Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones 9781841712321, 9781407352886

Contains a biography of Professor Barri Jones by Nick Higham and a bibliography of his published writings. Thirty-three

417 51 55MB

English Pages [369] Year 2001

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones
 9781841712321, 9781407352886

Table of contents :
Front Cover
Title Page
Copyright
Dedication
TABLE OF CONTENTS
List of Contributors
List of Illustrations
Preface
Barri Jones: A Life in Archaeology
Bibliography of the writings of Professor Barri Jones
A. Roman Britain
B. Public Policy and Archaeology
C. The Eastern Empire
D. North Africa
E. Methodology and Technology
F. Religion in the Empire and Beyond

Citation preview

BAR  S940  2001   HIGHAM (Ed.)  

Archaeology of the Roman Empire A tribute to the life and works of Professor Barri Jones

ARCHAEOLOGY OF THE ROMAN EMPIRE

Edited by

N. J. Higham

BAR International Series 940 B A R

2001

Archaeology of the Roman Empire A tribute to the life and works of Professor Barri Jones

Edited by

N. J. Higham

BAR International Series 940 2001

Published in 2016 by BAR Publishing, Oxford BAR International Series 940 Archaeology of the Roman Empire

© The editor and contributors severally and the Publisher 2001 The authors' moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781841712321 paperback ISBN 9781407352886 e-format DOI https://doi.org/10.30861/9781841712321 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 197 4 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd/ Hadrian Books Ltd, the Series principal publisher, in 2001. This present volume is published by BAR Publishing, 2016.

BAR

PUBLISHING BAR titles are available from:

EMAIL

PHONE FAX

BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

For Driana, Tudor and Angharad

Professor Barri Jones

ii

Contents Contributions

iii

List of Contributors

v

Lists of Illustrations

v1

General Abbreviations

vm

Preface

1x

Barri Jones: A Life in Archaeology, by Nick Higham

1

Bibliography of the Writings of Barri Jones, compiled by Anthony Birley, Nick Higham, David Mattingly and Paul Holder

6

Contributions: A Roman Britain Northern Britain I .Anthony Birley 2.Ben Edwards: 3.Richard Gregory: 4. W S. Hanson and Lorna

5.Mike Nevell: 6.David Shatter: 7.David Woolliscroft:

ii

'The Anavionenses' 'Reginald Bainbrigg, scholemaister, and his stones' 'Living on the Frontier: Iron Age-Roman Transitions in South-West Scotland' Sharpe: 'The relative information value of aerial photography and geophysical survey: a case study from the Clyde valley' 'The Edge of Empire: Late Prehistoric and Romano-British Settlement in North West England. A Study in Marginality' "'Agricolan" is an overworked adjective' 'The Roman Gask Project, Interim Report 1995-2000'

15

25 35

51

59 75 85

Southern Britain

8.Paul Holder: 'Observations on the Roman Military Diploma of 14 April AD 135 from Wroxeter 9.John Manley and David Rudkin: 'Excavations at Fishbourne Roman Palace 1995-1999: an interim statement' 1O.Bill Putnam and Hugh Toller: 'A Difficult Roman Road in Deepest Wales' 11. Hugh Toller: 'Cambrian Antiquity- The Romanists' 12 Adrian Waddelove: 'The River Dee and the Deva Harbour: A Reassessment'

95 105 117

123 131

B Public Policy and Archaeology 13.John Walker: 14.John Williams:

'Some Myths of Public Service' 'PPG 16 and the Romano-British Landscape of Kent'

141 147

'The Roman Destruction of Phalasarna' "'Romans Always Conquer": Some Evidence of Ethnic Identity on Rome's Eastern Frontier' 'Khibert Khaw: a Roman Town and Fort in northern Jordan'

155

C The Eastern Empire 15.Elpida Hadjidaki: 16.John Healey: 17.David Kennedy:

iii

167

173

18.Henry MacAdam:

D

'Philo ofByblos and the Phoenician History: Ethnicity and Culture in Hadrianic Lebanon'

North Africa

19. Graeme Barker: Floodwater Farming, Roman Imperialism and Desertification in Libya and Jordan 20.Paul Bennett, Andrew Wilson and Ahmed Buzaian: 'Euesperides, the first Benghazi' 21. Chris Hunt and Dave Gilbertson: 'Romano-Libyan Agriculture in the Tripolitanian Predesert: Mediterranean Perspectives 22.David Mattingly: 'Farming the Desert: Two contrasting Libyan Case-Studies

E

205

219 237 265

Methodology and Technology

23.Bob Bewley: 24.David Bird: 25.Keith Maude: 26.Denys Stocks: 27.Janet Webster: 28.Peter Webster:

F

189

'Developing Aerial Survey: A Tribute to Barri Jones' 'Aspects of Roman gold-mining: Dolaucothi, Asturias and Pliny' 'Roman Force Pumps' 'Roman Stone-working Methods in the Eastern Desert of Egypt' 'An Inlaid Iron Brooch from Rhyn Park' 'Earth, Fire and Water: The Making and Marketing of Roman Samian Ware'

257 265 277

283 287 289

Religion in the Empire and Beyond

29.Joanna Bird: 30.Nick Higham: 31.Michael Jones: 32.Tom Rasmussen: 33. Val Worden:

Censers, incense and donors in the cult ofMithras 'Bancornaburg: Re-visiting Bangor-is-y-coed' 'Early Christian Archaeology in Europe: Some Recent Research' 'Bacchus triumphs on the North Downs: the Polesden Lacey sarcophagus' Image and Symbol in Severan State Portrait Cameos

iv

303 311

319 335 343

List of Contributors Professor G.W.W.Barker: P. A Bennett: Dr. R. A Bewley: Dr. D. G. Bird: J. Bird: Professor A R. Birley: Dr. A M. Buzaian: B. J. N. Edwards: Professor D. D. Gilbertson: Dr. R. A Gregory: Dr. E. Hadjidaki: Professor W. S. Hanson: Professor J. Healey: Dr. N. J. Higham: Dr. P. Holder: Dr. C. 0. Hunt: M. J. Jones: Professor D. Kennedy: Dr. H. I. MacAdam: J. Manley: Professor D. J. Mattingly: K. Maude: Dr. M. D. Nevell: W. G. Putnam: Dr. T. Rasmussen: D. Rudkin: Ms. L. Sharpe: Dr. D. C. A Shorter: D. A Stocks: H. Toller: A.C.Waddelove: J. Walker: J. Webster: Dr P.V. Webster: J. Williams: Dr. A Wilson: Dr.D.J.Woolliscroft: Valerie Worden:

School of Archaeological Studies, University of Leicester Director, Canterbury Archaeological Ttust English Heritage Surrey County Archaeologist Archaeological Consultant Vindolanda Trust Lecturer in Archaeology, Garyanis University, Benghazi, Libya Lancashire County Archaeologist, retired. School of Conservation Sciences, Bournemouth University Palaeoecological Research Unit, School of Geography, University of Manchester Deputy Director of Maritime Antiquities, Ministry of Culture, Athens Department of Archaeology, University of Glasgow Department of Middle Eastern Studies, University of Manchester Reader in History, School of History and Classics, University of Manchester John Rylands University Library of Manchester Geographical Sciences, University of Huddersfield City Archaeologist, Lincoln I University of Nottingham Classics and Ancient History, University of Western Australia Robbinsville, New Jersey, USA Chief Executive, SussexpastlSussex Archaeological Society School of Archaeological Studies, University of Leicester Instructor, School of Art History and Archaeology, University of Manchester Greater Manchester Archaeological Unit, University of Manchester Principal Lecturer in Archaeology, Bournemouth University, retired. School of Art History and Archaeology, University of Manchester Director, Fishbourne Roman Palace, a property of the Sussex Archaeological Society Department of Archaeology, University of Glasgow Senior Lecturer in Roman History, University of Lancaster Clo School of Art History and Archaeology, University of Manchester Accountant and Amateur Archaeologist Public Relations Manager, Cheshire County Council, Chester Director, Greater Manchester Archaeological Unit Clo Centre for Lifelong Leaming, Cardiff University Centre for Lifelong Learning, Cardiff University Head of Heritage Conservation, Kent County Council University Lecturer in Roman Archaeology, Institute of Archaeology, University of Oxford The Roman Gask Project, School of Art History and Archaeology, University of Manchester School of Art History and Archaeology, University of Manchester

V

List of Illustrations 2.1 2.2 2.3 3.1 3.2 3.3 3.4 4.1 4.2 4.3 4.4 4.5 4.6 5.1 5.2 5.3 5.4

5.5 5.6

5.7 6.1 6.2 6.3 6.4 7.1 7.2 7.3 7.4

7.5 7.6 8.1 8.2 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 9.11 10.1 12.1 12.2 12.3 12.4 12.5 14.1 14.2 14.3 14.4 14.5 15.1 15.2 15.3 15.4

I. Hayton's record of Reginald Bainbrigg's stones, 1722. Reginald Bainbrigg's stones in the Broadclose wall, Appleby-in-Westmorland Diagram of the Bainbrigg stones, Appleby-in-Westmorland Unenclosed and palisaded settlements Pre-Roman/Roman enclosed settlement Pre-Roman Iron Age agricultural remains Roman military sites Map of the Clyde valley and the location ofBurnfoot Farm Aerial photograph ofBurnfoot Farm enclosures The resistivity survey data from Burnfoot Farm The resistivity data from Burnfoot Farm after the application of a high-pass filter Magnetic survey data from Burnfoot Farm Interpretative plot of the enclosures at Burnfoot Farm Beeston Castle The Iron Age Settlements of North West England Comparative Plans of Some Late Iron Age and Romano-British Enclosures in North West England Legh Oaks I & II. A pair oflate prehistoric and Romano-British farmsteads The Location of the Trenches at Great Woolden Hall The Main Structures Excavated at Great W oolden Hall The Romano-British Settlements of North West England during the 2nd Century AD Find-spots ofpre-Neronian aes-coins The Roman advance into north west England The Roman advance under Petillius Cerialis Roman consolidation under Agricola and in the later Flavian/Trajanic period England The Roman Gask Frontier Gask Tower Ditch Diameters Mains ofHuntingtower Small Roman Temporary Camps near the Gask system Cuiltburn: 1995/7 Excavations Gask Ridge Signalling Roman Military Diploma from Wroxeter: the outer face Roman Military Diploma from Wroxeter: the inner face Main plan ofFishbourne Roman Palace Aerial view showing the location of the 1990s excavations in front ofFishbourne Roman Palace Phase 1 - slight remains of a rectangular timber building Phase 2 - Building 3 with two rows of postholes to the north Some of the masonry foundations of Building 3 Phase 3 - Building 3, with the aqueduct cutting through the northern flanking wall Phase 4 - Building 3, with three further rows ofpostholes to the north Phase 5 - Building 3, with the central and courtyard pits Phase 6 - Drainage channel, after Building 3 had been demolished Area B - plan and section One possible reconstruction of Building 3 The Roman Road from Caersws to Bala The Regional Setting of Chester (after Waddelove 1986) The Deva Hinterland The Roodee's Eastern Side Watergate Square The Gas-Works Site Location ofWesthawk Farm, Monkton and Thurnham Geophysical Survey ofWesthawk Farm The Excavated Area Bat Westhawk Farm The Romano-British Village at Monkton Thurnham Roman Villa, Interpretative Plan Map of West Crete with Hellenistic coastal towns Map of Phalasarna Drawing ofrectangular tower and adjoining gateway Representation of rectangular tower as it appeared in antiquity

vi

31-32 33 33 48 48 49 49 56 56

57 57 58 58 70 71 71 72 72 73 74 80 81 82 83 91 91 92 92 92 93 103 104 111 112 112 113 113 114 114 115 115 116 116 122 138 138 139 140 140 152 152 153 153 154 161 161 162 162

15.5 15.6 15.7 15.8 15.9 15.10 15.11 15.12 16.1 17.1 17.2 17.3 17.4 17.5 17.6 17.7 17.8 17.9 17.10 17.11 17.12 18.1 19.1 19.2 19.3 19.4 19.5 19.6 19.7 19.8 19.9 19.10 20.1 20.2 20.3 20.4 20.5 20.6 20. 7 20.8 20.9 20.10 20.11 20.12 20.13 20.14 20.15 20.16 20.17 21. 1 22.1 22.2 22.3 22.4 22.5 22.6

Photograph of sherd from Athenian crater with red-figured representation ofa youth Drawing ofcircular tower Photograph of quay in military port Drawing of quay in inner basin Photograph of public road Photograph ofbasins in 'industrial area' Red-figure painting from a burial urn representing flying Eros between two maidens Photograph of Roman catapult stone The 'Romans always conquer' inscription Sketch map of northern Jordan Khirbet Khaw: location map Khirbet Khaw: air photograph of 11 April 1953 Khirbet Khaw: air photograph of 1990 Khirbet Khaw: air photograph of 1998 Khirbet Khaw: the west gate of the fort in 1998 Khirbet Khaw: plan of the visible remains of the fort and outline of the caravanserai Khirbet Khaw: two architectural pieces on the surface inside the fort in 1998 Kh. es-Samra: plan of the site drawn from the air photograph and showing the location and form of the fort Comparison of the plans of 4th century AD forts: (a) Umm el-Jimal; (b) Da'ajaniya; (c) Avdat 1; (d) Kh. Khaw Kh. el-Khalde: plan of the caravanserai Khirbet Khaw: tombs recorded by Briinnow and von Domaszewski Byblos, Phoenicia and the eastern Mediterranean Palmer's pen-and-ink sketch of the ancient town ofShivta Tripolitania, northwest Libya Romano-Libyan floodwater farming systems in the Wadi Gobbeen in the Libyan pre-desert Barri Jones, Bob Bewley and Peter Hayes examining the sediments exposed by a cut through wall 1250 in illustration 19.3 The location of Wadi Faynan within its region Looking northeast across the ancient field system in Wadi Faynan The survey area of the Wadi Faynan Landscape Survey A field system on the northern side of the Wadi Faynan A field map of part of the ancient field system in Wadi Faynan Section through a Roman-period water conduit channel General map ofCyrenaica Euesperides (Benghazi). RAF photograph, taken in 1948 Euesperides (Benghazi). G.R.H. Wright's plan of the city Euesperides (Benghazi). General plan of the Benghazi area Euesperides (Benghazi). G.R.H. Wright's plan of Site A Euesperides (Benghazi). Excavations in progress in 1969 Euesperides (Benghazi). The upper city (Sidi Abeid) Euesperides (Benghazi). Schematic plan of Areas Hand Euesperides (Benghazi). Area H Euesperides (Benghazi). Area L Euesperides (Benghazi). Area L. Stone quarry with rock cut tombs Euesperides (Benghazi). Area J showing section cut through road surfacings Euesperides (Benghazi). Area P mosaic Euesperides (Benghazi). Area P. Work in progress with new cemetery Euesperides (Benghazi). Area Q. Work in progress Euesperides (Benghazi). Updated plan ofEuesperides Euesperides (Benghazi). Sketch plan of Benghazi Stone mulch and laths protecting a young fig from animal damage, Formentera The Libyan Desert The ULVS survey area: distribution of all sites recorded in the Wadi Sofeggin/ZemZem region Romano-Libyan settlement sites and wadi walls in the Wadi Umm el-Kharab UL VS Settlement models The Wadi el-Agial, heartlands of the Garamantes Detailed map of the Wadi el-Agial around Genna

vii

163 163 164 164 165 165 166 166 172 183 183 184 184 185 185 186 186 187 188 188 188 204 213 213 214 214 215 216 216 217 217 218 227 227 228 229 229 230 230 231 231 232 232 233 233 234 234 235 236 243 252 252 253 254 255 256

22.7 23.1 23.2 23.3 23.4 24.1 24.2 24.3 24.4 25.1 25.2 28.la 28.lb 28.2 28.3 29.1 29.2 29.3 29.4

Model ofGaramantian settlement in the Wadi al-Agial The Roman Decapolis city at Jerash, Jordan Aldoth Bronze Age barrow Undecorated collared urn, Longworth's Primary Series (Bronze Age) Raise Howe signal tower Detail of the opencast face at Puerto del Palo Mining china clay by a form of ground sluicing in 19th century Cornwall Comparative sketch plans of (left) the mine at Brana la Folgueirosa and (right) the mine at Dolaucothi Detail of the main opencast face at Las Medulas Generic Roman force pump Examples of Force Pumps from excavated contexts The underside of the top of an inkwell The basal interior of an inkwell A small section of a form 29 mould from La Graufesenque Mould from La Graufesenque, probably for a lagena, 'signed' by Martialis in the decoration Mithraic vessels with pierced rims Detail of No. 1 from Cologne Handled Riiucherkelch from Friedberg A: figure placing a pot on an altar; B: lion, snake and crater group from the Heddernheim Mithraeum III tauroctony; C: lion, snake and crater group from a samian beaker 31.1 A weekend foray from Manchester 31.2 The sixth century baptistery at Albenga, Liguria, N Italy, external view 31.3 The sixth century baptistery at Albenga, Liguria, N Italy, internal view 31.4 The sixth century ivory seat (cathedra) of Bishop Maximian 31.5 Map from the programme of the 1994 congress at Split and Poree, Croatia, listing the venues of earlier congresses 31.6 Aerial view of the excavated church (fifth century and later), in the lower city at St. Bertrand-de-Comminges, Haute-Garonne, France 31. 7 Plan of the earliest churches in the forum at Lincoln 31. 8 The sequence of the episcopal group at Geneva 31.9 The more southerly of the double churches of the episcopal group at Salona, near Split 31.10 The early church at Cimiez, Nice 31.11 The reconstruction of the cella memoriae, excavated beneath Bonn cathedral 31.12 Part of the mosaic in the main apse of the sixth century basilica at Poree 32.1 The sarcophagus in the west portico at Polesden Lacey 32.2, 3, 4. Details of the front of the Polesden Lacey sarcophagus 32.5. Left end of the Polesden Lacey sarcophagus 32.6. Sarcophagus in the Walters Art Gallery, Baltimore 33.1 Portrait Cameo of the Severan Family, 198-209 AD 33.2 Portrait Cameo of Julia Domna, early third century AD 33.3 The Berlin Tondo portrait of the Severan family 33.4 The Apotheosis of Caracalla 33.5 The Spirit of Alexandria Mosaic

256 261 262 262 263 274 274 275 275 280 281 301 301 302 302 309 310 310 310 328 328 329 329 329 330 330 331 331 332 332 333 340 341 342 343 353 353 354 354 355

Widely Used Abbreviations CIL Corpus Inscriptionum Latinarum RIB Roman Inscriptions of Britain: Collingwood, R. G. and Wright, R. P. (eds.), 1965, Oxford, Oxford University Press

viii

PREFACE Barri Jones decided, early in 1999, to take early retirement from the Chair of Archaeology at the University of Manchester, the decision to take effect in the late summer of 2000. As news of this gradually emerged, a group of his friends and ex-students began to discuss appropriate means of celebrating his career and marking his remarkable contribution to archaeology both in Britain and abroad. It was decided to host an informal conference at Manchester in September 2000, the papers from which would then be published as a Festschrift. Contacts were made and promises of both papers and attendance poured in from three continents. Barri's sudden and tragic death in July 1999 led to the idea of a conference being abandoned and the Festschrift has now to take the form of a Gedenkschrift, that is a volume of essays offered in his memory. There is some consolation in the fact that our intentions had had to be shared with Barri at the very beginning of 1999 (only because I was scared that his ever-full diary would otherwise fail to accommodate the event). In a letter dated 4 February, Barri wrote that he was 'deeply grateful that such a move should emerge and that it should emerge from this [Manchester] University'. He was deeply touched that so many colleagues, collaborators and friends wanted to gather in his honour, and extraordinarily pleased when I told him all those who were involved and the arrangements which had been put in train. He immediately, of course, set about adding to their numbers. Very many of those involved have been able to contribute a paper to this volume, while others have wished it every success but have been prevented by circumstances from actively contributing. Such has been the goodwill and keenness to participate that several of the papers really lie outside the Roman Empire, either in terms of geography or period. All, however, have some link to Barri himself, and reflect his interests and his encouragement of others. I trust that the extraordinarily wide range of the contents will be forgiven, therefore, as providing a fair reflection of Barri's own work and his vision of archaeology as a discipline. This volume is dedicated to Barri's memory, in gratitude for his inspirational teaching, his charismatic academic leadership and his human warmth as a friend and as a colleague.

Nick Higham

ix

X

Barri Jones: A Life in Archaeology Nick Higham with Anthony Birley, David Mattingly and John-Peter Wild Barri Jones was born on April 4 1936 at St Helens, Lancashire, to Emlyn and Phyllis Margaret Jones, both teachers and both Welsh-speakers, who gave him as Welsh a string of forenames as it is possible to imagine - Geraint Dyfed Barri. However, to his later regret, they stopped speaking Welsh at home, so as to bring up their only child as an English-speaker from infancy. Their removal to Buckinghamshire led to Barri being educated at High Wycombe, where he attended the Royal Grammar School. It was there, as a teenager, that he first developed the love affair with Roman archaeology which would dominate his life, using his school holidays back in Wales to puzzle out the routes of lost Roman roads and locate camps and even a hitherto unknown fort. Not content with a mere schoolboy interest, he raised funding from Buckinghamshire Education Committee and led groups of senior pupils, past and present, from his school to carry out survey work on Roman roads, and sent in details for academic publication (see Journal of Roman Studies 46: 120-1; 47: 198). It was this, in a sense, which put him on the academic road of Roman scholarship. That is not to suggest he was a swot: Barri was an all-rounder in every sense, a leading schoolboy cadet and a very able and keen rugby player at school level, who never thereafter lost his enthusiasm for the game or his commitment to the Welsh national team.

and which was later published by the British School (in 1962 and 1963). With his doctoral thesis complete, Barri stayed on in Italy to carry forward the British School's survey of Apulia. It was here that he first immersed himself in the study of aerial photographs, which had been amassed in their thousands by John Bradford to support the survey, from training flights undertaken during the Second World War. Amongst his successes were the identification of an unexpected, major neolithic settlement in the region of Foggia, which was eventually published in Neolithic Apulia, and the systematic study of the landscape evidence for the Gracchan land reforms of the mid-second century BC - which brought him the benign attention of the great, late, Arnold Toynbee. Working in Rome also led to his involvement in the Fezzan expedition, which proved the first step in an involvement in the archaeology of North Africa, and especially Libya, which lasted up to his death. In the footsteps of Richard Goodchild, in the late 1960s Barri excavated at Tocra and Euhesperides, where he directed excavations to identify the remains of the ancient Greek city. The coup of September 1, 1969 caused excavations at Euhesperides to be closed, since it is a suburb of Benghazi, but Tocra continued, albeit without its Libyan workforce, meaning the Europeans undertaking all the manual work as several still recall. But Barri had other irons in the fire, for he had already come up with a new identification for the lost site of Hadrianopolis (he tellingly named his firstborn child Oriana several months before the season began). Despite the coup, fieldwork enabled him to prove his theory right, searching out and locating the lost city by tracing its aqueduct for seven kilometres right up to its source, the Jons aqua in 'the eyebrow of the Gebel', as the first part of the escarpment was called locally. An eyewitness recalls that 'it was a truly Dionysian scene as we got there: a large Jersey-type bull grazing under a vast, ancient fig-tree, which was growing next to the spring.'

Barri studied as a classics scholar at Jesus College, Oxford, where he received encouragement to take further his interest in archaeology from Professor (later Sir) Ian Richmond. He had been offered a lectureship in Wales even before his finals, but, with Richmond's support, he began a doctorate and won a scholarship (1959-62) to the British School at Rome, where he joined the South Etruria survey then being undertaken by John Ward-Perkins, Director of the School. It was at Rome that Barri began to attract wider attention and to establish himself as a rising star within archaeological circles. Such was his charisma that he succeeded in uniting the various British Rome scholars - whose interests were generally quite disparate to take part in a grand mass fieldwork outing in southern Etruria in 1961. He had a natural aptitude for fieldwork which so impressed Professor A. W. Lawrence of Cambridge University (younger brother of T. E. Lawrence) that he persuaded the Seven Pillars of Wisdom Trust to provide him with a car to enable him to get around more easily. His exceptional ability to 'read' an ancient landscape and to appreciate the possibilities offered by topography came on top of an enviable gift for languages and the natural Welsh charm which gave him a rare ability to enthuse even the most suspicious of 'locals' to assist him. The result was a study of Capena and its region, for which he was awarded his D.Phil from Oxford University,

From Apulia, Barri returned to the UK in 1964 and, indeed, almost to his birthplace, to take up a lectureship in archaeology and ancient history in the Department of History at the University of Manchester. There had been a succession of distinguished figures preceding him, but none except Hugh Thompson had stayed at Manchester over long. Barri proved the exception, rekindling his interest in British archaeology while still retaining his activities around the Mediterranean. With Manchester as his base, he threw himself once more into the Roman archaeology of Wales. He worked at the site of Caerau, which he had himself discovered, and the fort at Caersws

1

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

(with the late Charles Daniels, with whom he had formed a lasting friendship at Rome, and Bill Putnam). He excavated the Roman town of Carmarthen and, most famously, he re-examined the Roman gold mines of Dolaucothi, where he had already begun work before going to Italy (Jones et al. 1960). This interest in ancient mining was never sated, with fresh publication on Dolaucothi later on with Peter Lewis and further thoughts on the central issues even in the 1990s (with Keith Maude), and exploration of Roman mines in Rio Tinto and Las Medulas in Spain (Jones 1980).

archaeology of the Roman World from his appointment in 1964/5. Subsequently, a small of group of archaeologists were gathered around him and an honours degree in Ancient History and Archaeology had its first intake (of two) in 1969. Barri was an assured and charismatic teacher from the very beginning, with a rare ability to engage even the most disinterested of his listeners and to share with them his enthusiasm, knowledge and understanding, almost invariably conveyed via a mass of plans and photographs. Barri was a great believer in learning by doing, and his students found themselves sucked into weekend fieldwork and excavation in the local countryside, often being asked to take on tasks which they had never done before. From early on, he deployed a mixed group of postgraduates and undergraduates, many of whom went on to careers in archaeology. His own research students today make up a significant proportion of the archaeological profession in Britain and even beyond. The new degree programme was not just an educational experience but also an extended family, with Barri in particular offering frequent access to his own home which I recall helping to wall-paper - and even to his own bathtub at the end of a winter's day digging in Cheshire although I do remember being fourth to use the same water! Additionally he was a stalwart of the department's staff five-a-side team and a squash player who was always ready to compete against the students, albeit his technique in both areas relied rather heavily on skills learnt on the rugby pitch. From 1965 to 1968, he was also a tutor in Needham Hall, actively engaged in other aspects of student welfare.

Closer to home, he made himself an expert on the Romans in North-West England (Jones 1968) and began a campaign of excavations designed to further understanding of the comparatively neglected archaeology of the region. He worked first with Hugh Thomson at the prehistoric hillfort of Mam Tor and the Roman fort of Brough-onNoe, continuing the latter programme as a training excavation in partnership with John-Peter Wild from 1967. He worked with Peter Webster in 1965-66 at Whitchurch, then over the years at Northwich, Ribchester (1970), Lancaster (1971: published with David Shorter, 1988) and Manchester (1972, 1980). At Manchester, Barri was among the first to demonstrate that Roman archaeology still survived the rapid development of the city in the nineteenth century. The reconstruction of the Roman fort gate at Liverpool Road, and the University of Manchester (later Greater Manchester) Archaeological Unit are both monuments to his ability to wrest the past from the postindustrial landscape. The experience of digging in the built-up environment of (particularly) Manchester alerted Barri to the need to take action in advance of continuing urban regeneration nation-wide and he threw his energies behind the drive to protect Britain's archaeological heritage. He established his own archaeological consultancy, Archaeological Surveys, long before archaeology was widely thought of in commercial terms, which undertook excellent work in, inter alia, Warrington, and served him for a time additionally as a convenient and rapid publisher. From 1974 to 1978 he acted as secretary of Rescue, the charitable trust established to campaign for legislation and to profile the desperate need for action. Those who knew him during that period will recall the almost frenetic activity in which he was engaged, and the pressure which mounted on his professional and personal life. That said, it was very largely the influence of Rescue which led to the transformation of archaeology in Britain from the largely amateur status which it had hitherto enjoyed to develop a statutory service, regionally based and tied into the processes of planning approval. These changes were in no small part due to Barri's political skills, energy and dedication to a cause in which he had absolute faith. He subsequently published his own vision of the process in 1984, as The Past Impeifect.

Barri's high profile and many achievements meant that there was a growing risk that he would be lost to promotion elsewhere. When he was short-listed for the chair at Durham in 1971, R. E. Smith, then Professor of Ancient History, wisely had him promoted from lecturer, to become the first Professor of Archaeology at the University of Manchester. During the 1970s, Barri's commitment to aerial photography in Britain really took root. He had already been active in using vertical photographs in Italy, of course, and had flown in Wales - there are some rather murky runs of oblique photographs in the files dating from the 1960s. In the late 1960s and early 70s he began flying in the North West, later extending this programme to include much of Scotland and the Marches. Aerial photography had long been a valuable tool in mapping the ancient landscape in free-draining areas of (particularly) southern Britain, but in the north the limited campaigns of Professor J. K. St. Joseph had focused almost exclusively on the Roman military sites. It was Barri, alongside Dennis Harding at Durham, who seized the opportunity to build on the pre-war work of the Royal Commission in what had been Westmorland (now Cumbria), initially photographing the upstanding monuments of the limestone fells around Kirkby Stephen (Jones 1975). The extraordinary drought conditions of the summers of 1974

Alongside these activities, Barri was central to the development of archaeology as an undergraduate discipline at University level. At Manchester he taught the

2

Barri Jones: A Life in Archaeology

and 1975 led to this campaign being extended to the Solway Plain, and the discovery of a bumper crop of ancient sites of all sorts (Higham and Jones 1975). Largescale excavations of two Roman-British farm sites followed (Higham and Jones 1983), funded by the government, which created the first type-sites of rural settlement for the western land frontier of Roman Britain. All this activity spilled over into Barri's teaching: he had always used large amounts of aerial photographs but in 1976-77 he began teaching aerial photography as a subject in its own right, being one of the very first so to do (Bewley, this volume).

Roman Cumbria owes almost everything to his persistence and vision. Barri's involvement in the archaeology of so many aspects and regions of Roman Britain, combined with his topographical sense and geographical talents, made him a natural to produce (with David Mattingly) the first real Atlas of Roman Britain (1990, 1993), which pulls together in a single volume information and understandings concerning every aspect of the Roman province. This attractively produced volume has since become a standard work of synthesis, particularly for the military history and cultural assimilation of Britain into the Roman Empire. It was a matter of enormous frustration to Barri that the publisher allowed this volume to go out of print. In response he initiated discussions which will lead to its successful reprint by Oxbow Books in the course of 2001.

Barri's flying programme had demonstrated the enormous potential of aerial photography even in northern Britain and stimulated much of the subsequent commitment of government funding to the systematic flying which has since become such a significant part of the archaeological framework. On a regional level it revolutionised knowledge of sites in an area where the late Professor George Jobey, doyen of rural settlement archaeology in the North East, often used to remark that he had so little evidence that his best option was to place his caption and key on Cumbria when drawing a map of northern Britain. The result in academic terms was the writing of a volume in the Tribes of Roman Britain series on The Carvetii (Higham and Jones 1985) which could offer a completely new oversight of Roman and indigenous settlement in a frontier setting.

The pace and productivity of Barri's research in Britain would have exhausted most academics. Not Barri, who maintained his links with the archaeology of the Mediterranean throughout. In 1971, he and John Little published their work on the coastal settlements of Cyrenaica, finally publishing Hadrianopolis in 1974, and in 1973 he contributed to a multi-period volume with a paper on southern Etruria. In 1980 he took up a senior research fellowship at the British School at Rome, but he had already by then begun, in 1979, the most significant collaboration of the second half of his career, co-directing (with Graeme Barker) the UNESCO Libyan Valleys Project (1979-89) in Tripolitania. This pioneer project brought together a large contingent of British archaeologists, many of whom were Barri's ex-students, whose collective research over the ten year timeframe of the project set new standards for large-scale survey in semi-arid zones (see Hunt and Gilbertson, Barker, and Mattingly, this volume). The results were published via a series of around thirty articles, mostly in Libyan Studies, and then synthesised into a consolidated, two-volume report which Barri jointly wrote and edited, Farming the Desert (1996), which was awarded the prestigious James Wiseman Book Award by the Archaeological Institute of America in 2001. Even after the project had officially ended, Barri maintained his own commitment to Libyan archaeology by collecting and editing (along with R. Kronenburg) the writings of the late J. B. Ward-Perkins, which were published in 1993 as The Severan Buildings at Lepcis Magna.

Barri's interest was always in the interaction of military and civilian, Roman culture and provincial setting. His development of this aerial photographic programme took him into Scotland, where the northern side of the Solway revealed almost as complex a palimpsest of archaeology as the southern (Jones 1979; Jones and Walker 1983; Gregory, this volume). Further north, he flew the Moray plain, with similar results, and Inchtuthil, following up once more with work on the ground. In the Marches of Wales, among numerous lesser settlements he discovered and then trial excavated was the massive Roman marching camp at Rhyn Park, and latterly he sought the site of Caratacus' final stand against the Roman invaders. In the Isle of Man, in a programme initiated jointly with myself, systematic flying for the first time massively increased knowledge of the island's archaeology, identifying whole landscapes of unsuspected fields and farms and numerous ploughed out burial mounds. Among these was one of the densest groups of crop marks ever recorded, on the very edge of the runway of Ronaldsway Airport (Popular Archaeology 3.9, 1982; 5.8, 1984; Higham 1993: 270).

Barri's genius particularly suited working in the desert. He loved it and shared his delight with many others. He was, for example, a brilliant desert navigator - learning to read the extraordinarily unvaried landscape and memorise the minor landmarks within it. However, partly because he was always surveying the horizon, he combined this skill with a driving style that can best be described as experimental, causing wear and tear on the vehicles and skulls of those sitting in the back. Accelerating into large potholes and mud-wallows was his speciality. Despite his

Barri never took his eye off the Roman military, however, for all his interest in the wider landscape. His flying on the Solway first alerted him to the possibility of early westward extensions to Hadrian's Wall and he pursued this system relentlessly year by year, as cropping regimes slowly offered the opportunity (Jones 1976, 1982, 1991, 1993). Today our understanding of the coastal defences of

3

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

enormous energy and restlessness, however, he was remarkably patient while in North Africa, prepared to accommodate local idiosyncrasies and all the numerous small-scale disasters - such as running out of water or food, arrest (usually for photographing in prohibited areas) or the theft of key equipment.

success. No-one who has ever dug with Barri is likely to forget his terrier-like pursuit of post-holes - often across a newly cleaned surface just before it was due to be photographed - which would then leapt out to form meaningful building plans under his urgent probings. Barri was a man of action and always led from the front, with his hands-on approach and a damp flannel or linen handkerchief pulled out from some safe place to mop his brow. When it came to writing up his discoveries his great speed and restlessness occasionally led errors to creep into such details as map references, and he was sometimes inclined to over-interpret what he had found. Yet the sheer quantity of his output, and his capacity to break wholly new ground, would have broken any lesser being.

