthian export trade (oil, perfumes, wine) has been attested through the presence of both fineware containers and Corinthian transport amphorae throughout the Mediterranean, while imported transport amphorae and other items unearthed in Corinth illustrate the regular supply of products from the Aegean, the Levant and the CentralWestern Mediterranean. The mid-5th century BC irregular building, excavated in the late 1970s by the ASCSA in the area of the South Stoa (south-western corner of the Roman Forum), which is the focus of this paper, exemplifies in the best way the commercial circuits in the city. Perfectly located in the intersection of two major roads, it comprises diverse rooms distributed around a central courtyard. The first publication of the PAB assemblage underlined the commercial role of the building, mainly due to the presence of the amphorae in the open courtyard, but also in other contexts in the surroundings (Williams and Fisher 1976; Williams 1978, 1979 and 1980). According to excavation reports, two phases in the life of the building could be identified: a simple establishment with some commercial dealings during the first phase (shortly after 470 BC), which later was renovated and turned into a prosperous business devoted to the merchandising of local and imported foodstuff. This second phase developed probably into the beginning of the last quarter of the 5th century BC, and is represented by two strata.

486

Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence…

Figure 1. The central area of Ancient Corinth, with the temple of Apollo at the forefront and the skyline of Acrocorinth in the background (above); Plan of the excavated remains of the Punic Amphora Building (after Williams 1980) (bottom left); and recreation of the building in its second phase (model developed by R. Belizón Aragón ARQ-TECH for the Corinth Punic Amphora Building Project). The finds recovered from the PAB offer an exceptional dataset due to the good state of preservation and the large quantity of amphorae and fish remains uncovered. The amphorae and the fish remains were mainly found connected as part of the consecutive pavement layers of the courtyard, mixed with other Greek imports (wine amphorae and fineware pottery) as well as some ‘holemouth jars’ (identified as T-1451/T-4226 Carthaginian amphorae in Ramon 1995: 146). Archaeometric analysis on the western Punic vessels (Maniatis et al. 1984) indicated the existence of two different groups of fabrics for the western amphorae, suitable for wet and dry contents, and suggested diverse production areas from the Atlantic area of the Strait of Gibraltar region. A brief account of the fishbone material was included in the preliminary publication of the archaeological assemblage (Zimmermann-Munn 2003). Fish remains, essentially consisting of packs of scales, scarce

vertebrae and cranial bones, were primarily attributed to tuna and gilthead seabream. The association of the fishbone remains with Punic transport amphorae was interpreted as evidence of considerable trade of tárichos between the West and the Eastern Greek cities as early as the mid-5th century BC (Zimmermann-Munn 2003). The extent of those trading links has been further substantiated both by the discovery of more western Punic amphorae at Olympia and Athens and by several mentions in 5th century BC Greek textual sources (Sáez Romero 2014). 3. The Punic amphorae at Corinth in the 5th century BC The preliminary study of the PAB contexts underlined the presence of significant quantities of ‘Maña Pascual A4’ (Ramon T-11213) amphorae from the Punic western coastal

487

Antonio M. Sáez Romero and Tatiana Theodoropoulou

cities. These vessels were used to transport the famous tuna tárichos from τὰ Γάδειρα (Gadir, present-day Cadiz). Hundreds of Punic amphorae were crushed once empty, and used to build the overlapping layers of the floors in the central courtyard. As mentioned above, the first archaeometric analysis of the western Punic vessels (Maniatis et al. 1984) suggested the existence of two different groups of fabrics for the western amphorae, pointing to diverse production areas. But the lack of evidence for western Punic pottery workshops in the early 1980s did not allow for a more accurate provenance study. Since the early 1980s, only sparse additional information regarding the PAB amphorae assemblage has been published, with the focus on the typology and origin of the ‘hole-mouth jars’ or ‘Maña D’ vessels (ZimmermanMunn 2003: 202). The pottery study carried out between 2014 and 2019 extended to the examination and classification of most of the material from the PAB, and thus allowed to establish the typology of Punic and Greek amphorae present in the building. With respect to the Punic transport vessels, we confirmed that most of the individuals belong to variants of Ramon’s group T-11210 (Ramon 1995: 234–237). The latter shows the morphological diversity of a series fired in various workshops in the Strait of Gibraltar region. Additionally, the presence of containers manufactured in Punic workshops of western Sicily has been confirmed (Fantuzzi et al. 2020). These are variants of types T-1451 and/or T-4226, presenting slight differences with each other. The PAB contexts, abandoned around 430/420 BC, suggest that Carthaginian and western Sicilian amphorae were shipped to Corinth from the mid-5th century BC, but we cannot be sure if they shared the same routes and ships with the western amphorae, and whether they were carrying salted fish or a variety of products. In any case, the western imports seem to be quantitatively dominant (c. 95,5% of Punic imports, c. 360 vessels), and only a few (16) Carthaginian individuals have been identified. Besides the typological and stratigraphical study of the amphorae, the 2014-2019 project included extensive sampling in order to conduct new archaeometric analysis of their fabrics, which constitute the main specificity of PAB and, in general, of the remains of the Classical period at Corinth. The laboratory work was carried out in 20182019 at the Fitch Laboratory of the British School at Athens, directed by L. Fantuzzi and E. Kiriatzi (Fantuzzi et al. 2020). The samples included amphorae from all macroscopically observed fabric groups (i.e., four main western groups with up to eleven variants, and two Central Mediterranean clusters), in order to compare them with samples of ceramic workshops excavated in the Bay of Cadiz, the Andalusian Mediterranean coast and the northern Moroccan shore. The importance of this chemical and petrographic record is multifold. It offers a definite identification of the production workshops of the amphorae found at Corinth — and, therefore, of the

provenance of the salted tuna. Thanks to the significant number (c. 200 individuals) and variety of the samples, it also aspires to constitute an international reference collection for the study of the western Punic amphorae. 3.1. The Western Punic amphorae During the 2014-2019 research program, around 360 transport vessels produced in the Strait of Gibraltar area (the so-called ‘Círculo del Estrecho’) have been studied and compared with samples from Malaga, Morocco and Cadiz pottery workshops. Thanks to this preliminary classification it was possible to further refine the origin of the vessels. Most of the individuals present typical typologies and fabrics of the Ramon T-11210 group (Ramon 1995), widely attested in pottery workshops and fish-salting industries in the region during the 5th and early 4th centuries BC. All western individuals can be classified as distinct variants of the T-11210 series (Figure 2), but in the light of the distinctive ceramic fabrics observed, it seems quite clear that many western Punic cities would have been involved in the salt-fish trade with the Greek world. The archaeometric study (Fantuzzi et al. 2020) has confirmed that at least two thirds of the amphorae consumed in the PAB came from the Bay of Cadiz (Gadir, later Roman Gades), while the rest of the western transport vessels were produced in workshops located in Malaga or on the coast of Velez-Malaga, while only a few samples are outliers from unidentified workshops. During the 5th century BC, Gadir was a prosperous city, involved in the intense international commercial networks linking the Atlantic with Carthage, Sicily, Greece and the Near East. The salted fish products from this Atlantic port were indeed famous during the Classical period in mainland Greece as literary sources testify (e.g., Hippocrates, De internis affectionibus 25; Eupolis, Márikas 199; Aristophanes, Batrach 470; Nikostratos apud Athenaeus, Deipnosophistae 3. 118e; cf. López Castro 1997; García Vargas and Ferrer 2006 and 2012), and amphorae from Gaditan workshops have also been identified at Athens and Olympia (on the Gaditan kiln sites and fish processing facilities, see Ramon et al. 2007; Sáez Romero 2013 and 2014). But the shipments that arrived at Corinth included a wider variety of western fish preserves, and it seems very likely that other regional coastal cities developed their own particular recipes for salted fish and fish sauces, as suggested by López Castro (1993). A significant quantity of the amphorae found in the PAB had been fired and probably filled in the Malaga coastal area and other Mediterranean workshops of the Mediterranean coast of southern Iberia. Some vessels were possibly shipped from Algeciras Bay (Carteia) or the coast of present-day Morocco, in port hubs such as Lixus or Kouass, as evidence of local amphora production suggests (Kbiri Alaoui 2007; Sáez Romero 2011). Finally, few early examples of the T-11216 variant have been identified, probably also

488

Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence…

Figure 2. Western Punic T-11213 amphorae from different contexts of the PAB, including rims and upper parts (1-4), handles (5) and spikes or bottom parts (6-7).

