The Archaeology of Malta

767 16 34MB

English Pages [472]

Report DMCA / Copyright


Table of contents :
Series information
Title page
Copyright information
Table of contents
List of illustrations
List of tables
List of boxes
Abbreviations and Some Common Maltese Terms
1 Malta’s Archaeological Past
Environmental Setting of the Island Cultures
The Learned Scholar, Antiquarian, Artist and Collector
History of Archaeological Investigation
New Archaeological Directions during the Twentieth Century
2 The First Settlers and Farmers
Pre-Neolithic Human Presence
Early Neolithic Island Colonisation
Ghar Dalam Phase (5000–4300 BC)
Caves and Habitation Sites
Domestic Houses and Dwellings
Early Neolithic Economy
Ghar Dalam Material Culture
Lithic Industry
House Closure and Ritual Deposition
Grey Skorba Phase (4500–4400 BC)
Aspects of Grey Skorba Material Culture
Ceramic Production
Lithic and Other Implements
Personal Adornment
Red Skorba Phase (4400–4100 BC)
Architectural Remains: A Question of Function
Ceramic Production in the Red Skorba Phase
Lithic and Other Implements
Personal Adornment
3 The Culture of the Megalith Builders
Domestic Contexts
Żebbuġ Phase (CA. 4100–3800 BC)
Material Culture in the Żebbuġ Phase
Ceramic Production
Lithic and Other Tools
Funerary Contexts
Mġarr Phase (CA. 3800–3600 BC)
Material Culture
Ġgantija Phase (CA. 3600–3200 BC)
Domestic Contexts
Ceramic Production
Other Cultural Artefacts
Burial Sites
Saflieni Phase (3300–3000 BC)
Tarxien Phase (CA. 3000–CA. 2500 or Possibly 2200 BC)
Ceramic Industry during the Tarxien Phase
Aspects of Material Culture
Megalithic Structures
Technical Aspects of Megalith Building
Quarries, Engineering, Logistics
Construction Techniques
Functions of the Megalithic Buildings
Spatial Distribution, Territories and Social Structures
Use of Space
Exterior Space Associated with the Monuments
Interior Space within the Monuments
Sacred Spaces
Belief Systems and Ritual Practice in Neolithic Malta
Possible Shades of Shamanism
Altered States of Consciousness: Hallucinations and Iconography
Life-Cycle Symbolism
Upright Figures: Phallic Symbols, Pillars and Baetyls
Individuality: Body Shape and Adornment
Red Ochre and Colour Symbolism in Malta
Dualism and Triadism: Expansion of the Megalithic Structures
Funerary Contexts: Caves, Graves and Hypogea
4 Pushing Boundaries at the End of the Megalithic Building Period
Field-Furrow Scars or Cart Ruts?
Geomorphology of the Maltese Bedrock
A Strategy for Ancient Field Production
Intentional Closure of Neolithic Sites
Reasoning behind Deliberate Destruction of Structures
5 New Directions
A Cultural Interlude: Thermi Ware (Grey Ware)
Tarxien Cemetery Phase (CA. 2400–1500 BC)
Tarxien Cemetery Phase Settlement Sites
Tarxien Cemetery Cultural Assemblage
Ceramic Production
Figurative Artefacts
Dolmen Sites, Standing Stones and Grave Markers
Tarxien Cemetery Phase: Economic Aspects
Aspects of the Early to Middle Bronze Transition
The Borġ in-Nadur Phase (1500–CA. 1000/700? BC)
Fortified Sites
Non-Fortified Sites
Cave Sites
Aspects of Middle Bronze Material Culture
Ceramic Industry
Borġ in-Nadur Economic Activities
The Closing Years of the Middle Bronze Age
6 East Meets West
The Late Borġ in-Nadur Phase
Borġ in-Nadur Funerary Practice
Baħrija Phase
Ceramic Tradition
The Need for Cultural Definition
Sites with Bahrija Finds
On the Eve of Cultural Change
Phoenician Interests in the West
The Nature and Date of Phoenician Colonisation in the West
Chronological Issues for Malta
The Process of Phoenician Colonisation
Phoenician Funerary Practice
Cultural Traits and Economic Activities
7 Melita and Gaulos during the Punic Period
Land-Use and Settlement Patterns
The City and Other Urban Centres
Rural Contexts
Punic Rural Round Towers
Harbours and Port Facilities
Punic Culture Traits
Ceramic Industry
Beliefs, Rituals and Sacred Precincts
The Temple of Astarte-Juno
The Temple of Melqart-Hercules
Gozo: The Evidence for Punic Sacred Precincts
Funerary Contexts
Symbolism in Tomb Architecture and Contents
Symmetrical, Paired and Tripled Iconography
False Doorways and Empty Niches
Elixirs within the Tomb Setting
Invocation of a Divine Presence in Funerary Contexts
8 Malta’s Place in the Roman World
How ‘Roman’ was Roman Malta?
Politics and Government
Land-Use and Settlement Patterns
Urban Gozo
Malta’s Rural Settlements
The Bathhouse
Material Culture: Pottery and Other Manufactures
Maritime Trade and Harbour Infrastructure
Religious Practice and Sacred Sites
Mdina: Temple of Apollo
Mtarfa: Temple of Proserpina
Mithraism and the Mystery Cults
Other Cultic Associations with Mithraism
Funerary Practice: The Transformation of Punic Burial Sites
Closing Remarks
Appendix A Malta Radiocarbon Dates
Appendix B Gazetteer of Sites in Malta
Index of Features and Cultural Phases at Sites
List 1: Għar Dalam Sites
List 2: Grey Skorba Sites
List 3: Red Skorba Sites
List 4: Żebbuġ Sites
List 5: Mġarr Sites
List 6: Ġgantija Sites
List 7: Saflieni Sites
List 8: Tarxien Sites
List 9: Megalithic Structures
List 10: Sites with Tarxien Cemetery Material
List 11: Dolmen and Cairn Sites
List 12: Stone Circles
List 13: Menhir Sites
List 14: Borġ in-Nadur Sites: Fortified Settlements (Including Naturally Strategic Sites)
List 15: Borġ in-Nadur Sites: Non-Fortified Bronze Age Settlements or Minor Finds
List 16: Caves Used in the Bronze Age
List 17: Rock-Cut Pit (Bell-Shaped or Silo Pit) Locations
List 18: Sites with Baħrija Pottery
List 19: Fortified Towers
List 20: Grape-Pressing Pans
List 21: Punic- and Roman-Period Houses
List 22: Olive Oil Processing Sites
List 23: Rock-Cut Cisterns
List 24: Roman-Period Bathhouses
1 Malta’s Archaeological Past
2 The First Settlers and Farmers
3 The Culture of the Megalith Builders: The Late Neolithic of Malta
4 Pushing Boundaries at the End of the Megalithic Building Period
5 New Directions: The Appearance of the Axe-Bearers
6 East Meets West: Phoenician Mariners, Merchants and Settlers
7 Melita and Gaulos during the Punic Period
8 Malta’s Place in the Roman World
Recommend Papers

The Archaeology of Malta

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview


The Maltese Archipelago is a unique barometer for understanding cultural change in the central Mediterranean. Prehistoric people helped to reshape the islands’ economy, and when Mediterranean maritime highways were being established, the islands became a significant lure to Phoenician colonists venturing from their Levantine homeland. Punic Malta also sat at the frontline of regional hostilities until it fell to Rome. Preserved in this island setting are signs of people’s endurance and adaptation to each new challenge. This book is the first systematic and up-to-date survey of the islands’ archaeological evidence from the initial settlers to the archipelago’s inclusion into the Roman world (ca. 5000 BC–400 AD). Claudia Sagona draws upon old and new discoveries, and her analysis covers well-known sites such as the megalithic structures, as well as less familiar locations and discoveries. She interprets the archaeological record to explain changing social and political structures, intriguing ritual practices and cultural contact through several millennia. Claudia Sagona is a Senior Fellow in Archaeology at The University of Melbourne. She is the author of several books on Malta, including Looking for Mithra in Malta (2009), and the editor of several other books, including Ceramics of the Phoenician-Punic World (2011) and Beyond the Homeland:  Markers in Phoenician Chronology (2008). Her articles have appeared in academic journals including Anatolian Studies, Anatolia Antiqua, Ancient Near Eastern Studies, Mediterranean Archaeology and Oxford Journal of Archaeology. In recognition of her contribution to Malta, she was made an honorary member of the National Order of Merit of Malta in 2007.


Series Editor NORMAN YOFFEE, University of Michigan

Editorial Board SUSAN ALCOCK, Brown University TOM DILLEHAY, Vanderbilt University TIM PAUKETAT, University of Illinois STEPHEN SHENNAN, University College London CARLA SINOPOLI, University of Michigan DAVID WENGROW, University College London The Cambridge World Archaeology series is addressed to students and professional archaeologists and to academics in related disciplines. Most volumes present a survey of the archaeology of a region of the world, providing an up-to-date account of research and integrating recent findings with new concerns of interpretation. Whereas the focus is on a specific region, broader cultural trends are discussed and the implications of regional findings for cross-cultural interpretations considered. The authors also bring anthropological and historical expertise to bear on archaeological problems and show how both new data and changing intellectual trends in archaeology shape inferences about the past. More recently, the series has expanded to include thematic volumes.

Recent Books in the Series FRANCES F. BERDAN, Aztec Archaeology and Ethnohistory PETER MAGEE, The Archaeology of Prehistoric Arabia KOJI MIZOGUCHI, The Archaeology of Japan MIKE SMITH, The Archaeology of Australia’s Deserts A. BERNARD KNAPP, The Archaeology of Cyprus LI LIU AND XINGCAN CHEN, The Archaeology of China STEPHEN D. HOUSTON AND TAKESHI INOMATA, The Classic Maya PHILIP L. KOHL, The Making of Bronze Age Eurasia LAWRENCE BARHAM AND PETER MITCHELL, The First Africans ROBIN DENNELL, The Palaeolithic Settlement of Asia CHRISTOPHER POOL, Olmec Archaeology and Early Mesoamerica SAMUEL M. WILSON, The Archaeology of the Caribbean RICHARD BRADLEY, The Prehistory of Britain LUDMILA KORYAKOVA AND ANDREJ EPIMAKHOV, The Urals and Western Siberia in the Bronze and Iron Ages DAVID WENGROW, The Archaeology of Early Egypt PAUL RAINBIRD, The Archaeology of Micronesia PETER M. M. G. AKKERMANS AND GLENN M. SCHWARTZ, The Archaeology of Syria TIMOTHY INSOLL, The Archaeology of Islam in Sub-Saharan Africa

THE ARCHAEOLOGY OF MALTA From the Neolithic through the Roman Period

Claudia Sagona University of Melbourne

32 Avenue of the Americas, New York, NY 10013-2473, USA Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning and research at the highest international levels of excellence. Information on this title: © Claudia Sagona 2015 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2015 Printed in the United States of America A catalog record for this publication is available from the British Library. Library of Congress Cataloging in Publication Data Sagona, Claudia, 1956– author. The archaeology of Malta : from the Neolithic through the Roman period / Claudia Sagona, University of Melbourne.   pages  cm. – (Cambridge world archaeology) Includes bibliographical references and index. ISBN 978-1-107-00669-0 (hardback) 1.  Malta – Antiquities.  2.  Excavations (Archaeology) – Malta.  3.  Malta – History – To 870. I. Title.  II.  Series: Cambridge world archaeology. DG989.5.S235 2016 937′.85–dc23   2015009569 ISBN

978-1-107-00669-0 Hardback

Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party Internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

For the People of Malta. They know their islands are extraordinary, and it has been a delight for me to find out why.


List of Illustrations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . page x List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   xiii List of Boxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   xiv Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Abbreviations and Some Common Maltese Terms. . . . . . . . . . . . . . . . . . . . . .   xix 1 Malta’s Archaeological Past . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 The First Settlers and Farmers . . . . . . . . . . . . . . . . . . . . . . . . .   20 3 The Culture of the Megalith Builders: The Late Neolithic of Malta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .   47 4 Pushing Boundaries at the End of the Megalithic Building Period. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 5 New Directions: The Appearance of the Axe-Bearers. . . . . . . . . . . 134 6 East Meets West: Phoenician Mariners, Merchants and Settlers . . . . 171 7 Melita and Gaulos during the Punic Period . . . . . . . . . . . . . . . . . 218 8 Malta’s Place in the Roman World . . . . . . . . . . . . . . . . . . . . . . . 264 Closing Remarks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Appendix A: Malta Radiocarbon Dates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 Appendix B: Gazetteer of Sites in Malta. . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 Bibliography. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397 Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441



1.1 Map of Malta page 4 1.2 Map of Malta with key sites mentioned in the text. (Inset) Cross section showing the geology of northern Malta 5 1.3 Photograph reproduced from the Letters Book of the Archaeological and Geological Society of Malta 1860 9 2.1 (1–3) Plan of the Għar Ħasan cave site and possible prehistoric art. (4) Reputed Neanderthal taurodont teeth from Għar Dalam cave site 21 2.2 Għar Dalam cultural remains 27 2.3 Grey Skorba cultural remains 36 2.4 Red Skorba cultural remains 39 2.5 Red Skorba pottery 44 3.1 Zebbug pottery forms 51 3.2 Zebbug cultural remains 53 3.3 Mgarr cultural remains 56 3.4 (A) Kordin III megalithic building plan. (B) Possible menhir at Xemxija. (C) Menhir at Kirkop 57 3.5 Ggantija cultural remains 60 3.6 Ggantija pottery 62 3.7 Plan of the prehistoric Xemxija tomb site and details of tombs 5 and 6 65 3.8 Saflieni-phase pottery forms 66 3.9 Tarxien-phase pottery jars 68 3.10 Tarxien-phase pot and bowl forms 69 3.11 Tarxien pottery bowl, ladle, lid and tray forms 70 3.12 Extract from T. Zammit’s Archaeological Field Notes no. 11, 23.8.1915, p. 37, depicting convex-concave worked-stone bosses 72 3.13 Tarxien cultural remains 75 3.14 View and plan of the Ggantija (Gozo) megalithic structure 79 3.15 Views of the Ħagar Qim and Mnajdra megalithic complexes 80 3.16 Features at the Misqa Tanks located near Ħagar Qim and Mnajdra 82 3.17 Common hallucinated patterns seen in trance states and comparable designs in Maltese prehistoric contexts 97 x

Illustrations 3.18 Neolithic figurines 3.19 (1) Plan of Ħal Saflieni hypogeum. (2 and 3) Plan of the stone circle and burial chambers in the Xagħra Circle hypogeum, Gozo 3.20 (1–3) Bowl from Ħal Saflieni. (4) Relief decoration from Bugibba depicting a maritime scene. (5) Boat graffiti in the Tarxien Neolithic complex 4.1 Western sector of Malta showing the Burmarrad catchment relative to rock-cut groove (field furrow) areas and prehistoric sites 4.2 (1–4) Possible ancient fields identified in Table 4.1. (5–7) Rock-cut features at Naxxar, Xemxija and Skorba 5.1 Thermi-Ware pottery 5.2 Tarxien Cemetery pottery 5.3 Tarxien Cemetery pots, jugs, conjoint and tray shapes 5.4 Tarxien Cemetery figurines 5.5 Dolmens at Ta’ Cenc area, Gozo and Misraħ Sinjura, Malta 5.6 St. George’s Bay rock-cut vats and evidence for Bronze Age basketry in Malta 5.7 Borg in-Nadur plan of site and view of the Bronze Age wall 5.8 Wardija ta’ San Gorg, view of the Bronze Age site, house footings and rock-cut pits 5.9 Plan of and view to the site of Baħrija 5.10 Borg in-Nadur pottery forms 5.11 Borg in-Nadur pottery forms and decorative features 5.12 Borg in-Nadur pottery, painted and white gritty cooking wares 6.1 Late Borg in-Nadur pottery, a cache found in a Mtarfa bell-shaped pit in 1939 6.2 Late Borg in-Nadur phase, thin-walled variant fabric from jars and Baħrija pottery forms 6.3 (1) Phoenician ship. (2–4) Gold medallion and bronze lamp holder from tombs and an inscribed agate fragment from Tas-Silg 6.4 Phoenician tombs near Dingli Cliffs and Xemxija 6.5 Clay sarcophagus from Għar Barka and nine main rock-cut tomb plans of the Phoenician-Punic periods 6.6 Melitan archaic phase I pottery forms 6.7 Melitan archaic to early phase I showing the development of pottery forms 6.8 Phoenician wine-drinking pottery kit and shallow, rock-cut grape pressing pans near the Misqa Tanks 6.9 Plans of the Phoenician-Punic sacred precinct at Tas-Silg and reconstruction of a three-pillared altar (altar 45) in the complex 7.1 (1–11) Melitan late phase I to early phase II pottery forms. (12–14) Evidence for the Tanit symbol in Malta. (15–21) Evolution of the twin-nozzle lamp in Phoenician-Punic tombs. (22–24) Other lamp forms 7.2 The Phoenician building in Zurrieq showing the Egyptianising cornice

xi 99 102

112 117 120 136 142 143 145 148 150 155 157 158 163 164 166 179 181 192 202 203 207 210 212 214

219 227

xii Illustrations 7.3 (1–12) Melitan phase III pottery forms. (13) Map of ancient sites in the vicinity of Mdina and Rabat 7.4 Plans of a structure in Saqqajja I, Rabat, and of the Punic-Roman villa in Zejtun 7.5 Plan of Punic and Roman house remains at Ta’ Kaccatura (Birzebbuga) with view across the underground cistern and plan of the Tad-Dawl house 7.6 Plan of a Late Punic-Roman house at Ras ir-Raħeb and view from Baħrija to the site 7.7 Plan and view of the house and olive works complex at San Pawl Milqi 7.8 Ancient round towers at Ta’ Gawhar, Ta’ Cieda and Ta’ Wilga 7.9 An elaborate Rabat tomb found in Triq Ferris and carvings in a Phoenician-Punic tomb in Qrendi 7.10 Melitan phase III–IV pottery forms 7.11 Melitan phase III–IV pottery forms 8.1 Plans of the Domus Romana in Rabat and of warehouses and cisterns in Marsa 8.2 Bathhouses from the Roman period of influence at Ta’ Baldu, Ramla Bay, Gozo and Għajn Tuffieħa 8.3 Melitan phase VI pottery forms 8.4 Melitan phase VI pottery forms 8.5 Traces of Mithraic religious practice in the islands 8.6 Salina Hypogeum (southern complex), St. Paul’s Bay, Malta, and the ground plan of three inter-connected caves in Qasam il-Gewwieni

229 232

234 236 238 240 245 246 247 270 278 281 282 287 291


1.1 1.2 4.1 6.1

Radiocarbon dates from cultural contexts in Malta Radiocarbon dates from environmental contexts in Malta Agricultural potential of ancient fields in Malta Phases of Phoenician-Punic Malta based on funerary evidence

page 16 17 125 176



Chapter 1 The Lay of the Land Giovanni Francesco Abela (1582–1655) Chapter 2 Possible Palaeolithic Cave Art at Għar Ħasan Indications of Source Preference within the Lithic Industry


page 3 8 20 33

Chapter 3 Dimples, Pellets and Bosses as Decorative Features Evidence for Offshore Contacts in the Neolithic Period Population Estimates for the Late Neolithic in Malta Colour Symbolism: An Ethnographic Account

71 73 85 107

Chapter 4 Creation of Ancient Fields, Run-Off Farming and Economic Gain Ethnographic Accounts of Soil Manufacture in Malta in the 1830s

124 127

Chapter 5 Disc Figurines of the Tarxien Cemetery Phase Purple Dye from the Murex Seashell

144 152

Chapter 6 Phoenician and Punic – Melita and Gaulos – ‘nn (Anna) Language and Writing The Legend of Anna, of Phoenician Royal Lineage The Chronological Development of Rock-Cut Tombs in Melitan Malta

174 194 195 202

Chapter 7 Indicators of Continuity, Prosperity and Stability Conveyed in Diodorus Siculus’ Description of Malta in the First Century BC Quintinus (Jean Quintin d’Autun) and the Punic-Roman Temple Sites CIS I, 132: Decree by the People of Gaulos

220 249 254

Chapter 8 Tenets of Mithraism



On Monday, 7 January 2008, I paused from my work to look into the distance. Pottery was dancing in front of my eyes after hours spent at the University of Malta working on the finds from the site of Tas-Silg. As I scanned the tangle of stone buildings and rooftops, it struck me that, yes, Malta is small geographically, especially from my perspective as an Australian, but in terms of the ancient experience, these landmasses were never insignificant.The tightly clustered Maltese islands invite first impressions, but an intriguing and complex history will be unmasked with even the most cursory of enquiries. As is so often the case with families that have settled in distant places, there is a yearning to rediscover homelands. This strong link first brought my family and me to Malta. As with others familiar with Malta’s ancient past, front and centre in our imagination was the Neolithic period, with its monumental constructions. But William Culican, my mentor at The University of Melbourne, had always spoken so highly of Malta and so passionately about the ancient Phoenicians who made the archipelago their home that I had set out to learn more. My aim was to write an article on the colonial interests of these ‘merchant venturers’ as he referred to them. Well, the work grew beyond an article, and I was firmly captured – hook, line and sinker – by an area of research that is, happily, still as enthralling today as it was two decades ago. As time has gone on, I have been steadily drawn to every aspect of the islands’ ancient past. So it is that I have undertaken this survey of the archaeology of Malta. Aside from books aimed at a general audience, we are in need of a detailed synthesis that brings together the data from the earliest human colonisation to the Roman period. This book is written to fill that gap. The reader may grapple with new landscapes and names that are often quite difficult, but every effort has been made to overcome such unfamiliar territory. This study will surely not be the last word on the topic. But if it is received in the spirit in which it is written – to encapsulate how people have lived in Malta throughout the xv

xvi Preface millennia and to trace clearly each step of their cultural development and adaptation – then it would have served its purpose. During my research on ancient Malta, I have endeavoured to go back to the original sources and to place the islands in the wider Mediterranean setting. From this approach, the evidence often presents other angles and aspects overlooked in the past, and I have not shied from presenting different interpretations. I am greatly honoured by Norman Yoffee’s invitation to write for the Cambridge World Archaeology series. As it happened, I had been contemplating a synthesis on Malta for some time, and this book is the sum of information gathered since I first visited the islands over twenty years ago. To that end, my efforts would be the poorer without the help of many scholars, colleagues and friends. In terms of all things archaeological, and in one way or another, they have kept me informed of new discoveries, facilitated site visits, provided me with research opportunities, given notice of new literature or of the islands’ archives and kept me in touch with current trends. They put me back on the research track when I veered off course. In various ways and for their help, I  am grateful to the following (note that workplace affiliations may have changed over the years):  Government departments: Mr. P.  Gauci and Mr. Albert Callus, Private Secretaries to the Parliamentary Secretary, Ministry of Justice and the Arts, in Malta; the Honourable Clemente C.  Zammit (former Maltese Consul General in Melbourne). Staff of the superintendence or museum departments in Malta: Tancred Gouder,† Anthony Pace, Nathaniel Cutajar, Kenneth Gambin, Reuben Grima, Katya Stroud, Sharon Sultana, Suzannah De Pasquale, Theresa Vella and MariaElena Zammit. Museums: Gillian Varndell (Curator, Neolithic Collections, Neolithic of Britain and Europe, Department of Prehistory and Europe, British Museum), Marianne Eve and Marta Mroczek, British Museum, the trustees of the British Museum for their permission to reproduce some Maltese pottery (Figures  2.3, 2.5, 3.9, 3.11, 5.3, 5.11 and 6.2) and Sergei Kovalenko (Pushkin Museum). Universities: Anthony Bonanno, Mario Buhagiar, Anthony Frendo, Horatio Vella and Nicholas Vella (at the University of Malta); S. Baron (Register, Parliament House Art Collection, Canberra); the Italian Mission to Malta and Giulia Recchia (and for the photograph in Figure  6.3, no.  4; Università di Foggia), Maria Pia Rossignani† (Università Cattolica di Milano), David Trump (Cambridge), B.  Rawson and E.  Minchin (Classics Department, Australian National University), Piers Crocker and Gary Stone (formerly of the Australian Institute of Archaeology, Melbourne), Elizabeth Pemberton (formerly of The University of Melbourne) and Ken Sheedy and Clare Rowan (the Australian Centre for Ancient Numismatic Studies, Macquarie University). Libraries: Joseph Busuttil (former Chief Librarian, National Museum of Valletta), Joseph Caruana (Officer in Charge of the National Archives of Malta, Santo Spirito, Rabat), The University of Melbourne, particularly staff

Preface of the university library and the inter-library loans department, who keep the window on the world of scholarship firmly open. Maltese holdings in private museums: Canon John Azzopardi (former Curator of the Cathedral Museum, Mdina, and St. Paul’s Parish Museum, Wignacourt College, Rabat), Fr. Victor Camilleri (St Agatha’s Museum, Missionary Society of St Paul, Rabat), the Maltese Dominican Province, Rabat, and Christiane Ramsay Pergola, Baroness of Tabria (Custodian of the Scicluna Collection in the Palazzo Parisio, Naxxar). Custodians of other collections: Joseph Attard Tabone, Joseph Sammut,† Mary Sammut, the Custodian of the ‘Courtyard’ Collection, Fr. Joseph Fsadni, Helen Mahr, Carmel Fsadni and Mario Fsadni (Custodians of the Chevalier Paul Catania Collection), Daniel Micallef and Pauline Micallef (Rabat) and Fr. F. Abdilla and the staff at the Qormi Parish Church. Those who assisted me regarding past projects, friends and colleagues: Maxine Anastasi, Kevin Borda, Joseph Calleja, Kristian Checuti Bonavita, Joseph Borg (Melbourne), Josef Briffa, SJ, Anton Bugeja, Keith Buhagiar (also for his photograph in Figure  8.5, no.  5), Michelle Buhagiar, Celine Farrugia, Rebecca Farrugia, Patrick Galea, Joseph Huber (Melbourne), Josric Mifsud, Francisco Núñez Calvo, James Sacco, George Said-Zammit, Hanna Stöger, Davide Tanasi, Ernest Vella, Prof. Vella Bonavita (Perth) and Isabelle Vella Gregory, among many others. I owe a debt of gratitude to the Maltese community in Melbourne, which has continued to support my interest in Malta and has helped my research in all sorts of ways. I think it only fitting to mention the Demanuele family and the staff at the Grand Harbour Hotel in Valletta, who have made my many visits to Malta all the more enjoyable. I am very grateful to my colleague and friend Abby Robinson, who read the manuscript closely. The work is definitely better for her editing and for her valuable comments. Jarrad Paul’s help in preparing the tables concerning radiocarbon dates is very much appreciated. Unless otherwise stated, illustrations are my own or adapted by me for this book. To my daughter Amadea, I love the way you keep me grounded in the present when I could so easily get lost in the past. And to Tony Sagona, my strongest supporter – we are still travelling the archaeological road together after forty years – thank you just does not seem enough. My sincere thanks go to all of the editorial staff at Cambridge University Press for their assistance and guidance in preparing this book for publication.




a.k.a. also known as asl above sea level għajn spring għar cave, cavern GIS Geographic Information System g nien garden ħal village kbir large l/o limits of (on the outskirts of a region) MAR 1904–2002 Annual Report on the Working of the Museum Department (title varies), compiled by the curators and/or museum directors. Malta:  Government Printing Office. Superseded by Superintendence of Cultural Heritage’s Annual Reports, from 2003 (available on-line). MEPA Malta Environment and Planning Authority. MEPA’s website provides online pdf versions of government notices (prefixed: gn) appearing in the Gazzetta tal-Gvern ta’ Malta regarding scheduled properties (many of archaeological interest) as well as access to land survey sheets. NB T. Zammit’s Archaeological Field Notes, nos. 1–16 (1905–30), handwritten records largely concerning field work in Malta, held in the National Museum of Archaeology,Valletta. tal or ta’ of the triq street, road TSG96 prefix code for the site of Tas-Silg (1996) wied valley xagħra rocky ground



Relief, apprehension and curiosity may have occupied the thoughts of the first people to set foot on Malta’s shore sometime around the late sixth millennium BC. There was risk in the sea crossing, but the goal, a landmass which barely registered on the horizon from their homeland, probably in Sicily, must have been within their maritime capabilities. Thus conquered, this journey may have been made often by early seafarers, often enough to gauge that the islands could support their way of life before they made the commitment to stay. From that point forward, the human occupation of the Maltese Archipelago has contributed to a remarkable tableau shaped by endurance, ingenuity and adaptation.1 Most travellers to the archipelago today find their way to historic Mdina and Rabat, two towns sitting cheek by jowl on high ground in the heart of the main island – one of its most historic and picturesque locations. Around this central vantage point, visitors can see the expanse of Malta stretching to its shores.Their first impression might be that the landmass is not great and that its landscape, verdant in winter but parched by harsh summers, shows the unmistakable imprint of lives lived through countless passing generations. Looking east, encroaching urban areas are a compact hotchpotch of yellow limestone buildings that dazzle the senses in the midday sun. They advance inland from the coast as an unstoppable tide of development. Roads are soon lost from sight among houses but can be traced as they thread their way through rural areas north and west.Though the roads beckon to be followed, Malta’s western reaches are much less travelled. Close to Mdina, the valley dips and rises west to the Mtarfa plateau, and beyond that, the island tilts upward. Agricultural production takes place within the confines of small land holdings clearly defined by dry-stone ­walling  – stone is never in short supply in these islands. The demarcation between urban zones and the countryside is often abrupt, and remarkably, given the small size of the islands, it is still possible to have a sense of the remote in some pockets of the countryside. Travellers seldom make their 1


The Archaeology of Malta way to the western edge of the island, but what they miss is the dramatic cliff edge at land’s end plunging into the vast sea which stretches out to an unbroken horizon. If visitors travel to the north island, Gozo, there is less development, the land is weathered to a series of distinct plateaus and the impression of the rural dominates the senses. ENVIRONMENTAL SETTING OF THE ISLAND CULTURES

Past visitors to the archipelago were equally struck by its harsh appearance, but usually any negative first impressions dissipated as the traveller explored the islands. An overriding spirit of enquiry, engendered by the Maltese landscape, was captured by Sir Richard Colt Hoare in 1819: Nothing can be more uninteresting than the first aspect of this territory, to those who enter it on the land side. An extent of country, rather hilly than mountainous; thousands of stone walls, dividing and sustaining little enclosures, formed like terraces; villages so numerous as to bear the appearance of a continued town; and the whole raised on a barren rock, with scarce a tree to enliven the dusky-tinted view; such are the objects which first meet the eye of a traveller. This cheerless scenery struck me the more forcibly, after having quitted so recently the fertile and verdant regions on the opposite coast of Sicily. Indeed, I could nowhere find a parallel to it, except in some parts of the dreary Contea di Modica.Yet this was the spot which poetry and fiction had assigned as the voluptuous abode of Calypso! However, nature under all her forms, and in all her productions, frequently veils beneath an unpromising exterior, singularities, and even excellencies, which awaken curiosity, and raise admiration. Such was the case at Malta.2

The geology of the Maltese Archipelago proved to be just as alluring to early travellers and scholars as its ancient human history.3 Comprising three main islands  – Malta, Gozo and Comino  – its area spans a total of only 316 km2 with additional uninhabited minor landmasses. Located in the central Mediterranean region, the position of the archipelago has long been recognised as economically, politically and strategically advantageous. Insularity must have been a defining quality for inhabitants who called the islands their home.4 The sea forms both a barrier and a means of connectivity through maritime communication to not-so-distant neighbours: from shore to shore, Sicily lies some 80.5 km from Gozo; from Malta to Pantelleria, it is around 224 km; to Linosa, 132 km; to Lampedusa, 161 km; and to Tunisia, 303 km. Sea levels dropped by some 120 m during the last glacial period (from 70,000 BP), at which time Malta and Sicily were linked to Italy in a land mass referred to as the Hyblaean Plateau. If people did frequent these southern reaches of the European continent, however, they also withdrew into the large territory to the north rather than risk becoming isolated.5 Around 12,000 BC, whatever narrow land bridges remained were permanently submerged, isolating Malta. Land links to Africa that might have afforded the passage of human and animal

Malta’s Archaeological Past migrations are unlikely on present evidence, with a 40-km stretch of open water forming a perpetual barrier.6 Geological components of the islands consist of layers of limestone – Lower Coralline, Globigerina and Upper Coralline – laid down as marine deposits. Blue clay (or marl), over which lies a greensand (sandstone) layer, is trapped between the Globigerina and Upper Coralline strata. Globigerina is softer than the other limestone deposits and varies in hue from whitish to rich yellow. It is the preferred building material in Malta, and a number of large quarries can be found on the islands which exploit these deposits and vary in thickness from 20 to 200 m.7 This stratum, too, has been subdivided according to depositional layers into Lower, Middle and Upper Globigerina. Small amounts of chert (the term is used for the lesser-quality pale to very dark grey flint with a granular texture),8 which forms in the local Middle Globigerina limestone deposits, were used for tools, but the islands are devoid of high quality-flint. All other materials, such as obsidian (volcanic glass), basalt and metals, found in ancient contexts were imported. THE LAY OF THE LAND

Malta was shaped by a series of uplifts and subsidence to form a tilted landmass, higher on the western side of the island, with dramatic eroded coastal cliffs rising abruptly to over 200 m above sea level and gradually sloping down to the east (Figure 1.1). The landmass is fractured by a series of smaller fault lines which lie approximately north-east–south-west, particularly evident in the western half of the island (Figure 1.2, inset). A major division is formed by the Great Fault line (along which are built British fortifications known as the ‘Victoria Lines’) in Malta spanning from Fomm ir-Riħ in the west to the eastern shore at the foot of Madliena Tower; north of this, smaller fractures have shaped the islands into a series of ridges and valleys (from south to north, Bingemma, Bidnija, Wardija, Bajda, Mellieħa and Marfa Ridges). In southern Gozo, a fault line along a similar bearing formed the Ta’ Cenc cliffs.9 As limestone decays, it forms fertile red- to brown-coloured terra soils. Historically, farmers have also made soil by breaking up the stone and adding organic material in the course of intensive land-use practices.10 Generally, soil cover is thin, half a metre or less, with the deepest deposits accumulating in the valley floors. Erosion has had a significant impact, exposing the bedrock over large swathes of land in the west of the island. Deep cores and other environmental samples have facilitated reconstruction of the climatic conditions from ca. 45,000 years BP in the central Mediterranean region, which is likely to have some bearing on climatic conditions in the archipelago. Though the chronological depth of these analyses exceeds the scope


Figure 1.1. Map of Malta: (A) View eastward to In-Nuffara (in 2007); (B) view west to Victoria (in 2009); (C) Fomm ir-Riħ Bay, looking north from Ras ir-Raħeb, with geological strata visible in the cliff face (in 1998); (D) looking west across rocky landscape towards the Misqa Tanks area, near the prehistoric sites of Ħagar Qim and Mnajdra (in 1991); (E) view eastward from Mtarfa to Mdina and surrounding fields (in 2013) (photographs by A. and C. Sagona; map by C. Sagona). 4

Figure 1.2.  Map of Malta. Gozo: (1) Marsalforn, (2) Ta’Kuljat, (3) Ramla Bay, (4) Għajn Abdul, (5) Il-Mixta, (6) Santa Lucija, (7) Kercem, (8)  Victoria, (9)  Tac-Cawla, (10) Santa Verna, (11) Xagħra Circle, (12) Ggantija, (13) In-Nuffara, (14) Il-Hagra l-Wieqfa menhir, (15) Ras il-Wardija, (16) Xlendi Bay, (17) Xewkija, (18) Tal-Knisja, Mgarr ix-Xini, (19) Ta’ Cenc, (20) Mgarr ix-Xini, (21) Għar ix-Xiħ, (22) Għajnsielem, (23) Mgarr, (24) Xatt l-Ahmar. Comino: (25) Santa Marija Bay. Malta: (26) Il-Latmija cave, (27) Mellieħa Bay, (28) Mellieħa, (29) Xemxija, (30) Burmarrad, (31) Tal Qadi, (32) Għajn Tuffieħa, (33) Qala Hill, (34) San Pawl Milqi, (35) Golden Bay, (36) Għajn Tuffieħa 5

The Archaeology of Malta


of this study, of interest are findings concerning the Holocene period, which witnessed the rise of agriculture and animal husbandry in lands to the east, enabling human colonisation of remote island locations in the Mediterranean Sea. The Postglacial period (ca. 10,450–9700 BP) ushered in a warmer, humid climate, a result of which was that forest cover of deciduous and oak trees increased in the central Mediterranean region. Colder and dryer events punctuate the record at around 8400–8000 BP and ca. 6600–6000 BP, but the trend towards more arid conditions prevails and is especially marked in North Africa.11 Whether Malta experienced the same degree of forestation is less certain, but pollen extracted from cores in the silted valley of the Burmarrad (from deposits 14 m deep) indicate that minor stands of Erica, Pistacia and Quercus were present in Malta in a mid-Holocene record characterised by dense scrubland. Analyses of cores taken from this and the Marsa regions have found that soil erosion from highlands led to silting in valleys and the mouths of waterways. Thinning of tree cover and loss of vegetation on slopes and higher areas may have been triggered by arid conditions or, equally possible, such changes may have resulted from land clearing after people colonised the island.12 Marl layers vary in thickness (up to 75 m in places13) and are especially apparent on the eroded coastal margins (see Figure 1.1). This clay source was exploited for pottery production and for architectural use in antiquity.14 The

Figure 1.2. (cont.) Bay, (37) Gnejna Bay, (38) Bidnija, (39) Qala Pellegrin, (40) Fomm ir-Riħ Bay, (41) Ras ir-Raħeb, (42) Il-Kuncizzjoni, (43) Baħrija, (44) Mgarr, (45) Ta’ Ħagrat, (46) Skorba, (47) Zebbieħ, (48) Bingemma, (49) Nadur Tower, (50) Qallilija, (51) Mosta, (52) Fort Mosta, (53) Naxxar, (54) L-Ilklin, (55) Lija, (56) Attard, (57) Birkirkara, (58) Mtarfa, (59) Gnien is-Sultan, (60) Għajn Qajjied, (61) Għajn Klieb, (62) Nigred, (63) Mdina, (64) Domus Romana Museum, (65) Rabat, (66) Għar Barka, (67) Manoel Island, (68) Valletta, (69) Grand Harbour, (70) Kalkara, (71) Birgu (Vittoriosa), (72) Kordin I, II, III, (73) Marsa, (74) Paola, (75) Tal Ħorr, (76), Ħal Saflieni, (77) Tarxien megalithic structure, (78) Santa Lucija, (79) Buleben, (80) Bidni, (81) Zebbug, (82) Dingli Cliffs, (83) Buskett Gardens, (84) Għar Mirdum, (85) Wardija ta’San Gorg, (86) Girgenti, (87) Laferla Cross, (88) Siggiewi, (89) Ta’ Wilga, (90) Misraħ Sinjura, (91) Qrendi, (92) Misqa Tanks, (93) Mnajdra, (94) Ħagar Qim, (95) Mqabba, (96) Ħal Millieri, (97) Tad-Dawl, (98) Malta International Airport, (99) Luqa, (100) Bir Miftuħ, (101) Ħal Kirkop, (102) Safi, (103) Zurrieq, (104) It-Torrijiet, (105) Wied Moqbol, (106) Tal-Bakkari, (107) Ta’ Gawhar, (108) Għar Ħasan, (109) Ħal Far, (110) Zejtun, (111) Birzebbuga, (112) Għar Dalam, (113) Ta’ Kaccatura, (114) Borg in-Nadur, (115) St. George’s Bay pits, (116) Pretty Bay, (117) Marsaxlokk Bay, (118) St. George’s Bay, (119) Tas-Silg, (120) Il-Magħluq Harbour, (121) Delimara Peninsula (map by C. Sagona). (Inset) Cross section showing the geology of northern Malta, its stratified bedrock and fault lines (after Bowen-Jones, Dewdney and Fisher 1961; K. Buhagiar 2007).

Malta’s Archaeological Past impermeable clay layer, with the overlying greensand, traps rainwater that percolates down from the surface to form perched aquifers, which erupt as vital springwater sources. These springs compensate for the relative scarcity of surface water. Aquifers were further tapped at intervals by farmers in rural regions who developed subterranean water galleries, akin to qanat technology, that Keith Buhagiar suggests might date back to Roman and Islamic times in the islands.15 None of the natural water sources can be classed as rivers; rather, weathered valleys (Maltese ‘wied’) may carry water in streams, creeks or wadis, but because farmers have redirected the spring flows, many are dry. Large pits, cisterns and water channels which have captured and controlled water throughout the human occupation of the islands have been documented and will be discussed in the pages that follow. THE LEARNED SCHOLAR, ANTIQUARIAN, ARTIST AND COLLECTOR

The merits of this central Mediterranean location were recognised by ancient classical scholars; Malta may not have been at the forefront of political forces in the wider known world, but the islands did not go unnoticed. Of the Phoenician settlers, their Punic descendants, Hellenised neighbours and inhabitants of the not-so-distant Roman heartland who influenced the course of history for the main islands of Malta and Gozo, there is some written account. Such ancient texts sketched a broad historical framework for the times in which they were written. But the Bronze Age and Neolithic inhabitants before them, the nameless generations, had to wait until the modern era for their history to be uncovered. Classical commentaries, notable among them works by Diodorus Siculus, Cicero and Ptolemy, concerning the islands that they knew as Melita and Gaulos formed a starting point for historical enquiry as early as the sixteenth century AD. In reality, a time before the Phoenician never really concerned the deliberations of the antiquarians. Theirs was a quest to accommodate history as it emerged from biblical, Greek and Latin texts. Early chronicles of the islands can be threaded with fanciful and legendary events, antediluvian notions and mismatched associations between historic accounts and prehistoric ruins.16 By the seventeenth century, an increasingly mobile elite emerged, for whom personal enlightenment could be measured not just by their writings but also by accumulated objects of interest which they gathered during their journeys. In many respects, it would be otiose to focus too deeply on the individuals who penned the first histories of the islands, as numerous studies have already done so with encyclopaedic fervour. Nonetheless, antiquarian accounts and artistic representations are still valued for their depictions of urban centres when the islands were less developed.17 Some will feature throughout this work simply because the written accounts or sketches by these early scholars carry the


The Archaeology of Malta


reader back to a time when development was modest and ancient ruins were still obvious in the landscape, before urban expansion at best hemmed in some relics of the past and at worst obliterated them. Regarding Malta, the starting point for the early modern era is Quintinus (Jean Quintin d’Autun), whose description of the islands, classed as the earliest written account, drew on his personal observations made between 1530 and 1536.18 What survived of early collections of antiquities, particularly those assembled by Giovanni Francesco Abela, later formed the nucleus from which the National Museum of Archaeology grew.19 GIOVANNI FRANCESCO ABELA (1582–1655)

Abela’s intellectual pursuits led him to amass an eclectic array of imaginative curios of less reliable attribution and genuine artefacts from tombs, as well as objects he obtained from digging in ancient sites.The collection was effectively the first museum in Malta, housed in his Villa di San Giacomo in Marsa and arranged aesthetically with provenance factored into the order of artefacts.20 Woven into the fabric of the collection and threaded through his treatise, Malta Illustrata: Della Descrittione di Malta Isola nel Mare Siciliano con le sue Antichita, ed altre Notitie, was the basis for lasting recognition of Malta’s cultural heritage among scholarly society not just in the islands but also throughout Europe. Abela’s book was the first systematic account of Malta’s known history. Despite, or perhaps because of, the acclaim given to Abela’s private museum, his collection was plundered in the decades after his death. What artefacts remained came finally under state care in 1811 during British rule and were first housed in the library, the site of the fledgling national museum collection in Valletta.21 Judging by the sheer quantity of artefacts sourced in Malta, the most remarkable aspect is that the tombs, the source of many objects, appear to have remained largely intact until the 1600s. A significant part of the scientific enquiry turned to the natural world. Although fossil bones appeared in early displays, including Abela’s, they were explained as evidence of the giants written into classical narratives, notably the Cyclops of Homer’s Odyssey and giants in Old Testament tradition.22 By the nineteenth century, such interpretations could no longer be sustained in the light of decades of intense scrutiny of fossil remains which had exposed a rich faunal array of extinct species or ancestral remains of living animal species in fossil-rich deposits of Sicily and Malta. Giants’ teeth, under close examination, were identified as belonging to pigmy elephants. Thomas A. B. Spratt and Andrew Leith Adams, in particular, while on duty in Malta (the former within the Royal Navy, the latter in the Royal Army Medical Corp), actively explored, identified and excavated caves or other fossil-bearing deposits.23 Adams also made some observations regarding the archaeological sites

Malta’s Archaeological Past

Figure  1.3. Photograph reproduced from the Letters Book of the Archaeological and Geological Society of Malta 1860 (society scrapbook). Members of the society (including two women) photographed among the stones of the unexcavated site of Ħagar Qim. (Courtesy of the National Library of Malta.)

in the islands, recorded in his Notes of a Naturalist in the Nile Valley and Malta, published in 1870. HISTORY OF ARCHAEOLOGICAL INVESTIGATION

A strong Maltese identity grew within the political disquiet of the eighteenth and nineteenth centuries AD. The islands benefitted from the development of port infrastructure, which, in turn, became the focus of growing urban hubs, but the population chafed under the constraints of foreign rule. Malta was governed by the knights of the Order of St. John (1530–1798), experienced a brief French interlude under Napoleon (1798–1800) and subsequently became a British possession (officially from 1814 to 1964). Each regime tried, to varying degrees, to mould the island population to comply with foreign sensibilities on socio-political and religious levels. Throughout this time, the value of the archipelago as a strategic and economic mid-point in the Mediterranean steadily grew, and with growth came increased prosperity.24 Wider trends in scholarly pursuits and growing local sentiments towards cultural heritage crystallised in 1865 with establishment of the Archaeological and Geological Society of Malta (Figure  1.3).25 Although the society lost momentum in the course of the next decade, its members, nonetheless, contributed significantly to the identification and documentation of the islands’



The Archaeology of Malta fixed and movable archaeological heritage. Most importantly, the association comprised a cross section of society – Maltese and European – whose names appear in numerous accounts of discoveries made in Malta and Gozo. From its inception, key figures in the movement appear to have both stimulated the government’s recognition of its obligations and acted upon the resulting official concern for the historic and archaeological record. In 1881, the Council of Government established the Permanent Commission for the Inspection of Archaeological Monuments. Antonio Annetto Caruana’s contribution, in particular, should not be overlooked. Although dated now, his lengthy government-commissioned reports, written during the 1880s and 1890s, do give a reasonably clear account of the known archaeological sites and an overview of the ancient artefacts which had found their way into government and private hands.26 No less important during this time was the quest to document inscriptions found on the island. Among them was a pair of monuments with bilingual texts that were instrumental in the decipherment of Phoenician alphabetic writing by Jean-Jacques Barthélemy, facilitated by the accompanying Greek version. One of these monuments was eventually deposited in the Louvre after it was given to the French king, Louis XVI, in 1780 by Emanuel De Rohan, Grand Master of the Order. Caruana was forthright in pointing out the failings of past governing bodies to preserve historical sites, condemning the loss of artefacts of national importance which had been spirited away from the islands, and scathing about the continuing trend for discoveries of the day to remain in private hands.27 He made concerted efforts to track down ancient objects held in various government departments, found during the course of building the islands’ defences and infrastructure, so that they could be brought together under the protection of the National Library.28 While it could be said that through the publication of his reports, Caruana effectively put the British government on notice not to repeat the failings of the past, it is also the case that the sense of protecting ancient remains was matched in Britain itself, resulting in the 1882 Ancient Monuments Protection Act. This movement was carried to the lands of the Empire and was manifest in, for instance, the Archaeological Survey of India and the establishment of an Archaeological Department in Burma.29 Notwithstanding his positive steps towards consolidating and safeguarding the cultural heritage of the island, Caruana was a product of his time and a theological education; history to him was bound to the biblical narrative. Depending on one’s interpretation, Caruana either could not or would not consider a period in Malta earlier than the Phoenicians, who figured in the biblical accounts. But no amount of Christian zeal could sustain the link he made between the prehistoric megalithic structures in Malta and Gozo (ca. 3600–2500 BC) and the Phoenicians of the first millennium BC.30

Malta’s Archaeological Past Scientific rigour gradually transformed the approach to ancient world studies and archaeological fieldwork.The likes of Henry Rhind in 1856 had already started to question the Phoenician involvement in the construction of the lobed buildings which we now know belong to a prehistoric era.31 Around the turn of the twentieth century, Albert Mayr, in particular, brought the logic of archaeological interpretation of his time into his discussions concerning the ancient remains of the archipelago. Malta was the focus of much of his research, and Mayr’s work benefitted directly from his surveys of the ancient sites and examination of private and museum collections during visits to Malta in 1897–1898 and 1907. Nonetheless, he was influenced by Arthur Evans, who following prevalent diffusionist theory, looked to Mycenae for antecedents of the spiral iconography decorating some of the megalithic monuments and pillars (or baetyls) as a focal point of ancient devotion.32 Notions of Aegean influence endured until radiocarbon analysis provided a means of absolute dating that plunged the megalithic lobed structures deep into a prehistoric, Neolithic past. NEW ARCHAEOLOGICAL DIRECTIONS DURING THE TWENTIETH CENTURY

Within the first half of the twentieth century, a time of political and economic instability that resulted in two world wars and a depression, it is probably fair to say that the direction of Malta’s allegiances mattered not just in the wider arena but also at home among its population as it moved resolutely towards independence. There was a groundswell that looked to shake off British rule and rekindle links with Italy, a historic relationship characterised by cultural affinities nurtured among the intellectually elite in Malta.33 In light of these trends, British scholars realised that they were not alone in recognising the great prospects for fieldwork and sustained research programmes on the islands (focussed firmly on the prehistoric remains). They became vocal in their calls to reinvigorate interest in the islands, lest they lose to the machinations of international political rivalry the potential to excavate independently on British-held territory in Europe and their historic advantage in Malta.34 Work undertaken by Luigi Maria Ugolini from 1924 to 1935 comprised a systematic documentation of the prehistoric holdings in the museum in Malta and an architectural survey of the monuments. A member of the Italian Fascist Party from 1924, Ugolini was appointed Inspector of Excavations and Archaeology in Italy (1930), and his research in Malta in 1931 was directly funded by the state at Benito Mussolini’s instigation. Ugolini’s was a significant contribution to the study of Malta’s prehistory,35 but his untimely death in 1936 and the spectre of war ordained that the extensive archive of data he and his colleagues had amassed languished, forgotten, in the Museo Nazionale Preistorico Etnografico ‘Luigi Pigorini’ until it was rediscovered in 2000.36



The Archaeology of Malta Albert Mayr’s publications, followed closely by Erich Becker on Christian and Jewish iconography in the islands’ funerary contexts,37 and Ugolini’s research were catalysts for renewed fieldwork, largely organised through the British School at Rome under the directorship of Thomas Ashby (1906–26). Such interest in the islands must be measured against the significant role Sir Themistocles Zammit played in the first quarter of the twentieth century. The distinguished career of this learned Maltese doctor was multi-facetted. Renowned in the field of medical research, he was both the first director of the National Museum, when it became an institution independent of the library, and an active field archaeologist.38 He was the type of true scholar not only in depth of knowledge and range of accomplishments, but in the generosity with which he placed his knowledge at the disposal of others [His Excellency, Sir Harry Luke, The Officer Administering the Government, 28 February 1936].39

It is clear from those who acknowledged his help in their publications that Zammit was a scholar of note and, most importantly, a facilitator – perhaps the most positive and productive aspect of his illustrious career. It was largely through his assistance that the work of Ugolini and numerous visiting scholars was undertaken. No less remarkable and to his credit was his scholarly recognition of notable women in the field, and after meeting Margaret Murray (1863–1963) in England in 1920, he encouraged her to conduct excavations in Malta, principally at the prehistoric site of Borg in-Nadur. His invitation was well founded. She was then in her late fifties and had years of fieldwork and active research still ahead of her (well after retirement), but her contribution to Malta’s archaeological record developed during five seasons in the island is noteworthy. Her publications on the cultural remains found at Borg in-Nadur and the Bronze Age of Malta were comprehensive, and although they are now dated, in some respects they were unparalleled for many decades.40 Accompanying Murray were her colleagues and friends Edith Guest, who worked on skeletal remains and photography, and Gertrude Caton-Thompson (1888–1985), who had been asked by Zammit to work at Għar Dalam, close to the site of Borg in-Nadur.41 Only in very recent times, some eighty years after Zammit, have women of Maltese nationality begun to break through the glass ceiling of a field which has been dominated by men. Although the intertwined issues of nationalism and politics can be detected in the archaeological pursuits of those who carried out research and fieldwork in Malta, Zammit was not preoccupied by such political interests.42 Preserved in the numerous field diaries he kept (see, e.g., Figures 3.12 and 5.1, nos. 10–12) and throughout his annual museum reports to the government are notices concerning the presence of prominent expedition members, independent scholars or others who assisted him in the field, which clearly indicate that Malta was a hub of archaeological activity.43 Fully aware of the significance

Malta’s Archaeological Past of the islands’ exceptional ancient past, Zammit brought discoveries to a wide audience through a range of scholarly reports and publications. He embraced the principle that the holdings of the museum were of international significance, an asset that needed safeguarding for the future, and he made the collection available to researchers.44 Overall, there is a strong sense that he was participating fully in the international domain. He challenged entrenched views by looking to a Neolithic age for the megalithic buildings, an aspect that Ugolini worked to substantiate.45 Zammit was not alone in bringing Malta’s heritage into focus, and notable among his countrymen who were involved in archaeological investigations at this time were Giuseppe Despott (1878–1936), Napoleon Tagliaferro (1857–1939) and Paul F. Bellanti (1852–1927).46 After Themistocles Zammit died in 1935, acting directors were appointed to the Museums Department who shepherded the collections through World War II. John Ward-Perkins briefly took up the position of the inaugural chair in archaeology at the then Royal University of Malta in 1939, an appointment which has been seen as a direct counter to the Italian cultural influence infiltrating Malta.47 War had a huge impact on the island generally, and the archaeological remains did not escape. Some sites came to light in the course of preparations for conflict, during works on the airfield, military barracks and the digging of shelters; others were presumably lost ‘unnoticed and unrecorded’ in the flurry of quarrying and building activity that took place during the war.48 Dr Joseph George Baldacchino was appointed director of the museum from 1947 to 1955, and after him, Charles, Zammit’s son, took up this post, having formerly served in various curatorial capacities in the archaeological section of the museum between 1932 and 1958.49 At the end of the war, a survey of the damage was made that included the ancient sites of Kordin and Tarxien, as well as damage to Roman-period buildings at Ta’ Kaccatura near Birzebbuga. Reconstructed pottery was reduced to sherds again in the damp storage conditions of the museum basement.50 A post-war programme to survey and bring structure to Malta’s prehistoric record came with funding from the Colonial Welfare and Development Fund. This project was undertaken by a large research team, guided by Stuart Piggott and John Ward-Perkins, and was brought to its conclusion by John D.  Evans (Research Fellow, Pembroke College, Cambridge).51 During this period, David Trump (1958–63) and Francis S. Mallia (1959–71) were Curators of Archaeology at the Museum while its collections were ordered and catalogued and museum displays were put in place. Excavations undertaken at Skorba, Baħrija, Ta’ Ħagrat, Kordin III and Santa Verna (Gozo) in the 1960s by Trump provided vital missing chronological links in the developing prehistoric sequence. Their combined efforts, assisted in no small part by the staff of the National Museum, established the prehistoric cultural scheme for the islands, a sequence that took shape from the 1950s (with variations in nomenclature and cultural adjustments over that time).52 The basic cultural framework came



The Archaeology of Malta with the publication of Evans’ Prehistoric Antiquities of the Maltese Islands in 1971. Essentially, the sequence remains the accepted model. Whatever perceived colonial overlay nurtured this British initiative, it was severely challenged by the arrival of the Italian Missione archeologica a Malta on the eve of Malta’s independence in 1964. The Italian excavation programme spanned 1963 to 1970, albeit initially focussing on the later remains (Phoenician, Punic and post-Punic periods) on the island at the sites of Tas-Silg and San Pawl Milqi; in the second season, excavations also took place at Ras il-Wardija in Gozo.53 Hitherto, interest in the first millennium had been severely overshadowed by the preoccupation with the prehistoric period. The likes of Ward-Perkins actively lobbied against expanding research into this area, opting to maintain the strong focus on the prehistoric record which characterised the initial parameters of the funding.54 Not until the 1990s did independent research projects target the hundreds of Phoenician, Punic and Roman-period tombs that had been steadily documented throughout the twentieth century in particular. And these studies, I believe, were conducted outside of any contemporary political agendas.55 Development of radiocarbon dating put an end to any lingering diffusionist notions, sustained largely by distant cultural comparisons.56 Based on the new, scientifically generated absolute dates, it was found that many cultures in Western Europe which had formerly been linked with Near Eastern civilisations were significantly earlier. It was clear that distinct regional cultures had emerged independently. In an overview of these advances in archaeological research which appeared in the landmark study, Before Civilization: The Radiocarbon Revolution and Prehistoric Europe, Colin Renfrew clearly and sharply reset the research agenda.57 A  new chronology was emerging that effectively stopped in its tracks the accepted mode of interpreting the past. As Renfrew pointed out, what remained were reasonably well-defined, area-specific cultural sequences, acceptable though circumscribed relative chronological links and restricted regional studies that were still valid.58 He outlined the challenge that remained: ‘We are left with an alarming void – a mass of well-dated artifacts, monuments and cultures, yet with no connecting interpretation of how these things came about and how culture change took place.’59 The wave of fundamental change also affected Malta’s long-accepted place in the ancient world. John Evans may have begun his work programme from the position that cultural trends permeated the Mediterranean from the Aegean and came specifically into Malta through Sicily, but by its end, he accepted the new chronological paradigm, reassessing the data accordingly. On the basis that Malta’s prehistoric cultures were not born of distant Aegean regions, Evans’ response was to mull over the implications of insularity and isolation on independent cultural development, which he explained in a paper with the evocative title,‘Islands as Laboratories for the Study of Culture Process’.60 Regarding

Malta’s Archaeological Past the prehistoric finds in Malta, Renfrew tackled the issue of cultural organisation and the implications of spatial distribution of the monuments in regard to territories. By looking to other regions, which developed ‘temple cultures’, he came to similar conclusions to those of Evans – that such independent cultural trajectories were ‘both possible and natural’.61 The door was opened to new theoretical models by which to interpret archaeological evidence for ancient cultures. Malta, with its remarkable prehistoric sites, has continued to provide particularly fertile ground for the application of new ways of interpreting the past – but more of this in the pages that follow. The necessity of accruing new dates to the existing body of data Renfrew saw, in the 1970s, as an essential way forward. Although radiocarbon dating has been available since the 1960s, the technology has not been applied as a matter of course to new finds in Malta, especially in the Bronze and Iron Age contexts (probably due to the high cost of processing rather than the absence of suitable samples). Even within the prehistoric era, much hinged on a handful of absolute dates from a few well-defined levels.62 Noteworthy is recent work carried out in the Xagħra Circle burial ground in Gozo that has generated a suite of nineteen new radiocarbon readings which have suggested some refinements for some phases (Tables 1.1 and 1.2 and Appendix A).63 From the closing decades of the twentieth century, the nature of governance of the islands’ ancient heritage has been thoroughly reviewed, culminating in the 2002 Cultural Heritage Act overseen by the Superintendence of Cultural Heritage organisation, which was established at the same time. The act applies to immovable and movable as well as intangible heritage in regard to the ongoing ­protection, conservation and promotion of a wide spectrum of Malta’s cultural assets. Within the organisation, there is a Committee of Guarantee, which liaises between relevant cultural heritage and the government agencies and Heritage Malta, overseeing operation of the provisions set out in the act and management of heritage assets and museums.64 No less important was establishment of the Malta Environment and Planning Authority (MEPA) in 2001, within which is the Heritage Planning Unit (HPU), which manages data and identifies heritage assets for protection.65 The resulting restructuring has permeated all aspects of the Maltese cultural record, and the digital age has made the workings of these agencies transparent. State of the Heritage Reports appear annually and on the web, replacing the Annual Report on the Working of the Museum Department. Regrettably, the Bulletin of the Museum was short-lived (1929–35); otherwise, detailed publication of past and new finds and fieldwork remains the one area which has not been developed. Malta’s unique record has been internationally recognised, with the prehistoric sites of Ħal Saflieni, Ggantija, Ħagar Qim, Mnajdra, Tarxien, Skorba and Ta’ Ħagrat falling under the UNESCO Convention Concerning the Protection of the World Cultural Heritage (established in 1972).66


The Archaeology of Malta


Table 1.1.  Radiocarbon dates from cultural contexts in Malta (for detailed listing see Appendix A) OxCal v4.2.3 Bronk Ramsey (2013); r:5 IntCal13 atmospheric curve (Reimer et al. 2013) R_Date OxA-3750 R_Date SUERC-4391 R_Date SUERC-4390 R_Date SUERC-4389 R_Date UBA-10378 R_Date UBA-10383 R_Date OxA-3571 R_Date OxA-8197 R_Date OxA-3573 R_Date OxA-3575 R_Date OxA-3569 R_Date OxA-3574 R_Date BM-808 R_Date OxA-3570 R_Date OxA-8165 R_Date OxA-3566 R_Date OxA-3567 R_Date OxA-3568 R_Date OxA-5039 R_Date OxA-5038 R_Date OxA-3572 5500



4000 3500 3000 Calibarated date (Cal BC)




Hand in hand with almost two decades of legislative reform and unprecedented restructuring of museum and cultural heritage practices, significant trends in research have led scholars to revisit key sites that were excavated decades ago. Documentation generated during fieldwork, museum archives and excavated material in museum stores, as well as the sites themselves, have

Malta’s Archaeological Past


Table 1.2.  Radiocarbon dates from environmental contexts in Malta (for detailed listing see Appendix A) OxCal v4.2.3 Bronk Ramsey (2013); r:5 IntCal13 atmospheric curve (Reimer et al. 2013) R_Date Beta-208961 R_Date SacA-11658 R_Date Beta-200517 R_Date BM-710 R_Date BM-711 R_Date Poz-42682 R_Date Poz-42443 R_Date BM-141 R_Date Poz-42442 R_Date BM-143 R_Date BM-712 R_Date BM-101 R_Date SacA-11663 R_Date BM-100 R_Date BM-147 R_Date BM-145 R_Date BM-148 R_Date BM-142 R_Date Poz-42441 R_Date SacA-11665 R_Date Beta-200518 R_Date BM-216 R_Date Beta-200519 R_Date Poz-42444 R_Date SacA-11668 R_Date BM-378 R_Date SacA-11669 R_Date Poz-42439 8000







Calibrated date (calBC/calAD)

1000 1calBC/1calAD 1001

The Archaeology of Malta


been reassessed in the light of current archaeological practices and theoretic interpretations. Considering that many of these early investigations underpin the broad archaeological sequence constructed for the Maltese Archipelago, this focus on achieving higher definition and clarifying past discoveries is extremely worthwhile; Borg in-Nadur,67 Ta’ Ħagrat and Skorba68 have been the subject of such revisions. Similarly, archaeologists continue to revisit sites with a view to refining chronological sequences and acquiring additional data in line with current archaeological practice, for example, the Xagħra Circle (Gozo), Tas-Silg and the Zejtun villa sites. Despite the early archaeological exploration of the islands and their dominant place especially in general studies of the prehistoric period, many gaps in research have begun to be addressed only recently. Lithic material from the islands often was dismissed, possibly as it did not compare favourably with offshore industries.69 Fine sifting was carried out at Borg in-Nadur in the 1920s by Margaret Murray, and some debitage from stone tool production was recovered, but such fine resolution was otherwise under-represented.70 Recent research into lithic material from past and modern excavations represents a marked step away from perfunctory listings and a focus only on the formal tools.71 There is, however, a similar need for research to be carried out on the entire prehistoric bone tool industry, as well as food-production implements (or tools used for other functions) such as pestles, mortars and grinding stones of any period.72 Excavation reports, too, though often slow to eventuate, now include a suite of archaeological and scientific analyses; most notable is the Xagħra Circle (Gozo) report that appeared in 2009 and Tas-Silg (northern and southern sectors).73 The Xagħra Circle presented the first major analysis of human remains from a mass burial ground of the prehistoric period; it is the long-awaited counterbalance to the loss of data from the early excavations at Ħal Saflieni, about which we know comparatively little. There have been no major studies on human remains from Phoenician-Punic tombs in the islands. There is a backlog of unpublished material on Phoenician-Punic tombs and later catacombs excavated after the 1970s, mostly recovered through salvage operations. Field surveys have been undertaken; most remain unpublished or under-published, but finds from the joint Belgo-Maltese Survey Project in the north of Malta are very promising.74 SUMMARY

The archaeological record for the Maltese Archipelago is complex and one of extremes. Research has been lopsided, dominated by investigations into cultures that produced monumental prehistoric architecture. Such sites were easy targets for the budding field of archaeology, and later, they presented ideal case studies for successive theoretical models which sought to explain how people lived and structured their communities. Fragile habitation deposits and

Malta’s Archaeological Past signs of intensive land use over the millennia have left a challenging record. At times tightly packed stratigraphy occurs in key locations (such as Skorba and Tas-Silg), while other sites are marked by meagre evidence that people once lived there. Ancient accumulations of human settlements which are characterised by mounds (tepes, tels, tells or höyüks) in the Near East are not apparent in the Maltese landscape. Modern human land use reflects a dramatic pattern of concentrated urban zones encroaching into age-old rural areas, which are characterised by small agricultural plots defined by dry stone walls. It is challenging to conceptualise a very different prehistoric terrain where settlements were smaller but probably more evenly dispersed across the islands or a scene where Phoenician-Punic urban development formed in central Malta and Gozo, interacting with outlying hamlets that grew around commercial hubs. The early years of archaeological discussion were artefact based (especially focussed on pottery) in Malta, just as they had been in other regions.75 Such an approach has tended to be broad, typologically driven and, in Malta, hampered by a dearth of site-specific chronological discussions. Pursuit of past cultures in Malta throughout the history of its archaeological investigations has been influenced by the desire to compartmentalise major cultural phases defined for the islands, but the path of change is rarely clear-cut. In the modern era, the plethora of scientific analyses has opened up an array of possibilities to delve into the human story, from what filled the pot for the daily meal to the bigger picture of the genetic heritage of the population. Nonetheless, the definition of ancient cultures in the first instance rests on the material remains. While there is a trend towards downplaying the role of typological studies of artefacts in the scheme of building human history, environmental research, chemical, residue and use-wear in artefact analyses, recognition of spatial and chronological contexts of finds, identification of sources of raw material and research into production techniques continue to gather momentum.76 International interest in the islands’ prehistoric past has been reasonably steady over the course of the last century, but finds from the full spectrum of human occupation, spanning Bronze and Iron Ages as well as times of Roman influence, have continued to accumulate, and these need to be brought into the discussion to make full sense of each phase in its turn. What follows is an account of this remarkable record of human history.




Evidence of the Neanderthals in Malta is tenuous at best, limited to three taurodont molars which can be characteristic of their dentition. They were found in the Għar Dalam (‘Cave of Darkness’) complex, which was also rich in deposits of animals long extinct on the island (Figure 2.1).Taurodont dentition, however, can be found among modern human populations.1 Discovered in 1917 and 1936, these human finds are not without problems regarding context and reporting of scientific analyses.2 Decades later and further afield, two teeth from a single human individual, which seemingly had some archaic Neanderthal features, were found in a rock shelter at Fontana Nuova, Sicily, excavated in 1949. At the time, the associated industries in Sicily were assigned to the Aurignacian of the early Upper Palaeolithic, and a re-examination of these finds revived interest in the Maltese human specimens from Għar Dalam.3 It is now considered, however, that Sicily has no signs of Neanderthals.4 Palaeolithic occupation of Italy, generally, is thought to have been interrupted; some signs of Aurignacian point to settlement around 30,000 BP on the mainland followed by a period of apparent abandonment until Gravettian settlers made their mark from 26,000 BP. Not until after 15,000 BP, however, are there indications of the late Pleistocene (final Epigravettian) assemblage that heralds the permanent resettlement of Sicily.5 POSSIBLE PALAEOLITHIC CAVE ART AT GĦAR ĦASAN

The possible evidence for a Neanderthal presence in the cave at Għar Dalam is only fleeting and far from substantiated. Even less certain is cave art at Għar Ħasan, which also has been tenuously linked to the Palaeolithic period (Figure 2.1) and the identification of a lithic industry of this date at any site.6 Based on palaeogeographical research of the central Mediterranean region, it remains a possibility that people trekked as far as Malta during the glacial 20

Figure  2.1. Għar Ħasan:  (1)  possible cave art, animals, hand and human figure (bottom); the images are in red, brown and black hues (after Savona Ventura and Mifsud 2000, p.  29, pl. 10; Anati 1980); (2)  plan of the Għar Ħasan cave site (after Shaw 1950; Mifsud and Mifsud 1997; Savona Ventura and Mifsud 2000, p. 10, fig. 2); (3)  animals painted on the cave wall (after Savona Ventura and Mifsud 2000, back cover); (4)  possible Neanderthal taurodont teeth from Għar Dalam cave  site (after Fedele 1988, fig. 7). 21

The Archaeology of Malta


maximum of ca. 18,000 BP when it was the natural southward extension of lands from the current coastal region of Syracuse. The land passage was inundated some time after 14,000 BP, but certainly by 9000 BP, leaving the Maltese Archipelago uninhabited.7 EARLY NEOLITHIC ISLAND COLONISATION [O]‌n all the major Mediterranean islands, with the telling exception of Mallorca, the arrival date for domestic sheep and goat kept pace with the adjacent mainland. In the space of some four millennia, the Mediterranean sea therefore underwent a critical transformation  – from an impediment enforcing tortuous terrestrial patterning, to a highway for lateral transmission.8

For people to reach the remote islands of the Mediterranean, they had to achieve a reliable level of seaworthy boat-building technology and to have acquired a keen knowledge of the maritime environment: its seasonal conditions, tides and currents. In terms of capacity, those first vessels needed to convey a core supply of seed for reliable founder crops, a basic toolkit and utensils for daily work and domestic animals (sheep, goats, cattle and pigs), as well as the human cargo, enough to make a new settlement viable.9 Geographically, the Maltese islands would appear to have remained virtually isolated prior to ca. 5200 BC. Yet, considering the likely level of human technical know-how, the archipelago probably fell within the maritime capabilities of people millennia before it was actually settled.10 It has been demonstrated that Cyprus, for instance, was colonised ca. 10,500–9000 BP by people after a 60-km water crossing in boats capable of transporting goats, sheep, cattle and pigs.11 Groups practising a highly mobile hunter-gatherer economy, however, who required relatively large territories per band were less likely to adapt to the limitations of small island environments.12 Domestication of plants and animals was a pivotal threshold in human economic development requiring significant social adjustments, but it was a change that offered new opportunities. By harnessing aspects of the management of these two resources, people engineered a buffer against adverse environmental fluctuations and, importantly, achieved a relatively reliable means of food production with predictable outcomes that formed the basis from which to infiltrate and colonise less favourable areas, distant islands included.13 The Maltese Archipelago was ripe for exploitation, and it was well within the colonists’ maritime and economic capabilities. Any number of factors could have taken people to the islands initially and perhaps temporarily at first, until staying permanently seemed a better option than making the journey to their homeland. Animal domestication was a process that played out possibly in a series of independent experiments around 9000 to 8000 BC in lands east of the Mediterranean referred to as the ‘Fertile Crescent’, a territory arching over

The First Settlers and Farmers the upper reaches of Iran and Iraq and south-eastern Turkey. It followed plant domestication in the Levant by around a millennium. Independently, some crop developments have been suggested for European locations:  barley and lentils in Neolithic Greece and legumes in southern France.14 Skeletal remains of early domesticated goats in the Fertile Crescent demonstrated a marked decrease in size relative to their wild counterparts (shown genetically to be Capra aegagrus and Ovis orientalis, respectively15), which was considered to be largely due to culling practices. Sheep display similar traits of domestication not long after in the archaeological record. Hunting sites tend to reflect that larger males of a species were targeted for maximum gain, leaving female animals as breeding stock; conversely, bones from domestic herds are characterised by smaller, female animals past their reproductive prime.16 Selective breeding as well as separation from wild stocks may also have contributed to morphological changes.17 Goats, in particular, are more easily herded overland with immigrant human populations, as well as being tolerant of sea voyages; they adapt well to hot climates and browse a wider range of plant foods, making them an ideal component of colonisation processes.18 Among the other animals drawn into human farming economies were pigs, which were domesticated in south-eastern Turkey around 8500–8000 BC and cattle in the upper Euphrates region during the eighth millennium,19 but both took longer to infiltrate neighbouring regions. Throughout the seventh to sixth millennium, significant island colonising events occurred, coinciding with the growth of economies based on food production in lands bordering the Mediterranean rim.20 An ‘enclave colonisation model’ entailing ‘small groups of farmers in discrete dispersal jumps’21 encapsulates the likely manner in which the Maltese islands were settled. Impetus to colonise the archipelago would appear to have arisen in the southern reaches of Italy and Sicily, from where characteristics of the Stentinello pottery industry were carried to Malta. Stentinello was named after the type site in Sicily and is a variant of Impressed Ware. It was decorated with punctured, linear designs executed with the notched edge of the cockle shell, Cardium edule. Cardial pottery of Neolithic Europe, a related ware type, is an indicator of the cultural dispersal apparent in coastal regions from the Iberian Peninsula, France and Italy to North Africa from 5800 to 5400 BC.22 The new island landscape of the Maltese Archipelago probably presented a degree of similarity to the colonists’ homeland but, initially at least, little human competition for its resources.23 Smoke from large fires, glimpsed from the southernmost reaches of Sicily or by fisherman at sea, might have become a more regular sight after colonisation. It is likely that over the course of time, the first settlers were joined by others and equally possible that return journeys were made regularly, in optimal conditions, to gather resources not found in their new island home or to reconnect with people from an increasingly distant past.


The Archaeology of Malta


Mark Patton draws Malta into the wider discussion concerning Mediterranean island colonisation using criteria based on visibility, destination-area-todistance ratio and biogeographical models. On the scale of area-to-distance ratio (A/DR; 237 km2 surface area/250 km to the nearest mainland), Malta rated 0.9. Similarly, using a variant equation, the island rated 0.1 on target-width-to-distance ratio (T/DR). Both determinations are low, pointing to the relatively high degree of difficulty the archipelago posed as a destination for ­colonisation.24 The incentive to settle in such remote locations may lie in any number of motivations – social, political or economic – but the ability of people to remain permanently rested on available basic natural resources and an economy capable of adapting to a new environment. Nonetheless, Malta eventually became part of the wider phenomenon of Neolithic colonisation of the Mediterranean made feasible by proven advances in farming and maritime know-how.25 Geomorphological studies of the Burmarrad region in the north of Malta suggest that the first settlers encountered coastland environments sculpted into natural inlets and bays resulting from increased sea levels towards the mid-Holocene (ca. 5000 BC). Shallow and sheltered waterways that somewhat increased opportunities to harvest marine resources further inland may have facilitated infiltration of the island’s hinterland. In time, after the permanent appearance of people, the original landscape was altered. Introduced agricultural practices perhaps teamed with a climate trending towards less humid and more arid condition had a negative impact on the fragile environment. This ultimately led to the siltation of low-lying coastal estuaries and a corresponding degradation of hilltops.26 Clear signs of settled human populations appear in the Early Neolithic, a broad term encapsulating the Għar Dalam, Grey Skorba and Red Skorba phases in Malta. Sites from these early phases of human occupation of the islands were established in open areas, but caves, too, were occupied. GĦAR DALAM PHASE (5000–4300 BC)

Caves and Habitation Sites From the first attempts to make sense of the Palaeolithic cave art in Europe, there has been recognition that humans ventured into the deepest and darkest underground cave recesses for purposes which transcended the needs of habitation at its most basic level. Although some caves were exploited simply as shelters and as burial sites, others were gathered into the less tangible, more complex human exploration of a perceived supernatural realm. At times, naturally darkened, resonant and challenging places – unpredictable in the plan of their sculptured passageways and chambers – presented awe-inspiring qualities. They embody the most evocative setting for the phenomenologist’s foray into

The First Settlers and Farmers the ancient human sensory condition. ‘Secret places’, ‘mysterious’, ‘fantastically atmospheric’, ‘strange’, ‘magical’ and ‘fearful’ are responses garnered together by recent authors to strike an empathetic chord with the ancient human experience within cave settings.27 Notions of shamanism were grafted onto an explanation of European Palaeolithic figures painted on the walls that morphed from human to animal form, seeing them as expressions of altered states of consciousness in which, under certain conditions, hallucinations can occur at every sensory level. Such an interpretation has been hinted at for some of the paintings in the Għar Ħasan cave (Figure 2.1).28 Given the right stimulus, humans, then and now, can experience a similar range of heightened sensory events that transcend chronological and cultural boundaries. The key to recognising the universal nature of these sensations lies in how people have attempted to portray such events in their artwork. The caves provided an enduring canvas for such expression. It is an aspect which will be explored for later Neolithic cultural phases, especially in regard to the underground burial complexes and Ħal Saflieni in particular. Notwithstanding the possibilities of human use of caves and rock shelters, those frequented in Malta and Gozo during the Early Neolithic period, on face value, tend to have habitation deposits (no cave art, no figurines if we discount zoomorphic handles or artefacts which could not be explained as mundane). Indeed, the first of the cultural periods in Malta takes its name from significant, though somewhat mixed finds in Għar Dalam. Though this large, naturally formed site is best known for its skeletal remains of long-extinct Pleistocene fauna, it appears to have been used for human habitation from the time of the first settlers in the islands. Excavated in the 1920s, the stratigraphy of Għar Dalam yielded large amounts of Neolithic pottery fragments.29 The Impressed Ware closely resembled Sicilian Stentinello pottery; the strong association between the pottery industries of Malta and its northern neighbour has been detailed by others.30 While the contexts were far from clear, pottery collected in the Għar Il-Mixta caves in Gozo, on the western flank of the Għajn Abdul Plateau, similarly has affinities with Stentinello – ‘purer Stentinello pedigree, than the Għar Dalam ware’31 – bearing in mind that this broad pottery industry is characterised by regional variations within the central Mediterranean sector. Defining links to offshore cultures producing such Cardial or Impressed Wares is further complicated by the considerable chronological span of Neolithic wares, the limited data concerning settlement patterns and the typological development of the pottery industry during that time in the regions where it is found, including Malta.32 Strong links have been suggested between Malta’s Għar Dalam culture and Monte Kronio in Sicily; others look to the central southern coast nearer to Palma di Montechiaro (Piano Vento, Ciotta) as well as west, closer to Stentinello itself and the island of Ognina. Connections are also made to the


The Archaeology of Malta


Impressed Wares of the southern Italian region of Calabria. Conversely, Għar Dalam pottery imports which have been identified in Syracuse at Vulpiglia point to ongoing two-way, inter-island journeys.33

Domestic Houses and Dwellings From the site of Skorba came discrete deposits spanning the Early Neolithic period, as well as later megalithic structures that were constructed over disused or reclaimed parts of the settlement, a site of clear longue durée probably on the scale of a hamlet or small village. Only portions of the earliest levels spanning from the Għar Dalam through to the Mgarr phase could be accessed around the perimeter of the later monuments. Carbon dates from Skorba suggest that people were well established by 5000 BC after a foundation period of uncertain length during which population numbers grew and the island resources were harnessed effectively.34 Of constructed dwellings in the open, the clearest excavated evidence is primarily limited to a single partially preserved structure at Skorba, though reported pottery and lithic scatters, especially in Gozo, may be the fragile relics of habitation zones in other regions. The pottery finds in the Għar Dalam house are late in the phase and represent a transition to Grey Skorba.35 Only part of the bottom row of stones and a pebble and earthen floor remained of what was interpreted as a single-storey oval house. A reused quern at the centre of the floor may have functioned as a pillar base that perhaps supported the roof (Figure 2.2).36 Significantly, the dwelling lay within but approximately 35 m away from a faced, rubble-filled wall,37 between 60 and 80  cm wide, resting on bedrock (any stratigraphic connection between house and wall was severed by the later megalithic structures).This substantial wall is likely to have functioned more as a boundary than as a defensive structure in the Għar Dalam phase. Impressed mud lumps found among the debris, characteristic of wattle and daub, perhaps formed the upper structure or were from houses that left no other trace.38 The surviving wall segment – a length of 11 m exposed through excavation – was virtually straight. Għar Dalam period deposits (two ‘rich’ strata of the period were uncovered in trench AF) had accumulated against it on the northern side, probably equating with the interior. Clive Vella, however, has suggested that the finds in the deposit appear as refuse rather than remnants from a domestic floor level.39 The reasoning behind such a boundary wall may stem from homeland practices, which separated habitation areas from the natural environment. In southern Italy and Sicily, for instance, either constructed walls or substantial ditches encircled houses clustered together into compounds from 1 to 28 ha in size. Underlying the Skorba structure are indications of community co-operation mustered to undertake its construction, perhaps a response to

The First Settlers and Farmers

Figure 2.2. Għar Dalam cultural remains: (1–13) pottery forms (after Trump 1966, figs. 18–20); (14) oval house remains in the western sector of Skorba (after Trump 1966, fig. 10); (15 and 16) animal heads from Għar Dalam (drawn by C. Sagona after Evans 1971, pl. 32:7–9); (17–19) shell, bone and obsidian objects from Skorba (after Trump 1966, figs. 21–22).


The Archaeology of Malta


growing competition for land and the need for clear territorial demarcation. If these factors were the case, and not merely the grafting of established cultural practices onto a new setting, then they imply that the islands may have been rapidly populated, the destination of numbers of settlers wanting to make permanent land claims.40 Għar Dalam finds are known at other locations which were later developed into sites with megalithic structures: in Gozo at Xewkija41 and at Santa Verna and in Malta at Ta’ Ħagrat (Mgarr) and Kordin III, where a small amount of pottery was recorded in excavations in the 1960s.42 The last three sites were excavated by David Trump to clarify the cultural sequence, especially the chronological position of the Mgarr phase in the Late Neolithic. In regard to Għar Dalam pottery, Evans also lists Borg in-Nadur, where one or two pottery fragments of this phase were found.43 Tac-Cawla (in suburban Victoria, Gozo) had remains assigned to the Għar Dalam, and also observed in the area were Zebbug phase and later Neolithic pottery scatters. Pottery from spoil heaps at Tac-Cawla  – ‘hundreds and hundreds of pieces’  – lithic material (some imported flint and obsidian) and bone were collected at modern building sites.44 Reckonings were made based on the stratigraphy and deposits exposed within foundation trenches cut for new houses. Four archaeological levels were observed within nearly a metre of soil overlying bedrock; all were assigned to the Għar Dalam phase. Of the pottery, a range of both fine and coarse wares was identified (some reputedly imported).45 Little more can be said about the cultural development that unfolded at the site at the stage; however, a project started in 2014 under the acronym FRAGSUS (Fragility and Sustainability in the Restricted Island Environments of Malta) does include an excavation component at the site.46 Otherwise, a dearth of systematic field surveys and excavations or inadequate reporting of past archaeological investigations seriously limits what can be said of the known Għar Dalam sites. Veronica Veen and Adrian van der Blom’s observations, especially in Gozo, point to numerous locations that have small numbers of Early Neolithic sherds (Appendix B, List 1). At Ta’Kuljat (Gozo), Early and Late Neolithic sherd scatters also have been reported47; both Ta’Kuljat and Tac-Cawla reputedly have some Stentinello-like fragments. In Tas Srug, Xagħra (Gozo) were possible remains of a Għar Dalam–phase house, and adjoining fields have Zebbug-phase pottery.48 A ‘dense scatter of prehistoric potsherds’ was located on undeveloped land in Birkirkara, on Mrieħel Ridge (Malta). Such material was otherwise unknown in this region.49

Early Neolithic Economy The current evidence for food production in the Maltese Neolithic period is limited. From Skorba, agriculture included grain crops, barley and wheat

The First Settlers and Farmers (mostly Emmer [Triticum dicoccum] and a few grains of likely Club Wheat [T. compactum]), as well as pulses, notably lentils, in Għar Dalam contexts.50 Flint sickle blades with use-wear sheen and querns made of local Coralline limestone with concave upper faces point to food processing.51 In the absence of hard evidence and from a phenomenological perspective, Robin Skeates reminds us that the grain harvest also provided straw which could have been used for livestock (fodder), for construction (mud brick and thatch) and for household items (pottery, basketry and bedding).52 The presence of spindle whorls may indicate that the hair and wool of goats and sheep were being processed into some form of yarn. Only two awls survive in sound contexts to indicate that bone tools formed part of the Għar Dalam toolkit.53 Wood from the Judas, Hawthorn and Ash trees was gathered for domestic fires during the Għar Dalam phase, the carbonised remnants of which were found west of the megalithic structures, inside the settlement boundary wall (in area AF). Marine mollusc shells which were common among the food refuse at Skorba were probably harvested from the nearest coastal regions under 3.2 km away at St. Paul’s Bay to the north and Gnejna Bay on the west coast.54 Bones of domesticated sheep and goats (59 per cent of bone remains), cattle (7 per cent), pigs (34 per cent) and a dog were recovered from levels at the site. The fact that living, perhaps immature animals were initially brought to the island to be raised implies the use of sizeable and stable watercraft into which the animals were trussed. A recent study by John Borg has nudged along our understanding of animal remains from Skorba, and more is heralded to come.55 Most interesting is his identification of bones from ‘exceptionally large’ cattle, probably of wild stock, a pattern he suggests is repeated at other Neolithic sites in the archipelago. At face value, this may indicate that a wild herd or herds roamed prehistoric Malta. An alternative, although less convincing theory is that hunting expeditions were conducted offshore and returned with cattle meat still on the bone. It has been shown that herds in isolation need not be vast to be sustainable. A herd of forty-nine feral cattle in Chillingham Park, Northumberland, in northern England was found to be ‘viable and fertile’ despite total isolation and inbreeding for over 300 years.56

Għar Dalam Material Culture Both coarse and fine handmade wares have been identified within the Għar Dalam pottery industry (Figure 2.2).Whereas fabric colour hues can vary from dark to pale grey or, less often, brownish shades, the paste tends to be refined, but some gritty particles can be present. Pottery in the finer category tends to have pale surfaces, from off-white through greys to light yellows. The vessel surfaces are prone to defacing through deterioration. Known shapes are limited to baggy jars with narrow, near-vertical or slightly inward-sloping


The Archaeology of Malta


necks (Figure  2.2, nos. 5–8) and bowls of hemispherical (Figure  2.2, no.  3), trunco-conical (Figure 2.2, no. 4) and open, squat forms (Figure 2.2, nos. 1 and 2). Most bases were probably rounded (Figure  2.2, nos. 1–3, 6, and 8), though flat and pedestal shapes (Figure  2.2, no.  9)  have been documented. Handles vary somewhat at Skorba: the baggy jars could be fitted with small, pierced lugs on the neck or shoulder (Figure 2.2, nos. 5 and 7) and vertical, tunnel-like handles on the widest part of the body (Figure 2.2, no. 8); wide straps (Figure  2.2, nos. 10 and 11), ledges and small knobs also were found (Figure 2.2, no. 4). Zoomorphic handle fragments have been recorded at Għar Dalam itself – possibly a ram and a cow’s or bull’s head (Figure 2.2, nos. 15 and 16), and from Skorba, a cow’s head.57 Caroline Malone illustrates numerous examples of Cardial/Impressed Wares from Sicilian sites which are akin to Maltese Għar Dalam pottery; the range, in essence, conforms to a similar restricted-shape repertoire, but the Sicilian examples vary more widely in the decorative expression of linear, zigzag, dashed and zones of tightly packed hatched lines that may mimic fine basketry.58 Designs may have been enhanced by in-fill using gypsum mixed to a white paste which in most cases does not survive in Malta. Two categories of coarse ware were determined. Vessels could be made of a gritty, dark fabric, which could be streaky and imperfectly burnished on the surface or left plain and drab. Simple grooves, occasionally forming patterns (sometimes incorporating applied knobs), and fingernail impressions in lines decorate a few coarse-ware pots.Vessels built of this ware tended to be thicker in the wall. A second variety was described as a soft and friable ware that was rarely decorated. It was modelled into bowls, with either straight or curved walls, and heavy pots with necks sloping inward. Thick, pedestalled bases were common, but flat and round bases were not well represented. Handles were equally heavy and thick, in keeping with the vessels’ proportions, and they took various forms: vertical and horizontal strap handles, some pushed into the wall, forming a tunnel effect (Figure 2.2, no. 13), elongated, pierced lugs and solid knobs.59

Lithic Industry Cryptic, and normally rendered more obscure rather than less by archaeological discussion, stone tools are still worth bothering with. They were an essential component of ordinary life, providing the basic cutting edges for most tasks, and they provide concrete witness of an extensive and intriguing trade system.60

Aside from locally occurring chert which can be found eroding from limestone in Malta,61 all flint and obsidian artefacts were imported, as were greenstone for ornamental or functional axe heads and basaltic stones used for grinding. Pumice could have made its own way, floating on the tide to Maltese

The First Settlers and Farmers shores after eruptions of Mt. Etna in Sicily . Prehistoric Malta was certainly part of an inter-island network of trade in flint and obsidian.The starting point in the châine opératoire of lithic use must have been the impetus to acquire what could not be obtained in the archipelago; economic advantages but also socio-political aspects might have stimulated the decision to acquire exotic lithic material.62 The goal of procuring flint and obsidian, whether by offshore suppliers close to the source or by inhabitants in Malta contemplating a sea crossing, required planning and preparation; there is no realistic reason to discount either option, although there is a tendency to downplay the maritime abilities of cultures in Malta in the literature.63 What commodities were assembled by the Maltese islanders for exchange is not known, but food and non-durable manufactures such as basketry are likely. Chemical analyses have been carried out since the 1960s on obsidian from archaeological contexts and from known obsidian sources, each area with its own chemical fingerprint.64 Four main locations on Mediterranean islands were identified as having been exploited extensively in antiquity from ca. 6000 BC, with their obsidian distributed widely across the lands of the Mediterranean. Based on findings from GIS models related to obsidian sources, Robert Tykot argued that maritime distribution figured prominently in the process.65 Monte Arci (Sardinia) obsidian passed into northern Italy and Europe. Obsidian from Palmarola was carried to regions of Italy and Eastern Europe. From Lipari, distribution of the material overlapped to a degree with the Sardinian sources, but it circulated more widely in the southern and central Mediterranean regions. Pantelleria and Lipari were the principal sources of obsidian for Malta, as well as for North Africa and southern Italy.66 Obsidian from Skorba is thought to have come predominantly from Lipari during the Għar Dalam to Tarxien phases. Flint, however, seems to have been derived from the source at Monti Iblei in Sicily rather than the Perfugas outcrops of Sardinia; however, given the range of flint types found in archaeological sites in Malta, other, smaller quarries on the Italian mainland may have been exploited.67 In his study of the lithic industry in Malta, Clive Vella has methodically applied consistent classification based on a morphological typology (flake shatter, proximal flakes, bulky shatter, unimarginal tools, bimarginal tools and cores) with appropriate emphasis on debitage and waste flakes.68 The toolkit defined for the Borg in-Nadur site, for instance, points to scraping, perforation and cutting, and within these three functions, a somewhat traditional range of tools is documented (see Chapter 3).69 Nonetheless, with only patchy information concerning the excavation technique in some early fieldwork, selectivity in artefact collection and a dearth of data regarding specific contexts, conclusions, though indicative of trends, tend to encompass the broader Neolithic period, all limitations recognised by Vella himself.70 Even though the characteristics of the local chert mean that it does not lend itself to a level of refined knapping comparable to imported stones, it was an



The Archaeology of Malta important resource – a stone capable of producing a useful, heavy-duty cutting edge – as its persistent exploitation throughout the prehistoric period readily demonstrates. In fact, it may have been valued highly enough to be traded within Malta to areas apparently devoid of the stone.71 Other options for procurement of exotic materials were open to communities in Malta. Presumably export commodities (details of which are as yet unknown) were prepared in Malta, either for traders who arrived on its shores or to carry across the sea, closer to the sources, where they were to be exchanged for obsidian and flint. Careful preparation of watercraft and gathering of provisions, matched by well-timed journeys – mindful of seasonal maritime conditions – were likely part of the latter model. Underlying the presence of exotic materials in Malta is the complex ‘down the line’72 path of exchange systems from the source, via the hands of miners of quarry sites or collectors in the case of river-worn cobbles, through networks which brought raw materials to local villages, where they may have been partly processed for export.73 Given the distances involved in offshore trading, such lithic materials may have been partially prepared, trimmed of much of the cortex skin and fashioned into cores or tools ready for transport; two cores were certainly found in Grey Skorba contexts. These materials passed eventually to mariners for the sea journey to the Maltese Archipelago, among other Mediterranean locations. Once at its destination, such material was distributed and worked into tools. The choice of material was possibly task specific, with obsidian used for finer cutting tasks and flint for more robust activities.74 Vella argued that a structured social system (i.e., non-uniform or non-egalitarian) might have governed the distribution of exotic lithic material within the community.75 From their archaeological context in Malta, it may not be adequate always to class lithic artefacts as discards; to do so may mask a value placed on them as a commodity and obscure the final symbolic role that some lithics played, especially in the working area where the obsidian cores were located or in deposits within the Red Skorba buildings (discussed in due course). Grey obsidian from Lipari dominates the imported stone material in Malta and Gozo. In addition, people used greenish brown obsidian from Pantelleria, flint from Sicily and local chert for their lithic tools during the Għar Dalam Neolithic phase. Formal tools found in Skorba included two fine blades (Figure 2.2), an awl point and a scraper. Recent in-depth studies have been made of the lithic material from a number of sites excavated in the past, Skorba included.76 From the deposits connected to the boundary wall and drawing on David Trump’s original field notes, Vella indicated that tools were mainly of obsidian, comprising small debitage fragments which were discarded after ‘extensive use’; the lithics in the Għar Dalam house on the east of the site were also considered waste discards.77


Clive Vella noted unexpected differences between the lithic industries of Ta’ Ħagrat and Skorba, even though the sites are close, only 875 m apart: ‘Despite the obvious abundance of chert in the immediate area, the material that was introduced into Ta’ Ħagrat was of a better quality than at Skorba.’78 Very little has been written about the nature of the chert sources in Malta:  whether the material was mined or simply gathered as it eroded from the parent rock and just how abundant the resource was at the known outcrops.79 Selection of natural sources and manner of distribution evident in anthropological contexts is sometimes culturally value laden. Such might be the case in the ancient world, where, for reasons that may not be readily apparent now, one source is overlooked in favour of another. Restricted pathways through other people’s territories, negotiated or contested access to raw materials or exclusion zones because of land ownership or sacred overlays may have played a part; these aspects may have been the underlying variables between the Maltese sites that Vella observed. Perils which may not be apparent in the modern era, such as unhealthy or hazardous locations – for example, malaria-prone regions (which once afflicted Malta) – may have been avoided at certain times of the year. Specific qualities of a particular material at a given source could have been singled out above other options; resources can be revered in their own right and imbued with ‘symbolic properties’.80 Some locations also might have provided more than one commodity, as well as food and water supplies necessary while the resource was harvested, and hence, an expedition to such a region would have been more rewarding and more likely to succeed. Access to sites also can be tied into seasonality, a consideration for the obsidian sites of the Mediterranean islands during winter months, when the sea was less predictable. Availability of river-worn cobbles of obsidian or flint in offshore locations, for instance, might have been restricted by flooding at certain times of the year. It is possible that exotic artefacts in Malta were considered prestige objects, but they may have also embodied greater symbolism. That such hard-won lithics were left in houses at Skorba may point to a deeper cultural meaning and represent ritual placements; therefore, the record as it stands might be skewed due to selective deposition.

House Closure and Ritual Deposition We arrive at a point where some aspects concerning the end of the Għar Dalam house at Skorba are worth reconsidering. In particular, the inclusion of human bone in the fill – jaw fragments of two children (aged 7 and 4½) and


The Archaeology of Malta


adult skull fragments found nearby in a ‘refuse deposit’ linked to the house – may have a deeper symbolic meaning related to the deliberate closure of the house and its immediate environs.81 Certainly there is scope for reassessing in more detail the contents of all the variously dated houses at Skorba. Aside from one used limestone quern, it is unclear whether two bone awls and a few fine lithics were within the house deposit.82 A  single pierced seashell from this period at Skorba offers a glimpse of personal adornment.83 Formal decommissioning of a structure is a possibility encountered on a number of occasions in the archaeological record of prehistoric Malta (a theme addressed in Chapter 4). This practice may have initially travelled with the settlers from their offshore homeland. John Robb has drawn attention to use of the boundary ditches around Neolithic villages as ‘repositories of bone and bodies’ at locations in Sicily and southern Italy in order to link the individual and the community to their settlement and heighten the sense of ancestral connection to place.84 A range of items found within the deep cuttings might belong to ritual deposits rather than discards. These deposits, for the people who lived within the compound that the ditches defined, may have enhanced the role of the features. Lithics, querns, ceramics of various quality, anthropomorphic and zoomorphic fragmentary figurines, sea and land shells, animal bone and, notably, six pieces of human bone were found in the Stentinello cutting, possibly representing not an accumulation but a purposeful, discrete deposit. At Megara Hyblaea, a similar range of objects, as well as stone and obsidian cores, ochre, amulets, bone implements and over twenty fragments of human bone, was recovered from the ditch. The pattern is repeated at other sites, some with greater or lesser amounts of human bone.85 Clearly, the notion of defining the settlement is shared by Neolithic societies in Malta and by their northern neighbours. The sense of maintaining a link between the living and their ancestors, as well as of forging a lasting bond with their home and community, may equally have been expressed through ritual placement of token human remains and artefacts within houses brought to an end in Malta and within the communal setting of the ditch in southern Italy and Sicily. GREY SKORBA PHASE (4500–4400 BC)

Of the Grey Skorba period, only a line of irregular large stones, the remnants of domestic architecture, survived at Skorba. On either side were fugitive traces of related deposits (trench OD).86 Though the foundations consisted of stone, the upper walls appear to have been made of mud brick, of which little trace survived. To the north, vestiges of this period were recovered from pits identified simply as ‘refuse deposits’. No human skeletal remains of the period have been found.87 Grey Skorba communities are likely to have continued the established animal husbandry and agricultural

The First Settlers and Farmers methods of food production. The presence of two worn sickle blades suggests the harvesting of crops.88 ASPECTS OF GREY SKORBA MATERIAL CULTURE

Ceramic Production Of the Grey Skorba pottery industry, only a small number of basic shapes and diverse handle types have been mentioned in the literature (Figure 2.3, nos. 1–16). It is clear from David Trump’s report that signs of pottery ‘survivals’ – development and transition, Għar Dalam to Grey Skorba and Grey Skorba to Red Skorba – were present. Even within the few illustrated categories, he indicated that there were additional variations on the basic forms.89 As was recently undertaken by Clive Vella for the lithic industry (see next subsection), so the pottery finds and their contexts would benefit from a close reappraisal and more detailed publication. The Grey Skorba phase was contemporary with Serra d’Alto culture in Sicily.90 Nonetheless, Grey Skorba pottery is considered to be a local development from the previous Għar Dalam impressed ware. Vessels are built of dark grey clay, often peppered with crushed white, angular particles (Figure 2.3, nos. 21 and 22) which are quartzlike in appearance, perhaps comprising crushed chert recycled from the discards of the local lithic industry. Decoration, beyond a few examples of rudimentary incised, curved lines, is rare (Figure 2.3, nos. 3 and 11) – a marked departure from the previous phase – but the few designs herald the curvilinear motifs encountered in the following Red Skorba phase. Bowls can have a deep groove under the rim with white paste in-fill, a technique used to great effect on the striking red-slipped ware in the next phase (see Figure  2.5, no.  23). Well-burnished, unslipped surfaces tend to be various shades of grey in the finer category. Most frequent of the vessel shapes are deep conical bowls with walls that are straight to the rim (Figure  2.3, nos. 2 and 15). Larger examples appear to have a high or pedestalled foot, and some bowls have pierced horizontal-ledge handles. Smaller bowls have indications that resting surfaces were rounded. With the addition of high and wide thumb-plate handle attachments, the smaller bowls could have served as dippers, but the bifurcated tops (M-shaped) could have acted as rests for the handles of wooden spoons or other dining implements. Knobs and additional loop handles (Figure 2.3, nos. 12 and 15) were sometimes attached to the plates, while others could be fenestrated (Figure 2.3, no. 16). Further bowl forms have curved profiles. Deep bowls can have slightly off-set rims that angle out with rounded bases; other examples are squat, wide and shouldered bowls with flaring, simple rims and flattened resting surfaces (Figure 2.3, no. 3). Only one of several types of tall-necked, closed jars was illustrated (Figure  2.3, no.  1), though variations especially in rim form are



The Archaeology of Malta

Figure 2.3.  Grey Skorba cultural remains: (1–16) pottery forms; (17–20) objects of bone, tooth and shell from Skorba (after Trump 1966, figs 21, 23, 25, 26); (21 and 22)  detail of pottery fabric (reproduced with the permission of the Trustees of the British Museum, inv. nos. BM P1971–2–2–11T and BM P1971–1–1–63; photographs by A. and C. Sagona); (23) Grey Skorba house remains (after Trump 1966, fig. 8); (24 and 25) green-stone axe-shaped amulets; (26) ogival stone; (27–31) objects of bone and flint (after Trump 1966, figs. 8, 22, 26, 40).

The First Settlers and Farmers mentioned.91 Horizontal lugs that can be tubular and hollow, others that are false tubes simply impressed at the ends mimicking openings, or lugs that are solid throughout can be applied to the bowls. Ledge handles can be vertically pierced through (Figure 2.3, no. 6) or placed horizontally on the walls of vessels (Figure 2.3, nos. 7–10). Vertical, semi-recessed, tunnel-like handles have been documented among a coarser Grey Skorba Ware that is pinkish in hue, tempered with white gritty inclusions and usually left drab on the surface. Aside from recognising that they were attached to larger vessels, diagnostics were not identified among the Skorba finds.92 This recessed-handle form first appeared in the Għar Dalam repertoire (see Figure 2.2, no. 13). Another coarse-ware category was matt, unslipped and mottled in surface hues from reddish yellow to buff and grey to black. It was used in large jars, ovoid in shape, with thickened rims and flat resting surfaces (not illustrated in the publications). Applied ridges of clay appear to be decorative; in one example, the ridge is wavy and pendant from the rim of a hole-mouth jar (Figure 2.3, no. 11).93

Lithic and Other Implements Three bone awls (Figure 2.3, nos. 27, 29 and 30), a hollow bone tube (Figure 2.3, no. 17) and a grooved bone spatula or spacer (Figure 2.3, no. 28), possibly used in textile or netting, are all that survive to represent the bone-tool industry. Reused pottery roughly shaped into a disc and perforated for use as a spindle whorl or weight also points to an ongoing textile industry. Three broken flint blades with signs of use wear functioned as sickle blades that were presumably hafted (Figure 2.3, no. 31). Little mention has been made of working areas in Malta where discard waste fragments of obsidian or flint have been observed; it is possible that they may have existed but have simply been overshadowed by formal tools in the archaeological literature of the past. Pits in the north-east of Skorba, presumably in the open air, contained large amounts of lithic material, but the account of these finds was brief in the 1966 report.94 Clive Vella’s re-examination of the context of stone tools has been worthwhile for the added clarity it provides concerning the Grey Skorba period. It is possible that the pit area, however, represented working floors where tools were knapped and hence the concentration of large quantities of ‘lithic debris’, as identified by Vella – some 700–800 lithics predominantly characterised by ‘bulky shatter pieces’, unretouched tools, a lack of ‘over-exhausted lithics’ and two obsidian cores – although he suggests that this was ‘a dumping area of knapped waste’.95 The presence of the two obsidian cores in the deposit (Skorba locus VE4, approximately 47 m north-east of the only remnant of wall of the Grey Skorba–period house in area OD) was considered ‘surprising’, but their presence might reflect a lithic working area or at least a designated place where lithic material was, in a sense, stored. Aside from obsidian and flint, volcanic stones also were imported.96


The Archaeology of Malta


Significantly, from trench VE came a concentration of twenty-four stone objects (they are less frequent in the subsequent phase) averaging around 65 g and worked to an ogival shape which are generally described as ‘sling stones’. If they served as projectiles, the collection might imply conflict or perhaps a hunting strategy. The cluster does suggest that certain designated spaces were used to store or stockpile artefacts (Figure 2.3, no. 26).97 While it is beyond the scope of this volume, a comprehensive study of the entire pit deposits in the north-eastern sector of the site may reveal more complex cultural and perhaps ritual activities. Such depositional processes have been discussed by John Chapman for pits in the Neolithic and Chalcolithic Balkan settlements.98 Ritual activities rather than mere discards also were likely at Serra del Palco in south-central Sicily, where small red-ochre-filled damaged pots were found resting on ochre deposits within three pits; carbonised seeds were found at the sides of the cavities.99

Personal Adornment Body decoration included a pierced cockle shell, a cowrie shell and a cow incisor (Figure  2.3, nos. 18–20).100 Two fragments of ground-stone axe pendants made from an exotic dark greenstone and another piece of a paler green are among the earliest examples found in Malta (Figure 2.3, nos. 24 and 25). This exotic material outcrops in north-east Sicily.The fragments’ context came from within the remains of a house immediately west of the megalithic structures (in trenches OB10 and PC7, respectively).101 RED SKORBA PHASE (4400–4100 BC)

Architectural Remains: A Question of Function Two simple ovoid structures on the east side of the megalithic buildings at Skorba – designated the ‘North and South Rooms’ – were excavated, either side of which were stone paved courtyards (Figure 2.4, no. 1). Their shared wall seems to indicate that the two structures were built at the same time. All that remains are solid foundation walls in stone which may have supported daub or mud-brick upper structures.102 They appear to have stood within the protection of an encircling, faced and rubble-filled boundary wall  – a similar trait shared with the Għar Dalam–phase hamlet at the site. The larger North Room, especially, has been identified as a shrine, based on its architectural attributes and its contents, which date to the Red Skorba phase. Thus declared, the notion that it was a non-secular building has become entrenched in the literature,103 and the figurines it contained have been identified as representative of a fertility goddess ‘in whom this agricultural community strongly believed in their expectations of good harvests’.104 The

Figure 2.4.  Red Skorba cultural remains: (1) two rooms of Red Skorba date north-east of the megalithic buildings (after Trump 1966, fig. 9); (2 and 3) spindle whorls; (4–7) lithic tools made from chert; (8 and 9) obsidian tools; (10–13) pendants made of tooth, stone and shell; (14) ground-bone object; (15) a fragment possibly from a small model bed; (16–19) female figurine fragments (after Trump 1966, figs. 8, 9, 23, 26, 30, 31). 39


The Archaeology of Malta concept of a mother goddess associated with fertility, however, has been questioned in recent years.105 No human skeletal remains were found in Red Skorba contexts, and on the current evidence, we can discount a funerary function for the surviving architecture. The South Room had no signs of a doorway, which suggests that either the original floor was well above the bedrock surface and the door was higher than the level of the surviving foundation stones or possibly that entrance was gained to the South Room through the roof. The thickness of the shared wall is the second factor used to support the theory of a sacred function for these structures, as it was thought to be far more substantial (1.65–3 m wide) than is necessary for domestic architecture. The two buildings were thought to be redolent of later megalithic buildings.106 We have little evidence to judge whether all domestic architecture in Malta at the time was insubstantial or whether, conversely, sacred architecture had an element of monumentality prior to the Late Neolithic period. Greater wall thickness may have been necessary to construct an access stair or passage to the South Room that is now lost. Stone-faced and rubble-filled walls could be seen as a forerunner of the later monumental buildings, and indeed,Trump argues for continuity in aspects of architecture from the Red Skorba phase through to the Zebbug.107 A mitigating aspect against a domestic function concerns the lack of hearths in the rooms, but this also may indicate that such domestic activities took place outside in the courtyard areas or perhaps in a designated room which has not survived.108 The principal factor thought to weigh against the North Room serving a domestic function concerns its floor. Both rooms were excavated to bedrock, and the larger chamber was found to have a quite rough and deeply undulating floor.109 Though it would have been difficult to live on such a surface, it was equally less than ideal for any human-related activity, including religious activity. A considerable amount of pottery was packed into the depressions in the rock surface, including some substantial and joining fragments. Fill over the bedrock included lumps of mud brick and blue clay, which Trump thought were decayed remnants of the upper walls.110 Indeed, the rough bedrock was an unlikely living surface for the interior of the structures given that so much effort went into making level stone slab and cobbled exterior courtyards. It is possible that the pottery packed into the undulations, above which was a layer of blue clay – of which the nearest source lies over 1.5 km north-east and south of Skorba – which was devoid of ash and charcoal, was used to level the surface for a perishable floor that later eroded away in antiquity. This might explain why little pottery was found in the higher elements of the fill and towards the western end of the North Room111 and why few cultural remains were found near its entrance. Careful preparation of a domestic floor was documented in the excavations of a later house at Għajnsielem (or Mgarr) Road in Gozo. Longitudinal segments of oval structures preserved under dry-stone field walls

The First Settlers and Farmers were all that remained of this Late Neolithic (Ggantija, Saflieni and Tarxien) dwelling, about which the excavators noted, ‘the structure seems remarkably sophisticated for a domestic site’, with foundations, under-floor insulation and prepared torba (Maltese for a crushed-limestone plaster) living surfaces.112 Other interpretations are possible. If the concentration of artefacts was not accumulated through use of the buildings, considering that the floor was not functional and quantities of restorable pottery might indicate that the vessels were ritually broken, it could be suggested that the two areas were prepared with cultural material deliberately placed as a foundation deposit ‘immediately above the rock in both rooms’.113 Such a deposition was hinted at by Margaret Murray for a later megalithic structure at Borg in-Nadur. There a concentration of lithic material was found under the pavement of a structure referred to as a ‘dolmen’ (in reality, probably a niche within the complex), which Clive Vella viewed as a ‘caching of lithics … hidden or ritually deposited within the temples below the used floors’.114 Another instance was a ‘curious group of small votive terra-cottas’ found under an early floor in the Mnajdra complex.115 Alternatively, the concentration of artefacts may relate to a closing act on the eve of abandonment of the buildings.116 In either scenario, the artefact-rich deposits, in turn, were covered by blue clay. The finds, some common to both Skorba rooms, could be classed variously as habitation-related – a great quantity of pottery (some restorable), two globigerina limestone bowl fragments and abundant domestic animal bone (in comparable amounts to pottery); tools – quantities of chert and imported stone lithics; four pottery sherds modified into spindle whorls and one modelled, round and decorated clay whorl, all were damaged; possible objects of recreation  – twenty worked cow foot bones (plus others not worked) which might be gaming pieces; and ritual objects – five stylised female figurines (all damaged, four of clay, one of stone from the North Room and possibly part of a sixth very damaged; six skulls from goats with horns attached but the facial bones broken away).While exact context has not been stated, two of the horns were placed side by side on the North Room’s floor.117 Such preparation of a building for abandonment has been documented in Malta at the Tarxien megalithic complex. Themistocles Zammit, who excavated Tarxien, commented on the clean sand that filled the rooms. In the light of ritual practice in other sectors of the ancient world, the sandy deposit at Tarxien can be seen as part of a deliberate ritual act to close the building (discussed in Chapter 4). Ultimately, context is pivotal to understanding the presence of the figurines. Since their discovery, they have been taken as indicators of a sacred precinct proposed for the North Room, but such artefacts can serve in cult, magic, instruction or play118; indeed, any of the artefacts could straddle different functions or rituals which could have taken place. From within the Ggantija strata at Skorba came a few more broken Red Skorba figurines, suggesting that they were not limited to the North Room, but unfortunately, the exact context of


The Archaeology of Malta


these artefacts within the houses of this later phase – which are not identified as shrines – is not stated.119 Aside from the figurines, the rough floor and lack of internal features in the North Room do not support a sacred function. Furthermore, its structure, plan and scale fit well within the known Neolithic domestic house range.120 The stylised human figurines made of Red Skorba Ware and one carved from globigerina limestone originally comprised torsos fashioned without arms and stylised heads that had been broken off. Legs were not depicted beyond the hip lines and upper thighs (Figure 2.4, nos. 17 and 19). Two heads have been found that were formed from rods of clay flattened to oblique faces with impressed dots for eyes and mouth on one face and a small knob and punctured hole for the mouth on the other (Figure 2.4, no. 16). The figurine bodies bulge in an angular fashion at the back to depict the conical-shaped buttocks. Breasts appear as applied elongated ridges, and female genitals are represented by pubic triangles and incised lines (Figure 2.4, nos. 17 and 19). Further incised parallel lines that drape around the back and down between or on either side of the breasts might represent body decoration, jewellery or a simple form of garment.121 A rod of clay flattened to a disc shape and another poorly preserved conical fragment also may have come from figurines. A small artefact with legs is interpreted as a model bed (Figure 2.4, no. 15), drawing on examples from the later contexts at Ħal Saflieni, the most famous of which supports a reclining obese figure (see Chapter 3). As to the twenty objects fashioned from cow foot bones, as well as an unspecified number of unmodified bones found in the northern building, the proximal joint of the worked pieces had been ground off to varying degrees so that they could stand, and only one had an incised cross on the flat resting surface. Because they were found in the same room as the damaged female figurines, the excavator identified them as phallic symbols, drawing largely on evidence from the later Neolithic contexts, where female and phallic figurines were found within the megalithic monuments. Such objects could have served any number of the functions, including as gaming pieces (that one has a cross on the bottom might support this possibility), counting tokens or memory aids representing a tangible way to ‘document’ past events. Images of the few that have been published in the literature are not phallic in appearance (Figure 2.4, no. 14).122

Ceramic Production in the Red Skorba Phase Elements of the Sicilian Late Neolithic Diana ceramic industry, notably red-slipped and highly burnished surfaces with distinctive tubular handles, appeared in the repertoire of Red Skorba Malta.123 The chronological span for Diana cultural remains is based on radiocarbon dates from Grotto Cavallo (cal. 5238–4720 BC) and southern Italy that point to its beginning as early as

The First Settlers and Farmers the fifth millennium and extending to the mid-fourth millennium (Lipari, cal. 4000–3544 and 3775–3638 BC). The absolute date generated from Skorba (cal. 4221–3797 BC) does not contradict the Sicilian evidence.124 Although it has not been explored in the literature to any great extent, perhaps the repeated cross-fertilisation of ideas and technology between Malta and Sicily has much to do with maritime links more frequent than is usually credited to the ancients. As potters are likely to be women, much of the inter-island transference of cultural traits might relate to marriage and relocation negotiated during such ventures. Both the white, gritty fabric from which vessels are built and the limited shape range in the Red Skorba period appear in broad terms as developments on the previous phase (Figure 2.5). Interiors can be dark and smoothed. New to the industry is a thick red slip, often burnished to a high polish, on the exteriors of pots with simple but bold designs deeply incised into their walls and enhanced by white paste in-fill (Figure 2.5, no. 23). Designs tend to range from eyelike paired loops or circles (Figure 2.5, no. 18) to simple crosses (Figure 2.5, nos. 4 and 19), herringbone (Figure 2.5, no. 16) or meandering curved lines (Figure 2.5, nos. 1 and 3) on open conical bowls with straight walls and on sharply carinated shouldered bowls; designs are focused on the off-set in the line of the wall.125 On the latter bowl form, the upper wall of the vessel can lean out to varying degrees or remain near vertical to the lip; sizes range from a miniature to some 35 cm in diameter. Resting surfaces tend to be flattened, but rounded bases are recorded (Figure 2.5, nos. 2, 9 and 13), and somewhat impractical, overly high and narrow pedestalled forms (Figure 2.5, no. 1) have been documented which may have doubled as handles if the bowl was used as a lid. A  few deep forms tend to have more curved profiles (Figure  2.5, no. 22), and others, approaching biconical in profile (Figure 2.5, no. 10), recall bowls of the earlier phase (compare Figure 2.3, nos. 9 and 10); one example is in miniature (Figure 2.5, no. 4). These bowls can have quite distinct bold attachments:  flattened curvilinear handles cupped around small perforations in the wall (‘horse-shoe’ handles; Figure 2.5, no. 6); horizontal, narrow ledges (Figure 2.5, no. 22); and elongated, vertical lugs which are pierced from top to bottom and open out to spurlike ends (‘trumpet lugs’; Figure  2.5, nos. 8 and 10). Tunnel handles are again represented at Skorba (Figure  2.5, no.  7). Bulbous-bodied, closed and small jars have narrow off-set necks that can vary from conical through flaring to cylindrical; these, too, can carry a similar range of handles on the shoulders (Figure 2.5, no. 9), but all-over, slightly oblique lines decorating one jar body fragment were unusual (Figure 2.5, no. 5). High thumb-plates on rounded bowls continue, still formed into peaks at the top (M-shaped), but more exaggerated than in the Grey Skorba repertoire; however, the additional knobs and loop handles at the back appear to have been dropped (Figure 2.5, nos. 11, 13 and 14). New is a triangular thumb-plate that terminates in a small projection or hook angled inwards (Figure 2.5, no. 12).