Barri's achievements were widely recognised within the archaeological establishment and his vigour and vast knowledge were highly valued. He took up many burdens over the years, both inside and outside the University. For example, he acted as head of the Department of History (1973-76) and was naturally long-term head of the new Department of Archaeology once that had been created (1981-93, 1995-6). Amalgamation of Archaeology with Art History to form the new School of Art History and Archaeology (1996-) required all his political skills to make effective but it also released Barri from bureaucratic tasks which he never felt suited his own abilities. He was always far better employed as a strategist and an active researcher within archaeology than as an administrator, and found that the charm which worked so well in the field with volunteers, farmers and land-owners was less effective in the long campaigns necessary to achieve particular goals within the University environment. In the bigger world of archaeology, however, he made a major contribution as, variously, a Trustee for Rescue and for the Vindolanda Trust (to which he introduced Alan Bowman to such good effect, and which he supported through good times and bad), Vice President of the Council of British Archaeology (1984-7), Chairman of the Society for Libyan Studies (1984-88), a member of the Ancient Monuments Advisory Committee, English Heritage (1984-88) and as a Commissioner with the Royal Commission for Ancient And Historic Monuments (Wales) (1988-99). Further afield, Barri was called in as an expert adviser to help establish an archaeological service in Lesotho and Oman, whence he also accepted an invitation to visit Yemen.

These were minor idiosyncrasies in a man whom so many had learned to admire and whose lead was so frequently followed, even by his critics. He will be remembered as one of the great pioneers of aerial photography and of landscape archaeology, and as one of the principal architects of archaeology in Britain today. Perhaps it would be just to describe Barri as one of the last great archaeological explorers of the modem age, whose discoveries on several continents were driven by sheer enthusiasm and the restless energy to know. Among his admirers were a multitude of non-professionals, for Barri was one of the finest communicators in the world of archaeology, who could carry an audience with the very slightest previous knowledge into the frontline of contemporary research. He was unstinting in his willingness to address local societies and regional groups, albeit that his diary-keeping was not always up to the task and substitutes (myself included) had on occasion to appear in his place. He was a regular expert lecturer on Swan Hellenic tours of the Mediterranean. What is more, Barri had an extraordinary ability to raise public interest from negligible levels in support of his own research. When we were working together in Cumbria, several public events set up at Penrith with an impromptu committee of local volunteers, none of whom had previously met, were attended by hundreds, and his work in the Welsh Marches similarly contributed to the formation of the Border Counties Archaeological Group, which even had its own bulletin during the late 1970s.

That is not to say that Barri was always an easy person to work with. He was not always either punctual or reliable, in part as a necessary consequence of his overcommitment, but also because he found it virtually impossible to pass by a site which interested him, however late he already was for a scheduled meeting. I can recall awaiting his arrival outside a restaurant in Kirkby Stephen to join him as arranged for lunch - he was paying since I was an impoverished student - only for him to arrive so many hours late that lunch was 'off - all without a word of explanation. There was also his inimitable style of directing excavations. Like many of his generation, he was a self-taught archaeologist and he tended to allow his enthusiasm to dominate proceedings. He was very often right, of course, and he was a brilliant motivator, but 'Jones the trowel' - as he was known early in his career would rather share the excitement of discovery at first hand than look after such mundanities as recording and finds work. For these he regularly depended on his deputies, whom it must be said he generally appointed with great

Among the features which attracted these audiences were Barri's ability to master the local geography of an area and re-present it to an audience in such a way as to provide exciting new insights into their own past. His knowledge was encyclopaedic and his memory photographic. His lectures were invariably illustrated by a wealth of aerial photographs, the vast majority of which he had taken himself These included pictures taken with a kitesupported camera system, which Barri developed in partnership with the late John Allen, a lecturer in Aeronautic Engineering at the University of Manchester who also had a long-term commitment to Roman archaeology in the Marches. Such innovations were

4

Barri Jones: A Life in Archaeology

significant in enabling Barri both to photograph excavations and also to take pictures in the North African desert, where flying was rarely possible. Another aspect of the urgency with which Barri reached out to a wider public is evident in his editorship of first Popular Archaeology (later Archaeology Today: 1979-88) and then Minerva. In this role he was tireless in promoting the dissemination of recent research in a form which could be easily assimilated. He himself wrote extensively for all three titles, under several names, as ever illustrating his texts from his vast photographic store. Barri was the first to realise that his career was less organised than it might have been, and to remark that it had virtually ground to a halt, due to great difficulties in his personal life in the mid-1990s. He had long planned, for example, to publish a jointly written North West England to AD 1540, in the Longman Regional History Series, and had taken on the authorship of a post-colonial interpretation of Roman Britain for Tempus. It was in large part in expectation of sorting out his back-log of publications that he planned to retire slightly early, intending to move permanently to his cottage at Bwlch Y Cibau, where he had maintained a family base for some years to provide a weekend and holiday home for his youngest daughter who was still at school locally. It also offered a perfect centre from which to pursue his main hobbies, which he listed shortly before he died as 'Hill Walking, Flying and Photography in Light Aircraft [and] Watching Rugby'. Tragically, his plans were not to be. It was when hill walking in Snowdonia that Barri suffered a massive heart attack on July 16 1999, and, despite being air-lifted off the hill slopes, he died in the Bronglais General Hospital, Aberystwyth, the same day. He was buried on July 29. That Barri had a gift for friendship was amply demonstrated, for the village church was so full of mourners, some of whom had travelled hundreds of miles to attend, that many had to stand throughout the service. The world of archaeology has lost a giant whose departure has already been felt in numerous places. Many, both in and outside the discipline, have also lost a true friend who is sorely missed. BIBLIOGRAPHY Higham, N. J., 1993. The Kingdom of Northumbria: AD 350-1100. Stroud: Alan Sutton

5

Bibliography of the writings of Professor Barri Jones Anthony Birley, Nick Higham, Paul Holder and David Mattingly Note: the following chronologically structured bibliography includes Barri Jones' writings under the pseudonyms of GERONTIUS and DAVID BARRIE in Popular Archaeology, but not his numerous, published, unsigned contributions

1956 Merionethshire Expedition: Interim Report, Easter 1956. [Copy in Bodleian Library]

1957 Sam Helen: Report on the Search for a Lost Section of Roman Road between Aberyscir and the Llia Valley, Brecknockshire. Archaeologia Cambrensis 106: 56-63.

1958 (with R.D.Thomson) The Roman Fort at Caerau: a Roman site in north Brecknockshire. Bulletin of the Board of Celtic Studies 17.4: 309-315.

1959 Heol y Gaer: a new Roman Marching Camp in Brecknockshire. Transactions of the Honourable Society ofCymmrodorion 1958: 82-86. Roman Merionethshire: the Roman Road West of Caer Gai. Bulletin of the Board of Celtic Studies 18.2: 208-220.

1960 Veii: The Valchetta Baths (Bagni della Regina). Papers of the British School at Rome 28: 55-69. (with W.A.C. Knowles) Roman Merionethshire: The Dolddinas Camps. Bulletin of the Board of Celtic Studies 18.4: 397402. (with I.J. Blakey and E.C.F. Macpherson) Dolaucothi: The Roman Aqueduct. Bulletin of the Board of Celtic Studies 19.1: 71-84.

1961 New Roman Camp in Merionethshire. Bulletin of the Board of Celtic Studies 19.3: 254-255. Roman Montgomeryshire: Caersws: the Roman Road System. Bulletin of the Board of Celtic Studies 19.2: 177-192.

1962 Capena and the Ager Capenas. Pt. I. Papers of the British School at Rome 30: 116-207.

1963 Capena and the Ager Capenas. Pt. IL Papers of the British School at Rome 31: 100-158. Southern Etruria 50-40 BC: an Attack on Veii in 41 B.C. Latomus 22: 773-776.

1965 Ystradfellte and Arosfa Gareg: two Roman Marching Camps. Bulletin of the Board of Celtic Studies 21.2: 174-178.

1966 Review of A.E. and J.S. Gordon, Album of Dated Latin Inscriptions: Rome and the Neighbourhood. Vol. 2-3. Berkeley: California U.P., 1964-65. Journal of Roman Studies 56: 254. (with W.G. Putnam) The Earthwork at Cann Office (Llangadfan). Montgomeryshire Collections 59: 155-158. (with F.H. Thompson) Excavations at Mam Tor and Brough-on-Noe, 1965. Derbyshire Archaeological Journal 85: 123126.

6

Barri Jones: A Life in Archaeology

(with C.M. Daniels and W.G. Putnam) Excavations at Caersws, 1966: Interim Survey. Montgomeryshire Collections 59: 112-115.

1967 (with AH.A Hogg) The Roman Marching Camp at Esgairperfedd (Radnor). Bulletin of the Board of Celtic Studies 22.3: 274-276. (with F.H. Thompson and J.P. Wild) Manchester University Excavations at Brough-on-Noe (Navio), 1966. Derbyshire Archaeological Journal 86: 99-101.

1968 Manchester University Excavations 1967: 1 Brough-on-Noe (Navio). Derbyshire Archaeological Journal 87: 154-15 8. The Roman Camps at Y Pigwn. Bulletin of the Board of Celtic Studies 23 .1: 100-103. The Romans in the North West. Northern History 3: 1-26. Review of R.J. Smutny, Greek and Latin Inscriptions at Berkeley. Berkeley: California U.P., 1966. Journal of Roman Studies 58: 304. (with P.V. Webster) Mediolanum: Excavations at Whitchurch 1965-1966. Archaeological Journal 125: 193-254. (with C.M. Daniels and W.G. Putnam) Excavations at Caersws, 1967: Interim Survey. Montgomery Collections 60: 64-66.

1969 Beulah, Caersws I, Caersws II, Carmarthen, Clun, Whitchurch, Practice Camps, Bryn Glas. In V.E. Nash-Williams, The Roman Frontier in Wales. Second edition revised under the direction ofM.G. Jarrett. Cardiff: Wales U.P. Excavations at Carmarthen, 1968. Carmarthen Antiquary 5: 2-5. (with C.M. Daniels) The Roman Camps on Llandrindod Common. Archaeologia Cambrensis 118: 124-134. (with P.R. Lewis) The Dolaucothi Gold Mines I: the surface evidence. Antiquaries Journal 49: 244-272. (with J.P. Wild) Excavations at Brough-on-Noe (Navio), 1968. Derbyshire Archaeological Journal 88: 89-96.

1970 Aecae, Arpi, A(u)sculum Satrianum, Centuriation, Ergitium, Herdoniae, Salapia, Sipontum, Teanum Apulum. In N.G.L. Hammond and H.H. Scullard (eds.) The Oxford Classical Dictionary. Second edition. Oxford: Clarendon Press. Excavations at Carmarthen, 1969. Carmarthenshire Antiquary 6: 4-14. Ribchester Roman Fort. Archaeological Journal 127: 277-279. Roman Lancashire. Archaeological Journal 127: 237-245. The Roman site at Westbury. Montgomeryshire Collections 61: 96. Review of: Ordnance Survey. The Antonine Wall: Two-and-a-Half Inch Map. Southampton: Ordnance Survey 1969. Britannia 1: 321. (with P.R. Lewis) The Roman Gold-Mines at Dolaucothi. Carmarthenshire Antiquary 6: 90-103. (with P.R. Lewis) Roman Gold Mining in North West Spain. Journal of Roman Studies 60: 169-185. (with P.V. Webster) Derbyshire Ware: a Reappraisal. Derbyshire Archaeological Journal 89: 18-24. (with J.P. Wild) Manchester University Excavations at Brough-on-Noe (Navio), 1969. Derbyshire Archaeological Journal 89: 99-106. (with C.M. Daniels and W.G. Putnam) Excavations at Caersws 1968. Montgomeryshire Collections 61: 37-42.

1971 Excavations at Northwich (Condate). Archaeological Journal 128: 31-77. Fieldwork and Air Photography in Carmarthenshire. Carmarthenshire Antiquary 7: 3-40. Review of: A Lezine, Carthage-Utique: Etudes d'Architecture et d'Urbanisme. Paris: Centre National de la Recherche Scientifique, 1968. Journal of Roman Studies 61: 297-298. (with P.R. Lewis) The Dolaucothi Gold Mines. Bonner Jahrbucher 171: 288-300. (with P.R. Lewis) The Roman Gold Mines at Dolaucothi. Carmarthen County Museum. (with J.H. Little) Coastal Settlement in Cyrenaica. Journal of Roman Studies 61: 64-79. (with J.H. Little) Hadrianopolis. Libya Antiqua 8: 53-67.

7

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

1972 The Towy Valley Roman Road. Carmarthenshire Antiquary 8: 3-16. Review of: O.A.W. Dilke, The Roman Land Surveyors: an Introduction to the Agrimensores. Newton Abbot: David & Charles, 1971. Britannia 3: 373-374. (with S. Grealey and J.H. Little) Excavations at Castell Cogan 1971: an Interim Report. Carmarthenshire Antiquary 8: 1726.

1973 Civil War and Society in Southern Etruria. In M.R.D. Foot (ed.) War and Society: Historical Essays in Honour and Memory of J. R. Western, 1928-1971. London: Elek: 277-288. Crisis in Archaeology. Journal of Environmental Planning & Pollution Control 1.3: 50-63. (with J.P. Wild) The Deansgate Dig: Roman Manchester, Interim Report of the 1972 Excavations. Manchester: Manchester Univ., Dept.of Archaeology. (with J.H. Little) The Carmarthenshire Survey. Pt. 1: Excavations on the Roman Fort at Pumpsaint, Carmarthenshire: Interim Report 1972. Carmarthenshire Antiquary 9: 3-27. (with J. Manley) Delamere - Roman Road Section, Cheshire Archaeological Bulletin I: 8-10. (with S. Grealey and J.H. Little) The Carmarthenshire Survey Pt. II: Bron-y-Gaer. Carmarthenshire Antiquary 9: 29-32.

1974 Roman Manchester. (Editor S. Grealey.) Altrincham: Sherratt. British Antiquity 1973-1974: Romano-British and Related. Archaeological Journal 131: 398-408. (with P.R. Lewis) Ancient Mining and the Environment. In P.A. Rahtz (ed.) Rescue Archaeology. Harmondsworth: Penguin: 130-149. (with J.H. Little) The Carmarthenshire Survey. Pt. I: Excavations at Pumpsaint 1973: Interim Report. Carmarthenshire Antiquary 10: 3-16.

1975 Agriculture; Vernacular Architecture. In S. Grealey (ed.) The Archaeology of Warrington's Past. Manchester: Archaeological Surveys. British Antiquity 1974-1975: Romano-British and Related. Archaeological Journal 132: 322-331. The North Western Interface. In P.J. Fowler (ed.) Recent Work in Rural Archaeology. Bradford-on-Avon: Moonraker Press: 93-106. (with S. Grealey) Mote Hill and its Environs; Gazetteer of Sites. In S. Grealey (ed.) The Archaeology of Warrington's Past. Manchester: Archaeological Surveys. (with N.J. Higham) Frontiers, Forts and Farmers: Cumbrian Aerial Survey, 1974-5. Archaeological Journal 132: 16-53.

1976 Hadrian's Wallfrom the Air. Manchester: Archaeological Surveys. Aquae Arnemetiae (Buxton), Arosfa Gareg, Blaen-cwm-bach, Bremetennacum Veteranorum (Ribchester), Caersws, Dolaucothi, Esgair Perfedd, Lancaster, Llandrindod Common, Mamucium (Manchester), Mediolanum (Whitchurch), Melandra Castle, Moridunum (Carmarthen), Navio (Brough-on-Noe), Wilderspool, Ystradfellte. In R. Stillwell (ed.) Princeton Encyclopedia of Classical Sites. Princeton: Princeton U.P. Archaeology Report: Rhyn campaign base. The Times 22 Oct: 18 British Antiquity 1975-76: Romano-British and Related. Archaeological Journal 133: 271-277. Missed Opportunities in Carlisle. Rescue News 14: 15. Penny Wise - Pound Foolish. Rescue News 13: 16. Recent Work in Roman Wales. Proceedings of the Classical Association 73: 22-23. The Western Extension of Hadrian's Wall: Bowness to Cardurnock. Britannia 7: 236-243.

1977 British Antiquity 1976-1977: Romano-British and Related. Archaeological Journal 134: 397-403. The Colchester and Ipswich Aerial Survey. Aerial Archaeology I: 23-25. Hadrian's Wall from the Air. Country Life 161 March 3: 506-507.

8

Barri Jones: A Life in Archaeology

Italy. In M.I. Finley (ed.) Atlas of Classical Archaeology. London: Chatto & Windus: 96-134. Rhyn Park Roman Fortress: excavations 1977. Border Counties Archaeology Group Bulletin l. The Romans in North East Wales. Illustrated London News 265 February: 73-74. Review of D.R. Wilson (ed.) Aerial Reconnaisancefor Archaeology. CBA Research Report No. 12. London: Council for British Archaeology. Britannia 8: 470-471.

1978 High Heritage: North Walesfrom the Air. Oswestry: North Wales Newspapers. Rhyn Park Excavation Report '78. Manchester: [n. publ.]. Concept and Development in Roman Frontiers. Bulletin of the John Rylands University Library 61: 115-144. Patterns of Coastal Change. Country Life 163: 672. (with G.A.P. Thomas) The East Somerset Survey, 1975-77. Aerial Archaeology 2: 68-70.

1979 Aerial Photography in North Wales, 1976-7. Aerial Archaeology 4: 58-64. Cause for Concern. Popular Archaeology 1.1 (July): 40-41. Cause for Concern. Popular Archaeology 1.3 (September): 44-45. Day Out on Mendip. Popular Archaeology l.l (July): 15-17. Day Out: Offa's Dyke. Popular Archaeology 1.2 (August): 5-6. The Development of the Coastal Frontier. In B. Dobson (ed.) The Tenth Pilgrimage of Hadrian's Wall. Kendal: Titus Wilson: 28-29. The Future of Aerial Photography in the North. In N.J. Higham (ed.) The Changing Past: some Recent Work in the Archaeology of Northern England. Manchester: Manchester University, Dept. of Extra-Mural Studies: 75-87. Invasion and Response in Roman Britain. In B.C. Burnham and H.B. Johnson (eds.) Invasion and Response: the Case of Roman Britain. Oxford: British Archaeological Reports, British Series 73: 57-79. The Roman Evidence. In H. Crawford (ed.) Subterranean Britain: Aspects of Underground Archaeology. London: J. Baker: 85-99. Walking where Agricola once Stood. Chartered Surveyor 111.7: 270-273. The Western Stanegate. In B. Dobson (ed.) The Tenth Pilgrimage of Hadrian's Wall. Kendal: Titus Wilson: 27. (with P. Reynolds) Roman Manchester: the Deansgate Excavations 1978: an Interim Report. Manchester: Greater Manchester Council.

1980 Roman Manchester: Duke Place & North Gate, (Interim Report). Manchester: [n. publ.]. Archaeology and Coastal Change in the North West. In F.H. Thompson (ed.) Archaeology and Coastal Change. London: Society of Antiquaries: 87-102. The Hidden Frontier. Popular Archaeology 2.1 (July): 14-17. The Roman Mines at Riotinto. Journal of Roman Studies 70: 146-165. 11Tavoliere romano: l'agricoltura romana attraverso l' Aerofotografia e lo Scavo. Archaeologia Classica 32: 85-107. (with G.W.W. Barker) From Sofeggin to Zem Zem. Popular Archaeology 2.3 (September): 23-24. (with G. W.W. Barker) Libyan Valleys Survey. Libyan Studies 11: 11-36. (with D.J. Mattingly) Fourth Century Manning of the 'Fossatum Africae'. Britannia 11: 323-326.

1981 The Erosion of British Archaeology: People, Prospects and Priorities. B. Jones [ed]. Popular Archaeology 3.5 (November): 11-21. Italian Round-Up. Popular Archaeology 3.2 (August): 41-42. Review of J.D. Evans, B. Cunliffe & C. Renfrew (eds.) Antiquity and Man: Essays in Honour of Glyn Daniel. London: Thames & Hudson. Popular Archaeology 3.2 (August): 43. Review ofT.W. Potter, The Changing Landscapes of South Etruria. London: Elek, 1979. Antiquity 55: 143-144. (with G.W.W. Barker) The UNESCO Libyan Valleys Survey 1980. Libyan Studies 12: 9-48.

1982 Growing up with Rescue Archaeology. Portico: Journal of the Faculty of Architecture and Surveying: 5-11.

9

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Libyan Rock Art. Popular Archaeology 3.8 (February): 27-32. Rome's Northern Frontiers: view of the North. Popular Archaeology 3.12 (June): 40-44. Scapula, Paulinus or Agricola? Popular Archaeology 3.7 (January): 16-21. The Solway Frontier: Interim Report, 1976-1981. Britannia 13: 283-297. (with G.W.W. Barker) The UNESCO Libyan Valleys Survey 1979-1981: Palaeoeconomy and Environmental Archaeology in the Pre-Desert. Libyan Studies 13: 1-34. (with N. Higham) Manx Archaeology from the Air. Popular Archaeology 3.9 (March): 8-13. (with P. Lewis) Day Out: the Roman Gold-Mines at Dolaucothi. Popular Archaeology 4.1 (July): 10-14.

1983 The Development of Air Photography. Popular Archaeology 5.2 (August): 27-30. Excavations at Tocra and Euhesperides, Cyrenaica 1968-1969. Libyan Studies 14: 109-121. (with G.W.W. Barker) The UNESCO Libyan Valleys Survey IV: The 1981 Season. Libyan Studies 14: 39-68. (with C. Daniels) More to Come ...a Road Section at Gannochy, Perth. Popular Archaeology 4.10 (May): 22-23. (with N.J. Higham) The Excavation of two Romano-British Farm Sites in North Cumbria. Britannia 14: 45-72. (with P. Reynolds) Northwick- Excavations at Castle, 1983. Cheshire Archaeological Bulletin 9: 82-5. (with J. Walker) Either Side of Solway: Towards a Minimalist View of Romano-British Agricultural Settlement in the North West. In J.C. Chapman and H.C. Mytum (eds.) Settlement in North Britain 1000 BC-AD 100: Papers Presented to George Jobey, Newcastle upon Tyne, December 1982. Oxford: British Archaeological Reports, British Series 118: 185204.

1984 Past Imperfect: the Story of Rescue Archaeology. London: Heinemann. 'Becoming Different without Knowing It': The Role and Development of vici. In T.F.C. Blagg and AC. King (eds.) Military and Civilian in Roman Britain: Cultural Relationships in a Frontier Province. Oxford: British Archaeological Reports, British Series 136: 75-91. The Byzantine Bath-House at Tocra: a Summary Report. Libyan Studies 15: 107-111. Capital Rescue. Popular Archaeology 5.12 (June): 28-34. The Origins of Rome: from Myth to Metropolis. Popular Archaeology 6. l (July): 33-37. (with G.W.W. Barker) The UNESCO Libyan Valleys Survey VI: Investigations of a Romano-Libyan Farm, Part I. Libyan Studies 15: 1-44. (with N. Higham) From Ronaldsway to Ramsey: Air Photography in the Isle of Man. Popular Archaeology 5.8 (February): 7-13.

1985 Beginnings and Endings in Cyrenaican Cities. In G.W.W. Barker, J. Lloyd and J. Reynolds (eds.) Cyrenaica in Antiquity. Oxford: British Archaeological Reports, International Series 236: 27-41. Concluding Remarks. In D.J. Buck and D.J. Mattingly (eds.) Town and Country in Roman Tripolitania: Papers in Honour of Olwen Hackett. Oxford: British Archaeological Reports, International Series 274: 307-311. Introduction: Field Survey Today. In S. MacCready and F.H. Thompson (eds.) Archaeological Field Survey in Britain and Abroad. London: Society of Antiquaries: 3-7. The Libyan Valleys Survey: the Development of Settlement Survey. In D.J. Buck and D.J. Mattingly (eds.) Town and Country in Roman Tripolitania: Papers in Honour of Olwen Hackett. Oxford: British Archaeological Reports, International Series 274: 263-289. Welcome to Tunisia. Popular Archaeology 6.14 (December): 2-8. (with N.J. Higham). The Carvetii. Gloucester: Sutton. (with G.W.W. Barker) Investigating Ancient Agriculture on the Saharan Fringe: the UNESCO Libyan Valleys Survey. In S. MacCready and F.H. Thompson (eds.) Archaeological Field Survey in Britain and Abroad. London: Society of Antiquaries: 225-241. (with B.J.N. Edwards, P.V. Webster and J.P. Wild) Excavation on the Western Defences and in the Interior, 1970. In B.J.N. Edwards and P.V. Webster (eds.) Ribchester Excavations. Pt. 1. Cardiff: University College, Dept. of Extra-Mural Studies: 19-40.

1986 Rhodes: Classical to Crusade. Popular Archaeology 7.7 (August): 17-25.

Barri Jones: A Life in Archaeology

Roman Military Site at Cawdor? Popular Archaeology 7.3 (April): 13-16. Review ofS. Ireland: Roman Britain: a Sourcebook. London: Croom Helm. Popular Archaeology 7.5 (June): 42. (with D.J. Mattingly) A New Clausura in Western Tripolitania: Wadi Skiffa South. Libyan Studies 17: 87-96. (with P. Reynolds) The Duke Place Excavations on the Site of the North Western Comer of the Later Forts. In J.S.F. Walker (ed.) Roman Manchester: a Frontier Settlement. The Archaeology of Greater Manchester Vol. 3. Manchester: Greater Manchester Archaeological Unit: 13-20.

1987 Apulia. Vol. 1: Neolithic settlement in the Tavoliere. With contributions by C.D. Smith and D. Trump. Research Report of the Society of Antiquaries of London No. 44. London: Society of Antiquaries. Alexandria: "Next to Egypt" Archaeology Today 8.2 (March): 27-31. (with K.A. Brown) The Prehistoric Communities of Apulia. Archaeology Today 8.6 (July): 29-34. (with I. Keillar) Flights into Yesterday I: Aerial Photography in Scotland. Archaeology Today 8.6 (July): 22-26. (with K. Maude) Excavations at Dean Hall, Little Dean, 1985. Manchester Archaeological Bulletin 1: 38-41. (with K. Maude) The Solway Frontier 1984: Roman Frontier Studies. Manchester Archaeological Bulletin 1: 3-9. (with P. Reynolds) A Probable Roman Supply Depot at Llansantffraid-ym-Mechain (Powys). Manchester Archaeological Bulletin 2: 10-16. (with P. Reynolds) Roman Campaign into North Wales. Archaeology Today 8.7 (August): 21-26. (with B.J.N. Edwards and P.V. Webster) Trial Excavations in the Playing Fields, 1968-9. In B.J.N. Edwards and P.V. Webster (eds.) Ribchester Excavations. Pt. 2. Cardiff: University College, Dept.of Extra-Mural Studies: 13-27. (with I. Keillar and K. Maude) Excavations at Cawdor 1986: the Roman Site at Easter Galcantray. Manchester Archaeological Bulletin 2: 17-26. (with M. Nevell and P. Reynolds) Excavations at Castle, Northwich, 1983-1986. Manchester Archaeological Bulletin 1: 35-37.

1988 Searching for Caradog. Archaeology Today 9.4 (April): 36-39. Review of: Deplacements des Lignes de Rivage en Mediterranee d'apres les Donnees de l'Archeologie. Colloque International du CNRS, Aix-en-Provence, 5-7 Septembre 1985. Paris: Editions du Centre National de la Recherche Scientifique, 1987. Libyan Studies 19: 159-160. Review of R. Osborne, Classical Landscape with Figures: the Ancient Greek City and its Countryside. London: Philip, 1987. Antiquity 62: 189. Review ofL.F. Pitts and J.K. St Joseph: Inchtuthil: the Roman Legionary Fortress Excavations 1952-65. London: Society for the Promotion of Roman Studies, 1985. Britannia 19: 527-530. (with D.C.A. Shotter). Roman Lancaster: Rescue Archaeology in an Historic City, 1970-75. Brigantia Monographs. Manchester: Dept. of Archaeology, University of Manchester. (with G.W.W. Barker) Farming the Libyan Desert: The UNESCO Libyan Valleys Survey. Manchester Memoirs 126: 118131. (with N.J. Higham) Aerial Archaeology in the Isle of Man. Manchester Archaeological Bulletin 3: 58-63. (with R. Kronenburg) The Severan Buildings at Lepcis Magna. Libyan Studies 19: 43-53.

1989 The Development of Air Photography in North Africa. In D. Kennedy (ed.) Into the Sun: Essays in Air Photography in Archaeology in Honour of Derrick Riley. Sheffield: J.R. Collis: 29-4 7. Town and City in Tripolitania: Studies in Origin and Development 1969-1989. Libyan Studies 20: 91-106. The Western Stanegate. In C.M. Daniels (ed) The Eleventh Pilgrimage of Hadrian's Wall. Newcastle-upon-Tyne: Society of Antiquaries ofNewcastle-upon-Tyne: 92-95. (with R. Birley) Carvoran-Magna. In C.M. Daniels (ed.) The Eleventh Pilgrimage of Hadrian's Wall. Newcastle-uponTyne: Society of Antiquaries ofNewcastle-upon-Tyne: 41-43. (with E. Waddelove and AC. Waddelove) The Roman Settlement at Ruthin, Clwyd. Britannia 20: 249-254.

1990 A Brief History of Rome; The Etruscans and Rome; Roman Technology; Roman Architecture; The Roman Army. In B. Cunliffe (ed.) Swan Hellenic Cruise Handbook 1: Landscape and People.

11

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Casablanca; Rabat; Meknes; Marrakesh; Fez; Volubilis; Lixus; Tangier; Tipasa; Djernila (Cuicul); Constantine - Cirta; (with O Brogan) Annaba (Bane) - Hippo Regius); Tripoli (Oea). In B. Cunliffe (ed.) Swan Hellenic Cruise Handbook 6: The Western Mediterranean. Museums of the World: Presenting Libya's Past Minerva: the International Review of Art and Archaeology 1.1: 24-25. North East Wales: the Empty Quarter. In B.C. Burnham and J.L. Davies (eds.) Conquest, Co-existence and Change: Recent Work in Roman Wales. Trivium 25: 123-129. Ostia; Rome; Messina; Aquileia. In B. Cunliffe (ed.) Swan Hellenic Cruise Handbook 5: Italy and the Central Mediterranean. Searching for Caradog. In B.C. Burnham and J.L. Davies (eds.) Conquest, Co-existence and Change: Recent Work in Roman Wales. Trivium 25: 57-64. (with D.J. Mattingly). An Atlas of Roman Britain. Oxford: Basil Blackwell. Reissued Oxford: Oxbow Books, 2001. (with K. Maude) Dolaucothi: the Dating Problem. In B.C. Burnham and J.L. Davies (eds.) Conquest, Co-existence and Change: Recent Work in Roman Wales. Trivium 25: 169-171. (with K. Maude) Sifting the Sand: New Evidence for Core and Periphery at Lepcis Magna in Tripolitania. Manchester Archaeological Bulletin 4: 4-8. (with D.A. Welsby) UNESCO Libyan Valleys Survey: the Final Season 1989. Manchester Archaeological Bulletin 5: 5157. (with I. Keillar and J. Trust) Scientific Techniques: Radar Surveying. Minerva: the International Review of Art and Archaeology 1. 1: 18-19. (with E. Waddelove and A.C. Waddelove) The Roman Fort at Ruthin, Clwyd. Britannia 21: 299-302. (with E. Waddelove and A.C. Waddelove) The Roman Forts at Ruthin. Manchester Archaeological Bulletin 4: 32-38.

1991 Abertanat and Llanymynech: Survey and Excavations 1991. Manchester Archaeological Bulletin 6: 29-35. The Emergence of the Tyne-Solway Frontier. In V.A. Maxfield and M.J. Dobson (eds.) Roman Frontier Studies 1989. Proceedings of the XVth International Congress of Roman Frontier Studies. Exeter: Exeter U.P.: 98-107. Famdon: an Archaeological Opportunity. Manchester Archaeological Bulletin 6: 75.77. Tarradale: Investigation of a Cropmark Site near Muir of Ord, Ross & Cromarty. Manchester Archaeological Bulletin 6: 13-19. (with K. Maude) Dating and Dolaucothi. Britannia 22: 210-211. (with G.W.W. Barker, D.D. Gilbertson and D.A. Welsby) UNESCO Libyan Valleys Survey XXIII: the 1989 Season. Libyan Studies 22: 31-60.

1992 The Archaeology of Pre-Islamic Dhofar. Manchester Archaeological Bulletin 7: 44-51. Old Police House Camp, Bowness-on-Solway. Britannia 23: 230-231. (with K. Maude and J. McMorrow) Red Cat Farm: Investigation of a Crop Mark Site at Burscough Bridge, Lancashire. Manchester Archaeological Bulletin 7: 61-66.

1993 J.B. Ward-Perkins: The Severan Buildings of Lepcis Magna: an Architectural Survey. With a contribution by B. Jones and R. Ling. Edited by P. Kenrick. General Editor B. Jones. London: Society for Libyan Studies. Excavations on a Coastal Tower, Hadrian's Wall: Campfield Tower 2B, Bowness-on-Solway. Manchester Archaeological Bulletin 8: 31-39. Review ofH.M. Cotton and J. Geiger, Masada 2. The Yigael Yadin Excavations 1963-1965. Final Reports: the Latin and Greek Documents. Jerusalem: Israel Exploration Society, 1989. Journal of Roman Studies 83: 241-244. Review of D.H. French and C.S. Lightfoot (eds.) The Eastern Frontier of the Roman Empire: Proceedings of a Colloquium held at Ankara in September 1988. Oxford: British Archaeological Reports, International Series 143, 1989. Journal of Roman Studies 83: 241-244. Review ofB.H. Isaac, The Limits of Empire: the Roman Army in the East. (Rev. Ed.) Oxford: Clarendon Pr., 1992. Journal of Roman Studies 83: 241-244. Review of D. Kennedy and D. Riley, Rome's Desert Frontier from the Air. London: Batsford, 1990. Journal of Roman Studies 83: 241-244. Review of H. Solin and M. Kajava (eds.) Roman Eastern Policy and other Studies in Roman History: Proceedings of a Colloquium at Trarninne, 2-3 October 1987. Helsinki: Societas Scientiarum Fennica, 1990. Journal of Roman Studies 83: 241-244.

12

Barri Jones: A Life in Archaeology

(with I. Keillar and K. Maude) The Moray Aerial Survey: Discovering the Prehistoric and proto-Historic Landscape. In D.H. Sellar(ed.) Moray: Province and People. Edinburgh: 47-74.

1995 Farnhill: Excavations on the Solway Frontier, 1994. Manchester Archaeological Bulletin 9: 23-27. Review of F. Rakob (ed.) Simitthus I: die Steinbruche und die antike Stadt. Mainz: von Zabem, 1993. Classical Review 45: 203-204.