489

Antonio M. Sáez Romero and Tatiana Theodoropoulou

Figure 3. Examples of painted inscriptions (in red) documented in some of the T-11210 individuals from the PAB contexts. produced in workshops located in Malaga Bay. Major ports such as Gadir, Carteia, Malaka, Lixus, Maenoba, Abdera or Seks may have been involved in this commercial network that linked directly the Eastern markets with the Punic West — although the literary evidence only refers to the arrival and consumption of Gadir’s tárichos. During the primary processing of the pottery material we had the chance to come accross some exceptional finds found at PAB, namely the abundant graffiti engraved on Punic and Greek amphorae. During the 2015 study season, the presence of sets of tituli picti painted in red on some western Punic amphorae was identified thanks to a well-preserved individual of the type T-11216 (Figure 3). The study of additional material during the 2016, 2018 and 2019 seasons offered at least 15 more examples of these dipinti, painted on amphorae produced in different cities of the Strait of Gibraltar region. These finds are of major historical significance, as they constitute the oldest painted inscriptions known on this amphora series and a unique case for the Western Mediterranean Punic amphoras. Their discovery may provide new insights into the products transported by each variant of the Western Punic amphorae consumed in the PAB. Summing up, the group of western Punic amphorae from the PAB is a valuable dataset to piece together the main commercial trends that supplied the Greek mainland of western salted fish products. The data obtained during the 2014-2019 study program confirm that during the early and mid-5th century BC the western Punic cities of the Strait of Gibraltar region were shipping significant amounts of fish by-products packaged in T-11210 amphorae to the Central and Eastern Mediterranean Greek main cities. These products also arrived in Sicily and Magna Graecia (Ramon 1995; Castiglione and Oggiano 2011; Bechtold and Vasallo 2018), as well as Athens (Lawall 2006), Olympia (Gauer 1975) and — apparently in larger quantities — Corinth, where the western salt-fish were consumed and distributed through its port and in specialized kapeleia, such as the Punic Amphora Building.

3.2. Other Punic amphorae found at the PAB The recent study of the amphorae from the PAB and other excavated contexts in the vicinity of the building (e.g., the buildings located east to the PAB; see McPhee and Pemberton 2012) supports the hypothesis of a commercial struggle between the western Punic cities and the Carthaginian salted fish by-products. Together with the western imports from the Strait of Gibraltar area sensu lato, some other Punic products were traded in the PAB during the 5th century BC. Most of these amphorae were produced in western Sicily — most likely in Solus and/or Panormos — and perhaps, in minor quantities, in northern Tunisia. All these transport containers can be classified as Ramon T-1451 or T-4226 amphorae (Ramon 1995; Bechtold 2015), although most of the individuals are only represented by small portions of rims or handles and none of them preserved stamps, graffiti or painted inscriptions. The archaeometric analysis conducted at the Fitch Laboratory has also confirmed the provenance of the so-called ‘hole-mouth jars’, namely that some of the products traded/consumed in the PAB were produced in Sicilian ateliers (Figure 4). This is not surprising, as the increasing dataset of amphorae typologies and fabrics and the excavation of several fish-salting facilities and pottery workshops over the last decades in the western Sicilian Punic cities (Botte 2009) support the idea that both Punics and Greeks were involved in the fish-processing business in the region. It could further be suggested that the fish by-products were traded to Corinth and other cities in Greece and consumed from the 5th century BC onwards. Besides the troubles caused by the Peloponnesian War and the Athenian maritime blockade, the presence of these Carthaginian amphorae may add another significant clue to explain the abandonment of the Punic Amphora Building before the end of the 5th century BC. The transport of the Sicilian Punic amphorae during the late

490

Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence…

Figure 4. Carthaginian amphorae probably produced in western Sicily Punic workshops (1-3), and Chiot amphorae with red painted strips and inscriptions (4-6). Classical period to Corinth and the Eastern Mediterranean can be viewed within a troubled political and military situation between Carthage and the Greek cities of the Central Mediterranean, which culminated during the Carthaginian invasion of Sicily in 409-405 BC. An increased presence of fish-processing installations in NW Sicily also dates to the late 5th and the 4th centuries BC (Botte 2009). The growing influence of Carthage in the area during the second half of the 5th century BC may have contributed to the decline and deactivation of the commercial connection of the Gaditan area with the Greek East, and was probably beneficial for the development of the Palermo-Solunto fish-processing industry (Docter and Bechtold 2013). Several Central Mediterranean sherds found in some contexts at Corinth and in Athens also seem to support the continuation of Carthaginian long-distance trade towards the East during the late 5th and the 4th‑3rd centuries BC. It thus seems that there was a sudden substitution of western Punic imports by Carthaginian commodities around 425-

400 BC, as to this day no western amphorae dating from the 4th-3rd centuries BC have been attested at Corinth. 4. From bones to fish: species identification, body part representation One of the main goals in the 2014-2019 study program was the combination of the results of pottery analysis with the information provided by the actual remains of products stored and transported in them, i.e. fishbones. The PAB offers a unique opportunity to obtain a detailed insight into the latter, through the abundant and quite consistent fishbone assemblage recovered from the building (on an overview of relevant data from ancient Greece, see Theodoropoulou 2014 and 2018). The fish remains from the PAB were mainly recovered during the 1977-1979 excavations, scattered together with Punic amphora fragments in the courtyard and

491

Antonio M. Sáez Romero and Tatiana Theodoropoulou

porch, where the amphorae had probably been shattered and discarded, then covered with lime to form overlying floors. Fish remains have also been found in rooms 4, 5 and 6 of the building. No wet sieving was employed, though systematic dry sieving of the soils took place during the excavation. Previous publications of the PAB material (Zimmerman-Munn 2003) provided only brief accounts of the fishbones that supported the presence of salted fish products in the building, mainly batches of tuna scales, as well as some other fishbones attributed to seabream. The faunal analysis conducted in 2014-2015 at the archaeological museum of Ancient Corinth consisted in providing a complete taxonomic and anatomical identification of the marine remains, the recording of various marks, and allowed further refining issues of provenance, types of products and distribution. The assemblage is quite unique in its anatomical representation, as it mainly consists of batches of scales as well as free scales and other post-cranial elements, such as ribs/fins. In total, 35 packs of scales and 13 more fragments of packs were counted, cut in various shapes (rectangular, square, some triangular). Most of the chunks have sharply-cut edges, while at least five pieces preserve isolated cut marks on some scales. Furthermore, approximately 2000 scattered scales and a similar number of fins/spines of different sizes were retrieved across the building. Among the free scales, at least eleven specimens preserve various types of cut marks, all produced with a thin-blade instrument: a single longitudinal incision, two or several parallel incisions, or a combination of transverse (horizontal and vertical) cuts. In one case, some fin spines were found attached to the scales, while two pottery fragments preserved part of a pelvic or anal fin on their internal side.1 Other fish remains include a few vertebrae and various isolated cranial elements, as well as an association of right-side superficial cranial bones (gill to eye area) and upper suspensorium (upper jaw area) still in anatomical connection. With the exception of the latter, most frequent anatomical elements, i.e., scales and rib/fins are not easily identified to the family, genus or species level. This dominant group of remains has been identified as belonging to the Scombridae family, i.e., genus Thunnus sp., based on their morphological traits as well as size. However, some tuna species, particularly the albacore, present a high polymorphism of the scales within one individual, depending on their position on the body of the fish (Meunier and Sire 1981). This observation makes it particularly difficult to offer a secure species identification  Unfortunately, as the pottery and faunal remains used to build the overlapping floors of the courtyard were mixed, apparently without following any order, we cannot be sure about the connection between the amphora sherds and the fins as incovertible evidence of the original contents of the Punic transport vessels. This is beyond doubt very reliable chronological evidence, as it links both amphora sherds and fish remains, but it cannot be used as proof of the relationship of certain types with salted tuna chunks. 1