The Archaeology of Malta

Figure  2.5.  Red Skorba pottery:  (1–22) vessel shapes represented at Skorba (after Trump 1966, figs. 20, 27, 29); (23 and 24) detail of pottery surfaces on one Red Skorba sherd showing the red-slipped exterior with white-paste-filled incised-line decoration (23) and plain grey interior (24); (25) detail of pottery surface showing fine-grit temper (reproduced with the permission of the Trustees of the British Museum, inv. Nos. P1971–2–2–1A and P1971–2–2–11C; photograph by A. and C. Sagona).

The First Settlers and Farmers

Lithic and Other Implements Textile production at Skorba is evident from the Red Skorba period through to the Tarxien period.126 Re-worked pottery fragments had been fashioned into 3- to 6-cm-diameter discs and perforated for use as spindle whorls (Figure 2.4, no. 3). Another spherical pottery pierced ball probably served a similar function (Figure 2.4, no. 2). It was burnished and incised with a roughly executed all-over vertical zigzag pattern. All these whorls were damaged. Bone, Lipari obsidian, imported flint and local chert flaked tools were present in ‘small quantities’. Of these, a few fragments had use-wear sheen from use in sickles.127 Clive Vella’s examination of the small, general scatter of lithic material from the Red Skorba pavement identified a few obsidian bladelets sourced from Lipari. Though they were still functional tools, Vella could find no indication of hafting and suggested that they were handheld. He did not consider that tools were knapped in this area.128 Vella also drew a distinction between the two rooms in terms of lithic finds: in the North Room there was double the amount of imported lithic material, including two tools, with concentrations of unused chert fragments and tools in the eastern sector; in the South Room, imported stone reduced to small debitage was in the south-western area, and there were no functional tools. A ‘single flint tool’ positioned in the doorway and a broken sickle blade ‘twice the size of normal flint tools’ in the North Room, however, could be viewed as one of the ritual placements in preparation of the structure for use or for abandonment. Whether the room or rooms served a ritual or domestic role, the deposits over bedrock may reflect significant cultural and ritual practices.

Personal Adornment Shells that were pierced for suspension continue into Red Skorba contexts; from Skorba, cockle, winkle and cowrie shell pendants were found. The cowrie shell has clear signs of polish from wear and perhaps frequent handling (Figure 2.4, no. 13). A grooved cow tooth that was probably tied with a leather thong or string (Figure 2.4, no. 12), a ground and modified cow foot bone (Figure 2.4, no. 14) and a perforated slice of boar’s tusk also may have served as a pendants (Figure 2.4, no. 10). A fragment from a greenstone polished axe (Figure 2.4, no. 11) with a biconically pierced suspension hole and a second un-perforated oval-shaped fragment of imported mottled pale grey metamorphic stone also may have served as an amulet. SUMMARY

If the finds are viewed together, and despite the slight variations in their nature, there are strong and consistent themes. Grinding implements abound in the



The Archaeology of Malta deposits apparently in all periods, suggesting that cereals formed a large part of the diet of the early settlers; worn sickle blades also testify to agricultural activities. Domestic livestock was being exploited for its food value – bones are found in the deposits – and yarns spun on spindles were being produced in the houses. Manufacturing activities were being carried out, as indicated by the regular appearance of awls. Quantities of chert, presumably worked for sharp cutting edges, are common, but present, too, are better-produced blades of exotic material. Some cultural and ritual practices hark back to Neolithic homelands to the north. Houses were simple structures approaching ovoid in plan but united into communities by clearly defined boundaries. Structures within the hamlet may have served designated functions; certainly the absence of hearths in some rooms may indicate that cooking was conducted in other zones in the settlement. Signs of deliberate house closure are present. There is some repetition in the kinds of artefacts left in the houses in all periods at Skorba (from Għar Dalam through to Ggantija) which might indicate that certain items of personal and daily use became deliberate deposits when dwellings ceased to function, and these included tools (flaked stone and bone), shell pendants, spindle whorls and sometimes figurines. In the case of Għar Dalam structures, human jaw and skull bone fragments might be familial relics left when a house came to an end. That certain items are usually broken also might be the result of intentional acts and not just discards.


In the Maltese Archipelago there is always a perception of scale – of the finite resources of the land and a vast, bountiful sea  – and a sense that the sea is also an imperfect barrier. Against this background, the sheer monumentality and number of the prehistoric megalithic buildings1 in these small islands invites superlatives. Whether they are the first, and hence oldest, free-standing megalithic monuments, with greater antiquity than Stonehenge, has now been called into question by the finds at Göbekli Tepe, Turkey (9600–8000 BC).2 In any case, such claims can be a distraction from more challenging issues facing archaeologists who question the function of the buildings within society, the role they may have had in the landscape and the growing complexity of the local culture, social memory and motivation for building on such a scale. Funerary complexes are no less impressive, and these have to be taken into account when considering the ancient belief systems in Malta. So intriguing are the Late Neolithic sites in the islands that they have been drawn into a wide-ranging theoretical discussion as to their function and their geographic and environmental setting, the key aspects of which will be discussed in this chapter. DOMESTIC CONTEXTS

Traces of domestic sites and cultural remains which are likely to represent the mundane within the prehistoric phases are relatively under-represented in the archaeological investigations in Malta and Gozo. Five phases – Zebbug, Mgarr, Ggantija, Saflieni and Tarxien – which span the late fifth to third millennia, have been determined for the islands. They are largely based on distinctions in pottery defined by Evans and Trump from the 1950s to early 1970s.3 In broad terms, the period is sometimes referred to as the ‘Copper Age’ (or ‘Eneolithic’) in keeping with Malta’s European neighbours, although metal rarely figures in the island’s cultural remains of the time.4 The last three phases witnessed a period of architecture on a grand scale. This period is commonly known 47


The Archaeology of Malta as the ‘temple period’, but this terminology will be avoided here in favour of ‘Late Neolithic’ or the ‘megalithic building’ phase, less value-laden terms. Excavations at sites such as Ta’ Trapna (tomb site),5 Skorba (settlement site), Ta’ Ħagrat and Kordin (megalithic buildings) clarified the stratigraphic sequence for the islands.6 This fieldwork built on a foundation of early excavations, spanning the nineteenth and early twentieth centuries, and fledgling attempts to characterise the nature of the cultural phases in Malta by local scholars such as Caruana, Zammit and Tagliaferro,7 as well as non-Maltese researchers including Ugolini, Peet and Murray.8 As we have seen in Chapter 2, Skorba has the clearest published evidence of domestic dwellings clustered together into what probably amounted to a village settlement. Eventually, in the Ggantija phase, megalithic buildings were erected over these earlier structures, partly destroying their remains. Reuse of Skorba and other sites over hundreds of years points to their clear viability as habitats and to a deep social memory linking people to certain locations. In the early years of archaeological investigations in the islands, the large monuments were targeted. Only in recent years has the possibility of finding new, less prominent settlement sites via systematic survey and modern technology been carried forward through such ventures as the Maltese Temple Landscape Project (MTLP) focussed on Skorba (2009–14),9 fieldwork carried out by the Cambridge Gozo Project mentioned later and the FRAGSUS (Fragility and Sustainability in the Restricted Island Environments of Malta) programme (2014) discussed in Chapter  2. Recently, Kordin was subjected to a ground-penetrating radar survey by the Belgo-Maltese Survey team.10 There is still significant potential for the identification of additional domestic, funerary and other sites in the rural areas of Malta and even in the privately owned properties which surround urban sites such as Tarxien.11 Domestic locations include natural caves and built structures, but not as yet identified in great numbers.12 As in other regions of the Mediterranean, house structures tend to be round or oblong, but far more is known about settlements and houses in Sicily, for instance, than in Malta or Gozo.13 As part of the Cambridge Gozo Project, Caroline Malone, Simon Stoddart and their Malta-based colleagues provided a concise account of the known habitation areas of the Late Neolithic period and, importantly, a detailed account of the excavation of Għajnsielem structures in Gozo.14 Essentially, implements of food processing such as grinding querns and blades with sickle gloss15 and domesticated animal bone (goat, sheep, pig and cattle) indicate the continuation of mixed farming. The addition of occasional deposits of shells points to some exploitation of marine resources, which had long characterised the economic activities on the island. Indeed, some fixtures attributed to sacred functions within the megalithic building complexes may,

The Culture of the Megalith Builders in reality, have been used for grinding cereals. Spindle whorls, implements and items of personal adornment made of bone and shell indicate that domesticates and gathered resources were further exploited for production of secondary products. Though these economic aspects are discussed here, significant new developments may have occurred towards the closing years of the Late Neolithic; the evidence for these might lie in the so-called cart ruts, which are the subject of Chapter 4.

Z E BBUG PHASE (CA. 4100–3800 BC)

The Zebbug phase was first identified from tomb finds at Ta’ Trapna iz-Zgħira, and it represents a marked change in cultural practices. Whether the new phase was driven by an influx of people or by practices acquired through offshore contact is still not certain. In discussing insular archaeological contexts, Evans made the point that ‘one of the main focuses of interest is the effects of extreme isolation on communities through time, and one of the most fascinating aspects of this study is that of the rate at which cultural traits become lost.’16 In nearly every instance, however, there is a blurring of traditions; seldom do new cultural characteristics simply mirror those of a suggested origin, nor is there a complete abandonment of old, long-established practices in Malta (see Zebbug sites listed in Appendix B, List 4). New pottery techniques appearing at this time are thought to have hailed from the San Cono–Piano Notaro culture of the Early Copper Age in Sicily, though not all the shapes present in the Zebbug repertoire are found in Sicilian contexts,17 and new burial traditions are evident in Malta. Cultural similarities reflecting the influences of immigration, trade or other forms of interaction are equally possible. Nonetheless, continuity in terms of cultural memory and connection to place is suggested by the ongoing occupation of the settlement at Skorba. Trump identified transitional pottery at Santa Verna and Skorba over the so-called Red Skorba shrine (see Chapter  2) which he states ‘has nothing whatsoever in common with RSk [Red Skorba phase]’.18 His observation implies that if there was a period of overlap that might have generated cultural hybridity, it was not apparent. Among the traces of Zebbug dwellings at Skorba (western area) were remnant curved walls of mud brick and stone within which three floor levels were identified. Although no full floor level could be traced, the walls tended to be ovoid in plan. Depressions in the ground in two areas are thought to have served as hearths.19 Scatters of Zebbug pottery in rural areas of Malta may indicate the presence of other settlements, as yet unexcavated.20


The Archaeology of Malta



Ceramic Production Zebbug pottery from the Ta’ Trapna tombs (discussed later) and from the Ta’ Hagrat, Kordin III and Skorba sites is crucial to the definition of this period.21 Vessels are handmade, thin in the wall and moderately hard fired. Examples are usually well burnished to a high polish and pale grey to black in hue throughout. Contemporary with this pottery is a comparatively coarse yellow fabric which is mostly matt and rarely burnished; this fabric shows signs of less effective firing (Figure 3.1).22 Early shapes include wide and deep conical bowls, truncated to form a flat resting surface (Figure 3.1, no. 4), a type found at Santa Verna, Gozo (levels A9 and A10) and Skorba.23 Rims are simple and rounded. Designs comprise pairs of oblique parallel lines over the exterior body which are fringed with excised triangles on the outer edges (Figure 3.1, nos. 1 and 2). Interiors are roughly scratched with zigzag lines and undulating hatched bands which rise and fall around the mouth (Figure 3.1, nos. 2 and 3). These designs have been likened to vessels from Trefontane in Sicily.24 The shape repertoire for the floruit of the Zebbug phase is wide. A distinctive jar form is pear shaped with swelling neck and small mouth (Figure 3.1, nos. 29–31); these vessels often have opposing loop handles and equidistant pierced or solid knobs on the neck and/or widest part of the body. Sharply cut designs that decorate both ware types can be made pre- and post-firing. Designs often consist of two or more clusters of straight or curved lines that can be edged by small punctured triangles made with a stylus. Large jars tend to have horizontal lines dividing neck and body. Decorations often consist of roughly incised clusters of vertical lines – some forming elongated triangles left open (Figure 3.1, no. 31) or filled with cross-hatching (Figure 3.1, no. 23) – as well as V shapes in all-over designs25; patterns can incorporate the handles and lugs into the decorative scheme (Figure  3.1, no.  12). A  dashed line can border the straight lines, somewhat like human footprints or animal tracks (Figure  3.1, nos. 4 and 30). Curved designs cut one above the other appear as large and open scalelike designs; this is a common motif which could be interpreted as representation of simple structures (Figures  3.1, nos. 8, 12, 18 and 30, and 3.2, no. 1). These can be associated with anthropomorphic stick figures with hatched triangles for heads standing inside or outside the arches (Figure 3.1, nos. 12, 19 and 30).26 Other closed-jar forms appear baggy and wide with short, slightly flaring rims or are virtually hole mouthed (Figure 3.1, nos. 22–24). A few jar examples have wide but off-set necks (Figure 3.1, nos. 27 and 28) and tall, flaring necks (Figure 3.1, no. 24), and others have more sinuous profiles (Figure 3.1, nos. 25 and 26). Some vessels may have functioned as cups or jugs with single-loop handles (Figure 3.1, nos. 11–15). Bowls can vary: deep

Figure 3.1. Zebbug pottery forms: (1–4) bowl types from early in the Zebbug phase; (5 and 6) forms from Qala Pellegrin; (7–31) pottery shapes from the floruit of the Zebbug phase. Designs on no. 30 might depict field furrows and channel lines beside which are dashes that might represent the tracks of bipedal footprints (after Trump 1966, figs. 32–34; Evans 1971, figs. 27, 33; Malone et al. 2009, figs. 10.1–10.5). 51

The Archaeology of Malta


and round-based forms (Figure 3.1, no. 16) or conical with straight walls to the simple rim and truncated at the flat base (Figure 3.1, nos. 8 and 9). Other flat-based forms have curved walls and slightly flaring lips (Figure 3.1, nos. 17 and 18). One open bowl shape has two horizontal strap handles (Figure 3.1, no. 20); some have dimpled handle edges and dimpled lips.27 Strainers could be fashioned from conical bowls with regular perforations made in horizontal rows (Figure 3.1, no. 7). Pottery spoons also figure at this time and carry into the Mgarr phase: some have straight, featureless handles (Figure 3.1, no. 10); others have handles wich are curled into a ring.28 Two forms found at Ras il-Pellegrin (Gnejna Bay) were thought to represent a coarse, domestic pottery industry rarely found in the megalithic buildings.29 A jar with sinuous profile (Figure 3.1, no. 6) and a curved bowl-like lid with knobbed handles (Figure 3.1, no. 5) have been identified. Under a plaster-like thick slip, the fabric of Pellegrin Ware is over-fired and bright red to purple in hue; it was derived from the coarser yellow-ware type mentioned earlier, and vessels are not decorated.30 Pellegrin Ware first appears during the Zebbug but continues into the Mgarr and Ggantija phases at Skorba and into the Tarxien phase at Baħrija. Within the Zebbug phase, painted pottery also has been documented. Red paint on pale ground, thought to be an imported technique (although locally produced), was contemporary with vessels carrying incised designs. Occasionally, both techniques (paint and incised decoration) appear on a single vessel.31 Vessel sizes, especially among the jars, vary considerably, from miniature (Figure  3.1, no. 29) to large (Figure 3.1, no. 31).

Lithic and Other Tools Among the finds from Skorba were tools made from imported stones, including two axes polished and ground to a sharp edge, one complete and the other damaged (Figure  3.2, no.  3).32 A  whetstone made from exotic red material may have been used to keep the axe heads or other stone tools sharp,33 and a pumice fragment has a worn groove from sharpening an awl-like tool. Fewer Lipari obsidian and flint tools are present than in earlier phases.34 The bone tool industry included an awl fashioned from a sheep’s bone and a slice of cow bone well worn from use as a burnishing tool, possibly in pottery production. Two pottery spindle whorls were found at Skorba – not made from reworked sherds – and both were broken. An unspecified number of querns made of the local, hard Coralline limestone point to an economy that included agriculture and grain harvests.

Funerary Contexts Five shallow pits cut into the bedrock with roughly circular to oval plans were found at Ta’ Trapna iz-Zgħira, Zebbug. They contained human bones

The Culture of the Megalith Builders

Figure  3.2. Zebbug cultural remains:  (1)  pottery fragments with a design perhaps depicting an agricultural scene, found in tomb 5 in field Ta’ Trapna iz-Zgħira, Zebbug (after Baldacchino and Evans 1954, fig. 5:5); (2) Zebbug house in the western sector of Skorba; (3) ground axe from Skorba (after Trump 1966, figs. 8 and 35); (4) shell


The Archaeology of Malta


accompanied by personal items, lithic tools and pottery; animal bone, shells, beads and amulets also were present within the burial pits (Figure 3.2, no. 4). Pottery and fragments of limestone bowls were found in Ta’ Trapna tombs 3 and 5.35 The condition of the deposits was described as ‘utter disorder’,36 and it was thought that the upper structure of the tombs might have been quarried away, leaving only the shallow floor depressions. Recent excavation of an intact two-chambered tomb immediately south-east of the Xagħra Circle burial complex offers a more complete picture of funerary practice. In structure, the rock-cut tomb, accessed via a round opening to a vertical shaft, finds parallels to a type known as a forno (‘ovenlike’) in Sicily, southern Italy and Sardinia (Figure 3.2, nos. 8–10).37 Numerous burials (fifty-four or more adults and eleven children) were identified in a series of depositions showing signs of disturbance at each new burial; old burials were pushed to one side, and some materials were removed from the tomb and dispersed into the surrounding area.38 Remnants of other tomb chambers with traces of Zebbug deposits, which were probably part of an original Zebbug necropolis, were located at the sides of the later hypogeum complex.39 Finds associated with the burials were comparable to those in Ta’ Trapna. Furthermore, like tomb 5 at Ta’ Trapna, the Gozo tomb was marked by a menhir depicting a stylised human head of similar form and proportions (approximately 20 cm tall); further parallels are known in southern European and central Mediterranean regions (Figure 3.2, no. 7).40 Surprisingly, it has been demonstrated through stable isotope analysis on the skeletal remains of seven Neolithic people from the Xagħra Circle burial site – accelerator mass spectrometry (AMS) dated between 4348 and 2476 cal. BC (Zebbug to Tarxien phases) – that within the last ten or so years of those people’s lives, marine animals such as fish and molluscs did not figure in their diet. Instead, meat and milk products provided protein. It has been acknowledged, though, that the results might be misleading because of the small sample size and factors such as social elitism for those buried at the site, and the findings might not reflect the whole population of the islands.41 Certainly agriculture and animal husbandry must have continued as the main economic activity given the absence of abundant remains of marine or terrestrial wildlife. Colour held symbolic values for Neolithic communities in Malta and Gozo. The Ta’ Trapna menhir was coated in red pigment, and red ochre cakes as well Figure 3.2. (cont.) dome-shaped button with V-shaped drilled hole on flat surface, from tomb 5 in field Ta’ Trapna iz-Zgħira, Zebbug (after Baldacchino and Evans 1954, fig. 8:5); (5) ground axe from Zebbug-phase double-chambered tomb adjacent to the Xagħra Circle, Gozo (after Malone et al. 1995, fig. 20.6); (6) spindle whorl from Skorba (after Trump 1966, fig. 31:c); (7) menhir from the Zebbug tomb in the Xagħra Circle; (8–10) Zebbugphase double-chambered tomb adjacent to the Xagħra Circle,  Gozo (after Malone et al. 2009, figs. 7.7 and 10.46).

The Culture of the Megalith Builders as pots containing the material also were found in the graves there.42 Of the beads among the burials, some may have been deliberately coloured red, but others appeared to be caked with ochre, perhaps as a result of it being liberally rubbed over the body either in life or at the time of burial.43 In the east corner of the San Pawl Milqi site was another Zebbug burial, dug into an area of ‘soft rock’. A skeleton in flexed position on its left side was placed on a bed of arranged pebbles with an ochre deposit near the skull.44 It is still debated whether red ochre was among the known imported commodities, namely, obsidian and stone artefacts, which indicate an ongoing lively exchange system with neighbouring islands. Ochre can be found in Malta, but whether it was plentiful enough to match demand is uncertain. The mineral occurs, for instance, in Għar Ħasan, where paintings, of dubious Palaeolithic date, have been identified (discussed in Chapter 2). The possibility that the cave had more extensive ochre deposits, which were largely mined out, has not been scientifically investigated, but the floor is said to be covered with ochre-stained soil.45 Given that a cultural change occurred in the Zebbug phase, interaction between the islands may have included an exchange of people as well as skills and cultural practices. Imported wares with finely incised and cross-hatched triangles from Skorba and Santa Verna may have come to the islands during the Grey Skorba or Zebbug phase, but the contexts do present some problems.46 Polished axes and axe-shaped pendants were often (but not exclusively) made of greenstone, which may have had value at various levels: as an imported rare commodity, as a revered implement that required considerable labour to produce and for its green colour. Aspects of colour symbolism are not restricted to the Zebbug phase; indications appear within the earliest Neolithic contexts and continue well after, as considered later.47 MG A RR PHASE (CA. 3800–3600 BC)

Evans first mooted that Mgarr should be considered a separate phase,48 and this was seemingly confirmed by clear, though limited domestic deposits at Skorba.49 Only the sequence needed an adjustment; Mgarr is now positioned between the Zebbug and Ggantija phases.50 Nonetheless, the nature of this cultural unit is still debated:  whether it should stand as a distinct phase or should be viewed as a regional variant within the Zebbug phase.51 Perhaps because of the limited Mgarr finds, there are few indications of offshore contacts.52 In Skorba, house remains survived in the deeper deposits behind the rear wall of the western and later megalithic structure. They were described as “sub-rectangular” in plan (Figure 3.3, no. 14).53 The Kordin III megalithic site has earlier apsidal house plans clustered into a village; these have been tentatively identified as Mgarr based on the pottery present at the site but not positively linked to the domestic structures



The Archaeology of Malta

Figure 3.3. Mgarr cultural remains: (1–12) pottery forms (after Trump 1966, fig. 20; Evans 1971, fig. 34; Malone et al. 2009, fig. 10); (13) stone hammer (after Evans 1971, fig. 44); (14) remnant Mgarr domestic deposits at Skorba (after Trump 1966, fig. 8).

Figure 3.4.  (A) Kordin III megalithic building plan: (1–5) shaded area to the north are domestic structures, perhaps of the Mgarr phase (after Evans 1971, plan 17, and N. C. Vella 2004, p. 17); (6) elongated quern stone with several pits ground from use, placed across the threshold of a subdivided apsidal chamber; (7) paved forecourt. (B) Possible menhir at Xemxija. (C) Menhir at Kirkop. (A and C Sagona photographs.) 57

The Archaeology of Malta


(Figure 3.4, no. A).54 Cave sites also appear to have served as dwellings at Għar Dalam, Għajn Tuffieħa and Għar in-Nagħag,55 where signs of habitation include pottery, stone and bone tools and sometimes hearths with associated marine and terrestrial animal food remains (sites with Mgarr deposits; see Appendix B, List 5).

Material Culture Mgarr pottery is usually black to dark grey, though occasionally tending to yellow in hue, handmade and thoroughly fired (Figure  3.3). The shape repertoire has been subject to some revisions since Evans’ first scheme.56 Straight-walled cylindrical vessels with wide, flat bases or tops served as lids57 and bowls (Figure 3.3, nos. 11 and 12); these could have four small pierced lug handles on the lower edge, and another form was more splayed towards the flattened resting surface. Common forms were deep and baggy bowls (Figure  3.3, nos. 3 and 5), widest low on the profile with simple rounded rims, and large jars with off-set conical tall necks that lean into a featureless lip (Figure 3.3, no. 6). Another large jar type has a slightly off-set shoulder and long, wide neck which narrows to the rim (Figure 3.3, no. 4). The lip can be thickened, and two small or medium-sized loop handles can be attached at the off-set. Round-based bowls can have high loop handles (Figure  3.3, no. 9) or lug handles at the rim (Figure 3.3, no. 8). One jar form has a vertical short rim (Figure 3.3, no. 1). Painted pottery was found but is uncommon. Decorated strap handles were placed on cups and jugs. False string holes (not pierced right through the wall of the vessel) and pierced or solid angular lugs are represented. Shallow and wide cut-out linear designs can have chipped or lightly fringed edges. The addition of white paste in-fill, over which red ochre also could be applied, was a new practice.58 Rims can be decorated with small cuts in the lip. Curved lines are frequently encountered, but the effect is often like white or ruddled ribbons arching across the panels. Designs are cut pre- or post-firing into the otherwise highly polished surfaces.59 Panels were created using lines, as well as tracks formed by dashes. Convergent sweeping lines heralded patterns which would become a hallmark of the Ggantija phase. Linear designs that fan out from lug handles also continued as a popular decoration into the next period. The few tools remaining in Mgarr contexts at Skorba point to food production, including a flat-lipped and shallow stone tray, rectangular in shape with a gap on one side for pouring or funnelling out processed material. Two pebbles found in the same context may have been used as mortars, perhaps associated with the tray.60 Ta’ Ħagrat finds were equally scarce:  a hammer with ground waist (Figure 3.3, no. 13), a shark’s tooth and Lipari obsidian.61

The Culture of the Megalith Builders G G ANTIJA PHASE (CA. 3600–3200 BC)

Domestic Contexts Ggantija continues along the path of development, which had its origins in the Zebbug and transitional stages of the Mgarr phase. The Ggantija phase appears to have endured for centuries, and the monumental building works that appeared in Malta and Gozo suggest that it was a time of prosperity and stability. A few domestic structures investigated at Skorba were oblong in plan and built of 6-cm-thick mud bricks with a floor of torba62 and beaten earth (Figure 3.5). One doorway abutted and was blocked by the megalithic buildings, and this arrangement established the construction sequence: the houses predated the monumental structure at this site. In the so-called House of the Querns, the contents were found in situ, but the house had been burnt, and it is quite possible that it was purposefully brought to an end.63 Deliberate ending of a building’s use is suggested at various times, for example, by the deposit within the two previously discussed Red Skorba houses and by the final days of the Tarxien megalithic complex (see Chapter 4). Foreign influence is minimal in the Ggantija phase.64 The evidence of settlement sites in Gozo is meagre, but it nonetheless points to associations with fertile lands suitable for agriculture, reliable water supplied by springs and locations possibly within territories surrounding megalithic building sites. Two domestic structures at Għajnsielem, Gozo, were a chance find by Joseph Attard Tabone, who identified in situ prehistoric deposits at a building site (Figure 3.5, no. 8). The walls, though now largely missing, once edged a rock-cut depression that formed the floor of the larger oval house (Structure 1), which was 7.5 m × 5 m in size, and a smaller room was to the south. Pottery found within the Għajnsielem buildings spanned from the Ggantija through to the Saflieni and Tarxien phases.65 Other sites include Is-Srug in Gozo,66 where mud brick was used in the constructions. Pottery assigned to the Ggantija phase and bones from domesticated animals were recovered. At Ħagar Qim, there were traces of an oval house that had a floor cut from the bedrock and stone-built walls; the remains have been tentatively assigned to the Ggantija phase.67 Sherd scatters on and around cliffs to the north of a cluster of tombs at Xemxija (see later) also may be indicative of a settlement; west of the tombs are remnant large stones, possibly from a megalithic structure. The cave sites of Għar Dalam and Għar in-Nagħag (Gozo) contained some Ggantija pottery and artefacts. Prehistoric levels in the latter spanned the Mgarr, Ggantija and Tarxien phases, perhaps indicating a temporary campsite; a hearth was near the entrance, and charcoal particles and animal bone were found in the same context.68 A similar chronological span and range of finds – hearth, pottery, burnt animal bones, marine shells and bone tools – were


Figure 3.5. Ggantija cultural remains: (1) spindle whorl; (2) bone pendant; (3) anthropomorphic figurine; (4) possible table fragment; (5) carved stone plaque with triangular edge (all from Skorba, after Trump 1966, figs. 26, 31, 40, and 41); (6 and 7) house models from Ħagar Qim (see Evans 1971, p. 91, pl. 39, inv. no. Q/P1001; C. Sagona photographs); (8) Għajnsielem Road, Gozo, house plan (after Malone et al. 2009, p. 47, fig. 3:17); (9) domestic structures at Skorba (after Trump 1966, fig. 8). 60

The Culture of the Megalith Builders found in caves in the cliff face near Għajn Tuffieħa, the site of an extensive Roman-period bathhouse complex.69 It was during the Ggantija phase that the building of megalithic structures began in earnest (see later).

Ceramic Production Skorba produced some pottery that appeared to be transitional – well made and with precision demonstrated in the linear designs fringed by fine cuts.70 Grey-white paint over dark-surfaced pottery is also considered an intermediate ware type. Designs comprise wide curvilinear and linear bands, first observed in Mgarr contexts and in the later Ggantija House of the Querns at Skorba. Baggy deep bowls and round-based forms in the Mgarr phase also reflect a developing repertoire. A sharply carinated shoulder fragment from a large jar is represented too.71 The floruit of Ggantija pottery is handmade, relatively thin walled and varies in hue from black through greys to brown (Figure 3.6). Firing can produce a mottled surface with patches of red to yellow. One large jar – a unique find from Kordin III72 – has an off-set low on the body and a truncated conical neck that leans into a simple rim (Figure 3.6, no. 5); loop handles are attached at the off-set, and four equidistant lugs are on the shoulder. Some jar shapes have a wide, squat body with medium to tall conical necks and featureless rims; profiles can be deeply curved (Figure  3.6, nos. 12–14) or defined by sharp offsets. A hole-mouth form has a thickened rim and horizontal handles high on the body (Figure 3.6, no. 8), and another variation has a short vertical neck. One smaller pot has a carinated profile low on the vessel, above which the wall curves into a hole-mouth, and lugs are placed on the off-set (Figure 3.6, no. 3). Bowls can have well-formed disc bases (Figure 3.6, nos. 18 and 25), slightly omphalos (Figure 3.6, no. 20) or rounded resting surfaces (Figure 3.6, no. 26). Bowl rims can be pierced at four points to take string (Figure 3.6, no. 26) for suspension or to secure lids. The deep and baggy bowl continues from the Mgarr repertoire (Figure 3.3, no. 5; compare Figure 3.6, no. 6). Another deep form can have off-set or sharply carinated profiles and handles placed low on the wall (Figure 3.6, no. 24). Shallow bowls can have a raised ridge exaggerating the off-set in the profile, above which the wall swells before curving into the rim (Figure 3.6, no. 28) or tends to be somewhat vertical to the rim (Figure 3.6, nos. 27 and 29). Another shallow form is conical with a small loop handle at the rim (Figure 3.6, no. 11). More refined lids with string holes are derived from the flat-topped forms of the previous period (Figure 3.6, no. 10), and a deep cylindrical form, made to fit snugly over a bowl, also can be fitted with string holes on the flanged upper edge (Figure 3.6, no. 9). One simple small bowl has oblique holes at the lower edge (Figure 3.6, no. 15).



The Archaeology of Malta

Figure 3.6. Ggantija pottery (after Trump 1966, figs. 20, 36 and 37; Evans 1971, figs. 13 and 14; Malone et al. 2009, fig. 10).

The Culture of the Megalith Builders Highly burnished finishes are usual and can carry finely scratched and intricate designs over which liberal amounts of red ochre paste were applied; the scratched guidelines were made to hold and thus were originally obscured by the colour paste.The technique generally was not durable, and only remnants of red ochre in-fill usually survive. Various forms of ‘comet’ motif are found: some have a small circle at the confluence of the lines; on others, the narrowing lines are divided off into a triangle, in-filled with intricate hatched patterns (Figure 3.6, nos. 5, 9, 11, 20 and 23). Bold chequer patterns occur on carinated pots that have raised ridges into which small pierced lugs have been fashioned (Figure 3.6, no. 7). Chequer patterns and a development of the comet design which has small fine cuts fringing the tail lines might have occurred late in the Ggantija sequence, as they continue into the Saflieni phase.73 Pellegrin Ware continues at the type site in association with Ggantija pottery. Although Ggantija Ware was virtually absent at Tas-Silg, a case has been made recently for some surviving remnant stylistic features in the pottery, but such trends alone would not be enough to support an argument for a permanent settlement prior to the Tarxien phase.74

Other Cultural Artefacts Artefacts from Ggantija contexts at Skorba were mixed with earlier material (Figure  3.5, nos. 1–4). Nonetheless, a familiar range was recovered from the houses. Agricultural activities were evident in numbers of querns and rubbers; eleven grinding stones of local Coralline limestone from one Ggantija house earned it the name ‘Hut of the Querns’.75 A seafood element within the Ggantija diet is suggested by a pit in one of the houses containing 4 limpet and 120 Venus shells.76 Two spindle whorls (Figure 3.5, no. 1) and several bone awls, two made from bird bones, are among the surviving toolkit.77 Imported flint and obsidian as well as local chert flaked material are similar to the previous Mgarr industry, including two sickle blades, worn from use. Notwithstanding the problems surrounding the Ras il-Pellegrin site on a western promontory (excavated in the 1970s, it was never published, and no field notes are available), Clive Vella worked through the lithic remains and concluded that despite close proximity to local chert sources, the lithic industry had a significant amount of imported flint. As one reason for this, he suggested the influence of the excellent anchorage in the nearby Fomm ir-Riħ and Gnejna Bays, through which imported materials presumably entered the island. Indeed, exploitation and protection of these landing points may have been the motivation for a settlement there.78 Local chert, however, was clearly a valued commodity, and Ras il-Pellegrin equally may have been chosen as a settlement because of its proximity to outcrops.


The Archaeology of Malta


Burial Sites As for the Zebbug contexts, people were interred in rock-cut chambers, which were accessed through small round openings into one or more chambers. Partial or well-defined lobed or apsidal cavities were scooped out of the rock at the edges. Six examples were excavated in 1955 at Xemxija (St. Paul’s Bay) which contained predominantly Ggantija-period interments and pottery (Figure  3.7), but no detailed study of the skeletal remains or of the pottery and other artefacts including axe amulets has been made. A seventh chamber was cleared without documentation in 2001.79 Ggantija Ware figures among the mortuary contexts in the Ħal Saflieni hypogeum, although details of the excavation are lacking.80 SAFLIENI PHASE (3300–3000 BC)

Evans first suggested in the 1950s that the Maltese prehistoric repertoire might be broadened to include the category of Saflieni Ware, so named after the hypogeum site of Ħal Saflieni.81 His suspicion was confirmed by the occurrence of Saflieni-styled pottery at Skorba: stratified below Tarxien levels under the east ‘Temple’; over Ggantija levels in the post-abandonment fill in the ‘Red Room’, trench X; and in a pit in trenches QE and SE.82 He expressed some reservations that, in the absence of architecture and more substantial remains other than pottery, there was some room for doubt – doubt that is still prevalent.83 Trump also indicated that it was hard to distinguish the wares from what might be considered early Tarxien material but decided, based on the ‘pure levels’ with Saflieni Ware at Mgarr, Skorba and Santa Verna, that a sub-division seemed warranted ‘even if it remains little more than a transitional one’.84 Nonetheless, Trump held firm that it was ‘in a very real sense a transition between Gg (Ggantija) and Tx (Tarxien)’.85 Without good stratified contexts, much the same could be said about the stray examples that appear in secondary deposits at Tas-Silg. All things considered, fragments designated as Saflieni Ware from Tas-Silg do have different qualities  – light weight, hard fired but less compact than Tarxien Ware (Figure 3.8, no. 18). Overall, it does not seem unreasonable to classify some vessels as early forms of Tarxien Ware, in which case the Saflieni subdivision remains a useful sub-category by which to define a recognisable, probably chronological difference from the floruit of the Tarxien pottery industry.86 That such types are found outside burial deposits, as at Tas-Silg, suggests that Saflieni Ware is not solely a ‘funerary style’.87 Pottery shapes recall the previous Ggantija phase, for example, the deep bowls with flaring walls, bowls with virtually straight walls to the rim (Figure  3.6, no.  17; compare Figure  3.8, no.  10)  and the deep pots with angular handles placed low on the body (Figure 3.6, no. 21; compare

The Culture of the Megalith Builders

Figure  3.7.  (1) Tomb 5 entrance; (2)  plan of the Xemxija tomb site (after Trump 2002, p. 163); (3–5) tomb 5 plan and sections; (6) tomb 6 at Xemxija, entrance at top, broken roof in the lower area; (7 and 8) plan and section of tomb 6 (photographs by C. Sagona; plans and sections after Evans 1971, plans 28 and 29).


Figure  3.8.  (1–15) Saflieni-phase pottery forms (after Trump 1966, fig.  42; Evans 1971, figs. 10 and 13; Malone et  al. 2009, fig.  10); (16) detail of Saflieni pot surface with incised designs made post-firing (TSG96/220/14); (17) incised and spotted design with surviving white paste in-fill (TSG/2061/7); (18)­­pottery section showing well-levigated Saflieni clay (TSG96/69/28); (16–18 from Tas-Silg; photographs by C. Sagona). 66

The Culture of the Megalith Builders Figure 3.8, no. 7). New developments are extreme carinated forms such as bowls with high shoulders and some with rims angled markedly into the bowls forming anti-splash rims (Figure 3.8, nos. 12 and 13), which would carry over well into the Tarxien phase.88 Smaller deep pots, again displaying pronounced and angular off-sets at the widest part of the body, could be fitted with a foot similar to later Tarxien shapes (Figure 3.8, nos. 3 and 4; compare Figure 3.10, nos. 8 and 9. carinated forms, and 13–15, with high feet). Wall thickness can be thin relative to vessel size. Back-to-back curved and straight linear decorations can be linked with hatched fine lines or finely flecked spots (Figure  3.8, nos. 16 and 17). Surfaces tend to be polished. Rusticated applied decorations fan over the lower exterior walls in wide scallops on coarser deep jars. This technique also survives into the later Tarxien pottery repertoire, where such freely and haphazardly applied decorative layers are moulded into an elaborate array of all-over designs (Figure 3.8, no. 15; compare Figures 3.9, no. 15, and 3.11, nos. 13, 14 and 25). The lithic technology in Saflieni contexts at Skorba is akin to previous phases, with the exception of a new exotic grey flint, possibly sourced from the Hyblaean Mountains (Monti Iblei) in the south-eastern corner of Sicily near Siracuse. This stone was distributed far and wide in the Neolithic period.89 TARXIEN PHASE (CA. 3000–CA. 2500 OR POSSIBLY 2200 BC)

Essentially Evans’ landmark study still remains at the core of discussions of the prehistoric period in Malta. Since it was published, Xagħra Circle, Gozo, has provided new material on the Tarxien megalithic period, more has been found at Tas-Silg and contemporary wares from Borg in-Nadur have been revised recently.90 Traces of clear domestic sites in the Tarxien phase are limited: there were some finds of Tarxien date within the Għajnsielem house first occupied in the Ggantija phase91; a dwelling found at Tac-Cawla (southern Victoria, Gozo) appears to have been built in stone with torba flooring and a hearth92; domestic deposits in the same district came to light in 1995 and were associated with a stone wall described as ‘substantial’93; and a rounded house with torba flooring was located at Kuncizzjoni (diameter 7 m × 7 m).94 As with other phases in prehistoric Malta, caves also were frequented in the Tarxien phase, notably Għar Dalam, Għar in-Nagħag, Għajn Tuffieħa, Torri Falka and Ta’ Pergla iz-Zgħira. At Torri Falka, lithic material, spindle whorls and torba flooring were assigned to the Tarxien phase (see Appendix B, List 5 for Tarxien sites).95

Ceramic Industry during the Tarxien Phase As with most ceramic industries, both fine and coarse Tarxien Wares were produced (Figures 3.9–3.11). While various schemes have been devised for the


Figure 3.9.  Tarxien-phase pottery jar features: (1) tunnel handle from a jar with haphazard vertical incised lines; (2) angular so-called bridge-nose handle; (3) horizontal rows of flecked notches decorating the surface of a jar; (4) raised spot design in thick white paste ground; (5) incised hatched ‘comet’ motif; (6) incised linear design (1–6 reproduced with permission of the Trustees of the British Museum, inv. nos. (1) 1886, 5–1, 19; (2)  1852, 6–14, 26; (3)  1964, 12–6, 941; (4)  1923, 7–11, 41; (5)  1923, 7–11, 5; (6) 1923, 7–11, 17; photographs by A. and C. Sagona). Nos. 7–17, jar forms (after Trump 1966, figs. 20 and 36; Evans 1971, figs. 4, 9, 17–20 and 22; Malone et al. 2009, fig. 10). 68

The Culture of the Megalith Builders

Figure 3.10. Tarxien-phase pot and bowl forms (after Evans 1971, figs. 5, 7, 20, 21, 23 and 24).


Figure  3.11.  (1–22) Tarxien pottery bowl, ladle, lid and tray forms (after Evans 1971, figs. 10, 19, 21 and 22; Malone et al. 2009, fig. 1.14); (23) a decorative feature comprising flattened and finger-impressed rows of applied rolls of clay; (24) vertical lines roughly incised over the entire pot surface; (25) detail of fan design on lower wall and vertical lines above it; (26) haphazard finger-impressed jab marks made in zones over the surface of a vessel (23, 24 and 26 reproduced with permission of the Trustees of the British Museum, inv. Nos. 1852, 6–14, 9; 1886, 5–1, 8; 1964, 12–6, 947; 25 TSG96/303/24; photographs by A. and C. Sagona). 70

The Culture of the Megalith Builders range of ware types, essentially vessels are handmade from well-levigated clay. While fired adequately, in the finer vessels, clays generally remain quite dark at the core, and the walls of the vessels can be remarkably thin; coil manufacturing technique is likely. Dark grey and black surfaces are usual, although some red- and yellow-slipped wares do occur. Large tall-necked jars with narrow mouths can have distinctive tunnel handles built usually into the widest part of the vessel; these fixtures bulge markedly on the vessel interior (Figure 3.9, nos. 1, 7 and 9). Deep and sharply carinated bowls and pots were common, with equally distinctive ‘nose-bridge’ handles (Figure 3.9, nos. 2, 10, 12 and 16) or small pierced knobs high on the vessel wall. Other carinated deep pots had high, virtually horizontal shoulders which angled back into the mouth of the vessel, effectively serving as an anti-splash feature (Figure 3.10, nos. 4–6). Bowls were made in a range of forms (Figures 3.10, nos. 7–12, and 3.11, nos. 6–14 and 17–19), some with ladle-like handles (Figure 3.11, nos. 2 and 3) and others with a well-formed high ring foot (Figure 3.10, nos. 13–15). Steep-sided to shallow conical lids, some with a handle (Figure 3.11, nos. 20–22), spoons or scoops (Figure 3.11, nos. 1 and 4) and pierced sieves are also documented (Figure 3.11, no.  11). Evans saw antecedents of some Tarxien shapes within the Ggantija phase,96 and a number of the pottery shapes were made in varying sizes from quite large to miniature (e.g., Figure 3.9, no. 7; compare no. 8). DIMPLES, PELLETS AND BOSSES AS DECORATIVE FEATURES

One of the themes running through the Neolithic artistic expression in Malta is the use of spots. Drilled holes were bored into some large structural stones or fixtures within the monumental buildings (Figure 3.17, no. 1), and small applied pellets or puncture marks could cover some pottery surfaces (Figures 3.9, no. 4, and 3.17, no. 14). Both are eye-catching, and both can form all-over patterning. Significantly, a third technique can be teased from the archaeological record. Zammit documented several hundred stones, round in circumference, concave on one side and convex on the other, that were found in Tarxien; ‘a good number’ also were recovered from the Ħal Saflieni burial complex,97 and thirty-two were located in the Ħagar Qim Neolithic building.98 They were on average between 7 and 11 cm in diameter and around 5 cm deep. Significantly, the clusters of convex stones were found embedded in mortar-like matrix at the foot of large standing stones near doorways in Tarxien (Figure 3.12), and it is possible that they may represent decorative bosses that slumped to the ground over time: ‘a great number (over 50) of the concave hemispheres found in this place, not regularly laid down but massed & caked together as if in mortar’.99 In another location they may have formed a linear decoration in a possible threshold (apse-shaped room C, south-eastern entrance).100 (continued)



The Archaeology of Malta

Figure 3.12.  Extract from T. Zammit’s Archaeological Field Notes, no. 11, 23.8.1915, p. 37, depicting convex-concave worked-stone bosses in two lines across the threshold in the eastern sector of apse C in Tarxien; all but two (shaded) were found with the concave side facing down (held in the archives of the National Museum of Archaeology,Valletta).

Judging by the mortar on the concave side (probably a technique to improve the bond between the plaster and the stone boss), the convex surface probably faced out. As they were well formed on the convex side, Zammit also thought that this aspect probably represented the outer face.101 As a decorative technique, bosses set in mortar around a doorway parallel the drilled decoration around entrances at Mnajdra. It cannot be discounted that the decoration had greater symbolic meaning, hinting at artistic representations of spotted skirts on the statuettes in Malta, among them the famous sleeping figure from from Ħal Saflieni (Figure 3.18, no. 1), or possible allusions to trance states and the numinous (Figure  3.17, no.  6). From a phenomenological point of view, all these decorative techniques present a tactile quality that also might have been significant. Decorative elements on finer wares, as for all the Late Neolithic period, are frequently the means by which pottery is categorised (Figures 3.9–3.11). For the Tarxien period, scratched (Figure 3.9, nos. 5 and 6), often paired lines in sweeping curves with bifurcated ends (Figure 3.9, nos. 13 and 16) are coupled with jagged motifs (Figure 3.9, no. 13). These linear designs can be left open or further enhanced with fine hatching, usually angled this way and that in zones (Figure 3.9, no. 5). Other vessels have applied round and flattened pellets, effectively set into a thick ground of white paste in-fill (Figures 3.8–3.10). Less common are regular bands of small, flecked incisions covering the surfaces of vessels (Figures 3.9 and 3.10); small flecks also were used to fill around larger, more complex designs (Figures 3.10 and 3.11). Of the coarse wares, often referred to as ‘Sandy Pink Ware’, the fabrics tend to be grittier, less well levigated and fired to brown or reddish hues with a

The Culture of the Megalith Builders sandy texture. Large jars, smaller bowls and cooking vessels are made with these less-refined fabrics. Designs have a rustic quality. Clays liberally applied to the surfaces in thin or thick streaks can fan out in large petal-like shapes (Figures 3.9 and 3.11) or be daubed onto the surface in haphazard fashion.They were then either left as is or sometimes finger-impressed into bands of scale patterns (Figures 3.9, nos. 11 and 14, and 3.11, no. 23) or jabbed irregularly, forming a pitted surface (Figure 3.11, no. 26). Incised, roughly executed vertical lines can decorate most of the vessel surface (Figures 3.9, no. 1, and 3.11, nos. 12, 17–18 and 24). Only trays appear to have finger-impressed ‘pie-crust’ rims (Figure 3.11, nos. 15 and 16).

Aspects of Material Culture Of personal adornment there is a variety of designs among pendants and beads of bone and shell (Figure 3.13, nos. 7–10).102 None of the imagery from the monuments, however, depicts people wearing anything other than fringed, spotted skirts or simple loincloths. As we have seen previously, obsidian at Skorba was sourced largely from Lipari, with lesser amounts from Pantelleria. This trend is reversed at the Xagħra Circle in Gozo, but a downturn in obsidian usage generally is commensurate with wider Mediterranean trends in the later Neolithic phases.103 Imported flint persists in the archaeological record, which in itself indicates that Malta was not isolated, as is sometimes inferred from the literature.104 Close proximity of Borg in-Nadur to the sheltered Marsaxlokk Bay probably meant ready access to any lithic materials entering the island at that point, just as Ras il-Pellegrin seemed well positioned on the rim of Fomm ir-Riħ and Gnejna bays in the west of the island. Chert continued to be exploited for tools, some of which were deposited with burials in the Xagħra Circle complex.105 Chert recovered from Tas-Silg and Borg in-Nadur may have been acquired from a distant Maltese (as yet unknown) location and processed closer to the source.106 Grinding stones that continued to be found in ‘unabated numbers’ in Tarxien contexts remain virtually unpublished for the entire Neolithic period.107 EVIDENCE FOR OFFSHORE CONTACTS IN THE NEOLITHIC PERIOD

Contacts with other Mediterranean areas after the Red Skorba phase, once considered limited,108 have been re-evaluated recently by Clive Vella, drawing on lithic evidence, which at Ta’ Ħagrat ‘appears to “peak” in the Late Neolithic’.109 Persistent offshore trade is evident in the importation of flint material, filling the void created by a waning obsidian trade.110 Pottery, too, may have been a traded commodity, and a (continued)


The Archaeology of Malta


few exotic examples in Malta point to contact with the Piano Quartara (Lipari), Serraferlicchio and later Sant’ Ippolito (Sicily) cultures.111 Further inter-island connections mooted between Malta, Sicily and Sardinia concern transplanted cultural features (architecture and iconography) in the closing years of the Late Neolithic period; these links are explained from the perspective of immigrant populations abandoning Malta after cultural collapse, a scenario which is questioned in Chapter  4.112 The presence of Thermi-styled ware in Malta (see Chapter  5) in the Late Neolithic, might point to a shifting focus of contact with the distant Aegean through Italian networks.113

MEGALITHIC STRUCTURES I am uncertain who was the first scholar to speak, early in the last century or even earlier, in this context of ‘temples’. But they certainly have much to answer for.114

After some nine centuries of modest house construction in the islands, towards the end of the fifth millennium, this trend changed. Buildings with large apse-shaped rooms, somewhat clover-leaf in plan and massive stone slab construction, represent an astounding architectural leap in scale and technical know-how. Megalithic in every sense of the term, the structures burst onto the prehistoric scene accompanied by equally distinctive cultural facies. From the time of their discovery and excavation, the megalithic buildings have excited much popular speculation and scholarly theoretical discussion, more than any other period or archaeological feature in the Maltese landscape.115 The structures were quickly designated temples, yet the role they had in ancient island society might have been far more complex, influenced by and contributing to social definition, group interaction and economic stability, as well as embodying a religious and ideological focal point for the community that built them.116 There is also an element of the fortress in the massiveness of the walls and the scale of the buildings which might indicate that whatever or whoever was kept inside, or perhaps the revered internal space itself, needed protection.117 Accumulated knowledge associated with cultural relics and status items – symbols of authority, valued commodities or resources – may have been held there. It is possible that like substantial buildings of later periods and more complex societies, for example, the temples and palaces of other regions such as in Mesopotamia, the monuments in Malta also may have served as storehouses for mundane commodities such as food surpluses. We should not dismiss the possibility that interior spaces may have been used as dwellings, perhaps for religious or other hierarchy, people passing through rites of passage or members of society who were isolated from the general community for other cultural reasons.118 Indeed,

Figure  3.13. Tarxien cultural remains:  (1)  stone pendant; (2 and 3)  stone conical objects; (4) stone object; (5) pierced pottery artefact; (6) stone bowl (1–6 after Trump 1966, figs. 26 and 40); (7–10) pierced pendants. Tools:  (11–27) flaked lithic material from Borg in-Nadur as categorised by Clive Vella:  (11) core; (12) blade; (13) knife; (14–18) all-round scrapers; (19), thumb scraper; (20–24) side scrapers; (25) awl; (26 and 27) backed blades (after C. Vella 2011, figs. 6.4–6.7). 75

The Archaeology of Malta


Reuben Grima and Mario Vassallo fall just short of identifying the monumental buildings as dwellings: As is most eloquently attested at Skorba, many of these monumental structures were built in places that had already been used by the inhabitants for several centuries, primarily for settlement purposes.The choice of location, therefore, appears to have been dictated by the considerations that are foremost in any agricultural settlement, namely proximity to land suitable for agriculture, and a reliable supply of fresh water.119

Whatever their actual function, the megalithic structures are likely to have been the dominant feature in rural landscapes dotted with other, comparatively modest settlements. Andrew Sherratt argued, focussing on tombs, that building on a megalithic scale stemmed from the advance of farm-based economies in Europe.120 The change from hunting and gathering practices to cereal production tied people to place and required a sizeable workforce committed to tending and harvesting crops. The best chance of success was through the continuity of reliable familial units living in village communities. The massive stone monuments built by such groups ‘formed the basis for their symbolic construction of community’.121 Megalithic structures endured and, although modified over time as required, remained as permanent focal points for ritual activities, gaining historical significance with every passing year and embodying social memory and ancestral connections to the land. Such arguments seem equally valid for the megalithic structures in Malta and Gozo, especially those built over settlements already centuries old (Skorba, Kordin III, Ta’ Ħagrat in Malta and Santa Verna in Gozo).122 TECHNICAL ASPECTS OF MEGALITH BUILDING

Quarries, Engineering, Logistics In the broad-brush strokes charting human economic development, Bradley observed that hunter-gatherer communities of Mesolithic Europe lived within the natural world without seeking to modify their environment through farming or monument building. A fundamental shift in thinking – in ­ideology – emerged in the Neolithic period. ‘Power’, ‘control’ and ‘modification’ characterised how people began to shape their world and structure their economy – they were driving forces that were taken to their zenith in Malta during the Late Neolithic period. Only with the discovery of radiocarbon dating did the true magnitude of Malta’s past come into focus.123 Given the great antiquity of the structures, dating back to ca. 3600 BC and remaining in use to ca. 2500 BC, it was no longer possible to maintain a diffusionist viewpoint and look to the monumental architecture of the Aegean, Egypt or Britain for the origins of the Maltese Neolithic culture.124 Some argue for connections

The Culture of the Megalith Builders between Maltese and Sardinian architecture.125 In reality, if the engineering know-how and technology were imported from Sardinia, or from as far away as Egypt or Mesopotamia, as has been suggested, then we could expect signs of adapted or imported architectural styles as well. Others looked to Sicilian tomb architecture (megalithic façades and portal doorways which were cut through slabs of stone) and decorative motifs for connections with the Maltese structures despite the chronological differences.126 Instead, the archipelago’s lobed or apse-shaped rooms, budding one from the other within massive rubble-filled megalithic retaining walls, are unique. Malta’s megaliths appear as a spontaneous and indigenous response to local needs probably derived from a long tradition of rounded-to-ovoid domestic structures (Figure  3.5, nos. 8 and 9).127 Although no large village complexes have as yet been excavated, the Għajnsielem houses in Gozo and Skorba do suggest that rounded to ovoid rooms abutted each other in clusters. The basic polylobate plan appears somewhat clover shaped with additional paired rooms factored into some designs. Two to six rooms built around a central area or open corridor often follow a regular, generally symmetrical layout. In only a few cases do we have rectangular rooms, which do not conform to the basic prehistoric lobed plan. The complexity of the Debdieba (Luqa) site, however, with its chronological span of occupation extending into the Roman period and its subsequent destruction, leave room for doubt that the building was prehistoric.128 The other occurrence of prehistoric rectangular architecture is at Tas-Silg, where enclosed areas were either roofed over or left open to the elements.129 The megaliths were not always static buildings but show signs of experimentation, restructuring and development.130 It is possible that earlier and modest prototype buildings were levelled to make way for new, ambitious constructions.131 More than one complex could be built side by side (Ggantija, Mnajdra) or new ‘wings’ or complexes merged into existing structures (Tarxien, Ħagar Qim).132 The basic plan was, to a degree, distorted in later building phases.Why buildings were substantially altered, usually enlarged, remains to be seen, but this is an aspect that we will revisit later. After excavating six rock-cut tombs at Xemxija (predominantly of the Ggantija phase),133 which have shallow, scooped niches cut around the walls of the chambers, Evans saw these as the prototype, so to speak, of the lobed chambers of the above-ground monuments. He asked: ‘But why should a building designed to be a shrine or temple imitate the form of a tomb? The answer is clear – only if the rites to be celebrated there have to do with the dead.’134 The megalithic buildings of Malta thus became linked to funerary practice, and their builders were seen as practitioners of some form of death cult.135 In the absence of data concerning domestic sites available at the time (this would come in 1966), it was never suggested that the rock-cut graves might themselves have been inspired by lobed, rounded or ovoid house structures


The Archaeology of Malta


(compare Mgarr houses skirting the later megalithic structure; Figure  3.3, no. 14). That the tomb structures represented connections to the living and imitated settlement complexes is just as probable. The practice of modifying existing burial grounds to enlarge the burial complexes, such as in Xemxija tomb 5, is likely to have continued, resulting in the extensive cave complexes such as Ħal Saflieni and the Xagħra Circle, Gozo; in the latter, smaller Zebbug tombs were clearly incorporated into the complex.136 Two modest pottery house models representing apse-shaped rooms with door jambs found at Ħagar Qim in 1910137 were considered to represent the chambers of the megalithic structures, yet there is no air of the megalithic about the models (Figure 3.5, nos. 6 and 7). They do, however, mirror house plans such as those at Kordin III (Figure  3.4, nos. 1–5). The ancients were capable of representing unmistakable monumental structures with impressive façades (Figure 3.15, no. 2). There are stone-cut three-dimensional representations138 and possible depictions of structures in artwork and graffiti as well.139 After dwellings at Skorba, an enduring hamlet, were excavated and published in 1966, the link to funerary architecture should have been challenged. An ancestry for the megalithic buildings can be seen in the house plans of the Red Skorba phase, especially the simple ovoid rooms (whether house or shrine) with substantial rubble-packed walls (see Chapter 2), a building technique perfected for the monuments.140 Although indications of cultural change in the Zebbug phase are noted, they are generally thought not to signify a complete change in population.141 Remnants of a horseshoe-shaped Zebbug phase house (Figure 3.2, no. 2) and the oblong room with curved ends of the Ggantija phase (Figure 3.5, no. 9) at Skorba are, in a sense, divided into two curved or apsed areas by an entrance cut in the middle of one long side. The kidney-shaped tombs at Xemxija also conform to this general plan (Figure 3.7).

Construction Techniques The slightly inward-leaning stance of the individual massive blocks in the megalithic structures (Figure 3.14, no. 1) meant that they simultaneously rested on and also retained the packed-rubble fill between the exterior and interior walls, which are clearly evident in aerial photographs of Ggantija.142 Certainly greater strength was achieved for the outer stone casing wall at Ggantija, where a ‘header and stretcher’ technique was employed in the original structure. Here header blocks were placed through the wall at right angles to the stretcher blocks (Figure 3.14, no. 2). Erosion has occurred since archaeological excavation at the major sites. Seasonal rain has washed out the soil from between the larger cobbles and slabs, and this, as well as other environmental factors, has lead to substantial damage. Subsequently, significant interventions have been undertaken in an effort to halt the decay.143 Ħagar Qim and Mnajdra are now covered by canopies, and Tarxien is slated to follow (Figure 3.15, no. 3).Whereas

The Culture of the Megalith Builders

Figure 3.14.  (1) Rear wall of the Ggantija (Gozo) megalithic structure showing the header and stretcher placement of stone blocks (photograph by C. Sagona); (2) ground plan of Ggantija (after Evans 1971, plan 38a).


Figure 3.15.  (1) Alcove on the outside of the Ħagar Qim megalithic complex with standing pillar and truncated V-shaped stone architectural upright megaliths flanking the small alcove; (2) carving of megalithic building façade on one of the megaliths at Mnajdra; (3) Ħagar Qim monumental building façade with modern protective canopy; (4) interior of a room at Mnajdra (photographs by C. Sagona). 80

The Culture of the Megalith Builders the focus on these complexes has often fallen on the main structures, it should be noted that they frequently formed only a part – albeit the heart – of the overall complex. Compound walls (Skorba), open-courtyard areas (Ggantija, Kordin III [Figure 3.4, no. 7], Mnajdra and Tarxien), detached smaller-lobed and less monumental structures, curvilinear rooms and indistinct, poorly preserved stone walls, some from previous settlements, are recorded. These, too, must have played a role in defining spaces and the functions of the sites at various points in their history. The Misqa Tanks (Tal-Misqa) are possibly associated with the megalithic buildings. Located close to Ħagar Qim and Mnajdra, and perilously close to a recent stone quarry, this site is a significant water catchment facility with several large bell-shaped pits cut into an outcrop of bedrock (Figure  3.16, no. 1).144 In the absence of definitive archaeological deposits (removed long ago), it is difficult to link these facilities to the megalithic building period, but their proximity to the monuments is suggestive. It cannot be discounted that a Bronze Age site in the region was destroyed by quarrying activity, but Trump draws an interesting parallel to a similar cistern in the Ħal Saflieni underground burial complex (Figure 3.19, no. 3).145 Furthermore, four of the openings are large and oblong in shape, unlike Bronze Age bell-shaped pits (see Chapter 5).146 Numerous additional rock cuttings scar the surface: some are ancient channels that direct water to the openings, others are possibly petroglyphs (Figure  3.16, nos. 2 and 8)147 and still others are small pits that might be postholes or cup marks clearly clustering around the pit openings (Figure 3.16, no. 7).148 Structures built over the top of the pits seem entirely possible, whether winches to haul water, roofed shelters or constructions with some other purpose. No clear plan of the site has been published previously.149 Both the harder Coralline and softer Globigerina limestone were used for building the megalithic monuments.150 At sites such as Ħagar Qim and Mnajdra, evidence of nearby ancient, shallow quarries indicates that the blocks did not travel far to their destinations; proximity to stone resources may have been a key factor in the choice of site.151 It seems that outcropping bedrock which showed signs of natural horizontal fracturing may have been targeted for quarrying in the case of Coralline limestone and the simple technologies of splitting with wedges and leverage employed to cut and move the blocks.152 Using similar technology, Globigerina can be split from the parent stone, which tends to fracture along regular, linear paths from the point of pressure or impact.153 Abundant stone compensated for a lack of forests and precious soil deposits in the archipelago which could have been used for timber and mud brick.The persistent design suggests that there was an underlying architectural template, probably governed by cultural concepts of public and private spaces, discussed in due course. Stone balls (comparable to ten-pin bowling balls) and large notches cut into the feet of standing megalithic stones (e.g., Ħagar Qim, Tarxien) played a role in raising blocks onto their edges and pivoting upright stone blocks into



The Archaeology of Malta

Figure 3.16.  Features at the Misqa Tanks located near Ħagar Qim and Mnajdra: (1) plan of the site (C. Sagona):  (A)  a square and shallow grape-pressing floor and adjacent round collection pan of likely Phoenician-Punic date; (B) possible furrows or channels; (C) petroglyph of uncertain age; (D and E) grooves cut into the surface of the

The Culture of the Megalith Builders position. In the absence of abundant timber, it has been argued that the freshly quarried slabs glided along the balls as they were dragged from the source to the building site.154 Such a device possibly worked well on the barren bedrock surfaces on which structures were erected,155 but it is less easy to envisage if the blocks were hauled over pitted rock surfaces and softer field soils. Stone balls have been found in quantities at Tarxien, for instance, and others still remain wedged under the feet of standing megaliths.156 Engineering technologies including ramps, ropes, rollers, pulleys, levers and pivoting all have been suggested for elevation of the megaliths and for the gradual placement of upper courses of comparatively smaller blocks.157 Smoothed mallets and hammers made of the harder Coralline limestone would have been suitable to work the softer Globigerina, and such tools have been recovered from, for example, Ħal Saflieni and Tarxien.158 Essentially, the harder Coralline stone was used in its raw state, shaped but not dressed to any great degree, whereas Globigerina lent itself to grinding, carving and cutting into remarkably refined stone joints and dressed and decorated surfaces.159 Plaster and red ochre paint may have given the monuments a more uniform appearance, or perhaps they were enhanced with culturally significant motifs. Traces of red ochre on plaster were reported at Ggantija and Tarxien, as well as on wall plaster and the ‘heavily ochred’ flooring of one smaller room of Ggantija date immediately west of the megalithic structure at Skorba, which was designated the ‘Red Room’ for this reason.160 A few small models suggest that the apses were roofed over, but it is not certain that the entire structures were covered (Figure 3.15, no. 2).161 Surviving upper walls at Mnajdra with progressively inward-stacked stone blocks hint at a partial corbelled upper wall line to the roof (Figure 3.15, no. 4). Similarly, the carved façade with quasi-roofed area cut into the ceiling of the chamber in the underground Ħal Saflieni complex may mimic the partly corbelled interior walls of the above-ground monuments. Contemporary representations of the roof suggest that some kind of individual beams were employed162; whether these were timber planks, stone slabs, mud brick, straw bundles, matting, hides, cloth or combinations of these materials has yet to be proved. Even though Daniel Clark has provided some interesting statistical models for aspects of the engineering and construction of the Ggantija buildings, the archaeological evidence for roofing material which might support his hypotheses is wanting. Figure 3.16. (cont.) bedrock, which probably functioned as drainage channels into natural, shallow hollows; (F) deep channels cut to direct water into and around the cisterns; (2) grooves cut into the bedrock at point D; (3 and 4) slabs covering one of the cisterns showing the interlocking rebate edges on the long sides; (5) detail of central cistern with deep, inter-linking channels; (6) detail of one stone slab showing manufacturing tool marks; (7) view looking south-west across the site; (8) linear designs etched into the rock surface at E (photographs by C. Sagona).


The Archaeology of Malta


His estimates of timber (384 trees), wattle (384 m2) and daub (75 tons) and torba surfacing (38 tons) are for the southern temple alone. Timber and stone are likely to have been salvaged from an abandoned building in antiquity, whereas remnants of torba or wattle and daub probably would be left where they fell. Such material, however, would decay over time but might survive if burnt.163 Another possibility is that stone beams hidden in plain sight, such as those capping the Misqa Tanks, might be worth considering. The long sides on some of the Misqa slabs have been worked to form a stepped lip into which a corresponding rebate edge on the adjacent slab is fitted (Figure 3.16, nos. 3 and 4); the rebate forms a simple yet effective lock, holding the upper slab in place on a horizontal plane. Tool marks do not appear modern (Figure 3.16, no. 8), nor do they match well-preserved tool marks, for instance, in the Punic cistern at Ta’ Kacca tura (Birzebbuga), close to Borg in-Nadur. The design works well for the cisterns, as it might also have done over the chambers of the monuments. It is possible that the blocks were removed when the buildings were decommissioned and dragged to the cisterns at a later date (see Chapter 7), perhaps when the grape press was cut at the Misqa site during the Punic period or even more recently.164 Other roof beams may have been preserved at Borg in-Nadur, while reused stones at Tas-Silg and the dimensions and squared profile of a menhir on Kercem Road, Gozo, represent further possibilities.165 In any case, intentional removal of the roof would explain the dearth of relevant debris or architectural remnants in the buildings when they were brought to an end.

Functions of the Megalithic Buildings Interpretations concerning the role and function of the megalithic structures are varied. Discussions have assumed a phenomenological perspective,166 a recognition that people existed in, experienced and were part of the world around them and that a complex relationship was formed with the environment. This perspective is particularly evident in the excavation reports of the Xagħra Circle (Gozo), where close interpretation of the landscape, sites and artefacts of past lives has provided inroads into numerous aspects of the human experience, including the sensory (sight, smell, hearing, touch and taste) and emotional (feelings, hopes and aspirations). Extrasensory or heightened senses which might have been triggered during attempts to commune with a perceived spirit or supernatural realm have been considered.167 Increasingly, discussions are less about the monuments as remnant artefacts of a remote, intangible past and more about the people who built them, their motivation and the role of both builder and building within the landscape. Investigations strive to be comprehensive but also empathetic,168 as they take into account proximity to fields, water sources and catchment areas, associated geological resources, position relative to routes across the islands  – to the sea and to ports  – prominence in the landscape and relationship to settlements, burial

The Culture of the Megalith Builders grounds and so on.The possibilities will be surveyed in the discussion that follows, but understanding will always be hampered by the limitations posed by the buildings’ early excavation, the limited scope of the published reports and entrenched interpretations. Public and private spaces, formed within and outside the structures, as well as monumental entrances, thresholds and interior portal doorways, seem to have been particularly significant to the ancient builders. Foundation deposits were located beneath stones at the Tarxien complex, under altars and threshold stones and near doorways.169 From the permanent fitments within the rooms, the contents uncovered through excavation and decorative features surviving in stone (and possibly once painted with red ochre on plaster stuccoed surfaces), some headway can be made towards understanding how the monuments may have functioned in antiquity. Discussions address aspects of the society that built the structures, especially in regard to the social organisation, possible hierarchy and nature of the skilled workforce, and in turn, they invite speculation about population numbers in the islands during the fourth to third millennia.

Spatial Distribution,Territories and Social Structures Arguments for defined social organisation and territorial boundaries in Malta have focussed on the location of megalithic buildings. Renfrew used a Thiessen polygon model to determine distinct regions but took into consideration only some of the major megalithic complexes. By implication, his argument suggested that the sheer scale of the buildings pointed to a form of effective leadership required to orchestrate their construction, perhaps chiefdoms,170 with a hierarchy that was probably based on the inherited status of prominent families. The key individuals were likely to have officiated at ceremonies, organised gatherings and determined future directions of the community. Interaction with other groups on the island probably was managed through community leaders. Renfrew’s was the first realistic evaluation of social organisation, and it set the tone for future research. But there are at least thirty such buildings,171 to which the same model could be applied. If all these buildings were considered, the hypothetical subdivisions in Malta and Gozo would change significantly. Criteria focussed on geographical and environmental factors have resulted in more persuasive outcomes. POPULATION ESTIMATES FOR THE LATE NEOLITHIC IN MALTA

Our understanding of population numbers in Malta and Gozo has not progressed greatly since Colin Renfrew’s calculations in the 1970s. Working towards a definition of the social organisation which underpinned the (continued)



The Archaeology of Malta successful construction of monumental architecture, he estimated that the population for the archipelago was a round 11,000 based on the subdivision of the islands into six territories, each territory defined by a pair of megalithic buildings at its heart and each area home to some 2,000 inhabitants.172 An efficient and organised community, he argued, numbering between 500 and 2,000 people would have been needed for such ambitious constructions works. Using a basic occupation density configuration of 2 ha of land per person and 60 per cent of the available 4,020 ha capable of sustaining crops for Gozo alone, Daniel Clark estimated that the population would have comprised around 1,407 people in the Late Neolithic. His aim, like Renfrew, was to estimate how many able-bodied people and how many hours of toil would be needed to build the structures. It was estimated that over the course of fifty-five days 282 workers were able to construct the southern complex at Ggantija.173 Some fundamental data need to be available for realistic population numbers to be achieved. How many villages dotted the countryside, the scale of settlements, average size of the dwellings, environmental conditions, food resources and daily discards are just some of the factors that need to be considered, and most of these aspects are under-represented in the archaeological record for Malta.174 Progress can be made by drawing on ethnographic evidence of similar environments and comparable economic practices. Even recent studies of burial numbers and mortality rates within discrete locations such as the Xagħra Circle, Gozo, are thought to represent only a portion of community members, hindering demographic estimations.175 Attempts at determining the numbers living in the Late Neolithic settlements vary in their estimations, yet most recognise that the ambitious building programmes required a sizeable workforce and a degree of social complexity.176 ‘Prime-of-life’ men and women were needed to move huge stone slabs and cart quantities of rubble wall fill. A viable well-structured economy beyond subsistence level to support and sustain the workforce and their families would have been necessary, as well as a high degree of social cohesion, co-operation and organisation. The workforce must have comprised specialists in quarrying, stonemasonry, sculpture, rope and tool manufacturing and others with engineering skills. Presumably, this labour force, while gainfully employed in the community’s building tasks, would be taken away from their fields and food production, or the communities honed their building programmes to quiet times in the agricultural cycle. How often such an enterprise was undertaken is unclear, but it was surely not a frequent occurrence, which implies that the accumulated knowledge for carrying out such a project must have been safeguarded through many generations.

The Culture of the Megalith Builders Tallies of human burials supply some data concerning the population of the islands. Around 7,000 bodies (an estimate in itself because of the flawed recovery of human remains)177 were deposited in the Ħal Saflieni hypogeum in Malta through the course of its use as a crypt; Zebbug, Saflieni, Ggantija and Tarxien phases are represented in the complex. Presumably the hypogeum served the settlements in the immediate area of the nearby Tarxien megalithic monuments. It has long been considered that other such burial grounds must lay concealed elsewhere in Malta, perhaps commensurate with each of the known megalithic buildings.178 The findings of the excavations at the Xagħra Circle in Gozo (close to the Ggantija complex) suggest that the underground hypogeum contained only a portion of the population and that an alternative method of disposal of the dead operated concurrently within the ancient Gozitan society.179 As early as the 1920s, Margaret Murray recognised the significance of coastal locations.180 Reuben Grima’s recent investigations into the relationship of the megalithic structures to the topography, hydrology, slope of the land (or aspect) and proximity to the sea and to plain boundaries has clearly demonstrated that of all these environmental factors, it is the last two  – coastal ports and fertile plains – which are strong determining factors for the location of the structures.181 Another element defined by Grima, the availability of potable water (largely springs) in the harsh and arid climate of the archipelago, is to be expected, given that the structures were likely to fall within the territories of the settled communities that built them, and those, in turn, could only be viable with a ready water supply. Most indicative of the desired suite of environmental features is the stretch of the south-western coastline between Mnajdra and Kuncizzjoni. Devoid of fertile plains and coastal ports, it has no known megalithic buildings, even though springs dot the area.182 Similar negative evidence is even more pronounced in Gozo. Economic issues would appear to have been a factor in the choice of site. Fertile plains that stretched out before the monuments and ready access to the resources of the sea – that is, natural pathways to viable landing places – are likely to have been significant elements in shaping the social divisions of the islanders.183 The megalithic buildings were perched on the fringes of these zones,‘the liminal areas between the plains and the sea’.184 Cultural choice over how land was used, spacing between one community and the next, proximity to resources and competitive interactions which these may have generated might have played a part.185 Such consistent findings also imply that economic strategies and cultural practices were reasonably uniform throughout the islands for at least a millennium. Underlying the buildings’ locations at a cultural level may be issues of territoriality, boundaries between communities. The structures represented simultaneously gateways for and symbolic (or perhaps even physical) barriers against strangers. Megalithic structures that are close to the shores of Malta may have


The Archaeology of Malta


played a role as territorial markers. As Richard Bradley argued for such constructions:  ‘The new idiom was concerned with power. Monuments were constructed to dominate the landscape and to withstand the process of natural decay.’186 Any mariner approaching Malta from the only inlet on the south-west coast, at the foot of Magħlaq Fault, for instance, would have climbed to the top of the steep scarp to witness the spectacle of Mnajdra with Ħagar Qim higher still and an uncertain welcome by the inhabitants. Monumental structures in view of the coasts, ports and harbours have a long history and an enduring place in social memory; take, for example, the Moai of Easter Island, the Colossus of Rhodes, the Lighthouse of Alexandria or the Statue of Liberty. Placed on the periphery of fertile lands but on routes to the landing places on the coast might indicate that the monumental buildings functioned as landmarks for the island nations, as part of a network of territorial boundaries within the islands and entry points for newcomers to their shores. Within larger landmasses, for example, mainland Italy, population divisions between regions can be distinguished through cultural remains such as pottery.187 Some distinction in pottery has been discerned at Pellegrin, but there is considerable scope to investigate regional variations through a re-examination of past finds, which, on the whole, suffer from limited publication. Of the thirty or so monuments, only a few appear to align with astronomical features. The most impressive is Mnajdra, which catches the rays of the rising sun at the spring and autumnal equinox (21 March and 23 September, respectively). Ħagar Qim might have a lunar perspective. Rows of carved dots on upright stone blocks at Mnajdra arguably track the passage of stars, and a block carved with asterisk-like stars and crescents from Tal-Qadi has been linked to ancient astronomical observations.188 Combined axes of the known monuments tend to lie within a south-west to south-east arch.189 In reality, there appear to be a number of site-specific factors which may have come into play when the orientation of the buildings was considered. One study suggested that the monuments were oriented towards ancestral origins deep within the social memory of the population, but the variability in alignments is not so consistent as to support this notion.190 USE OF SPACE [Architecture] is a form of communication; it reflects the subtlety of social interaction among people. As a means to delay visual and physical access to the more personal areas, transitions also frequently bear a spiritual function: they mark the various stages of life and the passage from one world to another – neutral and sacred; exterior and interior; communal and familial; living and dead.191

Ethnographic studies of African complexes in the Burkina Faso (formerly Upper Volta) region (Gurunsi communities) indicate that village compounds

The Culture of the Megalith Builders have both domestic and spiritually significant spaces.192 These settlements offer some useful heuristic analogies concerning the use of space which can be applied to the megalithic structures in Malta. Encoded into the architecture at one level is a sense of familial or community territory. Careful manipulation of space is evident in the arrangements of rooms and courtyards to form corridors as well as boundaries controlling access or allowing free passage. As permanent structures, the complexes also follow a recognised, predictable and respected set of ground plans; the architecture embodies rules about how space is managed and who has access to it. In ancient contexts, Ian Hodder proposed a similar manipulation of space at Neolithic Çatalhöyük in Turkey.193 Habitual movement through domestic spaces was contingent on the subjective distribution of cooking, sleeping, gender-specific, ancestral and ritual areas. Architecture can convey a sense of defence, achieved by creating impenetrable zones, hidden interior areas and barriers between private and public spaces. Entrances are recognised as points of transition, thresholds between one zone and another, especially demarcated in the megalithic buildings in Malta, where rooms are accessed through portal doorways cut through solid-stone slabs. At the same time, from within the shadows of a building’s interior, surveillance of the open outer areas can take place. Privacy in African complexes can be achieved through deep shadows within building interiors; low levels of lighting should not be used as a criterion to assume limited function in the megalithic structures in Malta – it may have been a desired effect.194 Weather conditions and ventilation also governed which areas of the African houses were used. All these aspects and underlying guidelines regarding use of space could have been at play in the Maltese islands’ ancient structures.195

Exterior Space Associated with the Monuments Though there are those who focus on the private spaces within the Maltese structures, equally significant are the monumental exteriors. Massive door lintels, layers of blocks stacked metres high over several courses and substantial stone benches which acted as seating formed an impressive façade to the structures. Architecturally, the façades appear as backdrops to large open areas in front of the structures, where, in all likelihood, communal activities were carried out. Effort was expended on paving the courtyard in front of some structures (Santa Verna in Gozo, Kordin III and Mnajdra).196 These formal outer areas served as buffer zones between the outside and inner space. Roof areas and the wide wall tops of the monuments were part of the exterior space, and in the case of Tarxien, where a stone staircase on the interior of the building climbs to roof level, they appear to have been used. The tops of the buildings may have been important for observation over land, sea and the immediate courtyard area. It is likely that the carefully constructed exterior courtyards served first as a place where formal interactions could take place within the


The Archaeology of Malta


community and second as a mingling zone designed to control interactions between local visitors and non-islanders, but always favouring the ‘resident’ cultural group that owned the building. Against the monumental backdrop, economic transactions, exchange of goods, community meetings, decision making and celebrations could have been carried out. There is the notion that such occasions were public but also sanctioned by figures of authority or by a divine entity the buildings may have represented. Megalithic structures, associated burial grounds and the immediate environment were equally important in the ritual life of the ancient community. Such aspects were recognised by Caroline Malone and Simon Stoddart for the Ggantija–Xagħra Circle region. In their scheme, the monuments and the landscape figured in elaborate funerary rituals.197 Very little work has been done on the identification of isolated standing menhirs or perhaps lines of stones, which may have formed part of the prehistoric constructed landscape. Some standing stones survive, while others have been discounted as menhirs because they may have been part of destroyed megalithic buildings (Appendix B, List 13).198

Interior Space within the Monuments Once designated as temples, the features within the megalithic structures of Malta were assigned ritual functions: pillars became baetyls, tables were seen as altar stones, pierced stone blocks were equated with oracle holes in the manner of much later Greek temples,199 inner private rooms were the holy of holies and small rock-cut holes in the walls became tethering holes for sacrificial animals rather than fitments for tying door screens or curtains.200 If we accept that the monuments may have been linked to economic locations – on the periphery of viable arable land and with access to marine resources and landing places – then, by implication, they played an economic role.201 Not surprisingly, in and around the monuments were numerous grinding stones, both fixed and portable, for milling grain. In recent years, their function usually was ascribed to libation, but shortly after excavation, Zammit designated them as quern areas. Many show extensive signs of considerable wear from grinding; perhaps most marked is the long trough across a threshold in Kordin III with seven well-worn and deep grinding surfaces. Of this installation, Evans wrote: ‘Ashby says that a stone rubber was found in one of the compartments of this trough, and since the surfaces of the smaller compartments all showed traces of rubbing, he thought it possible that the trough was used for grinding grain.’202 But because of the overriding view that the buildings were temples, this common-sense interpretation was quickly dismissed. Pottery may have served as offerings in a temple, but in the light of well-worn food-preparation facilities, dining (whether ritual or not) also may have been a component of life around the monuments. Bone and shell discards, in the absence of well-defined contexts, may be used equally as evidence of sacred

The Culture of the Megalith Builders or profane activities.203 A  governing hierarchy may have facilitated payment for work in kind, through tribute or gifts of food surpluses, and the megalithic structures may have served as distribution points in such exchanges.204 Like the African compounds of Gurunsi communities, where certain areas were gender specific and other areas dedicated to food preparation, sleeping, storage and veneration of the divine, rooms in the Maltese structures may have served various functions.205 While not discounting that the megalithic buildings of the islands served a sacred role, not all rooms have overt signs of ritual or sacred function. Lobed chambers approached by bending down before passing through portal door slabs, as well as other areas that could be closed off by cordons and barriers formed by posts placed horizontally in rock-cut sockets in the doorjambs, point to controlled access to private spaces.206 Even the changing floor levels within the monument may have manipulated how space was used and controlled the passage of people through the buildings.207 The possibility that the megalithic buildings were habitations has been mooted in the past, but this notion is not generally accepted,208 although it has been suggested for lesser buildings to the north and east of the Ħagar Qim monumental structure,209 at Mnajdra,210 and around other complexes that may have served as housing. Nonetheless, there was ample evidence of animal and fish remains in the megalithic buildings; whether this constituted refuse from food consumption or remnants of offerings has not been scientifically determined. Monumentality and decorative features need not preclude the possibility that such structures were domestic, as has been argued for the even older megalithic structure at Göbekli Tepe in Turkey.211 In other contexts, the presence of hearths, spindle whorls, grinding stones, lithic tools and quantities of pottery would be taken as signs of domestic life, but in the context of the monuments, the function of such finds is usually seen as sacred.212 At the very least, some form of centralised economic activity, if not carried out at the sites then funnelling into the complexes from the hinterlands, seems possible. If house plans may be linked with tomb plans, as discussed earlier, then it should not be surprising that underground burial complexes mirrored above-ground megalithic complexes, especially considering Malone and Stoddart’s argument that burials in the Xagħra Circle complex might reflect only a portion of Neolithic society in Gozo.213 Ultimately, whether house or more complex structure (shrine or temple), both can have signs of ritual and mundane activities. Put simply, if the tomb mimics domestic architecture, then the elaborate hypogea might equally reflect a grander form of dwelling in the megalithic buildings.

Sacred Spaces Although other functions may have been associated with the megalithic monuments, they nonetheless exhibit aspects that are attributed by most



The Archaeology of Malta scholars to cult and ritual function.214 It is clear that the monuments occupy attention-focussing locations on open ground in prominent positions within the landscape and that they follow a specialised architectural plan, built on a monumental scale. Fixtures beyond what might be expected for average domestic use include very large and conspicuously decorated benches and repeated iconography seen in the decorative panels and in the statuary found in certain locations. Monumental thresholds, portal doorways, obscured rooms and hallways are generally interpreted as forming boundaries between the mundane and sacred realms. A  larger-than-life-size corpulent statue in Tarxien and the numerous full-bodied small statues from Ħagar Qim are described as deities. A fundamental disadvantage in Malta is the very limited understanding we have of the nature of more modest domestic structures for all phases and the daily activities that were carried out in them as a means by which to fully judge the monumental buildings and how they were used. Comparisons between domestic and sacred contexts, for example, underlie Verhoeven’s properties for detecting ritual in the archaeological record – referred to as ‘ritual framing’ – where the special qualities of place, architecture, objects, display and actions in ritual stand apart from the domestic and mundane.215 Notwithstanding this, the iconography in the megalithic buildings and a belief in a cosmology which encompassed life and death have long been suggested through the evidence from underground complexes, especially at Ħal Saflieni, where some chambers are cut to resemble the above-ground monuments. These aspects will be discussed in the pages that follow. The difficulty in identifying prehistoric cult within the archaeological record is widely acknowledged in the literature. What survives is only a small and durable portion of ancient cultural and religious expression. Questions raised concerning belief systems in Malta by Simon Stoddard and his colleagues are germane to the issue: ‘Can the habitual problems for the reconstruction of ritual be overcome in a prehistoric context with elaborate material culture?’ and ‘Can the repetitive structure and associated iconography be read without the aid of written texts?’216 The study of living traditional cultures can provide some analogies to possible ancient ways of life, hinting at the potential complexity of social behaviour and providing a means by which to move beyond the profound silence of cultural artefacts in a non-literate, ancient community. Of late, imaginative forays into the sensual, emotive and subjective have attempted to restore the human quality of ancient lives, but these accounts are destined to portray the phenomenologist’s own empathetic sensual, emotive and subjective responses to the archaeological record. The advantage of drawing on anthropological parallels over a phenomenological approach is the actual documentation of human behaviour rather than a substantially imagined one.217

The Culture of the Megalith Builders Unless there is hard archaeological evidence that even hints at an ancient practice, a phenomenological narrative runs the risk of weaving untenable threads into the fabric of academic discussion. One example is Christopher Tilley’s blending of ochre (of which there is ample proof) and honey use (for which there is no evidence whatsoever) into the ancient Neolithic belief systems in Malta. Linked together in his argument are Malta’s ‘honey-coloured’ stone and the stone’s decayed ‘honey-combed’ surfaces, but there is no indication that the ancients made any association between stone and honey. Instead, there is evidence that the megalithic slabs may have been plastered and painted with red designs. Malta’s ancient name, Melita, which may derive from the Greek word for honey, was a word applied to Malta centuries after the Neolithic period. To complicate the equation further, Malta was a Phoenician-Punic stronghold in the first millennium, and these settlers may have known the island by the Punic name, ‘nn.218 In any case, a connection between honey production and Malta in the classical tradition has now been discounted.219 Ultimately, one must bridge a chronological gulf, discounting significant colonising events in between and culturally distinct phases, to accept such a link.220 Overall, the megalithic buildings might have embodied layers of meaning matched by a range of functions from the mundane to the sacred.221 Edward Banning has challenged the archaeological community to rethink the role of the megalithic structures at Göbekli Tepe (Turkey).222 He has argued that there might have been significant overlap in the use of space within domestic settings where both ‘sacred and mundane’ rituals could have taken place. Domestic spaces can be richly endowed with symbolism while appearing unremarkable, and conversely, we should not overlook signs of the mundane in buildings with remarkable qualities. It cannot be assumed that clear distinctions were made by ancient societies between everyday life and spiritual belief in their use of buildings, their manufactures or their thoughts:  ‘the point is not that specialized shrines are incompatible with domestic ritual but that evidence for ritual or conspicuous symbolism does not automatically imply specialized temples.’223 The same reasoning can be applied to the megalithic buildings in Malta. BELIEF SYSTEMS AND RITUAL PRACTICE IN NEOLITHIC MALTA

Possible Shades of Shamanism There are bold mushroom hunters and there are old mushroom hunters, but there are no old bold mushroom hunters [cautionary proverb]

After the discovery of an enigmatic group of nine plaque or peg figures in the Xagħra Circle in Gozo, the possibility that the ancient inhabitants of Malta and Gozo practiced a form of shamanism has come to the fore.224 Grouped



The Archaeology of Malta together at the time of discovery and possibly once held in some form of bag, these small peg figures had sculpted human heads. Three others (one of them animal-headed) are carved on top of variations of stumplike stands (Figure 3.18, nos. 13–21). None of the full-figured obesity often portrayed in contemporary Maltese prehistoric contexts is evident in the group, but two figures wear the familiar pleated skirt (Figure 3.18, nos. 1–3; compare 17 and 18). A new way of interpreting this group’s function seems fitting, although the implications of shamanistic practices have not been fully explored in wider Maltese contexts. Notwithstanding the impact of Mircea Eliade’s classic work on shamanism in 1972 and the plethora of works exploring the possibilities of interpreting the past, Paul Bahn rightly cautions against the wholesale adoption of notions of shamanistic practices and acceptance of the universality of trance states, especially overly structured interpretations such as David Lewis-Williams and Thomas Dowson’s three stages of trance.225 Nonetheless, within the ethnographic accounts of societies that use hallucinogens or other stimulants to enter into trance states can be found ‘incontrovertible evidence of ritual performance, that is, repeated actions’ but not simply connected to the divine.226 In order to fully appreciate the implications of such a comparison for prehistoric Malta, the role of the shaman in living societies, the nature of trance states and how the term ‘shamanism’ is applied to ancient contexts will be considered. ‘Shamanism’ is a term derived from the Artic Tungus region and refers to women and men of the Siberian communities who have special knowledge, experience and capacity to interact with a perceived spirit world. They are acknowledged because of their ability to fall into a trance state, often accompanied by hallucinations at every sensory level, and a capacity to experience visions which they can recount to the wider community and interpret for them. Significantly, neuropsychological studies have argued that because of the common nature of the modern Homo sapiens brain, visual hallucinations seen when individuals are in trance states have common characteristics. Whether the ecstatic person painted the walls in the far reaches of the Lascaux cave in Palaeolithic Europe (ca. 15,000 BC) or lived in recent times, distortions of the senses which occur during trance events can have common traits, shared by humankind.227 Whereas we should not simply overlay Tungus cultural beliefs or modern experiences onto ancient societies, the term ‘shamanism’ in academic parlance has come to be more widely applied as a general term to explain a range of enigmatic cultural artefacts, possibly used by people in their attempts to connect with a perceived supernatural world. Specialised utensils or equipment that has been found in the ancient record and seems to fall outside domestic functionalism can be interpreted as shamanistic trappings. Artistic expression – especially common, recurring patterns in ancient iconography – likewise may have been inspired by altered states of consciousness experienced by the people who participated in such events.228

The Culture of the Megalith Builders Research on living communities presents a plethora of possibilities for considering past cultures. Studies have shown that in Tungus circles the shaman plays an important role in society. They often wear elaborate clothing and carry a range of magical charms and musical instruments, especially drums, as well as other paraphernalia. These individuals have specialised knowledge about medical herbs and psychoactive plants or animal products that help induce an altered state of consciousness. When in a trance state, the shaman is believed to bring about change on a spiritual and physical level. Shamans can gain insights into the future, delve into problems that face an individual or community, obtain blessings from benevolent spirits or, conversely, repel demons, commune with the dead to tap into ancestral knowledge or find lost or troubled souls as a means of healing the sick. They also conduct the necessary rites at funerals that help shepherd the spirit of the deceased into the afterlife.229 Aside from symptoms triggered by pathological conditions such as migraine or epilepsy,230 hallucinations can be induced in a number of ways, used alone or in combinations: rhythmic dance and vigorous movement including shaking of the head, light deprivation or flickering lights and rhythmic sounds such as chanting and drumming.231 The location in which such events take place and the underlying cultural beliefs and expectations of the participants – both the person in the trance and the audience involved – play a part in the outcomes. Hallucinogenic substances drawn from plant and animal species also were employed widely. Concoctions that were ingested varied, but ancient narcotics included the ‘magic mushroom’, often the white-spotted and red-capped Amanita muscaria,232 opiates and hemp, among numerous other psychoactive substances.233 Of possible plant sources, the Joint Pine (Ephedra fragilis), though now rare in Malta, is considered a native plant and is a source of ephedrine.234 Henbane does occur on the island, a plant Sherratt considered was used widely for shamanistic ‘spirit-flight’ in other cultures.235 Less likely is the ergot fungus, which grows on grains and is toxic to people and animals but can induce hallucinations as an adverse effect of ergot poisoning. Fermented alcoholic beverages could have played a role in trance events. A  variety of plants and barley, which was grown in ancient Malta, could have been used for such brews.236 Leaving narcotic substances aside, the monuments in Malta may have provided the necessary conditions for light manipulation and acoustic enhancement, which can both play a role in inducing altered states of consciousness. Rhythmic sounds and chanting as part of ritual practice do have some bearing on the Ħal Saflieni underground complex. A small niche cut in the wall of one chamber (the ceiling of which was painted freely in tendrils, spots and spirals; Figure 3.19, no. 1) will project deeply pitched utterances so that they resonate through the entire complex. The niche itself has three large red ochre spots painted on the interior. In a pilot study into the effects of sound on the human brain, Cook, Pajot and Leuchter237 demonstrated that variations in sound



The Archaeology of Malta frequency could bring about changes in brain activity.They addressed the basic question of whether ancient structures may have been designed to resonate in such a way as to affect the human brain in buildings such as Newgrange in Ireland. Their findings were positive: Activity in the left temporal region was found to be significantly lower at 110 Hz than at other frequencies. Additionally, the pattern of asymmetric activity over the prefrontal cortex shifted from one of higher activity on the left at most frequencies to right-sided dominance at 110 Hz. These findings are compatible with relative deactivation of language centers and a shift in prefrontal activity that may be related to emotional processing.238

The notion that stones in ancient monuments may have been positioned deliberately for acoustic effect has some potential for the complexes in the archipelago.239 In the Xagħra Circle in Gozo, less well preserved than Ħal Saflieni, it has been argued that a large stone bowl (now broken) in the complex may have served as a drum.240 A similar function could be argued for large bowls surviving in the above-ground monuments of Tarxien.241 In addition, outside the entrance at Ħagar Qim and at Tarxien are two obliquely cut large holes that intersect, leaving a bridge of stone – the original surface of the bedrock – over the top (Figure 3.15, no. 3, right foreground). It is possible that this bridge functioned as some form of percussion device or drum. A parallel can be drawn to New Guinea, where earthen drums were made from a bridge of hard-packed earth under which the soil had been dug away forming a hollow; the hard-packed bridge was beaten with the hands.242 The Ħagar Qim cuttings may have been positioned to take advantage of the possible acoustic qualities of the curved monumental façade of the building.

Altered States of Consciousness: Hallucinations and Iconography The key to detecting altered states of consciousness, particularly the visual component, lies in their ancient artistic expression. The open-eyed person in a trance state can see a range of forms that flicker across the line of vision, as if projected before him or her. These images can pulsate, zoom in and out or emerge into sight and then fall obliquely away. Anthropologist Gerardo Reichel-Dolmatoff, who experimented with the drug Banisteriopsis caapi, documented the progress of the event, describing the movement of motifs, ripples of colours that came and went and spinning and undulating patterns that appeared to him.243 The types of images which are commonly experienced include geometrical designs, sometimes open, hatched or cross-hatched, chequer patterns, dots, zigzags, stars, parallel lines and spirals, concentric circles or curves and back-to-back motifs.244 At times, the patterns were symmetrical, or they could transform into other random motifs (Figure 3.17, nos. 6–10).

Figure  3.17. Common hallucinated patterns seen in trance states. Designs from Maltese contexts: (1) spots on a Mnajdra megalith (photograph by A. and C. Sagona); (2)  decoration on a Tarxien bowl (after Evans 1971, fig.  21:7); (3)  spirals decorating the ceiling of room 20 in the Ħal Saflieni complex (after Evans 1971, fig. 14D); 97

The Archaeology of Malta


As the episode progresses, the subject can often try to make sense of the images he or she is seeing. At this point, the patterns will appear to transform into recognisable things: spirals become eyes, wavy lines appear as snakes, birds emerge from geometrical assemblages of triangles and zigzags become mountain ranges. Interpretations are culture specific and influenced by the person’s own life experiences and beliefs.245 The Tungus are more likely to identify bears and wolves common in their daily lives than giraffes or zebras.The recent illicit use of narcotics in the modern era has provided numerous examples of psychedelic imagery in art and song, inspired by drug-induced hallucinations. For the ancients of Malta, where the range of natural fauna was limited, their artwork includes small species of birds, fish and snails, as well as orderly rows of domesticated animals. Spiral and dotted motifs are a prominent element in the megalithic wall decorations, as well as on pottery, and are a common element of hallucinations. Whether these ways of ‘seeing’ followed a predictable three-stage path – a neuro-physical model – for the different types of visions or were random has been debated in recent years.246 Those in a trance episode can have evolving hallucinations, the sensation of cosmic flight, of seeing and passing through a vortex of lattice-like lines to emerge into a bright realm of strange and frightening composite animals. The shaman may have a sense of becoming the animals he or she sees or of morphing into the surrounding objects. The composite creatures of the Palaeolithic caves in Europe that appear as animal-headed people, for instance, and are discussed in terms of shamanistic hallucinations, are not present in Malta.247 Instead, a sense or morphing shapes is conveyed by other figurative representations.248 Examples include a polished human head emerging from a natural lump of chalcedony,249 bird necks and heads attached to featureless bodies250 and a human-headed snail.251 Seen in this light, the group of nine peg figures found at the Xagħra Circle complex takes on a deeper significance. It has been viewed as a collection of complete and semi-completed artefacts (Figure 3.18, nos. 13–18).252 Alternatively,

Figure 3.17. (cont.) (4)  radial pattern on a shallow-handled bowl found in a Nadur tomb, Rabat (after Evans 1971, pp. 107–8, fig. 12:3); (5) lid with back-to-back design from Ħal Saflieni (after Evans 1971, fig. 5:14); (6–10) central boxed pattern categories compiled through ethnographic research into trance states and modern experimentation with hallucinogenic substances (after Reichel-Dolmatoff 1972). Corresponding designs found in Malta: (11) painted spirals and spots on the ceiling in room 18 of the Ħal Saflieni underground complex (after Evans 1971, fig. 14C); (12) geometrical patterns from a Kordin III pot (after Evans 1971, fig. 9:3); (13) back-to-back spirals and spotted relief from Tarxien (photograph by A. and C. Sagona); (14) lid from Ħal Saflieni with ‘sunburst’ design (after Evans 1971, fig. 5:15); (15) ‘comet’ d­ ecoration on a Ggantija pot (after Evans 1971, fig. 28:4).

The Culture of the Megalith Builders

Figure  3.18. Neolithic figurines:  (1)  figure of a person wearing a fringed skirt lying on a bed from Ħal Saflieni (H: 6.8 cm; L: 12 cm; W: 6.8 cm, inv. no. S/P1000); (2)  two figures seated on a bed from the Xagħra Circle complex with detail of


The Archaeology of Malta


if these objects are all finished products, the depiction of abstract heads emerging from the partly carved stone matrix (Figure 3.18, nos. 13 and 14) or the more realistic heads on some figures (Figure 3.18, nos. 15–18), as well as an animal head with tusks – not quite boar, not quite dog (Figure 3.18, no. 21) – may represent a sense of metamorphosis (as experienced in trance states). Additional peg figures from the Xagħra Circle which were carved from cow tarsals have stylised human heads with facial features enhanced with red ochre and black pigment.253 A desire to depict spiritual metamorphosis also might explain why the contemporary corpulent statues from the Ħagar Qim megalithic building can be headless, not always through damage but with carved sockets to take a single or perhaps a series of different heads (Figure 3.18, no. 7).254 Whatever the motivation, removable figurine heads do point to a kind of dynamism for this type of artefact that is not often seen in the archaeological record. These figurines (and presumably who or whatever they represented) could change – and change dramatically. It is tempting to see in the range of human and animal figures among the Xagħra Circle figurative art the possible range of replacement heads used in the megalithic headless statues. Decorative features on the monuments have been the subject of numerous interpretations. Arthur Evans linked the Maltese prehistoric spirals with a ‘pan-European’ tree cult and argued that they had evolved through Mycenaean influence.255 The first radiocarbon determinations for Malta, well and truly pre-dating the Aegean examples, quickly quashed this notion.256 Many designs on pottery and the red-painted curvilinear and spotted ceiling of the Ħal Saflieni complex are redolent of hallucinated patterns. There is a

Figure 3.18. (cont.) the red ochre tendrils on top of the bed – legs, fingers and lips are also daubed in ochre (H: 12.8 cm; inv. no. SF742); (3) remains of a statue from Tarxien (H: 18 cm; inv. no. T/S28) with three registers, (a) the largely damaged main figure of a person or deity, (b) two figures preserved in the fringe of the skirt, and (c) a gathering of people carved under the skirt; the statue stands on a bedlike base (H:  18  cm; inv. no. T/S28); (4)  seated figure appearing to hold a disc to the body from the Xagħra Circle (H: ca. 7.5 cm; inv. no. SF775) (after Malone et al. 2009, fig. 10.68); (5) seated figure with arms on the lap revealing a scored torso (H: ca. 7.5 cm; inv. no. SF712) (after Malone et al. 2009, fig. 10.67); (6) statue from Ħal Saflieni (H: 38.9 cm; inv. no. S/S40); (7) statue from Ħagar Qim (H: 48.6 cm; inv. no. Q/S21); (8) head from Ħal Saflieni (H: 9.1 cm; inv. no. S/S38); (9) pebble amulet head from Tarxien carved with a human face (H: 4.5 cm; inv. no. T/S1); (10) similar head to no. 11, from Mnajdra (H: 8.5 cm; inv. no. Mn/P1000); (11) head from a figurine found in Tarxien (H: 7 cm); (12) head with bobbed hair style from Ħal Saflieni (H: 10.9 cm; inv. no. S/S39); (13–21) peg figures from the Xagħra Circle (after Malone et  al. 2009, figs. 10.60–10.63; inv. Nos. SF784/6 [13], SF784/4 [14], SF784/7 [15], SF784/2 [16], SF784/3 [17], SF784/9 [18], SF784/5 [19], SF784/8 [20], SF784/1 [21]); (1, 3, 6–12: photographs by A. and C. Sagona).

The Culture of the Megalith Builders fundamental difference in the artwork of the underground site compared to the above-ground monuments. Underground the circles, spirals, tendrils and dots on the walls and ceiling have a more chaotic appearance (Figure 3.17, nos. 3 and 11), whereas in the megalithic buildings the designs are controlled, regular and often repeating (Figure 3.17, no. 13). One reconstructed pottery bowl from Ħal Saflieni is worth noting in this regard, as it originally had eleven or twelve naturalistic animals  – bovines, perhaps goats with backward-curving horns and a central pig  – on one side and on the other a highly stylised interconnected row of hatched animals ‘hidden’ within a complex geometric pattern (Figure 3.20, nos. 1–3). The ability of the potter to move between a natural and a highly stylised representation of animals is clearly remarkable, but the dramatically different approaches apparent in the two designs convey a sense of a shifting from the naturalism of the real world to the abstract, perhaps motivated by a desire to portray the altered perception of trance states. Bonorva tomb 7 in Sardinia (dated to the Ozieri culture, 3800–2900 BC) offers some points of comparison. Despite the distance between the islands, these sites share common determinants; both probably witnessed funerary rituals carried out in underground chambers and a desire to reflect cultural settings which the deceased once experienced in life. Rather than direct cultural borrowing, however, the artwork may have been inspired by similar experiences (a shared human response to stimulus) resulting in altered states of consciousness. In each case, the above-ground architecture that the tomb mirrors is different: a house in Sardinia and a megalithic façade in Malta. In Sardinia, red ochre has been used to enhance a rock-cut tomb carved to resemble the interior of a house with rooms, ceiling beams and interior fixtures. The walls have been generously painted in large red ochre spirals, and in one area, a black-and-white chequer board adorns the wall. This representation of above-ground architecture, the paintwork (spirals and chequer board) and a funerary function find common ground in the Ħal Saflieni complex (Figure 3.19, no. 1). Statuary in Malta and Gozo from this period tends to represent quite passive postures, for example, some figures are standing with arms folded or hanging at the sides (Figure 3.18, nos. 6 and 7), others are seated (Figure 3.18, nos. 2, 4 and 5) and another lies prone on a bed as if asleep (Figure 3.18, no. 1), which also could be interpreted as reflecting a trance state.257 Another statue has a pair of figures sitting on a bed with double lines curling under their feet and over the top of the bed (Figure 3.18, no. 2); the lines and the legs have been enhanced with red ochre. One other figure on a bed from Ħal Saflieni is notable. It does not appear to be human but resembles a fish in shape, and it might be seen as another representation of spiritual metamorphosis.258 Surviving figurine heads found in contemporary contexts often display a similar tilt backwards with expressionless faces.These also might represent trance states looking beyond the mortal realm (Figure 3.18, nos. 2, 5, 8 and 11),259 a stance which has been documented in anthropological studies of such events.260 In design, these


Figure 3.19.  (1) Plan of the three levels at Ħal Saflieni indicating the scope for light to penetrate into the second level from the surface; arrows show points of access from one level to the next (after Cilia 2009, pp. 79, 86); (2) the original placements of stones in the circle around the underground burial chambers in the Xagħra Circle, Gozo (after the artist Jean Houel); darkened points show corresponding surviving stones; (3) plan of the burial chambers in the Xagħra Circle (2 and 3: after Malone et al. 2009, p. 9). 102

The Culture of the Megalith Builders figurines do not appear to be a development from the stylised examples found at Skorba dating to the Red Skorba phase (see Chapter 2). Nor are they akin to a figurine of Ggantija date, from the same site, which is a stumplike object that in profile has a rounded belly and steatopygpus buttocks; both body features have small holes pierced in them (Figure 3.5, no. 3).