1996 Squaring the Circle: Parallels for the Chester 'Elliptical' Building. Archaeology North West 2.4: 105-107. Review of M. Fulford and R. Tomber (eds.) Excavations at Sabratha, 1948-1951. Vol. 2: The Finds Pt. 2: thefinewares and lamps. London: Society for Libyan Studies, 1994. Classical Review 46: 334-335. Review ofH. Welfare and V. Swann, Roman camps in England. London: Royal Commission on the Historical Monuments of England, HMSO. Britannia 27: 489-491. (with G.W.W. Barker, D.D. Gilbertson and D.J. Mattingly) Farming the Libyan Desert: the UNESCO Libyan Valleys Archaeological Survey. 2 vols. Paris: UNESCO. (with I. Keillar) Marinus, Ptolemy and the Turning of Scotland. Britannia 27: 43-49. (with D.J. Mattingly and G.W.W. Barker) Architecture, Technology and Society: Romano-Libyan Settlement in the Wadi Umm-el Agerem, Tripolitania. In L. Bacchielli and M. Bonanno Aravantinos (eds.) Scritti di Antichita in Memoria di Sandro Stucchi. Vol. 2. Roma: L'Erma di Bretschneider: 101-113.

1997 From Brittunculi to Wounded Knee: a Study in the Development of Ideas. In D. Mattingly (ed.) Dialogues in Roman Imperialism: Power, Discourse and Discrepant Experience in the Roman Empire. Journal of Roman Archaeology Supplement Vol. 23. Portsmouth, R.I.: JRA: 185-200.

1998 (with M. & A. Webb and N. Bevan) 'Pleasant Libyan Acres': the Territory of Cyrene. In E. Catani and S.M. Marengo (eds.) La Cirenaica in Eta Antica: atti def Convegno Internazionale di Studi, Macerata, 18-20 Maggio 1995. Pisa/Roma: Istituti Editoriali Poligrafici e Intemazionali: 281-288.

1999 Preface: 'Until the lions learn to write, tales of hunting will always glorify the hunter'. In M. Nevell (ed.) Living on the Edge of Empire: Models, Methodology & Marginality, Late-Prehistoric and Romano-British Rural Settlement in North West England. Archaeology North West 3: 7-8. Conclusion: The North West and Marginality: their Fault or Ours? A Warning from the Cumbrian Evidence. In M. Nevell (ed.) Living on the Edge of Empire: Models, Methodology & Marginality, Late-Prehistoric and Romano-British Rural Settlement in North West England. Archaeology North West 3: 90-95. 2000

Aerial Archaeology around the Mediterranean. In M. Pasquinucci and F. Trement (eds.) Non-Destructive Techniques Applied to Landscape Archaeology. The Archaeology of Mediterranean Landscapes 4. Oxford: Oxbow Books: 49-60. 2001

(with D. Woolliscroft) Hadrian's Wallfrom the Air. Stroud: Tempus.

13

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

14

1

The Anavionenses Anthony Birley An honorific inscription now in the Vatican, found at Foligno in Umbria, ancient Fulginiae, presumably from a statue-base, supplies in line 3 an intriguing piece of information about Roman Britain: [---}..o, prae[f(ecto) coh]ortis, trib(uno) milit[um, p]raef(ecto) equit(um), censito[ri] Brittonum Anavion[ens(ium)J, 4 proc(uratori) Aug(usti) Armeniae mai(oris), ludi magni, hereditatium et a censibus, a libellis Aug(usti), praef(ecto) vigilum, praef(ecto) Aegy[pti}, 8 M Tammius Ce ... [---}

procurator of Greater Armenia straight afterwards (thus among many, e.g. Rivet and Smith 1979: 249; Birley 1981: 302; Rivet 1982: 321; Keppie 1989: 61). This was to reject or overlook the opinion of H.-G. Pflaum. He postulated a gap between Haterius' equestrian militiae and his procuratorships and higher prefectures, held in rapid succession ('excessivement rapide')-there were four separate posts, all at Rome, between Armenia Major, not before AD 114, and Egypt in 120. When Haterius became prefect of Egypt, Pflaum argued, he was probably aged about fifty, hence would have been well over forty as procurator of Armenia maior, a post which Pflaum rated fairly low in the hierarchy. If Haterius had still been an equestrian officer just before this, he would have been a rather elderly one-perfectly possible, of course (cf. E. Birley 1953: 135ff.); but Pflaum preferred to postulate an interruption before what became 'une tres belle carriere'. He showed, it must be added, that before Hadrian's reign junior census-officers such as Haterius did not enjoy the status of procurator (Pflaum 1960-1: 217-9, no. 95, cf. 176, no. 79). New evidence from Vindolanda indicates that Pflaum was right-at any rate, Haterius was in Britain a good dozen years before the annexation of Greater Armenia.

(Dessau 1892-1916: no. 1338=CJL XI 5213; not necessarily funerary, as Rivet and Smith 1979: 249; Rivet 1982: 321). The honorand's names and perhaps other personal details (e.g. tenure of some local office) are lost from the top; the cognomen (probably e.g. Celer) of M. Taminius, the man responsible for the monument, is not fully legible; and what followed this is missing, no doubt a phrase such as amico optima, indicating Taminius' connection with the honorand, and e.g. l(ocus) d(atus) d(ecreto) d(ecurionum). But what survives gives the complete career in the imperial service of a Roman knight who rose to be prefect of Egypt. He began with one post in each of the equestrian tres militiae. The units are not specified-but the third and last post, as prefect of cavalry, was probably coupled with what follows, the role as census-officer, censitor, of people designated in the genitive Brittonum Anavion[ens(ium)}, 'of the Anavion[ensian] Britons'. Thereafter he entered the administrative hierarchy, first as procurator of Greater Armenia. That country was only briefly a Roman province, which supplies a clear date: between AD 114, when it was annexed by Trajan, and 117, when Hadrian abandoned it (see e.g. Birley 1997b: 68f£, 78). Further, the inscription was found next to one honouring a Hadrianic senator: T.Haterius Nepos Atinas Probus Publicius Matenianus (Dessau 1892-1916: no. 1058=CJL XI 5212). This man was suffect consul in AD 134, having been governor of Arabia from c. AD 130, where he continued in office during the Jewish revolt of AD 132-6, which he helped to suppress, receiving an honorary triumph. He was later governor of Upper Pannonia (see now Eck 1999: 84-6, 89). These two items long ago made it clear that the former censitor must be the elder T. Haterius Nepos, abundantly documented as prefect of Egypt between AD 120 and 124. He was doubtless father of the cos. 134 (and perhaps also polyonymous; if so, one could restore the lost opening of his inscription as e.g. [T. Haterio ..f Nepoti Atinati ProbJo).

The new evidence, still not fully published, is a letter (found in 1993) from Haterius Nepos to one of the Vindolanda prefects, Flavius Genialis, a man already attested by several other writing-tablets, published and unpublished. Not all are definitely attributable to Genialis: doubtful cases are given a question-mark; incomplete or uncertain lettering is underlined: Tabulae Vindolandenses (henceforth TV) II nos. ?217 (period IV): a fragmentary letter beginning [ }.tilis Genial[i suo (this might be a different Genialis). 218 (period III): the address on the back of a partly preserved letter: FLAVIO GENIAL!. 219 (period II): the opening of a letter (no more preserved): Flavius f..roculus Geniali syfo} (the correspondent recurs in Inv. 93/1337, below). ?220 (period III): the address on the back of a letter, of which no more is preserved: [ ... ] GENIA[LI PRA]EFE[C]I[O] (?). 221-3 (all period III): the addresses read respectively: FLAVIO GENIAL[Il; FLAVIO GENI[ ](?); [GrnNIALI. ?224 (period III): the beginning of a letter (no more preserved): Licinius Asper G[eniali suo?}. 256 (period III): letter beginning [FJ lavius Genialis Ceriali suo salutem. this fragmentary but interesting text suggests that Genialis was a predecessor of Flavius Cerialis at Vindolanda. 301 (period II): the address of a short letter to Candidus, evidently slave of Genialis, whose title

The elder Haterius' census in Britain has generally been dated c. AD 110-112, on the assumption that he became 15

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

of prefect, although very faint, seems to be given: CANDIDO GENIALIS PRAEF a Severo ....i servo. ?303 (period I ditch): the address on the back of a fragmentary letter to a man called Albiso may read ALBISONI GE .... [ ], i.e. 'to Albiso, (slave) of Ge[nialis?] '.

least some of the flimsy leaf-tablets, would have been churned up during building operations. Genialis may, for example, have been prefect from the end of period II into the start of period III, c. AD 97/8 to c. 100; a slightly earlier tenure cannot be ruled out. Haterius calls Genialis 'lord, dearest brother', standard forms between brother-officers (cf now A.R. Birley, forthcoming). Genialis is known to have been a correspondent of Flavius Proculus (TV II 219), also /rater (Inv. 93/1337). This makes Prof[ulo] plausible as the man to whom Haterius wrote 'in the same words'. Genialis and Prof[ ulus?] were clearly to confer with Haterius at Coria, now identifiable as Corbridge (TV II: pp. 96f.), then garrisoned by a cavalry regiment. 1 Haterius was surely prefect of the Coria ala when he wrote to Genialis. The tone of his letter suggests that he was senior to Genialis. It is not improbable that he already had the census assignment, perhaps as early as c. AD 98. Or it could be dated slightly later, c. 100, but not much, at latest 104/5, when Vindolanda's period III ended-for the Anavionenses seem to have left a trace at Vindolanda itself in that period (see below).

Unpublished tablets (my own, provisional readings; full texts with commentary will appear in the next volume of Tabulae Vindolandenses, TV III, ed. by A.K. Bowman and J.D. Thomas): Inv. 93/1337 (period III?): the opening of a letter: Flavius Proculus Geniali suo. What little survives includes /rater, 'brother'. Inv. 93/1378 (period III?): part of a letter, of which the address reads: FLAVIO GENIAL! PRAEF. Inv. 93/1434+ 1449 (period IV): part of a largely illegible letter with the address GENTILI LIBERTO FLAVI GENIALIS (1434); below (1449) amico a Vegeto Genialis, 'to Gentilis, freedman of Flavius Genialis, (his) friend, from Vegetus (slave of) Genialis'. Finally, there is the letter to Genialis from Haterius Nepos, Inv. 93/1379 (period III?). This reads, in part (the opening greeting is missing and lines 1-4 are rather unclear): lines 5ff. tanto magis venturum Caris sicut•constituisti spero scripsi isdem verbis et Prof[ ulo?J, 'the more so do I hope that you will come to Coria, as you decided; I have written in the same words to Prof[ulus?].' The letter ends: vale domine /rater karissime, 'Farewell, my lord, dearest brother'. On the back is the address: FLAVIO GENIAL! PRAEF COH ab Haterio N_epot[e].

Before considering what they were doing at Vindolanda, one must ask who the Anavionenses were, where they were located-and what was involved when a Roman officer took a census among them. In his Britannia, Sheppard Frere was agnostic on the first two points: 'It is not certain what or where the Britons of Anavio were. The term Brittones implies northern Britain, and they are likely to have belonged to some area under direct rule, named from a fort ...The river Annan in Dumfriesshire was probably named ANAVA. .., but the region of Anavio cannot have been there, as southern Scotland lay outside the province between c. 105 and 140' (Frere 1967: 199 n.1; still in Frere 1991: 203 n.9). This interpretation was at once questioned by A.L.F. Rivet (Rivet 1969: 248£) and rejected by him in detail some years later (Rivet 1982). In their great work on the place-names, he and Colin Smith had meanwhile enshrined the location of Anavionenses 'in the region of the ...Anava' (a name known only from the Ravenna Cosmography, 108), viz. the river Annan (Rivet and Smith 1979: 249f., citing other works, including RIB I, p. 95).

The new tablet confirms that Flavius Genialis was indeed prefect, clearly of cohors VIII! Batavorum equitata, at Vindolanda in periods II, c. AD 92-c. 98, and III, c. 98104/5: for the chronology see R. Birley 1994, with the slight extension of period III noted by Bowman and Thomas 1996: 310f., discussing Inv. 93/1474A, expensa of a prefect who is certainly Flavius Cerialis. They convincingly interpret that text as showing that Cerialis was at Vindolanda in AD 102, 103 and almost certainly until mid-July 104. Soon after, perhaps not until 105, the Ninth Batavians left Vindolanda, never to return. Genialis was, it seems, a predecessor of Cerialis (cf. above on TV II 256). It is true that Genialis' correspondence has been found in levels assigned to at least three separate periods, II, III and IV. But III may be regarded as a continuation of II, with the praetorium and south gate, first built in II, the principal areas excavated of the Ninth Batavians' fort, having been reconstructed in III in almost the same position. Period IV, when cohors I Tungrorum (which had been at Vindolanda in period I) returned, saw a radical reconstruction: a barrack-block replaced the Batavians' praetorium. Naturally material from previous periods, not

Annandale is a substantial region. The distance from the river's source, beyond the Devil's Beeftub, north of Moffat, to its confluence with the Solway, just beyond Annan town, is a good fifty kilometres as the crow flies. Taking into account tributaries such as Evan Water, Moffat Water, Water of Ae and Dryfe Water, its breadth from east to west is at least thirty kilometres. Whether its people in the first century AD also controlled the areas immediately east and west of the lower reaches of the valley is not clear. 1 The 'Corstopitum' or Coria garrison was no doubt the ala Petriana, cf. RIB 1172, the tombstone of its signifer in Hexham Abbey; Birley 1961: 149f.; P. Bidwell, in Bidwell (ed.) 1999: 13, summarises present thinking-the 'Corstopitum' fort's period I lasted from c. 86-c. 105.

16

Anthony Birley: The Anavionenses

The country as far as Gretna on the north side of Solway or even up to the R. Esk might also go with Annandale. On the west, the R. Nith, Ptolemy's Novios and (emended) Novius of Ravennas, was once supposed to have given the Novantae their name. Rivet and Smith firmly reject this: '[t]he present ethnicon can hardly derive from the tribe's dwelling near the river Novius'. The Novantae were further west, in Galloway (1979: 425, cf. 428). Whatever the exact extent of potentially Anavionensian territory, it had significant ancient settlements, including impressive hillforts, notably Burnswark Hill, 'which must be a key site in any appraisal of the prehistoric archaeology of Annandale' (Royal Commission AHMScot. 1997: 9, and passim for other sites, a few mentioned below).

perhaps from York, in the early third century - that is exactly when the cult of Brigantia was being encouraged and there is no particular reason to suppose that the dedication reflects a local ancient identity with the tribal deity .... Alongside the latter group [Cartimandua's Brigantes] other tribes can be identified, which may at times have been confederates within greater Brigantia, but at some stage in the late pre-Roman period emerge with a separate and distinctive identity'. Names of other peoples in the supposed 'Greater Brigantian' area must be considered (sources in Rivet and Smith 1979, passim): two are registered by Ptolemy, Setantii in the Fylde and Gabrantovices on the Yorkshire coast; further there are, perhaps, combining and emending the 'Corstopitum' of the Antonine Itinerary and 'Corie Lopocarium' of Ravennas, the 'Sopites' or some such name around Coria-Corbridge (Rivet and Smith 1979: 322f., but see below); the rest are known from epigraphy, Textoverdi (or Textoverdi) in the South Tyne valley, Corionototae around Corbridge (?) and Carvetii of Cumbria. They are mostly interpreted as sections of the Brigantes (see esp. Hartley and Fitts 1988: 1£, who add the hypothetical 'Latenses' of the Leeds area; note also the suggestion, mostly ignored, that even the Parisi could also have formed part of the so-called 'Brigantian federation', A.R. Birley 1980: 164f., approved by Higham and Jones 1985: 9, 140 n.12). Yet there is really no direct evidence that the Brigantes did include all these 'sub-tribes' or 'septs'. Of course, the Romans re-drew boundaries, often ignoring natural and ethnic divisions, as Strabo complained (177, 4.1.1, on Gaul; 626, 13.4.12, on W. Asia Minor). The creation of the British Belgae is a case in point (Rivet and Smith 1979: 267). And who knows what inter-British conflicts had already eliminated independent peoples before Rome established the province? Four out of five of the British peoples who sent embassies to Caesar in 54 BC (De hello Gallico 5.21.1) sank without trace: his 'Cenimagni' are no doubt the Iceni; the Segontiaci, Ancalites, Bibroci and Cassi never recur (Rivet and Smith 1979).

This creates a problem, for Burnswark----or the Roman fort of Birrens four kilometres SE-has been claimed for a north-western extension of the Brigantes. The view has become canonical. Eric Birley restated it half a century ago, summarising earlier discussions (Birley 1953: esp. 3336). Its principal basis is the statuette of the goddess Brigantia from Birrens, inscribed Brigantiae s(acrum) Amandus arc(h)itectus ex imperio imp(eratum fecit?) (RIB 2091). At least two more dedications to the presiding deity of the Brigantes have been found outside the area for which Ptolemy gives Brigantian po leis (one at Corbridge, by a centurion of VI Victrix (RIB 1131) and another at South Shields, by a man called Congennicus (RIB 1053)), as well as in Yorkshire, the Brigantian heartland (RIB 623, 627, 628, 630). The dedication on the altar (long ago lost) from near Brampton, by a Caracallan procurator of Britain, is published as deae Nymphae Brig(antiae), RIB 2066. Yet the name is shown as BRIC (BR ligatured) in the only source, the Cotton Julius MS. Nymphs were normally water-spirits: one of these, called Bricia, is known at Luxeuil in Sequanian territory, CIL XIII 5426-whereas the powerful and robust Brigantia was surely no nymph or water-deity. The notion of a north-western pocket of Brigantian territory, which was to be cut off from the rest of the civitas by Hadrian's Wall, is accepted, with sometimes a caveat or slight hesitation, by (to give only a selection) Rivet and Smith (1979: 279), Maxwell (1980: 7), Rivet (1982: 321), Hartley and Fitts (1988: 1 f£, esp. 5), Hanson and Campbell (1986: 87), Hanson (1987: 57, 65, 91), Jones and Mattingly (1990: 277-280), Whittaker (1994: 81f.), the Royal Commission AHMScot. (1997: 169f£), Strange (1997: 26), Breeze and Dobson (2000: 32, 46). A rare sceptic is Todd (1981: 142 and 265, n.3); and Mann and Breeze (1987: 89) show welcome caution over the evidential value of the Brigantia dedication. It should be added that Higham (as he kindly reminds me) has also expressed well-grounded doubts about the majority interpretation of the Birrens altar (1986: 145-6):

Even Ptolemy gives only rough guidance on Brigantian boundaries: 'below the Selgovae and Votadini, stretching from sea to sea', with nine poleis, of which only Binchester, Catterick, Aldborough and York are certainly identifiable. The furthest place to the north with which he credits them is Epiacon, perhaps, but not verifiably, Whitley Castle near Alston; his Ouinnoouion is clearly Vinovia, Binchester, Co. Durham, hopelessly misplaced and its attribution to the Brigantes dubious (Rivet and Smith 1979: 120, 142, 360, 504f.). But his Koria, a polis of the Votadini, must be Corbridge (thus now Strange 1997: 30, Addendum, correcting his p. 21; he retains the spelling 'Curia', i.e. the variant reading in Ptolemy, Kouria-for Coria c£ above, on Haterius Nepos' letter to Genialis). Votadinian territory extended from Tyne to Forth (Rivet and Smith 1979: 508£). The Textoverdi and Corionotatae, unknown to Ptolemy-as were the Anavionenses and Carvetii-may have been small peoples in the northern

'A statuette from Birrens has been repeatedly cited as evidence that parts of the north shore of the Solway lay within the [Brigantian] tribal teritory (RIB2091). However, the statuette was dedicated by a legionary,

17

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Pennines, not necessarly subject to anyone, except to Rome.

architecti as one and the same man would make the Brigantia dedication Severan or Caracallan, whereas Birrens' most recent excavator found no evidence for third-century occupation. She treated with scepticism 'pottery identified as post-Antonine' in earlier excavations (cf. E. Birley 1937-8: 321, 326, 328 and figs. 27, 31.14, 33.1); cast doubt on the attribution to the fort of coins of Severus Alexander, Maxentius and Constantine I (c£ E. Birley 1937-8: 340, who also mentions a coin of Victorinus); and declined to accept any argument based on Birrens being the starting-point of Iter II in the Ant. It.: a Blatobulgio Castra Exploratorum m.p. xii (Robertson 1975: 286). The Ant. It., though mainly third century, is a compilation, it is true, which includes a few routes datable as early as Trajan and as late as Diocletian (Rivet and Smith 1979: 152f.). Still, given the limited nature of Robertson's excavations at Birrens (cf. Robertson 1975: fig. 2, facing p. 6, 'Trench Plan 1962-67, 1969'), it may be prudent to allow for Severan and later occupation.

Rivet comments that 'Amandus ...was an architectus and probably not himself a Brigantian, but why else should the goddess be honoured at Birrens?' (Rivet 1982: 321 n. 90). The answer is that as a military engineer Amandus had probably been seconded from the legion at York, and that he and C. Julius Apolinaris, centurion of VI Victrix at Corbridge (RIB 1131), had met the goddess at their base in the Brigantian heartland. Apolinaris coupled Brigantia, also called Caelestis, with Jupiter Dolichenus and Salus, and erected the altar 'by command of the god'-in practice no doubt by command of a priest of Dolichenus at York. At all events, Corbridge was Votadinian, not Brigantian (c£ above). Likewise, Amandus could easily have been at York, where Brigantia probably had a temple, when 'commanded' to honour the goddess (for a similar view, Mann and Breeze 1987: 89): her mural crown on the statuette suggests that she was a kind of city-goddess, indeed the tyche of the colonia-which would imply that the inscription is of the third century (cf. below). There is no need to infer that Birrens was in Brigantian territory. As parallel Mann and Breeze cite Sunuxsalis, taken to be presiding goddess of the Sunuci, a civitas to the west of the Ubii in Lower Germany, yet worshipped at Cologne, the former Ara Ubiorum, as well as in presumed Sunucan lands, the valley of the lnde and around Aachen (Mann and Breeze 1987: 89, not giving references; c£ e.g. CIL XIII 8248 and Ruger 1968: 99f.). A similar case is the dedication, between Novaesium and Gelduba, to Mercurius Arvemus (Dessau 1896-1916: no. 4591, cf. 4591-1, also in Germany), surely chief god of the Arvemi, a civitas a good 600 kms from the Lower Rhine.

Comment is required on Blatobulgium, interpreted as Celtic for 'meal-sack place' (Watson 1926: 411) or 'floursack'-a nickname based on Birrens having 'three granaries instead of the more usual two, and those of a very massive character' (K.H. Jackson in Robertson 1975: 3f.; approved by Rivet and Smith 1979: 269). But perhaps the name originally belonged to Burnswark Hill. Its profile, visible from very far off, from near Moffat in the north and from the central sector of Hadrian's Wall to the south-east, could well be said to look like a meal-sack, lying flat. Burnswark is impressive not just for its profile and the hill-fort on its summit, but for the Roman siegecamps on its north and south sides. The now orthodox interpetation is that the hill-fort was unoccupied when the siege-camps were constructed (Jobey 1977-78) and that they were for training (Davies 1972). This view is now enshrined in the volume on East Dumfriesshire (Royal Commission AHMScot 1997: 168f., 179f£). Still, Frere and St. Joseph (1983: 32ff.) were unconvinced, as was Keppie: 'Yet a case could still be made for a brief siege of the hilltop, say during the Antonine period, or even in Severan times, when recalcitrant elements had retreated to the old hillfort as a natural stronghold' (Keppie 1989: 67). Further light will surely be thrown on this question when the London Ph.D. thesis by Gwyn Davies on Roman siegeworks can be consulted (Davies 2000). In the meantime it should be recalled that George Jobey found evidence on the hill-top of 'some probably native occupation' datable to the 'late first or early second century AD'; the only coins were of Nero, Vespasian, Domitian, AD 87-9, 'not much worn', and Trajan (two). He cautiously wrote that the defences of the south camp 'appeared to post-date a Roman fortlet presumptively of late Antonine date' (Jobey 1977-8: 84f£, cf. 80)-but the provenance of the pottery evidence for the fortlet is very uncertain.

Amandus was once conjectured to be the same man as the trainee architectus Val(erius) Aman(dus), serving in the Lower German legion I Minervia in AD 208 and, it is supposed, transferred to VI Victrix, perhaps when Severus was on his way to Britain in that year. The trainee occurs on an inscription near Bonn, CIL XIII 7945, dated by [[duobus}} Aug[g}. cos., with duobus and the second g. of Augg. erased, clearly Caracalla and Geta, the latter, prematurely, as often, treated as an Augustus-there is no need to suppose the text was only put up after Geta's official elevation in late 209.2 Identification of the two The identification of the Birrens architectus Amandus with Val. Aman(dus) the trainee (ar(chitectante) ... discens) from near Bonn was proposed by Miller 1937; approved by E. Birley 1937-8: 279 and by R.P. Wright on RIB 2091. The editor, ad Zoe., dated GIL XIII 7945 to AD 210, altered by Miller, followed by Wright, to 209, although all accept that the consulate must be that of 208. 205, when Caracalla and Geta were previously consuls together, is theoretically possible, but more or less ruled out by what is known of the legate of I Minervia, C. Julius Septimius Castinus, named on the the same stone: see Alfoldy 1967: 51. Although Geta was not made Augustus until late 209, he was very often incorrectly given this rank on inscriptions long before this: numerous examples in Mastino 1981: 157f. Val. Aman(dus) need not have waited two years or more to write duobus Augg. cos. The inscription belongs to 208, surely (or even 205); and the trainee could have been taken by Severns on the expeditio Britannica, promoted to serve in VI Victrixand then ordered from York to Birrens, for rebuilding work. 2

Anne Robertson's excavations at Birrens produced a piece of evidence very relevant to Annandale. A small red

18

Anthony Birley: The Anavionenses

sandstone slab (35 by 52 ems) is incised with a figure interpeted as a dog or a serpent; above it is a crude inscription read by Richard Wright as CIST VMV CI LO MAbOMI, with VMV ligatured. He understood Cistumuci lo(co) Mabomi, '(Gift of) Cisturnucus from the Place of Mabomus', with Mabomi a variant for Maponi-i.e. Maponus, a known deity (Wright 1968: 209, no. 28 and pl. XIX. 2). This seemed to explain the curious passage in the British section of Ravennas, 108: sunt autem in ipsa Britannia diversa loca, ex quibus aliquanta nominare volumus, id est Maponi, followed by seven further names, 'there are moreover various places in Britain itself, of which we wish to name some, of Maponus, etc.'. Some thought these loca were 'tribal meeting-places'. Rivet and Smith, accepting Wright's reading, argued that they were 'just a collection of "odd places" ...which the Cosmographer had omitted to mention earlier. He may have been misled himself by *Locus Maponi, taking Locus to be a common noun'. Rather, 'it is in principle safer to consider that a British word is involved. This can only be *Zoe- "lake, pool" '. They then emphatically identified Locus Maponi as Lochmaben, 15 km west of Birrens, rather than, as has also been argued, with the 'Clochmabenstane' megalith, 2.3 m. high, close to the Solway shore, just south of Gretna, 12 km from Birrens. As they point out, cloch simply means 'stone' in Gaelic; 'stane' duplicates this (Rivet and Smith 1979: 212, 395). The 'Maben stone', presumed survivor of an original circle (Crone 1983; Royal Commission AHMScot 1997: 54, cf. 110, 112) was the traditional meeting-place of the Western March in the middle ages.

drastically ligatured. Whoever the dedicator was, anonymous on this interpetation, may just have intended to present a 'cistula to the god Maponus'. Keppie and Arnold (1984: 8£, no. 14 and pl. 5), noting Robertson's reading, note that the animal has a collar, and that 'in Welsh vernacular tradition dogs are associated with Maponus (or Mabon) in his capacity as a hunting god'-they also mention the 'intriguing possibility that the slab was a makeshift gravestone of the dog itself. Maponus is otherwise known by three altars from Corbridge (RIB 1120-1122) and one from Ribchester (583), all set up by Roman officers to Apollo Maponus; by one from near Brampton (2063), dedicated by four Germani to deo Mapono et n(umini) Aug(usti); and by a silver lunula from Vindolanda vicus, inscribed deo Mapono (Wright and Hassan 1971: 291, no. 12). This evidence has often been discussed (see E. Birley 1952-3; 1986: 55-58). Although the British cult probably emanated from a sacred centre, say at Lochmaben, Maponus is wellknown in Celtic literature as the youth Mabon, son of Modron (McQueen 1952-3; cf. E. Birley 1986: 55 n. 286a, a Celtic inscription from France naming Maponos). One may take it that Maponus was worshipped-as their chief god?-by the Anavionenses; and that Lochmaben was, if not their chef-lieu, at least a major centre of this people. Whether there was a fortified native settlement of the Roman period underlying the Brus stronghold at Castle Loch captured by Edward I in 1298 (Royal Commission AHM Scot. 1997: 197f£) is unfortunately unknown; and the impressive defences of Woody Castle, just north of Lochmaben, are completely undated (ibid.: 136f.).

Ralegh Radford long ago argued for Lochmaben as Locus Maponi and identified as traces of the 'temenos of the god' earthworks associated with the peel of Lochmaben above Castle Loch (Radford 1952-3; not discussed in Royal Commission AHMScot. 1997: 188, 203f£ On the nature of the loca, note that the interpretation dismissed by Rivet and Smith is re-stated by Mann 1992). Another suggestion may be added. Ravennas includes a second place, evidently in Scotland, that might evoke the god, Maporiton. Rivet and Smith (1979: 312) note that '[t]he name is built on British *mapo- 'boy, youth' ... and *ritu'ford' .' In spite of their caution (they suspect corruption from the preceding name, Tadoriton: ibid. 462) about associating Maporiton with Maponus, may not the place where the 'Maben stone' stands have been 'Maponus' ford'-across the Solway?

Much more could be said about Annandale during the Iron Age and the Roman period. Here it may just be noted that the Romans had already penetrated into this region in the early 70s. Doubts about pre-Agricolan activity have been removed by the dendrochronological dating of Carlisle to AD 72. It was Petillius Cerialis who initiated the invasion of Scotland (Caruana 1997; Shorter 2000; among other new evidence note the fort at Ladyward, between Lockerbie and Lochmaben, at the junction of the Dryfe Water with the Annan, unexcavated and hence impossible to date exactly, Frere 1990: 312f.). This means, for one thing, that the multae civitates who, having previously held their own (ex aequo egerant) against Rome, responded to Agricola's firm hand in his second campaign (AD 78) by 'putting aside their anger and handing over hostages' (Tacitus, Agr. 20.3) were not 'members of the Brigantian confederation'. 3 These multae civitates must be the-

To return to 'lo(co) Mabomi': scepticism is required about this crudely cut inscription. Anne Robertson offered 'an alternative reading ... CISTVM (or CISTAM) DIO MAPONI' (Robertson 1975: 95£, with fig. 25, no. 3 and pl. 10; approved by E. Birley 1986: 58 n. 292). The lettering is akin to cursive, not least the supposed B, a typical cursive form of P. The name should be read as MAPONI, even perhaps MAPONO, if a final O was botched. On what preceded, DIO is plausible-and DEO may have been intended. As for the remainder, one could perhaps read CISTVLAM, 'little box', with VLAM

The quotation is from Syme 1936: 153-many others have taken the same view on multae civitates; against, Birley 1973: 190, also criticising the interpretation of ex aequo egerant by Ogilvie and Richmond 1967: 219. On the endlessly discussed passage in Pausanias, 8.43 ('[Antoninus Pius] confiscated a large part of the territory of the Brigantes in Britain because they too began to invade with arms the Genunian region, the inhabitants of which were subject to Rome'), I follow Rivet and Smith 1979: 47, who convincingly argue that 'the text has been tampered with.' There was probably originally a reference a) to Lollius Urbicus' campaign in Britain and b) to trouble caused by the Brigantii of Raetia to their neighbours the Genauni (cf. e.g. Strabo 4.6.8, C 206, for these two 3

19

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

interpretation more plausibly puts Ouxellon at Rubers Law, 20 km S. of Trimontium, 8 km west of Dere Street (Strange 1997: 29). That leaves a big gap between Novantae and Selgovae. Why not? Ptolemy not only made an odd mistake by the 'turning of Scotland': as the discovery of the c(ivitas) Car(vetiorum) showed, his data was limited, or he was very selective. He missed out Anavionenses, Carvetii and a good few others.

many (surely more than two}-peoples of S. Scotland. This passage from Tacitus has been combined with Ptolemy's statement about the Brigantes 'stretching from sea to sea', with nine poleis, and Tacitus' earlier remark that the Brigantes were 'held to be the most numerous civitas in all Britain' (Agr. 17.1). The resultant interpretation has unjustifiably contributed to the notion of 'Greater Brigantia' which is here called into question. At all events, the Anavionenses-and their immediate neighbours-had had separate dealings with Rome for nearly three decades before Haterius N epos arrived, even on the earlier dating of his census.

As Haterius Nepos' operations can now be dated c. AD 100, this removes one complication, since Rome's abandonment of southern Scotland is dated to c. AD 105 (Hartley 1972: 15). Rivet was even prepared, because of the Birrens statuette, to extend Brigantian territory as far as the west side of Nithsdale, which he suggested as the dividing line between Brigantes and Selgovae;5 but, he went on, when what is now Scotland was evacuated, 'before the physical wall was built Brigantia was treated as a whole and ... all of it was kept under control-and taxed' (Rivet 1982: 320£). The new dating shows, it seems, that the censitor was, after all, in an area still directly controlled by Roman forts. On the other hand, with or without the 'Greater Brigantia' theory, it is hard to believe that he assessed Anavionenses for tax purposes. Censors and censitores counted heads in the first instance. The census of the Anavionenses was surely to find out how many young men, not taxes in money, they could produce. One may compare Tacitus on the Batavians, who supplied men for Rome's armies, instead of tribute, a privilege in their case (Germania 29); they could well have paid up in coin; not likely for the people of Annandale. They were to contribute soldiers-for numeri Brittonum in Germany.

