Transport Stirrup Jars of the Bronze Age Aegean and East Mediterranean (Prehistory Monographs 33) [Illustrated] 9781931534628, 1931534624

Table of contents :
List of Illustrations in the Text
List of Tables
List of Graphs
List of Figures
List of Plates
Acknowledgments
List of Abbreviations
1 Introduction
2 Typology
3 Scientific Backgound and Aims of the Analyses
4 Chemical Analyses
5 Petrographic Analyses
6 Interpretation of the Chemical and Petrographic Data
Chemistry, Petrography, and Typology: Geographical Associations
8 The Linear B Inscriptions
9 Chronology and Power
10 Trade
11 Catalog
Appendix
References
Concordance 1
Concordance 2
Index
General Index

Citation preview

Transport Stirrup Jars of the Bronze Age Aegean and East Mediterranean

Frontispiece. Inscribed transport stirrup jar (KOU01) from Kourion © The Trustees of the British Museum, London.

PREHISTORY MONOGRAPHS 33

Transport Stirrup Jars of the Bronze Age Aegean and East Mediterranean

by Halford W. Haskell, Richard E. Jones, Peter M. Day, and John T. Killen

Published by INSTAP Academic Press Philadelphia, Pennsylvania 2011

Design and Production INSTAP Academic Press Printing CRWGraphics, Pennsauken, New Jersey Binding Hoster Bindery, Inc., Ivyland, Pennsylvania

Library of Congress Cataloging-in-Publication Data Transport stirrup jars of the Bronze Age Aegean and east Mediterranean / by Halford W. Haskell ... [et al.]. p. cm. -- (Prehistory monographs ; 33) Includes bibliographical references and index. ISBN 978-1-931534-62-8 (hardback : alk. paper) 1. Pottery, Prehistoric--Aegean Sea Region. 2. Bronze age--Aegean Sea Region. 3. Aegean Sea Region--Antiquities. 4. Excavations (Archaeology)--Aegean Sea Region. 5. Pottery, Prehistoric--Mediterranean Region. 6. Bronze age--Mediterranean Region. 7. Mediterranean Region--Antiquities. 8. Excavations (Archaeology)--Mediterranean Region. I. Haskell, Halford W. GN799.P6.T73 2011 666'.68--dc23 2011033494

Copyright © 2011 INSTAP Academic Press Philadelphia, Pennsylvania All rights reserved Printed in the United States of America

To Hector Catling, to whom we are delighted to dedicate this volume, we offer our greatest thanks for his long-standing guidance and encouragement.

Table of Contents

List of Illustrations in the Text. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi List of Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii List of Graphs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv List of Figures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix List of Plates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxv List of Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxvii Chapter 1. Introduction, Halford W. Haskell, Richard E. Jones, and Peter M. Day. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1. Transport Stirrup Jars in Late Bronze III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2. Definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3. Construction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4. Decoration and Firing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.5. Hardness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.6. Distribution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.7. SJ Contents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.8. History of Scholarship. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.9. Current Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

viii

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

Chapter 2. Typology, Halford W. Haskell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2. Development of SJ Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.3. Typological Groups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Chapter 3. Scientific Backgound and Aims of the Analyses, Richard E. Jones and Peter M. Day. . . . . . . . . . . . . . . 23 3.1. History of Stirrup Jar Analysis and Methodology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.2. Aims of the New Analytical Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Chapter 4. Chemical Analyses, Richard E. Jones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 4.1. Sampling Strategy and Procedures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 4.2. Analytical Procedures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 4.3. Performance Characteristics of AAS and Comparability with OES. . . . . . . . . . . . . . . . . . . . . . . . . . . 30 4.4. Reference Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 4.5. Data Treatment and Classification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 4.6. Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Chapter 5. Petrographic Analyses, Peter M. Day. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.2. Petrographic Fabric Groupings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 5.3. Summary of Petrographic Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Chapter 6. Interpretation of the Chemical and Petrographic Data, Richard E. Jones and Peter M. Day. . . . . . . . . . . . . . 79 6.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 6.2. Chemical Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 6.3. Petrographic Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Chapter 7. Chemistry, Petrography, and Typology: Geographical Associations, Halford W. Haskell, Richard E. Jones, and Peter M. Day. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 7.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 7.2. Overall Pattern. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Chapter 8. The Linear B Inscriptions, John T. Killen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 8.1. The Material. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 8.2. Chemical Analysis Groupings versus Inscription Groupings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 8.3. Differences among Chemical Analysis Classes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 8.4. Implications of the Results for Cretan Geography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 8.5. Addendum: Inscriptions on Fine Ware Vessels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Chapter 9. Chronology and Power, Halford W. Haskell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 9.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 9.2. Sites and Chronology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 9.3. Summary of Dating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 9.4. Power. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 9.5. General Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

TABLE OF CONTENTS

ix

Chapter 10. Trade, Halford W. Haskell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 10.1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 10.2. Linear B Evidence. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 10.3. Historical Implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 10.4. General Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Chapter 11. Catalog. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Appendix A. LM/LH III Transport Class Stirrup Jars. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 Concordance 1. Sampled ISJs and Other Linear B Vases Ordered by Inscription Number. . . . . . . . . . . . . . . . . . . . 177 Concordance 2. ISJs and Other Linear B Vases Ordered by Catalog Number. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Index of References to Sampled Pieces in Chapters 1–10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 General Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Tables Graphs Figures Plates

List of Illustrations in the Text

Illustration 1.1.

KN32 (left) and TH05 (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Illustration 1.2.

Cap (MYC28#) on spout of MYC27 (left). Sealed cap and plug (right), House of the Oil Merchant, Mycenae. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Illustration 1.3.

Overlapping sections of base (interior shown at bottom) and body (exterior shown at top) of MYC13. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Illustration 1.4.

Black square dots indicate distribution of transport stirrup jars in the central and eastern Mediterranean. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Illustration 1.5.

Distribution of ISJs in the Aegean as indicated by site abbreviations. . . . . . . . . . . . . . . . . . . . . . . 4

Illustration 2.1.

PIG02 with small ring handle on shoulder at left. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Illustration 2.2.

PIG01 with horn on spout. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Illustration 2.3.

Typegroup Ia (MYC46). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Illustration 2.4.

Typegroup Ib (MYC78). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Illustration 2.5.

Typegroup IIa MYC89 (left) and MYC15 (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Illustration 2.6.

Typegroup IIb MYC72 (left) and MYC16 (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Illustration 2.7.

Typegroup IIc MYC74 (left) and MYC31 (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Illustration 2.8.

Typegroup III (TH07). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

xii

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

Illustration 2.9.

Typegroup IVa (MYC83). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Illustration 2.10.

Typegroup IVb (MYC50). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Illustration 2.11.

Typegroup VIII (TH18). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Illustration 2.12.

Typegroup IX (TH08). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Illustration 2.13.

Typegroup X (MYC52). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Illustration 2.14.

Typegroup XIa (MYC24). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Illustration 2.15.

Typegroup XIb (MYC20). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Illustration 2.16.

Detail photo of MYC20 showing a deep depression in the disc. . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Illustration 2.17.

Typegroup XIII (TH25). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Illustration 2.18.

Typegroup XIV (TH31). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Illustration 2.19.

Typegroup XV (TH37). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Illustration 2.20.

Typegroup XVI (TH53). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Illustration 2.21.

Typegroup XVII (TH55). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Illustration 2.22.

Typegroup XVIII (TH03). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Illustration 2.23.

Typegroup XIX (TH33). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Illustration 2.24.

Typegroup XX (KN05). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Illustration 2.25.

Typegroup XXI (KN32). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Illustration 2.26.

Typegroup XXIIa (ZYG04). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Illustration 2.27.

Typegroup XXIIb (ZYG10, left) and (ZYG09, right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Illustration 2.28.

Examples from the broader Groups A (top row), B (second and third row), D (fourth row), and E (fifth row). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Illustration 2.29.

Detail of x incised into surface of disc of KN05. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Illustration 2.30.

Detail of the pierced handle of TH04. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Illustration 8.1.

ISJs listed by site and chemical group. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

List of Tables

Table 1.

Catalog numbers of analyzed samples according to region and findspot.

Table 2.

Characteristics of the broad typological Groups A, B, D, and E.

Table 3.

SJs listed by typegroups, including pieces not analyzed but attributed on typological grounds.

Table 4.

SJs listed by broad group.

Table 5.

Composition characteristics of the Lefkandi brick standard as determined by the British Ceramic Research Association (BCRA), the Fitch Laboratory, and laboratories in Venice and Berlin.

Table 6.

Comparability of OES and AAS. OES1 (Catling and Jones 1977), drilled sample. OES2 (Catling and Millett 1965 but composition corrected according to factors in Jones 1986, table 2.5), same drilled sample as OES1. OES3 (WCISJ), new drilled sample. AAS (present study), fragment.

Table 7a.

Chemical compositions of the SJs, stoppers, and miscellaneous samples determined by AAS and expressed as percentage of element oxide.

Table 7b.

Chemical compositions of the SJs determined by OES, published in WCISJ and expressed as percentage of element oxide.

Table 7c.

Chemical (OES) compositions of the SJs from Thebes published by Catling and Millett (1965) and a SJ from Amnisos, previously unpublished. All compositions have been corrected according to factors in Jones (1986b, table 2.5) and are expressed as percentage of element oxide.

Table 8.

Reference material analysed by OES arranged according to region, findspot, type and date of pottery, and publication.

xiv

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

Table 9.

AAS reference material listed geographically (north to south in Greece and then Cyprus through Egypt).

Table 10.

Descriptions of reference material analyzed by OES or AAS, whose composition characteristics are presented in Table 11.

Table 11.

Composition characteristics of the reference material expressed as mean of the percentage of element oxide concentrations.

Table 12.

Mycenae: fine and coarse-textured pottery (CP = cooking pot, bs = body sherd); compositions determined by AAS and expressed as percentage of element oxides.

Table 13.

Chania: coarse and plain wares.

Table 14.

Hagios Giorgios: reference pottery and its compositional characteristics by AAS, expressed as percentage of element oxides.

Table 15.

Chemical work by NAA (unless otherwise stated) on Minoan pottery published since 1987 and arranged according to pottery date.

Table 16.

Three groups of SJs (including Linear-B-inscribed non-SJ vases and miscellaneous samples) isolated by principal components analysis (PCA).

Table 17.

Provenance of the main fabrics.

Table 18.

Composition characteristics of reference groups for relevant sites/regions.

Table 19.

Associations between SJs (and miscellaneous samples) and reference groups.

Table 20.

Associations for jars analyzed (by OES) in WCISJ.

Table 21.

Associations for those jars analyzed (by OES) by Catling and Millett (1965) and Catling and Jones (1977).

Table 22.

Other petrographic analyses.

Table 23.

Caps/plugs/stoppers analyzed by AAS, NAA, and petrography.

Table 24.

Classification of the inscribed jars associated with West Crete.

Table 25.

Other chemical analyses of SJs and stoppers.

Table 26.

The fabric groups, their suggested areas of provenance, and chemical associations.

Table 27.

Summary of typological, chemical, and petrographic results, and suggested associations.

Table 28.

Associations of typegrouped jars.

Table 29.

Associations of grouped jars.

Table 30.

Tabular summary of typological characteristics associated with each production zone.

List of Graphs

Graphs 1a–f.

Typegroup proportions: H(MD)/H and D/H.

Graphs 2a–e.

Typegroup proportions: H(MD)/H and D/H.

Graphs 3a–d.

Group proportions H(MD)/H and D/H.

Graph 4.

Comparison of major and minor element contents as determined at BCRA, Venice, and Berlin (all XRF), and RLAHA and FL (both AAS).

Graph 5a.

TH05: comparison of compositions (% oxide) determined by OES1 (striped), OES2 (gray), and AAS (black).

Graph 5b.

TH05: comparison of compositions (% oxide) determined by OES1 (white), OES2 (gray), and AAS (black).

Graph 6a.

Comparison of the mean and the 1-standard-deviation ranges of the group compositions of LH IIIB pottery from Myceanae.

Graph 6b.

Comparison of the mean and the 1-standard-deviation ranges of the group compositions of LH IIIB pottery from Thebes.

Graph 7.

AAS data of magnesium and chromium (left) and magnesium and calcium (right) oxide plots for fine (F), coarse (C with inverted triangle) and heavy (H with triangle) wares at Mycenae.

Graph 8.

AAS data of chromium and magnesium (left) and magnesium and calcium (right) oxide plots for plain (P) and coarse (C) wares from Chania.

Graph 9.

The distribution of Mn oxide contents in the reference material from Mycenae (left) and Chania (right).

xvi

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

Graph 10.

Discriminant analysis plot of composition groups in Crete representing Knossos (K full circle), Chania (plain [full square] and coarse [open triangle] wares), and Hagios Giorgios (HG open square).

Graph 11.

Discriminant analysis of AAS reference groups for Thebes (T open square), Mycenae (M triangle), Knossos (K open circle), and Chania (C full square).

Graph 12a.

Mg-Cr plot for OES data of the reference groups: Routsi (triangle); Menelaion, Sparta (inverted triangle); Mycenae (+); Knossos (circle).

Graph 12b.

Mg-Ca plot for OES data of the reference groups: Routsi (triangle); Menelaion, Sparta (inverted triangle); Mycenae (+); Knossos (circle).

Graph 13.

Element (oxide) distributions for all SJs.

Graph 14.

Principal component (PC) plot of the total SJ data set (excluding caps, stoppers, and plugs), suggesting two large overlapping groups, I and II, and a small group III (left).

Graph 15.

Principal component (PC) plot of the total SJ data set using Mg, Cr, Ni, and Ca contents only, revealing four potential groupings as shown with arrows.

Graph 16a.

Cr-Mg oxide plot of the total SJ data set.

Graph 16b.

Ca-Mg oxide plot of the total SJ data set.

Graph 17a.

Cr-Mg oxide plot for SJs found on Crete and in the Dodecanese.

Graph 17b.

Ca-Mg oxide plot for SJs found on Crete and in the Dodecanese.

Graph 18.

SJs found at Chania and the Rethymnon area.

Graph 19a.

Cr-Mg oxide plot for SJs found in Central Crete.

Graph 19b.

Ca-Mg oxide plot for SJs found in Central Crete.

Graph 19c.

Cr-Mg oxide plot for SJs at Knossos.

Graph 19d.

Cr-Mg oxide plot for SJs found at Kommos and Malia.

Graph 20a.

Cr-Mg oxide plot from SJs found in East Crete and the Dodecanese.

Graph 20b.

Ca-Mg oxide plot from SJs found in East Crete and the Dodecanese.

Graph 21a.

Ca-Mg oxide plot for SJs found in the Peloponnese.

Graph 21b.

Cr-Mg oxide plot for SJs found at Mycenae (ex. MYC65).

Graph 21c.

Cr-Mg oxide plot highlighting SJs found at Pylos and Sparta, Menelaion.

Graph 21d.

Cr-Mg oxide plot highlighting SJs found at Tiryns (triangle).

Graph 21e.

Cr-Mg oxide plot highlighting SJs found at Zygouries and Kythera.

Graph 21f.

Cr-Mg oxide plot for caps, stoppers, and plugs at Mycenae.

Graph 21g.

Ca-Mg oxide plot for caps, stoppers, and plugs at Mycenae.

Graph 22a.

Cr-Mg oxide plot for SJs found in central Greece.

Graph 22b.

Ca-Mg oxide plot for SJs found in cental Greece.

Graph 22c.

Cr-Mg oxide plot highlighting SJs found at Thebes.

Graph 22d.

Cr-Mg oxide plot highlighting SJs found in central Greece apart from Thebes.

LIST OF GRAPHS

xvii

Graph 23a.

Cr-Mg oxide plot for SJs found on Cyprus.

Graph 23b.

Cr-Mg oxide plot for SJs found in the Levant, Egypt, and Sardinia.

Graph 24a.