Life-Cycle Symbolism Despite Paul Bahn’s reservations, shamanistic practices gleaned from anthropological studies persist as a way of interpreting ancient cultural artefacts. They do so because the ‘physical constants’ of trance states  – brain function, neuropsychological response to stimulus and commonality of sensory reactions  – represent a viable tangible means of interpreting human cognitive processes, otherwise lost in a pre-literate past. Together the ancient and recurring expression through iconography, scientific analyses which identify hallucinogenic substances in archaeological contexts and specialised settings that may have triggered altered states of consciousness can point to past behaviours. Nonetheless, whether or not trance played any part in the ritual of the ancient people of Malta, what their megalithic buildings, manufactures and artistic expression meant in their spiritual lives  – the cultural overlay  – has invited broad-ranging interpretations. It may not be possible to condense here the many viewpoints present in the literature concerning what the megalith builders believed, but there is certainly a recurring theme concerning fertility, fecundity and renewal. Running concurrently with and enmeshed in the exploration of landscape and ancient societies are debates concerning the human body (body theory) as, at one level, a vehicle for social expression.261 At another level, the physical remains that bear the evidence of lives lived in a material environment also tell their story. Malta has provided tantalising indications of both. The figurative evidence in Malta has featured in the wider debates concerning gender in the ancient world. Figurines found in the megalithic buildings and the underground funerary complexes are usually classed as idols representing a deity. Despite the often sexually ambiguous nature of many figurines (e.g., Figure 3.18, nos. 2 and 5–7),262 they are frequently referred to as female in the literature.263 The focus has fallen on the few with obvious female attributes (breasts or a pubic triangle) as representing a ‘mother goddess’ connected to a fertility cult. Particularly notable among the examples is the full-breasted ‘Sleeping Lady’ found in the Ħal Saflieni burial complex, which is a masterpiece of lifelike modelling (Figure 3.18, no. 1).264 Marija Gimbutas went further, likening the early and basic megalithic ground plan to eggs and hence to the womb and later polylobed structures to the body shape of the goddess traced out as a building ground plan (Figure 3.14, no. 2).265 Faces on pottery, thought to represent the goddess in European contexts, are rare in Malta, and these tend


The Archaeology of Malta


to be stylised, human, rather than male or female.266 Four decades ago, however, Andrew Fleming questioned whether belief in an earth-mother goddess was as ubiquitous as believed, especially in regard to the megalithic Passage Graves of Europe.267 It is a continuing debate that attracts strong opinions, but taken at face value, the precise sex of the figures may not be the focus. Instead, the full-bodied figures in Malta convey a sense of abundance rather than hunger, economic success rather than failure and excess rather than scarcity. The docile posture of the figurines conveys a passive and calm air. There are no depictions of smiting gods, sacrifice, indications of warfare or weaponry.268 Rarely depicted is interaction between human figures or between human and animal figures.269 There is a dearth of explicit representations of genitals,270 intimacy, pregnancy,271 childbirth or children.272 Enigmatic twists of clay have been interpreted as foetuses, but it is more likely that they represent folded legs in the fashion of seated adult statues.273 Separate body parts were modelled: torsos from Tarxien and Mnajdra,274 as well as unambiguous legs with well-defined feet.275 In addition, there are figurines which suggest medical conditions and perhaps methods of treatment or healing rituals, which form another class of figurines found in the megalithic monuments and in funerary contexts.276 In nearly every instance, they are rustic and less formal in appearance, lacking the skill of production of the obese statues, which are often carved from stone. Equally poor in production were occasional examples of animal figurines.277

Upright Figures: Phallic Symbols, Pillars and Baetyls Stone, bone or pottery pillars found in ancient contexts on the islands are interpreted as phallic symbols. Those which terminate in a defined tip are evocative but assume a less sexual connotation when considered beside other figurines with anthropomorphic features278; they are not unlike the animal tarsal figures with unambiguous human heads from the Xagħra Circle.279 Phallic symbols or stylised pillar-like representations of human forms are imbued with a maleness that is otherwise absent among the figurative artefacts; no beards are shown on human faces, and figurines do not display male genitals. Though there are a number of free-standing single pillars,280 there are well-cut and finished carvings which display two or three pillars standing close together within a niche or cut in such a way as to fit into a niche. They were originally coated with red ochre.281 Many centuries later in Punic contexts (totally independent of Maltese prehistoric development), as we will see, double and triple pillars invite speculation about divine family units, parents and their offspring, with implications of fertility and reproduction, an interpretation that may have some relevance in these prehistoric settings. It should be noted that in an exterior alcove of the Ħagar Qim megalithic building is a standing stone behind a truncated triangular block that might

The Culture of the Megalith Builders indicate a female connection – the niche is likely to have formed a focal point of devotion, perhaps as an altar (Figure 3.15, no. 1). The large standing stones on either side of the central feature, though serving as architectural elements, may have some bearing on the overall, somewhat tripartite composition of the alcove and pillar.Three pillars also were found in a niche in the Borg in-Nadur complex, although it has been suggested that these might date to the Bronze Age occupation well attested across the site.282 Underlying the representation of morphing figures, whether inspired by trance states or not, might be an expression of people’s desire to manipulate the world in which they lived and to intervene in natural process. Rows of domesticated animals, fish, birds and tall plants grown in pots carved on stones within the monuments may reflect a preoccupation with fertility, with the cultivated and the natural environment.283

Individuality: Body Shape and Adornment Distinctions between the full-bodied figurines are subtle but nonetheless evocative. Andrew Townsend employed the term ‘ephebism’ to encompass the symbolic and non-natural depictions of prime-of-life people, both male and female, which may have been the underlying concept guiding prehistoric ­artists.284 Deeply pleated, full skirts are represented on some figures (Figure 3.18, nos. 1–3), while others are shown naked (Figure 3.18, nos. 4–6) or, in the case of a few, partly clothed in a flimsy loincloth covering their buttocks. How the hands and arms are positioned and whether the figures are standing, seated or lying down may be significant. Special treatment is afforded to the hair, which seems to indicate characteristics of an individual: it might be shaved at the front, increasing the sense of forehead (Figure  3.18, no.  1), plaited or worn in other short styles (Figure 3.18, nos. 2, 8 and 12). One figurine has a long, twisted plait of hair hanging down the back to the hem of the skirt.285 Considering the culturally specific setting in which figurines are found, megalithic buildings or funerary complexes, the differences may denote human qualities, social status, levels of spiritual attainment or readiness for interaction with the divine or spirit world; they may mark certain stages in the performance of ritual at a human level.286 If they represent the divine, then perhaps the differences mark shifting roles and differing levels of interaction with the community who revered them or changes in the natural environment and seasons. They also may represent not one divine entity but several, an aspect that will be discussed later. It cannot be discounted that the figures reflect an ancestral past.The concept of ancestor worship was sparked by Evans’ essay on the Mycenaean tree and pillar cult.287 Though often interpreted as representing a deity, it is possible that the statues may have depicted living people. One remarkable sculpture might convey cosmological concepts that straddle the mundane and divine realms



The Archaeology of Malta through three metopes, incorporating what was once a larger corpulent figure, found in Tarxien (Figure 3.18, no. 3).288 At the lowest level (Figure 3.18, no. 3c), a group of realistic figures is shown gathered together within a chamber; the one to the fore is seated, full-bodied and in the pose redolent of the figurines, with right arm to the side and left arm on the lap or across the body. Above this group are remnants of two standing robust figures who bar the way to substantial entranceways or who equally could be seen as standing within the vertical pleat lines of the skirt of an even larger figure that has been broken away (Figure  3.18, no.  3b). The skirted body and legs of this largest being probably did represent the divine in this instance (Figure 3.18, no. 3a).289 In this context, the smallest figures might depict revered individuals, living or deceased elders in the community, in realistic scenes that may have occurred around the megalithic buildings.290 If this is the case, this sculpture suggests that the upright stones comprising the megalithic buildings themselves might have symbolised the vertical segments – long strips, folds or fringing – of clothing, that is, the skirts depicted on numerous figurines. In this respect, the stones spotted with numerous pits drilled in their surfaces (Figure 3.17, no. 1) may evoke the spotted fringes of the figurines’ skirts, as worn, for example, by the famous ‘Sleeping Woman’ from Ħal Safieni (Figure  3.18, no.  1). Of the medium-sized figures within the skirt pleats, one could argue for their identity as revered ancestors or perhaps members of an ancient pantheon, minor divine beings governed by an overriding supreme deity. Figurines can have signs of possible body scarification, perhaps indicative of rites of passage or healing rituals. In particular, statues from the Xagħra Circle in Gozo, disproportionally large in the hip and thigh, can have arms folded tightly across the body with what seems to be a disc held flat covering the torso (Figure 3.18, no. 4).291 In others, one or both of the arms are placed forward, lying along the thighs or resting on the lap, revealing a deeply lined or scored abdomen. A few have incised lines on the back as well (Figure 3.18, no. 5);292 these lines do not appear to represent ribs or natural folds of skin.The statues can have the female attributes of modestly proportioned breasts. Curling under the corpulent legs of the pair of figures sitting on a bed from Xagħra Circle are double-line tendrils that continue over the top of the bed (Figure 3.18, no. 2).The legs and lines are heavily daubed with red ochre, which might reflect the shedding of blood by females, whether menstrual or perhaps postpartum. A parallel can be drawn to the double-lined tendril patterns often seen on contemporary pottery, again enhanced with generous quantities of red ochre (Figures 3.9, nos. 6 and 16, and 3.10, no. 12). In both instances of tendril designs (figurine and pottery), the lines terminate in V-shaped, bifurcated ends. As a recurring pattern, we can also look to unusual high-handled bowls that fork into two-pronged ends in the Red Skorba phase (see Chapter 2). In prehistoric Malta, this symbol might hail back to previous centuries, where the crown of the goat skull with horns attached forming a V shape was preserved

The Culture of the Megalith Builders in domestic and possibly a shrine setting at Skorba.293 Overall, the double curling lines, bifurcated prongs and red ochre might combine to signify the female, shedding of blood and the life forces this represents.This theory perhaps seems not so far-fetched when one considers the rich, often deep red and fertile soil in Malta and Gozo, to which such iconographic references might be directed. Other interpretations link spirals and curling motifs to water and the sea294 or to broad concepts such as beginnings (genesis or origins), continuity, regeneration, discovery, peace, tranquillity, strength or familial bonds.295 COLOUR SYMBOLISM: AN ETHNOGRAPHIC ACCOUNT

Ethnographic studies highlight the significance of colour in recent cultures. One account of ochre procurement in Australia may be useful in understanding the possible significance and implications of red colourants in non-complex society. Before European contact, indigenous Tasmanians mined and used red ochre liberally on their bodies and as fatty pomade in their hair. Men likened their ruddled hair to the new spring leaves of the trees and the implied health and virility. Men and women applied ochre to their cheeks, lips and body. For both, red ochre was linked to sexuality. Between 1829 and 1834, British immigrant George Augustus Robinson, with some Tasmanians, including a woman called Truganini, went into the countryside to find and eventually relocate the remnant members of indigenous Tasmanian groups. He was witness to the rituals and preparations surrounding access to ochre deposits. On one such occasion, much song, dance and storytelling took place, and courtship flourished among the people who were with him.296 Only after many months of living with the Tasmanians was Robinson led to the ochre source at Toolumbunner, and the Aborigines expressed delight: ‘The chief MANNALARGENNA is overjoyed at the prospect of seeing this celebrated place. He is in raptures. “By and by me see it, plenty, plenty.” ’297 Large skin bags were made in advance and filled to capacity at the mine, and certain rituals were observed at the site. One in particular involved collecting, heating and fracturing lunettes from the edge of naturally formed stone discs called ballwinny stones, which were harvested from the river at the foot of the mountain and used as grinding palettes for the ochre.298

Red Ochre and Colour Symbolism in Malta Extensive usage of red ochre indicates that it was a highly valued commodity among the ancient megalith builders of Malta, and it was employed widely in daily life (pottery and perhaps body paint), in spiritual expression (on the figurines and wall paintings) and in funerary contexts (painted ceilings and sprinkled on corpses).299 Imported commodities such as greenstone axes and



The Archaeology of Malta obsidian probably reflected the social standing of their owners, and ochre might have been another commodity brought to the islands.300 Laura Maniscalco301 discusses the finds of small ochre-filled pots at a site in Sicily where there is evidence of strong connections between the Sicilian ‘San Cono–Piano Notaro–Grotta Zubbia’ culture and the Zebbug period in Malta. She argues that the Maltese Zebbug pottery repertoire was possibly influenced by the Sicilian through the process of trade in ochre, which played an important role in the ritual of Malta at that time.302 It is possible that like the indigenous Tasmanians, the Zebbug community may have targeted certain, perhaps revered sources in Sicily and formed strong bonds with the Sicilian providers in whose territory the mines occurred. Interaction may have taken place for centuries, which was probably based on gift giving, barter and exchange. Colour symbolism does not end with the use of red ochre. Among non-complex societies, white, red and black, a colour triad, are a common combination, often reflecting cosmological beliefs and aspects of the physical world, the realities of daily existence (birth and transitions in life, death) and emotional expression. There may be no universal applications, but at the very least, the possibilities of deeper meaning can be suggested for the Maltese archaeological record. Green polished axes, for instance, were also imported. Some of them were possibly functional axes; others were small, pierced amulets and were worn. Robin Skeates argued for a metamorphosis of the polished axes from the functional ‘physical’ to the symbolic ‘conceptual’. They were ‘symbolically charged objects of social display, spiritual power and ritual sacrifice’.303 Malone and her colleagues suggested that their green hue might have carried symbolic values borrowed from the natural world to do with re-growth and renewal.304

Dualism and Triadism: Expansion of the Megalithic Structures ‘Mother goddess’ as a concept is a modern construct which has overshadowed other possibilities for interpretation of the archaeological record. Worth considering is that a fundamental shift in ideology occurred at some point in the megalithic building period.305 This is evident in the multiplication of buildings and also in iconographic representation of motifs which are duplicated or tripled. These changes might point to a shift from a cult with a single focus to pluralistic observation, or perhaps a more likely scenario is that the status of pre-existing minor cults increased, necessitating a corresponding addition of new buildings within the megalithic complexes. As David Barrowclough argues, the ‘cult involving greenstone pendants was but one of several cults that … co-existed during the Neolithic.”306 The general consensus has been that the elaboration of building design followed an evolutionary path from lobed, through trefoil, five-apse (Figure 3.14, no. 2), four-apse, to six-apse plans spanning the Mgarr, Ggantija, Saflieni (or early

The Culture of the Megalith Builders Tarxien) and Tarxien phases.307 There is no clearer path for defining the chronology of the buildings, which hampers this discussion. Many of the developments fall within the 400 years of the Ggantija phase (ca. 3600–3000 BC),308 but our understanding of the pottery that defines this period has yet to be further refined, and the current temple sequence is based on architectural typology.309 Nonetheless, what drove the developments may have been an increasingly complex cosmology. Returning to African Gurunsi communities, it was observed that old structures (admittedly in more easily worked materials) were destroyed and rebuilt to accommodate a son and his new family when he married.310 Could the building works have been made in Malta to accommodate an expanding belief system? If the buildings also were dwellings, could the developments have been driven by mundane needs, necessitated by evolving and growing family units? Views are many concerning the development of megalithic structures. Sherratt argues that inter-community rivalry may have been at play: ‘relative isolation nurtured a more extreme form of monumental principles exemplified in other areas, though the Maltese centres must be imagined as competing one with another, rather than with other groups, since there was apparently no alternative local ideology.’311 Further elaboration of monuments built by preceding leaders for self-aggrandisement is possible.312 Grima, too, suggests that a link existed between the size of the megalithic monuments and productivity of the land.313 Trump proposed that the alterations in the cases of Skorba, Mgarr and Kordin III created public outer areas and inner private spaces necessitated by altered rituals.314 Restructuring and reuse of stonework may indicate that neither the stones themselves nor the initial circumstances that inspired the constructions were revered in their own right; that is, once built, the monuments were not sacrosanct.315 It is clear that the monuments that survive to the present day are the final statement in their development. Major complexes such as Ggantija, Tarxien, Mnajdra and Ħagar Qim were originally built as single complexes and then were changed by partly dismantling the outer building wall to incorporate either a totally independent group of rooms in a new complex or rooms inter-connected with the pre-existing structure. There is also a sense of duplication in separate megalithic complexes built reasonably close together:  in a direct line, Ħagar Qim is 526 m from Mnajdra, Ta’ Ħagrat is 881 m from Skorba, Kordin I was 207 m from Kordin II and 629 m from Kordin III and Kordin II was 573 m from Kordin III. Inter-site visibility may have been a key factor in the location of these structures.316 Iconography depicting duplicated designs is represented by the remarkable twin figures sitting on a bed found in the Xagħra Circle hypogeum (Gozo), or perhaps they are a triad if we consider the small figure on one lap (Figure 3.18, no. 2). At Ħagar Qim, remnants of two full-bodied people standing side by side remain on one of the slabs now exposed on the outer wall of the complex


The Archaeology of Malta


and the small groups of figures stand within and under the skirts of the Tarxien statue (Figure 3.18, no. 3). Duplicated motifs on stele within the Tarxien complex represent two plantlike designs with tendrils. Panels also depict back-toback dual sets of spiral designs (Figure 3.17, no. 13). Dual and triple so-called phallic symbols cut into niches also may be significant in this regard. The motivation for the repeated pillar motif has not been investigated. One example stands out from the others. It depicts two pillar-like figures that carry more elaborate all-over decoration, and before them is a small pot or holed disc. Joseph Cilia317 argued that these represented artefacts of textile production: a spindle whorl weight and bobbins of some sort. Given the nature of the designs, it is possible that they are stylised representations of clothed human (or cult) figures. With the pot on the ground before them, the overall design recalls the iconography of the two figures sitting on the bed from the Xagħra Circle  – one with a pot, the other with a child or figurine on its lap. The niche pair is also not unlike the peg figures found in the Circle burial complex (Figure 3.18, nos. 13–18). That the figures represent people is further suggested by the grooved roof of the niche, which may represent the stones over a doorway or room in a megalithic structure.

Funerary Contexts: Caves, Graves and Hypogea Like the upper levels of the Xagħra Circle, Ħal Saflieni had remnants of megalithic constructions and rock-cut features that once formed a monumental entrance on the surface.318 It is possible that only some members of the community were buried in the major complexes such as Ħal Saflieni and the Xagħra Circle.319 A  second complex, for instance, was found in 1973 in the district of Santa Lucia, just under a kilometre south-west of the Tarxien megalithic complex.320 Signs of deliberate placement of certain bones – ‘bone flow’ – are likely at Bur Mgħez, a modified natural rock cavity in Tan-Naxxari quarry near Mqabba within which at least seventy individuals were interred.321 Tagliaferro describes skulls ‘lying on the large flat stones themselves’322 and others that were crushed by slabs. Long bones were gathered together in areas, with or without signs of order.323 Such ritual behaviour was observed at Xagħra Circle. In addition, the large stones originally may have formed shelving which collapsed. These stones, along with small pebbles, formed an arrangement associated with the skulls at the time of excavation,324 although Tagliaferro described instances where the stones seemed to have been used to prop up some bodies on their left side or placed in ‘crouching’ positions and to cover others.325 Fragmentary pottery was found in no particular observable order; very little lithic material and three shell pendants were recovered.326 From the evidence of the hundreds of burials in the Xagħra Circle, our understanding of funerary practice in the islands has been greatly enriched. Meticulous excavation has revealed the nuances of burial ritual.There are some

The Culture of the Megalith Builders indications of deposition according to the gender and age of the deceased and of ritual interaction between the living and their revered ancestors through rearrangement of skeletal remains. Symbolism reflecting the life cycles of their community and their belief systems, it has been argued, played out in funerary rituals, which may have passed over the landscape, along processional routes that moved between significant landmarks, the not so distant megalithic complex at Ggantija included. Reminders of life forces are embodied in green amulets and red-ochred artefacts left with the burials.327 Sifting through the evidence presented here, there are indications of how the megalithic builders perceived the passage of life and death. A well-reasoned and structured belief system emerges, with distinct concepts of predictable life processes and the uncertainties brought by death.This belief system is reflected in the megalithic buildings with their orderly, systematic ground plans, carefully controlled interior and exterior spaces and decorative motifs which reinforce concepts of order over the natural world, over the inhabitants’ productive island, surrounded by a bountiful sea. Symbols, though encoding the mysteries of their beliefs, the depths of which are only hinted at, are carefully presented in harmonious, repeating patterns. Neat rows of domestic animals as well as fish, almost waiting to be caught, are depicted (Figure 3.20, nos. 4 and 5); even representations of plants are controlled, or potted.328 Possible deities or revered ancestral spirits presided over areas within the monuments in the form of statues or were carried to the grave as figurines. Presumably such entities were thought to contribute to the well-being of their community on a day-to-day basis. Full-bodied, robust and overflowing with life-giving blood are notions conveyed by the statuettes, as well as by the curling linear iconography and the liberal use of red ochre in many aspects of the people’s lives. But rituals for the deceased brought them into a dark chaotic realm of unnatural echoing sounds and past ancestors whose remains eventually crowded the underground burial chambers. Some things may have been familiar to them, seen in everyday life, yet they are odd, almost inside out. The most striking example is the architectural façade at Ħal Saflieni, modelled on the above-ground monumental entrances but turned inward to face what appear to be the interiors of rooms with corbelled ceilings. Such a doorway is, in effect, false for the living but perhaps not so for the deceased if some form of spirit world was woven into the builders’ ancient cosmology. Underground, especially in the Ħal Saflieni complex, they encountered a puzzling, chaotic array of rooms and chambers that followed none of the clear, predictable, lobed plans that they could expect in the above-ground buildings.329 Monumental perilous stairs that end abruptly, unexpected deep pits and openings into deeper galleries become obstacles to avoid; there was no easy way to negotiate a path in this dark maze of chambers through concealed, light-deprived spaces. Decoration here is not at eye level on walls and doorways but overhead on ceilings and high on the walls. The paintings appear as a disarray of mixed images, curls,


Figure 3.20.  (1–3) Bowl or plate from Ħal Saflieni with different decorative techniques; highly stylised animals hidden within a repeating circular pattern on the exterior (1) and naturalistic animals on the interior (2) (inv. no. S/P.1) (after Evans 1971, p. 60, fig. 10:6); (4) relief decoration from Bugibba depicting a maritime scene possibly of a boat with a person standing in it catching a large fish (drawing by C. Sagona; photograph by A.  Sagona); (5)  boat graffiti scratched on megaliths at Tarxien; note that some smaller craft, particularly on the left, have inward-turned ends similar to the Bugibba carving (4) (after Fradkin Anati and Anati 1988, fig. 4). 112

The Culture of the Megalith Builders spirals and spots that merge randomly together. In this dim, unnatural environment, it seems a reasonable conclusion to draw that their religious experiences included elements of trance states that allowed them to interact with a perceived spirit world. Ħal Saflieni especially may reflect unnatural visions in paintwork that decorates a perceived realm of the dead, recreated in the three-dimensional galleries in this underground complex. Within their burial places and possibly in the fabric of the above-ground monumental structures and perhaps even their humble homes may lie encoded a blueprint of the megalithic builders’ cosmology. SUMMARY

As archaeological investigations on the islands have widened to include environmental and geomorphological aspects, so patterns defining site locations have emerged. In many respects, findings have pointed to economic factors for the site locations. Proximity to reliable sources of potable water and access to coastal zones and fertile plains were no less important factors underpinning the choices made by ancient communities. Skorba, in particular, was inhabited for centuries and was eventually capped by monumental structures. The megalithic buildings in Malta and Gozo have posed a challenge for generations of historians and archaeologists, both in the archipelago and further afield. We arrive at the point where there is no need to look beyond the islands’ shores for the origins of the Late Neolithic building technology. There is evidence that after centuries of working with stone – the principal building material quarried close to the megalithic sites  – the islanders acquired the skills they needed and a clear understanding of the limestone’s structural potential. In design, the structures clearly follow a basic lobed plan that grew out of rounded to ovoid domestic architecture, which may have been clustered or budded together in small hamlets. It is likely that in funerary architecture people attempted to mirror aspects of life; hence, smaller tomb chambers with lobed cavities might be modelled on house plans. Similarly, the large underground hypogeum at Ħal Saflieni was cut, over time, to resemble the above-ground megalithic buildings rather than the other way around – namely, that megaliths mimicked the grave.330 With this simple shift in perspective, the emphasis falls on funerary contexts reflecting life and fostering a connection between the living and the dead rather than on a preoccupation of the living with death cults. Driving much of the debate surrounding the emergence of the distinctive Late Neolithic culture are issues of why the islanders chose to build on a monumental scale, what function the buildings served and the nature of the society that built them. In this regard, there are no clear answers. The ancient megalithic builders had a developed sense of their environment, their place in the landscape. This spatial concept extended to a sense of the natural island



The Archaeology of Malta boundaries formed by the sea, but of the internal territorial divisions we know virtually nothing. There is a pressing need to locate and thoroughly investigate settled areas not characterised by monumental structures if the debate is to move forward. Discussions have ranged around diverse aspects of Maltese prehistoric belief systems. Some have concerned themselves with notions of mother goddesses, whereas others have applied definitions of the sacred. Private and public spaces have been explored. In this regard, cycle-of-life processes and the connectedness between people, as well as those concepts of places for the living and the dead, have been dealt with to some degree. Whereas belief systems and ritual are the main drivers of inquiry into the function of the monumental buildings, aspects of the mundane should not be overlooked. Terminology such as ‘temple building’ and ‘temple period’ locks the field into a way of thinking and perpetuates a marked division between sacred and mundane that may not have been so clear-cut for the ancient islanders. After centuries, the megalithic structures fell from use, sometimes abruptly, sometimes after serving as shelters of a different kind. People probably continued to eke out a living as they always had, but the monumental buildings, it would seem, no longer embodied the same cultural focus. How the closing years of the Late Neolithic unfolded and the new directions societies on the islands took are the subject of Chapter 4.


The megalithic builders and their forebears had successfully dwelt in the Maltese Archipelago for around two millennia, forming cultural focal points within geographically defined areas.1 It is questionable whether the islands were abandoned when the Neolithic way of life faltered, as has been long argued. Notwithstanding that transition periods are often the hardest to detect, there are at least signs that the final groups for whom the monuments held meaning closed them down in a manner which they held culturally significant. Thereafter, people simply went about their daily lives, albeit on a much more modest scale; their presence remains virtually imperceptible, most likely because their manufactures still carried the hallmark of Neolithic practices. The widely held view maintains that the end of the Neolithic came abruptly, prompting an effective depopulation of the islands around 2400 BC. The causes range from invasions from Sicily or beyond, inter-community rivalry and island-based social ruptures through climatic disaster and environmental mismanagement of the islands’ resources to over-population.2 The result, it is claimed, was social, economic and political collapse in the archipelago.3 This extreme scenario, however, is argued from a patchy archaeological record – a dearth of clearly stratified cultural sequences, a limited number of radiocarbon dates and a rather black-and-white interpretation of occupation deposits. It remains to be seen whether cultural remains of any interim phases will be detected in the future. Such material is likely to fall into the ‘grey’, not quite Tarxien Neolithic and perhaps not entirely Tarxien Cemetery, but categorised as one or the other in past reports largely because the archaeological sequence of Malta has yet to reach maturity.4 What came later in terms of Early Bronze Age traits, a pottery type known as Thermi Ware and the Tarxien Cemetery cultural phase will be the subject of Chapter 5. A significant restructuring of cultural practice appears to be the hallmark of the post-Tarxien phase. Certainly the monumental buildings ceased to be the hub of daily life and the heart of community organisation they once represented. From our perspective, and based on a record garnered largely 115

The Archaeology of Malta


from fledgling archaeological investigations, this change seems sudden, but the process may have been drawn out. Contemporary archaeological theory has shifted the focus away from abrupt, irreversible and utter catastrophe to a notion that communities, in times of stress, have a capacity to absorb disruptions and are resilient and adaptable.5 A constant, unchanging environment is not the standard within which a society lives. It is the human lot to evolve and change depending on any number of localised or external forces. In response to economic or political crises, people usually modify behaviour while maintaining key elements of their ways of life, especially how they provide for their daily sustenance or preserve elements of their belief systems. Stepping from the Late Neolithic period, Malta entered the Bronze Age.This period is characterised by the appearance of new cultural features, with significant funerary deposits, dolmen constructions, perhaps clusters of bell-shaped pits and hints of rough stone foundations and mud-brick buildings now dissolved into ill-defined deposits in a few locations. With the cultural changes came metal. Copper axes and blades were lavish inclusions within hitherto unprecedented cremation-pot burials. Part of this narrative concerns land use. In this chapter, we will explore two notions. The first considers the evidence for the late Neolithic people pushing the boundaries of agricultural zones into marginal areas.6 If there was a need to expand food production, whether this was driven by, for example, over-population, diminishing land fertility through mismanagement, environmental downturns or greater food production demanded by a governing hierarchy, or possibly an insidious combination of factors, what evidence can be marshalled to better understand this process? The second aims to crystallize the archaeological evidence for the end of the Neolithic period. Long thought of as abrupt, new excavations and revision of past explorations are changing our understanding of cultural change in the islands. There is evidence that communities in Malta made a decision to shut down the long-established practices which had endured up to the Tarxien phase.What came after has yet to be fully defined, but the ongoing excavations at Tas-Silg (north) and the reappraisal of the Borg in-Nadur site may point the way. FIELD-FURROW SCARS OR CART RUTS? It is usually possible to coax more resource production by applying capital and technology, increasing labor, applying energy subsidies, and making production more knowledge-intensive. Irrigation, fertilization, and mechanization are all ways to increase production, as is putting in more hours or tilling more land.7

A great debate has grown around the so-called cart ruts – ubiquitous grooves eroded into many rocky surfaces (Figure  4.1). Arguments usually look to

Pushing Boundaries at the End of the Megalithic Building Period

Figure 4.1. Western sector of Malta showing the Burmarrad catchment relative to rock-cut groove (field-furrow) areas. Sites are (1) Qammieħ, prehistoric settlement; (2) Qammieħ, likely prehistoric settlement with some Bronze Age pottery; (3) Għajn Zejtuna; (4) Bronze Age sherd scatter; (5) Mellieħa, Wied Ta’ Manna; (6) Manikita,


The Archaeology of Malta


explanations that concern some sort of vehicle carting stone or other produce back and forward across stretches of land. The inherent implications of related nomenclature are the subject of extensive discussions: wheel or runner, cart or sled, width of axle gauge, means of hauling, nature of goods hauled, deliberately cut or worn into bedrock, date of use and ‘routes’ followed. Each generation of scholars has tackled the function of the cuttings, and their views are presented in numerous academic works, popular books and websites.8 The Maltese evidence is often compared to similar markings found from Europe to Azerbaijan.9 Major studies have appeared, none are comprehensive, and most consider that these cuttings are the product of vehicles. The following discussion explores a different scenario based on a single question.What if the superficial appearance of the grooves has been misleading and they resulted from a very different function associated with agricultural land use? In essence, there are several factors that need to be considered in the argument for the rock-cut scars to equate with vehicle tracks:  the attributes of ancient vehicles and mode of traction (pulled along by people or animals), the nature of the bedrock into which they are cut, how the grooves were cut and the feasibility of wheels or runners operating in the grooves.Threaded through most discussions is a clear premise that the grooves are vehicle tracks; otherwise, the minutiae of the arguments vary considerably.

Figure 4.1. (cont.) Borg in Nadur sherds; (7) Ix-Xagħra tad-Dar il-Bajda, Bronze Age sherd scatter and rock shelters; (8) Xemxija megalithic ruins, Xemxija tombs, stone alignments and menhir; (9) Il-Qortin Bronze Age, possible fortified site; (10) Bugibba; (11) Il-Qarraba, strategic Middle Bronze site and scattered sherds; (12) Fawwara, Ras il-Gebel, Bronze Age site with sherds and lithic scatters; (13) Qala hill Bonze Age fortified settlement; (14) Mgarr stone circle; (15) Ta’ Lippija, cairn; (16) Ta’ Ħagrat, Mgarr megalithic building; (17) Skorba, Mgarr megalithic building; (18) Il-Qolla, Bronze Age sherd scatter; (19) Gebel Għawzara, possible Bronze Age fortified settlement; (20) San Pawl Milqi, Bronze Age sherds; (21) Tal-Qadi, Neolithic settlement and megalithic ruins; Tal-Qagħadi, Burmarrad, megalithic building remains; (22) Mosta, prehistoric settlement; (23) Mosta, dolmen and remnant megalithic in field walls; (24) Ras ir-Raħeb; (25) Baħrija, Bronze Age site; (26) Kuncizzjoni, megalithic building Il-Qlejgha; (27) Bingemma, Bronze Age sherd scatter; (28) Baħrija-period settlement; (29) Torri Falka; (30) Qallilija, Bronze Age remains; (31) Mtarfa pits; (32) Rabat-Mdina, various Bronze Age deposits; (33) Wied l-Imtaħleb; (34) Zebbug tombs, Ta’ Trapna iz-Zgħira; (35) Għar Mirdum; (36) Wardija ta’ San Gorg, Bronze Age remains; (37) Wardija ta’ San Gorg, Bronze Age fortifification and other sites; (38) Mnajdra; (39) Ħagar Qim. (A) Location of field 5; (B) field 1; (C) field 2; (D and E) fields 3 and 4 in Table 4.1 (see Figure 4.2) [map by C. Sagona; furrow clusters have been documented in Magro Conti and Saliba (2007) and Weston (2010); others have been plotted from Google Earth satellite images; Burmarrad catchment after Marriner 2012, fig. 2).

Pushing Boundaries at the End of the Megalithic Building Period From the outset, it has to be noted that chronology is an issue, and the features do not lend themselves to precision dating.10 If one maintains the position that vehicles were involved, then scrutiny should fall on the evidence for carts, wagons, sleds and animal harnesses in the ancient world, which have been the subject of detailed, far-reaching research outside of Malta.11 Essentially, depending on what period scholars pin the rock-cut grooves to, it is equally important to their arguments that the grooves represent tracks for an assessment to be made of what vehicles were available at the time. Vehicle use was cutting-edge technology in the ancient Near East and Europe, where there is direct evidence of carts and wagons in the fourth millennium BC, when such technology was developed. Vehicles in prehistoric times were big, cumbersome carts with solid wheels weighing a considerable amount. In some cultures, there is clear evidence that vehicles and the animals that pulled them were emblems of status and hence were sometimes incorporated into burial practices.12 It was not until the Roman period that roads were well made and level enough to enable the hauling of great weights. In addition, it was during Roman times that harnesses were suitably modified to distribute the weight across the animal’s body so that it could pull heavy loads efficiently. The areas with clusters of rock-cut grooves in Malta fall far short of the technology and know-how emanating from Rome concerning road construction, the kind of road construction which allowed heavy vehicle use.13 The remote location of many furrows, in barren rocky regions with no striking connection with major Roman developed areas, further suggests that they predate the Roman period. Quarries at Misraħ Għar il-Kbir (its popular name, ‘Clapham Junction’, is intrinsically misleading and should be avoided), which are thought to date to the Roman period based on the size of the stone blocks that were removed, clearly cut through pre-existing furrows in a number of places and hence post-date them.14 Many furrowed areas are not near quarries, and when they are, the furrows are often in large numbers – far more than seem necessary if the ruts represented vehicle tracks. The appearance of rock-cut edges on ancient but chronologically distinct features in close proximity can be markedly different, and this can offer some perspective on the worn condition of the furrows (Figure 4.2, nos. 5 and 7). The presence of considerable wear – such as depth of furrow, smoothing of the sides and irregular pitting in the floor of the cutting – has been used as evidence for erosion caused by even minimal vehicle passage. Yet cut surfaces of Roman quarried blocks are sharper and not similarly eroded.15 At Xemxija in the north of Malta, Neolithic tomb entrances (predominantly of the Ggantija phase), which most would agree are older than the smoothed furrows nearby, have notably sharper edges (see Chapter 3). Differences in eroded surfaces, in



The Archaeology of Malta

Figure  4.2. Possible ancient fields identified in Table  4.1:  (1)  field 1, Qalillija, 35º53′37.40″N, 14º23′05.72″E, elevation 153 m; (2)  field 2, Ix-Xagħra ta-Xini, Bieb ir Ruwa (limits of Baħrija), 35º53′38.60″N, 14º21′43.72″E, elevation 224 m; (3) fields

Pushing Boundaries at the End of the Megalithic Building Period these instances, suggest that friction through repeated tilling of the land did in fact play a part in the degree of wear in the furrows, not just erosion.16 In other locations, rock furrows are cut by Phoenician-Punic tombs, pointing to a date prior to the early first millennium,17 and in a few instances, lines of furrows are cut by ‘silo’ pits; these are usually associated with the Bronze Age.18 A number of furrow sites are closest to Late Neolithic and Bronze Age settlements (Figure 4.1).19 Given these chronological indicators, many clusters of field furrows therefore are likely to belong to the prehistoric era of Malta, a time when neither roadways, nor vehicles, nor animal trappings were advanced enough – not just in Malta, but throughout the ancient world – to explain such features. In addition, there are no representations of vehicles, wheels, model carts or sleds in the archaeological record for Malta, nor representations of people using ploughs or draught animals. If we were to attempt an interpretation of ancient Malta’s artistic record, then it might be possible to see in the arching and linear designs on Zebbug pottery representations of field furrows and channel lines beside which are dashes that might represent the tracks of bipedal footprints (see Figure 3.1, no. 30). GEOMORPHOLOGY OF THE MALTESE BEDROCK

There are two recent schools of thought concerning the geomorphological qualities of the Maltese bedrock, an important aspect of this debate. Mottershead, Pearson and Schaefer at the University of Portsmouth suggest that erosion has played a significant part in the current state of the furrows (smoothness of the walls, irregular depths and undulating floors). Compression of the rock (lying under thin soil cover) owing to human agency was greatest when the bedrock was softened through saturation by rain. These processes helped to shape the cuttings, especially after the bedrock was exposed to erosion.20 In another study, Martyn Pedley (University of Hull) acknowledged the effects of erosion but considered that the grooves saw regular use and that they were largely clear of soil at the time.21 Yet, regarding the fundamental role of

Figure 4.2. (cont.) 3 (north) and 4 (south), Misraħ Għar il-Kbir, between the fields, 35º51′07.81″N, 14º23′52.88″E, elevation 228 m; (4)  field 5, Naxxar, 35º55′32.98″N, 14º26′16.94″E, elevation 64 m (1–4:  arrows show direction of rock-cut grooves) (1–4 based on Google Earth image 15.4.2013); (5)  field furrows at Naxxar [line A  clips a deep pit (H)  and small channel lines leading into D, and from the pit (C), additional features could be small sockets for an associated structure (F and G); E appears to be another depression in the rock where water could have accumulated]; (6) Xemxija rock-cut furrows over a stepped surface perhaps demarcating one field from another (photograph by A. and C. Sagona); (7) regularly spaced field furrows at Skorba (after Murray 1925, pl. 33:4).


The Archaeology of Malta


human agency, the former view adheres to vehicle use as the ultimate cause of the furrows – put simply, the scars were largely formed by natural processes after people passed over the ground in a cart – whereas the latter finds that tool marks on the walls are, in the first instance, evidence that people cut the grooves for vehicle use. Both schools fall short of tackling the problems inherent in identifying vehicles as the ultimate cause of the furrows. Often the ground between the furrows is uneven and fraught with anomalies which present obstacles. For these, there is no comprehensive explanation to date for how animal traction pulling carts or wagons negotiated significant land slope and stepped areas (Figure 4.2, no. 6), nor how they crossed large pits (e.g., Figure 4.2, no. 5), especially as the transport has been linked to hauling and carting everything from stone to farm produce.22 It has been argued that the land was soil covered and that people ran their vehicles over the surface ‘blissfully unaware’ of the concealed rock surfaces,23 but this poses further questions, especially for multiple and regularly spaces grooves and for an array of additional features (pits, stepped terrain, side channels, single not paired grooves, different furrow spacing and abrupt furrow ends). Such rock-cut features associated with the grooves indicate that people were deliberately modifying the bedrock. This information suggests that the furrows are more than simple wheel marks and that the additional cuttings have no relevance for any notion of transport. Signs of tool marks in the furrow scars are clearly preserved in some cases.24 These anomalies should not be overlooked because they do not fit the entrenched vehicle-track explanation. With major European Union funding, The Significance of Cart-Ruts in Ancient Landscapes Project, under the direction of Herman Bonnici, focussed on rock-cut features in Misraħ Għar il-Kbir (Malta) and Padul (Spain).25 This venture has resulted in the most comprehensive and useful account of the features thus far published. Even so, these results are on too large a scale to consider smaller associated elements. In almost every documented example, no account is made of the numerous anomalies and associated features such as small trickle channels (Figure 4.2, no. 5, C and D), water catchment areas (Figure 4.2, no. 5, E and H) or even the gradient of the land which are part of the story.26 From its inception, the project name locked the results into the deeply entrenched mind-set of wheeled-vehicle tracks.27

A STRATEGY FOR ANCIENT FIELD PRODUCTION Ockham’s Razor – The principle (attributed to William of Occam) that in explaining a thing no more assumptions should be made than are necessary

[Oxford Dictionary]. On the basis of what is known about ancient vehicles, the passage of a four- or two-wheeled cart or dual-runner sled offers a weak solution to the genesis of the

Pushing Boundaries at the End of the Megalithic Building Period rock-cut grooves, especially for areas with large concentrations of furrow scars (e.g., locations in Figure 4.2, nos. 1–4). If clusters of such scarring had been found impressed in the soil of old land surfaces in Europe, few would question that they were evidence of ancient field furrows.28 Landscape studies by Grima, discussed in Chapter 3, determined that the megalithic structures were placed on the liminal zones of fertile plains and close to routes leading to safe anchorage on the coast. Others have argued that competition and increased demand for food may have stretched the islands’ resources during the course of the later Neolithic.29 It seems logical that new areas would be brought under cultivation because of increased demand (population growth or other cultural factors), either by farming in the marginal locations or by reclaiming degraded areas by building fields on rocky ground. Alternately, if such areas were originally viable and cultivated, exposure of the bedrock may have resulted from environmental mismanagement and/or climatic downturn. Exposed rock-cut features are evident in the lands around the megalithic buildings at, for instance, Skorba and Ta’ Ħagrat (Mgarr), where numerous furrow sites are cut over a wide arch to the north, spanning from west to east (Figure 4.1, nos. 16 and 17). Many more furrow scars are clearly visible in satellite images of the archipelago; in a number of cases the cuttings expand over the landscape well beyond the zones identified in published reports concerning the features. In short, there is extensive evidence of furrows in clusters across Malta’s exposed rocky surfaces that find a commonsense explanation as field scars but, by dint of their sheer numbers, struggle to support the notion of ancient vehicle passage. To understand fully the scarred rock faces, a concerted effort is needed to stand back from preconceived notions, to avoid function-loaded terminology (cart ruts, axles, gauges, roads, etc.) and focus on the individual circumstances of each cluster. If the farmers of ancient Malta were claiming rocky, non-productive ground for new fields, it is possible to draw some parallels from the ethnographic record, particularly from the Aran Islands, where the landmasses are small, rocky surfaces abound that are weather beaten and heavily eroded and water is scarce.30 Naturally, no two situations are identical, even within Malta, but the Aran Islands do offer some useful insights into agricultural practices and the making of soil and fields on otherwise rocky surfaces. Sand and mud was harvested from coastal margins on the Aran Islands at low tide. The rocky ground was covered in broken stones, and soil and seaweed were piled up into beds between deep furrows, a field technique known widely in Irish contexts as ‘lazy beds’. An essential element of agriculture is that fields are drained well. Some channels between the raised beds drained water from the farm plots to the shore in the Aran Islands.31 The depressions in between the beds were cleared regularly, and the soil was thrown back onto the raised beds. With regard to farmers making soil specifically in Malta, there are two accounts from the 1800s that indicate how it was done (see the box entitled, ‘Creation of Ancient Fields, Run-Off Farming and Economic Gain’).




Using the premise that rock-worn grooves occurring in clusters in Malta are ancient field scars (column F), the various characteristics of five such fields (column A) are represented in the Table  4.1. Gradient, to varying degrees, is present, and furrows tend to follow the slope up and down the rocky landscape (columns B–E).32 Presumably the furrows, relative to the incline, facilitated drainage or harnessed water (a practice known as ‘run-off farming’), especially when furrows can be shown to drain into cut pits, natural dissolution hollows or other catchment features. Susan Gregg has estimated wheat crop yields in antiquity, and one model she supplies for Neolithic Europe suggests that a 0.35- to 0.40-ha agricultural allotment per person would supply 55 to 65 per cent of the diet annually (column H).33 Climatic and environmental considerations, plant selectivity, maturation of a new field, level of manuring, region-specific farming know-how and individual or community needs are obvious variables that have to be factored in for Malta. Nonetheless, by implication, new fields on an otherwise barren rocky terrain must have resulted in a worthwhile degree of economic gain; such was their proliferation across the ancient Maltese landscape. Conversely, if this amount of land failed owing to erosion or abandonment, it may be possible to better determine how many people would have been directly affected. Bearing in mind that agricultural produce has been shown to have been people’s main dietary component (at least for the Brochtorff community in Gozo),34 such calculations could offer a significant avenue of research that has been overlooked, one which also could progress estimates of ancient population numbers in Malta and Gozo.35

In accounting for the sudden cessation of the rock cuttings in one area (limits of Siggiewi, near Laferla Cross), David Trump noted humorously,‘There seems to be a straight choice here between accepting that a soil layer has been swept away, or that the carts themselves became airborne.’36 In reality, he has described exactly what agricultural furrows would do at the end of the field, where the soil thins out to bedrock – they simply stop. Where there were no fields, where there was no hope of retaining or building soil, there are no field furrows scarring the rock. Cessation of furrow clusters also may mark where one person’s land ended and another’s began, or it may indicate larger territorial boundaries. The fossilised ancient Celtic landscape discerned for Figureheldean Down in Wiltshire offers useful insights into rural organisation relative to a settlement.37


Table 4.1.  Agricultural potential of ancient fields in Malta A








Site name in Malta with furrow clusters (see Figure 4.2)

Max. height of the field (asl)

Min. height of the field (asl)

Length of furrowed field

Average slope of long axis of field

Approx. no. of furrows (clearly visible)

Approx. area of ancient field

Model for Neolithic dietary requirements for wheat at 0.35–0.40 ha per person = 55–65% per individual (after S. A. Gregg 1988)

1. Qalillija (west of Mtarfa) (Figure 4.2, no. 1; see Figure 4.1, B) 2. Ix- Xagħra ta’-Xini, Bieb ir Ruwa (l/o Baħrija) (Figure 4.2, no. 2; see Figure 4.1, C) 3. Misraħ Għar il-Kbir (north side of hill) (Figure 4.2, no. 3; see Figure 4.1, D) 4. Misraħ Għar il-Kbir (south side of hill) (Figure 4.2, no. 3; see Figure 4.1, E) 5. Naxxar (l/o) (Figure 4.2, no. 4; see Figure 4.1, A)

164 m

142 m

228 m


220 m

179 m

9,349.4 m2 (2.310 acres) 0.934 ha 7,828.1 m2 (1.934 acres) 0.7826 ha

Food for a max. of 2.3 individuals

223 m

1-m drop per 5.7 m travelled 1-m drop per 59.6 m travelled

228 m

222 m

123 m


219 m

165 m

70 m

36 m

247 m

6,951.7 m2 (1.718 acres) 0.695 ha 7,412.7 m2 (1.832 acres) 0.741 ha 28,947 m2 (7.153 acres) 2.894 ha

Food for a max. of 1.7 individuals

228 m

1-m drop per 20.5 m travelled 1-m drop per 18.3 m travelled 1-m drop per 7.2 m travelled

Note: asl = above sea level; l/o = limits of a district.


≥12 ≥14

Food for a max. of 2.0 individuals

Food for a max. of 1.9 individuals Food for a max. of 7.2 individuals



The Archaeology of Malta Various field clusters reflected some chronological development, but of interest are features which served as boundary markers. Of particular significance are ditches that were traced cutting through the landscape for some kilometres; such ditches, as already noted, were characteristic of Neolithic settlements in Europe. It is possible that the long lines of sometimes single but usually paired furrows served a dual purpose in Malta as water channels that also effectively, if not intentionally, acted as demarcation lines between territories.Water within Malta’s arid locations may have been a highly regulated commodity that was channelled and stored; access is likely to have required some level of co-operation and negotiation, especially in marginal areas.38 Scientific studies do not argue against continual field cultivation and manipulation of water resources. Indeed, if the process of erosion can be triggered by the passage of a vehicle, weighed down or not, as they suggest, then continual clearing with farm tools or ploughing carried out by farmers working their land – repeatedly scraping tools along the same furrow lines over rock weakened by irrigation – would be equally wearing on the rock surfaces39: ‘when surface rocks are wetted by rain, they become so reduced in strength that they are very susceptible to erosion by applied forces.’40 At Tal-Ħotba, north-east of Zejtun, an area stripped of soil for building works revealed furrows, extensively and regularly scored into the bedrock, in close proximity to a Punic-Roman necropolis (fourteen tombs). Reputedly, the one did not impede the other, and it remains to be determined if the tombs pre- or post-date the furrows.41 Although the date assigned to the scoring of the bedrock in this particular area is uncertain, the scars of undeniable agricultural practice were etched into the rocky landscape. In short, unless the land is degraded and stripped of soil, we simply do not know how many existing farm plots covered such rock cuttings and whether some current fields in fact owe their existence to ancient activities.42 Most suggestive of agricultural practice are those areas with clusters of furrows occurring in approximately equidistant parallel lines (approximately 1.32 to 1.47 m apart), numbering from a few to twenty or more and over forty in the case of Misraħ Għar il-Kbir. But even a few grooves could indicate a field and/or drainage system.43 Whereas the rock beneath soil cover (especially when saturated) may have been compressed and later eroded, this poses more problems for advocates of vehicle use. If the furrows are not artefacts of agriculture and the rock surface was obscured by soil, as the University of Portsmouth project maintained, it is hard to explain how or why the ancients would have run vehicles in the same general direction but moving repeatedly across the surface a metre or so at a time until the landscape was scarred with well-spaced, regular, parallel rows.44 Scars as evidence of ancient fields and agricultural practice is a straightforward explanation.45 The significance of the furrow scars lies in what they represent in terms of prehistoric farming practices, but the cultural conditions during the time

Pushing Boundaries at the End of the Megalithic Building Period of their use and abandonment have yet to be determined. They may be evidence of farmers pushing production into marginal and vulnerable zones, driven by underlying economic need to expand arable land.46 Whether or not demands of an increasing population played a part, social issues concerning land-ownership and inheritance should be considered. Less tangible than population growth but equally possible is that some form of hierarchy (political or religious) made ever-increasing demands on the community in the Neolithic period, which pushed farmers into working more difficult areas. As we will see, geomorphological studies in Malta are indicating that an environmental downturn was well underway by the Bronze Age, probably resulting from the introduction of agriculture into the archipelago. It is possible that the rock cuttings are as yet a largely unrecognised part of this story. Field scars may offer a realistic means by which to determine the size of farm units in prehistoric Malta and to estimate crop yields.They ultimately may be the means by which to calculate population numbers in certain regions (see Table  4.1). This is a realistic line of research, one which could also offer some empirical evidence for regional territorial boundaries.47 Soil production and rehabilitation of rocky zones were common undertakings still current in the 1830s (see box entitled, ‘Ethnographic Accounts of Soil Manufacture in Malta in the 1830s’), but this practice has not been considered when trying to understand ancient agricultural practices in Malta.48


[1]‌ ‘The Soil is of a very peculiar nature. It is almost every where exceedingly thin, and in many places the rock is completely bare. But this rock, which consists chiefly of carbonate of lime with about seven per cent of alumina, is in general remarkably soft and crumbly, so that with very little expense of labour, it may be easily broken down and converted into a soil of extreme productiveness. In this way fields are every year reclaimed, and it is probable that much of the land at present under cultivation has been reclaimed in a similar manner. To this peculiarity of the soil the natives attribute, and apparently with justice, the dryness and general healthiness of their island.’49 [2]‌ ‘A countryman wishing to make a barren rocky surface a cultivated plot of earth, commences by breaking up the stones which lies [sic] on the surface, and for a depth of some six inches. This fine powder is carefully laid aside and mixed with the calcareous earth, which is invariably found under the first layer of stone – a half acre, which is the average size of a field, cleared in this way, is then covered with this artificial soil. By the assistance of manure, “and by its great aptitude in (continued)



The Archaeology of Malta its new form to the absorption of moisture from the atmosphere, its bulk very perceptibly increases, and soon forms a sort of concrete texture”. Water melons and cucumbers, requiring the least nourishment, are first raised, and will flourish the succeeding season – “corn is the usual growth of the third year”; and it is by this, and similar processes that by far the greater part of Malta and Gozo has been brought into a state of cultivation, and the soil been found so rich, that although only of a few inches depth, it will produce to the husbandman its two and three yearly crops, as a just reward for his toil and labours. It is a common conversation here with the countrymen, of their “ever-producing soil”; and a most happy thing it is for the Maltese, for had not nature ordained it so, many more instances would be noted of the death of the poor from absolute starvation.’50

The width between the furrows is kept to a uniform distance, usually governed by the length of the handle of the field tools used to weed and dig by hand (shovels, hoes, spades, etc.), a distance which can comfortably be spanned by field labourers tending the crops.51 Herein lies the basis for the misconception that the furrows in Malta are paired, with the space between representing the span of a vehicle axle.52 Clusters of rock-cut grooves are scars from intensive, hand-worked fields, where tools such as the ard were dragged by farmers year after year, perhaps numerous times a year, up and down the fields to turn soil. Intersecting lines that cross over clusters of furrows, which can be traced for some distance in some locations, are likely to have been water channels tied in with the practice of run-off farming techniques. Judging by the manner in which similar fields are manually tended in rural areas of modern-day eastern Turkey, for instance, such ditches, crossing the line of the field furrows, can be effectively manipulated to carry water to and from selected furrows simply by blocking channels with clods of earth or clearing them. Water run-off from these channels, presumably rich in nutrients, in the case of Malta was caught in rock pools and deeper rock cavities into which these long stretches of furrows can run or which they clip at one edge (Figure 4.2, no. 5).53 Significantly, geomorphological studies in the Burmarrad inlet in the north of the island, which was fed from the catchment area of the Wied Qannotta, Wied il-Għasel and Wied Għajn Rihana seasonal watercourses, have detected rapid sedimentation attributable to human activity in the region.54 Within or very close to this sector are, among other settlements, the Skorba and Ta’ Ħagrat Neolithic sites. This process of erosion began with the appearance of people in Malta and, by the Bronze Age, was advanced. Pollen analyses and high levels of charcoal particles point to fire-stick farming, which encourages

Pushing Boundaries at the End of the Megalithic Building Period natural fodder plants for livestock (Pistacia shrublands).There was evidence also of deforestation and intensive agricultural practice consistent with ‘terracing and planting of hillsides’ that ‘significantly impacted upon soil erosion, with high rates of Holocene sedimentation recorded during the Neolithic period’.55 Similarly, extensive terracing ‘from the Bronze Age’ is seen as an attempt to halt erosion.56 From the analysis of the Marsa sediments, pine was a nil presence within the Bronze Age sediments, olive (Oleaceae) fluctuates and various molluscs point to severe rain and corresponding erosion events.57 Environmental study of the mollusc remains in the Xagħra Circle (Gozo) for the Early Bronze Age (Tarxien Cemetery) contexts supports a degree of stability of the existing open vegetation, which stemmed from activities in the early Neolithic; however, given the enclosed nature of the site, the evidence may reflect localised conditions.58 When the furrow sites are factored into the equation, they complement these findings rather than impede our understanding (Figure 4.1). At this stage, the data point to a downturn in economic productivity with over-exploitation of the environment. These trends may have contributed directly to the redundancy of the megalithic buildings, or they may reflect socio-economic adjustment in the years that followed their decline. Traditional practices may have no longer held relevance within the pan-Mediterranean arena, which was witnessing the increasing mobility and interaction of people from distant lands. For better or worse, these trends were felt equally in the archipelago.59 How long the fields, especially in the extreme locations, endured or whether they had already failed is harder to determine, but environmental data derived from cores made in Marsa and in Salina Bay (both in Malta) and in Santa Marija Bay (Comino), as well as samples analysed from archaeological contexts at Tas-Silg, suggest that there was a ‘collapse in cereal production, most probably caused by regional aridity around 4300 cal. BP’.60 Ultimately, if we adopt a macro and site-specific approach to research into the rock-cut grooves in Malta, we could harness a new and large body of data for economic practice in antiquity drawing on a growing body of data concerning climatic and environmental conditions.61 INTENTIONAL CLOSURE OF NEOLITHIC SITES

Turning to the clear signs of cultural change, the most obvious is that surrounding the megalithic buildings. Some sites were abandoned; others may have served as backdrops to more modest dwellings and, in the case of Tarxien, as a burial ground. As the notions of total collapse and depopulation no longer seem convincing, the archaeological community is currently wrestling with the problem of what came next. Towards the solution is recognition of any interim, possibly hybrid cultural remains in the archaeological record as explaining unmistakable offshore contacts whether these meant foreign



The Archaeology of Malta elements settling Malta and/or the islands’ indigenous folk tapping into the wider Mediterranean trade network and identifying the character of the culture, in all its facets, in the immediate post-Neolithic period.There is no doubt that absolute dating will play a pivotal role in defining the age of deposits and where they fit into the scheme of change; the more dates that are generated, the better will be our piecing together of the evidence. The proposal, explored here, is that the Late Neolithic way of life was brought to a deliberate end within an established ritual framework which finds precedence, as we have seen, in the archaeological record in Malta. Although deposits within houses and within the megalithic buildings are often considered refuse, the discards of the past, there are indications of intentional placement of artefacts as part of ritual acts marking the termination of a building’s use. Deliberate closure is hinted at in the Għar Dalam house (Skorba) with the presence of small amounts – tokens – of fragmentary human remains of three individuals (children and adult) among cultural material. The same can be said of the transported ‘clean’ clay over a range of artefacts in the Red Skorba rooms and for the burning of a Ggantija house at Skorba with contents still in situ, including eleven grinding stones. Purposeful cessation of the Xagħra Circle site in Gozo also has been suggested, based on the internal built features which were pulled down at the end of the Tarxien phase along with signs of deposits that were possible closure levels.62 The lack of roofing material at the megalithic building sites also may be an artefact of decommissioning the ancient structures, particularly if stone capped them, as the artistic representations and models suggest. Aside from the long stone slabs with rebate interlocking ledges at the Misqa Tanks (see Chapter 3), it is possible that other long and, in this case, well-dressed stones found at the Borg in-Nadur site are remnant limestone roof beams removed at the closure of the megalithic building and reused at the site during the Bronze Age. One stone within house 2 excavated by David Trump appears to have traces of a rebate edge on the side; two others were found in 1955 when trenches were cut for pipes, and a standing menhir on Kercem Road, Gozo, is also a possibility.63 Most striking is the treatment of the Tarxien megalithic buildings. Only in recent years has the significance of an approximately 50-cm layer of sand in the complex come under review.64 Trump commented that the layer in the Tarxien building was sterile and that it did not ‘tally with other known natural deposits from Malta’.65 Nor did the sterile sand extend beyond the interior of the building.66 Fire damage indicates that the building may have been burnt before placement of the Tarxien Cemetery burials,67 and the megalithic structure may have been subjected to damage prior to transporting the sandy deposit.68 Even so, the focus on this layer entered the literature primarily for its chronological implications, whether it accumulated naturally over time or was due to a more sudden event such as seasonal flooding – especially in regards to determining

Pushing Boundaries at the End of the Megalithic Building Period the date of the Bronze Age burials placed over the top. In the absence of absolute dates, a slow or rapid accumulation of silt is significant for those looking to find the date of the later Tarxien Cemetery burials.69 Instead, drawing on the evidence of other ancient cultures (discussed later), it is likely that this layer was created through backfilling of the site in antiquity. Possible intentional destruction of parts of the Tarxien complex also has been mooted.70 In addition, every ‘receptacle and niche’ in the Tarxien complex was crammed with bones.71 Zammit assumed that the remains were refuse that had decayed in situ – the remnants of some form of ritual. It would seem that one final action was to load the building with food or animal bones from past significant occasions (ritual meals or offerings, or perhaps the remains represent the wholesale slaughter of animals raised for ritual purposes associated with the building).72 The effort required to backfill the site would have been considerable, suggesting that the communities had the human resources needed to carry out the work. REASONING BEHIND DELIBERATE DESTRUCTION OF STRUCTURES

Instances of site abandonment and ritual destruction of buildings have been well documented elsewhere in both the archaeological and ethnographic records.73 A form of ritual cleansing, such destruction has been interpreted as the ‘death’ of domestic quarters and the final statement in the function of the dwelling. In ethnographic contexts, death of the owner is often the trigger for a house to be brought to an end.74 Architectural evidence at the famous Neolithic site of Çatalhöyük in Turkey suggests ‘a remarkable degree of continuity’ between building phases at the site (i.e., levels VIII to VIA, ca. 5880 BC) that spanned hundreds of years.75 But in level V, continuity was no longer a priority.The site appears to have suffered partial destruction by fire.The buildings that were not burnt were swept clean, purposely filled in and built over without a time lag between the phases.76 Such efforts also were taken for religious sites. Considerable labour was brought to bear to cover most of the ancient site of Tepe Nush-i Jan (western Iran). Here the monumental Central Temple, dated to ca. 750 BC, was carefully filled into a height of 7 m with shale and small stones carried in from local outcrops. Additional soil and stone layers filled the remaining cavity, and a final cap of mud brick covered the wall tops.77 Marc Verhoeven documented further instances of deliberate destruction of buildings in antiquity. Level 6 at Tell Sabi Abyad (Syria) shows signs of abandonment and destruction by fire.The absence of victims in the destruction debris suggests that warfare and accident were not the cause of the conflagration. Wattle-and-daub Neolithic dwellings from the central Balkans usually were burnt, and it is argued that they were torched in closing rituals.These houses had signs of high-temperature burning, flashpoints


The Archaeology of Malta


evident in at various locations in the buildings and a lack of human remains, as well as signs that the structures were knocked down. A number of Pueblo sites in North America were burnt at the time of abandonment, even though the earth building material is by nature not easily set alight. At Chodistaas Pueblo, the burnt rooms were later filled with broken pottery thought to have been the artefacts of the domestic life of the former occupants. When the settlement at Duckfoot, Colorado, was abandoned, seven corpses had been placed in pits, their deposition again a ritual act rather than the result of some disaster.78 SUMMARY

The abandonment of the megalithic structures in Malta must have coincided with some fundamental change in the society that used them, and the end of the buildings was brought about following a similar underlying belief system applied to houses in past generations. Many influences might have acted as stimulus for change, and economic issues may well have been a factor. If the rock-cut grooves are economic indicators of, at first, pushing the boundaries of land productivity through reclamation of marginal zones and, subsequently, of their eventual abandonment or failure, they also may hold one key to understanding the final years of the Neolithic period in Malta. At this point, various interpretations are possible. One might be that Neolithic population numbers exceeded viable subsistence levels in the islands, even with marginal lands converted into new fields. A seasonal downturn of any type could have had a serious impact on such regions and the people who depended on those fields. There is evidence that a new economic trajectory overtook the island during the Early Bronze Age (discussed in Chapter 5), and it cannot be discounted that people walked off the fields and turned their back on the old way of life because they had an economic alternative. The structured fashion of Neolithic closure events has barely registered in the discussions for the island, yet indicators are preserved in the archaeological record.79 Such termination signs draw into question the group of robust statues neatly stowed under a step in the Ħagar Qim megalithic complex. These artefacts may no longer have served a purpose or had social meaning and thus were disposed of in a planned and culturally acceptable fashion.80 Twinned statues preserved only from the lower legs down were reused, facing into the wall fill at Ħagar Qim (no longer on view). Although this carved block was walled up during the building’s active phase, probably when the building was remodelled, the concealed sculptures may indicate that deliberate destruction of iconography was carried out as a Neolithic practice.81 Mutilation of a stone-carved full-bodied statue found in Tas-Silg appears to have been a purposeful act, not carried out by later Phoenician occupants of the site, nor an artefact of damage from recent ploughing. Indeed, any notion of ritual continuity spanning from the Late Neolithic through the Bronze to the Iron Ages

Pushing Boundaries at the End of the Megalithic Building Period cannot be sustained on current evidence; overt signs of religious activity have not been demonstrated, although use of the site in the Bronze Age is not in dispute.82 It is likely that the statue is testimony to the decommissioning of this site during the cultural transition. Perhaps a similar fate lies behind the badly damaged statue in the Tarxien complex, split to the right foot and located well below the surrounding existing standing stones with no fragments left near it.83 Collapse of architecture as well as intense burning of some areas with contents left in situ might relate to closure events or possibly natural disasters such as earthquakes in Tas-Silg; the destruction is dated ‘not later than the end of the Late Neolithic’.84 The sterile soil in the Tarxien complex may not be evidence of a chronological hiatus prior to its reuse as a burial site. A notion of a more prolonged Late Neolithic decline that witnessed sections of the megalithic structures fall derelict while certain areas continued in use has been proposed for Tas-Silg and may well apply to other locations.85 This process has been taken, first, as indicative of ‘ideological and social crisis’ and, second, as the result of transformations stimulated by the arrival of technologically savvy groups.86 A sense of east meets west is conveyed in the boat graffiti scratched on to one of the stones at Tarxien, where presumably local small canoe-like craft crowded around a comparatively large long boat with cabin (Figure 3.20, no. 6). This ancient scene conveys the sense of the end of an era, but it also heralds an unstoppable tide of change sweeping the region.87 Signs of an enduring and viable population, nonetheless, may be masked by an adherence to cultural practices that were not made redundant: pottery technology, the lithic industry, agricultural practices and modest house constructions all may have perpetuated earlier Neolithic traits to varying degrees. Cazzella and Recchia have taken up the notion of a period of transition in the late third millennium, an interim period between the floruit of the Tarxien phase and the fully formed Tarxien Cemetery.88 This presence may have been small scale, and its somewhat circumscribed range of grey pottery and presumably less durable commodities may have been the product of trading ventures rather than settlement, but they were signs of what was to come.



Whether megalithic complexes were deliberately closed, such as those in Tarxien and Ħagar Qim, or like other sites were simply abandoned, sometime around 2400 BC, the Neolithic way of life, its remarkable belief systems and cultural expression, was ruptured. Reuse of some sectors of Tas-Silg and Skorba did take place, but these people left behind a different footprint of building activity on a comparatively modest scale. Quite distinct from previous functions were burials made within the megalithic structure at Tarxien. Environmental downturn and stresses placed on agricultural lands may have taken their toll on the stability and apparently non-combative character of the former Neolithic communities, disrupting inter-settlement relationships and territorial demarcations that were once well honed. A new cultural horizon linked to an influx of people into the islands’ coastal margins may have contributed to a disruption of the Neolithic equilibrium, severing long-standing networks to the interior.1 This may have been a trigger for change that reverberated into the hinterland. Factors were thus set in motion that led to the redundancy of the megalithic structures and the way of life they represented. Increasing social mobility is apparent in the Mediterranean from the late third to the first millennium, and Malta, too, was touched by offshore influences in the period after the megalithic buildings ceased to serve their initial functions. New populations probably occupied the archipelago, and at various times throughout the Bronze Age, Malta was involved in trade. The appearance of metal axes and knives, as well as other signs of cultural practices new to the islands, points to undeniable offshore contacts.2 We may interpret the evidence in Malta as an indication of a wave of immigration during the Bronze Age, or it may signal that a largely indigenous population in transition took on alternative economic strategies, whether they were pushed to do so by environmental conditions or were influenced by newcomers to their shores.3 Emerging from the archaeological evidence, especially from recent excavations in the northern sector of Tas-Silg, are traces of an interim phase with links to the past but clear signs of a new cultural trajectory.We stand 134

New Directions: The Appearance of the Axe-Bearers at a threshold of understanding in regard to this phase that, once crossed, will fundamentally change our perception of the end of the Neolithic, especially in regard to an enduring population. What follows is a survey of the latest trends and the establishment of the subsequent Early and later Bronze Age cultures. A CULTURAL INTERLUDE: THERMI WARE (GREY WARE)

Owing to the early appearance of Thermi Ware (or grey ware) in the Late Neolithic and ongoing sporadic finds into Bronze Age contexts, scholars see it as a marker of both transition and an enduring population. The ware is less likely to bookend a hiatus in the culture sequence than to signify continuity. An interim phase that includes the presence of grey Thermi Ware is a subject gathering momentum.4 Examples have occurred in sufficient numbers and in enough reputedly sound contexts to demonstrate that this kind of pottery made its earliest appearances in the Neolithic. It has been found from within secure deposits at Skorba (assigned as early as the Ggantija phase, which spanned 3600–3200 BC) as well as in Tarxien phase and Tarxien Cemetery deposits (Figure 5.1).5 Examples came from the Tarxien megalithic structure: a small complete bowl fixed onto a high conical stand was found in soil at the rear of an elaborate altar (Figure  5.1, no.  4); one fragment was behind the right foot of the oversized statue in the Tarxien complex and on the floor (Figure  5.1, no.  10), and other fragments were in the northern entrance to space O in niche Q (Figure 5.1, no. 12) and in niche B further north into the complex (Figure 5.1, no. 11).6 More fragments of this ware were recovered to the east of the Tarxien megalithic structures. Dark-faced pottery from this area was both plain and decorated, and Zammit considered it to have an archaic appearance.7 Other examples from the southern sector of Tas-Silg are not the first finds of this ware at that site. Similar fragments in the northern zone were reported by the Italian Archaeological Mission to Malta.8 Evans considered the pottery type to be imported. The Tas-Silg examples, however, do not seem to be made of imported fabric, and chemical characterisation has confirmed that one example at least is of likely local production.9 There is every possibility that hybrid forms of Late Neolithic and Thermi-styled wares were produced on the island.10 Significantly, recent excavations in the northern sector have identified continued use of the site in post-Neolithic times, during what is described as a ‘Thermi–Tarxien Cemetery phase’ (see Chapter 6). The nature of this presence has yet to be fully documented, but it appears to reflect a shift in function away from the cultural and religious practices of the megalithic builders. To date, no overt signs of religious activity have been reported in the Thermi–Tarxien Cemetery deposits.11 Thermi and Tarxien Cemetery wares also were present in Bur Mgħez, a cave used as a burial ground in the Ggantija and Tarxien periods.12



The Archaeology of Malta

Figure 5.1. Thermi-Ware pottery: (1) Tas-Silg (north) (after Cazzella 2002, fig. 2:7); (2 and 3)  Tas-Silg (south), TSG96/2061/9 and TSG96/2134/10 (after Sagona 2015, figs. 1:11:2–3 and 1:152:2–3); (4)  complete pedestalled bowl (inv. no. T/P314) from behind an altar in the Tarxien megalithic building (after Evans 1971, fig. 24:2); (5 and

New Directions: The Appearance of the Axe-Bearers A limited number of forms characterises Thermi Ware, which comprises variations of distinctive, truncated conical and biconical vessels that are handmade, usually from compact and heavy, coarse paste (Figure  5.1, nos. 1–9). A characteristic thickened rim tends to be flat or sloped obliquely into the bowl; some have a slightly re-entrant inner rim edge. Dot-filled triangles are incised on the rim top and less often pendant from the outer rim (Figure 5.1, nos. 1, 2 and 5–7). Adaptation of the decorative motif of spotted triangles could be argued for examples found by Tagliaferro at Ħal Saflieni13 or for the conical bowl with everted rim and continuous spotted triangle motif under the exterior rim recovered by Trump from hut 2 at Borg in-Nadur, the sequence and date of which are currently under revision (Figure 5.1, no. 9).14 The design placed on the exterior of the Borg in-Nadur bowl is not common in Malta, although Cazzella does illustrate an example with a thickened rim with a decoration on the inner lip and on the exterior, as well as a truncated conical bowl from Rudine (Kosova) with the puncture-filled bands on the exterior.15 Indeed, one of the fragments from Tas-Silg (south) has incised spotted triangle designs on the interior rim and exterior wall in the fashion of Thermi-styled bowls (Figure 5.1, nos. 2 and 7).16 Occasionally, chequer patterns with alternate dotted zones (Figure 5.1, nos. 2 and 7) and large spotted triangles on the exterior walls are found (Figure 5.1, no. 4). One example sketched by Zammit from Tarxien appears to depict a complex decoration of stylised horned bull and a bird sitting on its head emerging above a triangle and spotted designs (Figure 5.1, no. 11)17; bird and bull is a motif known in later Aegean contexts.18 As is often employed in Bronze Age Malta, white paste infill enhanced the design, especially effective on the Thermi dark-faced surfaces. Manifestations of similar pottery have been traced as far away as Troy on the west coast of Turkey, the site of Thermi in Lesbos, which gives the ware its name, and coastal sites bordering the Adriatic Sea and Aegean. Nonetheless, direct contact from this far east is generally discounted. Chronologically, this ware appeared around 2300–2200 BC in southern European locations, though antecedents of the type can be seen significantly earlier, in the opening years of the Aegean Early

Figure 5.1. (cont.) 6) Skorba examples from areas OA2 and NA2 (after Trump 1966, fig. 44:h–i); (7 and 8) as for nos. 2 and 3; (9) deep conical bowl with everted rim from Borg in-Nadur, hut 2 (after Trump 1961, pl. 15; reproduced courtesy of The Prehistoric Society); (10) pottery fragment found near the foot of the oversized corpulent statue in the Tarxien prehistoric megalithic building complex; (11) elaborate decoration on Thermi-type sherd from space B, west of passage TE in Tarxien depicting a bird on the head of a horned animal; (12) a sherd from inside niche Q in Tarxien (10–12:  T.  Zammit, notebook no. 13, pp. 100, 101 and 216; held in the archive of the National Museum of Archaeology,Valletta).



The Archaeology of Malta Bronze Age.19 Closer to Malta, specimens have been found in sites in the east of Sicily, for example, Castelluccio, in Lipari (Capo Graziano), and the Apennine region of southern Italy, and significantly, at Coppa Nevigata, the earliest known site of murex purple dye production (see below). This pottery style as it appears in Malta has its closest parallels in the Sicilian site of Ognina.20 Thermi Ware shared a similar distribution path to bossed bone plaques, one of which, from the deposits east of the main Tarxien complex, remains a unicum to Malta.21 Whether the spread of this pottery tradition represents a small-scale migration of people or it was dispersed via trade is an on-going debate, and finds within the late third millennium place Malta early in the process.22 Given the longevity of the decorative technique and limited range of shapes of Thermi pottery, as well as its extensive distribution across the coastal lands bordering the Mediterranean, its enduring popularity might lie in a restricted function; that is, it is not so much an indicator of new populations as of the dispersal of a popular practice. At Tarxien, the complete bowl held burnt matter.23 Its location, concealed behind an altar, points to both the presence of such ware during the late Neolithic Tarxien phase and a possible ritual significance. Such finds might indicate that far from being static or isolated, the megalithic builders embraced trends permeating the region possibly as early as the Ggantija phase but reaching significant proportions by the close of the Tarxien phase.Yet that the wares persist in small numbers at these sites and in habitation remains to the east of the Tarxien complex may be indicative of a post-Neolithic interim culture. TARXIEN CEMETERY PHASE (CA. 2400–1500 BC)

The Early Bronze Age culture is known as Tarxien Cemetery after the type-site where distinctive pottery was first found in significant quantities, in a necropolis of cremation burials within a disused megalithic building. In the initial stages of the scheme devised by John Evans, this phase was designated ‘period IIA’.24 Funerary sites with Early Bronze remains at Tarxien itself and likely habitation deposits, although much decayed, at Xagħra Circle remain the strongest indicators of the period (dated 2400–1500 BC).25 Otherwise, Tarxien Cemetery finds are rare amongst excavation material at prehistoric sites, infrequent in field survey collections and equally poorly represented in Gozo.26 Opinions are polarised concerning the nature of the change from which the Tarxien Cemetery culture emerged.27 Overt links between this culture or its manufactures and the former megalithic builders are limited, and imported artefacts within the cremation deposits clearly point to offshore contacts. Tarxien Cemetery pottery is akin to wares found in Ognina (Sicily), where Thermi Ware, too, is present in contemporary contexts. A relationship between the two wares has long been acknowledged,28 but the issue

New Directions: The Appearance of the Axe-Bearers of a possible Maltese colony in Sicily, although deliberated in the past, is now generally discounted.29 It seems likely that the population living in the islands by the establishment of the Early Bronze Age contained an element of indigenous people still focussed on working the land and living in modest rural houses, as well as merchant-settlers, possibly from southern Italy, who saw the economic potential of engaging a local workforce and harvesting local resources, marine murex dye shells included. Two radiocarbon date ranges have been generated for Malta in the Tarxien Cemetery phase, and depending on how one interprets the cultural sequence – whether there is continuity from the Neolithic or a cultural hiatus  – the high (ca. 2400 BC) or low range (ca. 2000 BC) is accepted.30 The end of the Neolithic, a possible interim phase and the appearance of the Early Bronze Age need to be accommodated within this chronological window. It may be that across the islands, the process of change was not consistent – delayed in some pockets, abrupt in others – and definition may come with a concerted effort to generate radiocarbon dates at every opportunity. It could be argued that reuse of the old Neolithic sites for burials, particularly at Tarxien and in the Ħal Saflieni hypogeum,31 might reflect deep ancestral memories, not forgotten and still revered. Nonetheless, the choice to shut down monumental structures points to a people already undergoing significant change. If these structures served a religious function, then presumably ritual expression also was under revision, especially if the former establishments relied on substantial economic investment that was no longer considered worthwhile or sustainable. In a broad sense, the social order that once focussed on central monuments and on the sites where people (perhaps elites) were laid to rest in large communal hypogea was rejected. Artistic expression through the well-proportioned statues, spots and spiral designs also was made redundant. Any signs of cultural flux and the embryonic stages that saw out Neolithic manufactures and witnessed the evolution of the Tarxien Cemetery Wares, perhaps emerging from Zammit’s archaic handmade grey ware and Thermi pottery traditions, have yet to be defined. Tarxien and Tarxien Cemetery wares found in the Torri Falka cave habitation deposits, a 76-cm grey layer over a well-beaten earth floor (with no obvious strata within the level), are suggestive.32 Some sherds came from the Bur Mgħez cave site, a burial ground in the Neolithic (Ggantija and Tarxien periods). Also suggestive are the finds preserved from the eastern sector at Tarxien; both sites would be well worth closer examination to fully determine the nature of these deposits. Fundamental to this problem is the issue of what replaced the Neolithic way of life, aspects of which are reviewed later. This is a period that would benefit greatly from the generation of absolute dates, even from old collections where short-life samples (human or animal bone, seeds or other organic matter) may have been preserved in the national collections.


The Archaeology of Malta


Tarxien Cemetery Phase Settlement Sites Despite some locations reflecting ongoing occupation, albeit with apparently clear distinctions between phases, the nature of Tarxien Cemetery architecture, other than decayed mud brick, is practically unknown.33 Relatively insubstantial houses close to a few megalithic sites and deposits within caves at Torri Falka, Għar Dalam and Xagħra Circle, Gozo, appear to be habitation sites.34 In the 1880s, excavation (probably not much more than clearing accumulated debris) was carried out at the Borg in-Nadur site35; further investigations were made by Margaret Murray in the 1920s36 and by David Trump in 1959.37 Trump documented Tarxien Cemetery material inside the Bronze Age boundary wall, some sherds within the matrix of the wall itself and pottery in two layers of domestic deposits under the houses of later Bronze Age date (thirty-two fragments have been identified in the museum collection).38 The lower level he identified as having solely Tarxien Cemetery finds (phase 1, cultural period IIA, after Evans), but the upper was ‘a mixture of both phases’, that is, of Tarxien Cemetery and Borg in-Nadur remains (phase 2, cultural period IIB1).39 Recent illegal excavations in the eastern sector of Borg in-Nadur revealed substantial wall lines (of the later Middle Bronze period) and a metre of ash deposit that has been assigned to the earlier Tarxien Cemetery phase.40 Early Bronze Age pottery from the site found during Margaret Murray’s excavations was recently re-evaluated after it was rediscovered in the museum stores in Valletta.41 It has now been demonstrated with the publication of additional sections and a reappraisal of the field notes that the architectural sequence may hold, as determined by David Trump.42 Nonetheless, while problematic issues span the Bronze Age into post-Tarxien Cemetery contexts, the pottery that Trump designated as the latest in the Bronze Age sequence, in situ within house 2, has traits that appear to hail from the earlier Tarxien Cemetery period, perhaps late and degenerated forms. Dot-filled triangles pendant from the lip of a conical bowl, for instance, constitute a design linked more convincingly to Thermi Ware than to decorative techniques on late Borg in-Nadur pottery (cf. Figure 5.1, no. 9 with nos. 1, 2, 5 and 7). Material piled into the second structure (house 1) was considered to be refuse, perhaps a sign of deliberate closure of the dwelling, but its chronological value may not be clear-cut.43 As we are dealing with the fundamental building blocks of the Bronze Age cultural sequence for this period in Malta, the range of wares and their chronological place in the scheme are significant (aspects discussed for the Borg in-Nadur phase later). During excavation (in the 1920s–1930s) of the Roman-period bathhouse complex at Għajn Tuffieħa, quantities of prehistoric and Bronze Age sherd material were recovered.The natural supply of water was no doubt an incentive for settlement.44 Little more has appeared concerning the earlier occupation, with the focus of archaeological investigations falling on the Roman-period

New Directions: The Appearance of the Axe-Bearers finds which dominated the site. Traces of Tarxien Cemetery remains occurring in the vicinity of a few megalithic structures that have come to light after Evans’ 1971 study generally have been underplayed as a transient presence.45 In the southern sector of Tas-Silg, sporadic finds of Tarxien Cemetery pottery have been documented.46 Recent excavations in the north, however, have identified deposits belonging to the Tarxien Cemetery phase, indicating that the site had not been abandoned.47 At Skorba, a dwelling was built within and part using the standing stones of the western prehistoric structure. A wall defining the area of the house was added, and a bench abutted the curved wall of the apse. Pottery fragments were found, and one fragment was in the matrix of the wall. On the basis of a figurine foot in the deposit (Figure 5.4, no. 13) and the lack of a hearth, Trump suggested that the room might have served a religious function.48


Ceramic Production Vessels in the Tarxien Cemetery phase are handmade using a coil technique from generally medium-coarse and compact very dark grey-black fabric. Larger jars can have high conical necks sloping out from the juncture with the body (Figure 5.2, nos. 2 and 3) or slightly swelling and angling in towards the rim with two loop handles (Figure 5.2, no. 8). Other jars and pots display the same trend towards off-set, everted rims (Figure 5.2, nos. 4–6). Irregular rims that dip and curve have been likened to helmets in shape (Figure 5.3, nos. 2, 6, 8 and 9). Bowls vary from deep and conical (Figure 5.3, nos. 11 and 12) to more rounded and squat or carinated (Figure 5.3, nos. 3–6). A high foot can be added (Figure 5.3, no. 3), but most vessels rest on a flat base (Figure 5.3, nos. 10–12). A series of jugs or cups can have handles that are impressed down the length, creating a wide groove (Figure 5.3, nos. 8, 9 and 13); these tend to be attached at the rim and sometimes are drawn higher than the mouth of the vessel before sloping to the body of the pot. Occasional askoi (a type of asymmetrical vessel; Figure 5.2, no. 12) and small, conjoined pots are also known (Figures 5.2, no. 13, and 5.3, no. 17). Aside from the distinct and new range of shapes for Malta,Tarxien Cemetery Wares are often burnished and carry intricate and distinctive geometric designs in tightly packed formations all over the exterior (Figure 5.2, nos. 5, 6 and 10). Groups of parallel lines often border decorative zones comprised of cross-hatched triangles and chequer patterns, zigzags, dot-filled triangles and chevrons (Figure 5.3, nos. 18–21). Fine serrations along the edge and hatched lines within the designs would appear to have been produced with a shark’s tooth or shell edge (Figure 5.3, nos. 18 and 20).49 Although usually lost, enough



The Archaeology of Malta

Figure 5.2.  (1–14) Tarxien Cemetery pottery (after Evans 1971, figs. 25 and 26;Trump 1966, fig. 20; Malone et al. 2009, figs. 10 and 19).

New Directions: The Appearance of the Axe-Bearers

Figure  5.3. Tarxien Cemetery Pottery:  (1–17) pots, jugs, conjoint and tray shapes (after Murray 1934, pls. 22, 24, 25 and 27; Evans 1956, fig. 3; Evans 1971, figs. 4, 25, 26 and 29; Malone et al. 2009, fig. 10.19); (18–21) details of elaborate incised designs with


The Archaeology of Malta


traces of white paste survive to indicate that the decorative features received this further highlight (Figure 5.3, no. 19).

Figurative Artefacts Although the non-ceramic artefacts of this period are largely from funerary deposits found in the Tarxien complex, much can be gleaned from the array of objects documented.50 Among the substantial finds from funerary contexts is an array of disc-bodied (approximately 9 cm in diameter) figurines with stylised pointed heads but reasonably defined legs.The pose of the figures suggests that they are sitting on a stool, which forms a prop behind the legs (see box entitled, ‘Disc Figurines of the Tarxien Cemetery Phase’).


George Azzopardi has argued that figurines with disc-shaped bodies might represent the swaddled or bound state of corpses prior to cremation (Figure 5.4, nos. 1–4 and 11–13).51 This is a fascinating prospect supported by bundles of fine, ochre-coloured cloth found carbonised among the cremated remains at Tarxien:  ‘these masses of burnt fabrics are of a dark or light reddish yellow colour, readily distinguished among the grey ashes.The original dye must have been an iron ochre, for, on analysis, the ashes show that metal in considerable quantity.’52 If Azzopardi is correct, the array of elaborate decorative bindings depicted on the figurines might be significant, reflecting familial or social distinctions of the deceased, somewhat matched by some designs decorating the pottery (cf. Figures 5.2, no. 5, 10, and 5.3, nos. 18–20). It is possible that a second type of figurine, more naturalistic with clearly defined breasts on the body disc but none of the elaborate swaddling designs signified a female association (Figure 5.4, nos. 6, 9 and 10).Whether this accompanied a female cremation or was a representation of a deity at this stage cannot be determined. A second example from the Tarxien deposits has a more naturalistic torso, no breasts and stumps for arms (Figure 5.4, no. 7). Disc figures are now known from Skorba, Tarxien, possibly Tas-Silg and in the Xagħra Circle (Gozo).53

Figure 5.3. (cont.) some traces of white paste infill (19) and fine lines of holes (18 and 20) made with the edge of a shark’s tooth (18–21: reproduced with permission of the Trustees of the British Museum, inv. Nos. 1923, 7–11, 62; 1923, 7–11, 70; POA 199–20; 1923, 7–11, 64; photograph by A. and C. Sagona).

New Directions: The Appearance of the Axe-Bearers

Figure  5.4. Tarxien Cemetery figurines:  (1–4) decorated disc figurines from Tarxien (after Zammit 1930, figs. 3, 4, 7 and 8); (5) Tas-Silg (south) possible figurine stand or prop fragment (inv. no. TSG96/2066/1) (Sagona 2015, fig. 1:10); (6) female


The Archaeology of Malta


Tarxien Cemetery pottery, too, can have an anthropomorphic quality: legs and faces or eyes formed by small knobs, flattened and slightly dimpled applied discs often in pairs (Figures  5.2, nos. 11 and 13, and 5.3, no.  1).54 Once the corpse was cremated, quantities of wheat and pulses, as well as various cultural artefacts including beads, fish vertebrae, shells, animal teeth, small bird figures, animal amulets and carved vessels, as well as tools, hafted bronze points, bronze axes, small pots and so on, were placed – unburnt – into the urns.55 Hollow and simply decorated bird leg bones fashioned into tubes have been likened to similar artefacts appearing in both the Aegean and further east into the Caucasus.56 Like the ubiquitous use of stone for flaked tools in antiquity, however, these tubes may reflect the practicality of the bone structure for use as beads and fine tool handles, which was recognised throughout the ancient world, rather than technological transference between cultures. Outside of funerary contexts, one other artefact recently published by Davide Tanasi from the Borg in-Nadur site might be classed as figurative.57 Nothing similar has been recorded in Malta before. It does bear some resemblance to a more complete fenestrated object from Monte Grande in Sicily, which Robert Leighton identified as a ‘group of schematized figures’, classed among other cult objects from Early Bronze Age domestic contexts (Figure 5.4, no. 8).58

Dolmen Sites, Standing Stones and Grave Markers A desire for substantial grave markers might be reflected in the reuse of the Tarxien Neolithic building. At least thirteen or so independent dolmen sites, possibly serving a funerary function, have been located in the Maltese Archipelago, but the actual number could be greater.59 Evans compared them to similar structures in southern Apulia in the coastal region of Otranto (possibly related to the Cellino San Marco culture), where pottery akin to Tarxien Cemetery has been identified.60 In Malta, dolmen design is simple – smaller

Figure 5.4. (cont.) disc-bodied figurine from Tarxien with moulded breasts and punctures for eyes, nostrils and mouth (author’s drawing after Evans 1971, pl. 55:10); (7) figurine without signs of gender; stumps for arms; simple puncture marks for eyes, nostrils and mouth; pierced ears for rings (author’s drawing after Evans 1971, pl. 55:10); (8) fenestrated stand with incised base (after Tanasi 2011a, fig.  4:10, BN/P74); (9)  disc fragment with breasts from Skorba (after Trump 1966, fig. 43); (10) disc figure fragment from the Xagħra, Gozo (after Malone et al. 2009, fig. 10:70, SF75); (11 and 12) figurine stump legs from the Xagħra, Gozo (after Malone et al. 2009, fig. 10:70, SF816, SF41); (13) decorated figurine of bound legs with defined ankle bones (after Trump 1966, fig. 43).

New Directions: The Appearance of the Axe-Bearers stone piles or two or more large uprights support a horizontal capping stone, forming a low cavity or niche underneath, the depth of which can be enhanced by shallow basins cut into the bedrock; they vary in size from under 1 m to a horizontal slab nearly 4 m across (Figure 5.5). Dolmens are often (but not not always) placed on prominent high ground, in positions where they can be seen clearly from the valleys below and from distant locations. These stone structures rest on rocky slopes that are often devoid of soil or substantial vegetation cover. Linear designs and holes cut into the capping slab have been documented in a few examples (Figure  5.5, nos. 2–4); these features can be shared by Italian examples.61 A funerary function is generally proposed for the dolmens based on the role of similar structures found in Europe. One of three examples at Ta’ Ħammut (Naxxar, Malta, in the hinterland of Qalet Marku)62 was excavated by John Evans, and a quantity of Tarxien Cemetery sherds (of the types in Figure 5.3, nos. 2, 3, 6 and 9, bowls with everted rims, and Figure 5.3, no. 14, a small juglet) were found in a disturbed deposit devoid of human remains. The small cavity formed by the large stone slabs could have accommodated a cremation.63 If, in fact, human remains were once present in these monuments, they eroded away long ago. Evans also found some Tarxien Cemetery pottery fragments in a Wied Moqbol cairn (a small variation on Figure 5.3, no. 12, with a concave resting surface and a handle probably from a jug like Figure 5.3, no. 15, attached at the rim and shoulder).64 A  tray fragment with central dividing wall was made from Tarxien Cemetery burnished ware (Figure 5.3, no. 16).65 Dolmens can occur alone or be gathered together in small clusters such as those at Naxxar (three examples). Several remain in Gozo at Ta’ Cenc in various states of decay (one is called ‘Id-Dura tal-Mara’). Three were once in the region of Ta’ Fuq Wied Filep: one of them, now lost, was perched on the western side of Wied il-Għasel ravine,66 opposite two well-preserved examples at San Pawl Tat-Targa. Two dolmens are in Ta’ Għewra.67 We cannot be sure whether at the time of their construction the stones were above ground or were covered by a mound of earth.68 Either way, they may have served the dual function of a resting place for the deceased and a cultural marker in the landscape, a constant reminder of familial or other ties to the land. It is not beyond the realm of possibility that human remains, whether interred or cremated, were placed in the dolmen with the intention that they would be exposed, to be broken down and scattered by the forces of nature. Standing stones found in Malta have parallels in Italy, notably in Otranto, where they are associated with dolmens (see Figure  3.4, B and C).69 Unfortunately, in Malta, they have suffered over time; examples found in Kirkop and at Sant Anglu ta’Ħal Far no longer stand in their original positions, and others are likely to have been destroyed, escaping documentation.70 In the past, some standing stones proved, after excavation, to be elements of megalithic buildings (Appendix B, List 13).



The Archaeology of Malta

Figure 5.5. Dolmens: (1) Ta’ Cenc area; (2–5) Misraħ Sinjura, Malta, plan (3), lines and hole carved in the capping stone (2, eastern edge; 4, southern side), general view showing modern structure on top of the dolmen (photographs by A. and C. Sagona; plan after Evans 1971, plan 44).

New Directions: The Appearance of the Axe-Bearers

Tarxien Cemetery Phase: Economic Aspects Within the urns of the Tarxien Bronze Age cemetery period was evidence of an agriculture continuum: wheat, barley, peas and horse beans (a form of broad bean). Ample marine resources also were exploited, and the vertebrae of large fish were strung as beads, which suggests that sea-worthy craft were in use for deep-water fishing.71 Bones indicated that animal husbandry included the raising of sheep, goats, pigs and some cattle, as for previous cultures.72 Among the remains of these domesticates were dog bones.The absence of significant excavated settlement sites limits what can be said of daily food-production activities. Nonetheless, Trump argues for a commercial level of manufacture within the economy which may have included metallurgy in Malta.73 Sophisticated boring implements with bronze tips were found within the burial deposits. Exotic lithic material suggests that the focus of contact shifted from Sicily and its obsidian source at Lipari to the outcrops on the island of Pantelleria.74 Ochre, as we will see, may have played a part in the textile industry. Rock-cut and bell-shaped pits with curved or flat bases are a characteristic of the Bronze Age period in Malta, serving a range of site-specific functions.75 These features vary in size from 1.5 to 4.5 m deep. Constructed within and around habitation sites, pits do appear in a few Early Bronze Age contexts. Those at the water’s edge at St. George’s Bay, it has been argued, served an industrial use as vats employed in the production of murex shell dye.76 This purple dye source was highly valued in antiquity; its popularity was brought to a zenith by Phoenician traders during the first millennium in the wider Mediterranean region. Importantly, the earliest known evidence for murex dye works, dated to the late nineteenth century BC, has been found in Coppa Nevigata in Apulia, Italy,77 whereas in the eastern Mediterranean region, the processing of purple dye is placed later; in Crete, it can be assigned to sometime around 1700–1600 BC (middle to late Middle Minoan period).78 Before modern development of the shoreline at St. George’s Bay, this cluster numbered some seventy or more pits. The surviving archive plan in the National Museum illustrates their placement and profiles (Figure 5.6, no. 1).79 This is the clearest evidence that the Bronze Age economy may have included a significant commercial interest focussed on dye production within a textile industry. Spoliation due to proximity to the water’s edge alone would discount their function as food and water storage installations. It should be borne in mind that channels, presumably for fluids, funnel to the openings of numerous examples (Figures 5.6, no. 6, and 5.8, no. 3). Although pits of this kind are generally assigned to the later Bronze Age,80 this particular group could well originate from the Tarxien Cemetery period. Possibly pointing to the early date for the dye vat complex are three isolated examples inside the nearby Borg in-Nadur settlement wall (eastern side) and one found outside the wall



The Archaeology of Malta

Figure  5.6.  (1) St. George’s Bay rock-cut vats, sketch plan with the full range of pits recorded in an archive plan made prior to 1920 when the modern road was constructed and vats were still visible at the water’s edge (plan by the author after Grima 2001, fig. 11:8, from plan CD 100A/62; Sagona 1999, fig. 2); (2) view of the

New Directions: The Appearance of the Axe-Bearers (western side) (Figure 5.6, no. 7).81 In addition, a significant ash layer attributed to the Tarxien Cemetery period overlays bedrock.82 An association between the Tarxien Cemetery finds and the substantial ashy deposits (not linked with burials other than at Tarxien itself, and even there, cremation is thought to have been carried out elsewhere83) might point to industrial activity, whether dye works or other manufactures, in a number of locations. Certainly heating over a number of days played an important part in the purple dye process, which could explain the metre or so of ash deposit at Borg in-Nadur (see box entitled, ' Purple Dye from the Murex Seashell’).84 If the pits can be linked to the Tarxien Cemetery deposits, then they help to explain the success and development of the settlement. Already noted are textiles amply represented in the cremation burials at Tarxien. Quantities of coloured cloth incinerated on the funerary pyre may indicate a lavish supply – probably produced locally – on the island. Reddish yellow coloured fabric suggests that ochre pigments, possibly sourced in Sicily, were applied to the cloth used in the funerary ceremony. Over thirty, predominantly ceramic spindle whorls and weights from Tarxien Cemetery period contexts were found at Tarxien.85 Enigmatic anchor-shaped (or T-shaped) objects first appear in Tarxien Cemetery contexts (and continue into the Middle Bronze Age). They have upturned ‘flukes’, and some can have slight thread-worn grooves; hence they have been interpreted as weaving implements. It is possible that they were handheld devices, possibly for hanging string, wool or netting, and would have served well to hold spun thread if it had to be dipped in dye. First appearing in the Early Bronze Age Macedonian region, they have been located in the Balkans, central Greece and Corinthia, largely in domestic contexts with no apparent geographical preference (coastal, riverine or inland).86 Early Bronze Age Malta falls into line with trends in fabric production elsewhere, where it developed as an important industry and an indicator of status.87 Such issues might be at play in regard to the bundles of cloth placed with the cremation burials in Tarxien, likely to have been as valued as metal work. What has yet to be established is whether the islands of Malta could sustain the necessary numbers of sheep for large-scale wool production or whether part of the textile manufacturing process used plant fibre.There is no reason to discount the possibility that merchants (local or offshore) imported yarn and Figure 5.6. (cont.) vats; (3) detail of vat opening reddened by fire; (4) fine-coil basket imprint probably from a vessel base from Tas-Silg (TSG96 2112/9); (5) bowl with painted basket pattern from Birkirkara found in 1967 (see Sagona 2002, [55], fig. 4:9); (6) similar to no. 5, from Mtarfa, found 26 August 1939 (see Sagona 2002 [299], fig. 69:4); (7) vat cut on the west side of the Borg in-Nadur site, outside the wall line (plan, drawings and photographs by. C. Sagona).


The Archaeology of Malta


thread to be woven into cloth or lengths of woven cloth to be dyed on the island, closer to a source of murex shells or other dye substances. PURPLE DYE FROM THE MUREX SEASHELL

Purple dye was obtained from large numbers of murex seashells that were harvested in baited nets from waters up to fifteen fathoms deep. The glandular tissue, which produced a small amount of dye, was extracted quickly, salted for three days, and then the pulp was steeped in vats with water; the less water used, the more intense was the colour. Heating and skimming of unwanted waste matter played a part in the process.88 The malodorous, rotting matter was an inherent part of dye works; for this reason, vats were usually distant from settlements. Hues obtained from various species of shell differed from reds to blues to purple, and these were further altered by the addition of other chemicals, exposure to sunlight and air and the quality of the fabrics being dyed.89 The need to control the amount of air and light in the dyeing process may explain why channels and smaller vent holes were cut near the mouths of some of the vats in the St. George’s Bay group (Figure 5.6, no. 1) and rebate edges were made to receive lids.90 Channels leading back into the vats probably served to collect precious dyestuff as it was squeezed out of cloth.91 Most of the dyes needed a mordant to aid absorption of the colourant and to fix them into the fabric, improving both colour and lustre. Common ancient mordants were wood ash, leaves, roots (especially the perennial madder root), tannin and urine.92 Experimental work, however, has found that murex dye is colour-fast.93 Despite the difficulty in producing murex dye, it was highly sought after because of its brilliance, resistance to fading and durability.94 Given the early date for murex purple dye production in the central Mediterranean, and until more evidence comes to light, it can only be suggested that Malta may well have been a significant player in the early development of the murex purple dye industry; just how early hinges on fixing a date for the St. George’s Bay pits. This industry may have travelled along maritime links with the Apulian region already demonstrated via distribution patterns of Thermi-styled pottery, bossed-bone plaques and, later, dolmens and anchor-shaped objects.95 A textile industry on a commercial scale is certainly supported by the data as they stand. If we return to the question of what replaced the Neolithic way of life, especially after it was so definitively put to rest, a different path can be traced than utter collapse and abandonment of the islands. No longer defined by a strong archaeological footprint centred on the Neolithic monumental

New Directions: The Appearance of the Axe-Bearers structures within defined territories, the Early Bronze Age population has left a less distinct settlement pattern. Although still in the realm of speculation, the long-established Neolithic networks, whether controlled by families, religious hierarchy or dominant landowners, may have been undermined by new commercial opportunities. By choice, significant numbers of indigenous Maltese islanders might have shifted their economic focus to the production of cloth and later to dyed fabrics. The industry may have gathered momentum with a strong focus at St. George’s Bay, where dye vats are found in great numbers. The settlement at Borg in-Nadur, in particular, would appear to have steadily prospered, developing into the Middle Bronze period, when it was eventually defined by a settlement wall. Signs of wealth are reflected in the permanent disposal of valuable items such as bronze axes and blades, as well as abundant artefacts, some exotic, of personal adornment, in the cremation burials at Tarxien. It is likely that the impetus for change came from abroad and with it new cultural trends that influenced religious practices and technology, but it is likely that signs of an enduring indigenous population survive in the domestic contexts, so ill-defined at this stage. Co-existence of indigenous people with new settlers may explain the varying modes of burial practice, from reuse of the Tarxien megalithic structure, to internments within the existing hypogea, to other practices possibly marked by dolmens.96 How long such changes took to solidify is not clear, but these events may have occurred within a few generations. In Malta, a similar acceptance of change would be repeated when the Phoenicians arrived in the islands during the first millennium, but more of that later. ASPECTS OF THE EARLY TO MIDDLE BRONZE TRANSITION

What happened at the end of the Tarxien Cemetery phase is just as problematic as the closing years of the megalithic building period. Questions of continuity between the Tarxien Cemetery period and the following Middle Bronze, Borg in-Nadur, phase remain. On the basis of David Trump’s excavation, there was a clear interplay between the closing years of the Tarxien Cemetery and later Borg in-Nadur phases. What this means in terms of cultural succession also must be teased from the archaeological record as it stands. Occupation of this site was apparently without hiatus, although an influx of new settlers has long been argued for in the Middle Bronze Age.97 Some pottery forms, for instance, carry over into the next phase. A distinct segmented tray from Wied Moqbol (cf. Figures 5.3, no. 16, and 5.10, no. 1) is one survivor.Tanasi also comments on a period of transition at Għar Mirdum, which he dates broadly to the last half of the sixteenth century and first half of the fifteenth century BC. Like so many cultural phases in Malta and Gozo, the story remains only partly told because the archaeology of the islands is so fragile and under-represented in publications.98


The Archaeology of Malta


THE BORG IN-NADUR PHASE (1500–CA. 1000/700? BC)

Broadly referred to as the Middle Bronze Age of Malta, the Borg in-Nadur period (named after the eponymous site) is well represented across the islands. Many locations retain significant remains from the period, including smaller rural houses or hamlets, cave sites and well-defended settlements.99 A distinctive new pottery industry infiltrated the islands at this time and is generally taken to indicate another wave of settlers. In the absence of signs of hostility at the local level, the transition appears to have been peaceful, with clear indications of both Tarxien Cemetery and Borg in-Nadur people co-existing together at Borg in-Nadur. There is still a possibility that changes in the material culture were stimulated by increased maritime activity, and new ceramic styles which appeared in the islands might not necessarily equate with an extensive migration of people to their shores. Nonetheless, a number of sites during the Middle Bronze Age that were fortified or were chosen because they were naturally more defendable might indicate that the island was subject to periodic hostilities. This falls in line with trends in southern Italy and Sicily, especially for coastal sites.100 Of the religious and funerary practices of the day we know virtually nothing, and the little that can be teased from the archaeological record will be surveyed here.

Fortified Sites In more vulnerable settings or driven by a desire to secure access points, inhabitants increased the defensibility of their homes by building settlement walls. The best surviving example is on the hillock at Borg in-Nadur. It would seem that over much of the low coastal rise, Bronze Age remains can be found around and possibly reusing, standing stones of a substantial megalithic complex of the Tarxien period.101 Limited excavations have been conducted north-west of the prehistoric structures (Figure 5.7, no. 1). The extent of the Bronze Age settlement within the perimeter wall was approximately 3.4 ha (or 8.5 acres), and the population is thought to have been large (Figure 5.7, no. 2).102 The buildings excavated in the 1880s are poorly documented103; thus dwellings uncovered by David Trump remain our principal source of published information. In his 1960s investigations, it was found that an interim phase  – phase 3  – witnessed the extinction of Tarxien Cemetery culture and its replacement by Borg in-Nadur traditions. In Trump’s scheme, a fourth phase produced house 1, comprising preserved stone foundations of an oval structure (3.5 m × 7.5 m) with torba floor.The contents were clearly domestic: hearth, lower (quern) and upper (roller) grinding stones and pottery (cultural period IIB2). House 2 in phase 5 was partially excavated and was found to be relatively more elaborate, with a hearth in the room and a fireplace apparently blocking the entrance of house 1, which, it is argued, probably had been abandoned at

New Directions: The Appearance of the Axe-Bearers

Figure 5.7.  (1) Borg in-Nadur plan of site (after Evans 1971, plans 1 and 2; N. C.Vella et al. 2011, fig. 3.4); (2) exterior of the Bronze Age wall (north-east side) (photograph by A. and C. Sagona).



The Archaeology of Malta that time.104 A long dressed stone with shallow rebate depression facing upwards lay across the 3.5-m-wide room, which approached rectangular in plan with a torba floor. The stone would appear to have been reused and might have once formed part of the nearby megalithic structure. Although evidence of upper structures was not preserved for either house, they are presumed to have been wattle and daub. Much the same array of items was left in situ in house 2 as in house 1: a quern and roller as well as a better represented but noticeably different phase of pottery (designated cultural period IIB3).105 The nature of the pottery in house 2 has some features more akin to Tarxien Cemetery than to Late Borg in-Nadur. Entrance into both houses was originally gained through one of the long sides of the buildings, as was also the case for the dwellings exposed in the 1880s.106 Regardless of the heralded reappraisal of the 1960s excavations, neither house may, in reality, represent definitively the last of the Bronze Age sequence (see Chapter 6).107 As already discussed, illegal digging in the eastern sector of the site in 1998 and development works in 2001 uncovered further traces of the settlement wall, and the published photographs clearly indicate a major structure. These wall remains were erected over an earlier house with a small oven, also of (unspecified) Bronze Age date.108 Overall, reports of the wall leave some ambiguity in regards to its date. In 1998, ‘investigation of these remains established conclusively that they date to the Borg in-Nadur phase.’ In the reappraisal of the 1960s excavations, a wall shown in Trump’s original plans which stood parallel to his demarcated ‘defensive wall’ line, as well as associated levels banked up against it, were dated to the Tarxien Cemetery period. It remains to be seen whether this structure represents the inner face of the principal settlement wall.109 Il-Wardija ta’ San Gorg (limits of Siggiewi, Dingli Cliffs at Gebel Ciantar) was excavated by Tancred Gouder in 1973, but the findings were not published.110 Remnants of a large stone fortification wall protected the natural and effectively the only land approach to the Bronze Age village.The site is stepped as it slopes sharply to its southernmost point, and steep cliffs characterise its naturally defended location. Built with large blocks, the wall appears to have had a few building phases, some of them recent (Figure 5.8, no. 1). Within the settlement walls, fourteen bell-shaped pits cut into the bedrock, several with rebate openings, are still clearly visible (Figure 5.8, no. 3). Small channels cut across the exposed rock face fed water into the pits. Two stepped rock-cut rectangular depressions, rounded at the corners, formed part of the footings of small Bronze Age dwellings; the interior house walls were edged by low and narrow benches (Figure 5.8, no. 2). This region has significant archaeological remains that include dolmens and the cave site of Għar Mirdum (see later); otherwise, all the features of the peninsula settlement have been assigned to the Middle Bronze Age.111 Natural steep cliffs form defences for the site of Il-Qlejgħa tal-Baħrija (Figure  5.9, nos. 1 and 2), which has been archaeologically excavated by

New Directions: The Appearance of the Axe-Bearers

Figure  5.8. Wardija ta’ San Gorg:  (1)  view looking south to the Bronze Age site and beyond to the small uninhabited island of Filfla; (2) one of the rock-cut house depressions with interior bench; (3) rock-cut pits and channel to the north of the house (photographs by A. and C. Sagona).



The Archaeology of Malta

Figure  5.9.  (1) Plan of Baħrija (after Trump 1961, fig.  4); (2)  view south-east to Baħrija from Ras ir-Raħeb (photograph by C. Sagona).

New Directions: The Appearance of the Axe-Bearers Thomas Eric Peet (1909) and David Trump (1959). The site covers an area of approximately 7.5 ha.112 Peet identified two to three semi-rock-cut house footings with torba flooring, roughly rectangular in plan, and benches like those at Wardija ta’ San Gorg.113 Remnants of wattle and daub and fired bricks also were recovered. Baħrija pottery, some decorated with relief oblique bars, which can be a characteristic of the late Borg in-Nadur, and numerous small finds were located in the deposits.114 Trump excavated pottery-rich strata,115 which were assigned to the Bronze Age; no architecture was found. Baħrija pottery, considered contemporary with the later Borg in-Nadur (IIB3), appeared in the upper stratum at Baħrija overlaying a deposit from a lengthy Borg in-Nadur-period occupation.116 In recent years, Maria-Elena Zammit carried out a field survey.117 Ancient remains on high ground of the adjacent Mtarfa-Rabat-Mdina plateaus also can be classified as naturally fortified sites typical of other Bronze Age locations in Malta. Notice of Bronze Age discoveries have appeared over the years by Themistocles Zammit and, notably, by Nathaniel Cutajar in recent times. Finds under Santa Margerita Cemetery were made in 2006; more were identified within the walls of Mdina (1960 Vilhena Palace courtyard118); and bell-shaped pits, cut through and exposed, lie under Mdina’s fortification walls (further pits are west in Mtarfa).119 At 24–25 Inguanez Street, Mdina, house renovations permitted archaeological investigations beneath the existing floor level that revealed Borg in-Nadur pottery fragments. They were sealed under a mortar floor of the Late Antiquity period.120 To what extent, if at all, the Borg in-Nadur settlements occupied the intervening valley, or whether it was otherwise used, remains unknown. As the area is still characterised primarily by open farm plots, the potential remains for future investigation, especially employing non-invasive radar and standard field survey techniques. Aside from these sites, little can be said of finds at Il-Qolla (west of Burmarrad), where some Borg in-Nadur sherds, bell-shaped pits and traces of a boundary wall were identified on a small plateau.121 Middle Bronze Age sherds were identified at San Pawl Milqi, and it has been suggested that a naturally strategic site might have been at the top of Gebel Għawzara.122 Sections of a fortification wall were recorded at Fawwara (Ras il- Gebel) and Qala Hill, which are located near Għajn Tuffieħa and San Martin, respectively, on the Wardija Ridge. Il-Qarraba occupies a peninsula between the bays of Għajn Tuffieħa and Gnejna; its steep sides and sole access via a narrow bridge of land make this a naturally well-defended site.123 A small tower-like structure was located in the north-eastern sector, but no other ancient structures have survived. Some sherd scatters and flint material have been documented in the fields near the site.124 A degree of visibility between sites may have been a factor in site location, especially those along the Wardija Ridge.


The Archaeology of Malta


Non-Fortified Sites Of the numerous non-fortified sites, the small rural settlements that grew around the farming communities of the interior, some are represented by only small sherd scatters, others by rock-cut pits or remnant architecture. Although predominantly of the Punic-Roman period, at the Carlo Diacono Girls’ Lyceum site in Zejtun, two Bronze Age bell-shaped pits containing Borg in-Nadur pottery were found during ongoing archaeological investigations.125 Ninety fragments of Middle Bronze Age pottery and limited cultural pockets were identified in the Xagħra Circle, Gozo.126 The two fragments which are illustrated in the final report (bowl rim and thumb plate, similar to Figure 5.11, no. 10) appear to belong to the floruit of the industry. A dry-stone wall on the south-eastern perimeter of the site hints that a settlement may lie to the south that is yet to be investigated.127 John Linton Myres drew attention to pottery held in the Valletta museum from tombs in Bingemma. He also mentions sherd scatters in the field terraces in the valley, including handmade examples with dark core and red burnished slip, which are likely to be Bronze Age.128 Pottery of Borg in-Nadur type has been found in the material recovered in both the northern and southern sectors of Tas-Silg in recent years. From Ħal Millieri came seventeen sherds, although none were diagnostic.129 Salvage investigations in response to council drainage works at It-Tokk, Victoria, Gozo, also retrieved Bronze Age pottery fragments.130

Cave Sites Caves may have served as dwellings or perhaps as the setting of rituals, as yet poorly defined, in the Bronze Age.131 Għar Mirdum (‘Buried’ or ‘Subsided Cave’ in Maltese) was investigated by members of the Dingli Speleological Expedition under the supervision of Francis S. Mallia in 1964–5. This important site is recorded briefly in the museum report of the day and on an informative website compiled by one of the chief investigators, Paul Calleja-Gera.132 The cave succumbed to natural erosion and now comprises chambers located deep within the cliff wall that are linked by passages and crawlspaces.133 The excavators thought it served as a habitation site because of the quantities of bone recovered: ‘Where potshards were excavated, numerous bones (animal) turned up as well.This is a fact for every corner of the cave.’134 Signs of burning and carbon also were noted throughout the archaeological deposits. The site has yielded Bronze Age material in accessible areas, though this was possibly disturbed from its original position by the rock fall.135 Davide Tanasi’s recent reappraisal suggests that the chronological span of the site, based largely on pottery typology, is considerable though fleeting for some periods: Zebbug, Tarxien Cemetery, Borg in-Nadur (IIB1, IIB2, IIB3), Punic and Roman.Very little of the pottery (3,508 fragments), predominantly of the floruit of the

New Directions: The Appearance of the Axe-Bearers Borg in-Nadur Bronze Age,136 has been illustrated thus far, and the complex deposits are fraught with difficulties. Despite the less than ideal contexts, some significant artefacts were recovered, which indicate that the community that frequented the cave were not impoverished. Indeed, a bronze blade and a decorated bone handle are unique to the island.137 With regards to the number of sites in Malta, little is known of the finds made in other cave locations. In effect, most of the smaller sites have received only cursory mention, and although the data in existing archives and museum collections are under review,138 the available accounts rarely move beyond stating to which period the finds belong, and full reports of field projects are rarely prompt. Inside the Għar Dalam cave occupation, debris was found in the upper level, called the ‘surface layer’: ‘The cave floor contains on the surface Phoenician and Roman remains, with Bronze Age and Neolithic potsherds in plenty, and but little deeper down.’139 Għar il-Friefet, a cave site south-east of Għar Dalam, was excavated in 1963.Two trenches exposed stratigraphic deposits containing Borg in-Nadur wares as well as bones from domesticates (cow and goat).140 A ‘considerable amount of Bronze Age pottery’ was found in Għar ir-Riħ, located in cliffs on the southern coastline between the district, Il-Mara and Għar Ħasan.141 In Wied ta’ Manna in Mellieħa, a small cave had been used for habitation during the Bronze Age. A tomb (ill-defined in the brief report) and a cistern also were noted.142 Il-Mixta, along the north-eastern edge of Għajn Abdul plateau (limits of Santa Lucija, Gozo), is the site of two ‘roofless’ caves where excavations revealed Borg in-Nadur habitation deposits.143 ASPECTS OF MIDDLE BRONZE MATERIAL CULTURE

Ceramic Industry While the general periodisation of the Bronze Age in Malta is known, the on-going revisions of past excavations and current work at Tas-Silg (north) continue to refine the pottery repertoire.144 The potter’s wheel would not arrive in Malta until Phoenician settlement, and all vessels were built by hand in the Bronze Age. One dominant Middle Bronze Age category takes its name from the type site of Borg in-Nadur. This distinctive range generally has thickly slipped red surfaces (though some vary in red hues, others have blackened and blotchy patches; Figure 5.11, nos. 1–6) and incised decoration which was enhanced with white paste in-fill (probably a gypsum and milk paste mixture; Figure 5.11, nos. 5 and 6).145 Notable is the tendency for the red slip to craze into fine cracks all over the surface (Figure 5.11, no. 2). Likely serving as the finer wares of the culture, the repertoire comprises a large range of jugs (Figure 5.11, nos. 15–18), cups (Figure 5.11, nos. 7–10 and 11–13) and bowls (Figure 5.11, nos. 29–32). High-swung handles (Figure 5.11, nos. 11–13) and vertical thumb plates (Figure  5.11, nos. 8–10), as well as



The Archaeology of Malta high bases, are characteristic features of the cups (Figure 5.11, nos. 26–38). One noteworthy form is the bi-conical pedestalled bowl, not only shaped like an hourglass drum but commonly decorated with incisions and punctures that mimic the lashings used to secure leather over the open ends, top and bottom (Figure 5.11, no. 24).146 Whether such vessels doubled as drums or merely imitated them remains moot, but archaeological reconstructions of similar vessels in European contexts suggest that the latter is possible.147 Large jars can be closed with off-set, narrow necks angling into the jars (Figure  5.10, no.  16)  or conical and flaring necks (Figure  5.10, nos. 2–4). Pans or trays are shallow and flat based with straight (Figure 5.10, nos. 11–13 and 15)  or curved walls (Figure  5.10, no.  14). Deep pots have un-featured (Figure 5.10, no. 10) or slightly flaring rims (Figure 5.10, nos. 7–9). Some lids (Figure 5.12, nos. 4 and 5), and two types of lamps have been documented – a shallow vertical-sided dish (Figure 5.12, no. 2) and a pedestalled shallow bowl (Figure  5.12, no.  1); both have handles for ease of carrying. Pottery spoons, though rare, are found in Middle Bronze Age contexts, and it is possible that greater numbers were made in wood (Figure 5.12, no. 3). Indeed, certain decorative features, especially applied small pellets, may mimic nails (Figure  5.11, no.  4), and hatched lines might resemble string (Figure  5.11, no. 3). Middle Bronze Age pottery comprises more than one ware type often overlooked in the archaeological record in Malta for the prehistoric period. Recent studies have given attention to the coarse wares, but there is scope to define the categories more fully in the future.148 Tas-Silg pottery which is contemporary with the common red-slipped Middle Bronze Age wares similarly warrants sub-divisions; the types are all handmade but distinct. A painted category of pottery is often decorated with thick, roughly applied red spots covering the interior and bands of paint around the inner rim; occasionally linear designs are on the exterior (Figure 5.12, nos. 6–14). Another category is White Gritty Ware, thick-walled and with a large quantity of distinctive, angular quartz-like particles throughout the coarse-textured paste (Figure 5.12, no. 15), an aspect of the pottery also observed by Peet at Baħrija.149 This fabric is largely used in cooking ware such as open pots with loop handles and large jars sometimes decorated with small knobs near the rim (Figure 5.12, nos. 16–20). A drab and coarse ware used for domestic vessels also has been documented as used for household utensils: cooking pots, pans, high-handled bowls or cups, sieves and spoons.150 Fragments of sieves with closely pierced holes over most of the vessel wall have been documented at Borg in-Nadur and Tas-Silg. In the former site, they have been recently assigned to the Tarxien megalithic period, and in the latter, they fall within the Bronze Age (Tas-Silg, period II).151 Trays, some segmented (Figure 5.10, no. 1), and possibly lamps are made of coarse ware and are known from Borg in-Nadur contexts.152

New Directions: The Appearance of the Axe-Bearers

Figure 5.10.  (1–16) Borg in-Nadur pottery shapes: tray divided into two segments, open trays and various jars (after Evans 1971, figs. 2 and 3; Tanasi 2011a, figs. 4.24, 4.25, 4.32 and 4.34).