with standard zooarchaeological analysis. For this reason, a sampling of scales and other superficial body parts is scheduled for ZooMS analysis (Zooarchaeological Mass Spectrometry), in collaboration with the University of York. Further remains belonging to the Scombridae family included 2 vertebrae from medium-sized Thunnus sp. individuals (80-100cm), 2 caudal vertebrae and 1 urostyle of little tunny (Euthynnus alletteratus), as well as few posterior cranial bones and bones from the branchial arch, also attributed to the Thunnus genus. Besides Scombridae, other fish families are also present among the studied material: • Gilthead seabream (Sparus aurata): at least seven individuals have been identified (based on premaxillaemaxillae counts). As noted above, the remains also include one articulated right-side operculum/jaw zone and one almost complete set of hyomandibular zone, as well as 2 articulated pre-caudal rachis. • Other coastal fish identified in the material include: thinlip grey mullet (Liza aurata, 2 vertebrae), seabass (Dicentrarchus labrax, 1 cranial), white grouper (Epinephelus aenaeus, 1 cranial bone, 5 articulated pre-caudal vertebrae, one cut vertebra) and wrasse (Labrus viridis, 1 vertebra). • Furthermore, 73 Clupeid vertebrae (Sardina pilchardus), associated to several hundreds of tiny scales from the same contexts. Finally, two invertebrate remains have been identified among the fishbone material: 1 gastropod shell Nassarius cuvierii, and 1 fragment of sea-urchin plate incrusted on a Sparus aurata premaxilla. The remains could be associated with the presence of this fish in the assemblage, as adult gilthead seabreams are carnivorous and often feed on various invertebrates, the remains of which can be found in the stomach contents of captured fish. In the following concluding section, the observations made on the archaeo-ichthyological material are presented in conjunction with the synthesis of pottery data as well as relevant literary/historical references, in order to address issues related to the type of products, transport and consumption practices in classical Corinth during the 5th century BC. 5. Discussion. From the sea to the tavern: products, consumption and trade The fish remains from the PAB provide solid material evidence to suggest that diverse types of fish and fish by-products were transported, traded and consumed in the building centrally located in the classical city. The PAB, probably a modest two-floor building, was a vibrant specialized establishment that offered a selected variety of imported high-quality, presumably expensive, wines

492

Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence…

and fish. Among them, the renowned Punic salted tuna coming from beyond the Pillars of Heracles would have been one of the most sought-after commodities (García Vargas and Ferrer 2012; concerning the price of fish in Classical Athens, see Acton 2014: 232). The discovery of dozens of Corinthian and Attic glazed drinking vessels, particularly kylikes, kotylai and skyphoi, imply that both local and imported pottery vases were used to serve wine, whereas local plain and painted wares would have been used to set the table, cook and serve food, particularly fish. All things considered, this may have been an exclusive tavern, located not far from the main public areas of the city, in which the Corinthian elite would taste luxurious food and show off their wealth (the opsophagoi pejoratively mentioned in some literary sources; cf. Davidson 1995; García Vargas 2008). Coming to the nature of the fish products, the recent faunal analysis allows suggesting that more than one type of products are possibly present. The scarcity of tuna remains other than the batches of scales described above (cf. below suggestions for other body parts), seems to confirm that the main tuna product sold and consumed in the PAB were fish fillets, many of which had their scales left on the skin (λεπιδωτòν), although the presence of scale-free fillets (τιλτòν) — which would thus leave no remains — cannot be excluded. The shapes of the sliced packs identified fit nicely with known types of salted products, mentioned in Greek and Latin texts: triangular (τρίγωνον), square (τετράγωνον) and cubic (κύβιον) (Figure 5). These distinctions may also point towards the anatomical location of these fillets on the body of the fish, such as the lower part of the body, the belly (ὑπογάστρια), or the shoulder, for which ancient Cume was famous (ὠμοτάριχος — however the word also refers to the method of pickling in some references; on fish products from Cume, see Mylona 2008; Botte 2009). These are the most praised parts according to Pliny (NH IX, 48) and Athenaeus (Deipn. 3.120b-121c). The presence of fins in the material suggests that some fillets would have come from areas of the fish body with fins (probably the pectoral, pelvic and anal), which would have been left on the fillets during processing. The presence of cranial bones and branchial arch bones from Thunnus sp. is somewhat problematic, as usually these parts should have been removed from the fish prior to the preservation process, but some could have been partly left on slices situated near the head. However, the possibility of freshly caught tuna also being supplied by Greek fishermen and consumed along with the salted fish products in the Corinthian tavern cannot be discarded. Although ancient authors often make further distinctions of quality according to the size of tunas (e.g. θυννίδες, salted young tunnies), the state of preservation of the scales does not allow for age reconstructions based on the observation of growth annuli (on the method, see Wheeler and Jones 1989: 139–148). The size of both the packed scales and the scattered scales do not allow

adding further information regarding the age/size of the fish. Even though almost all size ranges have been recovered, the morphological particularities of tuna scales do not allow for a straightforward correlation of small scales/small tuna individuals vs. bigger scales/ bigger individuals. As different body parts of a fish produce scales of different size (Meunier and Sire 1981), various sizes could belong to the same individual, i.e., bigger scales from near the belly and smaller ones from near the tale or head. Further issues related to the processing of fish products as well as transport are not easy to distinguish through faunal analysis. For example, the degree of salting of the PAB products could have been variable according to the descriptions preserved in ancient literary sources (slightly salted ἀκρόπαστον, half-cured ἡμιτάριχος/ ἡμίνηρον, fully salted τέλειος), differences that would have affected their taste and texture (Curtis 2001: 245; Mylona 2008) but cannot be observed on the fish bones or scales (Theodoropoulou 2018). The variety of shapes and sizes of the scale batches isolated from their containers does not allow to suggest whether the containers were filled with one specific size or with chunks of different types of fillets, or if the specific shape/size of the chunks was linked to a particular region in the Punic West. In order to obtain further insight into the type of the product and transport modes, an experimental archaeology project has been developed (the details and results of the experiment are presented in Sáez and Moreno 2017). The experiment aimed at reproducing salted tuna chunks and making estimations about various issues, such as volumetric features, relative weight of each slice before and after salting, duration of the salting processes, as well as assessing conservation problems. The experiment started on early June. A plastic container was filled with alternate layers of tuna slices with their scales left on and salt. The shape and size of the slices (8 × 4.5 × 13cm) was close to the size of the packs of scales found at the PAB. Each slice of this size weighted (432g) right after cutting the piece. After two weeks — i.e., the estimated minimum duration of the trip from the Straits of Gibraltar to Corinth by ship — under a thick layer of marine salt, the pieces looked darker in colour and tasted very salty, similar to the modern Andalusian mojama, a delicacy of filleted salt-cured tuna typical of the Huelva and Cadiz region (Figure 5). The slice size reduced to c. 7.5 × 3.5 × 11.7cm and 328g. Left in the brine one more week, there were no important changes in the flavour, size, weight or external colour of the chunks and, surprisingly the product was not completely dry (personal observations made by ASR). The recreation of ancient tuna tárichos has been developed within the framework of a broader project that includes the reconstruction of kiln sites and fish-processing infrastructures through 3D virtual models, in order to calculate their productive potential from a quantitative

493

Antonio M. Sáez Romero and Tatiana Theodoropoulou

Figure 5. Triangular (A) and rectangular (B-C) examples of scale packs recorded in the courtyard deposits of the PAB. Also, images of the replication of the salting process imitation the shape and size of chunk C: fresh cut of a larger piece of bluefin tuna from Cadiz coast (D) and salted fillet (E).