Mg-Cr oxide plot for the reference compositions at Knossos (inverted triangle), Chania plain ware (circle), and Chania coarse ware (open circle).

Graph 24b.

Mg-Ca oxide plot for the reference compositions at Knossos (inverted triangle), Chania plain ware (circle), and Chania coarse ware (open circle).

Graph 25a.

Mg-Cr oxide plot of OES data of the reference compositions for Knossos (full circle), Kommos (open circle), Palaikastro (triangle), and Chania (inverted triangle).

Graph 25b.

Mg-Ca oxide plot of OES data of the reference compositions for Knossos (full circle), Kommos (open circle), Palaikastro (triangle), and Chania (inverted triangle).

Graph 26.

Mg-Cr oxide plot of AAS data for the reference compositions for Knossos (full circle), Mycenae (+), Thebes (open triangle), and Chania (inverted triangle).

Graph 27.

Mg-Ca oxide plot AAS data for the reference compositions for Knossos (full circle), Mycenae (+), Thebes (open triangle), and Chania (inverted triangle).

Graph 28a.

Mg-Cr oxide plot of AAS data for the reference compositions for Rhodes (+), Knossos (full circle), Thebes (open triangle), and Chania (inverted triangle).

Graph 28b.

Mg-Ca oxide plot of AAS data for the reference compositions for Rhodes (+), Knossos (full circle), Thebes (open triangle), and Chania (inverted triangle).

Graph 29a.

Mg-Cr oxide plot of AAS data for the reference compositions for Maroni (triangle) and Kalavassos (inverted triangle) on Cyprus, superimposed on the compositions presented in Graph 28.

Graph 29b.

Mg-Ca oxide plot of AAS data for the reference compositions for Maroni (triangle) and Kalavassos (inverted triangle) on Cyprus, superimposed on the compositions presented in Graph 28.

Graph 30.

Mg-Cr oxide plot for the reference samples at Mycenae (coarse: inverted triangle; fine: full circle; “heavy”: triangle) and MYC02LinB, MYC03, MYC06, MYC07, MYC09, MYC10, MYC12, and MYC17.

List of Figures

Figure 1.

Stirrup jars from Cape Gelidonya and Ialysos.

Figure 2.

Stirrup jars from Ialysos.

Figure 3.

Stirrup jars from Knossos.

Figure 4.

Stirrup jars from Knossos and Kommos.

Figure 5.

Stirrup jars from Malia and Mycenae.

Figure 6.

Stirrup jars from Mycenae.

Figure 7.

Stirrup jars from Mycenae.

Figure 8.

Stirrup jars from Mycenae.

Figure 9.

Stirrup jars from Mycenae and Karpathos (Pigadia).

Figure 10.

Stirrup jars from Karpathos (Pigadia) and Sparta Menelaion.

Figure 11.

Stirrup jars from Thebes.

Figure 12.

Stirrup jars from Thebes.

Figure 13.

Stirrup jars from Thebes.

Figure 14.

Stirrup jars from Thebes.

Figure 15.

Stirrup jars from Thebes.

xx

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

Figure 16.

Stirrup jars from Thebes and Uluburun.

Figure 17.

Stirrup jars from Uluburun and Zygouries.

Figure 18.

Stirrup jar from Zygouries.

List of Plates

Plate 1.

Stirrup jars from Athens, Cyprus, Dhenia, and Eleusis.

Plate 2.

Stirrup jars from Episkopi, Bamboula, and Ialysos.

Plate 3.

Stirrup jars from Ialysos.

Plate 4.

Stirrup jars from Ialysos.

Plate 5.

Stirrup jars from Ialysos and Knossos.

Plate 6.

Stirrup jars from Knossos.

Plate 7.

Stirrup jars from Knossos.

Plate 8.

Stirrup jars from Knossos.

Plate 9.

Stirrup jars from Kourion, Minet el Beida, and Mycenae.

Plate 10.

Stirrup jars from Mycenae.

Plate 11.

Stirrup jars, caps, and plug from Mycenae.

Plate 12.

Stirrup jars from Mycenae.

Plate 13.

Stirrup jars from Mycenae.

Plate 14.

Stirrup jars from Mycenae.

Plate 15.

Stirrup jars and sherd from Mycenae.

xxii

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

Plate 16.

Stirrup jars from Mycenae.

Plate 17.

Stirrup jars from Mycenae and Karpathos (Pigadia).

Plate 18.

Stirrup jars from Karpathos (Pigadia) and Palaikastro.

Plate 19.

Stirrup jars from Pylos.

Plate 20.

Stirrup jars from Pylos and Thebes.

Plate 21.

Stirrup jars from Thebes.

Plate 22.

Stirrup jars from Thebes.

Plate 23.

Stirrup jars from Thebes.

Plate 24.

Stirrup jars from Thebes.

Plate 25.

Stirrup jars from Thebes.

Plate 26.

Stirrup jars from Thebes.

Plate 27.

Stirrup jars from Thebes.

Plate 28.

Stirrup jars from Thebes.

Plate 29.

Stirrup jars from Thebes.

Plate 30.

Stirrup jars from Thebes and Zygouries.

Plate 31.

Variability in Fabric 1 SJs, with frequent quartzite, schist, and phyllite from Chania (KH09, KH11, KH12, KH19, KH22, KH39, KH40). Other examples of Fabric 1 SJs from Gla (GLA02), Knossos (KN12), Kommos (KO25), Malia (MA01), and Mycenae (MYC99).

Plate 32.

Examples of Fabric 1 SJs from Mycenae (MYC101), Thebes (TH86), and Uluburun (ULB07). Early Bronze Age ceramics, comparable to Fabric 1: a. Chania Kastelli 00/51; b. Chania Kastelli 00/110; c. Debla 00/11; d. Mitatoulia 00/18. Geological clay samples from Chania area (comparable to Fabric 1): e. S6B/2001; f. S6A/2001; g. S12A/2001; h. S12B/2001. Fabric 2 SJ with frequent quartzite, schist, phyllite, and common microfossils from Chania (KH52).

Plate 33.

Fabric 2 SJ with frequent quartzite, schist, phyllite, and common microfossils from Mycenae (MYC83). Fabric 3 SJ from Chania (KH38). Fabric 4 SJs with metamorphic rocks and microfossils from Hala Sultan Teke (HST01m), Chania (KH48, KH50, KH57), and Knossos (KN20). Fabric 5 SJ with cataclastic rock fragments from Thebes (TH01). Fabric 6 with volcanic rock fragments from Uluburun (ULB05). Fabric 7 SJs with chert from Uluburun (ULB06) and Mycenae (MYC87). Fabric 8 SJ with chert and serpentine from Enkomi (ENK12).

Plate 34.

Fabric 8 SJs with chert and serpentine from Enkomi (ENK13), Knossos (KN09), Kommos (KO16), Malia (MA05), Thebes (TH31), and Uluburun (ULB03). Fabric 9 SJs with siltstone, igneous rocks, and chert from Iria (IR02), Kommos (KO13), and Mycenae (MYC59). Fabric 10 SJs with siltstone, igneous rocks, chert, and microfossils from Iria (IR05, IR07) and Uluburun (ULB01).

Plate 35.

Fabric 10 SJ with siltstone, igneous rocks, chert, and microfossil from Uluburun (ULB08). Fabric 11 SJs with rounded schist, phyllite, siltstone, and altered volcanic rock from Kommos (KO06, KO21, KO22) and Mycenae (MYC43). Fabric 12 SJ dark siltstone with polycrystalline quartz from Chania (KH42). Fabric 13 SJ with red siltstone from Chania (KH56). Fabric 14 SJs with dark mudstone and siltstone from Chania (KH07, KH28, KH55), Malia (MA06), and Mycenae (MYC20).

LIST OF PLATES

Plate 36.

xxiii

Fabric 14 SJ with dark mudstone and siltstone from Mycenae (MYC66). Fabric 15 SJs with metamorphic rocks, altered volcanic stone, and calcimudstone from Chania (KH47) and Kommos (KO09). Fabric 16 SJs with clay pellets from Chania (KH51) and Mycenae (MYC38). Fabric 17 SJ with clay pellets from Mycenae (MYC104). Fabric 18 with coarse igneous fragments from Kommos (KO05m). Fabric 19 SJ, fine calcitic from Mycenae (MYC39). Fabric 20 SJ with frequent quartz from Chania (KH32). Fabric 21 SJ from Chania (KH01). Fabric 23 SJ with biotite schist from Kommos (KO08). Fabric 24 SJ from Kythera (KY01).

Acknowledgments

The study of Aegean coarse ware stirrup jars presented in this volume has several origins. Aspects of these jars formed the basis of Haskell’s doctoral dissertation and subsequent research. Following the publication in 1980 of Catling et al.’s study of the Linear B inscribed jars that included their chemical and inscriptional analysis, Haskell, Catling, Jones, and Killen planned a larger program of analysis. This was accomplished in the Fitch Laboratory at the British School at Athens between 1986 and 1991. The petrographic element of the study became possible in the later 1980s as Day, at that time Fellow in Ceramic Petrology in the Fitch Laboratory, had, with the aid of the Michael Ventris Award, completed a re-examination of a number of coarse ware stirrup jars found at Mycenae originally studied by John Riley (1980). With L. Joyner, Day carried out the petrographic analyses at the Department of Archaeology at Sheffield University. Financial support for the project has been provided from several sources: the University of London’s Institute of Classical Studies Michael Ventris Award for Mycenaean Studies to Day in 1990; and the National Endowment for the Humanities and Fulbright Commission at the American School of Classical Studies at Athens, and the Cullen Faculty Development Fund of Southwestern University, to Haskell. We have many acknowledgments to record. At the institutional level, we are indebted to the British School at Athens, its Council and Fitch Laboratory Committee; the Archaeological Service of the Greek Ministry of Culture; and the American School of Classical Studies. For permission to sample material in Greece, we thank the following: several Ephorates of Prehistoric and Classical Antiquities—1st (Athens), 2nd (Kythera), 4th (Nauplion), 5th (Sparta), 7th (Olympia), 9th (Thebes), 22nd (Rhodes), 23rd (Herakleion), and 25th (Chania); the National Archaeological Museum, Athens; Dr. K. Demakopoulou (Mycenae); Prof. S. Iakovidis (Gla); Dr. M. Vlazaki (Chania); Dr. Y. Tzedakis and Ms. E. Padopoulou (Armenoi).

xxvi

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

Dr. E. Hallager, Dr. B. Hallager, Prof. V. Watrous, Prof. J. Shaw, Dr. C. Morris, and Dr. A. Farnoux assisted us with the sampling at sites on Crete; Dr. E.B. French at Mycenae; Dr. T. Marketou on Rhodes; and the late Prof. P. Åström at Hala Sultan Tekke. For permission to study, photograph, and draw jars in Greece, and for assistance with this work, we thank the following: Dr. E.B. French at Mycenae; Dr. H.W. Catling and the late E.A. Catling (for drawings) at The Menelaion, Sparta; Dr. C.K. Williams II at Corinth; Dr. J. McK. Camp at Athens (Agora); the late Dr. S.A. Immerwahr at Athens (Agora); and Dr. V. Aravantinos at Thebes. We thank also the Department of Greek and Roman Antiquities, British Museum (London); the Department of Antiquities, Ashmolean Museum (Oxford) and Prof. M. Vickers; the Department of Antiquities of Cyprus and Prof. V. Karageorghis; the Department of Oriental Antiquities, Musée de Louvre (Paris), and Dr. A. Caubet; Dr. L. Vagnetti and the late Dr. L. Ceruti (Antigori); Prof. E. de Miro (Cannatello); San Antonio Museum of Art and Dr. C. Picón; the University of Pennsylvania Museum of Archaeology and Anthropology (Philadelphia) and the late Dr. G.R. Edwards; Prof. M. Benzi (Ialysos); l’Ecole françiase d’Athènes and Prof. J. Driessen (Malia). For access to Iria jars, we thank Prof. Y. Lolos, Mr. Ch. Pennas, and the late Mr. N. Tsouchlos. As regards the Gelidonya wreck, for permissions, information, and support we thank Prof. G. Bass; for the Uluburun wreck, Profs. C. Pulak and J. Rutter. We are very grateful to Prof. C. Knappett for allowing us access to his internal report on stirrup jars at Malia and the thin sections of his study. At the British School at Athens, Jones is grateful to Ms. E. Louka for considerable technical assistance in the Fitch Laboratory, and Ms. H. Clark for administrative assistance. At Southwestern University, Ms. M. Bliss provided invaluable assistance with the digitizing of photographs and drawings, and Ms. K. Lessard provided additional assistance beyond the call of duty. We are grateful to our respective institutions for their support and patience: Classics Program, Southwestern University, Texas (Haskell); Department of Archaeology, Glasgow University (Jones); Department of Archaeology and Prehistory, Sheffield University (Day); and Faculty of Classics, Cambridge University (Killen). For discussion and advice, our thanks go to Penelope Mountjoy, Lisa French, Erik and Birgitta Hallager, Nicolle Hirschfeld, Hans Mommsen, Jonathan Tomlinson, Vassilis Kilikoglou, Evangelia Kiriatzi, Jeremy Rutter, John Bennet, and Caroline Jackson. Jones thanks Gerwulf Schneider and Lorenzo Lazzarini for access to their data for the Lefkandi clay standard. And finally, to Hector Catling, to whom we are delighted to dedicate this volume, we offer our greatest thanks for his long-standing guidance and encouragement.

List of Abbreviations

Chronology CM

Cypro-Minoan

EIA

Early Iron Age

EM

Early Minoan

MM

Middle Minoan

LB

Late Bronze Age

LC

Late Cycladic

LH

Late Helladic

LM

Late Minoan

ProBA

Protohistoric Bronze Age

Chemical Analyses AAS

atomic absorption spectrometry

ALCA

average link cluster analysis

CV

coefficient of variation

DA

discriminant analysis

DF

discriminant function

ICP-ES

inductively coupled plasma–emission spectrometry

MS

mass spectrometry

NAA

neutron activation analysis

OES

optical emission spectroscopy

PCA

principal components analysis

WCA

wet chemical analysis

XRF

X-ray fluorescence analysis

Findspots of Cataloged Analyzed Samples AKM

Akanthou, Moulos

AM

Amnisos

ANG

Angeliana

ANT

Antigori

AR

Armenoi

ATH

Athens

xxviii

TRANSPORT STIRRUP JARS OF THE BRONZE AGE AEGEAN AND EAST MEDITERRANEAN

HISKP

Helmholz Institut für Strahlen und Kernphysik der Universität Bonn

NCSR

National Centre for Scientific Research “Demokritos,” Athens

RLAHA

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

CAN

Cannatello

CYP

Cyprus

DHE

Dhenia

EL

Eleusis

ENK

Enkomi

EP

Episkopi, Bamboula

GEL

Gelidonya

GLA

Gla

HST

Hala Sultan Tekke

IAL

Ialysos

IR

Iria

KAZ

Kazaphani

KH

Chania

KN

Knossos

KO

Kommos

KOU

Kourion

KY

Kythera

LK

Lefkandi

MA

Malia

MAM

Mameloukas Cave

MIN

Minet el Beida

MYC

Mycenae

OR

Orchomenos

Petrographic Analysis

PIG

Karpathos (Pigadia)

PE

petrographic examination

PK

Palaikastro

PPL

plane-polarized light

PYLA

Pyla, Kokkinokremos

XP

crossed polars

PYL

Pylos

RS

Ras Shamra

Pottery Typology

SID

Sidmant

LoD

light-on-dark

SM

Sparta Menelaion

DoL

dark-on-light

TAH

Tell Abu Hawam

dwl

deep wavy line (octopus derivative)

TH

Thebes

ddwl

TI

Tiryns

double deep wavy line (octopus derivative)

ULB

Uluburun

SJ

transport stirrup jar

ZYG

Zygouries

ISJ

Linear B inscribed stirrup jar

Measurements avg.

average

ca.

approximately

D

diameter

est.

estimated

g

gram

H

height

m

meter

µ

micron

MD

maximum diameter

mg

milligram

mm

millimeter

µm

micrometer

ppm

parts per million

Laboratories

References

BCRA

British Ceramic Research Association, Stoke-on-Trent

CIV

FL

Fitch Laboratory, British School at Athens

Sacconi, A. 1974. Corpus delle iscrizioni vascolari in lineare B (Incunabula Graeca 57), Rome.