The Archaeology of Malta

Figure  5.11. Borg in-Nadur pottery; details of surface and decorative features:  (1)  red-slipped fragment with typical flaking surface (TSG96/158/8); (2)  crazed thick red-slipped surface (TSG96/2163/155); (3)  incised decoration imitating string (BM 1923, 5-9-68 [22]); (4) incised grooves with small nail-like pellet

New Directions: The Appearance of the Axe-Bearers

Borg in-Nadur Economic Activities Grinding stones were recovered from the houses at Borg in-Nadur, which indicate that food processing was carried out in domestic quarters.153 Bell-shaped pits cut into the rock in many Middle Bronze Age sites are often drawn into the discussion concerning food production and water storage, but there is no proof that the large numbers of pits in some sites point to commercial levels of grain production as an export commodity.154 Other large clusters – over forty have been located at Baħrija – equally could have served alternate semi-industrial functions, perhaps in the textile or basketry industry. Though some pits may have served to hold potable water, inter-connected pits at Bir Miftuħ, for instance, are cut into porous Globigerina.155 Signs of plastered surfaces in some cavities served to overcome such flawed bedrock. There are pits that were reused during the Phoenician period for tombs.156 Bell-shaped pits have been reported over the years and as recently as 2008, when five were found during extension work at Luqa Cemetery.This brought the tally to around twenty-two pits, some plaster lined, which have been found in that area.157 Roman-period refuse and other deposits were found in a few, but by and large the pits had been cleared of contents. One (no. 16) excavated in 1966 was found to contain ‘a quantity of Bronze Age sherds and stone artefacts, one possibly a mortar’. In another pit (no. 17), emptied by the landowner, were four notched pillars thought to have formed the supports for a horizontal loom. For decades, the soil deposits of one pit at Tal-Mejtin (a district in the same region) excavated by David Trump remained the only source of information regarding the ancient Bronze Age environment.158 Pottery from this pit was classed as Late Bronze Age, but around seventeen pits were identified at this location, one with Tarxien Cemetery material. It should be recalled that deposits in pits could have a ritual value; this was certainly possible in Malta’s Neolithic phases. Such depositions may indicate deeper social memory that spanned cultural change and might hint at a degree of population continuity. Impressions on the resting surfaces of coarse pottery at Għar Mirdum and Tas-Silg indicate that tightly woven, medium-fine-coil baskets or matting probably were made on the island and were used as a base on which pottery was built (Figure 5.6, no. 4). Some pottery also can be painted to imitate basketry (Figure 5.6, nos. 5 and 6). Part of the preparation of the many straws and

Figure 5.11. (cont.) of clay (BM 1923, 5-9-15); (5)  incised and punctured decoration with white paste in-fill imitating string lashings (BM 1923, 5-9-53); (6) as for 5 (BM 1923, 5-9-68 [21]); (7–44) pottery shapes (4–6: reproduced with permission of the Trustees of the British Museum, photographs by A. and C. Sagona; 7–44: after Evans 1971; Tanasi 2011a, figs. 4.17, 4.23, 4.24, 4.26–4.28, 4.30–4.31 and 4.35).



The Archaeology of Malta

Figure 5.12. Borg in-Nadur-phase pottery variations from Borg in-Nadur: (1) lamp; (2)  shallow lamp; (3)  spoon from Tas-Silg (TSG96/2112G/4); (4)  lid from Borg in-Nadur (BN/P153); (5)  lid from In-Nuffara (NNF60/P/09/18); (6–14) painted

New Directions: The Appearance of the Axe-Bearers fibres used in basketwork can involve soaking the raw materials in water to make them more pliable, and fibres could be dyed for use in elaborate designs; the rock-cut pits could have served a role in these processes. Dyes can be drawn from many natural sources, not just marine, and such activities might have taken place across the island, for example, in regions such as Mtarfa, where clusters of pits, from several to forty or more, were documented in a few locations across the plateau.159 A continuing textile industry is suggested, not just by the rock-cut pits but also by the spindle whorls and loom weights, including T-shaped (or anchor-shaped) implements which appeared on the island during the Tarxien Cemetery period and are also found in Middle Bronze Age contexts. Maritime activity, a topic fundamental to the study of an island community, has been rather overshadowed. Some doubt the ability of Borg in-Nadur communities to conduct inter-island trade. To discount the ongoing maritime abilities of the ancient Maltese islanders at any time during the Bronze Age or to suggest that populations were not large enough to sustain offshore trade seems an unnecessary and overly cautious approach, yet such are the arguments that continue to be made.160 This situation would imply a loss of important technological know-how that ferried settlers to the islands in the first place and also that the population did not anticipated a declining supply of suitable timber (if this was ever completely the case) and take steps to avert the problem. Moreover, to suggest that the ancient Maltese somehow lost their seamanship skills overlooks the island’s geographical situation at the hub of the Mediterranean, close to Sicily and North Africa, and most likely a port of call at the very least. That Malta stood outside the general trend in Mediterranean mobility is unlikely.161 A population that was organised enough to build sizeable hilltop settlements, defended by substantial walls, probably also was capable of conducting maritime ventures. In the constrained interpretation, which says that it was not, Malta becomes an endpoint rather than holding a place within a ‘chain of networks’. It views Malta as a cultural vacuum into which the accomplishments of the wider Mediterranean filtered while it offered little in return. Mariners crisscrossed the seas exchanging objects and more importantly, they conveyed technology, skills and knowledge. The region was a changing arena, moulded Figure 5.12. (cont.) bowls from Borg in-Nadur (6), In-Nuffara, NNF60/P/2009/1, NNF60/P/2009/2 (7 and 8), Ħal Saflieni (9–12) and Tas-Silg (13–14); (15) detail of Borg in-Nadur White Gritty Ware from Tas-Silg (TSG96/2101/17); (16–20) Cooking pots made of White Gritty Ware (TSG96/2112/23,TSG96/303/29,TSG96/2112/5,TSG96/2101/17); [after Tanasi 2011a, figs. 4:38, 4:40 and 4:41 (nos. 1 and 2, 4 and 5); Tanasi 2008–9, fig. 8:a-b (nos. 7 and 8); Sagona 2015, figs. 1:19, 1:20, 1:160 and 1:245 (nos. 3, 13, 14 and 16–20); Evans 1971, fig. 43 (no. 6); Murray 1925, pl. 25 (nos. 9–12)].


The Archaeology of Malta


and defined by cultural developments, perhaps even by the integration of people from neighbouring islands through migration or inter-marriage.162 Arguments that Malta’s inhabitants did not or could not negotiate a way off the island imply an unequal standing with their immediate neighbours and within the greater Mediterranean. The proposition that the Maltese were trapped, deliberately blocked by Sicilian neighbours,163 is out of kilter with the archaeological evidence. Within the limitations of the archaeological record in Malta for the Bronze Age, any comparison with the ample evidence available for the contemporary period in Sicily needs to be approached with caution. Rather than dismissing Malta as out of the way and marginal at this time, off the beaten track for Mycenaean and Cypriot traders, other factors may be at play. Some sectors of eastern Sicily would appear to have had a strong impulse to acquire exotic goods, reserving certain imported vessels (at least) and possibly the commodities they contained for funerary events.164 Pottery from the Aegean and Malta figured among the imported array included in burials presumably of the elite sector of Sicilian societies. The comparative dearth of imported pottery (Aegean or otherwise) in Malta simply may indicate deeper and more conservative cultural traits and local economic forces. We are at a disadvantage in the later Middle Bronze Age (Borg in-Nadur period) because of an almost total absence of burials in the archaeological record, but imported wares are not common at the known settlement sites.165 Essentially, pottery was an export commodity from Malta, especially in the Middle Bronze Age, but what the home market sought in return is less tangible. A separate range of forces in which imported Maltese manufactures played a role drove Sicily’s ancient markets. The archipelago might not have been ‘excluded’ from the wider trade networks of the Mediterranean but rather worked within an economic niche of its own design.166 Of the funerary practices during the floruit of the Middle Bronze Age, we know virtually nothing:  skeletal remains were found in Għar Mirdum and burials were mentioned at ‘Ta’ Ħammut, Wied Moqbol, Racecourse Street (Gozo) and Ta’Vnezja’.167

The Closing Years of the Middle Bronze Age It may be more than coincidence that a repeated range of goods, most notably the tools of daily sustenance – upper and lower grinding stones as well as pottery – were left in houses at Borg in-Nadur at the end of their occupation. Similarly, querns, weights and T-shaped objects which were sometimes left in the pits may represent symbolic placements. The same also might be said of a cache of artefacts found in one of a number of pits in the Mtarfa Plateau. Devoid of a burial, the pit contained significant pottery, including a Phoenician-style twin-nozzle lamp deposited in the cavity. Though chronologically important,

New Directions: The Appearance of the Axe-Bearers the deposit also may embody a symbolic gesture (the chronological significance of this find is discussed in Chapter 6). Ultimately, one has to question why the houses were abandoned and the original function of the pits ceased. Deposits may point to a thread of enduring cultural practice, perhaps sustained through the historical memory of an indigenous population. A  shift away from past traditions and changes in economic activity with the appearance of Phoenicians on the islands may have been a contributing factor. Recent re-evaluations of inter-island links between Sicily and Malta have made significant progress, especially based on finds from Paolo Orsi’s past excavations. Pottery of Borg in-Nadur type has been identified at nearly a dozen sites in south-eastern Sicily, from Catania (itself among the cluster) to Calafarina in the south. Cozzo del Pantano in particular has been the subject of recent revision. Although the issue of direct imports versus local wares imitating the Maltese type continues to be debated, it may well be that prominent sites such as Thapsos in Sicily acted as a hub for maritime traders, including the Maltese islanders.168 Maltese Bronze Age pottery made its way to Sicily and vice versa, and a few Mycenaean fragments in Malta indicate that the islands were not isolated from the wider commerce of the Mediterranean. One Mycenaean IIIB (ca. 1300–1230 BC) fragment came from Borg in-Nadur, but from less than clear contexts.169 A few other fragments have come from Tas-Silg.170 One body sherd was found in secure contexts in clear association with Borg in-Nadur material of a type designated by Trump as IIB1, that is, with Red-Slipped Borg in-Nadur Ware with its distinctive burnished surfaces.171 Presumably, aside from pottery, perishable products made locally were exchanged but have not survived in the archaeological record. Cultural and chronological links to the east of the island and to Sicily are present among the finds in Għar Mirdum. A pedestalled cup or lamp with loop handle spanning from the rim probably to the foot (missing) and with the rim barely pinched to form simple nozzles was recovered. It finds parallels in hut 2 at Borg in-Nadur, excavated by Trump and assigned to his IIB3, and in two examples from the Għar Dalam cave site that had three nozzles on the rim and signs of burning.172 Significantly, from grave 1 at Thapsos in Sicily, excavated by Paolo Orsi, came an example of this pottery form dated to the Sicilian Middle Bronze Age, Thapsos II (TE IIIA2), which falls between 1400–1380 and 1310–1250 BC.173 Such evidence calls into question the long-standing chronological assumptions for the period that favour low dating. SUMMARY

Throughout the Bronze Age there is evidence that the lands of the wider Mediterranean felt the rise and fall of regional economic pressures. Malta was not immune to these trends, and some local communities met perceived threats, whether island based or from offshore raiding parties with fortified sites. What



The Archaeology of Malta role indigenous peoples played, whether active or passive, in the infiltration of Levantine settlers into the islands and coastal regions of the Mediterranean rim is a story that is still unfolding. The transition from one dominant culture to the next in Malta’s archaeological record appears rather seamless. There are no destruction levels, nor human remains left where they fell in their villages, nor caches of weaponry; instead, there are unmistakable signs of enduring indigenous practices set against a wave of unstoppable change. Products such as textiles, pottery and possibly basketry may have been made for offshore markets. As there is ample evidence of a continuing cloth industry, it could be argued that its products may have been a commercial incentive for the Phoenicians to target the island in the opening years of the first millennium BC. By this time, the Borg in-Nadur communities were well established throughout the islands, and the Levantine settlers appear to have quickly infiltrated, unopposed, into the hinterlands. It is this complicated but intriguing story of adaption for both groups that is the subject of Chapter 6.


For pre-literate cultures living around the coastal rim and on the islands of the Mediterranean, their history must be pieced together from archaeological data region by region. In the east, though, written accounts tell a story of troubled times. Turmoil in the 1200s witnessed the emergence of bands originating in the Aegean and possibly the western Mediterranean who, along with local groups, set their sights on the lands along to the Levantine coast to Egypt. Which nations comprised these bands of militants is still the subject of debate, although there is a significant emphasis placed on an Aegean component which carried with it to the lands it would eventually settle the hallmarks of Mycenaean culture; hence, the migrants are often referred to indiscriminately as the ‘sea peoples’.1 Political instability ensued in the region, and the numbers of these sea peoples may have swollen with members of displaced and disaffected groups from the areas they infiltrated. This was a time of economic disruption, driven perhaps by drier environmental conditions and famine that began in the late thirteenth century BC; some scholars refer to economic collapse with a corresponding Aegean ‘Dark Age’ at the end of the Late Bronze period which affecting the eastern Mediterranean.2 The marauders were not just pirates at sea, content with plunder gained from sporadic coastal raids. By the reign of Ramesses III (ca. 1194–1162 BC), their numbers included women and children travelling overland in wagons,3 lending weight to the argument that some bands of warring invaders grew to include people from local regions. Egyptian monuments depict fully armed warriors with distinctive headdresses, equipment and clothing characteristic of their nations in a flotilla of seaworthy vessels – a veritable navy – and significant land and sea battles that were fought to defend its frontiers. Egyptian and other city-states in the region provide intriguing lists of perpetrators, documenting the nature of their aggression. Texts indicate that within the ranks of the sea peoples there may have been Sardinians (Shardana)4 and North African Libyans. Eventually, groups appear to have settled along the Levantine coast. Claims that in el-Ahwat in Manasseh, 171


The Archaeology of Malta southern Levant, the ruins of a fortlike structure are similar to the nuraghi of Sardinia have been met with scepticism on cultural and chronological grounds.5 Other evidence concerns the Philistines (the Peleset), an Iron Age confederacy whose pottery bears strong links to Mycenaean wares and who stamped their name on Palestine. Despite the complex, mixed and mobile nature of populations in the Iron Age Levant, it has been possible to pinpoint some localised outcomes. Ayelet Gilboa and Ilan Sharon, for instance, argue that Tel Dor (southern Levant) served as home to one group, called Šikila, who were among the sea peoples’ contingency.6 How distant Malta figured during this time is unclear. It is unlikely that the islands of the archipelago had the resources to contribute in any systematic way to the raiding parties. If unrest also afflicted lands in the central Mediterranean, then the Bronze Age people of the Maltese islands may have been preoccupied with their own defences against a maritime threat. Certainly, Sardinian bronze figurines (broadly dated from the thirteenth century BC and after) are suggestive of that nation’s militant character.7 Perhaps the impetus for Bronze Age fortified settlements in Malta was the circumstances which engulfed their northern neighbours. Nonetheless, issues of isolation8 are less relevant during the middle to late second millennium in the central Mediterranean as a whole; the extent of offshore commercial interaction in the final decades of Bronze Age Malta is not clear, but at Baħrija (Malta) influence from the Calabrian Fosse Grave culture is evident in the pottery repertoire.9 Maritime commerce and well-travelled sea routes can be traced in the central region well before 1200 BC. As we have seen, during the Bronze Age of Malta, contact with Mycenaean traders is marked by a few artefacts found in settlements, artefacts which may have passed through the entrepôts of Sicily, where Aegean pottery is found in greater amounts. In a wider context, Sardinian-Tyrrhenian trade is evident throughout the Villanovan and Etruscan periods.10 Suffice it to say that hand in hand with the influx of people into the Levant, whether they were bent on hostility, immigration or trade, came their knowledge about resources and potential wealth, especially in the distant lands of the greater Mediterranean. The significance of this intelligence about the West should not be underestimated as a catalyst for the events that followed.11 How the end of the Bronze Age in Malta unfolded has been the subject of speculation for some time. The lack of clear interface contexts, within the limited data for the period, has fostered some erroneous notions that the islands were abandoned, becoming effectively terra nullius for the next influx of colonists.12 Indeed, life in the Maltese Archipelago was about to change. Sturdy seagoing ships brought a new range of people to its shores whose distant homeland lay over 1900 km east, along the Levantine coast. This was the vanguard of Phoenician explorer-traders, who led the way for a wave of settlers who would make Malta and Gozo theirs in a process of colonisation that played out over the course of perhaps many decades (ca. 1000–750 BC).

East Meets West Because of the patchy archaeological record, signs of hybridity take on greater than usual significance for the islands at this crucial time of change; they are in a sense ‘proof of life’, indicators of an enduring local population and cultural interaction with Levantine travellers. Borg in-Nadur people probably continued to live in and around their ancestral settlements. The environment was much as it had been throughout the Bronze Age; the island was largely deforested, with its population reliant for daily sustenance on small herds of domesticates, marine resources and a staple of reliable, hardy crops such as barley and pulses. A commercial arm to the island economy is also feasible. An assessment of the indigenous culture close to the arrival of the Phoenicians is the focus of this chapter. THE LATE BORG IN-NADUR PHASE

At issue is the nature of the Late Bronze Age culture leading up to the arrival of the Phoenician settlers. It is a pressing question in Maltese archaeology. Within the later part of the Bronze Age, a minor sub-phase or cultural identity known as ‘Baħrija’ also was present in Malta. As yet there is no unequivocal settlement context (ideally, a floor level) with unambiguous late Borg in-Nadur or Baħrija material occurring alongside foreign, early Phoenician wares or their local imitations. The nature of the evidence is such that more than one interpretation of the surviving data is possible. At the heart of the problem are the following: the sequence and date of the two oblong houses at Borg in-Nadur13; the late Bronze Age levels at Tas-Silg (north), which are said to form a complete sequence but are still under investigation and not fully published14; possible contemporary levels at Mdina (also unpublished)15; a tumulus with Bronze Age and Phoenician wares in Gozo, found in 1923 but yet to be analysed16; the Mtarfa pit site in Malta, which held a cache of undeniably late Borg in-Nadur pottery and an early Phoenician lamp17; and the date of and possible role played by the Baħrija culture. No small part of the argument is the nature of the Phoenician infiltration of the western Mediterranean. Scholarship has steadily recognised that this involved a long process of human interaction leading to eventual systematic colonisation by the Phoenicians. Malta has clear, perhaps unmatched evidence of this process.The path through these issues follows. After David Trump’s identification of levels at Borg in-Nadur and Baħrija as the last in the Bronze Age sequence, it was assumed that this period came to an end with the Phoenician occupation of the island. Trump’s excavations were considered to have filled the sequential gap between the Bronze Age and the Phoenician settlement period, traditionally placed in the mid-eighth century BC (Trump’s IIB3 and IIC).18 At the time, in the 1960s, the best date for this event in Malta came from a Phoenician tomb found in 1950 at Għajn Qajjied. It held two datable imported Greek cups (Figure 6.8, nos. 7 and 8),


The Archaeology of Malta


but otherwise the Phoenician-Punic sequence at that time was poorly understood. As it was generally considered that the Phoenicians settled the western Mediterranean in the eighth century BC and not before, Trump’s date for the end of the Bronze Age was thus placed around 750 BC. There are still no absolute radiocarbon dates in Malta for this period; datable material has been found throughout the years, but the opportunity has never been taken. The date for the Għajn Qajjied tomb may have endured, but since then, it has been clearly shown that the Phoenician pottery industry reflected in this tomb had already undergone significant evolution on the island.19 There are numbers of tombs with Phoenician-style wares that display the progress of that evolution, which logically predate 750 BC. In essence, the tombs hold evidence of the Phoenician infiltration of the island prior to the main colonising period of the Mediterranean. In the scheme devised for the Phoenician-Punic period in the archipelago based on the funerary evidence, the floruit of the systematic colonisation period has been termed the ‘established Melitan phase I’, and the evidence of Levantine contact prior to that time has been coined the ‘archaic Melitan phase I’ (an Orientalising period equating with trading contact and sporadic settlement).20


‘Phoenicia’, ‘Phoenicians’ and their Punic Descendants The term ‘Phoenician’ is an artificial construct coined by ancient Greek authors when referring to these Levantine traders and settlers; to Greeks, their combined land was Phoiníke and its people phoínikes. The term ‘Phoenician’ signified the colour purple-red and the ‘purple folk’, a reference to their acclaimed trade in purple cloth. After the sixth century BC, the western Phoenicians are generally known to modern scholars as ‘Punic’ from the Latin term Poeni, a broad term for Phoenicians encompassing the people, cities and settlements under the Carthaginian hegemony.21 When the archipelago enters the literate world, with the arrival of the Phoenicians from the Levant, we learn the names at that time of the two largest islands. Malta in Greek sources was known as Melite and in Latin texts as Melita. Gozo was Gaulos or Gaudos to the Greeks and Gaulos or Gaulus in Roman circles.22 These words are thought to derive from Punic mlth and gwl (for Malta and Gozo, respectively).23 The culture exhibits a unique blend of probably local and eastern elements which evolved along a distinct path, as it did at many of the Phoenician colonies in the central and western Mediterranean. Melita and Gaulos were not purely Phoenician, nor a carbon copy of Carthage, the most influential of Phoenician settlements.

East Meets West In time, Greek and Roman influences also would be felt in Malta. ‘Melita’ and ‘Melitan culture' are used here to cover the similar cultural developments in Malta and Gozo through the first millennium BC and into the early centuries AD.The term ‘Melita’ reminds us that the archipelago developed culturally along its own path, despite the ancient territoriality, politics and offshore influences, and it is applied generally to the cultural sequence comprising six phases devised for the islands in funerary contexts (see Table 6.1). Melita also embodies the notion that the islands had moved on from the anonymity of prehistoric culture.

‘nn (Anna) The ancient Punic name of ‘nn is found on coinage minted in Melita (Malta) in the late third to early second century BC discovered, for instance, in the excavation of Għar ix-Xiħ in Gozo.24 Some consider that ‘nn was the ancient name of Malta, whereas others link the name to the tradition of Anna and her flight from Carthage (see box ‘The Legend of Anna, of Phoenician Royal Lineage’). As the legend may have originally concerned the town of Cyrene and its king, Battus (and it was from Cyrene that Malta borrowed iconography for its own coinage), the identification of ‘nn as Malta’s name may not be valid.

If Borg in-Nadur house 2 is as late in the Bronze Age sequence and in date as Trump suggests, then it is remarkable that there is no known (or at least no published) Phoenician material (imported or local hybrid forms) at all from the entire Borg in-Nadur settlement. Indeed, there is very little later Punic pottery from the site (the Punic people were the cultural descendants of the Phoenician settlers) (Table 6.1).25 Notwithstanding these issues, it is premature to make definitive statements beyond outlining the problems as they stand, as the full publication of the village at Borg in-Nadur, as excavated by David Trump, is pending. Already an array of additional excavation sections and other data, not published before, have thrown new light on the excavation. Critical to the sequence is the nature of the deposits within the two structures and lying over the shared party wall of the houses. These layers, distinguished by Trump as distinct stratigraphic units, have been amalgamated in the recent reappraisal of 2011, and the meshing of these layers does alter the interpretation considerably.26 We will have to wait for the full final report to determine whether this conflation is justified. In ­addition, the Italian team working in the northern sector of Tas-Silg has heralded an independent development, a new Bronze Age sequence for Malta.27 In view of the current state of research, the best evidence for the closing years of


The Archaeology of Malta


Table 6.1.  Phases of Phoenician-Punic Malta based on funerary evidence Phase

Approximate date range

Broad cultural development

Melita I (Archaic)

ca. 1000–750 BC


750–620 BC

 Late Melita II

620–600 BC 600–500 BC

Melita III

500–410 BC 410–300 BC

Melita IV

300–100 BC

Melita V

100 BC–ca. 50 AD

Melita VI

ca. 50 AD–ca. 2nd century AD

Archaic phase I: period of trading contact and sporadic settlement (Orientalising period) Established phase I: fully fledged Phoenician colonisation Late phase I to early phase II Period of introversion: minimal foreign influence ‘Classic’ Punic: early phase III ‘Classic’ Punic: late phase III to early phase IV Incipient Roman influence: covers the Roman conquest of Malta in 218 BC Increasing Roman influence on the local pottery repertoire; culturally Punic Roman (increasingly dominant)–Punic (dwindling cultural facies)

the late Bronze Age – a ‘snapshot’ of the late Borg in-Nadur pottery repertoire in its final form – remains the material from Mtarfa pit. Pottery from recent salvage work in Mesquita Square, Mdina, has indicated that a settlement was located there, but only preliminary notice of the finds has been given. Remains of a substantial ashlar building or possibly a Punic fortification wall (see Chapter 7) were uncovered above Borg in-Nadur levels.28 One vessel which has been published29 does point to the Melitan phase I archaic period, and mention has been made of a fragment of Sicilian Finocchito Ware that dates after 730 BC, found with Phoenician red-slipped pottery of the early seventh century.30 Although this is not the earliest evidence for Phoenicians in Malta, the finds do demonstrate that their cultural influence moved quickly into the interior of the island in contexts that were not just funerary.31 Further evidence in Palazzo Xara (Mdina) comprised 3 m of deposits, including beaten-earth floors, hearths and domestic handmade wares, as well as imports, which Antonia Ciasca examined and dated broadly from the late eighth to early sixth century BC.32 More finds came from public works activities in Villegaignon Street adjacent to the convent of the Carmelite order.33 On the eve of World War II, in 1939,Ward-Perkins excavated a small number of bell-shaped pits on the central plateau of Mtarfa. These features can still be seen in a small reserve but are now hemmed in by high-density urban development. A few pits were found to contain archaeological deposits.Though the pits themselves probably dated to an earlier phase in the Bronze Age, some

East Meets West of the remains were clearly funerary in nature. One contained quantities of pottery, and although the find is not without controversy, it is the clearest indicator of the nature of the late Borg in-Nadur pottery repertoire. When it was excavated, Ward-Perkins considered it undisturbed. At the lowest level was a Phoenician-style lamp, basically a shallow bowl with two wick nozzles pinched in the rim, a typical form of the Phoenician lamp in the West, which he considered in situ: It had evidently been rifled in antiquity, but in the lowest levels were found a great quantity of broken pottery and a number of animal bones. The pottery formed a homogenous group, and it will ultimately be possible to reconstruct a number of vessels. These are of characteristic late Bronze Age fabric and include the remains of two large cinerary urns with handles and a number of smaller vessels decorated with white incrustation. The most important find from the lowest levels of the deposit was a lamp of characteristically Punic form which in the absence of any other trace of intrusive material must undoubtedly belong to the original tomb furniture. The existence of fabrics transitioned from the Bronze Age to the Punic period has already been recognised by Dr J. G. Baldacchino but this is the first case in which the transition has been archaeologically demonstrated.34

Some time later, in his report in the journal Antiquity,Ward-Perkins gave a different interpretation of the finds in the pit: ‘It had been rifled in Punic times, but it still contained, in addition to a lamp left by the robbers, a quantity of broken Bronze Age pottery.’35 Whether the lamp is an integral part of the pottery cache or was introduced into it determines how we interpret the deposit; it reflects on the chronology of the material, the manner of its deposition and aspects of interaction between local and foreign peoples at this time. The presence of the lamp in such a context would suggest that it was funerary; lamps are certainly a common inclusion in Phoenician burials on the island. As Ward-Perkins did not record any human remains, the material must be described as a pottery cache, though among the pottery were fragments of a large baggy urn of a type that in other tomb contexts was used for cremated remains. Nonetheless, the fragmentary vessels the pit contained were rightly considered representative of the final years of the Bronze Age in Evans’ prehistoric sequence of Malta.36 Significantly, the lamp and most of the other pottery in the cache were built of the same chalky reddish yellow fabric.37 Despite the passage of time,Ward-Perkins’ statement that this was the first indication of a Bronze Age–Phoenician transition is still pertinent. In isolation, this Mtarfa pit is significant. When considered with the other pits in the immediate vicinity, as well as in light of its general location, it has wider implications.38 The original site probably was associated with a Bronze Age settlement, and clusters of bell-shaped pits can be seen in other sites such as Baħrija and Wardija ta’ San Gorg.They may have been used for water or food



The Archaeology of Malta storage, but equally, they may have served a role in manufacturing processes (dye, textile or basketry works) of the settlement (see Chapter  5). Although the plateau is now largely capped by modern urban development, there may be some scope to detect signs of habitation in surrounding slopes and fields. Whatever original function the pits served in the daily life or economy of the Bronze Age, it had ceased by the time the Phoenicians arrived. Clearly, the Bronze Age culture had experienced some fundamental change by the time the Phoenicians infiltrated the island’s interior. As the Mtarfa find is the best example of the late Borg in-Nadur repertoire, it is worth pointing out its characteristics. Pottery shapes in the pit range from large baggy jars, some with narrowing vertical necks (Figure 6.1, no. 3) and others with horizontal handles on the widest part of the body (Figure 6.1, no. 2). Jars have open and wide-hole mouths (Figure 6.1, nos. 4 and 5); handles can be placed under the rim and high on the body. Small jugs, perhaps better described as tankards, have a swelling, cupped neck with strap handle from neck to shoulder (Figure 6.1, no. 1).Various bowl fragments are present in the pit; most are shallow with gently curved walls (Figure 6.1, nos. 10–14); and others are deep and conical (Figure 6.1, nos. 6–8), the latter continuing the fashion of hourglass-shaped biconical forms common at the height of the Borg in-Nadur period (e.g., Figure 5.11, no. 24). Trays are wide and shallow with straight sides and horizontal basket handles (Figure 6.1, nos. 18 and 20); one example has a small depression on the rim, perhaps used as a spoon rest (Figure 6.1, no. 19). An unusual angular leg also was present in the pit (Figure 6.1, no. 22) which may have come from a small table-like stand, of a type known in the Thapsos tombs of Sicily.39 Compared with the repertoire at the floruit of the Borg in-Nadur Bronze Age phase (see Figures 5.10–5.12), it is clear that some forms had evolved, but overall the range had contracted. Deep conical bowls with splayed handles rising above the rim (Figure 6.1, no. 16) would appear to further a tradition begun in the Tarxien Cemetery period (cf. earlier forms, Figures  5.3, no.  9, and 5.11, nos. 11–13). Other features have fallen from use, such as bowls with a thumb plate at the top of the handle (see Figure 5.11, nos. 8–10), high-footed bowls (see Figure 5.11, nos. 26–28) and stemmed shallow bowls that functioned as lamps (see Figure 5.12, no. 1). New forms of lids (Figure 6.1, nos. 15 and 21) are represented. A pyramidal loom weight in clay also was found, testament to an enduring textile industry (Figure 6.1, no. 9). Vessels are modelled by hand from clay that is soft and eroding (Chalky Reddish Yellow Ware), so soft that the fragments shed an orange-coloured dust when handled. Firing is generally inadequate, resulting in a dark-wall core but reddening towards the surfaces. Clays have a moderate to large amount of gritty inclusions in the paste. Surfaces still carry a red slip, which is often mottled with brown to blackened patches from the kiln. Unlike during the

East Meets West

Figure 6.1.  Late Borg in-Nadur pottery, a cache found in a Mtarfa bell-shaped pit excavated by John Ward-Perkins in 1939 (after Sagona 1999).



The Archaeology of Malta floruit of the period, the surfaces tend to be left dull and matt rather than burnished. The earlier burnished forms tend to have a finely crazed surface, a characteristic generally not present on the later dull wares. Decoration certainly had moved away from the incised horizontal and oblique clusters of lines on the rim, base and obliquely across the vessel so typical of the established Borg in-Nadur phase. Bowls in this late phase, when incised, had lines at the rim, often incorporating an off-set zone, likened to a swallow’s tail in shape (Figure 6.1, no. 13).White paste in-fill still can be used to enhance the incised designs. Small pellets of clay, perhaps mimicking metal rivets or nails, can decorate the walls or outer edges of handles (Figure 6.1, no. 5). In the absence of actual metal vessels, such decorative features may indicate that they were at least known to the local community in Malta and that the islands were drawn into the wider trade in metal wares in the Mediterranean. Applied bars are well represented (Figure 6.1, nos. 7 and 10), as are small solid lugs (Figure 6.1, no. 11) or loop handles (Figure 6.1, nos. 12 and, 14). These applied decorations are found on only a few sherds from Borg in-Nadur.40 Sherds from a hard-fired fabric (Thin-Walled Variant Ware) were found among the sherds of the pit. They came from a large, wide jar with distinctive roughly incised and white in-filled decoration comprised of hatched triangles and haphazard rows of zigzags (Figure 6.2, nos. 1 and 2). The kind has been found at Borg in-Nadur among the early material excavated by Murray and recently at Tas-Silg.41 On present evidence, it would appear to be a local variant, one possibly limited to a narrow time frame late within the final years of the Bronze Age but as yet not represented in early Phoenician contexts. Hence, as a cultural and chronological marker, these examples offer useful points of connection from site to site within the islands. Despite the absence of clear contexts, at least we can state that it is no longer viable to see Malta as having been in a cultural hiatus, depopulated and free to be occupied by Phoenician settlers.42 There are clear signs that the local Borg in-Nadur people endured and that their manufactures gained new inspiration from abroad. Equally erroneous is an approach that considers the entire Borg in-Nadur period as homogeneous from its appearance to the point when Phoenicians first set foot on the islands. Current research is tackling the nature of the stratigraphic evidence, reviewing the closing years of the Bronze Age cultural sequence and the associated chronological implications. A positive move forward is the reassessment of the excavated material from Borg in-Nadur.43 No less significant is the continuing research of Clive Vella, who is systematically working through the surviving collections of flaked tool material from the excavations in Malta, Borg in-Nadur included.44 Pottery in the last Borg in-Nadur phase also could carry a red slip. If the colour was due to ochre, as this was possibly sourced offshore, this red pigment could have continued as a sought-after exotic commodity. Fortifications

East Meets West

Figure 6.2.  (1 and 2) Late Borg in-Nadur phase, thin-walled variant fabric from jars, found at Tas-Silg (TSG96/2107/3) and Mtarfa (after Sagona 1999, fig.  5:3). Baħrija pottery: (3) Jug with incised linear decoration (B/P6); (4 and 5) bowls with up-swung handle; (6)  bowl rim fragment from Tas-Silg (TSG96/2067/1); (7)  typical excised decoration on the exterior of a pottery fragment that would have been filled with white paste (reproduced with permission of the Trustees of the British Museum, inv. no. POA 199/7); (8 and 9) high-shouldered bowls; (10) bowl with elaborate excised decoration (B/P1); (11) shallow plate from Baħrija; (12) flat lid, (B/P7); (13) crudely formed lamp with three loosely pinched nozzles on the rim (B/P4) (6 and 7: Sagona photographs; drawings after Sagona 1999, fig. 5:3; drawings after Evans 1971, figs. 10, 11 and 41).


The Archaeology of Malta


built during the floruit of the Borg in-Nadur, if not at its very beginning, still endure, but to what degree they were actively maintained during the life of the Bronze Age settlements has yet to be proved. Over the span of the Borg in-Nadur period, the evolution of the pottery industry can be traced from its first appearance through the established years representing the floruit of the culture to the final stage, by which time the industry had lost some if its quality. Whether or not we accept the re-evaluation of the Borg in-Nadur finds, it is likely that the simple ovoid form of house construction continued; certainly nothing more substantial has been identified in the archaeological record dating to the end of the Bronze Age in Malta (see Figure 5.7, no. 1). Similarly, it cannot be determined whether the dye vats at the water’s edge at Borg in-Nadur were still functioning when the Phoenicians arrived on the island, but these masters of purple dye production may have recognised the potential of murex shellfish living in shallow waters off the coast. Murex shells, the waste product of the industry, however, have been found in the upper levels at Tas-Silg, which may indicate that the Phoenicians produced purple dye in Malta.45 BORG IN-NADUR FUNERARY PRACTICE

No formal burials can be assigned strictly to the Borg in-Nadur phase from its established years through to the years just prior to Phoenician contact. We can, however, make some inferences using the earliest of the Phoenician funerary evidence. It is possible that the Late Bronze Age people marked the graves of significant individuals with shallow tumuli associated with stone circles, as suggested by an example in Strada Corsa (now Republic Street), Victoria, Gozo, which has no identified parallel on the island.46 Stone circles of the prehistoric period, likely to predate the Late Bronze Age, have been documented (Appendix B, List 12). Such constructions are unlike Levantine funerary traditions, and the practice may have local origins. No rock-cut chamber or shaft (typical of Phoenician-Punic traditions) was reported in association with a burial; instead, the remains were found under a stone slab. Similar tumuli have been found in Andalouses (Oran Coast, Algeria),47 and grave circles were recorded at Acholla (el-Alia) in Tunisia.48 The latter may have further significance when considering that the tradition documented by Stephen of Byzantium in the sixth century AD identified Acholla as a colony founded by the Phoenicians of Malta (in the third century BC; see later).49 In addition, a pit burial at Qallilija (Malta) was dug at the foot of a megalithic stone which probably served as a grave marker.50 The upright was one of several that were removed prior to 1924, before Paul Bellanti investigated the site. He thought that a second stone had been disturbed when the pit was dug, indicating that the upright stones may have predated the burial. Certainly the pottery urn with its crescent decoration found in the Qallilija cremation burial is a form

East Meets West that carried over into later tombs which show Phoenician influence. In other words, stone markers of this type might be a Bronze Age practice that was current up to the arrival of the Phoenicians and for a short time afterwards. Like so many periods in Malta’s archaeological record, the data are scarce and often limited by their early discovery and reporting. If such markers were common to all the islanders, then we could expect there would be more of them in the archaeological record. Overall, the absence of Borg in-Nadur burials is curious and might point to funerary practices that left little or no trace. We know nothing of religious life in the late Borg in-Nadur; no temples and no figurines have been found. Religious continuity at Tas-Silg has yet to be proved. Some motifs on the pottery may have significance in this regard. During the height of the Borg in-Nadur, the thumb plates sometimes have stylised designs: spurs that resemble horns, dots and lines that may have a symbolic significance. Crescent motifs, too, can be found on some of the pottery in the final years of the Bronze Age and into Phoenician contexts (Figure 6.1, no. 6; cf. Figure 6.6, no. 6). Evidence shows that belief systems in Malta after Phoenician settlement eventually took on strong characteristics of Levantine religious practice, but they were not unadulterated, and it may be possible that certain elements hailed from the local culture rather than from the new one. BAĦRIJA PHASE

Some scholars argued that the Baħrija cultural interlude brings the story of the Bronze Age to its end – to the arrival of the Phoenicians.51 Basically, if scholars envisage a Phoenician presence in Malta no earlier than the mid-eighth century, then they lower the date for the Baħrija culture accordingly. Though hybrid late Borg in-Nadur–Phoenician pottery can be detected in tomb contexts, the same cannot be said for the Baħrija remains. As the Phoenician presence in Malta would appear to date to the ninth century at least, we could expect some tell-tale sign of co-existence of the cultures on the island, but as yet no unequivocal published context has both Baħrija and Phoenician (or hybrid) wares together. There is no context in the published record of the excavated houses at Borg in-Nadur, nor in the Mtarfa pit, which indisputably held the Late Bronze Age material on the island. The absence of cultural overlap may indicate that the Baħrija phase predated the Phoenician period, but the evidence is far from conclusive. Trump did note that the distinctive dark-faced pottery appeared alongside Borg in-Nadur pottery. It is likely that Baħrija culture is a regional sub-group of the Bronze Age, contemporary with the Borg in-Nadur, probably occurring in the later (but not final) years of the culture. On the basis of the finds to date, the Baħrija phase had minimal cultural impact on the island outside the type site; fragments are found in very small quantities at Tas-Silg, Borg in-Nadur (Murray excavation) and Għar Dalam.52 This distribution may point to trade between Baħrija and these sites


The Archaeology of Malta


in the south-eastern part of the island, perhaps involving the distribution of Maltese chert, found in deposits immediately west of Baħrija.53 Nonetheless, as for the floruit of the period, more definitive data concerning the final years of the Bronze Age are under revision.54 The Bronze Age levels at Baħrija unfortunately were devoid of architecture. Trump turned the focus of his fieldwork to Baħrija (see Figure 5.9, nos. 1 and 2) after excavating the houses at Borg in-Nadur. Dark, highly decorated pottery was first recognised at Baħrija during Peet’s excavations, and the greatest concentration of Baħrija cultural artefacts was found there.55 As we have seen, the site had Borg in-Nadur levels (pottery, rock-cut pits, house footings and other finds) indicative of a long Bronze Age cultural presence.56 Only estimates have been mooted for the Baħrija phase chronology, and these range from the eleventh century to the eighth century BC.57

Ceramic Tradition Baħrija pottery is black at the core with slipped and black-burnished surfaces. Most distinctive are deeply excised and intricate geometric designs comprised of zigzag, linear and punctured patterns often all over the exterior surfaces of vessels (Figure 6.2, nos. 5–7 and 10). Shapes so far identified in the record are hemispherical bowls (Figure 6.2, no. 10), shouldered bowls with a carinated profile (Figure 6.2, no. 8) and flaring lipped bowls with high-swung handles (Figure 6.2, nos. 4 and 5). Flat lids (Figure 6.2, no. 12), jugs with slightly swelling necks (Figure 6.2, no. 3), plates (Figure 6.2, no. 11) and shallow-handled lamps with rims pressed to form multiple lamp nozzles (Figure 6.2, no. 13) are also present in the repertoire. Although the fabric and striking designs stand apart from the late Borg in-Nadur industry, some similar shapes  – jugs, flat lids and high-handled bowls  – appear in both. Strainer spouts from marine contexts and from Peet’s and Trump’s Baħrija excavations are not local forms but may reflect Sicilian influence.58 Links to Sicily have been identified for the Baħrija phase, as they have for previous Bronze Age phases, notably with Thapsos and Grotta Calafarina.59

The Need for Cultural Definition Renewed excavations at Tas-Silg north (2003–present) have uncovered deposits dating to the final years of the Bronze Age and the interface with Phoenician settlers (Figure 6.9, no. 2), finds which included some Baħrija remains.60 These excavations also have targeted prehistoric levels north-east of the extensive Punic temple architecture, the platforms on which the Punic structures were erected and beside the prehistoric megalithic ruins first exposed in the Italian excavations of 1963–70 (Figure 6.9, no. 1).61

East Meets West Giulia Recchia and Alberto Cazzella have suggested that the Borg in-Nadur–Phoenician interface period should be termed the ‘Baħrija phase’.62 While recognising the need for clear terminology, this particular term might lock this chronologically problematic and geographically circumscribed phase into the sequence. Unless the deposits have a dominant Baħrija character, distinct from Borg in-Nadur, such an attribution might give the phase more prominence than is warranted. Borg in-Nadur influence is strong in sites such as Mtarfa, and hybrid pottery can be detected in the archaeological record in early Phoenician contexts, which need to be considered.The current evidence supports the possibility that the Baħrija cultural tradition does not figure during the interface period. In many respects, this is an ‘Orientalising period’ for Malta, a cultural turning point that witnessed an influx of Levantine social, political, religious and funerary practices. If a term must be found for it, then this  – Orientalising period  – would be appropriately generic. Whether the ‘Baħrija phase’ to which Recchia and Cazzella refer overlaps with the opening years of Melitan I (as determined for funerary contexts) will require clear evidence of Phoenician cultural contact.

Sites with Baħrija Finds Soon after the Italian excavations began in the 1960s, however, speculation grew that the prehistoric megalithic structure had continued to function as a sacred precinct into the Phoenician-Punic period.63 Plaster flooring smoothed against sections of megalithic apsidal wall suggests that Phoenician builders incorporated the prehistoric structure into their temple design, but this does not imply that the prehistoric building, already ancient when the Phoenicians set foot on the island, had continued to function as a sacred site for the Early or Middle Bronze Age people or, if it did, that the same beliefs persisted.64 Sockets cut in the threshold of the ancient ruins at Tas-Silg might have served as querns for grinding grain, although the Globigerina limestone is relatively soft. There is no unequivocal proof that the pits served as libation holes in prehistory or that they served as sockets for standing pillars in the Punic period.65 The recent finds and their contexts are yet to be published and to withstand the test of scholarly scrutiny; until this has taken place, claims of an unbroken line of enduring religious function at Tas-Silg have to be viewed with caution.66 Despite the limitations of stratigraphy, the pottery from the southern sector of Tas-Silg does contain prehistoric wares, some from pockets of sealed in situ deposits and others from secondary contexts. But the range of prehistoric wares is reasonably clear: Ggantija (negligible finds), Tarxien Temple period, Tarxien Cemetery (negligible finds), Baħrija (negligible finds) and Borg in-Nadur Bronze Age ware. Scarce are pottery finds of the distinctive late Borg in-Nadur fabric.67


The Archaeology of Malta


Other sites with Baħrija remains are limited. A  probable domestic settlement at Qallilija did produce Phoenician- or Punic-period architecture and pottery, none of which has been comprehensively published. A brief account is given of distinct deposits, apparently with stratified Neolithic, Baħrija and Phoenician (Punic?) pottery among the finds. A house was built over a layer of dark organic soil that contained potsherds and animal bones from domesticates (sheep and pig), presumably prehistoric material.68 Black-burnished Baħrija pottery with distinctive excised, white paste in-filled designs was identified among the earlier finds (the architecture, possibly Phoenician-Punic, is discussed later).69

On the Eve of Cultural Change In summary, when Phoenician sails first appeared on the horizon, the islanders were settled into a peaceful way of life. They seem to have welcomed the newcomers to their shores. Perhaps they had already heard about men of Sidon and Tyre; their boats were not the first foreign crafts, and they would not be the last to moor in their harbours.70 Initially, coastal areas might have acted as zones of mingling between local and Levantine interests.71 Long-haul mariners could count on a level of hospitality and productive interaction at certain ports of call on their journey. As we have seen, the Bronze Age people in Malta had on-going dealings with Sicily, procuring red ochre and other commodities, at a time when Mycenaean traders were visiting Sicilian towns. Concepts of safe harbour for mariners, exchange and gift giving were probably familiar to them, already honed through their pre-existing offshore trading contacts. Such codes of behaviour paved the way for meaningful communication and, in time, commercial interaction – one basic reason for the Phoenician forays into the West. In other words, the people of Malta were not living in isolation but were already engaged in the economic activities of the wider region and experiencing the opportunities for social and cultural exchange this presented. The island dwellers probably offered Phoenician mariners hospitality:  food, water, shelter and respite on land from a journey already months long. Knowledge of Mediterranean lands travelled with all seagoing merchants, Phoenicians included, who played a role in the dissemination of information, ideas and know-how.72 These themes, developed in Homer’s Odyssey, are likely to have reflected general attitudes of welcome to newcomers not exclusively directed at Greek mariners; the epic tales reverberate with the telling of stories by sailors in foreign lands.73 For people in areas that were hitherto isolated or those who were cautious of Phoenician traders, Herodotus described how merchants from Carthage operated: The Carthaginians also say that there is a country of Libya beyond the Pillars of Hercules where there are men.When the Carthaginians go there,

East Meets West they unload their cargo and lay the wares out neatly on the beach; then they go on board their ships and raise a smoke. The natives, seeing the smoke, come down to the sea-shore and put gold alongside the wares; then they withdraw to a distance. The Carthaginians come ashore and look; if they think the gold is enough to buy the goods, they take it and leave; but if they do not, they go on board and wait. The others come and put down more gold, until the sellers are content. Neither side is unfair; the one will not take the gold until it equals the goods in value, nor the others take the goods until the sellers have taken the  gold [Herodotus, Book IV, 196].

Such strategies of interaction developed a foundation of mutual trust and laid down social conventions to which both parties could relate which would form the basis of future dealings.74 Many of these protocols and tactics may have been employed when the Phoenicians stopped in Malta. Important to the mariners were the islands’ numerous safe harbours, new markets for their goods and the potential of a safe staging point for future onward journeys; it was at least another 1,700 km to the rich mineral resources of Iberia. Naturally defendable locations such as offshore islands and peninsulas were typical of sites to which the Phoenicians laid claim (e.g., Cadiz, Motya and so on).There is no reason to believe that the Phoenicians did not recognise these qualities in Malta and Gozo. The fact that in time the islands became strongly Phoenician in character, to the exclusion of Greek colonising interests, indicates that they did. And thus, in Malta and Gozo, were sown the seeds of Phoenician interest which germinated into fully fledged settlements of some significance – in time, one of many bordering the Mediterranean rim.

PHOENICIAN INTERESTS IN THE WEST [C]‌olonialism should first and foremost be conceived in terms of a number of specific colonial situations which share a set of interrelationships.75

To appreciate fully who the Phoenicians were and how astounding their infiltration of the Mediterranean lands was, we need to turn to developments in their homeland. For written evidence, we are reliant for the most part on classical authors, from rival political and economic cultures, who distilled nuggets of information from now-lost Phoenician texts. Although Phoenician written histories were reputedly comprehensive, none has survived intact, a legacy of their use of papyrus, which does not last long unless kept in dry conditions. Additional inscriptions on clay or stone continue to be added through archaeological discovery, but they are seldom extensive. Whether Greek or Roman, Assyrian or biblical, accounts of the Phoenicians and their Punic descendants



The Archaeology of Malta may suffer from inherent bias and selectivity that still taint current accounts, encapsulated by Gaia Servadio: In Sicily the Greeks built temples, roads and defensive fortifications. They also brought with them that sense of aesthetics which expressed itself through music and theatre, sculpture and painting, and because of which a cooking vessel became a work of art, a dwelling architecture. In short, they brought Culture. … The Punics, a practical people but with little artistic talent, recognized the superiority of the Greeks as designers and creators; the Greeks, for their part, admired the Punics’ ability as traders but also – rather like English gentlemen – considered ‘trade’ a bit below their status. They looked down on the Punics, describing them as oily and short, too cunning and solely interested in making money. Unlike the Greeks, the Punics had no interest in historical events or cultural transmission. They preserved scant written accounts of the deeds of their ancestors, and their gods were hideous monsters, inhuman and rarely represented. Astarte was utterly unlike her beautiful sister Aphrodite, and the handsome Heracles looked quite different from Melqart. No opportunity for Phidias to emerge in that world!76

Several prominent, autonomous Phoenician city-states grew in economic standing in the region after 1200 BC.77 The largest towns, such as Berytus, Byblos, Sarepta, Sidon and Tyre, hugged the coastal fringe of the eastern Mediterranean in the region equating largely with modern-day Lebanon. A  typical Phoenician city was built on an offshore island or a promontory from where coastal ports either side of the peninsula could be used depending on weather conditions. These commercial hubs were nestled against rural hinterlands which supplied the staple needs of their communities, though supplemented by imported commodities as the populations grew.78 Ethnic kinship between different groups was evident in their use of the same north-western Semitic language or dialects of it. Their domestic lives, as reflected in architecture and fictile objects, show clear affinities. If there was a sense of a Phoenician nation, it grew out of the strong cultural links forged between the city-states and lively competitive forces in the marketplace. As a people, they saw themselves as Canaanite (rather than Phoenician) and at an individual level as, for example, Tyrians or Sidonians belonging to their specific home state. Consummate and intrepid traders, they were craftsmen in all manner of luxury goods, metal smiths, purveyors of fine textiles and foodstuffs and lumberjacks who exploited the cedar forests in the mountains of Lebanon, their homeland. Collectively known as Phoenicians, they achieved significant economic standing in the Near East, ultimately turning to the coastal regions of the greater Mediterranean for commercial expansion and settlement. The history of maritime activity in the second millennium provided a backdrop for Phoenician exploration. If the classical authors are to be believed, the two centuries leading up to the turn of the first millennium BC, the period following the invasions of the ‘sea people’, witnessed greater confidence in maritime

East Meets West travel between the Levantine coast and the lands bordering the Straits of Gibraltar.79 Official trading ventures organised by Levantine states, especially Sidon and Tyre, in conjunction with biblical kingdoms to fabled Tarshish were surely underpinned by reliable intelligence about a destination abounding in wealth and rich in mineral resources. Arguments locate Tarshish on the distant Iberian Peninsula, around Huelva, where silver and gold, mined in the Pyrite Belt, were processed.80 Recently, lead-isotope analysis of silver artefacts from the Levant point to an origin for the metal in Sardinia.81 Greeks, too, were also looking westward for commercial expansion, and disentangling identities, the ethnic origins of those who brought exotic merchandise found in the archaeological record, is complex.82 Mounting evidence from central and western Mediterranean regions indicates that Phoenician infiltration began with what can be broadly termed a ‘pre-colonising’ or ‘pre-colonial period’. Terminology in this regard has come to the fore, as it can reflect on how we perceive the nature of Phoenician settlement of the West. In defining the colonising or colonial experience, scholars are now actively searching for signs of both local people and foreign interests in locations where Phoenicians lived and traded. Of particular interest are indications of hybridity, where the mingling of two populations may have generated blended cultural traits.83 Hybridity can take different forms from site to site, but domestic pottery is especially indicative. New pottery shapes appearing in a local household repertoire can herald a fundamental shift in cooking or dining habits (usually a conservative marker of persistent cultural practice), influenced by interaction with Phoenician traders and settlers. Other indicators of change can be apparent in, for instance, architecture or religious or burial practices. Much current theory is working toward a greater understanding of the processes involved. It strips away the inherently negative connotations of colonisation by empires of the modern era, involving ‘violence, political domination, and economic exploitation of indigenous peoples’,84 which are not typical of the ancient Phoenician colonising experience.85 Terms such as ‘colony’ and ‘colonisation’ can carry associations with the empirical domination of nineteenth-century territories, and they have fallen from favour in the literature.86 Nonetheless, they are valid and accurate terms which can be qualified to exclude such negative associations concerning the unequal standing of distinct groups within a play for economic and/or political power; therefore, ‘colonisation’ is used here in the broad sense of a people moving from one area to settle in another and ‘colony’ for the place in which they chose to live. Sidonians may have led the way in infiltrating the western Mediterranean markets, but because long-term settlement was not their guiding force, they left minimal signs of their presence in the region.87 As exploration progressed, Phoenician traders, especially Tyrians, gained a foothold in many sectors of the Mediterranean coastal rim. Seasonal merchant staging posts (such as at Mogador in Morocco88 or Nora in Sardinia89) were established,



The Archaeology of Malta and market-based enclaves harvested the resources of the interior by tapping into existing indigenous networks. With some success, archaeological research has looked for the origins of specific Phoenician evidence in the east.90 The number of known archaeologically investigated settlements occupied by the Phoenicians around the Mediterranean is growing. In Malta, extensive necropoles remain as primary testament to an enduring Phoenician culture that infiltrated the entire archipelago, arguably from the time of their fledgling ventures into the West. Domestic sites are known, but by dint of their early excavation or limited publication, no direct in-depth assessment of many cultural traits, such as architecture, can be made. Initial points of contact were used as ports of call in the long westward journey and contributed to the restocking of ships for their onwards or return ventures, the sailors acquiring agricultural and livestock surpluses from the local markets and eventually from their own colonies as they gained a foothold in new regions.91 The process of Phoenician expansionism appears to have had three stages, which took some two to three hundred years to unfold: initial voyages of contact, with sporadic trade through barter; establishment of semi-permanent or permanent trading outposts (emporia); and eventually, a diaspora of people from the homeland, in a systematic and deliberate wave of immigration with a view to re-settlement. How this Phoenician westward movement unfolded is complex and contingent on a wide variety of local factors, including the influences of indigenous cultures with whom the settlers came into contact.92 We do have some indications that in the early days of settlement there was a two-way exchange, where information and cultural features were shared. In time, each Phoenician settlement took on its own characteristics while also remaining recognisably Phoenician. A  level of on-board maritime customs, perhaps shaped by cosmopolitan crews, also may have contributed to the character of the earliest Phoenician culture abroad. In general terms, local people, mariners and Phoenician settlers together influenced the direction society took in each of the zones touched by Levantine interests; there is evidence of inherent diversity and adaptation, as well as the rejection of certain characteristics. A Phoenician cultural package in the West is never an unchanged carbon copy of the homeland states, nor a facsimile of Carthaginian culture, seeping out from the North African coast after that site grew to prominence. In order for such long-distance travel and eventual colonisation to be undertaken, certain criteria had to be in place. The distance of some 4,000 km from the Levant to the Spanish coast took around eighty to ninety days to complete in a sailing craft, and these early voyagers were compelled to stop periodically to replenish food and water and to seek dry land for extended periods to avoid the perilous conditions of the Mediterranean winter. Sailing generally was restricted to summer and early autumn, spanning May-June to September-October.93 The journey west would not become economically

East Meets West viable until a measure of certainty was assured through developments in boat-building techniques and a commensurate level of navigational skills was acquired. Navigation through astronomical observations facilitated deep-water sailing rather than being bound to coastal observation in daylight hours.94 A  number of Phoenician boats have been depicted in the Assyrian palace reliefs of the seventh century BC, some hauling felled timber, some naval with impressive battering rams and others carrying soldiers and passengers. The Phoenician sea-going gaulos (or ‘tub’) had a wide, round and deep hull capable of holding large amounts of cargo (Figure 6.3, no. 1). It was powered by sail on a central mast and a bank of sailors under the deck at oars when required.95 Major powers – Assyrians to the east, Egyptians to the south and Anatolian kingdoms to the north – pressed against the peripheries of their territories. These wealthy nations readily absorbed the highly prized luxury items produced in Phoenician workshops. Iron, bronze, silver and gold flowed into the coffers of Assyrian neighbours, steadying the political balance in the region and permitting Tyre, in particular, to operate as an autonomous state with relative economic freedom. Such transactions involved crossing borders and negotiating international frontiers. Invasion and conquest were ever-present threats. In this cultural interplay, the Phoenicians honed their skills of communication, negotiation and compromise. It was perhaps inevitable that they would look westward across the Mediterranean for commercial expansion. THE NATURE AND DATE OF PHOENICIAN COLONISATION IN THE WEST

Chronology is an intensely debated aspect of Phoenician expansionism. Contact between the Levantine coast and Mediterranean lands has been traced to the thirteenth century BC, but in whose hands it was that artefacts travelled west remains problematic. A cylinder seal found last century at Vélez Málaga (southern Spain) and dated to the fourteenth century takes on added significance as one of the pieces of evidence confirming early Levantine contact.96 In addition, a small bronze figurine of a smiting god, probably shipped out from Ugarit, one of the other Levantine coastal towns or Cyprus and recovered in the waters close to Selinus or Selinunte and Sciacca (southern coast of Sicily), dated from the fourteenth to the thirteenth century BC and is likely to have been carried to its marine resting place by Levantine, possibly Sidonian, traders.97 Recently announced was the discovery of a cuneiform inscription dedicated to a deity incised on a fragment of brown and white banded agate from a crescent-shaped amulet found in the on-going excavations at Tas-Silg (north; Figure 6.3, no. 4).98 It probably came from Nippur in Mesopotamia and is dated to the thirteenth century BC. Despite the difficulties of its context, this amazing find represents the westernmost occurrence of cuneiform. It may have come to the island via Cypriot Mycenaeans or other merchants who are



The Archaeology of Malta

Figure 6.3.  (1) Phoenician ship; (2) gold medallion found in 1909 from a Phoenician tomb in Buskett Gardens, Malta (Sagona 2002 [62]); (3) bronze top from a tall lamp holder from a Phoenician tomb in Għajn Qajjied, found 22 September 1950 (after Baldacchino 1953, fig.  6, pl. 13:b); (4)  agate fragment with finely cut cuneiform inscription from Tas-Silg in  2010 (photograph courtesy of Giulia Recchia and the Italian Archaeological Mission to Malta).

East Meets West known to have frequented the central Mediterranean and who had commercial dealings with Sardinia from the thirteenth century BC onwards.99 Lixus in Morocco (ca. 1180 BC),100 Cadiz in Spain (reputedly established eighty years after Troy fell, ca. 1110 or 1104 BC)101 and Utica in Tunisia (ca. 1100 BC)102 were notable among the first colonies founded by the Phoenicians, according to classical sources, although the early dates remain controversial and not clearly substantiated by archaeological investigations. Thucydides implies a sequence of pre-colonisation in the central Mediterranean, with Phoenicians living in numerous locations in and around Sicily before the second half of the eighth century BC and prior to the Greek presence from 734 BC: There were Phoenicians living all around Sicily. The Phoenicians occupied the headlands and small islands off the coast and used them for posts for trading with the Sicels. But when the Hellenes began to come in by sea in great numbers, the Phoenicians abandoned most of their settlements and concentrated on the towns of Motya, Soloeis, and Panormus103

Imported Phoenician red-slip pottery does not figure in great amounts in all their settlements, though it remains a clear indicator of contact between East and West; this is true both of Malta and of sites outside the archipelago.104 The foundation date of Carthage is usually placed around 814–813 BC, based on classical sources. Radiocarbon dates taken from animal bone excavated in Bir Messaouda on the eastern slopes of Byrsa Hill support this date.105 The possibility of an earlier settlement from around 1215 BC has been mooted on the basis of Late Helladic IIIB pottery found at the site, hence reviving the statements of some classical authors who placed its foundation before the Trojan War.106 Nonetheless, that the Elissa of Phoenician royal heritage in classical tradition is said to have launched her colonial plans for Byrsa Hill in Carthage with the help of the inhabitants of Utica indicates that Utica was well established by the ninth century.107 Phoenician colonies were already founded on the Iberian coastline by the late ninth century BC, most notably Huelva.108 Sites in the southern coastal region of Andalucía (Spain) which were established in the eighth century (750–700) BC include Morro de Mezquitilla, Chorreras, Almuñécar (Sexi), Adra (Abdera), Toscanos, Cerro del Villar and Malaga (Malaka).109 Finds from Alcáçova on the Lisbon Peninsula are dated as early as the eighth century BC.110 In Sardinia, the effects of Phoenician traders and settlers were felt as early as the mid-eighth century BC, when a distinct regional ceramic repertoire evolves from both the indigenous nuraghic and eastern Mediterranean industries.111 Much the same pattern of cultural convergence has been observed on the west coast of Spain, where pottery bore traits of both local and Phoenician industries.112 The presence of hybrid pottery forms and the complexity of assigning ethnic affiliations to their features were equally evident in Lixus (Morocco), where the early levels have been dated from 820 to 770 cal. BC.113


The Archaeology of Malta


Recent excavations in the south-western sector of Motya, a low-lying island in a shallow lagoon off the west coast of Sicily, indicate that it was settled in the first half of the eighth century BC.114 There can be little doubt that Phoenicians themselves were present in the Maltese islands and not just the products of their numerous industries; this is not a case where pottery alone is used to signify a people.115 Apart from classical texts that list Melita and Gaulos as Phoenician lands, the written language of the Phoenicians has been found in inscriptions referring to funerary and religious practice. Pottery from the temple at Tas-Silg carries dedications to Phoenician gods, and their supplicants have Phoenician names. Unmistakably Levantine architectural techniques were used in the construction of rectilinear Phoenician buildings which have no known precedents from the prehistoric periods on the island.


A lasting contribution to human history is the Phoenician system of writing, from which the modern linear alphabet developed. It is quite likely that the traders who plied the Mediterranean disseminated the Phoenician alphabet.116 Whereas scribes specialised in the craft of writing and the preservation of all manner of data, it may not have been their exclusive domain. The ability to read and write might have been one of the skills needed in Phoenician workshops. A  range of artefacts including metal work (coins, jewellery and vessels), as well as gemstones, ceramic pots and other manufactures, carries inscriptions. Instrumental in the decipherment of the Phoenician script by Jean-Jacques Barthélemy in 1764 were bilingual texts in Phoenician and Greek on what are commonly known in the literature as cippi, pillars or candelabra from an uncertain context in Malta.117 They came to light through antiquarian activities and were formerly in Giovanni Francesco Abela’s personal museum. The inscriptions indicate that they were offerings by brothers  – Abdosir and Osirxamar – to the god Melqart (Hercules), lord or Baal of Tyre, dated to the later Punic period, around the time when Malta was drawn into the Roman world (second century BC).118 The modern Maltese language owes much to Arabic as it was spoke in north Africa, a legacy of the Arab conquest (AD 870), and subsequently elements of Sicilian Italian, among other European languages, filtering into the island from Norman times. While attempts have been made to link Maltese to the Phoenician language (possibly driven by nationalistic agendas within colonial overlays of the eighteenth and nineteenth centuries), this is not the case.119

East Meets West


THE LEGEND OF ANNA, OF PHOENICIAN ROYAL LINEAGE The Carthaginians flee wherever their wandering takes them, like bees sometimes swarm when they’ve lost their queen. A third harvest had gone to the threshing floor, And a third vintage to the hollow vats: Anna was driven from her home and left her sister’s walls in tears. First she paid her last respects. The soft ashes soaked up perfume mixed with tears, And got a lock of hair from her head. Three times she said ‘Farewell’, and three times pressed the ashes to her lips, and sensed her sister’s presence. After she found a ship and fellow refugees, she sailed full speed ahead, looking back at her sister’s dear walls. Near barren Cosyra is the bountiful island of Malta, pounded by the waves of the Libyan strait. Anna headed there, counting on ancient ties of hospitality. Her host Battus, the king and a wealthy man. After he had learned the fate of each of the sisters, he said, ‘This land, tiny as it is, is yours.’ And he would have dutifully offered hospitality to the end, but he feared the great power of Anna’s brother Pygmalion.123 The sun had reviewed the zodiac twice, a third year was passing, And Anna had to find a new land for exile. Her brother threatened to attack. Detesting warfare, the king told Anna, ‘We are not warlike, you flee to safety.’ She fled as ordered and entrusted her ship to wind and waves: her brother was rougher than any sea. Ovid Fasti, Book 3.555–80124


Building on the archaeological evidence to date, Malta appears to have been one of the earlier points of Phoenician contact and perhaps an important port of call in their infiltration of the Mediterranean. If the classical sources can be trusted concerning the colonisation of the West, then Malta could have been gathered into this process as early as the closing years of the second millennium BC (some time after 1100). A well-established Phoenician presence in Malta  – probably included among the ‘islands around Sicily’ in Thucydides’ account – may have been why Greeks did not claim Malta in their own westward colonisation of the region.120 With the appearance of Phoenicians, Malta and Gozo effectively moved into the Iron Age. Around Malta’s coastline were coves where ships could be beached, and natural harbours abounded. An important Phoenician shipwreck dated


The Archaeology of Malta to ca. 700 BC has been found recently off the coast of Gozo; the ship had sailed from Sicily.121 Among its cargo were numerous storage jars. Diodorus Siculus praised Malta’s ports in the first century BC. Classical sources hint at the hospitality the islanders offered to foreigners, which may have been pivotal in the initial success of the Phoenicians in the region. Notwithstanding that Ovid composed his poetic account centuries after Anna reputedly fled from Carthage to Malta following the death of her sister, Elissa (or Dido), there is the sense that the island offered safe haven in an apparently well-established Phoenician territory. Anna became a refugee, dependent at first on the hospitality of King Battus of Melita. Whether the story is pure myth or there is a kernel of a historical event tangled up in mixed identities,122 underlying the tale is a chronological sequence:  Malta was already well established with an entrenched ruling class when Carthage was newly settled. Ovid indicates that a stable and hospitable population lived in Malta, willing to provide shelter and sustenance for travellers. Such perceptions must have been acceptable – valid – to Ovid’s audience. Even though Battus in Ovid’s work feared the military threat of the homeland king, Malta’s distance from neighbouring lands and position out of the political limelight of the greater Mediterranean may have been advantageous. For the same reasons, several centuries later, Cicero would contemplate exile in Malta when he came under pressure from Rome.125 King Battus’ identity has attracted some debate, and his historicity in connection with Malta is usually dismissed; it is generally accepted that he was monarch over Cyrene. Nonetheless, classical sources are not our only evidence for social hierarchy in Malta.126 Michel Gras, Pierre Rouillard and Javier Teixidor127 argued that the early Phoenician tombs showed signs of wealth and an elite element; a ruling aristocracy does not seem unlikely. Antonio Caruana, writing in the late 1800s, also commented on the concentration of wealthy tombs in the Rabat region, between Ta’ Saura Hospital and the Dominican Church, which had been opened previously by antiquarians Giovanni Francesco Abela and Giovanni Antonio Ciantar.128 If a ruling class governed Malta, we should not assume that it was strictly of Phoenician origin. Despite the dearth of skeletal analyses for the Phoenician-Punic period in Malta, one study indicated that both foreign (i.e., Levantine) and local populations were buried in the islands’ Phoenician tombs, indicating that the population was well integrated from an early stage.129 The very nature of the substantial rock-cut tombs in tightly packed, designated burial grounds indicates that the society included a wealthy element. Gaps between the tombs, evident in the cemetery plans, represent the negative imprints of pathways between the tomb shafts as well as open spaces where above-ground monuments or perhaps shrines and altars once stood.130

East Meets West Although archaeological investigation of Acholla, Tunisia, has focussed on the Roman remains,131 the tradition that it was established as a colony of Malta may further indicate a long history of Phoenician presence in the archipelago. The implication of this tradition is that the Phoenicians in Malta were sufficiently established and had an economy ripe enough to support a significant offshore enterprise. Moscati also speculated that Malta might have had a hand in the foundation of the settlement of Pantelleria.132 What such satellite settlements meant in terms of the economic standing and social stability of the sponsoring state should not be overlooked. We have seen that for colonisation to be contemplated, certain criteria had to be in place. The colony’s success or failure depended on the early reconnoitre of the prospective site, then a stable homeland supply line and the intrinsic value of the sustainable resources which had inspired settlement in the first place. Regrettably, absolute dating through radiocarbon analyses derived from the islands during the Melitan period has not been carried out to date. The Phoenician-Punic sequence for Malta is currently pinned on the development of tombs, which have been found in great numbers over the last two or more centuries. Using the evidence of single or limited-use burial chambers recovered through archaeological investigations, it has been possible to trace both the development of tomb plans and the pottery sequence for the Phoenician-Punic period.133 Relative chronological determinations for the pottery sequence are largely based on well-dated imported ceramic finds from sound contexts. Such instances are significant but few. The very earliest Levantine occupation might not have amounted to more than a few small dwellings, more temporary than permanent, until the early traders could be sure of a long-term future in the region. Such early footprints in new areas would be quickly swept away by later, more substantial structures as the settlements grew in size and developed to reflect Phoenician economic success. In this regard, the largely funerary evidence in Malta of this interface period has greater implications for the colonisation processes in the Mediterranean than has hitherto been recognised. Here influence can be charted in the Levantine pottery forms and a repertoire which reflected hybrid styles and a few enduring Bronze Age shapes prior to fully fledged colonisation of the islands. Although some general parallels can be drawn to non-indigenous pottery types, there are few well-dated artefacts in tombs of this time which are able to point to a more definitive date for the early contact period. Within the cremation burial at Buskett Gardens were several silver loop earrings, a decayed medallion and a scarab finger ring which bore the name Menkheperre.134 Assigned without explanation to the sixth or seventh centuries BC in the original brief report of the find,135 the scarab’s date remains ambiguous. If it refers to the royal king otherwise known as Thutmose III, then a date of ca. 1504–1450 BC is too high for even the earliest reckoning of the Phoenicians



The Archaeology of Malta in the West. An alternative might be the high priest of the same name who assumed a kingship role during the Tanite period of the twenty-first dynasty in the south of Egypt. It is clear that the chronological path of succession for this dynasty is fraught with problems.136 Though it is tempting to accept this connection and so date the pottery from this tomb to 1045–992 BC – that is, towards the end of Menkheperre’s influence and in better synchronism with the Maltese data – the evidence is far from straightforward. Jaeger’s study of this specific category of scarab demonstrates the longue durée of examples which bear the name Menkheperre, from the eighteenth to the twenty-fourth dynasties and beyond.137 In addition, the name could refer to the deity Amun138 rather than a king, lengthening the time over which such artefacts were produced and thus further diminishing the scarab as a useful and precise dating tool. Notwithstanding the problems, the possibility that the scarab is from the turn of the first millennium, a viable date for these early tombs, is worth bearing in mind.


The modus operandi of Phoenician settlement has been described in the classical texts for Spain, especially concerning the establishment of temples in newly settled areas. Only after rituals had been performed successfully at new locations did the Phoenicians persevere in establishing a permanent foothold in the area; three such rituals were carried out before Gadir (Cadiz) was founded on the Spanish coast.139 There the Phoenicians quickly established a temple that embodied an administrative, religious and social focal point, as well as acting as a symbol of permanence for Phoenician interests in lands still held by indigenous populations. Similarly, the establishment of temples in Greek colonising events was, to Kopcke, motivated by ‘a kind of calculating political thinking’, one that has resonance for Phoenician practices. In their Levantine homelands, temples and the gods they represented were emblems of social bonds, of affiliations at a familial and state level. Phoenicians may have infiltrated distant Mediterranean communities through religious pacts, integration of elite local elements and persuasion of indigenous groups, drawing on long-established practices in their homelands140: [T]‌he trading quarter constitutes one of the most widespread Phoenician institutions in the East. These are small permanent establishments situated in towns and markets abroad and from the eleventh century they have two elements in common: a temple nearby and their location outside or close to the walls and gates of the towns.141

Sacred architecture earlier than the fourth century is limited in Malta and in the Phoenician-Punic world generally, often compromised by later Roman-period

East Meets West structures or levelled to make way for them; this is true of temples at Carthage, Tharros (Sardinia) and Eryx (Sicily), for instance.142 One temple site in Motya, associated with the cothon (a port facility), can be traced from its establishment ca. 750–650 to the late fourth century BC, with some reconstructions during that time.143 The significance of an early Phoenician presence at Tas-Silg within a Bronze Age cultural landscape has yet to be determined. If the Phoenicians settled Malta in a similar fashion to their infiltration of the Iberian Peninsula and other Mediterranean regions,144 then the temple of Hercules-Melqart in particular (the location is uncertain) and the temple dedicated to Astarte (later Juno) at Tas-Silg, on the hill above the bay of Marsaxlokk, may have been of paramount importance to a successful colonising event. As already mentioned, the Phoenician temple complex in Tas-Silg was established over and incorporated the remnants of a much earlier Neolithic megalithic building. Amalgamation of the ancient prehistoric relics with the temple appears to have been a deliberate act, and considering the significance placed on establishing a temple as part of colonising practices, the ancient stones would appear to have been absorbed into that process. It is a leap to suggest that the megaliths still held some religious significance for the successive Bronze Age inhabitants and even more so for the Phoenicians. The latter may have focussed on the ruins – in essence, claimed them – to gain legitimacy for their long-term intentions to occupy the islands. Ranged against continuity is the likelihood that the Late Neolithic people chose to shut down the megalithic structures and bring to an end at least that aspect of their culture, as previously outlined. A deliberately and badly damaged corpulent statue found toppled on its back in Tas-Silg may be part of the closure narrative.145 In most locations, any connection between the contemporary Borg in-Nadur communities and the Phoenician settlers is hard to piece together. Finds such as two Bronze Age rock-cut pits containing Borg in-Nadur pottery in the Punic-Roman villa site at Zejtun,146 less than 2 km away from Tas-Silg, might indicate the presence of an indigenous community with whom the Phoenicians interacted. Though the substantial temple structures and the platform upon which they were erected testify to Tas-Silg’s significance throughout the first millennium, evidence of the earliest Phoenician presence is limited to two small chapels (structures 42 and 46) and an altar with three sockets possibly for baetyls (structure 45), evident in a few blocks around in the east of the site (see Figure 6.9, nos. 3 and 4).147 Reused masonry, including plain gorge cornice (or cavetto cornice) blocks, probably belonged to an early Phoenician temple structure. Small amounts of imported and locally copied red-slipped pottery also point to a Phoenician presence. On one level, the location of early Phoenician tombs indicates that communities once lived in the regions where they are found, but it also points to



The Archaeology of Malta ancient roads through the islands. Phoenicians may have used more than one point of entry into the island where they maintained an enduring inland presence. Melitan phase I tombs and, presumably, associated settlements, most as yet unidentified in the landscape, were once dotted along the ancient pathways or roads from the coast into the interior, namely • From Mellieħa Bay or St. Paul’s Bay to Mellieħa; • From Għajn Tuffieħa Bay, around the bay itself, leading to the interior, to Mtarfa, possibly incorporating Mdina-Rabat and intervening lands, as suggested by necropoles at Għajn Klieb, Qallilija, Qlejgħa, Għar Barka and lands to the south of Rabat, especially near Buskett Gardens; • From Marsaxlokk Bay towards the religious precinct of Tas-Silg overlooking the bay and west to Zurrieq, the likely site of a Phoenician settlement; • From Marsa Creek, a southern inlet of the Grand Harbour, along likely roads leading south to Paola and east to Qormi; and • Infiltrating into Gozo through Mgarr Harbour and Xlendi Bay. Architectural remnants, rock-cut cisterns and associated habitation debris in lands bordering the northern slopes of the Qlejgħa–Ta’l-l’Isperanza Valley represent Phoenician-Punic settlements. Structures were found in the hills of Qlejgħa north of Mtarfa and over the Wied il-Qlejgħa watercourse in 1906148 which included large stone blocks (some with signs of ornamentation), a deep cistern still fed by a spring or groundwater and a scatter of ‘Phoenician’ pottery in the fields. In the excavated cistern, numerous pottery fragments were described as both very rough and fine, and it was thought that the area was once a ‘populous centre’.149 People in the settlement probably used tombs on the hilltops in the immediate region; as these contained remains from the earliest Phoenician Melitan phase I through the Punic phases to the Roman period (one tomb had Roman-period sherds tossed into the shaft), we can assign to the area a very long history of occupation.150 A settlement further west of Mtarfa on the slopes of the Qallilija district has already come to our attention for its Late Bronze Age (Baħrija) deposits. Found in 1912, its architecture comprised a length of regular linear wall, white plaster flooring, cobbled surfaces and a stone block embedded in the floor which might have been a base for a pillar used to support the roof. Within the partially excavated building were two stones hollowed into troughs and a third trough cut into bedrock. Two large jars, one with four ‘ears’ (handles?) on the shoulder, were thought to be of Punic date and were fixed into place with rocks, each containing grinding stones. Standing stones described as ‘menhirs’, destroyed by farmers who worked the land, could have been remnants of house walls using the opus africanum building technique known at Phoenician-Punic

East Meets West sites, but as underlying layers were prehistoric, actual menhirs are also a possibility. Three spindle whorls were found in a deep hole associated with the wall.151 A number of rock-cut tombs also have been documented close to these locations.152 Archival photographs of the better-preserved pot suggest that it belonged to the late Borg in-Nadur or Melitan archaic phase I. PHOENICIAN FUNERARY PRACTICE

The Phoenicians of Tyre (Tyre al-Bass) in Lebanon, for instance, buried their dead in pits in designated burial grounds; the cremated remains, stored in large urns, and a requisite range of pottery also were deposited in the graves. Rituals which involved burning are evident in connection with some of the burials. Yet, in Malta, there is some variety in the nature of the burials of the early Melitan phase I period which might reflect both local and Phoenician influences at play. In time, Phoenician funerary architecture, most notably shaft and chamber tombs, became the standard for the archipelago, although the plan and internal fixtures of the graves show a steady evolutionary path before catacombs replaced them in the Roman era. Bronze Age rock-cut and bell-shaped pits were reused for burials at Mtarfa.153 These could be further modified with the addition of small chambers to the side – the pit effectively became the shaft of the grave. In addition, simple pits,154 natural caves and rock clefts were used for burials in the Melitan phase I.155 Anthropoid clay coffins also were found in a substantial rock-cut tomb in the region of Għar Barka,156 and they point to a degree of opulence, as they did centuries later in Sidon (Lebanon). One held human remains and an iron ring.157 Rustic in the quality of modelled human female features, it falls into line with other sarcophagi found in the Levant from the thirteenth century and later Phoenician contexts which drew on the Egyptian mummy-form anthropoid coffins for inspiration (Figure 6.5, nos. 1 and 2). Inspired perhaps by a Philistine tradition of fashioning anthropoid clay coffins,158 stylistically and chronologically the Maltese example is closer to the forms found in thirteenth-century Deir el-Balah and tenth-century Palestine (Tell el-Fara, Lachish and Beth Shan)159 than to the later marble examples from Sidon, which date from the sixth century to the fourth century BC. It is worth noting that a Philistine anthropoid pottery sarcophagus was found in Neapolis, Sardinia.160 Based on the evidence of the material remains in the graves, it has been possible to trace a path through the variations of formal rock-cut tombs (e.g., Figure  6.4, nos. 1 and 2). At first rounded161 or apsidal in floor plan, tombs eventually took on a square plan, cut to large proportions for the shaft, often with a few small steps cut down one side for access. The variety of round, apsidal and square combinations indicates a degree of experimentation. Until regular square floor plans were adopted, chamber roofs were usually domed (Figure 6.5, nos. 3–12).



The Archaeology of Malta

Figure 6.4.  Phoenician tombs: (1) tomb chamber exposed by a road cutting close to Dingli Cliffs, Malta; a small lamp niche can be seen on the right wall edge; (2) tomb shaft and chamber entrance at Xemxija, Malta (photographs by C. Sagona).

Large sealing slabs, fitted into a rebate edge in the doorjambs, closed tombs. Internal fixtures that were cut into the chamber appear to have become more common as time passed.162 Niches for lamps, a persistent element in the Phoenician burial rites of the islands, were cut into the tomb walls. The head of the deceased could be propped up by a rock or pillow cut at the top of the bier. This bed could be further elaborated with small pits at the midpoint or near the head of the corpse and shallow ledges left at the edge.The rectangular trench cut into the floor served to add extra depth to the chamber, as few were high enough to stand upright in, but the trench also might have carried a symbolic significance.163 Pottery was frequently placed in the trench, and smaller items, especially cups and bowls, were positioned near the head and knees of the interred remains. Skeletons were placed on their backs on the bier and pushed aside for later burials if the chamber was small. THE CHRONOLOGICAL DEVELOPMENT OF ROCK-CUT TOMBS IN MELITAN MALTA

Based on the earliest burials in the rock-cut tombs in Malta, the following scheme charts the broad evolution of shaft and chamber constructions (Figure 6.5, nos. 3–12)164: • Unstructured pits and rock shelters and reused prehistoric ‘silos’ or bell-shaped pits (plan 1a–b); • Round chambers (plan 2a–d), Late Bronze Age–archaic Punic phase I tomb structures, at least as early as 900 BC to the seventh century BC (Figure 6.5, no. 3); (continued)

East Meets West

Figure 6.5.  (1) Detail of the clay sarcophagus foot with butterfly-shaped clamp sockets found in a tomb in Għar Barka in 1797; (2) clay mummy-form sarcophagus depicting a shrouded female (Sagona 2002, p. 349, fig. 29); (3–11) nine main rock-cut tomb



The Archaeology of Malta • Apsidal to oval chambers (plan 3a–f), phases I and II (Figure 6.5, no. 4); • Apsidal chambers with elaborate features (plan 4a–f), eighth to seventh centuries BC (Figure 6.5, nos. 5 and 6); • Squarish chambers with elaborate features (plan 5a–d], established phase I (Figure 6.5, no. 7); • Square to rectangular chambers (plan 6a–c); • Squarish chambers with ‘unstructured’ central trench or early central axis trench (plan 7a–c), from phases III and IV; • Narrow rectangular shaft and chamber (plan 8a–e), from phases III and IV (Figure 6.5, nos. 8 and 9); • Rectangular chambers with side trenches (plan 9a–d), phase III; and • Rectangular chambers with central axis trenches (plan 10a–i), from phasee III and IV (Figure 6.5, nos. 10 and 11). In the final stage of rock-cut tombs, niches were cut in the bier to take the head of the corpse (Figure 6.5, no. 12). By this stage, Roman cultural influences dominated the islands, and with the growth of urban centres came the use of catacombs for burials.

Both cremation and interment of adults were practised throughout the Phoenician-Punic period. Cremation sites have been recorded at Tad-Dlam in the Tac-Cagħqi area of Rabat. Another ashy deposit, with remnants of funerary pottery, was located at Nadur in 1924.165 Of these sites, very little has been documented. Malta’s funerary record indicates that from the opening years of Phoenician presence, two forms of ritual were in place: the first concerned adults who had presumably passed through a rite of passage in their early teens and the other, for infants and children who had not. In Malta, burial chambers rarely contained infants and children; the chambers were clearly designed for adults.166 Only under Roman influence (Melitan phase IV and later) did the tombs hold some children’s remains, signifying a shift in cultural practice, and these were neither in large numbers nor cremated.167 No division in adult funerary practices appears to have been made between the sexes, although grave goods may reflect some distinctions. Whether the individual’s arrival into adulthood was marked by ritual and ceremony – a rite of passage – is not known.The evidence from Malta foreshadowed the recent analyses of Carthaginian infant remains, Figure 6.5. (cont.) plans of the Phoenician-Punic period showing the evolution of shafts (right) and chambers (left); (12) final rock-cut chamber form with head niches cut into the end of the bier falls within the Roman period of influence (T = trenches cut in the floor of the chamber; B = bier or bed where corpses were placed; S = steps cut down the sides of shafts; HN = head niches) (after Sagona 2002, figs. 29 and 333).

East Meets West cremated and buried in urns in designated child burial sites called ‘tophets’. The study found that it was highly unlikely that child sacrifice through incineration  – one of the reputed infamous practices of the Phoenicians in the West – played a part in the infants’ demise as the remains included instances of stillbirth and premature and perinatal deaths.168 From the analysis of cremated remains at Tharros (Sardinia), Fedele found that about half the infants were cremated with an animal, usually an immature lamb or goat, and that the children had remained still during the cremation process, as if already dead or sedated.169 Although many discussions accept as fact that infants were sacrificed,170 the evidence is mounting that this practice, if it ever occurred in Phoenician society, was exaggerated in classical literature,171 a problem compounded in the recent era by biblical comparisons172 and by graphic descriptions such as that in Gustave Flaubert’s novel, Salammbo (1862). The implications for early Phoenician culture in Malta are such that it is worth dwelling on the reputed practice of human sacrifice at this point, even though scholarly argument also concerns later Punic evidence, especially from Carthage, where the concentration of human remains in the Salammbo tophet burial ground has attracted much attention. There are evocative reports of child-only burials in Malta. A deep pit in Rabat was filled with jar burials containing small animals and birds at what may have been a tophet site. Edward Charlton records no human infant remains, but two inscriptions found nearby in 1820 mention the molk (mlk baal) sacrifice.173 This ritual, Gras, Rouillard and Teixidor argued, referred not to human sacrifice but to animal offerings in association with burials of infants who had died presumably from natural causes.174 Nonetheless, they make a point about the validity of child sacrifice: ‘Why would the formula mlk baal have signified a ritual of child sacrifice in the 8th cent. in Malta, if not a single literary or archaeological indication informs us of the existence of this ritual at Tyre?’175 Unfortunately, the Rabat discovery was an early find and received only cursory reporting176; it is possible that human remains may have gone unrecognised or were edited out of the record. The actual location of this tophet was reputedly at the Saura Hospital site, but archival sources suggest that the burial pit lay in the Tal Virtu region.177 In later Punic Melitan contexts, remains of at least three infants interred (not cremated) in modified storage jars were found in a rock fissure in Qallilija. One jar had the letter M scratched on it, probably in reference to the molk ritual (or possibly signifying the deity Melqart), and another jar contained two children’s skulls. Animal remains were not reported owing to the poor preservation at the site.178 A  simple burial pit in Ta’ Vnezja, Mosta, held a modified storage jar split to take the interred remains of two children (approximately 2½ and 9 years of age).179 Offshore comparisons can be found. Aside from Carthage, distinctions between child and adult burials were reported in Puig des Molins, Ibiza.180 In Phoenician contexts at Achziv (southern Levant), ‘children were



The Archaeology of Malta never cremated in the Iron Age (or in other periods).’181 Rather, some children were buried outside tombs separate from adults who were cremated. Ethnographic evidence suggests that human sacrifice occurs within a society at times of great stress, and overall, it is a rare phenomenon.182 It would be anomalous if in the absence of modern medicine and over the millennium of Phoenician-Punic burials in Malta children had not died of natural causes. Diseases common to children are likely to have resulted in a high mortality rate in ancient Malta, as they did in other cultures.183 This begs the question, if not the tophets, where are the burial sites of infants who died of natural causes? CULTURAL TRAITS AND ECONOMIC ACTIVITIES

The funerary evidence indicates a pre-colonisation period in Malta where the pottery especially displays archaic qualities and traits of the indigenous repertoire in hybrid forms. One-handled pots (Figure  6.6, no.  10), trays (Figure 6.6, nos. 21 and 22) and various handmade jars embody local characteristics carried on from the floruit of the Borg in-Nadur period. Some shapes are probably of local inspiration, such as the large baggy urns with horizontal handles (cf. Figure 6.1, no. 2 handle and no. 3 jar top, with Figure 6.6, no. 5) and others with tall necks (Figure 6.6, nos. 6 and 7). Early Melitan vessels of the archaic phase I can mimic basketry in their lattice and crisscross painted decorations, reflecting a local rather than Oriental trend (Figure 6.6, nos. 25 and 26). These archaic Melitan phase I tombs do not have transport amphora, which may reflect the initial Phoenician trading pattern; in Tyre al-Bass burials, for instance, cinerary urns, jugs and smaller vessels prevail.184 Amphorae are linked to the burgeoning Phoenician trade in wine which drove international commerce in the greater Mediterranean in the eighth century BC and appears to have gone hand in hand with extensive immigration from the homeland to the settlements.185 Evidence for wine drinking does appear in Malta at this later date (equating with the Melitan I established phase), but more of this later. On the strength of current evidence, this lack of transport jars in the tombs can be seen as another indicator of their early date. Outside Levantine pottery shapes (usually locally copied), imports generally do not feature in these archaic tombs. The date of the funerary sequence for Melitan I has been determined relative to the tombs of the later colonising period of the eighth and seventh centuries BC; these are dated by the imported Greek vessels which do feature in the funerary repertoire by this time. This dearth of Aegean imports in the early tombs would appear to have wider implications. It might be a characteristic of early Phoenicians in the central Mediterranean corresponding to Thucydides’ account that first Phoenicians and then Greeks settled the region. Lorenzo Nigro also has noted that almost no Greek pottery or imitations were found

Figure  6.6.  Melitan archaic phase I pottery forms:  (1) asymmetrical askos; (2 and 3)  trefoil mouthed jugs; (4)  jug of likely Bronze Age inspiration; (5–7) large urns; (8 and 9) beakers (often called ‘thistle-headed beakers’); (10 and 11) handled pots, a local form from the Bronze Age; (12–19) bowl and plate forms; (20–22) cooking pots and pans; (23) broad-rimmed plate; (24) twin-nozzle lamp; (25 and 26) bowls with painted basket designs; (27) cup with four handles on the rim modelled on Greek forms (after Sagona 2002, fig. 339). 207


The Archaeology of Malta in the lowest levels (phase IX) uncovered in the renewed excavations (2002 and after) at Motya, Sicily, in the remains of a temple close to the Phoenician cothon (or port facility).186 Pottery fragments dated the level to the first half of the eighth century BC.187 Imported vessels, notably painted two-handled cups, did not appear in the archaeological record in Motya until the mid-eighth century BC. A similar influx of imported items has been noted for contemporary levels in Carthage (early Punic I, 760–675 BC).188 Whether the dearth of imported Greek pottery in the early tombs reflects the cultural nature of the first Phoenicians to infiltrate the islands (possibly Sidonians), the dominance of a local culture, which might not have valued imported wares, or simply the serendipity of archaeological deposition remains to be seen. Red-slipped vessels are a feature, often referred to as the ‘calling card’ of Phoenician wares appearing in the West. One form found at Tas-Silg (north and south), the handmade shallow bowl with everted, thickened lip (Figure 6.6, nos. 16–19), as well as a small amount of other sherd material, appears to be of the archaic Melitan phase I.189 The bowls are dated to the eighth century BC, largely based on the sequence Hermanfrid Schubart devised for similar bowl types found in Spain.190 Although the scheme has a general relevance to the colonisation of the West by the Phoenicians, the archaeological record for the period in Malta still must be viewed as comprising broad brushstrokes, not yet adequate for drawing precise parallels, and we should be wary of simply super-imposing the Spanish record on other regions.191 As red slip was employed in the closing years of the Bronze Age, this decorative treatment is less helpful in Malta for identifying a Phoenician presence unless good diagnostic fragments are at hand. Nonetheless, this bowl form is one of the early shapes which indicate the arrival of the Phoenicians in the region.192 Actual imported Phoenician pottery is rare, but it readily stands out in the repertoire by virtue of its relatively refined, pale clays and wheel production. A few examples from tomb finds and some sherds from Tas-Silg are quite distinct from local products. The Tas-Silg pieces originated from Tyre or Beirut and have been dated to the second half of the eighth century BC.193 New vessel shapes suggest fundamental shifts in dining practices and domestic habits under the influence of newcomers to Malta’s shores, particularly broad-rimmed plates with deep floors that ultimately hark back to Greek influence on the Phoenician repertoire. Wide-rimmed plates are not an indigenous form, and as Jaime Vives-Ferrandiz has argued for similar plates in Spain, they reflect a shift in repertoire and imply an introduced new domestic mode of consumption of solid foods.194 Such domestic traits can take time to change and evolve. An aspect of this can be seen in the endurance of the practice of hand building vessels from clays that would appear to have been local to Malta. Wheel technology, well established among potters in the Levantine homeland,

East Meets West was not immediately introduced to the islands, suggesting that potters were not among the ship’s crews who made the initial contact. The potters of Malta – probably in cottage industries rather than commercial workshops – continued their traditional practices for some time but borrowed new shapes from the Phoenicians. In short, against the backdrop of Late Bronze Age manufactures and the initial stages of the Melitan phase I, we have the introduction of entirely new shapes into the island repertoire, and during the Melitan established phase I (ca. 750–620 BC) came an increase in imported wares, along with significant numbers of Phoenician settlers.195 Many of the trends in pottery production as the industry became more established can be demonstrated through a tomb find at Għajn Qajjied in Malta.196 With only two interments of the early period in the tomb and an array of datable imported pottery of Aegean and Levantine types, as well as local forms, it is pivotal to the relative chronology of the Melitan phase I sequence in Malta. A  third placement of a cremation urn and associated pottery was significantly later (Melitan phase IV) and can be readily separated from the early deposits. The tomb illustrates a society in a state of flux. In addition, the chamber and the funerary deposits display a degree of opulence. Within the large-proportioned rock chamber was a stone slab bier which held the human remains. Fine Greek cups, probably the personal items of the deceased and showing signs of use, included a proto-Corinthian cup (Figure  6.8, no.  7), variously dated to between the eighth and mid-seventh centuries,197 and an eastern Greek bird bowl (Figure 6.8, no. 8). The only bronze, a three-pronged lamp holder, found in Malta came from this tomb, but it is a type more commonly found in Cypriot contexts, dating from within the seventh to sixth centuries (Figure 6.3, no. 3). Silver and bronze bracelets and rings, an imported bead and four iron loops, possibly from a wooden bed, also were among the finds.198 Items of personal adornment (e.g., Figure 6.3, no. 2) are not found in great numbers in the Melitan tombs when compared with sites such as Tharros (Sardinia), where the tombs were rich in such artefacts.199 In any case, as the funerary ritual progressed to a reasonably rigid format over the first millennium in Malta, such small trinkets did not seem to play a persistent role. The Għajn Qajjied tomb is pivotal to the dating of the Phoenician Melitan established phase I, and importantly, there are tombs which on the grounds of the typology of their contents predate it (within the archaic and early phase I) (Figures  6.6 and 6.7). Chronologically, the implication is simple. Burials in the Għajn Qajjied tomb have Phoenician-style pottery which represents an evolved repertoire; by implication, there was a path of earlier pottery development on the island. This line of ceramic evolution has been clearly shown through analysis of the Phoenician-Punic tombs excavated to date.200 It is important to look beyond the Għajn Qajjied find – which is not the earliest


Figure  6.7. Melitan archaic to early phase I  showing the development of pottery forms:  (1–4) large urns; (5)  piriform jug; (6)  mushroom-topped jug; (7)  variation on the previous form with flattened disc top; (8–10) wide- and narrow-necked jugs; (11–13) thistle-head beakers; (14 and 15) bowl forms; (17) two-handled cup modelled on Greek forms; (18) twin-nozzle lamp; (19) Single-nozzle lamp, rare in Malta (after Sagona 2002, fig. 340). 210

East Meets West tomb showing Phoenician influence – for the evidence of a pre-colonisation phase in Malta.201 Archaic pottery shapes not present in the Għajn Qajjied tomb because they had fallen out of use include thistle-headed beakers (e.g., Figure  6.6, nos. 8 and 9), large and baggy urns (e.g., Figure  6.6, no.  5)  and some bowl forms. Leaving the precise dating of the burials aside, as they are the subject of quite involved discussions, and instead shifting the focus to the pottery sequence,202 the Għajn Qajjied tomb demonstrates the tenacity of local cultural practices in the endurance of ceramic manufacturing techniques and the hallmarks of the fully fledged colonisation period. By the time of the second interment, wide-rimmed plates in the Għajn Qajjied tomb were wheel made and no longer slipped red all over; the floors were left in reserve. Only a few fragments of earlier handmade red-slipped plates were left in the chamber, with earlier pottery presumably cleared away before one or both of the surviving burials took place.203 Many shapes – transport amphorae (Figure 6.8, no. 5), pear-shaped flasks (Figure  6.8, no.  1), piriform jugs (Figure  6.8, no.  9), olpe or dipper juglets (Figure 6.8, no. 6),204 tripod bowls (Figure 6.8, no. 2) and fine drinking cups (Figure 6.8, nos. 7 and 8) –directly reflect Levantine traditions and the practice of wine drinking in particular. Pear-shaped flasks and tripod bowls often have been found together in other contemporary Phoenician settlements, and they are now thought to have served the specific function of mixing herbs or spices used to flavour wine.205 Recent analysis of Phoenician flasks from five sites in southern Levant indicated that cinnamon was once contained in them, a product which was harvested in southern India and Sri Lanka and was possibly one of the additives used with wine.206 The sequence of minor variations in this flask shape falls between the first half of the eighth and the last half of the sixth centuries BC, a useful chronological marker wherever the flasks occur in the Mediterranean region.207 Tyre al-Bass provides the earliest secure context for the larger piriform jug shape with globular body, which falls within the transitional Early to Late Iron Age.208 Núñez Calvo considered that this shape, too, was connected to the introduction of wine drinking and the changes it stimulated in cultural artefacts.209 It is no surprise that the Għajn Qajjied tomb contained two transport jars, wide at the shoulder and tapering to a rounded or somewhat pointed base, with two loop handles on the shoulder, of the type documented in other locations.210 The appearance of this wine-drinking kit is matched by the pollen record in Malta (taken from Marsa sediments); ‘grape seed [and] fragments of vine (Vitis vinifera) were found at levels corresponding to the Late Bronze Age and the Phoenician-Punic periods, indicating that viticulture may have started in Malta significantly prior to the Roman period.’211 The growing presence of vessels tied to wine consumption in the region has wider implications in regards to the nature of settlements. Wine is a commodity now considered strongly linked to Phoenician commercial enterprises. With increasingly stable settlements and growing migrant populations,



The Archaeology of Malta

Figure 6.8.  Phoenician wine trade: (1–8) pottery wine-drinking kit from the Għajn Qajjied tomb comprising large storage jars (5), trefoil and disc-topped jugs (3 and 4), small dipper jugs (1 and 6) teamed with tripod mixing bowls (2) and fine drinking cups (7 and 8), in this case two imported Greek cups which are also key to dating early

East Meets West the taste for wine appears to have rapidly stimulated a wide-ranging trade in this commodity.212 This trend is matched in Iberia from the eighth century BC, not just in the settlements of Levantine traders, but within the indigenous communities of the interior.213 The inclusion of disc-topped jugs (the equivalent of the ‘mushroom-topped jugs’ of other Phoenician settlements) (Figures  6.7, nos. 6 and 7, and 6.8, no.  4)  becomes a common feature  – perhaps a requirement  – of the early Melitan phase I funerary pottery. The disc-topped jugs in tombs, often teamed with a trefoil-lipped piriform-shaped jug (e.g., Figure 6.7, no. 5), first appeared in homeland Tyre, in the transition between the Early and Late Bronze Age.214 Carinated bowls with vertical walls to the rounded lip appear in Melitan tombs which have antecedents in the homeland from the first half of the eighth century BC.215 Variations of the form carry into the late Melitan phase III. The presence of the wine-drinking kit in funerary contexts also has implications for Phoenician ritual practice; it was carried to the grave, where it reflected on the deceased’s social status and on the family, who could afford this luxury commodity. Grape pressing vats (Figure 6.8, no. 10) also point to viticulture on the island, well entrenched by the Punic period. Small amounts of Juglans pollen from the Marsa sediments may indicate that walnut was introduced to the island around 950 BC, as it was in Sicily, perhaps by Phoenician traders.216 Increased agricultural activity, especially the growing of cereal crops, reached ‘unprecedented high levels’ in Malta during the Phoenician-Punic periods, aided by advantageous climatic conditions in the region.217 Olive cultivation, too, first appeared in Malta during the Phoenician period. Some pottery forms may have had a domestic function, but we have no excavated dwellings dating to the interface or initial Melitan phases in the archaeological record by which to judge; such sites witnessed considerable modifications over time. As the repertoire moved out of its archaic beginnings,218 vessels which ceased to appear may have done so because they lost their relevance in Phoenician daily or ritual lives. In time, a number of shapes that ended were tied to wine drinking, which may point to the simplification of this practice, possibly as it became available to a wider, non-elite element of society. A relatively recent parallel can be found in the consumption of tea in Figure 6.8. (cont.) tombs in Malta (after Baldacchino 1953, figs. 3–5); (9) another jug form that could be used for wine drinking is the biconical or piriform jug with narrow trefoil rim which mimics bronze forms; the pictured example was found in Mtarfa on 22 March1927 (after Sagona 2002, fig. 58:4); (10) testimony to local viticulture are the rock-cut shallow grape presses linked by small draining holes to the collection vats; the pictured example is near the Misqa Tanks in the vicinity of Ħagar Qim and Mnajdra (photograph by C. Sagona).



The Archaeology of Malta

Figure 6.9.  (1) Area of recent excavations at Tas-Silg, marked on the main plan with an asterisk (see 2), where Bronze Age deposits have been identified (after Cazzella et  al. 2007); (2)  plan of Tas-Silg north and south of the Zejtun-Delimara Road;

East Meets West Europe. Considered a luxury and an expensive commodity in the seventeenth century AD, the demand for tea stimulated much of the long-distance trade with the East. A set of high-status utensils was made for tea storage (caddies, often kept under lock and key), preparation (kettles to heat water, strainers, pots, milk jugs, sugar pots and teaspoons) and drinking (cups and saucers). Times of day were set aside for tea drinking which have endured in everyday parlance (tea time, afternoon and morning tea), which marked the ritualised nature of its consumption; sets of cups and saucers embody the notion that people assembled to consume the brew in a social ritual. Once trade routes were well established, along which reliable quantities of tea were transported, prices dropped, and tea became widely available to the general community; the paraphernalia required to brew and drink tea also diminished to a core of essential items. Though other contemporary tombs of the Melitan established phase I contribute to our understanding of the period, one from Mellieħa is worthy of particular notice. Of all the tombs in Malta, this find has the greatest number of unequivocal direct links to the homeland and to Tyre al-Bass in particular.219 A  large urn with handles from rim to shoulder (Figure  6.7, no.  4),220 a hemispherical bowl (Figure  6.7, no.  14)221 and a disc-topped round-mouth jug with neck ridge (Figure 6.7, no. 7)222 are shapes which find strong parallels in homeland contexts. The neck-ridge jug shape in Malta rarely has the downturned lip seen on the ‘mushroomed-topped’ examples in other Phoenician settlements. Instead, its everted lip is similar to the square-cut rim forms from Tyre.223 Two piriform trefoil jugs fulfiled the trend to supply the deceased with two jug types – trefoil and round-mouth shapes.224 The influence of imported forms such as the Mellieħa urn may have been the inspiration for a series of tall-necked urns with handles attached under the rim and at the shoulder which appeared in Malta during this period (Figure 6.7, no. 3). They are in marked contrast to the urns found in the archaic phase, which are handmade and baggy in shape (Figure 6.6, nos. 5–7).The plate in Mellieħa with a pronounced re-entrant inner lip is similar to examples from other sites in the central and western Mediterranean, less so to the homeland forms.225 A single-nozzle lamp (Figure 6.7, no. 19) in the group is extremely rare in Malta but can be matched to the predominant Levantine type. The single-nozzle form is represented in the Iberian contexts of the eighth century BC226 and in Carthage, among other Figure 6.9. (cont.) contours measured in metres (after Bonanno and Vella 2015); (3)  reconstruction of a three-pillared altar, altar 45, found at Tas-Silg, showing additional pit in the background possibly contemporary with the altar (after Ciasca 1993, fig. 6); (4) detail of altar 45 with three sockets probably for pillars (as in 3) and partially covered by later paving slabs (after Cagiano de Azevedo et al. 1973, fig. 2).


The Archaeology of Malta


sites. Yet it is noticeably absent from among the urn burials at Tyre al-Bass in the 1997 campaign.227 Lamps take on an increasingly symbolic role in funerary contexts, an aspect of their history, which will be discussed in Chapter 7. Perhaps linked to lamps are pedestal incense burners with attached cup and saucer, which have been found in other Melitan established phase I  tombs and are likely to have performed a ritual function. In subsequent phases, pedestal cups may have served a similar role (see Figures 7.1, no. 9, and 7.11, nos. 17–20).228 While we can suggest, at least in this instance, a strong link between Malta and homeland Tyre, the wider archaeological evidence reflects how quickly indigenous elements and imported influences merged in this new cultural setting, perhaps within three or four generations. Clearly, the deceased within the Għajn Qajjied and Mellieħa tombs were not part of the first wave of people to settle the central and western Mediterranean. SUMMARY

In their fledgling contacts with Phoenician mariners, the local Maltese islanders could not have anticipated that more of them would follow and eventually in such numbers as to make an indelible mark on the island culture. Phoenicians targeted local communities who would have facilitated important points of supply to meet their demands for onward journeys and, later, the requirements for permanent settlement of Phoenicians in their midst. By subsuming significant remains of prehistoric buildings into their newly founded temples, the Phoenicians claimed a connection to the land. If we adopt a phenomenological stance, then we might imagine that they spun their historical narrative to suggest that they had the right to claim the buildings. The newly founded Phoenician temples declared permanence, embodied a place for religious devotion that may have intrigued and then captivated the local population and presented a formal focal point where the wealth garnered by the Phoenicians accumulated. Where Phoenician remains are found in the archaeological record, there is often also some trace of indigenous Borg in-Nadur people and enough signs of enduring local practices to recognise cultural hybridity. The Phoenician occupation of the islands in all likelihood imprinted itself on existing Bronze Age settlements; the simple ovoid houses were likely wiped away by new buildings, which, in turn, would be rebuilt or remodelled under Roman influence. Some serious thought needs to be given to the interface period, if no evidence of the archaic Melitan phase I is forthcoming from the large settlement at Borg in-Nadur or others like it. The implications of some site abandonment by Bronze Age folk in the face of Phoenician infiltration of the island are intriguing. As yet, there are no signs of hostility, but a pattern is emerging where some sites fell into disuse while other locations swelled with the Levantine settlers.

East Meets West At this time, Malta was firmly established on the international maritime routes of the Mediterranean and could no longer float along untouched by socio-economic and political happenings in other regions. In 573 BC,Tyre suffered a defeat at the hands of Neo-Babylonian interests which infiltrated the Phoenician homeland to the very shores of the Levantine coast. For the central and western Mediterranean colonies, mostly well established by this time, their links to the homeland were effectively cut. These events coincided with the rise of Carthage as an economic power, the greatest of Phoenician colonies in the West. This turning point in Phoenician fortunes, around 550 BC, is often marked in scholarly works by the adoption of the term ‘Punic’, which is generally applied to descendants of the original Phoenicians in the West.229 How Malta fared in the face of these events will be the subject of Chapter 7.



