Amphora Research in Castrum Viall on Brijuni Island 9783700179726, 3700179723

Citation preview

AMPHORA RESEARCH IN CASTRUM VILLA ON BRIJUNI ISLAND EDITED BY TAMÁS BEZECZKY

BEZECZKY_Text.indd 1

18.03.19 14:39

ÖSTERREICHISCHE AKADEMIE DER WISSENSCHAFTEN PHILOSOPHISCH-HISTORISCHE K LASSE DENKSCHRIFTEN, 509. BAND

ARCHÄOLOGISCHE FORSCHUNGEN Band 29

BEZECZKY_Text.indd 2

18.03.19 14:39

AMPHORA RESEARCH IN CASTRUM VILLA ON BRIJUNI ISLAND CONTRIBUTORS TAMÁS BEZECZKY, PIERO BERNI MILLET, MICHEL BONIFAY, CLAUDIO CAPELLI, HORACIO GONZÁLEZ CESTEROS, SÁNDOR JÓZSA, MARTINO LA TORRE, ALEXANDER SCHOBERT, GYÖRGY SZAKMÁNY

EDITED BY TAMÁS BEZECZKY

BEZECZKY_Text.indd 3

18.03.19 14:39

Angenommen durch die Publikationskommission der philosophisch-historischen Klasse der Österreichischen Akademie der Wissenschaften: Accepted by the publication committee of the Division of Humanities and Social Sciences of the Austrian Academy of Sciences by: Michael Alram, Bert G. Fragner, Andre Gingrich, Hermann Hunger, Sigrid Jalkotzy-Deger, Renate Pillinger, Franz Rainer, Oliver Jens Schmitt, Danuta Shanzer, Peter Wiesinger, Waldemar Zacharasiewicz

Redaktion: Tamás Bezeczky / Katharina Preindl

Umschlagabbildung: Aerial photograph of Castrum villa (Courtesy of the Brijuni Museum)

Diese Publikation wurde einem anonymen, internationalen Begutachtungsverfahren unterzogen. This publication was subject to international and anonymous peer review. Peer review is an essential part of the Austrian Academy of Sciences Press evaluation process. Before any book can be accepted for publication, it is assessed by international specialists and ultimately must be approved by the Austrian Academy of Sciences Publication Committee.

Die verwendete Papiersorte in dieser Publikation ist DIN EN ISO 9706 zertifiziert und erfüllt die Voraussetzung für eine dauerhafte Archivierung von schriftlichem Kulturgut. The paper used in this publication is DIN EN ISO 9706 certified and meets the requirements for permanent archiving of written cultural property.

Alle Rechte vorbehalten. ISBN 978-3-7001-7972-6 Copyright © 2019 by Österreichische Akademie der Wissenschaften, Wien Satz und Layout: Andrea Sulzgruber, Wien Druck und Bindung: Wograndl Druck GmbH, Mattersburg https://epub.oeaw.ac.at/7972-6 https://verlag.oeaw.ac.at Made in Europe

BEZECZKY_Text.indd 4

18.03.19 14:39

V

CONTENTS Foreword (Pülz) .................................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . In memoriam Tamás Bezeczky (1949 – 2018) (Krinzinger) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . Introduction and acknowledgements (Bezeczky) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . List of figures . ....................................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . List of plates . ........................................ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . .

 VII  IX  XI XIII  XV

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS CHAPTER 1 Istria and the amphora workshops (Bezeczky) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1 Historical background .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 2 The Fažana workshop .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 The owners of the workshop .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 The amphora stamps .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 The production capacity of the villas and the distribution of the Fažana amphorae .. . . .. . . . . . . . . . . . . CHAPTER 2 The villas of Brijuni (La Torre – Bezeczky) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1 Verige Bay (Val Catena) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Kolci (Monte Collisi) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 3 Castrum . ................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHAPTER 3 The early Christian churches of Brijuni (Schobert) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 The Church of St. Mary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 2 The early Christian church within the walls of the Castrum and in Verige Bay . . . . . . . . . . . . . . . . . . . . . . . 3 The Church of St. Peter .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 4 The Bishop of Cissa Pullaria (Cessus) and his residence .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . .

3 3 5 6 6 7 11 11 15 17 29 29 34 34 36

THE AMPHORAE CHAPTER 4 Italian and Istrian amphorae (Bezeczky) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Italian wine amphorae from the late Republican and early Roman period .. . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Olive oil amphorae . ..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 3 Istrian fish sauces amphorae .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Northern Adriatic amphorae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Miscellaneous Italian amphorae .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . Conclusion ................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . CHAPTER 5 Hispanic amphorae (González Cesteros – Berni Millet) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Iberian amphorae production and the Adriatic region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Identification problems in distinguishing Lusitanian and Baetican amphora types .. . . .. . . . . . . . . . . . . 3 The Hispanic amphorae from the Brijuni Castrum villa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion ................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . CHAPTER 6 Eastern Mediterranean amphorae (González Cesteros) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . The eastern Mediterranean amphorae and the north Adriatic region . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . The eastern Mediterranean amphorae types from the Brijuni Castrum villa .. . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion ................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . .

41 41 42 44 45 45 46 47 47 51 52 53 55 55 63 69

BEZECZKY_Text.indd 5

20.03.19 19:08

VI

Contents

CHAPTER 7 African amphorae (Bonifay – Capelli) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1 Typology and origin of the African amphorae .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1.1 African amphorae of Punic tradition .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1.2 Roman-African amphorae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1.2.1 Classical Roman-African amphorae .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 Late Roman-African amphorae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1.3 African imitations of non-African types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 1.4 Storage jars (?) . .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 African imports to Brijuni in the northern Adriatic context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . Conclusion ................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHAPTER 8 Amphora stamps and inscriptions (Berni Millet – Bezeczky) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . CHAPTER 9 Catalogue (Bezeczky – Berni Millet – Bonifay – Capelli – González Cesteros – Józsa – Szakmány) . ........ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PLATES 1-22 .............. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71 71 71 71 71 72 73 73 74 76 77 79 101

EPIGRAPHY CHAPTER 10 Calendar graffiti on Dressel 20 amphorae (Berni Millet) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 The manufacturing stages of a Dressel 20 amphora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 2 Nominal graffiti . . ........ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 3 Calendar graffiti ........ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 The calendar graffito from the Castrum of Brijuni . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Asiaticus: another paradigmatic case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . 6 General reflections ..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Updated chart of calendar graffiti in order of the date of the year .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHAPTER 11 Inscriptions on laterite finds from Brijuni (Berni Millet) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. Epigraphy on laterite (tegulae, imbrices) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . B. Epigraphy on dolia . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. General index ........... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PLATES 23-34 . ........... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

125 126 126 128 132 135 139 144 147 150 168 169 173

MICROPETROGRAPHY CHAPTER 12 Micropetrography of the Fažana amphorae (Szakmány – Józsa) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Previous works on raw material determination .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . Methods and strategy . .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Results ....................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Brief thin section micropetrography of amphorae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Detailed thin section micropetrography of possible raw materials .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . Detailed comparative thin section micropetrography of Fažana amphorae .. . . . . . . . . . . . . . . . . . . . . . . . . . Discussion and conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . Summary . .................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . PLATES 35-37 . ........... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

187 187 187 189 189 189 191 194 196 197

CHAPTER 13 Summary (Bezeczky) . ..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 Sažetak (Bezeczky) .......... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 APPENDIX The Fažana amphora stamps (Bezeczky) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . INDICES .. . . ........................................ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BIBLIOGRAPHY . . .............................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . CONTRIBUTORS . ............................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

BEZECZKY_Text.indd 6

209 219 223 237

18.03.19 14:39

VII

VORWORT

Die neue Publikation zu den Amphoren aus der Castrum Villa von Brijuni gibt Anlass zu einigen forschungs­ge­ schichtlichen Reflexionen: Am Anfang der Forschungen zur Archäologie von Brijuni stand vor mehr als hundert Jahren Anton Gnirs, Kurator der altösterreichischen Zentralkommission für Istrien, Rektor der österreichischen Real-Schule und seit 1909 Direktor der Staatlichen Antikensammlungen in Pola. Im Auftrag von Paul Kupelwieser, dem damaligen Eigentümer der Insel, führte er seit 1904 eine Reihe von Ausgrabungen durch und legte auch das castrum frei. Er erkannte im Übrigen als Erster die wesentlichen Funktionen der Ruine, die später von einer befestigten Siedlung überbaut werden sollte. Es folgten die Grabung des Museums von Pula durch die kroatischen Archäologen Štefan Mlakar und Anton Vitasović mit einigen Präzisierungen zu den Bauphasen und den landwirtschaftlichen Produktionsabläufen in der villa rustica. Außer der im Museum von Pula aufbewahrten Feinkeramik wird das gesamte Fundmaterial der Grabungen heute im Depot des Museums des Nationalparks Brijuni gelagert, wobei die Amphoren Kollegen Bezeczky zur Publikation übergeben wurden. Aus den Fragmenten zahlloser Amphoren, den bewährten Transport- und Speichergefäßen – sie trugen in der Regel an den Henkeln eingestempelte Hinweise auf den Produzenten und Verwalter – ergeben sich für den Fachmann faszinierende Einblicke in die ökonomischen Strukturen der römischen Landwirtschaft und in die Handelswege und kulturellen Beziehungen im gesamten Mittelmeerraum. Ebendiesen Fragestellungen hat sich Tamás Bezeczky ausgehend von seinen 1998 publizierten Forschungen zu den Laecanius-Amphoren in den letzten Jahrzehnten mit ganzer Kraft gewidmet, neue Methoden weiterentwickelt und dabei wesentliche Resultate erarbeiten können. Das vorliegende Werk bietet zunächst einen kompakten Überblick zu allen römischen Villen der Insel, im Speziellen aber zur Castrum Villa mit ihren zahlreichen Bauphasen.

BEZECZKY_Text.indd 7

Berücksichtigung findet zudem die kleine Kirche, die im späten 5. Jh. n. Chr. etwas nördlich des castrum errichtet wurde. Bezeczky führte also – seinem wissenschaftlichen Selbstverständnis entsprechend – die Forschungen nicht alleine durch, sondern bettete diese in ein internationales Expertenteam ein, womit er dem Fundmaterial, das chronologisch vom 1. Jh. v. Chr. bis in das 7. Jh. n. Chr. reicht, und der geographischen Verteilung der Produktions­stätten praktisch im gesamten Kulturkreis des Mittelmeeres gerecht werden konnte. Ausgangspunkt der vorliegenden Studien waren die Produktionsstätten in den Latifundien der Familie Laecanius, die später in den Besitz der Kaiser Vespasian, Domitian und Trajan übergingen, ehe die Produktion unter Hadrian eingestellt und durch iberische Güter ersetzt wurde. Daher sind neben den adriatischen Produktionsstätten vor allem die iberischen, aber auch die afrikanischen Amphoren von signifikanter Bedeutung. Zum ersten Mal werden im bislang oft undifferenziert dargestellten Raum des „Eastern Mediterranean“ auch Gruppen von spezifischen Importwaren bestimmt und einzelnen Orten zugewiesen. Eine große Hilfe stellte bei diesem Unterfangen die Amphoren-Datenbank des Autors dar. Für die petrographische Bestimmung der Materialzusammensetzung der Laecanius-Amphoren haben auch geologische Untersuchungen zum besseren Verständnis der Herkunft des Tons wesentlich beigetragen. Die nun posthum vorgelegte letzte Publikation von Tamás Bezeczky ist ein ganz hervorragendes Beispiel für seine Arbeitsweise. Das konzise Konzept mit seiner detaillierten Zielsetzung wurde in allen wesentlichen Punkten von ihm bestimmt und gemeinsam mit ausgewählten Fachleuten verschiedener Disziplinen umgesetzt. Bezeczky gab die Gesamtgliederung vor und verfasste die archäologischhistorischen Überblicke zur Insel und alle Beiträge zu Laecanius. Auch alle Zusammenfassungen im Werk tragen seine Handschrift. Dass der Autor, durch seine schwere Krankheit geschwächt, die Fertigstellung seines bereits zum Druck eingereichten Werkes, an dem er bis zuletzt gearbeitet hat, nicht mehr erleben konnte, ist von besonderer Tragik.

18.03.19 14:39

VIII

Vorwort

Es war für das Institut für Kulturgeschichte der Antike der ÖAW eine selbstverständliche Pflicht, für den geregelten Abschluss des Publikationsprozesses zu sorgen. Abschließend sei den Autoren gedankt, durch deren Beiträge das vorliegende Werk zu einer umfassenden wirtschafts- und kulturgeschichtlichen Studie wurde. Dankend hervorgehoben seien Horacio González Cesteros für seine Korrektur­lesungen, besonders des umfangreichen Kataloges, sowie Frau Andrea Sulzgruber für den Satz und die graphische Gestaltung der Publikation. Zu danken ist zudem Frau Katharina Preindl für die redaktio-

BEZECZKY_Text.indd 8

nelle Weiterbetreuung des Manuskriptes sowie dem Verlag der ÖAW und seiner ehemaligen Geschäftsführerin Frau Mag.  L. Triska, die die notwendigen Rahmenbedingungen für die Drucklegung sicher­stellte. Besonderer Dank gebührt ferner dem Fonds zur Förderung der wissenschaftlichen Forschung (FWF), welcher das Projekt – wie alle vorangegangenen Forschungen des Autors – getragen hat, sowie dem Holzhausen-Legat der ÖAW für die maßgebliche Unterstützung bei der Finanzierung des Drucks. Wien, 2018

Andreas Pülz

18.03.19 14:39

IX

IN MEMORIAM TAMÁS BEZECZKY (1949  –   2 018)

Da es dem verantwortlichen Herausgeber sowie Hauptautor Tamás Bezeczky nicht mehr vergönnt war, den Abschluss und die Publikation des vorliegenden Werkes zu erleben, soll der Einleitung des Autors eine kurze Würdigung seines Lebenswerkes vorangestellt werden. Nach der Ausbildung zum IT-Ingenieur an der TU Budapest, die er parallel zur Lehrtätigkeit an einer Techniker-Lehranstalt absolvierte, und nach einer mehrjährigen Verpflichtung am Institut für Konservierung und Methodik der Museologie wurde Bezeczky im Jahre 1982 Kurator der Computerabteilung am Ungarischen Nationalmuseum und inskribierte sich gleichzeitig am Archäologischen Institut der Universität Budapest. Er stellte ein Forschungsprojekt zur Bernsteinstraße zusammen und engagierte sich an der Ungarischen Akademie der Wissenschaften für das Thema „Computer Application in the Archaeology“. Seine Dissertation Roman Amphorae from the Amber Route in Western Pannonia erschien 1987 in Oxford. Im Vorwort der Dissertation legte Bezeczky seine interdisziplinären Methoden und archäologischen Ziele programmatisch dar. Die damit manifeste Doppelbegabung war bestimmend für seinen weiteren wissenschaftlichen Weg, den er mit konsequenter Leidenschaft gegangen ist: computergestützte Forschungen

BEZECZKY_Text.indd 9

zu römischen Amphoren mit interdisziplinären Methoden zur Bestimmung von Material und Produktionsort. Schon als Student war Bezeczky auf vielen Grabungen und Fundorten in Ungarn, Jugoslawien und Italien anzutreffen. Nach Österreich kam er erstmals im Jahr 1986. Er war mehrere Jahre als Mitarbeiter bei der Ausgrabung am Magdalensberg in Kärnten tätig und mit der Bearbeitung der Amphoren betraut. Ein von der ÖAW in Aussicht gestelltes Stipendium, welches dann vom Ministerium leider nicht bewilligt wurde, war neben familiären Überlegungen einer der Gründe, Anfang der 1990er Jahre von Budapest nach Wien zu übersiedeln. Die persönlichen Lebensumstände der Familie waren in diesen Jahren sicherlich nicht leicht. Dennoch arbeitete Bezeczky zielstrebig an der Veröffentlichung seiner Resultate, die grundlegende Publikation zu den Amphoren vom Magdalensberg erschien 1994 in Klagenfurt. Im Sommer des gleichen Jahres lernte ich Bezeczky im Grabungsdepot von Carnuntum persönlich kennen und schätzen. Kiste für Kiste, Regal für Regal durchforstete er nach „seinen Amphoren“, dokumentierte Formen und Stempel und untersuchte die Gefäße auf Graffiti sowie aufgepinselte Hinweise. Er war für sein erstes Laecanius-Projekt unterwegs, das von der Soros Foundation (Ungarn) und in der Folge von der ÖAW und dem ÖAI gemeinsam finanziert wurde. Somit bot sich dem Autor die Möglichkeit zur vertieften Erforschung der bekannten Produktionsstätten in Fazana und Brijuni (Istrien), die mit den Latifundien der senatorischen Familie des Laecanius zusammenhängen. Die vor Ort erzeugten Amphoren als „Einweggebinde“ sind Beleg für die im ganzen Mittelmeer verbreiteten landwirtschaftlichen Produkte und bieten durch die Stempelreihen nachvollziehbare Einblicke in die Organisation der Herstellung und des Handels. Die viel beachteten Ergebnisse des Projektes wurden 1998 an der ÖAW publiziert. Mit den darauffolgenden Aufenthalten als Visiting Fellow am Department of Archaeology der Universität Southampton verdichteten sich die internationalen Kontakte Bezeczkys in besonderer Weise.

18.03.19 14:39

X

In memoriam Tamás Bezeczky (1949 – 2018)

Tamás Bezeczky und Sándor Józsa im September 2012 auf Brijuni (© G. Szakmány)

Als Grabungsleiter von Ephesos war es für mich naheliegend, den international anerkannten Experten 1998 in das Forscherteam einzuladen. Die Metropolis Asiae mit dem großen Handelshafen und dem weltweit berühmten Artemision erwies sich als idealer Ort für die Forschungen zur Entwicklung und Verbreitung der römischen Amphoren des östlichen Mittelmeerraumes. In mehr als 10-jähriger Arbeit schuf Bezeczky in der Tat ein Standardwerk, dessen Finanzierung dankenswerterweise über den gesamten Zeitraum dieses Projektes – nach kompetitiv erfolgten Antragstellungen – vom FWF getragen wurde. Über die traditionellen Bearbeitungsmethoden der Funde hinaus konnten mit geologischen und petrologischen Methoden mehrere lokale Produktionsstätten definiert und deren Entwicklung über einen größeren Zeitraum dargestellt werden. Daneben ergab sich für die modernen wirtschaftshistorischen Fragestellungen zu antiken Handelsrouten und Warentransporten eine Fülle von Importkeramik, woraus eine Vielzahl wirtschaftlicher und kultureller Kontakte nachgewiesen werden konnte. Von ganz besonderem Interesse war dabei die kontextuelle Betrachtung des gesamten Fundmaterials aus den bereits abgeschlossenen Grabungen auf der Tetragonos Agora, sodass auch das chronologische Gerüst der ausgewiesenen Typen und Varianten abgesichert werden konnte. Die Ergebnisse wurden 2013 in

BEZECZKY_Text.indd 10

den Forschungen in Ephesos (FiE XV/1) vorgelegt. In diesen Jahren wurde für die übersichtliche Verwaltung aller Informationen zur großen Zahl der Objekte und für die systematische Dokumentation der zahlreichen Para­meter von Bezeczky gemeinsam mit Peter Hornung auch eine spezielle Datenbank entwickelt. Sie ist in der Einleitung des Werkes ausführlich beschrieben und konnte in der weiteren Folge verdichtet und erweitert werden. Sie ist heute ein wertvolles Werkzeug für alle Spezialisten der Amphorenforschung. Mit seinem letzten großen Projekt, das wiederum vom FWF finanziert wurde, kehrte Bezeczky nach Istrien zurück und setzte es sich zum Ziel, die Amphoren aus einer kleineren Villa des Laecanius (sog. Castrum Villa) an der Westküste der Insel Brijuni zu bearbeiten. Das Fundmaterial dieser Villa war den Forschungen Bezeczkys in den 1990er Jahren nicht zugänglich gewesen. Die schiere Menge der Fragmente und das breite chronologische Spektrum erforderten eine neue Arbeitsweise. Wie in der folgenden Einleitung zur Publikation dargestellt wird, organisierte Bezeczky ein internationales Team und brachte das Projekt auf diesem Wege zu einem erfolgreichen Abschluss. Neben diesen mehrjährigen Forschungen, die jeweils mit einer Monographie erfolgreich abgeschlossen wurden, umfasst die Bibliographie Bezeczkys über sechzig Einzel­ titel, Rezensionen und daneben eine Fülle von praxisorientierten Forschungsleistungen, von denen die Datenbanken zu Ephesos und Brijuni sicherlich die wichtigsten sind. Tamás Bezeczky ist zu früh von uns gegangen, aber er hat ein reiches Erbe hinterlassen. Es ist den laufenden und zukünftigen Forschungen zu den römischen Transport­ amphoren zu wünschen, dass an diesem Erbe weiter gebaut werden kann. An den Schluss darf ich mit freundlicher Erlaubnis einen Satz von Frau Anna Bezeczky stellen: Ja, mein Mann hatte eine verkürzte Lebensdauer, aber ein volles Leben! Es wird einem selten zuteil, dass Arbeit, Beruf, Berufung und Hobby gleichzeitig zutreffen. Er hatte dieses Phänomen inne, dadurch war er ein glücklicher Mensch! Wien, 2018

Friedrich Krinzinger

18.03.19 14:39

XI

INTRODUCTION AND ACK NOWLEDGEMENTS

The Austrian Science Fund (FWF) accepted the research project The “Laecanius amphorae in Brijuni” (P 23684) in 2011. The goal of the project was to publish the new Laecanius amphorae from the Castrum villa in Brijuni Island. The Laecanius amphora stamps and the villas of Brijuni (Bezeczky 1998a) described the amphorae which were available in the mid-1990s. But at that time it was not possible to study the amphorae in one of the Laecanius villas (Castrum). The many thousand finds (amphorae, tegulae, fine ware, glass, etc.) discovered during the villa excavations, with a few exceptions, were kept in the archaeological collection of the Brijuni National Park. Rarely do we have the opportunity to analyze a group of amphorae from the first century B.C. to the seventh century A.D.  found in such a relatively small area (1 hectare). This made it clear that all of the excavated amphora groups needed to be published, not only those produced by the Laecanius family. An international team was formed, composed of the specialists who study food production in different parts of the Roman Empire and they contributed individual chapters. In this volume, the reader will find a selection of the amphorae from the villa. The finds were grouped by place of production (Adriatic, Iberian, Eastern Mediterranean and African), and the articles are also published in this order. Various objects have already been published from the villa, primarily the architectural elements found during the excavations, tiles with stamp (e.g. works by Štefan Mlakar, Branko Marušić, Vlasta Begović, Ivančica Schrunk, Anton Vitasović, Robert Matijašić) and ceramic finds (see papers by Philipp M. Pröttel, Verena Vidrih Perko and Mira Pavletić). Our team examined all of the currently available amphora finds and selected the rim, base, neck and handle pieces of the most important amphora types to be presented. All of the pieces of the collection which bear a stamp or inscription have been included. Some 90 amphora types are represented, which contained olive oil, wine, fruit and fish products. One must differentiate between locally made and imported amphorae. The Castrum villa was a place where olive oil and wine were produced. It is necessary to

BEZECZKY_Text.indd 11

clarify what was produced when. In addition, there are two detailed analyses. One of them is concerned with a rare graffito on a Baetican  amphora, the other with the abundant tile stamps. The new archaeometrical analyses of the Fažana amphorae started, they provide new data on the origin and method of production of the amphorae. The discovery of the villas on Brijuni Island and the amphora workshops in Fažana can be attributed to a lucky accident. In 1893, during the Austro-Hungarian Monarchy, a wealthy industrial magnate, Paul Kupelwieser, bought the Brijuni Islands. He was determined to build a holiday resort, or in today’s parlance, a wellness and recreation center. With the help of Robert Koch, the famous microbiologist, he eradicated malaria on the islands. During construction works for the roads and the hotel building foundations, Kupelwieser’s workers discovered several Roman ruins. He contracted Anton Gnirs, a teacher in the “Marinerealschule” of Pula, to lead the archeological excavations. Gnirs discovered several Roman buildings on the islands, Fažana and in Pula. After the Second World War, the Yugoslavian president Josip Broz Tito took a liking to the islands. One of his luxurious homes was in the vicinity of the Castrum villa. Later Tito himself was the driving force in the excavation of this villa. After his death, the islands were declared a National Park and parts of it were opened for the public. The story of the Castrum villa is closely connected to the Laecanius family. Ancient and modern authors and the excavations provide a great deal of information about this family’s history, economic activity, villas and amphora workshop. It is clear that their estates and the income that these estates produced went to the emperors after the last member of the family died without an heir. Less is known about the amphora workshop and villa in Fažana after the second century. Although the excavations brought a great deal of objects to light, our knowledge is incomplete in many ways. For instance, we do not know who used the villas in the midand late Roman periods. Regarding the excavated walls and objects, it is unclear which period the published maps belong

18.03.19 14:39

XII

Introduction and acknowledgements (Bezeczky)

to. Various rooms in the villas were altered on several occasions in the Roman period as well as during the reconstruction following the excavations. We only provide a short description of the Fažana workshop and the villas in Brijuni in this book, as some knowledge of them is necessary to understand the Castrum finds. There may be some overlap between individual chapters, but this is unavoidable in order to present the information coherently. We also refer to the rich bibliography for further information. I would like to thank the directors of the Brijuni National Park, Sandro Dujmović, Eduard Kolić, and Mira Pavletić and the Croatian Ministry of Culture for granting permission to study and publish the amphorae. I am grateful for the support I have receive from the directors, past and present, of the Austrian Academy of Sciences, Institute for the Study of Ancient Culture, Friedrich Krin­zinger and Andreas Pülz. I would like to take this opportunity to thank Piero Berni Millet, Michel Bonifay, Claudio Capelli, Horacio González Cesteros, Sándor Józsa, Alexander Schobert, Martino La Torre and György Szakmány for their work. Thanks are also offered to Claudio Capelli, Sándor Józsa and György Szakmány; they prepared the petrological (thin section) analyses and the photomicrographs (1x1.3  mm) of the amphorae. The petrological description of the amphorae can be found in the catalogue. The technical assistance of Anna A. Nagy and Ozren Grozdanić is very much appreciated.

BEZECZKY_Text.indd 12

I am especially grateful to Péter Hornung for creating the FileMaker database. The amphorae, which are in an extremely fragmented state, will be uploaded and saved in the database. Special thanks are due to Ágnes Bezeczky who translated parts of the manuscript (Chapters 1, 2, 4 and 13). The photos were taken by Andreas Pülz, Martino La Torre and Tamás Bezeczky. For their permission to reproduce pictures which appear in the text we are grateful to Brijuni National Park. The line drawings were pepared by Tamás Bezeczky, Piero Berni Millet and Horacio González Cesteros. Thanks are also due to Helka Németh who made the digital amphora drawings for publication. The drawings and photos are actual size on the plates. The scale of the rubbings is 1:1. Numbers in bold refer to the catalogue in the chapters on the amphorae. While I was working on the amphorae, the archaeologists of many museums generously helped me: David Peacock (†), Paul Arthur, Martin Auer, Marianna Bressan, Silvia Cipriano, Ines Dörfler, Smiljan Gluščević, Jana Horvat, Simon Keay, Ida Koncani Uhać, Vladimir Kovačić, Goranka Lipovac Vrkljan, Stefania Mazzocchin, Péter Pánczél, Elena Quiri, Federica Rinalani, Florian Schimmer, Eleni SchindlerKaudelka, Katalin Vanicsek, Paola Ventura, Péter Véninger, Reinhold Wedenig and Susanne Zabehlicky-Scheffenegger. Vienna, 2017

Tamás Bezeczky (†)

18.03.19 14:39

XIII

LIST OF FIGURE S

Chapter 1 ISTRIA AND THE AMPHORA WORKSHOPS 1.1 1.2 1.3 1.4 1.5

Map of Istria (Tabula Imperii Romani). Map of Fažana workshop and kiln. Dolium from Brijuni harbour, the position of the dolia in the Verige and Castrum villas. Complete Laecanius amphora from Magdalensberg. The distribution of the Fažana amphorae.

Chapter 2 THE VILLAS OF BRIJUNI 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18

Map of Brijuni Islands and Fažana. Map of the villa in Verige and the terrace building. Map of the terrace building. Map of Kolci villa. Castrum villa after Mlakar 1975-76 and Begović – Schrunk 2007a.  Castrum villa, after Mlakar 1975-76. Archive pictures from the excavations. The first Roman villa. Map of the first Roman villa. Map of the second Roman villa. The restored presses in the second villa. In the lacus, spicae and opus signinum wall. The oil cellar and fragmented dolia in the second villa. Mill stone and a mola olearia. Map of the third Roman villa. The smaller cellar and the dolia. Map of the Late Roman and Byzantine villa. Wine presses installation.

Chapter 3 THE EARLY CHRISTIAN CHURCHES OF BRIJUNI 3.1 3.2

The Church of St. Mary. Column of the main nave with impost and Greek cross.

BEZECZKY_Text.indd 13

3.3 3.4

Monolithic platform of the early Christian altar. Triumphal arch supported by two monolithic columns with capitals and imposts. 3.5 Capital in the sanctuary with crux coronata emblems. 3.6 Presbytery with septum. 3.7 Hospitium (domus presbyterorum), northern annex. 3.8 Room with double apses (cellae memoriae), southern annex. 3.9 Sarcophagus, narthex. 3.10 The Church of St. Peter, main entrance. 3.11 Peutinger Table pars V.  Chapter 10 CALENDAR GRAFFITI ON DRESSEL 20 AMPHORAE 10.1 A Dressel 20 globular amphora with a stamp, tituli picti and graffito. 10.2 Stages in the manufacture of the Dressel 20 amphora. 10.3 Nominal cursive graffiti. 10.4 Correlations between stamps and binominal graffiti on pottery from Villar de Brenes. 10.5 Calendar graffito from the first group. 10.6 Three calendar graffiti from the second group. 10.7 Calendar graffito by Lucrio with a consular date of 23 June 158 A.D. 10.8 Calendar graffiti from the fourth group. 10.9 An unpublished graffito from the fourth group. 10.10 The large globular body of the Dressel 20 amphora from Brijuni with an ante cocturam calendar graffito on its base. 10.11 The disposable upper part of a Dressel 20 amphora found in Nijmegen and Augst and Kaiseraugst. 10.12 Calendar graffito on the base of a Dressel 20 amphora from Brijuni. 10.13 Brijuni calendar graffito. 10.14 Geographical distribution of the Asiaticus graffiti. 10.15 The Altenstadt graffito. 10.16 The Lincoln graffito. 10.17 The Wall graffito.

18.03.19 14:39

XIV 10.18 10.19 10.20 10.21 10.22 10.23 10.24

List of figures

The Brough on Noe graffito. The Saint-Thibéry graffito. The Reims graffito. The Béziers graffito, the Vindolanda graffito. Melander graffiti: (1) Adelfa, (2) Testaccio. Definite dates for all the months of the year. Titulus δ and the stamp on a Dressel 20 amphora manufactured during the time of Caracalla (212-217 A.D.) and used in the year 224 A.D. 

Chapter 12 MICROPETROGRAPHY OF THE FAŽANA AMPHORAE 12.1 Geological map of Istria with field sampling locations. 12.2 Terra rossa covers the Mesozoic limestone in the coastal zone of Istria. 12.3 Outcrop of flysch rock series from the interior of Istria, north of Zajci. 12.4 Raw material transported from Trieste bay.

Chapter 11 INSCRIPTIONS ON LATERITE FINDS FROM BRIJUNI 11.1 The final letters EVPOR visible in the retrograde stamp on imbrex. 11.2 Post cocturam capacity graffiti on the dolia from Val Catena.

BEZECZKY_Text.indd 14

18.03.19 14:39

XV

LIST OF PL ATE S THE AMPHORAE Lamboglia 2 (nos. 1-3), Dressel 6A (nos. 4-8), ante Dressel 6B (nos. 9-10) amphorae. Plate 2 Dressel 6B (nos. 11-24) amphorae. Plate 3 Dressel 6B (no. 25), Fažana 1 (no. 26), Fažana 2 (no. 27), Porto Recanati (nos. 28-29), Forlimpopoli (nos. 30-31), Aquincum 78 (nos. 32-33) amphorae. Plate 4 Dressel 2-4 (nos. 34-36), Dressel 7-11 (no. 37), Beltran IIA (no. 38), Dressel 14 (no. 39), Dressel 20 (nos. 40, 42-45), Almagro 50A/Key XXII (no. 46), Almagro 51A-B (nos. 47-48) amphorae. Plate 5 Dressel 20 (no. 41), Almagro 51C (no. 49), Beltran 68 (nos. 50-51) amphorae. Plate 6 Gauloise 4 (no. 52), Mid Roman 1 (nos. 53-54), Palatine East 1/LRA 1 (nos. 55-56), Crete TRC 2 (no. 57), Rhodian (nos. 58-60), Knidian (no. 61), Koan with pinched handle (no. 62), Koan (nos. 6364) amphorae. Plate 7 Koan (no. 65), Dressel 5 (no. 66), Crétoise AC4 (nos. 67-69), Agora G 199/Pinched handle (no. 70), Agora F 66 (no. 71), Agora M 125 (no. 72), Agora M 279 similis (no. 73), Agora M 126 (nos. 74-76), Agora M 240 (no. 77), Kapitän I (no. 78) amphorae. Plate 8 Kapitän II (no. 79), Agora M 273 (no. 80), Late Roman Amphora 1 (nos. 81-86), Late Roman Amphora 2 (nos. 87-88), Late Roman Amphora 3 (nos. 89-90), Ephesus 56 (no. 91) amphorae. Plate 9 Ephesus 56 (no. 92), Agora M 307 (nos. 93-96), Late Roman Amphora / Gaza (nos. 97-99), Late Roman Amphora 13 (no. 100), Samos Cistern (no. 101), Beirut (no. 102), Cretan AC1C / MRC 3 (no. 103), M 235/6 (no. 104) amphorae. Plate 10 Late Roman Amphora 7 (no. 105), Sinope C (nos. 106-107), Hammamet 3 (no. 108), Tripolitanian III? (no. 109), Africana I A/B (no. 110), Africana I variant? (no. 111), Africana II A? (no. 112), Africana II B, Pseudo-Tripolitanian? (no. 113), Africana II C3 (no. 114), Africana II C (no. 115), AfPlate 1

BEZECZKY_Text.indd 15

Plate 11

Plate 12

Plate 13

Plate 14

Plate 15 Plate 16 Plate 17

Plate 18 Plate 19 Plate 20

ricana II D (no. 116), Africana II/III transitional (no. 117), Africana II variant (no. 118) amphorae. Africana II variant (no. 119), Africana III A (nos. 120-122), Africana III A/B (no. 123), Africana III B (nos. 124-126), Africana III C (no. 127), Spatheion 1? or Keay 35A (no. 128), Spatheion 1 (nos. 129-130), Spatheion 1 late (no. 131) amphorae. Spatheion 3A (no. 132), Spatheion 3B (no. 133), Spatheion 3C (nos. 134-136), Keay 3/5 or 64 (no. 137), Keay 8B (no. 138), Keay 59 and Keay 8B (no. 139), Keay 11B var. Keay 1984, Fig. 172.2 (no. 140), Keay 35A (no. 141), Keay 35B (no. 142), Keay 36 (no. 143), Keay 55 (nos. 144-145), Keay 57 (nos. 146, 147) amphorae. Keay 57 (no. 148), Keay 57-56-55 (no. 149), Keay 62 Q or Albenga 11/12 (nos. 150-152), Keay 62 (nos. 153-154), Keay 62 or 61? (no. 155), Keay 61 (nos. 156-157), Keay 61A/D (nos. 158-159), Keay 61? (no. 160), Keay 61C (no. 161), Keay 8A (nos. 162-163) amphorae. Keay 34 (nos. 164-166), Keay 1B ? (no. 167), Storage jars? (nos. 168-169), Miscellaneous (nos. 170171, 173-178) amphorae. Miscellaneous (nos. 172, 179-184) amphorae. Lamboglia 2 (nos. 2-3), Dressel 6A (nos. 5, 7), Dressel 6B (nos. 11-19) amphora stamps. Dressel 6B (nos. 20-22), Beltran 2A (no. 38), Dressel 20 (no. 40), Late Roman Amphora 1 (no. 86), Late Roman Amphora 13 (no. 100), Hammamet 3A (no. 108), Africana III A/B (no. 123). Photomicrographs Dressel 6B (nos. 9, 11-17). Photomicrographs Dressel 6B (nos. 18-22), Beltran 68 (nos. 50-51), Hammamet 3A (no. 108). Photomicrographs Africana I A/B (no. 110), Africana I variant (no. 111), Africana II/III transitional (no. 117), Spatheion I ? or Keay 35A (no. 128), Spatheion I late (no. 131), Spatheion 3A (no. 132), Spatheion 3B (no. 133), Spatheion 3C (no. 134).

18.03.19 14:39

XVI

List of plates

Plate 21 Photomicrographs Spatheion 3C (nos. 135-136), Keay 3/5 or 64 (no. 137), Keay 11B var. Keay 1984, Fig. 172.2 (no. 140), Keay 35A (no. 141), Keay 36 (no. 143), Keay 57 (nos. 146-147). Plate 22 Photomicrographs Keay 62 Q or Albenga 11/12 (nos. 150-151), Keay 62 (no. 153), Keay 61? (no. 160), Keay 34 (nos. 164-166), Keay 1B? (no. 167).

SIAN (no. 12i), L·PETR·AVIT (no. 13a), L·PETR (nos. 13b1, 13b2), M·SERI (nos. 14a1, 14a2, 14a3, 14a4). Plate 34 M·SERI (nos. 14a5, 14a6, 14a7, 14b1, 14b2, 14b3), SISENNAE (no. 15), L·ST·IVSTI (no. 16), TROSI (nos. 17a1, 17a2, 17a3), dolium stamp ARIA(…) (no. 18), dolium graffito LXIII(…) (no. 19).