One other matter needs mention. In their valuable monograph on the Carvetii, Nick Higham and Barri Jones hedge their bets. At first, they accept that they were one of the 'sub-tribes or septs that made up the Brigantian confederacy' and hence make Carvetian territory extend around the head of the Solway Firth into the Birrens area (Higham and Jones 1985: 9ff.). Later, referring to Haterius Nepos' operations, assigned to c. AD 112, they note that: 'Annandale's population appears to have undergone a census, implying incorporation within the Roman system, whether as an outlying independent tribe or not' (25-my italics). However, later still the Anavionenses are again 'a sept' (7 5)-a sept of a sept, it might be comented. 4 Ptolemy notoriously got the geography of Scotland wrong in a major respect by turning North Britain to the east; the northernmost part of England was also distorted (there is an extensive literature: see now Strange 1997). He gives three peoples for what must be Galloway, Borders and Lothian (plus E. Northumberland), each with several poleis: two for the Novantae, four for the Selgovae and three for the Votadini. Only Trimontion (Eildon Hills) of the Selgovae and Bremenion (High Rochester on Dere Street) of the Votadini are definitely identifiable (Rivet and Smith 1979: 139£, 276f., 475). Ouxellon of the Selgovae has been placed at Ward Law, near Caerlaverock at the mouth of the R. Nith (Rivet and Smith 1979: 483f.). This would stretch the Selgovae a long way, giving them Annandale as well as Upper Tweeddale. The latest

Before being sent across the North Sea, the Brittones needed basic training. The Anavionenses-and no doubt drafts from other north British peoples-are surely among the 'naked(?) Britons', with 'very many cavalrymen', who 'do not use swords', the 'Brittunculi', a previously umecorded and certainly derogatory term, who 'do not stay in the saddle to throw javelins', revealed by a Vindolanda writing-tablet of period III, c. AD 98-104/5 (TV II no. 164). They were being trained by the Batavians, whose officer was not impressed with them. The Anavionenses may now be identified at Vindolanda, in another text from period III (Inv. 93/1475A), a mass of fragments from an account perhaps recording distribution of rations. It includes many individual names, e.g. Vatt[o} Trev[er], Verecun[dus}, Sim[p}lex, Crescens and Senecio; Genia[lis} also occurs, not necessarily the prefect Genialis. There is also the entry Anavion[---}. Of course, this could be a man called e.g. Anavion[us] (a name not otherwise attested). But it could well denote the Anavion[enses}. Another name in this document might conceivably be of a further-otherwise unknown-British people: Segosi (occurring at least twice); there may even be a mention of the Celtic hunting-dogs newly prized by the Romans at this

peoples; further references in Hind 1977: 232). Two separate items were somehow abbreviated and conflated in the MS. The trouble in the 'Genaunian region' almost certainly had nothing to do with Britain. Alternatively, as Hind 1977 suggests, in a full and fair discussion, perhaps Pausanias knew that the Brigantes had attacked one oftheir-southernneighbours, well within the province, but muddled them with the Raetian Brigantii, and named one of their neighbours instead of e.g. the Cornovii. 4 Higham and Jones 1985 seem to have misled Whittaker a little, in his influential monograph on Roman frontiers. One map (Whittaker 1994: 42) shows the Carvetii stretching across the Solway Firth. Later he comments (referring to Jones 1982: 294f. and Higham and Jones 1985: 74-5): 'The stakes that were set up by either Trajan or Hadrian along the Solway River in Carvetii territory were obviously not there to divide anyone, since an inscription tells us that the Carvetii were simultaneously being assessed for a tax census north of the Solway', Whittaker 1994: 82-but on p. 95 'the Anavionenses of North Solway were given a census in the Trajanic period'. Also unfortunate is his idea (82; it is shared by Hind 1977: 232 n. 13, it seems), that 'the [Carvetian] center at Car-lisle ... reflects their name'; Car- is surely just from Celtic Caer-; cf. Caerleon, Carmarthen, Caerwent, etc.

Rivet 1982: 320f. does not refer to the identification ofUxellum of the Selgovae as Ward Law put forward in Rivet and Smith 1979: 483f., where at 455 the Selgovae are only given 'the Upper Tweed basin'. 5

20

Anthony Birley: The Anavionenses

time, vertragi (c£ Arrian, Cynegetica 2£). The Segosi may, like the vertragi, also be hunting-dogs, a breed normally spelled segusi. Had the Anavionenses, worshippers of Maponus the huntsman (and note the suggestion about the animal on Birrens Maponus-stone ), brought some hounds with them? (The Vindolanda prefect Flavius Cerialis was a passionate devotee of the chase, c£ TVII 233; several unpublished tablets further document his hunting activity). That a lunula of this god was found in a third-century context at Vindolanda may be called just coincidence.

12499), L. (XIII 12500). The standard view (e.g. Gillam 1984: 287£; Southern 1989: 133-4) is that the extra names derived from the places and rivers where the men were stationed: Elantienses from the R. Elz, close to N eckarburken where they are attested, Gr., expanded to Gr(inarionenses) from Grinario, the ancient name for Kongen, L. from the assumed name of the R. Lein, hence L(unenses). It is surely just as possible that the Britons brought these names with them-and thus that the n(umerus) Br(i)t(tonum) Cal on a tile-stamp from Ohringen were Brittones Cal(edonii). That idea has been dismissed in the past (c£ Southern 1989: 133 for the standard view), but recently revived (Kerneis 1999: esp. 378f£). It deserves consideration-and one may also ask whether the n.Brittonum A. at Niederbieber might have been Brittones A(navionenses) (Southern 1989: 132 does not discuss the A.). The Gurvedenses, for that matter, might have been from the Water of Girvan (ancient name unknown, but cf. Ptolemy's Vindogara, Watson 1926: 32; Rivet and Smith 1979: 501f.); and some if not all the other names could likewise come from N. Britain. So many ancient names of rivers, places-and peoples-in Britain are still unknown. 6

On the face of it, the Anavionenses left no direct trace in Germany. Over three dozen inscriptions, mainly from the Upper German limes, attest numeri Brittonum or Brittones; a few are dated-the earliest with an exact year of AD 146, CIL XIII 6511. It was long supposed that these were North Britons deported after the reconquest of southern Scotland in AD 139-142 (because of Historia Augusta, Antoninus Pius 5.4: Britannos per Lollium Urbicum vicit legatum alio muro cespiticio summotis barbaris ducto, 'he conquered the Britons through his legate Lollius Urbicus, another wall, of turf, being erected, the barbarians having been removed' -to Germany, so it was inferred). Then D. Baatz' excavations at Hesselbach, a known Numeruskastell of the Brittones, produced a different picture (Baatz 1973). His conclusions were spelled out by Sheppard Frere: 'the evidence suggests that the Brittones had been [at Hesselbach] from the inception of Period 2 ([AD] 115-30). But the plan of Period I is in all essentials identical, and Baatz shows that there is strong support for the view that the British unit was present then too (i.e. from [AD] 95100)' (Frere 1974: 495). The question was discussed further by John Gillam: building on Baatz' results, and citing his own earlier work and George Jobey's research on 'native' sites (Jobey 1974), he convincingly dismissed the old idea (going back to E. Fabricius, G. Macdonald and R.G. Collingwood), that whole populations were deported from southern Scotland in the Antonine period. There is no evidence of depopulation there in this period; and, as he pointed out, the total numbers serving in each unit in Germany was relatively small (Gillam 1984: 288ff.; his summary of the German evidence, 287f. and 294, is incomplete). Baatz' chronology was modified in detail by Barbara Pferdefirt, who brought down the beginning of Hesselbach's period I to at earliest c. AD 100-1 (Pferdehirt 1986: 279; 308-312); and the position was re-stated in Pat Southern's valuable survey of all known numeri (Southern 1989: esp. 94-8, 132-4). She argues, against Gillam, that the numeri Brittonum were first raised under Trajan rather than by Agricola, and, following Baatz, that further drafts were perhaps supplied under Hadrian (Southern 1989: 96).

Not many British civitates took their names from rivers; c£ above on the Novantae-not from the R. Novius. One can at least note the Setantii, from the Seteia, which is perhaps the Mersey (Rivet and Smith 1979: 356-7; Higham 1993: 31, a little hesitantly). These 'Merseysiders' are of course regarded as a 'Brigantian sept', but geography and later history render this less than compelling. All the same, there are examples in other Celtic areas: notably the Sequani from the R. Seine; and, with the prefix Ambi-, denoting 'both sides of the valley', e.g. Ambarri from the R. Arar; Ambidravi from the R. Drave (or Drau); Ambisavi from the R. Sava; Ambisontes from the R. Salzach (Isonta). The Anavionenses take their place in this company. Their territory was perhaps not all that large-compared, for example, with the Gallic Remi, who also underwent a census conducted by an (ex?-)equestrian officer in the Trajanic (or perhaps Flavian) period (D. Julius Capito, tribune of II Adiutrix and then censor civitatis Remor(um) foederatae, Pflaum 1960-1: 175f., no. 79). There is no need to assume that they did not have a separate identity. The evacuation of southern Scotland, c. AD 105, was probably made necessary by troop withdrawals resulting from Trajan's Second Dacian War, which broke out in June that year. The Ninth Batavians left Vindolanda about then (c£ above), and can be identified in the area of that

Inscriptions and tile-stamps show that very often these British units in Germany used further names, sometimes very abbreviated: e.g. Brit. Mur., evidently Murrenses (CIL XIII 6471, c£ 6454), Elantienses (XIII 6490, 6498; AE 1986. 523), Triputienses (XIII 6502, 6511, 6514, 65176519, 6606, c£ 6599), Gurvedens[es} (XIII 7343), A. (XIII 7749, 7752+add., p. 132); Cal. (XIII 12498), Gr. (XIII

Kemeis has some other, remarkably adventurous suggestions, e.g. deriving Murrenses from Muridunum (Carmarthen) and Lunenses from the R. 'Lynn [sic] en Caledonie'-and discovers the Novantae among the (former) inhabitants of the Agri Decumates by emending the Novarii in the Laterculus Veronensis 253, XV, Seeck (where some names are admittedly rather corrupt). Note also that this scholar, rejecting the arguments of Gillam 1983, Jobey 1974 and Southern 1989, seeks to reinstate the old theory of mass deportations to Germany after the campaigns ofLollius Urbicus (Kemeis 1999: 376f.). 6

21

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

the Brittunculi (cf Jones 1997 and he had planned a book about them). I was confident that he would greatly improve the finished product; and we had talked of jointly exploring South West Scotland when we were both OAPs. Much of what is here presented has already been aired briefly (Birley 1997A: 276f.; 1997B: 135f£; 1998: 302ff.; 1999: 45, 47). The fuller version may not necessarily convince any more than the earlier sketches: it must be conceded that the Anavion[enses} restored in Vindolanda Inv. 93/1475A-not to mention their hounds-are rather speculative. Whether the editors of TV III are prepared to accept this interpretation remains to be seen, but after seeing the grim reaper at work these past few years I was reluctant to wait any longer before giving some unpublished tablets an advance airing. At least, I hope that this discussion of the Brigantes and their neighbours, especially the Anavionenses, may arouse some interest and generate discussion.

war at the right time. 7 Other troops were probably sent overseas from Britain too-including numeri Brittonumto fill gaps in Germany after several units from there were sent to the Danube front. In particular, the legion XI Claudia at this time left Vindonissa in Germania Superior, where Trajan strengthened the frontier in the Odenwald, as elsewhere (Pferdehirt 1986: 30ff.). Removal of troops from the northern frontier zone in Britain may well have caused new unrest, not least among the Anavionenses and their neighbours, no longer under direct control. The effective frontier now, rather than the Solway-Tyne or 'Stanegate' line, may be thought to be a Stanegate-Dere Street system, including the Votadini in the defended area, but excluding Selgovae and Anavionenses (not unlike the line later dividing England and Scotland-or for that matter the frontier after Hadrian, with the outpost forts of the Wall's Varland). Disturbances beyond this (if not in the Brigantian Pennines as well) could well have lasted for over a decade, exacerbated by news filtering back of maltreatment of British conscripts. The Brittunculi writingtablet is a first sign of a contemptuous Roman (or Batavian) attitude. The draft protest letter, perhaps intended to be handed to Hadrian (the addressee is called 'Your Majesty') on his visit in AD 122, written by a man who had been severely beaten by centurions, shows a further development: not only was he innocent, he sayshe was 'from overseas', a tra(n)smarinus. By implication 'natives', nasty little Britons, could be flogged, and no doubt were (TV II 343, with the interpretation in Birley 1997A: 276f.; 1997B: 135£). This may have helped to provoke the serious revolt in Britain that broke out or was already under way at Hadrian's accession. One probable casualty can now be identified, the centurion T. Ann[ ], perhaps acting-commander of the First Cohort of Tungrians, 'killed in the war' and buried at Vindolanda (Birley 1998). It is to be hoped that more information may come to light about this war. Had the Tungrians been sent into Annandale? Had the Anavionenses re-occupied Burnswark? It would be nice one day to find out about a siege-say by Pompeius Falco----of Burnswark. This must probably remain in the realm of historical fiction.

BIBLIOGRAPHY

Alfoldy, G., 1967. Die Legionslegaten der romischen Rheinarmeen. Epigraphische Studien 3, Beihefte der Bonner Jahrbucher, Band 22. Cologne and Graz: Bohlau Verlag. Baatz, D., 1978. Kastell Hesselbach und andere Forschungen am Odenwald-Limes. Limesforschungen, Band 12. Berlin: Gebriider Mann Verlag. Bennett, J., 1997. Trajan: Optimus Princeps. A Life and Times. London and New York: Routledge. Bidwell, P. (ed.), 1999. Hadrian's Wall 1989-1999. A summary of Recent Excavations and Research prepared for The Twelfth Pilgrimage of Hadrian 's Wall, 14-21 August 1999. Carlisle: Cumberland and Westmorland Antiquarian and Archaeological Society and the Society of Antiquaries of Newcastle upon Tyne. Birley, AR., 1973. Petillius Cerialis and the conquest of Brigantia. Britannia 4: 179-190. Birley, AR., 1980. Review of Ramm 1978. Northern History 16: 254-5. Birley, AR., 1981.The Fasti of Roman Britain. Oxford: Clarendon Press. Birley, AR., 1997A Supplying the Batavians at Vindolanda. In Groenman-van W aateringe et al. (eds.): 273-280 Birley, AR., 1997B. Hadrian, The Restless Emperor. London: Routledge. Birley, AR., 1998. A new tombstone from Vindolanda. Britannia 29: 299-306. Birley, AR., 1999. The Vindolanda writing-tablets. In Bidwell (ed.): 37-48. Birley, AR. forthcoming. A band of brothers. Equestrian officers in the Vindolanda tablets.In Electrum 5, ed. E. Dabrowa, Crac6w. Birley, E., 1937-8. Excavations at Birrens, 1936-1937. Proceedings of the Society of Antiquaries of Scotland 72: 275-347.

This tribute to Barri Jones, whom I was lucky to know as a close, generous and deeply valued friend for almost forty years, was originally offered-to provoke debate-for the conference to mark his retirement. Barri was also for sixteen years my closest colleague at Manchester and for over twenty-five years a loyal and active fellow-Trustee at Vindolanda, where he is deeply missed. He was passionately interested in both sides of the Solway and in 7 The Ninth Batavians have been identified by two tile-stamps found in Rumania, at and near the fort of Buridava on the R. Olt, which guarded the Red Tower Pass from Wallachia into Transylvania. The evidence is fully discussed by Dietz 1982, who assumed, understandably (not yet aware of the Vindolanda evidence), that they were sent there from Raetia (where they were later based); this point was corrected by Strobel 1987: 275f.-whose own dating of the Ninth Batavians' move from Vindolanda to the Dacian front likewise requires modification, as it was based on the assumption that they were sent to the Danube area for the First Dacian War, AD 101-2. Buridava was part of the enlarged Lower Moesia from AD 102-118, cf. Bennett 1997: 95f.

22

Anthony Birley: The Anavionenses

Hanson, W.S. and Keppie, L.J.F. (eds.), 1980. Roman Frontier Studies 1979. Papers presented to the 12th International Congress of Roman Frontier Studies. 3 vols., Oxford: British Archaeological Reports, International Series 71. Hartley, B.R., 1972. The Roman occupation of Scotland; the evidence ofsarnian ware. Britannia 3: 1-55. Hartley, B.R. and Fitts, L., 1988. The Brigantes. Gloucester: Alan Sutton. Higham, N.J. and Jones, G.D.B., 1985. The Carvetii. Gloucester: Alan Sutton. Higham, N.J. 1986. The Northern Counties to ADJ000. Harlow: Longman. Higham, N.J. 1993. The origins of Cheshire. Manchester: Manchester University Press. Hind, J.G.F., 1977. The 'Genounian' part of Britain. Britannia 8: 229-234. Jobey, G., 1974. Notes on some population problems in the area between the two Roman walls. Archaeologia Aeliana, fifth series, 52: 17-26. Jobey, G., 1977-8. Burnswark Hill, Dumfriesshire. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, 3rd series, 53: 57-104. Jones, G.D.B., 1982. The Solway frontier. Britannia 13: 282-297. Jones, G.D.B., 1997. From Brittunculi to Wounded Knee: A study in the development of ideas. In Mattingly (ed.) 1997: 185-200. Jones, (G.D.)B. and Mattingly, D., 1990. An Atlas of Roman Britain. Oxford: Basil Blackwell Ltd. Keppie, L.J.F., 1989. Beyond the northern frontier: Roman and native in Scotland. In Todd, (ed.): 61-73. Keppie, L.J.F. and Arnold, B.J., 1984. Corpus Signorum Imperii Romani. Corpus of Sculpture of the Roman World. Great Britain. Volume I, Fascicle 4. Scotland. Oxford: Oxford University Press for the British Academy. Kerneis, S., 1999. La Bretagne rhenane. Note sur les etablissements bretons dans les Champs Decumates. Latomus 58: 357-390. MacQueen, J., 1952-3. Maponus in mediaeval tradition. Transactions of the Dunifriesshire and Galloway Natural History and Antiquarian Society, 3rd series, 31: 43-57. Mann, J.C., 1992. Loca. Archeologia Aeliana, 5th series, 20: 214-216. Mann, J.C. and Breeze, D.J., 1987. Ptolemy, Tacitus and the tribes of North Britain. Proceedings of the Society of Antiquaries of Scotland 117: 85-91. Mastino, A., 1981. Le titolature di Caracalla e Geta attraverso le iscrizioni (indici). Studi di storia antica 5. Bologna: Cooperativa Libraria Universitaria Editrice Bologna. Mattingly, D. (ed.), 1997. Dialogues in roman Imperialism: Power, Discourse and Discrepant Experience in the Roman Empire. Portsmouth: Journal of Roman Archaeology Supplementary Volume 23. Maxwell, G.S., 1980. The native background to the Roman occupation of Scotland. In Hanson and Keppie: 1-13.

Birley, E., 1952-3. Maponus: the epigraphic evidence. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, 3rd series, 31: 39-42. Birley, E. 1953. Roman Britain and the Roman Army. Collected Papers. Kendal: Titus Wilson & Son, Ltd. Birley, E. 1961. Research on Hadrian's Wall. Kendal: Titus Wilson & Son, Ltd. Birley, E., 1986. The deities of Roman Britain. In Temporini and Haase (eds.): 3-112. Birley, R., 1994. Vindolanda Research Reports, New Series, Vol. I. The Early Wooden Forts. Introduction and Analysis of the Structures. Bardon Mill: the Vindolanda Trust. Bowman, A.K. and Thomas, J.D., 1994. The Vindolanda Writing-Tablets (Tabulae indolandenses II). London: British Museum Press. Bowman, A.K. and Thomas, J.D., 1996. New writing tablets from Vindolanda. Britannia 27: 299-328. Breeze, D.J. and Dobson, B., 2000. Hadrian's Wall, 4th edition. London: Penguin Books. Caruana, I., 1997. Maryport and the Flavian conquest of North Britain. In Wilson (ed.): 40-51. Crone, A., 1983. The Clochmabenstane, Gretna. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, 3rd series, 58: 16-20. Davies, G., 2000. The Archaeology of Roman Siege Warfare. Unpublished Ph.D. thesis, London: Institute of Archaeology. Davies, R.W., 1972. The Romans at Burnswark. Historia 21: 99-113. Dessau, H., 1892-1916. Inscriptiones Latinae Selectae. 3 vols., Berlin: Weidmannsche Verlagsbuchhandlung. Dietz, K., 1982. Ein Beitrag Riitiens zum zweiten Dakerkrieg Trajans. Nochmals zum Militiirdiplom aus Oberstimm. Germania 60: 183-191. Eck, W., 1999. The Bar Kokhba revolt: the Roman point of view. Journal of Roman Studies 89: 76-89. Frere, S.S., 1967. Britannia. London: Routledge & Kegan Paul. Frere, S.S., 1974. Review of Baatz 1973. Britannia 5: 4946. Frere, S.S., 1990. Roman Britain in 1989. I. Sites explored. Britannia 21: 304-364. Frere, S.S., 1991. Britannia, paperback version of 3rd edition, London: Pimlico. Frere, S.S. and St Joseph, J.K.S. 1983. Roman Britain from the Air. Cambridge: Cambridge University Press. Gillam, J.P., 1984. A note on the numeri Brittonum. In Miket and Burgess (eds.): 287-294. Groenman-van Waateringe, W., van Beek, B.L., Willems, W.J.H. and Wynia, S.L., 1997. Roman Frontier Studies 1995. Proceedings of the XVIth International Congress of Roman Frontier Studies. Oxford: Oxbow Monograph 91. Hanson, W.S., 1987. Agricola and the Conquest of the North. London: B.T. Batsford Ltd. Hanson, W.S. and Campbell, D.B., 1986. The Brigantes: from clientage to conquest. Britannia 17: 73-89.

23

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Miket, R. and Burgess, C. (eds.), 1984. Between and Beyond the Walls. Essays on the Prehistory and History of North Britain in Honour of George Jobey. Edinburgh: John Donald Publishers Ltd. Miller, S.N., 1937. Note on an inscription from Birrens. Journal of Roman Studies 27: 208-9. Ogilvie, R.M. and Richmond, I.A., 1967. Cornelii Taciti de vita Agricolae. Oxford: Clarendon Press. Pferdehirt, B., 1986. Die romische Okkupation Germaniens und Ratiens von der Zeit des Tiberius bis zum Tode Trajans. Untersuchungen zur Chronologie siidgallischer Reliefsigillata. Jahrbuch des RomischGermanischen Zentralmuseums 23: 221-320. Pflaum, H.-G., 1960-1. Les carrieres procuratoriennes equestres sous le Haut-Empire romain, 4 vols. Paris: Librairie Orientaliste Paul Geuthner. Radford, C.A.R., 1952-3. Locus Maponi. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, 3rd series, 31: 35-8. Ramm, H., 1978. The Parisi. London: Duckworth. Rivet, A.L.F., 1969. Review and discussion of Frere 1967. Journal of Roman Studies 59: 247-250. Rivet, A.L.F., 1982. The Brittones Anavionenses. Britannia 13: 321-2. Rivet, A.L.F. and Smith, C.C., 1979. The Place-Names of Roman Britain. London: B.T. Batsford Ltd. Robertson, AS. 1975. Birrens (Blatobulgium). Edinburgh: T. & A. Constable Ltd. for the Dumfriesshire and Galloway Natural History and Antiquarian Society. Royal Commission on the Ancient and Historical Monuments of Scotland, 1997. Eastern Dumfriesshire: an archaeological landscape. Edinburgh: The Stationery Office. Ruger, C.B., 1968. Germania Inferior. Untersuchungen zur Territorialund Verwaltungsgeschichte Niedergermaniens in der Prinzipatszeit. Beihefte der Bonner Jahrbucher, Band 30, Cologne and Graz: Bohlau Verlag. Shorter, D.C.A., 2000. Petillius Cerialis in Northern Britain, Northern History 36: 189-198. Southern, P., 1989. The numeri of the Roman imperial army. Britannia 20: 81-140. Strange, A., 1997. Explaining Ptolemy's Roman Britain. Britannia 28: 1-30. Strobel, K., 1987. Anmerkungen zur Geschichte der Bataverkohorten. Zeitschrift far Papyrologie und Epigraphik 70: 271-292. Syme, R., 1936. Flavian wars and frontiers. Cambridge Ancient History 11, Cambridge: Cambridge University Press: 131-187. Temporini, H. and Haase, W., (eds.), 1986. Aufetieg und Niedergang der romischen Welt. II. Prinzipat. 18.1. Berlin and New York: Walter de Gruyter. Todd, M., 1981. Roman Britain 55 BC-AD 400. Brighton: Harvester Press. Todd, M., (ed.), 1989. Research on Roman Britain: 19601989. London: Britannia Monograph Series no. 11. TV II see Bowman and Thomas 1994

Watson, W.J., 1926. The Celtic Placenames of Scotland, being the Rhind lectures on archaeology (expanded) delivered in I 9 I 6. Edinburgh and London: under the auspices of the Royal Celtic Society, William Blackwood & Sons Ltd. Whittaker, C.R., 1994. Frontiers of the Roman Empire. A Social and Economic Study. Baltimore and London: The Johns Hopkins University Press. Wilson, R.J.A. (ed.), 1997. Roman Maryport and its Setting. Essays in Memory of Michael G.Jarrett. Maryport: Cumberland and Westmorland Antiquarian and Archaeological Society, Extra Series, Vol.28, on behalf of the Trustees of the Senhouse Museum. Wright, R.P., 1968. Roman Britain in 1967. II. Inscriptions. Journal of Roman Studies 58: 206-214. Wright, R.P. and Hassan, M.W.C., 1971. Roman Britain in 1970. II. Inscriptions. Britannia 2: 289-304.

24

2

Reginald Bainbrigg, scholemaister, and his stones Ben Edwards It was one of Barri Jones's characteristics that, whenever

corresponded with Camden, and many of his contributions survive in the Cottonian MS. at the British Library (hence Haverfield's paper) and he seems to have travelled along Hadrian's Wall twice; once in the same year (1599) as that of the visit by Camden and Sir Robert Cotton, when the latter pair famously avoided the central sector because of its lawlessness; once two years later. It does not seem that Bainbrigg and Camden met, or at least no evidence has been adduced to show that they did. That Bainbrigg's opinion was one to be valued on the subject of Roman inscriptions is shown by the inclusion in one of the letters in BM Cotton Julius FVI of the following in relation to an inscription from Ribchester (RIB 595): 'I would here how mr banebricke doth expo[u ]nd it. mr kampden hath had it all rede' (£301 olim 286).

someone in the field of archaeology who was new to him came to his notice, he wanted to know more about that person. Thus, at a meeting or conference, the appearance of an 'unknown' as a speaker would always elicit the question 'Who is this?'. And so it should be; we should all be as aware as possible about the people on whose work our own is partly based - a principle no less valid in the case of those long since dead than of those in front of our eyes. It is easy to read the names of these past workers and to ignore their persons and antecedents, but every now and then the study leads us down interesting and littletrodden paths. Reginald Bainbrigg (generally so spelt, though the name is the same as Bainbridge; contemporary versions such as Banbrick are common) was a case in point for the writer, and it was in the course of work on the life and activities of William Camden (Edwards 1998) that his name came to notice. Francis Haverfield, though he used Bainbrigg as a Cumbrian peg on which to hang a Camden paper in the Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society (Haverfield 1911), said little enough about him.

These activities are of interest and have been documented; another activity related to them has been alluded to, but has been less studied. This was Bainbrigg's habit of collecting inscribed stones. Of these, some were actual Roman inscriptions which he was able to acquire; others were copies of Roman inscriptions, in the sense that Bainbrigg's reading of the Roman inscription was chiselled onto stone; and, in addition, he was much addicted to erecting contemporary inscriptions, sometimes as labels to his Roman (genuine or otherwise) inscriptions and usually in Latin. Some of each of these classes of inscription have survived, and it is the purpose of this paper to examine them in order to see what light they throw on Roman Britain and, incidentally, on Reginald Bainbrigg. This work was begun, as has been said, in relation to a study of William Camden, and the mention of the stones in that study led to Mr. W. D. Shannon's working independently on Bainbrigg's stones, and it must be said at once that most of the conclusions here described are as much the result of his work, which he has generously passed to me, as of mine, though responsibility for the final interpretation is mme.

By profession Bainbrigg was, like Camden, a schoolmaster, and both men held the post of headmaster. However, while Camden's time in this post was spent at the metropolitan Westminster School, Bainbrigg's was spent in the remoteness of Appleby Grammar School in Westmorland. He was a local man, having been born in the nearby village of Hilton, in or about 1545. He is assumed to have been a pupil at the school he later headed, but he did not go up to Oxford until the very late age, for the sixteenth century, of 27. We do not know the reasons for this delay of ten years or more from the normal, but he was a member, like so many Westmorland men, of Queen's College, whose statutes gave preference to men from that county. He graduated B.A. in 1576, proceeding to M.A. in 1579. The following year he was licensed to teach in the grammar school at Appleby, as its second headmaster under the Charter granted by Queen Elizabeth in 1574. Still living in 1612, he probably died in 1613, in which year probate of his will, drawn up in 1606, was granted (for a summary of his life see Hinchcliffe 1974a: 30-35).

Before turning to look at the stones which survive today at Appleby, it will be useful to consider the evidence for their existence in the past. The printed sources are four in number - Camden, in Britannia (various editions); John Horsley in Britannia Romana (1732); Richard Gough in British Topography (1780); and RIB (Collingwood and Wright 1965). In addition there is a very important unpublished source, namely a list made by I. Hayton. He, in 1722, made a record of the stones, which is inserted into Richard Gough's copy of Britannia Romana, now in the Bodleian Library, Oxford (Gough Gen.Top. 128, f£4344r.).

That he was a scholar by inclination seems certain. It is said that his Latin was flowing and his Greek passable and that he had a working knowledge of Hebrew. He amassed, and left to the school at his death, a fine library (Hinchcliffe 1974a: 131-136; 1974b; 1996) which included many medical works in addition to more likely ones. But his importance for our purpose is that he was interested in antiquities, and particularly in Roman inscriptions. He

Camden received information from Reginald Bainbrigg, and some of this is preserved in BL MSS. Cotton Julius F.VI and F.X, made accessible in print by Haverfield 25

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

(1911). Camden did not quote any figures in referring to Bainbrigg's collection. In Philemon Holland's translation (Camden 1637, 761) he says merely that Bainbrigg 'hath exemplified [i.e. copied] for mee many antique inscriptions, and brought some hither to his garden (my italics). Gough is much more specific in his second edition of Britannia (1806), but his figures will be considered when we come to Gough's other evidence.

says, 'Some [inscriptions] after ye ruin of ye old school were used to build walls', adding, 'No. 6, 9, 10 have been lately stolen by ye famous John Huetson or some such Genius'. Richard Gough, half a century later, says, 'very few are now to be found of thirteen fixed on the wall of Appleby school by R. Bainbrigg, master there in Camden's time' (1780, II: 15). This sounds as though it was derived from Hayton's list, but by the time he came to edit Britannia, Gough had different information available, even though Hayton is almost certainly the source of the statement 'At the school here . . . were bred bishops Barlow, Bedel and Smith ... '. Gough's information on Bainbrigg's stones was derived, he said (Camden 1806, III: 410), 'From a MS. paper in my possession signed H.T. probably Hugh Todd' He begins, however, with a specific statement: - 'The curious collection of inscriptions made here by Mr. Camden's friend Bainbridge [sic] were dispersed or stolen at the rebuilding of the school in this [?19th] century, except those to be specified. So little consolation have collectors who labour for incurious or ungrateful posterity!'.

In point of time, the next witness is I. Hayton. Although he referred to himself as 'hypodidascalus', which Haverfield glossed as 'assistant master', he does not appear in the lists either of headmasters or of 'ushers' of Appleby Grammar School (Hinchcliffe 1974a: 137-139). He made a record of a number of Latin inscriptions, giving no more precise location than 'in schola publica Aballabensi exposita'; Hayton, like Bainbrigg on occasion, indentified Appleby as Aballaba, in fact the Roman name of Burgh by Sands. Even this is imprecise, since one of the stones was 'Fixt in a wall at Alston More'. Haverfield (1911: 348) credits Hayton with having 'catalogued twelve Roman stones, besides some others ... of Bainbrigg's own devising'. In fact, Hayton catalogued thirteen Roman inscriptions, which he numbered 1-13, and three of Bainbrigg's own, two of them numbered 14, 15 and the other unnumbered (see illus. 2.1). Hayton ends 'These are all ye Inscriptions yt remain . . of a Collecr'formd by a most curious antiquary with great expense & pains'. One is led to wonder for what purpose Hayton made his record. The date is too early for its use by Horsley, whose book appears to have been prepared in short time (see p.v of the Introduction by Eric Birley to the 1974 reprint of Horsley 1732). A possibility would seem to be that it was drawn up for Edmund Gibson's second edition of Britannia, which actually appeared in 1722. Gibson orginated from Bampton, near Shap, and not so far from Appleby that he might not have known ofBainbrigg's collection, or even of Hayton himself

Hugh Todd's statement is clear and specific - and, of course, incapable of being tested. 'Upon a front of a little building made of stone by Mr. Reginald Bainbrigg in 1602, he being then the schoolmaster of the freeschool of Appleby, are placed divers stones having Roman inscriptions upon them; ... They are placed in two ranks, twelve in the one and eight in the other. There is one also above towards the roof; and another on a coyne-stone looking towards the south. There are two more also on the front of the school-house, viz. on each side of the door one.' The total, it will be noted, is twenty four - far more than anyone else suggests. Dr. Todd, however, does not help us in considering Roman inscriptions, for he says 'Of some of these Mr. Camden hath printed the copyes in his Britannia, which were communicated to him by this schoolmaster, as he acknowledgeth', while he himself proceeded to record the wording of three of Bainbrigg's own inscriptions and that of none of the Roman inscriptions. He had said that 'The letters upon some of these [inscriptions on the little building of 1602] seeme to have been cut deeper, by direction of the sayd Mr. Bainbrigg, in respect they were allmost worne out by time', but then finishes with 'The other stones [presumably other than the three Bainbrigg inscriptions] which are placed on the front of this building, all except two, I take to be copies made by the said Reginald Bainbrigg, from severall originals which he found in sundry places of this country and the parts adjacent; most of which are published by Mr. Camden, ifnot all'. This 'MS. paper ... signed H.T.', was, like Hayton's note, preserved by Gough; the reference is Bodl. MS. Gough Gen.Top. 36, fol.139r. All in all, Dr. Todd was a most frustrating witness; but Gough was no better, for he signally failed to 'specify' the inscriptions which were not dispersed or stolen.