LIST OF ABBREVIATIONS

FS

Furumark (1941b) Shape

FM

Furumark (1941b) Motif

VIP

Raison, J. 1968. Les vases à inscriptions peintes de l’âge mycénien et leur con texte archéologique (Incunabula Graeca 19), Rome.

WCISJ

Catling, H.W., J.F. Cherry, R.E. Jones, J.T. Killen. 1980. “The Linear B Inscribed Stirrup Jars and West Crete,” BSA 75, pp. 49–113.

Special Catalog Number Formats AKM01

Bold indicates cataloged SJ.

ATH01

Italicized catalog number indicates SJ illustrated by a photograph in the plates.

GEL02

Underlined catalog number indicates SJ illustrated by a drawing in the figures.

IAL02

Italicized and underlined catalog number indicates SJ illustrated by both a photo (in the plates) and a drawing (in the figures).

xxix

KH23LinB

Catalog number with LinB indicates vase (not SJ) inscribed with Linear B.

ENK06m

Catalog number with m indicates vessel of unknown or non-SJ type.

MYC22#

Catalog number with # indicates stopper (cap or plug).

Varia cat. no.

catalog number

exc. no.

excavation number

HOM

House of the Oil Merchant, Mycenae

HWM

House of the Wine Merchant, Mycenae

in prep.

in preparation

inv. no.

inventory number

pers. comm.

personal communication

1

Introduction by Halford W. Haskell, Richard E. Jones, and Peter M. Day

1.1. Transport Stirrup Jars in Late Bronze III The transport stirrup jar (SJ) was a vessel type used extensively in the Late Bronze Age III Aegean world (Ill. 1.1). Found in a variety of contexts from domestic deposits to cargoes in ships’ holds (and only rarely in tombs), the type was used both to transport and to store liquid commodities in bulk. The peak of the production and exchange of this jar type was in Late Bronze (LB) IIIA:2–IIIB, corresponding to a controversial period on Crete and to the time of economic expansion on the Greek mainland. On Crete, SJs can be considered a regular component of the coarse ware repertoire of Late Minoan (LM) IIIA:2–IIIB, appearing at most major centers on the island. However, it is their presence in large numbers in storerooms on Crete and in the mainland that has attracted special attention, as an indicator of commodity movement and perhaps of centralized storage and control of goods. It is impossible to obtain an accurate count of extant transport SJs, since such jars appear in large basement deposits, and time and interest on the part of individual

excavators toward restoring and counting coarse wares yield, at best, only rough estimates. At Knossos, for example, we know that Evans and Mackenzie discarded mounds of coarse ware sherds, many of which no doubt represented SJs (Hallager 1977, 60–71). A very rough count of LB III jars reported from contexts throughout the eastern Mediterranean yields well over 500. A ratio of some 1:3 of inscribed SJs (ISJs) versus uninscribed SJs is far too high, because ISJs are far more likely to be recognized and reported than are uninscribed jars. Perhaps the main reason why these vessels have fired the imagination of archaeologists, however, is their broad distribution at coastal sites in the eastern Mediterranean, as well as their presence in the cargoes of the Uluburun, Gelidonya, and Iria shipwrecks. These transport jars clearly were part of the extensive exchange networks within the Aegean and beyond. While it is true that most SJs moved within the Aegean, significant numbers found in Cypriot and Levantine contexts attest to a type and scale of commercial activity that complements the betterknown exports of Aegean fine wares (see Gale, ed., 1991; Sherratt 1994). Beyond the long-accepted links with the Near East, SJ examples found in South Italy, Sicily, and

2

HALFORD W. HASKELL, RICHARD E. JONES, AND PETER M. DAY

Sardinia shed light on a rapidly emerging sense of Aegean relations with the West (Hallager 1985; Watrous 1992, 182; Vagnetti 1993). Since transport SJs represent a significant element in Aegean exchange, tracing their origins and movement can provide a major source of information regarding possible production centers and trade routes. It is for this reason that our study concentrates on the determination of provenance of these jars and the subsequent tracing of exchange routes, be they through specific, directed shipments or a more informal movement of goods across the landscapes and seascapes of the Mediterranean. Transport SJs are suited to such an analytical provenance project. Indeed, Linear B inscriptions on certain pieces (ISJs) suggest a high level of administrative involvement. At least in some instances of their exchange, SJs reflect large-scale transportation of commodities, taking their place in a very lively movement of goods as part of Late Bronze Age Aegean productive economy and exchange. It is in this context that the identification of production centers and export routes is critical for a full understanding of economic and political conditions in the Late Bronze Age. The study of SJs provides another opportunity— namely, to develop a fully integrated research design that accounts for all major aspects of this rich ceramic resource. As much as the vessels have been an important focus for scholars studying the nature of Late Bronze Age society—touching as they do on issues as disparate as ethnicity and identity in the Mycenaean world, through to the nature of palatial economy at this time— they also represent a cause célèbre for the pottery analyst in the Aegean. The long-standing interest in their origins has led to their analysis and reconsideration at almost every phase of the development of the physicochemical study of archaeological ceramics. They have often been regarded as a test case for the success of such techniques. It is true that the methodological changes

Illustration 1.1. KN32 (left) and TH05 (right).

that have taken place in the present long-term project have often led to an uneven coverage of all aspects of the vessels. However, the combination of so many independent features of SJs (morphology, decoration, chemical composition, petrography, and inscriptions, where they exist) means that this project is unique in its coverage of so many aspects of the ceramic class, which can shed light on their classification and origin. Previous laboratory-based analysis of SJs (beginning with Catling and Millett 1965 and culminating with WCISJ) concentrated on the more than 170 jars of the extant corpus that bear painted Linear B inscriptions. This was in response to scholarly interest in the implications of the inscriptions themselves, particularly as they relate (or do not relate) to the date(s) of the Knossos Linear B tablets (e.g., Palmer 1972; Palaima 1984; Hallager 1987). Nevertheless, most transport SJs were not inscribed, and both inscribed and uninscribed versions must be viewed together as a single class. While transport inscribed SJs have been studied intensively for some 30 years, and useful conclusions have resulted, the current study is the first to embrace the class of SJs as a whole and is certainly the largest study of its kind undertaken in Aegean Bronze Age archaeology. Furthermore, there has been considerable controversy regarding the validity of the techniques employed in previous SJ analyses and of the methodologies employed for the interpretation of the scientific data (e.g., McArthur and McArthur 1974; Wilson 1976). In the current study, previous analytical work is reviewed and critiqued. Weaknesses in earlier work included an insufficient number of samples and an inadequate range of assumptions, both of which compromised the interpretation of analytical data, often based too simplistically on a notion of “objectivity” in scientific analysis. Prime among the assumptions was an expectation of local provenance for most coarse ware ceramics in the Late Bronze Age. We have addressed these fundamental weaknesses, and in light of problems posed by earlier work, we present new methodological approaches to the material. Our effort is an interdisciplinary, collaborative archaeological project, embracing typological, chemical, petrographic, and epigraphic approaches. The results of the chemical and petrographic work, far from being relegated to appendices, constitute primary parts of this study and are integrated fully into the overall project. By establishing the origins and distribution of transport SJs, we ultimately aim to place these vases within their historical context. It is here that the authors’ collective expertise has been brought to bear. Initially, jars were classified into typological, chemical, and petrographic groups independently by Haskell, Jones, and Day, respectively. The three collaborators then compared and critiqued one another’s results, at

INTRODUCTION

3

times supplying confirmatory evidence (both positive and negative). As a last stage, results were synthesized and geographical associations assigned where possible, with final attributions made by all collaborators working closely together. Finally, Killen incorporated these results into his work on the inscriptions.

1.2. Definition The transport SJ corresponds to Furumark’s FS 164 (big domestic). Its height is generally around 0.40 m. It is a closed shape with a narrow “false neck” that is permanently capped by a clay disc; on large SJs this neck is always hollow. The true spout is set into the shoulder. On either side of the false neck is a vertical handle running from disc edge to shoulder; on Late Bronze III SJs these two “stirrup”-shaped handles are nearly always the only handles (on earlier SJs, a third handle, either disc-shoulder or ring on shoulder, is fairly common [Haskell 1985, 223]). Regarding capacity, FS 164 jars at Mycenae were measured directly, yielding an average of 12–14 liters (Ventris and Chadwick 1973, 58–60, 393–394). These figures correspond to estimates by Haskell for complete or restorable jars from other contexts based on treating the vase as two truncated cones (for the method, see Warren 1972, 144–145). Its narrow spout is comparatively easy to close securely (Ill. 1.2), a feature necessary for a vessel sent overland or overseas (Haskell 1981, 230, 232; Watrous 1992, 143–144). Stoppers consist of two parts, a clay plug fitting into the spout and a clay cap pressed over the plug and around the outside of the spout. The clay cap is then impressed with a seal. In order to pour the contents of the jar, this cap had to be broken, and therefore it documents a vessel’s final use, an important consideration for vessels that were no doubt used more than once.

1.3. Construction On some partial or partially restored SJs, evidence of construction is evident on the jars’ interior (Ill. 1.3). The body comprises three sections: the lowest is from the base up one-third to one-half of the way to the maximum diameter; the next is from that point to the shoulder; and the final section is the false neck. Overlapping sections were thinned somewhat and the contact surfaces scored; often the join was incompletely smoothed on the exterior.

Illustration 1.2. Cap (MYC28#) on spout of MYC27 (left). Sealed cap and plug (right), House of the Oil Merchant, Mycenae.

Illustration 1.3. Overlapping sections of base (interior shown at bottom) and body (exterior shown at top) of MYC13.

1.4. Decoration and Firing Apart from the white color on light-on-dark (LoD) jars, the colors on both dark-on-light (DoL) and LoD jars were achieved with the iron reduction technique, using a fine-textured clay enriched with more iron than was in the clay of the body. Red colors would result from an overall oxidizing atmosphere in the kiln, while for black/brown colors a three-phase firing sequence was likely: oxidizing-reducing-oxidizing. During the reducing phase, the decorated area alone would sinter such that its darkened color would be impervious to the final oxidizing phase, while the more porous body would respond to that final phase; the result would be dark decoration on a light-colored body. The LoD jars must have been coated all over with the decorative medium of the fine-textured, ironenriched clay, and again the three-phase firing would have been carried out. Based on analytical work on Early to Late Minoan painted pottery, candidates for the white

4

HALFORD W. HASKELL, RICHARD E. JONES, AND PETER M. DAY

paint on LoD jars could be talc, as found on Kamares Ware (Noll 1982), or kaolin (Stos-Fertner, Hedges, and Evely 1979), or a lime silicate (Noll 1982; Betancourt et al. 1984). While talc would have been available to potters in the Mesara, kaolin or the more available lime silicate would have been favored elsewhere on Crete and beyond.

the potters were exploiting their empirical understanding that medium calcareous clays gave a stable microstructure when fired over this temperature range. By contrast, the low-calcareous fabrics tended to be lower fired, that is, below 800°C, a result that harmonizes with the observation made above about the LoD jars. At any rate, there seem to be no grounds for supposing a functional basis to the two decorative types, DoL and LoD (pace Jones 1984b, 27).

1.5. Hardness Whether by drilling or crushing a fragment at the time of sampling, it was often possible to obtain a qualitative estimate of relative hardness. Among DoL jars, the majority could be subjectively regarded as having a hardness consistent with being well fired, but there were several with a friable fabric, and a smaller number that had been fired very hard. The LoD jars, especially those at Thebes, were notable in being poorly fired, many of them having a red, crumbly fabric. Generally speaking, the cross-sectional view of the SJs appeared to be no different from what is commonly encountered among the larger shapes of the Minoan and Mycenaean repertoire—that is, many examples that had been uniformly fired, but some examples with a gray/dark core indicated incomplete firing. The fabric of transport SJs has often been described as “oatmeal” in reference to the gritty quality of the clay (coined first at Troy: Blegen, Caskey, and Rawson 1953, 305). Day and Kilikoglou (2001) recently made use of microstructure examination (see Maniatis and Tite 1981) to estimate the firing temperature range of plain and decorated LM IA pottery found in and around the kiln at Kommos. Among the medium calcareous pottery (6%–10% CaO), some of which was decorated, a majority was fired in the range 850°–1050°C; it is possible

Illustration 1.4. Black square dots indicate distribution of transport stirrup jars in the central and eastern Mediterranean.

1.6. Distribution Transport SJs are found over a wide geographical area (Ill. 1.4). Most come from Aegean contexts, where the transport SJ is a standard feature of domestic basement deposits. They are found only occasionally in tombs on Crete, and very rarely in tombs in the Greek mainland. SJs were exported east from Aegean production centers to Anatolia, Cyprus, the Near East, and Egypt, where a large proportion is found in tombs. Very few SJs were sent west to South Italy, Sicily, and Sardinia. ISJs are restricted to Aegean findspots (Ill. 1.5). They are found primarily in administrative/palatial contexts where Linear B was in use. That the transport SJ was intended for trade is reinforced by the fact that it was exported immediately after its invention on Crete (see Haskell 1985, 223, 225). Hamilakis (1996, 23) has suggested that the invention of the SJ on Crete coincides with the beginning of systematic oil production.

Illustration 1.5. Distribution of ISJs in the Aegean as indicated by site abbreviations.