Chapter 11 INSCRIPTIONS ON LATERITE FINDS FROM BRIJUNI

Chapter 12 MICROPETROGRAPHY OF THE FAŽANA AMPHORAE

Plate 23 [M·ALBI·R]VFI (no. 1), IMP·AVG·G[ER] (no.  2), L·BARBI·L·L·EVPOR[IS] & L·BARBI· L·L·EVPOR[I] (no. 3), Q·G·NICEP (nos. 4a1, 4a2, 4a3, 4a4, 4a5), C·CEIONI·MAXI (nos. 5a1, 5a2), Q·CLODI·AMBROSI (no. 6a1). Plate 24 Q·CLODI·AMBROSI (no. 6a2, 6a3, 6a4, 6a5, 6a6), A·FAESONI·AF (nos. 7a1, 7a2, 7a3, 7a4). Plate 25 A·FAESONI·AF (nos. 7b1, 7b2, 7b3, 7b4, 7b5, 7b6). Plate 26 A·FAESONI·AF (nos. 7b7, 7b8, 7b9, 7b10, 7b11, 7b12, 7b13, 7b14, 7b15, 7b16). Plate 27 A·FAESONI·AF (nos. 7b17, 7b18, 7b19, 7b20, 7b21, 7b22, 7b23-Gnirs 1908, 7b23-Matijašić 1987, 7b24, 7b25). Plate 28 A·FAESONI·AF (nos. 7b26, 7b27, 7b28, 7b29, 7b30, 7b31, 7b32, 7b33, 7b34, 7b35, 7b36, 7b37). Plate 29 A·FAESONI·AF (nos. 7b38, 7b39, 7b40), C·FLAVI (nos. 8a1, 8b1, 8b2, 8b3, 8b4, 8c), L·FVLLONI (nos. 9a1, 9a2-Gnirs 1908, 9a2Matijašić 1987, 9a3, 9a4, 9a5, 9a6, 9a7), C·IVLI·AFRICANI (no. 10). Plate 30 C·LAECANI·P·F (no. 11a ), LAEC (nos. 11b1, 11b2, 11b3-Gnirs 1904/1910, 11b3-Matijašić 1987, 11b4, 11b5, 11b6, 11b7, 11b8, 11c1, 11c2, 11c3, 11c4). Plate 31 LAEC (no. 11d), PANSAE·VIBI (nos. 12a1, 12a2, 12a3, 12a4, 12a5, 12a6, 12a7). Plate 32 PANSAE·VIBI (nos. 12a8, 12a9, 12a10, 12a11, 12a12), PANSIANA (12b, 12c1, 12c2, 12d1, 12d2). Plate 33 TI·PANSIANA (nos. 12e, 12f ), C·CAESAR· PANS (no. 12g), NER·CLAVD·P (no. 12h), PAN-

Plate 35 Polarizing microscopic photos 1. Fired terra rossa soil (IST 12), 2. Chert grain in calcareous sandstone (flysch, IST 64b), 3. Chalcedony sponge spicule in calcareous sandstone (flysch, IST 64b), 4. Calcareous sandstone (flysch, IST 64b), 5. Micritic claystone (flysch, IST 31a), 6. Empty-shelled foraminifera in recent marine sediment (IST 16d), 7. Boring sponge traces on calcareous fragment in recent marine sediment (IST 17), 8. Opalic sponge spicule (acicular) in recent marine sediment (IST 17). Plate 36 Polarizing microscopic photos 1. Opalic sponge spicule (sterraster) in recent marine sediment (IST 16), 2. Thin-shelled clam in recent marine sediment (IST 67), 3. Fired recent marine mud (IST 68), 4. Texture of typical amphora (Castrum 13), 5. Fired terra rossa (IST 12), 6. Chert grain in amphora (Castrum 14), 7. Microsparite filled globigerinida foraminifera in amphora (Castrum 18), 8. Micrites and mollusc skeleton fragment in amphora (Castrum 12). Plate 37 Polarizing microscopic photos 1. Opaque filled globigerinida in amphora (Castrum 16), 2. Chalcedony sponge spicule and terra rossa inclusion in amphora (Castrum 13), 3. Empty-shelled foraminifera and opalic sponge spicule in amphora (Castrum 18), 4. Thin-shelled clam in amphora (Castrum 14), 5. Opalic sponge spicule (acicular) in amphora (Castrum 16), 6. Opalic sponge spicule (sterraster) in amphora (Castrum 16), 7. Boring sponge traces on calcareous fragment in amphora (Castrum 12), 8. Terra rossa soil fragment in amphora (Castrum 22).

BEZECZKY_Text.indd 16

18.03.19 14:39

SHORT HISTORY OF THE WORK SHOP AND THE VILL AS

BEZECZKY_Text.indd 1

18.03.19 14:39

BEZECZKY_Text.indd 2

18.03.19 14:39

3

1

ISTRIA AND THE A MPHOR A WORK SHOPS



Ta m á s B e z e c z k y

1

HISTORICAL BACKGROUND

Istria is a large peninsula located in the northeastern part of the Adriatic region. The Roman Senate considered the possibility of a military occupation of Istria several times during the second century B.C. This was one of the motives for founding the town of Aquileia.1 The town, which was built at a strategically advantageous location, became the most significant trade centre at the head of the Adriatic. It also served as an important military base for the northern campaigns. It was in the interests of the Aquileian troops as well as the Italian merchants to develop a political and economic relationship with the areas north of the Alps. They had an excellent connection with the kingdom of Noricum, as evinced by the fact that the king Voccio sent three hundred Norican cavalrymen to support Caesar in the civil war against Pompey.2 The respectable Italian settlers who came in the wake of Julius Caesar’s conquest promoted the Romanization of the region; Tergeste and Pola were granted the rank of colonia. Octavianus-Augustus’ interest in the region started after the war against the Iapodes in 35-33 B.C. The advantages of the peninsula were discovered during this period. The ties between Italy and Istria developed quickly. As a result, the peninsula became part of Regio X (Venetia et Histria) between 18 and 12 B.C.3 Some believe that Istrian oil production replaced the production in Apulia during the late Republican period. The Apulian region went through an economic crisis, the result of which was a decline in oil production in Brindisi.4 The political changes offered economic opportunities for the new elite. Some of the investors probably came from southern Italy. The traces of several villas have been discovered from Tergeste to Pola.5 The owners of the villas were mainly senators and members of the Roman elite.6 The Roman proprietors of

3 4 5 6 1 2

Livy 40, 34. Caesar BC 1, 18, 5. Degrassi 1953, 54. Baldacci 1967-68, 14; Manacorda 1995, 177. Matijaišić 1988, 27-71, T 2-3. Tassaux 1983-84, 193-229.

BEZECZKY_Text.indd 3

Fig. 1.1 Map of Istria (Tabula Imperii Romani).

the villas cultivated, harvested, and processed the olives (the ancient sources regarded Istrian olive oil as one of the best on the market) and wines.7 They also had their own presses, 7

Pliny the Elder mentions the olive oil from Istria and Baetica in his NH (15.9). Between 77 and 79 when he wrote the book, olive oil was only produced in Istria; production in northern Italy had already ceased. Thus, he praised only the oil of Venafrum. A little later, in about A.D. 102, Martial (12.63) mentioned the olive oil from Istria and Baetica in one of his epigrams. He also considered the Venafrum oil to be the best. Pausanias (10.32.19) in his travel book compared the Tithorea oil with the ones from Istria and Baetica. This description from the second century may have relied on either Pliny or Martial (or some other source), because at that time only Baetican or

18.03.19 14:39

4

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 1.2 Map of Fažana workshop and kiln, I. Map of Fažana, after Gnirs 1910: A - Sv. Marija Church. B - The site where the amphorae were found. C - The site of clay depots. D - Harbour. E - Amphora kiln. F - Ruins of the ancient villa rustica. G - Piazza del Duomo. H - The site where the Laecanius amphorae were found. K - Church, 1-2-3 Excavation 1990-1991; 4 - Trench with three floor levels (Bezeczky 1998a, 4); II. Map of the excavation 2007-2009: after Bulić – Koncani Uhač 2011: T1 - Plan of the amphora kiln 1, T2 - Plan of the amphora kiln 2.

BEZECZKY_Text.indd 4

18.03.19 14:39

CHAPTER 1 – Istria and the amphora workshops (Bezeczky)

storage cellars, and ceramic workshops. However, there are only two excavated amphora workshops in the region. One figlina has been found near Poreč (Parentium) in the bay of Loron,8 and the other is in the north of Pula in Fažana. Stone inscriptions and roof tiles point to other workshops at Novigrad, near Piran, and near Trieste, which may have produced amphorae.9 2

THE FAŽANA WORKSHOP

Anton Gnirs’ excavations at the beginning of the twentieth century found a kiln site under the modern buildings, marked it on a map and published the stamped amphorae.10 However, it seems probable that the description and the map were made only in connection with a new water main (plumbing). The excavations of 1990-1991 used similar methods. The excavated floor layers provided entirely new information regarding the possible existence of a villa to the south of the workshop.11 Since this area is separated by a cistern from the workshop and the kiln, it is possible that there was a villa to house the workers of the workshop. This area may have been isolated from the workshop. Gnirs mentioned the ruins of a press and a dolium north of the kilns.12 They must have been related to the production of olive oil. Another villa is supposed to be here but there is no precise information. Our knowledge increased considerably when the reconstruction of Fažana’s road network began in 2007. The excavations

African oil was still available on the markets. Galen of Pergamon also mentioned Istrian oil at the end of the second century (Buonopane 2009, 27-28). Finally, it is worth mentioning the letter of Cassiodorus (Variae 12, 22-24) from 537/538. He writes about the plentiful harvest of oil, wine and wheat in Istria. They were shipped to Ravenna. However, we do not know what kind of vessels were used. Nor is it clear which part of Istria he was writing about. 8 A team consisting of Croatian, French and Italian members have been working since 1994 in Loron. They unearthed the ruins of the workshop and four chambers of the kiln. No villas have been excavated so far. The amphora stamps have the names of well-known persons. The owners of Loron were T. Statilius T·F·Taurus Sisenna, later the wife of Emperor Claudius MES·CAE (Mes(alinae) Cae(ari uxoris) and Calvia Crispinilla. The emperors (Domitian, Nerva, Traian and Hadrian); Marion – Starac 2001, 97-118; Tassaux 2001, 511512; Manacorda 2010, 217-227. 9 Tassaux 2001, 512-517, with detailed bibliography; TITAC (Titatia Tertia), THAL (L. Quinctus Thallus), L·TERENTI (), C·ALTEN (C. Altenus or Altenius), P·C·QVIR, TRAVL ·ET· CRIS, CRISPINI, T·A·F· CRISPINAE; Koper: Bezeczky 1985, P·ITVR·SAB (P. Iturius Sabinus); Cipriano 2009, 178. 10 Gnirs 1910a, 79-88; Gnirs 1910b, 102; Gnirs 1911b, 37-38. 11 The rescue excavation was in 1991, see Bezeczky – Pavletić 1996, 143-148, Fig. 3 and Bezeczky 1998a, 4, Fig. 3. 12 Gnirs 1910a, 81.

BEZECZKY_Text.indd 5

5

reports mention two kiln chambers, although most of the ruins are below modern buildings.13 There are three periods of the Fažana workshop:14 - In the first period (from ca. 40 B.C.  to A.D.  78), the workshop was the property of the Laecanius gens. The Laecanius family died out without a direct heir in 78 A.D. - In the second period (from the Flavian period to the period of Hadrian, from A.D. 78 to 138) both the property and the workshop were taken over by the emperor Vespasian and were integrated into res privata. - During the third period (ca. the last third of the second century to the early third century), the amphorae of M. Aurelius Iustus were produced. THE PRODUCTS OF THE WORKSHOP The excavations make it clear that the workshop produced dolia, amphorae, stoppers, tiles, spicae, heating pipes, and clay lamps. The most important products of the workshop were the amphorae. Only two of the kiln chambers are known (Fig. 1.2, II). The material composition of the vessels sheds light on the production technology. The petrological analyses have shown that terra rossa, although  available in the vicinity, was not the basic inclusion.15 The bulk of the raw material consisted of flysch. Only a small quantity of terra rossa and recent marine sediment were mixed or added as temper in Fažana. The geological study of the peninsula shows that flysch is available in large quantities in the Koper/ Trieste bay. The Laecanii had huge properties near Trieste (Materia). It seems probable that this area provided the flysch necessary for the production. The amphorae filled with olive oil were shipped to the north Adriatic ports and Aquileia. Then the empty ships may have been filled with flysch as ballast.16 The types of ships the Laecanii used are not known, which makes it almost impossible to guess the quantity and frequency of the shipments. It was during the sailing season lasting from April to October that the necessary quantities of flysch had to be collected. The organization of the shipments must have been a major operation. Olive oil was stored in dolia in the villas and only before shipping was it filled into amphorae. Bulić  – Koncani Uhač 2009, 286-292; Bulić  – Koncani Uhač 2011, 123-128. 14 Bezeczky – Pavletić 1996, 147-148; Bezeczky 1998a, 4. 15 Józsa et al. 2016; see Szakmány  – Józsa in this volume; previous works: Mange  – Bezeczky 2006; Mange  – Bezeczky 2007; Bezeczky – Mange 2009. 16 See Fig. 12.4, Szakmány – Józsa in this volume. 13

18.03.19 14:39

6 3

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

THE OWNERS OF THE WORKSHOP: THE LAECANIUS GENS17

1. The first known member of the family in Istria was Publius Laecanius. He was born circa 90 B.C.  2. The son of Publius Caius Laecanius was also among the founders of Pula. His Dressel ante 6B amphorae and roof tiles have the C·LAECANI·P·F stamp. 3. His son Caius Laecanius was probably decurio. His roof tiles have the LAEC, his amphorae the C·LAEC·A stamp. He was probably born between 55 and 35 B.C.  4. His son Caius Laecanius Bassus was probably a knight. His amphorae have the C·LAEC·BASSI stamp. His amphorae are the first to have two stamps on the rim. (FELIX·SER). The Magdalensberg layers date them to 10-5 B.C.  5. Senator Caius Laecanius Bassus was praetor urbanus in A.D. 32 and consul suffectus in A.D. 40. The fragment of the sodales augustales claudiales mentions him as pater in the year 64. 6. His son C. Laecanius Bassus was consul ordinarius in 64 A.D. The junior Laecanius died of anthrax in 78, which is known from a remark of Pliny’s.18 7. The Laecanii adopted C.  Laecanius Bassus Caecina Paetus. He was consul suffectus in 70 and proconsul of Asia Minor (in Ephesus) in 78.19 He had properties near Minturnae, which are known from the inscriptions of his freedmen. 8. His son C. Laecanius Bassus Caecina Flaccus died in Brindisi when he was 18. 9. C. Laecanius Bassus Paccius Paelignus is also known. He had a dedication to Laecanius (6). There is an important stone inscription from the period of Claudius that documents the property of the senator Laecanius pater (5). The inscription (CIL, V.  698 = ILS 5889 = Inscr. It. X. 4, 376) mentions a property in northern Istria at Materija next to the Rundictes tribe (today part of Slovenia).20 This area used to be hills covered with woods, suitable only for animal farming and the timber industry. The Laecanius family must have had other properties. In Pula the Val San Pietro finds may be pieces of evidence.21 There is a stone inscription in Sv. Mihovil Bajolski (San Michele di Bagnole) near Vodnjan (Dignano) which was I shall follow the way Tassaux 1998, 83 has reconstructed the Laecanius family tree. 18 Pliny NH 26.5. 19 Taeuber 2011; Bezeczky 2013, 221. 20 CIL, V. 698 = ILS 5889 = IIt X 4, 376; Slapšak 1977, 122-128; Tassaux 1982, 248; Milotić – Petrak 2012, 297-310. 21 Gnirs 1910b. 17

BEZECZKY_Text.indd 6

offered to Iupiter by C. Laecanius Ialysos (CIL, V. 14 = Inscr. It. X. 630). There is another stone inscription in Guran, also near Vodnjan (Dignano), which mentioned C.  Laecanius Amycus and his matrimony Phorbe.22 The names Ialis (or Iali) and Amycus occur on a number of amphora stamps and it seems reasonable that the persons mentioned in the inscription and the one who signed the amphorae are the same. The work of the workshop managers/vilici had to be organized. This may have been the job of C.  Laecanius Menander, who was secretary to both the senators (CIL, V. 8142 = Inscr. It. X. Reg. X, 1, 114). THE EMPERORS Paolo Baldacci mentioned that the figlina in Fažana was taken over by the emperors after the family died out.23 The fact that the estate was taken over by the emperors can be established from the evidence the stamping system offers. The production of the amphorae with the emperor’s stamp was discontinued during Hadrian’s era. M. AVRELIVS IVSTVS There is only one amphora with this name. M.  Aurelius Iustus may have been a freedman who was responsible for the workshop and the production in the Castrum villa. There is a votive altar dedicated by M.  Aurelius Iustus to the goddess Flora in the cemetery near the villa.24 We have no more information about this person. He may have been a tenant (conductor) who paid rent to the Imperial treasury. 4

THE AMPHORA STAMPS

We have to update our knowledge and prepare a distinct chronology of the Laecanius and Imperial amphorae. Rescue excavations took place in Fažana, in the course of which new stamps were found and the kiln was excavated.25 Many amphorae with Laecanius and Imperial stamps have been found outside Istria since 1998.26 The list of the ca. 1 500 Matijašić 2001, 342-344. Baldacci 1967-68, 34. 24 ILJug (Inscriptiones Latinae quae in Jugoslavia) 1204; Mlakar 1979, 23-24; Starac 1995, 138, note 27. 25 Paić – Bulić 2008, 17-40; Bulić 2009, 257-270. 26 Belotti 2004; Bezeczky 2005b, 49-50; Bezeczky 2014; Bulić 2009, 264; Cipriano 2003; Cipriano 2008; Cipriano  – Ferrarini 2001; Džin  – Šalov 2008; Gabucci  – Quiri 2008; Gostenčnik 2002; Gugl 2003; Gugl 2004; Mazzocchin 2006; Mazzocchin 2007; Mazzocchin 2010; Mazzocchin 2013; 22 23

18.03.19 14:39

CHAPTER 1 – Istria and the amphora workshops (Bezeczky)

stamps can be updated almost on a daily basis. They provide us with an increasingly precise understanding of the Fažana workshop production. In the late Republican period, an amphora rim with a single stamp C.LAECANI·P·F belongs to the earliest Istrian Dressel 6B form (ante 6B). The shape of the Laecanius amphorae changes somewhat and the classic Dressel 6B shape develops during the Augustan period. Every amphora produced in the Laecanius workshop at that time had two stamps on the rim. The stamp of Laecanius is at the centre (C·LAEC·BASSI), with the second stamp (FELIX·SER) above the handle. Later, from the Tiberian to the early Claudian period, the C·LAEK·BASSI stamp uses the letter “k” instead of a “c”. The Laecanii were the definitive producers and transporters of Istrian olive oil for over one hundred years. From the Tiberian to the early Claudian period (Magdalensberg II) there are 41 different stamp types. The C·LAEK·BASSI stamp and its variants combined with the workshop manager / vilicus names (e.g. A, ADEL, ARCI, BAR, BARB, CAESI, CAR, CLARUS, COM, COMI, EUCHARISTI, FA, FAV, FELIX, FVI, HER, HERME, HOM, IALIS, IALI, L, OPI, OPTA, SPERATUS, SYNT, VIAT). Post Magdalensberg period I we know of 15 stamp types. The workshop managers’ / vilicus names are different from the previous period (e.g. AMETHYSTI, AMYCUS, BAR, CRESCENTIS, DI, EUCHARISTI, FAL, MARTI, NICOMEDE, PIERI). Post Magdalensberg period II the amphorae produced then can be dated with some precision. Before the death of Laecanius, we know of 4 vilicus / workshop manager stamps (CLYMEN, DATI, PAGANI, PTOLEM). These stamps also co-occur with the stamp of the emperor Vespasian. These are the pairs of stamps used during the reign of Vespasian: IMP (CLYME, CLYMEN, DAT, PAGANI, POLL), IMPE·VESP (BARNAE, PAGANI), IMP·VESP (COLONI, PAGANI), IMP·CAES·VESP (DAT, PTOLEM), IMP·CAES·VESP·AVG (CLYMEN). Titus: IMP·T and IMP·T·CAES·AVG; (BERENTS), IMP·T·CAES·AVG (PRIMIGEN); Domitian: IMP· DOMITI and IMP·DOM (LESBI); only one stamp is known from the time of Nerva: IMP·NERVE; Trajan: IMP·TRA (the stamp cannot be read with certainty:

Paić  – Bulić 2008; Pesavento Mattioli et al. 2000; Quiri 2009; Quiri 2011; Sakl-Oberthaler 2000; Schimmer 2009; Schindler-Kaudelka 2000; Starac 2001; Starac 2010a; Tiussi 2008; Tiussi 2010; Toniolo 2011; Vidošević 2003.

BEZECZKY_Text.indd 7

7

MANI?; SERV, and there is an incomplete stamp …VS); Hadrian: IMP·HAD.27 The stamp of M.  Aurelius Iustus is very different from the stamps of the Laecanian and Imperial amphorae. This is a stamp in hollow lettering, in the genitive case. Only one similar stamp was found during the new Fažana excavation: F♠A with a pattern of leaves.28 The list of the stamps is in the APPENDIX. 5

THE PRODUCTION CAPACITY OF THE VILLAS AND THE DISTRIBUTION OF THE FAŽANA AMPHORAE

As we shall see later, three of the villas with olive cellars have been excavated on Brijuni. The dolia in the cellars may be suitable for finding out the volume of the olive oil production. There are a number of fragmented dolia in the cellars. The capacity calculations were based on a complete dolia exhibited on the island (Fig. 1.3). Its form and material corresponds to the pieces in the cellae olearia in the Verige and Castrum villas. The capacity is approximately 1750 litres.29 Gnirs found numbers on the side of some of the dolia.30 The numbers may have referred to the quantity of the filled amphorae. However, we do not know when or who wrote these numbers on the dolia. The villas thus had the following capacities: Verige / Val Catena 56, Kolci / Monte Collisi 100 and Castrum 48 dolia. Accordingly, the island produced about 357  000 litres of olive oil per annum, plus the unknown quantities of Fažana and Dragonera. This is, of course, a rough estimate. Around 9 400 amphorae would be needed to transport 357 000 litres of olive oil. It should also be added that the production may have fluctuated from year to year. The capacity of the Dressel 6B amphorae was 37.9-39 litres. However, only a few complete amphorae are known.31 Published from a private collection by Starac 1994-95, 162, T 12.1. Paić – Bulić 2008, 28, no. 34. 29 I am grateful to Martino La Torre for the information: the ceramic capacity is approximately 403 l, which, if the gross density is 1.6, means that its weight when empty was around 645 kg. Thus, the dolia stood on a gravel bed to dry, and the convex shape of the floor was preserved. 30 Gnirs 1908b, 179; Bezeczky 1998a, 71-72; see also Berni Millet in this volume. 31 Ca. 12 congius = 39.3 litre. Complete amphorae Novaria: Bezeczky 1998a, nos. 3, 195, 197, 361 and 405; Carreum Potentia: Riva 1987, 92, Fig. 26, Pl. Va. 4 and Fig. 25, Pl. Va, 6a-b = Bezeczky 1998a, nos. 439 and 624; Opitergium: Cipriano – Ferrarini 2001, 49,

27

28

18.03.19 14:39

8

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 1.3 Dolium from Brijuni harbour, the position of the dolia in the Verige and Castrum villas (Drawing: La Torre 2017, Photo: Bezeczky 2014).

BEZECZKY_Text.indd 8

18.03.19 14:39

CHAPTER 1 – Istria and the amphora workshops (Bezeczky)

9

Fig. 1.4 Complete Laecanius amphora from Magdalensberg; stamps: C·LAEK·B / FELIX·PET (Drawing: La Torre 2017, Photo: Bezeczky 2014).

The height of the amphorae was 87-92  cm (Fig.  1.4). The outer diameter of the 604 amphorae studied was 12-17 cm. The most common diameter of the rims was 14-16 cm. This is true of 85 % of the vessels. This also means that the potters followed the model quite closely. We have little information about the size of the Fažana 1 and Fažana 2 amphorae. Only the upper parts are still around. The rim of the Fažana 1 amphorae was 12-14 cm, while that of the smaller Fažana 2 amphorae was 8-9 cm. We don’t have complete amphorae in Fažana and Brijuni, which makes their capacity unknown.

nos. 51, 54, 57, 58, 62, 65, 78 and 91; Patavium: Pesavento Mattioli et al. 2000, 37-38, Figs. 4, 7; Cipriano  – Mazzocchin 1998, 365, no. 5, T 1, 5; Ateste: Cipriano – Mazzocchin 1998, 365, no. 2, T 1, 2; Vicentia: Mazzocchin 2013, 133, 25.11e; Magdalensberg: Schindler-Kaudelka 2000, 390, G1.

BEZECZKY_Text.indd 9

The volume of production can only be guessed at on the basis of the amphorae which were found in more than 75 sites over the last 100 years. The Dressel 6B amphorae can be found in northern Italy along the Po valley all the way to Torino.32 In Raetia, the transportation route of the amphorae from Fažana can be traced to Curia, Bregens, Oberstim. In Noricum and Pannonia, along the Danube, Istrian oil was used all the way up to Aquincum in the north and Rittium in the south. In more central areas, the most important places where the amphorae have been found Magdalensberg, Aguntum, area Siscia, Sirmium and settlements along the Amber Route.33

Gabucci – Quiri 2008, 68-69; Mazzochin 2013, 78-81; Cipriano 2009, 173-185; Bezeczky 1998a, 75; Bezeczky 2014, 247252. 33 Bezeczky 1998a, 75, Fig. 47. 32

18.03.19 14:39

10

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 1.5 The distribution of the Fažana amphorae (Drawing: Bezeczky 2015).

BEZECZKY_Text.indd 10

18.03.19 14:39

11

2

THE VILL AS OF BRIJUNI



M artino L a Torre – Tamás Bezeczky

Pliny the Elder called the Brijuni Islands the Pullariae, a name also used by the Peuntiger Table.1 The Brijuni Islands are a group of fourteen islands. The largest island is Veli Brijun (579 hectares).2 Eight villas on the island of Brijuni and the surrounding islands were identified. Unfortunately, it was not possible to excavate more than three villas (Verige – Val Catena, Kolci – Monte Collisi and Castrum). Another villa on the mainland, 5 km north of Fažana, called Dragonera, was recently explored. The amphora stamps suggest that the villas formed part of the Laecanius family’s properties.3 Another villa that was excavated in 1909-1912 by Gnirs is situated ca. 700 m south of Fažana Center surrounding a 60 m-deep bay called Val Bandon almost directly opposite the large complex in Val Catena on the Brijuni island.4 1

VERIGE BAY (VAL CATENA)

The Verige villa (Val Catena), located in a bay, was a luxurious maritime terrace villa (villa maritima). Its length (from east to west) measured approximately 450 meters. It may have been possible to close off the bay with a chain (catena) between the north and south moles, like in the Lion Harbor of Miletus in Asia Minor.5 The complex was composed of separate buildings standing in a row along the shoreline of the narrow bay, which is more than 800 meters long. We find the following buildings on the northern shore (from east to Pliny NH 3, 26, 151; Peutinger Tablet segm. IV.1; Geographus Ravennas (V, 24, 101 = no. 165). 2 Veli Brijun or Brioni Grande; in the text we used the current designation Brijuni. 3 A rescue excavation took place at the Dragonera villa located 5 km north of Fažana, near Peroj, Starac 2010b. This building is a typical villa maritima, located on the seashore. It has its own port as well as agricultural and residential sections. The southern wing of the villa, located on the shore, was intended for residential purposes. The western and eastern wings of the complex both contain facilities for pressing olives. There are amphora rims with Laecanius and Imperial stamps. 4 Gnirs 1911d, 155-186. 5 It is also conceivable that the term dates back to the Roman period, as the chain would presumably have long been lost in the period after Antiquity, Weisshäupl 1900, 204. 1

BEZECZKY_Text.indd 11

west): the north mole; an industrial area (fullonica?) with water reservoirs and a small bath; the large thermae (F), with the palaestra (G) and a vivarium (E) beneath the water. Further to the west, we find a diaeta, an atrium (H) and a long portico (K), with a cryptoportico behind it. This made it possible to access the facilities on the northern shore even when the weather was bad. Another diaeta (L) leads to the temple area (M-P-N) in the western part of the bay.6 The living quarters and the villa rustica on the southern shore were connected to it by a peribolos hall in the shape of a semicircle, which was behind the three temples. This hall completed the temple area architecturally. There was another cryptoportico beneath the hall.7 Gnirs found a coin from the year 41 A.D. in the stonework of the hall.8 The main building (domus), in the southern part of the bay, is what is called a terrace building (R).9 The building had two parts. The western wing housed the living quarters, while the eastern wing held the industrial area.10 A luxurious villa maritima was combined with a villa rustica here. Unlike a villa urbana, which generally did not look imposing from the outside and was organized around its courtyard, these villas were open to the surrounding countryside. The two peristylia were U-shaped and opened to the sea. The industrial area included the oil presses in room X, but only the channels were observable at the time of the excavation.11 These channels connected the presses and the vat Gnirs 1904, 145-146. There were more than a hundred Laecanius amphorae in the debris of room L. The amphorae had the following stamps: BARB, COM, COMI, IALIS and VIAT. All of them belong to the first period of the workshop. The fragments of late Roman amphorae (types 1 and 2) were found near the staircase of room F, Bezeczky 1998a, 55. 7 Gnirs 1908b, 169-170. 8 Gnirs 1904, 140. 9 Gnirs 1915, 100-129. 10 Schrunk – Begović 2000, 255-257, assume that the industrial area was built first, and the other parts of the villa were constructed in two later phases. 11 Gnirs 1906, 33; Gnirs 1908a, 137 mentions a wine press and wine cellar. He presented no arguments. By contrast, Rostovtzeff 1957, 237 speaks of oil production. Later Gnirs 1915, 119 mentions an oil cellar (cella olearia); Schrunk – Begović 2000, 256, mention a 6

18.03.19 14:39

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.1 Map of Brijuni Islands and Fažana (Drawing: La Torre 2017).

12

BEZECZKY_Text.indd 12

18.03.19 14:39

13

Fig. 2.2 Map of the villa in Verige and the terrace building (Drawing: La Torre 2017).

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

BEZECZKY_Text.indd 13

18.03.19 14:39

14

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.3 Map of the terrace building (Drawing: La Torre 2017).

(lacus, 7.25 × 1.25 m, max volume 4.44 m3). The lacus was at the end of the adjacent cellar. The walls of the vat were made of opus signinum. It had an opus spicatum floor. There were four rows of sunken dolia in the cellar. The columns which held the roof divided the cellar into two naves. The stylobate (the top edge of which has not been preserved) stood 30-50 cm above the threshold to the portico. The size of the cellar was 8 × 27 m. It contained 56 (perhaps 60) dolia. The narrow spaces on the bottom terrace of the villa rustica may have served as a place to store amphorae. There are two large water reservoirs (V) made of opus caemeticium on the top of the hill above the terrace-building. It used the water from the karst well of Gradina-bunar (Z).12 The well was across from the terrace building. A small nymphaeum was erected next to the well. The water was lifted with a bucket-chain system from the well and emptied into two large reservoirs. These reservoirs were 18-20 m above wine press and wine cellar. If the villa was the property of the Laecanius family, it was undoubtedly used for olive oil production. There are no amphorae for the production and export of wine. The same is true of the other Laecanius villas. Only Dressel 6B amphorae produced locally have been found. 12 Bezeczky 1998a, 56, at the well of Gradina-bunar there were Dressel 2-4, Lamboglia 2/Dressel 6A, Gauloise 4, Dressel 6B (with Laecanius stamps), Forlimpopoli, Keay XXV, Spatheion, Almagro 50 and LR 1 amphorae.

BEZECZKY_Text.indd 14

sea level, which provided sufficient pressure for the water to reach the reservoirs on the hill on the southern side of the bay above the main house. They were connected by lead pressure pipes, which were installed in a canal.13 The two reservoirs were constructed in two phases, from east to west; the buttresses of the western reservoir were added in a third phase. The eastern reservoir was divided into three sections. This was necessary because major cracks in its side walls had caused leaks. The new divisions were placed exactly where the cracks were and sealed them. Both reservoirs were emptied through an opening to the north. In addition to the reservoir directly on the karst spring on the opposite slope (Gradina-bunar), there were further water storage facilities: two reservoirs to the northeast of the buildings, above the north mole, which also had a karst well in front of them; another reservoir to the northwest of the diaeta; and the cistern underneath the peristylium of the villa’s industrial area. In addition to the agricultural/industrial infrastructure, there were three bathing facilities: a small bathhouse in the northeast, at what is now the shoreline,14 the large thermae across from the terrace villa and the small thermae to the west of the living areas. 13 14

Weisshäupl 1899, 81-82; Gnirs 1908b, 174-175. Gnirs 1915, 134.

18.03.19 14:39

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

15

Fig. 2.4 Map of Kolci villa, after Gnirs 1908 (Drawing: La Torre 2017); A - living area, B - courtyard, C - oil cellar, D  - kitchen, E  - oil presses, F - lacus.

Thanks to the two springs, both the villa’s agricultural section and its luxurious facilities, including the wells and thermae, were as independent as possible from rainwater. An enclosed garden and the south mole were to the east of the villa rustica. The buildings on the northeastern part of the bay (A-D) were regarded both as the industrial area of the villa15 and as an independent villa rustica.16 The function of these buildings is controversial. During the first period, the villa was fairly simple with industrial areas and two reservoirs.17 The glass and ceramic objects found here are fairly varied with respect to their chronology and type.18 In 1996-1997 Croatian and German archaeologists were involved in underwater research in Verige bay, next to the Austrian pier.19 Begović Dvoržak 1990, Fig. 1, no. 9. Gnirs 1915, 146-147; Matijašić 1982, 56-57. 17 Gnirs 1906, 29-32, 35-36; Gnirs 1915, 146-147. 18 Bezeczky 1998a, 57, Lamboglia 2, Dressel 6A, Dressel 6B, Knidian, Dressel 25 and Gauloise 4 amphorae were identified; Vitasović 2004, 93-98 described some tile stamps (M·SERI, [A·FAE]SONI[·] A[F] and L·PETRO), see Berni Millet in this volume. 19 Under the direction of Mario Jurišić and Helmut Bender; see Bloier 2012. Unfortunately Bloier’s work is little help. In his book, a num 15 16

BEZECZKY_Text.indd 15

2

KOLCI (MONTE COLLISI)

Gnirs found a villa rustica on Kolci hill (Monte Collisi), on the northwestern part of the island.20 The three wings of the villa were built around a central courtyard (B). The living quarters (A) were in the southwestern wing. The kitchen (D), the four oil presses (E), a vat (F) and the large cellar (C) were to the east. Three adjoining rooms were arranged to the southwest of the presses. Every press had a large (2.18 × 0.86 m) stone block.21 There were two rectangular holes in the stones for the wooden uprights (arbores). The diameter of the concave press beds (ara) was 1.9 m and it was made of small ceramic bricks (spicae). There was a stone channel ber of amporae are misidentified, which makes his conclusions doubtful. For example, some of the Fažana amphorae have not been identified properly: nos. 950, 952, 967 are not Dressel 6A, they are rather Fažana 1; no. 991 is Fažana 2; no. 960 is not ante Dressel 6B; nos. 979, 980 are not Dressel 6B; no. 1045 is Dressel 6B; no. 1090 is not Schörgendorfer 558; no. 1000 is perhaps Kapitän 2 and not Dressel 9 similis; no. 915 is not Agora M 315, it is rather Agora m 279; no. 1095 is not Knossos 36 but rather Ephesos 56. 20 Gnirs 1908a, 134-136. 21 Gnirs 1908a, 135, mentioned four presses, of which only three have survived.

18.03.19 14:39

16

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.5 Castrum villa after Mlakar 1975-76 and Begović – Schrunk 2007a (Drawing: La Torre 2017). 1 - Prehistoric settlement, 2 - Remains of the salt works, 3 - Late Antique masonry tombs, 4 - Remains of the sarcophagus in situ, 5 - Church of St. Mary, 6 - Tombs in the narthex of the church, 7 - Late Roman settlement, 8 - Remains of the structures alongside the salt works and quarries, 9 - Late Antique building, 10 - Location of the ara dedicated to the Goddess Flora, 11 - Quarry.

BEZECZKY_Text.indd 16

18.03.19 14:39

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

around it. This channel (12 × 4 cm) led the oil into the vat (lacus, 8.65 × 2.30 m, max. volume ca. 13 m3) with an opus spicatum base. The rectangular oil cellar in the eastern part of the villa was evenly divided into two naves and contained about 156 dolia, Relying on Cato’s estimates, there may have been 100 olive oil dolia.22 The water supply of these facilities is unknown. Some objects were published from the villa: a vessel with Barbotine decoration, one terra sigillata without any precise definition, many roof tiles and stamped Laecanius Dressel 6B amphora fragments.23 There were Spatheion and Africana III amphora fragments in the rubble of Gnirs’ excavation.24

17

The material contained in various plastic bags was mixed up in the course of storage and transportation. The fine ceramics and glassware were stored separately, and we were unable to examine them. An amphora bearing the stamp FELIX·PET is missing from the finds as well as the tegula with the stamp LAECANI·P·F.28 Due to the site’s long excavation process, the missing documentation of the process and the restoration work, which was carried out in conjunction with the excavation, the findings at the site can only be interpreted in a very limited manner. A review of the documentation and the published articles shows that we do not know enough about the history of the villa. Many questions remain unanswered.

3 CASTRUM The Castrum villa is located on the western coast of Brijuni, in a Madona bay (next to the small churches of St. Mary and St. Peter). Saline Bay is located to the southeast of the villa. A number of important articles were published in connection with the villa. The excavations in the Castrum villa were started by Gnirs, continued by Mario Mirabella Roberti, Štefan Mlakar and Anton Vitasović.25 Gnirs excavated St. Mary’s Church, near the Castrum villa.26 As the villa excavations continued, with breaks both short and long, for nearly 80 years, we do not have access to all of the data. After Gnirs, Mlakar wrote the first longer article, which appeared in Histria Archaeologica in 1976. He published a detailed map, which was revised by Vitasović following an excavation in the 1980s.27 The original documentation of the excavations was unavailable, so we could only use the information provided in the publications. However, we were able to study five thousand photographs of the excavations in the Archive of the Brijuni Museum. The layers were 1.6-1.8 m thick in many places, which means that they were not suitable for stratigraphic analysis.

Cato De Agr. 12, 10, mentioned a 240 iugera estate which produced oil needing 100 dolia for oil, 10 dolia for amurca, 10 dolia for wine, 10 dolia for grape pomace, 20 dolia for corn. 23 Gnirs 1908a, 136; Gnirs 1908b, 177; C·CEIONI·MAXI and […?PAN]SIAN[A], see Berni Millet in this volume; Bezeczky 1998a, 62, Cat. 59, 154 and 155, C·LAEK·B, C·LAEK·B / [CO]M, C·LAEK. 24 Found in the depot, Bezeczky 1998a, 62, inv. no. B 221. 25 Gnirs 1911a, 75-97; Mirabella Roberti 1935, 294-295; Mlakar 1975-76, 5-50; Vitasović 2005, 157-210; Vidrih Perko – Pavletić 2000, 263. 26 Gnirs 1911a, 75-93. 27 Mlakar 1975-76, 50; Vitasović 2005, 181. 22

BEZECZKY_Text.indd 17

THE FIRST ROMAN VILLA The earliest villa was found near the shore. Many researchers dated this villa to the second century and the first century B.C. They thought that this villa was probably destroyed in the middle of the century. According to one opinion – based on Mlakar’s excavations – there was a military camp on the site of the villa during the Istrian wars of the second century B.C.  “Before the conquest of the Histrian hill forts Faveria and Mutila on the Istrian mainland, and the destruction of Nesactium in 177 B.C., the Roman army came from Aquileia escorted by navy, which carried supplies. Madona bay, oriented towards the west and shielded from the view from the mainland, was the most logical location for the first Roman camp and naval base on Brijuni. In the first century B.C., during the civil war between Caesar and Pompey, such a military camp in Madona Bay may have had a significant function. It is possible that it suffered damage as a naval base in the course of battles between Caesar’s and Pompey’s navies in which the numerous Histrian population held the side of Pompey.”29 Based on the amphorae excavated at the Castrum site, it is difficult to identify the earliest villa as a military camp. It is not possible to determine the function of the rooms on the basis of the walls of the villa. The published map does not resemble a camp. Furthermore there are no amphora finds from the second century (e.g. Greco-Italic, Dressel 1, Brindisi amphorae). In a later article, the authors mentioned above described the villa in an entirely different way:30 “The earliest architectural finds in Dobrika Bay belong to a modest rustic villa, dated by Mlakar to around 129-79 BC.  Its Inv. no. BA1867 and BA 4132; Matijašić 1998, 253. Begović – Schrunk 2007a, 96. 30 Begović – Schrunk 2010, 252. 28 29

18.03.19 14:39

Fig. 2.6 Castrum villa, after Mlakar 1975-76, A-Area of the big Roman villa, B-Area of the oblong building with a circular corner and a rectangular risalit at the side, C-Area of the Late Antique and Byzantine buildings, D-Area of buildings from the Early Middle Ages and recently erected parts.