Horsley, according to Haverfield (1911: 348), 'saw only eight', which is certainly inaccurate. He records at least twelve, apart from a reference to 'two or three stones with inscriptions upon them which the masons had lately destroyed at Appleby', of which he was told. Bainbrigg himself had mentioned, to Camden, an inscription which was 'now sett in my new house at Applebie', another 'in aedibus meis applebeiae', and a third in his garden (hortis), a location which Camden himself mentions. Horsley says of one inscription 'Under this stone has been another inserted in the wall whence the inscriptions at Appleby were taken, though now they are lying loose upon the ground (my italics) (1732: 297), and of another 'It is at present built up in the end of the old school house at Appleby (ibid: 298). Finally, Horsley, after quoting Camden on Bainbrigg's having taken some inscriptions to his garden, refers to another as 'in the bridge on the left hand, as you go into the town' (ibid: 300). The destruction by masons to which Horsley refers is presumably the same as that recorded by Hayton, who 26

Ben Edwards: Reginald Bainbrigg, scholemaister, and his stones

It behoves us at this point to consider Hugh Todd, whose

1. This is at ground level at the extreme left of the group. There is a horizontal groove worked into it, about a quarter of the way up. If it ever carried an inscription, it has now disappeared, but the groove makes it unlike any other of the stones apart from No.IO. Among possible identifications are the original or a copy of RIB 734 (Bowes) or part of a Bainbrigg copy of RIB 1875 Birdoswald (original at Rokeby). 2. A somewhat taller stone to the right ofNo.1, without discernible inscription or other features. 3. Stone, almost the full height of the wall, and with a diagonal break. Copy of RIB 1202 (Whitley Castle). According to RIB, this slab was 'now (1945) at Appleby Grammar School'. The identification is confirmed by the fact that Bainbrigg sent Camden a copy of the text including a number of transcription errors. One of these, the letters BECISADIA in 1.2, which should be read as part of ARA]/BICI ADIA[BENICI, is clearly visible today, as are other parts of the inscription. Camden's own reading of the inscription, resulting from his having seen the inscripiton at Whitley Castle, is much more accurate. Hayton records Bainbrigg's copy as No.8 in his list, adding a final line 'AD ALSTENMORE' which was doubtless on Bainbrigg's copy, and is among varied evidence which shows that Bainbrigg did not attempt to deceive in his copies. 4. Reads DEVM TIME, and was recorded by Hayton as the final line of his No.14, but, as RIB observes under 2357* (falsa), it is cut on a separate stone and in larger lettering. 5. Stone recorded by RIB as No.2357* (falsa). It is an eight line inscription summarising the history of Appleby (= Hayton No.14). ABALLABA QVAM C· C· / FLVIT/ ITVNA STATIO FVIT/ RO· TEMP· M· AVR· AVREL/ HANC VASTAVIT F· F· / GVIL·R· SCOT 1176/ HIC PESTIS SJEVIT 1598/ OPP· DESERT MERCAT'/ AD GILSHAVGHLINE. This is expanded thus by RIB (2357* - falsa): Aballaba quam c(ir)c(um) I jluit !tuna statio fuit I Ro(mana) temp(ore) M(arci) Aur(eli) Aurel(i) I hanc vastavit f(erocissimus) I Guil(elmus) rex Scot(orum) 1176 I hie pestis scevit 1598 I opp(ido) desert(o) mercat(us) I ad Gilshaughlin. The general sense of the thumbnail sketch of the history of Appleby is not in doubt; the precise meaning of Bainbrigg's inscnpt10n slightly more uncertain:- 'Appleby, which the Ituna (Eden) flows round, was a Roman station in the time of Marcus Aurelius; it was laid waste by the most ferocious king of Scots, William, in 1176; the plague raged in 1598, and the market moved from the deserted town to Gilshaughlin'. The repeat of 'Aurelius' is probably an error for 'Antoninus'. Gough (Camden 1806;III: 410), having quoted the inscription from Todd, expands F.F in 1.4 as 'funditis', which is improbable. Todd read Gilshaughlin with a succeeding letter F, which Gough expanded as 'fuit', but which has generally been taken as a final E. The last line of the inscription is now invisible. Bainbrigg would probably have expanded 'Scot' as 'Scottorum'. Gough says that Gilshaughline is 'four or five miles north-west of the town [of Appleby] in Cliburn parish', while RIB gives the location of Gilshaughlin House as nine miles west-north-

information Gough used. He was born about 1657, graduated B.A. in 1677, proceeding to M.A. in 1679, B.D. and D.D. in 1692. Collated prebendary of Carlisle in 1685, he died in 1728. He is described as an antiquary and 'a man of wide intellectual interests'. As a result, he assisted, among others, John Stevens in his additions to Dugdale's Monasticon and Edmund Gibson in his editions of Camden's Britannia. Though obviously no more reliable than was demanded by the standards of his day, his evidence is not such as to be discarded. However, it has not proved possible to discover any further information which he may have had on Bainbrigg's stones. It seemed at first possible that such information was to be found in Todd's MS Bishopric of Carlisle, of which a number of copies exist. That in the Bodleian (St. Edmund Hall MS 7/2-3) is deficient in the Appleby Deanery section, while another copy (the Youdale copy) in private hands has this section in full but without adding further information. There are two places in Appleby today where some of Bainbrigg's stones survive. The first of these is that described as the Broadclose wall (in what is now called Chapel Street); the second is the present site of the Grammar School, at Battle barrow to the north of the town, to which the school moved in 1887 (Hinchcliffe 1974a: 76). Nomenclature is not always consistent, but it seems that the school, having been served by the chantry priest of the chantries within the parish churches of St. Lawrence and St. Michael Bongate, had independent premises in the Schoolhouse, later called the Garden House. This lay in School House Close alongside a road called 'the road leading to the Butts', now Chapel Street. Bainbrigg rebuilt the school early in the seventeenth century on the opposite (east) side of that road, and additions were made in 1671 and 1826. It was at the time of the move under Bainbrigg that the old school became known as the Garden House, and it was when the school moved to its present site in 1887 that 'the inscriptions which Reginald Bainbrigg had incorporated in the 'garden house' ... were collected together and set in the wall of Broad Close opposite the site ofBainbrigg's School' (Hinchcliffe 1974a: 76), where they remain. Later (how much?) 'the lintel [of the 1671 extensions to Bainbrigg's school] and some other dated stones were ... moved up to Battlebarrow, where they still await a suitable permanent home' (Hinchcliffe 1074a: 76). It must be said that the Broadclose wall is a far from ideal location for the collection. They seem to have no legal protection, being neither a Listed Building nor a Scheduled Ancient Monument, and the remark quoted above from 1974 still remains true for the stones at the present school site. We will now proceed to a consideration of the stones themselves. There are sixteen such stones in the Broadclose wall (illus. 2.2) and two more are at Battlebarrow. The former group are represented diagrammatically and numbered in illus. 2.3; the two other stones are numbered sequentially.

27

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

west of Appleby. The site, at NY 571245, is almost precisely 7½ miles (12 km.) in a straight line from Appleby Market Place. 6. Stone with an incised inscription set within a raised border and with raised bands between the lines. It would never be mistaken for a Roman inscription, and is not mentioned by RIB. It is, however, recorded by Todd on Gough's 'MS paper'. It reads D· O· M· / D· L· L· M· REGINALDVS/ BAINBRIG QVI DOCVIT/ HIC ANN· XXII· LET· S· 57· 1602/ H· M· S· V·P, and was mentioned by Haverfield ( 1911: 345) as enabling an approximation to be made for Bainbrigg's date of birth. It is presumably also the source of Todd's statement that the new school on the east side of Chapel Street was erected in 1602. It is possible but profitless to suggest expansions for the abbreviations in 11.1,2 and 5. 7. Stone reading clearly Q· S· S· S· A· P· / CRA WDVNDALE. Horsley (1732: 299) says that the abbreviations may be expanded 'Quae supra scripta sunt apud Crawdundale'. It was therefore a label for the following inscription - the Crowdundle Beck quarry inscription( s). 8. Copy of RIB 998, quarry inscription from the Crowdundle Beck. RIB says of this copy 'now (1945) ... 11.4,5 have weathered away'. The lowest line (4) of the inscription is indeed illegible, but any fifth line would be below pavement level. Hayton (No.4) records the fifth line as CN· OCT· COT· COSS, presumably implying a consular date, as does Horsley (1732, 299), quoting Camden and Bainbrigg's copy, but suggesting that the original was on a separate rock from RIB 998. Bainbrigg himself, in BL Cotton Julius FVI, £334, olim 316, adds V to COSS. 9. Stone bearing the clear inscription H· M· EST·/ GALLACI, interpreted by Horsley (1732: 298) as 'Hoc monumentum est Gallaci'. The name Galacum was thought by Horsley, following Camden, to apply to Wheallep or Wheallop Castle i.e. Kirkby Thore. This stone is therefore a label for RIB 759, now below it, but above it in Horsley' s time. 10. RIB 759. This stone, (Hayton No.2) seems to be generally accepted as the original dedication to Belatucadrus. Bainbrigg himself says (BL Cotton Julius FVI, £333, olim 315) 'This stone was found in Whellep castle in Kirbethore and now sett in my new house at Applebie'. 11. This stone has clearly had three lines of text, and possibly a fourth, but is now virtually illegible. Mr. Shannon, in his consideration of these stones, suggested that it might be part of a Bainbrigg copy of RIB 2060 (Bowness on Solway) cut down to fit the width of the stone below, but I cannot read sufficient letters on it to justify this suggestion. Hayton's sketch of Bainbrigg's copy of RIB 2060 makes it look somewhat different from Bainbrigg's drawing, preserved by Camden, and reproduced in RIB. Hayton has only about ten letters per line as against about thirteen, and five lines of text. In addition, as a final line, he shows BLATI BVLGII, as in several cases, a label for the rest of the inscription.

12. A stone bearing a six line inscription commemorating Robert Langton and Miles Spencer as founders of Appleby Grammar School. It is recorded (unnumbered) by Hayton, and is mentioned in several modem sources relating to the school. It reads ROBERTO LANG/ TON ET MILONI/ SPENCER QVV APPLEBilE F.F./ HANC SCOLAM/H·M·OB M·P·R·B·P· The last line Hayton explains as 'Hoc monumentum ob memoriam pietatis R. Bainbrig posuit'. He then goes on to add 'R. Langton & M. Spencer were both LL.D of Q[ueen']s C[ollege] Ox[ ford]. At this school were bred Barlow bp of Lincoln, Bedel of Kilmore & Smith of Carlisle', the latter part of which was quoted by Gough in his edition of Camden. Gough also added, from Hugh Todd, the additonal lines DE REPVBLICA BENE MERERE/ PVLCHRVM EST. The bishops were Thos. Barlow (1607-1691), Wm. Bedell (1571-1642) and Thos. Smith (1615-1702). 13. A piece of stone rather different in character from most of the others, being less pink and more obviously bedded. It carries the remains of two lines of ?text noticeably lacking in curved elements. In fact, at first sight it suggests a set of Roman numerals beginning I II III IV ... It is just possible that it is a stone reported by Horsley (1732: 192 N60, Westmorland VII, and 300) as being built into Appleby bridge, which carried a text beginning IXIT, but the second line of that text began with SD, and it does not seem possible to restore those letters on this stone. It is worth noting that Horsley's text, not quoted in RIB, read 'IXITIVC IMVI/ SDVN VIXIO/ OSI VIINC' or something like it, which, as Horsley observes (1732: 300) 'is so much effaced, as to be unintelligible'. Holdgate (1982: 81) says 'part of the inscription which once surmounted the arch [sc. of the bridge] on the town side is still visible in the wall of a nearby house'. This, which has been recently sought in vain, may have been a medieval inscription from the bridge or the Roman one with which we are concerned. It is at least interesting that the medieval bridge was demolished in 1888, about the time the Grammar School was moving to Battlebarrow and the Broadclose wall was being built. 14. Copy of RIB 2285 (from near Brougham, and now at Brougham Castle). There were various problems with the precise transcription of this milestone inscription, with Camden correcting Bainbrigg's obvious error in inserting D at the beginning of line 2, but following Bainbrigg in breaking CONSTA/NTINO thus rather than CONST/ANTING. Horsley (1732: 192N59, Westmorland II, and 192) recorded the inscription correctly; Bainbrigg's copy is correct apart from the meaningless D at the beginning ofline 2. 15. This stone, while obviously belonging with the others, and quite unlike the normal building stones of the wall, is effectively blank. 16. The last stone in the collection in the Broadclose Wall is set a little to the right of, and separate from, the remainder. It bears clear traces of an inscription, but it is difficult to make any sense of them. The possibility has been raised that it is either the original or a copy of part of RIB 734, from Bowes, set upside down. It appears from Horsley (1732: 192N62 Yorkshire II, and 304) that both 28

Ben Edwards: Reginald Bainbrigg, scholemaister, and his stones

5. F[oun]d at Blatum Bulgium / (Boulness). Ye A in ye 2 first/ lines wants ye transverse / stroke, a mark of ye age from I Severus to Gordian. 6. At Aesica (Netherby / on Usk [Esk]). 7. F[oun]d at Lanercost / Abbey. Another/ copy adds G.P.R.F. / ye meaning ofwh[ich] / is not easy to say. 8. Fixt in a wall at Alston More / near Whitley castle where it was / found. difft erent] from Camden's copy. 9. Found at Lanercost. 7 & 9 are copies of/ inscr[iption]s at Burdoswald where / ye Cohors LElia dac. I kept / garrison & not at Lanercost / wh[ich] was never a station. 10.11. Supposd to be floun]d at Risingham. 12.13 F[oun]d at Chester on ye wall. / [Q[uery] if ye last word meaning 2 / or 100 sest[ertii] vowd by ye erector]/ certainly yt it came from chester on ye wall. 14.15. This is all yt remains ofye Under Masters house on ye S. side of ye School. /Ye scite is now a garden. Hoc Monumentum ob I memoriam pietati R. / Bainbrig. posuit[.] R. Lang / ton & M. Spenser were both / LL.D of Q.C.Ox. I At this school were bred/ Barlow b[isho ]p of Lincoln./ Bedel ofKilmore & / Smith of Carlisle./ Bainbrig was Camden's transcriber.

the original and 'a copy ... upon another stone ... yet remains at the same place'. This might seem to strengthen the likelihood of at least one of the two having survived, but Hayton's illustration gives four lines of text on Bainbrigg's copy, the first of which, following a line of dashes to represent a preceeding illegible line, can be accommodated to Horsley's NOB CAES, but the last of which defies restoration. The remaining two stones are now at the Battlebarrow site. Both are cut on grey sandstone rather than the red used for those in the Broadclose wall. 17. The stone recorded in RIB as 2356* (falsa). It reads IVL· FRONTI/ NVS DOMVIT/ BRIGANTES/ ET SILVRES/ HVIC SVCCES/ SIT IVL AGRIC/ SVB VESP AS, and is tersely dealt with thus in RIB 'Bainbrigg summarizes (inaccurately) the achievements of two governors of Britain under Vespasian'. 18. A stone reading H· AED· R· BAN/ BRIG· HIPODIDA/ SCVLIS· D· D·. Hinchcliffe (1074a: 33) says 'By 1606 sufficient progress had been made for Bainbrigg to hand over the priest's mansion house to the undermaster - and this stage he marked by the following inscription: R. Bainebrig Hoc. Aed. Hipodidascalus D.D. in P.P. 1606'. The differences between this record and the surviving stone will be noticed, and, while it is possible that a line reading 'in P.P. 1606' has been trimmed from the stone, it is curious that it does carry an inscription on its lowest surface (i.e. the narrow side at right angles to the main inscribed face. The inscription seems to read (it is not completely visible) *ACOBOI, where the asterisk indicates a chip off the comer of the stone, and where, formally at least, the fourth letter could be R and the last L.

These are all ye Inscriptionsyt remain or can now be / found. Some after ye ruin of ye old school were used to build / walls. No. 6 9 10 have been lately stolen either by/ ye famous John Huetson or some such Genius so yt scarce any / fragm[en]ts remain of a Collection formd by a most curious / Antiquary with great expence & pains.'

NOTE I am grateful to Mr. David Mawson of Brampton for sending me much information on Hugh Todd, including a copy of his entry on Todd for the new DNB. The Headmaster and Governors of Appleby Grammar School permitted access to, and photography of, the stones at the school's present site, and the Bursar first showed me those in the town.

I am delighted to dedicate this little study to the memory of an old friend whose achievements were less than they ought to have been as a result of his untimely death, but whom I have long numbered among the most significant of my archaeological acquaintance. APPENDIX

BIBLIOGRAPHY

Transcript ofHayton's record

Camden, W., 1586. Britannia sive florentissimorum regnorum Anglice, Scotice, Hibernice, et insularum adiacentum ex intima antiquitate Chorographica descriptio. London: Ralph Newbery. Camden, W., 1637. Britain, or a Chorographicall Description of the most flourishing Kingdomes, England, Scotland, and Ireland, and the Islands adjoyning, out of the depth of Antiquitie . . . . translated newly into English by Philemon Holland Doctour in Physick. London: Andrew Crooke [or Joyce Norton and Richard Whitaker]. Camden, W., 1722. (ed. Gibson, E.) Britannia, or a Chorographical Description of Great Britain and Ireland, Together with the Adjacent Islands. London: James and John Knapton et al. 2 vols. Camden, W., 1806. (ed. Gough, R.) Britannia: or a Chorographical Description of the flourishing

'Inscriptiones, Tituli aliaq. Romanre Antiq[ uitat]is mon[ ument]a / ex variis Nationis Briganticre partibus conquista & in / schola publica Aballabensi exposita ad autographoru[m] / fidem diligenter exscripta 1722 per I. Hayton scholre / hypodidascalum. 1. Found at ye Confluence / of ye Eimot [Eamot] & Loder [Lowther]/ 1602 near May/ boro' Castle. 2. Fd at Gallagum (Wheallop Castle) / near Kirkby Thore. / Belatucader seems to have been ye tute / lar deity ofye Brigantes. He had another/ altar at Wigton inscribed / "Aurelius diatona deo Sancto Belatu / cadro aram ex voto posuit". 3. Bro[ugh]t from Bowes 1605 I & fixt up here. The coh. / 1 Thrac. appears to have / occupied yt place under / Severus & to have erected/ a bath whence ye place/ prob[ab]ly had ye name ofLavatrre. 4. Over a quarry at Crodundale. / 1500 paces from Kirbythore. The / 2 first lines damagd by ye root of/ a tree w[hi]ch grows over it. The beg[inning] & / end are ill cut, ye middle much/ fairer. These 2 consuls do not / occur together in ye fasti.

29

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Kingdoms of England, Scotland, and Ireland, and the Islands adjacent; from the earliest antiquity. London: John Stockdale. 4 vols. Collingwood, R.G. and Wright, R.P., 1965. The Roman Inscriptions of Britain, vol. 1, Inscriptions on Stone. Oxford: Clarendon Press. [Abbreviated as RIB throughout] Edwards, B.J.N., 1998. William Camden, his Britannia, and some Roman Inscriptions, lecture delivered at the Senhouse Roman Museum, Maryport, 27 October, 1998. Privately printed Gough, R., 1789. British Topography or, an historical account of what has been done for illustrating the topographical antiquities of Great Britain and Ireland. London: T. Payne and Son and J. Nichols. 2 vols. Haverfield, F., 1911. 'Cotton Julius F.VI: Notes on Reginald Bainbrigg of Appleby, on William Camden and on some Roman Inscriptions', Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, new series, 11, 343-378. Hinchcliffe, E., 1974a. Appleby Grammar School from Chantry to Comprehensive. Appleby: J. Whitehead and Son for the Governors. Hinchcliffe, E., 1974b. The Bainbrigg Library of Appleby Grammar School. Newcastle upon Tyne: Oriel Press. Hinchcliffe, E., 1996. The Bainbrigg Library of Appleby Grammar School. Newcastle upon Tyne: Working Paper, History of the Book Trade in the North, University of Newcastle upon Tyne. Holdgate, M., 1956 (3rd ed., 1982). A History of Appleby. Appleby-in-Westmorland: Dalesman Books for J. Whitehead and Son. Horsley, J., 1732. Britannia Romana: or the Roman Antiquities of Britain: in three books. London: John Osborn and Thomas Longman. (Reprinted, 1974, with introduction by Eric Birley, Newcastle upon Tyne: Frank Graham).

30

Ben Edwards: Reginald Bainbrigg, scholemaister, and his stones

JJ(LCA-

VA-Rf1

l£1Jvi1"~t

Fig. 2.1 I. Hayton's record of Reginald Bainbrigg's stones, 1722. (Bodleian Library, University of Oxford, Gough Gen. Top. 128, fols 43r and 43v-44r; reproduced by permission).

31

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

MS. GoughGeri.Top 128 Fo!.43v-44r

32

Ben Edwards: Reginald Bainbrigg, scholemaister, and his stones

2.2 Reginald Bainbrigg's stones in the Broadclose wall, Appleby-in-Westmorland, 1998. (Photo: author).

The Bainbrigg Stones, Appleby-in-Westmorland

n

~

19Mf7

18

6

10 12

0

FT.

5

0

M.

1.5

14

~

2.3 Diagram of the Bainbrigg stones, Appleby-in-Westmorland. (Author).

33

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

34

3 Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland Richard Gregory Barri's lifelong commitment to the analysis of Roman frontiers and to the relationships between Roman and 'native' populations spanned the limits of the Roman Empire. Its strongest expression, perhaps, was made in his seminal contributions to the later prehistoric and early historic archaeology of the Solway area and, in particular, the Cumbrian Solway plain. His research there was established through a series of aerial surveys undertaken between 1974-76 and many excavations that extended over two decades (c£ Higham and Jones 1975; 1985).

the physical limits of the Empire were implicitly defined by the Tyne-Solway isthmus. The effects of this sustained tactical interest on the indigenous communities of the region have been considered in numerous studies (inter alia; Curle 1913; 1932a; 1932b; Robertson 1970; Fulford 1985; 1989; Millett 1990; Jobey 1974; Jones 1979a; Higham and Jones 1985; Jones and Walker 1983; RCAHMS 1997; Wilson 1998). These effects are identified through the occurrence of individual artefacts, found in a variety of non-military contexts, which are presumed to form an element of Roman material culture. Alternatively differences in settlement and agricultural patterns, between areas located north and south of the frontier are taken to indicate different indigenous responses to the Roman presence. (cf. Piggott 1958; Breeze 1984; 1989; 1990). Fulford (1985; 1989), for instance, views the circulation of Roman material in the frontier zone, acquired by either reciprocal trade or diplomatic congress, as a phenomenon largely confined to the indigenous elite. In this context, the circulation of high quality goods is seen as a mechanism that actively promoted, and secured, an elite's position at the apex of the social hierarchy. Fulford (1985) also observes that, in the Scottish lowlands, an increase in native metalworking coincided with an increase in the movement of Roman goods into the region during the firstsecond centuries AD. This he imagines may have been due to the stimulating effects of the Roman military market on the local economy. Correspondingly, the apparent decrease in metalwork during the fourth century may be due, in some measure, to the withdrawal of troops from the frontier zone (Allason-Jones 1989). A similar economic argument is adopted by Macinnes (1989), who notes that metalwork, as well as glass ornaments and more mundane stone objects, may have been produced for Roman consumption and distributed via Rome's dealing with the indigenous elite. Numerous writers have suggested that these exchange relationships may have formed part of a deliberate frontier policy, implemented by Rome, to create a membra partesque imperii at times when the frontier had retracted. These buffer zones were, perhaps, perceived by both Rome and the 'peripheral' states themselves as a means of protection against concerted attack from outside, or against dispersed banditry from within, without the concomitant economic burdens which would result from direct annexation (Braund, 1984; Whittaker, 1989; Hanson, 1989; 1997).

One problem which stemmed from this research, and which constantly pervaded Barri's interpretations, was an apparent hiatus in the settlement record. The hiatus stretched from the middle of the mid/late first millennium BC to the development of Romano-British settlement sites at some stage following the establishment of the Roman military frontier. Inevitably, in Barri's eyes, the apparent lack of settlement continuity made redundant any attempt to interpret the Roman Iron Age landscape in terms of the Pre-Roman Iron Age landscape. It is with this issue in mind that this essay has been written. Its focus, though, is not on the southern Solway Plain, so critical to many of Barri's thoughts and arguments, but on the adjacent landscape found to the north of the Solway: an area roughly equating with the former counties of Dumfriesshire and Kirkcudbrightshire in South West Scotland. Much of its backbone depends, once again, on the results of Barri's own aerial survey, which was carried out between 1976-8. Given the nature of the data and its status as 'evidence' much of what follows is inevitably speculative, but it is dedicated with affection and admiration to the memory of Barri's contribution to this seemingly remote comer of the Empire. Introduction For the Romans the landscape of Dumfriesshire became synonymous with the concept of an intra provinciam, or frontier zone (inter alia; MacDonald and Barbour 1897; MacDonald 1923; Birley 1948; 1954; Wilson 1998). In this sense the region was not unlike other peripheral zones adjacent to the Empire in that it was, at least in parts, periodically occupied by Roman forces during the Flavian, Antonine and Severan periods. Its strategic position as a major route-way into Central and Northern Scotland, however, appears to have fostered a high and concerted level of both military and political interest. This is seen vividly in the remnants of Roman lines of communication, particularly along Annandale and Nithsdale, and in the remains of a series of forts and camps, a number of which increased in strategic importance as frontier outposts when

Settlement studies of the region during this period tend towards broad generalisation. Aerial photographic data acquired from both sides of the Solway in the 1970s, for example, led to the suggestion that, on the northern Solway

35

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

plain, settlement densities were thinner, sites were more heavily defended, and agriculture was less developed, during the second and third centuries AD, in comparison with contemporary settlement south of the Solway (Jones and Walker, 1983; Higham and Jones, 1985). These differing conditions are implicitly related to different levels of security, economic stimulation and, in turn, subsequent prosperity either side of the frontier. In turn this connects them to the influence and effects of the pax Romana and to the notion of Romanization, as envisaged in Haverfield's (1905) original definition (cf Ringley 1996), and subsequently adopted by Curle (1913) in a Scottish context. In essence, the presence of the Roman military in the frontier zone, either occupying the region or in close geographical proximity to it, is held to have led directly or indirectly to a broad progression, or civilising, of the indigenous elite. Once this elite had become acculturated the process permeated through the indigenous social strata by a process of emulation (Millett 1990).

by a number of 'tribal' groups. The precise geographical extent of the varying tribes, however, is not made clear, and this cartographic dilemma has formed the source of several discussions (inter alia Maxwell 1980; Hanson and Maxwell 1983; Mann and Breeze 1987). On balance, these seem to suggest that the N ovantae may have occupied the greater part of Galloway, extending eastwards, perhaps, to the fringes of Nithsdale (Maxwell 1980). Nithsdale and Annandale may have been occupied by the Selgovae (Mann and Breeze, 1987), while Brigantian territory could have encompassed a large area of the northern Solway plain (Breeze and Dobson, 1987, but see Birley, this volume). Within Brigantian territory, however, later references to the Carvetti and Anavonienses indicate the existence of smaller tribal septs (Rivet and Smith 1979; Higham and Jones, 1985). From the viewpoint of the Roman military, the literary evidence suggests that the landscape supported a number of tribal groups which may have been organised in some kind ofloose confederacy.

Recent conceptual and semantic debates on the validity of the term 'Romanization', however, together with the development of 'post-colonial theory', have questioned these types of interpretation (inter alia Webster 1996; Ringley 1996; 1997; Mattingly 1997; Barrett 1997). In particular, these critiques suggest a much greater degree of complexity in the relationships between Roman and 'native', or colonisers and colonised. Sometimes, for example, relationships appear to have been reciprocal, bestowing both positive and negative benefits to the parties involved. For the indigenes, this may have involved some degree of emulation but equally it involved both overt and covert acts of resistance (Webster 1996; Ringley 1996; 1997; Mattingly 1997). It is important to recognise, however, that such processes, and their outcomes, were heavily dependent upon local circumstance (Mattingly 1997). In effect the experiential nature of both Roman and 'native' identity was probably a product of specific regional and historical conditions (cf Barrett 1997).

This geo-political description can be elaborated, however, by consideration of the archaeological and palaeoecological records for the region as these directly relate to the evolving settlement pattern and to the nature of economics, agriculture, and ritual in the period leading up to Roman occupation. Taken together, this evidence seems to suggest that the indigenous cultures of southwestern Scotland were beginning to approach some kind of zenith in the later stages of the Pre-Roman Iron Age. Moderately dense and well-organised agrarian communities were settled in a variety of different physical habitats (RCAHMS, 1997; Gregory, 1998; in press). Settlement

In attempting to chart this evolving social pattern, the remnants of later prehistoric settlement offer by far the largest set of available data. In the uplands and in the area west of the River Nith this is found in the form of upstanding earthworks. On the north Solway plain and in the major river valleys draining into the Solway Firth the evidence consists largely of crop-marks. Where settlement can be identified, it appears to vary in form and character between enclosed and unenclosed types, as it does in other areas of lowland Scotland and north eastern England.

This contribution addresses the holistic effects of Rome on an indigenous community inhabiting a large swathe of south west Scotland. The area is defined arbitrarily as stretching from the fort at Birrens to the fortlet at Gatehouse of Fleet. It contains an apparent concentration of Roman military fortifications amidst a well developed Pre-Roman Iron Age landscape and consequently forms a useful 'laboratory' in which to examine the penumbra of Roman occupation.

The unenclosed settlements in the region consist merely of single isolated houses or larger conglomerations of houses. They are scattered sporadically through the uplands and lowlands but form a relatively small component of the overall settlement record for the region, though presumably the archaeological visibility of these settlement types may account for this apparent rarity (illus. 3.1). The complex structural histories that are evident at a number of these sites suggest a particularly dynamic form of settlement that was employed through much of the last two millennia BC, and the early first millennium AD. In some cases, unenclosed settlements are temporally isolated, but in other instances they appear to either precede, or post-

Indigenous society

Inevitably the basis of this analysis rests with the Late PreRoman landscape 'which met the eyes of the advancing troops of Agricola' (Keppie 1989: 63). Records of these early proto-historical contacts provide a starting point for a consideration of existing local conditions. The writings of Tacitus, and the cartography of Ptolemy indicate that on the eve of Roman military activity the area was occupied

36

Richard Gregory: Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland

date, more substantial, more easily recognisable, enclosed settlements which are a more enduring feature of the prehistoric landscape in consequence of their upstanding boundaries. They also appear to constitute the dominant settlement type in the area (illus. 3.2). The materials used in boundary construction, and in the construction of internal features in these enclosed settlement forms, was probably dependent on the availability of local resources, and hence the use of differing building materials is probably not of any particular chronological significance. Thus, depending on the geographical locale, houses were constructed in either timber or stone. Boundaries were defined either by timber palisades, ditches with internal earth or stone banks, stone walls, or, in some cases, merely a low bank enclosing a scooped settlement interior (RCAHMS 1914; 1920; 1994; 1997; Jobey 1971). At other locations, like the crannogs of the region, natural features such as water were utilised in the creation of a boundary (Munro 1882; Morrison 1985).

be defined in plan by twin-grooves and a medial bank (RCAHMS 1997). Chronologically, it is now generally accepted that settlement composed of ring-ditch houses, and possibly those houses defined by twin-grooves, probably date to the Early/Mid Pre-Roman Iron Age (Hill 1982a; 1982b; RCAHMS 1997). Though they have been located inside the boundaries of numerous enclosed settlements, recent survey by the Royal Commission in eastern Dumfriesshire has shown that these early house types are sometimes found either singly, or clustered, as unenclosed settlement (RCAHMS 1997). It is therefore difficult to demonstrate an intimate connection between early house types and settlement enclosure. It may, however, be significant that where these early house types are found within enclosed settlement, these settlements are generally located in defensive positions. The other house type favoured in the area used a single 'ring-groove' to support the outer wall of the round house. Excavations in south west Scotland, and in other areas of northern Britain, indicate that the larger examples of this hut type have a long currency. They appear to have been used from the second half of the first millennium BC, until the early centuries of the first millennium AD (Hill 1982a; Hill 1982b; Haggarty and Haggarty 1983; Raisen and Rees 1995; Rideout 1996; Gregory 1998; in press). The smaller and more ephemeral examples, however, may be placed in the Late Pre-Roman/Roman Iron Age due to their association with the scooped settlements of the region (cf. RCAHMS 1997), and many have been located within the boundaries of other smaller, non-defensive settlements of the area.

Inevitably, these enclosed settlements vary in size, morphology, and location, very much as they do on the south side of the Solway (e.g. Higham and Jones 1983). The internal area of the enclosed curvilinear settlements, although varied, is generally under 0.3 ha. A small proportion of the curvilinear forms also show signs of multivallation, although - in the absence of excavation the contemporaneity of these boundaries is not always clear. The geographical siting of the settlements may also be significant. In many of the univallate and multivallate curvilinear sites the topographical settings seem defensive in character and they are often referred to as 'forts'. In the majority of cases, however, the settlements are located in non-defensive positions (cf. Jobey 1971; RCAHMS 1997). In addition to curvilinear forms, rectilinear enclosed settlements are also widespread though they tend to be clustered in spatially discrete zones. On the whole, they survive as crop-mark sites on the coastal plain, or within the eastern river valleys where modem agriculture is most productive. It seems likely that these locations were also characteristic of the Pre-Roman Iron Age landscape.

Where the details of house types are less clear, enclosed settlements may be broadly classified in terms of internal area, geographical siting, morphology, and the architecture of the settlement boundary. For the purposes of dating, this allows comparison with the few examples that have been excavated in lowland Scotland, and northern England. Unfortunately, this kind of typological analysis throws little chronological light on some important settlements because the construction of some boundaries was both possible, and appropriate, over considerable periods of time. This was certainly the case with the curvilinear palisaded settlements. Both the available radiocarbon dates from northern Britain (cf. Gates 1983; Jobey 1985; Rideout 1996), and the apparent construction of later palisaded phases over earlier Pre-Roman/Roman Iron Age earthworks in south west Scotland (RCAHMS 1997; Gregory 1998), indicate that construction in this style was a continuous feature of enclosed settlement during much of the first millennium BC and the first half of the first millennium AD. The same also appears true for the crannogs of south west Scotland. Excavation over the last century, and a recent programme of radiocarbon dating, reveals that many of the crannogs were constructed during the Late Bronze Age/Early Pre-Roman Iron Age but continued to be occupied into the Roman Iron Age and beyond (cf. Piggott 1953; Barber and Crone 1993; Crone 1993).

Because there has been little excavation, and due to a lack of diagnostic finds and uncertainties with the radiocarbon calibration curve during the Early/Mid Pre-Roman Iron Age, the dating of these varying settlement types is problematic. It is difficult to formulate any precise schemes of settlement evolution of the kind typified by the 'Hownam sequence' (Piggott 1948; 1949; Feachem 1961; Hill 1982b). In more general terms, however, a broad chronology, based on the architectural detail of round house construction, might be suggested. Survey and excavation in southern Scotland and the north east of England indicate that, within many unenclosed and enclosed settlements, different styles of vernacular architecture were utilised in round house construction. In south west Scotland, for example, these include ring-ditch and ring-groove houses, alongside other unique architectural forms, such as round houses which appear to

37

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Other enclosed settlement types, however, do appear to be linked to discrete temporal periods within later prehistory and early protohistory and allow some assessment of the changing nature of early settlement in the region. The most important are rectilinear forms. Their construction appears to have been a notable feature of the last two centuries BC, and to have continued into the Early Roman Iron Age (Jobey 1962; 1973; 1977; 1978b; 1984; Jobey and Jobey 1988; Johnston 1994). Excavation at a series of sites in south west Scotland, notably Rispain Camp (Haggerty and Haggerty 1983), Carronbridge (Johnston 1994) and Hayknowes Farm (Gregory in press), appear to corroborate this chronology with radiocarbon dates which indicate occupation spanning the Late Pre-Roman/Roman Iron Age.

excavations are interpreted, Traprain Law (Jobey 1976; Hill 1987; Close-Brooks 1987). This period of abandonment, dating to the Mid Pre-Roman Iron Age, may also be a feature of other defensive settlements in the region. In the uplands of eastern Dumfriesshire, for example, a number of defensive sites, or forts, reveal evidence for abandonment, with a later period of occupation which is detected in the presence of smaller scooped settlements. These are thought to date to the Late Pre-Roman/Early Roman Iron Age and are usually found overlying the earlier fort boundaries (RCAHMS 1997). The same may also be true within the lowlands, where a similar pattern appears at many cropmark sites (RCAHMS 1997). Other large, defensive settlements, however, appear to have been occupied throughout the Mid/Late Pre-Roman Iron Age and, in some instances, into the Roman Iron Age. This has been confirmed at two sites in the area. First, partial excavation at the 'upland fort' of Castle O'er in Eskdale produced radiocarbon dates suggesting maintenance of the boundaries throughout the Roman Iron Age (Mercer 1985; RCAHMS 1997). Second, excavations at a large circular double-ditched settlement, or lowland fort, associated with an extensive series of agricultural enclosures at Hayknowes Farm near Annan, produced evidence for occupation extending from the Mid PreRoman Iron Age to the Late Pre-Roman/Early Roman Iron Age (Gregory, in press). Although tenuous, this evidence appears to imply that many of the larger univallate/multivallate defended sites, which show no obvious sign of abandonment in the form of superimposed small settlements, may also have been occupied during the Late Pre-Roman/Early Roman Iron Age. Many of those lowland sites, which share morphological similarities to the Hayknowes settlement, may, therefore, be contemporary with the smaller curvilinear and rectilinear settlements. If this chronological argument is correct it implies that a developed settlement hierarchy, possibly initiated at some stage during the Mid Pre-Roman Iron Age, was a dominant feature of the Late PreRoman/Roman Iron Age landscape.