INTRODUCTION

1.7. SJ Contents While our analytical study necessarily concentrates on the vessel fabrics, we must remember that the contents, rather than the vases themselves, were of value to those who handled the jars. The jars are utilitarian rather than aesthetic in nature. This is in contrast to fine ware vases, including fine ware stirrup vases, which, regardless of contents, were of considerable intrinsic value themselves. As such, the origins of specific transport SJs reflect agricultural production areas, manufacturing centers, and possibly value-added goods (e.g., perfumed oil) perhaps within a command economy, while their export reveals external political or commercial relations. As will be seen in the results of our analyses, most of these stirrup jars are associated with Crete, a critical point in considering the economic situation of the Argolid and Crete in LB III. This point will be developed further in Chapter 10, concerned with trade. So far, systematic attempts to determine the contents of SJs by organic residue analysis have not been conducted. This type of investigation is currently blossoming in the Aegean, as not only have the techniques of extraction and analysis improved over the last decade, but also, critically, the available databases have grown and the interpretative framework has become more robust (Tzedakis, Martlew, and Jones, eds., 2008). Early work is presented by Jones (1986b, ch. 11), and later work with greater relevance to post–Bronze Age periods appears in Biers and McGovern (1990). For more recent work in the Aegean, the studies of Evershed and coworkers at Bristol, who have highlighted the frequency of occurrence of beeswax in prehistoric ceramics (Evershed et al. 1997), are relevant. But before considering the few relevant results of organic residue analysis, we should examine the epigraphic and archaeological evidence. At Pylos, tablet Fr 1184 records that some 38 stirrup jars (ka-ra-rewe; that this term refers to SJs is shown by a tablet at Knossos [K 778], which shows the SJ ideogram) were to be used for what is quite clearly perfumed oil, because two personal names on the tablet are known elsewhere as perfumed oil makers (Bennett 1958, 40–41; Shelmerdine 1985, 24–25). Contextual evidence for stirrup jar use can be found at the palace. Over 80% of the SJs found at Pylos come from the northeast side of the main building, an area identified by Shelmerdine (1984, 86–95; 1985, 58–62) as a perfumed oil workshop, at least at the time of the final destruction, on the basis of Linear B and archaeological evidence. At Mycenae, the socalled House of the Oil Merchant may provide further evidence. Behind one of the pithoi in Room 1 was a tablet dealing with disbursements of oil (Bennett, ed.,

5

1958, 96–97; Ventris and Chadwick 1973, 218), and in the corridor directly outside of the room were some 30 stoppered SJs. It has been suggested that Room 1 was a perfumed oil workshop, because under one of the pithoi was an arrangement for heating (Marinatos 1958; Palmer 1959, 433 n. 3). The arrangement, however, is sufficient only for warming and not for the boiling required in the manufacture of perfumed oil. Wine is another possible commodity for SJs. At Pylos, two SJs (although not of the transport variety) were found in Room 105, a storeroom associated with wine by sealings found there bearing the ideogram for wine. The Uluburun shipwreck provides further, and probably anomalous, evidence (Bass et al. 1989, 11): transport stirrup jar ULB04 (H. 0.455 m) contained various types of seeds, faience and stone beads, and orpiment, some or all of which may be intrusive; in any case, its contents, whether intrusive or not, certainly suggest that it was not filled for the final time in the Aegean. Indeed, this SJ may be a reused jar returning to the Aegean. The vase provides, therefore, little information regarding Aegean uses of transport stirrup jars. In summary, the archaeological and epigraphic evidence is not decisive about the contents of most individual stirrup jars. It seems to favor oil, perhaps perfumed, as the most common Aegean commodity, but wine obviously cannot be excluded. Although other commodities occasionally may have been transported in SJs, oil and wine appear to be the most common commodities. We can now balance that view against what emerges from the large program of contents analysis, mentioned above, carried out on Minoan and Mycenaean pottery in shapes relating to the preparation, storage, and consumption of solid and liquid commodities (Tzedakis, Martlew, and Jones, eds., 2008). The results point to a wider variety of liquid commodities than had been expected; besides wine, sometimes resinated, and often containing oil to seal the surface, there is evidence of (barley) beer as well as mixed fermented beverages containing together wine, beer, and honey mead. Of two transport SJs from Chania (Kastelli) that were analyzed (Tzedakis and Martlew, eds., 1999, 173, with illustrations)—one of them with octopus decoration—one (no. EUM 262) seems to have contained traces of resinated wine, possibly as part of this mixed fermented beverage. The promising results obtained from this program should encourage a systematic analysis of the contents of SJs from as many contexts as possible within the Aegean and beyond; indeed, the study on Canaanite amphorae will offer a useful point of comparison (Serpico et al. 2003). For the moment, the science-based results serve to amplify and extend the rather limited archaeological/epigraphic evidence presented above (see also M. Serpico, no date).

6

HALFORD W. HASKELL, RICHARD E. JONES, AND PETER M. DAY

1.8. History of Scholarship Early scholarly interest in SJs glossed over questions of provenance and trade, focusing instead on specific questions on their form. Sir Arthur Evans and his contemporaries, for example, were interested in its invention and early evolution (Haskell 1985, 221–222 with references), but little attempt was made to trace its subsequent development or distribution. A. Furumark (1941b) treated in detail only the mainland version, which spans the LB IIIA–IIIC periods. J. Benson (1961–1963) was the first to confront directly the question of where transport SJs were manufactured. He attempted to assign origins for a group of SJs found at Kourion on Cyprus, considering this matter within the larger context of transport SJs found throughout the eastern Mediterranean. He ultimately concluded that the Kourion jars were made in mainland Greece, citing parallels from Thebes, a conclusion brought into question by later work. Nevertheless, his study remains important in that it thoughtfully posed appropriate questions regarding origins. In 1968, J. Raison published a corpus of Linear B inscribed vases (VIP); all but four of the 120 cataloged examples were inscribed SJs (ISJs). Raison also attempted to establish typological groups for the ISJs themselves on archaeological grounds, drawing comparisons with uninscribed SJs. He was thus one of the first to apply both epigraphical and typological criteria to ISJs. Raison ultimately concluded that ISJs were not traded, but rather manufactured and used at or near their findspots. Although we may dispute Raison’s conclusions regarding origins, his typological work remains valid and we have made use of it. Anna Sacconi followed in 1974 with her own corpus of Linear B inscribed pottery (CIV). She augmented Raison’s corpus with pieces discovered since the publication of VIP and with others from Tiryns described in an unpublished manuscript by K. Müller made available to her. Furthermore, she excluded a few pieces included in VIP that she felt bore signs that should be classified as pot marks rather than Linear B signs. Sacconi concentrated on the inscriptions themselves, leaving aside questions of vase typology and origins. ISJs figure prominently in debates regarding the political and economic structure of the Late Bronze III Aegean world. Unfortunately, these coarse ware vases do not lend themselves easily to assignments of origins by traditional archaeological techniques alone. With this in mind, Catling and Millett (1965) undertook a chemical analysis of selected jars from Thebes. Jars were provisionally assigned to various chemical groups; most were considered to be of East Cretan

manufacture, some of local manufacture, and two Peloponnesian. Methodological objections were raised regarding the interpretation of the data; there were philological objections as well, because some jars bear West Cretan place-names. Unknown to Catling and Millett at the time of their publication were results of new excavations at Chania, which were beginning to yield ISJs. Catling and Jones (1977) later published revised conclusions regarding the East Cretan jars, assigning a provenance in West Crete. These analytical studies were undertaken in response to doubts regarding the date of the Knossos Linear B palace, raised initially by L.R. Palmer (1963b, 1965). In questioning Sir Arthur Evans’s date of 1400 B.C.E., Palmer cited the thirteenth-century date of most ISJs. His thesis was that if ISJs were in fact manufactured on Crete, as suggested by West Cretan place-names on some jars, then the Linear B palace at Knossos must have been in existence in the thirteenth century B.C.E. to manage the ISJ business. In 1980, Catling et al. published a comprehensive analysis project covering extant ISJs (WCISJ). As with the earlier studies, mixed results were obtained, but overall nearly 80% of ISJs were shown to be of West Cretan manufacture. Included in this study for comparison were a number of uninscribed transport stirrup jars at Mycenae, from the Houses of the Wine Merchant and the Oil Merchant; the majority of these vases were also ascribed to West Crete. The entire collection of SJs from these houses was studied as well, both in terms of typology and of petrology (Haskell 1981; Riley 1981). The results of the WCISJ study inspired renewed discussions about the political circumstances of Late Minoan III Crete (Palaima 1984; Hallager 1987).

1.9. Current Study Our study fills out the picture presented by earlier, more narrowly focused work, by including a significant number of uninscribed SJs, found at a wide range of central and eastern Mediterranean sites, from Sardinia to Cyprus and the Near East. It has been our aim to analyze a group of transport SJs that is representative of the entire corpus. The uninscribed SJs have been selected to present as broad a range of context and geographic distribution as possible. All available ISJs have now been analyzed, because they are of particular importance in and of themselves. In addition, several SJ stoppers have been included to provide a check for the final use and possible reuse of SJs.

INTRODUCTION

Each analyzed piece has a unique reference comprised of the site (i.e., findspot) code (e.g., MYC = Mycenae) followed by the catalog number. This unique reference is used consistently throughout the text. Table 1 lists the pieces in a summary fashion according to region and findspot. Catalog numbers followed by a superscript LinB indicate other Linear B inscribed vases (not SJs); those with a superscript m refer to miscellaneous pieces. Stoppers are indicated by a superscript pound sign (#). The pieces are listed more fully in Table 27 and in the catalog (Chapter 11). They appear in alphabetical sequence by catalog number. In this manner, the sites covered also can be found in alphabetical order (except for Chania, which is listed under the abbreviation KH; and Karpathos, Pigadia, which is listed under PIG), and furthermore, the reader is immediately alerted to the findspot of each piece. Selected pieces from the catalog are shown as line drawings in Figures 1–18; where these pieces are referenced in the text and tables, the catalog numbers are underlined. Photographs of vessels are presented as Plates 1–30, and their catalog numbers are italicized throughout the text and tables. A catalog number that is both underlined and italicized has been both drawn and photographed. Nearly all of the pieces were analyzed by Jones while at the Fitch Laboratory of the British School at Athens, either by optical emission spectroscopy (OES) or atomic absorption spectrometry (AAS). Petrographic work has been conducted by Day and co-workers at the University of Sheffield. In Chapter 2, the analyzed SJs are classified into typological groups. Chapter 3 focuses on the analysis of SJs, its history, and methodological issues. Chemical and petrographic analyses are presented in Chapters 4

7

and 5, respectively, followed by Chapter 6, in which the scientific results are integrated and discussed. In Chapter 7, we integrate the typological and analytical results and propose geographical associations for the various sampled vessels, drawing upon all of the results collectively—that is, the results of the typological, chemical, petrographic, and epigraphic work. It is hoped that these results will have a wider significance than is addressed by this particular study, since the criteria for the classification should aid excavators and other scholars in assessing the significance of their own SJs. Chapter 8 comprises a discussion by Killen of epigraphical considerations. Chapter 9 is concerned with the chronology of the manufacture and movement of SJs, and with the identification of production centers. Included in this chapter is a discussion of sites where pieces have been found and their dates. The dates of specific pieces are, of course, critically important for assessing the historical implications of production and movement. We need only remember the controversy still surrounding the date of the destruction of the Kadmeion at Thebes to be reminded of the significance of this aspect. Chapter 10 addresses the implication of our results for trade and exchange within the Aegean and beyond. SJs certainly represent manufacturing and export on a fairly large scale. An understanding of the movements of these vases should provide information about political and/or commercial connections between areas, and the locations of significant production and export centers. Chapter 11 presents the catalog of analyzed SJs, stoppers, and a few other vases included for comparison. An appendix listing LM/Late Helladic (LH) III transport class stirrup jars follows after that. The list of references concludes the text.

2

Typology by Halford W. Haskell

2.1. Introduction In this chapter, the SJs analyzed in this study are assigned to typological groups. After an introductory section on the development of SJ design, the bulk of the chapter is devoted to defining typological groups.

2.2. Development of SJ Design As coarse wares, transport SJs do not lend themselves to close dating, and ideally dates should be derived from fine ware context pottery. Certain chronological trends can, however, be established, and we deal briefly with development of the form during the period under consideration here—namely, Late Bronze III (for the chronological development from Middle Minoan (MM) III to LM/LH IIIA:1, see Haskell 1985, 224–225). In LB IIIA early, the transport SJ is comparatively rare outside of Crete and certain islands, and apparently absent in mainland Greece (fine ware versions do occur, however, beginning in LH IIA: Haskell 1985, 228–229,

nos. 88–93); we must rely on specimens on Crete, Karpathos, and Troy. Body shape follows the general principles seen in fine wares of the period. Conical (e.g., PIG03; Sellopoulo: Popham, Catling, and Catling 1974, pl. 34:a), ovoid (MIN02), and piriform (Zafer Papoura: Evans 1905, 464, fig. 83) shapes occur. False-neck assemblies and spouts tend to be relatively small in comparison to later examples (MIN02, PIG03; Sellopoulo: Popham, Catling, and Catling 1974, pl. 34:a). Handles are now almost always two in number, and usually oval in section. Very occasionally a third handle, either in the form of a vertical handle from disc to shoulder (Troy: Blegen, Caskey, and Rawson 1953, fig. 330.18) or a ring handle on the shoulder, is seen (PIG02; Ill. 2.1; Troy: Zafer Papoura: Evans 1905, 464, fig. 83; Blegen, Caskey, and Rawson 1953, fig. 330.14); third handles are characteristic of LM I–II stirrup jars. Horns on the spout and holes through the edge of the disc (PIG01; Ill. 2.2), common earlier, are now almost entirely absent and certainly have lost their original function. Decoration may be fairly elaborate (Sellopoulo: Popham, Catling, and Catling 1974, pl. 34:a; Chania: Theophaneides 1948–1949, 14, fig. 24) to quite simple (MIN02, PIG02). Both Light-onDark (LoD) and Dark-on-Light (DoL) examples occur.

10

HALFORD W. HASKELL

Illustration 2.1. PIG02 with small ring handle on shoulder at left.

Illustration 2.2. PIG01 with horn on spout.

Occasionally the octopus motif is seen, with displayed body (PIG03). Shoulder decoration can be fairly elaborate (e.g., Katsamba: Alexiou 1967, 64, fig. 38, pl. 25:b; Chania: Theophaneides 1948–1949, 14, fig. 24). By Late Bronze IIIA:2–IIIB:1, transport stirrup jars are found in comparatively large numbers, even in mainland Greece. Typological development is best illustrated at Mycenae, where the House of the Wine Merchant and the House of the Oil Merchant provide well-dated deposits of LHIIIA:2 late and IIIB:1, respectively (Haskell 1981, 226, 230). Ovoid (MYC11, MYC49, MYC50) and piriform (MYC08) shapes continue in LB IIIA:2. In LB IIIB, ovoid (MYC15, MYC16, MYC18, MYC26, MYC27, MYC31) and conical-ovoid (PYL04, TH11, TH12, TH13, TH14–TH16) become the norm. Little development of transport SJs occurs after LB IIIB:1. Late Bronze IIIA:2–IIIB false-neck assemblies and spouts tend to be fairly tall. Handles are generally round or rounded oval in section, and sometimes flattened or strap. For the transport version, almost without exception they come in twos. Decoration now is very simple, usually consisting of bands or the octopus motif (with or without displayed body). Both LoD and DoL examples occur. Shoulder decoration tends to be nonexistent or greatly simplified. More complicated patterns are rare (e.g., TH74, with hatched curvilinear pattern work; KN32 and KN33 with antithetic curved patterns; and a Knossos SJ [Popham 1964, pl. 3:b] with flower). Painted Linear B inscriptions occur on transport stirrup jars probably no earlier than LB IIIA:2. For example, several ISJs at Chania come from LM IIIA:2 or LM IIIA:2/IIIB:1 contexts, and while the date of large deposit of ISJs at Thebes is much debated, in any case it is not likely to be earlier than LH IIIA:2 (see Raison 1977). The peak of ISJ manufacture seems to span the LB IIIA:2 and IIIB:1 periods. While a few ISJs do come from contexts as late as LB IIIB:2 or even later (e.g., at Chania and Eleusis), it may be that at least some of them are vases that have been reused for some

time. Inscriptions may appear on the body, shoulder, and exceptionally on the disc. Nearly all inscriptions are painted, with very few incised signs identified as Linear B (KH16, AM01[?]).

2.3. Typological Groups 2.3.1. Typegroups We now turn to a discussion of typegroups. By “typegroup,” we mean a group of SJs that may be associated based on typological criteria: body shape, arrangement of specific ceramic features such as false neck, disc, handles, and spout, decoration, and the content, placement, and arrangement of inscription (if any). Transport SJs, while not subject to as refined a typological analysis as are fine wares, nevertheless fall into a number of distinctive groups. These typological associations have been established independently from, and without reference to, chemical or petrographic groupings. As such, they contribute independent support (positive or negative) for chemical and petrographic groups. For the sake of completeness, included in the discussions here are jars clearly belonging to typegroups but not analyzed chemically and/or petrographically. Some typegroups have been established by the author, others by J. Raison (VIP, 38–39 n. 143–144, 40–41 n. 150, 43 n. 165, 44 n. 168, 61–87, 87–90, 91–93, 93–96, 96–101, 101–107, 110–117, 162–164), who made extensive studies primarily of inscribed jars but also of a fair number of uninscribed examples. Raison established his groups in part based on sign forms, and, as our typegroups are based primarily on body shape and decoration, we include in our groups only those jars sufficiently well preserved to confirm inclusion on these bases.

TYPOLOGY

This section falls into two parts. The first defines and discusses individual typegroups and in some cases subgroups within typegroups. Typegroups are designated by Roman numerals (I, II, etc.), subgroups by lowercase letters (a, b). Fuller discussions of context dates may be found in Chapter 9. The second part defines broad relationships that can be detected among certain typegroups. These groups, listed in Table 2, are designated by letters (Groups A, B, D, and E [Group C no longer exists; it became part of Group B]). Not all jars may be associated with a narrowly defined typegroup, either because they do not seem to have close typological counterparts or because they are too fragmentary for attribution. Nevertheless, many such singletons do display general affinities to certain typegroups within broader groups. An addendum at the end of the chapter contains a discussion of a group established by Raison almost solely on paleographical grounds. These jars are fragmentary and cannot be formed into a typegroup on typological grounds. Table 3 lists all jars by typegroup. Table 4 lists broad Groups A, B, D, and E.