18

BEZECZKY_Text.indd 18

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

18.03.19 14:39

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

19

Fig. 2.7 Archive pictures from the excavations (Photo: National Park Archive).

Fig. 2.8 The first Roman villa (Photo: Bezeczky 2014).

BEZECZKY_Text.indd 19

18.03.19 14:39

20

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.9 Map of the first Roman villa (Drawing: La Torre 2017).

remains were found along the coast, below the level of a larger rustic villa which used the two walls to construct its foundations, leaving rooms of the earlier villa in its substructure. The villa from the second century BC has a large production area, which comprises approximately one third of its entire space, testifying to its production function.” Vitasović formulated another very similar opinion:31 “The material remains and social circumstances date the villa back to the period between the second half of the second century BC and the first quarter of the first century BC, that is, to the period between 129 and 79 BC. Even though the facility has not been fully researched, the material remains and finds offer an insight not only into dating but also into the owners’ various activities and needs. The life necessities of the people who occupied the villa were determined by the economic resources ( fertile land, animals and the sea) which dictated the development of agriculture, livestock breeding and fishery. This antique enclave, which is relatively distant from any hill forts, existed

until the time of the civil war between Anthony and Octavian, when it was destroyed. In the first half of the first century AD, a big Roman villa, 51 × 58.7 meters in size, was built above, to the east and to the north of the above-mentioned villa.” The period of civil war between Caesar and Pompey can be connected more readily to the villa. However, as we do not have any information regarding the layers we do not know where the large quantity of Lamboglia 2 amphorae were unearthed. Regarding the coins, only a single AS from the Republican period has been found.32 The published terra sigillata refer only to the Augustan era.33 We do not know who used the first villa in the beginning and when it was abandoned.

32 31

Vitasović 2005, 208.

BEZECZKY_Text.indd 20

33

Miškec 2002, 177; Vidrih Perko – Pavletić 2000, 265. Vidrih Perko – Pavletić 2000, 263.

18.03.19 14:39

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

21

Fig. 2.10 Map of the second Roman villa (Drawing: La Torre 2017).

THE SECOND ROMAN VILLA The second villa was built above and behind the walls of the first building. It was a U-shaped villa, open to the sea. Its large courtyard was surrounded by a living area and the agricultural buildings. The various representations of the living area are not identical.34 There is a small cistern in the large central courtyard for collecting rainwater. There was clearly a well nearby, providing the inhabitants of the villa with water to drink. As described by M. Porcius Cato, many things were needed in a villa rustica which produced olive oil: presses and dolia, agricultural equipment (cart, iron plough, harrow) and stables.35 There was not a great deal of space available at the villa to house the work animals (oxen, donkeys, occasionally sheep) used for transportation and everyday work. The stables and other such buildings were presumably outside the walls of the

34 35

Begović – Schrunk 2011A, 378; Vitasović 2005, 163-166. Cato De Agr. 5, 2 and 12, 10.

BEZECZKY_Text.indd 21

villa. It is possible that some of their functions were fulfilled jointly with the other Laecanius villas (Verige and Kolci). Three presses were reconstructed at the highest point of the villa, next to a lacus and a cellar. We do not know what the original form of the presses was; we only have a reconstruction made of stones. The measurements of the thus reconstructed presses match those excavated at Kolci (Monte Collisi) exactly. The bottom of the lacus is covered with spicatum (12 × 6.5 × 2.1 cm). Its bottom band walls are made of 21 cm (in some places 23 cm) thick opus signinum walls, and this layer insulates it from the stone walls. To the mortar of this layer of opus signinum pozzuolana (Vesuvian area) and ceramic fragments had been added to ensure the mineral hydration to give a waterproof cover. The waterproof layer was needed, because it prevented the oil and amurca from leaking out of the lacus. This is exactly how the huge cisterns in the Castrum and Verige (Val Catena) villas and the basins near the Gradina-bunar well were insulated. There is a large cellar next to the lacus in the Castrum villa. It is 8 m wide and 23.3 m long on the inside. The walls

18.03.19 14:39

22

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.11 The restored presses in the second villa (Photo: Bezeczky 2014).

are 45 cm thick. The cellar is divided in the middle by a row of columns. The dolia were sunk into the floor in four rows, 12 in a row, with 48 pieces in total. One dolium bears the stamp ARIAL(...).36 Many modern authors who have written about the history of the villa assume, without any proof, that there was wine production in the first century A.D.37 Locally made vessels for transporting wine have not yet been found. However, the large quantity of Dressel 6B amphorae produced in the Fažana workshop and found in the villa are the most important evidence of oil production and transportation. These amphorae bear the stamps of the owners (Laecanius family and the Emperors). The situation is similar in the other two villas in Brijuni. All of the buildings, presses, cellars, dolia, and cisterns were made of the same materials and based on the same concept.38

The tegulae help to date the villa; they include various stamps from the Pansiana workshop (PANSAE·VIBI) and the stamps Q·CEIONI·MAXI and M·SERI. They were all produced in the middle of the first century B.C., similarly to the C·LAECANI·P·F tegula found here.39 One may thus assume that the Laecanius family already possessed the villa at the time. This is supported by a stamped Dressel 6B amphora (earliest form). The stamp on the amphora matches the tegula stamp perfectly (C·LAECANI·P·F). Further tegulae with the LAEC stamp can be dated to the early Augustan period, which indicates that the villa continued to be in use.40 Only a few of the coins are dated to the Augustan period.41 There are several mill stones inside and outside the cellar. Some of them show oval wear. They belonged to the “stone roller/broyeur á rouleau” type of mill. There are fragments of

See Berni Millet in this volume. Begović – Schrunk 2007a, 42 and 97 = Begović – Schrunk 2010, 255-256 = Begović  – Schrunk 2011a, 377; Vitasović 2005, 209. 38 According to the micropetrographic analysis by Józsa and Szakmány, the dolia of the Castrum villa show a variable composition and texture. The groundmass with fine-grained non-plastics is more or less similar in both types, but the quantity of coarse temper is very different (approximately from 0 %-15 %). The groundmass of both dolia types consists mainly of clay with a small amount of fine-grained quartz, muscovite, strongly fired globigerinida and carbonate grains. In most of the dolia, there is a considerable amount of non-plastic material which is very large in size and consists of angular siliciclastic material (dominantly quartz-quartzite, rhyolite, chert-radiolarite-

spiculite, amphibolite, granite, flysch originated [partly slagish] sandstone clasts) and well rounded calcareous rock and terra rossa grains. The groundmass of dolia is very similar (except for the lack of recent marine opalic spiculae) to the material of the main type of Laecanian amphorae, so their raw material must have a very similar source (northwest Istrian flysch). The extremely large non-plastic material was most probably artificially added as temper. This kind of material with such a composition and grain size is not known in Istria. 39 See Berni Millet in this volume. 40 Perhaps connected with the stamped amphora C·LAECA, Pais 1884, 1077.86. 41 Miškec 2002, 177, one Dp 17 B.C. (RIC 342), two As 15 B.C. (RIC 382) and 7 B.C. (RIC 431).

36

37

BEZECZKY_Text.indd 22

18.03.19 14:39

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

23

Fig. 2.12 In the lacus, spicae and opus signinum wall (Photo: Bezeczky 2014).

a mill (mola olearia) near them.42 These were used during oil production. The earliest finds from the villa are the Lamboglia 2 amphorae, which contained Adriatic wine, and the Dressel 6B (ante 6B) amphorae, which contained olive oil. It is not possible to determine to which phase of the villa they belong. Many Dressel 6A, Dressel 6B and Dressel 2-4 amphorae, as well as amphorae from Rhodes, Kos, Knidos and Crete were found.43 These can be dated to the period from the late first century B.C.  to the second century A.D.  Some coins are dated from Tiberian to the Hadrianic period.44 Few stamped Dressel 6B amphorae from the Tiberian– Claudian period have been found (11-17). We know of only a few amphorae with an Imperial stamp in the period from Vespasian to Trajan (18-22). However, these stamped amphorae prove that the villa was continuously in use. This is also supported by a number of unstamped Dressel 6B amphorae. The crisis during Hadrian’s reign is indicated by a Dressel 20 type stamped amphora (40) from Baetica. Why Mlakar 1975-76; Bezeczky 1998a, 65; Vitasović 2005, 191; Matijašić 2008, 292-293. 43 Vidrih Perko – Pavletić 2000, 265-266; see Bezeczky (Chapter 4) and González Cesteros in this volume. 44 Miškec 2002, 177-178, Tiberius two As 15-15 and 34-37, two As (Divus Augustus) 15-16 and 22-30, Caius one As 37-38, (Germanicus) two As 37-38 and 40-41, Claudius I one As 41-54, Julio-Claudian dynasty one As 27 B.C.-68, Vespasian or Titus one As 69-81, Traian two As 101-102 and 103-117, Hadrian one S117-138 and one Dp 117-138, (Sabina) one D 119-137 and one As 119-137.

42

Fig. 2.12

BEZECZKY_Text.indd 23

18.03.19 14:39

24

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.13 The oil cellar and fragmented dolia in the second villa (Photo: Bezeczky 2014).

Fig. 2.14 Mill stone and a mola olearia (Photo: Bezeczky 2014).

BEZECZKY_Text.indd 24

18.03.19 14:39

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

25

Fig. 2.15 Map of the third Roman villa (Drawing: La Torre 2017).

would an olive oil-producing villa have needed to import oil from far away? Clearly because local production stalled. Oil was also imported in amphorae from Baetica under the reign of Antoninus. The same is true of the African I A/B (110) and Africana II A (112) amphorae which were transported to the villa during the mid-second and mid-third centuries. THE THIRD ROMAN VILLA It is not known when the third villa was built. Inside the villa, there was another press-room and a new, small cellar on the northern wall. As with the other presses, we do not know their shape, because they are to be found underneath walls that were built later on. It is clear that the reconstruction was built during the excavations. The small cellar is located next to the large one. Its roof was held up by two columns. The form and material of the smaller dolia here differ from those in the large cellar. It is possible that these dolia contained seeds rather than oil or wine.

BEZECZKY_Text.indd 25

There are fewer fine ceramics and amphorae in the villa from last third of the second century.45 The votive altar stone dedicated by M. Aurelius Iustus to the goddess Flora in the cemetery near the villa can be dated to this period.46 It seems likely that M. Aurelius Iustus used the villa; he may have been a tenant (conductor) who paid rent to the Imperial treasury. The Fažana 1 type amphorae that he produced can be found both in the villa and in the workshop in Fažana. There is only one amphora with this name M. Aurelius Iustus on the stamp. Both the amphora and the stamp differ significantly from the traditions of the workshop. The small amphorae (Fažana 2) dated to the second or third century pose a mystery. They probably contained fish sauces. It is not inconceivable that salt production took place next to the Castrum villa (Saline bay), and that the salt was used in the production of

According to Girardi-Jurkić 1985, 92, the production of the villa stagnated for about a hundred years. Pröttel 1996, 165 and 168. 46 ILJug 1204; Mlakar 1979, 23-24; Starac 1995, 138, note 27; Bezeczky 1998a, 16. 45

18.03.19 14:39

26

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.16

fish sauces. Few small amphorae have been found on the territory of the villa, and there are no basins or buildings related to fish processing. There are some Cretan fragments of form MRC 3 and Agora M 125 with micaceous fabric and Agora M 240 – the fabric without mica – Kapitän 1 and Kapitän 2 amphorae from the Aegean, Africana IIC and Africana IID dated in the third century. The circulation of coins was continuous from Antininus Pius to the Valentinianus III period.47 Regarding the villa, they highlight “Probably in the same period a fullonica was installed in courtyard of the villa rustica in Madona bay. Among the finds were furnaces, a cistern and remains of presses for extracting liquid from the fabrics, spaces for drying fabrics and final processing.”48 The workshop was presumably used by an Imperial procurator. It may perhaps be from a later period, when the state manufacture of military clothing was connected to the Balfium Cissense Venetiae et Histriae, mentioned by the Notitia dignitatum 435.49

Miškec 2002, 178-187. Suić 1987, 201; Matijasić 1998, 241-251; Begović – Schrunk 2007a, 98; Tassaux 1998, 96-97. 49 ND Oc I.  53, Oc XI, 67 Procurator balfii Cissensis; Begović  – Schrunk 2007a, 98-99. 47

Fig. 2.16 The smaller cellar and the dolia (Photo: Bezeczky 2014).

BEZECZKY_Text.indd 26

48

18.03.19 14:39

CHAPTER 2 – The villas of Brijuni (La Torre – Bezeczky)

27

Fig. 2.17 Map of the Late Roman and Byzantine villa (Drawing: La Torre 2017).

THE LATE ROMAN AND BYZANTINE VILLA OR SETTLEMENT At the time, the villa was surrounded by polygonal walls to protect it from attacks. The walls were weakly fortified and had no towers. Based on the ovens which can still be identified, a little more than 30 families lived within the walls in the Byzantine era. The authors refer to historical events when they describe the late Roman/Byzantine villa. These references may provide important clues. However, the role of the villa can only be guessed at. Starting in the fourth century, the villa was presumably not used as a military site; it may have been closely connected to the neighboring St.  Mary’s and St. Peter’s churches.50 A number of wine presses were found in the villa, but they are also difficult to date as we do not have any information

regarding the layers.51 It is clear that they were used to produce a significant quantity of wine. In the villa there are very significant amphora finds from this era. The amphorae come from the Iberian Peninsula, North Africa, Italy, the eastern Mediterranean and the Black Sea.52 Most of the amphorae contained wine, olive oil and fish sauce. This raises the question of what the villa produced, and why it was necessary to import these commodities. Among the late Roman and Byzantine amphorae,53 there are many Aegean and North African terra sigillata.54 The owners of the villa may have been very wealthy. Or perhaps it was the ecclesiastic leaders of St. Mary’s and St.  Peter’s churches who ordered the expensive foodstuffs. For details of this period, see the next chapter.

Matijašić 1993a, 249. Vidrih Perko – Pavletić 2000, 266. See the Part II chapters in this volume. 53 For the late Imperial and Byzantine finds, see: Girardi-Jurkić 1985, 93-94, nos. 252-265. 54 Pröttel 1996, 98, Fig. 15-16, 168, Fig. 71. 51

50

The authors of this volume agree that the villa was never a military base. González Cesteros is an exception. He thinks the villa was used by the military from the late Roman period. The excavations have not produced any indication that the area was used by military forces in an organized manner.

BEZECZKY_Text.indd 27

52

18.03.19 14:39

28

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 2.18 Wine presses installation (Photo: Bezeczky 2014).

BEZECZKY_Text.indd 28

18.03.19 14:39

29

3

THE E ARLY CHRISTIAN CHURCHE S OF BRIJUNI



Alexander Schobert

1

THE CHURCH OF ST. MARY

The Church of St. Mary is situated approximately 100 m north of the fortified settlement (the Castrum). It was built on top of a small hill, 40 m from Madona Bay, with a view of Pusti Island to the west. Four separate phases of construction can be identified on the basis of the preserved architecture.1 The three-aisle-basilica (Fig.  3.1), which was probably built at the end of the fifth century A.D., was 23.74 m long and 10.8 m wide.2 The wooden ceiling beams supported a triangular gable without separate mono-pitched roofs for the side naves. In addition to the main gate on the western side of the building, the original structure also had two entrances in the side naves which led directly into the presbytery. In the interior of the building there are two rows of monolithic columns dividing the hall into the main and the side naves. These columns were 2.51 m high and had an impost without capital to reduce the pressure on the arches. The imposts (Fig.  3.2) were decorated with Greek crosses (crux immissa), similarly to Theodoric’s palace in Ravenna. In the side walls between the columns at a height of 3.8 m there were windows, measuring 90 × 140 cm. They illuminated the main nave indirectly from the side. Small double columns decorated with crosses divided the windows, which were vaulted by semicircular arches (bifora). The sanctuary faced east, which had no apse, and was lit by a large window in the back wall.3 Apart from the fact that the Church of St. Mary is somewhat narrower and has only one main entrance at the front, the layout and the position of the columns are identical to the Arian Basilica dello Spirito Santo in Ravenna. The latter was built soon after 493 and represents the first ecclesiastical building by the Goths in Ravenna. Furthermore, this monu-

On the individual phases of construction, see Vitasović 2003, 61-101, Figs. 2-5; the information regarding sizes is drawn from Begović Dvoržak et al. 2007, 232-238. 2 Bratož 1994, 39-47; Chevalier 1999, 105-117; Ujčić 2005, 91. 3 Begović Dvoržak et al. 2007, 234. 1

BEZECZKY_Text.indd 29

ment indicates the influence of the church of Sant’Andrea dei Goti in Rome, built 25 years earlier.4 Most of the monolithic platform (1.4  × 0.76  × 0.2 m) of the early Christian altar has been preserved, lying not far away from the central position where it stood in the sanctuary (Fig. 3.3). The pedestal has, in addition to the recesses for the columns in the corners, a small indent (13.5 × 8 cm) for relics.5 A capital fragment was also found which was used as the base of the altar plate. Gnirs carried out the first excavations in the Church of St. Mary in 1906-07 and was in favor of dating it to an early period based on the shape of the altar platform.6 This hypothesis is confirmed by finds from Ravenna to North Africa, where this altar type was found from the fourth-fifth century onwards.7 The fragments of altar partitions with recessed Greek crosses which were found in the sanctuary also seem to be older than the material of the fifthcentury building.8 The basilica was expanded in the middle of the sixth century. A 5.4 m-wide covered portico-like narthex was added to the western side of the structure.9 An ambo was placed directly in front of the sanctuary and a cistern (1.8 × 2.77 m) was installed behind the building.10 The sanctuary was now separated from the rest of the church by a triumphal arch Bovini 2006, 51-52. Glaser 1997, 244. 6 Gnirs 1911a, 84-86, Fig. 11; this has not been refuted until now, see Vitasović 2003, 77-84; Begović Dvoržak et al. 2007, 234-235. 7 The classical altar table was prevalent in the west, e.g. in the St. Felicianus Basilica/Pula, Sant’Apollinare Nuovo/Ravenna, but also in North Africa (Timgad, Tebessa) and Greece (Nea Anhialos, Tassos). See Braun 1924, 125; Mazzotti 1960, 238-246; Sodini 1984, 17-24, Figs. 14-23. 8 This type was common in the fourth century A.D. At the end of the fifth century A.D. one can see that relief crosses were preferred. Comparable samples have been found in Strobeč close to Salona and Bijaći in the Trogir-region. See Migotti 1992, 120; Begović – Schrunk 2007a, 123 Fig. 97.  9 Church architecture flourished again after the Goths withdrew. Some architectural changes happened in this period, as is evident from the episcopal complex of Salona. See Bratož 1996, 339-340. 10 Begović Dvoržak et al. 2007, 234-235. 4 5

18.03.19 14:39

30

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 3.1 The Church of St. Mary. A - Narthex, B - Entrance hall built in the main nave, C - Main nave of the basilica, D - Sanctuary, E - Court, F - Diaconicum, G - Secretarium/sacristy, H - Vaulted corridor, K - Cistern, L - Socle of the altar, M, N - Place of lecterns, T - Side aisle, U - Block altar of the fifteenth-c. church, V - Scola cantorum, W - Court (Photo: Pülz 2013, Drawing: La Torre 2017 after Gnirs 1911a).

BEZECZKY_Text.indd 30

18.03.19 14:39

CHAPTER 3 – The early Christian churches of Brijuni (Schobert)

Fig. 3.2 Column of the main nave with impost and Greek cross (Photo: Pülz 2013).

31

Fig. 3.3 Monolithic platform of the early Christian altar (Photo: Pülz 2013).

Fig. 3.5 Capital in the sanctuary with crux coronata emblems, Greek crosses and volutes fitted with a narrow trapezoid impost stone (Photo: Pülz 2013). Fig. 3.4 Triumphal arch supported by two monolithic columns with capitals and imposts (Photo: Pülz 2013).

(Fig.  3.4) supported by two monolithic, 4.2 m-high columns with capitals and imposts which were used as additional reinforcement for the additional pilasters. The large triumphal arch was approximately 6 m high. The capitals are decorated on all four sides with christograms, which were executed in the middle of the fifth century as crosses in the center of circular medallions (crux coronata). Thus, a second use is also obvious for the altar table and screen. There is proof of modification in the middle of the sixth century: two columns with bipartite capitals were added to the sanctuary.

BEZECZKY_Text.indd 31

They consist of a capital with crux coronata emblems, Greek crosses and volutes fitted with a narrow trapezoid impost stone (Fig.  3.5). The same type of capital was used in the Basilica of Sant’Apollinare in Classe, which was dedicated in 549. However, just a few years later, it became common practice to carve capitals with imposts out of a single block of stone.11 11

Zollt 1998, 62; Bovini 2006, 95-96; Marušić 1967, 33-35; Begović Dvoržak et al. 2007, 235.

18.03.19 14:39

32

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

Fig. 3.6 Presbytery with septum (Photo: Pülz 2013).

Fig. 3.7 Hospitium (domus presbyterorum), northern annex (Photo: Pülz 2013).

Fig. 3.8 Room with double apses (cellae memoriae), southern annex (Photo: Pülz 2013).

BEZECZKY_Text.indd 32

18.03.19 14:39

CHAPTER 3 – The early Christian churches of Brijuni (Schobert)

33

Fig. 3.9 Sarcophagus, narthex (Photo: Pülz 2013).

The expanded presbytery with septum directly in front of the sanctuary and the construction supporting the superstructure with arcades in the main nave are from the Justinian Era (Fig. 3.6). These features and the lack of a transept, which was common at the time, are characteristic of North African church architecture.12 This is explained by close trade relations with North Africa and troop movements in that region until 534. This process of acculturation reflects the range of the finds in the Byzantine Castrum which, at the turn of the fifth century, only occasionally includes individual items from Italy; the eastern Mediterranean and North African wares, in addition to the amphorae and particularly African red slip ware dominate.13 The annex on the northern side of the building, next to the sanctuary, also dates back to the construction phase around the middle of the sixth century. This annex (13  × 5 m), which was called a hospitium14 (domus presbyterorum), had two floors, a small entrance and narrow windows (Fig. 3.7). There was an atrium in front of the building surrounded by solid walls. On the southern side, in a straight line with the hospitium, a room (11  × 6 m) with double apses was added to the nave (Fig. 3.8). A door connected it to the presbytery and another atrium to the south also belonged to it.

The function of this room is unclear. It may well have been an oratory with alcoves containing relics (cellae memoriae), like similar spaces with two apses in the early Byzantine churches of Cilicia and Isauria or the hall with three apses in the Abū Mīnā Northern Basilica.15 The Church of St. Mary was constantly renovated during the period of the barbarian invasions. This is clearly reflected in the architectural decoration, which is similar to chip carved wood sculpture and to the use of spolia. The continuing decay of the villas near the southern pier is corroborated by the remains of an architrave portraying a hippocampus,16 which had a secondary use as a foundation for an altar screen. It was originally part of the temple of Neptune in Verige Bay. In the course of the third period of reconstruction in the ninth-tenth centuries A.D., the hospitium was turned into a dormitorium with a chapter house.17 By building a southern wing, the Benedictine monks walled up the second entrance to the Church of St. Mary and increased the width of the monastery walls by 0.2 m. The initial function of the narthex was lost and it was repurposed as a grave (Fig. 3.9). The monks entered the complex from the west using the atrium to the north.18 Marcus Samuelis, Commissarius Briorum was Hill 1996, Fig. 20, 34, 50, 59, 60; Jaritz 1982, 145-148. Gnirs 1911a, 92; Schrunk – Begović 2000, 274; similarly spolia from Val Catena were used for the Euphrasius Basilica in Parentium. See Ujčić 2005, 96-98. 17 Although Gnirs 1911a, 90-92 dates the decorative elements of the hospitium to the seventh century, findings from the oldest known Croatian Benedictine monastery in Rižinice near Solin, which was founded by Prince Trpimir in 852, suggest that Mlakar’s dating is more likely to be correct. See Vitasović 2003, 100. 18 Begović – Schrunk 2007a, 136. The coast north of the castrum had been used as a necropolis since the Republican period. See Begović – Schrunk 2007a, 124. 15 16

Three-aisle-basilicas with a raised clerestory were widespread in Africa. On the main characteristics of North African basilicas, see Gnirs 1911a, 89-90; Duval 1974, 353-368. 13 This could be due to newly set customs borders, which led to fine ceramics and foodstuffs from North Africa swamping the western part of the Mediterranean. For an analysis of the finds, see Bloier 2012, 210 Fig. 11. For more on the finds, see Begović – Schrunk 2006, 191-202; see Bonifay – Capelli in this volume. 14 Begović – Schrunk 2007a, 123-124. 12

BEZECZKY_Text.indd 33

18.03.19 14:39

34

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

responsible for the final maintenance work in the fifteenth century. He was buried in the grounds of the church.19 The settlement near Dobrika Bay was finally abandoned after the plague epidemic in 1348.20 2

THE EARLY CHRISTIAN CHURCH WITHIN THE WALLS OF THE CASTRUM AND IN VERIGE BAY

It has not yet been possible to verify the existence and location of an early Christian church within the Castrum walls. Gnirs found some architectural fragments with cross decorations in the Castrum as early as 1908.21 The 1976-1983 excavations directed by Mlakar and Vitasović unearthed further objects from the interior of a church, such as Corinthian capitals, the remains of pilasters, imposts and barrier fragments.22 However, these items were scattered, which makes it impossible to specify the exact location of the church on a map of the settlement.23 The discovery of a lintel with a crux coronata emblem and tendril motif decorations which had been reused in a press did not resolve the question either.24 The Church of St. Mary was renovated in the middle of the sixth century. The stones which were removed were probably used as spolia. A polycandelon was found inside a building (20 × 8 m) reinforced by pilaster-strips on the territory of the Castrum villa.25 It has three holes to accommodate the chain. Although churches were equipped with similar hanging lamps, this does not necessarily mean that this building was in fact a domus ecclesiae. There is also no proof that the southeastern, circle-shaped avant-corps is what remains of a church.26 Regardless of the classification of the material, the fragments of architectural decoration with carved Latin crosses and broadened ends indicate that they were produced at an early stage. Such decorations were fashionable in the Mediterranean region in the 4th and 5th centuries.27 If there was no church in the Castrum, this would suggest Vitasović 2003, 100. Schrunk – Begović 2000, 268-269; Begović Dvoržak 2001, 177-183. 21 Gnirs 1911a, 75-97. 22 Mlakar 1975-76, 10-34; Vitasović 2005, 157-210. 23 See Marušić 1990, 404-413, 405 (sitemap) for an analysis of the finds. 24 Marušić 1990, 406-408, Figs. 3 and 5. 25 The villa rustica from the early Imperial period (62.8 × 51.2 m) was u-shaped with an atrium in the middle. It was open to the bay. See Mlakar 1975-76, 41. 26 Marušić 1990, 405-407. 27 Migotti 1992, 120, 123 Fig. 17 a, b.

that the Church of St. Mary or a precursor was built earlier, before the walls of the fortress were erected. The southern side of Verige (Val Catena) Bay in the eastern part of the island seems to be a particularly convenient location for a church with a baptistery. Gnirs28 was the first to unearth parts of a luxurious maritime terrace villa,29 which formed the imposing center of the latifundium. A temenos with temples for Venus, Neptune and Mars (?) close to the building was erected in the west (see the previous chapter Fig.  2.2, M-N-P and E-F-G-H). The complex also contained a library, a balneum with a portico, an industrial wing with equipment for oil production, piscinae for fish production and a pier for docking to the northeast.30 The thermae of the port were probably repurposed as an early Christian basilica during the reign of Theodosius, when Christianity became the religion of the empire. The tepidarium with an apse could have been rededicated as a nave. The 1.2 m-deep pool lined with white marble in the former frigidarium seems to be used as a baptismal font.31 In 1903, Gnirs found capitals decorated with dolphins in this room; they were originally part of the temple of Neptune. Early Christian graves were found outside in the vicinity of the apse, which may confirm that the building was repurposed as a church.32 The transformation of the port thermae and the continued use of the olive oil presses illustrate the interaction between the two bays. However, the hub of the island ultimately shifted from the villa complex in Val Catena to the western side of the island by the beginning of the fifth century at the latest.33 3

THE CHURCH OF ST. PETER

An observation post with a Castrum and a church was constructed on the Petrovac hill (33 m above sea level), approximately 170 m northeast of the Church of St. Mary, in an area which is now densely wooded.34 The advantage of this location was that it enabled observers to get early warning of enemy ships and thus warn the port settlement in time.

19 20

BEZECZKY_Text.indd 34

Gnirs 1904, 145-146. On the villa maritima in Val Catena: Lafon 2001, 459; Schrunk – Begović 2000, 257-261; see La Torre – Bezeczky in this volume. 30 In 1996-97 a Bavarian/Croatian team carried out underwater research on the northern pier. On trade goods and their analysis, see Bloier 2012, 197-205. 31 Begović – Schrunk 2007a, 113; Džin 2011, 97. 32 Gnirs 1915, 137, 141. 33 On the statistics of the finds see Bloier 2012, 196, 210, 227. 34 Begović Dvoržak 2001, 183-187 for a detailed description of the structure. 28 29

18.03.19 14:39

CHAPTER 3 – The early Christian churches of Brijuni (Schobert)

35

Fig. 3.10 The Church of St. Peter, main entrance (Photo: Bezeczky 2013; Drawing: La Torre 2017, after Begović – Schrunk 2007a).

Parts of the excavated defensive structure form a half-circle complemented by a small buttressed fortress. The fortification and the rest of the structure form a single architectural unit. A rectangular building (4.6 × 5.5 m) is visible directly in front of the church; its walls are up to three-quarters of a meter high. The entrance to the tower, which presumably served as an observation deck, was on the southern side.35 The hall-church has a rectangular layout (5.82 × 7.78 m). Its front area slopes to the south. The preserved bottom part of the main entrance (2.05 × 2.03 m) together with the surrounding stonewalls makes a precise reconstruction possible (Fig.  3.10).36 A polygonal apse formed the outer end of the Church of St. Peter. On the inside, it was horseshoe-shaped with a 2.9 m diameter.37 It contributed to the structural balance of the gabled roof to the northeast together with the triumphal arch in the center of the spatial axis and the halfmeter-wide pilaster-strips of the outer walls.38 This type of construction reflects Ravenna architecture and proves that the church was built during the same period as the Claustra Alpium Iuliarum.39 The Church of St. Nicholas in Pula from the fifth/sixth century has a similar structure and dimensions.40

Begović Dvoržak 2001, 186. Gnirs 1911c, 41-44. 37 Unlike the semicircular apse, which was already used in the secular architecture of the Imperial period, polygonal apses make their first appearance in the beginning of the fifth century in urban centers. See Migotti 1992, 118-119. 38 Mohorovičić 1957, 491-493. 39 The “Adriatic limes” was composed of an unbroken line of walls and guard posts and served primarily to protect transportation routes, specifically to protect trade routes in the time of Justinian. See Šašel – Petru 1971. 40 Rudolf Egger had already pointed out that this type of church had been established relatively early, especially in the Alps-Adriatic region. See Egger 1916, 122; Chevalier 1999, 108. 35 36

BEZECZKY_Text.indd 35

The Church of St. Peter had another entrance to the north, which gave the clergy direct access to the presbytery. The sanctuary is slightly raised and separated from the hall by six altar partitions. The traverse under the floor, which had clearly recognizable hollows to hold the septum, was preserved along its entire length. Altar screen remains used as spolia can also be found in situ. The limestone slabs (75 × 85  × 7  cm) are decorated with Greek crosses surrounded by zig-zagging lines and an interlaced pattern. Based on the Carolingian style of the altar screens, renovations took place in the seventh/eighth century. Gnirs found some items with relief ornaments from the early Christian period, covered by the fragments of a triumphal arch, during his inspection of the area outside the church in 1910.41 There is a partially preserved polychromatic floor mosaic in the church. It is a 0.35 m-wide strip with red, white and blue tesserae laid out in a rectangular pattern. Intersecting circles containing Greek crosses are connected by wavy lines.42 Water was provided using a trapezoid cistern (4.5 × 2.6 m) on the northern side of the building. Walls were discovered a few steps away which served to further separate the temenos. This suggests that there was a civilian settlement nearby, as the church complex would not have needed more fortifications otherwise. The large number of tegulae fragments and clearly identifiable, unbroken walls on the northern side of the hill may corroborate this hypothesis.43 Gnirs 1911c, 42; Begović Dvoržak 2001, 184. Begović Dvoržak 2001, 185. A similar polychrome mosaic was laid out at the beginning of the fifth century in the ambulatory between the apse and the synthronon in the Episcopal Basilica at Zadar. See Vežić 2005, 38-41. On an overview for mosaics in Adriatic church centres, see Campanati 2008, 435-454. 43 This hypothesis is based on Marušić’s inspection of the site. He found the scattered remains of many roof tiles and wall fragments there. See Begović Dvoržak 2001, 186 and see Berni Millet in this volume. 41 42

18.03.19 14:39

36 4

SHORT HISTORY OF THE WORKSHOP AND THE VILLAS

THE BISHOP OF CISSA PULLARIA (CESSUS) AND HIS RESIDENCE

The island of Brijuni was called Cissa Pullaria in Antiquity, as one sees in the writings of Pliny the Elder44 and the Peutinger Table (Fig. 3.11).45 The island had a special administrative status going back to the Flavian period, and it was under direct Imperial control.46 This is corroborated by the inheritance of the landowners Laecanii. They owned the latifundia on Brijuni Island. When the last Laecanius died without an heir, all of his property came into Vespasian’s ownership.47 After the middle of the second century A.D., military pressure on Rome increased constantly. To understand the whole situation in respect of social and economic changes, the archeological finds from Brijuni had to be compared with material from Pula, but a complete publication about these amphorae is still pending.48 Geopolitical and cultural changes took place in the course of the fourth-fifth centuries A.D. following the battle of Adrianople in 378. There were changes in the power relations in the Adriatic region following the Huns’ destruction of Aquileia (452) and the end of Western Roman Empire in 476.49 The first defensive wall was constructed around the civilian settlement which sprang up around the villa rustica in Dobrika Bay in the fourth century. In the time when Brijuni was ruled by the Ostrogoths, Cassiodorus, senator and praetorian prefect of the province of Istria (533-537), mentioned the island in a letter. He noted that it was an important source of supplies for the capital Ravenna, comparable to the latifundia of Campania.50 The fortification in Dobrika Bay (Castrum) was probably constructed when Iustinian I began to reconquer Istria at the latest. The walls were 2.6 m wide. They were equipped with balustrades and battlements. The walls were approximately 5 m high, and large cut stone blocks were placed at the ends to add stability. The structure had five gates with little towers on top for extra protection on the side facing away from

Fig. 3.11 Peutinger Table pars V.

the sea. It was rectangular in shape, and covered one hectare along the coast.51 The name of the island was altered along with its legal status, and it is referred to as Census (Cessus) in records from Late Antiquity.52 Brijuni formed the important northern endpoint of the defensive system which followed the maritime route to Ravenna along the eastern Adriatic coast. This was significant, as these maritime routes were extremely dangerous in the sixth century. The island was also home to a military outpost which protected the episcopal centers of Pola and Parentium/Poreč.53 The secular power of the Church increased during the reign of Justinian I, and Brijuni must have had a bishop in this period. This is supported by a deed of gift dating from 534 A.D. in which Bishop Euphrasius of Parentium donated one-third of the salt production from Val Laura (in the southern part of the island) to the church of Brijuni.54 Documents from the Synod of Grado in 579 contain the first written record of an “Episcopus Cessensis”, referred to there as Vindemius.55 Vindemius was inclined towards the Lombards in the Three-Chapter Controversy. The Lombards followed Arianism, which meant that they doubted the consubstantiality of the Father and the Son, and found the discussions regard-

Gnirs 1911a, 94-95; Begović Dvoržak et al. 2007, 232; Matijašić 2012, 193-208. 52 Suić 1987, 201-213; Vicelja Matijašić, 2007, 48. 53 Many sixth century fortifications have been preserved in Croatia and Slovenia, which formed part of the Justinianian defense system (Claustra Alpium Iuliarum). They were erected at strategically important locations inland as well as on islands along the Adriatic coast. See Poulter 2013, 63-76; Ciglenečki 2009, 205-222; on the castrum see Mlakar 1975-76, 10-34; Begović – Schrunk 2009, 223-236; Begović – Schrunk 2011a, 375-390. 54 Schrunk – Begović 2000, 268. 55 Begović – Schrunk 2011b, 658. 51

Pliny NH 3, 151, 152. Križman 1979, 314-316, 321. 46 Brijuni only came under the control of the administration in Pula in Late Antiquity. See Suić 1987, 201-213. 47 Bezeczky 1998a, 15 and 68; Bezeczky 2014 and see also his article in Chapter 1. 48 On the eastern Mediterranean amphorae, see González Cesteros in this volume. 49 Matijašić 2012, 150; Bratož 1996, 331-351. 50 Cassiod. var. 12, 22; see Begović – Schrunk 2007a, 103-104. 44 45

BEZECZKY_Text.indd 36

18.03.19 14:39

CHAPTER 3 – The early Christian churches of Brijuni (Schobert)

ing the relationship between the divine and human nature of Christ absurd.56 The bishop was arrested in 588 in Ravenna.57 During his year-long incarceration, he had enough time to reconsider his opinion on Monophysitism. As a participant at the Synod of Marano in 590, Vindemius had already adopted the Catholic faith.58 There are records of another Bishop Cenetensis in the person of Ursinus, who participated in the sixth ecumenical council in 680 in Rome.59 This confirms that an episcopal center existed in Brijuni from the middle of the sixth to the end of the seventh century. The bishop’s seat was moved to Ruginium/Rovinj around 800.60 Before an independent diocese was established in Brijuni, the visiting bishop could have used the church in Verige Bay for baptisms (see chapter 3, Fig. 2.2, F). The villa maritima was an obvious choice for the episcopal residence. At the turn of the fourth-fifth centuries, when the settlement in Dobrika Bay became the main settlement of the island, either a little church with a residence attached was built in castellis or already the Church of St. Peter.61 It may have been the original location of the altar of the Church of St. Mary, as it is older than the preserved construction materials. The

Bratož 2000, 640-641. In his Historia Langobardorum (3, 26), Paulus Diaconus claims that the Patriarch had the schismatic Videmius ex Istria arrested. See Bratož 2000, 64. Suić 1987, 207 clings to his theory that Videmius was arrested in the bishop’s residence in Brijuni. 58 Marušić 1990, 403.