A crude chronological assessment may also be attempted for the remaining enclosed settlements. One type which seems particular prolific and which falls into this category is non-defensive in character. It is comprised of settlement bounded by a curvilinear ditch and/or embankment, which encloses a small internal area, generally under 0.3 ha. To this class of site may also be added the small settlements with a scooped interior, which are confined to eastern Dumfriesshire, and the small stone-walled settlements found west of the River Nith (Jobey 1971; RCAHMS 1914; 1920; 1994; 1997). Evidence from Long Knowe, Eskdale (Mercer 1981), and a palisaded element of a ditched and embanked settlement at McNaughton's Fort, Kirkcudbrightshire (Scott-Elliot et al. 1966), suggest the possible use of sites like these from as early as c. 400 cal. BC (2 sigma). The evidence from the scooped settlements at Boonies (Jobey 1974) and Upper Cleuch (Terry 1993), however, together with associated palynological data, suggests that the construction of these small units 'peaked' during the Late Pre-Roman/Early Roman Iron Age (RCAHMS 1997). In this context, it may be significant that, in the recent survey of eastern Dumfriesshire, the majority of these settlements were not found to be associated with early house types, though a number showed evidence for relatively small ring-groove huts located inside the settlement boundaries (RCAHMS 1997).

Agriculture

The larger settlement units, which include those in 'defensive' locations and those showing signs of multivallation, whilst possessing a potentially extensive chronology, reveal an interesting pattern of occupation which may hold some relevance in assessing the changing social landscape in later prehistory. At the largest of these sites - the contour fort surrounding the eminence of Burnswark Hill - limited excavation suggests initial occupation between c. 800-400 cal. BC (2 sigma). This was probably followed by a period of abandonment, and then succeeded by a more ephemeral reoccupation at some stage during the Early Roman Iron Age (Jobey 1978a). It may merely be coincidence, but this sequence of occupation is similar to that recorded at other large hilltop sites in southern Scotland, such as Eildon Hill North (Owen 1992), and, depending on how the confused

The settlement of the area, of course, reflects an intrinsic link with the land. Well-developed settlement required the support of a well-developed agricultural system, which was, in turn, dependent on an extensive regime of land clearance. Fortunately, the nature and intensity of later prehistoric agriculture and clearance is now beginning to emerge through advances in survey, excavation and palynology. Palynology offers a broad chronological framework for the sequence of land use during the PreRoman Iron Age, and it is significant that this sequence corresponds with the sequence of later prehistoric settlement deduced from archaeological evidence (Tipping 1994; 1995; 1997b; RCAHMS 1997; Gregory 1998). During the Early and Mid Pre-Roman Iron Age, clearance is recorded in both upland and lowland environments. The pollen evidence indicates that it was relatively minor in

38

Richard Gregory: Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland

scale, and was associated predominantly with a pastoral economy, although some limited cereal cultivation may have occurred (Edwards et al. 1991; Jones et al. 1989; Birks 1972; Rapson 1994; Tipping 1994; 1995). With the onset of the Late Pre-Roman Iron Age these clearances appear to rapidly intensify, particularly in the lowland landscape. On the northern Solway Plain, at Burnfoothill Moss (Tipping 1995) and Racks Moss (Nichols 1967), arboreal pollen decreased markedly. At Burnfoothill Moss this is thought to date between 131 cal. BC - cal. AD71, reaching a zenith at c. cal AD150 when very little tree cover remained (Tipping 1995). This intensive episode of clearance was probably connected to the extension of both arable and pastoral farming. In the uplands, at sites such as Long Knowe, a similar pattern has been recorded (Mercer 1981; Tipping 1997a), though this may have been related more closely to livestock rearing (Mercer 1981; RCAHMS 1997). These events are not peculiar to this region. They are duplicated at other sites surrounding the Solway, such as Walton Moss, Cumbria, where a rapid decrease in arboreal pollen is also recorded during the Late Pre-Roman Iron Age (Dumayne and Barber, 1994). Moreover, this intensive episode of clearance forms a recognisable element of many Late Pre-Roman Iron Age pollen profiles from an even wider area of northern England and southern Scotland (Tipping 1997b). At the broad level of the region, the ubiquity of this clearance may suggest a hierarchical social organisation working towards the same aims (Tipping 1997b). The settlement evidence for south west Scotland indicates that the area shared a similar hierarchy.

(cf similar evidence from north Cumbria: Higham and Jones 1975; Higham 1978, 1986: 119ft). The initial use of these large land division systems appears to be a feature of the Late Pre-Roman Iron Age. At Hayknowes Farm, near Annan, a date of 340 cal. BC - cal. AD20 (2 sigma) was obtained from a drove-way which formed one component of an extensive system of linear boundaries that were probably associated with a large Late Pre-Roman Iron Age settlement (Gregory, in press). The appearance of these landscape features neatly corresponds with the intensive period of clearance evident in the pollen record, and to a period when the hierarchical system of settlement in the region was approaching, or at, its zenith. Indeed, the 'stock enclosures' that are linked to many early settlement types may well have played an integral part in this process. It is tempting to view these enclosures as a direct reflection of the potential quantities of stock that a settlement unit could possess. Hence, if quantities of stock can be equated with degrees of wealth, or power, it may well have been that the accumulation of this valuable resource acted both as the initial stimulus for the development of a social hierarchy, and subsequently promoted and propagated individual settlements within this social milieu. Ideology and ritual Other, subtler, elements may also be identified lurking within this domestic landscape. Though constrained in some measure by the pragmatic circumstances of the PreRoman Iron Age, these elements relate to a set of ideologies which were specifically linked to wider, shared, cosmological beliefs. Once identified and separated, their interpretation is significant in understanding the ways in which Pre-Roman Iron Age people conceptualised their world. One aspect of this ideological reconstruction is an assessment of the ritual deposition of artefacts representing votive offerings of some kind. This took place in three general contexts: first, the deposition of objects in the natural landscape; second, the deposition of 'votive' objects discovered in settlement contexts; and third, the deposition of objects in localities that may have been specifically constructed as ritual enclosures.

The archaeological remains that directly relate to early agriculture also corroborate the palynological evidence (illus.3.3). These varying remains indicate that both arable and pastoral farming was practised by the communities of later prehistory. Arable farming, for example, has been confirmed by the discovery of cord rig in areas that escaped the obliterating effects of medieval agriculture. In lowland Scotland it is thought to date from the Pre-Roman Iron Age (Topping 1989), and this seems confirmed in Dumfriesshire by virtue of the topographic position of cord rigs in association with Pre-Roman Iron Age settlement forms, notably ring-ditch houses (RCAHMS 1997). This latter association, of course, may also imply that a proportion of the recorded cord rig dates to the Early PreRoman Iron Age. Although sparsely distributed, this style of arable cultivation may once have been an extensive feature of the Pre-Roman Iron Age landscape: thus cord rig has been documented in Eskdale, Annandale, Nithsdale, and further west on the Kirkcudbrightshire coast (Maynard 1994; RCAHMS 1997).

Both metal and wooden objects, often in the form of agricultural implements, are sometimes found in natural locations which are inappropriate to their functional character. The natural context varies, but there seems to have been a preoccupation with localities that would prove inaccessible once the act of deposition had been undertaken. Invariably, these take the form of 'wet' locales such as, rivers and lochs, or peat bogs. The precise meaning behind this style of deposition is not clear, but it may relate to some form of ritual consumption, using objects that were considered to be - in some measure exotic (Barrett 1985; Barrett and Needham 1988; Bradley 1990). In south west Scotland this style of deposition, and the depositional rules that it employed, appears to have been initiated during the Bronze Age and to have continued into the Pre-Roman Iron Age (Gregory 1998). Initially, the scale of deposition was limited. For instance,

A strong economic reliance on pastoralism is also suggested by the archaeological record, particularly with the discovery oflinear boundary systems, detectable either as cropmarks or upstanding earthworks, which may represent the remnants of stock enclosures (Halliday 1982; RCAHMS 1997). The larger more complex systems form what may be loosely regarded as fields which, in some cases, are associated with what appear to be drove-ways

39

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

the only metalwork assemblage which may date to the Mid Pre-Roman Iron Age was a bronze pony cap associated with a pair of bronze drinking horns discovered at Torrs Farm (MacGregor 1976). This hoard, however, replicates many of the early depositional themes in that it was deposited in a 'kettle bog' - an inaccessible location - and was probably exotic to the region (cf Stevenson 1966). The deposition of the drinking horns, if this is what the bronze objects represent, may indicate a conceptual link with feasting and with ideas about fertility.

in an arena that was set apart from domestic life and which was, in all probability, communal in nature. Indeed, it indicates that manufacture, such as metalworking, was, perhaps, regarded as a special activity which bestowed a significant role upon those individuals who participated in it. The discovery of this site suggests the possibility of similar enclosures within the region, but without the unique conditions of preservation found at Overrig their identification will prove difficult. The recently excavated enclosure at Albie Hill, Annandale, however, may be one candidate. Its form is certainly unusual for a later prehistoric settlement (Strachan 1999). Although severely plough-damaged, concentrations of pits were uncovered, one of which was found to contain a Pre-Roman Iron Age sickle blade. This again bolsters the suggestion that deposition was connected, in some measure, with concepts of fertility (Strachan 1999). There was also evidence of iron working in the vicinity of the enclosure, suggesting parallels with Overrig. Again, metalworking was, perhaps, considered special, or exotic, and therefore required a manufacturing context which was set apart from everyday domestic life, and one which was imbued with wider cosmological concerns.

Although there is a problem with the independent dating of the metalwork (cf. Hunter 1997), these processes intensify during the Late Pre-Roman Iron Age. A variety of metal and wooden artefacts were deposited, and again many of the deposits may have been connected to concepts of fertility. The discovery of a La Tene sword associated with a wooden shoulder yoke from Lochar Moss (Truckell 1964), and the ard beam recovered from a bog near Lochmaben (Fenton 1968), are cases in point. Within settlement contexts, many of these themes, which were probably centred on the agricultural cycle, were duplicated. The most obvious is the Early/Mid Pre-Roman Iron Age ard and stilt, discovered as a foundation deposit beneath a crannog at Milton Loch (Piggott 1953; Guido 1974). The deposition of a large stone inscribed with a series of petrogylphs from the easterly orientated porch of a Late Pre-Roman Iron Age round house at Hayknowes Farm may again represent a foundation deposit (Gregory, in press). Here the precise meaning and significance of the stone and of the act of deposition is not known. The elaboration of many easterly orientated entrances with significant deposits during the Late Bronze/Pre-Roman Iron Age, however, has been equated with wider cosmological issues, which may, again, connect to the agricultural cycle (cf. Parker-Pearson 1996; Oswald 1997).

Summary

By combining the available evidence an interesting developmental sequence may be tentatively constructed. The Early/Mid Pre-Roman Iron Age landscape (c.700 200 cal. BC) appears to have been dominated by defensive sites, or forts, of moderate size. Other settlement types included unenclosed ring-ditch houses and probably a proportion of the palisaded settlements recorded in the area. Identifiable agriculture associated with many of these settlements is seen in the remains of cord rig, the survival of which is confined predominantly to the uplands. What this apparent pattern of settlement represents in terms of Iron Age social structure is problematic. For example, little is known concerning the number of people inhabiting these settlements. Nor is it clear what proportion of the settlement area was given over to other activities, such as the containment of stock, or why considerable effort was expended in the construction of settlement boundaries. Were the boundaries constructed for defence or display, or were they a means of isolating, and hence defining, a certain kin group? Perhaps they were constructed for a combination of reasons. What seems clear, however, is the theme of social isolation, which the form of the settlements implies. This isolation was broken, perhaps, when different kin groups assembled for the construction of the larger hilltop sites, such as Burnswark, which then became important for communal social interaction within a relatively isolated social landscape. In this sense, sites such as Burnswark may be viewed not solely in a defensive light, but as 'symbols of the community' (Hill 1995: 53).

The final location where ritual deposition has been noted is within a peculiar class of site, which is best described as a ritual or ceremonial enclosure. The Late Pre-Roman/Early Roman Iron Age enclosure at Overrig, in Eskdale, although not yet published, may represent one such example. The site is heavily vallated with three concentric ditches, around which a later rectilinear boundary was constructed (Mercer 1995). Within the interior of the enclosure, excavation has revealed two circular houses associated with a trapezoidal boulder setting of unknown function. The internal perimeter was defined by a constantly waterlogged ditch and a timber palisade which had no visible entrances (Mercer 1995). Significantly the enclosure was located in a 'natural amphitheatre', allowing activities within it to be viewed from the outside by an assembled audience. The precise nature of the activities conducted there is unknown, but excavation suggests that the manufacture of metal and wooden objects occurred within the confines of the enclosure and the deposition of unusual wooden objects and cremated bone occurred in the waterlogged inner ditch (Mercer 1995). The site, therefore, appears to amalgamate ritual with specialised manufacture

With the advent of the Late Pre-Roman Iron Age (c.200 cal. BC - cal. AD 100), the form of the social landscape was transformed. Population levels seem to have increased

40

Richard Gregory: Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland

and may in part have been responsible for the apparent change in settlement. Smaller, non-defensive settlements proliferated, and although many of the fort sites were abandoned or modified, some were retained, such as the large bivallate sites in the lowlands. It may also be significant that at many of these centres large systems of land division can be identified. This suggests the formation of a more structured social hierarchy where particular settlements rose to prominence, controlling substantial agricultural resources and perhaps procuring labour from less dominant households for construction and maintenance of the settlement and its associated holdings. It is difficult to identify the reasons behind this apparent transformation, but it would only require the acquisition and accumulation of a valuable resource, or resources, by one kin group, or household, to initiate a competitive chain-reaction throughout the later prehistoric landscape.

presumably lines of communication, were constructed throughout the region. Military sites which can be dated to this period with a moderate degree of confidence, include Milton (Clarke 1947; 1948; 1949; 1950; 1951; 1952a), Dalswinton (Richmond and St Joseph 1956), Glenlochar (Richmond and St Joseph 1952), Birrens (Birley 1938; Robertson 1975), Barnhill, Beattock (Maxwell and Wilson 1987), and Gatehouse-of-Fleet (St Joseph 1983). Other forts in the region which may also date to the Early Flavian period are those found at Drumlanrig (Maxwell and Wilson 1987) and Ladyward (RCAHMS 1997). To these sites may also be added a proportion of the undated temporary camps. These camps, which include those at Annanfoot, Annan Hill, Kirkpatrick-Flemming, Ward Law, Ruthwell, Middlebie, Torwood, Lochmaben, Broadlea, Fourmerkland, Trailflat, Carronbridge, Waterside Mains, East Gallaberry, and Hangingshaw, may represent the initial forays of Cerealis and/or Agricola into southern Scotland.

These valuable resources probably related, in part, to agriculture, which appears to increase exponentially during this period. Agriculture, however, was not merely a pragmatic economic activity relating to subsistence. It appears to have permeated many, if not all, of the belief systems of the Late Pre-Roman Iron Age in this area. Much of the intentional deposition which is a feature of this period was probably connected to issues specifically centred on the agricultural cycle. It may, therefore, be no coincidence that the increase in the deposition of particular types of artefacts, and the apparent appearance of ritual enclosures, corresponds to increases in clearance and agricultural activity in general. Indeed, concerns with the productivity of the land, and the well-being of stock and crops, presumably played a prominent societal role, particularly during a period when the climate may have been deteriorating (Baillie 1993).

Following the withdrawal of the Roman forces from northern Scotland, Roman occupation in the south west continued in a more rationalised form. Hence, between c. AD90-105 only the forts at Glenlochar, Dalswinton and Milton, and possibly Drumlanrig and Ladyward, appear to have housed Roman contingents. Blanket occupation was short-lived due to changing imperial policy, which resulted in the development of the Trajanic frontier across the Tyne-Solway isthmus. Although most areas in south west Scotland were now viewed as peripheral, interest and contact was still maintained with an area which embraced the northern Solway plain. A tombstone from Foligno, for example, bears the inscnpt10n CENSITO(RI) BRITTONUM ANAVION[EN(SIUM)], suggesting that this small region was being assessed in terms of taxation, if Anava can be identified with the River Annan (Rivet and Smith, 1979; Birley, this volume). With the later consolidation of the Tyne-Solway frontier, however, direct military contact in the area returned, and a Roman garrison was housed in the newly-constructed Hadrianic fort at Birrens (Birley 1938; Robertson 1975). With the Antonine re-occupation of south west Scotland - whether this was fragmented into two distinct periods, or not (cf Hodgson 1995) - garrisoning of the region became a priority. Forts were rebuilt at Birrens (Birley 1938; Robertson 1975), Glenlochar (Richmond and St. Joseph 1952), Drumlanrig (Maxwell and Wilson 1987), and possibly Ladyward. At Carzield a new fort was established (Birley and Richmond 1942) and a series of smaller fortlets were constructed at Burnswark (Jobey 1978), Milton (Clarke 1949; 1952), Barburgh Mill (Breeze 1974), Durisdeer (Clarke 1952b), and probably Murder Loch, Sanquhar and Lantonside (Maxwell and Wilson 1987). With the close of the Antonine period, identifying Roman military occupation in the area is more difficult. Some of the larger temporary camps, such as those found at Kirkpatrick Flemming and Lochmaben, may relate to Severan campaigning, but little more can be said regarding their dates or the dates of many of the other temporary camps in the region. If the evidence from Birley's (1938) excavation is accepted there is also a

Other activities which may have been important for the propagation of dominant kin groups included the manufacture of specialised goods, such as metalwork. This manufacture appears to have required a specialised locale, which was immersed within a wider ritual arena that may have served as a forum where dominance could be asserted through the control of esoteric knowledge. Those who controlled both the rituals and the manufacture may well have been the same individuals who first acquired power through agricultural accumulation.

The influence of Rome During the latter half of the first century AD the indigenous communities of south west Scotland were alerted to the presence of Rome. Initially this may have been by hearsay, but, by at least AD 80, direct contact was established via the Roman military (illus. 3.4). The evidence for Roman military expansion into south-west Scotland has been well documented and debated, and need not be replicated in detail here (cf Hanson and Maxwell 1983; Hanson 1987; RCAHMS 1997). In outline, however, initial contact appears to have occurred during the Early Flavian period, when a series of camps, forts, and

41

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

possibility that Birrens was not abandoned at the end of the Antonine period and that it continued in use throughout the Severan period, and perhaps even into the fourth century.

beyond the life of the camp (Johnston 1994). The establishment of the Hadrianic frontier may also have caused some disruption (Higham 1986: 178-9), particularly if Brigantian territory extended into the northern Solway plain. The influence of the frontier, however, is difficult to assess and relatively little is known of the Pre-Roman Iron Age landscape on the southern side of the Solway (Higham 1982).

What is clear from the history of Roman occupation and interest in the region is that it operated over both differing temporal and spatial scales. In most areas direct contact with the military was confined to relatively short periods of time. Initially these may have lasted between 15-20 years, only to be punctuated by a hiatus of at least 30 years. Occupation resumed briefly during the Antonine period, perhaps spanning the years c. AD 140-163. These shortterm interactions, nonetheless, placed certain demands on indigenous society. The construction of military camps, forts, and lines of communications must have impacted upon indigenous claims of tenure, and on indigenous resources. In some instances they may have resulted in the actual displacement of indigenous people.

The other stresses, which were placed on the indigenes during these periods of direct contact, probably centred on issues of military logistics. The resources required to construct camps, forts, roads and associated structures were substantial, particularly in respect to timber (cf. Hanson 1978; Breeze 1984; 1989). There is, however, surprisingly little evidence in the pollen record of changes in patterns of clearance throughout the Roman Iron Age (Tipping 1997b). This ambiguity also extends to another aspect of military demand: the provisioning of the entrenched forces. As Breeze (1989) notes, the military attempted to obtain food locally. This, he suggests, led to an expansion in agriculture, indicated to some extent by the presence of Roman goods on 'native farmsteads'. In general, however, the evidence for agricultural expansion consequent upon the arrival of the Romans is weak (Breeze 1989), and the recovery of Roman artefacts and coins from indigenous settlement sites is rare (Macdonald 1918; 1924; 1934; 1939; Robertson 1950; 1961; 1970; 1971). It is possible, of course, that a proportion of military supplies, such as grain, was derived from south of the Solway, but this sits uneasily with other evidence indicating that the indigenous economy had reached a level which was readily capable of supporting transient occupation by Roman forces. These forces, perhaps, procured supplies through taxation in kind which may have been directed particularly to the livestock which was one of the most valuable PreRoman Iron Age resources in north west Britain (cf. Higham, 1989).

Using the archaeological record alone, it is difficult to ascertain the precise details surrounding these periods of possible conflict, but the available evidence suggests that the Roman military exhibited a range of differing attitudes towards the indigenes. At times, these attitudes may have involved some form of hostility if, for example, the indigenes were perceived as a threat, or if indigenous settlement occupied a location that was considered tactically important. In turn, hostile acts may ultimately have led to the forced relocation of indigenous settlement units. At a number of sites, although the precise contemporaneity of indigenous settlement and adjacent Roman fortification is not clear, this scenario seems likely. At Milton, for example, aerial photography has located an indigenous 'rectilinear' settlement - a type typical of the last centuries BC/early centuries AD - within the confines of the Roman fort (RCAHMS 1997). It is therefore conceivable that this settlement may have been destroyed during the construction of the Roman fortifications. Similarly, at Ward Law a ditch with associated titulus appears to link a Roman camp with a curvilinear settlement (Jones 1979b). The reasons for this, apparently unique, arrangement is not clear, but it suggests that the settlement was in some way incorporated into the workings of the camp. Again, if the occupation of the settlement was contemporary with this fortification, its curious juxtaposition probably resulted in the removal of its inhabitants. At Hayknowes a series of indigenous settlements and an extensive system of agricultural boundaries are found in close proximity to the Roman camp at Annanfoot. Excavation at the earliest of these settlements suggests that its abandonment dates to the late first century AD, and it is tempting to connect this abandonment with the establishment of the nearby Roman camp (Gregory in press). At other sites, however, there seems to have been no dislocation of indigenous peoples, to whom the Romans were, perhaps, indifferent. Thus, at Carronbridge, although a Roman camp and road were built directly adjacent to an indigenous rectilinear settlement, the settlement continued to thrive and was occupied well

These dislocations and appropriations must have been traumatic for the individuals involved. It is debatable, however, whether such short-lived interference episodes had any lasting effect on the fabric of indigenous society. The evidence suggests that the basic structure of PreRoman Iron Age life was retained, with little discernible change in settlement or economy even in eastern Dumfriesshire, where direct occupation, or at least interest, was sustained, from the Flavian through to the Antonine period (RCAHMS 1997). The political structure which had developed during the Pre-Roman Iron Age was also retained. Internal hierarchies continued and appear to have become more entrenched, as in the case of Castle O'er and its environs (Mercer 1995; RCAHMS 1997). This large site, which was occupied throughout the Roman Iron Age, was linked to a series of smaller settlement units and a large system of land divisions in some phases of its history. It may also have been intimately connected to the ritual enclosure at Overrig (Mercer 1995; RCAHMS 1997).

42

Richard Gregory: Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland

When set within a broader historical context, increased hierarchical control, although corresponding with the arrival of Rome, may represent a continued indigenous progression which was approaching its zenith after centuries of internal propagation during the Pre-Roman Iron Age. Of course, it is probable that many of these burgeoning seats of power were forced into some form of treaty obligations with Rome, and it is also possible that the Romans exacted taxation from them. Other forms of direct interaction, though, seem limited. There is no evidence, for example, for the presence of a vicus in the area, even though at Birrens the longevity of occupation could have led to one developing. It may be that the fort annexe acted as an arena for interaction and, in a sense, functioned as a 'proto-vicus'. As Robertson (1970: 202) notes, 'fort annexes were probably the main primary source of supply for the Roman material which found its way on non-Roman sites'. Correspondingly, the discovery of bronze terrets at Birrens may also imply the trade of indigenous metalwork to the Roman troops (Allason-Jones 1989). These piecemeal interactions, however, betray an overriding sense of caution on the part of the indigenes. The local people, while accepting and using Roman material culture, appear to have been reluctant to enter into close, or binding, economic and cultural relationships with the Roman military.

traditional ways of life. In fact, internal society had been gradually transforming over many centuries into a landscape dominated by hierarchical and agrarian values. Inevitably, such transformations may have resulted in an undercurrent of unease, manifest in recourses to tradition. The arrival of the Roman military must have encouraged these feelings. It may, therefore, be no coincidence that metalwork deposition increases at the beginning of the Roman Iron Age, with possible votive deposits including the Carlingwark Loch hoard, the Middlebie hoard, the Balmaclellan hoard and the Birrenswark Bit (MacGregor 1976). Similarly, it was during this period when evidence emerges of a more explicit link with the past, witnessed in the apparent re-use of traditional monuments. At Picts Knowe, a Neolithic henge was re-used and re-structured. The ditch was re-cut, after which a series of wooden objects, dating to the Roman Iron Age, were deposited at the base of the refashioned feature (Thomas 1998; Thomas, pers com). If the presence of Rome had any significant impact, therefore, it was probably in this strengthening of traditional, native, identity.

ACKNOWLEDGEMENTS This contribution has benefited greatly from the help and assistance of Dr ME Robinson who read early drafts of the manuscript. I would also thank Mr G Bowden who kindly prepared the figures.

In a similar manner, the use of Roman objects does not appear to imply any form of Romanization, in a traditional sense, but ironically the very opposite. Objects were used in familiar contexts where, perhaps, their meanings could become transformed. The use of objects in the home would be an obvious example of this, though the available evidence suggests that it was contexts connected to a wider religious or cosmological ethos which were more significant in this process. The large metalwork hoard deposited in Carlingwark Loch is a case in point (Curle 1933; Piggott 1953). In this example, although both Roman and Native objects are found in association, the rules and underlying themes of earlier Pre-Roman Iron Age deposition are retained (cf. Hunter 1997). The same may also be true of the Roman artefacts recovered from the ritual enclosure at Overrig (Mercer 1995). In these examples, the fact that the material was culturally Roman may have been of little significance. Its real importance may have been due to the fact that it was simply different. Thus, during the Pre-Roman Iron Age, it was items which were exotic, either in terms of their technological novelty or place of manufacture, that were significant to particular acts of deposition. In just the same way, perhaps, during the Roman Iron Age, items of Roman material culture may simply have added another element of exotica, which were then incorporated into traditional rituals and ceremonies (c£ Hunter 1997). As with the Pre-Roman Iron Age, it is also possible that it was the individuals who had access to these elements of exotica who also controlled other aspects of political and religious life.

BIBLIOGRAPHY Allason-Jones, L., 1989. Introductory remarks on the nature of Roman trade in the North of Britain. In van Oriel-Murray (ed.): 13-24. Baillie, M. G. L., 1993. Dark Ages and Dendrochronology. Emania 11: 5-12. Barber, J. W. and Crone, B. A., 1993. Crannogs: a diminishing resource ? A survey of the crannogs of southwest Scotland and excavations at Buiston Crannog. Antiquity 67: 520-33. Barrett, J. C., 1985. Hoards and related metalwork. In Clarke et al. (eds.): 95-106. Barrett, J. C., 1997. Romanization: a critical comment. In Mattingly (ed.): 51-64. Barrett, J. C. and Kinnes, I. A. (eds.), 1988. The Archaeology of context in the Neolithic and Bronze Age: Recent trends. Sheffield: Department of Archaeology and Prehistory, University of Sheffield. Barrett, J. C. and Needham, S. P., 1988. Production, circulation and exchange. In Barrett and Kinnes (eds.): 127-40. Barrrett, J. C., Fitzpatrick, A. P. and Macinnes, L. (eds.), 1989. Barbarians and Romans in North-West Europe: from the later Republic to late Antiquity. Oxford: British Archaeological Reports International Series 471. Birks, H. H., 1972. Studies in the vegetational history of Scotland II: Two pollen diagrams from the Galloway Hills, Kirkcudbrightshire. Journal of Ecology 60: 183217.

These limited examples imply a general reluctance on the part of the indigenes towards any radical alterations of

43

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Antiquarian Society, Third Series, 26 (1947-8): 13349. Clarke, J., 1950. The Roman forts at Milton (Tassieholm), near Beattock, Dumfriesshire. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series, 27 (1948-9): 197201. Clarke, J., 1951. Excavations at Milton (Tassieholm) in season 1950. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series, 28 (1949-50): 199-221. Clarke, J., 1952a. Milton (Tassieholm). In Miller (ed.): 104-10. Clarke, J., 1952b. Durisdeer. In Miller (ed.): 126. Close-Brooks, J., 1987. Comment on Traprain Law. Scottish Archaeological Review 4: 92-4. Crone, B. A., 1993. Crannogs and chronologies. Proceedings of the Society of Antiquaries of Scotland 123: 245-54. Curle, J., 1913. Roman and Native Remains in Caledonia. Journal of Roman Studies 3: 99-115. Curle, J., 1932a. Roman Drift in Caledonia. Journal of Roman Studies 22: 73-77. Curle, J., 1932b. An Inventory of Objects of Roman and Provincial Origin Found on Sites in Scotland not definitely Associated with Roman Construction. Proceedings of the Society of Antiquaries of Scotland 66: 277-397. Dumayne, L. and Barber, K. E., 1994. The impact of the Romans on the environment of northern England: pollen data from three sites close to Hadrian's Wall. The Holocene, 4: 2 (1994): 165-73. Feachem, R.W., 1961. Unenclosed platform settlements. Proceedings of the Society of Antiquaries of Scotland 93 (1960-1): 79-85. Fenton, A., 1968. Plough and Spade in Dumfries and Galloway. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 45: 147-88. Fulford, M. G., 1985. Roman Material in Barbarian Society c 200 BC - c AD 400. In Champion and Megaw (eds.): 91-108. Fulford, M. G., 1989. Roman and Barbarian: The Economy of Roman Frontier Systems. In Barrrett et al. (eds.): 81-95. Gates, T., 1983. Unenclosed settlements in Northumberland. In Chapman and Mytum (eds.): 10348. Gregory, R. A., 1998. Prehistoric landscapes in Dumfries and Galloway. Unpublished Ph.D Dissertation: University of Manchester. Gregory, R. A., 2001, in press. Excavation at Hayknowes Farm, Annan. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series, 75. Guido, M., 1974. A Scottish crannog redated. Antiquity 48: 54-6. Gwilt, A, and Haselgrove, C. (eds.), 1997. Reconstructing Iron Age Societies. Oxford: Oxbow Monograph 71.

Birley, E., 1938. Excavations at Birrens, 1936-1937. Proceedings of the Society of Antiquaries of Scotland 72 (1937-8): 275-347. Birley, E., 1948. Dumfriesshire in Roman Times. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series, 25 (1946-7): 132-50. Birley, E., 1954. Some Military Aspects of Roman Scotland. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series, 31 (1952-3): 9-21. Bradley, R., 1990. The Passage of Arms: An archaeological analysis of prehistoric hoards and votive deposits. Cambridge: Cambridge University Press. Braund, D., 1984. Rome and the Friendly King: The Character of Client Kingship. London: Croom Helm Breeze, D. J., 1974. The Roman fortlet at Barburgh Mill, Dumfriesshire. Britannia 5: 130-62. Breeze, D. J., 1984. Demand and supply on the Northern Frontier. In Miket and Burgess (eds.): 264-86. Breeze, D. J., 1989. The Impact of the Roman Army on North Britain. In Barrrett et al. (eds.): 227-34. Breeze, D. J., 1990. The Impact of the Roman Army on the Native Peoples of North Britain. In Vetters and Kandler (eds.): 85-97. Breeze, D. J. and Dobson, B., 1987. Hadrian's Wall (3rd edition). Harmonsworth: Penguin. Burnham, B. C. and Johnson, H.B. (eds.), 1979. Invasion and Response: The case of Roman Britain. Oxford: British Archaeological Reports, British Series 73. Champion, T. G. and Megaw, J. V. S. (eds.), 1985. Settlement and Society: aspects of West European prehistory in the first millennium BC. Leicester: Leicester University Press. Champion, T. C. and Collis, J. R. (eds.), 1996. The Iron Age in Britain and Ireland: Recent Trends. Sheffield: J.R.Collis Publications. Chapman, J.C. and Mytum, H. C. (eds.), 1983. Settlement in North Britain J000BC-ADJ000. Oxford: British Archaeological Reports, British Series 118. Clack, P. and Haselgrove, S. (eds.), 1982. Rural settlement in the Roman North. Durham: Council for British Archaeology, Group 3. Clarke, D. V., Cowie, T. G. and Foxon, A. (eds.), 1985. Symbols of Power at the time of Stonehenge. Edinburgh: HMSO. Clarke, J., 1947. Report on excavation at Tassieholm (Milton), Beattock, 1946. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series, 24 (1945-6): 10010. Clarke, J., 1948. The forts at Milton, Beattock (Tassieholm). Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series, 25 (1946- 7): 10-26. Clarke, J., 1949. Excavations at Milton. Transactions of the Dumfriesshire and Galloway Natural History and

44

Richard Gregory: Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland

Hill, J. D. and Cumberpatch, C. C. (eds.), 1995. Different Iron Ages. Oxford: British Archaeological Reports International Series 602. Hill, P. H., 1982a. Towards a new classification of prehistoric houses. Scottish Archaeological Review 1: 24-31. Hill, P. H., 1982b. Settlement and Chronology. In Harding (ed.): 4-43. Hill, P. H., 1987. Traprain Law: the Votadini and the Romans. Scottish Archaeological Review 4: 85-91. Ringley, R., 1996. The 'legacy' of Rome: the rise, decline, and fall of the theory of Romanization. In Webster and Cooper (eds.): 35-48. Ringley, R., 1997. Resistance and domination: social change in Roman Britain. In Mattingly (ed.): 81-100. Hodgson, N., 1995. Were there two Antonine occupations of Scotland? Britannia 26: 29-49. Hunter, F., 1997. Iron Age hoarding in Scotland and northern England. In Gwilt and Haselgrove (eds.): 108-33. Jobey, G., 1962. An Iron Age housestead at West Brandon, Durham. Archaeologia Aelianna, 4th Series 40: 1-34. Jobey, G., 1971. Early settlements in Eastern Dumfriesshire. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 48: 78-105. Jobey, G., 1973. A Romano-British settlement at Tower Knowe, Wellhaugh, Northumberland. Archaeologia Aelianna, 5th Series 1: 55-79. Jobey, G., 1974. Excavations at Boonies, Westerkirk and the nature of Romano-British settlement in Eastern Dumfriesshire. Proceedings of the Society of Antiquaries of Scotland, 105 (1972-4): 119-40. Jobey, G., 1976. Traprain Law: a summary. In Harding (ed.): 191-204. Jobey, G., 1977. Iron age and later farmsteads on Belling Law, Northumberland. Archaeologia Aelianna, 5th Series 5: 1-38. Jobey, G., 1978a. Burnswark Hill, Dumfriesshire. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 53: 57-105. Jobey, G., 1978b. Iron Age and Romano-British settlements on Kennel Hall Knowe, North Tynedale, Northumberland. Archaeologial Aelianna, 5th Series 6: 3-28. Jobey, G., 1984. A settlement on Boggle Hill, Thorneyburn. Archaeologia Aelianna, 5th Series 12: 241-2. Jobey, G., 1985. The unenclosed settlements of the TyneForth: a summary. In Spratt and Burgess (eds.): 17794. Jobey, I. and Jobey, G., 1988. Gowanburn River Camp: An Iron Age, Romano-British and more recent settlement in North Tynedale, Northumberland. Archaeologia Aelianna, 5th Series 16: 11-28. Johnston, D. A., 1994. Carronbridge, Dumfries and Galloway: the excavation of Bronze Age cremations, Iron Age settlements and a Roman Camp. Proceedings of the Society of Antiquaries of Scotland 124: 233-91.