494

Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence…

perspective. Physical and digital reproductions of the Punic amphora types have also been developed on a fullscale size. In addition to calculating of Gadir’s capacity standards, this part of the project tackled other key issues, such as the weights of empty and filled amphorae, the use of the interior coating, and the sealing systems (some preliminary results can be found in Sáez and Moreno 2017; García Vargas and Sáez 2018). With respect to the transported product, it has been estimated that a typical T-11213 amphora would contain c. 28kg of cured tuna fillets and an equivalent quantity of salt. With respect to how the clients of the PAB consumed the tuna fillets, several hypotheses can be proposed, as a significant set of cooking wares was documented in the building. The tuna slices may initially have been desalted, and then boiled, cooked or fried using the ceramic pots and pans that were found mostly in the porch and the main courtyard. The preparation process could have been quite similar to the one described by Athenaeus: ‘First comes this piece of horaion [tuna cut in cubes]; that cost two obols. I must wash it well. Then I will sprinkle seasoning in a casserole, place the slice in it, pour over it some white wine, stir it in oil and stew it until it is as soft as marrow, covering it generously with a garnish of silphium’ (Ath. Deipn. 3.117d; see also Plautus, Poen. 240-244). A salted and dried product, similar to modern Andalusian mojama, could have been another variant of tuna products consumed at the PAB, presumably cut into thin slices or small cubes. The chunks and the rest of the fish could have been roasted on a skewer, or even oven-baked in the open-air areas of the building and then consumed in the main rooms, however no traces of fire are left on the remains. After being cooked or roasted, the salty tuna dish would have been consumed inside the building, presumably accompanied by Chian wine, as indicated by the large amount of Chian amphorae found in the same contexts in the courtyard of the PAB (Figure 4, 4-6). Once more, Athenaeus provides a clear testimony of this profitable business model established in the heart of Corinth during the 5th century BC: ‘as we ate our salt fish many of us had a desire to drink’ (Ath. Deipn. 3.120b). The archaeo-ichthyological study of the recent study program also put forward the trade and consumption of fish products other than salted tuna slices. The presence of small Clupeid vertebrae may also be suggestive of the transport of small fish to Corinth, either salted or as fish-sauce, as early as the 5th century BC, which could be a ‘local product’, i.e., produced in the Aegean (cf. the so-called θασία ἅλμη in Aristoph. Acharnai, 671). The presence of scales associated with vertebrae could point to a salted rather than a fermented product (Figure 5). The remains of bream, mullet and grouper could potentially be related to a fish preparation, either solid or sauce-like, for which various parts of different species would be added, similar to later Roman examples (Djaoui, Piquès and Botte 2014; Bernal-Casasola et al. 2014), also suggested for some central Mediterranean Punic

finds (Wilkens 2000). The presence of these common Mediterranean species may also indicate that fresh fish (cf. above the supposed stomach contents associated with seabreams), locally caught, were also sold and consumed in the establishment. Of major interest is the association of various types of fish products with the presence of Punic amphorae from different parts of the Mediterranean. A key issue that yet remains unsolved regards the content of the Carthaginian vessels classified as ‘hole mouth jars’ in the preliminary reports, namely whether those vessels were also carrying fish products — and of what type — to Corinth or other commodities. To conclude, more than refining the types and contents of salted products found at the PAB, the integrated study of the transport vessels and associated fish remains, although still in progress, has allowed to clarify some key aspects in our understanding of the implementation and specificities of a ‘globalized Mediterranean market’. Two major observations stem from this first five-year project. First, the presence of different fabrics among the amphorae originating from the Strait of Gibraltar region provides a solid indication that Gadir was not the only city in that area participating in this commercial connection between the West and the East. Second, the presence of 16 amphorae produced in Punic cities of the central Mediterranean among hundreds of western individuals (see Ramon 1995: 145–146) suggests two rather than one major sources of Punic supply, i.e. the Strait of Gibraltar area and Western Sicily. This crucial observation based on the study of the PAB contexts is meaningfully inscribed in a wider discussion of major historical interest, many essential aspects of which need to be further developed. First, can the Sicilian amphorae be considered a sign of the emergence of a significant network of salted-fish factories in that area, already in the 5th century BC? Did their transport to Corinth continue after the abandonment of the PAB? And second, were these amphorae brought to Corinth as part of the same commercial network that seems to have connected the Atlantic and the Aegean in a direct way, or was it the intermediation of Carthage — or powers like Syracuse — that made the existence of this lucrative relationship possible? The continuation of research on the Punic Amphora Building is expected to offer exciting new information in the near future, which extends beyond debates on the transport of salted fish in the hundreds of amphorae documented in their contexts. Acknowledgements We would like to thank Charles K. Williams, Nancy Bookidis, Ioulia Tzonou, Guy Sanders and Christopher Pfaff (ASCSA) for their encouragement to implement this research program and support throughout the study of the PAB material. Our gratitude extends to the staff of the Corinth museum, Nicol Anastassatou, Manolis Papadakis, and Panos Kakouros, for their assistance with everyday logistics, access to the storerooms and

495

Antonio M. Sáez Romero and Tatiana Theodoropoulou

handling of the trays and materials. ASR would like to thank Ricardo Belizón (Universidad de Sevilla) and James A. Herbst (ASCSA) for their contribution to the project, and in particular to the graphic and three-dimensional documentation of the objects and the building. ASR also benefited greatly from sharing of data regarding the results of the archaeometric study and enriching discussions about amphorae with Leandro Fantuzzi (Universidad de Cádiz), Evangelia Kiriatzi and Noémi Müller (Fitch Laboratory, BSA). TT would also like to thank Philippe Béarez and François Meunier (MNHN, Paris) for providing useful information and literature about tuna morphology. The project has received funding from the Vice-Rectorate for Research of the University of Seville and the American School of Classical Studies at Athens (H.S. Robinson Corinth Research Fellowship 2016). Bibliography Acton, P. 2014. Poiesis: Manufacturing in Classical Athens. Oxford - New York, Oxford University Press. Bechtold, B. 2015. Le produzioni di anfore puniche della Sicilia occidentale (VII-III/II sec. a.C.). Carthage Studies 9. Bechtold, B. and Vasallo, S. (eds) 2018. Le anfore puniche dalle necropoli di Himera (seconda metà del VII - fine del V sec. a.C.). Bulletin Antieke Beschaving (Babesch) Supplement 34. Leuven, Peeters Publishers. Bernal-Casasola, D., Cottica, D., García Vargas, E., Toniolo, L., Rodríguez-Santana, C. G., Acqua, C., Marlasca, R., Sáez, A. M., Vargas, J. M., Scremin, F. and Landi, F. 2014. Un contexto exceptional en Pompeya: la pila de ánforas de la Bottega del Garum (I.12.8). Avance de un estudio interdisciplinar. Rei Cretariae Romanae Fautorum Acta 43: 219–232. Botte, E. 2009. Salaisons et sauces de poissons en Italie du sud et en Sicile durant l’Antiquité. Collection du Centre Jean Bérard 31; Archéologie de l’artisanat antique 1. Naples, Centre Jean Bérard. Castiglione, M. and Oggiano, I. 2011. Anfore fenicie e puniche in Calabria e Lucania: i dati e i problemi. In M. Intrieri and S. Ribichini (eds), Fenici e Italici, Cartagine e la Magna Grecia. Popoli a contatto, culture a confronto. Atti del convegno internazionale (Cosenza, 27-28 maggio 2008). Rivista di Studi Fenici 36 (1-2): 205–231. Curtis, R. I. 2001. Ancient food technology. Technology and change in history 5. Leiden - Boston, Brill. Davidson, J. 1995. Opsophagia. Revolutionary eating at Athens. In J. Wilkins, D. Harvey and M. Dobson (eds), Food in Antiquity: 204–213. Exeter, University of Exeter Press. Djaoui, D., Piquès, G. and Botte, E. 2014. Nouvelles données sur les pots dits « à garum » du Latium, d’après les découvertes subaquatiques du Rhône (Arles). In E. Botte and V. Leitch (eds), Fish and Ships. Production and commerce of salsamenta during Antiquity. Actes de l’atelier doctoral, Rome 18-22 juin 2012. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17: 175– 197. Arles, Errance - Aix-en-Provence, Centre Camille Jullian.