In terms of material culture, there is no line in the sand that differentiates the Phoenician from the Punic period in Malta. The tendency to compartmentalise Maltese history into ‘Phoenician’, ‘Punic’ and ‘Roman’ masks a persistent and evolving culture. Such issues may vex modern scholarship, but Malta continued along a stable and steady cultural path come what may in the political arena of the greater Mediterranean. The term ‘Punic’ (a Roman designation for their Phoenician competitors) is applied to the later Phoenician cultural centres of the central and western Mediterranean. Some use the term specifically in relation to the significant changes of the sixth century BC, whereas others apply it generally to the western colonies:  a Punic West born of a Phoenician East. Notwithstanding, this was a time of flux. Unrest in the eastern homelands saw Tyre succumb to the Assyrian king Nebuchadnezzar (573 BC). The sixth century in the greater Mediterranean region witnessed economic downturn in, and even abandonment of, a number of Phoenician settlements, perhaps hastened by the increasing emergence of coinage for commercial transactions.1 This monetary development undermined long-established exchange and barter systems between elite elements within indigenous groups and the Phoenician traders. In Malta, the trend can be observed in a degree of contraction in the pottery repertoire in funerary contexts in the Melitan phase II (Figure 7.1, nos. 1–11). There were fewer imports, suggesting reduced interaction with the homeland and the eastern Mediterranean in general. Equally, there is less variety in tomb pottery, and the quality of lamps is poorer. In the economic hard times that affected the ceramics industry, at least in terms of lamps, people may have fallen back on their own resources, producing what they needed at home. Of the ephemeral record, however, we can only speculate that the slowing of the economy was felt at all levels on the islands. Lying west and slightly north of Malta on the African coast, Carthage had grown into an unrivalled economic and political power, effectively the hegemonic city of the established Phoenician settlements in the region.2 Within 218

Melita and Gaulos during the Punic Period

Figure 7.1.  (1–11) Melitan late phase I to early phase II pottery forms: (1–3) large urns; (4–6) wide-rimmed plates; (7 and 8) bowls; (9) stemmed incense cup; (10 and 11)  twin-nozzle lamps (after Sagona 2002, fig.  342); (12) Tanit sign common to the Phoenician-Punic settlements in the central and western Mediterranean, but not Malta; (13 and 14)  pottery combinations of lamp, cover plate and urn which


The Archaeology of Malta


this structure, however, ancient Melita should be seen as an important Punic location, viable in its own right, not simply a vassal of Carthage. Significantly, textual resources, beginning with a fragment of Hecataeus in the first half of the fifth century BC, supplement the story of Malta in the first millennium.3 Malta was neither a short-lived Phoenician emporium nor a hardship post in the Mediterranean.4 By the sixth century BC, its population was substantial, and there were extensive burial grounds throughout the islands, especially within the lands around Rabat-Mdina. Several hundred tombs have been documented throughout the archipelago, and many of them have multiple burials (e.g., Figure 7.3, no. 13).5 In time, the archipelago was economically prosperous and politically stable enough for its Melitan traders reputedly to have established a colony in North Africa at Acholla; if this is not true, then at least there existed a perception that expansion was within the economic capabilities of Malta. If it is possible to generalise about distinctions between Phoenician and Punic phases in Malta, then it would be to highlight the greater sense of ‘establishment’ of the culture in the islands: construction of substantial rural complexes devoted to oil production on a commercial scale, a pottery industry which had matured to meet market demands with a standard wheel-made product, notable growth in urban centres, religious precincts reflecting the island’s prosperity and obvious signs of extensive maritime commerce. LAND-USE AND SETTLEMENT PATTERNS

Phoenicians and their Punic descendants did not restrict themselves to peninsulas and coastal regions in Malta and Gozo. Evidence for Punic structures outside the temple precinct at Tas-Silg is limited, perhaps due to flawed investigations or a failure as yet to published adequately. Nonetheless, what does exist confirms the impression portrayed by the classical sources. INDICATORS OF CONTINUITY, PROSPERITY AND STABILITY CONVEYED IN DIODORUS SICULUS’ DESCRIPTION OF MALTA IN THE FIRST CENTURY BC

Diodorus Siculus indicates that Malta was a prosperous hub, with significant maritime commerce conducted through its ports. Figure 7.1. (cont.) might have represented the Tanit symbol, example from Għar Barka, tomb 3, found 30 April1925 (after Sagona 2002, fig.  31); (15–21) the evolution of the twin-nozzle lamp in Phoenician-Punic tombs, from the earliest (15) diagonally down to the latest (21), a development decreasing in size; (22) single-nozzle lamp, rare in Malta; (23) single-nozzle small and closed mould-made lamp appearing in Punic, Melitan phase IV contexts and after; (24) rare twin-nozzle lamp covered by a strip of clay at the front; (after Sagona 2002).

Melita and Gaulos during the Punic Period [I]‌t possesses many harbours which offer exceptional advantage, and its inhabitants are blest in their possessions; for it has artisans skilled in every manner of craft, the most important being those who weave linen, which is remarkably sheer and soft, and the dwellings on the island are worthy of note, being ambitiously constructed with cornices and finishes in stucco with unusual workmanship. This island is a colony planted by the Phoenicians, who, as they extended their trade in the western ocean, found in it a place of safe retreat, since it was well supplied with harbours and lay out in the open sea; and this is the reason why the inhabitants of this island, since they received assistance in many respects through the sea-merchants, shot up quickly in their manner of living and increased in renown [Diodorus Siculus 5.12.2–3]. THE CITY AND OTHER URBAN CENTRES

At issue in understanding the cultural trajectory of the islands is the growth of urbanism and rural development in the Mediterranean. From the 1970s, interconnections between urban centres and their rural hinterlands have been debated and archaeologically tested drawing on a growing body of survey data in some Punic contexts.6 Definitions of what constitutes a city have tended to rely on a list of criteria by which to assess the scope of ancient establishments – universal criteria applied to all.7 More recently, discussions are focussing on features and trends specific to individual locations, as well as distinctions between the development of urban centres in homeland Phoenicia and in the western Punic settlements.8 Viability of a burgeoning urban centre hinged on political stability, social harmony and reliable economic ties to surrounding rural areas.9 It is clear that in time, urbanism reached a pinnacle in the form of a city also called ‘Melita’, but its proto-urban past and a detailed understanding of its spatial context, especially in regards to smaller urban or rural settlements, farmland, ports and communication systems, has yet to be written.10 Signs of a viable local community in the Bronze Age–Iron Age interface years, as we have seen, have been hard won from the surviving data. The situation is even more obscure as the first millennium progresses. Hints of indigenous settlement patterns have been detected, perhaps fossilised, in enduring cultural landscapes, but they have hitherto not been archaeologically proven.11 In turn, it is clear from later Roman-period sites that they often had earlier Punic occupations, whether the remaining structural elements date from the earlier period or were modified or replaced by later inhabitants. Early and erroneous classifications have led to some substantial structures being identified as exclusively Roman when, in reality, they came into existence during the Punic period. This is the situation for a number of structures, among them round towers in the south of Malta and substantial rural houses in, for example, Birzebbuga, St. Paul’s Bay and Zejtun. Punic-period pottery fragments litter fields and open ground; some might dismiss such evidence as



The Archaeology of Malta collateral, imported during soil improvements, but this is unlikely. It is more convincing to interpret the scatters as signs of archaeological sites surviving under long-established field plots. Existing historic hamlets and the roads which link them may have their origins deep in antiquity, as is evident at Zurrieq with its surviving Phoenician-Punic building.12 At what point do we say that land-use patterns are Punic or that they are Roman when intense and enduring occupation of economically profitable locations is apparent? Moreover, we should be mindful that preserved within the surviving Roman sites may be the blueprint of earlier Punic land-use and settlement plans.13 It is this possibility and the persistent downplaying of Punic achievements, which rarely emerge from the shadows of Rome, that will be explored in this chapter. No expansive settlements have been excavated in their entirety, and hence the scale of urban development can only be inferred from fragmentary data. Nonetheless, while archaeologists debate what constituted a city, ancient classical authors considered that urbanism was advanced in the archipelago. Significantly, as early as the mid-fourth century BC,14 Pseudo-Scylax indicated that Malta possessed a ‘city’ and a harbour town. Ptolemy (ca. 90–168 AD) documented two major locations, Melita city and a peninsula settlement (i.e., ‘Chersonnes’ or ‘Chersonesus’), as well as two important sanctuaries, the Temple of Juno (the Phoenician Astarte) and the Temple of Hercules (the Phoenician Melqart), the locations of which have attracted a great deal of scholarly interest over the centuries.15 Further to the ancient perception that a city had grown up in Malta, in the absence of physical evidence, some aspects of its characteristics can be determined anecdotally through classical references to the island. In particular, administrative and social unity is implied by the ancient sources. Malta’s renowned maritime trade must have thrived on significant surpluses, with an economy that had moved away from subsistence to market-oriented production. Specialisation within the labour force is documented, as well as an industrial sector associated with the city that produced fine textiles and other crafts, commensurate with other Punic urban centres in the West.16 The strong commercial identity of the urban centre is reflected through the minting of local coinage in the period of Roman domination.17 If Mdina-Rabat was the site of Melita city, it was located within a geographically defined territory, surrounded by the rural lands and linked by various roads perhaps to other minor urban centres and to coastal regions. In terms of distance from the city to points of anchorage which might have been exploited in Punic times, then following direct lines it lay approximately 7 km from Għajn Tuffieħa Bay, 6.8 km from St. Paul’s Bay, approximately 7 km from Burmarrad ria,18 under 9 km from the Marsa Creek region of the Grand Harbour and approximately 14.5 km from Marsaxlokk Bay. Given the strong maritime nature of Phoenician-Punic economies, the city might have received and distributed goods through all of Malta’s active ports, great and small.

Melita and Gaulos during the Punic Period The areas of Phoenician and Punic settlements around the Mediterranean rim offer some useful perspectives for the Maltese Archipelago. Homeland Tyre, an island settlement, was approximately 17 ha (40 acres) in size, Beirut 2 ha (5 acres) and Sidon and Arwad some 60 and 40 ha (150 and 100 acres), respectively.19 Towns in Phoenician Iberia and in Ibiza encompassed around 2 to 4 ha of land.20 The island of Motya off the west coast of Sicily is about 40 ha.21 By the sixth century BC, Carthage (Tunisia) had grown to be an impressive city. At the pinnacle of its power, it covered at least 55 ha (about 136 acres). Surrounded by fertile hinterland, Carthage reaped the benefits of its prime location on the maritime routes flowing east-west through the Mediterranean. The city bore the hallmarks of Phoenician culture: advanced harbours, large urban districts, cult centres dedicated to Phoenician gods, industrial precincts on the city outskirts and extensive burial grounds dedicated to adult or child burials.22 The Maltese Archipelago stretched across 316 km2 and comprised urban areas, smaller towns and productive hinterlands within a Punic cultural overlay. Rabat, if it proves to be the site of ancient urban Melita (city), was approximately 35.3 ha (87.4 acres) and together with Mdina, a likely sacred and/ or administrative precinct, covered some 42 ha (103.8 acres). Overall, the persistent portrayal of the archipelago as a minor entity in the Phoenician-Punic occupation of the western Mediterranean is a modern construct born of the current global perception; the ancient world perceived Malta differently. It is unknown whether the shape of the city’s development was dictated by the contours of the land or followed a regular Hippodamic grid plan. Examples of the former style can be found in Punic cities such as Tharros and Monte Sirai in Sardinia.23 Nor is there concrete data on population size in Malta during the Punic period; there is even less evidence for the size of the city’s population. Estimates of 5,000 people or more are generally considered requisite for a city, but the actual head count is in fact incidental because Melita was considered a city in antiquity regardless.24 For the ancient Greeks, a city implied the presence of centralised government, accommodated in designated buildings and administering a defined territory, a level of urban infrastructure and a public character evident in a theatre, a gymnasium, an agora and amenities such as a water supply. Size was not a determining factor.25 Attempts have been made to assess population numbers in Malta based on a ‘body count’ from funerary contexts and the extent of available arable land – estimates vary widely between 489 and 17,555 people spanning the course of the first millennium and into the first century AD.26 It is possible that rock-cut tombs catered to only the affluent in society; presumably the city also was home to poorer communities which disposed of their dead in more modest and less archaeologically visible ways. Nonetheless, the presence of large necropoles (by any Punic settlement’s standards) occupying the rocky high ground in Rabat, Mtarfa, Għajn Klieb and Qallilija, begs the question of where the urban buildings stood. Most scholars assumed that Melita city occupied the Rabat-Dingli



The Archaeology of Malta plateau, but the evidence is equivocal. It is unlikely that residential houses were constructed close to cemeteries; the Phoenician-Punic tradition was to have quite defined necropoles that stood apart from major settlements on high ground, beyond waterways, or on coastal shores.27 In time, some burial grounds were reclaimed. Smaller Roman residential quarters, at the site of the Domus Romana Museum in Rabat, for instance, were built over Phoenician-Punic tombs, not over older residential remains.28 Not surprisingly then, Nathaniel Cutajar has demonstrated through cumulative archaeological excavations that there is a dearth of Punic residential remains in Mdina from the early sixth century BC through the rest of the millennium.29 Nonetheless, it is evident that at least 2 m of archaeological deposits remain locked under Mdina’s historic buildings. Whatever habitations existed in the Phoenician period did not continue in this area after the sixth century BC. Ashlar structures were uncovered in Mesquita Square (Mdina), thought by some to belong to a fortification wall of Punic date. This wall defines an open public space, perhaps with administrative buildings at its heart, or the temple complex dedicated to the deity, Melqart-Hercules.30 The elusive Temple of Hercules, also possibly located in Zurrieq, is further discussed later. An administrative, often walled upper precinct and a residential lower sector are in keeping with known Phoenician settlement plans.31 In 2005, excavations in Cathedral Square (Mdina) brought to light a substantial masonry wall surviving to around 8 m in length, the blocks of which were hewn from Coralline limestone, each measuring 1.5 m long by 0.5 m wide.32 Based on the size of the blocks, the wall was attributed to the third to first century BC, thus belonging to the period of Roman influence.33 Antiquarian observers commented on architectural fragments that were still exposed in their day.34 In 1920, Zammit noted heavy masonry harvested in antiquity from a ‘pagan’ temple and used to construct a level platform for further structures outside historic Mdina, on its north-western side, in the vicinity and east of the Domus Romana site.35 Two other threads of evidence for an official building are provided by Phoenician-Punic statuary. A  statue which may have depicted a female dignitary or the deity Astarte was once fixed within the historic main gate at Mdina (a secondary placement).36 The doves on her chest are one of Astarte’s symbols.37 A  second fragment, of a bearded male head, possibly of similar date, was found in a cistern north of the platform mentioned earlier.38 Additional archaeological finds from the streets and squares of Mdina have been recovered through archaeological sondages related to modern municipal trenching and rehabilitation works (from 2002). Though the subject of numerous media bulletins, they still await publication.39 Rabat, too, was clearly occupied from Roman times and perhaps from the late Punic-Roman interface when the islands came under Roman influence, but irrefutable evidence of earlier occupation is wanting. This is evident at Saqqajja (the eastern sector of Rabat), where a Roman house was built on

Melita and Gaulos during the Punic Period bedrock.40 This district yielded substantial architecture, exposed during excavations in the 1960s, 1970s and 1980s, none fully published. Most has been assigned to the Roman period, but some finds were reputedly of Punic date.41 Burials on high rocky ground in the southern sectors of Rabat (largely south of the ditch defined by Triq Santa Rita), in Mtarfa, Għajn Klieb, Għajn Qajjied and Qallilija, might indicate the expansion of Punic settlement, whether the city itself, small urban satellites of the city centre or rural hamlets perhaps following the Għeriexem and Qlejgħa valleys flanking the Mtarfa and Rabat plateaus (Figure 7.3, no. 13).42 This possibility waits to be further tested archaeologically, and undeveloped farm plots are conducive to non-invasive survey work in the future. Antonio Caruana thought Għar Barka was the site of a Greek suburb. However, given that he laboured under the misconception that the prehistoric megalithic structures were Phoenician, he would not have considered that more regular structures, whether at Għar Barka or at other localities, could be Punic.43 Terracotta sarcophagi of Phoenician-Punic date, however, have been found in the Għar Barka district (see Figure 6.5, nos. 1 and 2).44 A ditch or moat just under 700 m long (25 m wide and 3.6 m deep) was cut across the width of Rabat. It is most evident in aerial photographs of the grounds behind St. Paul’s Church. The length of the street (Triq Santa Rita) traces the line of the ditch as it heads north-west across the plateau. A segment at the north-west end was exposed in 2007. Associated fortification walls defining the area also have been documented.45 Usually described as Roman, the ditch was cut in the Punic period, according to Anthony Bonanno,46 and thus delineated what was once the urban area. Ditches as a defence strategy are known elsewhere in the Punic world, such as the one dug around the perimeter of Carthage and its territory47 and others in various locations in the Phoenician homeland.48 Anomalies to account for, however, are the many rock-cut tombs that were found in New Street, Rabat (identified as the current Triq Navarra Gusman49), which spanned the Melitan phases III to VI, and other tombs under the Domus Romana near the citadel of Mdina which fall within the space defined by the ditch.50 The presence of these burial grounds might indicate that urban sprawl had engulfed presumably redundant burial sites, as it did in Carthage.51 It is also possible that the ancient wall veered eastward following the line of Triq Navarra Gusman, leaving the tombs on the outside. Considering these aspects, a tentative date for the ditch falls probably around the beginning of the first century AD or later. It is possible that the ditch served a primary role during the Roman period of domination as a quarry for monumental construction works perhaps within the Mdina walled precinct, if not for the fortification wall itself, and a secondary, defensive purpose after the rock was removed. For these reasons, more will be said of these defence works in Chapter 8. Another sizeable urban district may have grown in the Zurrieq region. Here a well-preserved building nestled within the grounds of the parish priest’s



The Archaeology of Malta house (Figure  7.2, nos. 1–5) is remarkable because it is considered by some to have survived intact from Phoenician-Punic times and is still in use some 2,500 years after it was constructed! An adjacent joining length of wall can still be traced from the street, incorporated into modern dwellings. The original and larger complex at Zurrieq was in ruins when Houel, Boisgelin, Caruana and Mayr made their assessments of the structure.52 Houel depicted the walls as free-standing on open ground (Figure 7.2, nos. 2, 4, and 5). One of the tell-tale architectural elements is the Egyptianising cornice – plain gorge cornice or cavetto cornice – that caps the walls (Figure 7.2, no. 1); such architectural features once decorated the Phoenician-Punic temple buildings at Tas-Silg.53 The size of the stone blocks, too, falls within a Punic range. Most accept that the building is an ancient structure, and this is the impression given by the watercolour sketches made by Jean-Pierre-Laurent Houel in 1782 (Figure 7.2, nos. 4 and 5).54 It was archaeologically investigated in 1938 and 1964. Excavations to the south revealed deposits that reached 1.66 m in depth, and the structure was found to be resting on a ‘stepped foundation on the edge of a vertical drop in bedrock, probably one side of an ancient quarry’; pottery extends from the Punic period to the modern era.55 An ancient cistern (perhaps a converted quarry) and a debris-filled well also were recorded under a nearby house.56 Thomas Ashby thought that the extant building might have been a rural dwelling, Phoenician in date and modified in Roman times.57 It also has been mooted that it is the famed site of the Temple of Melqart-Hercules; the merits of this suggestion are considered later. On current evidence, it seems likely that this building is a remarkable survivor from the Phoenician period. In addition to this structure, an ancient ashlar wall approximately 20 m in length, presumed to be of the Roman period and standing two to three courses high, still flanks the road Triq iz-Zurrieq (in Safi not far north-east of Zurrieq). This substantial wall contributes to the view that the region had a settlement of some size.58 It is possible that one of Houel’s etchings of 1782 depicted this Safi wall. In favour of this identification is the background landscape, with a town at the centre (possibly Mqabba) and high ground rising beyond, perhaps equating to the Dingli-Rabat plateau. Otherwise, the written description of his sketch appears to indicate that it was the ruins at Tas-Silg, ruins which Houel identified as the Punic Temple of Hercules.59 We now know that Astarte was the object of worship at Tas-Silg, not Hercules, as Houel supposed. Excavations beside the Safi wall in 1967 uncovered additional architectural elements: flooring, a room, T-shaped foundations and blocks with lewis holes. Poorly preserved pottery was noted but otherwise not described.60 Whether these structures fell within an urban area remains to be proved archaeologically, but it is likely given the scale of the buildings. Another strong indicator that this region was a significant Phoenician-Punic economic centre are the rural towers which ring the area to the east, standing between the Safi-Zurrieq region and Marsaxlokk Bay, from north to south at Ta’ Gawhar, It-Torrijiet and

Melita and Gaulos during the Punic Period

Figure 7.2.  (1) Top of the Phoenician building in Zurrieq showing the Egyptianising cornice (photograph by A. Sagona 2007); (2) plan of the building (after Houel 1782, pl.


The Archaeology of Malta


Tal-Bakkari. On the north-west side of Safi-Zurrieq stands Ta’ Wilga. More will be said about these structures later. It is worth pointing out that based on the current information, it seems that two of the oldest settlements which developed into urban centres  – Mdina-Rabat and Zurrieq  – are not in coastal locations. As yet, unlike for Rabat-Mdina, there is no evidence that a Middle or Late Bronze Age town preceded the Phoenician-Punic settlement at Zurrieq. RURAL CONTEXTS

At the height of its economic power, Carthage, along with many Punic settlements of the Mediterranean, was engaged in commerce over geographically wide-ranging regions, trading in a diverse array of commodities. Recognition of Carthaginian-Punic agricultural know-how is evident in the Roman seizure and translation of twenty-eight volumes on the subject, written by the Carthaginian agronomist Mago after the capture of Carthage in 146 BC.61 Within this context, the population of Malta may have had a degree of self-sufficiency, but imported transport amphorae in the islands and in marine locations indicate that they were keyed into the commercial food networks existing between Mediterranean lands.62 Organic remains from Tas-Silg point to pastoral activity: the raising of cattle, pigs, sheep and goats and presumably the manufacture of the secondary products they generated.63 Melita’s high-quality linen textile industry was long known from the classical references.64 Marine resources (e.g., shellfish, bony fish and sea urchins) also were exploited for food. Single murex seashells in tombs and in some numbers at Tas-Silg might indicate that purple dye was produced on the island, but whether the dye vats at the water’s edge at St. George’s Bay were exploited at this time is unclear.65 The environmental record for Malta taken from deep sediments at Marsa points to high levels of cereal production during the Phoenician-Punic period.66 Significantly, the greatest levels of olive (Olea) pollen occurred not in the Roman period but in the earlier Phoenician-Punic contexts. This trend is matched in North Africa and Iberia, where olive oil production can be traced to the eighth century BC.67 Considering this and, as we shall see, the Punic

Figure 7.2. (cont.) CCLIX, fig. 2); (3) drawing of the south-east side (after Bonanno and Vella 2000, fig. 3); (4 and 5)  the structure and additional walls as they appeared to the artist Jean-Pierre-Laurent Houel in the late  1770s (two images by Houel:  Greek House in Casal Zurico on Malta and Plan and Cross-Section of a Greek House, acquired by Catherine II from the artist courtesy of the Hermitage:  http://www.arthermitage .org/acquired-for-Catherine-II-from-the-artist.html).

Melita and Gaulos during the Punic Period

Figure 7.3.  Melitan phase III pottery forms: (1) amphora; (2) trefoil jug; (3 and 4) thick and large twin-nozzle lamps; (5) unguentarium; (6 and 7) plates; (8 and 9) cups modelled on Greek forms; (10–12) typical bowl forms (after Sagona 2002). (13) Map of Mdina and Rabat: X to Y: rock-cut ditch following the line of Triq Santa Rita;


The Archaeology of Malta


origin of farmsteads in Malta, a number of which have substantial olive works, it seems likely that oil production began in earnest well before Roman domination. Excavations in Gozo at Tal-Knisja also have found evidence that some of the terrace fields date to the Punic period.68 Similarly, traces of grape plant remains in environmental samples which coincide with a wine-drinking pottery kit demonstrate that viticulture appeared in the Late Bronze Age–Phoenician period, undoubtedly fostered by Levantine settlers.69 Dotting the archipelago are rock-cut pans, once used as pressing floors, with inter-connecting collection vats. These facilities were almost certainly for grape crushing and the production of wine (see Figure 6.8, no. 10).70 Associated small pits documented at some sites may have been used as stands for amphorae with round or pointed bases while they were being filled. The pressing floors resemble a similar range of facilities appearing in the Near East and dated to the late eighth century BC, though they probably originated much earlier in the Syro-Palestinian region.71 With the inhabitants of the Mediterranean lands showing a predilection for wine from the seventh century onwards, it is likely that such features were introduced during the Phoenician-Punic period and then may have seen many years of use. Confirmation of their antiquity came during excavations that began in 2005 in the area of Tal-Knisja (Mgarr ix-Xini, Xewkija, Gozo), which indicated that one such installation was used around 500 BC.72 Small clusters of treading pans along this valley suggest vineyards in close proximity and, given the modest scale of the facilities, that production probably was aimed at a local level. The pans were inexpensive and virtually indestructible installations, easily reused during each harvest season.73 Agricultural enterprises still included cereal crops, but during the Phoenician-Punic period, wine making and olive oil production were firmly established, with olive oil forming a significant commercial and export commodity. Waste by-products of agriculture and animal husbandry are likely to have provided fuel, particularly valuable in the absence of timber on the islands.74 Bonanno identified two types of villa sites which marked the Maltese countryside, ‘residential’ and ‘rustic’, which in his scheme equated basically with dwellings that had more opulent proportions and fitments than farmsteads Figure 7.3. (cont.) (1) reputed site of the Temple of Persephone, now marked by St. Nicholas statue in Oswald Street; (2)  Mtarfa bell-shaped pits; (3)  Mdina, St. Paul’s Square excavation; (4) Mesquita Square excavation; (5) Xara Palace square excavation; (6) Domus Romana Museum; (7) UCLA Rabat excavations1983–4; (8) Santa Margerita Cemetery; (9) massive Roman architecture uncovered in 2006; (10) 1956–7 excavations in Tal-Franciz field; (11) number 62 Triq San Pawl excavation of 2008; (12) Roman ruins, mosaic workshop; (13) Triq Nikol Saura, massive architecture; (14) Fiddien Roman villa (Bonanno 1977, site 5) (map after Vella 2012, fig. 1).

Melita and Gaulos during the Punic Period with dedicated workrooms for olive oil production or other commodities.75 In the past, the villas were generally assigned to the Roman period, even though a number have clear indications of Punic origins. An adherence to a rigid periodisation of the Maltese archaeological record is circumscribed and misleading.76 It is overly simplistic to view Rome as a conquering force in the face of which Phoenician heritage and the Punic way of life evaporated instantly. Inadequate reporting, few systematically investigated sites or concomitant field surveys, a restricted approach to archaeological interpretation and wanton destruction, as in the case of one house in Kercem (Gozo), have limited the scope of enquiry into the Punic rural landscape. No unaltered, typical Phoenician-Punic farmstead fossilised in Malta’s archaeological landscape has yet been identified largely because of the cultural continuum of many sites into the first millennium AD. But the evidence does exist to prove that rural Punic Malta is not just conjecture.77 In some cases, the evolving architecture of farmsteads reflects the hybrid nature of a persistent Punic building tradition, adapted to reflect Roman influence and taste. Those archaeologically investigated sites which have claim to Punic antecedents are outlined here. One rural house in Zejtun, within the grounds of a girls’ secondary school, was discovered in 1961 during building works (Figure  7.4, B). Excavations began in 196478 and were continued later by Tancred Gouder (1971–4)79 and by Anthony Bonanno (1976–7 and 2006–present).80 The site also has been prospected using ground-penetrating radar by the Belgium-Maltese Survey team.81 While it had a Roman period of use, confirmed by coins dated between 222 and 361 AD,82 it was clear from significant ceramic finds that the site also was occupied during the Punic period. Some of this pottery is perhaps as early as the Melitan phase II (Coarse Grey Ware), and significant quantities of Punic Crisp Wares fall clearly within late phases III to IV, according to the sequence constructed for the Punic funerary record in Malta (Figures 7.10 and 7.11 for phase III–IV Punic pottery types).83 One cooking-pot fragment was inscribed with ‘Astarte’ or possibly ‘Anat’, both Phoenician goddesses.84 Inscribed pottery is an unusual occurrence outside the sanctuary site of Tas-Silg, and it is likely that the pot was used in a household shrine.85 At the very least, the inscription suggests that the occupants were once culturally Phoenician-Punic, speaking (and possibly literate in) the language of their ancestral homeland and steeped in Phoenician-Punic religion. Floors of four rooms were covered with lozenge-shaped tiles, and the walls had painted plaster comparable to the Domus Romana in Rabat and the structures in San Pawl Milqi.86 Such decorative features were catered to Roman influences filtering into the islands. Yet, architecturally, the plan of the dwelling has more in common with Punic houses in places such as Kerkouane (Tunisia)87 than Roman villas. Rectangular buildings had clusters of rooms placed around a simple courtyard with an inward-looking perspective. Some were domestic, whereas others, to the



The Archaeology of Malta

Figure  7.4.  (A) Plan of structure in Saqqajja I, Rabat (after MAR 1965, fig.  2). (B) Plan of Punic-Roman villa in Zejtun in the grounds of a girls’ secondary school (after plans drawn by Keith Buhagiar in 1995 and Josef Briffa 13 August 2000): (1–4)

Melita and Gaulos during the Punic Period north, were equipped with installations used in the production of olive oil. It is likely that the establishment was once located close to olive groves. Unfired mud-brick construction in this Zejtun site and more than five Punic tombs to the east also point to a Punic rather than Roman origin for the complex.88 As the project is still ongoing, we await final word on the site.89 The Punic origin of the house site known as Ta’ Kaccatura (Birzebbuga), close to Borg in-Nadur, is indicated by pottery sherds and coinage found in the deposits.90 It is quite clear that a Punic building was demolished in the north-eastern sector to make way for a house with a central peristyle, but significant remains of the earlier structure survived. Evidence of the older sectors of the structure comprised ‘concrete’ (Figure 7.5, no. 1, G) and torba flooring (Figure 7.5, no. 1, D), rock-cut depressions which probably defined earlier rooms (along the north-west and north-east walls of rooms D–F in Figure 7.5, no. 1) and walls displaying the Punic opus africanum building technique (shaded wall lines in Figure 7.5, no. 1), which used a combination of large upright orthostats between which were smaller stone blocks.91 Rooms (Figure  7.5, no.  1, D–F) off a corridor (Figure  7.5, no.  1, C) and a doorway (Figure  7.5, no.  1, B) are all that remain of the original building, but rock-cut steps (Figure 7.5, no. 1, A) might indicate that much of the earlier Punic residential quarters lay at a lower level, a pattern reproduced in the later Roman period dwellings. Located in and around the Punic sector of the building were an olive press and other installations, perhaps indicating that oil production continued despite the building renovations of later years (Figure 7.5, no. 1, F and north of D). A tomb of Phoenician date also was found in the vicinity.92 As in the case of a number of sites of Punic origin, especially in rural contexts, very large square rock-cut water cisterns with simple square-sectioned pillars and massive slab roofing are located close by (Figure 7.5, nos. 1 and 2). A substantial rural structure associated with the olive oil industry was located between the towns of Luqa and Mqabba; it is now referred to as the ‘Tad-Dawl house in Ħal Kirkop’ (Figure 7.5, no. 3). Caruana investigated the site and identified the buildings as Greek; as we have seen, he considered the prehistoric megalithic structures with curved walls as Phoenician-Punic.93 Significantly, Caruana did draw attention to remnant cornice and ashlar blocks similar to those used to build the Phoenician-Punic Zurrieq building and Figure 7.4. (cont.) rooms with lozenge-shaped terracotta tile pavement; (5) raised area formed by ashlar blocks; (6) olive press base; (7) ashlar stone block; (8) production area; (9) olive press base; (10) storeroom; (11) water channel leading to the cistern (13); (12) dislodged ashlar stones; (13) water cistern; (14) water channel leading to the cistern (13); (15) pit measuring approximately 110 cm deep; (16 and 17) stone-covered channels (drains?) (after Bonanno [no date]).



The Archaeology of Malta

Figure 7.5.  (1) Plan of Punic and Roman house remains at Ta’ Kaccatura (Birzebbuga) with substantial underground cistern at the site immediately to the east; zones A–G

Melita and Gaulos during the Punic Period those employed in architecture at the Punic site of Tas-Silg. Massive stones from olive-crushing mills, as well as a complex of vats, tanks and pools to collect the oil, were installed in the northern sector of the complex. The architecture reflects a Punic house plan, consisting of clusters of rooms which essentially open into the building, creating private domestic space and enclosed production works. Caruana claimed that some smaller rooms in the south of the building were animal pens but gives no reason for this conclusion; they might equally have been domestic areas and storerooms. Two thresholds (Figure 7.5, no. 3, D1 and D2) were fitted with grooves along which door slabs comprising sizeable stone discs (2.43 m in diameter) were rolled – a sealing technique known from Levantine funerary contexts and one, in the context of this building, which may have acted to fortify the building against theft.94 Two large rock-cut water cisterns with simple squared pillars supporting massive roofing slabs mirror the Punic design at other rural houses already mentioned and also those near the round towers (Figure 7.5, no. 3, C1 and C2). Punic tombs in the region may well have served the occupants of this complex. The house was lost to large quarries in the region (immediately to the south of the Luqa airport and in the lands caught in the crook of the two runways). It is worth noting that Caruana’s published plan differs to one held in the National Museum archive which has been reproduced in recent publications. The latter appears to have oddly angled walls in the southern and eastern sectors which were rectified in the final form of the plan accompanying Caruana’s report.95 On the isolated headland of Ras ir-Raħeb (also referred to as Ras il-Knejjes) in the far western reaches of Malta (Figure 7.6, nos. 1 and 2), wall lines of a regular structure were exposed in 1961–2, incorporating two upright stones that date from the Neolithic period (Figure 7.6, no. 1, a).96 Stone for the site might have been quarried near the western side of the cliff edge. A boundary wall follows the arch of the precipice for at least 60 m.97 Arguments about the building’s function remain unresolved. Architecturally, it does not have features commensurate with a temple. Nor are there middens with accumulated ash, bone, pottery or other temple discards, nor scatterings of pottery inscribed with dedications to a deity, so common at Tas-Silg, although those kinds of deposits probably had a specific ritual function concerning the goddess Astarte. A few small female figurines and others which depict a Herculean-styled male with ragged lion’s cape were recovered, easily explained as items from a household shrine.98 Other finds included two Siculo-Punic coins, some Punic-period pottery and a carved ivory plaque depicting a boar of late Punic inspiration.99 A  large courtyard was covered with a cocciopesto floor (Figure  7.6, no.  1, b), Figure 7.5. (cont.) have remnants of the earlier Punic building; H is a peristyle feature added in the Roman period; (2) view across the cistern from the north-east corner (photograph by C. Sagona); (3) Tad-Dawl house plan with olive oil production facilities in the north and cisterns to the south (after Caruana 1888, fig. 17).


Figure 7.6.  (1) Plan of a Late Punic–Roman house at Ras ir-Raħeb, in the far west of Malta:  (a)  standing prehistoric megaliths; (b)  courtyard with Cocciopesto floor; (c)  room with tile floor; (d) stone with square-cut pits possibly connected to olive oil production. (2) View from Baħrija to Ras ir-Raħeb (photograph by C. Sagona). 236

Melita and Gaulos during the Punic Period comprised of crushed red pottery highlighted with white marble tiles. This type of flooring was found at Tas-Silg, but it is common to many Punic locations, including domestic sites.100 The ground plan, with corridors, remnants of rooms and water storage facilities with quantities of domestic Roman coarse wares, suggests that it was a house.101 A stone with square-cut pits (Figure 7.6, no.  1, d) was probably part of an olive press, but this has been disputed.102 Nonetheless, artefacts of Roman date and architectural elements indicate the longevity of occupation at the site. From 1963 to 1968, the Italian Archaeological Mission to Malta excavated the ruins of San Pawl Milqi, close to St. Paul’s Bay (Figure 7.7, nos. 1 and 2).103 Here, too, there were unmistakable signs of the Punic period extending back to the third century BC. Evidence such as stamped amphorae fragments ties the site to a Punic agricultural system.104 Ample pottery, especially in the drains (Figure  7.7, no.  2, C10, C13 and C14) and water cisterns (Figure  7.7, no.  2, -1- and -2-),105 stands as testimony to a significant Punic presence at the site. Fictile masks of Sicilian form link some occupation to the late fourth to early third centuries BC. Two Punic tombs, architectural elements such as a volute or Aeolic capital (variations of which were commonly used in Carthage and Kerkouane, North Africa), and an altar base of pre-Roman (i.e., Punic) date also point to earlier occupation.106 Reused stone blocks, old painted surfaces and the carved architectural elements indicate that the original house comprised ‘elegant residential quarters’,107 recalling Diodorus Siculus’ comments concerning the elaborate architectural features of Punic buildings in Malta. The house plan and nature of the milling installations are of the type documented in Punic North Africa rather than Roman Italy.108 As already seen at other sites, a representative number of Punic inscriptions from the complex are particularly indicative of the ethnicity of the people who lived and worked there. The language of the inscriptions, as well as the structure of their Phoenician-Punic names, refers to familial connections and reverence for Phoenician-Punic divinities.109 Fire probably destroyed the buildings in the third century BC. Other architectural remains are less well represented. Little has been recorded of a site at Ħal Far in the south of Malta, investigated in March 1922, but a large building complex comprising several rectangular and square rooms with well-made torba flooring and a passage between the chambers was documented. Olive oil production installations, a possible kitchen, pottery and several coins spanning the Punic and Roman periods were noted.110 The site was reburied, and the area became an aerodrome. A Phoenician house of large proportions, built of stone with several rooms, was found in Kercem (Gozo), but the owner of the property purposefully destroyed it; some pottery was recovered from the site.111 Zammit also recorded possible Punic architecture within the ruins of the Tarxien Neolithic complex, along with quantities of Punic pottery and artefacts, as well as some signs of inhumations ascribed to this period.112



The Archaeology of Malta

Figure  7.7.  (1) View over San Pawl Milqi looking east (photograph by A.  Sagona 1991); (2) plan of San Pawl Milqi, with shaded areas depicting remnant Punic features, namely, an open paved area, perhaps dating as early as the third century; wall lines 4a, 01 and 5a and their foundation trenches; intersecting walls OI and OII in the eastern

Melita and Gaulos during the Punic Period PUNIC RURAL ROUND TOWERS

Six round towers survive in various degrees of preservation in the main island of Malta, but none have been documented in Gozo. Though they were originally assigned, in turn, to a prehistoric, a Phoenician, and a Roman date,113 they are likely to be Punic, their use extending into the period of Roman influence. A climate of unrest in the central Mediterranean which sparked three Punic wars (spanning 264–146 BC) may have led to their construction and continued use. Control over Sicily, in particular, was contested in ongoing military and naval campaigns by western Greek interests and later by clashes between Rome and Carthage.114 Melita and Gaulos were, in a real sense, close to the front line. Fortifications were constructed around public and private establishments, such as the rural villa and its oil works at San Pawl Milqi and the temple at Tas-Silg.115 Some of the free-standing towers appear in an arch to the north-west and east of the Zurrieq-Safi region, which, as already hinted at, may have been a significant urban centre, perhaps with the fabled Temple of Melqart-Hercules at its heart. Given the towers’ locations, they may have stood in small hamlets close to communication routes leading into and out of the town. Much more work needs to be done to determine whether these structures should be viewed collectively as part of a combined defence system protecting the urban area;116 alternatively, they may have been strongholds designed to fortify and defend vulnerable rural regions exposed to plunder by hostile parties arriving in Marsaxlokk Bay to the south-east. Pottery recovered during excavations at Ta’ Wilga tower (in the Mqabba region, west of the Kirkop-Safi-Zurrieq area) (Figure 7.8, nos. 3 and 4) in 1910 was identified as Punic, although none was illustrated in the brief account of the investigations, and a Phoenician tomb also was located nearby.117 At Ta’ Gawhar (Kirkop) (Figure  7.8, no.  1), finds in stratified deposits included two bronze buckets and a double-bladed iron axe head, a gold wire earring and a carbonised bread roll which were buried under the roof material when the building burnt. A coin dated to the first century BC indicates that the building was still in use after the period of the Punic wars. An elaborately carved stone block served as a shrine, which stood on the roof given its position in stratigraphic sequence of floor deposits and roof

Figure 7.7. (cont.) corner; votive deposit (amphorae 1 and 2) and in pits (SSW perhaps associated segments of wall lines 6 comprising several large blocks); Punic tombs (2 and 5)  and deposits in canals C10, C13 and C14 and water tanks -1- and -2- right of area F and G (see Cagiano de Azevedo 1969a, pp. 94–5; 1969b, pp. 114–15; 1969c, pp. 120–4; plan after Bussuttil et al. 1969, fig. 7).



The Archaeology of Malta

Figure 7.8.  (1) Ta’ Gawhar; (2) Ta’ Cieda tower; I–III are excavation trenches; (3) Ta’ Wilga plan; (4) view of Ta’ Wilga (plans after MAR 1960, figs. 3 and 4; Ashby et al. 1913, fig. 26; photograph by C. Sagona).

Melita and Gaulos during the Punic Period debris.118 The other structure which has been archaeologically investigated is Ta’ Cieda (San Gwann) (Figure 7.8, no. 2), which shared a similar origin and function, although it continued in use well past the third century AD.119 Of Ta’ Cieda and the remaining sites, It-Torrijiet (Zurrieq), Tal-Bakkari (Ħal Far) and possibly Tas-Santi (south of L-Imgarr), we have little more than brief reports.120 The outer-perimeter tower walls vary in thickness:  1.70 m at the foundation narrowing to 1.10 m at the top for Ta’ Wilga, 2 m for Ta’ Cieda and 3 m for Ta’ Gawhar. The overall diameters were about 12.80 m (Ta’ Wilga), 14–16 m (Ta’ Gawhar) and about 25 m (Ta’ Cieda). As strongholds bolstering the security of prime rural landscapes, they may have served a dual role: to protect the wealth – possessions and produce – of local farm estates and to act as a line of defence for the town. At Ta’ Cieda, a large storage jar used for grain was found, as well as horseshoe fragments. Substantial Punic-style water cisterns were cut into the rock near the Ta’ Wilga, Ta’ Cieda, Tal-Bakkari and Ta’ Gawhar structures,121 indicating that people were living in close proximity to the towers. There is potential in the Kirkop-Safi-Zurrieq area, for instance, to document other features in the landscape which might help to define the role of the towers there. Additional wall lines and further structural remnants, possibly from another tower, also were reported near Ta’ Wilga.122 Numerous architectural relics also can be seen in the current field walls, along with massive accumulations of small cobbles occurring in numerous locations east of Safi-Zurrieq which far exceed what is generally observed for rural, dry-stone walling. Defensive towers can be found in other regions which were contemporary with or part of the Punic world.123 In the Baetis Valley region of Spain, towers with square floor plans, such as Turris Lascutana, stood on significant trade routes and in highly profitable agricultural lands. They date between 400 and 200 BC, earlier than the time suggested by classical authors who accredited Hannibal with their construction.124 In wider contexts, they served as lookout posts, for defence and possibly for border or coastal protection, as reflected by locations in North Africa and the islands of the Mediterranean, all under Carthaginian influence.125 In appearance, the Maltese examples strongly resemble towers built on the Greek island of Siphnos (Sifnos) in the fifth to third centuries BC; large ashlar blocks follow a round floor plan, there are internal wall subdivisions and at some sites the quality of the construction technique is mediocre, as dictated to a degree by the available stone resources.126 The close proximity of some Greek towers facilitated signalling from one to the other, also a possibility for the buildings in the regions of Ħal Far, Kirkop and Zurrieq in Malta. Within the Greek landscape, some towers served to protect the metal resources of the interior in times of unrest, some were located at the heads of water courses, thought to be points of refuge during pirate raids, and others were linked to farming compounds, especially in the third



The Archaeology of Malta ­century BC.127 It was a similar need to defend lands and property which likely drove the independent development of strongholds in Malta, whether round or square in plan, rather than direct cultural contact with other locations such as Spain or the Aegean. Aside from these architectural remains, the open terrain of the Maltese islands is littered with Punic pottery scatters; they are the ubiquitous cultural ‘noise’ in the landscape, reported in passing during many archaeological investigations. When Margaret Murray excavated the Santa Sfia site in 1921,128 only a few megaliths remained in situ from a prehistoric structure; the remaining slabs appeared to have been re-cut for a much later building, of which only a small length of straight wall survived. Murray, curiously, identified the later finds as medieval on the basis of architecture. Zammit was quite clear about the nature of the artefacts found there: The potsherds collected by Dr Guest up to the 6th [August] are all of a Punic or late Roman character mostly amphorae, plates and jars. No earlier fragments except one which might have been bronze age. 1 bronze nail, 1 copper ring or ear-ring, 1 coin probably Maltese with hall mark.129

The last item was eventually identified as a Punic coin of the sort dated to the second century BC with veiled female head on the obverse and three Egyptian figures (presumably Isis, Osiris and Nephthys) on the reverse.130 No medieval pottery or objects were recorded.The team moved its operations to a small mound in Tal-Bakkari not far from one of the round towers, where two small chambers (L: 6 m; W: 3.6 m and L: 6 m; W: 2.7 m), side by side, were thought to date to the Roman period and may have been used as Mithraea (see later). Here Punic and Roman pottery, a cistern and a Punic tomb also were located.131 Of other archaeological remnants, Baldacchino132 noted animal remains and Punic potsherds in a sondage excavated at Bur Mgħez in the 1930s, although the context was unclear. In investigations outside the Għar Dalam cave,133 building foundations with torba flooring lay 0.6 m beneath the surface. Another trench was situated 25.9 m from the entrance and some 3.35 m above the creek, where there were ‘numerous Roman or Punic potsherds in the debris.’ Some Punic pottery was recorded within prehistoric ruins at Borg in-Nadur, quite close to the Ta’ Kaccatura house site which was Punic. Punic wares were found in the field soils overlying the prehistoric megalithic building,134 and at Ta’ Ħagrat, remnants of a rectangular building, Punic pottery, a mould-made hollow figurine and deffun flooring likely to be Punic in date were found.135 On the southern coast, 0.8 km west of Binghisa Fort, at Wied il Mista, in a prehistoric cave dwelling, Punic lamps and other pottery fragments were recorded.136 Punic material was also documented in a recent detailed survey of the Bidnija and Burmarrad valleys by the joint Belgo-Maltese Survey Project.137

Melita and Gaulos during the Punic Period Gozo has a similar Punic footprint in the archaeological record. Salvage investigations in response to council drainage works at It-Tokk, Victoria, produced [a]‌few sherds immediately above the natural clay [that] indicated a first occupation in the Early Bronze Age. Above this, levels including a few floors and rough walls, contained sherds of Punic type.Amongst these were some coarse handmade, red-slipped and burnished ware not previously recognised.138

The Punic level was thin and without sign of structures. HARBOURS AND PORT FACILITIES

Evidence of Punic activity is found in the vicinity of most major bay areas in Malta and Gozo. Aside from the sacred precinct at Tas-Silg, many ruins and tombs have been reported in the lands around Marnisi Hill (north–north-east of Borg in-Nadur) and Delimara Peninsula, both overlooking Marsaxlokk Bay.139 Anthony Bonanno suggested that an artificial inner harbour of Marsaxlokk, called ‘Il-Magħluq’, might date to the Punic period.140 The concept of a constructed docking place, known as a cothon, finds parallels elsewhere in the Phoenician-Punic world, for instance, in Carthage (Tunisia), Mahdia (Tunisia), Motya (Sicily) and Nora (Sardinia).141 There is no evidence for Punic occupation in Valletta itself, but extensive Roman remains have been found in the Grand Harbour region at Marsa (see Chapter 8); indicative of Punic commercial interests in the area are coins spanning the third to the first centuries BC.142 Punic tombs in the Paola and Tal Ħorr districts suggest that any settlements were south of the harbour.143 On the north-west side of Valletta, Ashby reported Phoenician or Roman remains in the Villa Frere at Pietà ‘in the innermost recesses of the harbour’, in reference to Marsamxett Harbour.144 With the sheltered harbour opening on its northern side, the natural formation of Manoel Island flanked by Sliema and Lazzaretto creeks is very similar to the cothon in Carthage.Though typical of Punic sites elsewhere, the island is now capped by historic fortifications, and there is limited scope for survey work. Marine explorations to the south of the island produced predominantly mediaeval artefacts.145 On the west coast of Malta, Għajn Tuffieħa Bay and Gnejna Bay within Golden Bay provide sheltered coves where there is some possibility of beaching smaller vessels. Over a kilometre inland is the large Roman bathhouse (Għajn Tuffieħa), but Punic burial sites and earlier Bronze Age remains dot the area. Gozo, too, has some suitable bays for short-term anchorage and inlets at Xlendi, Marsalforn and Mgarr ix-Xini. These bays were once further inland but have silted up over the centuries.146 Commerce may have been conducted through the bay of Mgarr, where some evidence of ancient pottery was reported during dredging works in the 1970s.147


The Archaeology of Malta


Ceramic Industry After a time of experimentation and adaptation in the production of ceramics, a homogeneous class of hard-fired durable pottery emerged which can be broadly categorised as Punic Crisp Ware.148 Variations in kiln conditions, gritty inclusions (fine to coarse examples) and surface-slip colours are broad ranging, but shapes reflect the standardisation characteristic of wheel-made wares. The significant change came with introduction of the potter’s wheel. Production appears to have shifted from a cottage industry to workshops where vessels were produced in large quantities for the local market. Stages in pottery development continue to be discussed at length elsewhere, but basic trends in production can be observed.149 Red-slipped wares, popular in phase I, slowly dropped out of the repertoire. In their place came vessels which in many respects reflected a certain time of experimentation in production (Melitan phase II) (Figure 7.1, nos. 1–11) and were often built of coarse gritty clays. Pottery was adequately but not always thoroughly fired, leaving darker cores and greyish hues near the surface. Slips tended to be thin and ranged through very pale yellows to greys.Vessels are thick walled and functional, with little evidence of decorative features. In the Melitan phase III period, slipped surfaces on cups, bowls and lamps were often burnished (the sub-category coined as ‘Thick-Slipped Ware’) to render the vessels impermeable; this was perhaps a strategy to conserve the precious commodities they held: wine and oil (Figure 7.3, nos. 1–12). Otherwise and in later phases, pottery surfaces generally are matt with red-painted bands (bi-chrome and poly-chrome painted examples are virtually unknown in Malta) depending on the fashion of the day; free-hand painted decoration is rare throughout the Punic repertoire. Overall, unpainted vessels predominate.150 An oatmeal texture is common where the slip covers but does not obscure the underlying grit within the clay matrix of the vessels. There was no sudden change in traits. Rather, we find a gradual shift over the centuries in fashions of pottery recovered from tombs and at sites such as Tas-Silg. Influences from North Africa, the Aegean and the Italian Peninsula can be detected. The tomb pottery repertoire evolved to include a core of shapes such as large jars or urns, jugs with trefoil lip, bowls, some with near vertical rims, bowls with rounded walls, wide-rimmed plates and two-handled cups modelled on Greek forms (Figures 7.10 and 7.11), and lamps always have twin nozzles, though they reduce in size over time (Figure  7.1, nos. 15–21). The wine-drinking pottery kit became less defined. No longer used were the disc-topped jugs which were a hallmark of Phoenician wares; they are not found after Melitan phase II. Gone, too, are the thistle-headed beakers, piriform jugs with trefoil lips, tripod bowls and small pear-shaped oil flasks.151

Melita and Gaulos during the Punic Period

Figure 7.9.  (1) Stylised plan of the features in an elaborate Rabat tomb found in Triq Ferris – a sun burst painted on the ceiling, a carved relief human figure in the rear wall niche and floor plan cut by three parallel trenches and a round pit at the top of the central trench; (2) the appearance of the Triq Ferris tomb from an archival photograph (the location of the tomb is lost); (3) face carved into the rear wall of a Punic tomb chamber in Qrendi, which appears to have the left hand held up near the face and an earring in the right ear, as well as some definition of the hair; (4) one of two carved figures on either side of the same Qrendi tomb (after Sagona 2002, [399] fig. 133:1–2 and [602] fig. 219).

Remarkably, Punic Crisp Ware endures – though degenerated, with very gritty clay and thin walls – well into the Roman period. The shape range expanded through the course of the Melitan tomb phases IV–VI to incorporate foreign influences such as unguentaria and closed, bulbous-bodied flasks modelled on Greek aryballoi. Roman fine ware and cooking wares and transport vessels



The Archaeology of Malta

Figure 7.10.  Melitan phase III–IV pottery forms: (1 and 2) large storage jars; (3–9) various urn forms (after Sagona 2002).

Melita and Gaulos during the Punic Period

Figure 7.11.  Melitan phase III–IV pottery forms: (1–3) trefoil-mouth jugs; (4 and 5) round-mouth jugs; (6–8) flasks; (9) large unguentarium; (10 and 11) bowls; (12 and 13) cups with horizontal handles modelled on Greek forms; (14–16) plates; (17–20) stemmed incense cups or burners; (21) shallow baking pan; (22) twin-nozzle  lamp (after Sagona 2002).


The Archaeology of Malta


were imported from various workshops of the Mediterranean, and these, too, would be placed in the Melitan tomb phases V–VI, but such vessels also were present in domestic contexts. Local contemporary products fell short of the slick, thin-walled and lustrous red-slipped vessels which were imported in numbers to the markets of Malta. A fairly standard range of beakers and open bowls was produced on the islands and slipped with matt red clays over comparatively grittier fabric (see Chapter 8). Of the widely renowned Punic products, only scant traces survive in the archipelago to hint at Melitan daily life. Loom weights in Punic contexts reinforce the notion of a labour force which specialised in the production of fine clothing; murex shells harvested for purple dye may have been employed to raise the garments to a pinnacle of quality. Refined carved ivory fragments from boxes or furniture, rare instances of gold and, more commonly, bronze jewellery and clothing pins reflect a population in step with other Punic settlements.

Beliefs, Rituals and Sacred Precincts Although the bulk of Phoenician-Punic written history and mythology may be lost to scholarship, some headway nonetheless can be made into the culture’s belief systems through archaeological discoveries. In this regard, the Maltese evidence has been particularly informative. Finds made in the sacred precincts and the hundreds of tombs in Malta have provided ample signs of ritual practice.

The Temple of Astarte-Juno Tas-Silg has furnished evidence for early occupation, but its most significant remains belong to the Punic period. Massive structures survive from the Melitan Temple of Astarte, a sacred precinct of considerable standing in antiquity. Classical references document its accumulated wealth, twice plundered. One of Houel’s etchings of 1787 is of a substantial wall comprised of large rectangular blocks, surviving to five courses in some parts, which ran parallel to and set back from an unpaved road. The caption reads simply, ‘Reste du Temple d’Hercule situé près de Marsasirocco’. Other blocks intersecting at right angles indicate additional wall lines. It is generally considered that the ruins that Houel depicted were those at Tas-Silg, a view somewhat reinforced by his description of the area in which they were situated.152 As we have seen, an alternative identification might be the Safi wall mentioned earlier. Hundreds of inscribed vessels excavated from Tas-Silg which carry simple dedications to the Phoenician goddess Astarte, rather than Melqart/Hercules, provide unambiguous evidence of the object of worship there.153 In the southern sectors were very large accumulations of organic remains, discarded pottery and

Melita and Gaulos during the Punic Period occasional small relics of devotion (amulets and figurine fragments) which date largely from the Punic period. QUINTINUS (JEAN QUINTIN D’AUTUN) AND THE PUNIC-ROMAN TEMPLE SITES

After living for six years in Malta, Quintinus tackled the identity of ancient architecture in his Description of Malta (1536), drawing on the classical authors but supplementing his account with personal observations made on the island.154 In his efforts to locate the two renowned temple sites in Malta dedicated to Hercules (Phoenician Melqart) and Juno (equating with Astarte), documented by Ptolemy, Quintinus looked to the south of the island in lands bordering Marsaxlokk Bay, ‘the south-eastern harbour’.155 First linking possibly the site of Borg in-Nadur with Hercules,156 he went on to describe extensive ruins he associated with Juno: I think that the temple of Juno, as one can see from the remains which still exist, could be considered not only as one among the great, but also among the magnificent temples of antiquity; it is situated about half-way between the town and the castle. The ruins lie scattered through many acres of land; the foundations of the temple cover a large part of the harbour, even far out to sea, built there on the hill-top, on a plain, and sheltered from winds on all sides by very steep slopes. The name of the place is very difficult to pronounce except by the local Maltese tongue. On the hill-top there is a shrine dedicated to the Blessed Virgin, called ‘Ta’ Qort’ (Tal-Qortin).157

Those following after Quintinus thought that he meant the peninsula extending to the tip where Fort San Angelo stands overlooking the Grand Harbour.158 Giovanni Francesco Abela (1647) made a link to a scatter of ancient structures presumably surviving at that time on the peninsula.To what locations Quintinus was referring remains at the heart of the problematic tangle of historic suppositions. Anton Bugeja’s synopsis takes an important step forward in sorting through centuries of misunderstandings by drawing on the available historic documentation for the Tas-Silg Phoenician-Punic temple precinct in regard to the area’s past names.159 In essence, he confirms that the site of Tas-Silg was known as Tal-Kasar and that within this locality was a field called ‘Ta’ Birikka’ (Bir Ricca), writing that ‘Tal-Kasar stands for “Kasar”, “Chasar”, “Cassar”, “Cashar”, “Casar” or “Qasar” in the consulted documents. The toponym is taken to refer to a tower-like building.’160 This location ‘was a well-known landmark in quite a large area’.161 Furthermore, Abela, he suggests, was actually referring to the Tas-Silg location rather than Borg in-Nadur for the Temple of Hercules. Can we take Bugeja’s observation one step further and suggest that Quintinus before Abela also meant the Tas-Silg site? In that case, unlike Abela, Quintinus got the object of (continued)



The Archaeology of Malta devotion (that is Juno/Astarte) at Tas-Silg correct. In relation to his general comments that the ruins lay half-way between the town and the castle, the latter (the ‘castle’) could be a reference to the area’s toponymic association with a tower.The ‘harbour’ may indicate Marsaxlokk Bay with its ancient mooring facilities at Il-Magħluq. Quintinus’s ‘town’ thus could be one of the settlements closer to the bay, perhaps Zejtun. Ruins ‘far out to sea’ could be a reference to Delimara pointing south from Tas-Silg into the sea. It is not unreasonable to suggest that a smaller shrine was located on or near the site of the church dedicated to Santa Marija tas-Silg.162 This would mean that when Quintinus introduced the area of the ‘the south-eastern harbour’ in his narrative, he did not shift his focus from this general location for both temple sites. Notwithstanding that there is now ample archaeological evidence which associates the site of Tas-Silg (Tal-Kasar) with the goddess, the exact whereabouts of the Temple of Hercules is still unknown, but neither the prehistoric site of Borg in-Nadur nor the ruins fringing the Grand Harbour are likely candidates.163 If we consider Ptolemy’s information that Juno’s place of worship faced east and Hercules’s faced south, then the general region marked by the extant ruins within Safi-Zurrieq is a good contender for the latter. As happened at other Punic temple sites, such as in Eryx (Sicily) and Tharros (Sardinia), extensive rebuilding and remodelling were carried out at Tas-Silg over the centuries of its use. Scarce traces of Phoenician structures at Tas-Silg are preserved among predominantly Punic ruins – some in situ, some reused blocks. Additional features have been dated to later periods. In the fourth century BC, the original temple area, itself enclosing prehistoric stones of great antiquity, and the large courtyard, which was the focus of worship in the open air, were surrounded by a pillared portico. Platforms extended around the complex, providing level ground for additional courtyard and altar areas. These works may have included outer defences, as well as a possible tower on the complex’s north-eastern corner.164 A double Egyptian gorge capital with flat-cushioned top (the abacus) may have surmounted one of two pillars which stood at the temple entrance, a feature of temple façades depicted in contemporary iconography.165 Other ornately carved blocks indicate that the temple was further embellished at various times in its Phoenician-Punic past.166 Parts of its infrastructure were extensive, such as the subterranean water galleries, branching out from wellheads at the site, which were explored by the Cistern Exploration Project in 2012.167 The current rural setting of Astarte’s temple may not resemble the cultural landscape which grew around the site in the Phoenician-Punic period. After the final year of Italian excavations at Tas-Silg (1970), Michelangelo Cagiano de

Melita and Gaulos during the Punic Period Azevedo gave an account of aerial photography of the area that visually confirmed the layout of the temple but importantly also indicated substantial traces of architecture in the fields beyond the archaeological compound: a possible pathway and likely necropolis. Current satellite images also hint at possible traces of regular structures.168 One villa at Bir Ricca, in a field called ‘Tal Ħereb’ (Zejtun), and other remains at Ħal Ginwi169 were close to the temple.170 When viewed as a whole, house and other remains from the Punic to Roman period lie north-west to south-east along the road Triq Xrobb L-Għagin or close to it, from Tas-Silg leading directly to the complex in the school grounds at Zejtun. It is quite likely that further remains along this route have so far escaped detection. Ritual at Tas-Silg is implied through the uniform range of pottery shapes left at the site. From the beginning of archaeological investigations at Tas-Silg, it became clear that devotees were bringing predominantly small bowls and plates in great numbers171; in later stages, cooking pots were offered as well. Large amounts of organic remains found in the temple’s southern dumps may be remnants of animal sacrifices and food offerings (domesticates and marine sources). Ritual dining and food rations for temple staff also may have resulted in such refuse. Numerous pottery fragments were inscribed prior to firing, indicating a pre-determined function for the vessels. Some were marked with brief dedications to Astarte or perhaps Tinnit. Recently, the name ‘Milk-Astarte’ has been added to the divine names associated with the site.172 Although domestic sites are limited, rarely does the pottery from such contexts or from funerary deposits carry inscriptions. Plates were purpose made to hold offerings. In this regard, Anthony Frendo and Dennis Mizzi point out that just under half the inscriptions on pottery at Tas-Silg (south) are the X-shaped sign (Neo-Punic ⊃aleph or mem) which appears alone, but the letter may embody an abbreviation of Punic words relating to temple practices (MTN, MTT or MTNT – ‘gift’; MNH￵T – ‘offering’; MSKT – ‘libation’; McSR – ‘tithe’). Similarly, the letters T and LT, comprising just over a tenth of the remaining inscriptions, also point to the same notion (TR(W)MH and LTR(W)MH denoting general offerings). Other markings might indicate the nature of the offering held in the pottery.173 These signs convey a sense of well-defined offerings, perhaps routinely brought to the sacred precinct. Regular bowl sizes for food items such as grain might have equated with a standard measure suitable for a temple offering; other bowls were quite small and could have held a single fruit or some other commodity. Though one tends to think of food and liquid offerings in temple contexts, items of intrinsic value might equally have been donated. North or south, very few such artefacts were recovered from the temple precinct, but there were coins and gold leaf, as well as an amulet depicting a female head with fugitive traces of gilt, among other finds. We do know from classical sources, however, that the Temple of Astarte was particularly rich.


The Archaeology of Malta


Some have pondered why inscriptions figure so markedly on the ceramic finds. It might be that the written word, particularly in temple contexts, was a powerful tool in religious expression, especially durable when written on stone or as markings cut into the surfaces of pottery.174 At a glance, even if the supplicant could not read, the vessel and its contents were recognisable as dedicated to a religious ritual.

The Temple of Melqart-Hercules Silius Italicus (first century AD) documents rituals connected to the Temple of Hercules in Gadir (Cadiz), including hair cutting, celibacy, bare feet and wearing loose white clothing. Perpetual fire was maintained on the temple altars, but the sacred halls were devoid of images and figurines of the deity.175 The cult of Hercules is considered to have originated in Phoenicia, where he was equated with Melqart, otherwise known as Baal.176 His cult is associated with the dying and rising mythology entrenched in Near Eastern belief systems concerned with renewal and resurrection.177 As we have seen, Mdina has some claim to a sacred precinct, but other alternatives have been suggested. Accounts by the early antiquarians and ­ ­historians of Malta are further complicated by the mistaken identification of megalithic prehistoric buildings, notably at Borg in-Nadur, as Phoenician-Punic structures which, in turn, were seen as the Temple of Hercules. Underlying the problem are the incorrect co-ordinates given by the classical geographer Ptolemy based on his grid scheme.178 The co-ordinates listed are Melita Island Melita city and peninsula179 Temple of Juno Temple of Hercules

38 45 34 40 38 40 34 45 39 –– 34 40 38 45 34 35

In a broad sense, these locations place Astarte-Juno’s temple in the east, in the southern reaches of the Grand Harbour, and Melqart-Hercules’ temple to the west of it, in an approximate north-south line with Mdina-Rabat, but erroneously at sea. Using Ptolemy’s location as a guide, others have looked to the far west of the island and the region of Għajn Tuffieħa Bay.180 As noted, Nicholas Vella considered that the headland of Ras ir-Raħeb, south of that bay, was the location of the Temple of Hercules. A few figurines found there may be little more than icons of household worship, and there are no other signs of the area serving as a sacred precinct; a stone connected to olive processing does suggest that the site was similar to other mill sites in Malta (see later).181 If Silius Italicus can be taken at face value, then the figurines cannot be indicators of Hercules’ temple. Certainly in the Near East and in the Mediterranean colonies, pillars or baetyls were typical symbols of the deity in the temples. Others have suggested that both Astarte and Melqart may have

Melita and Gaulos during the Punic Period been worshipped in the same precinct at Tas-Silg.182 Iconography from the site, probably from the Roman period, does reflect the notion of consorts of a later introduced cult. These take the form of Serapis and Isis, who have between them a third figure, rich in symbolism of the rising solar divinity. In addition, elements of a male solar divinity, possibly Mithra, are present in a graffito at the sacred site.183 Nonetheless, the classical authors convey the sense that distinct geographical locations for the temples existed in Melita.184 Coinage minted in the islands, which bore a male head, is thought to have represented Melqart-Hercules.185 On current evidence, the building at Zurrieq mentioned earlier has virtually no vestiges of a sacred function, yet this site, too, has been identified as a standing remnant of the temple. Several strands of evidence need to be considered. The gorge cornice that tops the walls is an architectural feature used on Punic religious structures in other, offshore locations.186 Not all the additional walls shown in Houel’s plan (Figure 7.2, nos. 2, 4 and 5) have survived, but the cornice and well-finished structure depicted in his drawing suggest that an open courtyard stood outside the extant room. A second doorway is shown with one block forming half the lintel hanging over the opening and with an edge on the surviving doorjamb. Twin pillars, dual peaks or a mountain cleaved in two feature in the mythology surrounding the god Melqart-Hercules. In the iconography of his temples,187 however, links between the two inscribed cippi or pillars and Melqart-Hercules, the Baal of Tyre, are not restricted to dedications in Phoenician and Greek on the plinths. Acanthus leaves such as those decorating the pillars also can symbolise Hercules, especially his association with death and rebirth. In essence, acanthus signified his immortality in contrast with the mortality of other heroic figures threaded through classical literature.188 Acanthus leaves, incidentally, are woven into the wreath at the top of the central feature (itself rich in symbolism of renewal and birth) in the Serapis and Isis relief, mentioned earlier, from the temple at Tas-Silg. The cippi may have featured as twin pillars in temple furniture, in line with dual monuments in Melqart-Hercules’ cult centres at, for instance, Tyre and Gadir, but their actual find spot in Malta is far from certain.189 Construction of a temple dedicated to the Tyrian god Melqart pre-empted or went hand in hand with establishment of Phoenician colonies at Carthage, Lixus and Gadir, among other sites.190 It is likely that Malta figured in the Tyrian colonising process, and the temples to this deity around the Mediterranean were highly regarded in the ancient world. The present parish church of St. Catherine of Alexandria, very close to the reputedly ancient Zurrieq building, was constructed between 1632 and 1659. This was also the period spanning the compilation of Abela’s antiquarian collection, in which the cippi first appeared.191 It is not unreasonable to suggest that such ancient artefacts came to light in Zurrieq during the seventeenth-century church-building process and that the Hercules temple was located there.192


The Archaeology of Malta


Gozo: The Evidence for Punic Sacred Precincts One damaged Punic inscription from the island of Gozo (CIS I, 132, also referred to as ‘Melitensia Quinta’), variously dated from the fourth to second centuries BC, concerns the dedication of three religious artefacts and four temples in Gozo.193 The mention of rabs (‘magistrates’) in the text indicates that Gaulos was governed by senators, drawn probably from an elite or noble class.194 The sites of the temples dedicated to Sadambaal (S￳dmb’l) – a deity not known outside this inscription195 – Astarte and another two gods mentioned in the text, whose names are lost, are not known. CIS I, 132: DECREE BY THE PEOPLE OF GAULOS

1. “The people of Gaulos constructed and renovated three . . . [and] 2. the sanctuary  – the temple of S￳dmb‘l and the sa[nctuary  – temple of (name of deity) and the] 3. sanctuary – temple of ‘Aštart and the sanctu[ary – temple of (name of deity)] 4. At the pe(riod) of the r(ab)ship of the esteemed senators ’Arish, son of Y’l [. . . son of . . . and] 5. Šapput, son of Zybqm, son of ‘Abdesmoun, son of Y’[ln . . . and the (kind of offering)] 6. offered Ba‘alšillek, son of Hanno, son of Abdešmoun [and the (name of function) was] 7. Bl’, son of Klm, son of Ya‘ezer. The inspector of the quarry (was) Y’[l . . . son of . . .] 8. The people of Gaulos.”196

Aside from the substantial religious sanctuaries in urban settings documented in classical sources, smaller sanctuaries may have dotted the rural landscape. A sacred precinct built at the tip of Ras-il-Wardija in the far west of Gozo may have been dedicated to a Punic deity. Only segments of the roughly cut foundation stones of an unroofed walled area stood within an equally eroded enclosure. Pits cut into a stone block formed an altar in the entrance of the boundary wall (the pits are the principal indication of cult practice).When the structure was first excavated by the Italian Archaeological Mission, traces of painted and plastered walls were evident, as well as poorly preserved domestic pottery of the third century BC to the second century AD. But all these artefacts may have arrived at the site through secondary deposition, as indicated by coinage of the recent historic era found in the mix.197 Sometime in the first to second centuries AD, the cliff face above the building was cut to form the inner sanctum of a Mithraic temple.198 That this

Melita and Gaulos during the Punic Period solar divinity was worshipped at the site might indicate a link to earlier sun worship, in much the same way that a very similar Mithraeum (a Mithraic place of worship) was built close to the Temple of Apollo in Cyrene, North Africa.199 Devotion to male solar and female lunar deities in Gozo is also indicated by the iconography on coinage minted on the island from the first century BC (see Chapter 8).200 A modified cave, thought to be a sacred site known as ‘Għar ix-Xiħ’, in Gozo, was the subject of archaeological investigations from 2005 to 2010.201 Figurines, pottery fragments, possibly coins and organic material have been linked to rituals carried out at the site from the late Punic to post-Punic period, perhaps by mariners using the harbour. Although the magnitude of the temples cannot be determined from the text (CIS I, 132), neither of the small sites of Ras il-Wardija and Għar ix-Xiħ in Gozo would seem to qualify. The scale of the temples mentioned in the text was such that an inspector of the quarry site, from which the building material was obtained, was required.202

Funerary Contexts Rock tombs are the most prevalent Punic artefacts in the Maltese landscape, found in small numbers each year during building and development of urban infrastructure, as well as in the open countryside.203 In the absence of extensive architectural evidence, they stand as testimony to nearby, as yet undetected Punic settlements. Large necropoles were investigated during the eighteenth to twentieth centuries, especially around Rabat (Figure  7.3, no.  13). As we have already seen in rural contexts, houses with pre-Roman origins can have small numbers of Punic tombs or pit burials nearby (e.g., Zejtun,Ta’ Kaccatura [Birzebbuga], Tad-Dawl Ħal Kirkop and San Pawl Milqi). It would seem that country residents did not venture far to bury their dead. As already discussed (see Chapter 5), the Maltese Phoenician-Punic tombs, which held no infant burials until the Roman period of domination, draw into question the real nature of infant burial practices and raise the spectre of child sacrifice which haunts Punic studies, especially concerning Carthage. The rock-cut tombs were designed to hold adult remains, and children were buried – it would seem they were interred rather than cremated in Malta – in designated grounds, referred to as ‘tophets’ in current literature. There is a gradual increase in cremation of adults from the floruit of the Melitan Punic period (established phase IV, ca. 300 BC),204 although inhumation remains the preponderant method of burial. Only during phase II is evidence for cremation wanting.205 Tombs often held multiple burials, placed over time; genetic studies have not yet been employed to determine any familial connections between individuals in the chambers. Adults are usually placed on their backs in tombs with arms folded over the abdomen


The Archaeology of Malta


or simply placed at the sides of the body and legs straight and together.206 If the tomb had a bier, the corpse was laid on that; if not, the body was placed directly on the floor. Bodies were, however, positioned in line with the long axis of the chamber, as one would expect, but the axis does not seem to be cut along any set compass bearing. Instead, clusters of tombs would be aligned according to available space.207 In later tombs, sometimes the head of the corpse was near the entrance, but more often the feet were nearest the opening.208 There is no evidence that the corpse was prepared other than perhaps being loosely clothed or covered with a shroud secured with bronze pins, as the skeletons show no signs of constraints. Other than the scale of the rock-cut tomb itself, overt signs of wealth or luxury goods are limited; very rarely, gold jewellery, some bronze rings, armillas, pins, earrings and beads have been documented.209 It is likely that certain graveyards were sought after, given the crowding of tombs, perhaps because they were close to urban areas or other features of cultural or historical significance. Larger burial grounds point to enduring settlements in the Maltese landscape, as well as a high degree of cultural continuity, which persisted into the Roman period. Tomb structures continued to evolve gradually with alterations to a basic plan comprising a shaft and mostly single or double chambers as cemetery space and presumably individual needs dictated.210 By the Melitan phase IV, shafts up to 4 m deep, 0.6 m wide and 1.80 m long were cut into the bedrock. Chambers measure around 2 m long, 1.5 m wide and 1 m high. Many parallels can be found in the Phoenician-Punic world.211 Lamp niches, larger recesses, trenches in the floor, small round pits, biers and rare occurrences of human or animal reliefs in phases III–IV had a role in the funerary rites at the time of burial. Above-ground monuments may have been present in the main necropoles in Rabat; evidence for them exists in blank spaces appearing in otherwise tightly packed burial grounds and rare survivals of worked stones thrown or fallen into tomb shafts.212 Shafts were originally capped by slabs lying flat or placed in gable fashion, but few remain intact. Baetyl grave markers are rare in Malta. One in the National Museum of Archaeology,Valletta, depicting a person in a niche is stylistically very similar to finds in Carthage and might have been imported into the island: whether in antiquity or in recent times is not known.213 Another was reused in a wall at Tas-Silg.214 SYMBOLISM IN TOMB ARCHITECTURE AND CONTENTS

Symmetrical, Paired and Tripled Iconography Management of space would appear to have been as calculated in Phoenician-Punic funerary contexts as it was in their house designs. In a sense,

Melita and Gaulos during the Punic Period their beliefs were played out in the three-dimensional setting of the tomb itself; tomb architecture, the burial and pottery placements in the tombs were all vehicles for expression of ritual practice. While the tenets themselves of the belief systems are poorly understood, the physical evidence – persistent, repeated and present in varying degrees of elaboration – once figured in the rite of passage of the deceased.215 Based on the Maltese evidence, tombs clearly reflect a progressive march towards an array of built-in features by Melitan phases III–IV.216 Long depressions or ‘trenches’ were cut into the floors of the chambers, though with different orientations through the centuries; this feature likely held symbolic value rather than functioning as a drainage pit or sump.217 At first, the depressions were cut parallel to the inside of the chamber entrance; later, they were cut down the centre following the long axis of the chamber. Variations include two or three trenches or a single bier with one trench against a long wall (see Figure 6.5, nos. 3–7, cf. 10 and 11). Coffins in terracotta or stone are rare.218 During the floruit of the Punic period, tomb plans and fixtures exhibited symmetry, dualism or triads, enhanced by the central-axis trench (see Figures 6.5, no. 11, and 7.9, nos. 3 and 4).219 Symmetry often was maintained through continued use; one or more interments occupied each bier, and between the bodies a cinerary urn could be placed at the rear of the chamber. Significantly, in some cases the gradual placement of three urns displaying evolving jar shapes indicates the enduring nature of ritual practice over the lifetime of an individual or across generations within a family.220 On a few occasions the urns or other vessels were cemented to the floor, further suggesting that placement was significant.221 An example of three-figure symmetry is evident in a tomb from Qrendi, where a head is carved into the rear wall of the chamber (Figure 7.9, no. 3), and either side of the entrance are two small figures cut into the entrance wall of the shaft (Figure 7.9, no. 4).222 Although the carving of the head is indistinct, a hand does appear to be at the left side of the face. Hair and an earring have been reasonably well defined on the right side. The face might represent a deity, and the two smaller figures acolytes. Punic coinage in Malta also can depict an Egyptianised central deity with two attending female figures.223 In addition, triple images, or triads, are a common motif carved into stelae from burial grounds in the western Mediterranean, and multiple pillars or baetyls can figure in altars at sacred sites. Altar 45 at Tas-Silg, for instance, may have consisted of three baetyls (see Figure 6.9, nos. 3 and 4) fixed into sockets and secured with lead poured into the gaps; its base slab was worn smooth and reddened from heat damage, probably sustained through repeated ritual activity.224 As Miranda Green argued for religious imagery elsewhere, so the Punic triads might ‘reflect the totality of the family, as with some Egyptian or classical triads where mother, father and child represent the whole and continuance of human existence. Eternity is reflected here, and the endless cycle of


The Archaeology of Malta


regeneration.’225 In the case of triads, they are likely to symbolise a divine trinity, common to many Phoenician-Punic cities, but in varying combinations of favoured deities, such as Baal, Ashera and Astarte, who were bound into a cycle of resurrection and renewal. Or as Joseph Shaw has argued, a divine pluralism may be represented, where devotions were made to a deity in various guises; for example, there was syncretism between female divinities Astarte, Tanit and Ashera.226 It is noteworthy that burnt offerings may have been part of the ritual of construction, for remains of ash and bones were found crushed in the sockets, presumably impacted by placement of the upright stones.227

False Doorways and Empty Niches Doors and entrances were iconographically important. Within the chambers of some Melitan phase III tombs, large niches which were left empty may have represented false doorways, over which there was sometimes a smaller lamp niche. Punic stelae, especially from Carthage, can depict multiple rebate edges on empty niches which create the sense of a false doorway with recession and depth.228 A parallel could be drawn to the notion of an absent deity manifest in the vacant niche or the empty throne that, alone or with a removable divine symbol, became the focus of religious devotion in the Levant.229 Niches in tombs conveyed the concept of an entrance or gateway between an earthly life and an afterlife perhaps connected with the beth-el (dwelling place of the god). In only one instance was a coin placed in the mouth of the deceased, possibly signifying the spiritual journey and an anticipated monetary cost; on that occasion, the assimilation of Greek funerary practice or the burial of a Greek individual is likely.230

Elixirs within the Tomb Setting Small pits cut into the bier or floor of the chamber might have contained water or some other liquid or were intended to receive the same in a symbolic sense. The significance of fluid is suggested by cups found in pits cut into tomb floors in Attard and Qallilija,231 as well as cups placed near corpses.Tombs were nearly always provided with some vessels232; usually one or two jugs (or an equivalent pouring vessel such as an askos233) and a lamp were present. This range evolved to include amphorae, plates or bowls, small flasks and unguentaria and then contracted again at other times.234 In burials of Melitan phases I–II, at the height of the colonisation period, both disc-topped (or ‘mushroom-topped’) and trefoil-mouthed jugs were included among the grave goods. It is possible that the styles were gender specific, but without unequivocal proof of the sex of the deceased for most of the burials, no conclusions can be reached.235 The practice of placing trefoil- and disc-topped jugs in burials was also documented at Tyre al-Bass from the second quarter of the eighth century

Melita and Gaulos during the Punic Period BC. Chemical analyses on the disc-lipped jug from Tyre hinted that it had once contained honeyed water (hydromel). Based on the efficient pouring shape alone, it was suggested that wine was held in the trefoil-mouthed jug.236 Certainly wine drinking, a prestigious activity at the time, most likely used trefoil-rimmed jugs. In later Punic contexts in Malta, two trefoil-mouthed jugs were common placements after the disc-topped forms had petered out.

Invocation of a Divine Presence in Funerary Contexts By virtue of its range of features, one tomb from Rabat (Figure 7.9, nos. 1 and 2), in particular, offers a unique key to decoding the rock-cut features in less-endowed tomb chambers.237 A painted and carved figure in a simple priestly robe looms out of a recess in the rear wall.238 His right arm hangs at his side, clutching an infant or perhaps a vessel (the carving is indistinct), and his other arm holds a curved sceptre across his chest. Sceptres and jugs appear in representations of the deity, sometimes in his own divine hands and sometimes in the hands of his acolytes.239 The figure dominates the tomb. His face is painted in a mask of red ochre, and his head is topped by an Egyptian-style wig, painted black and formed by the shape of the niche itself.240 Red ochre and a disc pendant are on his chest. In front of him, three trenches and a pit were cut into the floor. Whitewashed walls were painted with a thin red band 13 mm wide just below the ceiling.241 This single line was connected to the central ceiling decoration of a rayed disc, in-filled with twelve alternating red and white petal shapes (stylised in Figure 7.9, no. 2), which, thus combined, could be viewed as the firmament through which the sun travels.242 Robe, pendant, sceptre, mask and ceiling decoration connect the figure in the niche to the sun deity. The iconography evokes many elements of the Phoenician solar divinity in the wider Levantine and Mediterranean depictions:  solar discs at the top of many funerary stelae in the West, on the foreheads of masks, on gems and on seals as well as figures present or absent in niches.243 Rather than the figure being apotropaic, in the sense of scaring away intruders into the tombs, or an image of a devil, or a depiction of grimacing death in human form, this particular Rabat representation strongly suggests that the ancients were depicting a divine entity visiting the tomb.244 He enters through a false door into the chamber, where presumably some form of interaction with the deceased takes place aimed at carrying the essence of that person to an afterlife. In the archival photograph of this tomb, traces of decoration appear on either side of the figure’s head. Unfortunately, the paintwork was poorly preserved and may not have been readily visible to the naked eye, but the figure would appear to have been set against a more elaborate backdrop. The curved lines of paint on either side of his head may be rays or streams of life-giving waters, features sometimes shown in connection with Baal-Melqart of Tyre.


The Archaeology of Malta


Strong connections between solar divinities and water were a persistent theme in concepts of renewal in ancient Near Eastern religion, a theme that may have been revived much later in Malta through the cult of Mithra.245 The colour red might have had symbolic value in Phoenician-Punic funerary contexts of the central Mediterranean. A  thick pomade of red cinnabar or clay and red ochre, for instance, was painted on the face of a corpse at Smirat, situated south of Carthage on the North African coast. The residue stained the skeleton after decomposition of the flesh.246 Tombs at, for example, Djebel Mlezza and Tuvixeddu, Cagliari, were painted with red motifs.247 That some highly symbolic value ought to be attributed to the Rabat tomb designs is obvious, but the choice of red might have carried its own meaning. Red colorant in ancient societies has been widely acknowledged especially for its potent association with blood. In the case of the North African sites, the use of red ochre on the corpse is considered an indigenous Libyan-Phoenician practice.248 No mention of such rubrica has been made in the literature concerning Maltese Punic tombs.249 We should recall that red ochre does not occur in Malta and must be imported. The mask on the Rabat figure, as well as other instances of carved faces in the Maltese tombs,250 may have greater implications for Punic religious beliefs. Derived from the Levant, masks and the ritual function they served spread to some western Mediterranean Phoenician sites.251 Fictile examples from the Dermech necropolis are some of the oldest masks in Carthage, dated to the end of the eighth to the early seventh centuries BC.252 Though masks occur in Ibiza, Sicily and Sardinia, they are rare in Malta.253 Many are often smaller than life size, depicting somewhat realistic through to gnarled and grotesque features; both male and female forms occur. Rather than being designed for humans, the Rabat figure in the niche conveys the impression that masks may have been fashioned for the deity in response to the perceived need for mortal men and women to be protected from the ‘radiant’ face of a god, particularly the solar divinity, so overwhelming that he was masked.254 The sense of covering the divine radiance before mortals in Near Eastern belief systems can be found in the biblical account of Moses after being in the divine presence on Mt. Sinai: Moses did not know that the skin of his face shone because he had been talking with God. And when Aaron and all the people of Israel saw Moses, behold, the skin of his face shone, and they were afraid to come near him. . . . And when Moses had finished speaking with them, he put a veil on his face; but whenever Moses went in before the Lord to speak with him, he took the veil off, until he came out; and when he came out, and told the people of Israel what he was commanded, the people of Israel saw the face of Moses, that the skin of Moses’ face shone; and Moses would put the veil upon his face again, until he went into speak with him [Exodus


Melita and Gaulos during the Punic Period A male Phoenician deity appears in other contexts, notably on gems and seals, with an animal mask, or bull-headed, or wearing a horned helmet. This iconography is reflected in one Rabat tomb where a bull’s head was carved into the niche in the rear wall.255 The plundered state of this chamber again deprives us of the complete record, but on the basis of the central trench dividing the floor into biers, the tomb is likely to have been cut after Melitan phases III–IV. In Malta, the tripartite Phoenician Tanit symbol (Figure  7.1, no.  12)  was constructed in a three-dimensional way using the pottery of adult cremation burials. Maltese cremations frequently have the repeated pottery combination of cinerary urn, cover plate256 and twin-wicked lamp (Figure 7.1, no. 14). It is plausible that the combination reflects the tripartite Tanit sign  – triangle (urn), horizontal crossbar (plate) and orb (lamp) (cf. Figure 7.1, no. 13). The Tanit symbol is likely to have had a broad meaning, embodying the fundamental elements of divinity generally257 and, in terms of Maltese funerary contexts, probably invoking the notion of resurrection or renewal through divine intervention. Twin-nozzle lamps played a special symbolic role, probably in connection  – through heat and light  – with the solar deity, evident in this three-part pottery combination. Significantly, the twin-nozzle lamp persists in Malta, even after single-nozzle forms (Figure 7.1, no. 23) were gaining popularity, including the special treatment given to some lamp niches in tombs (Figure 7.1, nos. 15–21). Lamp niches in Malta are mostly simple small cavities cut in the wall, but more elaborate features can be present which emphasised the lamp; for example, in one Mosta tomb a niche (and presumably the twin wicks of the lit lamp) form the eyes and forehead of a stylised face with chiselled projecting nose and slit mouth. Elsewhere, small stone stands are cut in niches to hold the lamp (Rabat, Triq il-Buskett), and pillars are cut in the rear chamber wall which accentuated the lamp niche on top (Ħal Far).258 Such symbolic themes are not restricted to Malta. The significance of the twin-nozzle lamp is indicated in the wider Mediterranean through, for instance, statuettes carrying lamps on their heads found at Isla Plana, Ibiza.259 A box with four figurines placed on a burial urn (urn 8) in a Tyre al-Bass grave contained some familiar emblems: a mask, a horse with rider carrying a shield and wearing a horned helmet (a signifier of divinity), a deity before a bowl (probably a reference to life-giving water) and a so-called temple façade figurine with a ramp. In regards to the last, rather than a temple façade, it may be a depiction of the portal to the underworld. An orb on the lintel and a baetyl on the ramp may symbolise one and the same thing – the presence of the solar deity and his daily ascent from the underworld to the sky.260 The al-Bass tomb may have been a simple pit burial, but the iconography of the objects left there conveys common themes. Artefacts in burials in Malta can reflect a sense of tokenism: objects brought to the tomb which represent the requisite funerary repertoire but are of poor quality, such as pottery seconds that were holed by shrinkage cracks, squashed


The Archaeology of Malta


into irregular shapes and sometimes carrying ancient mends or under-fired pots.261 Ritual is also hinted at through other finds. At Tharros (Sardinia), a small number of child cremations placed in urns contained a single seashell or a piece of coral, pottery sherds and a sea-worn pebble.These objects were usually placed near the bottom of the urns and were thought to have been part of a deliberate ritual act rather than being accidental inclusions.262 Blue nodules and pebbles also have been documented in the rock-cut tombs in Malta.263 From the accumulated evidence, it seems that Punic Melitans were placed in the tomb with necessary utensils around them – sometimes with clear, deliberate positioning – which would play a role in a rite of passage to a spirit life. In funerary contexts in Malta, the iconography of the tombs points to a firm belief in a solar deity, whose presence in the tomb was anticipated. It was his attribute as a solar god, a god of renewal, that was evoked; Šamaš, Melqart-Hercules, Baal of Tyre, Baal-Shamem or Baal-Hammon, Lord of the Brazier in Carthage, seems likely.264 Much of the elaborate iconography and many of the symbolic pottery combinations and tomb fixtures emerge during Melitan phases III–IV, the floruit of the Punic period around the late fifth to fourth centuries BC and after.This complexity was matched in the wider Punic ambit. In situations where burials were placed in the earth, such as in Carthage or Tyre al-Bass, stelae and figurines convey the iconographic signifiers of belief systems. In Malta, it would seem, where the rock-cut tombs could be cut with permanent symbolic features, that such markers were not necessary. SUMMARY

As Rome gathered momentum, expanding its interests in the Mediterranean, it ran headlong into conflict with Carthage, the greatest of the Punic cities in the West. By this time, the city had established strong commercial ties to markets of the Mediterranean rim. Over the course of three Punic wars, Rome battered the commercial and political powerbase until the city finally fell. Richard Miles encapsulates the demise of Carthage: Great Carthage drove three wars. After the first one it was still powerful. After the second one it was still inhabitable. After the third one it was no longer possible to find her.265

Once seized, Punic infrastructure in the Mediterranean actually contributed to the success of Rome. Rome subsumed the long-established Punic economic network, vast, predictable and profitable. The architectural remains tell a story of cultural endurance in Malta despite the dominance of Rome and the fall of Carthage in 146 BC. In their burial grounds, the people of Malta demonstrated the persistence of their distinct Punic cultural heritage and a remarkably tenacious belief system which lasted well into the first millennium AD. Society in the islands was separate enough

Melita and Gaulos during the Punic Period from mainland Europe for the daily way of life to continue unabated and strong enough to rebuild when attacked, leaving the clearest evidence of cultural hybridity and social evolution under the umbrella of Roman influence – a genuine Punic survival. Ethnically, the people of Malta retained their resilient cultural identity, as evident in Punic inscriptions found in sites which continued into Roman times. Baal was mentioned, for instance, in an inscription cut into the wall of St. Paul’s catacomb near an entrance from the Vicolo Catacomb (Ħal-Pilatu).266 There may be historic reasons why the Punic presence in Malta has been so consistently pushed into the background, but because it has been, the subtle changes in life under Roman domination in the relatively closed environment of an island community have remained obscure. Though Rome would finally end Punic economic and political dominance in the region, recent genetic studies of living populations have demonstrated a lasting human legacy of Phoenician-Punic colonisation in the lands of the Mediterranean, including the Maltese Archipelago.267 It was the growing popularity of new religions, especially Christianity, which put an end to old Punic beliefs in Malta. To what extent the Carthaginian conflicts with Rome affected life on the islands is difficult to assess, but this issue will be explored in Chapter 8.



How ‘Roman’ was the Roman world? [Kühr 2011]. What we find in the Roman Empire is a situation of greater complexity, in which individual agents have far greater choice over how they construct or present their identity within the context of Roman colonialism [Laurence 1998, p. 2]. HOW ‘ROMAN’ WAS ROMAN MALTA?

In 218 BC, at the outbreak of the Second Punic War, Roman Consul Tiberius Sempronius Longus sailed with his fleet to Melita from Sicily, prompting the Carthaginian commander in Melita, Hamilcar, to surrender with some 2,000 troops. The island and its urban centre followed suit. In many respects, it was impossible in the sudden surrender of Hamilcar and in the absence of Carthaginian military support that the Melitans could have mounted any realistic opposition to Rome. Next to no archaeological evidence survives for the documented presence of Carthaginian troops.1 Unlike Sardinia, there is no landscape, mountainous or otherwise, from which to stage resistance.2 Malta, thereafter, was administered under the umbrella of provincial Sicilian authority. Surrender was perhaps the better option for the inhabitants, having already suffered at Roman hands during the First Punic War,3 and annexation may have saved the islands from extensive destruction. Though the Romans may not have needed the islands for strategic purposes of their own, Rome’s severed ties with Carthage certainly meant that Melita and Gaulos could no longer be allowed to serve such a purpose in the hands of Rome’s enemies.4 Indeed, the islands’ merchants may have anticipated the possibility of surrender and seen a real opportunity for economic gain. As suppliers of commodities such as olive oil, as a cog in the long-established Punic trade networks of the Mediterranean, but most importantly with significant ports on the route between the Italian Peninsula and North Africa, it may have been in Rome’s


Malta’s Place in the Roman World interests to maintain the infrastructure of these central Mediterranean islands, especially in the climate of continuing war with Carthage. The nub of understanding how the Maltese Archipelago fared at the local level during and after the tumultuous years of the Roman-Carthaginian wars concerns ethnicity, identity and choices. Cultural change was unavoidable as the islands experienced first-hand what could be termed the ideological manifesto of Roman standards by which the territories were governed: recognition of central authority in Rome, development of the urban administrative centre and religious ritual performed at a state level.5 Rome established and maintained power through the interplay between urbanism and imperial control, administered through the governor as Rome’s representative. The inhabitants experienced public performance of rituals within a visual urban – often monumental  – framework, managed according to Roman ideals. Yet, within this setting, agency was in the hands of individuals, who effectively weighed up their options, drawing on their own understanding of the current socio-political situation, and modified their daily lives accordingly.6 If we consider aspects such as heritage, ethnicity, gender, status, wealth and life experiences, perceptions of self and any sense of ‘Roman-ness’ under these circumstances was likely to be complex in the provinces, Melita and Gaulos included.7 In comparison with earlier periods, our understanding of historical events during the late Punic to Roman period is improved by an increase in textual sources (classical works and public inscriptions8) and significant archaeological remains available to scholars. These were formative years for the islands, having slipped from the Carthaginian orbit to the Roman sphere. Themes of ‘victor and vanquished’ tend to dominate how the islands are portrayed in recent historical accounts after they were subsumed into the Roman world.9 Archaeological evidence, however, points to a multifaceted community in Melita, with clear indications of demographic continuity. It is a pattern shared by other Punic territories in the region.10 Again and again in this study, we return to the notion that the way the islands were viewed in antiquity was far removed from modern perceptions. Despite their central Mediterranean location, recent accounts tend to conform to a centre-periphery model, with the islands on the periphery, and the inherent inequality between the parties which this embodies, whether overt or implied in the literature.Yet the islands had been significant enough to warrant a permanent Carthaginian military presence. Considering that naval battles figure strongly in the history of warfare in the region, it may not be accurate to class the islands, with their superior harbours, as a ‘rather remote, almost insignificant corner’11 of the Roman Empire, even if they did not figure large in the histories of the day. Indeed, geographically, the islands were only a stone’s throw from Sicily and were well placed on sea routes to North Africa.



The Archaeology of Malta Melita was recognised in antiquity as a place of veritable sanctuary, removed from the hothouse of political life in Rome.12 Nonetheless, in Malta, Roman authority encountered a developed society, steeped in cultural practices familiar to the lands bordering the Mediterranean rim and already acculturated to standards shared by or acceptable to Rome. By contrast, Greg Woolf encapsulates conditions in Gaul (France): It is evident that, as a result of the extension of Roman power over Gaul, local societies there were drawn into a much more complex imperial world. Romans were more differentiated from one another in terms of wealth, occupation, experience and status than were Gauls who mostly lived in locally circumscribed societies ruled by warrior and religious élites and were mostly full-time agriculturalists.13

Possessing urban centres, acclaimed religious precincts and a stratified society, acceptance of ‘the structuring principles, the cultural logic of empire’14 may not have needed great adjustments to the way of life in Melita and Gaulos. Centuries of scholarly work have charted the emergence of Rome and its conquest of the ancient world, supplemented in the modern era with data gleaned from archaeology and the gamut of scientific know-how. Steeped in a Punic way of life, the people of the Maltese Archipelago appear to have retained a strong local identity which, as we have seen, traced much of its ethnic heritage to the Phoenician Levant. Notwithstanding, those living in rural and urban settings may have coped with change in different ways. One aspect of their daily existence is particularly enduring and indicative – coinage. Money that lay at the heart of commerce, found in every consumer’s purse, was decorated with emblems of the land, religious or official iconography. Though coinage serves a basic economic function, it also can carry significant expressions of identity at official and public levels. In Rome itself, coins carried portraits of those in power and emblems of state, but that was not necessarily so in Rome’s provinces.15 The right to mint coins under Roman dominion indicates that a degree of autonomy was granted to both Melita and Gaulos; the archipelago seems to have been rewarded for its shift in allegiance.16 The task of minting coins was delegated to resident provincial authorities, and it was their understanding of the political overlay emanating from Rome but also the continuing strong, local identity which shaped their choice of how coins would appear.17 Decisions which authorities made in Melita and Gaulos were driven by active recognition of home-based interest groups. Thus reflected to a degree was a managed identity, one that was acceptable to Rome while at the same time signalling the complexity of an insular yet manifold social composition and the influences filtering into the island from Hellenised neighbours, especially Sicily. As already noted, contemporary inscriptions such as the twin cippi dated to the second century BC carried dedications in Punic and Greek.18 Aside

Malta’s Place in the Roman World from money minted elsewhere and carried to the islands, coinage of Melita and Gaulos for the most part did not bear the portraits of Roman or provincial rulers or their family members; instead, it carried deep-seated regional themes on both the obverse and reverse of the coins.19 Symbolic representations clearly entrenched in local (Punic) heritage, religious iconography and Punic and Greek legends were stamped on the currency. Coins, in a sense, point the way to understanding life on the islands at this time. Any discourse must recognise the islands’ pre-existing Punic population and how it adapted under a new regime. Archaeologically, we have ample evidence of how Punic people living on the islands accepted Roman political and cultural norms. Aspects such as the adoption of Roman bathing traditions, expansion and renovation of rural complexes, development of some religious centres, urban construction on a large scale, growth of major port facilities and the advent of Roman-styled merchandise in local markets will be traced in this chapter. POLITICS AND GOVERNMENT

Significant political events in Rome, taking place over several centuries of its dominion, were marked in the archipelago by fundamental changes in how the islands were administered. But no more than broad brushstrokes can be applied to the historical narrative of the times. Melita and Gaulos were gathered first into a republic (up to 27 BC), where, like other Roman provinces, they could expect – and did get – a degree of fair administration along with legal recompense when it was flouted. Such an occasion in Melita was documented by Cicero, who acted as prosecutor in the legal case against the unjust rule of the Sicilian governor Caius Verres (73–71 BC). Verres plundered the Temples of Juno (Astarte) at Tas-Silg, effectively turned Melita town into a factory for fine textiles for his own gain and pursued Diodorus, a wealthy man born in Malta, in order to take possession of his prized silver cups.20 The image imperial Rome constructed for itself was one of benevolent administration with legislation which held its governors accountable for any abuses of their power.21 The governor’s role was ideally to uphold the ‘image of Rome’ and provide ‘just rule’; abuse of this position of absolute authority at the local level would be seen as un-Roman behaviour. Caius Verres had acted blatantly against of these tenets. While Roman ideals were promoted, local cultural traditions were tolerated, although political control accountable to Rome was paramount.22 Local government of both islands was aligned to republican principles and administered through a senate and a popular assembly. This structure was removed, however, under Augustus, with the emergence of the empire (spanning 27 BC–AD 476).23 During this time, it would seem that the emperor directly administered the islands through a procurator.24 Inscriptions suggest that local municipal


The Archaeology of Malta


authorities were again in place by the second century AD in both islands.25 Citizenship was granted to all free men of the empire in AD 212 (presumably in Melita and Gaulos as well) by Caracella, who intended to increase substantially the flow of taxes to Rome, payable only by citizens of the empire.26 Yet, despite these political machinations, it is possible to detect a lingering sense of Punic heritage into the second century AD in the archaeological record of the islands, but not much beyond that time. Certainly the introduction of new belief systems – perhaps Mithraism (early centuries AD) and then Christianity (fourth century AD) – put an end to Punic religious and funerary practices. LAND-USE AND SETTLEMENT PATTERNS In spite of the Roman conquest, Malta remained culturally a Punic centre . . . and it may never have been completely Romanised [Buhagiar 2007, p. 11].

Inscriptions found in Melita and Gaulos in the Roman period testify to the building programmes conducted by governing officials.27 Development of the urban centre featured strongly in the process of Roman imperialism. Already discussed in Chapter  7 was a deep ditch cut across Rabat, possibly sometime in the first centuries BC–AD (or later), that is likely to reflect such an agenda. Nor were religious institutions neglected. Roman-period inscriptions clearly testify to the refurbishment of sacred buildings in Melita and Gaulos. Massive building works can be charted in Melita (town), and a number of rural complexes underwent extensive rebuilding and remodelling. Day-to-day life saw the increased introduction of pottery styles and other artefacts from the Roman world and its colonies. In terms of commerce, the islands show signs of prosperity. Both urban and rural settlements appear to have thrived under the new regime. Undoubtedly the islands were home to a socially and economically stratified society. Hallmarks of Roman culture were adopted by the elite  – the urban wealthy and rural estate owners  – a strategy which did prove advantageous in Malta and reflected in the considerable development in town and country. By the late second century BC, during the republican administration, urban growth took place close to Mdina, the religious and perhaps administrative centre of the town from the sixth century BC. Around this time (ca. 125–80 BC), the substantial house (now the Domus Romana Museum) was built immediately to the south-west of Mdina.28 Drawing on Hellenistic architectural and decorative traditions, it featured a peristyle edged by sixteen stuccoed and painted Doric columns topped by an entablature with a repeating triglyph and metope design.29 Stylistically, the stuccoed and painted walls and fine mosaics have been assigned to the late second to first centuries BC, and its contents indicate that the house stayed in use into the fourth century. Mosaics covered the floors of its principal rooms and were traced

Malta’s Place in the Roman World out with small tesserae fragments or larger lozenge-shaped tiles displaying the careful manipulation of colours and geometrical design to create clever optical illusions of three-dimensional depth. Smaller detailed studies, the emblemata, incorporated into the floor designs were crafted in refined opus vermiculatum work (linear trails of tiny coloured fragments following the contours of the naturalistic designs), the panels of which may have been prepared in the workshops of Italy or as far afield as the Near East and imported to the islands.30 Ample proportions of the house, its mosaics and the objects found there reflect the opulence of the day, and these characteristics are mirrored in similar homes from Pompeii to Pergamum. Fixed in situ remains have been preserved on the site (now the Domus Romana Museum), and some finds from the house are displayed there (Figure 8.1, no. 1).31 The pinnacle of authority maintained a visual presence through sculptural likenesses of Roman rulers. Such iconography of the imperial family in Malta, as in many provinces, acted as symbols of Rome’s central authority.32 Several fine marble examples have been found in the grounds of the house, depicting the Emperor Claudius, possibly his daughter, Claudia Antonia, and a youthful Nero, among other figures, which are now damaged beyond recognition.33 Certainly the occupants of the building were from the elite of Melitan society, and it has long been speculated that the house was home to the governor of the island.34 Very little has been documented concerning the specific locations of finds from within the house. Several eroded coins fall within the third and fourth centuries AD.35 Of the ceramics found in the 1881 excavations, which were restricted to the front main rooms of the building, Caruana notes, ‘No remains of Table or Cooking utensils, or implements of worship, were found, but only a large quantity of fragments of earthenware vessels &c.’36 Despite the impressive proportions of the house, its unexcavated eastern sections were removed without documentation in 1889, when a road was cut into the bedrock; this cutting exposed sections through stairs and other rock-cut features. In subsequent years, 1920–5, T. Zammit reported that faience beads, bronze fishhooks, bone and ivory bodkins, spoons and hinging mechanisms, as well as animal bones (presumably the refuse of meals), were recovered through renewed archaeological investigations at the site.37 Zammit indicated that cartloads of pottery – said to have been ‘numerous all over the place’38 – were removed which ranged from very large storage jars and rough domestic wares to fine vessels including assorted black wares, some figured Hellenistic and Campanian black glossy (or glazed) slipped wares of the first century BC, Samian and pseudo-Samian, as well as decorated and plain Arretine. The Samian potters’ marks fell within the first century AD. Overall, Zammit noted, ‘Most of the red ware, however, was an imitation of Sigillata of the 4th century AD.’39 Glass vessel fragments, metal artefacts, figurines, masks, a strigil and lamps are also mentioned among the finds.



The Archaeology of Malta

Figure  8.1. (1) Plan of the Domus Romana in Rabat. Shaded areas indicate the subterranean water installations and cisterns (after Depasquale and Cardona 2005; Gouder 1983). (2)  Marsa warehouse and cisterns (after Barbaro 1794, pl. 1); the villa of G. F. Abela is indicated on the left.