11

and Kommos (KO25) from an LM IIIA context. The vases are ovoid and fairly slender (ratio of D/H avg. 0.69), and have fairly high maximum diameters (ratio of MD/H 0.53). The cream paint is fairly thick and of good quality. Bands adorn the bodies, and on a few, a horizontal wavy line is visible on the shoulder (MYC14, MYC78). A continuous band connects the bases of the handles, false neck, and spout. Typegroup I can be subdivided according to handle and disc decoration: Typegroup Ia (MYC46, MYC77, KN31[?]) has a line running up each handle, continuing as parallel oblique lines on the disc (cf. at Midea: Demakopoulou et al. 1997–1998, 63, fig. 18). Typegroup Ib (MYC14, MYC47, MYC78, MYC79–MYC82, KN30, KO25) has parallel oblique dashes on the handles and a spoked wheel on the disc (cf. from the NW House at Knossos a disc/handle fragment: Popham 1970, pl. 41:c). These two subgroups are linked together on the basis of body shape, quality of fabric, paint, and finish.

2.3.2.2. Typegroup II (Ills. 2.5–2.7; Graph 1b)

A banded LoD typegroup is associated with LB IIIA:2 contexts (for dates, see Ch. 9). This typegroup was first identified at Mycenae (Haskell 1981, 226–228, Groups 1 and 2): 13 vases from the House of the Wine Merchant (LH IIIA:2 late) can be attributed to this typegroup (MYC13, MYC14, MYC46, MYC47, MYC61, MYC76–MYC78, MYC79–MYC82, MYC88). Other examples are at Knossos (KN30, KN31), one of which (KN30) comes from a stratified context (LM IIIA:2),

Another banded LoD typegroup, of LB IIIB:1 date, is represented by jars found at Mycenae (MYC15, MYC16, MYC18, MYC26, MYC27, MYC31, MYC40[?], MYC72, MYC74, MYC89, MYC90, MYC91–MYC99, MYC103), Thebes (TH71, TH72), and Tiryns (TI19), a typegroup first recognized at Mycenae (Haskell 1981, 232, Groups 1–3; a total of 21 jars at Mycenae are of this typegroup). The jars are ovoid, sometimes irregularly so, with bulges marking the points where body sections were joined, are fairly tall (D/H avg. 0.65), and have relatively high maximum diameters (MD/H avg. 0.54). Most spouts bear a painted cross, and a few have a thin arch adorning the shoulder (MYC16, Ill. 2.6). A continuous band connects the false neck, handle, and spout bases. A SJ from Gla (GLA02) may possibly belong to this typegroup (closest to IIa; see below). GLA02 is said by the excavator (Iakovides 1989, 41) to bear part of a Linear

Illustration 2.3. Typegroup Ia (MYC46).

Illustration 2.4. Typegroup Ib (MYC78).

2.3.2. Typegroup Descriptions 2.3.2.1. Typegroup I (Ills. 2.3, 2.4; Graph 1a)

12

HALFORD W. HASKELL

Illustration 2.5. Typegroup IIa MYC89 (left) and MYC15 (right).

Illustration 2.6. Typegroup IIb MYC72 (left) and MYC16 (right).

Illustration 2.7. Typegroup IIc MYC74 (left) and MYC31 (right).

Illustration 2.8. Typegroup III (TH07).

B sign. Iakovides suggests ẉa ̣ Preserved are three vertical lines on the body, approximately 0.10 m long; the upper part of the mark is lost. Iakovides proposes that the possible wa here is an abbreviation for wa-na-ka-tero, as is postulated for the Eleusis ISJ (EL01). The sign on GLA02 is below one handle, an unusual position but paralleled by two DoL ISJs at Thebes (TH37, TH38). It is, of course, impossible to know what the mark looked like in its entirety, or in fact whether this mark stood alone; to the right the vase is preserved, so this may be a final mark, but to the left the body is lost. One may compare another jar, probably belonging with this typegroup, from Schliemann’s excavations of 1876 (Wace et al. 1921–1923, 21, fig. 5), which has an arch bisected by a vertical line, and also a fragmentary piece at Tiryns (VIP, fig. 27). Raison (VIP, 23–24 n. 112) states that TH76 is by the same potter as TH71 and TH72, but the decoration of TH76 is not sufficiently preserved to confirm its inclusion in our typegroup. Typegroup II, like Typegroup I, can be subdivided according to handle and disc decoration, this time into three subgroups. Typegroup IIa (MYC15, MYC18, MYC89) has a vertical line on each handle, terminating in a hook on the disc. Typegroup IIb (MYC16, MYC26, MYC72, MYC90, MYC91–MYC95) has oblique dashes on the handles and a spiral on the disc. Typegroup

IIc (MYC27, MYC31, MYC74, MYC96–MYC99, MYC103, TH71, TH72) bears horizontal dashes on the handles and a thin curl or spiral on the disc. As with Typegroup I, all of the jars within Typegroup II are nearly identical except for the variations in handle and disc decoration. Several jars of this typegroup were stoppered (Haskell 1981, 230, 232). The jars of this typegroup for which caps are well-enough preserved to identify the sealing (MYC16, MYC27, MYC31, MYC74; also Nauplion 5343) all bore caps impressed with the same seal (lionheaded daimon and dogs; Sakellariou 1964, no. 161). Stoppers and caps associated with some of these jars have been analyzed (MYC28#, cap of MYC27; MYC29# and MYC30#, cap and plug, respectively, of MYC31; MYC32#, cap of MYC16; MYC70# and MYC71#, cap and plug, respectively, of MYC15; MYC69#, plug of MYC18).

2.3.2.3. Typegroup III (Ill. 2.8)

Another LoD typegroup is represented at Thebes (TH07, TH09–TH11, TH13). All of these banded jars are inscribed on the shoulder. The inscription is ku-ruzo or ku-ja-ni where fully preserved. There is some

TYPOLOGY

paleographic evidence to help confirm the validity of the typegroup. Raison argued that most of these vases were probably inscribed by the same hand (VIP, 110–117). Killen has noted that the form of the ni on TH10 (and TH44) is unusual in that it lies on its side, an odd position not paralleled elsewhere (p. 97; also WCISJ, 89). The vases are of normal ovoid shape and proportions (D/H avg. 0.68; MD/H 0.55; e.g., TH07). Somewhat unusually, the band group at the top of the body zone surrounds the vase well below the handle bases. A single band connects the false neck, handle, and spout bases, and vertical lines on the side of each handle join at the disc edge. Sign heights average about 0.08 m; on one jar associated by Raison with this group (TH78), the signs are unusually large (0.12 m). This typegroup corresponds to Raison’s Group of Thebes 845 (TH11; VIP, 110–117). Raison included also the fragmentary vessels TH06, TH42, TH44, TH78, plus TH Z 856 and Z 959, based to a large extent on sign forms; these are left out of our typegroup as they are not sufficiently preserved.

2.3.2.4. Typegroup IV (Ills. 2.9, 2.10; Graph 1c)

At Mycenae are four analyzed (MYC11, MYC49, MYC50, MYC83), plus at least five not analyzed, banded jars from a LH IIIA:2 late context, the House of the Wine Merchant (Haskell 1981, 228, Groups 3–4, 227). These slender ovoid vases (D/H avg. 0.66) have maximum diameters above the median (MD/H avg. 0.52). The standard band connects the false neck, handle, and spout bases. This typegroup can be subdivided into two subgroups according to the disc decoration. Typegroup IVa (MYC11, MYC49, MYC83) has the vertical band on each handle terminating in an irregular spiral or circle on the disc, Typegroup IVb (MYC50) in a hook at the disc edge. Apart from these minor differences, overall typological criteria link all of these vases together.

Illustration 2.9. Typegroup IVa (MYC83).

13

2.3.2.5. Typegroup V Two analyzed inscribed jars at Tiryns (TI01, TI10) represent this typegroup. The typegroup was established by Raison (VIP, 162–164, Group of TI Z 1; also included by Raison are TI Z 2, 3, and 5, but these are too fragmentary to allow for detailed typological work). All of the jars of the typegroup bear the same inscription (u-pata-ro) and all inscriptions are on the shoulder. Only two signs are fully preserved; their heights are 0.07 m and 0.094 m. The jars are also similar in fabric and finish.

2.3.2.6. Typegroup VIII (Ill. 2.11; Graph 1d)

These banded vases (TH12[?],TH15, TH16, TH17, TH18, TH19, TH92) are tall, and ovoid in shape (D/H avg. 0.65; MD/H avg. 0.54). This is Raison’s Group of Thebes 853 (TH18; VIP, 61–87). Several vases also included by Raison are extremely fragmentary, and so we do not include them in our typegroup (TH14, TH20, TH40, TH41, TH43, TH45–TH47, TH77, TH80– TH86, TH88–TH90; Z 884, 957, 974). A single band connects the false neck, handle, and spout bases. A wavy band on each handle terminates in a circle at the disc edge. Inscriptions are on the body. A paleographical note suggests Cretan affinities. Palaima (1988, 113) notes that the ne of TH16 has affinities with Knossos tablet material, as well as with Class IV at Pylos (the Pylian class with strong Knossian affinities).

2.3.2.7. Typegroup IX (Ill. 2.12)

These banded, slender ovoid jars (TH08, TH21, TH22, TH27–TH29, TH52, TH60, TH66, TI04[?]; D/H avg. 0.68; MD/H [TH21] 0.55) correspond to Raison’s Group of Thebes 858 (TH22; VIP, 101–107). TH79 and Z 965, fragmentary pieces, are included by

Illustration 2.10. Typegroup IVb (MYC50).

14

HALFORD W. HASKELL

Raison based on sign forms. We add TI04. SJs of this typegroup have a wavy band on each handle joining across the disc. On most jars, a continuous band connects the false neck, handle, and spout bases, while on three jars (TH08, TH27, TH29) one loop connects the false neck and handle bases, and on one (TH22) an additional band connects the false neck and handles. The typegroup was formed by Raison primarily on the basis of sign forms, although the uninscribed vases included reflect typological and fabric links. All inscriptions are on the shoulder. Sign heights range from 0.045 m (TH28) to 0.07 m (TH29), with most at or near 0.06 m. The inscription a-do-we on TH08 links this jar with two at Tiryns (TI04, TI09). Raison (VIP, 177) notes that the signs of TI04 are identical to those of TH08. Despite the inscription, TI09 probably does not belong with this typegroup. Raison’s Group of TI Z 24 (VIP, 176–178) includes TI04 and TI09. Paleographic factors militate against TI09 as a member of this typegroup. The sign height of TI09 is 0.043 m, considerably smaller than signs on Typegroup IX jars. Furthermore, while Bennett (1986, 139) has noticed that there are fairly close similarities between sign forms and inscription placement of TI04, TI09, and TH08, he points out further that the do of TI09,

the vertical stroke of which bows to the right, seems significantly different. TI09 may be an imitation of the other two jars, manufactured with a different clay and inscribed by another hand (see also, Killen this study, p. 99). Another link with Tiryns comes with TH21 and TH22 (di-no-zo, perhaps a misspelling of no-di-zo [Killen 1980; WCISJ, 90; now also this study, p. 99]), to be compared to TI11, TI12, and TI14 in Typegroup VI. It has been argued that the odd loop on the no of these vases sets them apart and reflects affinities with similar renderings of the sign in the Knossos tablets (Bennett 1986, 137; Hallager 1987, 175). Palaima (1988, 113) notes also that the di of Hand 91 at Pylos, Class IV, has affinities with the rendering on these vases and at Knossos. Another inscription provides a link with Tiryns: the inscriptions of TH27–TH29, and TH79 are related by Bennett (1986, 139–140) to TI08.

Illustration 2.11. Typegroup VIII (TH18).

Illustration 2.12. Typegroup IX (TH08).

Illustration 2.13. Typegroup X (MYC52).

Illustration 2.14. Typegroup XIa (MYC24).

2.3.2.8. Typegroup X (Ill. 2.13)

Typegroup X can be dated to the LB IIIA:2 late period (MYC52 plus Nauplion 10974, House of the Wine Merchant). These slender, banded vases (MYC52, D/H

TYPOLOGY

0.71) have fairly high maximum diameters (MYC52, MD/H 0.55). One band connects the bases of the false neck and handles, and another the false neck and spout. The spout is surrounded by an irregular horizontal band. The shoulder and spout decoration is paralleled on several Typegroup XI jars. A broad band on each handle terminates in a circle at the disc edge, from which springs a spiral.

2.3.2.9. Typegroup XI (Ills. 2.14–2.16; Graph 1e)

Typegroup XI has a wide distribution. The best dated are those of the House of the Oil Merchant at Mycenae (Haskell 1981, 232–234, Groups 4 and 5; LH IIIB:1 end). These are very broad ovoid-globular or biconical vases (D/H avg. 0.76), with low maximum diameters (MD/H avg. 0.48). Analyzed examples are found at Episkopi, Bamboula (EP01), Knossos (KN02), and Mycenae (MYC17, MYC19, MYC20, MYC21, MYC23, MYC24, MYC42, and fragmentary MYC66). Although not analyzed, parallels come from Thebes (Archives Room, LH IIIB context: Spyropoulos and Chadwick 1975, 34, no. 5, photo 81), Midea (basement rooms, VIb, VII, destruction level [LH IIIB:2]: Demakopoulou and Divari-Valakou 1994–1995, 326–327, pl. II:A, B; cf. also at Midea a body fragment with similar double deep wavy line decoration: Demakopoulou et al. 1996, 20, fig. 5:a), and Mycenae (House of the Columns, Nauplion 13838: Wace 1949, fig. 110:f). A characteristic feature of these vases is a deep depression in the disc (Ill. 2.16; see here Hallager 1987, 187 n. 169). On some, a single band connects the bases of the false neck, handles, and spout, while on others one band connects the false neck and handle bases and another the false neck and spout. This typegroup can be subdivided into two subgroups: Typegroup XIa (EP01, KN02, MYC17, MYC19, MYC23, MYC24, MYC42, MYC66) bears a double deep wavy line around the body (cf. a smaller double deep wavy line jar in a Rethymnon tomb: Kanta 1980, fig. 86:7; and a body sherd at the Unexplored Mansion: Popham 1984, pl. 107:d), while XIb (MYC20, MYC21) (also Nauplion 13838: Wace 1949, fig. 110:f; VIP, 151) has plain bands on the body. The jar at Knossos (KN02) has the same inscription as its piriform, banded cousin at Rethymnon (AR01) (wi-na-jo) on the shoulder, and as the double deep wavy line jar at Midea (Demakopoulou and Divari-Valakou 1994–1995, 326– 327, pl. II:A, B). MYC20 and the nearly identical but fragmentary jar from the House of the Columns (Wace 1949, fig. 110:f) bear an otherwise unparalleled sign on the shoulder, consisting of two vertical strokes, the one to the left longer, with four or five horizontal strokes between (the closest parallel is sign *56, but on that sign

15

the vertical strokes are of equal length and the horizontal strokes number only three). In typological terms, this typegroup is extremely sound. The unusually broad proportions and low maximum diameter, the distinctive deep depression in the disc, and on most, the characteristic double deep wavy line inspire great confidence in the validity of the typegroup. Two jars found in the House of the Oil Merchant (MYC19, MYC24) still had stoppers in their mouths. On one (plug MYC25#, jar MYC24), the sealings are too worn for identification, but on the other (cap and plug MYC35# and MYC36#, jar MYC19) they represent a bovid motif (Sakellariou 1964, no. 160); this motif also occurs on the cap of the banded, torus-footed jar from this deposit, jar MYC73.