37

construction of the Church of St. Mary commenced when the Episcopus Cessensis was appointed, and as the majority of the parish had already received the sacrament of baptism, there was no need to set up a baptistery. For newly converted Christians, children or migrants, baptism by aspersion was a possible alternative. After the hospitium was completed, the bishop moved to the annex of the Church of St. Mary or used the terrace house on the Petrovac hill, which Gnirs described as a “monumental building with Ionic capitals” in 1911.62 The area has not been examined again since. However, the terrace house is the most probable location for the episcopal residence. On the one hand, it is located in between the churches of St. Peter and St. Mary, and on the other, the bishop would have been able to find safety quickly in the fortress in the event of an attack. The Castra provided shelter for the population, but they were not a military fortification. This is also demonstrated by the location of the two churches outside of the walls. The evidence presented here indicates that there was undoubtedly a bishop’s see in Brijuni which existed for more than two and a half centuries.

56 57

BEZECZKY_Text.indd 37

Schrunk  – Begović 2011a, 680; for more on the synod see Bratož 1998, 589-602. 60 Bratož 2000, 643-644. 61 See Begović – Schrunk 2010, 263 for a differing view. 62 Gnirs 1911a, 93. 59

18.03.19 14:39

BEZECZKY_Text.indd 38

18.03.19 14:39

THE A MPHOR AE

BEZECZKY_Text.indd 39

18.03.19 14:39

BEZECZKY_Text.indd 40

18.03.19 14:39

41

4

ITALIAN AND ISTRIAN A MPHOR AE



Tamás Bezeczky

Amphorae produced in Italy and Istria make up 27 % of all amphorae excavated at the Castrum villa. We can differentiate between two types, those made locally and those transported to the villa from other parts of Italy. To review Istria’s economic history, one must examine many Adriatic amphora types: Lamboglia 2, Dressel 6A, Dressel 2-4, Dressel 6B, Fažana 1, Fažana 2, Loron small amphora, Porto Recanati / amphora con collo ad imbuto, Schörgendorfer 558, flat based / Forlimpopoli, north Adriatic fish sauces amphorae, Aquincum 78, amphora from Grado and Bónis 31/5. A significant number of these amphora types were also found at the Castrum villa. Many studies examine where they were produced and what they contained, but current answers to these questions are not always convincing. Clear typological and petrographical descriptions are needed. The amphorae published from northern Italy do not always match the finds from Noricum and Pannonia. The recently published amphorae with stamps and painted inscriptions will be very helpful in determining their origins.1

great deal of research on this amphora since the publication of Nino Lamboglia’s article.2 Based on the published data, these vessels were produced in a number of places from Aquileia to Brindisi.3 Central Dalmatia was also recently mentioned as a production area, thus starting a lively discussion.4 It is necessary to examine, as soon as possible, the Adriatic amphora types mentioned in the introduction, the amphorae produced in known or presumed workshops, and geological samples from the surrounding area.5 The production of Lamboglia 2 amphorae began between the end of the second century B.C. and the emergence of the Dressel 6A amphorae in the final third of the first century B.C.6 Although it is impossible to specify when exactly Dressel 6A production began, it is likely to have been in the last third of the first century B.C.7 There are two stamps on the Castrum amphorae. The BAR·ME(…) stamp (2) has not been found elsewhere. The other Castrum stamp (3), VERSO, is widely

Lamboglia 1955, 262. Panella 1970, 117; Will 1970, 383-386; Tchernia 1986, 68-74; Empereur – Hesnard 1987, 33; Cipriano – Carre 1989, 85; Cipriano 1994, 207 Fig. 3; Sciallano – Sibella 1994; Bruno 1995, 83-92; Zupančić et al. 1998, 345-357; Monsieur et al. 2006, 310-313; recently Rizzo 2014, 120-123; Carre et al. 2014, 417-428 with complete bibliography. 4 Cambi 1989, 321; Lindhagen 2009; Lindhagen 2013; Kirigin et al. 2006, 193-194; Panella 2011, 96-97; Carre et al. 2014, 417428. 5 It would be helpful to use the petrographical analysis of the Lamboglia  2 amphorae found in Ephesus for further research. Bezeczky 2013a, 114-120, see Roman Sauer’s results: as previously known, Lamboglia 2 amphorae were produced in the workshops excavated near Brindisi (Apani and Giancola). This was proved using the petrological analysis of the amphorae found in Ephesus; fabrics P and P1 are characteristic of the area around Brindisi. Another fabric group (O) indicates a central Italian origin. Finally, the marine clay used for fabric group P2 is also similar to the various types of amphorae made in Italy. See also Sauer 2013, 120-124. 6 Tchernia 1986, 53-56; Grace 1965, 11; Baldacci 1972, 109; Van der Werff 1986, 103; Peacock  – Williams 1986, 100; Carre 1985, 211. 7 Carre 1985, 211; Van der Werff 1986, 103; Peacock  – Williams 1986, 100; Tchernia 1986, 134. 2 3

1

ITALIAN WINE AMPHORAE FROM THE LATE REPUBLICAN AND EARLY ROMAN PERIOD

LAMBOGLIA 2 The earliest dated amphorae found at the Castrum villa belong to the Lamboglia 2 type. Lamboglia 2, one of the best-known amphorae from the Adriatic coast, was widely used across the Mediterranean region. It is a robust vessel with thick walls. In fragmented form it is difficult to distinguish it from the type of amphora that later came to replace it, the Dressel 6A.  The two amphora types are generally made of the same fabric, as the Dressel 6A amphorae were made at certain Lamboglia 2 workshops. There has been a

1

The papers of the volume eds. Pesavento Mattioli  – Carre 2009.

BEZECZKY_Text.indd 41

18.03.19 14:39

42

THE AMPHORAE

known.8 The content of these amphorae was wine.9 A significant quantity of Lamboglia 2 amphorae was found at the excavation near Pula harbour.10 This is also one of the bestknown amphora types in Aquileia and in the Trieste area.11 The pieces transported from Aquileia are primarily found in Magdalensberg (Noricum), but they have been found occasionally in the Drava-Sava valley, conquered by Augustus.12 The form is common in the western Mediterranean, present in Greek regions and in North Africa.13 DRESSEL 6A Dressel 6A amphorae differ only slightly from the Lamboglia 2. The design of the handles is somewhat different, and the stubs of the Dressel 6A amphorae are pointed.14 The production of these amphorae began in the final third of the first century B.C.  and continued until the middle of the first century A.D. This is corroborated by the pieces found in the Carthage amphora wall, as well as the painted amphora inscriptions published from Rome and Magdalensberg which include the dates of consuls.15 Amphorae from the southern and western areas of Pannonia are a form of indirect proof that has barely been taken into account. The Dressel 6A amphorae are found in the military camps and settlements along the stretch of the Amber Road conquered by Tiberius (Siscia, Sirmium, Gomolava, Emona, Salla,

Desy 1989; Bruno 1995; Carre et al. 1995; Gabucci – Quiri 2008, 65; Brecciaroli Taborelli 1987, Pl. XLI 28.3. 9 Formenti et al. 1978, 95-100; Tchernia 1986, 53; Empereur – Hesnard 1987, 33. 10 Starac 2008, 123-168. 11 Carre – Cipriano 1985, 6-23; Cipriano – Carre 1989, 97-98; Verzár-Bas 1991, 182-194; Verzár-Bas 1994; Horvat – Bavdek 2009, 84-93; Auriemma et al. 2008, 171-173. 12 Maier-Maidl 1992, 95-103; Bezeczky 1994, 15; Bezeczky 1995, 156. 13 Tchernia 1986, 53-56, 69-74; Beltrán, 1970, 349-358; Grace, 1979, Fig.  36; Will 1987, 202-204; Will 1989b, 302-305, Figs. 7-11; Cipriano  – Carre 1989b, Figs. 14, 97-99; MartinKilcher 1993, 277-279; Martin-Kilcher 2005, 208, 212; Bruno 1995, 83-275; Župančić et al. 1998, 345-357; Ehmig 2003, 46; Lindhagen 2009, 95, Fig.  4; Maier-Maidl 1992,  95-103; Lund 2000a, 84; Monsieur 2001, 75, Fig. 12; Bezeczky 2005a, Lamboglia 2; Bezeczky 2013a, 114-120; Vidrih Perko 2006, 211; Auriemma 2006, 168-169; Mazzocchin 2013, 66-67; Rizzo 2014, 122-123; Carre et al. 2014, 417-428. 14 Baldacci 1967-68, 7-49; Buchi 1973, 531-637; Carre 1985, 211; Böttger 1992, Fig. 1. nos. 3-5, T. 96, no. 3 T. 97; MartinKilcher 1993, 277-279. 15 Delattre 1894, 89-110 = CIL, VIII.  3, 2200, dated between 43 and 15 B.C.; CIL, XV.  4582, A.D.  36; Carre 1985, 211-213; Piccottini 1997, 204.

Savaria, Devin).16 No Dressel 6A amphorae have been found in eastern Pannonia, conquered in the late Claudian period. We know of only one stamp in the Castrum villa, consisting of the letter V stamped on a rim (5). It has not been found elsewhere to date. A post cocturam graffito of the number XXIIII can be found on the rim of another amphora (7). The questions regarding the workshops where this amphora type was produced are similar to those regarding the Lamboglia 2 amphorae.17 It is important to note that these amphorae were not produced on the Istrian peninsula, even though some recent publications consider them to be Istrian.18 The inscriptions on the Dressel 6A amphorae discovered in Rome and published by Dressel mention wine,19 although the inscriptions of some amphorae from Salzburg and Italian sites refer to fish sauce.20 However, in all probability, these amphorae were generally used to store wine. They are fairly common in Aquileia, Noricum (Magdalensberg) and western Pannonia.21 They also reached Italy and the western Mediterranean, North Africa, Greece, the eastern Mediterranean, the Red Sea and India in fewer numbers.22 2

OLIVE OIL AMPHORAE

The olive oil produced in Istria is well-known. It was praised by ancient sources, which considered it equal in quality to olive oil from Baetica. Olive oil was also produced in northern Italy, and transported using Dressel 6B amphorae.23

8

BEZECZKY_Text.indd 42

There are a few pieces in the Claudian layers in Ostia, but they are missing from the Flavian layers: Carre 1985, 213; Bezeczky 1994, 130; Bezeczky 2013, 120. 17 See note above nos. 1 and 2. 18 Zevi 1967; Buchi 1973, 547; Will 1996, 208. It seems to be a misunderstanding to suppose that the Istrian amphorae can be found in Athens: Will 1997, 124. This is based on improper identification of the two types (Dressel 6A and 6B). 19 CIL, XV. 4582. 4583. 4653. 4852; Zevi 1966, 217. 20 Heger 1986, 132-135, Fig. 1; Carre et al. 2009, 216-217. 21 Carre 1985, 209-218; Cipriano – Carre 1989, 85-88; VerzárBas 1991, 195-196; Verzár-Bas 1994; Maier-Maidl 1992, 80-95; Bezeczky 1994, 22-34; Bezeczky 1995, 156. 22 Baldacci 1967-68, 27-29; Riley 1979, 151-156; Carre 1985, 211213; Tchernia 1986, 129-133; Cipriano – Carre 1989, 85-88, Figs. 17, 100; Maier-Maidl 1992, 86; Pesavento Mattioli – Zanini 1993, 47, Figs. 49-51; Martin-Kilcher 1993, 277-279; Bezeczky 1995, 156; Bezeczky 1998b, 228-230; Bezeczky 2013, 120-122; Bruno 1997, 517; Martin-Kilcher 2005, 208 and 212; Lund 2000a, 84; Rizzo 2003, 149; Lindhagen 2009, 96; Ehmig 2003, 47; Tomber 2005, 230; Vidrih Perko 2006, 211; Auriemma 2006, 168-169; Auriemma et al. 2008, 171-173; Carre et al. 2014, 417-428; Rizzo 2014, 123. 23 Some painted inscriptions in Magdalensberg mention Istrian oil, Piccottini 2000-01, 378; most recently, Carre et al. 2009, 217 16

18.03.19 14:39

CHAPTER 4 – Italian and Istrian amphorae (Bezeczky)

DRESSEL 6B (1A PHASE), ANTE DRESSEL 6B Few amphorae of the earliest Dressel 6B (1a phase) type have been found at the Castrum villa. These are primarily rim and base fragments. This is a transitional type, which is similar to both the Brindisi / Adriatic ovoid type and the classic Dressel 6B.  These amphorae were termed ante Dressel 6B by Alessandra Toniolo.24 Marie-Brigitte Carre and Stefania Pesavento Mattioli later provided a more precise description of the different variants of the Dressel 6B form, distinguishing between production in Cisalpine Gaul and Istria.25 They used late Republican stamps of people who held office at the end of the Republican period to date the ante Dressel 6B type.26 The distribution of the ante 6B outside northern Italy was limited, and even in the northern region they have only been found at Magdalensberg.27 The Laecanii workshop in Fažana also produced Dressel 6B - ante 6B (phase 1A) amphorae. Such an amphora, bearing the stamp C·LAECANI·P·F, was found in the Venice lagoon. This stamp matches a tegula stamp from the Castrum villa.28 The ante 6B type amphorae are important proof of the beginning of production in Fažana, invalidating all previous opinions, which dated the first amphorae from the Laecanii workshop to the Tiberian period.29 Petrological analyses suggest that the ante Dressel 6B amphorae (9-10) found in the Castrum villa were not produced in the workshop at Fažana or, if they were produced there, must have been the workshop’s earliest products. Different raw materials were used to produce these amphorae and the amphorae bearing stamps from Italy (AP·PVLCRI, SEPVLLI P·F, P·Q·SCAPVLAE). The analysis of the Castrum (9) amphora reveals that the composition of the particles in its material is somewhat similar to that of the ampho-

218 mentioned that the Dressel 6B amphorae occasionally contained other things, based on the painted inscriptions. 24 Toniolo 1991, 21-23. 25 Carre – Pesavento Mattioli 2003, 453-476. 26 Appius Claudius Pulcher was consul in 38 B.C.; one of the members of the Sepullius family, P. Sepullius Macer was quattuorvir monetalis in 44 B.C.  The Sepullius families are known from inscriptions in Padova; Buchi 1973, 585-588; Carre  – Cipriano 1985, 12; Zaccaria 1989, 481; Tassaux 2001, 502-503; Toniolo 1991, 168; Cipriano  – Mazzocchin 2000, 169-184; Cipriano  – Mazzocchin 2002, 307-311. 27 Bezeczky 1998b, 235. 28 Toniolo 2011, 189-192; Matijasić 1987, 164, no. 25; Matijasić 1998a, 253. Unfortunately, we were unable to carry out petrological analysis, as we did not have samples of either the amphora or the tegula. 29 Martin-Kilcher 2000, 508.

BEZECZKY_Text.indd 43

43

rae from northern Italy, but is also similar to the amphorae produced in the workshop at Fažana. DRESSEL 6B (2A–3A PHASE), CLASSICAL DRESSEL 6B The Dressel 6B (2a–3a phase) amphorae from Fažana are one of the best-known products of the Istrian peninsula.30 The workshop clearly produced amphorae in large quantities, with many potters collaborating, which lead to differences between the pieces belonging to individual batches. We know the rims best, as these were preserved because of the stamps. The few entire amphorae which have been published reveal the differences quite clearly: there are differences even between items that bear the same stamps. Few Laecanius and Imperial amphorae have been found at the Castrum villa. Regarding the Laecanius amphorae, the LAEK·A (11), [C·LAEK] / SYNT (12), [LAEK] / COMVS (13), LAEK (14), stamps belong to the Tiberian– early Claudian (Magdalensberg II, A.D. 14-45/50) group.31 Among the stamps which belong to the Claudian–Flavian period (Post Magdalensberg I, A.D. 45/50-78), CL[AEK] (15) matches the stamps found in Celeia and Poetovio.32 Only the letters E and K written in ligature have survived on one stamp, which can be grouped with several different stamps [CLA]EK (16). The PA[GANI] stamp (17) can be dated to post Magdalensberg II (A.D. 45/50-78) or the Vespasian (A.D. 78-79) period. The IM[P] stamp (18) also belongs to this group.33 An amphora with a particularly difficult to decipher stamp, LESBI (22), was produced in the Domitian period (A.D. 81-96).34 Three amphora rims with the IMP·TRA stamp were also found, and can be dated to Trajan’s reign (A.D.  98-117). The stamp of the workshop manager appears next to the IMP·TRA stamp (19) on one of the rims. Unfortunately, it is hard to decipher, as only the last two letters of the name or its abbreviation are visible … VS.35 The final Dressel 6B amphora was produced during the Carre – Pesavento Mattioli 2003, 462. Bezeczky 1998a, nos. 1, 507 and 611; Vidrih Perko – Pavletić 2000, 265. 32 Bezeczky 1998a, nos. 37-38. 33 Bezeczky 1998a, no. 648; Vidrih Perko – Pavletić 2000, 265. 34 Siscia inv. no. R 7935; Concordia, Cipriano 2008, 304-305, Fig. 1, [IMP]·DOM / LESBI. 35 There is one amphora in Vindobona with two stamps: IMP(eratoris) [Nervae] TRA(iani) (upside down, M-P letters in ligature) and SERV(...) with the S written backwards. The abbreviation SERV(...) can be interpreted in various ways: Servacus, …. Serus, or simply Serv(us), Bezeczky 2005b, 48. The Castrum workshop manager’s stamp is different from the previous one. 30 31

18.03.19 14:39

44

THE AMPHORAE

reign of Hadrian. The amphora finds prove that the villa was in use until Hadrian’s reign. During this period, the Istrian producers were unable to export enough olive oil to their traditional markets, perhaps because of bad weather – some of the trees probably froze – or some other unknown reason.36 From then on, the provinces of Raetia, Noricum and Pannonia imported olive oil from Baetica, transported in Dressel 20 amphorae. A stamp from a Dressel 20 amphora (40) has been found in Castrum villa. It was probably on an amphora used to transport olive oil from Baetica, to make up for the lack of local olive oil.37 FAŽANA 1, ISTRIAN DRESSEL 6B (4A PHASE) As already mentioned, the workshop in Fažana started producing Fažana 1 (Dressel 6B 4a phase) amphorae bearing the stamp of M. Aurelius Iustus in the last third of the second century A.D. With the appearance of this amphora, it is as though the form of the Dressel 6B had been forgotten. It had a short, straight rim, reinforced at the neck (26). The classic 6B form had a stub, while this form had a pointed spike. The whole amphora is rougher. While these amphorae initially appeared to be the latest variation on the classic 6B form, it now seems that they belong to a type of their own, which archaeologists call type Fažana 1. The stamp of M.  Aurelius Iustus is very different from the stamps of the Laecanian and Imperial amphorae. This is a stamp in hollow lettering, in the genitive case. Only one similar stamp was found during the new Fažana excavation: F♠A with a pattern of leaves.38 The quantity of the Fažana 1 amphorae found in the Castrum villa is one-tenth that of the Dressel 6B amphorae. However, they were produced for a much shorter period of time, and to date have only been found on the Istrian peninsula. M. Aurelius Iustus probably attempted to reconquer the lost markets.39 He was unsuccessful, and his amphorae can only be found in Istria.40

3

ISTRIAN FISH SAUCES AMPHORAE

FAŽANA 2, ISTRIAN DRESSEL 6B (4A PHASE) A small amphora which differs from all the other types produced in the workshop was given the name Fažana 2 (Dressel 6B 4a phase). The rim is chalice-shaped and the handle is round in section (27). Gnirs described these amphorae after excavating the workshop, and some fragments can be seen in the Brijuni Museum.41 The rescue excavation of 19901991 unearthed rim and handle fragments which belonged to these small amphorae.42 For the moment, there is not enough information to date the Fažana 2 amphorae, and we must wait for the detailed publication of the excavations. The clay used to produce Fažana 1 and 2 amphorae is similar to that used for the Laecanius and Imperial amphorae.43 This form is probably connected to the Bónis 31/5 ampho44 ra. Éva Bónis found this 32 cm-high amphora in the Emona cemetery (in grave no. 934, Grabfeld Wienerstrasse). She dated the finds and the amphora to the first-second century A.D. Sonja Petru later published the necropolis of Emona and, taking the rest of the finds into consideration, dated the amphora to the second century A.D.45 The amphora found in Iseum in Savaria can also be dated to the second century.46 The other small amphora published by Bónis is also from Emona, but from a different cemetery (Grabfeld Lenarčič). Based on the photograph, this amphora is most similar to the amphorae produced in the Loron workshop.47 Found in Pannonia, in the collapsed cellar of a villa in Baláca, it also belongs to this group. The demolished layer was dated to the end of the second century or the beginning of the third century.48 This suggests that the workshops in both Fažana and Loron started producing such amphorae around the middle or end of the second century.49

Gnirs 1910a, 81, no. 4; Bezeczky 1998a, 9, Fig. 8. Bezeczky – Pavletić 1996, 157, Pl. 3, nos. 9-10. 43 Bezeczky – Mange 2009, 253. 44 Bónis 1942, 232, T 31, no. 5, inv. no. 6397, grave no. 934. 45 Petru 1972, 88. 46 Bezeczky 1987, 78, no. 320, Fig.  30. Based on the thin section, Szakmány and Józsa are of the opinion that all of the inclusions are similar to the Fažana amphorae but opalic spiculae are missing. 47 Bónis 1942, 232, T 33, no. 7, inv. no. 6764, grave no. 70; Marion – Starac 2001, 117; Marion 2009, 284. 48 Palágyi 2007, 287. 49 It is possible to determine where the individual pieces were produced based on the petrological analyses. For new evidence of other workshops see Vidrih Perko – Župančić 2011, 154. 41 42

Bezeczky 1998a, 10. See in the catalogue no. 40 and also see González Cesteros – Berni Millet in this volume. 38 Paić – Bulić 2008, 28, no. 34. 39 Bezeczky 1998a, 16; Tassaux 1982, 254 he dated to the Laecanius period; later he (Tassaux 1998, 93) dated it to the end of the second century. 40 Vidrih Perko – Župančić 2011, 154. 36 37

BEZECZKY_Text.indd 44

18.03.19 14:39

CHAPTER 4 – Italian and Istrian amphorae (Bezeczky)

4

NORTHERN ADRIATIC AMPHORAE

AQUINCUM 78 There are rim fragments of the Aquincum 78 type amphora in the Castrum (32-33). One such amphora was found in 1978 in Aquincum (Pannonia).50 The formal characteristics of the amphora resemble the Dressel 6B type. The body is oval-shaped ending in a short stub. The thin, small handles’ cross-section is circular, buff in colour (10 YR 8/3). The dimensions of the Aquincum amphora are the following: diameter of the rim 10 cm, height 77 cm. Chronologically it was used between the end of the first to the end of the second century A.D. It is widely distributed throughout Pannonia (Poetovio, Salla, Arrabona, Mursella, Tokod, Aquincum, Mursa). Its place of production is currently unknown. 5

MISCELLANEOUS ITALIAN AMPHORAE

PORTO RECANATI / ANFORA CON COLLO AD IMBUTO A few Porto Recanati amphora fragments were found in the Castrum villa (30-31). The rim and the neck of this Porto Recanati type are funnel-shaped.51 Some of them have a single line on the rim, others have a horizontal groove. The finds in Italy make it possible to identify two sub-types.52 Few stamps are known on this type of amphora.53 It is certain that they were not produced in the Fažana workshop. Petrological analyses showed that there are particles of volcanic origin (clinopyroxene – augite and diopside) in the fabric of the Porto Recanati amphorae.54 Istrian amphorae rarely contain such volcanic inclusions, and only in trace amounts. The

Bezeczky 1997, 178. Mercando 1974, 174; Mercando 1979, 180; Carre 1985, 232235; Bezeczky 1987, 35-36. 52 Mazzocchin 2009, 193-195. 53 Bezeczky 2005b, 61; Mazzocchin 2009, 206-207. The stamped amphorae from Athens and Corinth do not belong to this type. The Dressel 6A amphorae bear the IVLI (palm) / PAVLINI stamps. 54 Similar volcanic inclusions are found in the Schörgendorfer 558 amphorae too. Though some archeologists thought that the form was produced on the Istrian peninsula, on the basis of the petrological analysis, the site of production should be located in northern Italy. Bezeczky 2005b, 56-57 and 63; Sauer 2005, 120-121 and 123124; perhaps in the region of Lake Garda. Pesavento Mattioli 2011 does not agree with this. Their precise origin is unknown but, based on their distribution, they may have been produced in central Italy. Research into these amphorae must continue. There are no Schörgendorfer 558 amphora in the Castrum villa but Vidrih Perko – Pavletić 2000, 266 suggested it. 50 51

BEZECZKY_Text.indd 45

45

Porto Recanati amphorae were clearly produced in northern, or perhaps central, Italy. What was transported in these amphorae? Olive oil can probably be ruled out, as it was readily available in Brijuni in the first and second centuries. Porto Recanati amphorae are widely known in Cisalpine Gaul and the northern provinces (Noricum and Pannonia).55 FORLIMPOPOLI / FLAT-BOTTOMED AMPHORA There are Forlimpopoli / flat-bottomed amphora fragments from Castrum villa. This means that local wine production cannot be ruled out at the villa. It is unclear whether the wine presses are from the second century or a later period.56 If there was a significant quantity of wine produced, as the number of wine presses suggests, then the wine could have been exported. This would resolve various questions concerning wine production in Brijuni.57 The petrological analyses indicate that most of the Forlimpopoli amphorae came to the Castrum villa from Italy. The analyses of a few pieces revealed that the Fažana workshop also produced flatbottomed amphorae, although the entire form of the vessel is unknown. Their fabric is similar to that used for the Dressel 6B.58 Workshops have been found in the  Forlimpopoli and Rimini region.59 The most recent discovery was in Crikvenica, on the eastern coast of the Adriatic.60 They were used from the first century A.D. to the end of the third century.61 The content of the amphora was wine. It is widely distributed in the Adriatic region and throughout the Mediterranean.62

Toniolo 1991, 29-31; Mazzocchin 2009, 206-207; Bezeczky 2005b, 61-63. 56 Matijašić 2008, 289-300. 57 Tassaux 1992, 252; Brun 2004, 53-54; Vitasović 2005, 209; Matijašić 2008, 289-300; Begović  – Schrunk 2007a, 42 = Begović  – Schrunk 2007b, 329-330 = Begović  – Schrunk 2010, 256 = Begović – Schrunk 2011a, 378. 58 The vessel fragments (handle and bottom) analyzed by Szakmány and Józsa were found during the excavation of the workshop and in the Castrum villa. 59 Aldini 1978, 236-245; Aldini 1981, 5-13 and 40-45; Stoppioni 1993; Carre 1985, 228-231; Williams et al. 2005, Forlimpopoli type; Rizzo 2014, 126-127. 60 Lipovac Vrkljan 2009, 309-314; Lipovac Vrkljan 2011, 9-18. 61 Carre 1985, 230; Williams et al. 2005b, Forlimpopoli type. 62 Carre 1985, 230; Tchernia 1986, 258-260; Williams et al. 2005b, Forlimpopoli type; Rizzo 2014, 127; Bezeczky 2005b, 48; Auriemma et al. 2008, 175. 55

18.03.19 14:39

46

THE AMPHORAE

DRESSEL 2-4

CONCLUSION

Quite a few Dressel 2-4 amphorae were found at the Castrum villa, the majority of which were produced on the Adriatic coast. Such amphorae were not produced in the Fažana workshop. Some items were brought from the Tyrrhenian coast. These amphorae from Campania (35), have a very characteristic fabric, with many “black sand” inclusions. The Dressel 2-4 amphora type was an imitation of the Hellenistic Koan vessels. It was produced at a number of sites (Italy, Spain, France, North Africa, Cilicia and Egypt), and its forms are quite varied.63 Places of production have been found in Rimini/Ravenna and Modena in Emilia Romagna. Stamps suggest that it was probably also produced in Trieste and, perhaps, northern Istria (Školarice).64 The production of the first Dressel 2-4 amphorae is dated to the second quarter of the first century B.C. Production was continuous until the second or third century A.D.65 This amphora type contained wine.66 They were primarily shipped to the western Mediterranean, but can be found in the northern provinces (Raetia, Noricum and Pannonia) and Asia Minor, as well as at Egyptian, Red Sea and Indian sites.67

From the end of the Republican period onward, Istria and the Brijuni Islands played an ever more significant role. Wine, olive oil and, perhaps, fish sauces transported in Adriatic amphorae are among the Castrum villa finds until the mid-Roman period. The production of the wellknown Istrian olive oil began at the end of the Republican period. Oil production and export was continuous until the period of Hadrian. A new situation arose when oil production ceased after Hadrian. We noted that it was replaced by Baetican oil, and amphorae containing African oil have also been found. New amphora types (Fažana 1 and 2) appeared, which clearly reflects changes in agricultural production. Wine was initially imported from Italy in Lamboglia 2, Dressel 6A and Dressel 2-4 amphorae. Later, wine was also produced on the island, although it is unclear when production started. The number of presses suggests that a significant quantity of wine was produced. If this was exported, one must wonder what vessels were used for this purpose. As we do not have chronological data, it is impossible to date the Forlimpopoli wine amphorae precisely, and their origin also requires further study. The Porto Recanati and Aquincum 78 amphorae, found in limited quantities, also pose interesting questions. We do not know much about agricultural production after the third century, but it is clear that oil production was continuous. We have no information regarding the quantities produced, but it is certain that oil was imported from the eastern Mediterranean in the Late Roman period.

SICILIAN AMPHORAE One fragment of a north Sicilian late Roman amphora (177) was also found at the villa. Such Crypta Balbi 2 amphorae are rare in the northern Adriatic region.68 The provenance of some of the amphorae could not be determined precisely. Two of them (53 and 54) have the label “Mid Roman amphorae” in the catalogue. They may come from Sicily too.

Panella 2001, 194; Williams et al. 2005a, Dressel 2-4. Žerjal 2008b, 464-465. 65 Williams et al. 2005a, Dressel 2-4; Bezeczky 2013, 129. 66 CIL, IV.  2556-2559, 5554, 5560-5561, 5592, etc.; Panella 1980, 256. 67 Tchernia 1986, 127-129; Panella 2001, 192-211; Williams et al. 2005a, Dressel 2-4; Tomber 2008, 40; Bezeczky 2013, 129-132; Rizzo 2014, 108-115. 68 Saguì 1998, 321; see González Cesteros in this volume. 63 64

BEZECZKY_Text.indd 46

18.03.19 14:39

47

5

HISPANIC A MPHOR AE



Horacio González Cesteros – Piero Berni Millet

1

IBERIAN AMPHORAE PRODUCTION AND THE ADRIATIC REGION

The Iberian Peninsula was a very rich region in Antiquity, famous for the production of huge quantities of basic foodstuffs and raw materials, mainly metals, grain, horses and wine, olive oil and fish sauces.1 The last three products were sent in amphorae throughout the whole Roman world and beyond its frontiers. Even if we talk in general about a Hispanic production or Hispanic products, the Iberian Peninsula was, and is still today, a very varied geographical area,2 with strongly marked contrasts between the north and the south, between the coastal and inland regions, and between the Mediterranean Levantine area and the Baetican and south Lusitanian coastal regions. These large differences made the Hispanic amphorae production widely varied. As containers used primarily for maritime or fluvial transport, amphora production was located in the coastal areas of ancient Hispania, or around some big rivers such as the Ebro in the northwestern part of Spain, the Tagus and Sado in the central area of Portugal and, above all, the Guadalquivir in southern Spain. The three Augustan provinces of the Iberian Peninsula, Hispania Citerior Tarraconensis, Baetica and Lusitania, exported their amphora products from the Classical period until Late Antiquity, together with Mauritania Tingitana, a region closely linked to Baetica since pre-Roman times.

HISPANIC PRODUCTS IN THE IMPERIAL PERIOD The strong complexity of the Hispanic foodstuff production processes and the variety of the products is manifested in a clear diversity between the rhythms and markets of the different Hispanic exports. Concerning the Adriatic region, so far as we know the Tarraconensis amphorae, mainly for wine products, are only represented by a probable Dressel 3-2 spike documented in Aquileia3 and another possible exemplar found in Altino.4 The presence of “Hispanic Dressel 2-4” amphorae is hinted at by some underwater finds along the Dalmatian coast,5 as for example within the mixed Spanish and south Italian cargo of the Pupak reef shipwreck (late first century A.D.), or in the Paržanj island near Hvar.6 In the Pupak reef material, a Dressel 2-4 stamp of HISP has been interpreted as a sign of a Hispanic production. We have no parallels for this stamp, and the form of this piece looks strange for a Tarraconensis Dressel 3-2. Therefore we are sceptical regarding whether it was produced in any Hispanic workshop. Within our material in Brijuni, we found not a single fragment of Tarraconensis amphorae. This total absence suggests that despite some sporadic pieces found in some contexts in the central and eastern Mediterranean,7 their main markets were always located in the western part of the Roman Empire, above all in southern and western Gaul during the late Republican and Augustan period,8 and on the Tyrrhenian coast and the cities of Rome and Ostia during the first century A.D.9 Verzár-Bass 1994, 389, Pl. 60 A1. Cipriano 2003, 242. 5 M. Jurišić (2000, 11) remarks that the amphorae from the ŠkoljićUnije islands, previously classified as Pascual 1, is not that, but a Portorecanati amphora type. 6 Jurišić 2000, 12-14, 70-71. 7 Adamsheck 1979; Hayes 1992; Reynolds 2000; Bezeczky 2008; Bezeczky 2013a, 139; González Cesteros  – Yilmaz 2013. 8 Laubenheimer 2015. 9 Tchernia – Zevi 1972; Tchernia 1986; Rizzo 2003; Panella 1973. The problem about finding a good calibration of the quanti 3 4

The image transmitted by ancient authors from the Roman period, including Pliny, Strabo and Columela, all show us how versatile the Iberian economy was. 2 In the third book of his “Geographica”, the Greek writer Strabo makes a very vivid ethnical and cultural division between the different parts of the Iberian Peninsula, contrasting a highly developed wine and oil-consuming Iberia, mainly the south and the Mediterranean coast, with a barbarian Iberia where the tribes were still living on butter and acorns. 1

BEZECZKY_Text.indd 47

18.03.19 14:39

48

THE AMPHORAE

Southern Spain, the Roman imperial province of Baetica, was an extraordinarily rich area that produced and exported an impressive agricultural surplus. This success is reflected in the huge quantities of Baetican amphorae found all around the Mediterranean, as well as in the central and northern European regions. The most important Baetican products exported in amphorae were fish-sauces and salted fish products from the coastal and lacustrine areas, wine or wine subproducts from all around the province, and vast quantities of olive oil from the Guadalquivir valley and nearby regions. It is important to differentiate between the Baetican production regions to elucidate the different impacts of Baetican amphorae imports in the northern Adriatic, and within the material from Brijuni. Baetican amphorae production and export are documented from the Classical period10 to the sixth century A.D.11 The liveliest export phase of Baetican wares occurred between early Augustan times and the third century A.D., based on the extraordinary quantities of fish-sauces and, above all, olive oil sent to the military markets in northern and central Europe, as well as to the civilian subsidized market in the city of Rome. The north Adriatic region shows a specific dynamic, unlike the Baetican privileged markets in the western Mediterranean and closer to its markets in the East. In the north Adriatic area, we find a good number of Baetican imports during the first two centuries A.D. Nevertheless, if we investigate more carefully we can observe varying commercial trends operating in the different parts of the Adriatic region. Despite the lack of any complete study for the Dalmatic area, it looks as if the Baetican products arrived in small quantities during the first and second centuries A.D. Based on the underwater and published evidence, we record the arrival of some Dressel 7-11, Beltrán IIA and IIB, Haltern 70 and Dressel 20.12 Further, from some inland places (but still close to the coast) such as in Burnum,13 Narona or in Split,14 we can find Hispanic Dressel 7-11 from the late first century B.C.  or early first century A.D.  The Guadalquivir valley products are still to be found in lower quantities than

the fish products in the first century A.D. The presence of Haltern 70 is only known through four exemplars found in the Pupak reef finds.15 In the excavations of Diocletian’s Palace in Split, a fragment of a rim of a first century A.D. Dressel 20 was found,16 but the presence of Dressel 20 is better confirmed by the discovery of a cargo of second-century A.D. Dressel 20 pieces in Spinut, near Split,17 many of them bearing stamps of the Antonine period. They were found together with some Tripolitana I amphorae,18 clearly indicating olive oil importation from overseas during the last twothirds of the second century A.D. Another Dressel 20 similar to the Spinut pieces is documented by Cambi as coming from the island of Žut,19 and a third-century Dressel 20 was also found on Stipanska Island,20 showing a regular enough presence of Dressel 20 in this area during the Antonine and Severan periods. The presence of Baetican fish-sauce amphorae and some sporadic second-century Dressel 20 is also documented at various places along the Istrian Peninsula, as for example in the waters off Rovijn, where some Dressel 7-11 and maybe Dressel 20 have been found;21 in Savudrija, where some Dressel 20 fragments are documented;22 in the Loron villa near Poreč, where Dressel 6B and other Istrian types were discovered, accompanied by some fragments of Dressel 7-11,23 and maybe in Rijeka, where a fragment of a Pompei VII is mentioned in the Principia excavation.24 Nevertheless, it is around Pula, the most important city and the Roman port in Istria, where the biggest quantities of Baetican amphorae within the Istrian Peninsula ought to be found. Indeed, a recent visit to the Roman amphitheatre in Pula revealed in the exhibition rooms some exemplars of late second and third-century A.D. Dressel 20, as well as a couple of exemplars of Baetican Dressel 7-11. Most of these amphorae must come from local excavations in Pula, probably in the harbour area. However a late Antonine or Severan Dressel 20, displayed separately from the other Hispanic pieces, comes

Jurišić 2000, 15, 71. Will 1989a, 63. 17 Cambi 1983, 363-381; Jurišić 2000, 22; Radić-Rossi 2007. 18 The non Dressel 20 amphorae from Spinut have normally been classified as Africana II. Nevertheless, the drawings presented by Cambi 1983 (Figs. 2 and 3) show without any doubt that they are Tripolitanian I forms. 19 Cambi 1983, 382. 20 Cambi 1983, 381-382. 21 Starac 2006. 22 Vidrih Perko – Župančić 2005, Figs. 4, 4-6. 23 Marion – Starac 2001, 120. 24 Višnjić 2009, 130. 15 16

ties of Tarraconensis amphorae in the city of Rome was already mentioned in a recent paper (González Cesteros 2015). We hope the most recent works of G. Rizzo, such as the recently published work about Ostia (Rizzo 2014) can at least partially fill this important vacuum. 10 Sáez Romero 2010. 11 Bernal 2000; Bernal – García Vargas 2008, 676. 12 Jurišić 2000; Glicksman 2005; Vidrih Perko 2006. 13 Borzić 2011. 14 Glicksman 2005, 1999.