Malta’s Place in the Roman World Of particular interest are clear signs of industry conducted in the grounds of the grand house or neighbouring dwellings, such as bone and ivory sewing needles and clay weights used in the manufacture of textiles.40 A number of damaged clay moulds were found in the 1922 excavations which were possibly for the manufacture of small discs or plano-convex ingots or used as crucibles; they showed signs of considerable heat damage and discolouration from the molten metal.41 Several small stone (marble?) balls may have been weights ranging in size over 130, 160, 180, 185 and 190 mm in diameter; another was oblong with a cross incised on one face.42 Together these finds suggest that at least part of the function of the site may have been economic. Of the numerous occasional finds in the Rabat-Mdina district, few substantial reports have appeared, but collectively they do point to the extent of the urban area built over the plateau and down into the valley floor during the Roman period of influence. The grand house was not isolated. When chronologically well-defined artefacts are wanting, the dimensions of regular architectural blocks are used as indicators of the Roman period.43 In 1909, Zammit established that structures were built over bedrock and that water cisterns were cut into the rock some 6 m under the glacis outside the citadel and the ditch to the north-east of the Domus Romana, close to the Greeks Gate.44 Lesser structures and narrow streets were uncovered immediately to the north-west of the house (built in the third and fourth centuries AD); these were constructed from stone harvested from older sites and finished with a mortar coating.45 Careful management of water is as apparent in the republic period as it is throughout the human occupation of the islands. Extensive channels, a few deep wells (about 12.2 m deep) and cisterns (about 4.5 m deep) were made in the grounds of the large house and underneath it, as well as elsewhere in this sector of Rabat (shaded areas in Figure 8.1, no. 1).46 On the northern side of the 1889 road cutting was a complex of structures found in 1920: a level platform of stone blocks, an L-shaped water cistern and a partially constructed room, the roof of which was supported by pillars on one side and fitted into the natural rock face on the other which spanned a natural fissure with spring water at its base. Rock-cut channels, water galleries and collection facilities were constructed in the immediate area, but the spring also continued into the valley to the Għajn Ħammiem fountain, where traces of a Roman bathhouse were identified.47 In front (south-east) of the Domus Romana on the Museum Esplanade, in what is currently a car park in the open square, Misrah Anton Agius, but was formerly a field known as ‘l-Għalqa ta’ Kola’, extensive architectural remains have been documented since 1957.48 Excavations were undertaken in 1983–4 by a team from the UCLA Institute of Archaeology in the same area. Finds including heavy masonry, a bottle-shaped cistern, a possible street flanked by houses, coinage spanning from Constantine (AD 309–37) to one local coin of the second to first century BC and pottery from the turn of



The Archaeology of Malta the centuries BC–AD are currently being processed for publication by the Domus Romana Melitensis Archaeological Research Project.49 Further, tiled flooring and architecture were uncovered during municipal re-surfacing works in 2009.50 Poorly preserved walls comprising square stone blocks came to light in a field called ‘Tal-Franciz’ in 1957, a location which is thought to have equated with the Santa Margerita Cemetery compound flanking the southern side of the road, Il Wesgha tal-Muzew, and opposite the Domus Romana Museum. Rock-cut installations which were remnants of olive presses were revealed there when the entire soil covering was removed from the field, exposing the bedrock.51 Additional large stone blocks were uncovered during building works to the rear of and underlying the cemetery, where the land slopes down to the valley.These finds attracted media attention in 2004, but little has otherwise been published; nevertheless, this substantial structure was assigned to the second and third centuries AD.52 More remains were found on Triq tal-Muzew in 1931 which included architectural and small finds, coins, glass and Samian as well as other pottery of the Roman period.53 Further sites lie south between the Domus Romana Musuem and Triq Santa Rita, which runs the course of the ancient defensive ditch of Melita. Caruana relates anecdotal accounts from 1867 of houses and mosaic floors in Triq Bir l-Iljun and in other unspecified locations in Rabat, known to exist between the 1830s and 1860s.54 Notice of archaeological finds between 2003 and 2009 in Doni and Old Doni streets – a well, a coin hoard that spanned the period AD 337–6055 and flooring – concerned remains dating from Hellenistic through Roman to post-Roman periods.56 Work carried out by Mevrick Spiteri and Edward Calleja in Triq San Pawl has entered the scholarly arena with detailed descriptions and copious illustrations. Despite the narrow confines of the three exposures, ‘ashlar wall alignments, floors, water channels and cisterns’ were uncovered within complex stratified deposits.57 Building works on Saqqajja Square (later designated Saqqajja I, south of Mdina, investigated in 1965–6 and 1969)  uncovered substantial remains of a Roman-period house oriented north-east to south-west with elaborate flooring supported by up-turned peg-based Roman amphorae, a technique already documented in Malta (see Figure 7.4, no. 1).58 The jars were set into clay applied over bedrock. One room was paved with red pottery fragments enhanced with white marble pieces, another with only white tesserae tiles and the third with stone chips set in mortar and pounded smooth. Under the third room was a large water channel linked to the plumbing of the house. A  bell-shaped cistern was cleared of mid-first to second-century AD pottery, including fine red wares (local copies and imported Eastern Sigillata B, Italian and African red-slip pottery), an inscribed lamp and the head from a female figurine. Significantly, none of the Rabat sites so far mentioned are said to have had underlying Phoenician or Punic domestic remains. On the basis

Malta’s Place in the Roman World of pottery recovered from the site, such as North African red-slip wares, the house appears to have come into use during the first century and remained so through to the closing years of the second century AD.59 Excavations at a second site (Saqqajja II) on Triq Nikol Sawra revealed architectural remains of the city fortifications, perhaps a gate or tower, 5 m below the current street level. Shops and dwellings sheltered by the modified portico of a building were found within the city wall. Imported coloured stones held in stone chests, a partially worked panel and shop furniture fitments indicate that one building within the group was a mosaic workshop; a small statue of Isis also was recovered.60 The evidence suggests that a 700-m-long ditch running the length of Triq San Rita, under St. Paul’s Church and still exposed in the garden behind the church, dates to Roman republican times or later (the chronological arguments for which can be found in Chapter  7). This cutting effectively isolated the residential areas from extensive burial grounds, which fan out west and south across the plateau and were still in use during the Roman period and long after. By tracing remnant segments of the fortification wall, Fr. Eugene Teuma has suggested that the wall itself was built 41 m from the edge of the ditch and may have reached thicknesses of 4.7 to 5.10 m.61 By the mid-seventeenth century AD, signs of the ditch in front of St. Paul’s Church were not apparent in two sketches made by Willem Schellinks. Both pictures were made from vantage points in the immediate area of the Piazza tal-Parrocca, one possibly near to the entrance of St. Paul’s Grotto, looking north-east to Mdina, and the other closer to Triq il-Vittorja, looking south along what is now Triq il-Kullegg.These fascinating glimpses of Rabat before recent historic developments show an undulating landscape with mounds suggestive of underlying ancient remains and exposed entrances to burial sites. Of the latter, Teuma suggests that one shown with a plaque and cross over the door became the Chapel of St. Mary Magdalena after a long period of use as a charnel house.62 The sketches also indicate the degree of clearing and levelling that took place to produce the modern landscape of Rabat, although the mound occupied by the pillar and cross (shown in both pictures) appears to have survived, now topped by a statue of St. Paul. The extensive burial grounds were already ancient in Roman times, having spanned the first millennium of Phoenician-Punic occupation. Nonetheless, other settled areas have been documented in these outer zones which date to the Roman or post-Roman periods. A  colourful mosaic fragment and limestone columns from an ill-defined ‘elaborate building’ were found in a field close to the Abbatija tad-Dejr catacomb.63 In Gnien is-Sultan, an outer district of Rabat (in a field called ‘Ix-Xagħra’64), significant scattered finds point to structures of Roman date, including slabs of ornate cornice possibly from a small shrine, wall foundations, remnants of a large water tank, coloured marble, tiles and mosaic fragments. Pieces of fine pottery, large


The Archaeology of Malta


amphorae and glass also were recovered in the habitation debris, as well as ‘heaps of ash’, animal bones and shell which had the appearance of ‘kitchen residue’.65 Such a location might represent well-appointed residences or a hamlet on the fringes of urban Melita. Certainly the concentration of necropoles in the region, many in use from Punic times into the Roman period of influence (Gnien is-Sultan, Għajn Klieb, Għajn Qajjied and so on) point to outlying settled areas. URBAN GOZO

Reports of finds of the Roman period in Gozo are few and poorly represented in publications. For descriptions of substantial Roman architecture within the citadel of Gozo, one must turn to the early antiquarian and historical accounts; the physical evidence is now reduced to a few fragments preserved in the Gozo cathedral and the Gozo archaeological museums and by private individuals. From the accumulated evidence, it is likely that Roman-period architecture lay under or near St. George’s Basilica: documented in 1937 were cisterns and a bronze coin hoard dated to AD 259–73,66 in 1976–7 building foundations,67 and in 2000 more foundations from an L-shaped building.68 The city walls or other sites rarely register beyond anecdotal accounts.69 Evidence of significant Roman-period remains were found in Victoria, in lands west of the citadel east of Triq L’Imgħallem (Foreman Street), some exposed in 1970s, more in 1995.70 Though the function of a major ashlar wall remains obscure (whether for fortification or revetment), the area did produce pottery-rich deposits and a large pit in clay with sherds of late Roman date. As in Malta, so it is in Gozo that when councils undertake work on infrastructure or residents build on their property, archaeological deposits are often encountered. This was demonstrated south of the citadel. A city wall most likely defined urban Gaulos, with necropoles falling outside the town’s southern-most limits, which followed the line of Triq Varjringa.71 Sewerage works that ran through It-Tokk (Independence Square) uncovered archaeological layers to a depth of 3.5 m which included numerous walls and floors, a burnt level thought to have been a house from which complete closed lamps and small pots were recovered and a cellar with fragmentary stamped amphorae, all falling within the Roman period. A  section drawing made by David Trump at the time (1961) has been reproduced recently.72 A  few years later, domestic animal bone and late Roman pottery, which appeared to be unstratified, were found during renovations in It-Tokk.73 Paving and some amphorae fragments were found in Via Kercem, Victoria.74 Observed in the grounds of a convent in Triq Palma in 1993 was a barrel-shaped cistern with pottery in its fill, which was probably linked to Roman-period dwellings.75

Malta’s Place in the Roman World MALTA’S RURAL SETTLEMENTS

While plunder and destruction reputedly characterised Roman dealings with the islands during the First Punic War, annexation as an outcome of the second war resulted in significant economic growth under Roman dominion. Development of the rural sector seems to have been no less significant than of the urban centres. Of the country villas so far identified in Melita and Gaulos – forty-seven locations have been documented to varying degrees – a number were fitted with installations for olive oil production; at other locations, only the large grinding stones or other fixtures have survived.76 Although Punic communities built villas and produced oil, some complexes were renovated or extensively renewed in the first centuries AD. The size of the olive-processing facilities at sites such as San Pawl Milqi indicate production on a large scale; whether fruit was supplied strictly by private groves or perhaps also by growers in the region remains unknown. Certainly at a number of villa sites the remnants of oil-production facilities and burial grounds have been documented on the slopes overlooking the Burmarrad flood plain as it extends inland (see Chapter 4). Given the numbers of known production sites, oil which was not destined for local consumption must have been transported overland to the closest port facilities for export.77 Great claims have been made for the substantial house and olive works at San Pawl Milqi. Some identify it as the country estate of the local dignitary, whereas others believe it to be the place where St. Paul stayed after the ship in which he was travelling (ca. 60 AD) was reputedly wrecked on the nearby rocky shore. Both traditions (this site as a governor’s home and St. Paul’s connection with Malta) continue to be vigorously debated.78 Archaeologically, the house reflects the fortunes that befell the island. From Punic origins and destruction by fire in the third century, the buildings witnessed rebuilding and continued occupation well into the ninth century. Although not restricted to rural areas (as we have seen from the Rabat finds at tal-Franciz mentioned previously), further oil-pressing installations were found in architectural remains at Ħal Għargħur in Triq San Gwann close to St. John’s Chapel.79 A few ancient remains survive in Birgu (Vittoriosa), occupying a peninsula stretching into the Grand Harbour. A  pink-granite column, two column bases, a substantial ashlar wall line and a millstone, probably from olive works, have been documented.80 On the basis of the millstone, Anthony Bonanno argued that Birgu may have been under cultivation, with a few corresponding rural dwellings in the region.81 Coinage at sites such as Ħal Far suggests continuity of occupation into the post-Punic period despite the poor preservation and early investigation of the site; surviving installations link it to thriving oil production in the island.82 At Bidnija east of a hamlet called ‘Ħal Dragu’, north-west of Mosta, only an olive-crushing stone and a few other fitments remain of the industry, but wall lines, tiles and pottery fragments also



The Archaeology of Malta were noted.83 From likely Phoenician-Punic origins, the house in Zejtun (in the girls’ school grounds) witnessed continued occupation into the Roman period. Coins dated between AD 222 and 361 have been recorded.84 This complex also was part of the olive oil industry which marked the Punic and Roman periods on the island, with fitments still evident at the site. Cisterns (one cylindrical) and channels comprised part of the water supply and plumbing of the building. Remains are limited to foundations and stone-paved floors, originally covered by a thin layer of soil once under cultivation. Ras ir-Raħeb (or Ras il-Knejjes) in the far west of the island is likely to date from the late Punic period into the Roman period and was perhaps involved olive oil production.85 Overall, four medieval church structures, including two at Ħal Millieri, testify to a historic settlement, now within a largely rural setting. Ashlar blocks of Roman dimensions appear to have been reused in the foundations of the Annunciation Church, as well as built into field walls in the area. An ancient olive-production facility is indicated by a large lower stone of an olive pipper in the courtyard associated with St. John’s Church. Nonetheless, the focus of the Ħal Millieri excavation was firmly on defining the medieval pottery sequence, and the Roman history of this location has yet to be further explored (see later for a discussion of the fictile remains).86 Not all sites have signs of olive oil works, but despite their poor preservation, a number have the remnants of substantial architecture and hallmarks of opulence, such as embellished architectural fragments, marble features and mosaic or tiled flooring.Towers in the Safi region, some of which clearly functioned during the Punic period, continued in use into the period of Roman influence. Other sites linked to this time are limestone quarries, which were exploited during the building boom of the Roman period. Large cisterns, some hewn in Punic times, others in the Roman period, continued to provide water for inhabitants in the rural sector.87 A bell-shaped cistern with Roman pottery fragments in the fill, for instance, was found in Triq in-Nutar Zarb (Attard).88 Ras il-Pellegrin, excavated in 1970–1, remains unpublished, but earlier field observations make brief mention of regular squared blocks, ‘abundant pottery’ and a rock-cut water cistern filled with debris (6 ft long × 2.9 ft wide), thought to be remnants of the Roman period.89 Indeed, in the absence of obvious habitation ruins, the cisterns are seen as indicators of occupied areas. In recent years (2006–12), the Cistern Exploration Project, Malta/Gozo has mapped and filmed a large number of water storage sites using remote sonar equipment; many belong to the recent historic era, but some are of considerable antiquity.90 As early as 1923, substantial ashlar masonry, including fluted columns from a house, had been uncovered on a farm in a locality known as Bir Ricca (already featured in the discussion of Punic Malta) in the field of Tal Ħereb (Zejtun) to the north of the Tas-Silg temple. A  large cistern located in the

Malta’s Place in the Roman World vicinity had already attracted attention. Domestic pottery, found in abundance, was assigned to the Roman period, but two twin-nozzle lamps  – a distinctive Punic type  – were among the fictile remains, possibly indicating pre-Roman-period habitation at the site.91 Although no ground plan or finds were published, nor were further investigations made, Zammit did note that additional archaeological evidence existed in the neighbouring fields (some no doubt referring to Tas-Silg).92 Indicative of the possibilities for further research in this region is the report that the foundation stones of the house survived to 76 cm under an overburden of 1 m of soil. Aerial reconnaissance of the Tas-Silg site revealed additional features in the surrounding fields, located directly north, which are still evident in satellite images.93 The data suggest that the Tas-Silg site stood within a landscape populated with some opulent houses in the Punic to Roman periods, built close to the temple buildings and overlooking one of the major ancient harbour facilities on the island. Although the sacred precinct may have suffered a downturn or redundancy under the Roman Empire (ca. first to fourth centuries AD), by the fifth century, the site had been re-invigorated by the construction of a Christian church. THE BATHHOUSE

Public bathhouses in both Malta and Gozo indicate that, as in other Roman regions, these establishments came to feature in daily life.94 This trend indicated a fundamental shift away from Punic traditions, where daily ablutions were conducted in the privacy of the home. Washing for a Punic resident was a strong cultural practice demarcating the transition from public to private space, and this practice is evident in the placement of washing facilities in better-preserved Punic domestic houses around the Mediterranean.95 A patron in the Roman-styled bathhouse could expect to advance systematically through a sequence of chambers which controlled heat – the frigidarium, tepidarium and caldarium  – as well as participating in exercise regimes, ball games or sweat rooms (sudatoria) which were offered in some establishments. Floors raised on multiple pillars to allow the passage of heated air, the hypocaust or suspensura, were pivotal to the successful operation of these establishments. A strong communal element marked the nature of the Roman bathhouse, facilitating social interaction. Classical writers of the first century AD, Celsus and Pliny the Elder, document the perceived health benefits of the public bathhouse.96 Għajn Tuffieħa was excavated in 1929 after ancient remains were discovered in the course of developing the regional water infrastructure. What proved to be a bathhouse complex of the Roman period (first to second centuries AD) had signs of opulence in the use of intricate mosaic and tiled floors (Figure 8.2, nos. i–vi) and the requisite range of purpose-built rooms, including a tepidarium, caldarium and furnace room with associated hypocaust installations (Figure 8.2, no. 3). A row of adjoining smaller ancillary rooms flanked a wide



The Archaeology of Malta

Figure 8.2.  Bathhouses from the Roman period of influence. (1) Small round example at Ta’ Baldu (after K. Buhagiar 2000). (2) Ramla Bay, Gozo, house and bathing facilities: (1–6) rooms of the villa; (7) entrance to the bathhouse; (8) passage to room 11; (9) large room isolated from the other areas; (10 and 11) dressing rooms with benches;

Malta’s Place in the Roman World corridor running east-west through the complex. Water cisterns, stone channels and clay and lead pipes were used to manage water throughout the site. A large lavatory with notched covering slabs which could accommodate up to nine people also was linked into the plumbing of the site through a channel that ran around the floor of the chamber. Further excavations, carried out in 1981, identified additional buildings and cultural remains. Private bath facilities have been recorded within larger estates, such as the waterside villa at Ramla Bay in Gozo (Figure 8.2, no. 2). This complex comprised at least six rooms in the dwelling and five more linked to the bathing facilities. Heating via elevated floors and ceramic flues set within the walls was generated by a furnace. Marble, coloured stones, painted plaster and some sculptures embellished some rooms.97 Other facilities have been documented to varying degrees in urban and rural settings from the early years of archaeological interest. Abela identified remnants of ancient bathhouses and fountains in Malta; one ruin near Għajn Ħammiem (Rabat) which may have been Roman in date was supplied with water from the spring.98 Caruana recorded baths embellished with mosaic and marble found in 1720 and baths with ‘monochromic’ mosaics in Mdina. Other remains were noted in the harbour area of Marsa. A smaller and simpler rural example has been documented in Ta’ Baldu near Dingli village (Figure 8.2, no. 1).99 MATERIAL CULTURE: POTTERY AND OTHER MANUFACTURES

Following a long and largely conservative evolutionary path, the common Punic Crisp Ware pottery tradition was reasonably consistent in character. Vessel shapes, however, came into and went out of fashion during the course of the centuries. Nonetheless, the evidence clearly points to continuation of the potting industry at the local level into the period of Roman influence.100 The tendency towards increasing grittiness and thinning vessel walls is especially evident at Tas-Silg. Designated Crisp Black Gritty, this ware is, on the whole, quite poor in quality. It would appear to have been widely used across the island, albeit represented by somewhat better examples than at Tas-Silg. Another category, Red Ware (Soft and Eroding), also might be a degenerative form of Crisp Ware which developed at the same time, perhaps in attempts to mimic imported Roman wares made of refined clays.101

Figure 8.2. (cont.) (12) bath; (13) room with marble floor; (14–18) hot rooms; (19) cold bath (after Ashby 1915, fig. 25). (3) Għajn Tuffieħa bathhouse: (i –vi) mosaic designs on the floors of the rooms – i in room E, ii in room L, iii in room I, iv in room K, v in room F and vi in room G (after Zammit 1930).



The Archaeology of Malta Funerary contexts from Melitan phaseVI in Malta illustrate the effects on the pottery market of moving into the Roman sphere of influence, trends which can be detected well before that time (from Melitan phase IV).102 Numerous forms of imported small cups and bowls, flasks and jugs sourced from the workshops of the Roman world appear, as well as vessels likely produced on the islands (Figures 8.3 and 8.4).103 Tas-Silg (predominantly in the north), San Pawl Milqi and Ras il-Wardija (Gozo), among other sites, have produced a range of Imported Roman fine wares such as tall beakers, mould-made decorated bowls and vessels with dimpled walls sourced from North African and Italian workshops.104 Such forms as the small bowl with vertical rim, often with a flange skirting the widest point, are well represented in collections amassed probably from burial contexts.105 At Ħal Millieri, excavations within the Annunciation Church, a fifteenth-century building, as well as in the area once occupied by the adjoining Church of the Visitation, substantially increased the range of data for the later Roman period.106 Pottery spanning the first century BC to the seventh century AD established the chronological depth of occupation in the region. Characteristic of the later potting industry are strong throwing ridges or rills, which find parallels in Roman-Byzantine pottery from other offshore regions.107 Stratified deposits at the Church of the Annunciation, Ħal Millieri, indicated that an earlier church structure had preceded it, and poorly preserved later Roman remains suggested that before that, it was once a habitation site. Indeed, ashlar blocks indicate that ancient architecture in the area had been reused in the church building.108 Similar to trends in other Mediterranean regions, a series of thin-walled and round-based cooking pots made of Red Bricky Ware was used widely in the early centuries AD.109 There are examples in the late Melitan tombs and at Tas-Silg, and they can be inscribed with dedications to the deity made prior to firing, which suggests strongly that they were manufactured locally; chemical analyses remain ambiguous in this regard.110 In Tas-Silg, what appears to be a degenerated form of this fabric, designated Late Red Bricky Ware, is probably the last of its type in the islands.111 The cooking pots find a wide distribution in the Mediterranean region, and variations in shape and rim form are numerous, which hints at chronological and regional development for the type. Many of them are fitted with a pair of opposing horizontal handles, some pushed close to the wall of the pot (Figure 8.4, no. 25). MARITIME TRADE AND HARBOUR INFRASTRUCTURE

Shipping and transport continued unabated into the Roman period. Marine finds have been documented regularly from 1959 and generally comprise transport amphorae and anchor components, as well as occasional remnants of prestige cargoes.112 Ancient graffiti depicting boats has been observed in

Malta’s Place in the Roman World

Figure 8.3.  Melitan phase VI pottery forms, strongly influenced by Roman ceramic types:  (1–7) urns and jars; (11–17) jugs; (19–22) small juglets; (23–28)  flasks (after Sagona 2002).



The Archaeology of Malta

Figure  8.4. Melitan phase VI pottery forms:  (1 and 2)  spouted flasks; (3 and 4) unguentaria; (5–23) various cups and bowls; (24 and 25) cooking pots; (26) a deep situla or bucket shape; (27) cooking tray or pan (after Sagona 2002).

a few locations, in the Tar-Ragħad Hypogeum, for example.113 Transport jars were entering the island over the centuries from a range of sources, which points to widespread trade extending from the east – the Levant and Aegean – through the central Mediterranean regions, including North Africa, Sicily and the Italian mainland, to the west and the Iberian Peninsula. As we have seen,

Malta’s Place in the Roman World some jars ended their journey serving as under-floor supports, whereas a few were left in tombs; we have fragments from domestic contexts and port storage facilities, and even the sea floor has yielded a considerable number. Together the evidence points to significant maritime activity associated with most points of good anchorage on both islands.114 Sections of the coastal waters have been archaeologically surveyed:  from Mellieħa Bay to Salina Bay in the north of Malta, the Grand Harbour, the northern Santa Marija and San Niklaw Bays in Comino, and southern Gozo, focussed at Xatt l-Ahmar and Xlendi Bay.115 Notice of a Roman wreck at Mellieħa Bay appeared in 1965; it became known as the ‘mortar wreck’ after its cargo of Italian mortaria of the first to third centuries AD, which was excavated and published by Honor Frost.116 Structures connected to maritime commercial activity were identified at Marsa as early as the 1640s.117 Ashby linked the southern area of the harbour with Ptolemy’s peninsula settlement, Chersonesus.118 Indeed, a complex of storage facilities which testify to commercial activity in the region – three ruined warehouses and cisterns  – was found in 1768 on Jesuits’ Hill (Ta’ Cejlu promontory). The remains are now covered by the Marsa power station (Figure  8.1, no.  2).119 These structures spanned the Roman and later periods; numerous Carthaginian and Melitan coins of the third to first centuries BC were recovered there, which might indicate an earlier Punic commercial foothold in the area. An overburden of more than 3 m of soil covered Roman-period structures, flooring, pits and pottery in a field bordering Racecourse Road, but these were restricted exposures in foundation trenches.120 Substantial architectural and archaeological deposits continued to appear periodically during building and other works in the 1940s, 1956 (Government Industrial Technical College, Coronation Gardens),121 1959 (Race Course Road in Marsa)122 and 1993 (road works at the power station).123 More recently, in 2005, a storm-water canal project exposed possibly the same ruins documented in early investigations and subsequently preserved under accumulated alluvium.124 Typical of their role in the storage and distribution of commercial goods in and out of the country, the finds included transport vessels (260 in one chamber).125 Tombs in nearby areas such as Paola indicate that a town probably stood in the region, but no remains of contemporary domestic structures have been found. Marsa’s commercial significance in the Roman period is not in doubt. Even though the location of the settlement of Chersonesus (the peninsula town) documented in Ptolemy is still unconfirmed, it is unlikely that he was referring to the Grand Harbour area. However, given the magnitude of the temple site at Tas-Silg, it probably stood in the vicinity of a substantially settled area. Houses, cisterns and tombs, as we have seen, border the coastal zones of Marsaxlokk Bay (from west to east, for example, Benghajsa, Ta’


The Archaeology of Malta


Kaccatura, Marsaxlokk [Bonanno’s site 11]126 and Bir Ricca), which points to well-established settlements in the region. Although these complexes were not linked into one large urban zone, they are likely to have formed a number of small towns or hamlets feeding into and out of the port and mooring facilities in the bays within Marsaxlokk, from the Punic period to the Roman period of influence. With key port facilities in Il-Magħluq within Marsaxlokk, it is hard to envisage that a settlement did not grow up around this commercial hub and on to the Delimara peninsula, crowned by the temple precinct; that this area was Chersonesus is possible.127 The downturn in fortunes at Tas-Silg, to the point that the temple buildings appear to have fallen into decay, may correspond with the commercial focus shifting away from the southern port it overlooked to the Grand Harbour region during the Roman period. At this stage, no massive storage facilities have been located in the Il-Magħluq area, and the increased volume of trade passing through Malta after the fall of Carthage likely necessitated the development of Marsa as a maritime commercial clearinghouse. The Roman practice of developing public infrastructure in its provinces saw the construction of other sacred precincts in the city of Melita, which may equally have contributed to the redundancy of Tas-Silg.128 RELIGIOUS PRACTICE AND SACRED SITES

The old Phoenician deities slipped from prominence or were assimilated into cults acceptable to Rome – hence the interlinked identities of Juno/Astarte and Hercules/Melqart in Malta. Formal religion in Roman contexts enforced the structures of the empire within communities which gathered for the performance of sacred ritual, and such occasions acted to bolster regional authority. Political and religious interests were often concentrated at the heart of towns. Urbanism and political power embedded in architectural expression and public spaces were no less significant than the religious rites.129 By the Roman period, temples and possibly administrative buildings were sited on the adjacent Mdina and Mtarfa plateau tops. In the 1600s, architectural remnants of grand proportions littered both areas; the Age of Enlightenment saw a number of fragments gathered into private collections and homes or reused in church buildings or buried when the locations were developed. By the 1800s, virtually all traces of ancient architecture in Mdina and Mtarfa had gone, and the same situation was repeated in Victoria, Gozo.What had not been taken or buried was destroyed.130 MDINA: TEMPLE OF APOLLO

A second-century AD inscription found in Mdina recognised the marble embellishments (columns and pilasters) to a temple made by a benefactor for whom a statue was erected. Significantly, the inscription records that he also

Malta’s Place in the Roman World installed a statue of Apollo, and this is taken to indicate that a temple to this solar deity once stood in Mdina.131 Caruana mentioned two ‘splendid marble buildings’, temple ruins, in Notabile (i.e., Mdina) in Strada Reale (i.e., Villegaignon Street) and a theatre.132 Various public works over the decades, as already noted, have uncovered deep deposits and significant architectural remains from more recent history back to the Bronze Age within the walls of Mdina.133 Discoveries made during projects in 2000–1 and 2005 generated a number of media bulletins, but archaeological reports of rescue excavations have remained at the ‘grey literature’ stage. Nonetheless, important glimpses of underlying deposits emerged at Vilhena Palace (the Natural History Museum),134 in St. Paul, Mesquita and St. Publius Squares and in Villegaignon Street. Skirting Mdina, during the addition of new buttresses below Vilhena Palace in 1963, three courses of reused Roman masonry were revealed along with a few fluted-column segments.135 MTARFA: TEMPLE OF PROSERPINA

A temple dedicated to Proserpina, goddess of the underworld and renewal, a deity who had no earlier link with the Punic era in Malta, once stood in Mtarfa. An inscription recording renovations of the temple indicates that it was already old and threatening to collapse around the turn of the first centuries BC–AD.136 By the 1880s, no trace of the structure could be found, although Caruana indicated that the site of the temple was marked by the statue of St. Nicholas. He states that Mtarfa was deserted: ‘the arrangements of the streets, and many tombs’ were still ‘scattered all around’.137 Architectural fragments have been preserved, as happened in Mdina, in museums and private collections, as well as through reuse in recent historical buildings.138 MITHRAISM AND THE MYSTERY CULTS

Until recently, Mithraism had not been traced in the archaeological record in Malta.139 This religion grew to prominence in Anatolia during the first century BC when the Commagene Kingdom emerged as a political force under King Antiochus I, who claimed ancestry back to the kings of Persia. It is likely that through connections between the Commagene Kingdom and Rome, belief in Mithra (or Mithras) was transmitted. A solar deity, Mithra also governed the life-giving waters and the abundance of nature. He was seen as the defender of good and destroyer of evil. As the keeper of truth, contracts and agreements were sworn in his name. In essence, Mithra was at the heart of a dualistic struggle between good and evil, between light and darkness, night and day. Cult doctrines – the Mysteries of Mithra – were taught to the initiated, who, after fulfilling seven stages of instruction, worked their way to eternal life after death.140 Driven by this tenet of renewal and hope, even for the most humble


The Archaeology of Malta


in society, Mithraism was readily adopted by soldiers, slaves, officials and citizens of Rome and its provinces. Along the journey that this religion took, striking characteristics and secret rites evolved. Wherever Rome advanced, so did the cult of Mithra, and by the second century AD, this manifestation of worship was sanctioned by the state. TENETS OF MITHRAISM

Renewal was visually encapsulated in taurochtonous (bull-slaying) scenes, which were placed in cult centres known as ‘mithraea’. The multi-layered iconography symbolised the astrological passage of time played out in a panorama of seven revered celestial bodies (including the sun and moon), stellar constellations and seasonal regeneration. Enduring and abundant nature is represented by trees, flowering plants, palm trees and fronds and ears of corn or wheat. Such scenes usually depict a raven (constellation Corvus, Latin for ‘crow’) as the messenger of Sol (the sun), who instructs Mithra to kill the bull (representing the moon). A second bird, the cockerel, was the herald of dawn; it chased away the demons and is often called the ‘Persian bird’. Celestial figures in Mithraic scenes also include the dog (Canis Major), the scorpion (Scorpio) and the lion (Leo), representing star constellations. Mithra is portrayed with two torch-bearing attendants: Cautes, with his torch held up symbolised dawn, and Cautopates, shown with his torch pointed down, represented sunset.Together the three fi ­ gures – Cautes, Mithra and Cautopates – represented dawn, noonday and sunset, the three cardinal points of the Sun. A Mithraic cult centre, as found in urban and rural contexts, comprised an assembly room with central aisle and benches on either side where ritual meals were consumed and ceremonies were carried out.141 The buildings often were decorated in a rustic fashion; walls were stuccoed to represent rough cave interiors, the night sky studded with stars was sometimes painted on the ceiling, or actual cave settings were left reasonably natural looking. Water was an important factor in site locations, and the sanctuaries could be built near the sea, close to springs or be provided with channels and basins. In the late Punic-Roman tombs of Rabat, a small number of figurines were found which show a caped figure on horseback, unmistakable depictions of Mithra (Figure 8.5, no 7). Coinage in the Pontic region of Turkey often depicts Mithra on horseback in front of an altar, sometimes with the torch bearers, with a raven and often under a tree and with a serpent (elements of this iconography might be preserved in Figure  8.5, nos. 5 and 6). Closer to the homeland of Mithraism, very similar figurines of Mithra on horseback came from Artachat in Armenia which date to the first century AD. Curiously, this particular Mithraic iconography is not very common outside the Middle East,

Malta’s Place in the Roman World

Figure  8.5. Traces of Mithraic religious practice:  (1)  stela depicting a Mithraic initiation scene found in Mdina; (2)  figurine of a cockerel found in tomb 4, New Street, Rabat; (3) impression of the Ras il-Wardija (Gozo), a possible mithraeum site,



The Archaeology of Malta although one example from Carthage has been published. Such depictions in Malta might suggest that the islands embraced the cult from an early date or that direct paths of transmission existed between Anatolia and Malta. In addition, figurines of cockerels (Figure 8.5, no. 2) also have been found in a few Rabat tombs – one together with a figurine of Mithra on horseback – which are likely relics of the cult. A carved lintel in the newly rediscovered catacomb at Ħal Resqun in Malta takes on greater meaning if viewed from a Mithraic perspective (Figure 8.5, no. 9).142 It depicts three human figures, two with hands up or out (a common Mithraic pose) and one with head pointed down; all are surrounded by animals and fish.While the sense of nature’s abundance is clear from these creatures, the three figures make sense if viewed as representing the three cardinal points of the sun, that is, if they are identified as the two torch bearers, Cautopates and Cautes, either side of Mithra. Hence, the upside-down figure of Cautopates can be explained as the setting sun and the half circle above him as probably a representation of a mountain behind which he will descend; the fish swimming nearby may be an allusion, too, to the fact that from a perspective in Malta, the sun sinks below the sea. Furthermore, solar and lunar devotion is reflected in coinage minted in Gozo depicting the bust of Luna over a lunar crescent (Figure 8.5, no. 10) and Sol (the sun god) as an armed soldier in front of a radiant orb on the obverse (Figure 8.5, no. 11). The assimilation of one solar divinity into another may have been a natural one, given the enduring promise of renewal embodied in the dying and rising god, Mithra, as also in the Punic solar divinity.143 That a place of worship, a mithraeum, was constructed close to an open-air Punic Figure 8.5. (cont.) illustrating the rock-cut benches and aisles, the outer compound wall and one of the rock-cut cisterns in front of an artificial cave cut in the cliff face; (4) one of two engraved gems made of chrysoprase found in Marsa (Malta), depicting the zodiac, seven revered astral bodies in Mithraism shown as stars, Luna the moon goddess and Sol the solar divinity on one side; the lion overcoming a bull (represented by its head) on the other side symbolises the sun overcoming the moon in the cycle of day and night; (5 and 6) carved wall in a Salina catacomb (see Figure 8.6, no. 2, complex at top left), defaced in antiquity, possibly depicting Mithra on horseback in front of an altar and a tree in the top left corner; (7) terracotta figurine of Mithra on horseback found in tomb 11, chamber A, in Triq Ferris, Rabat; (8) figure in an arched niche, possibly an unfinished representation of Mithra’s so-called rock birth, located in Għajn il-Kbir, north-east of Misraħ Għar il-Kbir; (9) carving in Ħal Reskun catacomb in Gudja (Malta) depicting the three positions of the sun – dawn, midday and sunset – in Mithraic terms, the figures of Cautes, Mithra and Cautopates, set in the abundance of nature, represented as animals and fish; (10 and 11) bronze coinage minted in Gozo depicting Luna above the crescent moon and Sol depicted as a soldier in front of a sun symbol (1–7 and 9–11 after C. Sagona 2009; no. 5; photo by K. Buhagiar; see also Sagona 2009, fig. 29; no. 8 photograph by A. Sagona).

Malta’s Place in the Roman World sanctuary at Ras il-Wardija (Gozo) also points to the persistence of Punic ­cultural memory well into the Roman period of domination. Two nearly identical cut gems which display Mithraic symbolism have been found in the port region of Marsa (Figure  8.5, no.  4). On one side, a lion standing over a bull’s head represents the solar divinity’s conquest of the moon. On the other side, Luna, with points of the crescent moon at her shoulders, stands beside Sol, equipped with rayed crown, whip and boots but otherwise naked; both figures are encircled by signs of the zodiac. Aside from these finds, a number of artefacts now held in private collections are likely to have come from one or more mithraea in Malta.144 As mithraea could accommodate only small groups of followers, they are found in considerable numbers throughout the Roman world. Possible sites have been identified in Malta, some purpose built, others in natural cave settings. Several appear to have been incorporated into the later catacomb systems, especially in Rabat.145 A unique and superbly carved marble plaque found in an unidentified location in Mdina carries a Mithraic initiation scene depicting (left) a father figure or Pater, the highest Mithraic grade, and (right) Miles, the soldier, another member of the cult’s hierarchy. Both are naked except for signs of their rank and identity. Between them is an initiate, a beardless youth dressed in Persian clothing – tunic, trousers, cloak and cap – typical of Mithra’s garb in taurochtonous and other scenes (Figure 8.5, no. 1).146 The significance of this find is deeper than the artistic merits of the work itself. As noted already, a temple to another solar divinity, Apollo, once stood in Mdina, which it should be recalled had been a sacred precinct in the Punic period. The evidence of such a relief carving, so heavily redolent of Mithraic practices, strongly suggests that a mithraeum also was once located in the vicinity. Likewise, a close association between Apollo and Mithra occurred at Cyrene in North Africa, where a mithraeum (called the ‘Cave of the Priests of Apollo’) was cut in a cliff face in the shadow of Apollo’s temple. Cyrene’s rock-cut mithraeum finds a striking parallel in Ras il-Wardija, Gozo (Figure 8.5, no. 3).147 The latter comprises a cavelike chamber cut into the cliff face with large wall niches and a second outer room; both were fitted with benches cut in the rock and altar places with cupped depressions in the rock. The mithraeum’s setting is spectacular, overlooking the sea and perched on a high cliff with rock-cut water storage facilities outside the chambers. Graffiti incised into the walls might represent the so-called rock birth of Mithra, simple depictions of a figure emerging from stone with arms held out or up.148 Its period of use is likely to fall between the early to late second through fourth centuries AD; however, the archaeological deposits removed by the Italian archaeological mission to Malta (1964–7) were secondary fill.149 To the south are the footings of what might be an older, constructed open-air, possibly Punic sanctuary with a cupped altar place.



The Archaeology of Malta A possible depiction of Mithra’s rock birth also can be seen in a rock-cut niche close to an extensive quarried area north-east of Misraħ Għar il-Kbir, in a location called ‘Għajn il-Kbir’ (Figure  8.5, no.  8), a site well worth detailed investigation.150 Another cultic shrine, possibly a mithraeum, was found in a region known as ‘Qasam il-Gewwieni’ (or Kasam-el-Geuieni), approximately 1.2 km south-east of Mdina, which received some attention in the nineteenth century (Figure 8.6, no. 3). Iconography in the chambers included an Aigikampos figure (the fish-tailed goat symbolic of the constellation of Capricorn), a carved person with a ‘long branched wand in his hand’, a figure likened to Harpocrates (perhaps one of the torch bearers) and depictions of the moon, a dog and two ill-defined animals. Rock-cut benches, an altar edged by trenches, possible circles decorating the walls and ceiling (perhaps representing celestial bodies) and water once flowing from a spring were features within three interconnected quasi-rectangular chambers.151 Margaret Murray considered ‘pagan’ a pair of chambers at Tal-Baqqari near one of the late Punic round towers. The name (bakar), which has been linked to cows and to terms for the dawn or morning, and persistent folklore concerning a golden bull found in the area are evocative of aspects of the Mithraic cult.152 OTHER CULTIC ASSOCIATIONS WITH MITHRAISM

Mithraic shrines sprang up throughout Europe, and some also display iconography of related and sometimes assimilated cults. Traces of such mystery cults have been identified in Malta, especially related to the winged Kronos, god of time, as well as Isis and Serapis, gods of a fertility cult which emerged in Egypt and depicted on a relief found at Tas-Silg.153 Solar iconography, laden with notions of renewal and regeneration, is woven into the centre of the carving of Isis and Serapis. It includes a flower ringed by an acanthus-leaf wreath which forms an orb springing from a water lily flower that in turn emerges from an egg shape. Acanthus leaves symbolise rebirth in connection with solar divinity. Water lilies (the blue Nymphaea caerulea)154 in iconography were borrowed from Egyptian contexts, where they can be depicted as flowers radiating out from solar orbs or in the hands of divine and royal figures. Water lily flowers carried connotations of renewal but also allusions to the potent hallucinogenic narcotic derived from the flower and its ability to enhance spiritual experiences. The egg shape on the stele might convey the sense of the cosmic egg seen in connection with Mithra and with Kronos.155 It may be significant from the point of view of ritual and symbolism at Tas-Silg that pollen from water lilies, an exotic species in Malta, has been identified in stratified soil samples, possibly once cultivated in a pool somewhere close to the site.156 Information is rather limited about other possible temple locations which have been identified in Malta and Gozo. Nearer the Grand Harbour region,

Malta’s Place in the Roman World

Figure  8.6.  (1 and 2) Salina Hypogeum (southern complex), St. Paul’s Bay, Malta; the catacomb appears to have reused an open courtyard with traces of rock benches,



The Archaeology of Malta substantial architecture comprising columns and a statue was found between 1.83 and 4.27 m deep in mud ‘in the basin at the foot of casal Paula’.157 Six white marble columns reputedly from a temple in the citadel in Victoria, Gozo, were reused in the wharf at Mgarr.158 Għar ix-Xiħ, a partially modified cave close to the bay, continued as a small shrine from Punic times into the Roman Republic period; the site was excavated between 2005 and 2008. Hearths and beaten earth floors dated to the second century BC.159 FUNERARY PRACTICE: THE TRANSFORMATION OF PUNIC BURIAL SITES

In Malta, cultural continuity is clearest in funerary contexts. Prolonged use of traditional burial grounds into the late Punic to Roman periods points to enduring social memory. It is a pattern which has been charted in other locations, such as Spain.160 Tombs of Melitan phase IV have the most graphic displays of symbolic features relevant to Phoenician-Punic belief systems (lamp niches, large empty niches, multiple floor trenches, instances of figurative art, elaborate pedestals related to lamp placement and well-cut and white-washed walls).161 Some tombs, already ancient by the Roman period and possibly linked to local elites in the colonial period (eighth to seventh centuries BC), were re-opened to take new burials, with no signs of disturbance in the intervening centuries.162 Layers of meaning are trapped within these depositions. For the phenomenologist, such evidence might evoke notions of a heightened sense of reverence for the person laid to rest in ancestral crypts, of a sense of belonging which stemmed from familial lineage possibly stretching deep into the past and of traditional links to the island and yet acknowledging a connection with the first wave of settlers to come from a distant homeland in the Levant. The associated rites, firmly Punic in nature and conducted by the grieving kin, convey a message shaped by an enduring sense of identity, one linked to place and ancestral heritage. Notably absent from the late Punic tombs, for instance, are single-nozzle lamps (see Figure 7.1, no. 23), even though they were present in non-funerary contexts in Malta. Instead, lamps continue to have twin nozzles, even though they had evolved over time to a simple non-flanged back, thin-walled, small and baggy form (see Figure 7.1, no. 21). Lamps, as we have seen, held symbolic and possibly religious significance, especially in funerary contexts (see Figure 7.1, nos. 13 and 14). For generations, rock-cut tombs were used exclusively for adult remains; only in the late Punic times were child

Figure 8.6. (cont.) which might have been a mithraeum (photograph by C. Sagona); (3) ground plan of three inter-connected caves in Qasam il-Gewwieni that probably were used as a cult centre, perhaps for Mithraic purposes (after Anon 1850, 1851).

Malta’s Place in the Roman World burials present, and then only on a few occasions. Resistance to change at the local level may be embodied in the renewed use of ancient burial grounds, especially in Rabat. Whether this translates into political but passive resistance to Roman dominion is harder to monitor, although such a scenario has been mooted for Sardinia, where Punic culture also lingered.163 Punic tombs fall into a pattern of a rock-cut deep shaft with narrow steps down one side or opposing footholds down the long walls of the shaft. Rectangular chambers have biers cut along the sides on which interred remains or cinerary urns were placed; the depth of the floor was increased by a trench cut in the centre and parallel with the long axis of the chamber, but symbolically this feature reflected adherence to deeply held belief systems.This type of tomb endures, but the final stages witness the appearance of head niches cut into a pillar of rock on the bier (see Figure 6.5, no. 12). Matching the likely population growth in the urban sector during the Roman period of influence, catacombs were hewn into the bedrock to cope with the concentration of people (Figure 8.6, nos. 1 and 2). Some complexes clearly broke into existing Punic tombs which have left traces fossilised within the cavities of the labyrinths.164 Others have possible hints of Mithraic influence, especially within chambers fitted with seating, well-made arched doorways that may have originally been cultic niches and rounded tables, presumably for ritual meals that carried on into later mortuary settings. It is likely that significant holdings in small ecclesiastical and private museums, characterised by large amounts of late Punic and Roman wares, both local copies and imported forms, were sourced from the late burial grounds and catacombs. Blown-glass bottles are well represented in the tombs, a production technique which is dated to the first century BC and after. Small clay unguentaria distributed widely throughout the Mediterranean appear in numbers from the Melitan phase IV tombs. They often show signs of damage made to release the contents of the bottles, which presumably held fragrant oils woven into the rituals performed at the burial of the deceased. SUMMARY

For the best part of a millennium, the Phoenicians and their Punic descendants shaped the cultural framework of their chosen home. They were drawn resolutely into the economic ambit of the Carthaginian hegemony along with many other Punic-held markets of the Mediterranean. Merchandise flowed into and out of the archipelago along well-honed maritime routes.The ubiquitous transport jars, manufactured in many lands bordering the Mediterranean, which came to rest in Malta, as they did elsewhere, stand as testimony to the far-reaching network to which the islands subscribed. Through annexation of lands allied with Carthage and the eventual defeat of Carthage itself, Rome inherited the Punic maritime commercial



The Archaeology of Malta infrastructure. If the shadow of an uncertain future passed over Malta and Gozo, it was fleeting. Any disruption to the population’s way of life and economic practices as a result of the conflicts was soon swept aside. Despite the might of Rome and its dominance over its territories, Punic characteristics persisted among the people living in Malta and Gozo. Surviving traits can be charted after 218 BC in the archaeological record of their sacred precincts, in funerary contexts, in continuing occupation of urban settlements and rural establishments, in coinage, in the language of some inscriptions, in people’s names and in pottery traditions. Continuity at the local level might be seen as a silent resistance to Rome. In reality, change had already been infiltrating the Maltese islands; it came hand in hand with commercial transactions. Perhaps underlying the accounts of pirates wintering in Malta and the notion of the islands as a place of refuge for political refugees is an element of non-conformity with Rome. Nonetheless, it was inevitable that aspects of Roman life would permeate the islands. It would seem from the archaeological record in Malta that after annexation, towns and country areas flourished. Major public building works were undertaken in Melita city, and harbour infrastructure was significantly developed in the inner reaches of the Grand Harbour at Marsa to cope with a marked increase in shipping. If the renovations of a number of villas are taken into account, the rural sector witnessed prosperity, albeit strongly influenced by Roman tastes. At the local level, fundamental changes came with the introduction of public bathhouses, and within the home were Roman fine tableware and cooking pots. What spelt prosperity for one may have disadvantaged others, and the south-eastern sector of Malta may have suffered a downturn when the focus of shipping shifted to the northern port facilities. The fabled sacred precinct at Tas-Silg seems to have become redundant; perhaps it was still visited by devotees who could not let go of the religion of their forebears, but they may have shared the site with new cults, of Isis and Serapis and perhaps of Mithra, which were shaped around familiar tenets of renewal. The pattern of new allegiances and economic gain was repeated in Gozo. This narrative may end in the stride of Roman influence on the island, but Malta was destined to see through the Middle Ages at first under Byzantine rule until that faltered with the Arab conquest of the islands in 870. After the Norman annexation in 1091, Christianity would again take hold in the hearts and minds of the inhabitants, who were governed by successive Sicilian monarchs.165 A  number of the archaeological sites that survive from these centuries are known to us from our examination of their earlier history.166 Architecturally, Malta and Gozo retain an array of structures from substantial fortifications, churches and chapels, as well as palatial urban to humble rural dwellings which included some cave sites.167 Rabat and Mdina, in particular, still convey a strong sense of their medieval past. A  Muslim burial ground

Malta’s Place in the Roman World was found over the remains of the Domus Romana in Rabat, and extensive ­catacombs remained in use through some of this period – and a few are linked to underground churches.168 Although there is a wealth of histories, documentation and artworks dealing with this period,169 the archaeological evidence noted in the literature is still quite patchy and remains to be woven into comprehensive studies.170



Some five-and-a-half thousand years passed from when people first dragged their boats ashore on the islands to the time when the Maltese Archipelago was gathered up in the Roman net cast over the central Mediterranean region. While much has been made of the Late Neolithic achievements – the megalithic monuments are an astounding legacy of a pre-literate stone age – every period and every phase has something to add to our understanding of how people lived in this island environment.The extraordinary claims made for the megalithic structures, bound as they are into ancient notions of the numinous, overshadow the remarkable achievements at all levels of people’s lives. Daily routines, rituals in the home and toils in the field are nonetheless known only in the broadest way. Although these islands occupy a small expanse of land, in terms of ancient territories, they were not inconsequential. For land masses devoid of mineral wealth, the succession of cultures made the best of what was on offer: reliable water supplies, fertile soil, easily worked stone for construction, the abundance of the sea, the advantageous central maritime location and significant mooring points along the coastlines. None of these qualities are insignificant. Circumscribed territory may influence how people now think about the islands’ past – it is difficult to escape our global perspective – but it does not appear to have curbed how people lived in Malta and Gozo in former times. The challenge for current researchers is to stand back and view the archipelago not in terms of its limitations but rather from the perspective of its strengths. As we have seen, a number of themes permeate research concerning Malta. Issues of cultural contact, isolation from or integration into economic networks of the Mediterranean, are ever present. It is irrelevant that imported objects are not found in great numbers in Malta. Such finds are recorded for all periods, whether obsidian, flint or crafted objects such as pottery and metalwork. They are clear indicators that the islanders were part of the extensive maritime activity to a greater or lesser degree over time. The intensity of occupation in certain locations throughout the centuries, one way or another, 296

Closing Remarks has compromised the archaeological record. Sites such as Mdina or Zurrieq, among a number of other historic locations, cap ancient deposits. Others are likely to have been swept away by natural forces, especially in the exposed high rocky regions. Conversely, some settlements remain obscured by the accumulated silt in the valley floors. Research has tended to focus on the monumental and the obvious sites in the landscape. How people felt, whether secure or vulnerable, in their island home is less clear. Skorba’s Early Neolithic settlement, hemmed in by a boundary wall, does represent a desire for clear territorial demarcation. Given the modest scale of the wall, it does not really seem an effective defence if there were a realistic threat. On one level, the later megalithic structures convey a sense of display, of self-assured confidence perhaps aimed at neighbouring settlements on the island. But those built in close proximity to the sea also may have presented an imposing façade, possibly directed at visitors to their shores. The Middle Bronze Age anxiety about Mediterranean neighbours may have stimulated the building of more substantial fortification walls and the location of settlements in naturally strategic locations. No evidence of destruction has been documented, yet some element of threat must have been present. The passage from modest habitations to monumental structures in Neolithic Skorba has invited much discussion which revolves around increasing social complexity, longevity of social order, rich belief systems and technological know-how. These cultural aspects are now coloured by accumulating data concerning environmental conditions and sophisticated geographical surveys which provide an expansive view of ancient landscapes. Even though such studies in Malta are relative latecomers to the wider archaeological scene, they are nonetheless most welcome and positive steps toward understanding human occupation of the islands throughout their history. On the strength of environmental core samples taken in a few locations, we now know that Neolithic farming may have had a negative impact on the delicately balanced ecosystem. Over time, it would seem that agricultural practices, which had already pushed the boundary of what the island could support, coupled with adverse climatic conditions, might have put an end to the Late Neolithic way of life. The archaeological record of the archipelago has proved to be a fertile landscape for new ways of interpreting the past. Increasingly, research is delving into the seemingly intangible realm of the senses. Evocative though the prehistoric megalithic monuments are, whether we can gauge the nature and magnitude of ancient sentiments is contingent on countless unseen and unknowable variables. For many decades, cultural change in Malta has been viewed as a series of breaks in the chain of occupation of the islands, of sharply distinct collapse episodes, after which new colonising events took place. Worth considering is the accumulating evidence for closure and possible intentional destruction



The Archaeology of Malta of some megalithic structures. Such behaviour paints a different picture, one tempered with choice. New cultural direction might be interpreted as social collapse of old practices, yet it has been demonstrated that signs of continuity are present in the archaeological record and that the local Neolithic communities made the decision to strike out along a new socio-economic path. An intriguing possibility is that structured closure of sites might be a concept that carried over into the Bronze Age with enduring populations. Whether ongoing adverse climatic conditions brought Tarxien Cemetery folk to Malta’s shores or whether they simply travelled over an ever-widening seascape of trade networks, the cultural trajectory in Malta took a new turn. Immigrants brought metal weapons with them in unprecedented quantities. Innovation included commercial enterprises aimed at offshore markets, notably extensive dye works and an associated textile industry. Some of these economic ventures brought imported artefacts to Malta which point to interaction with Sicily and southern Italy throughout the Bronze Age. There is no reason to believe that people in Malta did not participate actively in maritime enterprises. Inter-island commerce in the central Mediterranean tapped into distant eastern networks incorporating the Aegean, Cyprus and the Levant; even Egypt appears as a source of trade items. Evidence for continuity is present in the shift from Tarxien Cemetery to Borg in-Nadur culture at the end of the Early Bronze and the beginning of the Middle Bronze Age. By the time the Phoenicians made Malta their home in the first millennium BC, the islands already may have had a long history of commercial activity in the region. These Levantine merchants readily saw the value of the archipelago as a staging point and a hub in their westward infiltration of the Mediterranean, which ultimately went beyond the Straits of Gibraltar. Notwithstanding the lasting, all-pervading presence of the Phoenicians, during the early stages of their settlement in Malta there are markers of an enduring local Bronze Age pottery technology. Continued occupation of some sites, Mdina, for instance, suggests that the local late Borg in-Nadur population absorbed Phoenicians into their towns or tolerated them on the fringes of their settlements. Phoenician settlers, who eventually arrived in large numbers, however, propelled the island into a future saturated with Levantine people and their traditions. The apparently non-violent Phoenician occupation of Malta and Gozo and the archaeological record to date indicate that the resident islanders fully integrated with the Levantine settlers. In turn, the cultural practices of their Punic descendants would linger well into the Roman period of domination. The ethnicity of the people living in Malta and Gozo at this or at any time always lies at the heart of understanding the social processes at play. Cultural edges are not as clear-cut as how they have been portrayed in the literature. The issue of ethnicity also underpins how we view the surrender of the archipelago to Rome. Rome would reap the benefits of the extensive Punic trade

Closing Remarks systems into which Malta was firmly tied, but the island’s Punic inhabitants still remained Punic in many respects, perhaps for generations, before the old Levantine ways lost their relevance. Undoubtedly, the islands still conceal significant traces of their ancient record, as was so clearly demonstrated by the discovery in 2010 of the cuneiform inscription on an agate fragment found at Tas-Silg. In an instant, such discoveries can change how we think about the past, concerning the ingenuity of people not only as they adapted to different situations but also in regards to the connections they forged.This small but well-travelled object reinforces just how remarkable archaeology is in the Maltese islands.



Environmental (28 entries) Lab code Material Result (BP) 95.4% probability

Beta-208961 Plant fragments 1460 ± 40 536–659calAD

Lab code Material Result (BP) 95.4% probability

SacA-11658 Charcoal 1650 ± 30 330–433calAD

Lab code Material Result (BP) 95.4% probability

Beta-200517 Charcoal 2510 ± 40 796–509calBC

Site Lab code Material Result (BP) 95.4% probability

Tarxien BM-710 Beans 3286 ± 72 1700–1422calBC

Site Lab code Material Result (BP) 95.4% probability

Tarxien BM-711 Barley 3354 ± 76 1784–1496calBC

Lab code Material Result (BP) 95.4% probability

Poz-42682 Wood fragments 3655 ± 35 2139–1938calBC

Lab code Material

Poz-42443 Grain

(continued) 301

Appendix A

302 Environmental (28 entries) (cont.) Result (BP) 95.4% probability

3810 ± 30 2347–2190calBC

Site Lab code Material Result (BP) 95.4% probability

Tarxien BM-141 Beans 3880 ± 150 2711–1949calBC

Lab code Material Result (BP) 95.4% probability

Poz-42442 Charcoal 4010 ± 35 2620–2466calBC

Site Lab code Material Result (BP) 95.4% probability

Skorba BM-143 Charcoal 4380 ± 150 3381–2620calBC

Site Lab code Material Result (BP) 95.4% probability

Skorba BM-712 Charcoal 4478 ± 56 3359–3010calBC

Site Lab code Material Result (BP) 95.4% probability

Tarxien BM-101 Charcoal 4485 ± 150 3541–2874calBC

Lab code Material Result (BP) 95.4% probability

SacA-11663 Charcoal 4565 ± 30 3238–3108calBC

Site Lab code Material Result (BP) 95.4% probability

Ta’ Ħagrat BM-100 Charcoal 4640 ± 150 3695–3005calBC

Site Lab code Material Result (BP) 95.4% probability

Skorba BM-147 Charcoal 5000 ± 150 4080–3510calBC

Site Lab code

Skorba BM-145

Appendix A


Material Result (BP) 95.4% probability

Charcoal 5140 ± 150 4266–3652calBC

Site Lab code Material Result (BP) 95.4% probability

Skorba BM-148 Charcoal 5175 ± 150 4330–3695calBC

Site Lab code Material Result (BP) 95.4% probability

Skorba BM-142 Charcoal 5240 ± 150 4353–3712calBC

Lab code Material Result (BP) 95.4% probability

Poz-42441 Charcoal 5410 ± 40 4348–4226calBC

Lab code Material Result (BP) 95.4% probability

SacA-11665 2 Loripes lacteus (shell) 5445 ± 30 4349–4251calBC

Lab code Material Result (BP) 95.4% probability

Beta-200518 Charcoal 5730 ± 40 4689–4486calBC

Site Lab code Material Result (BP) 95.4% probability

Skorba BM-216 Charcoal 5760 ± 200 5081–4237calBC

Lab code Material Result (BP) 95.4% probability

Beta-200519 Seeds 5870 ± 40 4839–4654calBC

Lab code Material Result (BP) 95.4% probability

Poz-42444 Charcoal 6055 ± 35 5047–4848calBC

Lab code Material Result (BP) 95.4% probability

SacA-11668 Peat 6115 ± 30 5082–4949calBC


Appendix A

304 Environmental (28 entries) (cont.) Site Lab code Material Result (BP) 95.4% probability

Skorba BM-378 Charcoal 6140 ± 160 5386–4719calBC

Lab code Material Result (BP) 95.4% probability

SacA-11669 Charcoal 6500 ± 30 5525–5460calBC

Lab code Material Result (BP) 95.4% probability

Poz-42439 Charcoal 6650 ± 60 5661–5484calBC

Cultural (21 entries) Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3750 Animal bone 3580 ± 75 2140–1741calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle SUERC-4391 Human bone 3910 ± 40 2491–2283calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle SUERC-4390 Human bone 3920 ± 35 2489–2293calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle SUERC-4389 Human bone 4035 ± 35 2636–2471calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle UBA-10378 Human bone 4048 ± 28 2635–2478calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle UBA-10383 Human bone 4054 ± 24 2636–2488calBC

Appendix A


Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3571 Human bone 4080 ± 65 2780–2476calBC

Site Lab code Material Result (BP) 95.4% probability

Ħal Saflieni OxA-8197 Human bone 4130 ± 45 2873–2581calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3573 Human bone 4170 ± 65 2896–2580calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3575 Human bone 4225 ± 70 2940–2581calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3569 Human bone 4250 ± 65 3021–2831calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3574 Human bone 4260 ± 60 3024–2834calBC

Site Lab code Material Result (BP) 95.4% probability

Qala il-Pellegrin BM-808 Animal bone 4300 ± 60 3097–2856calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3570 Human bone 4300 ± 60 3097–2856calBC

Site Lab code Material Result (BP) 95.4% probability

Bur Mgħez OxA-8165 Human bones 4305 ± 65 3104–2849calBC


Appendix A

306 Cultural (21 entries) (cont.) Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3566 Human bone 4600 ± 65 3525–3262calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3567 Human bone 4860 ± 65 3794–3516calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3568 Human bone 5170 ± 65 4080–3796calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-5039 Human bone 5170 ± 130 4266–3703calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-5038 Human bone 5330 ± 100 4349–3966calBC

Site Lab code Material Result (BP) 95.4% probability

BR-C-X the Circle OxA-3572 Human bone 5380 ± 70 4348–4044calBC


This listing of archaeological sites is compiled alphabetically using common spellings of site names (this does vary) and according to the names generally used. Ħal Saflieni, for instance, appears first rather than Tarxien, the town in which the site is located. An index follows which lists sites in the gazetteer according to cultural period and feature types. Co-ordinates and elevation are provided using Google Earth at the end of each entry; the Google Earth reference targets the approximate centre of the feature wherever possible. These listings are not comprehensive because of the varying availability of data, and bibliographical details are provided as a guide. See also documentation on the website of the Superintendence of Cultural Heritage, National Inventory of Cultural Property of the Maltese Islands ( .mt/page.asp?p=21569&l=1). MALTA

Attard (Mosta Road): six bell-shaped pits, undefined remains; MAR 1938–39, §26–7, p. 5 (approx. 35°53’33.09”N, 14°26’22.26”E, elevation 79 m). Attard (Triq in-Nutar Zarb): bell-shaped cistern with Roman-period pottery fragments; MAR 1965, p.  2, fig.  1 (35°53’23.37”N, 14°26’37.53”E, elevation 65 m). Baħrija (Il-Qlejgħa tal-Baħrija): settlement, about forty rock-cut pits; Borg in-Nadur and Baħrija finds; Peet 1910; MAR 1960, p.  5; Evans 1971, p. 200; Trump 1966, 2002, p. 255; Mifsud 1995, table 20 (35°53’53.27”N, 14°20’04.06”E, elevation 152 m). Baħrija: cave site used in the Bronze Age; Planning Authority,‘Area of Arch. Importance’, Il-Qlejgha, l/o Rabat, ref. no. GF 0268/97 (35°53’47.41”N, 14°20’1.85”E, elevation 127 m). Benghisa (or Bengħajsa):  Roman-period house and scatters of masonry; Buhagiar 2007, §17, p. 189; MEPA_GN_241_1997 (no. 24) (35°48’36.37”N, 14°32’11.57”E, elevation 25 m). 307


Appendix B Bezbezija (an area south of Targa Gap and east of Mosta):  multiple-use tomb, Ggantija phase and architectural remains; Roman-period finds and house remains; MAR 1928–9, p. 4; 1929–30, p. 7, Evans 1971, p. 28, fig. 1 (466738) (35°54’53.46”N, 14°24’8.26”E, elevation 70 m). Bidnija region (Malta Survey Project area):  bathhouse with apses; multi-period finds; Docter et  al. 2012, pp.  123–4 and 140, fig.  19:5 (35°55’35.03”N, 14°24’28.70”E, elevation 79 m). Bidnija (Tal Bidni, east of the area Ħal Dragu, north-west of Mosta): olive oil installations; MAR 1911–12, §14, p. 4 (35°55’21.97”N, 14°24’9.67”E, elevation 87 m). Bidnija: a limestone vat of Roman date; Docter et al. 2007, fig. 15, site B21; Bonanno 1977, §14, p. 76 (35°55’34.11”N, 14°24’22.23”E, elevation 81 m). Bingemma: extensive burial chambers and small catacombs in a cliff face; prehistoric, Borg in-Nadur and later finds in region; Myres 1901, pp.  89–90; Buhagiar 1986, pp.  272–93 (35°54’13.56”N, 14°22’41.95”E, elevation about 178 m). Birgu (Vittoriosa): remnant oil press; Bonanno 1993, pp. 25–6, pls. 1.4 and 1.6–1.9 (35°53’15.61”N, 14°31’20.57”E, elevation 27 m). Birkirkara (on Mrieħel Ridge, vacant piece of land in Triq E.  Schembri):  Għar Dalam phase and prehistoric sherd scatter; MAR 1996, pp. 1–2 (approx. 35°53’32.96”N, 14°27’52.42”E, elevation 60 m). Bir Miftuħ: interconnected rock-cut pits; destroyed; Evans 1971, p.  200 (35°50’49.60”N, 14°29’54.89”E, elevation 79 m). Birzebbuga (Għaxaq):  dolmen, south-east of a Phoenician-Punic cistern, north-west of Borg in-Nadur; Bronze Age pottery was found in the dolmen; MAR 1914–15, p. 1; Evans 1956, p. 89; 1971, p. 195 (approx. 35°49’45.14”N, 14°31’27.88”E, elevation 24 m). Birzebbuga (St. George’s Bay shore, immediately east of Borg in-Nadur): about seventy rock-cut pits; MAR 1960, p. 5; Sagona 1999; Grima 2011, p. 366, fig. 11:8 (35°49’52.56”N, 14°31’51.72”E, elevation 3 m). Borg in-Nadur (area known as Xagħra Tal-Kaccaturi, Birzebbuga):  two possible Għar Dalam sherds; megalithic building period of Tarxien phase; Bronze Age (Tarxien Cemetery, Baħrija, Borg in-Nadur) finds; Murray 1923, 1925, 1929; Trump 2002, pp. 140–1; Evans 1953, p. 44; Tanasi 2011a, pp. 76–84, figs. 4:3–4:8 (35°49’52.27”N, 14°31’44.34”E, elevation 12 m). Borg in-Nadur: rock-cut grape-pressing pan (approx. 35°49’57.25”N, 14°31’34.83”E, elevation 17 m). Bugibba: Tarxien-phase megalithic structure; Cilia 2004, pp. 159–61;Trump 2002, p. 161 (35°57’16.87”N, 14°25’4.80”E, elevation 17 m). Buqana (Attard):  MAR 1910–11, pp.  3–4; 1926–7, p.  3; Evans 1971, p.  6; Trump 1966, p.  47 (approx. 35°53’51.49”N, 14°25’24.79”E, elevation 100 m).

Appendix B Bur Grad (or Ta’ Gernac at the limits of Gudja, Luqa and Tarxien districts): cistern; Zammit, notebook no. 4, pp. 101 and 103; MAR 1914–15, pp.  3–4; 1915–16; Buhagiar 2007, §28, p.  189 (approx. 35°51’56.88”N, 14°31’5.05”E, elevation 59 m). Burmarrad (district of Ta’ Diar id-Dwieb, field ‘Il Wilga ta’ fuk il Wied’ near the intersection of paths known as ‘Sqaq Tal Mqarqca’ and ‘Tal Latmia’): a largely destroyed Roman-period house site; Zammit, notebook no.  8, pp.  75–7; MAR 1925–6, p.  4 (approx. 35°55’51.99”N, 14°25’43.57”E, elevation 18 m). Burmarrad (and Wardija Road): scant remnants of a Roman-period house beside the road from Burmarrad to Wardija; water cistern; MAR 1926–7, p. 8; Buhagiar 2007, §40, p. 189 (approx. 35°55’56.51”N, 14°25’31.74”E, elevation 19 m). Burmarrad (in fields west of Burmarrad cemetery):  dense sherd scatter Hellenistic to Late Antiquity (35°56’30.88”N, 14°24’40.40”E, elevation 14 m). Burmarrad: three ashlar blocks south of the same location indicate a likely major classical site in the area perhaps deeply silted over; MAR 1992, p. 80 (35°56’29.62”N, 14°24’42.82”E, elevation 12 m). Bur Mgħez (located in Tan-Naxxari quarry, near Mqabba):  modified natural rock cavity; Late Neolithic and Tarxien Cemetery period burials; Tagliaferro 1911, pp.  147–50; 1912, pp.  143–50; MAR 1910–11, p.  4; 1911–12, p. 2; 1921–2, p. 4; 1922–3, pp. 2–3; Evans 1971, pp. 40–1; Cilia 2004, pp. 109, 144 and 203, see plan, p. 144 (35°51’2.39”N, 14°28’26.03”E, elevation 72 m). Fawwara: see Ras il-Gebel. Fiddien (Tal Għassiewi area near Mtaħleb): Roman-period house remains; MAR 1925–6, p. 4; Zammit, notebook no. 8, p. 60; Bonanno 1977, §5, p. 76; Buhagiar 2007, §5, p. 189 (35°52’52.28”N, 14°22’25.51”E, elevation 152 m). Gebel Għawzara: Bronze Age hilltop settlement; Bronze Age finds in the general vicinity (e.g., at nearby San Pawl Milqi); Buhagiar 1996, p.  20; Busuttil et al. 1969 (for 1968), p. 94; Malta Survey Project, Docter et al. 2012 (35°55’52.73”N, 14°24’18.09”E, elevation 114 m). Għajn Ħammiem (Għajn Hammam, Rabat): Roman house site; bathhouse, spring and water facilities; Abela book I, n. III; Caruana 1882, §98, p. 92; Zammit, notebook no. 7, pp. 16–17; MAR 1922–3, p. 7 (35°53’10.34”N, 14°24’0.19”E, elevation 177 m). Għajn il-Kbir: arched niche, possible Mithraeum site within an open quarry (35°51’17.42”N, 14°24’11.64”E, elevation 213 m). Għajn Qajjiet: Phoenician-Punic tomb site; Roman house remains; Buhagiar 2007, §6, p.  189 (35°53’09.35”N, 14°23’03.23”E, elevation 169 m).



Appendix B Għajn Tuffieħa (Mgarr): site in close proximity to a natural spring; a cave dwelling is in the region; Mgarr, Ggantija, Tarxien, Tarxien Cemetery pottery; substantial Roman-period bathhouse; MAR 1929–30, pp. 8–10; 1930–1, pp. 3–4; 1981, §5, p. 62; 1982, §4, p. 65; Zammit 1930a, pp. 56–64; 1931a, pp. 56–64; Evans 1971, pp. 29 and 231; Bonello et al. 1964 (for 1963), p. 20; Bonanno 1977, p. 76, site 3; Buhagiar 2007, §2, p. 189 (35°55’37.04”N, 14°21’26.94”E, elevation 84 m). Għajn Tuffieħa area (a.k.a. Ta’ Zammitello): stone circle, destroyed; MAR 1917–19, p. 12; Evans 1971, p. 39; Cilia 2004, p. 207 (photograph bottom right), back cover, site 49 (approx. 35°55’35.73”N, 14°21’25.28”E, elevation 80 m). Għajn Zejtuna (Mellieħa): Neolithic megalithic remains; coastal location at the water’s edge; destroyed; MAR 1934–5, p. 8; Evans 1971, p. 29; Cilia 2004, pp 208 (archive photo) and 328 (backboards); Trump 2002, p. 167 (approx. 35°58’3.19”N, 14°22’1.43”E, elevation 6 m). Għar Dalam (Birzebbuga):  cave deposits spanning Għar Dalam, Grey Skorba, Zebbug, Mgarr, Ggantija, Tarxien, Bronze Age, Baħrija finds; Issel 1866; Despott 1923; Sinclair and Keith 1924; Evans 1971, pp. 18–21 and 299 (35°50’11.40”N, 14°31’40.81”E, elevation 24 m). Għar Ħasan (or Għar Ħassan):  possible painted wall panels; Anati 1990, pp.  166–72; 1995, pp.  103–4; Mifsud and Mifsud 1997, pp.  145–51 (35°48’23.84”N, 14°31’3.77”E, elevation 36 m). Għar Mirdum: Tarxien Cemetery and Borg in-Nadur finds; later Punic deposit; Tanasi (in press), §4, fig. 16:e (MRD64/P/753); MAR 1964, p. 4, pl. 2:1–4; 1965, pp.  1–2; Evans 1971, p.  22; Calleja-Gera 2008 (approx. 35°50’47.41”N, 14°23’31.70”E, elevation 191 m). Għar in- Ngħag: used cave; Neolithic period; Cilia 2004, backboard, §7 (35°48’25.71”N, 14°31’38.91”E, elevation 31 m). Għar ir-Riħ (between Il-Mara and Għar Ħasan):  used cave in Bronze Age; MAR 1974–5, §5, p.  56 (approx. 35°48’27.01”N, 14°31’23.83”E, elevation 24 m). Għar tal-Friefet (south-east of Għar Dalam): a cave site; Bronze Age finds; MAR 1963, pp.  3–4, fig.  2; Evans 1971, p.  229 (approx. 35°50’4.88”N, 14°31’45.69”E, elevation 18 m). Għar tax-Xemgħa:  cave habitation site (35°56’50.88”N, 14°22’46.47”E, elevation 32 m). Għar Tuta:  habitation; Abela 1647, p.  97; Buhagiar 2007, §8, p.  189 (35°50’31.55”N, 14°25’5.54”E, elevation 211 m). Girgenti: Neolithic-period natural cave site; finds described as of the ‘Copper Age’ (Tarxien?); MAR 1965, p.  2 (approx. 35°51’31.72”N, 14°23’49.19”E, elevation 212 m). Gnien is-Sultan (district of Rabat):  Roman house and shrine; Punic tombs; (approx. 35°53’15.54”N, 14°23’22.30”E, elevation 156 m).

Appendix B Ħagar Qim: oval dwelling (east building); megalithic buildings; museum site; Evans 1971; Trump 2002, pp. 142–7; Malone et al. 2009, p. 54, fig. 3:3 (35°49’39.93”N, 14°26’31.54”E, elevation 128 m). Ħal Bajjada: remnants of a catacomb; MAR 2002, p. 176 (35°52’29.59”N, 14°23’46.99”E, elevation 211 m). Ħal Dragu: Roman house remains; olive oil installations; MAR 1911–12, p.  4; Buhagiar 2007, §33, p.  189, fig.  27 (approx. 35°54’44.66”N, 14°24’32.56”E, elevation 67 m). Ħal Far (Birzebbuga): catacomb, unspecified location found in 1999; megaliths in the area; menhir moved from original location; dolmen known as Il-Gebla Msaqqfa; MAR 1913–14, p. 2; 1999, p. 199; Zammit, notebook no.  4, pp.  68–72; Zammit 1930, p.  59; Evans 1971, pp.  198–9 and 229; Trump 2002, pp.  167 and 169 (approx. 35°48’51.95”N, 14°30’32.98”E, elevation 65 m). Ħal Far (north-west of St. Angelo’s chapel):  Bronze Age and possibly earlier wares also were recorded; later olive oil installations; architecture; water cistern; Zammit, notebook, no. 6 (16 March 1922), pp. 49, 50 and 53; MAR 1921–2, p. 4 (approx. 35°49’18.78”N, 14°30’20.47”E, elevation 77 m). Ħal Farrug (Luqa Parish): dolmen; MAR 1914–15, p. 1; 1936–7, p. 14; Evans 1971, p. 230 (35°51’38.90”N, 14°28’13.57”E, elevation 66 m). Ħal Għargħur (in Triq San Gwann close to St. John’s Chapel):  house and oil installations; MAR 1954–5, p.  4; Bonanno 1977, §19, p.  189 (35°55’26.51”N, 14°27’05.87”E, elevation 140 m). Ħal Gilwi: house and olive oil installation; Buhagiar 2007, §19, p.  189 (approx. 35°50’56.50”N, 14°33’11.58”E, elevation 33 m). Ħal Ginwi (Triq Xrobb l-Għagin, Zejtun): megalithic buildings; Ggantija, Tarxien; some Tarxien Cemetery pottery; Roman-period house remains, mosaic floors; MAR 1917–19; Zammit, notebook no. 13 (21 July 1917), p. 206; Laferla 1918, p. 235; Evans 1971, pp. 25–6; Cilia 2004, pp. 99–100; Trump 2002, p. 191 (35°50’53.90”N, 14°32’50.01”E, elevation 36 m). Ħal Kbir: post-Punic cistern; Buhagiar 2007, §10, p.  189 (approx. 35°49’49.45”N, 14°27’42.47”E, elevation 93 m). Ħal Kirkop: menhir; MAR 1913–14, p. 3; Zammit 1930, p. 60; Evans 1971, pp.  198 and 230; MEPA_gn571_94, plan (35°50’40.98”, 14°29’6.57”E, elevation 77 m). Ħal Millieri: Borg in-Nadur finds; oil installation, post-Punic; Blagg et al. 1990, pp. 43–53 and 105–6; MAR 1976–7, §3, p. 64; 1977–8, §§1–2, p. 63; 1979, §6, p. 60 (30°50’20.21”N, 14°28’15.47”E, elevation 91 m). Ħal Resqun: catacomb rediscovered in 2006 (Sunday Times of Malta, 27 August 2006); MEPA, gn628_08 (§ 26), plan 17; Superintendent of Cultural Heritage 2005, §2.4.5, p.  7 (35°51’10.23”N, 14°29’40.40”E, elevation 74 m).



Appendix B Ħal Saflieni: extensive prehistoric rock-cut hypogeum; museum site, UNESCO World Heritage listed; Zebbug, Saflieni, Ggantija and Tarxien finds; Tarxien Cemetery and Borg in-Nadur pottery fragments in bell-shaped tomb (found 1991); Tagliaferro 1910, pl. 9; Dukinfield Astley 1914, pp. 394–6; MAR 1991, p. 81; Evans 1971, pp. 59–60; Trump 2002, pp.  128–35; Tanasi 2008, fig.  14 (35°52’10.90”N, 14°30’25.04”E, elevation 54 m). Ħas Saptan: cistern; Buhagiar 2007, §26, p.  189 (35°50’01.46”N, 14°31’02.90”E, elevation 37 m). Ħlantun: post-Punic cistern; MAR 1916; Buhagiar 2007, §29, p.  189 (35°50’1.46”N, 14°31’2.90”E, elevation 37 m). Id-Debdieba (now Luqa airport runway): Megalithic structure, destroyed; finds spanned Ggantija, Tarxien, Punic and Roman periods; possible low-lying fortified Borg in-Nadur settlement; MAR 1914–15, p. 1; Ashby et al. 1916; Trump 2002, p. 190 (35°51’12.58”N, 14°27’55.24”E, elevation 100 m). Id-Debdieba: rock-cut pit; Punic finds; Ashby 1914; MAR 1914, p.  15.1 (35°51’3.15”N, 14°28’0.05”E, elevation 92 m). Il-Bezbezija (Il-Busbisja): possible house site; MAR 1929–30, p. 7; Buhagiar 2007, §34, p. 189 (approx. 35°54’25.58”N, 14°24’29.74”E, elevation 90 m). Il-Hazzien: possible post-Punic house; Buhagiar 2007, §11, p. 189 (approx. 35°49’46.82”N; 14°27’44.69”E, elevation 95 m). Il Ħofra (Gudja): menhir; other stones (destroyed) suggest that it was a megalithic building site; quantities of pottery and lithic material; Evans 1971, pp. 198 and 229 (approx. 35°50’53.55”N, 14°29’52.04”E, elevation 73 m). Il-Latmija (Mellieħa region):  cave site with Neolithic deposits; historic period troglodyte dwelling; Cilia 2004 (backboard); Skeates 2010, pp. 85 and 126, fig. 3:5; Buhagiar 2001, pp. 3–9 (35°58’38.87”N, 14°19’43.01”E, elevation 59 m). Il-Qortin (overlooking Wied Qannotta, close to L’Mdawwar Road): red-slipped Borg in-Nadur pottery and several rock-cut pits were identified at the site; Evans 1971, p. 200; Mifsud 1995, p. 50 (vicinity of Qolla, west of Burmarrad). It-Tafal ta’ Bingemma: grape-pressing pans; Bonanno 2008a, appendix (approx. 35°54’21.07”N, 14°22’41.13”E, elevation 166 m). It-Torrijiet (or Tat-Torrijiet, district of Tal-Ħlantun, Zurrieq):  round tower; MEPA_GN 1082_2009 (no. xiii), plan 13/35; Bonanno 2005, pp. 91 and 294–7; Trump 1997, pp. 91–2 (35°49’30.83”N, 14°29’21.23”E, elevation 110 m) It-Tumbata (Luqa):  possible megalithic remains; three inter-connecting rock-cut pits; MAR 1959–60, p. 3; 1960, pp. 4–5; 1961, p. 5; Evans 1971, pp.  24 and 200; Trump 2002, p.  167 (35°51’30.31”N, 14°29’1.45”E, elevation 79 m).

Appendix B Kavallerizza: settlement; Borg in-Nadur phase; Tanasi 2008, fig. 14. Kordin I: megalithic building; destroyed;Ashby et al. 1913, pp. 17–62;Trump 2002, p. 188 (35°52’52.58”N, 14°30’15.53”E, elevation 18 m). Kordin II: megalithic building; destroyed; Ashby et  al. 1913, pp.  17–62; Trump 2002, p. 188 (35°52’54.41”N, 14°30’23.29”E, elevation 41 m). Kordin III: one Għar Dalam sherd, Mgarr, Ggantija and Tarxien finds; megalithic buildings; Ashby et  al. 1913, pp.  17–62; Evans 1953, p.  44; Trump 1966, p. 19, fig. 17:a; 2002, pp. 136–7; N. C.Vella 2004, pp. 17–19 (35°52’37.60”N, 14°30’32.59”E, elevation 41 m). Kuncizzjoni (Qortin l-Imdawwar):  remnant megalithic building and Tarxien-period finds;Tarxien-phase houses; MAR 1938–9, pp. 6–9; Evans 1971, p. 109; Trump 2002, p. 166; Malone et al. 2009, pp. 52 and 54–5, table 3:3 (35°54’03.20”N, 14°20’24.09”E, elevation 152 m). Laferla Cross (Tas-Salib):  rock-cut grape pressing pan; Bonanno 2008a, 2011, p. 65 (approx. 35°51’0.16”N, 14°25’1.41”E, elevation 219 m). L-Iklin: Roman house remains; MAR 1975–6, §2, p. 60; Bonanno 1981a, pp.  213–20; Buhagiar 2007, §42, p.  189 (35°54’38.96”N, 14°27’15.45”E, elevation 117 m). L-Imselliet: olive works and tombs, Bonanno 1992, p.  74 (approx. 35°54’57.53”N, 14°23’43.27”E, elevation 80 m). Luqa Cemetery: rock-cut pits; see Tal-Mejtin (35°51’17.48”N, 14°29’42.49”E, elevation 72 m). Manikata: fortified site; Borg in-Nadur finds; Fedele 1988, site 8.9.87/17, p. 198, fig. 3:17, ref. 2, §5, p. 29; Weston 2010, §5, p. 29 (35°56’34.36”N, 14°21’35.21”E, elevation 80 m). Marnisi (Marsaxlokk area):  cistern; MAR 1933–4, pp.  9–10 (approx. 35°50’48.78”N, 14°31’50.47”E, elevation 57 m). Marfa (Mellieħa Ridge): stone circle, destroyed; referred to as ‘Marfa Middle’ by Veen and van der Blom 1992, site 17, pp. 65–6; they indicate that it was an apsidal building with Għar Dalam and Zebbug sherds; Evans 1971, fig. 1, ref. 408819; MAR 1917–19, p. 12; 1929–30, p. 5; Fedele 1988, p. 198; Mifsud 1995, p. 48, fig. 14; Cilia 2004, backboard, site 40 (35°58’51.23”N, 14°20’38.27”E, elevation 39 m). Marsa: Roman house structures; Barbaro 1794, pl. 1; Caruana 1882, §98, p. 92; 1896, p. 33; Bonanno 2005, p. 238 (35°52’54.97”N, 14°29’59.40”E, elevation 11 m). Marsaxlokk: Roman-period house; MAR 1931–2, p.  5; Bonanno 1977, §11, p.  76; Buhagiar 2007, §23, p.  189 (35°50’10.64”N, 14°32’39.18”E, elevation 6 m). Mdina: Bronze Age and Phoenician strategically located settlement; bell-shaped pits and other rock cuttings; Roman-period religious precinct; Sagona 1999, pp. 36 and 43, pl. 6 (35°53’4.34”N, 14°24’9.56”E, elevation 200 m).



Appendix B Mdina (De Piro’s house):  ancient pavements were exposed in the yard; MAR 1906–7, p. E7 (35°53’12.17”N, 14°24’15.12”E, elevation 190 m). Mdina (Mesquita Square):  archaeological deposits of the Bronze Age–Phoenician phases, excavated in 1998 (35°53’8.80”N, 14°24’9.27”E, elevation 199 m). Mdina (Natural History Museum,Vilhena Palace): post-Punic architecture (west); fortification (east); forecourt excavations rear courtyard (2000–1); Trump 1997, p. 106 (35°53’5.47”N, 14°24’13.13”E, elevation 199 m). Mdina (Triq il-Villegaignon, near Carmelite Monastery):  excavated dark-earth strata and pottery finds of the Bronze Age–Phoenician interface (35°53’11.92”N, 14°24’10.99”E, elevation 199 m). Mdina (Triq Inguanez):  excavation site; 4-m dark-earth deposits; Borg in-Nadur sherds were found in deposits under a Late Antiquity mortar floor; Phoenician layers; MAR 1993, p.  74; R.  Zammit, ‘Roman Wall Found in Mdina’, in Times of Malta 7 May 2005 (35°53’6.24”N, 14°24’13.54”E, elevation 199m). Mdina (Xara Palace Square):  3-m Phoenician deposit; three phases and architecture (35°53’6.27”N, 14°24’15.03”E, elevation 194 m). Mellieħa Ridge: see Marfa. Misqa Tanks (near Mnajdra and Ħagar Qim): inter-connected large rock-cut pits; Evans 1971, pp.  200–1; Trump 2002, pp.  152–3 (35°49’44.96”N, 14°26’07.06”E, elevation 121 m). Misraħ Miel (area of Castello Zammitello): house and olive oil production facilities; Bonanno 1977, §4, p. 76; Buhagiar 2007, §3, p. 189; 1992, p. 26 (35°55’14.71”N, 14°21’32.59”E, elevation 91 m). Misraħ Sinjura (Qrendi Parish): dolmen with petroglyph; Ashby et al. 1913, p. 8, pl. 1, fig. 1; Evans 1971, pp. 196, 197 and 231 (approx. 35°50’30.52”N, 14°27’2.21”E, elevation 100 m). Mnajdra: megalithic building; museum site, UNESCO World Heritage listed; Ggantija and Tarxien phases; a pit aligned with the building may have been the socket for a menhir; Trump 2002, pp. 148–51; Cilia 2004, pp. 126–41 (35°49’36.37”N, 14°26’10.61”E, elevation 83 m). Mosta (far northern sector):  two clusters of megaliths in boundary wall (approx. 35°55’2.70”N, 14°25’12.44”E, elevation 84 m). Mtarfa: Bronze Age and Phoenician-Punic finds; rock-cut pits; reputed classical Temple of Persephone (destroyed) but traditionally located at the site of St. Nicholas statue, Oswald Street; MAR 1924–5, p.  2; 1938–9, p. 12–13; 1960, p. 5; 1973–4, §5; Ward Perkins 1938–9, p. 34; 1942, p. 34; Evans 1971, pp. 107 and 200; Sagona 1999 (temple site:  35°53’25.15”N, 14°23’45.82”E, elevation 189 m; pits: 35°53’24.37”N, 14°23’37.98”E, elevation 191 m). Nadur (Gudja, near megalith Id-Dawwar site): dolmen; Evans 1971, p. 196 (35°50’8.09”N, 14°30’42.68”E, elevation 66 m).

Appendix B Paola:Triq il-Palma; cistern; MAR 1936–7, pp. 13–14 (approx. 35°52’11.99”N, 14°30’25.67”E, elevation 54 m). Qajjenza (Birzebbuga):  megalith structure destroyed; Bronze Age, Borg in-Nadur settlement; Mifsud 1995, table  19; Cilia 2009, §3, backboard; Tanasi 2008, fig. 14 (approx. 35°50’0.22”N, 14°32’10.64”E, elevation 21 m). Qala Hill: a naturally fortified location with significant spring, Borg in-Nadur period sherd scatters, a large number of bell-shaped pits within the settlement and a fortification wall; possible ditch; Mallia 1968, p. 60;Trump 1997, p. 142; Mifsud 1995, p. 49 (35°56’11.41”N, 14°22’43.51”E, elevation 93 m). Qaliet Marku (Naxxar): megalithic structure; Trump 2002, p. 167; MAR 1927–8, pp. 3–4; 1935–6, p. 18; Evans 1971, pp. 44 and 231, fig. 1 (499776) (approx. 35°56’29.64”N, 14°26’39.46”E, elevation 33 m). Qallilija (Rabat): tentatively identified as prehistoric megalithic site; Bronze Age, Borg in-Nadur, Baħrija finds; possible remnants of a dolmen; menhir; rock-cut pits; Trump 2002, p. 167; Bellanti pre-1924; Zammit, notebook no. 4, (26 September 1912), pp. 4, 5, 9–17; MAR 1912–13, pp. 3–4; Ashby et al. 1913, p. 126 (approx. 35°53’37.12”N, 14°23’5.17”E, elevation 155 m). Qarraba (Għajn Tuffieħa): a Bronze Age site with Borg in-Nadur period finds;Trump 1997, p. 142; 2002, p. 252; Mifsud 1995, p. 49 (35°55’38.44”N, 14°20’27.33”E, elevation 28 m). Qawra (along the undeveloped coast):  dense Punic sherd scatter; MAR 1997, p. 1 (approx. 35°57’25.12”N, 14°25’20.10”E, elevation 22 m). Qawra: catacomb complex, opposite Salina catacomb, MAR 2002, p. 176 (approx. 35°56’58.37”N, 14°25’12.77”E, elevation 31 m). Qolla (Il-Qolla; west of Burmarrad, San Pawl il-Baħrija, Bidnija):  settlement on plateau; rock-cut pits; Evans 1971, p.  232; Mifsud 1995, p.  60 (35°55’53.21”N, 14°23’50.05”E, elevation 87 m). Qolla: Borg in-Nadur sherd scatter on northern slopes of the plateau; Mifsud 1995, table 19 (approx. 35°56’0.50”N, 14°23’50.95”E, elevation 53 m). Qrendi: Sqaq il-Bagħal; possible megalithic building remains; Trump 2002, p. 167 (approx. 35°50’6.97”N, 14°27’17.49”E, elevation 106 m). Rabat (near the Church of St. Marija tal-Virtu): bell-shaped pits, possibly Bronze Age, never investigated; see Buhagiar 1979, p. 326, n. 4 (approx. 35°52’23.95”N, 14°24’25.80”E, elevation 195 m). Rabat (in a field called ‘tal-Franciz’ likely to be the location of Santa Margerita Cemetery):  Bronze Age settlement remains; MAR 1957–8, p. 9, pl. 4 (35°53’6.14”N, 14°23’58.91”E, elevation 192 m). Rabat (Triq Nikol Saura):  architectural remains (35°52’55.10”N, 14°24’7.02”E, elevation 197 m). Rabat (1909 excavation): buildings and cisterns; archaic sherds (presumably Bronze Age) found in a 6-m-deep sondage; MAR 1909–10, p. 7 (approx. 35°53’7.15”N, 14°24’3.40”E, elevation 194 m).



Appendix B Rabat (1920 excavation):  level platform; L-shaped cistern; pillars; spring (approx. 35°53’8.25”N, 14°24’0.83”E, elevation 188 m). Rabat (1983–4 UCLA excavations in the open square, Misrah Anton Agius; now a car park):  architectural remains; a bottle-shaped cistern (35°53’4.20”N, 14°24’3.47”E, elevation 198 m). Rabat (2006 excavations at a building site): stratified massive architecture; pottery (35°53’5.88”N, 14°23’57.53”E, elevation 191 m). Rabat (2007 excavation): exposed ancient defensive ditch (35°53’2.84”N, 14°23’45.09”E, elevation 187 m). Rabat (Domus Romana):  archaeological museum on site; Caruana 1882, §98, p. 92; Bonanno 2005, p. 236 (35°53’6.71”N, 14°24’0.98”E, elevation 193 m). Rabat (St. Paul’s Square): 2000–1 and 2005 excavations revealed deep deposits and architecture (35°53’10.64”N, 14°24’12.29”E, elevation 200 m). Rabat (currently a garage): mosaic workshop (35°52’56.29”N, 14°24’7.88”E, elevation 196 m). Ras il-Gebel (‘Tal Palma’, and Fawwara, Mgarr; rocky plateau, north-west of Zebbieħ):  tentatively identified as prehistoric megalithic site; stone circle; strategic Bronze Age settlement and finds; MAR 1914–15, p.  2; Trump 2002, p. 167; Cilia 2004, backboard, §48 (approx. 35°53’30.82”N, 14°23’11.77”E, elevation 123 m). Ras il-Pellegrin: Ggantija finds; Roman house site; Evans 1971, p.  216; C.  Vella 2010; Bonanno 1977, §2, p.  76; Buhagiar 2007, §4, p.  189 (35°55’00.52”N, 14°20’16.13”E, elevation 115 m). Ras ir-Raħeb (or Ras il-Knejjes): remnant megalithic stones still in situ; house remains, with olive press stone; only Punic and later pottery recovered from the site; MAR 1922–3, p.  5; 1962, p.  6; Bonanno 1977, §1, p.  76; Trump 1997, p.  127; 2002, p.  167 (35°54’23.19”N, 14°19’42.79”E, elevation 21 m). Safi (Triq Iz- Zurrieq):  substantial wall remains on the roadside (35°49’57.97”N, 14°28’54.33”E, elevation 109 m). Safi: Roman house remains; MAR 1966; Buhagiar 2007, §31, p. 189 (approx. 35°49’49.58”N, 14°29’14.87”E, elevation 98 m). San Pawl Milqi: Zebbug ochre burial; Borg in-Nadur pottery finds; Punic-Roman house; oil installations; tombs; cisterns; Mizzi 1879; Bonello et al. 1964 (for the 1963 season); Cagiano de Azevedo et al. 1965 (for 1964); 1966 (for 1965); 1967 (for 1966); Bozzi et al. 1968 (for 1967); Busuttil et al. 1969(for 1968); Buhagiar 1996, p. 20; 2007, pp. 48–54; 2007, §36, p.  189; Bonanno 1977, §15, p.  189; 2005, pp.  298–9 (35°56’0.08”N, 14°24’41.99”E, elevation 34 m). San Pawl Tat-Targa (limits of Naxxar, ‘Ta’ Fuq Wied Filep’):  dolmens; Ashby et  al. 1913, pp.  8–10, fig.  3; Evans 1971, pp.  196 and 231; Trump

Appendix B 2002, pp.  276 and 279; Sciberras 1999, p.  102; Cilia 2004, pp.  220–1; MEPA, gn781_11 (35°55’10.43”N; 14°25’53.57”E, elevation 84 m). Santa Lucija (Lellux street, about 800 m south of Ħal Saflieni; 1.1 km south-east of the Tarxien Neolithic complex): a small complex; ‘Copper Age’ finds; megalithic structure marking site; MAR 1973–4, §7, p. 51; Magro Conti ca. 1997 (35°51’43.71”N, 14°30’15.18”E, elevation 62 m). Santa Sfia: some remnant prehistoric large stones and a minor wall; excavated by M. Murray 1920; MAR 1921–2, p. 1. Skorba (Mgarr): settlement and megalithic structures; first listed as a menhir site; Għar Dalam, Grey Skorba, Red Skorba, Zebbug, Mgarr, Ggantija, Saflieni finds; reuse in the Bronze Age,Tarxien Cemetery, Borg in-Nadur phases; MAR 1914–15, p. 2; 1937–8, p. 2; 1961, pp. 1–4; 1962, pp. 3–5; 1963, pp. 2–3; Evans 1971, pp. 36 and 230;Trump 1966, 2002, pp. 156–9 and 251; Zammit and Mallia 2008; MEPA, gn764_98 (§2), plan (35°55’15.01”N, 14°22’39.95”E, elevation 116 m). St. Thomas Bay: house site; Bonanno 1977, §13, p.  76 (35°51’39.89”N, 14°34’30.95”E, elevation 3 m). Ta’ Baldu (Dingli): small bathhouse; Caruana 1882, §98, p. 92; K. Buhagiar 2000, pp. 41–51 (approx. 35°52’7.96”N, 14°22’1.89”E, elevation 178 m). Ta’ Berikka (Bir Ricca, or Biricca; in the field of ‘Tal-Hereb’, Zejtun, north of the Tas-Silg):  house architecture; water cistern; Punic and Roman finds; Zammit, notebook no. 7, pp. 59, 60 and 63; MAR 1923–4, p. 5; see also Buhagiar 1996, p. 5; 2007, §18, p. 189 (35°50’49.39”N, 14°33’05.40”E, elevation 41 m). Ta’ Bisqra (south-west slopes of the hill, west of the holiday village of Ta’ Bisqra): ‘large quantities’ of Ggantija, Pellegrin and a few ‘late’ Għar Dalam wares;Veen and van der Blom 1992, site 18, pp. 60 and 66 (approx. 35°57’52.79”N, 14°20’31.65”E, elevation 50 m). Ta’ b’Xejn (a.k.a. Ta’ Għammar, Zurrieq): standing stone and prehistoric pottery; destroyed; MAR 1913–14, p. 3; Evans 1971, pp. 199 and 232, fig. 1, ref. 553644 (35°49’17.09”N, 14°30’14.20”E, elevation 74 m). Tac-Cagħqi, Rabat: rock-cut pits; Phoenician-Punic tombs; later catacombs; Evans 1971, p. 200 (approx. 35°52’35.59”N, 14°23’51.18”E, elevation 212 m). Ta’ Cieda (northern location of San Gwann, Triq il-Korvu): round tower; Bonanno 2005, pp. 91 and 294–7;Trump 1997, pp. 91–2; MEPA, GN588_94 (§10), plan, p. 6815 (35°54’26.34”N, 14°28’48.41”E, elevation 75 m). Tad-Dawl (Ħal Kirkop/Mqabb):  rural house and olive oil works of Punic and later date; destroyed; Bonanno 1977, §8, p. 76 (35°51’2.50”N, 14°28’37.69”E, elevation 78 m). Ta’ Għewra: Gudja; dolmen site; Evans 1971, pp. 196 and 229 (35°49’57.70”N, 14°30’30.62”E, elevation 62 m).



Appendix B Ta’ Gawhar (Gudja-Kirkop):  round tower; rock-cut cistern; archaeologically investigated; Borg in-Nadur finds; MEPA_GN 1082_2009 (no. xii), plan 12/35; Bonanno 2005, pp. 91 and 294–7; Trump 1997, pp. 91–2 (35°49’58.06”N, 14°29’58.19”E, elevation 83 m) Ta’ Ħagrat (Mgarr):  some Early Neolithic finds; Red Skorba, Saflieni, Zebbug, Mgarr, Ggantija and Tarxien remains; megalithic building; also Tarxien Cemetery, Borg in-Nadur finds (one Baħrija pot sherd); Zammit 1929, pp. 5–25; MAR 1961, p. 4;Trump 1966, pp. 17–19, figs. 16 (plan) and 17:c, d, e, f (sections); 1953, pp. 47–8, pl. 8:16; cf. 1971, pp. 33 and 208–9; 2002, pp. 154–5; Zammit and Mallia 2008; Evans 1971, p. 33; Skeates 2010, p. 126 (35°55’6.70”N, 14°22’6.82”E, elevation 92 m). Ta’ Ħammut (or Qalliet Marku): vicinity of Magħtab (Naxxar Parish); site of three dolmen, 200 m inland, near Baħar ic-Cagħaq (Pebble Bay); MAR 1927–8; Evans 1956, pp. 86–8, fig. 1, pl. 7:1; 1971, p. 194; Cilia 2004, p. 201 (35°56’23.70”N, 14°26’51.00”E, elevation 14 m). Ta’ Ħlantun: see Tal-Liebru. Ta’ Kaccatura: Punic-Roman house; oil installations; well-preserved large cistern; Ashby 1915; Bonanno 1977, §10, p. 76; Buhagiar 2007, §16, p. 189 (35°50’8.71”N, 14°31’32.78”E, elevation 30 m). Tal-Bakkari: near Ħal Far; round tower, Punic and later usage; possible Bronze Age remains, Punic pottery and remnant architecture to the north near the late Antique early, Christian structures; bell-shaped pit; MEPA, gn588_94 (§8), plan, p.  6813; MAR 1921–2, pp.  3–4; Bonanno 2005, pp. 91 and 294–7; Trump 1997, pp. 91–2 (35°49’10.09”N, 14°29’45.75”E, elevation 92 m). Tal-Bidni (west of Marsascala, rocky area called ‘Xagħra tal-Bidni, Zabbar’):  megaliths; dolmen; post-Punic house; Evans 1971, pp.  197, 200 and 232; Buhagiar 2007, §24, p.  189; MAR 1914–15, p.  1; (megaliths: 35°52’3.30”N, 14°33’20.17”E; dolmen: 35°52’4.80”N, 14°33’19.02”E, elevation 43 m). Tal Brolli (Birzebbuga):  dolmen; Evans 1971, p.  229 (35°49’50.23”N, 14°31’22.33”E, elevation 36 m). Tal-Firminka (Safi): dolmen; (approx. 35°49’41.26”N, 14°29’9.94”E, elevation 103 m). Tal-Garda (Għaxaq): dolmen; Evans 1971, pp. 195–6 and 230 (35°50’22.51”N, 14°31’26.30”E, elevation 30 m). Tal Għassiewi: see Fiddien. Tal Ħlas road (district between Qormi and Zebbug, south of Wied is Sewda; in the field ‘Ta l’Isfar’): cistern near the chapel of the same name; Roman date; Zammit, notebook no. 4, p. 55; MAR 1913–14, p. 4; Bonanno 2005, p. 107 (35°53’1.43”N, 14°27’29.30”E, elevation 31 m). Tal-Ħorr (Gudja):  modified rock-cut pits, MAR 1924, p.  4; Evans 1971, p. 200 (35°51’1.19”N, 14°30’26.10”E, elevation 76 m).

Appendix B Ta’ Lippija (Mgarr):  possible megalithic building remains; Trump 2002, p.  167; Cilia, backboard, §46 (approx. 35°55’21.38”N, 14°20’48.83”E, elevation 85 m). Ta’ Lippija: rock-cut grape-pressing pans; Bonanno 1977, app.  2 (35°55’32.48”N, 14°21’20.23”E, elevation 92 m). Ta’ l-Itorbu (Għaxaq):  possible dolmens site; Cilia 2004, §15 backboard (approx. 35°50’19.57”N, 14°30’29.29”E, elevation 65). Tal-Liebru (a.k.a. Ta’ Ħlantun, Zurrieq Parish): dolmen site; MAR 1954–5, p. 1 (35°49’26.69”N, 14°29’23.46”E, elevation 108 m). Tal-Mejtin: Luqa; seventeen pits dated in the Early and Middle Bronze Age; MAR 1960, p.  5; Evans 1971, pp.  24 and 200 (35°51’17.11”N, 14°29’43.76”E, elevation 72 m). Ta-Qadi: megalithic buildings and settlement; menhir; Trump 2002, p. 160 (35°56’11.98”N, 14°25’13.92”E, elevation 17 m). Ta’ Qali (U.S. Embassy site): a megalith in the shaft of later Punic tomb; Bronze Age settlement site, seventeen bell-shaped pits and later Punic tombs; Micallef, in Times of Malta, 15 September 2007; other finds have been made at ‘tal Kali,’ Zammit 1911, p. 6 (35°53’26.68”N, 14°25’35.80”E, elevation 103 m). Ta’ Raddiena (off Għargħur Road, Birkirkara): curved wall probably from a lobed structure; Tarxien-period finds; megalithic stones can still be seen in a field wall beside the Birkirkara by-pass; MAR 1986, p. 68; also Trump 2002, pp. 168–9 (35°54’15.91”N, 14°27’56.67”E, elevation 68 m). Tarxien: cluster of prehistoric megalithic buildings; UNESCO World Heritage listed; Borg in-Nadur finds; Punic and Roman material was in the upper levels; Zammit, notebook no. 13 (28 June 1916), pp. 183 and 191–4; 1916, 1917, 1920, 1930b;Trump 2002, pp. 120–7; MAR 1958–9, p. 2, pl. 1:a–b; Bonanno 2007, pp. 109–10; Tanasi 2009, fig. 14 (35°52’09.15”N, 14°30’43.30”E, elevation 58 m). Ta’ Sinserna: cistern; post-Punic; Zammit, notebook no. 4, p. 77; Buhagiar 2007, §25, p. 189 (approx. 35°50’1.46”N, 14°31’2.90”E, elevation 36 m). Tas-Santi: south of L-Imgarr, remnant ashlar structure (possibly a tower); Bonanno 2005, pp. 91 and 294–7 (35°54’21.89”N, 14°21’22.91”E, elevation 135 m). Tas-Silg: megalithic complex; possible dolmen; Saflieni, Tarxien and other prehistoric finds; Tarxien Cemetery, Borg in-Nadur, Baħrija finds; Phoenician-Punic religious precinct; Rossignani, Airoldi and Grassi 2011; Bonanno and Vella 2015; Trump 2002, pp.  138–9; Cazzella and Recchia 2012, p. 84, fig. 6:3; Sagona 2015 (35°50’45.24”N, 14°33’07.40”E, elevation 44 m). Tas-Sittin: near Fawwara, Roman-period house site; Bonanno 1977, §6, p. 76; Buhagiar 2007, §9, p. 189 (35°50’31.65”N, 14°25’9.11”E, elevation 197 m).



Appendix B Tat-Tamla, Luqa:  megalithic site; Evans 1971, p.  230, fig.  1 (540681) (35°51’30.31”N, 14°29’1.07”E, elevation 79 m). Ta’ Trapna iz-Zgħira, (Zebbug):  tombs with Zebbug finds and one Mgarr pot; Baldacchino and Evans 1954, fig.  1; Evans 1971, pp.  166–9 (35°52’17.43”N, 14°25’33.38”E, elevation 121 m). Ta’ Vnezja: adult female and child burial with some ochre nodules, possibly Neolithic; Borg in-Nadur pottery in the museum stores from this location; Trump 2002, p.  262 (approx. 35°53’33.34”N, 14°24’44.21”E, elevation 128 m). Ta’ Wilga (Mqabba):  round tower, archaeologically investigated; MEPA, gn588_94 (§9), plan, p. 6814; Bonanno 2005, pp. 91 and 295; Trump 1997, pp. 91–2 (35°50’59.37”N, 14°27’34.70”E, elevation 83 m). Ta’ Zammitello (Mgarr): see Għajn Tuffieħa. Torri Falka (Mgarr): cave site; finds include lithics, spindle whorls and torba floor; destroyed by quarrying possibly in 1887;Tarxien,Tarxien Cemetery finds; MAR 1927–8, pp. 1–2; Evans 1971, pp. 39 and 231 (35°54’23.15”N, 14°24’3.11”E, elevation 139 m). Torri Falka (Falka Hill): site of a proposed quarry; surface concentration of pottery, Roman-period house remains, ashlar blocks, ancient quarry marks and stratified deposits; MAR 2002, p.  175; Superintendence of Cultural Heritage 2004, §5.5.6; 2005, §6.2, p.  14; Buhagiar 2007, §35, p. 189 (approx. 35°54’29.88”N, 14°24’00.17”E, elevation 145 m). Wardija Hill: house site; Bonanno 1977, §16, p.  76; Buhagiar 2007, §39, p. 189 (35°56’22.30”N, 14°24’17.68”E, elevation 56 m). Wardija Hill: house and olive oil installations; Bonanno 1977, §17, p.  76; Buhagiar 2007, §38, p.  189 (35°56’16.06”N, 14°24’14.10”E, elevation 55 m). Wardija Hill: house site, a Punic-Roman industrial site; Mariner et  al 2012, fig.  2A; Bonanno 1977, §18, p.  76; Buhagiar 2007, §37, p.  189 (35°56’30.34”N, 14°24’21.90”E, elevation 29 m). Wardija ta’San Gorg (limits of Siggiewi, Dingli Cliffs):  settlement, defence wall, rock-cut pits, house footings; wine-pressing vat; MAR 1960, p.  5; MAR 1972–3, §2, p.  72; Evans 1971, pp.  116 and 200; Bonanno 2008a; MEPA, gn588_94 (§11), marked on plan, p. 6816 (35°50’31.70”N, 14°23’56.12”E, elevation 222–151 m) Wied Filep (Naxxar, ‘Fuq Wied Filep’):  dolmens; see San Pawl Tat-Targa. Wied il-Għajn (Marsaskala):  menhir; MAR 1946–7, p.  6; Evans 1971, pp. 198 and 232; Cilia 2004, p. 209 (approx. 35°51’36.70”N, 14°34’10.50”E, elevation 24 m). Wied il-Għasel (vicinity of Fort Mosta): recorded dolmen site, destroyed (approx. 35°55’21.38”N, 14°25’35.83”E, elevation 81 m).

Appendix B Wied is-Sewda: cistern; MAR 1913–14, p. 4; Buhagiar 2007, §32, p. 189 (approx. 35°53’5.80”N, 14°27’19.74”E, elevation 37 m). Wied ix-Xagħra (terrace to the left at the head of the creek, between Għar il-Kbir and the locality of Girgenti); pottery of Skorba or Mgarr might indicate a Neolithic settlement; Fedele 1988, site 7.9.87/11, p.  196, fig.  2:11; (approx. 35°51’1.66”N, 14°24’0.13”E, elevation 219 m). Wied l-Mtaħleb (Rabat District): cave with Bronze Age remains; Evans 1971, p. 232; Mifsud 1995, p. 58 (approx. 35°52’37.83”N, 14°21’28.22”E, elevation 169 m). Wied Moqbol (to the east–south-east of Zurrieq, 600 m from Tal-Bakkari Church, Ħal Far Road):  cairn, stone circles; perhaps a classical or earlier site, large stone blocks reused in dry stone wall, found on proposed quarry site; MAR 2002, p. 175; Evans 1953, pp. 65 and 87–9, fig. 2; 1956, pp. 87–9, figs. 2 (plan), 3:4–6 (pottery finds);Trump 2002, pp. 261 and 279; Mifsud 1995, table 13; Tanasi 2008, fig. 14 (35°48’50.05”N, 14°29’6.36”E, elevation 56 m). Wied ta’ Manna: Mellieħa: cave site; Bronze Age finds; MAR 1922–3, §7, p. 4; Zammit, notebook no. 7, p. 16; Evans 1971, p. 230 (35°56’39.34”N, 14°22’20.00”E, elevation 36 m). Wied Znuber: dolmen; several megalithic circles; Murray 1923, p.  19 (approx. 35°48’41.76”N, 14°30’37.73”E, elevation 53 m). Xemxija: megalithic site; seven prehistoric tombs within an archaeological reserve, sites dated to the Mgarr, Ggantija and Tarxien phases; Borg in-Nadur Bronze Age pottery in the upper levels; Punic tombs; possible Roman road and latter historic remains; bathhouse; rock furrows; Evans 1971, pp. 112–16, plans 24–9 and 236–41; Trump 2002, pp. 162–5; Cilia 2004, pp. 164–5;Tanasi 2008, fig. 14 (35°56’58.74”N, 14°22’52.97”E, elevation 63 m). Xrobb il-Għagin: megalithic building eroded off a cliff edge; possibly of Tarxien date; Cilia 2004, pp.  99 and 102–4; Trump 2002, p.  189 (35°50’38.58”N, 14°34’5.16”E, elevation 23 m). Zebbieħ: menhir; see Skorba complex. Zejtun: Girls’ school site; three pits with Borg in-Nadur finds; Punic and Roman house; olive oil installations; MAR 1961, p.  5; 1972–3, p.  72; 1973–4, §4, p. 51; 1974–5, §3, p. 56; 1975–6, §2, p. 60; 1976–7, 2, p. 64; Abela 2012; Bonanno 1977, §12, p. 76; 2005, p. 304; Buhagiar 2007, §21, p. 189 (35°51’05.81”N, 14°32’08.39”E, elevation 66 m). Zurrieq (Triq il-Karmnu): extant structure on private property and length of substantial wall built into current house visible from the street; assigned a Phoenician-Punic date; Houel 1787, pp. 97–8; Buhagiar 2007, §30, p. 189 (35°49’48.77”N, 14°28’32.86”E, elevation 115 m).


Appendix B

322 GOZO

Borg il-Għarib (Għajnsielem, north of Xewkija to Qala road, Triq il-Qala):  remnant megaliths close to l-Imrejsbiet; MEPA, gn241_1997 (no. 21), map, p. 2432 (36°1’48.46”N, 14°17’4.30”E, elevation 92 m). Borg ta’ l-Imramma (Sannat):  some megalithic architectural elements; possible Għar Dalam finds; scattered Neolithic period remains; Veen and van der Blom 1992, p.  28; Cilia 2004, pp.  167 and 199, backboard, §8; Jaccarini and Cauchi 1999, p.  425; Trump 2002, p.  184 (36°1’10.47”N, 14°15’29.18”E, elevation 128 m). Dwejra (Fuq tal-Bniet, St. Lawrenz):  rock-cut basin, grape-pressing pan; MEPA, gn37_11 (§6), map 2:6; Jaccarini and Cauchi 1999, pp. 427–8 and 433; Bonanno 2008a (36°2’49.19”N, 14°11’56.83”E, elevation 79 m). Dwejra (Tal-Ibrag, St. Lawrenz):  rock-cut basin; MEPA, gn37_11 (§5), map 2:5; Jaccarini and Cauchi 1999, pp. 427–8 and 433; Bonanno 2008a (36°3’1.28”N, 14°11’38.62”E, elevation 42 m). Gebla ta’ Sansuna (Triq Gnien Imrik; Xagħra, Gozo, Xagħra): dolmen; Caruana 1896b, p. 142, pl. 2, fig. 2; Evans 1971, p. 193; Sciberras 1999, p. 105; MEPA, gn1082_2009 (§xviii), plan 18/35 (36°3’10.61”N, 14°15’37.35”E, elevation 110 m). Ggantija: museum site; UNESCO World Heritage listed; Zebbug, Mgarr, Ggantija, Tarxien, Tarxien Cemetery and a few Borg in-Nadur finds; Evans 1971, pp. 179–80 (36°02’50.13”N, 14°16’09.07”E, elevation 125 m). Għajn Damma, Xagħra: a number of prehistoric sites; megaliths, surface scatter of pottery; also a cluster of Xemxija-styled tombs; MAR 1995, p. 4 (approx. 36°3’20.43”N, 14°15’20.68”E, elevation 100 m). Għajnsielem (flanking Għajnsielem Road): house segment with Ggantija, Saflieni and Tarxien deposits; Borg in-Nadur sherd scatters and Roman finds; Malone et al. 1988, 2009, pp. 41–56;Tanasi 2008, fig. 14 (36°1’43.81”N, 14°16’48.07”E, elevation 98 m). Għajnsielem (on the north side of Triq L-Imgarr, Għajnsielem):  Borg in-Nadur pottery scatter; Malone et  al. 1988; Trump 2002, pp.  252–3; 1997, pp. 91–2 (36°1’44.86”N, 14°16’49.02”E, elevation 99 m). Għar Abdul (also known as Għar Il-Mixta, Kercem, west of Santa Lucija): cave sites around the cliff edges of the plateau, largely destroyed; Għar Dalam, Neolithic, Bronze Age (i.e., Borg in-Nadur pottery), Phoenician to Medieval and more recent historical finds; wells and cisterns; Bronze Age fortifications on the summit; some were destroyed and others are subject to erosion; MAR 1969, pp. 5–6; 1970, p. 6; Anati 1995, p. 105; Trump 2002, pp. 186–7 [approx. cave locations: (1) 36°2’52.06”N, 14°12’23.70”E; (2)  36°2’51.12”N, 14°12’23.92”E; (3)  36°2’51.00”N, 14°12’24.18”E; (4)  36°2’50.74”N, 14°12’24.19”E; (5)  36°2’50.57”N, 14°12’24.29”E; (6)  36°2’48.32”N, 14°12’25.95”E; (7)  36°2’47.62”N, 14°12’26.96”E, elevation 154 m].

Appendix B Għarb area, to the north on a hill slope: reputed Għar Dalam settlement, Skorba wares, Late Neolithic, Punic and Roman finds;Veen and van der Blom 1992, site 4, pp. 60 and 62 (approx. 36°4’17.49”N, 14°12’36.36”E, elevation c. 62). Għarb area (Triq Ta’ Maxwell, on south-west hill slope to the immediate north of road, Għarb area): reputed Għar Dalam–period sherds and flint material; Punic finds;Veen and van der Blom 1992, site 2 (listed as 124 m asl), p. 62 (approx. 36°3’41.06”N, 14°11’42.84”E, elevation about 118 m). Għar Il-Mixta: see Għar Abdul. Għar in-Nagħag (west of Santa Lucija): cave dwelling with stratified habitation deposits; Mgarr, Ggantija, Tarxien, Tarxien Cemetery (Thermi or Grey wares), and Punic and recent phases represented; MAR 1935–6, p. 19; Evans 1971, p. 21 (approx. 35°48’25.71”N, 14°31’39.09”E, elevation about 32 m). Għar ta’ Għejzu (Xagħra, Gozo, in field L’ Mdawra, vicinity of Ggantija flanking a plateau): cave site, a few sherds of Għar Dalam type; Ggantija finds and human remains; MAR 1933–4, pp. 6–7; Evans 1971, pp. 183–5 and 233; Cilia 2004, p.  183; MEPA, gn_357_98, plan; MEPA, gn588_94 (§2), plan, p. 6806; MEPA, gn853_10, plan, site 3;Veen and van der Blom 1992, site 11, p. 64 (36°2’50.90”N, 14°15’57.07”E, elevation 137 m). Id-Dura ix-xagħra il Kbira: dolmen; Ashby et al. 1913, pp. 10–11, fig. 4 (36°1’5.95”N; 14°15’32.35”E, elevation 120 m). Id-Dura tal Mara (Ta’ Cenc): numerous dolmen and prehistoric features in the region; Ashby et al. 1913, p. 10, fig. 3; Cilia et al. 2004, backboard, §9 (36°1’18.19”N, 14°15’19.32”E, elevation 128 m). Il-Hagra l-Wieqfa (Gebla l-Wieqfa Qala): (1) reputedly very eroded Għar Dalam sherds; (2) Triq il-Barbagann (a.k.a. Triq il-Tempiju); listed often as a menhir but is a Late Neolithic, sole megalithic structural remnant; Caruana 1896b, p. 142, pl. 2, fig. 1; Evans 1971, pp. 199 and 233; Trump 2002, p. 185; Cilia 2004, backboard (Gozo site 6); MEPA, gn290_98 (§3), plan 3;Veen and van der Blom 1992, site 13, pp. 60 and 64 (36°2’12.15”N, 14°18’32.19”E, elevation 121 m). Il-Hagra tad-Dawwara (Kercem Road):  possible prehistoric menhir with squared profile; Trump 2002, pp.  186–7 (pictured) (36°2’31.28”N, 14°13’48.09”E, elevation 90 m). In-Nahhalija: post-Punic house; Lewis 1977, p.  91; Buhagiar 2007, §1, p. 189 (35°56’21.96”N, 14°20’23.23”E, elevation 56 m). In-Nuffara: plateau-top settlement; one Tarxien Cemetery sherd, Borg in-Nadur pottery fragments and about twenty bell-shaped pits; MAR 1960, p. 5, fig. 1; Evans 1971, pp. 171 and 200; Trump 2002, p. 252; Tanasi 2008, fig.  14; (in press a), p.  7 (36°2’30.57”N, 14°16’24.27”E, elevation 133 m).



Appendix B It-Tokk (Victoria):  Borg in-Nadur and Roman-period finds; MAR 1961, p. 5; Evans 1971, p. 200; cf. Bonanno 2005, p. 213 (36°2’41.69”N, 14°14’20.68”E, elevation 100 m). Ix-xagħra il Kbira (Ta’ Cenc region): destroyed megalithic structure; possibly Bronze Age; Cilia et al. 2004, §7 (approx. 36°1’15.09”N, 14°15’33.38”E, elevation 120 m). Ix-xagħra il Kbira: dolmen; Ashby et al. 1913, pp. 10–11, fig. 4; Cilia et al. 2004, backboard, §10 (36°1’5.95”N; 14°15’32.35”E, elevation 120 m). Ix-Xaqqufja (near Birbuba area, on road from San Lawrenz and 1½ km from church):  reputedly a small scatter of Għar Dalam sherds; Late Neolithic and quantities of Punic pottery in the area; Roman house remains; Veen and van der Blom 1992, site 1, pp.  60 and 62; Buhagiar 1997, p.  116; 2007, §44, p.  189 (approx. 36°3’58.15”N, 14°11’34.36”E, elevation105 m). Kercem: two Neolithic round tombs found in 2009; on the Parish Hall building site; skeletons, tools, pottery were found; Superintendence of Cultural Heritage 2008, §6.2; notice in Times of Malta, 10 October 2010, 11 October 2009, 18 March 2010 (36°2’31.55”N, 14°13’39.10”E, elevation 108 m). L-Imrejsbiet (Għajnsielem, south of Xewkija to Qala Road):  megaliths, 403 m north-east of an excavated prehistoric house; MEPA, gn_241_1997 (no. 21), map, p. 2432; Trump 2002, p. 185 (36°1’46.85”N, 14°17’3.70”E, elevation 91 m). Maghqad ix-Xih (Xewkija): a dolmen was located on the site of the present Church of Xewkija; Jaccarini and Cauchi 1999, p.  425; after Agius de Soldanis, unpublished manuscript 1750, p. 82 (approx. 36°1’54.15”N, 14°15’39.63”E, elevation 113 m). Marsalforn: post-Punic house; Bonanno 1977, §22, p.  76; M.  Buhagiar 1997, p. 116 (approx. 36°4’17.96”N, 14°15’42.62”E, elevation 24 m). Mgarr ix-Xini, Xewkija:  nine grape-pressing pans bordering a gully; Jaccarini and Cauchi 1999,site 1,Il-Gandotta (36°1’33.71”N,14°15’48.54”E, elevation 68 m), site 2, Tal-Salvatur (36°1’25.16”N, 14°15’46.28”E, elevation 49 m), site 3, Tal-Salvatur (36°1’25.09”N, 14°15’48.23”E, elevation 48 m), site 4, Tal-Gruwa (36°1’31.42”N, 14°15’41.53”E, elevation 59 m), site 5, Tal-Gruwa (36°1’28.62”N, 14°15’45.78”E, elevation 54 m), site 6, Tas-Sabbora (Is-Sabbara) (36°1’28.13”N, 14°15’44.86”E, elevation 53 m), site 7, Il-Gandotta (36°1’32.92”N, 14°15’46.67”E, elevation 62 m), site 8, Il-Gandotta (36°1’32.97”N, 14°15’47.05”E, elevation 63 m), site 9, Tal-Gruwa (36°1’31.68”N, 14°15’39.18”E, elevation 60 m). Nadur (outskirts, Kenuna Plateau, west of historic tower): reputedly a single Għar Dalam sherd, unspecified Skorba finds; Veen and van der Blom 1992, site 12, pp. 60 and 64 (approx. 36°2’12.60”N, 14°16’58.83”E, elevation 133 m).