2.3.2.10. Typegroup XII Two relatively small (H. ca. 0.30 m) ISJs from Mycenae constitute the next typegroup (MYC05, MYC06). As observed by Aravantinos (1980), these jars share a similar fabric macroscopically, decorative scheme, and inscription (restored ka-ra-u-ko); Aravantinos notes the significance of the name Glaukos, which has mythological associations with Mycenaean Greece and Crete. A line runs up each handle edge; a circle runs around the disc.

Illustration 2.15. Typegroup XIb (MYC20).

Illustration 2.16. Detail photo of MYC20 showing a deep depression in the disc.

16

HALFORD W. HASKELL

2.3.2.11. Typegroup XIII (Ill. 2.17; Graph 1f)

Three jars of this typegroup bear single marks (TH23[?], TH25, TH26) whereas one does not (TH63?). The typegroup was established by Raison as the Group of Thebes 861 (TH25; VIP, 91–93). Raison also includes TH58, TH59, TH73. These are quite slender conical or ovoid vases (H/D avg. 0.69) with fairly high maximum diameters (MD/H avg. 0.57). Individual bands circle the false neck, handle, and spout bases (except for the single band on TH73), and the disc is adorned with a circle. The “inscribed” vases bear a single sign on the body of a circle bisected by a horizontal line (“ka”) (on TH26 the line is vertical), read by Sacconi (CIV, 226) as pot marks. The spoked wheel or isolated ka occurs at Tiryns (Döhl 1979, 63, 66, no. 107, fig. 10.107; Raison 1963, fig. 2) and Knossos (Raison 1963, fig. 1; VIP, 91, 223–224; Haskell examined the fragments at Knossos in June 1995 and found them to belong to an amphora, not a SJ). One can compare also a domestic SJ fragment from Mycenae (French 1969, pl. 18:b.5); here one cannot tell if this is an isolated motif (wheel), mark, or part of a series of marks/signs.

2.3.2.12. Typegroup XIV

side. On the jar on which it appears as a body inscription, the characters appear between undulations, one undulation compressed in order to leave room for the inscription. A Central Cretan affinity is suggested by the i on TH30: its peculiar embellishment is not typical for Linear B at mainland sites, but it is paralleled in Knossos tablets (Palaima 1988, 113; a parallel also exists with Class IV at Pylos, Hand 19, which itself has affinities with a Knossian tradition).

2.3.2.13. Typegroup XV (Ill. 2.19)

These banded conical-ovoid jars (TH37, TH38; D/H avg. 0.70; MD/H 0.52 [TH37]) have a single band enclosing the false neck, handle, and spout bases, and a wavy line on the handles; solidly painted crescents adorn the disc. This group corresponds to Raison’s Group of Thebes 873 (TH37; VIP, 93–96) and includes also Z 874 (Sacconi [CIV, 232] did not accept these marks as syllabic signs). A single mark appears on the body of each, which on the two restorable jars can be seen to be below the handle, a most unusual position. The signs resemble wa, but upside-down. This mark is not paralleled in Linear B, and its significance is uncertain. It does resemble quite closely a sign occasionally painted in red on fine

(Ill. 2.18; Graph 2a)

Three idiosyncratically inscribed SJs make up this typegroup (TH30, TH31, TH32). This typegroup is Raison’s Group of Thebes 867 (TH31; VIP, 87–90). Raison also includes fragmentary uninscribed jars TH61 and Thebes 904, 909, 929. These are ovoid jars tending toward globular (H/D avg. 0.71; MD/H avg. 0.51), decorated in the main body zone with a deep wavy line. In one case (TH31), the deep wavy line infringes on the shoulder zone. Shoulders are decorated with motifs that are unusually elaborate for transport stirrup jars. On those well-enough preserved, a single(?) band can be seen encompassing the false neck, handle, and spout bases. Vertical bands run up each handle, and the disc is adorned with a circle. These inscribed vases form a good paleographic trio. Sign heights range between 0.062 m (TH31) and 0.073 m (TH32). The inscriptions seem to reveal a lack of familiarity on the painter’s part with Linear B. E. Hallager (1987, 174) remarks on the oddness of the r ̣ụ on TH31 and TH32. Palaima (1988, 113) notes that at Pylos, the embellishment by Hand 19 (Class IV) is paralleled on TH30 and also on Knossos tablets. On two vases, the inscription is ị-r ̣ụ, and on the third r ̣ụ-ị, perhaps a simple misspelling. On the two vases on which the inscription appears on the shoulder (TH30, TH31), the inscription complements a curved line motif to the

Illustration 2.17. Typegroup XIII (TH25).

Illustration 2.18. Typegroup XIV (TH31).

TYPOLOGY

ware Mycenaean pottery found in the east, a correspondence noticed by Raison (VIP, 95; see Stubbings 1951, 46, nos. 7, 8; cf. Schaeffer 1936, 119, fig. 50 [VIII, XIV]; also Masson 1956, 244, fig. 212).

17

displayed octopus on TH55) occupy the upper two-thirds of the vase, including the shoulder. A single band connects the false neck, handle, and spout bases (TH54 has, in addition, individual base rings). Sidelines on each handle run into a circle on the disc.

2.3.2.14. Typegroup XVI (Ill. 2.20)

2.3.2.16. Typegroup XVIII (Ill. 2.22; Graph 2c)

This typegroup was established by Raison on the basis of certain decorative details (TH53, TH56[?], TH57, TH62; VIP, 38–39 nn. 143–144). All are displayed octopus jars, piriform-ovoid to ovoid, of normal proportions (D/H avg. 0.70; MD/H 0.51 [TH53]), with fairly slight features. All have a displayed octopus adorning the upper body and shoulder. False neck, handle, and spout bases are ringed individually. On the disc is a circle. The middle of the spout is adorned with a horizontal band.

Typegroup XVIII is made up of four large banded jars at Thebes (TH01, TH02, TH03, TH04; VIP, 43 n. 165, 44 n. 168). These are very broad jars (D/H avg. 0.80; MD/H avg. 0.56) which have individual rings around shoulder bases. Spirals or concentric rings, or solid paint, adorns the disc. TH02 and TH03 have fairly high maximum diameters and narrow bases. TH03 has a disc foot.

2.3.2.15. Typegroup XVII

2.3.2.17. Typegroup XIX

(Ill. 2.21; Graph 2b)

(Ill. 2.23)

Another typegroup of octopus and deep wavy line jars established by Raison shows somewhat less consistency in decorative details (TH54, TH55; VIP, 40–41 n. 150; also Thebes 892). These are relatively slender ovoid jars (D/H avg. 0.67; MD/H avg. 0.55). The main zones (a carelessly rendered deep wavy line on TH54, a one-eyed

Five jars bearing identical or nearly identical inscriptions and one uninscribed jar make up this typegroup (TH33, TH34, TH35[?], TH36, TH39[?], TH50[?]; Raison’s Group of Thebes 876 [TH39]: VIP, 96–101). All of the inscriptions, ta-de-so or ta-*22-de-so, appear on the shoulder. These are banded ovoid jars (D/H

Illustration 2.19. Typegroup XV (TH37).

Illustration 2.20. Typegroup XVI (TH53).

Illustration 2.21. Typegroup XVII (TH55).

Illustration 2.22. Typegroup XVIII (TH03).

18

HALFORD W. HASKELL

avg. 0.68; MD/H 0.48 [TH33]), with a torus or piriform foot, and with bands running along the handle edges and joining a circle at the disc edge, from which springs a spiral. E. Hallager (1987, 176; but see Killen 1980, 89–90) believes that these inscriptions reflect a common scribal tradition along with KH Z 5 (KH05) at Chania.

2.3.2.18. Typegroup XX (Ill. 2.24)

Typegroup XX comprises one nearly whole jar (KN05: D/H 0.76; MD/H 0.53) and four handle/false neck fragments (KN19–KN22). The whole jar is irregularly ovoid and decorated with plain bands. On the shoulder, the false neck, handle, and spout bases may be ringed independently (KN05, KN19[?]) or connected by one continuous band (KN20); the remaining fragments are too worn to determine shoulder decoration. Bands along each handle edge run on to the disc, between which are short horizontal dashes, creating a sort of laddered effect. On the disc surface of each is a painted spiral and an incised cross (an incised + appears on a typologically earlier, three-handled stirrup jar from the Unexplored Mansion: Popham 1984, pl. 119:c). The fabric of KN05 is, in macroscopic terms, quite coarse, more so than that of KN19–KN22. However, the painted spiral and incised cross serve to bind this typegroup together.

2.3.2.19. Typegroup XXI (Ill. 2.25; Graph 2d)

Two nearly identical jars at Knossos make up this typegroup (KN32, KN33); to this may be added the identical jar from Malia (Deshaynes and Dessenes 1959, pl. XLVII:6, 8). These tall ovoid jars (D/H avg. 0.74) have their maximum diameter just above the midpoint of the vase (MD/H avg. 0.52), and have in their main decorated zone a deep wavy line. In the broad shoulder zone are antithetic loops, a motif derived from the upper tentacles but now separated from the main zone by bands. On KN33, a continuous band links the false neck, handle, and spout bases. A wavy band runs up each handle to the disc, on which there is a circle. An unusual feature is the relatively low level of the two bands marking the top of the main body zone, placed well below the handles.

2.3.2.20. Typegroup XXII (Ills. 2.26, 2.27; Graph 2e)

Several jars from a single LH IIIB:1 late deposit at Zygouries, the Potters’ Shop, constitute this typegroup (ZYG01, ZYG03, ZYG04, ZYG05–ZYG08, ZYG09, ZYG10, ZYG11). These are broad (D/H avg. 0.86) jars, with fairly high maximum diameters (MD/H avg. 0.53). Most have one or more holes running through each

Illustration 2.23. Typegroup XIX (TH33).

Illustration 2.24. Typegroup XX (KN05).

Illustration 2.25. Typegroup XXI (KN32).

Illustration 2.26. Typegroup XXIIa (ZYG04).

TYPOLOGY

19

handle, to aid in firing. They are decorated with plain bands and have a sort of tassel design on the shoulder. The typegroup can be subdivided into two, based solely on size. One (ZYG01, ZYG03, ZYG04, ZYG05–ZYG08) is about 0.37 m high (estimated capacities, by truncated cones, about 12.5 liters, well within the normal range for transport stirrup jars), and the other (ZYG09, ZYG10, ZYG11) is about 0.57 m (capacity of nearly 50 liters).

2.3.3. Broader Groups While each typegroup represents a discrete group of nearly identical jars, certain typegroups may be associated with others in more general ways. Here we are able to include also single jars that do not belong to individual typegroups. Table 3 lists the jars within each broad group. Characteristics of the broad typological Groups A, B, D, and E are summarized in Table 2. Detailed discussions of each group follow here.

Illustration 2.27. Typegroup XXIIb (ZYG10, left) and (ZYG09, right).

Mycenaean influence (Kanta 1980, 289). Discs may be decorated with rings, spirals, or spoked wheels. Most Linear B ISJs belong to this group. Shoulder inscriptions on these jars outnumber body inscriptions by about 3:2. Characters are fairly large and sometimes carelessly rendered (shoulder inscriptions cluster around 0.06–0.08 m, while body inscriptions cluster around 0.08–0.13 m).

2.3.3.1. Group A (Ill. 2.28; Graph 3a)

Most Group A jars are ovoid or ovoid-piriform; often a prominent bulge can be seen in the lower body at the place of attachment. Group A jars are relatively slender (avg. D/H 0.68), with the maximum diameter generally above the midpoint (avg. MD/H 0.54). The false neck is fairly tall, with a cylindrical or concave profile. The disc is generally flat, sometimes with a slight bulge or depression (the central depression in MYC08 is unusual). Handles are round or rounded oval in section. The spout is fairly tall, either cylindrical or slightly concave in profile. Both DoL and LoD vases are common (Tzedakis and Kanta 1978, 26; Kanta 1980, 257; also Mavriyannaki 1967–1968, 169, fig. 4), in contrast to the rare LoD vases of all the other groups. Decoration is very simple, especially in LB IIIB, consisting of bands and a few more elaborate motifs. MYC08 is exceptional with its foliate bands on the shoulder, perhaps indicative of its relatively early date (LB IIIA:2); elaborate motifs are more often seen on Group B pieces (TH31, TH61, TH74). The octopus motif, while seen on fine ware stirrup jars of the Chaniote workshop (Tzedakis 1969, 399, 400, figs. 8, 10–14), is very rare on Group A transport stirrup jars. Most SJs of this group have a single band on the shoulder enclosing the handles, false neck, and spout (noted also by Kanta [1980, 294]), while a very few have these features ringed individually, the latter suggestive of

2.3.3.2. Group B (Ill. 2.28, 2.29; Graph 3b)

This group is best represented by SJs found in Central Crete, the Argolid, Thebes, and Cyprus. Relatively few ISJs belong to this group. Body shape and proportion (D/H avg. 0.73; MD/H avg. 0.52) vary far more than in Group A, where there is considerable consistency. Shape ranges from ovoid to piriform to biconical to globular; on the broader specimens, the maximum diameter may be at or below the midpoint of the vessel and the diameter-to-height quotient sometimes a relatively broad 0.80–0.85. The vases may have a significant thickening at the base. Occasionally a disc foot (TH91) (cf. the Linear B ideogram for stirrup jar at Knossos [K 778]: Bennett 1958, 40–41) or splaying base is seen. Two ISJs of this group are quite small for transport stirrup jars (Typegroup XII); they may be the smallest extant Linear B ISJs anywhere. The false neck is fairly tall, but sometimes somewhat shorter in scale than their Group A counterparts. Often a distinctive deep depression can be seen in the disc (Ill. 2.16), deep enough so that it may be considered a deliberate feature, and seen most often on specimens in both the Argolid and Central Crete. Handles are round to flattish oval in section, and spouts are either concave or cylindrical in profile. At times a neck ring may be rendered plastically at the base of the false neck (Ill. 2.29), a feature seen also in some Group D jars.

20

HALFORD W. HASKELL

Group A

MYC16

Group B

MYC83

TH07

TH18

MYC24

KN05

KN32

MYC20

TH25

TH31

TH68

TH74

Group D

MYC73

TH03

ZYG04

Group E

IAL02

IAL07

Illustration 2.28. Examples from the broader Groups A (top row), B (second and third row), D (fourth row), and E (fifth row).

TYPOLOGY

All Group B jars are DoL. Body decoration ranges from simple bands to the octopus motif, with or without displayed body. The octopus motif is quite common on SJs found in Central Crete and at Thebes. The octopus motif may appear as a displayed octopus, as a deep wavy line in one or more registers, or as a double deep wavy line in one register (Typegroup XI). Displayed octopus motifs seem less common at Mycenae, at least in LH IIIB. This may be due only to chance of preservation at Mycenae, since displayed octopus jars are known earlier (e.g., MYC55, LH IIIA:2 late); a medium-sized jar was also found in a chamber tomb (Athens 2772: Xenaki-Sakellariou 1985, 136, pl. 62 bottom: IIIA-type octopus, a jar with a pinched lug on the shoulder and perhaps a little earlier than the octopus jar from Karpathos [PIG03], as its octopus is more naturalistic, although dated by E. French [1965, 161] to LH IIIA:2 late). For single deep wavy lines on Argolid-found SJs, see specimens at Mycenae (French 1969, 79), Tiryns, (TI Z 9: VIP, pl. 168; CIV, pl. XI; Bau 3: Grossmann and Schäfer 1975, 71, no. 64, pl. 49:64), and Prosymna (Blegen 1937, fig. 120:227). The main zone may be terminated anywhere from the middle of the lower body to the maximum diameter, and often includes the shoulder zone. On jars with separate shoulder zones, the motifs can be fairly elaborate (e.g., TH23, TH31, TH61). Sometimes a continuous line connects the handle, false neck, and spout bases; at other times, one line connects the false neck and handle bases and another the false neck and spout bases; in still other cases, bases are individually ringed. Spouts are often adorned with a U-shaped band, irregular horizontal band, or horizontal band. Discs may be decorated with circles, spirals, crosses, or x’s. A very few Group B jars bear painted inscriptions. Two on Crete bear identical Linear B inscriptions, wina-jo (KN02, AR01), placed on the shoulder, as does a pair at Mycenae inscribed ka-ra-u-ko (MYC05, MYC06). A single sign not paralleled elsewhere appears on the shoulder of two Typegroup XI jars at Mycenae (MYC20; Nauplion Museum 13838: Wace 1949, fig. 110:f; the latter is now lost). Some of these painted signs are quite carefully rendered, while others are carelessly rendered and seem to be as much decorative as informational. Signs tend to be comparatively small in scale, in contrast to their Group A counterparts, and appear on the shoulder either opposite the spout or between the spout and a handle. Pot marks resembling, or in some cases identical with, Linear B ka occur in Group B (see Section 2.3.2.11, Typegroup XIII and notes). Variations of the normal ka are sometimes seen: some lack both strokes within the circle, others bear a vertical stroke, others a horizontal. It could be suggested that the ka is meant to represent the

21

first syllable of the Mycenaean word for stirrup jar, although why one would write “stirrup jar” on a stirrup jar is a little obscure. A prominent bulge often occurs in the lower body at the point of joining (e.g., KN05, PYL05). On some flat discs at Knossos, a cross or x is incised into the surface (e.g., KN05 [Ill. 2.29], KN19–KN21). It may be noted here that the jar at Amnisos (AM01) bears two incised “signs” on one handle which resemble the Linear B characters e-to.