BEZECZKY_Text.indd 48

18.03.19 14:39

CHAPTER 5 – Hispanic amphorae (González Cesteros – Berni Millet)

from an underwater find off the coast of Fažana opposite the Brijuni Islands.25 On the other side of the Adriatic, Baetican amphorae, mainly for fish products, reached the northern part of Italy uninterruptedly, constituting a significant portion of the non-Adriatic amphorae.26 The relatively important quantity of Dressel 7-11 that arrived here caused, we believe, local imitations to be spawned, developing a special type of Dressel 7-11.27 In the Veneto and Friuli regions proper the arrival of Hispanic fish sauce-type amphorae are documented only in small quantities in Aquileia,28 in Altino29 and in the Padova territory.30 A special case, though, is constituted by the substantial numbers of these products found in Verona, which cannot be taken as sporadic finds.31 This pattern of distribution can be rounded off with the addition of some places in Emilia-Romagna,32 such as Parma, Piacenza and Modena.33 In most of these last, the pointed Imperial-phase fish-sauce amphorae (such as Dressel 7-11, Dressel 12 and Beltrán IIA and B) can be found in good quantities, being reused for the construction of drainage systems and for the insulation of floors. On the other hand, the presence of Dressel 20 is not well documented and it looks like it was a very rare occurrence during the first century A.D. We think there are two reasons that explain this fact. The first one is that in most of the excavated contexts with Hispanic materials, the Hispanic amphorae were being utilized for some construction propose, for which the pointed amphorae types were more useful. Secondly, the high presence of Istrian Dressel 6B during the first and early second centuries A.D. squeezed all other kinds of olive oil containers out of the market. A good

Cambi 1983, 383. Pesavento Mattioli 2000. 27 This is a question not exempt from controversy. New analyses made by Capelli excluded the possibility of production of these strange Dressel 7-11 (quite close to Dressel 8) in the central and southern Adriatic, and probably they are not from a northern Adriatic production either, even if the last is not to be rejected out of hand. We think the formal features, the presence of ink inscriptions mentioning “Hispanum”, comparable to the Dressel 7-11 copies from Lyon, and the absence of these amphorae outside of northern Italy (with the exception of one exemplar found by Bezeczky in Poetovio) all suggest a production centre not too far away from their main distribution places in northern Italy. For the latest on this subject: Mongardi 2016. For Poetovio see Bezeczky 1993. 28 Verzár-Bass 1994. 29 Cipriano 2003. 30 Pesavento Mattioli 2000, 744. 31 Pianetti  – Modrzewska 1994; González et al. 2000; Pesavento Mattioli 2000. 32 Pesavento Mattioli 2000. 33 M. Mongardi, personal communication.

49

line of enquiry for the future would be to analyse the presence of Dressel 20 from the Hadrianic period, once the massive Istrian olive oil production came to an end and the Baetican Dressel 20 took over the old Istrian markets in Noricum and Pannonia.34 Aquileia is the best place for achieving a better understanding of this process.35 Some Dressel 20 have already been found there, some of them with stamps of the Antonine period,36 and as we have mentioned before, some shipwrecks with Dressel 20 from the second century A.D. exist in the Adriatic area. The westernmost province of ancient Hispania also exported some amphorae to Brijuni and the Adriatic shores. The beginnings of amphorae production in Roman Lusitania still remains a contentious subject. Even if an early amphorae production, imitating the most typical Baetican and Italian forms, was already under way in the second part of the first century B.C., a wide diffusion of Lusitanian types is not well documented until the second half of the first century A.D.37 The main form of this period is without doubt the Dressel 14, produced at a considerable range of kiln sites on Lusitanian terrain, especially from Peniche in the north to the River Sado in the south, and along the Algarve coast region.38 Other forms were developed in the second and third centuries A.D., above all the Lusitania 3, a flat-bottomed sort clearly linked to the Gauloise 4 and Dressel 30 of the second and third centuries.39 The Imperialphase Lusitanian amphorae from the first to third centuries A.D. had a great take-up in their own regional markets, as well as in some northern Hispanic areas such as Gallaecia40 and in some inland centres of western Baetica.41 Outside the Hispanic markets, they have been found in Sardinia, on the route to Rome,42 in Italy in the Campanian area and, above

25 26

BEZECZKY_Text.indd 49

Bezeczky 1998a, 10-11. In September 2013, together with T.  Bezeczky, we visited our colleague F.  Schimmer in Aquileia, where he kindly showed us the ceramic material from his surveys in the canal area, where some Dressel 20 from the second and third centuries A.D. had been recovered. We are very grateful to Dr. Schimmer for allowing us to see and mention his unpublished material. 36 Pesavento Mattioli 2000, 741. The stamps M.E.EVPRO and M.AEM.RVS have been classified as of Antonine production by Berni (2008, no. 1367 and 1279). 37 On the diffusion of Lusitanian Amphorae, see in the recent publication: Vaz Pinto et al. 2016. 38 Fabião 2008, 729-733; Vaz Pinto et al. 2016. 39 Diogo 1987; Fabião 2008; Quaresma  – Cordero Raposo 2016; García Vargas 2016. 40 Fernández Fernández 2016. 41 García Vargas 2016. 42 Porqueddu et al. forthcoming. 34 35

18.03.19 14:39

50

THE AMPHORAE

all, in Rome43 and its port in Ostia.44 They sporadically reach some places in the eastern Mediterranean, for example Beirut45 or Elaussia Sebaste,46 but in the Adriatic there were no real regular deliveries during the Roman Imperial period.47 HISPANIC PRODUCTS IN LATE ROMAN TIMES The export economy of ancient Hispania was transformed during the third century A.D., undergoing several changes that not only affected the main amphora forms, but also their markets and routes.48 During the Late Empire, the Baetican products still appear in large quantities in their primary markets in the western part of the Empire,49 but not at the overwhelming level they did before. In the eastern Mediterranean, even if the preliminary works of D.  Bernal50 have laid a foundation for further research, we still lack detailed studies about the Hispanic amphorae diffusion in the different eastern regions. The situation is best known in places such as Corinth,51 Ephesus,52 Gortyn,53 Beirut,54 Caesarea Maritima55 and Alexandria,56 which reveal a continuous, but minor, trickle of Baetican products from the Augustan era to the first half of the fifth century A.D. and even later, as the presence of some Keay 19 and Dressel 23d in contexts with material predominantly dating from the late fifth and sixth centuries A.D. in Ephesus have recently suggested.57 In the case of Lusitania, the production of their amphorae reached its peak in the fourth and fifth centuries A.D.58 and continued, at a diminished level during the sixth century too.59 For the northern Adriatic area, we can confirm a continuous presence of Hispanic amphorae (at this moment exclu-

sively Lusitanian and Baetican products) in contexts from the third and fourth centuries A.D.60 They disappear from the Adriatic progressively during the first half of the fifth century A.D.61 Olive oil amphorae from Baetica are very rare in our region during the late Roman period. We have only found references to Dressel 23 in northern Istria, as, for example, in the villa of Školarice near Koper in Slovenia,62 and maybe in Savudrija, in the coastal area of Croatia.63 In the Trieste and Venetia environs, we also find some references to a sporadic presence of Dressel 23.64 Forms related to fish products, such as Keay 16 and Keay 19, are present almost everywhere in the Adriatic, but always in small quantities.65 An example of the level of the influx of these kinds of amphorae in the Istrian region is the discovery of a complete Almagro 51C, probably of Lusitanian origin, in waters near Barbariga, some kilometres north of Fažana,66 and the recent discoveries in the harbour area of Pula.67 Together with these pieces in Aquileia are some Beltrán 72,68 and here again, and in Trieste and in our context in Brijuni, also some Beltrán 68. Both amphora types are interesting Baetican ones linked to fish products and wine exportation, from the third to early fifth centuries A.D.69 Regarding the Lusitanian amphorae, it remains true that an accurate differentiation amongst the various Baetican products is seldom met with in publications on the Adriatic and other Mediterranean regions, as has been mentioned several times and recently by virtually every scholar present in the workshop that took place in Troy in October 2013.70 In Auriemma et al. 2012. The practical absence of Hispanic amphorae in Classe (port of Ravenna) in contexts from the fifth century A.D. is quite remarkable. In the southern Adriatic, some small quantities of Almagro 50 and Keay 19 have been found in contexts from the fourth and fifth centuries (Volpe et al. 2010), but later they disappear completely from the Adriatic region. 62 Žerjal 2010, 704. 63 Vidrih Perko – Župančić 2005, Figs. 4, 5. 64 Maselli Scotti et al. 2003, 25; Auriemma et al. 2012. 65 Jurišić 2000, 22; Vidrih Perko 2006; Auriemma et al. 2012. 66 Višnjić et al. 2010. 67 We would like to thank the excavation director Ms. Ida Koncani Uhač for inviting us to see this interesting material from the second to early fourth centuries A.D. 68 Gaddi – Degrassi 2016, Fig. 5. 69 For Beltrán 72: Bernal 2016b; González Cesteros et al. 2016; for the Beltrán 68: Bernal Casasola 1996; Bernal 2009; Bernal 2016a. In Trieste they found a complete Beltrán 68 in the excavations of a well in the Roman forum, but it was identified, together with a complete Baetican and some Lusitanian Almagro 51C, as a Mauritanian Dressel 30 (Degrassi – Maggi 2011, Figs. 1, 3-5). 70 Vaz Pinto et al. 2016. 60

Rizzo 2003; Martin 2016. 44 Panella 1973; Rizzo 2014; Martin 2016. 45 Reynolds 2000. 46 Ferrazzoli 2010. 47 The presence of Dressel 14 is documented by some sherds at Aquileia: Gaddi – Degrassi 2016, 440, but it was not possible to determine if they were of Baetican or Lusitanian origin. 48 Bernal Casasola 2000a; Bernal Casasola 2000b. 49 González Cesteros 2010; Berni Millet – Moros Díaz 2012, 214-217. 50 Bernal Casasola 2000b; Oren-Pascal – Bernal 2000. 51 Warner Slane – Sanders 2005; Will 1983. 52 Bezeczky 2008; Bezeczky 2013a; González Cesteros 2012; González Cesteros – Yilmaz forthcoming. 53 Romero – Pascale 2001. 54 Reynolds 2000. 55 Oren-Pascal – Bernal 2000. 56 Will 1983. 57 This material is still being processed by one of us; it is hoped that we will soon be able to present it to the scientific community. 58 Fabião 2008, 740-743. 59 Fabião 2008, 743. 43

BEZECZKY_Text.indd 50

61

18.03.19 14:39

CHAPTER 5 – Hispanic amphorae (González Cesteros – Berni Millet)

most cases, scholars have published them en bloc, assigning the whole Hispanic production, with the exception of olive oil containers, to the Lusitanian. This problematic question will be tackled next. For now, we can only point out that, due to the common use of the same maritime routes,71 we should expect the presence of types like Almagro 50 (Keay 16, Keay 22), Almagro 51A-B (Keay 19), and Almagro 51C, that is, a dual presence of Baetican and Lusitanian pieces at the end markets, as well as occasionally finds of other Lusitanian forms, as, for example, the Keay 78 found in the Forum of Aquileia.72 2

IDENTIFICATION PROBLEMS IN DISTINGUISHING LUSITANIAN AND BAETICAN AMPHORA TYPES

The close connections between the Baetican and Lusitanian amphora types have their roots in the facts that both provinces are located on the Atlantic seaboard and in the economic expansion of Baetican production from pre-Roman times. A close constant economic relationship existed between both provinces, and the production of similar amphora types in both regions runs from the second half of the first century B.C.73 down to the latest documented exports in the sixth century A.D., or even later.74 The first Lusitanian Roman amphora types followed the manner of the ovoid amphorae, close to those produced in the Guadalquivir and Cadiz workshops during the second half of the first century B.C.75 Accordingly a Lusitanian version of types such as Haltern 70 and Dressel 7-11 in Augustan times has been proposed in several publications,76 and a limited production has been documented in the territory of Augusta Emerita.77 Nevertheless, before the Julio-Claudian dynasty the Lusitanian amphora exports were mostly limited to the provincial and nearby Baetican and Galician markets. It is during the second part of the first century A.D. that the containers from the Sado-Tagus region and the Algarve (this last region was even more closely linked to Baetica) start to be produced in huge quantities and reached customers beyond the peninsular borders.

Bombico 2016; Bernal 2016c. Vide infra. 73 Morais – Fabião 2007. 74 Bernal – García Vargas 2008, 676. 75 Morais – Fabião 2007; García Vargas et al. 2011. 76 Fabião 2008, 726. 77 Bustamante Álvarez – Heras Mora 2013. 71

51

The Lusitanian amphora production was dominated during the last part of the first century and the second century A.D.  by the Dressel 14. This amphora is the most typical Imperial type produced in Lusitania, but at the same time a Baetican variant exists too, with some typological peculiarities. In the 1970s scholars such as M.  Beltrán78 and A.  J.  Parker79 discriminated between them, calling them A and B. Parker named Dressel 14B as the Lusitanian version: still today this remains the most commonly used name for this variant. Even if scholars progressively observe the distinction between the two variants, there yet remains an unremitting confusion. This lack of clarity gets in the way of a better understanding of the productive and commercial dynamics of both provinces during the first and second centuries A.D. We think that the Baetican version was produced at a lower level than the Lusitanian Dressel 14, being a secondary type lagging behind other main fish-sauce vase-types from Baetica in the late first and second centuries A.D., such as Beltrán IIA and IIB. For the most important Lusitanian and Baetican amphora types of Late Antiquity, the same problems of duality and resemblance are even worse than during the Imperial period. Types such as Almagro 50, 51A-B and 51C were produced in a large number of kilns from today’s Galicia to eastern Baetica, resulting in major formal variations, as already noticed by Keay in his work about Late Antique amphorae in northeastern Spain.80 However, since the end of the 1990s the development and study of the very rich amphora production of Late Antique Hispania has clarified the typological differences and the individual commercial dynamics of Lusitanian and Baetican production. The work demonstrates not only the formal variants linked with each one of these regions, but also the pre-eminence of certain types among each production family. The Lusitanian potters might choose the Almagro 51C and some particular versions of the Almagro 50, such as the Keay 22 and the so-called “Sado 1”, as their principal and preferred types,81 while in the Baetican workshops the Keay 19 and the bigger Keay 16 were favoured. These generalities do indeed assist in giving a better distinction between both producers on a regional basis, but the reality is more complex still. We must consider the implications about the exceptional numbers of workshops involved in the production of amphorae, each producing several dif-

72

BEZECZKY_Text.indd 51

Beltrán 1970, 456-464. Parker 1977, 37-39. 80 Keay 1984, with all the variants proposed for forms 16, 19, 21, 22 and 23. 81 Fabião 2008, 736-743. 78 79

18.03.19 14:39

52

THE AMPHORAE

ferent amphora types at different times. A better knowledge of the fabrics and an improved familiarity with the Hispanic forms is the only safe way to establish a certain identification of Baetican and Lusitanian types. Another important typological misunderstanding affects some special Hispanic types retrieved at different Mediterranean sites: those classified as Mauritanian Dressel 30. This problem particularly impacts upon the Baetican Beltrán 68, a type produced in various Baetican workshops from the third to the early fifth centuries A.D.82 It has turned up here and there in Italy and the eastern Mediterranean, for example in Caesarea Maritima,83 Gortyn,84 Brescia85 and Aquileia, where at least three complete examples have been found.86 These amphorae were classified as Dressel 30 in Gortyn and Aquileia, as well as in a recent publication of the ceramic material from Caesarea,87 even though Bernal had already published them as Beltrán 68. In our opinion, the Hispanic amphorae of the Forum context in Aquileia have been badly misjudged. Not only was the Beltrán 68 published as a Mauritanian Dressel 30, but so also were a complete Baetican Almagro 51c/Keay 23,88 as well as some third/fourth century Almagro 51c/ Keay 23 (with a large everted rim and flat handles with central groove).89 Furthermore, in the old Forum excavation a Lusitanian Keay 78 was classified as a Dressel 23.90 It is necessary to make a careful revision of Hispanic amphora imports from the late Julio-Claudian period right through to Late Antiquity. Paying attention to their formal features and to the macroscopic characteristics of their fabrics constitutes the only way to formulate a better distinction between the Baetican and Lusitanian amphorae in their eastern and central Mediterranean places of import. We fully acknowledge that it is not an easy matter, because there are a lot of resemblances between types and subtypes and a perplexing range of Baetican and Lusitanian fabrics, but today good works of synthesis and internet tools exist

by means of which one can at least achieve a correct typological identification.91 3

THE HISPANIC AMPHORAE FROM THE BRIJUNI CASTRUM VILLA

After presenting a general overview of the Adriatic region and some problems related to the Hispanic production, it is time to analyse the Hispanic material documented during our campaigns on the island of Brijuni. Within the material from the Castrum where we worked on Brijuni, we retrieved a small quantity of Hispanic amphorae. Even if the bulk of local and Adriatic products outnumbers the finds of imports from other regions, the Hispanic material is still found in much lower quantities than the African or the Oriental containers. As was the case for the whole Istrian and eastern Adriatic region, the small number or lack of documented Hispanic amphorae till late Julio-Claudian or Flavian times can be partially explained by the existence of locally produced olive oil and fish products. Nevertheless, for olive oil, we find within the Castrum material some Baetican Dressel 20 (40-45). All of them are variants produced after the time of Hadrian’s reign, and are of the typical Guadalquivir Valley fabric. From the second century A.D., together with the almost complete amphora (41) with a graffito ante cocturam presented by P. Berni in this volume, we could recognize some other rim and handle fragments of Dressel 20, one of them bearing a stamp from the second half of this century.92 During the Severan period and the third century A.D., the Dressel 20 continued to arrive in Brijuni, but for the fourth and fifth centuries we lack any Dressel 23 or nearby Baetican olive oil types. This pattern seems to be repeated throughout the whole Adriatic region, where Dressel 23 are really rare. The explanation must be sought in the Adriatic’s own production and in the new sources of olive oil imports, such as, for example, Africa and the eastern Mediterranean.93 A very useful tool for any amphora specialist is the internet-site of the project “Amphorae ex Hispania. Paisajes de producción y consumo”: http://amphorae.icac.cat/. Here the archaeologist can find a detailed description with wide-ranging literature about the most important amphora types produced in the Iberian Peninsula and the Balearic Islands. 92 Catalogue no. 40, see Amphora stamps and inscriptions chapter. 93 Some oriental amphora types possibly linked to an olive oil content, such as the LRA 2, arrived in the Adriatic area, but their quantities are not really enough to represent a continuous supply of oil from this region. The African amphorae are well-documented from the third century A.D., but not all of the present types have to be linked to an olive oil content (see Bonifay – Capelli in this volume). 91

Bernal 2016a. 83 Bernal Casasola 2000b. 84 Portale – Romeo 2001. 85 Bruno 2002, no. 63-64. 86 Degrassi – Maggi 2011, Figs. 1, 3; Villa 2013. 87 Johnson 2008. 88 This piece should have been produced in the same Cadiz workshops as the Beltrán 68, as resemblances in shape and fabric declare (Degrassi – Maggi 2011, 263, Figs. 1. 3 and 4). 89 This kind of Almagro 51c has recently been found in some new excavations in the harbour area at Pula. 90 Verzár-Bass 1994, Pl. 60, AS 6. 82

BEZECZKY_Text.indd 52

18.03.19 14:39

CHAPTER 5 – Hispanic amphorae (González Cesteros – Berni Millet)

We do have some earlier Hispanic fish-sauce amphora exemplars in Brijuni. From the late first century A.D. there is a rim fragment of a late form of Dressel 7-11 (37), similar to some pieces found in the Neronian-early Flavian depot of “Estagnon” in Fos-sur-Mer.94 The fabric of our piece is buff in colour on the external surface, also with traces of a whitecream slip, and red brick colour at the break, with a hard and sandy texture: it is close to the red-coloured products of the Cadiz area. The Beltrán IIA base (38) must be dated from the same period or a little bit later. The fabric of this piece is buff in colour on the surface and at the break too, of a relatively soft texture with many small quartz inclusions, small voids, and some hematite inclusions. These petrological features suggest a different but nearby production area, maybe on the banks of the ancient Lacus Ligustinus or in the area around today’s Jerez de la Frontera. Also from the late first century or the first half of the second century A.D. is a rim that seems to be from a rough Dressel 14A (39).95 This type was produced mainly in the Cadiz-Algeciras and Malaga-Granada areas from the Julio-Claudian to the late Antonine periods.96 The fabric of our exemplar seems to be closer to some productions of the Malaga region: it is hard and coarse, buff in colour on the surface and at the break, with large quantities of small and medium-rounded quartz of different sizes, some sparse small red pieces and large lime-based inclusions; some big voids are seen at the break. The typological features seem close to the Malacitanian Dressel 14, such as were produced in the Calle Carretería workshop.97 We can argue that the zenith of the production of Dressel 14 in the Mediterranean Baetica is around the Flavian period and the first half of the second century A.D. This was probably the moment when this piece arrived at Brijuni. Wine from Baetica is also documented in Brijuni by the presence of two exemplars of Beltrán 68, one of them (51) undoubtedly produced in a workshop in the Bay of Cadiz in the late third or fourth century A.D. The fabric of this exemplar shows, without a doubt, its origin in the Bay of Cadiz, and is similar to the late Dressel 7-11 rim (37). It is still covered by a pale slip on the external surface, but the colour is red brick at the break and on the inside. The fabric is hard and coarse and has several large hematite inclusions.

Marty – Zaaraoui 2009, Figs. 9,3; 10,3. We would like to thank Dr. García Vargas for helping us with this form. 96 Bernal Casasola 1998; Mora Serrano – Corrales Aguilar 1997; García Vargas 2000, 83-84. 97 Rambla Torralvo – Mayorga Mayorga 1997, Figs. 4-5. 94 95

BEZECZKY_Text.indd 53

53

For the late Roman period we have some Hispanic fish amphorae in the material from the Castrum. Despite the absence of Lusitanian products during the first two centuries A.D. and the continuous arrival of Baetican amphorae then, during the fourth and probably fifth centuries the Lusitanian products replaced the Baetican imports being sent to Brijuni, and all of the documented fish amphorae from this period come from Lusitania. The presence of rims from two versions of the Lusitanian Almagro 51A-B (47 and 48) must be placed at the end of the fourth century A.D.  or, better still, during the fifth century A.D.,98 a time of high production and export from the Lusitanian workshops. The fabric of these two exemplars is different in colour and texture, but in both cases they are of a central Lusitanian production. The first Almagro 51A-B rim (47) exhibits a red colour with remains of a white slip; it is hard and coarse, with many small and medium-size rounded or sub-angular inclusions, most of them quartz of different colours; it also has occasional small mica flocks, some lime inclusions and a couple of red rounded inclusions, probably grog/chamotte. On the other hand, the other Almagro 51A-B, with a longer rim, is in a finer fabric, buff in colour, soft and sandy in texture and with a large number of small lime and quartz inclusions. It also has some occasional mica flocks and rounded and small red inclusions. This fabric is similar to the rim of the Almagro 51C (49), just the colour is a little bit different. The other Lusitanian amphora we have documented in Brijuni is a ring base of form Keay 22, a related type to the Almagro 50, produced in the Lusitanian workshops during the fourth and fifth centuries A.D.99 In this case the fabric is also close to the last two fragments, but it was fired at a higher temperature that ensured the disappearance of almost all the lime components, creating some large voids. The external surface is pale buff in colour and at the break it is buff and orange, presenting the small and medium-size red inclusions and occasional mica flecks typical of Lusitanian production. CONCLUSION In conclusion, we can argue that the Hispanic amphorae arrived at Brijuni from the late first to the early fifth centuries A.D. Even if they travelled together, the dynamics of the Baetican and Lusitanian imports suggest that the import of the products of these two Hispanic regions occurred at different times. The Baetican amphorae arrived mainly in the 98 99

Mayet – Tavares Da Silva 1998, 204-205; Fabião 2008, 740. Fabião 2008.

18.03.19 14:39

54

THE AMPHORAE

second and early third centuries A.D. This is the case of the olive oil containers that always appear after Hadrian’s principate, at the abrupt end, then, of the Istrian olive oil production.100 For the fish products it is important to stress the absence of early Dressel 7-11 forms. We know they arrived all the while in the whole Adriatic region, and even if they are not very numerous in the Istrian Peninsula, some examples of first century A.D. Dressel 7-11 do exist in Pula, Rovijn and Loron. For Brijuni, the earliest fish amphorae type is the late Dressel 7-11 rim from the Neronian or early Flavian period. Equally, one should underline the absence of late Roman Baetican fish amphorae types, such as Keay 16 and Keay 19, which is somewhat surprising as they are documented in other Adriatic contexts.101 The varied Baetican forms we have in Brijuni reflect the different places where these amphorae were produced inside the Baetican territory. All of the olive oil containers have the typical Guadalquivir valley fabrics, while the fish-amphora types show a multiplicity of fabrics from Malaga to Lacus Ligustinus, with a central focus around the Bay of Cadiz. In the case of the Lusitanian products, they arrived from the fourth century A.D.  and continued at least during the first half of the fifth century. All of them are fish amphorae types: they appear to replace the Baetican imports of the

100 101

Bezeczky 1998a, 10-11. Jurišić 2000, 22, Pl. 4; Auriemma et al. 2012.

BEZECZKY_Text.indd 54

previous centuries. Despite the different types documented (Keay 22, Almagro 51C and Almagro 51A-B), all of them seem to come from the same Lusitanian region: this is situated in central Lusitania, from the workshops around the Tagus and Sado confluences. Even if the Hispanic material arrived only in minor quantities at Brijuni, it nonetheless formed a continuous presence from the late first century A.D. to the fifth century A.D. This is more or less the same image we have gleaned from other Dalmatian and north Adriatic contexts, but the absence of some Baetican fish amphorae types of the Augustan and Julio-Claudian periods, as well as late Roman Baetican fish amphorae, is remarkable. On the other hand, the presence of Dressel 20 from the second and third centuries is symptomatic of the decline of the Istrian olive oil production from the late Hadrianic period: this pattern supports the theory of a reduced production of Istrian oil amphorae and the arrival of Baetican products in the north Adriatic and in the Danubian military markets.102 In the case of fish products, the local production must have been constant during the entire Roman period and beyond, but Hispanic products, first from Baetica and later from Lusitania, were imported into this particular place in the northeastern Adriatic at least until the fifth century A.D.

102

Bezeczky 1998a, 10-11; Bezeczky 2000.

18.03.19 14:39

55

6

E ASTERN MEDITERR ANE AN A MPHOR AE



Horacio González Cesteros

THE EASTERN MEDITERRANEAN AMPHORAE AND THE NORTH ADRIATIC REGION Archaeological finds all along the Adriatic region show the existence of a special link between this geographical zone and the eastern Mediterranean during Antiquity. The proximity to western Greece and the Aegean world through the Ionian Sea and then the Corinthian Gulf and Isthmus supported a continuous commercial activity, in both directions, documented well back into pre-Roman times.1 Not only amphorae imports were involved, but many other types of pottery and diverse goods, for example marble and other kinds of stone.2 Nevertheless, the commercial trends that linked the Adriatic region and the eastern Mediterranean seem to have taken differing paths between the north and south Adriatic, mainly during the late Republican and early Imperial periods. We can interpret this distinction as part of a trade directly from both regions into the Aegean area, something also easily observed through the different amphora types produced in each zone in Roman times. For example, in Apulia the main amphora factory, concentrated in the second and first centuries B.C., produced its own ovoid forms with close relations to the north Peloponnese amphorae that developed from the so-called Corinthian amphorae, produced in Corinth and Corfu from the Archaic to the Hellenistic periods. It is a commonplace observation that in the western Mediterranean research does not distinguish enough between the materials emanating from the different eastern Mediterranean regions. The traditional western viewpoint assigns the classification of all the eastern material within the broad term “eastern Mediterranean amphorae”. Perhaps this attitude can be understood for places such as Gaul, northwestern

There exists a huge bibliography about Greek imports in the region before the Roman conquest: AAD 1977; Brusić 1999; Auriemma 2006; Auriemma 2007; Slane 2008; Pizzirani 2013; Auriemma – Silvestrilli 2013; Ugarković 2016; etc. 2 Jurišić 2000, 39-41; Carre – Maselli Scotti 2001, 230; Glicksman 2005. 1

BEZECZKY_Text.indd 55

and central Europe or even the Iberian Peninsula, but not for the Adriatic region, where imports reach high percentages from the Hellenistic period, or even earlier. We believe it is somewhat important to refine the classifications, because we are dealing with a vast geographical expanse, with a huge cultural, political and geographical diversity.3 1

REPUBLICAN PERIOD

During the late Republican period the commercial relations between the Adriatic and the Aegean increased in a significant way. This was due to several factors, for example the widespread commerce linked with the Roman occupation of the northern Mediterranean territories throughout the second century B.C., or the establishment of a free port at Delos in 167 B.C. – a move that was a serious inducement for exchanges between Italy and the Aegean. So far as concerns the Adriatic region, an important factor for the economic and commercial growth was the colonization of broad areas all along the coast and the most important valleys, for example the Po, as well as the control of Illyrian piracy in the late third and second centuries B.C.  Archaeological work has brought to light the introduction of big quantities of some amphora types from the Adriatic to the Aegean during the Republican period,4 but the arrival of Aegean wares in the Adriatic area seems even more important, above all from the moment the new Roman province of Cisalpine Gaul started to be colonized and developed.5 This involves the late Hellenistic Greek amphorae found all along the Adriatic region,

A new study about the Cilician, Cypriot and Levantine pottery imports from latest excavations of Padova University in Aquileia has recently been published (Dobreva – Riccato 2015). We think it a good idea to focus more on the different eastern Mediterranean regions, even more so for sites producing such large amounts of material as Aquileia. 4 Grace – Savatianou-Pétropoulakou 1970; Tchernia 1986, 68-74; Lund 2000; Bezeczky 2006; Bezeczky 2008; Bezeczky 2010. 5 Baldaci 1972. 3

18.03.19 14:39

56

THE AMPHORAE

mainly Rhodian,6 but also others such as Koan, Knidian or Chian. 2

EARLY AND MIDDLE ROMAN PERIOD

During the Imperial period we can observe the presence of a high number of eastern amphorae, often the most significant amphora imports after the Italian,7 with a richer variety of forms and a bigger range of provenances. For the Istrian Peninsula they do turn up in several places, such as, for example, Pula,8 in the Roman villa of Školarice,9 Loron,10 or in Rovijn,11 but as yet the systematic and quantitative analysis of most Istrian contexts is inadequate. We are trying to fill this gap now, and hope that the most recent discoveries in the harbour area in Pula will bring more precision to the matter. Some amphora types from areas outside the Aegean, such as, for example, Palestine or Cilicia, have also been found all along the Adriatic, the northern area being no exception.12 This is a trend that developed still more from the second century A.D., reaching its highest point in Late Antiquity with the arrival of large quantities of Aegean, Levantine and Cilician vessels and the sporadic presence of products from other regions such as Crete, the Black Sea or Egypt. Nevertheless, from the time of Augustus to the mid-second century A.D., the bulk of the oriental amphorae in the north Adriatic area arrived from the Aegean or nearby regions. We also find new amphorae production regions represented in the northern Adriatic from early Imperial times. Even if amphorae from places such as Knidos, Kos, Chios and above all Rhodes13 (or nearby territories) continued to be imported into the Adriatic in good quantities, comprising the bulk of the eastern Mediterranean imports, other pro Jurišić 2000, 14; Tiussi 2007; Marengo – Paci 2008. It has been suggested that local imitations of Rhodian amphorae were produced in the late Republican period in northern Italy (Baldacci 1972, 125; Facchini 1997), based on the interpretation of single petrographical analyses from Calvatone (Facchini 1997) and other places. We prefer to exercise caution on this controversial subject. 7 Jurišić 2000; Pesavento Mattioli 1992; Pesavento Mattioli 2005; Cipriano 2003; Glicksman 2005; Cipriano – Mazzocchin 2011; Mazzocchin 2013. 8 Personal visit to the museum. 9 Žerjal 2008. 10 Marion – Starac 2001, 119-120. 11 Maric 2009. 12 For new finds in Aquileia: Dovreba – Riccato 2015, 132. 13 A new study about Imperial Rhodian amphorae on the Croatian coast is being carried out by P. Dugonjic, who presented new and interesting material from the early Imperial period, some of it from underwater finds: Dugonjic 2015. 6

BEZECZKY_Text.indd 56

duction centres such as Crete or the Asia Minor coastal area were also exporting their products to our region. Amphorae such as the Cretan AC4, AC1 and in lesser quantities AC3 and AC2, as well as the Asia Minor versions of the “onehandled jar”, including early Agora F 65/66 versions,14 are more or less normal on both sides of the Adriatic in this period. Other forms, for example the first and early second century Dressel 24 probably produced in the Chios/Erytrai region, have also been found, albeit but sporadically,15 as well as the alum amphorae from Milos.16 This last form is a specific import answering the needs of the local wool industry.17 A more problematic matter involves the small, but continuous presence of form Dressel 25 in some Adriatic places such as Trieste18 and, above all, in some Cisalpine cities.19 This type is present in the Dressel table made up mainly from the Castra Praetoria amphorae, a corpus which is primarily dated to the mid-first century A.D.  Its production has normally been linked with the late Corinthian and ovoid Apulian production due to its typological characteristics: this has recently been confirmed by finds in the northern Peloponnese, among them the pottery workshop of Aigion, where related forms were produced from the second until the late first centuries B.C.20 The epigraphic evidence found on these amphorae, both tituli picti and stamps, always appears in Greek. The fabric analysis carried out on some exemplars in Padua21 and in Ephesus22 determines their production on the Greek mainland, in the Corinthian hinterland or the northern Peloponnese. This new archaeological data allows us to definitely rule out Manacorda’s suggestion Berti 1990; Carre 2007, 591. Radić – Jurišić 1993; Jurišić 2000; Belotti 2008b. 16 Cipriano et al. 2005. 17 Hence the relatively large number of alum amphorae from Milos found in Padua: all linked with the important wool industry of this city (Pesavento Mattioli 2005). They turn up in other places in northern Italy, for example Novara in the Upper Po valley (Quiri – Spagnolo Garzoli 2015). 18 Auriemma 2007, 141. 19 Cipriano et al. 1997; Cipriano  – Ferrarini 2001, 76-78; Mazzocchin 2013, 87-88. This form is above all well-represented in Vicenza (Mazzochin 2013, 87-88), already making its way into the interior of the Po valley. Starac and Matijašić have also pointed out the presence of one rim of a Dressel 25 in a depot, mainly stocked with Lamboglia 2 and Dressel 6B, in Pula (Starac - Matijašić 1991). However, we are not sure this fragment can really be interpreted as a Dressel 25 or Class 24 of Peacock and Williams: it looks much more like a Brindisian form. 20 We would like to thank K. Filis for this important information and the permission to see this important material in Aigion. Filis 2016. 21 Cipriano et al. 1997. 22 Bezeczky 2013a, 91-93. 14 15

18.03.19 14:39

CHAPTER 6 – Eastern Mediterranean amphorae (González Cesteros)

about a production place around ancient Venafrum,23 based on the elevated numbers of this form present in Ostia. From the late second to the early fourth centuries A.D., good quantities of a varied range of eastern Mediterranean amphorae continued to arrive in the northern Adriatic. This chronological time-frame reflects an interesting period in the ceramic production all around the Roman world: it is marked by the diffusion of some amphorae types developed from old prototypes, but also by the creation of new forms that are going to form the basis for the typical productions of Late Antiquity. We find some good dated contexts of this period in the northern Adriatic, mainly in Trieste24 and in the Aquileia region,25 but also in Pula, where a high number of the amphorae from the museum, today exhibited in the Roman amphitheatre, certainly belong to the third and early fourth centuries A.D.  The material from the excavations near the harbour that we were able to see in the archaeological museum in May 2014 is also predominately to be dated to the late second and third centuries.26 During this important phase the region must have been under the direct influence of the new and significant role played by Spalato and Aquileia as important centres within the organisation of the Roman Empire. As for commerce with other regions, a significant increase in the African imports is to be detected, mainly terra sigillata and several types of amphorae: they appear in quantities equal to or even in excess of the oriental vessels. Concerning local amphorae production, from the second half of the third century that of the Adriatic drops to its lowest level: the region becomes definitely marginal in amphora production and export. The eastern Mediterranean forms continue to arrive in good quantities, above all the Aegean products. Nevertheless, the most important oriental imports of the Hellenistic and early Roman phases, namely the Rhodian and Koan amphorae, gradually disappear from the mid-second century A.D.  onwards, due to a general economic trend that affected the whole Roman world. The vacuum left by this event is filled by material from a variety of other geographical zones, mainly the Asia Minor coastal area and probably some nearby islands like Samos and Chios, where the different versions of the “one-handled jar” (Athenian Agora Manacorda 2012, 153. Auriemma  – Quiri 2004; Auriemma 2007; Auriemma et al. 2012. 25 Degrassi  – Maggi 2011; Auriemma et al. 2012; Villa 2013; Dovreba – Riccato 2015. 26 We would like to thank Ms. Ida Koncani Uhač for allowing us to see this material. 23 24

BEZECZKY_Text.indd 57

57

M 45; M 125-126; J 46; J 47; M 240; M 255)27 and other amphora types, such as the early versions of M 273 and above all the late Dressel 24 or related types, were originally manufactured. The Mediterranean coast of Anatolia also started to export amphora products regularly to the whole Roman world: even if they only arrived in a roundabout way in our region, they have nonetheless been found in almost every archaeological context. From the southern Anatolian regions comes a series of amphora forms with a common characteristic, namely a bifid or pseudo-bifid handle. Even if one admits to an Aegean production of the so-called Pseudo-Koan and pinched bifid types during the second century A.D., it is undeniable that they were also produced in regions such as Pamphilia, Isauria and Cilicia. The same can be said for the different variants of Agora G 199 and M 239, one of the most frequently encountered amphora types of the eastern Mediterranean from the late first to the early fourth centuries: they were probably also produced in Asia Minor28 and a Cypriot production has also been mooted.29 On the other hand, Agora M 54 and G 198 are typical forms from Isauria, Cilicia and Cyprus30 that also reached our region sporadically. Amphorae from other eastern territories only arrived in really insignificant numbers in the northern Adriatic during this period. Pieces coming from the Black Sea, mainly from Sinope, have been retrieved,31 but always in inconsequential amounts, likewise the Levantine amphorae. Even if their presence is important to show the arrival of these goods, we must admit that they were far from forming any part of the bulk of amphora imports into the most important places in northern Dalmatia, Istria and Cisalpina. On the other hand, some Levantine products had already arrived even in the first century A.D., thereafter being documented during the early and middle Imperial Roman period, such as, for example, the Carrot Amphorae Camulodunum 189 and the Kingsholm 117 found in some important centres in the northern Adriatic and Po Valley,32 and they continued in middle Imperial Within the different types of “one-handled jar”, more than just the typical Ephesian forms are known. Moreover, productions from other areas are also present in places such as Trieste (Auriemma 2007, 147) and Aquilea (Villa 2013, no. 68). 28 González Cesteros – Yilmaz forthcoming. 29 Hayes 1991; Lund 2000b; Williams – Lund 2013; Bezeczky 2013b, 165-167. 30 Rauh – Slane 2000; Reynolds 2005. 31 Belotti 2008a. 32 Carre 2007, 591-592; Auriemma 2007, 150; Cipriano – Mazzochin 2011; Mazzochin 2013, 75; Dovreba  – Riccato 27

18.03.19 14:39

58

THE AMPHORAE

times with some occasional arrivals of forms like Chalk 6, Agora K 108 and versions of baggy jars.33 Lastly, some Egyptian AE 3 of the Roman period were imported to Trieste,34 but the presence of other Egyptian amphorae of the Roman period is not documented in the region. In addition to the above patterns, the northern Adriatic region displays some individual characteristics concerning oriental amphora types imported in the mid-Roman period. One remarkable feature is the small incidence of one of the most exported oriental types, namely the Kapitän 2.35 Even if this form does appear in several places, their quantities are always trivial, quite unlike the situation in other Mediterranean regions, where it has been recovered in bigger amounts, especially on the Aegean and the Black Sea.36 Another intriguing matter is that of the presence of the so-called “Middle Roman Amphorae 8” in several northern Adriatic centres. The bibliography about amphorae in northeastern Italy commonly alludes to the frequent presence of this type during the late second and third centuries.37 Even if it is undeniable that most of these pieces could be linked with what Riley called MRA 8,38 his own definition of this form is quite problematic. More attention to its postulated production area in west Cyrenaica is recommended. Riley mentioned the relevance of this form at the Tocra Kiln site, based mainly on amphorae spikes.39 He also mentioned that the pieces found in Berenice include one of the fabrics docu-

2015, 117. In the Augustan shipwreck of Cape Glavat of Mljet Island, a probable Levantine amphora similar to the Augst 47 was found, together with something similar to the Pamphilian amphorae (Radić – Jurišić 1993; Jurišić 2000, Figs. 4, 8-9). We cannot say anything more about these pieces at the moment. 33 Auriemma et al. 2012; Auriemma et al. 2015; Dovreba – Riccato 2015, 117-118. The piece from the Aquileian well that appears in: Degrassi – Maggi 2011, 265, Figs. 2,3, is not an Agora K 108, but a baggy jar version of the third-fourth centuries A.D.  34 Auriemma 2007, Pl. 36, 100-102. 35 Verzár-Bass 1994, 401; Jurišić 2000, 21; Auriema  – Quiri 2004, 52-53. 36 Within the Aegean, a very good example of the high quantities of Kapitän 2 comes from several Ephesian contexts, dated to the end of the second and the third centuries A.D. (Ladstätter 2005; Ladstätter 2008; Bezeczky 2013a, 149). Moreover, they have also been found in huge numbers in almost every Mediterranean region and along the Danubian frontier. 37 Ferrarini 1993; Belotti 2008b; Auriemma et al. 2012. In a visit to the grain depot in Pompeii, we saw some amphorae that could also be similar to the so-called MRA 8 and could confirm the beginning of this production at least to the mid-first century A.D. However, this observation is not certain and more research must be done on this interesting topic. 38 Riley 1979, 193. 39 Riley 1979, 193.