Haggarty, A. and Haggarty, G., 1983. Excavations at Rispain Camp, Whithorn 1978-1981. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 58: 21-51. Halliday, S. P., 1982. Later prehistoric farming in southeast Scotland. In Harding (ed.): 74-91. Hanson, W. S., 1978. The organisation of the Roman military timber supply. Britannia 9: 293-305. Hanson, W. S., 1987. Agricola and the Conquest of the North. London: Batsford. Hanson, W. S., 1989. The Nature and Function of Roman Frontiers. In Barrrett et al. (eds.): 55-63. Hanson, W. S., 1997. Forces of change and methods of control. In Mattingly (ed.): 67-80. Hanson, W. S. and Maxwell, G. S., 1983. Rome's North West Frontier: The Antonine Wall. Edinburgh: Edinburgh University Press. Hanson, W. S. and Keppie, L. J. F. (eds.), 1980. Roman Frontier Studies 1979. Oxford: British Archaeological Reports International Series 71 : i. Harding, D. W. (ed.), 1976. Hillforts: later prehistoric earthworks in Britain and Ireland. London: Academic Press. Harding, D. W. (ed.), 1982. Later prehistoric Settlement in South-East Scotland. Edinburgh: University of Edinburgh Occasional Paper No.8. Hartley, B. R. and Wacher, J. (eds.), 1983. Rome and her northern provinces. Gloucester: Alan Sutton. Haverfield, F., 1905. The Romanization of Britain. Proceedings of the British Academy 2: 185-217. Higham, N. J. (ed.), 1979. The Changing Past: some recent work in the archaeology of northern England. Manchester: University of Manchester Department of Extra-Mural Studies. Higham, N. J., 1982. The Roman impact upon rural settlement in Cumbria. In Clack and Haselgrove (eds.): 105-22. Higham, N. J., 1978. Early Field Survival in North Cumbria. In Early Land Allotment in the British Isles: A Survey of Recent Work, edd. H. C. Bowen and P. J. Fowler. Oxford: British Archaeological Reports, British Series 48: 119-25. Higham, N. J., 1986. The Northern Counties to ADJ000. Harlow: Longman. Higham, N. J., 1989. Roman and Native in England north of the Tees: acculturation and its limitations. In Barrrett et al. (eds.): 153-74. Higham, N. J. and Jones, G. D. B., 1975. Frontier, forts and farmers. Cumbria Aerial Survey 1974-75. Archaeological Journal 132: 16-53. Higham, N. J. and Jones, G. D. B. 1983. The excavations of two Romano-British farm sites in North Cumbria. Britannia 14: 45-72. Higham, N. J. and Jones, G. D. B., 1985. The Carvetti. Stroud: Alan Sutton. Hill, J. D., 1995. How Should We Understand Iron Age Societies and Hillforts? A Contextual Study from Southern Britain. In Hill and Cumberpatch (eds.): 4566.

45

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Jones, G. D. B., 1979a. Invasion and Response in Roman Britain. In Burnham and Johnson (eds.): 57-70. Jones, G. D. B., 1979b. The future of aerial photography in the north. In Higham (ed.): 75-89. Jones, G. D. B. and Walker, J., 1983. Either side of the Solway: towards a minimalist view of Romano-British agricultural settlement in the North-West. In Chapman and Mytum (eds): 185-204. Jones, V. J., Stevenson, A. C. and Batterbee, R. W., 1989. Acidification of lakes in Galloway, south west Scotland: a diatom and pollen study of the post-glacial history of the Round Loch of Glenhead. Journal of Ecology, 77 (1989): 1-23. Keppie, L. J. F., 1989. Beyond the Northern Frontier: Roman and Native in Scotland. In Todd (ed.): 61-73. Macdonald, G., 1918. Roman coins found in Scotland. Proceedings of the Society of Antiquaries of Scotland 52: 203-76. Macdonald, G., 1923. The Romans in Dumfriesshire. Transactions of the Dunifriesshire and Galloway Natural History and Antiquarian Society, Third Series, 8 (1920-1): 68-100. Macdonald, G., 1924. Roman coins found in Scotland. Proceedings of the Society of Antiquaries of Scotland 58: 325-29. Macdonald, G., 1934. Roman coins found in Scotland including a hoard from Falkirk. Proceedings of the Society of Antiquaries of Scotland 68: 27-40. Macdonald, G., 1939. Roman coins found in Scotland IV. Proceedings of the Society of Antiquaries of Scotland 73: 241-5. Macdonald, J. and Barbour, J., 1897. Birrens and its Antiquities, with account of recent excavations and their results. Dumfries: Standard Office. MacGregor, M., 1976. Early Celtic Art in Northern Britian. Leicester: Leicester University Press. Macinnes, L., 1989. Baubles, bangles and beads: trade and exchange in Roman Scotland. In Barrrett et al. (eds.): 108-16. Mann, J. C. and Breeze, D. J., 1987. Ptolemy, Tacitus and the tribes of north of Britain. Proceedings of the Society of Antiquaries of Scotland 117: 85-91. Mattingly, D. J. (ed.), 1997. Dialogues in Roman Imperialism: power, discourse, and discrepant experience in the Roman Empire. Portsmouth: Journal of Roman Archaeology Supplementary Series 23 ): 5164. Mattingly, D. J., 1997. Dialogues of power and experience in the Roman Empire. In Mattingly (ed.): 7-24. Maxwell, G. S., 1980. The native background to the Roman occupation of Scotland. In Hanson and Keppie (eds.): 1-14. Maxwell, G. S. and Wilson, D. R., 1987. Air Reconnaissance in Roman Britain 1977-84. Britannia 18: 1-48. Maynard, D., 1994. Archaeological discoveries in the Dune System at Brighouse Bay. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 69: 13-34.

Mercer, R., 1981. The excavation of an earthwork enclosure at Long Knowe, Eskdale, Dumfriesshire 1976. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 56: 38-72. Mercer, R., 1995. Field guide for Castle O'er public walk. Privately Circulated Manuscript. Mercer, R., 1997. Kirkpatrick Flemming Dunifriesshire: An anatomy of a parish in South West Scotland. Dumfries: Dumfries and Galloway Natural history and Antiquarian Society Ann Hill Bequest 6. Miket, R. and Burgess, C. B. (eds.), 1984. Between and Beyond the Walls: essays on the prehistory and history of North Britain in Honour of George Jobey. Edinburgh: John Donald. Miller, S. N. (ed.), 1952. The Roman Occupation ofSouthWestern Scotland. Glasgow: Glasgow University Publication Series 83. Millett, M., 1990. The Romanization of Britain. Cambridge: Cambridge University Press. Morrison, I., 1985. Landscape with lake dwellings: the crannogs of Scotland. Edinburgh: Edinburgh University Press. Munro, R., 1882. Ancient Scottish lake dwellings. Edinburgh: Douglas. Nichols, H., 1967. Vegetation change, shoreline displacement and the human factor in the Late Quaternary history of south-west Scotland. Transactions of the Royal Society of Edinburgh, 67 (1967): 145-87. Oswald, A., 1997. A doorway on the past: practical and mystic concerns in the orientation of roundhouse doorways. In Gwilt and Haselgrove (eds.): 87-95. Owen, 0. A., 1992. Eildon Hill North, Roxburgh, Borders. In Rideout et al. ( eds.): 21-72. Parker-Pearson, M., 1996. Food, fertility and front doors in the first millennium BC. In Champion and Collis (eds.): 117-32. Piggott, C. M., 1948. Excavations at Hownam Rings, Roxburghshire. Proceedings of the Society of Antiquaries of Scotland 82 (1947-8): 193-225. Piggott, C. M., 1949. The iron age settlement at Hayhope Knowe, Roxburghshire. Proceedings of the Society of Antiquaries of Scotland 83 (1948-9): 45-67. Piggott, C. M., 1953. Milton Loch Crannog 1: a native house of the 2nd century Ad in Kirkcudbrightshire. Proceedings of the Society of Antiquaries of Scotland 87 (1952-3): 134-52. Piggott, S., 1958. Native Economies and the Roman Occupation of Britain. In Richmond (ed): 1-27. Richmond, I. A. (ed.), 1958. Roman and Native in North Britain. Edinburgh: Nelson. RCAHMS: Royal Commission on the Ancient and Historical Monuments of Scotland, 1914. Inventory of Ancient Monuments of the Stewartry of Kirkcudbrightshire. Edinburgh. RCAHMS: Royal Commission on the Ancient and Historical Monuments of Scotland, 1920. Inventory of Ancient Monuments of Dumfriesshire. Edinburgh.

46

Richard Gregory: Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland

RCAHMS: Royal Commission on the Ancient and Historical Monuments of Scotland, 1994. Glenesslin Nithsdale: An Archaeological Survey. Edinburgh. RCAHMS: Royal Commission on the Ancient and Historical Monuments of Scotland, 1997. Eastern Dumfriesshire. Edinburgh. Raisen, P. and Rees, T., 1995. Excavations of three cropmark sites at Melville Nurseries, Dalkeith. Glasgow Archaeological Journal 19: 31-50. Rapson, S., 1994. Pollen analysis at Brighouse Bay. Transactions of the Dunifriesshire and Galloway Natural History and Antiquarian Society, Third Series 69 (1994): 25-33. Richmond, I. A. and St Joseph, J. K. S., 1952. The Roman Fort at Glenlochar. Transactions of the Dunifriesshire and Galloway Natural History and Antiquarian Society, Third Series 30 (1950-2): 1-16. Richmond, I. A. and St Joseph, J. K. S., 1956. The Roman Fort at Dalswinton. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 34 (1955-6): 1-21. Rideout, J. S., 1996. Excavation of a promontory fort and a palisaded homestead at Lower Greenyards, Bannockburn, Stirling, 1982-5. Proceedings of the Society of Antiquaries of Scotland 126: 199-269. Rideout, J. S., Owen, 0. A. and Halpin, E., 1992. Hillforts of Southern Scotland. Edinburgh: AOC (Scotland) Ltd., Monograph 1. Rivet, A. L. F. (ed.), 1966. The Iron Age in Northern Britain. Edinburgh: Edinburgh University Press. Rivet, A. L. F. and Smith, C., 1979. The Placenames of Roman Britain. London: Batsford. Robertson, A. S., 1950. Roman coins found in Scotland (1939-50). Proceedings of the Society of Antiquaries of Scotland 84; 137-69. Robertson, A. S., 1961. Roman coins found in Scotland 1951-60. Proceedings of the Society of Antiquaries of Scotland 94: 133-83. Robertson, A. S., 1970. Roman Finds from non-Roman Sites in Scotland. More Roman 'Drift' in Caledonia. Britannia 1: 198-226. Robertson, A. S., 1971. Roman coins found in Scotland 1961-70. Proceedings of the Society of Antiquaries of Scotland 103: 113-68. Robertson, A. S., 1975. Birrens (Blatobulgium). Edinburgh: Dumfries and Galloway Natural History and Antiquarian Society. St Joseph, J. K. S., 1983. The Roman fortlet at Gatehouseof-Fleet, Kirkcudbrightshire. In Hartley and Wacher (eds.): 222-34. Scott-Elliot, J., Simpson, D. D. A. and Coles, J. M., 1966. Excavations at McNaughton's Fort, Kirkcudbright. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 43: 73-9. Spratt, D. and Burgess, C. B. (eds.), 1985. Upland Settlement in Britain: The second millennium BC and after. Oxford: British Archaeological Reports, British Series 143.

Stevenson, R. B. K., 1966. Metal-work and some other Objects in Scotland and their Cultural Affinities. In Rivet (ed.): 17-44. Strachan, R. J., 1999. Excavations at Albie Hill, Applegarthtown: Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 73: 9-16. Terry, J., 1993. Excavations of a farmstead enclosure, Uppercleugh in Annandale, Dumfries and Galloway. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 68: 53-86. Thomas, J. S., 1998. Pict's Knowe, Holywood and Holm. Current Archaeology 160: 149-60. Tipping, R., 1994. The form and fate of Scotlands's woodlands. Proceedings of the Society of Antiquaries of Scotland 124: 1-54. Tipping, R., 1995. Holocene evolution of a lowland Scottish landscape: Kirkpatrick Flemming. IL Regional vegetation and land-use change. The Holocene 5 (1): 83-96. Tipping, R., 1997a. Holocene evolution of a lowland Scottish landscape. In Mercer (ed.): 27-35. Tipping, R., 1997b. Pollen analysis and the impact of Rome on native agriculture around Hadrain's Wall. In Gwilt and Haselgrove (eds.): 239-247. Todd, M. (ed.), 1989. Research on Roman Britain 196089. London: Britannia monograph series 11. Topping, P., 1989. Early Cultivation in Northumberland and The Borders. Proceedings of the Prehistoric Society 55: 161-79. Truckell, A. E., 1964. The Archaeological collections of the Society. Transactions of the Dumfriesshire and Galloway Natural History and Antiquarian Society, Third Series 41 (1962-3): 55-66. van Oriel-Murray, C. (ed.), 1989. Roman Military Equipment: the Sources of Evidence. Oxford: British Archaeological Reports International Series 476. Vetters, H. and Kandler, M. (eds.), 1990. Aktendes 14 Internationalen Limeskongresses 1986 in Carnuntum. Wien: Osterreichische Akademie der Wissenschaften. Webster, J., 1996. Roman imperialism and the post imperial age. In Webster and Cooper (eds.): 1-17. Webster, J. and Cooper, N. J. (eds.), 1996. Roman Imperialism: Post-Colonial Perspectives. Leicester: School of Archaeological Studies, University of Leicester, Leicester Archaeological Monographs 3. Wilson, A., 1998. Roman Penetration in Strathclyde South of the Antonine Wall. Glasgow Archaeological Journal 20 (1996-97): 1-40. Whittaker, C. R., 1989. Supplying the System: Frontiers and Beyond. In Barnett et al. (eds.): 64-80.

47

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

0

A Palisade 0 Unenclosed settlement

3 .1 Unenclosed and palisaded settlements

0

Curvilinear single boundary t:. 0-4.99 ha .t. 4.99 ha+

Curvilinear multiple boundary (max. internal area) 0 0-4.99 ha • 4.99 ha+ Rectilinear Ill 0·4.99 ha ■ 4.99 ha+

3.2 Pre-Roman/Roman enclosed settlement

48

Richard Gregory: Living on the Frontier: Iron Age - Roman Transitions in South-West Scotland



Linearbounda~es

OCord

rig

3.3 Pre-RomanIronAge agriculturalremains

Dalswinton Glenlochar Birrens Barnhill. Beattock Gatehouse-of-Fleet Drumlanrig Ladyward Annan Foot Annan Hill

3.4 Roman military sites

49

11 12 13. 14 Middlebie 15 Torwood 16. Lochmaben 17. Broadlea 18. Fourmerkland 19. Trailflat 20. Carronbridge

24. 25. Burnswark

26. 27 28. 29. 30.

Barburgh Mill Durisdcer Murder Loch Sanquhar Lantonside

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

50

4

The relative information value of aerial photography and geophysical survey: a case study from the Clyde valley W. S. Hanson 1 and Loma Sharpe geochemical analysis of the formation of cropmarks. Both are now in their final phases.

Aims It is a commonplace that different methods of non-invasive

survey provide potentially different information about sites; hence the inclusion in any regional survey of a range of approaches. This paper sets out to quantify, using one particular case study, the different planimetric information obtainable from oblique aerial photography, resistivity and magnetometry survey.

Burnfoot Farm 2

The two main enclosures at Burnfoot Farm, Libberton (NGR NS991405) were first recorded from the air for Cambridge University Committee for Aerial Photography (henceforth CUCAP) as cropmarks in 1975. They were photographed subsequently several times by both RCAHMS and by the first named author between 1980 and 1995, though not always with the same quality of result or level of information recovery. Nonetheless, the site has been monitored from the air for some 25 years and the photograph taken in 1989 (illus. 4.2), when the field was under a barley crop, represents one of the occasions when visibility of the archaeological features was at its optimum. Others include CUCAP photography in 1975 (BVK 83) and RCAHMS photography in 1995 (C53 l 03).

The Clyde Valley Project

The first named author began aerial survey in the West of Scotland in the early 1980s with modest funding from the University of Glasgow and latterly from the RCAHMS. The original aim of the work was to address some of the issues of bias in the collection of aerial photographic data in Scotland by more regular flying in the western lowlands (c£ Hanson and Macinnes 1991: 157). Reconnaissance has continued thereafter more sporadically, linked to the varying availability of a suitable aircraft and pilot, and to the vagaries of weather conditions in the area. In 1995 a 5year programme was begun, funded by Historic Scotland, with the aim of elucidating the development of the landscape of the upper Clyde valley over time. This programme has sought to place archaeological sites discovered from the air into their wider context and assess the level of threat to elements of that landscape, particularly from agricultural activity.

The site complex lies in what is currently a pasture field to the north of and immediately adjacent to the present farm. Reaching down from the high ground in the east, the field lies on a terrace on the east bank of the River Clyde. The solid geology underlying this terrace is composed of unclassified andesite and basalt, overlain by boulder clay derived from around the nearby Southern Uplands Fault to give a sandy-clay subsoil. Although it appears flat in the aerial photographs, the field in fact undergoes fairly substantial topographic changes. Over half of the field is occupied by a series of three rounded undulations of fluvioglacial origin. The two large enclosures are situated upon two of these higher points, in what appears to have been a deliberate exploitation of the natural features. From here the ground falls away, steeply in places along the line of a geological fault between the andesite and basalt and the Old Red Sandstone deposits, to the river flood plain to the north-west. The current course of the river is only some 40m away from the western boundary of the field.

Between Biggar and Lanark, the Clyde begins to meander across an area of a lower ground widening the valley (illus. 4.1 ). The area is good arable land, rich in archaeological remains, both artefacts and sites. It represents, for example, one of the greatest concentrations of Neolithic axes in Scotland and one of the densest distributions of archaeological sites of all periods in the western lowlands. Many of these sites have been discovered from the air since the publication of the Royal Commission Inventory of Prehistoric and Roman Monuments in Lanarkshire (RCAHMS 1978), and intensive agricultural activity continues to erode their remains. Historically the valley is an important nodal point in the communications network of lowland Scotland. Roman roads cross it to the north and south and an important railway junction is situated at Carstairs. Nonetheless, the valley retains its essentially rural character.

Geophysical Survey Methodology

Two weeks of survey were conducted at the site by undergraduate students from the Department of 1 The first named author was introduced to aerial reconnaissance by Barri Jones while undertaking postgraduate research at Manchester under his supervision. He experienced his first solo flight at Barri's instigation and rapidly discovered how easy it is to lose track of exactly where you are when flying in a light aircraft. This paper is offered in memory of Barri and serves to demonstrate how the seed he sowed has grown and that the propagation process continues. 2 We are grateful to the farmer, Mr. David Russell, for ready access to his land in order to undertake the geophysical surveys.

The second named author was employed on a part-time basis to undertake the aerial photographic transcriptions and geophysical and other fieldwork for the project. The other half of her time was devoted to a PhD on the

51

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Archaeology, University of Glasgow, as a supervised training survey in March 1999 and 2000, involving both resistivity and magnetic survey. The resistivity survey was carried out with RM15 Resistivity meters, and twin electrode configurations with the mobile electrodes set 0.5m apart. The magnetic survey was carried out using FM36 fluxgate gradiometers. Data were dumped into the Geoplot 2.0 data processing package, and ultimately processed using Geoplot 3. All instruments and data processing packages used during the survey were developed by Geoscan Research Ltd.

approximately circular in shape, c.90m in diameter, defined by two widely-spaced ditches (enclosure 2) with possible entrances to both east and west. The first gives the impression of being later in date, as its northern ditch is the better defined where it coincides with the second enclosure. The optimal photographic record (illus. 4.2), which reveals the site as a series of positive cropmarks reflecting the presence of a complex of ditches, generally confirms this account. However, it suggests only a single entrance to circular enclosure 2 on the east side in addition to a major gap, at least in the inner ditch, to the west. But the photograph also reveals rather more. For example, within the double ditched enclosure (2) are at least two possible, though ill-defined, circular houses: one revealed as a dark, sub-circular mark towards the centre, the second partly defined by a narrow curving ditch up against the southern side of the inner enclosure ditch. In addition, the polygonal enclosure (1) is potentially double ditched, at least on its northern and eastern sides, and seems to overlap the arc of a third circular enclosure (6) in the southern comer of the field, though their respective enclosing ditches cannot be shown to coincide at any point. Enclosure 6 lies largely beneath the buildings of the modem farm and is defined by a single, broad ditch with an entrance to the north, though evidence of the line of the ditch is lost as it approaches the entrance to the field.

A 20m x 20m survey grid was established over the field and the survey commenced using sampling and traverse intervals of I .Om for both methods. Because the surveys were conducted by different students at different times, the machines may not always have been optimally balanced. This imbalance appears in places in the survey plots, primarily in the resistivity data, as slight mismatches in grid edges, giving an indication of the location of individual 20m grids. It has been considered preferable to leave the mismatches visible rather than to over-process the data to remove them. The incomplete grids noticeable on the resistivity plots arose where there was standing water at the entrance to the field in 1999 or as a result of instrument failure. Gaps in the magnetic data, again at the entrance to the field, were the result of the parking there of a combined harvester in 2000. Magnetic data collection was also stopped short of the grid edge all along the east side of the survey grid because the field is bounded by a metal fence.

In addition, a further possible enclosure is visible. Located between, and apparently overlapping, the two ditches on the north-east side of enclosure 2 is a further sub-circular enclosure (4) not dissimilar in either size or character to the clearer of the two possible circular houses within that enclosure. It is defined by a narrow arc of ditch which is clearest on the north west side.

All of the survey plots are grey scale plots. The resistivity survey results are presented in illus. 4.3 as a shade plot without any processing applied other than edge matching of certain grids and despiking. The latter is necessary occasionally when there has been poor electrical contact made between the mobile electrodes and the ground, caused by particularly stony areas and the use of inexperienced surveyors. Processing was completed with the plot being interpolated, a cosmetic effect applied to smooth the appearance of the data. It is applied to all of the plots presented here to remove the slightly pixelated appearance that often occurs at a I.Om survey interval. In illus. 4.4 the same data is presented, but a high-pass filter has been applied prior to interpolation. This processing tool provides a means of filtering out lower frequency readings which are often caused by underlying geological trends, leaving responses more likely to be caused by features of an archaeological nature. The magnetic data (illus. 4.5) is also presented as a shade plot, but without any further processing having taken place.

Resistivity Survey

Increased moisture content facilitates the passage of an electric current, so that ditches are generally seen as lines of lower resistance. Areas of high resistance indicate drier ground and usually reflect the presence of concentrations of stone or more compacted earth, as in the case of lines of walls or earthen banks. The results of the survey at Burnfoot do not entirely comply with the anticipated norm. It is quite clear from the aerial photographs that the various enclosures are revealed by positive cropmarks indicating that they are demarcated by ditches. Yet the resistivity survey records the same enclosures as lines of high resistance, a response that requires some explanation. However, this is not a unique phenomenon, for it has been noted on other sites, though it is not so common that its occurrence passes without comment. The first condition that must be taken into account is the fact that the cropmark was recorded in mid-late summer when the barley crops in Clydesdale are ripening, whereas the geophysical surveys were undertaken in March in relatively damp weather. Clark noted that under certain conditions of climate and seasonality, reversal of recorded

Aerial photography

The published record of the enclosures at Burnfoot (RCAHMS 1978, 146) refers to only two broad-ditched enclosures: one polygonal in shape, defined by a single ditch (enclosure 1), and a second to the north,

52

WS. Hanson and Lorna Sharpe: The relative information value of aerial photography and geophysical survey

anomalies could occur (1990: 48-56). In an extended experiment conducted over a series of 3 ditches of various sizes cut into chalk, he concluded that the size and the cross-section of the ditches had a greater influence on the resistivity anomaly recorded than the composition of the materials filling them. Negative anomalies of varying strength and clarity, the traditionally accepted response from a ditch, were recorded over the largest of his ditches all year round. The smallest ditch, however, revealed a positive anomaly of varying strength for the whole of the year, while the medium sized ditch showed an intermediate response, with an occasionally high resistance during the summer months. This was explained as being a response to the soil moisture conditions, with the large ditch acting as a reservoir. The small ditch was shown to be very responsive to changing climatic conditions and was assumed to be drying quickly, particularly if its filling was coarse textured, thus giving high resistance readings. Although these experiments were carried out in very different geological conditions from those encountered at Burnfoot, accounts of further investigations into resistivity responses in relation to seasonality suggest that the variations described by Clark are valid for other geological settings. (e.g. Clark 1990, 51; Al Chalabi and Rees 1962).

the characteristics of internal structures and may reflect the lines of enclosing banks. Augering indicated that there was a higher concentration of stone present in this area. There is no sign of either internal house postulated on the basis of the aerial photograph. The possibility that there was part of a further enclosure (3) in the field was first noted following the geophysical surveys of the site, particularly the magnetic data (see below). It is visible on the resistivity plot as a rather blurred semicircular area located against the south-eastern edge of the field and overlapping both inner and outer ditches on the south-east side of enclosure 2. The limit is defined by two lines of lower-resistance, clearest on the south side, indicating a possible double ditch. Slightly to the north of the centre of this enclosure is a large part of a small circle of positive resistance which may reasonably be interpreted as a house. Enclosure 4 is visible as a sub-circular structure defined by a narrow line of high resistance. It measures around 25m east-west and may possibly be identified as a house. It appears to overlie the ditches of enclosure 2, which lie approximately 15m apart at this point. There is a very high resistance feature of uncertain identification in the north eastern quadrant of its interior. Immediately to the north east of enclosure 4 is a further possible enclosure (5) of similar size and definition, though this lies outside the outer ditch of enclosure 2.

The transcription of the aerial photographs of the Burnfoot enclosures indicates that the average width of their ditches varies from 3-5m, which is comparable to Clark's smallmedium ditch. The field is likely to have been holding water at or above its capacity during the period of the surveys, as indicated by the waterlogging at the entrance in 1999, so that the ditches may actually have been holding less moisture as a result of their better drainage capacity. This would account for the high resistance anomaly relative to the background reading for the field. A contributory factor could be that there is a high percentage of stone in the backfilling of the ditches, since limited probing suggested that the field was very stony.

Magnetic survey

Because the in-filling of ditches tends to increase their magnetic susceptibility compared to the surrounding subsoil, they tend to show as positive anomalies in magnetic surveys, while features such as walls or rubble spreads composed of material that contains less magnetic minerals give negative readings. At Burnfoot, however, as with the resistivity survey, the opposite effect seems to occur, with ditches showing as negative lines. This is assumed to be due to the uniform field associated with silting and soil formation in the ditches set against the higher magnetically susceptible background of subsoils and drift geology derived from boulder clays from the Southern Upland Fault, and the underlying highly magnetic solid igneous geology. Similar situations have been observed in the Abbey complex on Iona (Clark 1990: 96) and at Balfarg henge in Fife (Munro and Papamarinopoulos 1978). Comparable examples have been identified by Doggart in Ireland, where the peat which had formed in ditches had a lower magnetic susceptibility that the surrounding soils and underlying drift geology (1983).

If we now turn to the interpretation of the resistivity survey (illus. 4.4), the polygonal enclosure (1) is relatively clear, with indications of two ditches on all three sides revealed, though the relationship with the double ditched enclosure (2) is less clear. In the west of the enclosure is an area of high resistance which is likely to represent a concentration of stone, either a structure, given its regularity, or perhaps more likely part of the rampart. There appears to be a break in the enclosure ditch centrally placed in the northwest side.

Virtually the whole of the double-ditched enclosure (2) is visible, suggesting overall dimensions of around 100m by 120m, and dimensions within the inner ditch of 60m by 75m. The enclosure is less regular, particularly on the north west side where the outer ditch appears to project down the slope. There are no obvious breaks for entrances. In the interior there are two areas of high resistance bordering the eastern and western arcs of the inner ditch, the former continuing in a less marked way between the inner and outer ditches. Neither areas have

Despite the fact that only some two-thirds of it was covered by the survey, the outline of the polygonal enclosure (1) is very clear, confirming the presence of two ditches on both the north east and south east sides. The relationship with the double-ditched enclosure (2) is quite explicit and gives the opposite impression to that obtained 53

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

from the aerial photograph. The outer ditch on the south side appears to cut across the wide outer ditch of enclosure 1, suggesting a possible link with what has been taken so far to be the line of the inner ditch of that polygonal enclosure. Similarly, the inner ditch of enclosure 1 gradually converges with, but still seems to overlie, the outer ditch of enclosure 2. Unfortunately, the survey did not extend into one of the more contentious areas of interpretation in the north east quadrant. Nonetheless the picture obtained is clear and suggests that enclosure 2 was of greater regularity than either the aerial photograph or resistivity plot indicate. There are also strong indications of a break in both inner and outer ditches in the south-east quadrant which may well indicate an entrance. The interior of enclosure 2 is magnetically quiet, with no clear indication of houses within it, though an arc of narrow ditch up against the southern side is faintly detectable, reflecting the line of one house postulated from the aerial photograph. High magnetic values were recorded mirroring the line of the ditches at various points around the circumference, both inside and outside the ditches. These are, however, simply edge effects caused by the return flux of the magnetic anomalies which are dipolar in nature (Keary and Brooks 1991: 162).

the double-ditched sub-circular enclosure appears, from the magnetometer survey, to post-date its polygonal neighbour. Both are likely to have been settlements of later prehistoric date, though only enclosure 2 provided any evidence of internal houses, and that only tentatively from the aerial survey with very limited support from the magnetometry. An Iron Age date for enclosure 2 would not be an unreasonable suggestion. The relationship between polygonal enclosure 1 and enclosure 6 could not be established, but since they occupy common ground they are unlikely to be contemporary. Given the provision of double ditches, the former seems unlikely to have served as an attached stockyard for a settlement based in the latter. Again primarily on the basis of the magnetometer survey, enclosure 3 appears to overlie and therefore post-date enclosure 2. The presence of a circular house in its interior clearly indicates its function as a settlement, again of later prehistoric date. The morphological similarity and close juxtaposition of enclosures 4 and 5 suggest their contemporaneity and linked function. Their smaller size, and the relatively high magnetic readings from the former, would indicate dwellings. On the basis of the resistivity survey the former appears to overlie and, therefore, postdate enclosure 2. The situation is reminiscent of the phenomenon frequently observed in field survey and occasionally in excavation in upland areas of Northumberland and the Borders where unenclosed stonebuilt houses can be seen to overlie the defences of small fortified enclosures (e.g. Jobey 1964: 52-4; Steer 1958: 103).

Enclosure 3 is quite clearly visible, except on the east side, as an arc of ditch which appears to cut both inner and outer ditches on the south-east side of enclosure 2, and so presumably post-dates it. The possible house in its interior is also relatively clear, at least on the north side. There is no indication in the magnetic plot of the line of the possible house, enclosure 4, lying between the ditches of enclosure 2. However, there is an area of high positive magnetic responses that corresponds with its position. Though this may be a response to the underlying geology, given the suggestion from both the aerial and resistivity surveys of a house here it seems more likely to represent some anthropogenic activity, the most obvious of which would be the presence of spreads of burnt or midden material. There are, however, no anomalies to mark the position of enclosure 5 beyond the outer ditch of enclosure 2.

Conclusion The probable unenclosed house, enclosure 4, is discernible on the resistivity plot and the aerial photograph, but only detectable on the magnetic survey as a concentrated area of higher readings, which thus provides useful complementary information. The second example, however, enclosure 5, can only be seen on the resistivity plot, though its clarity and similarity with enclosure 4 are sufficient to instil confidence. Enclosure 6 is sharply demarcated on the aerial photograph and has been detected magnetically in part, even though there were constraints on the survey in that area. Its complete absence from the resistivity survey almost certainly reflects the general dampness of that part of the field, for the problem of waterlogging near the entrance has already been alluded to above.

The parking of a combined harvester by the gate prevented the magnetic examination of two 20m squares in the comer of the field and consequently only a small section of the north part of enclosure 6 was recorded, adjacent to the south-western boundary of the field. The anomaly was, however, strong and clear.

As can be seen, the enclosures 1 and 2 are the most obvious features on the aerial photograph and both the magnetic and resistivity plots. They are clearest, however, particularly in terms of their interrelationship and the location of probable entrances, on the magnetic survey. Neither geophysical survey plot provides any support for the existence of internal houses, thus calling their existence into question, though the resistivity survey suggests the presence of internal banks. Enclosure 3 and its associated internal round house is visible on both geophysical plots,

Interpretation of the site complex (illus. 4.6) The combination of geophysical survey plots and aerial photographs reveals a series of up to six enclosures in the area of the modem field lying to the north of Bumfoot Farm. Enclosures 1 and 2 are separated on the ground by a deep gully, visible in all three media, which serves to mask or confuse the relationship between them. Nonetheless,

54

WS. Hanson and Lorna Sharpe: The relative information value of aerial photography and geophysical survey

though different elements feature more strongly on each plot, so that the full interpretation requires input from both. Indeed, superficial examination of the magnetic plot might lead to the conflation of the north side of the house and the south side of the enclosure to suggest a much smaller enclosure. With hindsight, as is often the case, traces of enclosure 3 are visible on the aerial photograph, but were not sufficiently clear to have postulated its existence on that evidence alone.

Munro, M.A.R. and Papamarinopoulos, S. 1978. The Investigation of an unusual magnetic anomaly by combined magnetometer and soil susceptibility surveys. In Archaeo Physika 10: Proceedings of the 18th International Symposium on Archaeometry and Archaeological Prospection, Bonn: 675-680. Newman, C., 1997 Tara: an archaeological survey. Dublin: Royal Irish Academy. RCAHMS 1978. Lanarkshire: An Inventory of Prehistoric and Roman Monuments. Edinburgh: HMSO. Steer, K.S., 1958. The Severan re-organisation. In Richmond, I.A. (ed.) Roman and native in north Britain. Edinburgh: Nelson: 91-111.