2.3.3.3. Group D (Ill. 2.28, 2.30; Graph 3c)

These jars are fairly distinctive in terms of body shape and proportions and in certain typological details. They may be conical, ovoid, or globular, and often are quite broad (see Sections 2.3.2.16, 2.3.2.20, on Typegroups XVIII, XXII; avg. D/H 0.85; MD/H 0.54). The handles are sometimes strap handles, a feature seldom seen in the other groups. At times, articulated neck rings are seen at the base of the false neck. Characteristically, handles tend to be pierced, presumably to aid in firing (Ill. 2.30; see Sections 2.3.2.16, 2.3.2.20, Typegroups XVIII, XXII, ATH01). The fabric may be relatively fine. In fact, most Group D jars seem to have been manufactured according to the principles of potting fine ware rather than domestic

Illustration 2.29. Detail of x incised into surface of disc of KN05.

Illustration 2.30. Detail of the pierced handle of TH04.

22

HALFORD W. HASKELL

wares. A notable exception is KN36, with its relatively coarse fabric and finish; its morphology suggested to Popham (1964, 16) that it was an imported piece. All are DoL. Decoration usually consists of simple bands. On the shoulder can be a simplified floral or tassel motif (e.g., ZYG04, Ill. 2.28). This motif is similar to FM 72.6–7, thought by Furumark to be a late motif belonging to LH IIIB and LH IIIC:1. It appears on stratified SJs at Mycenae, however, in LH IIIB:1 (Mountjoy 1976, 84). False neck, handle, and spout bases are generally ringed individually. Interestingly, no jar of this typological broad group is inscribed. All ISJs belong to Groups A and B.

2.3.3.4. Group E (Ill. 2.28; Graph 3d)

Eight (IAL02, IAL04, IAL05, IAL06[?], IAL07, IAL11–IAL13) of 13 analyzed jars found at Ialysos belong to Group E. Typologically, jars of this group are quite distinctive. The jars are very broad, tending toward conical-globular (IAL11) and ovoid-globular (IAL12 and IAL13; D/H avg. 0.82). The maximum diameters are near the midpoint of the vase (MD/H avg. 0.53). The bases are extremely narrow, and the false-neck assembly fairly small relative to the size of the body. They may be decorated with band groups (IAL13) alone, or with a series of horizontal wavy lines (IAL11). The shoulder may be decorated with a rough hatched pattern (IAL12).

2.3.4 Addendum: Raison Group of TI Z 11 A fairly large collection of fragmentary inscribed jars at Tiryns (TI02, TI03[?], TI11, TI12, TI14,

TI17[?]) defines another Raison Group (Group of TI Z 11 [TI02]: VIP, 169–175). Also assigned to this typegroup by Raison are TI Z 12 (probably), 13, 14, 16 (probably), 20(?), and 22(?). As is the case with most ISJs at Tiryns, these jars are too fragmentary for close typological analysis. It is difficult to confirm from the photographs published by Raison whether all vases cited belong, but we trust Raison’s eye, as he has made extremely accurate observations elsewhere (for example, one might be tempted to remove TI17 from this typegroup, but it has been left here on the authority of Raison). These banded jars bear the inscription nodi-zo (where fully preserved) hanging just below the two bands separating the shoulder from the body. Heights of fully preserved signs average 0.08 m. On the basis of the inscription, Killen has connected this group with two Typegroup IX ISJs at Thebes, TH21 (Z 857) and TH22 (Z 858), noting that the di-no-zo of the Theban pieces probably represents a misspelling of no-di-zo (WCISJ, 90). Another paleographic point provides further support for connections. The no of TI11, TI12, TI14 (also TI Z 11, 13, 14, and Typegroup IX jars TH21 [Z 857], TH22 [Z 858]) has a horizontal loop across the vertical strokes in the upper part of the sign. There are no parallels for this version in tablets from the mainland, but no as inscribed by Hand 103 at Knossos has horizontal strokes (not a loop) across the vertical strokes (Bennett 1986, 137; Hallager 1987, 175). Further subtle differences in letter forms within this typegroup have been pointed out by Bennett (1986, 137), which may show that not all were manufactured simultaneously; nevertheless the jars represent a distinctive paleographic tradition.

3

Scientific Backgound and Aims of the Analyses by Richard E. Jones and Peter M. Day

3.1. History of Stirrup Jar Analysis and Methodology 3.1.1. Introduction Laboratory-based study of coarse ware stirrup jars (SJs) has a long history, spanning over forty years from the first application of chemical analysis to prehistoric Aegean pottery by Catling, Richards, and Blin-Stoyle (1963). During this time, all aspects of the process of provenance determination—from sampling strategies to techniques of analysis, from data treatment to the framework of interpretation—have altered greatly. This is not just a matter of record in the history of the discipline, but of great relevance, as the SJs have become the cause célèbre of physico-chemical analysis of Aegean ceramics, relying heavily on one or more of these aspects in different phases of the work. In the past, the analytical study of SJs, especially those with Linear B inscriptions, had come to mirror the working methods and assumptions, implicit or explicit, that have formed the basis of pottery composition and provenance work over the last four decades in the

Aegean. For this reason, this chapter addresses some of the main issues of methodology while at the same time not attempting to review this large subject as an end in itself. While we acknowledge that if this provenance project were to be initiated today, its research design would be different, we aim to contextualize the development of ceramic analysis and provide the background to our own work. We will consider fundamentally different ways of looking at SJ provenance. Within this effort, since the main part of the laboratory-based work in this investigation began more than a decade ago, we critically review compositional and other data relevant to the SJs that have been obtained in the intervening period up to the present.

3.1.2. History of Stirrup Jar Analysis The first phase of laboratory-based study of SJs, conceived by Hector Catling and carried out by Mrs. Anne Millett at the Research Laboratory for Archaeology and the History of Art (RLAHA) at Oxford, was by any standard ambitious for its time, being one of the first major

24

RICHARD E. JONES AND PETER M. DAY

cases to which the RLAHA’s data bank of compositions characteristic of production centers in the Mycenaean and Minoan worlds could be applied. It consisted of the chemical analysis of a set of 25 jars from the Kadmeion at Thebes, most of them inscribed in Linear B. The conclusion reached in this inquiry, that the majority of the jars were probably products of East Crete, immediately proved controversial from several points of view—the lack of extant ISJs from Minoan sites in the east of Crete being perhaps the main one. The publication of Catling and Millett’s work in 1965, followed soon after in 1968 by that of Raison’s corpus of Linear B inscribed pottery (VIP), had the effect of raising the jars’ archaeological profile to a level that has scarcely diminished over the years. As a result, the laboratory component has assumed a comparable importance, becoming a frequently quoted example of the early use of chemical analysis in provenance studies of pottery in the Mediterranean world (Cherry and Knapp 1991, 94; Knapp and Cherry 1994, 34). With the excavations of the Minoan levels at Chania revealing the importance of that center in LM IIIB, it became necessary to reinvestigate the origins of the Theban jars, and to this end Catling and Jones’s (1977) study was able to show that the west of Crete was compatible as the region of origin for those jars originally analyzed by Catling and Millett. Much has since been written about the analytical procedure and methodological content of both these studies, the substance of which is summarized below and which, in any case, has been reviewed fully in WCISJ and by Jones (1986b, 477–493). The study in WCISJ formed the foundation and inspiration of the present investigation. In essence, that study applied the same general approach to the problem of the origins of the ISJs and the same analytical technique, optical emission spectroscopy (OES), as the previous studies, while also addressing many of its deficiencies— for instance, by increasing the number of test samples (77 ISJs from seven sites on the mainland and Crete, constituting 54% of the extant corpus, were analyzed, together with 19 uninscribed jars), by ensuring that reference data were available for most of the sites, and by adding the univariate treatment of the data to the main techniques of multivariate statistical treatment. The study in WCISJ comprised much more than an interpretation of the chemical data; it exploited the opportunity to correlate three independent dimensions of information on the jars: the archaeological context, the epigraphic content, and the chemical composition. In a research program that had both analytical and archaeological coherence, that study identified West Crete as the prime production zone of the ISJs, and pointed to other potential sources, thus paving the way for future work in terms of both methodology and archaeological questions. Its major conclusion, that

ISJ production was apparently largely a West Cretan phenomenon datable to LM/LH IIIB—found general, if mainly tacit, support among prehistorians. Furthermore, at a general level, the classification of the chemical data seemed to harmonize with the toponym groupings discernible in the reading of the Linear B inscriptions. The WCISJ study was in effect a vote of confidence in the chemical approach. There was discernible structure and consistency within the data; crucially, the composition characteristic of West Crete was apparently distinctive with respect to most, if not all, of the other possible source regions. The previous compositional overlap between the east and west extremities of the island was solved by the inclusion of better comparative compositional reference groups. Despite the shortcomings of the analytical technique—giving rise, for instance, to serious overlap in the composition ranges of some of the reference groups—the assignment of origin to the majority of the test samples was made with some confidence. However, problems remained, especially in the case of the main compositional overlap, that which made the discrimination of Central Crete, the Argolid, and Boeotia difficult in some circumstances. The interpretation in WCISJ of those samples belonging to this compositional type tended to rely on archaeological criteria; one of these criteria was a principle assigning a local origin if the sample composition was compatible with such an origin. This is perhaps one principle and practice in ceramics analysis that has been consistently challenged, and overturned since that study, as the extent and intensity of ceramic exchange has become clear (Wilson and Day 1994; Day and Wilson 1998). Such a reassessment of some analyses presented in WCISJ would be important, as a source area in the Argolid and Central Crete would have major implications for the interpretation of production, exchange, power, and even ethnicity.

3.1.3. The Present Study The present study forms a natural continuation of the earlier work and indeed builds explicitly upon it. Implicit in its original strategy was a strongly provenancedirected inquiry which could only benefit from the deployment of a technique of chemical analysis, in this case atomic absorption spectrometry (AAS), which would in turn make good (we hoped) the shortcomings of the OES-based studies. It was also based on the conviction that the ISJs were “the tip of an iceberg” of a larger phenomenon of transport jar exchange; if progress were to be made, the new study would have to redress the imbalance of the earlier work in its concentration on SJs that were inscribed. Thus the broader picture of

SCIENTIFIC BACKGROUND AND AIMS OF THE ANALYSES

the origins and distribution of SJs as a class, whether inscribed or not, was set as the goal, requiring the sampling of jars on a numerically large scale from as wide a range of geographical contexts as possible. The chemical component of this project was carried out intermittently over a long period, from 1986 to 1991, a time that saw profound changes in the approach to analytical ceramic studies in the Aegean and elsewhere, elements of which are reviewed in more detail in the next section. Prime among the factors in these changes was the recognition of the applicability of petrographic analysis to Bronze Age pottery. This recognition began during the 1970s (Matson 1972; Betancourt, ed., 1979, esp. 3–11) and has seen considerable development since then (Vaughan, Myer, and Betancourt 1995; Whitbread and Jones 2008). There was as well an increasing confidence in the study and interpretation of macroscopic fabric characterization and the application of other chemical and statistical techniques.

3.1.4. The Role of Petrographic Analysis In terms of SJ studies, the first wave of change was the active introduction of petrographic analysis to analytical programs. This was initiated by J.A. Riley’s research at Southampton University in the early 1980s, as part of a large-scale application of ceramic petrology to Aegean pottery. The contention was that petrography had as much to contribute to characterization and provenance work in the East Mediterranean as the chemical approach, which itself had long dominated the picture. The SJs were the perfect case in point, with the coarse texture of their fabric suggesting that they would repay examination on the polarizing microscope. In a series of articles, Riley (1982, 1983) extended the conviction that petrography had the potential to establish pottery provenance on a micro-scale, in some cases using characteristic inclusions to demonstrate a nonlocal provenance or to specify a place of origin (e.g., serpentine in North Central Crete). In 1981, he applied this approach to 62 SJs and 10 stoppers found at Mycenae, characterizing in detail the geological provenance of a variety of groups found in the material. In research aimed to build on Riley’s work, Day attempted to demonstrate the potential of petrographic approaches to reveal a range of information on provenance, exchange, and technology of prehistoric pottery in the Aegean. As part of this, such analyses were increasingly incorporated into the analysis of SJs, eventually adding substantially to interpretation of the chemical data. Therefore, the SJ analytical program, conceived initially as chemically based, increasingly

25

incorporated a petrographic component. As a first step, Day took responsibility for a reassessment of the petrographic analyses of material from Mycenae; then he analyzed a selection of the material already analyzed chemically, and finally, during the 1990s, he extended the geographical coverage to include jars found on two wrecks, Uluburun and Iria, and further analyses of material from Crete. Betancourt and co-workers also published the analysis of a SJ bearing an incised CyproMinoan sign, suggesting a provenance in Central Crete (Palaima, Betancourt, and Myer 1984).