Docter, R. F. and Bechtold, B. 2013. Two forgotten amphorae from the Hamburg excavations at Carthage (Cyprus, and the Iberian Peninsula) and their contexts. Carthage Studies 5: 91–128. Fantuzzi, L., Kiriatzi, E., Sáez Romero, A. M., Müller, N. S. and Williams, C. K. 2020. Punic amphorae found at Corinth: Provenance analysis and implications for the study of long-distance salt-fish trade in the Classical period. Archaeological and Anthropological Sciences 12: 179–200. García Vargas, E. 2008. Entre el consumo de lujo y el gusto popular: las salazones de la Iberia púnica y su romanización (ss. V-I aC). Una perspectiva histórica y cultural. In J. Napoli (ed.), Ressources et activités maritimes des peuples de l’Antiquité. Cahiers du Littoral 2(6): 87–108. Boulogne-sur-Mer, Centre de Recherche en Histoire Atlantique et Littorale, Université du Littoral Côte d’Opale. García Vargas, E. and Ferrer Albelda, E. 2006. Producción y comercio de salazones y salsas saladas de pescado del litoral andaluz en Época Fenicio-Púnica. Temas y problemas. In C. Tavares da Silva and J. Soares (eds), Simpósio Internacional Produção e comércio de preparados piscícolas durante a proto-história e a época romana no Ocidente da Península Ibérica. Homenagem a Françoise Mayet (Setúbal, 7-9 Maio 2004). Setúbal Arqueológica 13: 19–38. García Vargas, E. and Ferrer Albelda, E. 2012. Más allá del banquete: el consumo de las salazones ibéricas en Grecia (siglos V y IV a. C.). In B. Costa and J. H. Fernández (eds), Sal, pesca y salazones fenicios en Occidente. XXVI Jornadas de Arqueología FenicioPúnica: 85–121. Ibiza, Museo de Arqueología de Ibiza. García Vargas, E. and Sáez Romero, A. M. 2018. Todo el pescado vendido. Una lectura cuantitativa de la producción púnica y romana de sal y salazones en la Bahía de Cádiz. In J. Remesal, V. Revilla and J. M. Bermúdez (eds), Cuantificar las economías antiguas. Problemas y métodos. Instrumenta 60: 161–213. Barcelona, Universitat de Barcelona. Gauer, W. 1975. Olympische Forschungen VIII. Die Tongefässe aus den Brunnen unterm Stadion-Nordwall und im SudostGebiet. Berlin, De Gruyter. Glazebrook, A., and Tsakirgis, B. 2016. Houses of Ill Repute. The Archaeology of Brothels, Houses, and Taverns in the Greek World. Philadelphia, University of Pennsylvania Press. Kbiri Alaoui, M. 2007. Revisando Kuass (Asilah, Marruecos). Talleres cerámicos en un enclave fenicio, púnico y mauritano. Sagvntvm Extra 7. Valencia, Universidad de Valencia. Kelly-Blazeby, C. F., 2006. Kapeleion: Casual and Commercial Wine Consumption in Classical Greece. Unpublished PhD thesis, University of Leicester. Lawall, M. 2006. Consuming the West in the East: Amphoras of the western Mediterranean in the Aegean before 86 BC. In D. Malfitana, J. Poblome and J. Lund (eds), Old Pottery in a New Century. Innovating Perspectives on Roman Pottery Studies. Atti del Convegno Internazionale di Studi (Catania, 22‑24 Aprile 2004). Monografie dell’Istituto per i Beni Archeologici e Monumentali (IBAM) 1: 265–286. Catania, Consiglio Nazionale di Ricerche.

496

Salting and consuming fish in the Classical Mediterranean. A review of the archaeological evidence…

López Castro, J. L. 1993. La producción fenicia occidental de salazón de pescado. In II Congreso peninsular de História antiga (Coimbra, 18 a 20 de Outubro de 1990). Actas: 353–362. Coimbra, Universidade de Coimbra. López Castro, J. L. 1997. Los fenicios occidentales y Grecia. In F. J. Presedo, P. Guinea. J. M. Cortés and R. Urías (eds), II Reunión de Historiadores del Mundo Griego Antiguo. Homenaje al Profesor Fernando Gascó (Sevilla, 1821 diciembre 1995): 95–105. Sevilla, Scriptorium. Maniatis, Y., Jones, R. E., Whitbread, I. K., Kostikas, A., Simopoulos, A., Karakalos, Ch., and Williams II, C. K. 1984. Punic amphoras found in Corinth, Greece: an investigation of their origin and technology. Journal of Field Archaeology 11: 207–222. McPhee, I. and Pemberton, E.G. 2012. Corinth VII.6. Late Classical Pottery from Ancient Corinth: Drain 1971-1 in the Forum Southwest. Princeton, American School of Classical Studies at Athens. Meunier, F. J. and Sire, J. Y., 1981. Sur la structure et la minéralisation des écailles du Germon, Thunnus alalunga (Téléostéen, Perciforme, Thunnidae). Bulletin de la Société Zoologique de France 106(31): 327–336. Mylona, D. 2008. Fish-eating in Greece from the Fifth Century BC to the Seventh Century AD. A Story of Impoverished Fishermen or Lavish Fish Banquets? British Archaeological Reports International Series 1754. Oxford, Archaeopress. Mylona, D. and Nicholson, R. 2018. The Bountiful Sea: Fish Processing and Consumption in Mediterranean Antiquity. Proceedings of the international conference held at Oxford, 6-8 September 2017. Special issue of the Journal Maritime Archaeology 13(3). New York, Springer. Ramon Torres, J. 1995. Las ánforas fenicio-púnicas del Mediterráneo Central y Occidental. Instrumenta 2. Barcelona, Universitat de Barcelona. Ramon Torres, J., Sáez, A., Sáez Romero, A. M., and Muñoz, A. 2007. El taller alfarero tardoarcaico de Camposoto. Monografías de Arqueología 26. Sevilla, Junta de Andalucía. Sáez Romero, A. M. 2011. Alfarería en el Extremo Occidente fenicio. Del renacer tardoarcaico a las transformaciones helenísticas. In B. Costa and J. H. Fernández (eds), Yōserim: la producción alfarera feniciopúnica en occidente. XXV Jornadas de Arqueología FenicioPúnica (Ibiza, 22-26 de noviembre de 2010): 49–106. Ibiza, Museo Arqueológico de Ibiza. Sáez Romero, A. M. 2013. Talleres cerámicos en Gadir en época postcolonial ¿un modelo alfarero excepcional? In D. Bernal‑Casasola, L. C. Juan, M. Bustamante, J. J. Díaz and A. M. Saez (eds), Hornos, talleres y focos de producción alfarera en Hispania. Congreso internacional de la SECAH (Cádiz, 3-4 marzo de 2011). Monografías Ex Officina Hispana I (I): 215–249. Cadiz, Universidad de Cádiz.

Sáez Romero, A. M. 2014. Fish processing and salted-fish trade in the Punic West: New archaeological data and historical evolution. In E. Botte and V. Leitch (eds), Fish and Ships. Production et commerce des salsamenta durant l’Antiquité. Actes de l’atelier doctoral, Rome 18-22 juin 2012. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17: 159–174. Arles, Errance - Aix-enProvence, Centre Camille Jullian. Sáez Romero, A. M. and Moreno, E. 2017. Contando la historia. Experiencias de cuantificación y análisis volumétrico en centros artesanales púnicos de la Bahía de Cádiz. Archivo Español de Arqueología 90: 219–246. Sáez Romero, A. M., Theodoropoulou, T. and Belizón, R. 2020. Atunes púnicos y vinos egeos en una taberna de la Grecia Clásica. Resultados iniciales del Corinth Punic Amphora Building Project. In S. Celestino and E. Rodríguez (eds), Un viaje entre el Oriente y el Occidente del Mediterráneo. IX Congreso internacional de estudios Fenicios y Púnicos (Mérida 2018). Mytra 5: 799– 818. Merida, IAM-CSIC. Theodoropoulou, T. 2014. Salting the East: Evidence for salted fish and fish products from the Aegean sea in Roman times. In E. Botte and V. Leitch (eds), Fish and Ships. Production et commerce des salsamenta durant l’Antiquité. Actes de l’atelier doctoral, Rome 18-22 juin 2012. Bibliothèque d’Archéologie Méditerranéenne et Africaine 17: 213–227. Arles, Errance - Aix-enProvence, Centre Camille Jullian. Theodoropoulou, T. 2018. To salt or not to salt: a review of evidence for processed marine products in Greek antiquity and Aegean traditions. Journal of Maritime Archaeology 13(3): 389–406. Wheeler, A. and Jones, A. 1989. Fishes. Cambridge Manuals in Archaeology. Cambridge, Cambridge University Press. Wilkens, B. 2000. I resti ittici dal magazzino del IV-III sec. a.C. di Olbia. Rivista di Studi Punici 1: 81–88. Williams II, C. K. 1978. Corinth 1977: Forum Southwest. Hesperia 47: 1–39. Williams II, C. K. 1979. Corinth 1978: Forum Southwest. Hesperia 48: 105–144. Williams II, C. K. 1980. Corinth Excavations. Hesperia 49: 107–134. Williams II, C. K., and Fisher, J. E. 1976. Corinth 1975: Forum Southwest. Hesperia 45: 99–162. Zimmerman-Munn, M.-L. 2003. Corinthian trade with the Punic West in the Classical period. In C. K. Williams and N. Bookidis (eds), Corinth. The Centenary 18961996. Results of Excavations conducted by the American School of Classical Studies at Athens: 195–217. Princeton, Princeton University Press.