BEZECZKY_Text.indd 58

mented at the Tocra workshop, as well as local fabric from Berenice. Carre also pointed to the probable production of this form in some kilns in Apollonia.40 Even if it presents some resemblances with early forms of M 273, the MRA 8 seems to be a Cyrenaic production. The problem is that this amphora type has not been found in any significant quantity anywhere except at the northern Adriatic sites;41 or at least they have not been recognized or published as MRA 8 anywhere else. The MRA 8 is well-represented in several places in our region,42 including Trieste and Aquileia, always within a chronological range from the mid-second to the mid-third centuries A.D., a period for which we have textual and epigraphical evidence of the contacts between Aquileia and the Cyrenaica.43 Before passing to the Late Antiquity phase, there is also the important question about the contents of the eastern Mediterranean amphorae found in the whole Adriatic region during the early and middle Roman times. Most of the amphora types of these manufacturing contexts are linked to wine consumption, and held such well-known wines as the Rhodian, Koan, Knidian, Chian or the Ephesian brands. However other kinds of goods produced in the eastern regions were so transported too. Together with some dried fruits or dates carried in the Levantine Carrot Amphorae and Kingsholm 117,44 we also find some alum amphorae from Milos. Moreover, the high number of Dressel 24 and related types, mainly in the second and third centuries’ versions, as found in Trieste and other northern places, are significant too.45 These forms are connected to the olive oil export of the Aegean, as the finds in Monte Testaccio and the inscriptions mentioning oil have already confirmed.46 A titulus pictus on the neck of a Dressel 24 (Knossos 15) Carre 2007, 596. As an example, their presence in the Roman baths of Leptis Magna in Tripolitania is reduced to the level of some sherds (Bonifay  – Capelli 2013, Tab. II). 42 Belotti 2008b, 455; Auriemma et al. 2012. 43 Belotti 2008b, 456; Auriemma et al. 2016, 393. 44 Auriemma 2007, 150; Cipriano – Mazzochin 2011; Mazzochin 2013, 75. 45 In Aquileia, the Dressel 24 are well-documented too (Auriemma 2007, 142-145; Auriemma et al. 2012, 172; Auriemma et al. 2016), as well as in Oderzo (Cipriano – Ferrarini 2001). We should also mention again the presence of two early Dressel 24 types within the cargo of the first century Cape Glavat shipwreck (Radić – Jurišić 1993, 125). For a complete view of the diffusion of this form in the region: Belotti 2008b. 46 Their own presence in the Testaccio, as an olive oil amphorae deposit, as well as some tituli picti found there and mentioning oil as the content (Carreras Monfort 1999, 97-98; Remesal Rodríguez – García Sánchez 2007, 173-182), support this hypothesis. 40 41

18.03.19 14:39

CHAPTER 6 – Eastern Mediterranean amphorae (González Cesteros)

from Aquileia recently published by Auriemma and other colleagues,47 indicates liquamen ma… as the content.48 Even if an occasional export of fish products in Dressel 24 cannot be rejected, and some scholars have argued for this use based on both the typological division between Knossos 18 and Knossos 15 and some tituli picti mentioning supposed fish sauce content,49 we believe a secondary use of this particular piece is a more plausible option. The use of Latin and not Greek in the titulus, and a comparison with the amphorae from the Grado shipwreck,50 where wine amphorae were reused and contained a fish product, points in this direction. We agree with Belotti51 and believe we must assume that olive oil was the content of most of the Dressel 24 exported to the northern Adriatic, even if some exceptions existed. In this volume and in other publications, the theories that linked the abrupt decline of the Istrian amphora production and export with a tragic cessation of the Istrian olive oil production have been set out.52 In the pages dedicated to the Hispanic imports, we associate the arrival of Dressel 20 during the Antonine period with the halting of the massive Istrian olive oil production. Nevertheless, we have not documented the presence of Dressel 24 and related forms in Istrian contexts outside the Trieste area, neither is it the case at the Castrum and in the other excavations on the Brijuni islands. But equally no fragments of Cyrenaic MRA 8, also a possible olive oil container that arrived in the northern Adriatic region during the mid-second and mid-third centuries have turned up. We must assume that within a few years after the cessation of the Dressel 6B production the Istrian olive oil was again able to cover local demand. Thus it is probably to be expected that only small amounts of foreign olive oil amphorae were imported, and that, too, into the most important settlements, where indeed some Dressel 20 have

Auriemma et al. 2016, 392-393. Auriemma et al. 2016 published the titulus as liquamen mattum, which seems to us a little bit strange, since mattum means humid or impregnated and liquamen is already a liquid product. It could be a kind of mixed liquamen, impregnated with oil, wine or any other liquid. Another possibility is to read maturum and not mattum, being related to the quality of the fish product, but this does not seem exactly what is written according to the little epigraphical evidence existing on this point. However, we have not seen the piece and must believe in the author’s reading. We would like to thank E. García Vargas for comments on this subject. 49 Auriemma et al. 2016, 392-393; Dyczek 2008, 517-518, based on tituli picti, proposed a sporadic use of Dressel 24 for other contents. 50 Auriemma 2000. 51 Belotti 2008. 52 Bezeczky 1998a, 10-11.

been documented,53 probably to cover the above-average demand such larger places generated. On the other hand, the Istrian olive oil production must have been unable to supply the ancient markets associated with Dressel 6B anymore, neither the nearest, such as Trieste, nor the important cities of Po valley, even though, in the recent past, the small containers that continue the Istrian production during the second part of the second century and the beginning of the third century A.D. have appeared in small quantities in several places in northern Italy.54 3

LATE ROMAN AND EARLY BYZANTINE PERIOD

The circulation of these above amphora forms ceased in Late Antiquity, when new types appear in our region. As we have already seen, the presence of eastern forms is quite significant in early and middle Roman times and this situation continues during Late Antiquity, indeed it is now that they reach their peak, providing the majority of the overall amphora ensemble in several contexts. These imports arrived in huge numbers from the late fourth till the seventh centuries in the whole Adriatic region. Even so, during the fourth and early fifth centuries the high presence of African imports, quite important in places such as Trieste and its Slovenian hinterland,55 Aquileia56 or even in Ravenna,57 kept even quantities of oriental imports in second place, or sharing the first. It is from the second part of the fifth century till the seventh century or even later, due mainly to a massive presence of LRA 1B and LRA 4, when the productions from the eastern Mediterranean really take over, even penetrating into inland places in Friuli and Slovenia.58 In Late Antiquity, we discover almost every kind of eastern Mediterranean amphora in the Adriatic,59 but there are

See González Cesteros – Berni Millet in this volume. Auriemma 2007, 129-130. We must keep in mind the limited capacity of these containers. Archaeologists are not always in agreement about the content of such forms; even if they continue the typological line marked by the Dressel 6B, some tituli picti seem to refer to fish products. 55 Auriemma  – Quiri 2004; Auriemma 2007; Vidrih Perko  – Župančić 2005; Vidrih Perko 2006. 56 Auriemma 2007, 152; Villa 2013. 57 Augenti – Cirelli 2010. 58 Villa 1998; Vidrih Perko 2006; Auriemma – Quiri 2007. 59 If we compare the relatively good numbers of Cretan amphora types that arrived during the Roman period, their near absence now is the only remarkable change apropos the Oriental regions exporting their products to the northern Adriatic. It must probably be explained as part of the economic and commercial changes of Late Antiquity,

47

53

48

54

BEZECZKY_Text.indd 59

59

18.03.19 14:39

60

THE AMPHORAE

also changes in the main oriental suppliers to the northern Adriatic consumers. Cilicia-Cyprus and the Levant have by now already displaced the Aegean basin as the bulk importers. The colossal import numbers of LRA 1A and the even bigger quantities of the later 1B version,60 followed by different versions of the Palestinian LRA 4, demonstrate the special link that existed between the Adriatic and these two regions. This is further underlined by the fact that these two amphora types in all their different variants have been found in almost every place in our region.61 The LRA 1 seems to have been the only Cilician-Cypriot import in our area in Late Antiquity, but from the Levantine coastal region, together with the Gaza containers, other kinds of amphorae produced in today’s Israel and Lebanon also arrived. In some coastal places, with contexts dated in the fourth and fifth centuries, above all in Ravenna and Aquileia, the form Agora M 334 is present in limited quantities,62 alongside other types linked with the Carrot amphorae tradition. These containers must have been produced in Phoenicia, the M 334 above all in the Akko region.63 Maybe some LRA 4 had arrived earlier, but their presence seems to be important only from the fifth century onward. One must observe the high quantities of LRA 4 found in Aquileia64 and above all in Ravenna.65 The relevance of Ravenna as the Imperial capital and the main civilian, religious and administrative centre from the early fifth to the eighth centuries is well known: it makes Ravenna an advantageous point for the study of the commercial trends that took place in these centuries. Excavations carried on in the last decades in Classe, the main harbour of Ravenna, only a few kilometres from the city, brought to light huge quantities of pottery from this period.66 In some late fifth and sixth century contexts in Classe, large numbers of LRA 5/6 amphowhen the distribution of Cretan amphorae is mostly limited to the eastern regions. 60 In Aquileia Auriemma and Quiri proposed some years ago a much bigger presence of the earlier LRA 1A than the later LRA 1B (Auriema – Quiri 2007, 38), and this view has been reinforced by the latest excavations in the city (Dobreva – Riccato 2015, 115117). The cause for this situation must be linked to the decline in importance of this city during the late fifth and sixth centuries: newly published material underlines the decline of imports from the midlate fifth century, even if eastern Mediterranean and African amphorae do continue to arrive (Ceazzi – Del Brusco 2014). 61 Auriemma – Quiri 2007. 62 Stoppioni Piccoli 1983, 132-133; Cirelli – Cavanicci 2014; Auriemma 2007, 147; Dovreba – Riccato 2015, 121. 63 Reynolds 2005. 64 Dovreba – Riccato 2015, 119. 65 Augenti – Cirelli 2010. 66 Augenti – Cirelli 2010; Cirelli 2014.

BEZECZKY_Text.indd 60

rae have been discovered,67 but in other contexts in Ravenna and in other places in the northern Adriatic they seem to be just sporadic imports, with a few being documented in Aquileia,68 maybe in Trieste,69 certainly in its surrounding region (as for example in Koper),70 and more thinly in Friuli71 and in the military sites of Slovenia.72 Although the Levant and Cilicia-Cyprus were the main oriental export regions in the first phases of Late Antiquity, by the fourth and early fifth centuries we can also see the presence of different Aegean forms in the northern Adriatic. Even if already present on the Dalmatian coast, a complete piece of the form Agora M 235/6 deserves attention: it was found in the Wall depot of Salona, whose construction dates from the Theodosian period of government,73 but unfortunately no more pieces of this form have turned up in the northern areas,74 with the exception of one exemplar found in Brijuni. A possible Cretan origin has been proposed for this form,75 something that seems possible to us. But our personal experience in Ephesus, where the form is present within the harbour necropolis finds76 in different fabrics, leads us to hypothesise that it is still early to support this sole Cretan origin, but rather that the possibility of more than one production site must also be considered.77 From the late fourth to the end of the sixth centuries, the most widespread Aegean containers were the small LRA 3 produced in the Cayster valley and nearby areas.78 The diffusion of the different versions of these small amphorae reached its peak in the fifth and early sixth centuries, being retrieved in almost every place in the Mediterranean basin and even outside this region. The northern Adriatic is not an Augenti – Cirelli 2010, 609. Vezár-Bass 1994, 411; Auriemma – Quiri 2007, 47; Ceazzi – Del Brusco 2014; Dovreba  – Riccato 2015, 119-121; Dovreba – Riccato 2015, 119-121. As mentioned before, an early version of baggy-jar amphorae dated in the third-fourth century was found in a well in the Forum of Aquilea (Degrassi – Maggi 2011). 69 Auriemma 2007, 150. 70 Cunja 1996. 71 Auriemma – Quiri 2007, 47. 72 Vidrih Perko 2006, 216; Auriemma – Quiri 2007, 47. 73 Cambi 1989, 331-335; Mardešić 1994, 294-295. 74 These amphorae are also present in some southern Adriatic contexts, as in Puglia (Volpe et al. 2010, 645-646, Figs. 5, 3-4), and it is possible they have been mixed up with other types, mainly LRA 2, in other places. 75 Bonifay – Capelli 2013, 74-75. 76 González Cesteros – Yilmaz forthcoming. 77 M. Bonifay (personal communication) mentions a possible production site in the Peloponnese. This possibility must also be kept in mind, and we hope future discoveries will resolve the geographical origin of this form. 78 Bezeczky 2013a.  67 68

18.03.19 14:39

CHAPTER 6 – Eastern Mediterranean amphorae (González Cesteros)

exception, and they are commonly found there, which agrees with the already mentioned relatively common presence of their “one-handled” predecessors (vide supra). Together with the LRA 3 in the sixth century, another type produced in the Ephesian hinterland arrived in a few places in the northern Adriatic. The so-called Ephesus 56 is to be found in Trieste, where it has been confused with the LRA 3,79 and we have also documented several exemplars within our material in Brijuni. Their absence in contexts in Ravenna is strange, but we have to admit that they could have been muddled up with the more common LRA 3. Similar fabrics are also seen in some Agora M 273 amphorae, and mostly with the latest versions of this type named the Samos Cistern amphorae. These two connected forms are represented in several places in our region, but always in small quantities, especially so for the earlier M 273.80 It seems that the Samos Cistern form achieved more success within the Adriatic,81 where it can be found even in some inland sites in Friuli,82 Trieste and its Slovenian hinterland83 and in Classe.84 Another similar or related type to the M 273 is the Adamsheck 22/Torone VII85, found only in a sporadic way in our region.86 It seems to be produced above all during the fifth century; the exact production place is still an open question. It has been proposed that it might be the island of Icaria, west of Samos, on the basis of some tituli picti.87 Even if the evidence is not so clear, we have to keep in mind this hypothesis. Meanwhile a match for the characteristic calcareous fabric of these amphorae (with cream or whitish colour, low quantities of mica and of small red inclusions, together with occasional small rounded quartz grains) must be sought both in future discoveries on this island, and elsewhere in the eastern Aegean. Other production places cannot be ruled out as some variety in the fabrics exists, and also the massive

61

production that the form attained in the fifth century might be more than a small island could achieve.88 The presence of LRA 2 in the region is one of marked contrasts. The arrival of this form maintains the level of the relatively well-represented Dressel 24 variants of the late second and third centuries. However their numbers in some main centres of the region, such as Ravenna89 or Aquileia, are not particularly elevated,90 and only in Trieste does their number reach a higher volume.91 Nevertheless they seem to be more concentrated on the Slovenian route to Ljubljana,92 in the military places of Friuli93 or in Istria, where they have been found in Koper94 in similar quantities as in Trieste. Certainly they are present in good numbers in our Castrum, where they form one of the most important amphora types within the Late Antique material. Even if in some cases it is possible that some seventh and eighth century globular amphorae have unfortunately been classified together with the LRA 2 in Slovenia, Istria or Friuli, the quantities of LRA 2 that arrived from the late fourth to the late sixth centuries A.D. are really momentous. It has been suggested, and seemingly favourably within the scientific community, that the LRA 2 were containers directly linked with the supply of the military, on the basis of their extensive presence on the lower Danube border.95 The same can certainly be claimed for the cases of the early Byzantine military sites in the Adriatic region too: the big quantity of LRA 2 found in the Brijuni Castrum also supports this idea and, in our personal opinion, indicates that the Castrum was used as a military base during the early Byzantine period, even if the number of soldiers based there was not too high96. We are not sure if the LRA 2 were used as olive oil amphorae, but the fact that they developed from the latest versions of Dressel 24 and their rounded/globular size could indicate this function. The huge amount of this type found in the excavations of the Harbour Necropolis in Ephesus (González Cesteros  – Yilmaz forthcoming), and the intended future analysis on numbers of them, will be very useful in resolving this question. 89 Cirelli 2014. 90 Auriemma – Quiri 2007. This interpretation could change in the future for Aquileia, as recent excavations on the via Bolonia brought up an important number of LRA 2, in comparable numbers with other eastern Mediterranean amphorae (Ceazzi  – Del Brusco 2014). 91 Auriemma 2007; Ventura et al. 2014. 92 Vidrih Perko 2006. 93 Villa 1998. 94 Cunja 1996. 95 Karagiourgou 2001. 96 It is quite possible that it developed as a military fortress and should have been part of the Claustra Alpium Iuliarum: Mlakar 1975-76; Begović – Schrunk 2009; Begović – Schrunk 2011a. 88

See LR3 and Ephesus 56 in Bezeczky 2013a, 164-170; Degrassi et al. 2007, Fig. 1.8. Even if the spike of this exemplar is not of the usual sort, we must admit that this complete piece is an Ephesus 56 and not a LRA 3. 80 Vezár-Bass 1994, Pl. 71; Auriemma 2007, 145-146; Auriemma – Quiri 2007, 44. 81 Auriemma – Quiri 2007, 44-45. 82 Arthur 1990; Villa 1998. 83 Degrassi et al. 2010; Vidrih Perko – Župančić 2005. 84 Stoppioni Piccoli 1983; Augenti – Cirelli 2010, 609. 85 Adamsheck 1979; Papadopoulos 1991. 86 Vidrih Perko – Župančić 2005; Cirelli 2014, Fig. 5. 87 Arthur 1998. 79

BEZECZKY_Text.indd 61

18.03.19 14:39

62

THE AMPHORAE

Their presence in our region, as well as of some large cylindrical African containers, can also be helpful in determining the nature of the challenge that the local olive oil production faced in external imports during Late Antiquity. From the fourth century, the amphorae from the Black Sea increased their production levels, especially as the region of Sinope started to convey more amphorae to its customers, mainly in the Black Sea and eastern Mediterranean. The arrival of these products in the northern Adriatic region seems to be virtually at the symbolic level only. We have already mentioned the infrequent presence of Black Sea products during the early and middle Roman periods, and even if we can find some more exemplars from the fourth to the sixth centuries,97 they never attain any real consequence in the total number of pieces at any archaeological context.98 Their absence, above all in Ravenna, is quite significant. It typifies the almost complete absence of these amphorae in the western and central Mediterranean, the Adriatic being no exception. Finally, in some Adriatic sites the presence of a small number of Egyptian amphorae is documented. They form quite a rare group in our region, appearing almost as samplers, even in big commercial places such as Ravenna, where some LRA 7 have been found.99 They have also been retrieved in other places such as Ljubljana,100Aquileia,101 Trieste,102 in Friuli and in some points in the Venetian lagoon,103 but always in really small quantities. Other Egyptian forms of Late Antiquity have not been documented, except for some fragments of AET 3/Kellia 172 found in Aquileia.104 The effective absence of amphorae from Crete within the Late Antique imports also needs pointing out. Even if it is possible that some Cretan types have been misidentified, we agree that the Cretan exports seem to have a most limited geographical diffusion from the fourth century A.D.  The only documented exemplars are a rim fragment of a possible

Belotti 2008a. In the literature we have been using, we have only documented a possible C-Sinop III in the upper layers of the excavations on the northern side of the river port in Aquileia (Maggi – Urban 2001, Fig. 6). However, we cannot affirm it is really a Black Sea amphora and the authors did not mention anything about this piece in particular. 99 Stoppioni Piccoli 1983; Augenti – Cirelli 2010. 100 Vidrih Perko 2006. 101 Vezár-Bass 1994; Auriemma – Quiri 2007, 48; Ceazzi – Del Brusco 2014. 102 Auriemma 2007, 151; Degrassi et al. 2010, 580. 103 Auriemma – Quiri 2007, 48. 104 Vezár-Bass 1994, Pl. 64 AO 24; Auriemma – Quiri 2007, 48.

TRC 2105 found in Trieste,106 together with our piece in Brijuni.107 Before leaving the amphora forms found in the northern Adriatic, attention must be paid to two types whose ascription to a production place in the eastern Mediterranean is problematic. We are talking about the San Lorenzo 7, found in several northern Adriatic and Italian settlements, but always in small quantities,108 and the so-called “Vigna Barberini 13”,109 an amphora present in late sixth century contexts in Ravenna.110 The San Lorenzo 7 turns up in the central and eastern Mediterranean terrain; it is presumed to have been produced on the Aegean from the second half of the second to the late fourth or early fifth centuries.111 We are in broad agreement with this postulation, that seems reinforced by some analyses made in Milan,112 but due to the range of fabrics and the extensive diffusion of this type in Italian contexts, we prefer to leave the question open for future studies. The second and more problematic type is the so-called “Vigna Barberini 13” found only in Ravenna within the Adriatic region. Cirelli has proposed an oriental origin for this form, based on the abundant mica in the fabric and a certain distribution weighted towards the eastern Mediterranean. It is true that the type has resemblances with some Late Antique Cretan amphorae, mainly with form TRC 1 of Portale and Romeo,113 but we think it has been confused with the Cretan amphorae. For us it is a different form, much closer to the south Italian and Sicilian Late Antique flat-bottomed amphorae, such as, for example, the latest forms of Keay 52 or the Crypta Balbi types. The fabrics of the Crypta Balbi 2 amphorae are in fact quite similar to that commonly described for the “Vigna Barberini 13”, namely hard and of red or brown-red colour and with abundant mica. The petrological analysis carried out on some pieces found in Rome supports the macroscopic visual identification.114 However, the latest analysis of “Vigna Barberini 13” pieces found around the Palatine hill shows how difficult it is to determine a concrete production region, as both the Aegean

97 98

BEZECZKY_Text.indd 62

Portale – Romeo 2000. Degrassi et al. 2010, 580. 107 Vide infra. 108 Auriemma – Quiri 2007, 52; Dobreva 2012, 104-105. 109 Rizzo et al. 2004. 110 Augenti – Cirelli 2010, 609; Cirelli 2014, 543. 111 Arthur 1998, 171-172; Mongardi 2016. 112 Dobreva 2012, 104 on Lusuardi Sienna 1990, 384. 113 Portale – Romeo 2000, 422. 114 Pacetti 1998, 204; Casalini – Crespi 2010.

105 106

18.03.19 14:39

CHAPTER 6 – Eastern Mediterranean amphorae (González Cesteros)

and southern Italy remain possibilities.115 Lastly, the actual geographic distribution of our type is almost constrained to Italy, being found above all in Rome and Ravenna, and in Kellia within the eastern Mediterranean.116 In the Aegean area we have recently documented a piece coming from the ancient harbour excavations in Ephesos.117 Though but a footnote to the diffusion of Late Antique eastern Mediterranean amphorae in the northern Adriatic region, the above must be seen as part of a common trend developing from the late fourth century A.D. that affected the whole Mediterranean basin.118 Likewise, we believe that the large quantities of amphorae from these regions found in the Adriatic signify the continuous relations between our region and the eastern Mediterranean. The continuity of the same routes that connected the Adriatic and the Aegean from the Bronze Age on must surely be admitted, as, too, the increased volume of traffic from the fourth century A.D. It is beyond doubt that the rise of Ravenna as the most important urban centre in the whole region from the late fourth century and above all from the fifth to seventh centuries brought into being a new commercial infrastructure, which finds its new main redeployment centre in the port of Classe. Excavations carried on during the last decade have highlighted the spectacular number of warehouses and archaeological materials found in Classe, which can only be understood by its having developed the role of an important redistribution centre that probably supplied other main points in Cisalpina, Venetia and Istria. It is also important at this point to underline the special character of Ravenna, now being the centre of the Eparch and court, as well as one of the most important religious settlements with a powerful bishop. Even if well-represented along the Adriatic anyway, the numbers and percentage of LRA 4 and of LRA 3 in Ravenna contrast with other places in the northern Adriatic, among them Brijuni, where the number of LRA 1 is over ten times that of LRA 3 and LRA 4, with the LRA 2 taking second place in the oriental amphorae represented. Places close to Ravenna, such as Rimini, seem to present a similar but less dramatic version of this pattern: here the LRA 1 is still the most common type,119 or at least is present in a higher per Casalini – Crespi 2010, 104. Egloff 1977, type 171. 117 For us the form found in Thessaloniki by Papanikola-Bakirtzi 2010 and identified by Cirelli 2014, 543 as the same form they have in Ravenna is, in fact, something different with a rounded base and different rim: it must be linked with the Cretan production, as with forms TRC 1 or TRC 2 of Portale – Romeo 2000, 422. 118 Panella 1989; Reynolds 1993; Remollà 2000; Pieri 2005. 119 Negrelli 2008.

63

centage than in Ravenna itself. Another important centre in the region in the fifth and sixth centuries looks to have been Trieste, a main port from where products must have been redistributed to its Slovenian hinterland and the western Istrian coastal settlements, for example Brijuni.120 This special link between our region and the eastern Mediterranean territories continues beyond the chronological frame being discussed here: we can still observe the arrival of oriental products, again amphorae, to main centres such as the Venetian lagoon or Comacchio from the late seventh to the ninth or even tenth centuries.121 They are also found in Croatian waters even later still.122 THE EASTERN MEDITERRANEAN AMPHORAE TYPES FROM THE BRIJUNI CASTRUM VILLA Considering the broad diffusion of the eastern Mediterranean amphorae in the Adriatic from the Hellenistic period till Late Antiquity, it is not surprising that a good quantity of oriental amphorae have turned up within the ceramic material of the Brijuni Castrum.123 It is obvious that the Castrum follows the trends discussed for the complete northern Adriatic region, reminiscent of places such as Aquileia, Trieste or Ravenna in the late Roman and early Byzantine periods, but it also has some marked peculiarities of its own. Here it should be observed that the main eastern Mediterranean types are the most common of the amphorae found in Castrum, such as the Rhodian, Koan and Knidian during the Augustan and Julio-Claudian period, or the LRA 1 in Late Antiquity. It is worthwhile, though, to separate the oriental forms assignable to the early and middle Roman period from the Late Antique imports, because, even if during the Roman period their presence is greater than other kinds of import from outside the Adriatic region, it is in Late Antiquity that they really outstrip their rivals on the island, even the African imports. This feature has already been mentioned by Vidrih Perko and Pavletić.124 The phenomenon, without any doubt, occurs because the biggest quantities of amphorae in the Castrum belong to the late Roman and Byzantine occupational phases, a time when

115 116

BEZECZKY_Text.indd 63

We have already pointed out the value of a precise study of the amphorae from the Roman and Byzantine periods in Pula: comparing it with other main places along the northern Adriatic, above all with Trieste and Ravenna, would probably be most instructive. 121 Gelichi – Negrelli 2008. 122 Brusić 1976; Brusić 2010; Zmaić 2013. 123 Bezeczky et al. 2015. 124 Vidrih Perko – Pavletić 2000. 120

18.03.19 14:39

64

THE AMPHORAE

a military unit was probably accommodated in the Castrum and a religious community in the nearby church.125 The late Roman and Byzantine period is a time when the Adriatic territories do not seem to have exported their surpluses in amphorae, as opposed to Roman times when Istria was an amphora production region. Even if amphorae were not produced on the island (we have not yet found any remains of kilns and debris linked to pottery production), it is certain that olive oil and other commodities were produced there.126 On the other hand, amphora production is attested in the workshop of Fazana during early Imperial times, which lies just on the other side of the Brijuni channel and is one of the natural ports for the archipelago.127 Within the oriental material of the Castrum we find material from all kinds of provenances and products, with the greatest import of the Imperial period belonging to wine amphorae from the Aegean region, while in Late Antiquity it is now the wine from Cilicia-Cyprus, from the Aegean and Levant that represents the bulk of the imports. The absence of oriental amphorae linked with olive oil during the Roman period is noteworthy, though the LRA 2, from the eastern Mediterranean with its possible olive oil contents, is represented in the Castrum during the Late Antiquity. This pattern contrasts with that of other places in the region, mainly Trieste and Aquileia, where amphorae such as the Dressel 24 and related forms, or the MRA 8, also a possible olive oil container, are commonly encountered. Due to the stratigraphical problems inherited from the earlier unprofessional excavations that took place in the Brijuni Castrum,128 no clear distinction between the material of the early and later periods is feasible. One must rely directly on the typological distinctions for a better chronological understanding. One would, however, expect an increased presence of amphora imports during the latest occupational phase, as the so-called Castrum acquired a different role at this time, transforming it from a productive pars rustica of a Roman villa to a military fortress.129 The local historical development of Istria in Roman times made the whole region into an olive oil production and exporting area, with good established markets in northern Italy and central Europe;130 it exported many more amphorae and goods than it brought in in balancing or luxury products. This situa See Schobert in this volume. Bezeczky 1998a; and Bezeczky (Chapter 4) in this volume. 127 Gnirs 1910; Gnirs 1911b; Bulić – Koncani Uhač 2011. 128 Vidrih Perko – Pavletić 2000. 129 See Chapter 3.1. 130 Bezeczky 1998a; and Bezeczky (Chapter 4) in this volume.

tion, though, changed from the mid-second century A.D.: a higher amount of imported produce is to be expected from this time on, something that is confirmed for the Late Roman times. EARLY AND MIDDLE ROMAN PERIOD The material from the Castrum includes amphorae coming from the eastern Mediterranean, running from the late first century B.C.  to the early seventh century A.D.  As already mentioned, some of this attained a very high profile in the Late Roman and Byzantine phases, even though of minimal status beforehand; moreover, typical forms for the second and third centuries A.D. are still only found in quite small quantities. Eastern products, however, arrived here continuously during Roman Imperial times, from many different regions, although the most represented are the well-known Aegean Imperial productions. From the Augustan period, the arrival of eastern Mediterranean amphorae at the Castrum is observed. Among them, the most important types produced in the Aegean basin at this time are present, as they are in other places in the northern Adriatic. However, a key difference exists with elsewhere in northern Italy, Slovenia and Croatia: namely the “elevated” number of Koan and Knidian amphorae compared to the Rhodian,131 a pattern completely different from that seen in other places, above all in the Cisalpina. Maybe this difference is only due to chance, but equally the dynamic of the Castrum and the Brijuni islands may not have been exactly the same as in other places. The presence of a small Augustan “one-handled jar” of form Agora F 65/66 within our material has been documented. Writers of early Imperial times mention the quality of the wine produced in Ephesus and nearby places,132 where this vase form was normally produced, but in our exemplar the fabric seems not to be Ephesian but rather from some other place in Asia Minor or the nearby islands, where these containers and their later versions with two handles were also produced. The absence of other oriental amphorae linked with high quality wines for this period, above all of Chian amphorae, argues against any luxury imports in the

125 126

BEZECZKY_Text.indd 64

We have to remember that the typology of some of the Rhodian amphorae found in Brijuni places them in the second half of the first century A.D.  132 Ladstätter 2008, 181-182; Bezeczky 2013a, 25. 131

18.03.19 14:39

CHAPTER 6 – Eastern Mediterranean amphorae (González Cesteros)

Castrum,133 though maybe they are located in other places on the island, like the luxury villa of Val Catena. The later versions of Koan amphorae, frequently known as Dressel 5, are present in Brijuni, pointing to the continuous arrival of these wines during the end of the first and the second centuries A.D. However, not all the fabrics found in these pieces can be assigned to the island of Kos – a broad Aegean and even Pamphilian and Cilician production is to be expected. Equally, some of the Rhodian fragments seem to be late versions, probably produced during the second part of the first century A.D., as, for example, the catalogue entries (58 and 59). Other products from the Aegean and nearby areas are present too. A few fragments of AC 4, a typical Cretan form based on the Rhodian types of the first and early second centuries,134 have been found in Brijuni. It is not always easy to distinguish this kind of amphora from the Rhodian Imperial amphorae, especially in cases of big modules. The handle fragment (69) corresponds to a small module with a really prominent horn. It can be classified without any doubt among the small-size AC4 produced in the late first century A.D.  The arrival of other oriental containers, for example some “one-handled” vessels from Ephesus (forms M 125 and M 126), and also the handle fragment (70) of Agora G 199 that presents a rough, soft, probably Cilician, fabric, can also be dated to the late first and second centuries A.D. These forms were produced over a longer period: they are normally to be found in second and early third centuries contexts too. Now, Cretan products continued to arrive at Brijuni, but just sporadically: a couple of fragments of form MRC 3 of Portale and Romeo (103). But it is again the “one-handled” Asian container that constitutes the most important import of the late second and third centuries. Some different versions of these small containers, above all the form Agora M 240 with the typical Ephesian fabric, are known. The presence of an Agora M 279 similis with another kind of fabric without mica requires comment: this must be from a manufacturing locale that excludes big areas of Asia Minor, such as Ephesus and the Meander valley. A base (78) of type Kapitän 1 and some rims and handle fragments of Kapitän 2 were recovered. The fabric of Based on the epigraphic evidence of Delos in the second century B.C., Tchernia 1986, 105, mentions that the Knidian wines were four times as expensive as the Koan. We cannot be sure that the wines of Knidos were high end wines, but it must be admitted that they were not usual wines. 134 Marangou 1995, 84. 133

BEZECZKY_Text.indd 65

65

the Kapitän 1 seems to be of northern Asia Minor origin, or from the nearby islands like Chios or Lesbos. It is red in colour, medium-hard and laminated with small quantities of mica and other small stone inclusions. On the other hand, the Kapitän 2 fragments are of two different fabrics, one of them is the most typical, that is brick red in colour, hard and coarse with big grey and white inclusions, while the second is smoother and finer, with temper smaller-sized and lower in quantity (79). As already considered, almost any place in the northern Adriatic region has yielded big quantities of these two forms. While the situation is more or less normal for the Kapitän 1, in case of the Kapitän 2 it must be regarded as an exception within the broad dispersal of this type seen in most of the Mediterranean regions. LATE ROMAN AND EARLY BYZANTINE PERIOD It is in this period, between the fourth and early seventh centuries A.D., when the products coming from the eastern part of the Mediterranean reached their zenith, that they and the African imports constituted the bulk of the amphorae found in the Castrum. These significantly increased quantities of oriental amphorae must find their explanation in the presence of soldiers and the church personnel in Brijuni. Such quantities of African and oriental amphorae are usually believed to reflect the Byzantine occupation of the central and western Mediterranean carried out by Justinian during the central decades of the sixth century. In our opinion, this is an extremely simplistic explanation of a broader and more complex economic process that requires a more nuanced and multiple approach: not only will the Byzantine presence in Italy or Africa have played a vital role, but also the growth in output of northern African produce and that in most of the eastern Mediterranean territories,135 not to mention the close commercial links that were always practicable between the Aegean and the Adriatic. One important change that took place apropos the Late Antique eastern types is the replacement at that time of the Aegean basin by the Cilician and Cypriot areas as the most important source regions supplying amphora products to the Castrum. It is the same process already remarked upon and that involved the whole northern Adriatic area, where the Cilician and Levantine products always hold premier place among the imports of this phase. The sheer volume of LRA 1 pieces makes the relevance of their imports for the

135

Bonifay 2004; Pieri 2005.