Appendix B Qala, Gebla l-Wieqfa: see ll-Ħagra il-Wieqfa. Qala, north of il-Wileg: said to have Neolithic remains, including a few Għar Dalam sherds and Zebbug finds; Veen and van der Blom 1992, site 14, pp. 60 and 64 (approx. 36°2’21.05”N, 14°19’12.25”E, elevation 126 m). Qala (east of the Immaculate Conception National Sanctuary Church): reputedly the site of a Neolithic megalithic building;Veen and van der Blom 1992, site 15, pp. 60 and 64 (36°2’4.50”N, 14°19’20.56”E, elevation 108 m). Qortin ta’ Ghajn Damma: Borg in-Nadur finds; Tanasi 2008, fig.  14 (36°4’12.54”N, 14°15’55.94”E, elevation 95 m). Ramla Bay (Ramla l-Hamra): large house and elaborate bathhouse facilities; submerged causeway; MAR 1910–11, pp. 11–12; Ashby 1915, pp. 70–4; Bonanno 1977, §21, p. 76; 1980; 2005, pp. 235, 342 and 343; Buhagiar 2007, §43, p. 189 (30°03’40.35”N, 14°16’56.54”E, elevation 8 m). Ras il-Wardija: open-air Phoenician-Punic structure (temple?) with altar; rock-cut chamber probably a Mithraeum; Sagona 2009; Trump 2002, p. 252 (36°2’10.85”N, 14°11’11.32”E, elevation 28 m). Santa Verna (ix-Xagħra District): scant signs of an early Neolithic presence under the threshold of a Late Neolithic megalithic structure; Red Skorba, Saflieni, Zebbug, and Grey Skorba remains; Ashby et al. 1913, pp. 105–23; MAR 1961, p. 5; Trump 1966, pp. 19–20, fig. 17:b; 1997, pp. 160–1; 2002, p. 182; Evans 1971, p. 189; cf. 1953, pp. 47–8, pl. 8:14–15; MEPA, gn_357_98 (§4), plan; MEPA, gn_588_94 (§3), plan, p. 6808; Cilia 2004, pp. 196–7 (36°2’44.64”N, 14°15’31.03”E, elevation 142 m). Ta’ Blankas (Xewjika):  dolmen and menhir site, destroyed; Cilia 2004, backboard, §26 (approx. 36°1’43.92”N, 14°15’19.31”E, elevation 84 m). Tac-Cawla: suburban Victoria; settlement with Għar Dalam, Zebbug, and Late Neolithic finds;Veen and van der Blom 1992, pp. 19–30 and 63; Cilia 2004, pp. 167 and 203; Malone et al. 2009, pp. 54–5, table 3:3;Trump 2002, p. 186 (approx. 36°2’10.43”N, 14°14’33.60”E, elevation 95 m). Ta’ Kuljat: Early and Late Neolithic sherd scatters, Bronze Age settlement defence walls; Veen and van der Blom 1992, pp. 60 and 62, site 5; Trump 2002, p. 28; photographs of the Ta’ Kuljat defence wall can be see at the Megalithic Portal website: .php?sid=10383 (36°3’30.32”N, 14°14’26.69”E, elevation 162m). Tal Ħamrija (Xewkija District): MAR 1906, §12, p. E3; a stone olive press was taken to the Gozo Library. Caruana (1899, p. 223) also mentions oil works in this region (approx. 36°2’1.82”N, 14°15’33.44”E, elevation 98 m). Tal-Knisja: see Mgarr ix-Xini; rock-cut grape pressing pan, excavated; Superintendence of Cultural Heritage 2008, §6.1 (approx. 36°1’34.86”N, 14°15’47.73”E, elevation 70 m). Ta’ Marziena megalithic buildings on the Munxar-Sannat road: Ggantija pottery finds; MEPA, gn290_98 (§1), plan 1; Jaccarini and Cauchi (1999),



Appendix B p.  425; Trump 2002, p.  183; Cilia 2004, pp.  167 and 196 (36°2’0.09”N, 14°14’24.30”E, elevation 96). Ta’ Pergla iz-Zgħira (or Tal Marri, Xagħra):  cave habitation site with Ggantija and Tarxien period finds; MAR 1913–14, p.  2; Evans 1971, p.  186; Trump 2002, p.  185 (approx. 36°3’42.00”N, 14°16’16.52”E, elevation 76 m). Tas-Salvatur (a district called ‘tal-Ħorob’ or Herheb, Xewkija District): (1) Roman wall, numerous pottery finds; MAR 1950–1, p. 18; (2) (in a field wall) olive pipper now in the Gozo Museum; 1958–9, p. 4; M. Buhagiar 1997, p. 116 (approx. 36°1’59.35”N, 14°16’14.40”E, elevation 103 m). Tas-Srug (Is-Srug; lands called,‘Il Qortin tas-Sruc’, Xagħra): rock shelter and surrounding area; Early and Late Neolithic (later ‘Copper Age’; Ggantija phase) pottery; possible hut remains; adjoining fields have Zebbug-phase pottery; MAR 1938, §4–5, p. 9; 2002, p. 176; Evans 1971, pp. 185 and 233; Veen and van der Blom 1992, site 7, pp. 60 and 63; Malone et al. 2009, pp. 54–5, table 3:3 (approx. 36°3’24.02”N, 14°15’18,62”E, elevation 91 m). Ta’ Tingħi: rock-cut pit; tentative date the Bronze Age; Jaccarini and Cauchi 1999, p. 427 (36°1’39.94”N, 14°15’25.55”E, elevation 76 m). Victoria (west side of St. George’s Church): cylindrical cisterns, late Roman period; MAR 1936–7, pp. 14–15 (36°2’37.62”N, 14°14’21.15”E, elevation 108 m). Victoria, Triq Palma (Franciscan Sisters, Heart of Jesus, House of Charity): a barrel-shaped cistern; Bonanno 2005, pp. 212–13 (36°2’34.30”N, 14°14’24.78”E, elevation 105 m). Xagħra Circle (Brochtorff Circle):  hypogeum; stone circle; Zebbug, Mgarr, Ggantija, Saflieni, Tarxien and Tarxien Cemetery, Borg in-Nadur finds; Malone et al. 2009, §6:10, pp. 88–9 and 93; Trump 2002, pp. 176–81 (36°02’47.03”N, 14°15’54.11”E, elevation 137 m). Xagħra,Triq Ta’ Ħamet: Borg in-Nadur finds;Tanasi 2008, fig. 14 (approx. 36°2’35.37”N, 14°16’2.16”E, elevation 93 m). Xewkija: open settlement of Għar Dalam phase; Evans 1971, pp.  191–2; Jaccarini and Cauchi 1999, pp. 425 and 428; Mallia 1985, pp. 156–8;Trump 2002, p. 186 (approx. 36°1’53.98”N, 14°15’43.66”E, elevation 112 m). Xewkija: megalithic building; floor A, destroyed; Zebbug and Ggantija deposits (36°1’53.60”N, 14°15’42.08”E, elevation 109 m); (3)  floor B, destroyed (36°1’53.98”N, 14°15’43.66”E, elevation 112 m). Xewkija: Roman-period house and olive oil installations; Bonanno 1977, §20, p.  76; M.  Buhagiar 2007, §46, p.  189 (approx. 36°1’49.35”N, 14°15’31.69”E, elevation 100 m). Xewkija: circular grain pit plastered with clay and lined with pottery; used as a Roman-period pottery ‘dump’; MAR 1977–8, §5–6, p. 63 (general area: 36°2’1.82”N, 14°15’33.44”E, elevation 97 m).

Appendix B COMINO

(1) Western central zone, Borg in-Nadur Bronze Age and Roman-Byzantine sherd finds; house remains; Fedele 1988, p. 193, fig. 1, site 5.9.87/4. (2) House remains, MAR 1931–2; Buhagiar 2007, §47, p. 189. (3) Tomb sites, Punic and later; MAR 1912–14, p. 4. INDEX OF FEATURES AND CULTURAL PHASES AT SITES

List 1: Għar Dalam Sites Malta: Birkirkara; Borg in-Nadur; Għar Dalam; Il-Latmija; Kordin III; Skorba; Ta’ Ħagrat Gozo: Borg ta’ l-Imramma; Għar Abdul; Santa Verna; Tac-Cawla; Ta’ Kuljat; Tas-Srug; Xewkija Unverified attributions of sites to the Għar Dalam Phase: (Malta) Marfa;Ta’ Bisqra; (Gozo) Għarb area; Għar ta’ Għejzu; Il-Hagra l-Wieqfa; Ix-Xaqqufja; Nadur

List 2: Grey Skorba Sites Malta: Għar Dalam; Skorba Gozo: Santa Verna

List 3: Red Skorba Sites Malta: Skorba; Ta’ Ħagrat; Wied ix-Xagħra Gozo: Santa Verna

List 4: Zebbug Sites Malta: Buqana; Għar Dalam; Ħal Saflieni; Ras il-Pellegrin; San Pawl Milqi; Skorba; Ta’ Trapna iz-Zgħira Gozo: Ggantija; Tas-Srug; Xagħra Circle; Xewkija

List 5: Mgarr Sites Malta: Għar Dalam; Għar in-Nagħag; Għajn Tuffieħa; Kordin I, II and III; Skorba; Ta’ Ħagrat; Ta’ Trapna; Wied ix-Xagħra Gozo: Ggantija; Xagħra Circle

List 6: Ggantija Sites Malta: Bingemma; Għajn Tuffieħa; Għar Dalam; Għar in-Nagħag; Ħagar Qim; Ħal Ginwi; Kordin III; Ras il-Pellegrin; Skorba; Xemxija


Appendix B


Gozo: Ggantija; Għajnsielem; Għar ta’ Għejzu; Ta’ Marziena; Ta’ Pergla iz-Zgħira; Tas-Srug; Xagħra Circle

List 7: Saflieni Sites Malta: Skorba; Tas-Silg Gozo: Għajnsielem; Santa Verna; Xagħra Circle

List 8: Tarxien Sites Malta: Bugibba; Għajn Tuffieħa; Għar Dalam; Għar in-Nagħag; Ħal Ginwi; Ħal Saflieni; Kordin III; Kuncizzjoni; Mnajdra;Tarxien;Torri Falka; Xrobb il-Għagin; Xemxija Gozo: Ggantija; Għajnsielem; Ta’ Pergla iz-Zgħira; Tac-Cawla; Xagħra Circle

List 9: Megalithic Structures Malta: Borg in-Nadur; Bugibba; Għajn Zejtuna; Ħagar Qim; Ħal Far; Ħal Ginwi; Id-Debdieba; Il Ħofra; It-Tumbata; Kordin I; Kordin II; Kordin III; Kuncizzjoni; L-Iklin; Mnajdra; Mosta; Qajjenza; Qaliet Marku; Qallilija; Ras il-Gebel; Ras il-Pellegrin; Ras ir-Raħeb; Skorba; Sqaq il-Bagħal; Ta’ Ħagrat; Tal-Bidni; Ta’ Lippija; Ta-Qadi; Ta’ Raddiena; Tarxien; Tas-Silg; Tat-Tamla; Xemxija; Xrobb il-Għagin Gozo: Borg il-Għarib; Borg ta’ l-Imgramma; Ggantija; L-Imrejsbiet; Ix-xagħra il Kbira; Qala, Gebla l-Wieqfa; Santa Verna; Ta’ Marziena; Xewkija

List 10: Sites with Tarxien Cemetery Material Malta: Borg in-Nadur; Bur Mgħez; Għajn Tuffieħa; Għar Dalam; Għar in-Nagħag; Għar Mirdum; Ħal Ginwi; Ħal Saflieni Hypogeum; Skorba; Ta’ Ħagrat; Tas-Silg; Torri Falka Gozo: Ggantija; In-Nuffara; Xagħra Circle

List 11: Dolmen and Cairn Sites Malta: Birzebbugia; Ħal Far; Ħal Farrug; Misraħ Sinjura; Nadur; Qallilija; San Pawl Tat-Targa; Ta’ Għerwa; Ta’ Ħammut; Tal-Bidni; Tal Brolli; Tal-Firminka;Tal-Garda;Ta’ l-Itorbu;Tal-Liebru;Tas-Silg;Wied il-Għasel; Wied Moqbol; Wied Znuber Gozo: Gebla ta’ Sansuna; Id-Dura tal Mara; Ix-xagħra il Kbira; Maghqad ix-Xih

Appendix B

List 12: Stone Circles Malta: Għajn Tuffieħa; Mellieħa Ridge; Ras il Gebel; Ta’ Zammitello; Wied Moqbol; Wied Znuber Gozo: Marfa; Xagħra Circle

List 13: Menhir Sites Malta: Ħal Far; Ħal Kirkop; Il Ħofra; Mnajdra; Qallilija; Ta-Qadi; Ta’ b’Xejn (Ta’ Għammar); Ta’ Qali; Wied il-Għajn; Zebbieħ Gozo: Il-Hagra l’Wieqfa; Il-Hagra tad-Dawwara; Ta’ Blankas

List 14: Borg in-Nadur Sites: Fortified Settlements (Including Naturally Strategic Sites) Malta: Baħrija; Borg in-Nadur; Gebel Għawzara; Għar Abdul; Għar in-Nagħag; Qolla; Il-Qortin; Manikata; Mdina; Mellieħa; Mtarfa; Qala Hill; Qarraba; Qolla; Rabat-Mdina; Ras il-Gebel; Wardija ta’ San Gorg; Wardija ta’ San Gorg region Gozo: Għajnsielem; In-Nuffara; Ras il-Wardija; Ta’ Kuljat

List 15: Borg in-Nadur Sites: Non-Fortified Bronze Age Settlements or Minor Finds Malta: Il-Qliegħa tal-Baħrija; Bidnija; Bingemma; Ħal Millieri; Kavallerizza; Qajjenza; Qolla; Qallilija; San Pawl Milqi; Skorba;Ta’ Gawhar;Ta’ Qali;Ta’ Vnezja; Tarxien; Tas-Silg; Zejtun; Wied Moqbol; Xemxija Gozo: Għajnsielem; In-Nuffara; It-Tokk; Qortin ta’ Ghajn Damma; Għajnsielem; Triq Hamet; Xagħra Circle Comino: Western central zone

List 16: Caves Used in the Bronze Age Malta: Baħrija; Għar Dalam; Għar tal-Friefet; Għar ir-Riħ; Għar Mirdum; Ħal Saflieni; Wied l-Mtaħleb; Wied ta’ Manna Gozo: Il-Mixta

List 17: Rock-Cut Pit (Bell-Shaped or Silo Pit) Locations Malta: Attard; Il-Qliegħa tal-Baħrija; Bir Miftuh; Birzebbuga; Id-Debdieba; Il-Qolla; Il-Qortin; It-Tumbata; Luqa Cemetery; Mdina; Mtarfa; Qallilija; St. George’s Bay; St. Marija tal-Virtu; Tac-Cagħqi; Tal Ħorr; Ta’ Qali; Tal-Mejtin; Wardija ta’ San Gorg; Zejtun Gozo: In-Nuffara; Ta’ Tingħi


Appendix B


List 18: Sites with Baħrija Pottery Malta: Baħrija; Borg in-Nadur; Għar Dalam; Qallilija; Ta’ Ħagrat; Tas-Silg

List 19: Fortified Towers Malta: It-Torrijiet; Ta’ Gawhar; Tal-Bakkari; Tas-Santi; Ta’ Wilga; Ta’ Cieda

List 20: Grape-Pressing Pans Malta: Borg in-Nadur; It-Tafal ta’ Bingemma; Laferla Cross; Ta’ Lippija; Wardija ta’ San Gorg Gozo: Dwejra (Fuq tal-Bniet, St. Lawrenz); Dwejra (Tal-Ibrag, St. Lawrenz); Mgarr ix-Xini; Tal-Knisja

List 21: Punic- and Roman-Period Houses Benghisa; Bezbezija; Burmarrad; Falka Hill; Fiddien; Gnien is-Sultan; Ħal Dragu; Ħal Ginwi; Ħal Millieri; Il-Bezbezija; Il-Hazzien; L-Iklin; Marsaxlokk; Misraħ Miel; Ras il-Pellegrin; Ras ir-Raħeb; Safi; St. Pawl Milqi; St. Thomas Bay; Ta’ Berikka; Tal-Bidni; Ta’ Kaccatura; Tarxien; Tas-Sittin; Wardija Hill; Wied Moqbol; Zejtun

List 22: Olive Oil Processing Sites Malta: Bidnija; Birgu; Ħal Far; Ħal Għargħur; Ħal Gilwi; Ħal Millieri; L-Imselliet; Misraħ Miel; Ras ir-Raħeb; St. Pawl Milqi;Tad-Dawl;Wardija Hill; Zejtun Gozo:– Tal Ħamrija; Tas-Salvatur; Xewkija; Zejtun

List 23: Rock-Cut Cisterns Malta: Attard; Ta’ Berikka; Bur Grad or Ta’ Gernac; Burmarrad and Wardja road; Ħal Far; Ħal Kbir; Has Saptan; Qrendi; Marnisi; Pawla; Rabat; Ta’ Kaccatura; Tal Ħlas road; Ta’ Sinserna; Mdina; Rabat; Tas-Silg; Wied is-Sewda Gozo: Għadira ta’ San Rafflu; Triq Palma; Santa Lucija;Victoria; Xewkija

List 24: Roman-Period Bathhouses Malta: Bidnija; Deyr Handul; Għajn Ħammiem; Għajn Qajjiet; Għajn Tuffieħa; Kordin; Marsa; Marsaxlokk; Mdina; Rabat; Ta’ Baldu; Xemxija Gozo: Ramla Bay


Chapter 1  Malta’s Archaeological Past 1 ‘Malta’ is used in a general sense for the Maltese Archipelago, as it is in the international arena. When specific islands within the group are referred to, this will be made clear in the text. 2 Hoare 1819, pp. 492–3. 3 For early geological accounts and research, Zammit Maempel 1989; Fenech 2007, p. 6. 4 Knapp 2007, pp. 45–8. 5 Leighton 1999, p. 14; Hunt and Schembri 1999, pp. 41–75, for the earlier, Quaternary environment. 6 Leighton 1999, pp. 14–15. 7 Schembri 1997; Fenech 2007, pp. 18–20. 8 For the qualities of local Maltese chert, C. Vella 2008b, §3:3. 9 Fenech 2007, pp.  6–10 and 18–20. Of more recent tectonic movements suggested for the southern sector of Malta, an often-cited indicator is human manufactures, the so-called cart ruts that run into the sea at a few points around St. George’s Bay, Birzebbuga (e.g., Fenech 2007, p.  10; Weston 2010, §50, p. 108; Magro Conti and Saliba 2007, pp. 88–9).These are interpreted as having sunk below sea level due to subsidence of the bedrock, but the fact that the standing stones of the prehistoric monument at Borg in-Nadur, a mere approximately 172 m to the west, show little signs of titling that should theoretically match the slope of the ruts, calls the theory into question. Such notions were current in Adams’ time; see Adams 1870, pp. 249–50.

10 Fenech 2007, p. 22. 11 Ibid., pp. 16–18. 12 Burmarrad cores, Djamali et  al. 2013, pp. 367–80; Marriner et al. 2012, pp. 1–14; Marsa cores, Fenech 2007, pp. 6–7; Marsa and Salina Bay in Malta, Santa Marija Bay in Comino, Carroll et al. 2012, pp. 24–40. Before the recent pollen analyses from the greater Mediterranean region and from the Maltese Archipelago, the main source of environmental data was the meagre pollen evidence from a bell-shaped pit cut in the Bronze Age at Tal Mejtin (Luqa District), which first indicated that by this time (third to second millennium BC), trees had largely disappeared from the landscape; MAR 1959–60, p.  3; 1960, p. 5; 1961, pp. 5 and 8; Trump 1961, p. 257; 1966, p. 51, n. 5; Evans 1971, pp. 24 and 226. Environmental studies concerning other archaeological projects: Xagħra Circle, Gozo, Schembri et  al. 2009, pp.  17–39; Tas-Silg (south), Fenech and Schembri (2015)]; Hunt 2000, pp. 111–14; Hunt (2014)]; Tas-Silg (north), Fiorentino et  al. 2012. Regarding field systems in the Mistra Valley, Hunt and Vella 2004–5, pp. 57–65. Schembri (1997) gives a concise summary of the current climatic and environmental conditions in Malta. 13 Fenech 2007, pp. 6–8. 14 Tanasi (in press b) fig.  16:c, depicting a miniature vessel of unfired clay from Għar Mirdum; clay fill levels and wall material at Skorba appeared to comprise imported marl,Trump 1966, pp. 13–14; see Chapter 2.


332 15 K. Buhagiar 2007, pp.  103–31; see especially fig.  12, depicting strings of water galleries along the valley slopes of Wied ir-Rum, Dingli; Fenech 2007, pp.  26–8; Zammit 1924b, pp. 1–46, concerning historical water supplies and springs in Malta. 16 Leighton 1989, p. 187. 17 Bonici Cali 1961, pp.  76–7; Freller 2008, p. 84. 18 H. C. R.Vella 1980a, 1980b. It also should be noted that Malta featured in many early maps, Gambin 2008. 19 Biographies of Abela can be found in, for instance, Bonnici Cali 1953; Maltese Historical Society 1961, collected essays. 20 The site of Abela’s house, essentially the first private museum in Malta, was clearly marked in Barbaro’s plan of Marsa (1794, pl. 1); the villa was depicted in the foreground of a lithograph illustrating the Grand Harbour by Pierre Gabriel Berthault (dated ca. 1740), a copy of which is held in the British Museum (no. 1917,1208.4248); Bonnici Cali (1961, pp.  76–81) encapsulates the written accounts of the museum as it appeared in Abela’s time. 21 K. Gambin 2003, p.  9; Cutajar 1995, pp.  67–8; MAR 1904, pp.  1–2; Caruana 1899, pp. 1–2. 22 Leighton 1989, pp. 183–4 and 188. Biblical references to post-creation giants appear in Genesis 6:4 and Numbers 13:33, which mention the Nephilim, an obscure term generally taken to mean powerful people of abnormal stature and of part-divine, part-human descent. 23 Zammit Maempel 1989, pp.  23–94, concerning Thomas Abel Brimage Spratt, 1811–88); pp.  171–234, for Andrew Leith Adams; with comprehensive listings of their publications. 24 Bradford 2002, pp.  186–9; Blouet 1967, pp. 159–68; Freller 2010. 25 Bugeja and Sagona (in preparation). 26 For a biography of Antonio A.  Caruana, see Caruana Galizia 1997. 27 N. C. Vella and Gilkes 2001, p. 355, after sentiments expressed in Caruana 1881, p.  23, dispersal of artefacts offshore and in private hands; 1882, p.  10, concerning flaws in J. G. Vance’s 1839 work at Ħagar Qim, published in 1842; p. 18, neglect of monuments under the Order of St. John;

Notes to Pages 7–11 p. 90, archaeological finds kept in private collections; p. 168, benefits of ‘merging’ of private collections into a single location; 1898, p. 4,‘spoliation’ of ancient sites; 1899, pp.  1–2, lamenting objects of national interest removed from Malta. 28 Sagona 2002, pp. 7–9. 29 Grima 2011, pp. 342–3. 30 Concerning the biblical bias in Caruana’s historical interpretation, Vella and Gilkes 2001, pp. 354–6; cf. Caruana Galizia 1997, pp. 14–27. 31 Rhind 1856, pp.  397–400; A.  J. Evans (1902, pp.  41–4), summarises the flawed arguments linking the Phoenicians to the megalithic structures. 32 Notions of cultural diffusion westward from the Aegean:  A.  J. Evans 1901, §29, pp. 196–200; Anon 1922, p. 27; Mayr 1901, pp. 705–9. Stöger (2000, pp. 3–9) provides a very useful survey of Albert Mayr’s publications, which concern many aspects of Maltese archaeology  – virtually an untapped field in his day – from Christian and Phoenician burial grounds to numismatics and the prehistoric monuments. Prevailing diffusionist theory dominating early twentieth-century histories in Malta, N. C. Vella and Gilkes 2001, p. 362; such concepts endure in later works, e.g., Evans 1959, pp. 138 and 162–4; Mayr’s diffusionist views applied to the entire western Mediterranean were soon under attack in the absence of evidence, Peet 1910, p. 142. Concerning Mayr’s insights regarding the prehistoric site of Borg in-Nadur, Bugeja 2011, p. 34. 33 N. C. Vella and Gilkes 2001, pp. 356–7; at the heart of the nationalistic debate was the nature of the Maltese language in education. 34 G. Clark 1989, p.  133:  ‘Apart from Gibraltar, the scene of Dorothy Garrod’s first independent excavation, the only chance for Cambridge prehistorians to dig in Europe on British soil was offered by the Maltese islands.’ 35 Ugolini was assisted in his efforts by Italian architect Calro Ceschi and artist Igino Epicoco. Maltese expertise came from La Ferla (engineer), Arturo Zammit and F.  Doublet (draughtsmen), as well as A.  Vassallo (architect); see M.  Tabone in

Notes to Pages 11–15 Pessina and Vella 2005, p. 7, and p. 13 in the same volume. 36 The archive is currently being prepared for publication by Andrea Pessina (formerly at the Luigi Pigorini Museum) and Nicholas Vella (University of Malta), Pessina and Vella 2005, p. 9. One stage in the publication of Ugolini’s archive is a new edition of Ugolini’s volume, Malta:  Origini della Civiltà Mediterranea (2012, with an English translation and introductory essays concerning Ugolini’s life and work by Pessina and Vella). See also N. C. Vella and Gilkes 2001, pp. 368–72. 37 Becker 2009; concerning Erich Becker, Fenech 2001, pp. 8–15. 38 T. Zammit was museum director from 1903 to his death in 1935; N. C. Vella and Gilkes 2001, p. 359. 39 MAR 1935–6, pp. 1–2. 40 Murray 1963, pp.  129–34; for her publications on Malta, Murray 1923a, b, 1925, 1928a, b, 1929, 1934. On Murray’s life:  Drower 2006a, pp.  109–41; Whitehouse 2013, pp.  120–7; Sheppard 2013, pp. 197–222. 41 Gertrude Caton-Thompson in Malta, Drower 2006a, p. 122; 2006b, p. 356. 42 Ellul-Micallef 2013, vol. 1, pp. 245–7. Other accounts of Zammit’s achievements, N. C. Vella and Gilkes 2001, pp. 358–60. 43 Among those who stepped into work with Zammit or on his behalf were Lewis Upton Way and Harris Dunscombe Colt; N. C.Vella and Chapman 2001, p. 52. 44 Zammit in MAR 1904, p.  3, concerning the establishment of the museum:  ‘The importance of the antiquities of Malta is recognized by archaeologists and a great service will be done to the island if the relics of old times are better taken care of, for the future.’ 45 Pessina and Vella 2005, p.  19; N.  C. Vella and Gilkes 2001, p. 358. 46 Concerning Napoleon Tagliaferro, Zahra 1980, p.  721. Paul Bellanti’s research focussed largely on the Phoenician-Punic tombs in Malta, Bellanti 1913, pre-1924, 1924. Zammit and Bellanti did not always agree and appear to have had an ambivalent workplace association, Ellul-Micallef 2013, vol. 1, pp.  270–2; vol. 2, pp.  194–6. Concerning the Bellanti family,

333 W.  Zammit 2011. Regarding Giuseppe Despott, Curator of Natural History (1922–33), MAR 1932–3, p. 1; K. Gambin 2003, p. 42. 47 N. C. Vella et al. 2011, p. 258; Clark 1989, pp. 132–3. 48 Baldacchino in MAR 1946–7, p. 1. 49 MAR 1955–6, p.  1; Bugeja 2006, pp. 23–45, on Baldacchino’s life and career; K. Gambin 2003, pp. 41–2. 50 MAR 1938–9, pp.  12–13; 1946–7, p.  1; Bonnici Cali 1961; K.  Gambin 2003, pp. 26–9. 51 Evans 1971, p.  v; N.  C. Vella and Gilkes 2001, p. 372; Skeates 2012. 52 MAR 1953–4, p.  2. The development of the prehistoric sequence, its terminology and numerical codes can be found in MAR 1959–60, p. 1; Trump 1961, 1966; Evans 1953, 1959, 1971. While much has been made of the political background, nationalism and steps towards Maltese independence, the foreword to Evans’ 1971 book clearly states the significant role the Maltese archaeologists and scholars from other disciplines  – a veritable Maltese academic ‘who’s who’  – had in the process; Evans 1971, pp. v–vi; cf. N. C. Vella and Gilkes 2001, pp.  372–5; N.  C. Vella et al. 2011, pp. 257–8. 53 As we will see in the chapters that follow, prehistoric megalithic remains at Tas-Silg broadened considerably the scope of research at this site. 54 N. C.  Vella and Gilkes 2001, pp.  372–4. Though substantial preliminary reports appeared soon after each season, a final statement has yet to appear concerning the excavations. 55 Vidal González 1996; Said-Zammit 1997; Sagona 2002; M. Buhagiar 1986 and 2007, for the Roman and Byzantine funerary remains. 56 N. C. Vella et al. 2011, p. 258. 57 Renfrew 1970, 1973, pp. 103–8, concerning the crumbling foundations of diffusionist chronological interpretation severed by what he termed the chronological ‘fault line’. 58 Renfrew 1973, p. 106. 59 Ibid., p. 109. 60 Evans 1973b, pp. 517–20. 61 Renfrew 1973, pp. 147–66.

334 62 Evans 1961, p. 143;Trump 1963, pp. 302–3; 1966, pp. 48–9; 1995–6, pp. 173–7; 2004c, pp.  231–41; Renfrew 1972, pp.  141–5; Skeates 1999–2000, p.  185; Fenech 2007, pp.  35–48. Radiocarbon readings for the Italian region can be found in Skeates and Whitehouse 1994. Compare the scope of dates available in the 1990s for Neolithic Italy, brought together in a comprehensive study, Skeates 1994. Scientific dating has not been exploited for post-Neolithic contexts. 63 Malone et al. 2009, pp. 341–6. 64 For legislation concerning cultural heritage and related organisations, Pace 2001, 2003. The Malta Centre for Restoration unit, which was established by the 2002 legislation, ceased to exist following an amendment to the Cultural Heritage Act in 2005; its role was taken over by Heritage Malta. Mission statements by the Superintendence of Cultural Heritage can be found on their website:  http:// www.culturalher .asp?p=21406&l=1 (accessed January 2013). 65 See MEPA website and links to ‘Heritage’: heritage-home (accessed January 2013). 66 Pace 2001, p. 233. 67 Tanasi and Vella 2011. 68 Zammit and Mallia 2008. 69 Baldacchino and Evans 1954, p.  19:  The ‘remains of the stone and flint industries, with one notable exception, are not of great importance.’ 70 Murray 1923a, p.  31:  ‘The whole was sifted by hand so that every particle of flint and scrap of pottery should be found. In this way a number of small flint chips were obtained, some not larger than 1/16 of an inch across; they appear to be the débris left by a workman when making or sharpening a flint implement.’ She also published some of the lithic finds from the Borg in-Nadur site and noted find spots around the megalithic structure; C.  Vella (2011, pp.  191–2) reassesses the Borg in-Nadur lithics. 71 Recent lithic studies concerning Malta: C. Vella 2008b, §4:0; 2009, pp. 85–7; 2011, p.  173; Malone et  al. 2009, p.  242; Cazzella and Moscoloni 2005, pp. 272–84; Moscoloni and Vella 2012, pp. 65–81;Vella and Moscoloni (in press).

Notes to Pages 15–22 72 Lemorini 2012, pp. 84–9. 73 Xagħra Circle Report, Malone et al. 2009; Tas-Silg (south), Bonanno and Vella 2015; Tas-Silg (north), the most recent accounts appeared in Scienze dell’Antichità (for 2012, see articles by Amadasi Guzzo; Bonzano; Cazzella and Recchia; Copat, Danesi and Recchia; Fiorentino, D’Oronzo and Colaianni; Lemorini; Mayer; Moscoloni and Vella; Notarstefano; Piacentini; Rossignani; Semeraro). 74 Docter et al. 2012. 75 Malone 2003, pp.  272–3, for Neolithic studies in Sicily focussed on ceramics. 76 Malone 2003 p. 272; Robb and Michelaki, pp.  161–81. Residue studies:  Tas-Silg (north), Notarstefano 2012, pp. 119–30. Chapter 2  The First Settlers and Farmers 1 Schuurs 2013, §2.12, pp.  47–8; Cartmill and Smith 2009, fig. 7:20. 2 On the discovery of the taurodont teeth: Despott 1918, pl. 19:B; 1923, p. 18; Keith 1918, pp. 404–5; 1924, pp. 251–60. For review of the literature concerning the possible Palaeolithic human remains in Malta:  Mifsud and Mifsud 1997; Savona Ventura and Mifsud 1999; Frendo 1999b. 3 For Fontana Nuova, Sicily: Bernabo Brea 1950, pp.  115–43; Chilardi et  al. 1996, pp. 553–63; Guzzard 2008, p. 39. 4 Lo Vetro and Martini 2012, pp. 21–2. 5 Chilardi et al. 1996, pp. 553–63. 6 Anati 1990, pp.  166–72; 1995, pp.  103–4; Mifsud and Mifsud (1997, pp.  145–51) note that the painted designs are evident in a number of locations throughout the cave. The comment by Evans (1953, p. 60) concerning the archaic (Mousterian) qualities of lithics found in the much later Neolithic contexts cannot be used as evidence for the Palaeolithic period; cf. Mifsud and Mifsud 1997, pp. 122 and 125. Absent still is surviving pre-Neolithic lithic evidence from secure contexts; see Skeates 2010, pp. 79–80; Trump 2002, p. 25. 7 Shackleton et  al. 1984, pp.  307–14, figs. 1 and 4. 8 Broodbank 2006, p. 214. 9 Ibid., p. 216; Pace 2004a, pp. 19–24. 10 Broodbank 2006, p.  218; after Cherry 1981, p. 47. Strasser et al. (2010, pp. 145–90)

Notes to Pages 22–28 argue that the exploitation of coastal lands in Plakias (south-western Crete) during the Lower Palaeolithic points to the maritime capabilities of early human groups by 130,000 BP; Simmons 2012, pp. 895–7. For Palaeolithic maritime crossing and settlement in Sardinia, Dyson and Rowlands 2007, pp. 17–27. 11 Zeder 2008, p. 11599; Malone 1997–8. 12 Guerrero Ayuso (2001, p.  138) discusses island colonisation relative to the Balearic Islands. 13 For the development of farming and domestication of animals:  Price 2000; Makarewicz and Tuross 2012, pp. 495–6. 14 Zeder et  al. 2006, p.  139; Zeder 2006, pp. 105–17; 2008, pp. 11597–9. 15 For DNA analyses of sheep, Bruford and Townsend 2006, p.  306; of goats, Luikart et al. 2006, pp. 294–305. 16 Zeder 2008, p. 11597. 17 Zeder 2008, p.  11598; Makarewicz and Tuross 2012, pp. 495–505. 18 Luikart et al. 2006, p. 294. 19 Bradley and Magee 2006, p.  317; Rosenberg and Redding 1969, pp. 55–64. 20 Patton 1996, pp. 59–62. 21 Broodbank 2006, pp. 214–15. 22 Bellwood 2005, p.  75; Price 2000, pp.  11–16. Trump (1966, p.  22–3) indicated that ‘true cardial impressions’ were not present in Malta to that date; the Ta’ Mixta material has stronger links to offshore types, see n. 41. 23 Skeates 2010, pp. 81–91. 24 Patton 1996, pp.  35–58 and 104–11, table 3:2. Island biogeographic models: A/DR after MacArthur and Wilson 1967; T/DR after Held 1989. 25 Broodbank 2006, p. 217. 26 Marriner et al. 2012, p. 10. 27 Whitehouse 1992, 2001; Skeates 2010, pp.  208–10; 2007, pp.  90–5; Tanasi (in press a). 28 Savona Ventura and Mifsud 2000, p.  32, after Anati 1990, 1995. 29 Sinclair and Keith 1924, p.  266; Despott (1923) concerns the excavations at Għar Dalam and has some illustrations of the pottery found in the cave; Evans 1971, pp. 208–9. 30 Malone 2003, p. 275, fig. 12; Tanasi 2011b, pp.  283–4; Evans 1971, pp.  208–9; Magri 1976–7, pp. 510–18.

335 31 Bonanno 1993c, p.  222; seven of the Il-Mixta sherds are illustrated, fig.  4; also Bonanno 1990, p.  14; Veen and van der Blom 1992, p.  18, fig.  13. Other caves which might have been occupied in the prehistoric period were destroyed. 32 Malone 2003, pp.  237 and 242; Leighton 1999, pp.  61–2. Although a sequence for the Għar Dalam phase is suggested by the four levels at Tac-Cawla, Gozo, on current evidence, little more than broad statements can be made for the early Neolithic in Malta. 33 Malone 2003, p. 275; Guzzardi 2008, pp. 39–48, figs. 2:2 and 2:3 (Vulpiglia pottery); Bonanno 1993b, p. 222, after Maggi 1976–7; Veen and van der Blom 1992, pp. 31–8; Sluga Messina 1988, pp. 169–71. 34 Skorba charcoal samples from context FB6 (lab. nos. BM–378 and BM–216) give dates that are ‘indicative’ of the broad time spans for the period but not specific enough to pin down a date for island colonisation; Trump 1995–6, pp.  174–5. Skeates (2010, p.  126) pushes the date to 5200 BC for the beginning of the Għar Dalam phase; Leighton 1999, p.  73:  the ‘14C date (5266–4846 cal. BC) represents a terminus ante quem, rather than the arrival date of the first settlers’;Veen and van der Blom 1992, p. 16, to 5500 BC. 35 Trump 1966, p. 11, §5c; 2002, p. 31. 36 Trump 1966, p. 10, §5c, fig. 8, pl. 13:a, b. 37 Ibid., p.  10, §5b, fig.  8, pl. 12:a, b; 2002, p. 29 (pictured); C. Vella (2008a, pp. 78–9) discusses this boundary wall, drawing on Trump’s original field notes in respect to depositional contexts of lithic material. 38 Trump 1966, p.  10. These building techniques (wattle and daub on stone foundations) find antecedents at, for instance, Piano Vento in Sicily, Leighton 1999, p. 73. 39 Trump 1966, p. 10, §5b, p. 24; cf. C. Vella (2008a, pp. 78–9), who cites a lack of floor material and the presence of charcoal, burnt seeds, animal bone and spent tool debitage in the deposits associated with the Għar Dalam–period boundary wall. 40 Ditched and walled settlements in Neolithic Sicily, Malone (2003, pp.  240 and 252–5) discusses numerous examples of ditched and walled settlements in Sicily and Italy; also Leighton 1999, pp.  69–70; Whittle 1994, pp. 149–50; Flemming 1982.

336 41 Likely Għar Dalam finds at Xewkija: Magri 1906, pls. 1:1, 2:3, 5, 69, 4:29 and 8:10; Evans 1971, pp. 191–2. 42 Għar Dalam finds at Santa Verna: Trump 1966, pp. 19–20, fig. 17:b; Evans 1971, p. 189 (three sherds). For Ta’ Ħagrat:  Trump 1966, pp. 17–19, figs. 16 (plan) and 17:c–f (sections); Evans 1971, p. 208; Zammit and Mallia 2008. 43 See Trump 1966, pp.  17–21, table  1, for the adjustment to the Maltese cultural sequence; cf. Evans 1953, p.  44, fig.  2, for the earlier, erroneous placement of the Mgarr phase following Għar Dalam; Evans 1971, p. 34;Trump 1966, p. 19, fig. 17:a, for Kordin III investigations. 44 Veen and van der Blom 1992, p. 19. 45 The problematic Tac-Cawla investigations (see Skeates 2010, p. 126) were carried out by Veronica Veen and Adrian van der Blom (1992, pp. 19–30), and some of the finds are illustrated. Given the limitations of their fieldwork and reporting, claims about cultural development need to be viewed with caution. Similarly unverified are attributions of some sites in the archipelago to the Għar Dalam phase in their gazetteer (pp.  59–67; see Appendix B, List 1); the authors indicate clearly that they had no access to the museum stores (p.  15). Background to their fieldwork can be found at english/background.html (accessed 23 September 2013). 46 Stoddart 2014, pp. 20–4. 47 Veen and van der Blom 1992, pp. 60 and 62, site 5; Trump 2002, p. 28. 48 MAR 2002, p.  176; Malone et  al. 2009, pp. 54–5, table 3:3. 49 For Għar Dalam finds at Triq E. Schembri, Birkirkara, on Mrieħel Ridge, MAR 1996, pp. 1–2. 50 Trump 1966, §9, p.  10, carbonised grain came from Għar Dalam deposits south of sector SB; Helbaek 1966, app. IV. 51 Trump 1966, p. 29, for Grey Skorba–period sickle blades; 2002, pp. 34–5. 52 Skeates 2010, pp. 85–6. 53 Trump 1966, p. 24, fig. 21:b. 54 Ibid., p. 24, n. 1. 55 Percentages of animal bone at Skorba after Borg 2008, pp. 87–8;Trump 1966, §9,

Notes to Pages 28–31 p. 24; Gandert 1966, app. III. All are broad tallies for the site rather than culturally defined. 56 Visscher et al. 2001, p. 303; 2010. 57 Għar Dalam pottery discussions: Despott 1918, pl. 7, figs. 1 and 2; 1923, pp. 22–4, pl. 1, figs. 1 and 2, pl. 2, fig. 1; Evans 1953, p. 44, pl. 7:1–9; 1959, pp. 41–4, pl. 35:1–f; 1971, pp. 208–9, pl. 63:5–7; Trump 1966, pp. 21–4, figs. 18–20; 2004a, pp. 251–2. Other ‘new’ reconstructed shapes reported in Veen and van der Blom (1992, p. 22, fig. 17) can be, for the most part, accommodated within the five forms identified by Trump at Skorba (bowl rims, pedestal bases and baggy jar rims). An open plate and deep nearly vertical-sided open pots with rounded or squared shape have otherwise not been documented for this phase; however, vessels with squared mouths are known at Italian sites in the Middle Neolithic period; see Malone 2003, fig. 13:6–9. Zoomorphic handles are similar to examples found in Sicily; e.g., Leighton 1999, p. 61, fig. 29, from Stentinello. For the Maltese examples, Trump 1966, pp. 22 and 24 (not illustrated; damaged and a surface find); Evans 1959, pl. 72; 1971, inv. no. GD/P3, for both Għar Dalam examples, pp. 20 and 209, pl. 32:7–9. 58 Malone 2003, p. 276, fig. 12; cf. decorated sherds from Tac-Cawla, in Veen and van der Blom 1992, p. 25, fig. 20. 59 Trump 1966, p. 23, figs. 18:d, f and 19:a, h (see Figure 2.2). 60 Robb 2007, pp. 186–8. 61 For outcrops of chert, for instance, in the west of Malta on the southern slopes of the promontory, below Baħrija:  MAR 1938–9, §35, p.  6; C.  Vella 2008c, p.  84; 2009, p. 88; 2010, fig. 1. 62 Discussion of prehistoric exchange mechanisms:  Tykot 2003, pp.  59–86; Sherratt 1982, pp. 13–26. 63 Apsects of the lithic châine opératoire are discussed in C. Vella 2008a, pp. 82–6. On changing patterns of cultural interaction in the Mediterranean, Copat et  al. 2010, pp. 41–64. 64 Cann and Renfrew 1964; Cann et  al. 1969; Hallam et al. 1976; Bigazzi and Radi 1981; Mello 1983; Ammerman et al. 1990; Bigazzi et al. 1981; Phillips 1992; Petrassi

Notes to Pages 31–37 and Zarattini 1997; Tykot 1996, 1997, 1998, 2003; Tykot and Ammerman 1997. 65 Tykot 2003, pp. 69–75; Ammerman 1979, pp. 95–110. 66 Appearance is often used as a factor in determining origin of obsidian artefacts in Malta: uniform, either free of impurities or streaked with white, grey-black and clear from Lipari; slightly opaque and greenish in hue from Pantelleria, C. Vella 2008b, §3:2. For Pantelleria obsidian sources:  Francaviglia 1988, pp.  109–22 (see references);Vargo 2003. For obsidian sources in Lipari and Vulcano, C.  Vella 2008b, fig.  3. Wider Mediterranean trade in obsidian, Robb 2007, pp. 192–3; C.  Vella 2009, pp.  99–101, for the shifting patterns of contact between Mediterranean cultures; offshore connections between Malta and obsidian sources: C. Vella 2008a, pp. 82–6; Trump 1966, pp.  49–50. Concerning quantities of obsidian in archaeological contexts in Malta: cf. Trump 1966, p. 50, table 4; C. Vella 2008a, pp. 79–82, figs. 3–7. 67 Sources of flint, Robb 2007, pp.  112 and 186ff. Concerning Sicilian lithic resources and the nature of lithic studies from preand post-1960s, Karimali 2005, pp. 186–7. C.  Vella 2008b, §3:3, n.  26, for Italian quarries in Mount Lessini (Veneto) and Gargano (Puglia). Chemical characterisation has not been conducted on flint lithics in Malta, C. Vella 2009, p. 89. 68 Vella’s typology is best outlined in his 2011 publication (see Chapter 3). 69 C.Vella 2011, p. 181, figs. 6:4–6:7. 70 Ibid., pp. 173 and 191–2. 71 Chert usage at Tas-Silg, where sources in the south-east of the island have not yet been located, C. Vella 2015, §4:6. 72 Tykot 2002. 73 Significantly, Robb (2007, p.  196) notes that obsidian was distributed as both blades and cores. For the Maltese evidence, C.  Vella 2008a, p.  84. Regarding preparation and importation of flint material, C. Vella 2010, pp. 11–12; 2008c, p. 89. 74 On differing use of lithic types, C.  Vella 2008b, §7:0; 2008c, pp. 90–1. 75 C. Vella 2008a, p.  85; 2008c, pp.  90–2; 2010, p. 12.

337 76 Pits of Grey Skorba date were excavated in trenches QE, OE and PE, located in the north-east of the site; C. Vella 2008a, pp. 77 and 80–1; 2008b. 77 C. Vella 2008a, pp.  79–80. Trump’s field notes are held in the National Museum of Archaeology, Valletta. C.  Vella 2008c, p. 87, table 6:1, a quantity of chert debitage came from the boundary-wall area. 78 C. Vella 2009, p.  100, on variations between lithic assemblages at Ta’ Ħagrat and Skorba. 79 K. Buhagiar 2007, p.  358. Malone et  al. (2009, §10.3.2, p. 242) indicate that chert was ‘readily found in outcrops and pebbles’. C.  Vella (2009, p.  88) mentions chert outcrops at Wied Marsalforn, Wied Saqqajja and Wied Ħemsija documented in the nineteenth century, but the last two sites have not yet been located. Wettinger (2000, p.  577) lists Wied Ħemsija as a locality near Misraħ Kola. 80 Karimali 2005, p. 187. 81 Mangion 1966; Trump 1966, pp.  10–11; trench EF was about 8 m south-west of the Għar Dalam house. 82 Trump 1966, p. 24, Għar Dalam objects, figs. 21:b (awl) and 22:a (obsidian blade). 83 Ibid., p. 24, fig. 26:e. 84 Robb 2007, p. 96. 85 Leighton 1999, pp. 67–72. 86 Trump 1966, pp. 10–11, pl. 22:a, p. 16 (for stratigraphy); C. Vella 2008a, p. 77. 87 Trump 2002, p. 31. 88 Trump 1966, p. 29. 89 Grey Skorba pottery development, Trump 1966, pp. 25–30; 2002, p. 30. 90 Leighton 1999, p.  73, Serra d’Alto falls within Sicily’s Middle Neolithic (p. 62). 91 Trump 1966, p. 25, fig. 23. 92 Ibid., p. 26. 93 Ibid., p. 28, but pictured with Red Skorba Wares in fig. 29:g. 94 Trump (1966, pp.  2 and 28–9) identified quantities of chert flakes in the Grey Skorba phase forming an amorphous industry with rare instances of more complex reworking and limited signs of use wear; see also MAR 1961, pp.  3–5; 1963, pp. 2–3; Evans 1971, pp. 36–9 and 209–210. 95 C. Vella 2008a, p.  81, see also pp.  76–8 and 80–1, fig.  4 (the deposits were presumably, although it is not clearly stated,


Notes to Pages 37–42

one somewhat of the plan of the Copper Age temples, but they seem not to have been linked together in any way, as the rooms of the latter were.’ 107 Trump 2002, p. 31. 108 Trump 1966, p. 13. 109 See ibid., p. 12; photograph of the North Room uneven floor in pl. 16:b; Malone (2003, p. 257) considered that the structure both in ’form and scale’ fell within a domestic range. 110 Fill over bedrock in the North Room is described in Trump 1966, p. 13. 111 Trump (1966, p.  14) notes that pottery ‘became noticeably less frequent towards the western end of the North Room and very much scarcer higher in the fill everywhere.’ 112 Malone et al. 1988, p. 300; 2009, p. 51. 113 Trump 1966, p. 14. 114 Ritual caching of lithic artefacts, C.  Vella 2011, p.  191; also discussed in N.  C. Vella et  al. 2011, p.  52. Murray (1923a, p. 26) noted that Neolithic pottery and lithic tools were under the floor of the north-east apse at Borg in-Nadur:  ‘Under the slab were found pieces of neolithic [sic] pottery, a small flint implement (pl. XVI, 16) and a scrap of obsidian’; cf. Evans 1971, pp. 8–9. 115 Ashby 1910, p. 159. 116 Trump 1966, p.  14. C.  Vella (2008a, p.  77) also argued that the deposit, which he viewed as ‘refuse’ accumulated in the rooms ‘possibly within a short time span’. 117 Concerning finds from the Red Skorba rooms, Trump 1966, p. 14, pl. 18:a, the pair of goat horns in situ: pp. 33–4, fig. 30, pls. 26–7, figurines; p. 34, fig. 31:a, b, spindle whorls; p. 34, fig. 30:g, pl. 28:a, worked cow bones. 118 For the function of figurines in prehistoric contexts:  Voigt 1983, pp.  186–95; Lesure 2002; Mabry 2003. 119 Trump 1966, p. 40; Evans 1971, p. 211. 120 Malone 2003, p. 257. 121 Red Skorba figurines are depicted in Trump 1966, p. 33, fig. 30;Vella Gregory 2005, pp.  28–31, figs. 1–5 and 7–12 (fig. 6, a modified tooth described as a phallic symbol, is likely to be grooved for suspension; see Figure 2.4, no. 12).

from pits); 2008b, §6.1.2: ‘Although this core was discarded, it was still in an utilizable stage’; 2008c, p. 87, table 6:2, for similar sentiments: ‘This area of the village must have been an extensive dump over a short period of time.’ For the obsidian cores from Skorba VE4, Trump 1966, p.  28, pl. 23:b; the core from Pantelleria weighed 1.7 kg and the other from Lipari weighed 400 g. 96 Evans 1977, p. 20; Trump 1966, p. 30. 97 Trump 1966, pp.  29–30, pl. 23:d (sling stones). Area A  at Tas-Silg appears to have been used as a knapping zone, away from the prehistoric buildings at the site, C. Vella 2015, §4.2.1, §4.2.2, §4.2.3 and §4:7. 98 Chapman 2000, pp.  64–5 and 77; the matter of pits as lithic workshops is discussed for Kompolt-Kistér (central Hungary). 99 Maniscalco 1989, pp.  537–41; the Serra del Palco finds are linked to the San Cono-Piano Notaro–Grotta Zubbia pottery styles contemporary with the Maltese Zebbug phase. 100 Trump 1966, p. 28, fig. 26b, c, g. 101 Ibid., p.  16, fig.  14, for discussion of the stratigraphy in this section; for greenstone artefacts, Leighton and Dixon 1992, pp. 179–200; Skeates 2002, pp. 13–22. 102 Trump 1966, p. 28, where daub is mentioned, and p. 13, where nodules of mud brick are addressed; for the courtyards, pp. 11–12, fig. 9, pls. 14:b and 18:b. 103 MAR 1963, p.  2; Bonanno 1993b, p.  220:  ‘Here, in time, two huts of the same type, but slightly larger assumed the function of shrines for non-secular, if not exactly religious purposes’; Evans 1976–7, p. 131. 104 Bonanno 1993b, p.  220. Also Trump 1966, pp. 33–4, pls. 26–27; Evans 1976–7, p. 131. 105 See Chapter 3 for a discussion concerning the concept of a fertility goddess; Fleming 1969; Meskell 1995; Lesure 2002. Interpretations of figurines in Malta can be far ranging, Malone et al. 2009, p. 298. 106 Evans 1976–7, p. 131: ‘The form of the huts, and their juxtaposition, reminds

Notes to Pages 42–49 122 Cow foot bones from Skorba pictured in Trump 1966, pl. 28:a; 2002, p. 43. 123 Leighton 1999, p.  73, fig.  32 (Sicilian Diana Ware); Trump 2002, p. 30. 124 Red Skorba radiocarbon sample (BM–148, wood charcoal) from the Red Skorba ‘shrine’, Trump 1995–6, pp.  175–6; Malone et  al. 2009, p.  342; Leighton 1999, p.  65; Malone 2003, table 1. 125 Other Red Skorba designs (palm leaf, ear of corn, vertical dashes) are mentioned in Evans 1971, p.  211, but have not been illustrated. 126 Spindle whorls and textile production, Trump 1966, fig. 31, found in Red Skorba, Zebbug, Ggantija and Tarxien deposits in Skorba. 127 Trump 1966, pp. 34–5. 128 C. Vella 2008a, p.  81; 2008c, pp.  87–8. This author also has noted a dearth of evidence for hafting at Tas-Silg, C. Vella 2015, §4.5.1. Chapter 3  The Culture of the Megalith Builders: The Late Neolithic of Malta 1 Traces of over thirty megalithic structures survive; see Appendix B, List 9. Grima and Vassallo 2008, p. 8. 2 Renfrew 1973, p.  147; 2004, pp.  10–11. Some scholars debate that the Göbekli Tepe buildings were not free-standing but were partly sunken into the earth. Nonetheless, a number of upright megalithic stones were free-standing in the interior of the buildings. Similarly, the structures at Nevalı Çori,Turkey (8600–7900 BC), also have free–standing internal upright megaliths:  Schmidt 2011, pp. 41–83; Hauptman 2011, pp. 85–138. 3 For the evolution of the prehistoric pottery sequence:  Evans 1953, 1971; Trump 1961, 1966. Bonanno (1993c) has a good summary of the works written on the prehistoric period to that date. Later applications of the pottery sequence: Stoddart et al. 2009a, p. 79; Bonanno et al. 1990, pp. 190–205. 4 Evans 1971, p.  224. Similarly, metal is not that common in contemporary cultures of Sicily, Leighton 1999, p. 87. Concerning terminology:  Bonanno 1993a, p.  37; Cultraro 2008, pp. 5–6.

339 5 Zebbug-phase tombs at Ta’ Trapna iz-Zgħira, Zebbug and Baldacchino, Evans 1954, pp. 1–21. 6 Evans 1971, pp. 207–28; Trump 1966. 7 Early studies and reports of prehistoric Malta include works by Maltese scholars: e.g., Caruana 1882, 1898, 1899; Zammit 1916, 1917, 1920, 1930b; Tagliaferro 1911, 1912, with T. Ashby (Ashby et al. 1913). 8 European scholars who made inroads into Maltese archaeology:  e.g., Ugolini 2012 (1934); Peet 1910; Murray 1923a, b, 1925, 1929, 1934. 9 The MTLP was a joint German Archaeological Institute, University of Bournemouth, University of Malta project, Heritage Malta and Superintendence of Cultural Heritage project. 10 A collaborative Belgium-Malta Survey Project in the north of Malta has used systematic field survey and ground-penetrating radar to good effect; Docter et al. 2012; Malta Survey Project 2008; available at: .mt/arts/classics-archaeo/projects/maltasurvey (accessed 22 November 2011). 11 Additional structures and deposits of the Neolithic period in fields west and north of the Tarxien site, MAR 1997, p.  2. Concerning the megalithic structure in Triq N. Tagliaferro (formerly Sqaq Sardinia), see Superintendence of Cultural Heritage 2008, §3.1.4. 12 Houses had not been detected in the early 1900s, Ashby et al. 1913, p. 13. 13 For a survey of settlements and domestic structures in Sicily, Leighton 2009, pp. 100–1 and 116–21, figs. 54–5. Examples of settlement sites in Sicily: Monte Racello, Orsi 1898, pp. 204–06, fig. 15; around nine ovoid houses or huts (capanna) formed a small village at Sante Croci hill, Orsi 1926. 14 Prehistoric houses in Malta: Malone et al. 1988, 2009. The Għajnsielem (Gozo) site was identified by Joseph Attard. 15 Querns were particularly prevalent in Ggantija contexts, Trump 1966, p.  15; a sickle blade was evident in finds from, for instance, the Għajnsielem house, Gozo, Malone et al. 1988, p. 301. 16 Evans 1977, pp. 12–13. 17 San Cono–Piano Notaro, Sicily, contemporary with the Zebbug pottery

340 tradition: Trump 1966, p. 36; Evans 1971, pp. 212 and 214; Leighton 1999, p. 93. 18 Trump 1966, p. 36; 2004, p. 254: Zebbug shows ‘no local continuity’; Cultraro 2008, p. 10. 19 Trump 1966, §5f, p. 14. 20 Trump 2002, p. 32. 21 Evans 1971, p. 212, for a synthesis; Trump 1966, pp. 35–8. 22 Evans 1971, p. 212; Cultraro 2008, pp. 9–10. 23 Early Zebbug pottery: Trump 1966, p. 36, fig.  34 (see Figure  3.1, nos. 1–4); Evans 1971, p. 213. 24 Similarities between Zebbug and San Cono–Piano Notaro culture wares of Trefontane, Trump 1966, p. 36. 25 Malone et  al. 2009, p.  221, fig.  10.2.H; p. 222, fig. 10.3.A. 26 Baldacchino and Evans 1954, p. 9, figs. 6:4 and 7:10. 27 Trump 1966, p. 35, not illustrated. 28 Spoons of Zebbug date, Trump 1966, p. 36, not illustrated. 29 Trump 1966, p.  36, fig.  32:i, j (see Figure 3.1, nos. 5 and 6); Evans 1971, p. 213. 30 Trump 1966, p. 36 31 Ibid., pp.  35–6, fig.  33:b, f; 2004a, p.  256; Evans 1971, p. 213. 32 Trump 1966, pp. 36–7, fig. 35, pl. 29:b; the polished axes are said to have been made from igneous or metamorphic stones. Evans 1984, for ‘temple period’ offshore contacts. 33 Not all the finds from Skorba have been illustrated in the final report; see Trump 1966. 34 Trump 1966, pp. 37 and 49–50, table 4: not all of the artefacts mentioned by Trump were illustrated in the report of 1966; Evans 1971, p. 213. 35 Baldacchino and Evans 1954, limestone bowls from Ta’ Trapna tomb 3, p. 6, tomb 5, pp.  13–14, also p.  19. Pierced shells from Skorba (settlement), Trump 1966, p.  37; from Ta’ Trapna (tombs), Evans 1971, p. 213. Cf. a range of finds from the Zebbug tomb, Xagħra Circle site, Malone et al. 1995. 36 Evans 1971, pp.  166–9, plans 31–5. Baldacchino and Evans 1954, report of the five Zebbug tombs; p.  19 lists European parallels. 37 Funerary evidence (a forno tombs) in Sicily is summarized in Leighton 2009, pp. 93–9,

Notes to Pages 49–55 fig. 43; tombs of the Monte Claro culture (Cagliari), Atzeni 1967, pp. 157–79. 38 Malone et  al. 2009, pp.  95–103; 1995, pp. 303–45. 39 Malone et al. 2009, pp. 82–3, 96, figs. 6:4 and 6:5. 40 Zebbug stylised human statue menhirs, Xagħra Circle, Malone et  al. 2009, pp. 282–3, figs. 10:46 and 10:47; p. 283 for comparative examples; Malone et al. 1995. Ta’ Trapna iz-Zgħira, Zebbug: Evans 1953, pl. 9:3; Baldacchino and Evans 1954, tomb 5, pp. 13 and 20, pl. 3. 41 Richards et al. 2001. 42 Menhir at Ta’ Trapna coated with red ochre, Baldacchino and Evans 1954, p. 13, pl. 3. 43 Baldacchino and Evans 1954, p. 1: the Ta’ Trapna burials were ‘profusely impregnated with red ochre’; burials with red ochre cakes, tomb 1, p.  4, tomb 2, p.  4, tomb 4, p.  11, tomb 5, p.  14; red ochre–encrusted beads, tomb 4, pp. 10–11, tomb 5, p.  14; pottery containing or stained with red ochre, tomb 3, p. 6, tomb 4, pp. 6–7. Skeates (2008, p. 8, fig. 1) illustrates a grinding stone with ochre residue on the interior found in one of the Zebbug tombs, displayed in the National Museum of Archaeology, Malta. 44 Busuttil et  al. 1969, (for 1968)  San Pawl Milqi, tomb 1, pp. 94, 101 and 123–4, fig. 7, pl. 22:1–3, 5–6. 45 Attard Montalto et al. (2012, pp. 1094–102) identify likely sources for red ochre pigment in Malta and present their analyses of samples from archaeological contexts in the islands; a red ochre outcrop in Għar Hasan is illustrated in Savona Ventura and Mifsud 2000, p. 20, pl. 5. Cf. Maniscalco (1989), who presents cultural evidence that ochre may have been a commodity traded between Malta and Sicily in antiquity. 46 Cultraro 2008, p.  10, figs. 1:2:d–g, i and 1:4:5–6. 47 Malone et  al. 2009, pp.  253–60 and 363, figs. 7:8, 10:30, 10:31 and 10:32. 48 Evans 1953, pp. 45–8. 49 Trump 1966, p. 38. 50 Evans 1971, p. 214. 51 Concerning the nature of the Mgarr phase:  Cultraro 2008, pp.  10–11; Malone et al. 2009, pp. 226 and 228.

Notes to Pages 55–71 52 Limited offshore contact in the Mgarr phase:  Evans 1971, p.  215; Cultraro 2008, p. 10. 53 Trump 1966, §5g, p. 14. 54 N. C. Vella 2004, pp. 17–19. 55 Malone et  al. 2009, p.  53, table  3:2, with references. 56 Evans (1953, pp.  45–8, pl. 7:10–16) fig.  6 has been largely revised by Trump 1966, p. 38, fig. 32 and corrected by Evans himself in 1971, fig. 34. 57 Designs on the flat surface of some point to their use as lids, Evans 1971, p. 215. 58 Trump 1966, p.  38; 2004, p.  256; Evans 1971, p. 214. 59 Evans 1953, p. 46; 1971, p. 214. 60 Trump 1966, p. 38, illustrated only in situ, pl. 19:b; the tray measured 46 cm long × 22 cm wide × 5 cm thick. 61 Evans 1971, p.  215; not all the recorded artefacts of the Mgarr phase have been illustrated in the reports. 62 Torba comprised ‘ground limestone, clay and residual pottery which was collected and ground, and used specifically for filler, a tradition that has continued in Malta until recent times’; Malone et  al. 2009, §3.3.5, p. 49. 63 Trump 1966, p.  15:  ‘It would be tempting to suppose that it was deliberately destroyed to clear the site for the temple.’ 64 Evans 1971, p. 217. 65 Malone et  al. 2009, pp.  41–56; 1988, pp. 297–301. 66 Malone et  al. 2009, pp.  54–5, table  3:3, with references. 67 Ibid. 68 Evans 1971, p. 21. 69 Malone et al. 2009, p. 53; Evans 1931, p. 56; MAR 1929–30, pp. 8–10. 70 Transition Mgarr-Ggantija pottery was also identified among the material from tomb 2 at Xemxija, Evans 1971, p. 217. 71 Evans 1971, p.  217. Trump 1966, pp. 38–40, figs. 38–39, for Mgarr-Ggantija transitional wares. 72 Evans 1971, p.  216, fig.  35:14 (see Figure 3.6, no. 5). 73 Evans 1971, p. 216. 74 Tas-Silg (north):  Copat et  al. 2012, pp.  47–8. Tas-Silg (south) produced very little Ggantija pottery, Sagona 2015, pp. 7–10. 75 Trump 1966, §5h, p. 15.

341 76 Evans 1971, pp. 37 and 39. 77 Trump 1966, p. 40. 78 C.Vella 2010, p. 11; 2008c, p. 84. 79 Evans 1971, pp.  112–16, plans 24–9, and appendices by G.  Pike and S.  Rogers, pp. 236–41; Trump 2002, pp. 162–5; Cilia 2004, pp. 164–5. 80 Evans 1971, p. 59. 81 Evans 1953, p. 60: ‘Completely new, however, are the curious “Saflieni bowls” ’; Evans 1971, p. 218. 82 Trump 1966, §6, p. 17, §15, p. 41. 83 Evans 1971, pp.  218–19:  Nothing is yet known of buildings of the Saflieni phase, though the abundance of ‘Saflieni pottery, especially Saflieni bowls, in the Hypogeum is noteworthy. It seems likely to have been a short phase, but there is no means of dating it except through the preceding and succeeding phases’; more recently, Malone et al. 2009, §6.7, p. 85. 84 Trump 2002, p.  223; 1966, §15, p.  41; cf. Malone et al. 2009, pp. 85–6 and 229–31. 85 Trump 1966, p. 41; 2002, p. 223. 86 Trump 1966, p. 41; Tas-Silg, Saflieni-type pottery, Sagona 2015, § 87 Cf. Malone et al. 2009, §6.7, p. 85. 88 Trump’s ‘semi-hole-mouthed jar’ (1966, § 15, p. 41), after Evans 1953, p. 59, fig. 9:52. 89 Robb 2007, pp.  112 and 188; Trump 1966, p. 42. 90 Tarxien period, Evans 1971, pp.  219–24. New data from, Xagħra Circle, Malone et  al. 2009; Tas-Silg, Sagona 2014; Borg in-Nadur, Tanasi 2011a, pp. 76–84; additional material from Margaret Murray’s excavations of the 1920s is forthcoming; see Briffa and Sagona (in preparation). 91 Malone et  al. 2009, p.  49; Trump 2002, pp. 207–8. 92 Malone et  al. 2009, pp.  54–5, table  3:3; Veen and Van der Bloom 1992, §6, p. 63. 93 Trump 2002, p.  207, referring to investigations by Mark Horton and Nathaniel Cutajar. 94 Malone et al. 2009, pp. 52 and 54–5. 95 Evans 1971, pp. 39 and 231; MAR 1927–8, pp. 1–2; Malone et al. 2009, p. 53, table 3.2. 96 Evans 1971, p. 216. 97 Zammit 1930, pp. 84–5, pl. 23:5. 98 Convex stones at Ħagar Qim, Evans 1971, p. 94, inv. no. Q/S23, pl. 66:8. 99 Zammit, Archaeological Field Notes no.  13 (1915–19), 20 September 1915, p.  90.


See  also three cupped stones sketched on p.  88 in space L (a small alcove in front of the stairs that lead to roof level); p. 89 indicates that a number were found 'mostly embedded in a kind of mortar made naturally in the course of burning fires close to limestone which in fact is turned into lime and which water turn into mortar. One of the stones has a hole at the side’; 21 September 1915, p. 91: ‘To the east of the space [i.e., space L] numerous sherds were found and also numerous concave hemispheres of which two have a hole at the side, one hole goes completely through the stone, the other is about 5  mm deep (with two small sketches)’; 23 September 1915, p.  98:  ‘mass of concave hemispheres in corner L’. More were found in the 1916 season of excavation: 18 May 1916, pp.  134–5:  ‘At the foot of the upright on the left forming one of the jambs of the doorway in the passage of O to the north which was cleared deeply to observe how far it [was] broken, Neolithic sherds and 24 hemispherical stone were found.’ See also 30 June.1916, p. 153, one example; 22 July 1916, p. 168, near area L, ‘a mass … enough to fill a large box. These are mostly, made of a bluish [clayey] stone matted together, often with lime, ashes and soil.’ 100 Zammit 1930, Tarxien plan, ‘second temple’, between, pp.  4–5; Zammit Archaeological Field Notes no. 11 (for 1915), p.  37, 24 August 1915 at point L:  ‘On what may be a threshold a number of stone discs, of the ordinary pattern [one is sketched] are arranged in a line, as far as visible today. Many other discs have been met in the vicinity’, see also p. 38, where the apparent orderly arrangement of ten stones is noted (a few photographs were taken at the time). 101 Zammit (1930, pp. 84–5) suggested that they could have been cups used for holding food; Evans (1971, p.  94) discounted this possibility as they are too shallow. 102 Zammit et al. 1912. 103 Trends in obsidian use in the prehistoric era, C. Vella 2011, p. 188; 2014, §4.5; for the Xagħra Circle, Malone et  al. 2009, §, p. 253.

Notes to Pages 71–76 104 Imported lithic material in Borg in-Badur, C. Vella 2011, p. 188. 105 Xagħra Circle lithics, Malone et al. 2009, pp. 242–53. 106 Chert sources and usage relative to the Tas-Silg site, C. Vella 2014, §4:6; at Borg in-Nadur, C. Vella 2011, pp. 188–9, n. 37. 107 Trump 1966, §16, p. 43. 108 Evans 1977, p. 21; Trump 2002, p. 212. 109 Imported lithic material appears in significant amounts at Ta’ Ħagrat; C. Vella 2009, p. 100. 110 For fluctuations in obsidian trade, Robb and Farr 2005, pp. 35–41. 111 Trump 2002, pp. 212–13; also Evans 1971, p. 223, regarding Neolithic contacts. 112 For example, Bonanno 1993b, p.  224; Trump 2002, p. 241. 113 Cazella et  al. 2007, p.  255. Thermi and other exotic wares also were found in small numbers in Gozo at the Xagħra Circle; Malone et  al. 2009, pp.  238–40, fig. 10:19. 114 Renfrew 2007, p.  8, concerning the megalithic buildings in Malta. 115 Renfrew 1973, pp. 156–7. 116 Trump 2007. Düring 2002, p. 226, who defines the function of the religious and domestic quality of the buildings under the title ‘ritually elaborate buildings’; see also Düring 2001, pp.  1–18; Turnbull 2002. 117 Contra Evans 1996, p. 39: ‘They do not seem to be well-adapted for defensive purposes, while the apparent lack of interior lighting and the internal arrangements do not favour the idea of any kind of residential use.’ 118 Megalithic structures as dwellings or as intrinsic parts of settled areas have been discussed from time to time: Ashby et al. 1913, pp.  5–6; Murray 1923a, pp.  24–5; Cardona 2008a, p. 62. 119 Grima and Vassallo 2008, p.  8. Pace (2004b, p. 150) also described Ta’ Ħagrat as ‘almost hut-like in appearance and dimensions’; see also Grima 2007, p. 38. 120 Sherratt 1990, pp. 147–9; 1995b. 121 Sherratt 1990, p. 150. 122 Megaliths over earlier settlements, Trump 1966, p. 19; 2002, p. 137; Grima (2004, p.  333) argued that these earlier settlements ‘were already important foci of activity’, which had a firm hold

Notes to Pages 76–78 in social memory; N.  C. Vella 2004, pp. 17–19. 123 Renfrew 1973, pp. 109ff. 124 Evans 1901, 1902. 125 Lilliu 1970; Bonanno 1993b, p. 224; Tusa 1991. Conversely, despite the chronological problems and differing functions, it has been mooted that Sicilian tombs were modelled on the Maltese megalithic structures: Terranova 2008, figs. 4:2–27. 126 Evans 1901, p. 198, drawing on, for example, Orsi 1892, pp. 69–70, pl. 6. Connection was made with tombs of Sicily in regard to plan and to the Sicilian window tombs (‘tombe a fenestra’). The Maltese megalithic architecture predates the Sicilian evidence by centuries. Portal doorways are found in various contexts in Europe; see Fleming 1969, pp. 248–9; Leighton 2009, p. 128; Bradley 1989, pp. 68–75, on deaths and entrances; 1998, p. 40. 127 ‘Indigenous’, meaning simply (in this volume) those already living in the area, whether historically immigrants themselves or not. 128 Ashby et al. 1916, pp. 1–6; Trump 2002, p. 190. 129 Cazzella et al. 2012b, p. 601. 130 Caruana 1896a, p. 30, for Kordin; Clark 2004, pp. 367–8, for a discussion of architectural technology. Alterations had possibly been undertaken at Borg in-Nadur, Vella, Zammit and Bugeja 2011, p.  53, who also suggested that the scheme of temple development needs to be revised on the basis of the diverse characteristics of the buildings (n. 36); Di Salvo 1988. 131 Grima 2004, p. 333. 132 T. Zammit (no date, p. 4) acknowledged that the apse plan and central corridor of the Ħagar Qim complex were modified from the usual form. 133 Reports of the Xemxija tombs:  MAR 1955–6, p.  2; 1956–7, p.  2; Evans 1971, pp.  112–16, plans 24–9; see also Trump 2002, pp. 162–5. 134 Evans 1959, pp.  90–1; cf. Pace 2004a, pp.  28–9; Trump 1999, p.  94; Stoddart 2007, pp. 54–60. 135 Proponents of megalithic structures deriving from funerary contexts: Evans 1976–7, p. 143; 1977; Trump 1981, fig. 2.

343 136 Zebbug funerary deposits in the Xagħra Circle, Malone et  al. 2009, pp.  95–102; Ħal Saflieni, Pace 2004c, p. 23. 137 Depictions of clay room models:  Torpiano 2004, p.  348 (bottom); Trump 2002, p. 81; Evans 1971, inv. no. Q/P1001, p. 91, pl. 39:15–16. 138 Illustrations of stone megalithic building models, Torpiano 2004, p. 348 (top) small stone model from Ta’ Ħagrat; p.  354 (top) stone fragment with graffiti of possibly two building façades from Skorba; p.  355 (top) amulet representing a façade from Tarxien, Trump 2002, p. 120. 139 Artistic representations of likely megalithic buildings, Trump 2002, pp. 192–5; three-dimensional model, Evans 1971, pls. 33:11–12, 47:7–11 and 51:6, possible architectural references under the hem of a robust sculpted figure, pl. 48:3–5; possible niches or doorways, pls. 41:  8 and 50:9–11. 140 Development of megalithic mode of building from domestic architecture: Evans 1953, pp. 41–94;Trump 2001, pp.  14–15; 2002, pp.  87–9; Pace 2004a, pp. 28–9. 141 New influences in Zebbug phase: Evans 1971, p. 214; Trump 2002, pp. 30–1. 142 Torpiano 2004, p. 347. 143 Partial collapse of the megalithic structures at Mnajdra in April 1994 is documented in the media of the day; see also MAR 1994, p.  485; 1996, p.  1; at Ħagar Qim in 1996, MAR 2001, p. 166; Tarxien in 1999, MAR 2001, p.  166; in 2003, Superintendence of Cultural Heritage 2005, §3.3, p. 8. Efforts to halt the natural erosion have been undertaken at the most vulnerable sites, overseen by the Scientific Committee for the Conservation of the Megalithic Temples (formed in 2000); Stroud and Chetcuti 2008, pp.  95–110. Mnajdra also was subjected to graffiti attack (26–27 October 1996)  and wanton, large-scale vandalism; MAR 2001, pp.  165–6. Both Ħagar Qim and Mnajdra are now protected within high-security compounds, and a number of other sites also were upgraded in this regard; MAR 2001, pp.  166–7; 2002, p. 175.

344 144 These large and deep cisterns at Tal-Misqa are linked by cut tunnels, and the stone from which they are cut is compact and water-tight Lower Coralline; Mifsud 1995, pp.  56–7. Additional cisterns (of unknown age) are reputedly located closer to Ħagar Qim; Cilia 2004, p.  142. A  large, bell-shaped cistern also has been reported at Ta’ Birikka north of the Tas-Silg site, documented in Bugeja 2015, §2.4. Rock-cut and inter-connected shallow pans at Tal-Misqa have been identified as wine presses probably of Phoenician-Punic date (Figure  3.16, no.  1:A); Bonanno 2008, pp.  1–18; Jaccarini and Cauchi 1999, for other examples in the archipelago. 145 Trump 2002, pp. 152–3. 146 Excellent aerial photographs by Daniel Cilia appear in Trump 2002, p. 152; Cilia 2004, pp. 142–3. 147 It is likely that while some carvings appear to be ancient, some lines do appear to be recent and were possibly added or modified around the time the petroglyphs came to public attention with an article in the Sunday Times of Malta, ‘Malta:  Ancestral Home of the Ancient Egyptians:  3.  Prehistoric Petroglyphs Found Near Misqa “Tanks” on the Maghlaq Plateau’, 22 October 2000, pp. 56–7. 148 The Misqa Tanks and their associated cuttings are clearly evident in Cilia’s aerial photograph of the site (2004, p. 142). 149 The site has been surveyed, and the plans are held in the University of Malta; N. Vella, personal communication, 2011). The plan presented in Figure 3.16, no. 1, was prepared by the author from Google Earth and photographic data. 150 Grima 2004, pp.  329–35; see also geological map relative to megalithic structures on p. 328; Torpiano 2004, p. 357. 151 Quarries close to megalithic monuments can be observed at Ggantija, Ħagar Qim, Mnajdra and Ras ir-Raħeb. For a discussion of quarries and quarrying: Clark 2004, pp. 372–3; Grima 2004, pp. 331–2 and 336. 152 Clark 2004, pp. 372–3. 153 Ibid., p. 373.

Notes to Pages 81–84 154 Stone balls as rollers: e.g., Zammit 1994, pp.  15–20; Pollini 1988, pp.  133–51, fig. 10; Torpiano 2004, p. 357. 155 Bedrock surface as the foundation of the megalithic structures, Grima 2004, p. 330. 156 Zammit 1980 (1929), p.  10; Bonanno 1988, pp.  104–7, fig.  1, depicting the numerous stone balls at Tarxien. 157 Pollini 1988, pp. 133–51, figs. 1–6. 158 Evans 1971, p. 66, pl. 66:6, inv. no. S/S.45, mallets from Ħal Saflieni; p. 146, pl. 66:5, inv. no. T/S.41, waisted hammer head from Tarxien; Grima 2004, p. 336. 159 Grima 2004, p. 336. 160 Traces of plaster on architecture: Zammit 1917, p. 272; Trump 1966, pp. 8, 9 and 17 (the Red Room at Skorba); Evans 1971, p. 175; Grima 2004, p. 337. 161 Representations of roofing occur on a small building model from Ta’ Ħagrat, pictured in Trump 2002, p.  192; Cilia 2004, pp. 152 and 370; on the interior of a rock-carved niche with two standing figures and a representation perhaps of a small pot from the Tarxien complex, Cilia 2004, p. 57 bottom, middle; Grima 2001, p. 51. 162 Reconstructions of the possible roofing are numerous; for example, Ceschi 1939; 1970; Piovanelli 1988, pp. 125–31, fig. 1 (for Tarxien); Pessina and Vella 2005, pp. 37–8; Cilia 2004, p. 152; Clark 2004, pp.  370–1; Torpiano 2004, pp.  355–7; Trump 1966, pp. 6–7; 1999, p. 96; 2002, pp. 196–7. 163 Clark 2004, pp.  371–2; Torpiano 2004, p. 357. 164 For grape-pressing pans, see Chapter  7. The stone beams are generally thought to be Punic (Cilia 2004, p. 142), but the masonry does not conform to the usual quality of Punic work. If these long and narrow blocks were shown to be Late Neolithic in origin, then it could be asked, do such blocks lie hidden in plain sight at locations such as the Punic sacred precinct of Tas-Silg, built over a megalithic complex? Cazzella et  al. (2012b, p.  602) document the reuse of some megalithic blocks in Tas-Silg. 165 Possible megalithic roof beams at Borg in-Nadur:  MAR 1955–6, p.  7;

Notes to Pages 84–89 photograph in Vella, Zammit and Bugeja 2011, p.  46, fig.  3:1, two long lengths shown lying parallel and were found under 45.7 cm of soil (beam 1: 4.24 m long × 30.4 cm wide × 22.8 cm thick; beam 2:  4.74 m long × 25.4  cm wide × 25.4  cm thick). Early and Middle Bronze Age sherds were found in the soil around the beams. Another possibility is another long stone reused in Bronze Age house 2 (Vella, Zammit and Bugeja 2011, p. 55, figs. 3:7 and 3:8). Although this did have a well-dressed rebate depression on the upper surface (p. 59, fig. 3:1), it may well have served a similar function. This stone measured 2.9 m long × 30 cm wide × 30 cm thick and was likened to the former long beams (p. 60, n. 42). The Kercem Road (Gozo) menhir is illustrated in Trump 2002, p. 187. 166 Tilley 1994; 2004, pp. 87–146. 167 Malone et  al. 2009, §14.7.1, pp.  380–3; Skeates 2010, pp. 222–4. 168 Imagination and empathy have been harnessed in the pseudo-anthropological recreation of the ancient processes of burial across the Gozitan landscape in Malone et al. 2009, §14.7.1, pp. 380–3. 169 When possible, Zammit lifted (and later replaced) stones, finding a diverse array of objects underneath (1920, pp. 192–4). 170 Renfrew 1973, pp.  152–9, fig.  33; further discussions on social organisation in Trump 2002; Cazzella and Recchia 2006, p. 699. 171 Although Renfrew’s 1973 essay focussed on five territorial divisions and listed only fourteen megalithic structures (fig.  33); current estimations identify some thirty buildings in Malta and Gozo; Evans 1959, p.  84; Grima and Vassallo 2008, p. 8. 172 Renfrew 1973, pp.  153–5; see also Renfrew and Level 1979, ‘x-tent’ model. 173 Clark 2004, p.  377; Malone et  al. 2009, p. 369. 174 Hassan 1978; Kolb et  al. 1985; Chamberlain 2006, 2009. 175 Xagħra Circle burial numbers, Malone 2009, p. 369. 176 Sherrat 1990, p.  150; Renfrew 1973, pp. 155–9; Evans 1977, p. 23.

345 177 Evans 1973a, p.  218; most of the bone material had been thrown away by the time Zammit took over the excavations. 178 Trump (1997, p. 14) speculates on other burial grounds – ‘as many as fourteen’ – yet to be discovered; Evans 1973a, p. 218. 179 Malone et al. 2009, pp. 365 and 369; tallies of individuals represented by skeletal material differ ranging from 350–450 to 450–850. 180 Murray 1923a, vol. 1, p.  25:  ‘They [the megalithic buildings] are built on the open hillside, visible on every hand especially from the sea.’ 181 Grima 2002, pp. 25–8; 2004, pp. 327–45; 2007, pp.  36–8; Grima and Vassallo 2008, p. 8. 182 Cf. maps indicating proximity to springs (Grima 2004, p.  341), to sea (Grima 2004, p. 344) and to plains (Grima 2004, p. 342). 183 C. Vella (2010, p.  3; 2011, p.  174) draws distinctions between coastal and inland sites; the former he considers richer in imported lithics because they were close to and dominated the points of entry to the island (e.g., Ras il-Pellegrin and Borg in-Nadur overlooking bays with good anchorage). 184 Grima 2004, p. 345. 185 Robb and van Hove 2003; Robb 2007, pp. 75–118; 2001, pp. 175–202. 186 Bradley 1998, p. 34. 187 Malone 1985; Perlès, 1992; Patton 1996, p. 151. 188 Micallef 2001, pp.  36–44; Ventura 2004, pp.  306–25; astronomical connections to the megalithic building orientations, Cox 2001, pp. 24–37; 2008, pp. 1.7–1.8. 189 Grima and Vassallo 2008, p.  8; Trump 2002, pp. 199–201. 190 Stoddart et  al. 1993, pp.  16–17; cf. Cazzella and Recchia 2007, pp. 67–8. 191 Bourdier and Minh-ha 1985, p. 29. 192 Bourdier and Minh-ha (1985) give plans of villages with designated structures for specific functions  – e.g., fig. D3:F is a shrine  – as well as men’s and women’s areas and places for storage and working. 193 Hodder 2006, 2007, pp. 105–20; Hodder and Cessford 2004, pp. 17–40. 194 Bourdier and Minh-ha 1985, p.  29; aspects of visibility in and around the

346 megalithic structures are conveyed in Anderson and Stoddart 2007, pp.  41–4, pls. 2–9; Gheorghiu 2001, 2003. 195 Malone 2007, pp. 31–2. 196 Ashby et  al. 1913, pl. 5 (Kordin III) depicting the cobbled courtyard, pl. 20 (Mnajdra), pl. 28 (Sant Verna, Gozo). 197 Malone et al. 2009, pp. 361–84. 198 An exposed block at what is now identified as the Skorba megalithic complex was originally called a menhir, Trump 2008, p.  27. A  possible alignment of three standing stones at Ta’ Bir Miftuħ is in an area where quarrying may have eliminated further remains, MAR 1991, p. 81. For Qala, Gebla l-Wieqfa (ll-Ħagra il-Wieqfa, Gozo), Trump 2002, p. 185. 199 Murray (1923a, vol. I, pp.  22 and 41)  describes stone pillars as ‘baetyls’ or bethel stones  – equating with the dwelling place of the deity  – which were found at Borg in-Nadur, (see also p.  25). Oracle stones were mooted for Kordin by Caruana (1896a, p.  30). See also Malone 2007, p. 28. 200 Hort 1921, pp. 167–8, fig. 2, who argued that the V-shaped holes were used to restrain animals awaiting sacrifice. Holes in stones identified as screen or door fitments, Anderson and Stoddart 2007, p. 43. 201 Bradley 1998, p. 37. 202 Evans 1971, p. 73, after Ashby et al. 1913, pp. 42–3. 203 The ruins on Corradino Hill yielded ‘plentiful’ bone from domesticates and shellfish, as well as pottery and hard-stone and flint tools: MAR 1908–9, §6, p. E3; ‘Charred bones and teeth of sheep, oxen, pigs, and dogs have been repeatedly picked up … in many of the chambers of Hagar Qim and Mnaidra … in abundance,” Caruana 1896a, p. 43. 204 Megalithic structures as distribution points for food and drink, Renfrew 1973, pp.  155–9 (tied into concepts of chiefdoms and some form of central control over economic production); Stoddart et al. 1993, p. 8. 205 Bourdier and Minh-ha 1985, pp. 17–19. 206 Hort (1921, pp.  167–8) argued that the minor door-jamb holes were being used to secure doors at the openings of the buildings.

Notes to Pages 89–91 207 Grima 2001, pp. 48–65. 208 Schuchhardt 1928, pp.  427–31; Murray (1923a, vol. I, p.  24) dismisses the idea that the apsidal buildings were habitations; Ugolini 2012 (1934), pp.  305–9; Scarre 1998, p.  73:  ‘Perhaps we should emphasize here that there is no doubt they were temples – there is no domestic debris, for example, to suggest that anybody lived in these impressive complexes.’ Fenech (2007, p.  109) suggests that building megalithic structures may have been a response to failure of past, more flimsy habitations and sacred spaces after extreme weather conditions. 209 Zammit (no date), p. 5: ‘A mass of disjointed blocks are [sic] seen some still standing in their place and some disarranged. Originally, these must have formed a series of chambers, possibly dwelling places of the attendants of the temple.’ Murray 1923a, vol. 1, p. 25: ‘It is probably true that the buildings round them may be priests’ houses though at present there is no exact evidence to that effect.’ Superintendence of Cultural Heritage (2008, §3.1.1) reported further ‘post-holes, rock-cut trenches and levelled surfaces’ around the monument as well as a ‘large circular rock-cut feature of unknown function’ with in situ prehistoric stratified deposits which lay to the north of the Ħagar Qim complex, found in 2008 in preparation for the installation of the protective canopy. 210 Ashby 1910, p.  158:  ‘at Mnaidra [sic] it was found that besides the two main parts of the structure there were some subsidiary buildings, which, though less massive, were of considerable importance; they were perhaps devoted to domestic uses, inasmuch as a very large quantity of pottery was found in them.’ 211 Banning 2011. 212 Finds of grinding stones in the prehistoric sites, e.g., Evans 1971, p. 35, from Ta’ Ħagrat, inv. no. Mg/S.11, querns, grinding stones, pounders, hammer stones, ‘digging-stick’, stone weights and abundant lithic material; from Kordin III; p. 73, a trough with deeply ground pits and an upper grinding stone was found in the complex; from Ħagar Qim, p. 93, see ‘Tools and weapons’, where listed are

Notes to Pages 91–95 whorls, mortars, grinding stones, pestles, pounders, sling stones, pumice stones and lithic material (‘a large number of implements’); from Qortin l’Mdawwar (a.k.a. Kuncizzjoni), p.  109, a quern, grinding stones, mortar and small pebble grinders. Most of these artefact categories from the prehistoric sites have not been analysed in any depth; cf. Malone 2007, pp.  23–34, where ritual and cult are central to interpreting the artefacts found in the megalithic structures. 213 Malone et al. 2009, p. 369. 214 Criteria used for determining ritual in the archaeological record:  Renfrew 1985; 1994, pp.  47–54; Verhoeven 2002, concerning ritual framing ‘defined as the way in which people and/or activities and/or objects are set off from others for ritual, non–domestic, purposes. Ritual framing is mainly achieved by creating a special place, a special time, and by the use of unusual objects: e.g., a stage is set up, special clothes are worn, distinctive objects are put on display’ (p. 6); Malone 2007, pp.  23–34; Barrowclough 2007, pp. 45–53; criteria have been captured in table 8:1. 215 Verhoeven 2002, p. 235. 216 Stoddart et al. 1993, p. 3. 217 Skeates 2010, pp.  84, 131–2, 139–40, 145–6, 154–5, 196–7, 205–7, 214–16, 222–4 and 235–6. Recent interpretations have ventured into virtual reality: Flynn 2000, 2005a, b, 2007. 218 ‘Melita’ as a Greek term associated with honey: Mayr 1909, p. 23; Seltmann 1946, p.  82; Culican 1971, p.  2. One instance of honey use in the late Punic-Roman period has been identified at Tas-Silg through residue analysis, Notarstefano 2012, p. 127. For the Punic word ‘nn, see Chapter 6. 219 Bonanno 2005, p. 147. 220 Tilley 2004, pp. 87–146. Malta’s possible production of honey or at least trading in honey during the classical period is not certain, Cassia 2008, pp. 149. 221 C. Vella (2010, p.  3) argued for distinguishing between the megalithic monuments based on the lithic assemblages: ‘If even at the material culture level a difference in relative “richness” and function is observed, then it should seriously

347 be considered that there is the possibility that not all monuments were meant for the same rituals or purposes.’ 222 Banning 2011, pp. 619–60. 223 Ibid., p. 625. 224 Scarre (1998, p.  80) speculates on the use of mind-altering substances in temple ritual:  ‘When fires or aromatic herbs (perhaps even hallucinogens) were burned in hearths or braziers’; Vella Gregory 2005, pp.  34 and 62–71; Savona Ventura 1999, pp.  110–15; concerning the Xagħra Circle peg figures as shamanistic: Cilia 2004, p. 283; Malone et al. 2009, p. 298. 225 Eliade 1972; Bahn 2010, pp.  75–8; Lewis-Williams and Dowson 1988, pp.  201–45 (with replies, pp.  217–38); 1993, pp.  55–65; Lewis-Williams and Pearce 2005. Renfrew (2007, p. 11) also questions whether shamanism should be used in the explanation of the archaeological record. Furst 1972; Pearson 2002; Price 2001; Sherrat 1991, 1995; Wallis 2003. 226 Stoddard et al. 1993, p. 3. 227 Studies of Palaeolithic cave art and shamanistic practices:  Lewis-Williams and Dowson 1988; Lewis-Williams 2002a, b; Dronfield 1993, 1995. 228 Discussions concerning wide-ranging artistic expression perhaps stimulated by trance states and narcotic use in antiquity, e.g., Stein 2006, pp.  19–34, figs. 9–75; 2011. 229 Oppitz 2004, pp. 20–3. 230 Sacks 1995, pp. 51–98, pls. 1–7. 231 Tilley (2007, pp.  129–30) discusses the effects of light deprivation in connection to Tarxien. 232 Sagona and Sagona 2011, concerning the magic mushroom Amanita muscaria of the Magi. 233 Evans Schultes 1972, pp. 3–54, for a synthesis of ‘mind-altering’ substances used particularly in New World contexts; Goodman, Lovejoy and Sherratt 1995a. 234 Blamey and Grey-Wilson 1993, §29, p.  30, Joint Pine (Ephedra fragilis) is listed and well-illustrated on Malta Wild as a very rare indigenous plant, protected by law:  http://www fragilis.php (accessed 5 October 2013).

348 235 Sherratt 1996. 236 Beer drinking, Sherratt 1987, 1995a. 237 Cook et al. 2008, p. 101; Devereux 2009, pp. 225–31. 238 Cook et al. 2008, p. 96. See also Devereux and Jahn 1996; Devereux and Jahn 2003. 239 Grima and Vassallo (2008, p.  6) have discussed the natural acoustic qualities of the Mgarr Valley in relationship to the Ta’ Ħagrat monuments. Acoustic qualities of ancient buildings:  Watson 2001; Watson and Keating 2003; Cook et al. 2008. 240 For the stone bowl as a possible drum:  Malone et  al. 2009, p.  192; Stoddart 2002. 241 Large stone vessel (perhaps a drum) in the Tarxien megalithic building, Zammit 1917, p. 269, pl. 33, figs. 2 and 3. 242 Earth drums are discussed in Kunda 1979, p. 58; Blades 1970, p. 44. 243 Reichel-Dolmatoff 1972, p.  91:  ‘It is oblique in my field of vision. It goes from top left to bottom right … now this whole scheme tilts, and moves about … almost 45 degrees … Hexagons, all like a ceiling full of hexagons, some tilt 35 degrees.’ 244 Sagona and Sagona 2011, pp.  394–404; Reichel-Dolmatoff (1972, p.  111) refers to the Tukano of north-west Colombia, in the Amazon region, using the drug Banisteriopsis caapi, an Amazonian jungle plant, the bark of which is used to prepare a narcotic drink; Evans Schultes 1972, pp. 33–40. 245 Lewis-Williams and Dowson 1993, p. 55. 246 Ibid., P.  56; cf. Helvenston and Bahn 2003, pp.  213–16 and 220–3; note responses to Helvenston and Bahn by Bradshaw 2003, p. 216; also reviewed by Chippindale 2003, pp. 218–19. 247 Only a few very poor paintings are preserved in Għar Ħasan; see Chapter 2. 248 Vella Gregory (2005, p. 34) drew a parallel between the stick figures on Zebbug pottery with ‘no specific human form’ and shamanistic representations in South African contexts. 249 Polished head emerging from stone from Tarxien, Evans 1971, pl. 49:5; other indistinct human depictions from Mnajdra, Evans 1971, pl. 41:18; a virtually

Notes to Pages 95–104 featureless hand from Ħagar Qim, Evans 1971, pl. 41:11–12. 250 Bird pendants: Zammit et al. 1912, p. 11, pl. 9:12; Evans 1971, pl. 37:9–10. 251 Anthropomorphic snail figurine from Xagħra Circle, Vella Gregory 2005, pp. 11 and 96–7. 252 Malone et al. 2009, p. 304. 253 Vella Gregory 2005, pp. 72–5. 254 Obese headless figures, Evans 1971, pls. 39:17–20 and 40:1–9; from Ħagar Qim, interchangeable heads, Vella Gregory 2005, p. 20. 255 Evans 1901, p. 198. 256 Ibid., pp.  196–200; Dukinfield Astley 1910, p.  157; cf. Renfrew 1973, who argues against the old diffusionist theory on the basis of the 14C dates. 257 Evans (1973a, p.  217) also entertained the idea that the figurines represented ‘a rite actually carried out within the Hypogeum, in which “priestesses” attempted to induce prophetic dreams by sleeping among the remains of the ancestral dead.’ 258 Evans 1971, pl. 37:3–4. 259 Heads titled backward and expressionless faces, Evans 1971, pl. 36:13 (from Ħal Saflieni), pls. 48:14 and 49:12 (from Tarxien); Vella Gregory 2005, pp.  26–7 and 40–1, examples from Xagħra Circle and from Tarxien  – the author considered them to be gazing at the sky. 260 Thompson Drewal (1986, p.  66) of the African Yoruba ritual possession in trance:  ‘They lick their lips in an agitated fashion. Their upper torsos drop, the heads roll back, and their eyes roll upward’;Thompson Drewal 1992, p. 183. 261 Body theory:  Boric and Robb 2008; Vella Gregory 2005. 262 Vella Gregory 2005, pp.  20 and 180; Townsend 2007, p. 73; Monsarrat 2004. 263 Fleming 1969, p.  251; see also Meskell 1995; Lesure 2002. 264 Figurines interpreted as mother goddess, Gimbutas 2001. 265 Gimbutas 2001, p. 95. 266 Human faces on pottery in Malta, Vella Gregory 2005, pp.  174–5, figs. 169–71 and 173–4. 267 Fleming 1969, pp.  247–61:  ‘The idols have become articles of faith for those

Notes to Pages 104–107 who believe in mother-goddess hypothesis; but were they such for their makers?’ (p. 252); contra, for example, Daniel 1958. Discussions on the earth mother concept as it has been applied to Malta: Hutton 1997; N. C. Vella 2007b, pp.  61–71; Townsend 2007, pp.  78–80; Rountree 2003, pp. 25–43; Malone 1999, pp. 148–63. 268 Evans 1973, p. 219. 269 One embracing couple figurine came from the Tarxien megalithic building; Vella Gregory 2005, p. 142. 270 Explicit depictions of breasts or genitals, Vella Gregory 2005, p.  112–14, fig.  86 (from Mnajdra). 271 Figurines depicting pregnancy, Vella Gregory 2005, p.  22, fig.  86 (from Manajdra). 272 A possible infant on one lap of a pair of figures from Xagħra Circle, Vella Gregory 2005, pp. 56–61, fig. 28. 273 Twists of clay from Mnajdra, Vella Gregory 2005, p. 115, figs. 87–91; cf. legs on other figurines, Vella Gregory 2005, p. 105, fig. 84, p. 172, fig. 163. 274 Torso figurines, Vella Gregory 2005, pp. 116–17, fig. 92 (from Tarxien), fig. 93 (from Mnajdra). 275 Legs and feet figurines, Vella Gregory 2005, p. 173, fig. 158. 276 Figurine pierced by numerous shell fragments at sixteen points on the body from Tarxien, Vella Gregory 2005, pp. 106–11, fig. 85, also fig. 86 (from Mnajdra). 277 Animal figurines from Ħal Saflieni, so poor in production that they are described as ‘wretched’, Zammit et  al. 1912, pp. 6–7, pl. 6:2, 8. 278 Human-headed or anthropomorphic pillars, Vella Gregory 2005, p.  171, figs. 154 and 155, with noses from Santa Verna and Mnajdra. 279 Vella Gregory 2005, pp. 72–5, figs. 38–44. 280 Single phallic-like pillars, Vella Gregory 2005, pp. 166–69, figs. 142, 144–155, from Tarxien,Tal-Qadi, Mnajdra, Ggantija and Xagħra Circle (Gozo); Bonanno 1992b. 281 Paired or triple ‘phallic-like’ pillars,Vella Gregory 2005, p. 167, fig. 143a, b, p. 170, fig. 153. 282 Murray 1923a, pp.  22, 25; cf. Vella, Zammit and Bugeja 2011, p. 53.

349 283 Ridley (1976, p. 25) discusses depictions of grasses in terms of the worship ensuring the harvest/food-procurement cycle and comparing plant depictions with those of the Aegean. 284 Townsend 2007, pp. 72–80. 285 Figurine with plait from Ħagar Qim, Vella Gregory 2005, p. 21. 286 Hairstyles symbolic of ritual in shamanistic practices of Yoruba ritual (West Africa): Thompson Drewal 1986, p. 67; 1992, p.  182, a shaven head gave the appearance of an expanded head, representing the alluded-to effects of the trance state. 287 Evans 1901, §29, pp. 196–200. 288 Vella Gregory 2005, pp. 137–138, fig. 109. 289 Figures sheltering under the skirt of a greater figure or deity:  Zammit 1930c, pp.  53–5; Evans 1971, pl. 48:3–5 (from Tarxien); Vella Gregory 2005, pp. 136–9, figs. 109 and 110. 290 Reverence for ancestors in regards to prehistoric Malta appears in Malone et  al. 2009, p.  353; comments by Colin Renfrew. 291 Figurines holding discs from Xagħra Circle, Gozo, Vella Gregory 2005, pp. 18 and 40 (right), 41 and 78 (fig.  56), 80 and81 (figs. 55 and 56), 82 (fig.  57), 86 (fig. 64) and 91 (fig. 75). 292 Lined or scored torso on figurines from Xagħra, Vella Gregory 2005, pp.  23, 24 and 40 (right), 87 (fig. 67), 89 (fig. 70) and 90 (fig. 73); from Tarxien, Vella Gregory 2005, p. 116 (fig. 92); from Mnajdra,Vella Gregory 2005, p. 117 (fig. 93). 293 Goat horns at Skorba, Trump 1966, pl. 18:a, from the North Room floor in area YD, pl. 28:b, from the so-called shrine. 294 Spirals and running curl designs as representing water, Grima 2001, pp. 55–9. 295 The meaning of spirals and curling patterns,Vella Gregory 2005, p. 102. 296 George Augustus Robinson’s journals can be found in Plomley 1966; visits to plentiful ochre deposits on Swan Island (off the north-eastern tip of Tasmania) and the Tasmanians’ song, dance, storytelling and courtship, pp. 274–84 (15–27 November 1830). 297 Plomley 1966, p. 895 (3 July 1834).

350 298 Concerning the preparation for and visit to the Toolumbunner ochre mine in the Gog Ranges, Tasmania, Plomley 1966, pp. 895–908 (3–25 July 1834); Robinson had been informed of the mine as early as 1931 (12 December 1931), pp. 549 and 582, n.  75; see A.  Sagona (1994) for an account of the rediscovery and excavations carried out at the Toolumbunner ochre source. 299 Red-painted statues and figures from Ħal Saflieni, including the ‘sleeping lady’, are clearly illustrated in, for instance, Vella Gregory 2005, pp. 48–52, figs. 20 and21, pp. 56–7, fig. 28. 300 Attard Montalto et  al. (2012, p.  1101) identify ochre sources in Malta, but they note that some of the colorant could have been sourced offshore:  ‘It is possible that similar materials from a near-identical geological environment (e.g., Sicily perhaps?) may have been an alternative source.’ This is certainly suggested on the grounds of cultural artefacts discussed in Maniscalco 1989. 301 Maniscalco 1989. 302 Ibid., p.  541. It has been demonstrated that ochre can be found in the Maltese Archipelago: Attard Montalto et al. 2012, pp. 1094–102. 303 Skeates 2002, p.  13; Barrowclough (2007, pp.  49–53) argued that greenstone axes ‘were worn by adherents of a cult’ (p. 50); Leighton and Dixon 1992, pp.  179–200, for a discussion of stone sources in Sicily and southern Italy. Murray 1929, pp.  28–30, also discusses the pairing of buildings in Malta. 304 Malone et al. 2009, §14:3, p. 363. 305 Trump 1966, p.  47, in regard to the motivation for structural changes to the megalithic buildings that they were ‘in accordance with some change in ritual’. Evans 1971, p. 81, referring to at Ħagar Qim:  ‘The main group is irregular in shape and plan, and gives the impression of a spontaneous growth as more and more rooms were needed for the elaborate worship.’ 306 Barrowclough 2007, p. 53. Murray 1923, p.  23, records three standing upright stones in the small apse. 307 Trump 1966, p. 47, table 2, revising Evans 1959, pp. 84–134; Evans 1971, p. 218.

Notes to Pages 107–113 308 Malone et al. 2009, p. 342, table 12.1, for new radiocarbon determinations; Vella, Zammit and Bugeja 2011, p.  53, n.  36, recognised the need for revision of the megalithic building period. 309 Evans 1971, pp. 215 and 218. 310 Bourdier and Minh-ha 1985, pp. 45ff. 311 Sherratt 1994, p. 199. 312 Ibid. 313 Grima 2008, pp. 49–54. 314 Trump 1966, p. 47. 315 Grima (2004, p. 337) opts for a utilitarian rather than symbolic approach by the Maltese builders to their building materials; cf. Pearson 1998, discussion of the symbolic quality of the stone itself in the construction of British monuments drawing on ethnographic parallels with Madagascan communities. 316 Vella 2008b, §2.2 (not paginated). 317 Cilia 2000, pp. 11–22. 318 Notice of the conservation and archaeological work carried at Ħal Saflieni appeared in MAR 1991, pp. 81–2; 1992, p. 80; Pace 2000, 2004. 319 Malone et al. 2009, p. 369. 320 Magro Conti ca. 1997; MAR 1973–4, §7, p. 51. 321 See ‘Discussion’ by Zammit in Tagliaferro 1912, p.  150; 1911, pp.  147–50; MAR 1910–11, p. 4; Evans 1971, pp. 40–1; Cilia 2004, pp. 144 and 203, see plan, p. 144. 322 Tagliaferro 1911, p. 149; 1912, p. 147. 323 Tagliaferro 1912, p. 147, the long bones ‘as a rule, found lying horizontally in the direction of the axis of the cave (E.N.E.) or in a perpendicular direction, were in some cases lying without any order, or even heaped pèle mèle’; Tagliaferro 1911, p. 149. 324 Ashby et al. 1913, p. 12; Tagliaferro 1912, p. 148. 325 Tagliaferro 1911, p. 149. 326 Tagliaferro 1912, p. 148. 327 Malone et al. 2009, pp. 364–84. 328 An altar at Ħagar Qim has a potted-plant motif repeated on all four sides:  Evans 1971, inv. no. Q/S44, p. 93, pl. 41:8; rows of domestic animals from Tarxien, pls. 17:5–6 and 18:3–4. 329 Skeates 2007, p.  93; Whitehouse 2007, pp. 97–105. 330 The persistent notion of death cults and the megalithic buildings imitating

Notes to Pages 113–121 tombs was first suggested by Evans 1959, pp. 90–1. Chapter 4  Pushing Boundaries at the End of the Megalithic Building Period 1 Grima and Vassallo 2008, p. 6. 2 Reasons for change at the close of the Late Neolithic period: ‘ruthless extermination’ of the inhabitants, Evans 1959, pp. 168–9; issues of aggression are discussed in Magro Conti 1999, pp.  191–208; environmental disasters and over-population: Bonanno 1993a, p. 44; Trump 2002, pp.  238–41; internal social unrest, Bonanno et al. 1990, pp. 202–3. 3 The notion of complete collapse is encapsulated in Evans 1971, p.  224; 1973a, pp.  218–19; Bonanno 1993a, pp.  35–45; 1993c, p. 224; 1986a, pp. 40–1;Trump 1976. Cf. discussions concerning the nature of collapse in ancient societies in Yoffee and Cowgill 1988. 4 Evans (1971, p. 152) is a case in point: ‘The grey pottery is, of course, quite distinct from any of the native wares in use in Malta during the Copper Age and technically has more in common with the Cemetery pottery, so that in 1953 I treated it as belonging to the Cemetery’ (Evans 1953, p. 68); also Trump 1966, p.  46, of Thermi Ware, ‘Sherds previously were nearly always associated with TC [Tarxien Cemetery] material, though never in closed contexts.’ 5 Concerning cultural resilience:  Tainter 1988, 2006; Yoffee 1988; McAnany and Yoffee 2010; and concerning ‘overshoot and collapse’:  Catton 1980; Diamond 2005; also Ellenblum 2012, p. 18. 6 While pointing out that such linear scars were the result of agricultural practice, I originally interpreted them as evidence of a failed economy, as much due to over-exploitation of marginal land as to a downturn in environmental conditions.While maintaining the view that the field scars are agricultural artefacts, I  have modified the socio-economic argument I  put forward in 2004 (Sagona 2004b) largely because of the research undertaken for this book. 7 Tainter 2006, p. 72. 8 Concerning rock-cut furrows (a.k.a. ‘cart ruts’), the principal recent discussions with historical overviews and bibliography concerning the topic appear in MAR 1959–60,

351 p. 2; Hughes 1999; Magro Conti and Saliba 1998, 2007; Cardona 2008a, pp. 48–9, 68–70; Trump 2004b, pp. 379–97; Weston 2010. 9 Cf. Hughes (1999, §2, p. 66), who sees the ruts as unique to Malta. I  am grateful to Ronnie Gallagher, who shared his extensive knowledge and photographic record of the rock-cut furrows in Azerbaijan after my article on the topic appeared (Sagona 2004b); see discussion of other rock-scarred areas in Magro Conti and Saliba 2007, pp. 169–76. 10 Sagona 2004b, pp. 45–60; Cardona 2008a, pp. 68–70. 11 Concerning ancient vehicles, harnessing and animal traction:  Piggott 1983, 1992; Sherratt 2006; A. Sagona 2013. 12 Wheeled-vehicle development, Bakker et  al. 1999, pp.  778–90; A.  Sagona 2013, actual preserved wooden wagons in burial mounds (kurgans) and clay models ca. 3500–2500 BC. Mischka 2011, pp. 329–758, for early evidence for wheel vehicles (tracks 1.10– to 1.20-m apart) at 3420–3385 cal BC in Flintbek, Germany. 13 See, for example, Roman road technology in Turkey:  French 1980, 1983, 1998; Mitford 1998; cf. arguments favouring Roman-date scars in Malta:  Bonanno 1994; Magro Conti and Saliba 1998. 14 Quarries cutting pre-existing furrows north-east of Skorba, Mgarr:  Cardona 2008a, §§8A and 8B, pp.  49, 66–7, figs. 4 and 5; at North Baħrija, Weston 2010, §27, p. 66; east of Nadur tower, §31, p. 72; Ta’ Baldu, §40, pp.  87, 161; Misraħ Għar il-Kbir (a.k.a. ‘Clapham Junction’), §44, pp. 96–7. 15 Cf. Degree of wear on the sharper quarried edge than on nearby furrows in Weston 2010, pp.  72–3, §31, near Nadur Tower; on the isolated pedestal of stone left after quarrying at Misraħ Għar il-Kbir, §44, p. 97. 16 Cf., e.g., photographs of Neolithic tomb 5 (Figure 3.7, no. 1) and nearby furrows at Xemxija in Weston 2010, §3, pp. 26–7. 17 Furrows cut by Phoenician-Punic tombs occur at Mtarfa, Magro Conti and Saliba 2007, p.  75, figs. 78 and 79; also Trump 2004b, p. 391; Cardona 2008a, p. 70, fig. 37; Weston 2010, Mtarfa, §35, p. 80; Tat-Targa, §36, p. 82; also discussion pp. 163–8. 18 See Weston (2010, §35, Mtarfa, p.  80), who argues that the groove overruns the

352 silo despite the obstacle this would have presented for any vehicle (which is the premise he supports); a bell-shaped pit at Birzebbuga also cut one rock-scar line, Magro Conti and Saliba 2007, pp.  88–9, fig. 95. 19 For discussion of the chronology of the ruts: Bugeja 2001; Sagona 2004b, table 1, pp.  47–8. Furrows close to prehistoric megalithic sites are noted in Cardona (2008a, pp.  48–9) and Weston (2010, pp. 152–71). 20 Mottershead et  al. 2008, pp.  1065–79, ‘Geomorphological evidence suggests that the trackways were formed during a period of soil erosion.’ Others also accept that erosion has played a part in the appearance of the grooves; Dawkins (1918) looked to a natural cause through weathering of the soft limestone; Mottershead, Farres and Pearson 2010; Weston 2010, pp. 172–200. 21 Pedley 2007, pp. 56–9, concerning limited erosion at Misraħ Għar il-Kbir of approximately 1.5 cm in the floor of the furrows since their manufacture, pp. 57, 59 and 70; on tool marks (‘mason’s marks’), pp.  64, 68–9 and 71; on regular use of the furrows, pp. 64–6; on lack of soil cover in the grooves when in use, pp.  64, 65 and 71; also Pedley et al. 2002. 22 Some documented anomalies:  furrows intersecting and flowing into dissolution (or doline) pits; at Birzebbuga, Magro Conti and Saliba 2007, pp.  87–8, figs. 92–95; at the limits of Naxxar, T’Alla w’Ommu, San Pawl tat-Targa, Magro Conti and Saliba 2007, p. 82, fig. 85; also Trump 2004b, p. 392. Weston (2010) does document some associated features:  furrows cutting doline pits at Manikata (§5, p. 30), one clipping the edge of a pit and others running into a pit at Naxxar (§18, pp. 51, 52 and 186); channel cut between furrow and cut pit at Naxxar (§18, p. 174); clustered cuttings that flow into three natural basins, Kalkara (§39, pp. 85 and 192); anomalous modifications to furrows at Port Burmarrad (§8, p.  33); at Tat-Targa, (§36, p.  175); single groove (i.e., not paired) at Busewdien (§13, p. 42). Furrows over stepped terrain, Trump 2004b, p. 393, at il-Qallilija, Mtarfa and Dingli cliffs;

Notes to Pages 121–122 p.  395, at San Pawl tat-Targa, north-east of Skorba (in an archaeological buffer area). Cardona 2008a, §6 and §8 b, p. 49, at one site there are two areas with different furrow spacing: §6 at 1.35 m apart; a few metres east is another group 1.45 m apart (Cardona notes that this is more in line with other sites); figs. 4 and 5, furrows (§8b) are clearly cut by quarry (§8a). Furrows truncated at cliff edges:  Magro Conti and Saliba 2007, Qala Pellegrin, limits of Mgarr, p. 216, NW_0039; Weston 2010, at Ta’ Baldu, §40, p. 87; Għar il-Kbir, §43, p.  91, where the cave roof has collapsed, cutting the furrows; Għar Zurrieq, Trump 2002, p.  268, for grooves that ‘appear to take off into space’; Trump 2004b, p.  385, at Ta’ San Gakbu, above Mtaħleb valley; p. 389, at Qala il-Pellegrin; near Manikata, ‘ruts appear to stop at the lip of the cliff. … The loads they carried presumably completed their journey by rope’, Trump 1997, p. 146. 23 Mottershead, Farres and Pearson 2010, p.  228; also Mottershead, Pearson and Schaefer 2008. 24 Schneider (2001, p. 16), among others, also has identified the unworn beginning of furrows showing clear signs of pecked tool marks; also Zammit 1928, p.  252; Evans 1934, pp. 339–40; Trump 1997, pp. 33 and 145; Hughes 1999, pp.  70–1; Cardona 2008a, p. 70; at Misraħ Għar il-Kbir, Magro Conti and Saliba 2007, p. 57, fig. 47, tool marks in long, curving shallow gouges are likely to be plough marks; also p. 68, for tool marks seen as evidence of deliberate manufacture and re-cutting furrow lines. Contra Western 2010, pp. 172–3, who found no such evidence. 25 Project Director Hermann Bonnici (Chief Architect, Restoration Directorate of the Project Design and Implementation Department [Restoration Unit], Ministry for Resources and Rural Affairs); principal partners in the Significance of Cart-Ruts in Ancient Landscapes Project (carried out October 2004 to September 2005)  were from Heritage Malta, the Restoration Unit, the University of Malta and the Malta Environment and Planning Authority (MEPA) with APORTECO of Spain (such features are in Caminos de

Notes to Pages 122–126 los Molinos in Padul). Although detailed maps were made of the case-study area in Malta, most of the plans in the publication are not at the same resolution (Magro Conti and Saliba 2007). Another volume, by Gordon Weston (2010), is worthy of mention, listing fifty sites with some recognition of associated features. 26 Omission of other features found in close association with rock-cut furrows certainly changes the archaeological interpretation of them. This was made very clear in Magro Conti and Saliba’s two versions of one major suite of features at Misraħ Għar il-Kbir. In 1998, it was drawn with three quarried areas (of likely Roman or later date) clearly cutting pre-existing furrows, but in 2007, the plan was re-drawn curiously omitting the quarries, and no mention of quarries was made in the description of this cluster (as was the case, e.g., with their entry for Qallilija, l/o Rabat [2007, p. 222, NW_0062]), nor is the cutting of furrows by quarries made clear in the discussion, pp. 129–33; cf. Magro Conti and Saliba 1998, fig. 1, and 2007, p. 223, entry NW_0064, Misraħ Għar il-Kbir, l/o Rabat. This sequence of two quite different and unrelated ancient activities (furrow and quarry) has direct bearing on the chronology of the features and on their function. In addition, the detailed plan (MEPA 2005, scale 1:1000) of the same area gives quite a different impression of the site; cf. Magro-Conti and Saliba 2007, p. 18, map 1, and p. 48, map 6i; that given in Mottershead et al. (2008, p. 1068, fig. 3) is different again, and none of the plans tallies with the scars captured in Google Earth images dated 15 April 2013. 27 Mottershead et al. 2008, p. 1068, ‘there is widespread agreement that the paired ruts are ultimately related to the passage of vehicles.’ 28 There are many documented examples of field scars in archaeological and ethnographic contexts that offer excellent parallels to farming and irrigation practices:  e.g., Smith et  al. 1968, pp.  353–67, Lake Titicaca region; Megaw et  al. 1961, pp. 200–15; Bradley 1978, p. 41, Gwithian, Bronze Age Beaker culture settlement, Cornwall; Barker 1985, 2005.