497

Conclusion It is difficult to speak of a ‘conclusion’ in such a fluctuating and developing field of research. The delay in publication means that the aims of the Roman Amphora Contents International Interactive Conference in October 2015 — to publish the latest research in the field — have been compromised by the more recent publication of several case studies and the launch of ambitious multidisciplinary research programmes. The question of amphorae contents, particularly during their intense diffusion in the Roman period, has underpinned all studies devoted to these artefacts, which are so abundant on all archaeological sites in the Mediterranean and beyond. It is therefore no coincidence that this conference has been dedicated in honour of one of the main pioneers of such studies in Spain, Professor Miguel Beltrán Lloris. Other prominent specialists were also welcomed, those who laid the foundations for the study of Roman amphorae and their contents in the second half of the 20th century, in particular our colleagues and friends Fanette Laubenheimer, Stefanie Martin-Kilcher and André Tchernia. These same specialists were the first to ask this destabilising question: what is the real importance of amphorae in the overall transport of foodstuffs in Roman times? Indeed, it should not be forgotten that ‘las ánforas son sólo una parte del comercio’ (see above Beltrán Lloris). Even if we exclude the case of the dolia, whose traces as transport containers can be found in shipwrecks (and which even generated the production of new amphorae in locations where wine arrived in bulk), and small ceramic containers, whose quantitative importance can be demonstrated every day by archaeological research, what is the economic weight of the amphorae, whose capacity rarely exceeds 30 litres, compared to the gigantic quantities of products transported in barrels?1 Moreover, barrels, and skins, are better adapted to land transport than amphorae. Nevertheless, as noted in the foreword, referring to the 1986 Siena conference, we cannot deny the importance of amphorae for the ancient economy and therefore for the economic history that archaeologists and historians are trying to recreate. Even if amphorae are not the only vector of maritime trade in Roman times, they are a concrete indicator of this trade, and we need to know more about their contents. In this field, today as yesterday, things still come in the form of questions:  See the example of Vindolanda: 2 900 litres of oil and wine in 50 amphorae in comparison, but 11 000 litres of wine in 10 barrels (Marlière and Torres Costa 2005). 1

The link between typology and content, can it be trusted? In the chapters by S. Martín Kilcher and A. Tchernia the reader will be able to reflect on the concepts that have concerned researchers since the beginnings of amphorology: monovalence (amphora x = content x), bivalence and multifunction. This is a key aspect for studies of the ancient economy. Our opinion, and that of the authors within this monograph, is that in Roman times each type of amphora was conceived at a specific time and in a specific region to store a particular type of product: otherwise, how do you explain the diversity of the typology? Thus, monovalence (each type or family of amphorae = certain content) must have been the normal situation from at least the Augustan times onwards. There are of course exceptions to this rule, such as the Haltern 70 which are bivalent, taking into account their use for the transport of olives and wine derivatives (mainly the thermally reduced grape must known as defrutum or sapa, much appreciated for their high potential as a preservative). There may have been moments of experimentation, as happens in the 1st century BC with the well-known ‘ovoids’, until the relationship between container and content was systematized in the different productive areas. Regarding polyvalence (versatility), this system seems typical of the period preceding the RomanRepublican era, in which there are very few types or even a single family, as is the case with the Phoenician-Punic world and other pre-roman communities using transport amphorae. Polyvalence of contents could also occur at the end of the Roman imperial system of distribution of goods (centred on the ports of Rome) as demonstrated by the African so-called spatheia of the first half of the 5th century (Bonifay, supra). Tituli picti: are they useful for the determination of content? The painted inscriptions alluding to the contents constitute, without doubt, potentially the most useful tools in relation to the history of research on these topics. However, there are very few examples in the Roman world, if we compare them with the number of known amphorae (a ratio of perhaps 1/1000 would be an approximate indicator, with exceptions such as at the well-known Monte Testaccio at Rome). The most abundant examples, in the well-known Dressel  20, do not provide information on the content, as is well understood. The two most extensive series of tituli picti are those of the early Imperial garum Baetican amphorae (studied by E. García Vargas and by P. Berni and E. Botte) and the Late Roman dipinti, especially in various families of Eastern amphorae (synthesized by J.-L. Fournet). Mentions of vinum and oleum are quickly recognized; as well as some fish products such as garum, liquamen, muria

Roman Amphora Contents. Reflecting on the Maritime Trade of Foodstuffs in Antiquity: 499–504

Conclusion or allec, although there is no exact consensus between the semantic nature of all these fishery terms, which could refer to diverse products with varied ingredients;2 although it is not disputed that these are all products made with marine resources (although terrestrial meat is used in some cases, such as garum mixtum). Also interesting are the references to products of less ‘direct’ interpretation, such as laccatum, recently considered by some authors as lacertus, alluding to young mackerel;3 or the cordyla or ‘young tunas’ of some Baetican amphorae, interpretations derived from very convincing and coherent philological studies, but which do not find an empirical demonstration (to date, no amphorae with these inscriptions associated with bones of these species have been found). When we have the epigraphic data and the archaeozoological evidence in the same amphorae, as happens with Dr. 21/22 of the Bottega del Garum in Pompeii, the interpretations are complicated, giving the impression that various products were made with similar fish (Bernal-Casasola et al. 2014). So the more information we have available, the more complex the interpretation becomes, as always happens in historical-archaeological investigations. This is one of the aspects that needs to be developed in the future.

as literary sources seem to show, or the Provençal coastline for the transport of pitch in reused amphorae)? Is it permanent or more characteristic of certain periods (e.g. Late Antiquity, according to the example of Yassi Ada)? T. Peña outlines several avenues of research to identify these reuses: superimpositions of tituli picti or unusual tituli picti, multiplicity of visible or chemical traces of contents, evidence of deterioration and repair of the amphorae or their stoppers. Ahead of further research on this issue, we should be cautious about the contents of the various types as one moves away from the coastline. It is clear that an amphora found in a rural site far from the coast is less likely to show evidence of its original contents than an amphora found in a port city, and the latter is itself less informative than an amphora found in a wreck. This questioning of the validity of amphorae to inform us about their primary content according to where they were found should lead us to be even more cautious in our economic and historical interpretations. It should also encourage us to better select the amphora samples submitted for chemical analysis, giving priority to wrecks and port contexts. Nevertheless, shipments such as the famous Grado wreck contents should not, for the time being, lead us to immediately ‘throw the baby out with the bathwater’.

Another element that should be introduced in this debate is the assessment of whether the tituli picti were occasional or whether, on the contrary, all or most of the amphorae were labelled in Antiquity. This is a much debated question, such as attributing its widespread presence on the Dr. 20 at Rome’s Monte Testaccio to the uniqueness of these productions, fiscally controlled by the Roman annona. As Emilio Rodríguez Almeida proposed at the time — followed by other colleagues such as P. Berni — the most logical explanation is that all the amphorae had tituli picti at the time of their labelling as archaeological contexts in good conservation conditions seem to demonstrate, such as the aforementioned case of Pompeii or on some well-known Mediterranean wrecks.

Chemical analyses: a question of interpretation? Can we trust them?

As we will see later, interdisciplinarity is needed to better face the problem of amphorae contents, and tituli picti should be related to the analysis of the content of the vessels (solid or liquid). The reuse of transport amphorae: exception or routine? T. Peña (see above) raises an important issue, that could also apply to barrels, to call into question the relevance of amphorae as evidence of the nature and direction of trade. The evidence for the reuse of amphorae for storage and transport is undeniable; there are numerous cases of this from written sources and archaeology, including on wrecks. The question is how widespread this phenomenon is. Is it universal or confined to certain regions (e.g. Egypt,  Recently analyzed, for instance, in Grainger 2021: especially 5–6, 115–130. 3  A recent interpretation of this product in Djaoui 2016. 2

Organic residue analysis has been applied for several decades to investigate the content of amphorae. Condamin and Formenti were real pioneers on the subject and today we can confirm some of the data they obtained with more reliable techniques. Although much time has passed, we still need to discuss the achievements and the problems of organic residue analysis. Organic residue analysis relies on the identification of what are called ‘biomarkers’, that are the chemical fingerprint of the ancient substances that were in contact with the ceramic matrix (in the case of amphorae). However, it is not easy to identify these biomarkers and the papers presented here demonstrate that a lot has still to be done. An important issue is that fresh substances differ from ancient ones because of degradation due to different causes that are mainly related to the production (i.e. burning of wood to produce pitch), transformation of the fresh substances as well as degradation and post-depositional processes. All these factors make the identification of the original contents of amphorae based on organic chemical analysis a difficult and still debated subject. However, we are far better positioned than a few decades ago, as several projects have shown. While until a few years ago we still had to rely on indirect information provided by the presence/absence of a visible pitch coating to address amphorae contents (mainly wine vs oil), we can now directly identify the residues of the substances preserved. One of the greatest achievements over the last few years is the possibility