3.1.5. Methodological Issues The deficiencies of the early chemical work on the SJs can be summarized as insufficient comparative data, inadequate statistical techniques, and insufficient explanation of the deductive stages. Wilson (1976, 56) demonstrated that on strict statistical criteria, “the jars cannot all be reasonably assigned to the groups suggested by Catling and Millett,” and he went on to write, “of the provenance of the jars among the sites so far examined, the most likely origin would be east Crete. And that is a very different conclusion from that presented by Catling and Millett.” McArthur and McArthur (1974) criticized the inherent statistical bias of the RLAHA’s program of composition and provenance work in the Aegean, resulting from the nonrandom selection of samples and the inflexible use of an insufficiently wide range of elements for measurement (nine elements determined by OES). A critique of poor reproducibility of data was the focus of J.T. McArthur’s (1978) assessment of Catling and Millett’s and Catling and Jones’s studies (for estimates of accuracy and precision, see Jones 1986b, table 2.2), Furthermore, McArthur questioned the validity of the composition differences between East and West Cretan pottery as established by Catling and Jones and, on a more general level, recommended that as a laboratory’s chemical database expanded, the identification of composition groups considered characteristic of a particular region should be regularly monitored. While the investigation in WCISJ made good some of the criticisms of the early work, problems remained, notably in the overlap of some of the crucial reference groups. This was viewed primarily as a product of an inadequate technique of chemical analysis. Another issue that has come under scrutiny is the nature of the reference material employed in the early SJ studies, including that in WCISJ (French 1991). At the mainland sites, this material was biased heavily toward fine wares (although at some of the Cretan sites, such as Palaikastro, Knossos, and Chania, a better balance of fabrics was achieved). At the time, the

26

RICHARD E. JONES AND PETER M. DAY

justification for relying uncritically on fine ware reference groups was that at many sites the coarser fabrics— apart from, say, cooking wares—could not normally be differentiated from the fine wares with respect to the nine elements measured. The late 1980s brought more systematic critiques of the chemical approach to provenance determination in the Aegean. This coincided with the generation of a substantial amount of new chemical data for prehistoric pottery by neutron activation analysis (NAA) (by groups based in Manchester, Bonn, and at the National Centre for Scientific Research [NCSR] “Demokritos” in Athens; see Ch. 4, Sec. 4.4.2). Outside the realm of chemical analysis, and perhaps more instrumental in medium-term change in research design, was the development of the role of petrographic analysis in Aegean prehistoric pottery studies. The success of early work in providing detailed accounts of production, distribution, and even technology, within particular geographical regions (e.g., Crete: Day 1991, 1995a, 1997; the Cyclades: Vaughan 1990; Vaughan and Wilson 1993; the Peloponnese: Whitbread and Jones 2008), raised expectations of ceramic analysis, while field archaeologists started to have renewed faith in the power of macroscopic assessment of fabric (Moody 1985; Haggis and Mook 1993). In addition, a number of large chemical and isotopic data sets for other artifacts were produced, which raised similar questions to those of chemical data for pottery. Notable among these were the elemental and isotopic data for bronzes. Knapp and Cherry’s thorough critique of provenance studies (1994, 17–40), which arose out of the British Academy–funded Research Project “Science and Archaeology: Bronze Age Trade in the Aegean and Adjacent Areas” (see Gale, ed., 1991), was of great importance in encouraging a critical approach to such data and demonstrating the value of an informed, archaeological contextualization of these analyses. There were a number of difficulties encountered by chemical studies of prehistoric Aegean materials that can be summarized here. Provenance studies using chemical analysis are controlled analytical and statistical exercises. Within that framework, the tighter the control on the many variables, natural and human (or behavioral), and the larger the corpus of reference data, the more secure in, principle, will be the statements about origin. In practice, however—particularly with prehistoric pottery studies—the demands from this formula may become counterproductive because the realities of pottery making are increasingly left unaccounted for. The result from classifying the chemical data—the group or the cluster of samples—has definable statistical meaning but may be abstract in terms of precisely what it refers to in the pottery production sequence or hierarchy. Its archaeological

significance may therefore be lost, since so much effort has gone into creating objective conditions with which to identify that grouping or cluster (Jones 1994b). In this way, chemical analysis runs the danger of becoming a means to a narrowing end, losing contact with the very questions that it is trying to resolve. Furthermore (and again, particularly for the prehistoric period), the criteria for selecting reference material for chemical analysis sometimes carried assumptions that were archaeologically dubious or biased. And finally, the archaeological strategy behind chemical analysis could at times be too dominant, paying scant attention to what the results of analysis were saying: that important as it was to have a wide and comprehensive coverage of production centers for characterization purposes, the relative lack of geological variation in some areas would mean that the chances of chemically differentiating individual centers would be slim; at best there would be interregional discrimination. In brief, as the study discussed below by Day et al. (1999) put it, compositional space was too simplistically interpreted as geographical space. For the SJs themselves, this point takes on added significance, since there is an additional and potentially more serious difficulty, one that goes to the heart of the provenance principle (for a recent statement on this principle, see Wilson and Pollard 2001). One of the basic tenets behind this principle— namely, that the composition groups identified in the course of examining the chemical data are a reflection of spatial variability—does not or need not apply consistently to the SJs. During the 1990s, therefore, there was a distinct trend away from the provenance-directed analysis concerned simply with a single ceramic class and treating it in isolation. Instead, researchers moved toward analyzing the broader picture of production at a given site or producing region (Day 1988a). It was a matter of reordering priorities. Only in this way could a reasonable framework of the dynamics of production be assembled, or indeed a clear picture drawn of which ceramic fabrics were or were not local to any given archaeological site—something of prime importance for creating valid reference groups for comparison with samples of unknown origin. This new approach has inevitably been driven by the need to understand the ceramic materials, their nature, location, and treatment by the potter within a well-defined area or region. Such needs have been met by combining in differing ways petrographic analysis, clay prospection, experimentation with clay materials, and, where appropriate, the input of ethnographic evidence. Nowhere in the Aegean (or indeed farther afield) has this new direction been explored and developed more extensively than in Minoan Crete, through the work of one of us, Day, and co-workers. The primacy of the

SCIENTIFIC BACKGROUND AND AIMS OF THE ANALYSES

petrographic approach can be seen in the studies of the production and distribution of Neopalatial pitharia in the Knossos region (Day 1988a, 1988b) and in the Mesara, and of Neopalatial coarse and fine wares in East Crete (e.g., Day 1995a). But more relevant to the present study is the interplay between petrographic and chemical analysis of two Early Minoan (EM) pottery classes (Dark Gray Burnished and Dark-on-Light wares) from a number of areas of Crete (Day et al. 1999). Having securely defined these wares stylistically and petrographically, it was possible to assess whether chemical similarity (or dissimilarity) was a function of origin, technology (paste recipe), or diagenesis, on their own or in combination. This step forward signals the convergence of the standpoints whose polarization, according to Knapp and Cherry (1994, 16), has dogged the progress of provenance work in the East Mediterranean Bronze Age: on the one hand, those who expect an objective answer in terms of provenance, and on the other, those who emphasize the domination of human aspects of raw material choice and manipulation. It is through an understanding of the latter that one can move forward realistically to assess the provenance of any given ceramic group. Regarding the techniques of chemical analysis, NAA has been predominant in the last decade or more, although the last few years have seen the emergence of inductively coupled plasma-emission spectrometry (ICP-ES) as an additional technique generating data for Late Bronze Age pottery, with thirteen elements currently determined at the Fitch Laboratory. In this connection, recent multi-technique (NAA-ICP-XRF [X-ray fluorescence analysis]) and inter-laboratory comparative and standardization exercises, as part of the GEOPRO consortium project, are relevant: see Hein et al. (2002) and Tsolakidou and Kilikoglou (2002). Of these, the latter is perhaps the more relevant since it was carried out on 35 EM cooking pots. It concluded that determination of the 48 elements by ICP-ES and ICP-mass spectrometry (MS) was sufficiently precise and accurate, with some exceptions related to the procedure used for sample preparation. Ti, Zr, Cr, Lu, and Yb were problematic, perhaps because of the acid dissolution method, but classification of the data proved the same irrespective of the analytical technique. For the coarser samples, ICP and XRF appeared better than NAA because they determined major, as well as the minor and trace, elements. All the main instrumental techniques of ceramic analysis, including AAS and, in a minor role, ICP-ES, feature in a large program of multi-laboratory analysis (carried on in the early 1990s) of a ceramic standard prepared at the Fitch Laboratory (FL) at the British School at Athens in 1985, the so-called Lefkandi Brick standard (Jones 1994b); Jones (forthcoming) has presented its results.

27

3.2. Aims of New Analytical Work Having reviewed the overall background to the present inquiry, it is time to formulate more concisely its aims, scope, and development. 1. As already explained, the project was conceived as a logical continuation of the previous studies, and thus was planned with chemical characterization alone in mind. Almost all samples were submitted to chemical characterization, which spanned the period 1986–1991. It was only during the project’s early stages that the need to incorporate petrographic analysis was recognized, and this was in response to the developments in methodology mentioned above, as well as the encouraging results that petrographic analysis was generating at that time. Thus petrographic analysis was perforce integrated into a project without, on the one hand, benefiting from influencing its design and yet, on the other hand, as the results in Chapter 5 make clear, contributing very significantly to the project’s outcome. In the first instance, this has meant that this technique was initially restricted to those jars for which the original sample taken for chemical analysis was also amenable to thin-section preparation. Day began in 1989 with a reinvestigation of the Mycenae jars previously examined by Riley, proceeding in 1991 to many of the SJs from Crete already analyzed chemically, some from Thebes (Day and Haskell 1995), and ending with those from the Iria wreck (Day 1999) and Cannatello in Sicily in 1995 (Day and Joyner 2005). Since the completion of the project, Day has continued study of other SJs at Kommos, and particularly their relationship with the “short-necked amphora” found at that site, as well as Canaanite and Egyptian transport jars, in collaboration with J. Rutter and V. Kilikoglou. 2. The project intended to sample for laboratory analysis representative numbers of typologically identified SJs from as many findspots within the Aegean and beyond as possible and to include ISJs not previously considered. The sampling was to include recent finds (e.g., those from the Iria wreck found in 1990). At Mycenae, in addition to the jars, we also sampled caps, plugs, and stoppers, with a view to understanding something of the sequence of events beginning with the original sealing of the jars’ contents, to their decanting and the jars’ possible subsequent

28

RICHARD E. JONES AND PETER M. DAY

reuse. As described elsewhere (Ch. 9, Sec. 9.2.2.7), Mycenae is unique as a findspot, because so many SJs were found with their stoppers and (seal-impressed) caps. This feature, which was partly explored by Riley (1981), has been integrated into the present study. 3. The chemical and petrographic data presented in Chapters 4 and 5, respectively, have been treated in a flexible manner, consistent with the constraints on interpretation outlined above. But underlying the data treatment has been the identification of well-defined chemical composition groups and petrographic fabric groupings. Knowledge of the association of samples and their identity within a chemical group and within a fabric grouping constitutes information whose value should not be underestimated. The next step associates a group or grouping with reference material with sufficient confidence that allows some statement about origin to be suggested. As has long been recognized, negative statements about origin may be made more easily than positive ones. In Chapter 6, the independently reached statements about possible origin are compared and, where possible, a consensus is reached. Finally, Chapter 7 examines the correlations among the respective stylistic, chemical, and petrographic classifications, with a view to establishing a consensus on the groups’ possible origins.

4. The importance of Mycenae and Chania as two principal findspots of ISJ/SJs is recognized in two ways: first, by integrating the data from two techniques of chemical analysis, AAS and NAA, and from petrographic analysis, and second by examining the chemical compositions of a range of fabrics related or similar to those of the SJs. Some of the chemical and petrographic data for SJs from Mycenae (House of the Oil Merchant) have already been published (Day 1995a; Jones 1995; Tomlinson and Day 1995). 5. Although issues of origin dominate this study, the technological characteristics of SJs have been investigated, if only indirectly. The main aim has been to relate these to what is already well known about the production of large vessels of the Late Bronze Age in the Aegean (e.g., that outlined, perhaps most relevantly, in the account of production technology in the kiln at Kommos [Shaw et al. 2001], where the earlier oval-mouthed amphorae, which are clearly a precursor of the SJs, were produced), as well to complement the work of Leonard et al. (1993) on (small) decorated SJs. Functional analysis in the form of organic residue analysis was excluded from the study because expertise at the time was lacking and resources were insufficient to do justice to the large number of samples that would be necessary to reach meaningful conclusions.

4

Chemical Analyses by Richard E. Jones

4.1. Sampling Strategy and Procedures The two basic criteria for selection were to sample SJs as extensively and intensively as possible, according to the circumstances of availability and accessibility. In practice, the sampling strategy varied from site to site: where the number of examples at a given site was limited—say, fewer than 10—effort was made to sample all of them, irrespective of type. At sites where the SJ finds were more numerous, there was room for some selectivity. At Thebes, specific ISJs were sought, as well as some balance between the main decorative classes, DoL (including octopus) and LoD, while attention was paid—at Mycenae, for instance—to the combination of jars, their stoppers, and caps.

4.1.1. Whole Jars Completely preserved jars were sampled with an electrically driven 2-mm-diameter tungsten carbide drill

head, in the base in four or more different positions, with at least 200 mg collected for chemical analysis only.

4.1.2. Restored Jars Where the jar had been restored whole, it was possible (with the assistance of a museum conservator) to detach a body sherd, from which a small fragment could be removed with pliers. After removal of glue or plaster of paris, this fragment was suitable for petrographic analysis. Following a further check that any residue, glue, or plaster of paris had been fully cleaned off, the weathered surface and any paint or slip were removed either with the drill head or with a steel scalpel blade; the cleaned fragment was then crushed to fine powder in an agate mortar for chemical analysis. At least 300 mg of powder was collected. Where the jar was partial, commonly with the upper half remaining, a small fragment could be broken off to provide a suitable sample for both chemical and petrographic analysis. In some instances (e.g., MIN03),

30

RICHARD E. JONES

however, only drilling was feasible, and here great care had to be exercised to prevent any contamination of the collected powder during drilling.

4.1.3. Sherds A fragment was broken off, providing a suitable sample for both chemical and petrographic analysis. Samples taken from large sherds were preferred. An unanalyzed fragment was always retained for reference purposes in the Fitch Laboratory. (FL) For each analysis, the form of the sample, either a drilling or a fragment ground to powder, was recorded. In a number of cases, it was possible to take multiple samples by drilling and by detaching a small fragment. During sampling, it was sometimes possible to make observations on the relative hardness of the fabric, firing history, and construction detail.

4.2. Analytical Procedures All the chemical analyses were carried out in the Fitch Laboratory at the British School at Athens by atomic absorption spectrometry (AAS), using a Phillips SP2900 instrument. The analytical procedure was first published by Liddy in 1989 and later described in greater detail (Liddy 1996; see also Brodie and Steel 1996). Sample preparation involved drying in an oven at 110°C for several hours, followed by fusion of 25 mg of finely powdered subsample with lithium metaborate in a platinum crucible. The resulting glass was dissolved in dilute nitric acid. Calcium, magnesium, iron, manganese, chromium, nickel, sodium, and potassium were analyzed in an air-acetylene flame; aluminium, titanium, and silicon in a hotter acetylene-nitrous oxide flame. The quantitative determinations of these elements (in their oxide form) were made with reference to manually generated calibrations prepared with a number of ceramic standards and so-called Pot mix standards covering the full elemental concentrations ranges (oxides of Al 10%–30%, Ca 0%–30%, Mg 0.5%–10%, Fe 3%–15%, Ti 0.5%–1.5%, Na 0.2%–2%, K 0.4%–4%, Mn 0.010%–0.3%, Cr 0.010%–0.3%, Ni 0.005%–0.1%). The ceramic standards consisted of well-characterized Greek pottery of long standing (Jones 1986b, tables 2.2, 2.3): Mycenae 22 and 23, Knossos 38, and New Knossos Standard; these were supplemented by the Lefkandi Brick standard as discussed in the next section (Jones 1994a; forthcoming) and British Ceramics Standard 269.

Silica determinations varied between ca. 40% and 65%. Taking the median of ca. 55%, when summed with the other 10 element contents, gave totals ranging from 90% to 100%. But some 20% of samples gave summed total compositions in the 80%–90% range, suggesting that a significant number of silica determinations were underestimating the true value. Although it was possible to identify the cause of the problem—attributable to incomplete fusion and the consequent presence of small amounts of undissolved silicate in the sample solution— it proved difficult to remedy the problem satisfactorily. As a result, silica contents are not quoted in the composition tables. The Fitch Laboratory’s routine procedure was to analyze batches of up to 30 samples, excluding the six ceramic standards. That the SJs were collected and subsequently analyzed over a long period of time (from the end of 1986 to 1991) had two consequences. First, the SJs were not analyzed in successive batches, and some batches comprised SJs alone, while other batches included samples other than SJs. Second, as was the case with the OES analyses (reported by WCISJ, 61–63), the performance characteristics of which the analytical procedure was capable, in particular the long-term reproducibility, had to be monitored frequently. The relevant data on these issues are presented in the next section.

4.3. Performance Characteristics of AAS, and Comparability with OES The performance characteristics of AAS at the Fitch Laboratory have been estimated in a number of ways, first with respect to a large inter-laboratory and intertechnique analysis carried out by the writer (Jones 1994a). In 1983, the FL decided to add to its series of ceramic standards to use in its chemical characterization work on Greek pottery. The laboratory was interested in a Greek clay that was available commercially and therefore in bulk, whose composition had the following requirements: CaO