353 29 Stoddart et  al. 1993, pp.  8–9; Richards et al. 2001, pp. 253–6. 30 Sagona 2004b; O’Dowd 1994; O’Sullivan 1977. 31 Robinson 1986, p.  123. Traditional Irish farming techniques, Bell and Watson 1986. 32 Pedley 2007, p. 68. 33 Gregg 1988, p. 145. 34 Crop yields after Overton and Campbell 2006, table VIII. Maltese Neolithic diet, Richards et al. 2001. Zorn (1994, pp. 31–48) gives a useful survey of strategies which can be taken to estimate population sizes and associated economic issues. 35 Population estimates have been attempted in the past:  Renfrew 1973, pp.  153–5; Grima (2008, p.  52) estimated that ‘the core agricultural areas could together support a maximum of around 4,700 persons’, a tally determined 70 per cent of the islands’ landmass (p.  51), not factoring in the possibility of reclaimed marginal lands, as suggested here. 36 Trump 1997, p. 122; see examples of abrupt cessation of furrow lines at Tal-Palma in Weston 2010, §12, p. 40. 37 Applebaum 1954, pp. 107–14. 38 Keith Buhagiar (2007, pp.  103–31) has shown evidence of water management in historical contexts in Malta. 39 Mottershead et  al. 2008, pp.  1074–7; cf. rock-cut grooves as field furrows, Sagona 2004b, pp. 45–60. 40 Mottershead et al. 2008, p. 1070. 41 Information concerning the Tal-Barrani, Zejtun, finds outside of popular media: MAR 1965, p. 4; 1966, p. 6; 1993, pp.  75–6; for illustrations of tombs and field furrows, the latter identified as vine or olive grove trenches, see (accessed March 2013). Pace et  al. 2012, pp.  66–7, fig.  4. For the notion that these rock-cut grooves are the relics of various periods, Cardona 2008a, p. 69. 42 An excellent example of parallel field scars running in different directions in three adjacent fields, within standing dry stone walls, can be seen in the east of Gozo, at the tip of Tal Blata (location of the central field 36°1′53.31″N, 14°19′23.53″E, elevation 79 m).

354 43 Zammit 1928, p. 18; Evans 1971, p. 2020; Trump 2002, p. 266. For plans and illustrations of clustered furrows, Magro Conti and Saliba 2007:  Tal-Wej, limits of Naxxar (p.  197, C_0019); Xewkija (p.  199, GZ_0029); L-Imgiebaħ, limits of Mellieħa (multiple entries for the general area:  pp.  211, 270–4, NW_0002, NW_0255, NW_0256, NW_0257, NW_0258, NW_0261, NW_0262); Tal-Faccol, Zebbieħ limits of Mgarr (p.  220, NW_0052); Qallilija, limits of Rabat (p.  222, NW_0063); Misraħ Għar il-Kbir, limits of Rabat (p.  223, NW_0064); Ix-Xagħra ta’ l-Għattuqa, limits of Rabat (p. 226, NW_0073); Għar Ilma, Mtaħleb, limits of Rabat (p.  232, NW 0090); Wied il-Busbies, Mtaħleb, limits of Rabat (p.  233, NW_0093); Ix-Xagħra tax-Xini/Bieb ir-Ruwa, limits of Baħrija (p.  234, NW_0095); Ta’ Paris, limits of Rabat (p. 235, NW_0096); Gnien Buswedien, limits of Mgarr (p. 237, NW_0122); L-Imsieraħ ta’ l-Imtarfa, Mtarfa, limits of Rabat (p. 247, NW_0205); Ix-Xagħra tal-Qortin limits of Mellieħa (p.  248, NW_01210). Sites in Trump 2004b:  Laferla Cross Hill, San Lawrence, pp. 380 and 383; Bisqra, Xemxija, Zebbieħ and Qalillija, p.  382; Ta’ Mrejnu (Mgarr), the Nadur clusters north and south of Bingemma road and Mtarfa, p. 383. 44 Cf. Weston 2010, pp.  189–205. Cardona (2008, p. 60) comments, ‘It is questionable how any type of vehicle could pass over the same line often enough to create a considerably deep rut’; however, the suggestion that a feature such as a permanent wall guided ancient vehicles to keep them on the tracks that resulted in deep grooves is a superfluous, circular argument. 45 Proponents of vehicle use do acknowledge some field scars, Magro Conti and Saliba 2007, p. 123, fig. 137a–b. 46 Renfrew 1973, pp.  152–5; Richards et  al. 2001, p. 253. 47 Cf. The first attempt at defining population numbers, social structures and territorial boundaries in prehistoric Malta by Renfrew (1973, pp. 147–66). 48 One exception is Bugeja (2011, pp. 21–3), who discusses farmers levelling areas to produce new fields at Borg in-Nadur,

Notes to Pages 126–129 as depicted by Jean Houel in an eighteenth-century drawing (fig. 2:2). 49 Anon 1830, p. 154. 50 ‘A.’ 1838, p. 782. 51 Sagona 2004b; O’Dowd 1994; O’Sullivan 1977. 52 Cf. Reconstructions of possible vehicles for which there is no archaeological proof:  Lewis (1977, p.  63, fig.  14) reports having found worn stones (plough shears?), although their whereabouts have not been stated; Trump 2002, pp.  284–5; 2004b, p.  388; Magro Conti and Saliba 2007, p.  104; Weston 2010, pp.  113 (sled), 114 (travois) and 118 (cart). 53 Irrigation and other agricultural practices were observed by the author in the rural sectors of the highlands of eastern Turkey during the course of twenty years of field work. Weston (2010, pp.188–9) has documented one case where a furrow has been permanently blocked by a stone. Rather than accidental (‘a piece of loose stone appears to have fallen into and partially blocked a rut’, at San Gwann, §24, p. 189), it may be rare surviving evidence of water manipulation via a cut channel (or furrow) in antiquity. Indeed, Weston comments on the stone ‘forming a dam against rainwater’ (2010, p.  188). Longer single or double furrows that cross furrow clusters or travel across greater stretches of land may have functioned as water channels, at Il-Qattara, Magro Conti and Saliba 2007, p. 80, fig. 83 (single cutting); others are described as water channels, mentioned but not illustrated in their catalogue, e.g., Ta’ L-Imsaqfa, limits of Rabat (p.  239, NW_0131). Mottershead et  al. (2008, p.  1066) have classed the clusters as ‘regional route nodes’ despite the lack of settlement or other occupation feature that might concentrate such scars in one area. 54 Marriner et al. 2012, pp. 10–12. 55 Ibid., p.  10, fig.  8. Fenech (2007, p.  113) indicates from cores taken at Marsa that cereal pollen pointed to human-induced environmental changes characterised by open karstland in the Neolithic period, possibly as early as the Għar Dalam period. She suggests that ‘people were responding to environmental changes that

Notes to Pages 129–131 were possibly mainly caused by extreme weather events and climate and were making the best of the island’s resources in a struggle for survival’ (p. 114); see also §8.3, pp.  107–9; increased charcoal particles in the Marsa samples also was noted (p. 109). 56 Schembri et al. 2009, pp. 18 and 36–9. 57 Fenech 2007, §8.3, p. 109. 58 Schembri et al. 2009, pp. 38–9. 59 Richards et al. 2001, p. 254. 60 Carroll et al. 2012, p. 38. For Tas-Silg analyses, Fiorentino et  al. (2012, pp.  169–84) indicate that the climate was drier in what the Italian project refers to as the ‘Thermi period’, or the Early Bronze Age (p. 174). 61 Environmental studies:  Djamali et  al. 2013; Marriner et  al. 2012; Fenech 2007; Fenech and Schembri 2015; Hunt 2000, 2014; Hunt and Vella 2004–5; Schembri et al. 2009. 62 Stoddart et  al. 2009b, pp.  155 and 191–2, concerning the collapse of the ‘shrine’ in the Xagħra Circle, trapping all ritual paraphernalia it contained; Cutajar et al. 2009, p.  213; see also minutes of the ‘Temple Debate symposium’ (2004) in Malone et al. 2009, pp. 357–8. 63 The possible remnant roofing slabs from Borg in-Nadur can be seen in Tanasi and Vella 2011, p. 38, fig. 2:10 (in house 2) and pp. 46–7, fig. 3:1. 64 For the excavated stratigraphy at Tarxien as it came to light, Zammit, NB no. 11, for 1915, p. 79; NB no. 13, p. 159, ‘One would have expected some Neolithic objects near the floor but the masses of light soil handled is hopelessly free from objects.’ A few of Zammit’s field notebook entries are transcribed in Evans 1971, pp.  150–1. For the excavation publication, Zammit 1916, pp.  135–6. Zammit (1920, p.  184) in the third report of the Tarxien excavations indicated that the sterile layer permeated the whole complex: ‘For about 2 ft. (60  cm) from the ground, this soil is of a grey silty material, light, and homogenous”; Zammit 1930b, pp.  45–7, pl. 14. Pace 2004b, p.  63:  ‘How this massive burial [of the Tarxien megalithic building] came about is now difficult to reconstruct. Was this the sort of debris that one associates with the abandonment and disrepair

355 of buildings? Was it the result of deliberate destruction or the result of some natural calamity, such as a seismic tremor?’ (also p. 69). Murray 1934, p. 1, pl. 4:A. Bonanno et al. (1996) looked to social change during the Tarxien Cemetery period, but not all the authors held the views expressed; cf. Bonanno (1993a, p. 37), who did recognise the artificial quality of the level but interpreted the deposit as a Bronze Age fill, purpose-made to receive cremation burials: ‘I find the debate on the significance of this “sterile” layer quite futile’ (p.  37). Other discussions:  Vella 1999, p.  229; Patton 1996, p. 110. 65 Trump 1976, p. 606. 66 Evans 1971, p. 149. 67 Evans 1971, p. 150: ‘On balance, it seems likely that some fires took place in the cemetery area after the temples had gone out of use.’ 68 Evans 1971, p. 149. 69 Cf. Trump 2002, p.  286, who indicates the problems of determining natural causes for the sterile layer:  ‘The silt layer was taken to imply a long period of abandonment, but this should not be pressed too far since Malta’s short, heavy rainstorms are capable of shifting large quantities of soil quite rapidly’; also Patton 1996, p. 110. 70 Pace 2004, p. 63. 71 Murray 1923a, vol. 1, p. 25: ‘Prof. Zammit is also of [the] opinion that no human beings could have endured in their dwelling-places the stench of decaying animal matter, such as must have filled the Tarxien building from the bones with which every receptacle and niche were crammed.’ Zammit 1917, p.  264. J.  G. Vance, who cleared Ħagar Qim in the late 1830s, observed that signs of fire – a large amount of charcoal and ash – were confined to one chamber, a few rooms had copious amounts of pottery and nine statuettes and he noted ‘great quantities’ of animal and bird bones in the debris within the complex. Significantly, ‘a few human remains’ are also listed among the bone, which could be interpreted as intentional depositions of relics, as we have seen at Skorba in earlier contexts, Vance 1842, pp. 230–6.

356 72 How deep and lasting social memory and ritual practice were on the island might be hinted at by the two large ox horns found on the floor of room M at Tarxien, which seem to be an echo of the six pairs of goat horns found in one Red Skorba house (see Chapter  2); two were side by side on the floor. Tarxien ox horns, Zammit 1930b, p. 25, pl. 6, fig. 2. 73 Verhoeven 1999, pp.  62–3; Wilshusen 1986; Matthews et al. 1996. 74 Verhoeven 1999, p.  62, after Allen 1891; Beals 1934; Kent 1984. 75 Hodder and Cessford 2004. Düring (2002, p. 222) posed the question, ‘Could it be that the link with the past was no longer important in the way it had been before?’ 76 Concerning preparations prior to destruction of houses at Çatalhöyük, Mellaart 1966, p. 172: ‘These houses had not been destroyed by fire and were swept clean before they had been filled into make room for houses of Level V’; see plan p. 184, fig. 8. Hodder and Cessford (2004, pp.  32–3) documented the dismantling of timbers and thoroughly cleaned and back-filled rooms. 77 Stronach 1968, pp.  183–5; Roaf and Stronach 1973, p.  137; Stronach et  al. 1978, p. 10. 78 Verhoeven 1999, p.  61, Tell Sabi Abyad, Syria:  ‘However, no bodies were found inside the buildings of Level 6 (apart from the two skeletons on the roof of building V, which were most likely not victims, but part of a mortuary ritual). … if indeed hostilities resulted in the fire, people seem to have escaped from the village’ (p.  62, n. 27). 79 Vella 1999, p.  229:  ‘although I  strongly believe that the suggestion of intentional mutilation on the statue from Tas-Silg is a plausible one, it is difficult to establish it in a definitive way even when other sites are brought into the discussion.’ 80 Trump (2002, p.  98) suggested that the four Ħagar Qim statues were ‘apparently secreted for safe-keeping’; MAR (1949–50, pp.  1–2) indicates that a fifth badly damaged statue (Q/S22) was found in the dump of the 1839 excavations; Evans 1971, p. 85, pl. 40:9 (note p. 92 lists erroneous find spot).

Notes to Pages 131–135 81 MAR (1913–14, p.  1) concerns the discovery of the walled-up statue remnant at Ħagar Qim; Evans 1971, plan 18A, for the location of the sculpture and clear signs of remodelling of the building in the south-western sector. N.  C. Vella (1999, p. 228) indicates the difficulty in interpreting when the Tas-Silg sculpture was damaged; cf. Cazzella and Recchia 2007, p. 68, who maintain that religious continuity occurred at the site, albeit in a new form. 82 N. C. Vella 1999, pp. 229–30; Cazzella and Recchia (2007, p. 68) state that continued use of the site is “indisputable”; Cazzella and Recchia 2012b, 15–38. 83 Zammit 1916, p. 133, pl. 15:2; N. C. Vella 1999, pp. 228–9; cf. Evans 1971, p. 120. 84 Cazzella and Recchia 2012b, p. 28: chamber M, apse IVA, corridor IVB, steps G in forecourt J all show signs of destruction that was possibly intentional, and no attempt had been made to restore the structures. 85 Cazzella and Recchia 2007, pp.  68–70; 2012b, p. 28. 86 Cazzella and Reccia 2012a, p. 82, fig. 6:2. 87 These notions also were expressed in Cazzella and Recchia 2012a, p. 82. 88 Ibid. Chapter 5  New Directions: The Appearance of the Axe-Bearers 1 Grima (2004, p.  345) argues for a spatial organisation and networks across the islands. 2 Copper axes were found in numbers among the cremated remains of the Tarxien Cemetery period; see illustrations in Murray 1934, pls. 4–6; Zammit 1930, pl. 17; Magro Conti 1999, pp.  191–204, 208; Trump 2002, p.  246; Cilia 2004, original archive photographs reproduced on pp. 210 and 214. 3 Evans (1959, p. 168) argues for a campaign of extermination by people of the Bronze Age; cf. Trump (1997, pp.  21–2, 48), who argues for abandonment of the island. Change is attributed to the collapse of the ‘temple period’ and the immigration of Tarxien groups: Bonanno 1986, pp. 40–1; 1993c, p. 224, among others. 4 Trump 1966, p.  46, concerning Thermi Ware: ‘This ware is so far the only thing

Notes to Pages 135–137 to span the gulf between the Copper Age [i.e., Late Neolithic] temple civilization and the Bronze Age cultures which replaced it’; Malone et al. 2009, §, p. 239; Evans 1956, pp. 97–9, fig. 6; 1984, pp.  151–2, 496–7; Cazzella 2002, p.  139; Cazzella et  al. 2007, pp.  245–9; Cazzella and Recchia 2012a, pp.  80–2; 2012b, pp. 28–32; Copat et al. 2012, pp. 49–51. 5 Trump 1966, p. 46; cf. Evans 1984, p. 496, who does not identify any of the Tarxien Cemetery material as Thermi Ware (‘This ware is not found in the Tarxien Cemetery in Malta and was treated as something separate by the excavator, Zammit.’). 6 Thermi-styled ware was found in the Tarxien complex: Zammit 1930b, type 19, p. 72, pl. 21 (two type 19 examples); C. G. Zammit in MAR 1956–7, p.  2, pl. 5, the pedestalled bowl (Figure  5.1, no.  4)  was described as a lamp ‘of hard fired and polished grey ware (pl. V) … recovered from the soil in the thickness of the wall at the back of the hollow altar stone in the same temple’; also in Trump 1966, pl. 31:d. Cazzella and Recchia (2012a, p.  82) suggested that a hole was purposefully cut through the altar to house the complete pedestalled bowl. Locations of Thermi Ware from Tarxien:  near the oversized statue in the Tarxien complex, Zammit, notebook no.  13 (1915–19), pp.  100–1 (sketched; see Figure 5.1, no. 10); in space O, p. 101, ‘this type of pottery found along the floor of space O and of the Northern entrance of O’; from space B, p. 101, ‘Black sherd with an archaic pitted and incised design. From space B to the west of passage TE’ (sketched; see Figure 5.1, no. 11); from inside niche Q, (12 June 1918), p. 216,‘a few black sherds’ (decorated fragment, sketched; see Figure 5.1, no. 12); eastern limits of the ruins (8 August 1918), pp. 232–3. Zammit’s Tarxien site plans can been seen in his 1916 (fig. 1) and 1920 (pl. 13) publications. 7 Evans 1971, p. 152, after Zammit, NB no. 13 (8 August 1918): ‘The furthest E limit of the ruins cleared further. Rough & archaic sherds thickly covered with caked soil found at about 2 ft from the surface … black sherds with deeply cut squares and dots’ (p. 232); (10 August 1918): ‘More rough sherds were obtained from the E remains. Many of the pots were rough,

357 thick, of a greyish colour, black in fracture, showing very poor baking. It is the roughest type of pottery so far met with at Tarxien. … Many are ornamented with simple lines’ (p. 233). Murray (1934, pls. 16:15, 18–22 and 35:1–13) depicts a range of fragments that appear to fall within the Thermi Ware category. 8 Thermi-like fragment from Tas-Silg (north), Mallia 1965, p.  74, pl. 33:2; from Tas-Silg (south), Bonanno 1999, pp. 220–1, fig. 2, bottom (inv. no. TSG96/2061/9), pl. 1, bottom; Sagona 2000, fig. 9:4; 2014, figs. 1:11:2–3 and 1:152:2–3. 9 For the results of chemical characterisation that has confirmed that one ‘Thermi-styled’ example is of likely local production, Mommsen et  al. 2006, p.  84. Evans (1953, p.  68) was adamant that examples from other prehistoric contexts of the earlier Tarxien ‘temple period’ in Malta ‘have a distinctly foreign appearance’; cf. Trump 1966, p.  46, who suggested that Thermi Ware was probably produced (‘copied’) locally, especially in the Early Bronze Age of the island. 10 Copat et al. 2012, p. 51. 11 Cazzella et  al. 2007, p.  253:  ‘The temple’s rear access, and probably a large part of the rest of the temple, were still in use during the Early Bronze Age, even if the ideology of the worship had changed’; 2012b, p. 602; Cazzella and Recchia 2012b, pp.  28–32 (note that these authors refer to the Borg in-Nadur phase as the Late Bronze Age and the Baħrija phase as the Early Iron Age). The authors maintain that ‘ “symbolic activities” were still carried out at the site during the Early Bronze Age, the evidence of which will appear in a forthcoming publication’ (p. 32, n. 36). 12 Evans 1971, pp. 41–2; reports in Tagliaferro 1911; 1912; MAR 1910–11, p.  4; 1911–12, p. 2; 1921–2, p. 4; 1922–3, pp. 2 and 3. 13 Punctured spots as infill of triangle and other motifs, Tagliaferro 1910, his pottery class 19, pp. 15–20, pl. 9:2, 4. 14 Trump 1961, pl. 15:middle right; Sagona 2008, p.  494, fig.  4:5–6; see pp.  489–503 for a discussion of the sequence and dating of huts 1 and 2 at Borg in-Nadur; 2011, pp. 397–432. Recchia also has questioned the hut sequence at this site; see Recchia and Cazzella 2011, pp. 381–2.

358 15 Cazzella 2002, p. 140, from Tas-Silg, fig. 1:7; from Rudine (Kosova), fig. 1:6. 16 Sagona 2015, §, figs. 1:11:2–3 and 1:152:2–3; also published in Bonanno 1999, pp. 220–1, fig. 2, bottom (TSG96/2061/9), pl. 1, bottom; Sagona 2000, fig. 9:4. 17 A motif not unknown in Maltese prehistoric contexts, Evans 1971, p. 142, pl. 47:5, illustrating a sherd with a bird perched on horns in relief decoration from Tarxien (TP 78). 18 Bird and bull motif on a Mycenaean pot from tomb 83, Enkomi (Cyprus) in the British Museum, inv. no. GR 1897.4-1.1150 (vase C416); crater, Mycenaean pictorial-style decoration, from Klavdia, British Museum, inv. no. GR 1898.1020.15; Mycenaean strainer-spout jug (1250–1225 BC), attributed to painter 20, in the Getty Villa Museum, inv. no.  85.AE.145; Karageorghis 1956, pp. 144–5, figs. 1 and 2 (Pierides Crater no.  42), figs. 3 and 4 (British Museum C416, mentioned previously); Benson 1961, pp. 340–1, pl. 103, fig. 13 (Pierides 42), fig. 14 (C402). 19 The appearance of Thermi Ware is now placed in the final centuries of the third millennium BC, based on finds from southern Italy, Dalmatia and the south-western Aegean; Cazzella and Recchia 2012a, pp.  80–1; see Rambach 2001; 2004, p.  1208, reporting on finds at Olympia; and to Forenbaher and Kaiser 2000. 20 Thermi-styled ware from Troy, Thermi and Lipari: Trump 1966, p. 46; see notes 3 and 4; 2002, p.  271. From Castelluccio, Cazzella 2002, p.  143, fig.  2, map; Cazzella and Recchia 2012a, pp.  80–1, fig. 6:1. Decorated truncated conical and bi-conical bowls from Lipari, Bernabò Brea and Cavalier 1980, pp.  530–5, pls. 130:1–5, 131:1, 132:2 and 133:5 (among other examples). In southern Italy, Evans 1953, p. 68; 1984, pp. 496–7. From Ognina south of Syracuse excavated in the 1960s: Palio 2008, pp. 71–80, figs. 5:4 and 5:6, after Bernabò Brea 1966; Cazzella and Recchia 2012a, p.  82. Wider surveys of Thermi Ware distribution: Cazzella 2002, pp. 139–52, with map on p. 142 and illustration of shape range in fig.  1; Recchia and Cazzella 2011, pp.  378–81; Cazzella

Notes to Pages 137–140 and Recchia 2012a, p.  80; Palio 2008, pp. 76–80. 21 Bossed bone plaque from Tarxien: Zammit 1920, fig.  19; 1930b, pl. 25:2; Evans 1956, pp.  99–100, fig.  7; 1971, T/B1, p.  148, fig. 50 (from ash deposits near the entrance to room 30 on plan 30A); Bonanno 2008c, p. 32. Bone plaques from other locations are documented in Cazzella and Recchia 2012a, p. 82; they list new finds at Grotta del Pipistrello Solitario and Grottaglie in Apulia; Leighton 1999, pp. 144–5, fig. 71. 22 Cazzella and Recchia 2012a, pp. 80–2. 23 A smoke-blackened surface is clearly evident in the photograph appearing in Trump 2002, p. 249. 24 Evans 1953, p. 44, fig. 2; pp. 65–9, fig. 10; Trump 1961, p. 259. 25 Malone et  al. 2009, §6:10, pp.  88–9, 93, 207–18; pottery §10:2:9, pp.  238–40, fig.  10:19. No clear signs of architecture but significant pottery finds were documented at the Xagħra Circle site. 26 Trump 2002, pp. 252–3. 27 Advocating population replacement in the Early Bronze Age, Bonanno 1993a, p. 40. Proponents of enduring indigenous elements, Caroline Malone and Simon Stoddart in Bonanno et al. 1990, p. 202. 28 Evans 1971, p. 224; Bonanno 1993a, p. 42. 29 Evans 1984, p. 496; Bernabò Brea 1968–9; Trump 1966, p. 46; cf. Cazzella and Recchia 2012a, p. 80; Bonanno 1993a, p. 42. 30 Trump 1995–6, pp.  174 and 177, sample number BM-141 (2711–1949 cal. BC), OxA-3750 (2140–1741 cal. BC), BM-711 (1784–1496 cal. BC) and BM-710 (1700–1422 cal. BC); see Tables  1 and 2 and Appendix A. The samples came from Tarxien Cemetery itself in Malta and the Xagħra Circle, Gozo. Malone et al. 2009, pp.  341–6, ­tables  12:1 and 12:2; Cazzella et al. 2007, pp. 243–5, fig. 1; Copat 2012, pp. 51–2; Evans 1971, p. 224. 31 Evans 1971, p. 59. 32 MAR 1927–8, pp.  1 and 2; Evans 1971, p. 39. 33 Trump 2002, p. 215. 34 Tarxien Cemetery cave finds:  at Torri Falka, MAR 1927–8, pp. 1 and 2; at Għar Dalam, Evans 1971, p. 20: ‘All the Bronze Age phases are represented’; at Xagħra Circle, Gozo:  Cutajar et  al. 2009, p.  215; Bonanno 1993a, p. 38; Tanasi (in press).

Notes to Pages 140–144 35 The 1880s work may have been carried out by members of the Permanent Archaeological Commission rather than by A.  A. Caruana, as is usually stated; Bugeja 2011, pp.  27–32; Vella, Zammit and Bugeja 2011, pp. 54–5. 36 Murray 1923a, part 1; 1925, part 2; 1929, part  3; excavations were conducted between 1921 and 1927. 37 Trump 1961; 2002, p.  254; reassessed in Vella, Zammit and Bugeja 2011, pp. 54–68. 38 For Tarxien Cemetery pottery from Borg in-Nadur, Tanasi 2011, pp.  84–8, figs. 4:9 and 4:10. 39 Trump 1966; 2002, p.  251; revised in Vella, Zammit and Bugeja 2011, §3.3, fig. 3:14, pp. 54–66. 40 C. M. Buhagiar 2000, p. 45;Vella, Zammit and Bugeja 2011, p.  47, n.  5:  ‘Magro Conti specifies that behind the wall the Borg in-Nadur phase deposit lay over a 1-m-thick ash layer of Tarxien Cemetery phase date'”; see also Figure 3.2. 41 Tanasi 2011a, pp. 84–8, figs. 4:9 and 4:10. 42 I am grateful to my colleagues for their recognition that the data available in print contained ambiguities and for the revision they carried out; Vella, Zammit and Bugeja 2011, pp. 54–66. 43 Trump 1961, p. 255: ‘After its abandonment, the hut ruins were used for the dumping of domestic rubbish, including much pottery of standard Borg in-Nadur type, here called IIB2.’ 44 Concerning the Għajn Tuffieħa water supply: MAR 1929–30, p. 10; 1930–31, p. 4. 45 Trump writes of Tarxien Cemetery finds: ‘Occasional sherds of this type can be found amongst the material from other temple sites, but they are few in number, have no secure contexts, and are therefore unreliable as evidence.They probably represent no more than casual visits  – picnic parties, perhaps’ (2002, p. 251); or as a ‘squatter’ presence, Bonanno 1993a, p. 37. 46 Tarxien Cemetery pottery at Tas-Silg (north), Cagiano de Azevedo et  al. 1965 (for the 1964 season), fig.  33:2; (south), Sagona 2015, §, figs. 1:10:4–11, 1:11:1, 4–8, 1:144:4, 1:149:6, 1:150:1–4, 1:151:1–4 and 1:245:7. Notions of abandonment of Tas-Silg during the Tarxien Cemetery phase, Vidal González 1998, p.  111:  ‘It is also important to remember that, at the

359 end of the Temple Period (2500 BC), this site was abandoned during the Tarxien Cemetery phase (2500–1500 BC).’ 47 Copat et al. 2012, pp. 51–4. 48 Trump 1966, pp.  7 and 8, 17, 48 and 49, fig.  3; Tarxien Cemetery figurines, pp. 43–4, fig. 43:a–c (see Figure 5.4, nos. 9 and 13); 2002, p. 251 (see photo). 49 The implement used to impress fine dots in the Tarxien Cemetery pottery was the edge of an ark shell or a shark’s tooth: Zammit 1930b, pp. 76–7; Murray 1934, p. 2, pl. 38:4–7. 50 Murray (1934) published a significant amount of Bronze Age cultural remains. Finds in sites such as the Ħal Saflieni hypogeum tend to be minimal or under-represented in publications (see Appendix B, List 10). 51 Azzopardi 2006–7, pp.  9–17, fig.  2, pl. 2. The disc-bodied figures were published in Zammit 1916, pl. 18:1–2; 1930b, pp. 53–4, figs. 3–10, pl. 16; Murray 1934, pl. 11. 52 Evans 1971, p. 150. Zammit 1916, p. 136, the first report of the Tarxien excavations (also transcribed in Evans 1971, pp. 150–1). Concerning burnt fabric in the Tarxien Cemetery deposits: Murray 1934, pl. 38; Zammit, NB no.  13 (1915–16), Wednesday, 22 September 1915, p.  95:  ‘in many cases the bodies were enveloped in a thick cloth dyed a deep yellow, or perhaps a reddish colour turned yellow through burning’; Friday, 24 September 1915, pp.  102–3, artefacts from point B in passage TE (see p.  83):  ‘small masses of yellow ochre which seems to have impregnated some tissue’; p.  119 from TE at the E end of space O: ‘Tissue of some sort of thread in a mass found on ashen sherd at the same place’; p.  122 from passage TE at the juncture with room O:  ‘The ashes in connection with this group of objects contained masses of woven fabric of a yellowish red colour, bright and golden when first found in the damp soil. A bit of cloth was seen to surround the head of the humerus, burnt to a white ash. The ashes of that fabric are very difficult to handle & to preserve. Small bits were successfully mounted in Canada Balsam on a glass microscopial [sic] slide & show well the woven threads. Each thread is

360 made of two threads. The fabric has two threads per millimetre.’ A small sample is illustrated in Murray 1934, pl. 38:8. 53 Disc-bodied figurine fragments from Skorba:  Trump 1966, p.  44, fig.  43:a–c; from Tarxien (thirteen examples): Zammit 1930b, pp. 48–54, figs. 4–10, pls. 15:1 and 16; from Tas-Silg:  stand or prop fragment, inv. no. TSG 96/2066/1, Sagona 2015, fig.  1:10; from Xagħra Circle:  Malone et  al. 2009, p.  305, foot and leg fragments, figs. 10:70:SF41, SF345, SF816 and possibly SF489; disc body fragments, figs. 10:70:SF456 and SF75; possible stylised rodlike head fragment, fig.  10:70:SF489. General discussions: Trump 2002, pp. 258–9; Cilia 2004, pp. 213, 216–17 and 219. 54 Pottery with eye decoration, Murray 1934, ‘owl-faced’ pottery designs, pls. 12:1–2, 13:2–4, 19:4, and 28:7A and 14A. Applied legs, Murray 1934, pl. 28:3–3A; also shown in Trump 2002, p. 247. 55 Zammit 1916, pp.  136–8, pls. 16:3–19:1; Murray (1934, p. 3) indicates that the axes were copper with 12 per cent tin; Murray illustrated many of the finds, pls. 7–9. 56 Rahmstorf 2010, pp. 227–79, fig. 8. 57 Tanasi 2011a, pp.  87–8, fig.  4:10, BN/ P74, which he describes as a ‘model’; originally published in Murray 1929, pl. 22:221, but the drawing has not represented the cut-out walls. 58 Leighton 1999, p. 141, fig. 69:7. 59 Tanasi and Vella 2011, p. 414: ‘We are still lacking a comprehensive survey of dolmens, traditionally associated with the early Bronze Age (Tarxien Cemetery phase); it is clear that their distribution along the margins of major topographic features, including deep-sided wadis, plateaus and plains, begs explanation.’ 60 Evans (1971, p. 224) links the dolmen of Malta to those found in Otranto (in the region of Giurdignano, Minervino di Lecce and Melendugno), south-eastern Italy; after Jatta 1914; Palumbo 1956; see discussion by Cazzella and Recchia 2012a, p.  84. Other examples may exist in Malta and Gozo; recent survey work by Heritage Malta, for instance, in the region of Wardija ta’San Gorg has apparently identified many dolmen (Nicholas

Notes to Pages 144–149 Vella, personal communication, 2013). General discussions of the Maltese dolmens: Ashby et al. 1913, pp. 8–10, figs. 1–5; Evans 1956, pp. 85–101; 1971, pp. 193–7; Sciberras 1999, pp.  101–6; Cilia 2004, pp. 220–1; Trump 2002, pp. 276–9. 61 An early photograph of Misraħ Sinjura appears in Ashby et  al. 1913, p.  8, pl. 1, fig. 1; Mifsud 1995, p. 37. 62 Illustrated in Trump 2002, p. 261. 63 Evans 1956, pp.  85–101. A  shallow rock-cut pit is thought to be a remnant of a dolmen at Ta’ Ħammut, pictured in Cilia 2004, p. 221; Trump 2002, p. 277. 64 Wied Moqbol cairn is one of three to the south-east of Zurrieq, 600 m from Tal-Bakkari church, Ħal Far Road; Evans 1956, pp. 87–9, fig. 2;Trump 2002, pp. 261 and 279. 65 Evans 1956, p. 88, fig. 3:6. 66 Ashby et al. 1913, p. 9; the dolmen was destroyed when Fort Mosta was built, fig. 2. 67 Evans 1971, pp. 196 and 229. 68 Cazzella and Recchia 2012a, p. 84. Safi is a dolmen region; some are intact, and remnants of others might survive in field walls; shallow earth mounds might be of archaeological interest in this regard. 69 Evans 1971, p. 198. 70 Qallilija [343], Sagona 2002, p. 32, n. 51. 71 Fenech 2007, §3.5.3, p.  54, §8.3, p.  109; Trump 2002, p. 255. 72 Trump 2002, p. 255. 73 Ibid., p. 256. 74 Pantelleria as the dominant supplier of obsidian in the Bronze Age, C.  Vella 2015, §4:7. 75 Coles and Harding (1979, p. 193) compared examples in Malta with those in the Milazzo culture (Sicily) contemporary with the Mycenaean sherd finds in Malta. 76 Sagona 1999, pp.  23–60, for the St, George’s Bay dye vats. 77 Reese 2005, pp. 110–11, after Minniti 1999, pp.  177–98. I  am very grateful to David Reese for drawing my attention to his surveys of the murex dye industry throughout the ancient Mediterranean regions. 78 Reese 1987, pp.  203–6, for detailed evidence of purple dye industry in the east Mediterranean; 2005, p.  119; see also

Notes to Pages 149–152 Stieglitz 1994.Although Davide Tanasi has argued for a Cretan origin for four pottery fragments found in Borg in-Nadur and Baħrija, he places them at the time of Phoenician infiltration of the islands. Otherwise, there is little indication of any direct links between Crete and Malta, Tanasi 2009b, pp. 519–38. Note that a fifth fragment (fig. 2a–b, BN/P75) is likely to belong to a medieval form of closed lamp cover well known on the islands, a form discussed by Murray herself (1929, p. 21, pl. 33; not mentioned in Tansasi’s work); see also 1925, p. 26, where she explicitly mentions the modern date of the fragment. Another example reputedly from a tomb in Gozo and pictured in Murray (1929, pl. 32) is tentatively assigned by her to the Bronze Age. Certainly the shape of the Borg in-Nadur fragment is closer to the local and later product than either of the comparative materials illustrated by Tanasi from Crete (2009b, fig.  3:a–f). The early date and limitations of the twentieth-century reports on the Borg in-Nadur site remain problematic for the contexts of the finds. For examples of the medieval form: Sagona 2003, The Courtyard Collection, inv. no.  1248, p. 240, fig. 101:17; Dalli 2006, cover (from the Gozo Archaeological Museum). An example without provenance in St. Agatha’s Museum, Camilleri 1994, inv. no. N2, p.  12. It needs to be noted that among the medieval pottery repertoire is a thickly slipped and sometimes white-painted ware that has the potential to be mistaken for wares in the earlier Bronze Age industry; an example is pictured in Dalli, 2006, p. 244. 79 Concerning the St. George’s Bay rock-cut pits, Adams 1870, pp.  249–50, pl. 7:4. The archive plan is reproduced in Grima 2011, p. 366, fig.11:8 (the plan dates to ca. 1920–1, ref. CD 100A/62). 80 Evans 1971, p. 200.Tarxien Cemetery pottery (cultural period IIA) also came from a bell-shaped pit found at Tal Mejtun (Luqa), MAR 1960, p. 5. Murray 1923, p. 22, mentions spray from the nearby coastline reaching the ruins at Borg in-Nadur. 81 Vella, Zammit and Bugeja 2011, p.  47:  ‘Three unrecorded rock-cut silo

361 pits have also been noted within the perimeter of the wall’; see also n. 5. 82 The ash layer at Borg in-Nadur was apparent in illegal diggings in more recent years, and some contemporary Early Bronze Age deposits were identified in Trump’s 1960s excavation. C. M. Buhagiar 2000, p. 45: the boundary wall was built over ‘the remains of an earlier domestic structure, probably including an oven’; Tanasi and Vella 2011, p. 5, fig. 1:3, plan of the Borg in-Nadur site; see ‘silo pits’;Vella, Zammit and Bugeja 2011, p. 47, n. 5. As yet there is no evidence that the ancient Bronze Age wall extended down to the water’s edge. Indeed,Vella, Zammit and Bugeja (2001, p. 47) indicate that the wall ‘encircled the entire hilltop’. 83 Evans 1971, p. 150. 84 Reese 1987, p.  204; after the dye was extracted, it was ‘mixed with salt and water and boiled for several days’. 85 Evans 1971, p.  151; Zammit 1930b, pp. 72–3. 86 Carington Smith 1992, pp.  692–4, pl. 11:38, nos. 2800 and 2801 (from Nichoria, south-western Greece). 87 Textile production in European contexts, Reese 1987; widespread trading networks dealing in high-status commodities, textiles included, Kristiansen and Larsson 2005, pp. 108 ff. 88 Jensen and Jensen (1965, p.  8) indicate that decay lessens the quality of the dye substance; Reese 1987; Ruscillo 2005, for experimental work done with murex dyes. 89 Ziemba, Akatay and Schwartz 1979, pp. 17–27; Reese 1987, p. 203. 90 Barber (1991, p.  241) summarises the equipment, installations and refuse from the process of producing dyes and colouring fabrics:  vats, tanks, channels, loom weights, containers of ashy material and pierced stones have been found at various dye works; Mifsud 1995, p. 54. 91 Barber 1991, pp. 240–2. 92 Ziemba, Akatay and Schwartz, 1979, p. 19; Barber 1991, pp. 235–9. 93 Ruscillo 2005, p. 105. 94 Leggett 1944, pp. 64–9. 95 Models of cultural exchange are explored in Cazella et al. 2007, p. 248; Cazzella and

362 Recchi 2012a, p.  84, fig.  6:1a. Imports of faience beads from Egypt also were found in the Tarxien Cemetery cremation urns: Stone 1971, pp. 235–6; Murray 1934, p. 4, pl. 41:6 (burnt), 7 and 8. 96 The various forms of burial practice in the region:  Cazella, Pace and Recchia 2007, p. 252; Cazzella and Recchi 2012a, p. 84. 97 Evans 1971, p. 225. 98 Sagona 2008; Għar Mirdum Early Bronze–Middle Bronze transition, Tanasi 2011, §8. 99 Mifsud 1995. 100 Fortified sites in Apulia:  Cazzella 1991, pp.  49–60; Leighton 2009, pp.  120–1, fig. 54:B, an enclosure at Monte Grande (Sicily). 101 For a reappraisal of Murray’s 1920s finds, Tanasi 2011a, pp. 76–84, figs. 4:3–4:8. 102 Trump 2002, p. 255. 103 Of the 1880s excavations, little is known. Bugeja (2011, p.  32) has made a good case for the Permanent Archaeological Commission as the principal investigators rather than A. A. Caruana, although he was a member of the organisation at the time. 104 Vella, Zammit and Bugeja 2011, pp. 59–60, fig. 3:14. 105 One house found in the 1880s also held querns and rollers, Vella, Zammit and Bugeja 2011, pp. 61 and 64, fig. 3:12. 106 Vella, Zammit and Bugeja 2011, p.  61, fig. 3:12. 107 Sagona 2008, pp.  490–504; 2011, pp. 397–414. 108 For the additional exposure of the fortification wall at Borg in-Nadur:  MAR 1998, p.  2; C.  M. Buhagiar 2000, p.  45; more indications of fortification appeared three years later, MAR 2001, p. 167. 109 It is hoped that the reappraisal of Trump’s excavations will provide new data on the date of the boundary wall. Until then, see Vella, Zammit and Bugeja 2011, p.  62, fig.  3:13b, stratigraphic unit 20 (wall), units 16, 18 and 19 (accumulated deposits). Evans 1971, pp. 14–15, general discussion of the architectural finds. 110 Wardija Ta’ San Gorg settlement, MAR (1972–3, §2, p. 72) reports only that work was undertaken at the site. A reappraisal of the site is underway by the University of Malta; see Tanasi (in press), §8.

Notes to Pages 152–160 111 Recent survey work by Heritage Malta has apparently identified numerous dolmens in the Wardija ta’ San Gorg region (N. Vella, personal communication, January 2013). Bronze Age remains are documented by MEPA, gn588_94 [§ 11], marked on plan, p. 6816. 112 Trump 2002, p. 255. 113 Peet 1910, pp.  150–1; Evans 1971, pp. 105–7. 114 Peet 1910;Veca 2011. 115 Trump 1961, pp. 253–62. 116 Concerning the 1959 excavation,Trump 1961, 2002, p. 255. 117 M.-E. Zammit 2006. 118 MAR 1960, pp. 5–6;Trump 1961, p. 260; Mdina sherds were classified as IIB1 (the floruit of the Borg in-Nadur). 119 Galea Debono 2006:  ‘The 2005 excavations carried out by the Superintendence in Mdina, and those carried out this summer at Santa Margerita in Rabat, have demonstrated, without a doubt, that a considerable settlement existed in the area at least 400  years prior to the arrival of the Phoenicians in Malta,’ citing N.  Cutajar. Bell-shaped pits are reported in Mtarfa, Ward-Perkins 1938–9; 1942; MAR 1973–4, §4, p.  51, four ‘silos’ found in the British Services zone, one with finds dated from the first century BC to second century AD. 120 MAR 1993, p. 74. 121 Il-Qolla, west of Burmarrad, Evans 1971, p. 232; Mifsud 1995, p. 60. 122 Buhagiar 1996, p. 20; after Busuttil et al. 1969 (for 1968), pp. 94 and 105; Mifsud (1995, p. 48) mentions the inter-site visibility as well as strategic views to the coast of some fortified sites; Mallia 1968. 123 Qarraba, Bronze Age site, Trump 2002, p. 252. 124 Mifsud 1995, p. 48. 125 MAR 1973–4, §4, p. 51. 126 Stoddart et  al. 2009a, pp.  89–90 (contexts 1279, 1289 and 1291), fig. 6:15; for the pottery finds, pp. 241–2, fig. 10.19:V and W.  Unfortunately, the excavation report is limited in the discussion and illustrations of the Middle Bronze pottery finds. 127 Stoddart et  al. 2009a, p.  90 (context 1279), fig. 6:15.

Notes to Pages 160–162 128 Myres 1901, pp.  89–90; he also makes mention of earlier Neolithic wares around the Bingemma necropolis. 129 Blagg et al. 1990, p. 53. 130 It-Tokk, Victoria, Gozo:  MAR (1961, p.  5) lists Early Bronze Age material among the finds; Bonanno (2005, p. 213) notes the potential of Trump’s field notes concerning the site. 131 Discussions of cave use in Malta during the prehistoric period: Trump 2002, pp.  253–4; Malone et  al. 2009; Skeates 2010, pp.  208–37; Tanasi (in press), §2; Moyes 2012. 132 Reports on Għar Mirdum excavations: MAR 1964, p. 4, pl. 2:1–4; MAR 1965, pp.  1 and 2.  The finds that were recovered are the subject of recent research being carried out by Davide Tanasi (in press). The site produced ‘car loads’ of ceramic finds. A lecture by F. S. Mallia is transcribed in Calleja-Gera 2008. It should be noted that a later cache of pottery, typically funerary in nature, was deposited in the Punic period (probably Melitan phase IV), centuries after the Bronze Age, and that this cache does not figure in the discussion at hand. 133 MAR 1965, pp. 1 and 2. 134 Calleja-Gera 2008. 135 Ibid. 136 Tanasi (in press), §4, concerning the pottery in Għar Mirdum:  ‘Just a few fragments belonging to the Tarxien Cemetery phase or to a final stage of the Borg in-Nadur period were identified.’ 137 MAR 1965, p. 1; Tanasi (in press), especially §4, fig.  13 (finds), §6 (chronology). Trump (2002, p.  262) does make reference to a possible adult female and child burial in Ta’ Vnezja (north-eastern Mdina) with no certain connection to Borg in-Nadur pottery held in the museum stores, simply identified as coming from this district. These burials probably belong to the Neolithic period as they were found with scraps of Neolithic wares and ochre nodules. 138 David Tanasi, in particular, has been recently evaluating the existing museum collections and data relevant to the Bronze Age in Malta:  Tanasi 2008, 2008–9, 2010, 2011a and in press.

363 139 Sinclair and Keith 1924, p. 266. 140 MAR 1963, pp. 3 and 4, fig. 2. 141 MAR 1974–5, §5, p. 56. 142 MAR 1922–3, §7, p.  4; Zammit, NB no. 7, p. 16. 143 MAR 1969, pp. 5 and 6. 144 Tanasi 2011a, pp.  88–133. In essence, red-slipped Borg in-Nadur pottery has been subdivided into four sub-categories of what Tanasi terms ‘fine wares’ that he maintained (at least to 2011)  corresponded to David Trump’s IIB1, IIB2 and IIB3 categories (Trump 1961). Virtually new to the scheme is recognition of semi-fine and coarse wares. 145 Tanasi 2011b, p.  295, after Paternoster et  al. 2008, p.  60. The latter, by dint of their chemical analyses, argue that the practice of using an organic binder with plaster evident in pottery decoration (whether egg or milk) has its origins in prehistory. Milk is still used by tradesmen to slow down the drying time of plaster so that it can be worked – perhaps an advantage for the ancient potter who had intricate designs to embellish with the paste. 146 Sagona (in press c). 147 Ibid. For pottery designs imitating drum lashing:  Tanasi 2011a, fig.  4:14:BN/ P43/60, BN/P48, BN/P142a, BN/P40 and BN/P53; fig.  4:17:BN/P13, BN/ P43.41 and BN/P45a; fig. 4:18:BN/P40; fig. 4:19:BRG/010/88, BN/P108c, BN/ P45h and BN/P43/1; fig.  4:21:BN/ P122 and BN/P48. 148 Tanasi 2011a, §4.4.1, pp.  88–93, figs. 4.11 and 4.12; Sagona 2015, §  ‘Borg in-Nadur Slipped Ware’, §  ‘Painted Variant’, §  ‘White Gritty Ware’, § ‘Drab and Coarse Ware.’ 149 Peet 1910, p. 155: ‘The clay is never entirely pure and, especially in the large vases, contains a considerable proportion of small fragments of quartz.’ White Gritty Ware from Tas-Silg, Sagona 2015, §, figs. 1:20:1–5, 1:21:1–3, 1:157:6, 1:158:2 and 1:245:11. 150 Borg in-Nadur, drab and coarse ware, from Tas-Silg, Sagona 2015, §, figs. 1:18:5–9, 1:19:1–7, 1:157:7, 1:158:3–5 and 1:159:1–3.

364 151 Sieve fragments from Borg in-Nadur, Tanasi 2011a, p.  84, fig.  4:8; but, as noted in the text, this site is currently under extensive revision, and dates of finds may be adjusted accordingly; from Tas-Silg, Sagona 2014, §  ‘Borg in-Nadur, Drab (and Coarse) Ware’, fig.  1:19:5–7, dated according to the stratigraphic sequence at this site. 152 Segmented trays from Xagħra Circle, Gozo:  Malone et  al. 2009, p.  242, not illustrated; cf. BN/P12 from Borg in-Nadur in Evans 1971, fig. 2:1. 153 Trump 1961, pp. 254–7. 154 Trump 2002, p. 291. General discussions of the bell-shaped pits, MAR 1960, p. 5. 155 Mifsud 1995, p. 56; a clay-lined pit was excavated at In-Nuffara, Gozo, MAR 1960, p. 5, fig. 1. 156 At Mtarfa, some bell-shaped pits have signs of reuse for later burials, while a few contained pottery but had no signs of skeletal remains:  Sagona 2002, pp. 242–4, see tombs listings; Evans 1971, p. 201. 157 F. Vella 2008. For past investigations, Evans 1971, pp.  24–5; MAR 1959–60, p.  3 (Tal Mejtin, Luqa, near the cemetery); MAR 1960, p.  5 (sites with bell-shaped pits); MAR 1961, pp. 5 and 8 (Tal Mejtin cistern and pollen sample); MAR 1966, pp.  3 and 4; Trump 1961, p. 257 (Baħrija pits); 1966, p. 51, n. 5 (Tal Mejtin pollen sample). 158 MAR 1966, pp. 3 and 4; the land owner at Tal-Mejtin had observed crop variations and determined where the pits were located, which, in the past, he had emptied and begun to reuse as water cisterns. Pollen from the archaeological investigations was analysed by H.  Godwin (Cambridge University); MAR 1961, p. 8, pl. 1:B, app. 1. 159 Mifsud 1995, p. 55. In a small open area at the corners of Triq Ellul and Triq San Antonio in a new housing estate, one can still access around ten pits cut into a layer of Upper Coralline bedrock but capable of holding water long term. On preparation of raw materials for basketry drawing on indigenous American traditions, Tufton Mason 1988, pp.  44–54. Coiled basket impressions in Middle Bronze Age contexts from Borg

Notes to Pages 162–169 in-Nadur:  Tanasi and Vella 2011 (CD files associated with the publication, kindly supplied by the authors), BN/ P303; Tas-Silg: Sagona 2015, fig. 1:159:1. 160 For example, Vella, Zammit and Bugeja 2011, pp. 262–8; pp. 262–3: ‘It is difficult to assess whether the Maltese communities scattered on their hilltop settlements in both islands could achieve the minimum thresholds to deploy longboats, even communally, and the likelihood is that they probably could not.’ Hard evidence of maritime ventures is limited. A Tarxien Cemetery bowl from the wreck site at Xlendi Bay was thought to be a casual loss overboard, Trump 2002, p. 251. 161 For discussion of human mobility, Vander Linden 2007, pp.  349–50; issues of human agency where individuals or groups work for their own or mutual benefit, Joyce and Lopiparo 2005, pp. 367–74. 162 Agency in such discussions can form a veneer over the deeper issues of how people contribute to social change, Dobres and Robb 2000, p. 4. 163 Vella, Tanasi and Anastasi 2011, p.  263, concerning Maltese maritime capabilities, which ‘Thapsians were keen to restrict if not control.’ 164 Leighton 1999, pp.  144–6; Tanasi 2003, pp. 559–611; 2010, pp. 103–11. 165 Some burials were made in older Neolithic burial chambers at Xemxija, but they were badly decayed and were not studied in detail,Trump 2002, p. 164. 166 Cf. Tanasi 2011b, p. 284. 167 MAR 1960, p.  5, concerning Middle Bronze Age burials. Pottery of Borg in-Nadur type was found at in Ta’ Vnezja, north-east of Mdina, but it is unlikely to be connected to the ochre burial found in this region, which is associated with Neolithic pottery and lumps of ochre, Evans 1971, pp. 28, n. 1, and 230; cf. Trump 2002, p. 262. 168 Revision of the complete finds from tombs 13 and 23, which contained Borg in-Nadur pottery, at Cozzo del Pantano can be found in Tanasi 2011b, pp. 283–337; see also a listing with reference to the Sicilian sites with Borg in-Nadur wares and map (fig.  9:4).

Notes to Pages 169–175 Guzzardi 2008, p. 40, after Orsi 1895, for Thapsos T 6, col. 101, fig. 7; T 22, cols. 108–10, fig. 15; T 64, cols. 135–6, fig. 52; for Milocca-Matrensa, Orsi 1903, T 6, pp. 136–49, pl. 10:3, 5. 169 Mycenean IIIB fragment from Borg in-Nadur, Taylour 1958, pp. 3, 79 and 80. 170 Mycenaean fragments from Tas-Silg (south):  Sagona 2015, §, figs. 1:121:7 and 12, 1:123:6 and 1:223:5; (north): Mallia 1966, p. 50, fig. 35:20, a painted rim sherd from a bowl or cup described as LH IIIA or LH IIIB. 171 Sagona 2008, pp. 496–7; 2011, p. 410. 172 Examples of footed lamps from Għar Dalam are illustrated in Murray 1934, pl. 22:16–17; Evans 1971, p. 20, GD/P.1 and GD/P.2, pl. 32:10–11; 1953, pl. 13:9. 173 Orsi 1895, p. 95, fig. 3; the lamp is illustrated with a grooved up-swung handle fragment characteristic of Tarxien Cemetery wares; Tanasi 2008, fig. 1, for the chronological chart and fig.  50 for the chronological range of pottery forms from Malta and Sicily. He illustrates four from Maltese contexts (Għar Dalam and Borg in-Nadur), figs. 38 and 39. Chapter 6  East Meets West: Phoenician Mariners, Merchants and Settlers 1 Gaston Maspero coined the term ‘people of the sea’ (1896, ch. 4); cf. Drews (2000), who argued that these marauders should be classed as largely Sardinian pirates, intent on plunder; Voskos and Knapp 2008, p.  677; Killebrew 2008, pp. 54–8; Killebrew and Lehmann 2013. 2 Kaniewski et  al. 2010, 2011, p.  1:  ‘The Sea Peoples symbolized the last step of a long and complex spiral of decline in the ancient Mediterranean world’; Leriou 2011. 3 Egyptian land and sea battles, Nelson 1930, pls. 32–34 and 49 (on land), pls. 37–39 (at sea). 4 The connection between the Shardana sea peoples and Sardinia has largely been discounted in recent years; Dyson and Rowlands 2007, p. 101. 5 Zertal 2001, 2002. Cf. Finkelstein 2002; Finkelstein and Piasetzky 2007. 6 See Gilboa 2005, pp.  47–78; 2006–7; Sharon and Gilboa 2013, pp. 393–468.

365 7 Dyson and Rowland 2007, p. 76. 8 Evans 1977, pp. 12–26. 9 Trump 1966, p.  259; offshore influence was evident in excised and painted decoration of his IIC2 phase pottery. 10 Dyson and Rowland 2007, pp. 102–3. 11 Arruda (2009, p. 115) also considered that the infiltration of the Portuguese region was controlled and targeted;‘it appears that from the very beginning there was a desire to reach certain areas – in this case, ones located at the center of the Portuguese western coast. This pattern demonstrates, I  believe, that these arrivals were not at all random, haphazard occurrences but, on the contrary, were determined by a set of goals and a previously established project.’ 12 Frendo 1995, pp. 114–21. 13 Discussed in passing in Chapter 5; Trump 1961, pp. 253–62. 14 Cazzella and Recchia 2006, 2012b, pp. 15–38; Recchia and Cazzella 2011. 15 C. M. Buhagiar 2000. 16 Zammit, 1922–4, NB no. 7, pp. 41–2 (13 August 1923); MAR 1923–4, p. 2; Trump 1966, p.  4, n.  2. Azzopardi (2007, p.  14, fig. 2) adheres to a fifth century BC date assigned in the MAR publication, but the material has yet to be analysed. I saw the material in the National Museum of Archaeology, Valletta, but the pottery had not been cleaned in the intervening years since excavation, Sagona 2002, Victoria, Gozo [699], pp.  1122–3. Since then, repeated attempts to access the finds have failed. For a hypothetical reconstruction of the stone circle arrangement, Bonanno 2005, p. 46. 17 Ward-Perkins 1938–9, pp.  12–13; 1942, p.  34; Evans 1971, p.  107; Sagona 1999, pp. 25–37. 18 The chronology as adhered to in Tanasi and Vella 2011, p. 8, table 1:1. 19 Sagona 2008, pp. 504–20; 2002. 20 Sagona 2002, pp. 21–49. 21 Dietler 2009, pp.  6–7; Amadasi Guzzo 2011, pp. 119–20; Bonanno 2005, p. 14. 22 M. Buhagiar 2007, p.  11. Concerning Gozo, H. C. R. Vella 1995. 23 Bonanno 2005, p. 36. 24 For the Punic word ‘nn:  Mayr 1895, p.  8; Seltman 1946, p.  82, pl. 5:1a–1b; Coleiro 1965, pp.  117–27; Filigheddu 2007, p.  158. The word ‘nn found on

366 coinage: Frey-Kupper 2015, table 7:6) lists a coin minted in Melite (Malta) in the late third to early second centuries BC, found in the excavation of Għar ix-Xiħ in Gozo; Bonanno 2005, p. 37 (middle), p. 159 (top right), p.  187 (coin with inscription pictured); M. Buhagiar 2007, p. 11; Acquaro and de Vita 2009, p. 92. 25 Murray (1923, pl. 12:95) published a single shallow cup fragment, a type well known from the Punic tombs, but it was a shape Phoenicians adapted from the Greeks. 26 See Vella, Zammit and Bugeja 2011, pp. 62–6, fig. 3:13c; distinct levels 4, 5 and 6 are treated as one unit. 27 The current Tas-Silg (north) sequence is discussed in Ciasca and Rossignani 2000, p. 54; Cagiano de Azevedo et al. 1973 (for 1970), p. 104. A comprehensive account of the entire Borg in-Nadur site is in preparation (Tanasi and Vella, personal communication, April 2013), which includes new chemical analyses which have been carried out on pottery samples from Trump’s excavations. See also Tanasi 2011a, p. 134. As yet, they have not indicated whether Phoenician cultural remains were found in unequivocal association with late Bronze Age deposits; none have been published. 28 I am grateful to Nathaniel Cutajar, then Acting Superintendent of Cultural Heritage (Malta), for his discussion of the finds from his excavations in the Rabat-Mdina area (Pottery Workshop, Malta, January 2007). C.  M. Buhagiar (2000, p.  46) indicates that salvage work was conducted at Mesquita Square, Xara Palace and St. Paul’s Street, all within Mdina; notice of excavations appears in MAR 1998, p.  2. Bonanno (2005, p. 48) notes that carbonised remains may indicate that wood was used for some construction. 29 Cutajar 2001, p.  79, fig.  1, left, depicts a carinated bowl from the excavations in Xara Palace, Mdina, probably of form I3, a shape appearing in the archaic to early phase I; see Sagona 2002, p.  168, fig. 340:16. Also illustrated (Cutajar 2001, fig.  1, right) is a double-cordon handle from a small vessel; such handles were attached to piriform jugs, form I1; see, e.g., Sagona 2002, fig. 340:5.

Notes to Pages 175–177 30 The imported sherd is discussed in Semeraro 2002, pp. 506–7, fig. 12b; N. C. Vella 2005a, p. 440. 31 Antonia Ciasca (1982, p. 139) also argued that Phoenician engagement with the indigenous Late Bronze Age community brought them to Rabat-Mdina and to the early settlement of central Malta. Sagona 2002, pp.  29–39, for archaic tomb finds in the Rabat region that were found in Buskett Gardens [68–69], Mtarfa [295], Għar Barka [121], Għajn Klieb [98] and Rabat [459], among other sites; see list in Sagona 2002, p. 39; N. C. Vella 2007a, pp. 72–3. 32 MAR 1998, p. 2; Nathaniel Cutajar, personal communication, 2007, lecture delivered at the pottery workshop; Bonanno 2005, p.  259; the excavated area under Xara Palace Hotel is pictured. 33 Bonanno 2005, pp. 48, 161 and 259. 34 Ward-Perkins 1938–9, p. 12. 35 Ward-Perkins 1942, p.  34. We can only speculate why Ward-Perkins changed his opinion from identifying Mtarfa as a secure context without ‘any trace of intrusive material’ (1938–9, p. 12) to a disturbed cache with an intrusive lamp ‘left by robbers’ (1942, p.  34). Perhaps the change of heart had to do with maintaining the traditional line of argument of his day that Phoenicians arrived in the western Mediterranean in the seventh century BC. See Vella (2005a, pp. 438–44), who adheres to the low chronology and so rejects the initial impression given by Ward-Perkins and accepts his later interpretation. 36 Evans (1971, p.  227) clearly stated that the Mtarfa pit represented the late Borg in-Nadur phase, and a few of the shapes from this pit were represented in his typology for the Borg in-Nadur period as a whole:  ‘On the other hand Phase 3 [i.e., Trump’s IIB3] would seem to have survived, if we take literally the evidence of the material found by Ward-Perkins in Silo I at Mtarfa’; Trump 1961, pp. 261–2. 37 Sagona 2002, ch. 3, pp.  77ff, for Phoenician-Punic ware types including Chalky Reddish Yellow Ware (p. 78). 38 For tombs found on the Mtarfa plateau, Sagona 2002, pp.  876–93, entries [282–304].

Notes to Pages 178–185 39 Leighton 1999, p. 173, fig. 91:8, after Orsi 1895, pl. 5. 40 For reappraisal of Murray’s excavations and a few applied bars and crescents on pottery fragments, Tanasi 2001, with CD photographic files that accompanies the text, (ch. 4) pl. 15:BN/73 (lid), pl. 20:BN/ P133c and BN/P133d and pl. 21:BN/ P135c (vessel fragments). Later investigations, Trump 1961, pl. 14, bottom left. 41 Identification of the Borg in-Nadur, Thin-Walled Variant Ware was first made by Evans 1953, p.  71; 1971, pp.  225–6. Thin-Walled Variant Ware from Tas-Silg (north), Cagiano de Azevedo et  al. 1966 (for 1965), fig.  35:14–15; (south), Sagona 2015, §, figs. 1:21:4–7 and 1:160:1; from the Mtarfa pit:  Ward-Perkins 1939, p.  12; 1942; Evans 1971, p.  225; Sagona 1999, p. 32, fig. 5:3; 2002, tomb entry [295], pp. 29–32, 886–90, figs. 62–68; from Borg in-Nadur, Murray 1923, pl. 12:98–101. Evans (1953, p.  71; 1971, pp.  225–6) also argued for subdividing the Borg in-Nadur wares to include this thin-walled and hard-fired ware. 42 Brusasco 1993; Gómez Bellard 1995, p. 452; Bonanno 1993b, pp. 236–8. 43 Tanasi and Vella 2011. 44 Lithic material from Margaret Murray’s excavations at Borg in-Nadur:  Murray 1923a, p. 44, pl. 16; 1923b, pp. 65–6; 1925, pp.  30–3, pls. 23 and 24; 1929, p.  11, pls. 18 and 19; reappraisal in C.  Vella 2011, pp. 173–94. 45 Fenech and Schembri 2015, § for Hexaplex trunculus and Bolinus brandaris; Cagiano de Azevedo et al. 1966 (for 1965), p. 171. 46 Strada Corsa (Racecourse Road),Victoria, Gozo, near Villa Rundle, found 13 August 1923, inspected 18 August 1923, Zammit, NB no. 7, pp. 41–2; MAR 1923–4, pp. 2–3. 47 Vuillemot 1965, pp.  259–82, especially Tumulus V; these were dated to the seventh and sixth centuries BC. 48 Anziani 1912, pp.  247–55, figs. 3–5, and p.  259, a grave circle around a shaft opening. 49 Concerning Acholla as a Melitan colony, Stephen of Byzantium (Stephani Byzantii), Ethnica 152: Frendo 1995, p. 114; Bonanno 2005, pp.  41–2. Excavations

367 beyond the impressive Roman remains are limited in Acholla: Gozlan and Bourgeois 1981; Gozlan 1984–5; Picard 1947, 1950, 1953, 1968. 50 Pit burials in Qallilija [343], Sagona 2002, p. 32, n. 51, p. 916, fig. 99:7. 51 Bonanno 1993d, p. 45. 52 Trump 1966, p.  44:  ‘the relative importance of this phase [Baħrija] has been much reduced.’ 53 C. Vella commented, ‘Although chert outcrops have not been identified in south-eastern Malta, I  suggest  – with due caution – that sites in this corner of the island, including Borg in-Nadur and Tas-Silg, were procuring their chert from other areas, and therefore, selected “better” quality chert’ (2011, pp.  188–9); presumably, this is applicable to both Late Neolithic and Bronze Ages (§6.3, p. 177); chert sources, C. Vella 2010, fig. 1. 54 Vella, Tanasi and Anastasi 2011, pp. 265–8. 55 Excavations at Baħrija by Thomas E. Peet, MAR 1909–10, pp. E4–5; Peet 1910, pp. 141–63. 56 Trump 2002, p. 255. 57 Trump 1966, p. 44; Recchia and Cazzellla 2011, p. 389; ‘the chronology of the Baħrija phase appears to fall between the eleventh and eighth centuries BC.’ 58 Strainer spouts from jugs, from Baħrija:  Peet 1910, p.  158(c), pl. 14:28, 32; Trump 1966, p.  259, among ‘foreign Baħrija’ IIC2 wares; from marine find (reputedly Salina Bay), Sagona 2003, no.  537, p.  135, fig.  49:2–3; Tanasi 2011b, p. 266, fig. 9:6:1–5. 59 Guzzardi 2008, p. 40, after Orsi 1898, for Grotta Calafarina. 60 Recchia and Cazzella 2011, pp.  375 and 381–2. Of the Italian Archaeological Missions to Malta (1963–70), preliminary reports were published each year which offer some insights into the finds, substantial architecture and rituals at the site. The University of Malta’s excavation report is in press and will appear in the Ancient Near Eastern Studies monograph series (Peeters). 61 Recchia and Cazzella 2011, pp. 373–8, figs. 1 and 2. 62 Recchia and Cazzella 2011, p. 391. 63 Cagiano de Azevado et al. 1967 (for 1966), p. 125: ‘The fanum consists of the ancient

368 megalithic temple which was still in use in Phoenician and Punic times.’ Bozzi et al. 1968 (for 1967), pp.  107–8, who noted reuse of temple sockets, apparently during Phoenician-Punic times. Cagiano de Azevado et al. 1973 (for 1970), p. 104; Ciasca 1988a, p.  206; cf. Bonanno 2007, pp. 108–9. 64 See Frendo 1995, pp.  114–21; Bonanno 1999, pp. 210–12; 2004, p. 286. 65 Shaw (1989, pp.  178–9) also argued that the pits cut in the Tas-Silg threshold were unlikely to be sockets for pillars, rather pivots for doors; cf. Ciasca 1976–7, p. 169, fig. 2. 66 Cf. Buhagiar 1996, pp. 6–7; Frendo 1995, pp. 114–21. 67 Sagona 2002, p.  78, for the Chalky Reddish Yellow Ware of the late Bronze Age; 1999. 68 Zammit, NB no. 4 (26 September 1912), pp. 4–5 and 9–17; MAR 1912–13, pp. 3–4. 69 Baħrija pottery found:  Ashby et  al. 1913, p.  126, postscript; Qallilija is on high ground, now a rocky landscape, and was once walled, MAR 1912–13, pp. 3–4. 70 Cutajar 2001, p. 80, after Antonia Ciasca, who argued for a high level of integration between the Bronze Age and Iron Age communities. 71 Sagona 2004a, p. 245. 72 Kopcke 1992, pp. 103–13. 73 Homer’s Odyssey:  e.g., Book VII, the encounter with the Phoenicians. 74 Kopcke 1992, p. 109. 75 Van Dommelen 1998, p. 34. 76 Servadio 2000, pp.  29–30. Phoenicians are often measured against Greek counterparts: Ridgway 1994–5, p. 77; Markoe 2000, p. 10. 77 Aubet 2001, pp.  29ff; Lipinski 2004; Markoe 2000, pp. 192–206; Gilboa 2005, pp. 51–2. 78 Phoenician city-states exchanged food staples for timber and skilled labour in the tenth century BCE; see II Samuel 8; Kings 5:15, 24 and 25, and 9:11–14. 79 Aubet 2001, p.  111. On the Portuguese Atlantic coast, Arruda 2009, pp. 121–3 (for contact as early as the eleventh and tenth centuries BC). 80 González de Canales et al. 2010, pp. 136–63; López-Ruiz (2009, pp.  255–80) offers a more circumscribed view of interpreting

Notes to Pages 185–190 the location of Tarshish (Tartessos in Greek texts); Lipinski 2004, pp. 225–65, for a comprehensive discussion of this mercantile destination; Albright (1941, p.  21) suggested that the name ‘Tarshish’ may have meant ‘mine’ or ‘smelting works’. 81 Thompson and Skaggs 2013. 82 Dietler 2009, pp.  5–13; Rouillard 2009, pp. 131–40. 83 Tamar Hodos summarises the debate on colonisation terminology (2006, pp.  13–18). Van Dommelen 1997a, 1998a, b, 2006; Dietler 2009, pp. 20–35. 84 Van Dommelen 2002, p. 121. 85 See Ridgway 1988–9, p. 134; 1994–5, p. 79, for issues of interaction between Sardinian cultures and foreign settlers. 86 I do acknowledge that the term ‘colonisation’has fallen from favour (Van Dommelen 2002; Semeraro 2002), although in central and western Mediterranean circles it continues to be used (e.g., Perez and Lopez-Ruiz 2006, pp.  90ff). I  think in regards to the Punic world, ‘colonisation’ and ‘pre-colonisation’ are not employed with the political overtones which characterise the study of colonial events of empires of the modern era. 87 Peckham 1998, pp. 349–51; Fletcher 2004, pp. 51, 59–61 and 66. 88 Pappa 2008–9, pp. 56–8. 89 Lipinski 2004, pp. 244–6. 90 Aubet 2002d, pp.  86–7; Niemeyer 1990, pp.  469–89; apart from Carthage, other colonies followed ‘strictly a traders’ strategy, designed to facilitate the advancement of their mercantile interests’ (p. 487). On specific Phoenician city-state origins, Aubet 2001, pp.  213ff; Macintosh-Turfa 2001, p.  264, on Sidon and Kition in Cyprus as origins of Phoenicians, rarely Aradus. 91 Numerous agricultural implements at Cerro del Villar indicate that the grinding of wheat was carried out at the Phoenician site on a domestic and a commercial scale, Aubet 2002a, pp. 91–5. Also, Aubet (1997, pp. 11–22) discusses the economic activity at the site. 92 For settlement processes in various regions: (concerning sites in Ibiza) Ramón 2002, pp.  148–52; (Motya, Sicily) Nigro 2010, pp.  1–48; Caltabiano and Spagnoli 2010, pp. 117–49.

Notes to Pages 190–196 93 Aubet 2001, pp. 166–72. 94 Strabo Geography 16.2.24; Baity 1973, p.  417. Two Phoenician shipwrecks off the coast of Ashkelon were the first found in deep water (dated to the eighth century BC), Ballard et al. 2002, pp. 151–68. 95 Markoe 2000; Fletcher 2004, pp. 67–74. 96 López Castro 2008, pp. 273–87. 97 Moscati 1988b, p.  48. For a possible lower date for the object, in the eighth century BC:  Falsone 1993b, pp.  45–56; Leighton 1999, pp. 226 and 228. 98 The cuneiform text in Malta:  Mayer 2011, pp. 141–53; 2012, pp. 91–6; Cazzella et  al. 2012a. The inscribed stone was found in fourth- to third-century BC contexts, but this may have been a secondary deposition owing to ancient construction works at the site, Sagona and Sagona (in press). 99 Dyson and Rowlands 2007, pp. 96–101; Begemann et al. 2001, concerning copper trade in Sardinia and links to Cyprus from the second millennium. Crete, too, has been shown to have shifted its trading focus to the central Mediterranean during the thirteenth century BC, Watrous, Day and Jones 1998, pp. 339–40. 100 Foundation of Lixus (Morocco), Pliny the Elder Natural History 19:63. 101 Foundation date of Cadiz (Spain): Strabo, Geography 1:3.2; Pliny the Elder, Natural History 19.63; Pomponius Mela, Description of the World 3:46; Velleius Paterculus, History of Rome I:2.3. See discussion by Mederos Martín (2005, p. 306). 102 Foundation of Utica (Tunisia): Pliny the Elder, Natural History 16.216. 103 Thucydides, Book 6.2, writing in the fifth century BC. 104 Guirguis 2010b, p. 175; red-slipped finds from the earliest settlements in Sardinia are rare and difficult to identify as either local or imported products. Examples of imported red slip in Malta from Tas-Silg (south), Sagona 2015, §161, figs. 1:121:2 and 4, 1:221:2–6, 1:222:3–7 and 1:248:13; one fragment is from a piriform jug with the exaggerated pinched rim formed into two peaks above the lip of the vessel, probably from Bithia, Sardinia.

369 105 Docter et  al. 2005, pp.  557–77; 2008, pp. 379–422. 106 Mederos Martín 2005, p.  311, drawing on evidence of a Mycenaean stirrup jar at Carthage published by Rakob 1996, p. 53. Classical sources for an early foundation date:  Greek historian Filisto of Siracuse, in Jacoby 1929, vol. IIB, no. 556, fr. 47; Eudoxus de Cnidus, in Jacoby 1954, vol. IIIB, fr. 360; Appian, Punica 1,1. 107 Dietler 2009, p. 7. 108 Huelva Phoenician emporium, González de Canales et  al. 2004; 2006, pp. 13–29; Phoenician colonies in Iberia, Almagro-Gorbea 2001, p. 239. 109 Aubet 1997, p. 11. 110 Arruda 1993, pp. 193–214. 111 Guirguis 2010a, pp. 1207–8. 112 Vives-Ferrándiz 2008, pp. 249–52. 113 Aranegui et al. 2011, p. 301. 114 Nigro 2010, p.  7. The renewed excavations (2002 and after) have uncovered temple remains associated with the cothon, an artificial port or docking facility. 115 Issues of ethnicity concern many regions, and arguments have a wider validity: cf. Leriou 2011, pp.  254–5, concerning the problems of identifying immigrant Aegean people in the Cypriot archaeological record through pottery alone; Boardman 2005, influences present in Al Mina. 116 Röllig 1992, pp.  95–102; Markoe 2000, pp. 110–14. 117 One pillar is now in the Louvre; they were published in Renan 1881, vol. I, §CIS I, 122 and CIS I, 122 bis, pp. 150–2. 118 Gubel and Gatier 2002, §178, p.  158; Falsone 1993a, p.  246; early documentation concerning the cippi in Malta, Ellul-Micallef, vol. 2, pp.  182 and 198, n. 9. 119 Brincat 2011; Smith, pp. 46–7. 120 Bonanno 2005, p.  24. The Phoenician occupation of Motya sometime between 770 and 750 BC has recently been confirmed, Nigro 2010, pp. 14 and 42, table 1. 121 The Phoenician shipwreck discovery received a lot of media attention in 2014; e.g., ‘Ancient Phoenician boat believed found in Maltese waters’, Times of Malta, Monday 25 August.2014. The wreck

370 was found in the course of the Groplan Project led by Tim Gambin, University of Malta. 122 Aubet (2001, p.  215) tends to accept, at least, the Carthage foundation story as based on fact:  ‘there are too many coincidences between the eastern and the classical sources to allow us to think that the story of Elissa had no historical basis.’ Niemeyer (2004, p. 45) also argues emphatically against discarding the literary sources, which point to an early date for the Phoenicians in the western Mediterranean. See Chapter  7 for later connections between Cyrene and Melita. 123 Pygmalion or Pummayyaton reigned 820–774 BC. 124 Ovid (Nagle 1995 translation). 125 Cicero contemplated taking refuge in Malta during the civil wars in the second half of the first century BC, Cicero Letters to Atticus, vol. 1, 3.4; vol. 2, 10. 7–9 and 18. 126 H. C. R.Vella 1980a, pp. 19–21 and 52–74, nos. 52–59, concerning comments by Jean Quintin d’Autun (AD 1536), who accepted that Battus was historical; also, Seltman (1946, pp. 83–5), who identified the veiled female head on third-century BC coins in Malta as Anna of Ovid’s poem (see box); cf. Culican 1971, p.  10; Bonanno 2005, pp. 17 and 19. 127 Tomb evidence of wealth in Malta, Gras et al. 1991, pp. 145 and 147, based largely on a tomb in Għajn Qajjied; see Sagona 2002, entry [105], pp. 181–3 and 808–12. 128 Caruana 1898, p.  69; these wealthy tombs contained luxury imported items including Greek figured vases. 129 Analyses of human remains in Malta are few:  Thurnam 1867, pp.  488–99; Schwidetzky and Ramaswamy 1981, pp. 108–38, especially p. 123. 130 Necropoleis in Rabat, Tac-Cagħqi Hill, Sagona 2002, p. 456, fig. 136; pp. 490–91, figs. 170 and 171; p. 492, fig. 172. 131 Punic necropolis at Acholla, Anziani 1912; 1947–55 excavations of Roman levels and the elaborate mosaics found there:  Picard 1947; 1950, pp.  131–2; 1953; 1968.

Notes to Pages 196–201 132 Moscati 1968, p. 189: ‘We have no precise information about the period when Acholla was founded, but it seems reasonable to assume that this happened before the power of Carthage absorbed the surrounding region, which would presumably be before the seventh century.’ 133 Sagona 2002. 134 Sagona 2002, Buskett Gardens [68], p. 786, fig. 7:20–31. 135 C. G. Zammit 1937–8, pp. 3–4. 136 Wente 1967, especially pp.  158–9; Hagens 1996, pp.  154–63; Niwinski 1979, pp. 58–68. 137 Jaeger 1982. 138 Peck 1987, p. 235; Schulman 1984, p. 410. 139 Strabo, Geography 3.5.5. 140 Kopcke 1992, p.  111; Aubet 2002c, pp. 228, 230–1; Sagona 2004a, pp. 246–7; Belén Deamos 2009, pp. 193–228. 141 Aubet 2000, p. 97. 142 Markoe 2000, pp. 179–80. 143 Nigro 2010, pp. 4–5. 144 Temples in early Phoenician settlements:  Antas (Sardinia) a temple site in the Punic period located near the nuraghic settlement, Moscati 1969; at Kition-Kathari (Cyprus), an early Phoenician sanctuary was sited within a disused Late Bronze Age building:  Karageorghis 1971, pl. 51; 1976, p. 98, fig. 18, pls. 69 and 70. 145 N. C. Vella 1999, pp. 225–39. 146 Bonanno 2005, p. 304. 147 Ciasca 1993, p. 230. 148 Settlement remains found in the il-Qlejgħa (‘il-Kligħa’) region, north of Mtarfa, 3 August 1906:  Zammit, NB no.  1, pp.  18–19, with sketches of the irregular, roughly rectangular cistern (measuring L:  6.7 m, W:  1.82 m, D: 1.82 m); MAR 1907, §7, p. E2; Ashby 1915, p. 48. 149 MAR 1906–7, §7, p. E2. 150 Tombs at Qlejgha that can be dated:  Sagona 2002, pp.  941–6, phase I  [372]; phase III [369]; phase IV [374], [376] and [377]; phases IV–VI and Roman influence [370] and [378]. 151 Settlement remains found in Qallilija in 1912, north-west of Mtarfa on the northern side of the Qlejgħa valley: Zammit,

Notes to Pages 201–205 NB no.  4 (26 September 1912), pp.  4, 5 and 9–17; MAR 1912–3, pp.  3–4. For the large, once-standing blocks, Zammit, NB no.  4, p.  4:  ‘There is a long field with large blocks of stones all around. Several menhirs existed in this field which were destroyed by the farmers.’ The site has already been noted for material of ‘Neolithic’ and Baħrija type, although the stratigraphy is far from clear in the accounts of the site; a dark organic soil layer under the Phoenician house contained (prehistoric?) pottery and bone from domesticated animals. 152 Tombs at Qallilija:  Zammit, NB no.  4, (26 September–18 November 1912), pp.  1–31; MAR 1912–13, pp.  7–10; Sagona 2002; Ashby 1915, p. 48. 153 Reused bell-shaped pits:  Sagona 2002, pp. 242–4; Bonanno 2007, pp. 110–11. 154 Pit burials, Qallilija [343], Sagona 2002, p. 32, n. 51. 155 Burials in natural rock cavities, e.g., Marsa [201]:  Sagona 2002, pp.  848–9, fig. 42:4–11; MAR 1917–19, p. 12. 156 Abela 1647, p. 153; Bulifon 1698, raccolta IV, pp.  119–32, on other Għar Barka tomb finds with items of value. Denon (1802, p. 18, pl.V) illustrates the substantial rock-cut tomb shaft and chamber, which held a box-shaped sarcophagus on four legs and mentions tomb finds at Għar Barka; the chamber had three small wall niches and appeared to be roofed over with large stone beams. Concerning the tomb find:  Mayr 1905, p. 469, pl. 1:3; Said-Zammit 1999, pp.  403–10; Sagona 2002, entry [121], pp. 819–20. 157 Caruana 1882, p. 29. 158 Harden 1971, p. 102. 159 Hölbl 1989, pp. 135–7 and 144; cf. Gouder (1978, p. 177) assigns the sarcophagus to the fifth century BC. 160 Bartoloni 1997, pp.  97–103; 1998, pp. 139–42. 161 Archaic Melitan phase I, roundchambered rock-cut tombs: Sagona 2002, plans 2A–E, pp.  244–6, with listings of known examples. 162 Sagona 2002, pp.  248–9, tomb plans 4A–E and 5A–D, usually had a full

371 complement of lamp niches, biers, rock pillows, floor trenches, steps and shelves. 163 Alternative explanations for this feature include drainage for the tomb, Mayr 1905, p. 468. 164 See Sagona (2002) for a detailed discussion of the Melitan Phoenician-Punic tomb sequence. 165 For the Nadur [307] site, Sagona 2002, p. 278. 166 Ibid., p. 293. 167 Ibid. See index ‘infant and child burials’ for the occurrence of children’s remains in the tombs; pp.  252–4, tomb plan 8.  In a Rabat tomb (Sagona 2002 [403]), the remains of two babies (one premature) were found in association with a Romano-Punic cooking pot and numerous small trinkets and implements. Unfortunately, the precise location of this tomb within Rabat was not specified, but its late date is clear. 168 Schwartz et al. 2010. 169 Fedele 1983, pp. 639–43. 170 E.g., Brown 1991; Stager 1980, pp. 1–12; 1982, pp. 163–6; Stager and Wolff 1984, pp. 30–1; Smith et al. 2013, pp. 1191–1207; Xella et al. 2013, pp. 1199–1207. 171 Classical references to child sacrifice, Diodorus Siculus, The Library of History 20.14.4–7:  ‘There was in their city [Carthage] a bronze image of Cronus extending its hands, palms up and sloping toward the ground, so that each of the children when placed thereupon rolled down and fell into a sort of gaping pit filled with fire.’ 172 Concerning child sacrifice in biblical references, Stavrakopoulou 2004. 173 For the tophet site:  Charlton 1861, pp.  131–3; Molk inscription:  Melitensis Tertia, Corpus Inscriptionum Semiticarum I, 123a (eighth to early seventh centuries BC); Caruana 1882, pp. 37–9; Gras et al. 1991, pp. 160–1, fig. 15 – ‘[votive] stele molk baal which has erected Nahum to Baal Hammom, Lord, for He has heard the voice of his words.’ Sagona 2002, pp.  275–7, sacred ‘wells’, for a discussion of this likely tophet site in Rabat. 174 Gras et al. 1991, pp. 127–76. The associated animal sacrifice to Baal Hammom

372 was made as a request to the divine for another child. 175 Ibid, pp. 151–61. 176 Charlton 1861, p.  133:  ‘Canon Bonici, at Malta, possesses a fine and quite perfect Phoenician inscription of six lines, which was found near the hospital at Rabbato, in an excavation like tank, which contained, also, a large number of vases filled with the bones of animals and birds.’ 177 N. Vella, personal communication, 11 January 2013; Sagona 2002, pp.  275–7, for information concerning infant burial sites in Malta. 178 The burial cave was located at Gebel Majjn, Qallilija [367]:  Sagona 2002, pp.  264 and 940–1; MAR 1933–4, pp. 8–9; MAR 1935–6, p. 23 (minor reference). One jar was ‘roughly handled by the workmen’ and the contents lost. The bones in the grave were generally badly decayed owing to the damp conditions and could not be preserved. Other modified storage jars that were used for infant burials can be found in the private collections of the islands, suggesting that the Qallilija site was not unique, Sagona 2002, p. 635, fig. 315:3. 179 Mosta, Ta’ Vnezja [240]:  Sagona 2002, p. 862; MAR 1935–6, p. 23. A modified amphora from Zejtun (Sagona 2002, [748]) dates to the late third century AD and belongs to a North African series, which often has a shaved finish, as the one from this tomb. The skull of a child and the skeletal remains of a youth aged around twelve to fourteen years also were found in the Zejtun chamber. Another tomb at Tarxien contained a peg-base amphora in association with phase V pottery. Bellanti also documented a shaft tomb at Nigred, Rabat (Bellanti, before 1924, pp. 205–6; Sagona 2002 [455]), which appeared to have had burials in modified amphorae, accompanied by a modest number of small vessels. Generally, the descriptions suggest that at least three child burials were made here, tentatively dated to phase IV or phase V. 180 Jar burials of interred children were found in plot 47, Vía Romana, Puig des

Notes to Pages 205–208 Molins, Ibiza, in 1990 and 1991, dated to the sixth century BC, Ramón 2002, p. 138. 181 Mazar 2001, p.  158, concerning child burial at Akhziv. 182 Stager (1980, pp. 1–12) discusses various socio-economic reasons for child sacrifice; Stager and Wolff 1984, pp. 30–1. 183 Among the life-threatening diseases known to have plagued Maltese populations in recorded history is malaria, occurring in 1919 and 1941 around the salt marshes of Salina, Gatt 1996, p.  78; Brown 1984, pp.  209–35, and Webster 1996, pp. 42–3, for discussions of malaria in ancient Sardinia. Brothwell and Sandison 1967, on ancient diseases. 184 Sagona 2002, pp. 29–39; 2008, pp. 511–12; Tyre al-Bass, Aubet 2004. 185 The magnitude of the wine trade in the eighth century BC can be seen in the two shipwrecks laden with amphorae found off the coast of Ashkelon, Israel, Ballard et al. 2002, pp. 151–68. 186 Nigro 2010, pp. 14 and 41; see also the chronological chart for the Motyan excavations, table 1. 187 Nigro 2010, p. 8. 188 Bechtold and Docter 2010, pp.  87 (table 1) and 91. 189 Red-slipped bowls (archaic Melitan phase I) recovered by the Italian Archaeological Mission to Malta, Bozzi et  al. 1968 (for campaign year 1967), pp.  107–8, fig.  13:1–2. Ciasca (1999, p.  75, fig.  6) depicts a collection of red-slipped bowls, an early form (not of the reserve floor type of the established Melitan phase I); the type also was found in the University of Malta’s excavations at Tas-Silg south (Figure 6.6, nos. 17–19), Sagona 2015. 190 Ciasca 1999, p.  75, after Schubart 1976, pp. 179–96. 191 Mitchell (1987, p.  51) also cautioned against applying the Spanish sequence to the material (of uncertain context) from Tharros. See also Bernardini (2008) in regards to the Sardinian evidence. Similar plates in Carthage are discussed in Docter et  al. 2008. Although there is the potential to establish the chronological development

Notes to Pages 208–211



194 195 196 197

of the various plate forms found in Phoenician-Punic contexts in Malta, it can only be achieved with a clear and chronologically long-ranging, stratified sequence. Cf. Red-slipped bowls from other contexts:  Tyre al-Bass, Núñez Calvo 2004, fig.  75:17; Carthage, Bechtold 2007, pp. 335–7, figs. 154 and 155; Nigro 2010, pp.  17–18, fig.  17, red-slipped bowl (MC.07.1690/4 from sondage E) from Motya in the preparatory layers under the cothon temple dated to the second half of the eighth century BC. I am very grateful to F. J. Núñez Calvo for this identification made at the Pottery Workshop in Malta (‘Ceramics of the Phoenician-Punic World’, January 2007). The sherd 2109/7 is a red-slipped plate fragment; two other fragments (inv. nos. TSG96/2145/9 and 10), probably from the same vessel, have been identified. Vives-Ferrandiz 2008, p. 251. See Sagona (2002) for a detailed discussion of the Melitan Phoenician-Punic ceramic sequence. See Sagona 2002, Għajn Qajjied [105]. The dating of the proto-Corithian cup varies:  Dunbabin in Baldacchino 1953, pp.  39–40, for a late eighth-century BC date; Semararo 2002, pp. 492–3, cat. no.  1, dated to 690–650 BC; similarly, N. C.Vella 2005a, p. 439, after Semeraro. Other contemporary Melitan established phase I tombs with datable Greek imports include Buskett Gardens [62], which contained a proto-Corinthian cup with a painted ray pattern around the base, dated to the early seventh century. Buskett Gardens [63] held two seventh-century proto-Corinthian cups. Bellanti describes one of these cups as having ‘broad red stripes’ (Bellanti, before 1924, p.  56). Examples of similar locally made proto-Corinthian cups with handles lifted up and near-vertical rims were found at Motya, Ciasca et al. 1978, p.  14, pl. 6:2 middle, from tomb 55, p.  14, item b, an example without a foot; tomb 119, p. 46, pl. 31:4 left, identified as of the proto-Corinthian, sub-geometrical style, dated to the

373 beginning of the seventh century BC. An imported proto-Corinthian cup was present in Mtarfa [293], falling within the first quarter of the seventh century BC. 198 Baldacchino 1953, p.  34, pl. 13:b; the iron loops in the chamber possibly came from a wooden coffin. 199 Barnett and Mendleson 1987, p. 32. 200 Sagona 2002. 201 Cf. Bonanno 2007, p.  106; N.  C. Vella 2005a. 202 Concerning the date of the Għajn Qajjied tomb, Sagona 2002, pp.  39–49; cf. N. C.Vella 2005a. 203 Għajn Qajjied, remnant archaic, handmade and red-slipped plates:  Sagona 2002, p. 811, nos. 22 and 23; Baldacchino 1953, type H3, items b and f. Presumably they were damaged or decayed and were partly cleared away when the second interment took pace. 204 Guirguis 2010b, pp. 182–5; olpe or dipper juglets are found in the central Mediterranean (Sardinia, Sicily and Malta); Bartoloni (1996, p.  94) considered that they were derived from Cyprus. 205 Tripod bowls connected to wine consumption:  Vives-Ferrándiz 2008, pp.  254–5; Botto 2000, pp.  63–98. On pear-shaped flasks:  Culican 1970a, pp.  5–11; Ramón 1982, pp.  17–41; Groenewoud and Vidal González (1996, pp.  197–205) illustrate four examples without context in Maltese private collections (St. Agatha’s and Wignacourt Museums); from Carthage, Briese 2007, p.  324, fig.  145, nos. 1883 and 1886, appeared to be imported. 206 Namdar et al. 2013. 207 Orsingher 2010, pp. 37–69. 208 Núñez Calvo 2010, p. 57. 209 Núñez Calvo 2010, p.  58; the piriform jug shape mimics metal forms circulating widely in the Near East and Mediterranean. 210 Amphorae from Għajn Qajjied, Sagona 2002, figs. 23:1, 3 and 25:1–2; cf. Carthaginian finds, Bechtold and Docter 2010, pp.  85–103, fig.  4; Sagona (in press b). 211 Fenech 2007, p.  98, also p.  109. Environmental samples from Tas-Silg

374 (north) have indicated that grape seeds were present in the Early Bronze Age at the site; see Fiorentino et  al. 2012, pp. 178–9. 212 Sagona 2015; Núñz Calvo 2010, p.  58; Guirguis 2010b, pp.  181–4; Botto 2010, p. 154. 213 Wine consumption in Iberia:  Buxó 2008, pp. 145–54; 2009, p. 158; the ‘production of wine seems to be one of the distinguishing elements of Phoenician colonization.’ 214 Núñez Calvo 2010, pp.  57–8; Fletcher 2006. 215 Nigro 2010, pp.  18–19, fig.  19, second and third from the top (MC.07.1835/36 and MC.07.1835/66) in the opening years of phase VIII in Motya; parallels are drawn to Carthage, Sidon and Ras al-Bassit. 216 Fenech 2007, pp. 98–9 and 109. 217 Fenech 2007, p. 110. 218 Sagona 2002, archaic to early Melitan phase I, disc-topped jug forms, fig.  340:7–9; forms I:3a, I:3b and I:3d; trefoil jug forms, fig.  340:5–6, forms I:1, I:2b. 219 Unfortunately, we have to rely on the archival photograph and a sketch of the Mellieħa [215] pottery; Sagona 2002, p. 853, fig. 40:15–21. 220 The Mellieħa urn is designated form 1:5 within the Maltese scheme of urns (Figure  6.7, no.  4). Núñez Calvo 2004, p.  290, Tyre al-Bass, urn form Cr F1f of Phoenician type: U.10–1, p. 463, and U.42–1, p. 460, both assigned to period IV, dated from the second half of the eighth to the beginning of the seventh centuries BC or perhaps even slightly earlier; cf. U.55–1, p.  457, assigned to period III and dated around the second quarter to the mid-eighth century BC; 2010, pp. 53–6, fig. 3. 221 Núñez Calvo 2004, pp. 332–3; see hemispherical bowls, subtypes CP2a and CP2b. The Mellieħa example would appear to be a red-slipped type rather than painted. Its slightly incurving lip suggests that it belongs to the latter category, which Núñez places in period IV, dated to the ‘standard Iron Age II … second half of the VIII and beginning of the VII century BC’ (pp. 362–3).

Notes to Pages 211–217 222 The ridged-neck jug (the rim is damaged) from Mellieħa would appear to be red slipped and burnished. The closest shape to the Maltese example from the Tyre 1997 cemetery is the square-cut rim jugs (Ja2), which can have red-slipped surfaces, although red- and black-banded necks are more frequent; see Núñez Calvo 2004, pp. 307–10, figs. 163, 164; numerous parallels are listed on p. 310.The category is assigned to period III, dated from the second quarter to the mid-eighth century BC (p. 363). 223 Núñez Calvo 2004, pp.  65, U.4–2 and U.4–3, and 138, urn 4, formal group II, painted bi-chrome ware; some of the examples from Tyre also approach flat instead of down-turned, everted rims, Seeden 1991 (urn 5 burial) no. 18, figs. 35 and 36. 224 I am grateful to Dr Nicholas Vella (University of Malta) for supplying me with a copy of the drawing on the back of the original photograph of the Mellieħa tomb group. It depicts a bi-conical trefoil jug of the type illustrated in Figure 6.7, no. 5, with the hand-written comment, ‘Tomb furniture found at Melleha [sic.] Jan 1912. 2 of this type of enochoe [sic] where [sic] also found.’ 225 Maas-Lindemann (2005, p.  110) made similar observations; cf. plates from Tyre al-Bass in fig. 2:c–k and the Junon tomb, Carthage, in fig. 3:e–m. 226 Single-nozzle lamps in the Iberian Peninsula, for example, from Cadiz: Ruiz Mata 2002, pp. 178–9, fig. 7:18; Huelva, González de Canales Cerisola et  al. 2004, pp.  80–1, pl. 54:12–20. It is a homeland variety; see, e.g., those of type 3 from Tyre, Bikai 1978, pp.  18–20, pl. 7:6, described as ‘quite low in profile, the rim being quite sharply formed and it is often burnished.’ 227 The Tyre al-Bass burials were made in pits as opposed to rock-cut chambers in Malta, which may have driven a fundamental difference in the funerary repertoires; F. Núñez Calvo, personal communication, 2007. 228 Incense cups in Maltese tombs, Sagona 2002, pp. 215–18. 229 Van Dommelen and Gómez Bellard 2008, pp. 2–3.

Notes to Pages 218–223 Chapter 7  Melita and Gaulos during the Punic Period 1 Decline in sixth-century Iberian contexts:  Aubet 2002b, p.  224; Fernández Jurado 2002, pp. 259–60. 2 Aubet 2001, pp. 341–6. 3 Hecataeus fragment 341:  ‘An island over against Carthage’; Braun 2004, p. 329. 4 Concerning the diverse range of Phoenician settlements, van Dommelen 1997b, p. 246. 5 Sagona 2002, pp. 278–80, app. 1, pp. 745–6; Said-Zammit 1997, pp.  3–6 and 36–42. Said-Zammit (2001, pp.  117–46) focusses on burials in the Rabat (Malta) region, where at least 372 tombs have been located. 6 Nature of urbanism:  e.g., Finley 1973; Snodgrass 1991; Osborne and Cunliffe 2005. Studies concerning urban and rural development in Phoenician-Punic regions: van Dommelen 1997b, pp. 243–78; Arruda et  al. 2007; van Dommelen and Gómez Bellard 2008. 7 Listed criteria for determining cities:  e.g., Childe 1950; Osborne 2005, pp. 5–8; Smith 2009, pp. 1–20. Niemeyer (2002, pp. 31–48), after F. Kolb’s six-part definition presented in Die Stadt im Altertum (1984), examined the concept of ‘city’ for the Phoenician site of Toscanos (Spain). 8 Site-specific determinations of cities and urban centres, van Dommelen 2005. 9 Morris 1991. 10 N. C. Vella 2007a. 11 See Sagona 2002, pp. 269–73. 12 Caruana (1882, pp.  21ff) argued that the existing old towns may have great antiquity: ‘Some of these villages were mixed up with the Phoenician centres, and in my present state of knowledge, I am not quite able to draw a line of demarcation between them’ (§101, p. 93). 13 Cf. Spanò Giammellaro et  al. 2008, pp. 151–6. 14 Pseudo-Scylax Periplous [Circumna­ vigation] 111, writing in about the mid-fourth century BC; also Cicero, Verrine Orations; in Melita (island), ‘there is a town, also called Melita’ (first century BC); Livy, The History of Rome (21.51.1–3), mentions a town in Melita (the island) in the first centuries BC-AD.

375 15 Ptolemy, Geography 4.3.13; Ashby 1915, pp. 25–7 and 43; Lipinski 2004, p. 378. 16 Cicero, The Verrine Orations, in the first century BC; Verres turned the town of Melita ‘into a factory for the weaving of women’s dresses’. 17 Punic and other coinage circulating in Melita and Gaulos:  Bonanno 2005, pp. 96–8; Coleiro 1965, pp. 117–27, pls. 50 and 51; Gouder 1973, pp.  1–16; Sammut 2001, pp.  1–6, pls. 1–7. Two coin hoards spanning the mid-third to mid-second centuries BC represent the currency exchanged throughout North Africa and the central Mediterranean. One hoard was from Mqabba:  Zammit, NB no.  5, p.  83 (who provides details concerning context); MAR 1921–2, p.  6; Jenkins 1983, pp.  19–36. The other hoard was from unknown contexts in Malta, Visonà 1990, pp.  169–92. Coins from Tas-Silg (north):  Perassi 2004–5, pp.  371–86; Novarese 2006, pp.  49–79. Twenty-eight coins from Tas-Silg (south), Frey-Kupper 2015, with comprehensive bibliography. 18 Marriner et al. 2012, pp. 10–13. 19 Bikai 1989, p. 206; Markoe 2000, pp. 72–3. 20 Ramón 2002, p. 150. González Prats et al. 2002, pp. 122–3; La Fonteta was considered one of the ‘principal Phoenician cities’ at about 8 ha. Aubet (2002d, pp. 80–1) notes that Morro de Mezquitilla spanned 2 ha. 21 Aubet 2001, p. 232. 22 Markoe 2000, pp. 72–108. 23 Markoe 2000, pp.  83–95, for Phoenician urban development. 24 Osborne 2005, p. 1. The current population of Rabat is over 11,000 (31 December 2010, Demographic Review 2010, National Statistics Office Malta). 25 Pausanias, Description of Greece 10.4.1 (second century AD); on defining cities and urban development:  Morely 2008, pp.  122–6; Osborne and Cunliffe 2005; Van Dommelen 2005, pp.  143–67; Finley 1977, pp. 305–6; Di Vita 1990. 26 Said-Zammit 1997, pp.  36–42; Fenech 2007, §3.4, pp. 51–2. Such population estimates are tenuous at best. Complicating the determinations are reuse of tombs and the clearing away of old burials. Changing economic directions saw the steady growth of maritime trade, which delivered unknown quantities of food items to the

376 islands during the Phoenician-Punic to Roman periods. The population was not restricted to island-based subsistence. 27 Markoe 2000, pp. 73, 173–4 and 188–91. 28 MAR 1924–5, p. 3: ‘Another cave which had all the appearance of an original Phoenician rock-tomb served the purpose of a cellar. It was, naturally enlarged and conveniently arranged for use as a cellar but the primitive shape could not be mistaken [see also p. 5]’; another tomb was found in excavations extended south of the Domus Romana: Zammit 1924–5, p. 5; Zammit, NB no. 8, pp. 12 and 14. 29 Cutajar 2001, p. 81. 30 MAR 1960, pp.  5–6:  excavations in the rear courtyard of the Vilhena Palace recovered largely second-century AD Roman ‘heavy masonry’; Cutajar 2001, pp.  80–2; Bonanno 2005, p. 87. More lengths of fortification walls were found behind Vilhena Palace in 1962 and 1963, but these were identified at the time as Roman; MAR 1963, p. 7, pl. 3; Mayr 1909, pp. 142–3. 31 Markoe 2000, p. 72. 32 Blagg et al. 1990, p. 44; Roman architectural blocks measured about 1.62 and 1.26 m long by 50 cm high at Ħal Millieri. 33 MaltaMedia News bulletin, 6 May 2005:  ‘Excavations reveal Mdina’s unknown aspects’; available at:  http:// exec/view.cgi?archive=8&num=5952; accessed 27 November 2011); Stroud 2001, pp. 10–11. 34 See documentation in, e.g., Ashby 1915, pp. 30–47; Buhagiar 2007, pp. 14–20. 35 Excavations were extended north-east of the Domus Romana Museum site; Zammit 1922, p. 132, fig. 1. 36 Mayr thought that the statue represented Astarte and that a second temple to her was in Mdina, Mayr 1909, p. 127, fig. 31; Ashby 1915, p. 33, fig. 3. 37 Culican 1971, p. 4. Olmos and Fernández 1987, pp.  214–16, figs. 2–4, for a thymiaterion stand depicting Astarte holding a dove found in Albacete (Spain). For an eastern origin (perhaps in Cyprus) of Aphrodite and on-going syncretism between the Greek goddess and Astarte, both of whom are represented by doves, Budin 2004, pp. 95–145. Miller

Notes to Pages 223–225 Ammerman 2001, pp.  338–9, for model doves from the sanctuary of Aphrodite at Santa Venera, Paestum, Italy. Bonanno (2005, p.  163) suggested that the statue from Malta depicted Isis. 38 Excavations carried out by T. Zammit and discovery of the sculpture, Zammit, NB no. 5, for 20 September 1920, p. 68; MAR 1920–1, pp. 3–4; Zammit 1922, pp. 131–3, with photographs of the extended excavations to the north-east of the Domus Romana site. Bonanno 2005, pp.  213–14; the sculptured head is depicted on pp. 214–15; cf. the female statue on p. 163; both have a very similar treatment of the hair, which falls in ringlets depicted with simple scoring of the locks, but the female figure shows the signs of later influence in the drapery. 39 N. C. Vella 2007a, pp. 72–3. 40 MAR 1965, p. 3. 41 Cutajar (2001, p. 81) mentions some finds as Punic; Bonanno 2005, pp.  215–17, as Roman. Notice of finds in the Saqqajja area: MAR 1932–3, pp. 8–9; 1965, p. 3, pl. 1; 1966, pp. 5–6; 1969, p. 5; 1970, §§16–18, pp. 5–6. 42 Architectural remains have been identified in a field towards Buskett. A  large stone block had four ‘Phoenician’ coins under it, MAR 1905, §12, p. 2. 43 Caruana 1888, p.  450. See Bujega (2011, pp.  17–34), who charts the confusing course concerning the identification of Tas-Silg, Borg in-Nadur and the Temple of Hercules through antiquarian and more recent historic accounts, which largely stem from an erroneous interpretations of Abela (1647) and Ciantar (1772). 44 Sarcophagi from Għar Barka and elsewhere in Malta and Gozo:  Abela 1647, p.  153, libro II, notitia II; Caruana 1882, p. 29; Mayr 1905, pp. 478–81. 45 MAR 1932–3, pp. 8–9; Ashby 1915, p. 30; see earlier references to the wall in n. 2. 46 Bonanno 2005, p. 87; 1992a, p. 19. 47 Lancel 1997, p.  263, drawing on Appian, History of Rome (Libyca) 54, and Eumachos Neapolitanus (see Jacoby 1923, §178 F2); Lancel 1999, p. 178. 48 Sidon’s triple-ditch defences, Markoe 2000, p. 94, after Diodorus Siculus Library of History 16.44.5–6.

Notes to Pages 225–228 49 Identification of the current Triq Navarra Gusman with New Street was made by Canon J. Azzopardi, former keeper of the Mdina Cathedral Museum, Sagona 2002, p. 968, n. 454. 50 Punic tombs in the Domus Romana museum site, Rabat:  Zammit, NB no.  6, pp. 7 and 14–15; Sagona 2002, entry [461], found 1921, p.  1005, fig.  167:1; Zammit, NB no.  8, pp.  12 and 14; MAR 1924–5, p. 3; Sagona 2002, entry [462], found 1924, p. 1005, fig. 166:2. 51 Mayr 1909, p.  141, figs. 32:c (ditch) and 5 (Saura). The latest datable evidence from a tomb in this group is Rabat, New Street [436] in Sagona 2002, pp. 256–7 and 992–5. Lancel (1997, pp.  220–1) discusses the pressures placed on burial grounds by urban sprawl in ancient Carthage, which saw the cemeteries of Dermech and Douimès reclaimed, although graves were carefully relocated at Carthage. 52 Houel (1782, ch. IV, pl. CCLIX) drew a plan and views of the additional standing-wall lines, which he considered Greek; they later were partially demolished. Boisgelin (1805, p.  55) states that ‘it has only one window and a folding door.’ Caruana 1882, §101, p.  94, where Houel’s sketch is reproduced on p. 95, and Caruana followed his line of reasoning that the structure was Greek; Mayr 1909, pp. 89–90, fig. 30. 53 Hölbl 1989, pls. 19:1–2, 20 and 21:1–2; Bonanno 1996a, p. 47; 2005, p. 88. Tas-Silg cornice fragments:  Cagiano de Azevedo et al. 1967 (for 1966), pl. 21:2; Cagiano de Azevedo et al. 1972 (for 1969), p. 73, pl. 2:1; Cagiano de Azevedo et al. 1973 (1970), pls. 20:1 and 40:1–2. Cardona in his unpublished master’s thesis (2010) and his recent study of architectural elements at Tas-Silg (Cardona 2015) indicates that remnant gorge-cornice fragments also were found in Mtarfa, thrown into a silo pit (MAR 1973–4, §5, p. 51). 54 Historic houses of the modern era are also finished at the roof with gorge cornicing; for example, a large house on the north of the Pjazza san Pawl square, Mdina. Further sites in Valletta, including a three-storey building at Triq Sant Anton, north-east of the Upper Barakka Gardens; a building in

377 the open area around the Grand Harbour Hotel and the British Hotel, down the steps of the Batterija towards Victoria Gate; Triq il Lvant (East Street), nos. 8 and 11; the Straight Street monument. Also, Triq Sir Ugo Mifsud, Lija and a building of the recent past on the road from Rabat to Baħrija at the turn off to the Għajn Klieb necropolis. 55 MAR 1938–9, pp. 2–5; the Zurrieq building was investigated by R.  V. Galea and Charles Zammit. Further investigations are reported in MAR 1964, pp.  6–7. Gouder 1978, p.  183; a sondage was excavated around the building, but the results have not yet been fully published, and the few comments available suggest that they were inconclusive. Houel’s depiction of village paths around the structure also in