18.03.19 14:39

66

THE AMPHORAE

Adriatic region136 starkly obvious here, just as it does for the whole Mediterranean area and beyond. The LRA 1 in fact achieve a much higher percentage in Brijuni than in any other place studied in our region, accounting for more than 20 % of the total amphorae. This percentage is similar to that determined by Vidrih Perko and Pavletić in their preliminary study.137 The most probable explanation exists, once again, in the unique nature of this settlement during Late Antiquity.138 Most of the LRA 1 belong to the typical forms of the late fifth and sixth centuries and should be placed within the form LRA 1B of Pieri.139 This in itself is nothing new, because, firstly most of the amphorae imports are placed in the late fourth to early seventh centuries and, secondly, form 1B was the variant of the LRA 1 that got exported most. The fabric of these pieces is typical of material from the Cilician and Cypriot areas, presenting different colours, but being hard or medium-hard, sandy or coarse with a number of white and yellowish limestone pieces, as well as other grey and sometimes red inclusions. There is one amphora rim and handle (85) which could come from the Soli workshop (Turkey), with its calcareous fabric.140 The LRA 2 hold a quite elevated position within the Castrum material (87-88), apparently on a scale greater even than in Trieste or Koper. The argued link between this particular form and the army supply as documented on the lower Danube border141 could also apply for the Castrum and the northwestern Adriatic territory. The only problem we find with this hypothesis is that of the presumed content of these containers. If it was olive oil, which probably it was at least in part, then it raises the question of what was happening to the local Istrian production at this time. Was Istria not able even to supply the soldiers and military staff settled on its borders anymore? There is no easy answer to this question. The remains of three Late Antique oil-presses inside the Castrum,142 be they used for oil, wine or both, do not help us in coming to a better understanding of the Cirelli 2007, 47. Vidrih Perko – Pavletić 2000. 138 It has been suggested that there is a connection between some amphora types such as LRA 2 (Karagiourgou 2001) or LRA 13 (Diamanti 2010a; Diamanti 2010b) and the needs of the early Byzantine State, as is expressed in the annona, whether military or civilian. It could account for the elevated number of LRA 1, but the question is difficult to answer since this is the case wherever one looks in the Adriatic. 139 Pieri 2005, 75-77. 140 Autret et al. 2010, 203-207. 141 Karagiorgou 2001. 142 Matijašić 1993a, 249-251.

elevated quantities of amphora imports from this period. Most of the LRA 2 pieces belong to the latest variant documented in the late fifth and sixth centuries: they have taller necks and handles, not really rounded bodies and have wavy incised lines in the upper part of the body walls. Nevertheless a good number of pieces from the earlier versions of the late fourth and early fifth centuries were recovered too. We also have to point out that within our material not a single globular amphora from the late seventh century and later has turned up, although we do have the presence of an LRA 13 with an Imperial stamp (100) and some late African forms documented. The fabric of most of the LRA 2 fragments is hard, a coarse reddish-brown on the external surface and more yellowish on the inner face. It presents some white limestone inclusions, sometimes large, and also some voids of different sizes. The source of this fabric must be placed in the Argolid and Boeotia, where some kiln sites have already been found.143 However, some pieces with a finer fabric occur, with small or medium white, red and occasional grey inclusions: these probably come from Chios or the nearby territory in Asia Minor.144 A special piece is no. 104. It is an almost complete rim fragment, with part of the neck and a handle, of the form Agora M 235/6, a type that recalls the LRA 2 in shape. We have already mentioned the small presence of this fourth and early fifth century form in the Adriatic: the piece at Brijuni seems almost exceptional. The fabric of this exemplar is light red, hard and fine, sandy, with some red, grey and white quartz inclusions. We are not able to affirm the provenance and the continent of this amphora, but we propose an origin in Crete or on the Greek mainland.145 Within the same family of containers belongs a neck fragment with an incomplete stamp of a Byzantine Emperor (100), for which the best-known parallels are similar stamps found in Alexandria,146 Istanbul Saraçhane and Yenikapı,147 Athens,148 Geronissos island (Cyprus),149 Kellia,150 Tocra151 and the island of Kos,152 where this amphora form was produced.153 With the publication of the discoveries from the

136 137

BEZECZKY_Text.indd 66

Rudolph 1979; Gerousi 2014. Opait – Tsaravopoulos 2011; Arthur 1990, 286. 145 Vide supra. 146 Sztetyłło – Borkowsky 1986. 147 Hayes 1992; Kara 2015. 148 Grace 1949; Opait – Diamanti 2014. 149 Connely – Wilson 2002, 274-275. 150 Bonnet – Cattin 1999, 538-539. 151 Bentaher 1994, 231. 152 Diamanti 2010a; Diamanti 2010b; Diamanti 2012. 153 Also produced in Paros: Diamanti 2016. 143 144

18.03.19 14:39

CHAPTER 6 – Eastern Mediterranean amphorae (González Cesteros)

Halasarna workshop on the eastern coast of Kos, the number of stamped pieces documented has significantly increased. Our stamp presents the bust of an emperor holding a sceptre in the left hand and probably the item of clothing called a mappa in the right hand. Around the emperor’s bust a partially preserved inscription is to be seen. We can distinguish some Greek capital letters in the lower right part of the stamp and maybe in the upper left too, but in this last position they have been effectively erased. These letters can be read as: (--) ΟΛ(------) and must correspond with ΠΤΟΛΑΙΜΕΟΥ. The best parallel examples are two stamps from Athens, recently published by Opait and Diamanti.154 The Athenian exemplars preserve more of the stamps and were stamped on the neck too. The authors mention the similarities with others stamps found in Kos and date them to the last decades of the sixth century,155 before the reign of Phocas, who appears on stamps too, but is always represented with a large beard, which is apparently not the case in our piece. This piece, classified as an LRA 13, a term not without controversy,156 seems to correspond with a product of the well-known kiln site of Halasarna (modern Kardamaina). Both the epigraphic evidence and the visual characteristics of the fabric (red-brown, hard, fine and micaceous, with dark grey and quartz inclusions) correspond to the description given by Diamanti about the production of this and other kinds of amphorae in the Halasarna workshop157 much more closely than to items produced in other Late Antique workshops in Kos described by Didioumi.158 In the northern Adriatic region, in excavations in Byzantine Comacchio, a seventh or early eighth century globular amphora produced in some Aegean workshop was recovered: it bears a circular stamp with two letters on the neck.159 Although the stamp on this piece probably serves the same purpose as that mentioned above, we think the different

Opait – Diamanti 2014, 55, Figs. 4 and 9. Opait – Diamanti 2014, 55. 156 The problem about this denomination comes from the Berenice publication. Riley argued that within this type there was a broader variation from the late sixth and seventh centuries, which drew its descent from both the sixth century LRA 1 and LRA 2. In some publications we find the term LRA 13 used for the late globular amphorae (Demestica 2005), as well as for smaller and thinner containers derived from the LRA 1 (Diamanti 2010). It is true that this late form seems to have taken some characteristics of both LRA 1 and LRA 2 types, but it would be more appropriate to make a clear division between the two types (we would like to thank Dr. Diamanti for the opportune conversation about these types during the Rei Cretariae Congress 2014). 157 Diamanti 2010a; Diamanti 2010b, 1. 158 Didioumi 2014. 159 Gelichi – Negrelli 2008, Fig. 11.1. 154 155

BEZECZKY_Text.indd 67

67

type of amphora and the different form of the stamp suggest a different production place. According to Diamanti, stamps of this kind are only to be found on amphorae produced in the late sixth and early seventh centuries. They represent some kind of state control linked with both production and commercial activities, probably under Imperial administration through the control of officers linked to the Eparch: perhaps the Kommerkiarioi, a group of public officers with different economic duties – for example being in charge of taxes, or controlling the imports and exports with other territories.160 From the late sixth to the early ninth centuries, the Kommerkiarioi had stamps with the bust or silhouette of the emperor (as is discerned in our fragmentary example),161 which they used to certify the good quality and guarantee the commercial trafficking of the stamped products.162 This fragment is a good example of the continuity of Aegean imports into Brijuni. Within the Aegean, a similar range of production centres and amphora types may be observed in early and middle Roman times, but now the most important places are no longer the islands of Kos, Rhodes and nearby coastal territories in Asia Minor. Despite the stamped amphora fragment already mentioned, a Dodecanesian provenance for any more Late Antique amphora fragments, as based on the visual characteristics of the fabrics, cannot be vouchsafed: even if some Samos Cistern pieces seem to have similar fabrics to some early Roman Koan amphorae.163 However, other territories in the eastern Aegean, such as the central coastal band of Asia Minor, especially the territory around Ephesus and the nearby islands, and also Chios and the continental territories close to it, all seem to be well-represented within the Castrum assemblage. The small wine containers produced in Ephesus and other places in western Asia Minor and on some nearby islands continued to arrive in the form of LRA 3 to Brijuni during Late Antiquity.164 Within our material are a good number of these small containers in their different versions and different sizes, but always with the typical dark brown or a reddish soapy and highly micaceous Ephesian fabric (89-90). They are the third oriental type, but far fewer in number than the LRA 1 and LRA 2. The relevance of the Ephesus hinterland

Diamanti 2010a; Opait – Diamanti 2014, 55. Probably some late sixth, or better early seventh-century emperor: Maurice (582-602), Phocas I (602-610), Heraclius (610-641). 162 Oikonomidès 1989; Diamanti 2010a. 163 Vide infra. 164 Imitations of LRA 3 and Ephesus 56 have recently been found on the island of Kos, in the kiln site of Mastichari (Didioumi 2014; Bezeczky 2013a, 167). 160 161

18.03.19 14:39

68

THE AMPHORAE

in the supply of the Castrum goes beyond the presence of the LRA 3 and is enhanced by a goodly quantity of that other important Late Antique Ephesian form, the so-called Ephesus 56 amphora (91-92).165 We spoke in the first part of this article of the remarkable absence of this form in the northern Adriatic region, being only documented by one complete exemplar from Trieste,166 where it has been wrongly classified within the LRA 3. Even if amphorae of this form have been misjudged and put in with the more diffused LRA 3, we still believe the relatively elevated number of Ephesus 56 found in Brijuni is a true exception – again to be connected with the possible military unit settled there and the ecclesiastical community of the church. From the eastern Aegean region another amphora form came to Castrum during late Roman and Byzantine times. The Agora M 273 produced during the fourth and fifth centuries A.D.  is well-represented here, as well as their latest version, the so-called Samos Cistern amphora, typical of the sixth and early seventh centuries.167 Within our material a high number of M 273 and related forms were counted, with different fabrics that suggest various production places within the Aegean area. Some of them show mica and are a dark brown-reddish colour, similar to the Ephesian production, but in other cases the mica is absent and the fabric is more calcareous, with reddish and yellowish colouring and some dark inclusions. The big formal variations in size and fabric, found mainly in the M 273 corpus, make it difficult to achieve a clear comprehension and distinction of all the different variants existing. Nevertheless some pieces must be categorised within a subdivision of the M 273 and described separately. This concerns the form identified for the first time by Adamsheck from some upper parts found in the Corinthian harbour of Kenchreai,168 and better documented some years later by Papadopoulos in the Chalcidian locale of Torone from an almost complete piece.169 For this reason the most appropriate name for this form, clearly connected with the M 273, ought to be “Adamsheck 22/ Torone VII”. As already mentioned, an Icarian origin has Bezeczky 2013a. Degrassi et al. 2007. 167 We are not in agreement with those scholars who have tended to extend the production of the Samos Cistern type till the second part of the seventh century A.D. For us this form, first documented in a mid-sixth-century context in the Heraion of Samos (Isler 1969), is mainly found in sixth- and early seventh-century contexts and their presence in some mid- and late seventh-century findspots could be partially residual. 168 Adamsheck 1979. 169 Papadopoulos 1991; Papadopoulos 2001, 566-568, Fig. 155. 165 166

BEZECZKY_Text.indd 68

been suggested for this subtype typical of fifth century contexts in the eastern and central Mediterranean. We can only confirm that within the Castrum material the small quantity of these amphorae all have the typical calcareous, white or yellowish fine fabric. Unlike other places in the northern Adriatic and as in the case of the relative frequencies of the LRA 1 and LRA 2, the Samos Cistern group is present in the Castrum at a lower level than their predecessors had been. Only a small quantity of fragments of this late form has been documented at Castrum. The fabric varies from one piece to another. Piece 101 has a fine and hard fabric, red at the core and light brown outside, with some small red and colourless inclusions, similar to some Koan amphorae. Even if the late Cretan amphorae did not make any impact on the supply of the northern Adriatic, a couple of pieces have been documented within the material of the Castrum. One rim must be related to form TRC 1 or TRC 2 of Portale and Romeo (57, 178 and 179). On the other hand, piece 180 looks to belong with the group named TRC 4, as presented by the two Italian scholars, and produced from the fifth to early seventh centuries.170 At present no parallels of this late form have been found in our region. The fabrics of these two pieces are quite different, even if they both present a regular quantity of small limestone inclusions. After Cilicia and the Aegean, the Levant is the other main oriental region that supplied amphorae to the Brijuni islands. Unlike most places in the northern Adriatic, almost all the amphorae documented from this region in this period are Palestinian LRA 4; types such as Agora M 334 or LRA 5/6 have not been found. Equally the presence of LRA 4 within our material is much more modest than in the main northern Adriatic places, above all Ravenna, where they form the most important amphora type.171 In Brijuni the number of LRA 4 is less than the LRA 3, and so they comprise the fourth most important oriental type of this period. Most of the LRA 4 belong to forms of the sixth or early seventh centuries, included in Peri’s group B; they present the typical Gaza-Palestinian fabric, but some pieces are finer, more reddish and with fewer inclusions (99).172 In the first part of this article, the minor relevance of other oriental production regions in the northern Adriatic was pointed out, particularly Egypt and the Black Sea, whose products arrived just occasionally in this area. In Bri Portale – Romeo 2000, 422. Cirelli 2007; Cirelli 2014. 172 Pieri 2005, 105-107. 170 171

18.03.19 14:39

CHAPTER 6 – Eastern Mediterranean amphorae (González Cesteros)

juni, some fragments of Egyptian amphorae can be inserted within the group of the LRA 7, namely the AE 7.1 group of Dixneuf, dated mainly between the end of the fourth to the mid-seventh centuries.173 Nevertheless it is possible that piece 105 could be an early version of the group AE 7.2, dated from the seventh to the end of the tenth centuries. It is really difficult to identify both subtypes of LRA 7 from just small fragments, since some parts of the amphorae, for example handles and spikes, were probably handmade.174 The alluvial fabric of our pieces makes it easy to determine the production place as within the Nile valley; they also present several voids made by the disappearance or burning out of organic components, mainly straw. The Late Antique products of the Black Sea had even less relevance than the Egyptian imports in our region, and normally they are not found in excavations. In the case of the Castrum there are some documented pieces that are undoubtedly connected with the Sinopean production. They belong to some amphorae of the so-called “Sinopean Carrot amphorae” group, recently classified as lying within the C-Sinope group as a result of the excavations of the workshop of Demirci near Sinope.175 The Castrum pieces belong to form C-Sinope II, with their different variants. The fabric of these pieces is not always homogeneous, but they do have some common features, mainly the use of black volcanic stone fragments as temper, typical of the Sinopean area. The presence of these fragments, even in an unimportant quantity, must be stressed, since they are not found in the north Adriatic contexts. Even if it is not easy to associate a product with every Late Antique oriental amphora with certainty, it seems quite clear that most of the imports from the eastern part of the Mediterranean that arrived at the Castrum were wines. Only the LRA 2 are possibly linked with an olive oil content. No oriental type found in Brijuni can be associated with fish products. Nevertheless, some amphora forms, for example the LRA 1 and the LRA 2, can transport different goods. CONCLUSION The staggering number of eastern Mediterranean amphorae and the variety of forms present in the Brijuni Castrum seem to be in concordance with the general pattern of the north-

Dixneuf 2011, 154-173. Dixneuf 2011, 155-156. 175 Kassab Tezgör 2010. 173 174

BEZECZKY_Text.indd 69

69

ern Adriatic, a region that maintains a special link with the eastern regions by the maritime routes of the Aegean. Nonetheless, some peculiarities of this material draw attention to themselves: probably they exist thanks to the special character of the site, situated in the estates of a very rich family that would later become Imperial property during the early and middle Roman period, and form a possible military and ecclesiastical base during Late Antiquity. Its nearest most important Istrian city is Pula, only around 10 kilometres from the island. The shipping of the amphorae supply from this important harbour, as well as from Fažana, is the most logical option. Nevertheless the singularity of Brijuni seemingly ensures the occurrence in the amphorae spectrum both of other forms and also of small variations in the respective percentages present, compared to other places in the region, such as, for example, Aquileia or Trieste. A complete publication of the material in Pula could help solve several questions, as well as providing a deeper appreciation of the social and economic development of the whole Istrian Peninsula. During early Imperial times the amphora imports from regions outside the Adriatic have no perceivable relevance at Brijuni. The existence of several presses on the island demonstrates the production of wine in this period and, we are sure, of olive oil too. The large number of Dressel 6B documented must be assigned to this foodstuff production. Within this trend, the oriental amphorae of early Imperial times are the most represented imports coming from outside the Adriatic region. This pattern looks to conform to the norm as seen in other places in the northern Adriatic. However, there is one main difference with places such as Aquileia, Trieste or the most important Cisalpine cities: at Brijuni, the number of Rhodian amphorae is outnumbered or at much the same level as the Koan and Knidian imports. The abrupt end of Istrian amphora exports during the Antonine period had an extraordinary impact on life at Brijuni: much less material was brought in now, due to the end of the Dressel 6B production and their gradual substitution by other types, mainly Fažana 1 and Fažana 2, that nonetheless never reached the massive volume of their predecessors. Even if it is true that deposits of the late second and third century layers in the Castrum are not so well represented as others, nonetheless the number of amphorae assignable to this period decreases proportionately. The oriental forms still count as one of the most represented overseas import groups: they make up a diverse group of eastern Mediterranean amphorae from different regions conveying a varied range of but one product. All the types documented are wine amphorae; forms linked with olive oil, like the Dressel 24

18.03.19 14:39

70

THE AMPHORAE

or the MRA 8, commonly found in some of the best excavated places in the northern Adriatic, do not appear at Brijuni. This fact, together with the arrival and maintenance of new Istrian amphora types that could be linked with olive oil, could indicate the limited revival of the Istrian olive oil production, some years after the abrupt interruption that probably took place during Hadrian’s reign.176 Within the oriental amphorae, most of the forms came from the Aegean basin, but from the end of the first century A.D. forms appear from other regions such as Crete or Cilicia. Nevertheless, the most represented types are the several versions of “one-handled jar”, a typical production from the Cayster valley and Ephesus region, but also produced in other workshops throughout the Asia Minor territory and nearby islands. The Late Antique material shows a completely different world-view for the Brijuni islands. The administrative, military and religious changes that took place in the whole Roman world from the late third or early fourth centuries A.D. also have their effect on the architectonical and material remains of the so-called Castrum. Even if it is possible that the Castrum still produced wine during this period and maybe olive oil too, as the existence of some presses from this period surely argues,177 it is undeniable that its function has been transformed from a productive villa rustica to the quarters of a military unit. Some hundred meters from the Castrum, a church is destined to be built in the fifth century or even earlier.178 It is in Late Antiquity that the oriental amphora imports are at their zenith. Even if this chronological phase is probably overrepresented, because of the nature of the settlement at that time, nonetheless the volume of imports from outside the Adriatic is now much higher than earlier on.

Several oriental regions exported their amphorae products as far as Brijuni during this period, with Cilicia, the Aegean and the Levant being the most important among them. The huge quantities of LRA 1 overwhelm all other oriental imports into second place. This dominance represents the big difference compared with other places in the region, but there are some smaller particularities in the percentages too. The LRA 2 are really well-represented within the material and the Ephesian products too. The presence of some Ephesus 56, an almost non-existent form in the northern Adriatic, is exceptional and constitutes a central difference from any other context in the region, at least up until now. Other products, such as late Cretan and Black Sea amphorae types present in Brijuni, are also unexpected in the region, but their number is really insignificant in the Castrum. Lastly, the Levantine products, mainly the LRA 4, are found in Brijuni too, but they are not worthy of comparison with numbers in other places in the northern Adriatic, especially with Ravenna, which seems to have been their main market. The presence of one fragment of the neck of a probable LRA 13 with a Byzantine Imperial stamp and dated to the very late sixth century A.D. is the latest oriental form documented with certainty in the Castrum. The amphora study of Brijuni contributes to the economic and social history of Istria; other studies will hopefully continue, taking up the baton. We see Brijuni, and Istria generally, not only as an producer of amphorae and olive oil production to a major degree during the Roman period, but also as a lively market on a most important trade route in the Adriatic – one open to the whole Mediterranean world, with especially close ties to the eastern part.

Bezeczky 1998a, 10-11. Matijašić 1993a, 249-251. 178 See Schobert in this volume. 176 177

BEZECZKY_Text.indd 70

18.03.19 14:39

71

7

AFRIC AN A MPHOR AE



Michel Bonifay – Claudio Capelli

In this chapter, we are dealing with containers attributed to Roman Africa, namely the Roman provinces of Mauretania Caesariensis, Numidia and Africa Proconsularis, gathered from the beginning of the fourth century onwards into the dioceses of Africa (at the same time as the new provinces of Sitifensis, Byzacena and Tripolitania were created), and corresponding to the modern-day (coastal) territories of Algeria and Tunisia, and of the western part of Libya. Sixty-one sherds of African amphorae have been studied, belonging to thirty different types, and dated from the midsecond century to the seventh century A.D.  An accurate graphic and photographic documentation together with a visual examination of each sherd followed by a binocular examination of all amphora micro-samples and by a petrographic analysis of a series of selected thin sections allowed us to specify the typology and the origin of most of these containers.1 Associated to the chronology and the content hypothesis currently suggested for each of these types, this amphora study sheds new light on African foodstuff distribution during the Roman and Late Antique times in the northern Adriatic. 1

TYPOLOGY AND ORIGIN OF THE AFRICAN AMPHORAE

1.1 AFRICAN AMPHORAE OF PUNIC TRADITION A single sherd can be attributed to this family. The foot 108 is very characteristic of type Hammamet 2E or 3A of the late fifth-start of the sixth century.2 Similar palm marks incised The whole graphic and photographic documentation is due to Tamás Bezeczky, our thanks to him. African amphora sherds were extensively examined and discussed in Brijuni by Bezeczky and Michel Bonifay in June 2013. The petrographic study of the micro-samples and the thin sections of African amphorae was performed by Claudio Capelli in Genova in 2014 and interpreted in collaboration with Bonifay in Aix-en-Provence in 2015. In the catalogue, the thin section descriptions are to the work of Capelli. The paper was written in 2015. 2 Bonifay 2004, Figs. 50.1 and 51.3. 1

BEZECZKY_Text.indd 71

before firing are already known on the same amphora type in Tunisia.3 The fabric fits very well with an origin in the northern Hammamet gulf, probably in the region of Sidi Jdidi, where a huge workshop has recently been evidenced.4 1.2 ROMAN-AFRICAN AMPHORAE 1.2.1 CLASSICAL ROMAN-AFRICAN AMPHORAE No Tripolitanian amphorae were found, except a single dubious sherd (109), similar in shape to type Tripolitana III (?), but the fabric is not typical for a Tripolitanian origin. A first, small group includes classical African amphorae of types Africana I and II of the mid-second to mid-third centuries. Among these the Sullecthum/Salakta products are well attested, with sherds of Africana I A (110) and II A (112). The attribution to this workshop is very easy, even with the naked eye, due to the very distinctive, brick-red and/or grey fabric, rich in small white inclusions (lime and micro-fossils inclusions, alongside the usual rounded quartz grains)5; types Africana I A and II A were star products during the second-third centuries in Sullecthum, as demonstrated by the most recent workshop surveys.6 Other production regions are not completely absent from the Brijuni assemblage. No. 111, provisionally identified as a variant of Africana I (?), could originate from a northwestern zone of Africa Proconsularis, perhaps the same zone where the type Ostia XXIII was produced,7 while no. 113 is typologically very similar to the Central-Tunisian production of Africana II ‘pseudo-Tripolitanian’.8 Wadi Arremel: Bonifay 2004, Fig. 8.17. Sidi Jdidi: Mukai 2010, Fig. 170, context 169.2. From now see Mukai 2016. 4 Mukai 2014: workshop of Tefernine. 5 Capelli et al. 2006. 6 Nacef 2015a. 7 Bonifay et al. 2015, 189-210. Hypothesis advanced more from a typological rather than a petrographical point of view. 8 Bonifay 2004, 114 and Fig. 110, type 24. In this case, petrography cannot help. 3

18.03.19 14:39

72

THE AMPHORAE

A second group, later in date, brings together Africana II types of the mid/second half of the third century. The Nabeul origin of type Africana II C (114 and 115) is confirmed by the petrographical macroscopic observation. This geographical attribution permits us to classify the foot 115 as Africana II C instead of Africana II D, because this first type is very characteristic of the Nabeul amphora production.9 No. 114 could be a late variant C3 of the same type, not predating the beginning of the fourth century. Type Africana II D was produced in several cities of the central Byzacena coastline, at least in Hadrumetum, Leptiminus and Salakta, but no. 116, according to the petrography, more likely originates from Thaenae, where the production of this type has also been evidenced.10 Finally, the origin of base 118 and rim/neck 119 cannot be determined.11 The third, most furnished group is devoted to type Africana III or Keay 25, from the fourth century. A first sherd (117), transitional between type Africana II D and type Africana III A, remains of unknown African origin. Within the variant Africana III A (= Keay 25.1), two workshops already encountered, Thaenae (?) (120) and Nabeul (122 and 123), have been identified while no. 121 remains undetermined. When the sherds are small, it is difficult to use the typology because type Africana III A was produced almost everywhere in Africa, and we have to deal with very discrete morphologic variations. By contrast, the variant Africana III B (= Keay 25.3) is considered as being typical of Nabeul.12 Indeed, among the three sherds attributed to this type, two (124 and 125) seem to have been produced in Nabeul, while the last one (126) does not. Lastly, a single sherd (127) can be identified as Africana III C (= Keay 25.2), originating from Nabeul, according to the petrography and the typology. 1.2.2 LATE ROMAN-AFRICAN AMPHORAE SMALL-SIZED LATE ROMAN-AFRICAN AMPHORAE Nine sherds have been attributed to the so-called ‘spatheion’ type13 of the fifth century. The first one (128), dubiously classified as spatheion 1, and originating from Nabeul, could also

be identified as a large-sized Keay 35A amphora. Sherds 129 to 131 are more characteristic of type spatheion 1. According to the typology and the petrography, two of them (130 and 131) come from Nabeul, the latter probably being later in date. Traditionally included in the ‘spatheion’ family, even if more probably completely distinct in shape and in date, the so-called ‘miniature spatheia’ (type 3), dating back to the end of the sixth and the seventh century, are well represented in Brijuni. A first sherd (132) is difficult to classify within the current typologies, while its fabric does not fit with any known references, except some Algerian ones. No. 133 could be identified as a variant B of type spatheion 3, with a petrography very similar to the amphora production of northwestern Zeugitana (namely between Utica and Tabarka). Lastly, a Nabeul origin is confirmed for the three examples of spatheion 3C (134-136), the latter being the less characteristic from a petrographical point of view, for such an allocation of origin. MEDIUM-SIZED LATE ROMAN-AFRICAN AMPHORAE A new type of late Roman-African amphora is currently being identified from rim/neck fragments scattered around the western part of the Mediterranean,14 characterised by frequent epigraphic marks vertically incised on the neck before firing.15 One of these finds comes from the Sidi Zahruni workshop in Nabeul,16 the city where, it seems, according to the petrography, that almost all these amphorae where produced. During the current investigation on this type, some complete exemplars have been identified,17 none of them being published until now.18 This amphora type is about 85  cm high for a body diameter of 26  cm, so looks like a medium-sized amphora. We provisionally propose identifying this new type with the fragmentary rim/neck published by Simon Keay under his type 3/5.19 One sherd (137) might belong to this type, with a classic Nabeul-zone B fabric.

Capelli et al. 2016, 288. Portus: Franco 2012, Fig. 5.8. Aquileia: Cipriano – Carre 1987, Fig. 19; Bueno et al. 2012, Fig. 6. 16 Bonifay 2004, Fig. 20.5. 17 For example, in the Mercati Traianei collection at Rome (inv. no. ANF 347, 348, 350) and in the storeroom of the Aquileia museum. 18 See, however, a possible comparison with a complete exemplar found in Algeria and published without scale: Salama 2006, Fig. 23. 19 Keay 1984, Fig. 42.4. 14 15

Bonifay 2004, 115. Bonifay 2004, 33. 11 In this case, the typology, which is not clear, could not help. 12 Bonifay 2004, 122. 13 On this erroneous but universally admitted denomination, see Bonifay 2004, 125. 9

10

BEZECZKY_Text.indd 72

18.03.19 14:39

CHAPTER 7 – African amphorae (Bonifay – Capelli)

LARGE-SIZED LATE ROMAN-AFRICAN AMPHORAE The large-sized late Roman-African amphorae are the most numerous. Two first sherds (138 and 139) belong to type Keay 59/8B, with a very distinctive pink, compact fabric, and beige, smooth surface. The rim 138 illustrates a variant of type Keay 8B dating back to the second half of the fifth century,20 while the foot 139 could belong to the one or the other of the two sub-types Keay 59 or 8B. The origin of this Keay 59/8B type has been attributed to south Byzacena, based on the discovery of two workshops.21 Another sherd (140) belongs to a barely documented type, classified by S. Keay as a variant of his type 11B, even if the rim profile is very different from the main (Tripolitanian) type.22 The petrography cannot help us to localise the production, but some similarities are observed with the fabrics of the previous Keay 59/8B type. The presence of both these types in the survey of the Kasserine region23 could be an argument for a common, south Byzacena origin. A fourth sherd (143) is a Keay 36 amphora, with a fabric entirely compatible with the normal fabric of the Keay 27 and 36 amphorae. According to the distribution of these types in Africa, an origin from the Carthage region or the southern Mejerda valley seems likely.24 Then, the Brijuni assemblage provides a large series of large-sized amphorae originating from Nabeul, including types Keay 35A (141), 35B (142), 55 (144 and 145), 57 (146-148), 55-56-57 (foot 149), from the fifth century-first half of the sixth century. Most of these amphorae, easy to classify within Keay’s typology, appear to come, according to the petrography, from zone B of the Nabeul workshops, active between the fifth and the seventh centuries.25 On the other hand, as elsewhere, type Albenga 11/12 or Keay 62Q amphorae (150-152), from the second half of the fifth century-first third of the sixth century, are difficult to classify, and their origin(s) remain unknown until now. By contrast, the latest productions from the the mid-sixth to the end of the seventh century are easier to characterise. Except for two feet of type Keay 62 or 61 (155 and 160), the first one perhaps attributable to the Nabeul workshop and the second one of unknown origin, all these amphorae

See, for example, Pieri 1998, Fig. 198.280-282. Bonifay 2004, 132. 22 Keay 1984, Fig. 172.2. 23 Neuru 1990, Figs. 16.j and k. 24 Bonifay et al. 2012, 241. 25 Bonifay et al. 2010. 20 21

BEZECZKY_Text.indd 73

73

appear to have been produced in the Sahel region of Tunisia. The Moknine workshops are represented by examples of Keay 62 (153 and 154) and Keay 61 A/D (158-159) amphorae. The latter are characteristic of the ‘classic’ fabric, evidenced in the workshop dump excavated at the beginning of the 90s,26 while the first do not seem to fit with the new workshop dump recently discovered by Jihen Nacef in the same city.27 Another sherd (161) must be classified as type Keay 61C, and subsequently attributable, from both a typological and petrographical point of view, to the workshop of Henchir Chekaf in the hinterland of Salakta.28 The type Keay 8A is also present (162-163) and its origin in the Sahel region seems probable, even if it is not possible to specify the workshop itself. The production of this type is evidenced in Leptiminus29 and near Thapsus.30 Lastly, a southern Sahel provenience (the region of Rougga?)31 has been confirmed for the three sherds of type Keay 34 (164-166) discovered in Brijuni. 1.3 AFRICAN IMITATIONS OF NON-AFRICAN TYPES A single sherd (167) could be attributed to a variant of type Keay 1B, which is a fourth-century late derivative of type Gauloise 4, perhaps produced in the territory of modern-day Algeria. However the fabric of this sherd is quite different from the normal fabric of this type.32 1.4 STORAGE JARS (?) Two fragments of massive feet (168 and 169) do not seem to belong, because of their size, to ‘normal’ transport amphorae. They may fit better with African storage amphorae, well known in the African territory,33 and occasionally travelling, as evidenced by examples found in Rome34 and even in a shipwreck along the Croatian coastline.35

Bonifay 2004, 35; Gandolfi et al. 2010, 37-38. Nacef 2015b. 28 Bonifay 2004, 140; Nacef 2007; Gandolfi et al. 2010, 37. 29 Dore 2001, Fig. 1.65, no. 18. 30 Nacef 2014, Fig. 3.26-33. 31 Bonifay 2004, 143. 32 Capelli – Bonifay 2007, 554-555. 33 Bonifay 2013, 539 and Fig. 1. 34 Contino – D’Alessandro 2014, 328 and Figs. 9-12. 35 Huguet 2012, 130 and Fig. 50.2. 26 27

18.03.19 14:39

74

THE AMPHORAE

2 AFRICAN IMPORTS TO BRIJUNI IN THE NORTHERN ADRIATIC CONTEXT

MID-SECOND CENTURY TO END OF THE THIRD CENTURY

A first, previous survey on African imports to Brijuni had demonstrated that African amphorae represented about 40 % of the whole late Roman amphorae, while they were completely absent from the early Roman contexts.36 In consequence, if we want to gauge the imports of African amphorae to Brijuni, it is important to periodise these imports, and to look for comparisons with other northern Adriatic sites as well as other African ceramic imports. The specific history of the Roman villa and the Byzantine ‘castrum’ could also help with the interpretation of this documentation.

As early as the mid-second century, the new type Africana I reached the northern Adriatic in such quantities that they could be reused for the packaging of local salted fish, as shown by the Grado shipwreck.43 Following the drop in the production of Histrian olive oil, and in spite of the massive import of Spanish oil containers Dressel 20, the need for olive oil, not only for local use but also for redistribution towards the north, ensured an outlet for the African production surplus. As in many places in the Mediterranean and also within the Grado shipwreck cargo, the Salakta products are the first ones to be attested in Brijuni (110). The same port of Byzacena also provided other foodstuffs, packaged in Africana II A amphorae, possibly salted fish or even wine (112). The presence of the Sullecthum amphorae along the eastern Adriatic coastline, associated with central Tunisian ARS A/D44 and C, has sometimes been interpreted as evidence of a direct connection between Byzacena and the Adriatic, but the petrographical observations on the late Antonine and Severan amphorae from Brijuni show that some other African regions (111?, 113) were exporting their olive oil surpluses as well. The situation seems to change from the mid-third century onwards, with the emergence in the Brijuni assemblage of new exporting cities, like Nabeul and Thaenae, and with the supremacy of other foodstuffs rather than olive oil: probably salted fish and/or wine packaged in Africana II C (115) and II D amphorae (116). These types are well attested in Aquileia.45 But it does not mean that African olive oil did not reach the Adriatic anymore. Africana I B amphorae are present in Aquileia46 alongside Tripolitana III amphorae,47 including some late variants,48 evidence of continued imports of oil from the Severan period onwards.

BEFORE THE MID-SECOND CENTURY Although the villa existed from the first century B.C.,37 no African amphorae found in Brijuni seem to predate the midsecond century A.D.  Almost the same situation prevails in the underwater finds along the Croatian coastline.38 Of course, the percentages of African amphorae are low everywhere during the late Republican period and the beginning of the Empire, even in Rome where the African (wine?) containers generally represent less than 10 % of the total amount of the amphorae in the mid-first-century A.D.  contexts.39 But on the other hand, the African imports of oil, olives and wine become well attested during the time of Flavian not only in Rome, but also in Pompeii, and in Arles, with the types Tripolitana I, Ostia LIX and Schöne-Mau XXXV. It is likely that the gap in early African amphorae will soon be filled, at least in part, by new research in the northern Adriatic. Some Flavian types are already attested in Zaton40 and in Aquileia,41 suggesting some imports of African wine and salted fish (?). Probably, the inhabitants of the ‘Castrum’ villa of Brijuni did not need to consume these imported foodstuffs, although they already felt the need to use some African table- and cooking wares.42

Vidrih Perko – Pavletić 2000, Figs. 2 and 3. Begović – Schrunk 2011a, 376. 38 Pešić 2013, 1210. 39 Reynolds 2010, table 1a. 40 Pseudo-Dressel 2/4 (wine) from western Tripolitania. Personal obs. in 2013, thanks to Dr. S. Glušcević. 41 At least three complete exemplars of type Uzita 52.10 (salted fish) in the collections of the Museo Archeologico Nazionale. Personal obs. in 2013, thanks to Dr. P. Ventura, Director of the Museum. 42 Vidrih Perko – Pavletić 2000, 264 (ARS forms Hayes 8A and 9A), and 265 (early ACW form Hayes 198). 36 37

BEZECZKY_Text.indd 74

Auriemma 2000. Some sherds have been found in the ‘Castrum’ villa: Pröttel 1996, Pl. 47. 12-13. 45 Africana II C: Verzár Bass 1994, Pl.  66, AA 15. Africana II D: Verzár Bass 1991, Pl.  38, AA 29; Cipriano – Carre 1987, Fig. 20. 46 Carre et al. 2007, 608 and note 7. 47 An example of Tripolitana III with a stamp AVGGG in Aquileia: Cipriano – Carre 1987, Fig. 18. 48 Verzár Bass 1994, Pl. 66, AA 2. 43 44

18.03.19 14:39

CHAPTER 7 – African amphorae (Bonifay – Capelli)

75

FOURTH CENTURY TO BEGINNING OF THE FIFTH CENTURY

FIRST HALF OF THE FIFTH CENTURY TO FIRST HALF OF THE SIXTH CENTURY

Abundant in the Brijuni assemblage, the fourth-century Africana III amphorae are also the most widespread African containers throughout the Roman world. These amphorae generated the highest percentages of African imports in Rome,49 and provided impressive quantities of exemplars in the northern Adriatic50, often reused in drainage systems.51 The hypothesis of a wine content, put forward for these amphorae according to their typology, their systematic internal pitch lining, and the ‘universal’ pattern of their distribution, has recently been confirmed by a series of chemical analyses,52 even if another content could be possible for variant B. The ‘Castrum’ villa – or already ‘settlement’ – of the fourth century imported a good deal of African wine (at least 120-123, and possibly also 124-127). Africana III amphorae also are “the most common find in the underwater sites in Croatia”.53 Africana III wrecks are particularly abundant on the eastern side of the Adriatic, which is considered the ‘upward’ direction of sailing. One of the most important cargoes is illustrated by the Sobra wreck, from the middle-third quarter of the fourth century, with an assemblage of Africana III A and B, some of them bearing a stamp PGT attributable to the city of Nabeul.54 Two examples of the same stamp have been found in Aquileia.55 These observations are consistent with the Brijuni assemblage where most of the Africana III amphorae appear to originate from Nabeul. The comparison between the Brijuni and Aquileia assemblages extends to some rare types, such as a variant of Keay 1B (167) attested at both of these sites.56 At the same time, Aquileia and the ‘Castrum’ villa/settlement were still being supplied with African table- and cooking wares. In Brijuni, high-quality C4 ARS Hayes 53A with applied decoration are present, alongside cruder D1 plates Hayes 59 and 61A with stamped decoration of styles A(i) and (ii).57

The period comprised between the first half of the fifth century and the first half of the sixth century is a time of massive inflow of African amphorae into the fortified settlement of Brijuni. Paul Reynolds already noted this trend in the whole Adriatic from the mid-fifth century onwards, explaining it by the ‘major shipping route to Ravenna’, and the ability of cities and settlements located along this route to benefit from these exports.58 A lot of different types are present in Brijuni, already wellknown in different places on the Mediterranean, sharing in common a strong presence of Nabeul products: spatheia type 1, Keay 3/5, 35A and B, 55, and 57 (128-131, 137, 141-142, 144-145, and 146-148), for example in Catalonia where they were first classified,59 in southern Gaul,60 in Rome,61 and even on the Black Sea.62 In some places, such as in Brijuni, they are associated with the south Byzacena types Keay 59/8B (138-139), northwestern types Keay 27/36 (143), and the unprovenanced type Keay 62Q (150-152). All these amphorae are well attested along the eastern side of the Adriatic, in Albania,63 Croatia and the Slovenian Karst region64, as well as in Aquileia,65 Ravenna66 and along the Po Valley.67 A major problem posed by this documentation is that we know little about the content of these containers. This uncertainty is further increased by the fact that the spatheion type was possibly a cargo space-filler with an interchangeable content of oil, fish sauce or wine. Reuses of these amphorae are also possible, as demonstrated by the case of the warehouse of Classe, which burned down at the end of the fifth century.68 In Brijuni, these amphorae are accompanied by huge amounts of African tableware, with a pattern very similar to the one known in the northern Gulf of Hammamet and in

Reynolds 2010, 99. See also Auriemma – Quiri 2007, 34. Keay 1984. 60 Marseille: Bonifay et al. 2011. 61 Panella et al. 2010, 66-67. 62 Opaiţ 1997-98. 63 Butrint: Reynolds 2004, Fig. 13.155. Byllis: Bonifay – Cerova 2008. Durres: Hoti – Capelli – Piazza 2008; Shkodra 2005, Fig. 10.2, 4-5. Shkodra: Hoxha 2003. 64 Vidrih Perko 1997. 65 Collection of the Museo Archeologico Nazionale, personal obs. in 2013. 66 Antiquarium of Classe, personal obs. in 2006. 67 Brescia: Brogiolo 1999. Milan: Lusuardi Sienna 1990. 68 Pecci et al. 2010. 58 59

C. Panella in Anselmino et al. 1986, Table 3; Panella – Saguì 2001, 766-767. 50 Carre et al. 2007, Fig. 4. 51 See, for example, Maurina 1999. 52 Woodworth et al. 2015, 41-57. 53 Pešić 2013, 1208. 54 According to recent petrographical analyses. 55 Verzár Bass 1994, Pl. 76, AB 7-8. 56 Verzár Bass 1991, Pl.  38, AA 42; Verzár Bass 1994, Pl.  69, AA46. 57 Pröttel 1996, Pl. 48, 54.6-11, 85-87, 89-91. 49

BEZECZKY_Text.indd 75

18.03.19 14:39

76

THE AMPHORAE

Nabeul: forms Hayes 50B.61, 61C, 87A, Fulford 27,69 suggesting a direct connection between the Cape Bon region of Tunisia and the northern Adriatic for both the amphorae and the tableware. Another interesting feature of the fifthcentury African imports in the northern Adriatic is the presence of huge quantities70 of flanged bowls, Fulford 22-23, again produced in Nabeul.71

including some tableware, ARS forms Hayes 105, 106 and 109.74 Does this character imply that, at this time, the Brijuni ‘Castrum’ had become a Byzantine military settlement as well?75 No precisions about the contents of these amphorae can be obtained from the current state of the research. CONCLUSION

MID-SIXTH CENTURY TO LATE SEVENTH CENTURY The last period of African imports to Brijuni mainly includes products from coastal Central Byzacena: Keay 62 and 61 from Moknine, Keay 61C from the Salakta hinterland, and Keay 8A from unknown Tunisian Sahel sources (153-154, 156-159, 161, and 162-163), alongside miniature spatheia of type 3C from Nabeul (134-136). This kind of assemblage has sometimes been interpreted as a signature of the Praetorian Prefecture organisation72 and of the specific supply of the Byzantine territories. At this time, the supply of Brijuni is identical to the supply of the Byzantine place of Koper,73

Pröttel 1996, Pl. 52.21, 52.8-9, 50.14, 47.8. Brijuni: Pröttel 1996, Pl.  52.17. Aquileia: Verzár Bass 1994, Pl. 45-46. 71 Bonifay 2004, 255 and Fig. 139. 72 Reynolds 2010, 127. 73 Cunja 1996.