500

Conclusion of identifying grape/wine and possibly other fruit biomarkers. This has led to an important change in the interpretation of the contents of amphorae where a wine/ related content was only deduced from the presence of an organic coating identified either by chemical analyses (Heron and Pollard 1988) or preserved as a visible lining. Although some amphorae analysis still investigates only the organic coating of the amphorae, analyses are now generally much broader. The identification of fish biomarkers allows for a better understanding of the content of ancient amphorae that are seen to be much more complex than previously thought. Another step forwards in organic residue analysis is that we are finally able to analyse series of amphorae. Although some of the papers presented here still show the analysis of only a few pieces, five years after the RACIIC Conference we are more and more aware that if a number of analyses are carried out on one amphora type, that can lead to the identification of exceptions in the content of a specific type, and thus more systematic analyses are essential. If until recently the only exception to this situation was Garnier’s PhD, luckily the analysis of series of amphorae is becoming a more and more common habit, thanks to the possibilities offered by different projects. A first example that deserves to be cited, although it was never officially financed, is the intended CORONAM Project, which put on the table an international, interdisciplinary and systematic analysis of different amphorae types that relied on the contribution of different scholars — many present at the RACIIC Conference — that led to a first publication on Late Roman African amphorae (Woodworth et al. 2015). Later, national and European projects were financed which allowed the broadening of the spectrum of amphorae analysed. Among them was the RACAMed Project financed by the Spanish Ministry and directed by P. Reynolds and A. Pecci — many of the scholars who participated in this conference were involved; for Early Medieval amphorae, analyses were carried out in the framework of the NeuMed project (dir. R. Hodges); the Sicily in Transition project (dir. M. Carver and A. Molinari); and the POMEDOR project (dir. Y. Waksman) focused on Medieval pottery. As some of the papers presented here pointed out, debate on the interpretation of the results of the analyses is still very active. However, we are now confident that when accurate checks of post depositional, sampling and laboratory contamination are carried out we can obtain important data on the content and the coating of the vessels and amphorae in particular. In this regard, the more substances we are able to identify, the more we have to pay attention to the interpretation, and we have to take care not to over-interpret the results. To avoid this, it is important to carefully select the laboratories performing the analyses, as they must

demonstrate reliability of analyses and experience in the identification of residues and contamination, and in the interpretation of results. A very interesting aspect that organic residue analysis has highlighted is the identification of several substances in the same amphora samples. This is certainly due to the presence of coatings and content. However in many cases it can be the results of a sum of activities (coating, use, reuse, re-cycling and discard) and we have to remember that chemical analysis cannot provide a ‘stratigraphy’ of the absorption. Therefore, we have to leave open the different possibilities of interpretation of a mixture of coatings (sometimes also involving a mixture of substances), contents (a single content or a mixture) and potential re-uses (again involving single substances or mixtures). Although this leads to problems in the interpretation of the original content (first content) of the amphorae which is frustrating for traditional scholars, in fact it broadens the scenarios for further investigation and shows that the reality is much more complex than the neat classifications that archaeologists would like to achieve. Again, to solve this problem we need to integrate the research of the different specialists — taking into account the context where the amphora was last used and found by the archaeologists, the discussion of the results of the analyses with the different specialists in amphorae typology, the archaeology of the provenance areas, provenance of the amphorae, epigraphists, chemists and so on — to achieve the most accurate interpretation. Good samples to demonstrate the methods and interpretations of organic residue analysis are those coming from shipwrecks (see above Cibecchini; Koutsouflakis). They provide situations in which, in most cases, both organic coatings and solid contents are preserved. When no solid content is preserved we usually assume that a liquid one (usually wine or oil) could have been the original content of the amphorae. The systematic study of organic coatings combined with organic residue analysis will allow to confirm the old hypothesis that when a thick pitch lining is present, usually wine or fish products are preserved in the amphora. Although some exceptions exist, this still seems to be a common pattern, but it must still be verified with a greater number of analyses. On the contrary we know that the absence of visible pitch cannot exclude the presence of wine/garum in an amphora. Again shipwrecks allow us to confirm whether the analyses can identify the residues of preserved solid contents (fruit, fish, meat…) and to study the relationship between type, inscription and content, without major post-depositional contamination and degradation issues. However we must not forget that even in shipwrecks there are amphorae that were re-used and therefore we must always pay attention to the general context (type of amphora, tituli, provenance of the amphora, provenance of the ship, organic residues…).

501

Conclusion Finally, as mentioned in this publication, samples cannot be given to the ‘chemists’ as if they were a black box. Many mistakes and over interpretations are produced in this way. The collaboration among the different specialists should start from the beginning of the project, as the more fruitful results can be obtained only through a joint strategy, the discussion of the advantages and disadvantages of the different choices and through good sampling, which must be performed according to the questions raised by the project and following good practices (sampling must be ‘clean’, not bringing any contamination in the sampling phase in the field, or in the laboratory, and must provide the possibility of checking for post depositional contamination). As the title of the conference suggests, specialists need to be more interactive. State of the question: very contrasting depending on the region One of the initial objectives of the RACIIC congress was to try to produce a summary table with the general consensus on the contents attributed to each amphora form. In the literary ‘jungle’ in which we move every day, this type of synthetic document is more and more essential to clarify the state of the art for the reader. In the end, this was not done, for various reasons. The first is the need for consensus by at least a significant part of the scientific community and today, that is still not possible. By comparing the articles in this monograph the reader will notice that there are various points of view on the same question. To give a couple of paradigmatic examples, in the case of one of the most important families related to the commercialization of Hispanic garum, Dr. 7/11, tituli picti and contents related to grapes and watered down wine are also known; or in the case of the internationally traded Dressel 1, the wine packaging par excellence, for which the imitations from the south of Hispania and some points of the Tyrrhenian coast (Albinia) and perhaps of Sicily suggest their primary use — not their reuse — for the packaging of salsamenta (cf. the contributions of Beltrán, Bernal-Casasola et al. and Menchelli). Sometimes the different authors do not even agree on the definition and geographical origin of certain types (e.g. the Keay 41, defined as a Hispanic piriform amphora or as an African cylindrical amphora, depending on the authors: see above Bernal-Casasola et al. vs. Bonifay). The production of this table of amphora contents (if it will ever be possible to produce such a schematization of the data obtained) is one of the pending subjects for the future and at the same time one of the conclusions of this congress: the need for an international Round Table to generate an instrument of consensual work on the contents of the main amphora series with an Atlantic-Mediterranean diffusion. First, it will be necessary to define these series, and secondly, the diverse direct (inscriptions, solid contents, organic residue analysis) and indirect criteria to identify their content (typology, productive geography of the workshops).

Another element highlighted in the various contributions presented at this congress is our diverse degree of knowledge of the palaeocontents of Roman amphorae in the different geographical areas or provinces of the Empire. The best known productive areas are those of the Western Mediterranean: Hispania (divided into the three main areas, Baetica, Lusitania, Tarraconensis, with differentiated problems), Galliae, Italy (and Sicily) and proconsular Africa with its corresponding late Roman provincial subdivision. Everybody is aware that this situation is the result of three aspects: the productive and exporting importance of these regions; the extensive historiography of the studies, inaugurated by Heinrich Dressel; and the efforts of the latest generation of archaeologists who have dedicated a good part of their respective academic careers to these topics (Fanette Laubenheimer in France is a good example of this). However, most of the Eastern Mediterranean regions are still poorly characterized, as is the case in general with the Aegean, the Anatolian coasts or the SyrianPalestinian corridor, with honourable exceptions such as Crete and isolated cases such as Beirut (treated in this monograph by Reynolds and Woodworth) or the late Roman productions manufactured in the micro-region of Ephesus (Bezeczky supra) or in the Black Sea (Opaiţ supra). But much more problematic are other regions that we know were important in the markets and whose products traded in amphorae are completely unknown to us, as is especially the case with the well-known amphorae of Mauretania Caesarensis, many of them stamped; or those coming from Tingitana, which we still do not even know how to recognize typologically with any certainty, although findings from recent years are improving the situation (Bernal-Casasola et al. 2019). Behind these imbalances are other important questions, such as the tacit consideration that Aegean and Eastern amphorae in general are wine containers, for which we have good indicators (tituli in some cases and resin coatings in others); but several articles in this book show that oil and salted fish should also be considered (see Reynolds, Opaiţ, Koutsouflakis). Thus, efforts for the next generation of archaeologists will have to focus on two issues: qualitative questions for western Roman productions (types of fish sauces, filling in the geographic gaps); and investigations to try to scientifically verify the palaeocontents of Eastern productions, by productive micro-regions, following the example of the strategy applied to Cretan amphorae analyzed in a diachronic key by various authors (Marangou-Lerat 1995; PoulouPapadimitriou and Nodarou 2014); or the case study of Egypt, synthesized in the pages of this book by D. Dixneuf. What, then, does the RACIIC Congress contribute to the international debate on the palaeo-content of amphorae? It is the reader who must answer this question after digesting the various contributions presented in

502

Conclusion this monograph, which is important because of the contribution of international amphorology specialists. We highlight three aspects. The first is updated regional syntheses on the main producing areas by the most internationally renowned researchers, alone or in collaboration: Lusitania (I. V. Pinto, R. Morais, C. Fabião, C. Oliveira, S. Gabriel), Hispania Citerior / Tarraconensis (R. Járrega, A. Ribera), Baetica (D. Bernal-Casasola, E. García Vargas, A. M. Sáez Romero, H. González Cesteros), Galliae (F. Laubenheimer), Tyrrhenian Italic and Sicily (S. Menchelli), Upper Adriatic (M.-B. Carre, S. Pesavento Mattioli), North Africa (M. Bonifay with J. Nacef), Egypt (D. Dixneuf), Eastern Mediterranean (P. Reynolds, T. Bezeczky) and the Black Sea (A. Opaiţ). Secondly, a battery of case studies allow us to approach the appropriate methodology. These can be used as a manual of good practices for the characterization of the amphora contents. They combine adequate sampling with archaeobotanical, archaeozoological, epigraphic, palynological and archeometric studies of a diverse nature to try to answer the question about the content of the analyzed amphora series. These case studies include one of the first Mediterranean preserved examples, such as the western Punic amphorae recovered in Corinth (A. M. Sáez Romero, T. Theodoropoulou); to the unique case of Pompeii (D. Bernal-Casasola, D. Cottica, R. Marlasca, C. G. Rodríguez Santana, E. García Vargas); passing through the rich and exceptionally well-preserved material from the underwater context of various Mediterranean areas such as the Aegean (G. Koutsouflakis), Sardinia (I. Sanna, L. Soro, C. Nervi); and on to the unique finds of the wrecks Arles-Rhône 3 in the south of France and the Bou Ferrer on the Spanish coast of Alicante (G. Piquès, N. Rovira, M. Tillier, F. Cibecchini, D. Djaoui, C. De Juan). Heuristic problems of some type or family are also dealt with, such as the case of the late Punic amphorae of the Ramon T-7.4.3.3 type (M. Luaces); or the containers with halieutic products recovered in the cetariae of the Murcia coast in southern Tarraconensis (A. Quevedo, M. Sternberg, J. D. Hernández-García). We hope that these works serve as inspiration to the reader when facing the analysis of the amphora contexts under study. Thirdly, these pages demonstrate that any amphora with palaeocontent remains is a ‘treasure’, an exceptional find. Perhaps for any researcher focused on Classical Archaeology or Late Roman Archaeology this statement will sound strange, given the thousands of amphorae that we find daily in excavations and stored in our museums. Of these, less than a hundred have reached us with any physical content; with macro-remains and tituli picti only dozens; and of the latter those subjected to archaeometric analysis of organic contents less than 10. Very few Roman amphorae fulfill the tetralogy: preserved inscriptions related to the content, physical remains, archaeometric analysis and also, fundamentally, well studied and published results. Incredible, yes,

but equally true. Hence, one of the objectives of this International Congress of Cadiz and of the publication of these Proceedings is to raise awareness of the importance of these findings, and of their preservation and study. We hope that in the coming decades we will be able to offer a much more complex and articulated panorama of the Economic History and maritime traffic of Rome, aspects for which amphorae play a strategic role. Finally, a fundamental aspect that RACIIC has highlighted it is the need for interdisciplinarity. Indeed, one of the important lessons of this congress is the confirmation that in order to advance the research it must be interdisciplinary. In the case of semi-solid or solid products (salted fish, allec, meat, olivae, fruits…) the research will have to be carried out by specialists of amphorae typologies, epigraphy, and the corresponding specialists (ichthyologists or more broadly archaeozoologists, carpologists…). In all cases the ideal situation would be to work with chemists, to carry out organic residue analysis to investigate the liquid products (oil, wine, garum, organic contents that are not visible, preserving agents such as vinegar…) that could be the content itself or were used to better preserve the solid content. A theme that was not mentioned in the conference, but of which we are all aware is that all the studies of the content should be coupled with the analysis of the provenance/origin of the amphorae. This also implies the involvement of specialists (geologists, archaeometrists). We consciously avoided this at RACIIC because of the very wide research on the subject that would have gone beyond the time available at the conference, but we are fully aware that only the combination of provenance and content will allow more accurate investigations of the themes of use and re-use of amphorae and their original content. Cádiz – Aix – Barcelona - Toulon, May 2021 Darío Bernal-Casasola, Michel Bonifay, Alessandra Pecci and Victoria Leitch Bibliography Bernal-Casasola, D., Cottica, D., García Vargas, E., Toniolo, L., Rodríguez-Santana, C. G., Acqua, C., Marlasca, R., Sáez, A. M., Vargas, J. M., Scremin, F. and Landi, F. 2014. Un contexto exceptional en Pompeya: la pila de ánforas de la Bottega del Garum (I.12.8). Avance de un estudio interdisciplinar. Rei Cretariae Romanae Fautorum Acta 43: 219–232. Bernal-Casasola, D., Díaz, J. J., Bustamante Álvarez, M., Pascual, M. A., Fantuzzi, L., Retamosa, J. A. and Ghottes, M. 2019. Tamuda y las ánforas mauritanas occidentales. Primeros apuntes tipológicos y arqueométricos. HEROM. Journal on Hellenistic and Roman Material Culture 8: 155–210.

503

Conclusion Djaoui, D. 2016. The myth of ‘Laccatum’: a study starting from a new titulus on a Lusitanian Dressel 14. In I. V. Pinto, R. R. de Almeida and A. Martin (eds.), Lusitanian Amphorae: Production and Distribution. Roman and Late Antique Mediterranean Pottery 10: 117–127. Oxford, Archaeopress. Grainger S. 2021. The Story of Garum. Fermented Fish Sauce and Salted Fish in the Ancient World. London-New York, Routledge, Heron, C. and Pollard, A. M. 1988. The analysis of natural resinous materials from Roman amphorae. In E. A. Slater and J. O. Tate (eds.), Science and Archaeology, Glasgow 1987. British Archaeological Reports British Series 196: 429–448. Oxford, BAR publishing. Marangou-Lerat, A. 1995. Le vin et les amphores de Crète de l’époque classique à l’époque impériale. Études Crétoises 30. Athens, École française d’Athènes. Marlière, E. and Torres Costa, J. 2005. Tonneaux et amphores à Vindolanda: contribution à la connaissance de l’approvisionnement des troupes stationnées sur le mur d’Hadrien (II). In A. Birley and J. Blake (eds.), Vindolanda, The Excavations of 2003/2004: 214–236. Hexham, Vindolanda Trust.

Poulou-Papadimitriou, N. and Nodarou, E. 2014. Transport vessels and maritime trade routes in the Aegean from the 5th to the 9th c. AD. Preliminary results of the EU funded “Pythagoras II” Project: the Cretan case study. In N. Poulou-Papadimitriou, E. Nodarou and V. Kilikoglou (eds.), LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean. Archaeology and Archaeometry. The Mediterranean: a Market without Frontiers. British Archaeology Reports International Series 2616 (II): 873–883. Oxford, Archaeopress. Woodworth, M., Bernal‑Casasola, D., Bonifay, M., Garnier, N., Keay, S., Pecci, A., Poblome, J., Pollard, M., Richez, F. and Wilson, A. 2015. The content of African Keay 25/ Africana 3 Amphorae: Initial results of the CORONAM Project. In C. Oliveira, R. Morais and A. Morillo Cerdán (eds.), ArchaeoAnalytics. Chromatography and DNA analysis in archaeology: 41–57. Esposende, Município de Esposende.

504