Even if this documentation is no longer contextualised, a careful integrated archaeological and petrographical approach to the individual sherds leads to a better understanding of the supply of the Brijuni Islands with African foodstuffs from the mid-second century to the end of the seventh century. This assemblage represents an important milestone for the study of the African amphorae between the sites recently surveyed in Albania (Butrint, Byllis, Shkodra), and the main cities of the northern Adriatic (Aquileia, Ravenna), and probably the first one to have been extensively studied with the specific aim of specifying the origins of the amphorae within the African territory.

69 70

BEZECZKY_Text.indd 76

74 75

Pröttel 1996, Pl. 52.1-2, 4-6. Begović – Schrunk 2011a, 384.

18.03.19 14:39

77

8

A MPHOR A STA MPS AND INSCRIPTIONS



Piero Berni Millet – Tamás Bezeczky

The catalogue contains the inscriptions and stamps on the amphorae found at the villa. However, we wanted the reader to be able to study them all together, as in the 16-18 tables. 2. Lamboglia 2, rim, neck, no. 533 (152) – 3960, Pl. 1 and 16. Stamp: BAR·ME[---], on the rim. This stamp has not been found elsewhere. P.  Desy mentions the BARNAES and G.  Volpe another BARNA (retro) stamps.1 The Tarentum [---]BARN stamp is probably fragmentary.2 However, the Delos BARN stamp could be connected to the Castrum stamp.3 3. Lamboglia 2, rim, no. 656 (118), Pl. 1 and 16. Stamp: VERSO, on the rim. The other Castrum stamp, VERSO, is widely known.4 Brunella Bruno distinguished two VERSO stamp types, but the Castrum version is somewhat different from both.5 In Vercelli dated in second part of the first century B.C.6 5. Dressel 6A, rim, neck, no. 140 (54) – 25973, Pl. 1 and 16. Stamp: V, V(…), on the rim. This stamp has not been found elsewhere. 7. Dressel 6A, rim, neck, no. 140 (153) – 25972, Pl. 1 and 16. Numeral graffito on the rim, cut after firing: XXIIII, perhaps ‘twenty-four’.

Laecanius (nos. 11-17) and Emperor’s (nos. 18-22) production, in Fažana workshop. Stamp - Dominus stamp: the Laecanius or Emperor’s stamp Stamp* - vilicus or workshop manager stamp 11. Dressel 6B, rim fragment, no. 1493 (01), Pl. 2 and 16. Stamp: [LA]EK·A, [La]ek(ani) A(...).7 Occurrence: Magdalensberg, Mantua, Novaria, Este, and Verona. 12. Dressel 6B, rim, neck fragment, handle, no. 38 (25)  – 2129, Pl. 2 and 16. Stamp*: [S]YNT, [S]ynt(…).8 Occurrence: Dertona, Muntia, Patavium (2), Vercellae, Vincentia (2). 13. Dressel 6B, rim fragment, no. 1326 (611), Pl. 2 and 16. Stamp*: COMVS, Comus, S retro.9 Occurrence: Magdalensberg. 14. Dressel 6B, rim, no. 1243 (110), Pl. 2 and 16. Stamp: LAE, Lae(kani) . 15. Dressel 6B, rim, no. 719 (117), Pl. 2 and 16. Stamp: C·L[AEK], C. L[aek](ani) . Occurence: Pula, Celeia, Poetovio.10 16. Dressel 6B, rim, no. 1279 (123), Pl. 2 and 16. Stamp: [LA]EK, [La]ek(ani) .

Desy 1989, 179; Volpe 1990, 245, Fig 241, no. 3 = Blanc-Bijon et al. 1998, 55, no. 610. 2 Desy 1989, 26, no. 46. 3 Desy 1989, 59, no. 327; CIL III. 7309,35. 4 Desy 1989, nos. 116, 118, 281, 299, 302, 358; Bruno 1995, 110112; Carre et al. 1995, 47, no. 96 bis; Gabucci – Quiri 2008, 65; Italy (Milano, Ivrea, Altino, Luni), Albania (Durrhachium, Kryemdhej), Greece (Delos), Spain (Cabezo Augudo, San Jordi A, Villaricos, Barcelona). 5 Bruno 1995, nos. 110-111. 6 Brecciaroli Taborelli 1987, 118. 1

BEZECZKY_Text.indd 77

17. Dressel 6B, rim, no. 1371 (121), Pl. 2 and 16. Stamp*: PA[GANI], Pa[gani].

Bezeczky 1998a, 95, no. 1; Mazzocchin 2013, 121, nos. 25.1a-d. Perhaps Syntrophus and as Syntychus, Mócsy et al. 1983, 279; Pape 1884, 1459; Bezeczky 1998a, 204, no. 507; Mazzocchin 2013, nos. 25.14a-b. 9 Mócsy et al. 1983, 86; Bezeczky 1998a, 232, no. 611. Pl. 32 and 56. 10 Bezeczky 1998a, nos. 37-38. 7 8

18.03.19 14:39

78

THE AMPHORAE

Occurence: Pullaria ins.(Verige), Aguntum (2), Iuvavum, Augusta Taurinorum, Concordia, Dertona, Opitergium (2), Patavium, Cervignano.11 18. Dressel 6B, rim fragment, no. 1476 (648), Pl. 2 and 16. Stamp: IM[P], Im[p](eratoris) [Vespasiani].12 Occurence: Fažana, Pullaria ins., Dragonera, Aguntum, Aquincum, Concordia, Flavia Solva, Tokod, Augusta Taurinorum. 19. Dressel 6B, rim, no. 1338 (120), Pl. 2 and 16. Stamp: IMP·TRA, Imp(eratoris) Tra(iani), upside down Stamp*: ...VS.13 Dominus stamp occurence: Mediolanum, Vindobona, Savaria. 20. Dressel 6B, rim, no. 1426 (122), Pl. 2 and 17. Stamp: [IM]P·TRA, [Im]p(eratoris) Tra(iani), upside down. Occurence: Mediolanum, Vindobona, Savaria. 21. Dressel 6B, rim, no. 12 (116) – 26017, Pl. 2 and 17. Stamp: [I]MP·T[RA], [I]mp(eratoris) T[ra](iani). Occurence: Mediolanum, Vindobona, Savaria. 22. Dressel 6B, rim, no. 500 (119) – A 10627, Pl. 2 and 17. Stamp*: LESBI (reverse). Occurence: Concordia, Patavium (2), Siscia, Savaria, Augusta Taurinorum.14 38. Beltran IIA, base, no. 10 (6) – 16027, Pl. 17. Numeral graffito cut after firing: V, perhaps ‘five’. 40. Dressel 20, handle, no. 1271 (125), Pl. 17. Stamp: palm branch (ramus palmae) M[---]. Guadalquivir production, similar stamp (ramus palmae) MAF (ramus palmae)15 The palm branch was in vogue during Hadrian’s period. Occurence: Augst. 41. Dressel 20, body, base, no. 1554 (115), Pl. 5 and Ch. 5 Figs. 10, 12 and 13. Graffito cut before firing: Asiatici / sec( )  × k(alendas) mar(tias) Alb( ) + II. Date: Second third of the second century A.D.  Bezeczky 1998a, nos. 490, 652-659; Cipriano  – Ferrarini 2001, 167-168; Bezeczky 2014, 250. 12 Bezeczky 1998a, 240, no. 648. Pl. 34 and 56. 13 Bezeczky 2014, 251. 14 Bezeczky 2014, 251. 15 Martin-Kilcher 1987, 94, (1388), ST 2a. 11

BEZECZKY_Text.indd 78

Similar graffito occurence: Germany: Altenstadt (Hessen), England: Lincoln (Lincolnshire), Wall (Staffordshire), Brough on Noe (Derbyshire), Vindolanda (Bardon Mill), France: Saint-Thibéry (Hérault), Béziers (Hérault), Reims (Marne). See Berni Millet in this volume. 86. Late Roman Amphora 1 (LR 1), rim, neck, handles, no. 1342 (136), Pl. 17. Graffito cut after firing: A( ). 100. Late Roman Amphora 13 (LR 13), neck, no. 659 (148), Pl. 17. Stamp: Emperor’s bust and inscriptions: (…)OΛ(…).16 Occurence: Alexandria, Kellia (Egypt), Istanbul, Yenikapı (Turkey), Athens, Kos (Greece), Tocra (Libia), Geronissos (Cyprus). See González Cesteros in this volume. 108. Hammamet 3A, base, no. 1315 (135), Pl. 17. Stamp: Palm branch, similar stamps in Hammamet golf.17 See Bonifay – Capelli in this volume. 123. Africana III A/B, neck, handle, no. 865 (190), Pl. 17. Graffito cut after firing: K( ). 172. Miscellaneous, neck, no. 1356 (195), Pl. 15. Stamp: [---]VR. Date: Augustean-Julio-Claudian period. Comment: Marseille, it is a Gauloise 2 type.18 182. Miscellaneous, body fragment, no. 846 (199), Pl. 15. Red dipinto: K( ). 183. Miscellaneous, body fragment, no. 734 (200), Pl. 15. Graffito on the body cut after firing: Υε(…). Ephesus amphora fragment, the type is uncertain (one-handled jar, LR 3 or Ephesus 56?). 184. Miscellaneous, body fragment, no. 38 (26)  – 2130, Pl. 15. Nominal graffito on the body cut after firing: IVLIVSSIFA(…). The problem is how to cut the words: Iuli Ussi Fa[---] ? or Iulius Sifa[---] ? Is it a Roman graffiti? Diamanti 2010, 8. Bonifay 2004, 20, Fig. 8, 17-19. 18 Similar stamps in Laubenheimer 1985 and Bertucci 1991. There is a VAL stamp in Nijmegen. 16 17

18.03.19 14:39

79

9 C ATALOGUE

Tamás Bezeczky – Piero Berni Millet – Michel Bonifay – Claudio Capelli – Horacio González Cesteros – Sándor Józsa – György Szakmány

ABBREVIATIONS Name - simple name

DNB - diameter at the junction of neck and body

No. - number or inventory number

DB - diameter of the body

Stamp - dominus stamp

DF - diameter of the foot

Stamp* - vilicus/workshop manager stamp

H - height of the fragment

Dim - dimension measured

HT - height of the amphora

D - diameter of the rim

S - section of the handle

V - thickness of the rim

Colour - Munsell Colour

HD - height of the rim

SF - surface colour

DN - minimum diameter of the neck

VC - visual characteristics

HN - height of the neck

Date - date range

ITALIAN AND ADRIATIC Lamboglia 2 1. Name: Rim, neck No.: 18 (09) - 3561 Context: Room Cist.X Dim: D = 17 cm, V = 2.9 cm, HD = 4.3 cm, H = 10.4 cm Colour: light red (2.5YR 6/6) VC: Hard, very fine-grained, micaceous fabric. Sparse scatter of small red iron ore and fine-grained limestone particles. Date: late second to late first century B.C.  Origin: Italy, W coast of the Adriatic 2. Name: Rim, neck No.: 533 (152) - 3960 Context: Room H (RR) Stamp: BAR·ME(…), on the rim. This stamp has not been found elsewhere. Dim: D = 16 cm, V = 2.5 cm, HD = 5.1 cm, H = 9.2 cm Colour: light reddish brown (5YR 6/6), SF: buff (7.5YR 7/4) VC: Hard, fine-textured fabric, slightly micaceous, sparse scatter of small red iron ore. Date: late second to late first century B.C.  Origin: Italy, W coast of the Adriatic 3. Name: Rim No.: 656 (118) Context: ? Stamp: VERSO, on the rim

BEZECZKY_Text.indd 79

Dim: D = 18 cm, V = 3 cm, HD = 3.4 cm, H = 5.5 cm Colour: reddish yellow (5YR 6/6) VC: Hard, fine-textured fabric, sparse scatter of small red iron ore, few reaction rims surrounding voids which held limestone. Date: first quarter of the first century B.C.  Origin: Italy, W coast of the Adriatic Dressel 6A 4. Name: Rim No.: 18 (10) - 3563 Context: Room?, 128 Dim: D = 19 cm, V = 3.2 cm, HD = 5.4 cm, H = 8.3 cm Colour: buff (5YR 7/3) VC: Hard, fine-textured fabric. Date: late first century B.C. to mid-first century A.D.  Origin: Italy, W coast of the Adriatic 5. Name: Rim, neck No.: 140 (54) - 25973 Context: Courtyard, 106 Stamp: V, V(…), on the rim Dim: D = 18 cm, V = 2.4 cm, HD = 4 cm, H = 7.5 cm Colour: reddish yellow (5YR 6/6) VC: Hard, sandy fabric. Scatter of small red iron ore and fine-grained limestone particles. Date: late first century B.C. to mid-first century A.D.  Origin: Italy, W coast of the Adriatic

18.03.19 14:39

80

THE AMPHORAE

6. Name: Base No.: 18 (11) - 3567 Context: Room?, 128 Dim: DF = 3.9 cm, H = 25 cm Colour: buff (5YR 7/4) VC: Hard, fine-textured fabric. Sparse scatter of small red iron ore, few reaction rims surrounding voids which held limestone, microfossil. Date: late first century B.C. to mid-first century A.D.  Origin: Italy, W coast of the Adriatic 7. Name: Rim, neck No.: 140 (153) - 25972 Context: Courtyard, 106 Graffito: XXIIII on the rim, post cocturam Dim: D = 16.5 cm, V = 2.4 cm, HD = 4.7 cm, H = 6.5 cm Colour: light red (2.5YR 6/6), SF: buff (5YR 7/4) VC: Hard, fine-textured fabric. Sparse scatter of small red iron ore and fine-grained limestone particles, microfossil. Date: late first century B.C. to mid-first century A.D.  Origin: Italy, W coast of the Adriatic 8. Name: Rim, neck, handle No.: 1025 (106) Context: ? Dim: D = 16.5 cm, V = 2.2 cm, HD = 4.3 cm, DN = 11 cm, S = 3.8 × 4.4 cm, H = 39 cm Colour: reddish yellow (5YR 6/6) VC: Hard, fine-textured fabric, voids, sparse scatter of small red iron ore and fine-grained limestone particles. Date: late first century B.C. to mid-first century A.D.  Origin: Italy, W coast of the Adriatic Dressel 6B earliest form (1A phase) / ante Dressel 6B 9. Name: Rim, neck, handle No.: 221 (61) - 3742 Context: Cist. X, 192 Dim: D = 19.5 cm, V = 3.2 cm, HD = 4.6 cm, H = 17.2 cm Colour: reddish yellow (5YR 6/6) VC: Hard, very fine-grained, slightly micaceous fabric. Sparse scatter of small, fine grained limestone particles. Thin-section: Massive brownish red matrix, different sized holes. Non-carbonatics: moderately sorted, very fine-grained (15-60 μm), medium amount, few coarser (100-180  μm) mainly quartz. Many micas, accessory minerals. Few globigerinoids (?), limonitic concretion, terra rossa. Raw materials: mainly clayey flysch, more terra rossa. Date: mid-first to the end of the first century B.C.  Origin: North Italy or Istria? 10. Name: Base No.: 69 (42) - 3898 Context: Room K, 190 Dim: FD = 8.1 cm, H = 8 cm Colour: yellowish red (5YR 5/6) VC: Hard, very fine-grained fabric. Frequently scattered rounded limestone inclusions, and microfossils. Date: mid-first to the end of the first century B.C.  Origin: North Italy or Istria?

BEZECZKY_Text.indd 80

Dressel 6B Laecanius production, Fažana 11. Name: Rim frgm. No.: 1493 (01) Context: Room 203 C Stamp: ..EK·A, (C) [La]ek(ani) (Bassi), A(...). Dim: D = 15 cm, V = 2.1 cm, H = 13.7 cm Colour: light red - reddish yellow (2.5 YR 6/6 - 5 YR 6/8) VC: Hard, very fine-grained fabric. Frequently scattered rounded carbonate particles, and microfossils, shells. Thin-section: Massive brownish red matrix. Non-carbonatics: very well sorted, very fine-grained (15-70  μm), medium amount, mainly quartz. Micritic carbonate: abundant, coarser-grained (40-200 μm), well sorted, rounded. Many globigerinoids. Few micas, limonitic concretions, bivalves, forams, opalic spiculae, terra rossa. Raw materials: mainly flysch, little terra rossa, little recent marine material. Date: Tiberian / early Claudian period 12. Name: Rim-, neck frgm., handle No.: 38 (25) - 2129 Context: Room D Stamp*: .YNT, [S]ynt(...) Dim: D = 14 cm, H = 10.4 cm. Colour: light red (2.5 YR 6/6) VC: Hard, very fine-grained fabric. Frequently scattered rounded limestone particles, and frequent microfossils. Thin-section: Massive brownish red matrix, few elongated holes. Non-carbonatics: very well sorted very fine-grained (15-80 μm), large amount, mainly quartz. Micritic carbonate: abundant, coarser-grained (40200 μm), well sorted, rounded. Many globigerinoids, accessory minerals. Few micas, bivalves, forams, limonitic concretions, opalic spiculae, terra rossa. Raw materials: mainly flysch, little terra rossa, little recent marine material. Date: Tiberian / early Claudian period 13. Name: Rim frgm. No.: 1326 (611) Context: 221 Stamp*: COMVS, Comus, S retro Dim: D = 15 cm, V = 2.2 cm, H = 6.9 cm Colour: reddish yellow - yellowish red (5 YR 6/6 - 5/6) VC: Hard, very fine-grained fabric. Frequently scattered rounded limestone particles, and frequent microfossils. Thin-section: Massive brownish red matrix, few large holes. Non-carbonatics: very well sorted, very fine-grained (15-80 μm), low amount, mainly quartz. Micritic carbonate: abundant, coarser-grained (40200 μm), well sorted, rounded. Many micas, globigerinoids, limonitic concretions. Few opalic spiculae, terra rossa. Raw materials: mainly flysch, little terra rossa, little recent marine material. Date: Tiberian / early Claudian period 14. Name: Rim No.: 1243 (110) Context: 101 Stamp: LAE, (C) LAE(kani) (Bassi) Dim: D = 14.5 cm, V = 1.8 cm, H = 4.8 cm Colour: reddish yellow (5YR 7/6), SF: light red (2.5YR 6/6)

18.03.19 14:39

CHAPTER 9 – Catalogue (Bezeczky et al.) VC: Very fine-grained fabric, a dissolved secondary carbonate. Thin-section: Mixture of non-calcareous brownish red and calcareous brownish grey matrix. Non-carbonatics: abundant, well sorted, very fine-grained (15-100  μm), mainly quartz. Micritic carbonate: few, coarse-grained (generally 100-400  μm maximum 1000  μm). Many globigerinoids. Few micas, limonitic concretions, terra rossa. Raw materials: limy and non-calcareous clayey flysch, very little terra rossa, very little recent marine material. Date: Claudian / Flavian period 15. Name: Rim No.: 719 (117) Context: ? Stamp: C.L…, C. L[aek](ani) (Bassi) Dim: D = 14 cm, V = 1.9 cm, H = 8.1 cm Colour: reddish yellow (5YR 6/6) VC: Soft, larger iron ore particles, and microfossils, few reaction rims surrounding voids which held limestone, irregularly fired. Thin-section: Massive brownish red matrix. Non-carbonatics: very well sorted, very fine-grained (15-60  μm), medium amount. Micritic carbonate: coarser-grained, medium amount. More micas, globigerinoids, large bivalves, limonitic concretion, terra rossa. Raw materials: mainly flysch, more terra rossa, very little recent marine material. Date: Claudian / Flavian period 16. Name: Rim No.: 1279 (123) Context: 103B Stamp: …EK, (C) [LA]EK(ani) (Bassi) Dim: D = 14 cm, V = 1.9 cm, H = 6.8 cm Colour: reddish yellow (5YR 6/6) VC: Soft fabric, iron ore particles, few reaction rims surrounding voids which held limestone. Thin-section: Massive brownish red matrix. Non-carbonatics: very well sorted, very fine-grained (15-60  μm), medium amount, few coarser (100-140  μm) mainly quartz. Micritic carbonate: coarser-grained (40-200 μm), well sorted, rounded. Many globigerinoids. Few micas, limonitic concretions, opalic spiculae. Raw materials: mainly flysch, little terra rossa, little recent marine material. Date: Claudian / Flavian period 17. Name: Rim No.: 1371 (121) Context: ? Stamp*: PA…, Pa[gani] Dim: D = 14 cm, V = 2 cm, H = 6 cm Colour: reddish yellow (5YR 6/8) VC: Soft fabric, scattered rounded limestone and iron ore particles. Thin-section: Non-massive, inhomogeneous pale brownish red and darker reddish brown matrix, much fine porous. Non-carbonatics: very well sorted, very fine-grained (15-60 μm), large amount, few coarser (100140  μm), mainly quartz. Few globigerinoids, limonitic concretion, terra rossa. More micas, opalic spiculae. Raw materials: mainly flysch mixed with some terra rossa, little recent marine material. Date: Claudian / Flavian period

BEZECZKY_Text.indd 81

81

Imperial production, Fažana 18. Name: Rim frgm. No.: 1476 (648) Context: 203 B Stamp: IM[P], Im[p](eratoris) Dim: D = 14 cm, V = 2.3 cm, H = 7.5 cm Colour: reddish yellow (5 YR 6/6) VC: Soft, fine-grained fabric, scattered rounded limestone particles, and microfossils. Thin-section: Massive brownish red matrix, few large holes. Non-carbonatics: very well sorted, very fine-grained (15-70 μm), medium amount, few coarser-grained (100-140 μm) mainly quartz. Micritic carbonate: abundant, coarser-grained (40-200 μm), well sorted, rounded. More limonitic concretions. Few micas, globigerinoids, recent foram, opalic spiculae. Raw materials: mainly flysch, very little terra rossa, little recent marine material. Date: A.D. 78/81 - 81 19. Name: Rim No.: 1338 (120) Context: 111-112 Stamp: IMP·TRA, Imp(eratoris) Tra(iani), upside down Stamp*: ...VS.  Dim: D = 14 cm, V = 1.9 cm, HD = 5 cm, H = 6.2 cm Colour: light red (2.5YR 6/8) VC: Hard fabric, scattered rounded limestone particles, iron ore, angular voids. Thin-section: Massive inhomogeneous brownish red and darker reddish brown matrix, few various-sized holes. Non-carbonatics: very well sorted, very fine-grained (15-60 μm), medium amount, mainly quartz. Micritic carbonate: abundant, coarser-grained (40-150 μm), well sorted, few coarse (generally 250-400 μm) rounded. More micas, globigerinoids, opalic spiculae. Few forams, limonitic concretions. Raw materials: mainly flysch, very little terra rossa, little recent marine material. Date: Trajanic period 20. Name: Rim No.: 1426 (122) Context: 209? Stamp: ..P.TRA, [Im]p(eratoris) Tra(iani), upside down Dim: D = 15 cm, V = 2 cm, H = 6.1 cm Colour: reddish yellow (5YR 6/8) VC: Soft fabric, scattered fine brown, white particles. Thin-section: Massive brownish red matrix. Non-carbonatics: very well sorted, very fine-grained (15-70  μm), large amount, few coarser (100-140 μm) mainly quartz. Micritic carbonate: coarser (generally 40-200  μm), abundant, well sorted, rounded. More globigerinoids, opalic spiculae. Few micas, forams, molluscs, limonitic concretions. Raw materials: mainly flysch with very little terra rossa, little recent marine material. Date: Trajanic period 21. Name: Rim No.: 12 (116) - 26017 Context: 143 Stamp: .MP.T.., [I]mp(eratoris) T[ra](iani)

18.03.19 14:39

82

THE AMPHORAE

Dim: D = 14 cm, V = 1.9 cm, H = 5 cm Colour: light red (2.5YR 6/8) VC: Soft fabric, scattered bigger black and brown, white particles. Thin-section: Massive brownish red matrix. Non-carbonatics: very well sorted, very fine-grained (15-70  μm), medium amount, mainly quartz. Micritic carbonate: coarser-grained (40-150 μm), abundant, few large (100-400  μm), well sorted, rounded. More micas, globigerinoids. Few limonitic concretions, terra rossa, opalic spiculae. Raw materials: mainly flysch, little terra rossa, little recent marine material. Date: Trajanic period 22. Name: Rim No.: 500 (119) - A 10627 Context: 447 Stamp*: LESBI (reverse). Dim: D = 16.5 cm, V = 2.5 cm, H = 6.7 cm Colour: light red (2.5YR 6/8) VC: Very hard, micaceous, very fine fabric, scattered fine brown, white particles. Thin-section: Massive inhomogeneous brownish red and darker reddish brown matrix, few elongated holes. Non-carbonatics: very well sorted, very fine-grained (15-80  μm), large amount, few coarser (100-300  μm), mainly quartz. Micritic carbonate: coarser-grained (40-100 μm), abundant, well sorted, rounded. Few micas, globigerinoids, molluscs, limonitic concretions, terra rossa, opalic spiculae. Raw materials: mainly flysch with little terra rossa, little recent marine material. Date: Domitianic period 23. Name: Rim, neck No.: 1347 (124) Context: Room behind the press Dim: D = 15 cm, V = 2 cm, HD = 5 cm, H = 10.5 cm Colour: reddish yellow (5YR 7/6) VC: Soft, fine fabric, scattered fine brown, microfossils, few reaction rims surrounding voids which held limestone, microfossil. Date: Tiberian to Hadrianic period 24. Name: Rim, base No.: 85 (46) - 3390 and 3381 Context: Room K Dim: D = 15 cm, V = 2.2 cm, HD = 5 cm, DF = 4.5 cm Colour: red (2.5YR 5/6) VC: Hard, fine, Strange Fažana fabric shows frequent white and brownish particles, microfossils, few reaction rims surrounding voids which held limestone, microfossil. Date: Tiberian to Hadrianic period 25. Name: Rim, neck, handle No.: 104 (49) - 23032 Context: 37? Dim: D = 14.6 cm, V = 2.1 cm, HD = 5 cm, S = 3 × 3.1 cm, H = 27.5 cm Colour: red (2.5YR 5/8) VC: Soft fabric shows frequent white and brownish particles, frequent microfossils, shells. Date: Tiberian to Hadrianic period

BEZECZKY_Text.indd 82

Fažana 1, Fažana, Istria 26. Name: Rim, neck, handle No.: 1 (1) - 16286 Context: 104 ? Dim: D = 10.5 cm, V = 1.6 cm, HD = 7.6 cm, H = 14.8 cm Colour: reddish yellow (5YR 7/6) VC: Hard fabric, scattered fine brown, white particles, few reaction rims surrounding voids which held limestone. Date: last third of the second to early third century A.D. Fažana 2, Fažana, Istria 27. Name: Rim No.: 1370 (164) Context: Dolia 4 Dim: D = 10.3 cm, V = 1.4 cm, H = 10.4 cm Colour: reddish yellow (5YR 6/8) VC: Hard fabric, scattered bigger dark brown and white particles and iron ore. Date: last third of the second to end of the third century A.D.  Porto Recanati 28. Name: Rim No.: 994 (104) Context: ? Dim: D = 16 cm, V = 2.5 cm, H = 13.5 cm Colour: red (2.5YR 5/8) VC: Hard, fine, frequent microfossils, scattered brown and white particles. Date: from the early Claudian period to the second century A.D. Origin: Italy, W coast of the Adriatic 29. Name: Rim, neck, handle No.: 1419 (151) Context: ? Dim: D = 13 cm, V = 1.5 cm, HD = 8.3 cm, H = 28 cm Colour: light reddish yellow (5YR 6/6) VC: Hard, fine fabric, scattered bigger brown and white particles, different fabric from the other Porto Recanati amphora no. 28. Date: from the early Claudian period to the second century A.D. Origin: Italy, W coast of the Adriatic Forlimpopoli 30. Name: Neck, handle No.: 623 (80) Context: ? Dim: D = 9.3 cm, V = 1.7 cm, HD = 1.2 cm, H = 12.7 cm Colour: very pale brown (10YR 7/4) VC: Hard, fine fabric. Thin-section: Dark brownish grey, partly fired calcareous and matrix. Noncarbonatics: moderately sorted, abundant, very fine - fine-grained (generally 15-100  μm), few coarser (until 300  μm) mainly quartz. Much micas, strongly fired biotite. Carbonate: more fine-grained strongly fired, few large (250-700 μm) rounded silty limestone clasts (flysch).

18.03.19 14:39

CHAPTER 9 – Catalogue (Bezeczky et al.) Date: first century to the end of the third century A.D. Origin: Italy, W coast of the Adriatic 31. Name: Neck, handle No.: 1267 (170) Context: Room 102 Dim: D = 9 cm, V = 1.1 cm, HD = 1.2 cm, S = 1.9 × 5.4 cm, H = 8.1 cm Colour: buff (10YR 8/4) VC: Hard, fine fabric, small reddish inclusion, mica, small voids. Date: first century to the end of the third century A.D. Origin: Italy, W coast of the Adriatic Aquincum 78 32. Name: Rim No.: 1527 (163) Context: Room 219 Dim: D = 10 cm, V = 1 cm, HD = 2.6 cm, H = 7 cm Colour: buff (10YR 8/4) VC: Hard, very fine inclusions. Date: mid-first century to the end of the third century A.D. Origin: Istria? 33. Name: Rim No.: 1310 (127) Context: Room 221 Dim: D = 8.8 cm, V = 1.1 cm, H = 4.1 cm Colour: buff (10YR 7/3) VC: Hard, very fine, some bigger iron ore and limestone inclusions. Date: mid-first century to the end of the third century A.D. Origin: Istria?

83

Dressel 2-4 34. Name: Handle No.: 31 (15) - 1489 Context: Street, 137 Dim: S = 2.3 × 4.7 cm, H = 16.2 cm Colour: buff (7.5YR 7/4) VC: Hard, fine fabric, iron ore and limestone inclusions, frequent white reaction rims around voids, indicating where the limestone grains once existed. Origin: Italy, Adriatic region 35. Name: Handle No.: 386 (87) - 25128 Context: Room GL Dim: S = 2 × 4.5 cm, H = 16.3 cm Colour: red (2.5YR 5/6) VC: Campanian, black sand fabric. Date: first century A.D. Origin: Campania, Vesuv region 36. Name: Rim, neck, handle No.: 1308 (142) Context: Room 105/B2 Dim: D = 12 cm, V = 1 cm, S = 1.8 × 3.2 cm, H = 14.5 cm Colour: reddish yellow (5YR 6/6), SF: reddish yellow (5YR 7/6) VC: Hard, very fine fabric. Red iron ore, brownish inclusions, irregularly fired. Date: first to third century A.D. Origin: Italy, Adriatic region

SPANISH Dressel 7-11

Dressel 14A

37. Name: Rim No.: 455 (66) - 9251 Context: Room A, 111, Dim: D = 15.5 cm, V = 3.5 cm, HD = 2.8 cm, H = 4.2 cm Colour: buff (7.5YR7/4), SF: white (2.5Y 8/2) VC: Cádiz fabric, sandy, hard. Date: second half of the first century A.D. Origin: Cádiz region

39. Name: Rim, neck, handle No.: 1341 (157) Context: Room 231 Q Dim: D = 19.8 cm, V = 2.2 cm, S = 4 × 6.6 cm, H = 6.1 cm Colour: buff (7.5YR 7/4) VC: Hard and slightly coarse fabric. Date: late first and second century A.D. Origin: Málaga?

Beltran IIA

Dressel 20

38. Name: Base No.: 10 (6) - 16027 Context: Room?, 156 Graffito: V (…), post cocturam Dim: DF = 3.8 cm, V = 0.8 cm, H = 9.1 cm Colour: buff (5YR 7/2) VC: Hard, fine fabric, fine red, colourless and white inclusions, mica. Date: second half of the first century A.D. and second century A.D. Origin: Cádiz region

40. Name: Handle No.: 1271 (125) Context: Room 215 Stamp: palm branch M(…), ((ramus palmae)) M[---]. Dim: S = 4.3 cm, H = 12.3 cm Colour: light reddish brown (5YR 6/4) VC: Guadalquivir, very hard, coarse. Date: first half of the second century A.D. The palm branch was in vogue during Hadrian’s period. Martin-Kilcher 1987, 94, ST 2a, MAF ((ramus palmae)) on Antoninian handle (Hadrian or Antoninus Pius). Origin: Guadalquivir Valley

BEZECZKY_Text.indd 83

18.03.19 14:39

84

THE AMPHORAE

41. Name: Body, base No.: 1554 (115) Context: ? Graffito: Asiatici / sec( ) x k(alendas) mar(tias) Alb( ) + II, ante cocturam Dim: DB = 52 cm, V = 2.5 cm, H = 58.5 cm Colour: light brown to light grey (7.5YR 6/4 to 6/2) VC: Hard, medium-grained Guadalquivir fabric. Date: second third of the second century A.D.  Origin: Guadalquivir Valley 42. Name: Rim No.: 1378 (156) Context: Room 204 Dim: D = 20 cm, V = 4.4 cm, H = 4.5 cm Colour: light brown (7.5YR 6/4) VC: Guadalquivir, very hard, coarse fabric. Date: second third of the second century A.D.  Origin: Guadalquivir Valley 43. Name: Rim, neck No.: 923 (89) Context: Room 220 Dim: D = 16 cm, V = 3.8 cm, HD = 3.2 cm, H = 5 cm Colour: light reddish brown (5YR 6/3) VC: Guadalquivir, very hard, coarse fabric. Date: end of the second-beginning of the third century A.D.  Origin: Guadalquivir Valley 44. Name: Rim, neck No.: 1391 (126+201) other part 1375 box Context: Room 233, Garden Dim: D = 19 cm, V = 3 cm, H = 8.2 cm Colour: reddish yellow (5 YR 7/6). VC: Guadalquivir, hard, coarse fabric. Big quartz inclusions. Date: mid-third century A.D.  Origin: Guadalquivir Valley 45. Name: Rim No.: 1407 (158) Context: Room 234 Dim: D = 15 cm, V = 2.7 cm, H = 5 cm Colour: reddish yellow (7.5YR 6/6) VC: Hard, fine, sandy fabric. Date: third century A.D. Origin: Guadalquivir Valley Almagro 50A/ Keay XXII 46. Name: Base No.: 626 (81) Context: ? Dim: DF = 4.9 cm, H = 8.8 cm Colour: buff (5YR 7/4) VC: Lusitanian very fine fabric, reddish, dark brown and colourless inclusions. Date: fourth and fifth century A.D. Origin: Lusitania

BEZECZKY_Text.indd 84

Almagro 51A-B 47. Name: Rim No.: 140 (52) - 25940 Context: Courtyard, 106 Dim: D = 8.9 cm, V = 1.3 cm, HD = 2.8 cm, H = 4 cm Colour: red (2.5YR 5/6), SF: buff (7.5YR 7/4) VC: Lusitania (Tejo/Sado or Algarbe) hard, coarse fabric, reddish, dark brown and colourless inclusions. The microscope shows frequent white reaction rims around voids, indicating where limestone grains once existed. Coated. Date: late fourth and fifth century A.D. Origin: Lusitania 48. Name: Rim No.: 140 (168) - 25939 Context: Courtyard, 106 Dim: D = 8 cm, V = 0.9 cm, HD = 0.8 cm, H = 7.1 cm Colour: buff (7.5YR 7/4) VC: Lusitania (Tejo/Sado or Algarbe) soft, fine, sandy fabric, fine reddish inclusion, some voids. Date: fifth century A.D.  Origin: Lusitania Almagro 51C 49. Name: Rim, neck, handle No.: 34 (21) - 16159 Context: 36? Dim: D = 11 cm, V = 1.5 cm, H = 4.3 cm Colour: dark buff (7.5YR 7/2) VC: Lusitanian hard, coarse, sandy fabric, red and colourless inclusions. Irregularly fired. Date: fourth century A.D.  Origin: Lusitania Beltran 68 50. Name: Rim, neck, handles No.: 1221 (60) Context: Courtyard Dim: D = 12 cm, V = 1.6 cm, HD = 3.2 cm, S = 2.9 × 5.2 cm, H = 15 cm Colour: red (2.5YR 5/4) VC: Hard, very fine-grained sandy fabric common with quartz inclusions, limestone or voids that once contained this material. Thin-section: Ca-rich matrix. Very abundant, moderately sorted inclusions (