Though the application of geophysical survey is an increasingly common approach in any large scale landscape study (e.g. Newman 1997), its validity in Scotland has been seriously questioned in the past. The different information obtained from the three modes of survey examined here serves to underline the validity of such an approach and justify more extensive application of geophysical survey to augment and clarify what is known from aerial reconnaissance. Nor is it simply a case of different information coming from different media, for the appearance of less certain anomalies on more than one medium can serve to reinforce their identification. Though aerial survey will continue to be by far the most costeffective means of enhancing our knowledge of the distribution of archaeological sites in a lowland context, geophysical survey is a relatively inexpensive means of extending that database, at least when compared to excavation. It can serve not merely to augment the information from photography, as here, but to extend beyond it by considering the apparently blank areas between known settlement foci, particularly in areas less susceptible to aerial photography in terms of soil type or agricultural regime.

BIBLIOGRAPHY Al Chalabi, M.M. and Rees, I.A., 1962. An experiment on the effect of rainfall on electrical resistivity anomalies in the near surface. Bonner Jahrbucher 162: 266-271. Clark, A. J., 1990. Seeing Beneath the Soil: Prospecting Methods in Archaeology. London: Batsford. Doggart, R., 1983. The use of magnetic prospecting equipment in Northern Ireland. In Reeves-Smyth, T. and Hamond, F. (eds.), Landscape Archaeology in Ireland, Oxford: British Archaeological Reports, British Series 116: 35-45. Hanson, W.S. and Macinnes, L., 1991. The archaeology of the Scottish lowlands: problems and potential. In Hanson, W.S. and Slater, E.A. (eds.) Scottish Archaeology: new perceptions. Aberdeen: Aberdeen University Press: 153-66. Jobey, G., 1964. Enclosed stone-built settlements in north Northumberland. Archaeol. Aeliana 4th series 92: 4164. Keary, P. and Brooks, M., 1991. An Introduction to Geophysical Exploration. (2nd Edn.). Oxford: Blackwell Scientific Publications.

55

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

N

i

Tinto

Hill

Key: Railway line ---Scale:

Dungavel Hill

o___________

_c_.5miJes

4.1 Map of the Clyde valley and the location ofBumfoot Farm.

4.2

Aerial photograph ofBumfoot Farm enclosures, 1989 (Copyright W.S. Hanson) 56

WS. Hanson and Lorna Sharpe: The relative information value of aerial photography and geophysical survey

~

130.00

12t67

11133 105.00 86,67

80.33 80.00

7167 133>33

55.00 46,67 38,:33 30.DOohm Palette· redgrejlb.ptt

o--140m

Size: x 1/2

-3SD Mean 3SD

Mimimum 30

4.3

Ma>:imum Contrast

13D

Units

Absolute

1

The resistivity survey data from Burnfoot Farm.

301)0 25.83

21.67 17.50 13.33

N

9.17 5.00 0.83 .3.33 .750 ·1167 ·1583 -20.00 ohm

Palette: redgreJb.ptt

n 11111111111"""'

1 40m

Sizex112

~

·3SD Mean 3SD

ClipParameteri· Mimimum -2[f M.~wirnum 30

Crntraet IJnits

1 Abiolute

4.4 The resistivity data from Burnfoot Farm after the application of a high-pass filter.

57

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

5.DD 4.00

3.00 2.00 1.1)(1

O.DO -1.00 -200

-3.00 ·4.00 -5.00 -5.00 ·700 nT

Palelle : redg,eyb.p~

OM""""

\4Dm Size:, 1/2

-3SD Mean 3'3D

Mimim11m hfoximum

-7 5

Contrast Units

Absolute

4.5 Magnetic survey data from Burnfoot Farm.

BURNFOOT

I

I 0

JO

1w -

50

-

4.6 Interpretative plot of the enclosures at Burnfoot Farm.

58

IOOm

5 The Edge of Empire: Late Prehistoric and Romano-British Settlement in North West England. A Study in Marginality Mike Nevell Age) population who made their living from subsistence farming, in what has been characterised as a climatically and agriculturally marginal area (Nevell 1999a & 1999b). What has emerged from the wider landscape research of the 1980s and 1990s is an increasing need to understand the origins and ultimate destination of Romanization in North West England. Now under dissection is the interplay between subsistence farming, the landscape and the regional climate. This process has added value in that this example has some potential to provide models and parallels applicable to other episodes of transition in the region, such as the shift from hunter-gathering to farming, or more particularly the later change from a rural to an industrial society.

Introduction

The Roman period, together with its associated social and cultural changes, has left a rich archaeological record across much of western and southern Europe, northern Africa and the Middle East. In Britain it is a period that is perceived as being well known and extensively studied. At least within North West England (however that modem term is defined) this era is marked by the appearance of large military sites and, for the first time, associated urban centres. It was Barri Jones who first introduced me to Roman archaeology when I was an undergraduate at Manchester. Later, as one of his postgraduate students, he encouraged my interest in the Romano-British occupation of North West England, particularly the problem of site identification and location which became the subject of my PhD. His breadth of vision and knowledge widened my own horizons and that of my research, even if his supervision was sometimes haphazard. Finally, as a colleague his continued enthusiasm for my own work and that of the archaeology unit at Manchester (which he helped to found) remained a constant source of support. His restless energy was a vital factor in promoting continued research into the Romano-British landscape around Manchester and in the North West in general, firstly through a conference entitled Living on the Edge held in Manchester in November 1995, and secondly through his two contributions to the subsequent book of papers, based loosely on that event, which I was editing (Jones 1999a, 1999b; Nevell 1999). One of my last memories of Barri is his hasty arrival on the morning the volume was due to go to press with the long promised preface in his hand. Written in his typically classical and erudite style it gave the volume that broader vision and context which Barri was so often the first to recognise.

This paper presents an overview of some recent attempts to provide a theoretical framework for both the location and interpretation of late prehistoric and Romano-British rural sites within the region, centred around the theme of marginality. It was an issue first addressed by Barri Jones in 1979 but it has taken the archaeological community of North West England two decades to pursue the problems of site location and shallow settlement hierarchy during the late prehistoric and Romano-British period which he first identified (Jones 1979). Professor Sydney Pollard has demonstrated the impact of two main types of marginality, which he terms on the one hand economic marginality and, on the other, social/political (Pollard 1997: 10-7). In N orth-westem terms we may be able to see the impact of each during the late prehistoric and Romano-British era. Firstly, I propose to focus on economic marginality, which is more about the natural features of a region than its political make-up. In pre-industrial, non-urbanised societies this economic marginality was expressed in how good the land was for cultivation. Typically there were three types of landscape which made regions marginal in Europe; mountains, forests, and fen or marshland. North West England has all three in abundance. However, some of these marginal regions became highly productive once industrialisation took hold. Therefore, we should be wary of dismissing such areas as always being economically marginal just because they were marginal for subsistence agriculture (Dark and Dark 1997). Secondly, it is necessary to examine the social/political tensions between the centre and the periphery. The issue here is the tension between the centre as an overriding influence, which seeks to open up, subject, and/or colonise the fringe, and the potency of the fringe community, which might alternatively come to dominate the centre (Millett 1990). These stresses could be expressed physically as much as intellectually and in the

As long ago as 1979 Barri Jones drew attention to the fact that Romano-British rural settlement in North West England, essentially the non-military sites, remained a largely unknown quantity (Jones 1979). In part this was due to the concentration of over two centuries of scholarly study on the more visible Roman period sites, the Roman forts. In part it was due to the difficulties of applying twentieth century rapid site location techniques such as aerial photography and field walking in a region dominated by mossland, grasslands on the extensive clay terraces and widespread urban sprawl south of the River Ribble. Since that observation, and despite two decades of research (Collens 1994; Higham 1980; Higham and Jones 1985; Jones 1979; Matthews 1999; Philpott 1994), we are only just beginning to investigate the impact of the Roman military presence on the indigenous late prehistoric (Iron

59

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

cereal cultivation. The recovery in temperatures after c 150 BC are likely to have restored the limits for cereal cultivation to their mid-twentieth century levels, between 200m and 250m AOD, by the time of the Roman conquest of the region in the 70s AD. There was continued improvement into the third century AD. Therefore, for most of the first millennium BC, all year round settlement above c. 110m AOD within the region would appear to have been filled with risk. Such figures are undoubtedly crude but it is not so much the absolute temperatures that are important in the assessment of the impact of climatic change on the local environment, but rather the pattern of fluctuation, and this is likely to have approximated to these figures.

North West, which lay at the extreme north-western edge of the Roman Empire, may be recoverable from the archaeological data. Before we can assess how climatic and social marginality might have affected the archaeology of the late prehistoric and Romano-British period in North West England, we have to arrive at a consistent geographical definition of the area under discussion. The focus of this paper is the landscape to the south of the Cumbrian massif and to the west of the Pennines. Yet even this zone is topographically fragmented, being dominated by a series of river valleys running east to west into the Irish Sea: the Dee, Gowy, Weaver, Mersey, Alt, Douglas, Ribble, Wyre and Lune. Several of these rivers are separated by prominent ridges or hills, especially north of the Ribble where the Lancashire plain is reduced to a narrow strip a few kilometres wide. However, some coherence can be seen in the catchment area of the River Mersey and its estuary. This catchment area runs westwards from the Pennines as far as Liverpool, Chester and the Wirral, defining an area that is roughly bowl shaped. The Mersey Basin, as geographers have long called this area, encompasses most of the land south of Wigan and north of Nantwich, and includes the Gowy, Weaver, Sankey and Mersey rivers - an area roughly 80km by 70km. It is surrounded on three sides by hills: the Rossendale uplands and its outliers around Wigan to the north and north-west, and the Pennines to the east and south-east as far as Congleton. It is this area that is the primary focus of the rest of this paper.

An Eco-deterministic Model for Settlement: the Late Prehistoric Period Several researchers have demonstrated that the palaeoenvironmental deposits from the Mersey Basin preserve evidence for climatic change during this period (Brayshay 1999; Cowell and Innes 1994; Hall et al. 1995; Nevell 1999a). They also show something else, which is evidence of anthropogenic changes to the vegetation, in the form of woodland clearance episodes and the occurrence of cereal pollen. This evidence allows us to postulate the impact of changing settlement trends on the landscape, and this in turn can be used to construct an eco-deterministic model of settlement trends within the region for the centuries before and after the Roman conquest. Seven lowland diagrams available from this period show broadly similar developments in the regional vegetation of the lowlands of the Mersey Basin, with the two earliest episodes of woodland clearance assignable to the mid to late first millennium BC, separated by a short phase of woodland regeneration. The first of these episodes occurred immediately after the early first millennium BC recurrence surface, dated to the period 795-595 BC, and was characterised by sustained woodland clearance and an absence of cereal pollen, suggesting pastoral farming (Nevell 1999a; Cowell and Innes 1994; Hall et al. 1995; Leah et al. 1997; Ogle, Robinson and Shimwell 1997: palaeo-environmental sample sites are from Chat Moss, Holcroft Moss, Knowsley Moss, Godley Brook, Lindow Moss, Risley Moss, Simmonswood Moss). A brief period of forest regeneration was followed by a second phase of woodland clearance within the Mersey Basin during the late first millennium BC. This was characterised by a period of highly intensive agricultural activity, involving major deforestation, high levels of weed pollen and, for the first time, the introduction of cereals (and possibly hemp/hops) in high quantity. This period of intense land use has been dated at Lindow Moss, in eastern Cheshire (SJ 8200 8050), to the period after 430-250 BC (340 +/- 90 BC; BM 2401). Samples from Simmonswood Moss on Merseyside show a similar pattern of clearance, dated after the period 790-257 BC (2380 +/- 80 BO; Birm-1221; Cowell and Innes 1994). The same pattern is apparent in

Climatic Marginality

A Model for Climatic Marginality I have already discussed elsewhere, at some length and detail, the theoretical background for the influence of climatic marginality upon settlement during the first millennium BC and early first millennium AD within North West England (Nevell 1992 & 1999a). Using the models pioneered by Parry (1975), it is it possible to argue that within the Mersey Basin, and elsewhere west of the Pennines, climatic instability, and in particular fluctuations in annual mean summer temperatures, affected the altitudinal limit on cereal cultivation. This in tum limited the range of permanent settlement, particularly during the first millennium BC, but less so during the early centuries of the first millennium AD. Within North West England in the mid-twentieth century, the marginal limit for cereal cultivation lay between 200m and 250m above ordnance datum (henceforth AOD: Crowe 1962, 44). Around 1200 BC the altitudinal limit for cereal cultivation could, theoretically, have been as high as 460m AOD. By c. 150 BC, however, the fall in the annual mean summer temperature, which may have been as great as c. 2.5 degrees C., could have reduced this limit to as low as 110m AOD (Lamb 1982; Nevell 1999a). Certainly land above this level would have been highly marginal for

60

Mike Nevell: The Edge of Empire

Chat Moss and Holcroft Moss in northern Cheshire and western Greater Manchester (Nevell 1999a; Birks 1964, 1965; Hall et al. 1995; Leah et al. 1997).

75%. Extensive upland woodland clearance is also indicated in this period from Featherbed Moss, although the evidence from here suggests a decline in upland activity above c. 300m AOD after the mid-third century AD, with regeneration of the woodland beginning around AD 280 (Tallis and Switsur 1973: 744). Similar regeneration of woodland began around AD 290 at Deep Clough in central Rossendale (Tallis and McGuire 1972: 727).

These lowland pollen samples therefore indicate a twophase development of agriculture in the Mersey Basin during the mid- to late first millennium BC. In contrast, three diagrams from the uplands of the region indicate sustained forest clearance from the mid-first millennium BC in the Rossendale area, but only from the late first millennium BC elsewhere in the southern Pennines (Bartley 1975: 378; Tallis and McGuire 1972; Tallis and Switsur 1973: palaeo-environmental samples from Deep Clough, in the upper Irwell valley; Rishworth Moor and Featherbed Moss).

Climate and Settlement Trends To summarise, therefore, the palaeo-environmental evidence for the period c. 795 BC to AD 526 reflects the climatic decline of the early to mid-first millennium BC and the subsequent recovery of the late first millennium BC and the first few centuries of the first millennium AD. Furthermore, this evidence can be used to model variations in human activity, indicated by three phases of clearance, which culminated at the end of the first millennium BC and beginning of the first millennium AD with the first occurrence of large quantities of cereal pollen coincident with sustained forest clearance. This pattern of human activity would appear to match the climatic cycle of this period, although whether the two are directly related is difficult to prove: superficially the link seems strong. Within this context it may be significant that the few upland univallate hillforts or palisaded enclosures of the southern Pennine uplands - such as Almondbury, Castercliffe, Mam Tor and perhaps Portfield, (Combes 1982; Combes and Thompson 1979; Cunliffe 1991: 344-52; Varley 1976), appear to have been abandoned by the middle centuries of the first millennium BC (Kenyon 1991: 28; Hart 1984: 73-5). By the last quarter of the first millennium BC none of the largest hillforts in the North West were occupied. Thus, at the hillfort of Portfield, in the Ribble Valley to the north of the Mersey Basin, the main period of use for the defences belonged to the years 750-500 BC (Beswick and Coombs 1986, 175-6). Similarly, at Castercliffe (also in the Ribble Valley) radiocarbon dates for the ramparts centred on 510 +/- 70 BC (S 286; Coombs 1982: 127-8). In Cheshire the ramparts at Maiden Castle were dated to c. 390 BC (British Archaeological Abstracts 88/510) and the main occupation of the hillfort at Beeston Castle spanned the years 765 to 257 BC (Ellis 1993: 85-6).

An Eco-deterministic Model for Settlement: the RomanoBritish Period Seven dated pollen diagrams are available from the Mersey Basin for this period and all record major and sustained woodland clearance over many centuries at the end of the first millennium BC and during the first centuries of the first millennium AD. These clearances appear to be broadly chronologically coincident across the Basin, and form the third significant period of palaeo-environmental disturbance after the recurrence surface of 795-595 BC. The end of this third phase of clearance activity is marked by a second recurrence surface which is radio-carbon dated to the years 326-526 AD (Godwin and Willis 1960: 62-72). Five pollen diagrams indicate a major and sustained upsurge in agricultural activity, associated with significant amounts of cereal pollen and widespread tree clearance, in the centuries immediately before the c. 326-526 AD recurrence surface, at Chat Moss A & B, Holcroft Moss, Knowsley Park Moss and Lindow Moss I (Nevell 1999a; Cowell and Innes 1994; Hall et al. 1995). This evidence is supported by two palaeoliminological studies showing increased soil erosion in this period in Cheshire at Peckforton Mere, near the central Cheshire Ridge, and at Rostherne Mere in northern Cheshire, where this episode began sometime between 366 BC and AD 60 (Leah et al. 1997; Schoenwetter 1982). The upland pollen diagrams from the fringes of the Mersey Basin also indicate an upsurge in activity during this period, but of a different nature. The pollen diagrams from Deep Clough, at 340m AOD, and Rishworth Moor, at 410m AOD, both indicate the continuance of the substantial woodland clearance seen towards the end of the first millennium BC. Both are characterised by the dominance of grass pollens indicative of an open landscape perhaps used for pastoral farming (Bartley 1975: 378; Tallis and McGuire 1972: 723). The extent of upland woodland clearance in this area by the beginning of the Roman period is indicated by a pollen sample carbon-dated to the years 50 BC to AD 110 (30 +/- 80 AD: GaK 2025), which shows that tree pollen accounted for only 15% of the total dry land pollen, shrub pollen 10% but grass pollen

The palaeo-environmental and archaeological evidence would seem, therefore, to provide some support for the theory that the 110-250m zone in the southern Pennines and its foothills was the most agriculturally marginal area. However, the lack of a comprehensive network of dated palaeo-environmental samples from across North West England means that other marginal areas may not be represented in this data. The highly localised palaeoenvironmental evidence from Tatton Park, for instance, which lies on claylands at c. 60m AOD, shows late prehistoric clearance activity but regeneration in the Roman period, hinting that it may be possible to recover smaller niche environments with finer data (Higham and 61

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Cane 1999). One further zone of agricultural marginality highlighted by the North West Wetlands Survey is the large basin mosslands of the region, which were not conducive to early settlement in the same way as the Somerset Levels and the Fens. Additionally, coastal change in this period, particularly around the Wirral peninsular, the Mersey Estuary and along the Lancashire coast line, would also have affected settlement potential, particularly below the Sm contour line.

Beeston (4ha) can be viewed as true hillforts. How accurate such an assumption might be is open to question, although anthropological parallels (Matthews 1999) would suggest that such a settlement hierarchy may be recoverable from size alone. The point here is that the earthwork and cropmark enclosures from the Mersey Basin can be treated as broadly one category of sites. This gives us over 60 sites in the Mersey Basin which, morphologically, may belong to the late prehistoric and Romano-British periods, although only 22 have produced excavation evidence from this period (Table 5.1). Of these, 13 can be shown to be late prehistoric in origin; 16 enclosures can be shown by excavation to have Romano-British phases; and six have both late prehistoric and Romano-British phases (Duttons Farm (pers. comm. Dr R. Philpott), Brookhouse, Great Woolden Hall, Irby, Mellor and Rainsough). There are a further eight enclosures where various types of fieldwork have failed to provide a date, although a late prehistoric or Romano-British origin is strongly suspected (Arthill, Bradley, Burton Point, Giant's Seat, Helsby, Little Lever, Oakmere, Rhodes Green, and Woodhouses).

Social Marginality: The Archaeological Evidence For Settlement The Nature of the Evidence

The archaeological evidence for settlement during the first millennium BC and early first millennium AD is dominated by two monument types: small enclosed sites, usually less than 2ha in area, and larger, usually unenclosed, sites above 3ha in area. The best known of these settlement sites are the enclosures, which can be divided into those surviving as upstanding earthworks and those that have been ploughed-out but which are recovered as cropmarks. Most of the earthworks were first identified and catalogued by Forde-Johnson (1962). He used the hillfort model in his interpretation of these earthworks, comparing them with the better known sites of the Welsh Marches and South-West of England. Using his criteria there are thirteen hillfort type sites that lie in the modem counties of Cheshire, Greater Manchester, Lancashire and Merseyside, of which eleven are situated within the Mersey Basin (Beeston Castle, Bradley, Burton Point, Castlesteads, Eddisbury, Helsby, Kelsborrow, Maiden Castle, Oakmere, Rainsough, and Woodhouses). These sites ranged in size from 0.lha (Burton Point) to 4ha (Beeston Castle), and had a mixture of single and multiple ditches and banks as defences.

Three topographical sub-groups can be tentatively identified within this group. Firstly, promontory settlements, examples of which are beginning to be found along the escarpment edges of the river valleys of the Mersey Basin. Dated examples are known from Castlesteads, Great W oolden and Rainsough, but other potential examples include a double-ditched cropmark site at Giants Seat in the Irwell Valley, and the cropmark ditched enclosure at Little Lever in the Irk valley. Secondly, hilltop sites along the western Pennine fringes and along the Central Cheshire Ridge (Beeston Castle, Eddisbury, Kelsborrow, Maiden Castle, Mellor and Hangingbank). Thirdly, niche sites on or near to the boundary between two different soils types (Duttons Farm, Irby, Halton Brow, High Legh, Legh Oaks I & II, Tatton Park and Winwick).

The ploughed-out cropmark enclosures were identified through survey work during the 1980s and 1990s by archaeologists from Chester, Liverpool and Manchester (Collens 1994 and 1999; Jones 1999b). Within the Mersey Basin these number over 50 and more can be expected. These sites are characterised by small single and double ditched enclosures, usually less than 2ha in area, of a type familiar in southern Britain from the late first millennium BC (Nevell 1999a; Collens 1994, 1999). The cropmarks range in size from 0. lha to 2.8ha. Stylistically there is no difference between the cropmark sites and the earthwork sites traditionally identified at hillforts, other than topographical location (the earthworks usually lie in the 110-250m AOD zone) and the presence of earthwork banks and ditches in the latter. Within Forde-Johnson's own work he made a distinction between true hillforts above roughly 2.5ha in area which he supposed acted as central places and the home of a local chieftain, and smaller sites which he regarded as farmsteads. If we apply this criterion to the Mersey Basin, then of the earthwork enclosures only Kelsborrow (3.3ha), Eddisbury (3.5ha) and

The second category of sites comprises 11 settlements all bar one of which appear in the last quarter of the first century AD and are associated with the Roman conquest and occupation of the Mersey Basin: Castleshaw, Chester, Heronbridge, Manchester, Melandra, Meols, Middlewich, Nantwich, Northwich, Tilston and Wilderspool. These Romano-British civilian settlements are characterised in a number of ways. Firstly, they are large compared to the late prehistoric sites. They range from roughly 3ha (Melandra and Northwich) to over 50 ha (Chester), although the area of a number of sites remains unclear (Castleshaw, Meols and Nantwich) and even for the other sites the figures quoted in this paper are only estimates. This great variability in size reflects the individual development of these sites during more than three centuries of Roman occupation. Secondly, they have some internal order through the presence of streets.

62

Mike Nevel!: The Edge of Empire .

Name of Site Beeston Castle Brookhouse, Halewood Castle steads Court Farm Dutton's Farm Eddisburv Great Woolden Halton Brow Hanging bank High Legh Irbv Kelsborrow LelZhOaksI Lel!b Oaks II Maiden Castle Mellor Rainsough Saltnev I Saltney II Ochre Brook, Tarbock Winwick

Area in Hectacres 4.0ha ? I.I ha ? ? 3.5 ha 1.1 ha 1.68 ha 1.23 ha 0.6ha 0.88 ha 3.3 ha 0.1 ha 0.31 ha 0.7ha 2.2ha 0.96 ha ? ? ?

l.27 ha

-

---

-

--- - - - -- -

Period if Activitv LBA - Mid-Tron Age Iron Age -RB Late Iron Age RB Iron Age-RB Iron Age Late Iron Age - RB RB RB RB Iron Age -RB iron Age Iron Age RB Iron Age Late Tron Age - RB Late Iron Age - RB RB RB RB RB

---

-

-----.

Heieht above Sea Level I00mAOD 20mAOD ll0mAOD 20mAOD !SmAOD lSOmAOD 20mAOD 70mAOD 230mAOD 60mAOD 60mAOD 170mAOD 60mAOD 60mAOD 2llmAOD 220mAOD 65mAOD l0mAOD !0mAOD 20mAOD 25mAOD

-------

Number of Ditches 2 2 1 I ? 3 2 1 2 I I 2 1 I 2 2 1 I I I I

Table 5.1 Late prehistoric and Romano-British Rural Settlements in the Mersey Basin

settlement may have extended across as much as 30ha. It was dominated by salt processing and had streets running from the south-eastern and north-eastern gateways of the old fort site and side streets off those (Petch 1987; Higham 1993; Matthews 1999; Shorter 2000a,b).

Thirdly, they have Romanised buildings in the form of rectangular or winged timber-framed, half-timbered or stone buildings. Finally, these sites produce vast quantities of Roman material, typically pottery but also brick, glass, masonry and tile, the latter four materials occurring for the first time in the Mersey Basin.

The purely civilian settlements, with the exception of Meols, appear to have developed after the Roman conquest in the AD 80s and AD 90s and they continued in occupation until the fourth century. They appear to have had a wider economic base than most of the vici, with crop processing, metalworking, leather processing, pottery manufacture and salt production attested at these sites. Whilst the size of Meols and Nantwich is unclear Heronbridge and Tilston became substantial settlements of c. 13ha and c. 8.5ha respectively. Wilderspool grew into a large settlement perhaps covering as much as 30ha, comprising stone buildings, quayside buildings and an industrial zone to the south at Stockton Heath. Although all three sites had a primarily linear pattern of development, the largest, Wilderspool, appears to have had a more developed street pattern. The lack of a confirmed military presence at these sites, combined with a certain regularity of street pattern and Romanised buildings (particularly at Wilderspool from where large stone architectural fragments have been recovered) has led to their identification as 'small' Roman towns (Petch 1987).

Within this group of Romano-British settlements we can distinguish between those associated, at least initially, with a Roman fort (Castleshaw, Chester, Manchester, Melandra, Middlewich and Northwich) and those without an obviously direct military origin (Heronbridge, Meols, Nantwich, Tilston and Wilderspool). Most of those settlements attached to the Roman forts of the region, which traditionally have been identified as vici, began in the AD 70s and AD 80s and ranged in size from c. 3ha to c. 7.5ha. Dominated by industrial activities linked to the Roman forts (usually metal working and pottery manufacture) the buildings within these settlements were arranged along one or two roads leading from the fort gateways, producing a characteristic strip pattern of development. Three of these sites were abandoned in the mid-second century AD when their forts were dismantled (Castleshaw, Melandra and Northwich: Redhead 1999; Webster 1971; Petch 1987). The other three sites appear to have continued in occupation throughout the Roman period. However, whilst Manchester, along with the more northerly vici at Burrow-in-Lonsdale, Lancaster and Ribchester, grew to only c. 7.5ha (Walker 1986), those at Chester and Middlewich became much larger settlements. The settlement that grew outside the walls of the legionary fortress at Chester, usually referred to as a canabae, grew to over 50ha in area, the largest such site in the region. It had a complex grid of streets to the east, south and west of the legionary fortress and imposing structures such as the amphitheatre and bath complex. At Middlewich the

Meols is the most unusual of these purely civilian settlements, since its origins lie in the late prehistoric period. Lying on the northern coast of the Wirral peninsula, the site has been largely washed away by coastal erosion and survives mostly as a large collection of finds, spread along nearly 1km of coastline west of Dove Point. These have recently been reviewed by Matthews ( 1996), who proposed a long period of occupation from the 63

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

middle of the first millennium BC to the late medieval period. The finds assemblage from the site is the most exotic from any of the late prehistoric settlements of the region, consisting almost entirely of coins, which elsewhere only occur as stray finds. These include two silver coins of the Coriosolites, a tribe based in northern Brittany, three Carthaginian drachmas of the third and second centuries BC and a very worn gold coin of uncertain Celtic origin. In addition there are two swanneck pins of typical Iron Age form (Longley 1987: 104). The Roman material is less exotic, but far more extensive, including over 63 coins, metalwork and quantities of pottery including amphora fragments.

pig bones in Phases II and IV, and rotary quern fragments from Phase III hint at a mixed farming economy. The final phase of activity at Great W oolden Hall (Phase IV) was represented by second century AD, local, Romano-British wares from the plough soil and from the final fill of the inner ditch; this latter context also produced a radio-carbon date of AD 100-320 (cal AD 210 +/- 110, GrN 16851). The gap between Phases III and IV may suggest a hiatus in occupation, at least in this part of the enclosure. How far Great W oolden Hall and the other extensively investigated enclosures in the Mersey Basin (Brookhouse, Court Farm, Irby, Legh Oaks II and Mellor) genuinely reflect the late prehistoric and Romano-British rural settlement pattern is unclear. The double or single ditched compound of less than 2ha in area, often curvilinear in plan form and containing one or more circular buildings, appears to have been the most common form of settlement type in the centuries immediately preceding the Roman conquest. However, larger, true hillfort-type settlements were known (although all appear to have been abandoned by the first century BC) and unenclosed sites may also have existed. Therefore, it seems probable that during the late prehistoric period the Mersey Basin lay on the interface between the main settlement types of the Iron Age. A hillfort dominated zone lay to the west and south, villages and open settlements to the south-east, and enclosed homesteads of the north and north-east. Although the number of sites so far recovered is too low to give anything other than an indication of potential settlement densities, a concentration of 12 cropmark and excavated enclosure sites around Warrington in the lower Mersey Valley (Collens 1999; Nevell 1999a; Cowell and Philpott 2000) suggests that we may be dealing with intensive valley occupation in localised areas.

Late Prehistoric Settlement Trends

What does this growing body of late prehistoric and Romano-British settlement data reveal about the social structures of this era? In particular, is it possible to see any evidence for social marginality by contrasting the evidence for the two periods? One of the best known of the late prehistoric enclosed farmstead sites is Great W oolden Hall (SJ 691 936), a promontory, double-ditched enclosure in the Glazebrook valley between Salford and Warrington discovered by aerial photography by the editor and excavated by GMAU in 1986-8. Until the results from Irby are published and more sites are extensively excavated the finds and overall phasing of this site are our best guide as to the trends likely to be visible on the other enclosure sites of the Mersey Basin (Nevell 1989, 1992, 1999c). The earliest activity at Great W oolden was represented by a small assemblage of flint recovered from fieldwalking activities over the enclosure and from the excavations themselves. This material would seem to fit a date sometime in the late Neolithic or early Bronze Age, although it is not clear whether this activity was little more than ephemeral.

The issue of the nature of this late prehistoric society has begun to be addressed over the last decade (Higham 1993; Matthews 1994; Nevell 1999b; Cowell and Philpott 2000). In particular Keith Matthews has proposed an anthropologically based model of settlement hierarchy, where the size of the social grouping is directly reflected in the size of settlements in the landscape (Matthews 1999). Thus, the 60 plus enclosures of the late prehistoric and Romano-British period confirmed and suspected within the Mersey Basin, fall into four broad size groupings which can be fitted into Matthews' model. These bands are sites between O.lha and 0.4ha, sites between 0.4ha and 0.9ha, 0.9ha to 2ha and sites larger than 2ha, perhaps corresponding to Matthews, seventh to fourth levels of settlement; that is small and large family farmsteads, hamlets and villages. Intriguingly, by the late Iron Age all the traditional style hillfort settlements, those to be associated with Matthews' fourth level of settlement which performed as central places, had all been abandoned, leaving the landscape dominated by sites such as Great W oolden Hall and Irby, which were arguably large family

The major period of activity (Phases II to IV) were associated with the ditches of the enclosure, which appears to have begun in the latter part of the first millennium BC. This took the form of four structural episodes spanning the first century BC to the late second/early third century AD, starting with a series of rectangular pits in Phase I. A ditched compound containing a hut circle was designated Phase II, being succeeded by an oval, palisaded compound, with a hut, in Phase III (two circular features were located elsewhere within the enclosure by geophysical survey and it is possible that these may represent other structures from Phases II and III). This was finally replaced by a further series of pits in Phase IV. Phases II and III were dated, by radio-carbon samples, to 120 cal BC - cal AD 80 (40 BC +/- 25, GrN 16849) and 65-15 cal BC (20 BC+/- 100, GrN 16850). The acidic conditions of the site meant that very little palaeo-environmental material survived. However, the presence of burnt sheep bones in Phases II and III, burnt

64

Mike Nevell: The Edge of Empire

farmsteads. Such a shallow hierarchy in the late Iron Age, with only two or three levels of settlement, would appear to fit Kosse's definition of a simple chiefdom (Kosse 1990). This may even suggest that the Mersey Basin, if not the whole of the North West, had been 'taken over' by one or more external force(s) in this period (Higham 1987). Therefore, immediately prior to the Roman conquest, North West England, and the Mersey Basin in particular, was characterised by a shallow settlement hierarchy comprising ditched, often oval, enclosed farmsteads, with concentrations of such sites along several major river valleys of the area (the Irwell, Mersey and Sankey). This gave the region its own unique late prehistoric character.

Matthews have stressed the impact of the supply needs of the Roman army on the native population during from the AD 70s to the AD 150s, when military numbers peaked, and this may be reflected in the upsurge in clearance activity noted in the palaeo-environmental evidence and in the increase in absolute farm numbers suggested by the archaeological material (Higham 1993; Mathews 1999; Cowell and Philpott 2000). A similar impact on the numbers of native farmsteads has been observed in Cumbria (Higham and Jones 1985; Jones 1999b). A new element to emerge in the rural settlement pattern during this period was the villa, or Romanised farmstead. Whilst the winged building at Eaton-by-Tarporley remains the only fully excavated example in the region the remains of stone structures associated with plaster, tile and hypocaust fragments at Crewe Hall and Tattenhall strongly suggest these were also villa sites, whilst the fragmentary remains of Roman-period stone buildings are known from Daresbury, Frodsham and Kelsall, and such structures strongly suspected at Ashton and Poulton. Higham is surely right in suggesting that the Eaton-by-Tarporley villa represents the home of a member of the local indigenous elite who adopted the Roman life-style (Higham 1993). Eaton does not appear to have had any Iron Age antecedents but we should perhaps expect to discover these on the other probable villa sites in central and western Cheshire. Their existence may be linked to the Chester legionary fortress and its canabae. Doubt remains about the extent of its prata legionis, although a boundary in the east along the River Gowy seems highly likely and the settlement at Heronbridge may well have lain outside (Higham 1993, Mason 1987, Matthews 1994, Petch 1987). How much of the Wirral Peninsula lay within its bounds is unknown. The impact of c. 6000 legionaries and the population of the canabae might be expected to be visible in an increase in the numbers of farms around the legionary fortress. Mathews has argued strongly that the prata legionis alone would have been capable of feeding the garrison (Matthews 1999). Ironically, none of the newly identified enclosures and villa/potential villa sites lie within 10km of the fortress, although the villa/potential villa sites do form a ring beyond this limit to the south and east which might be significant in defining Chester's prata legion is.

Romano-British Settlement Trends

The settlement pattern of the Romano-British period is far more complicated than the shallow hierarchy of the late prehistoric era. However, the dominant form of settlement type in the region during the Roman era remained the defended enclosure. By the end of the 1990s, 19 ditched enclosures of less than 2ha in area, interpreted as farmsteads, had been positively identified as RomanoBritish through excavation, of which eight had late prehistoric origins (Table 1). These 19 enclosures had a single ditch, usually enclosing a rectangular compound which contained one or more buildings. On some of those sites which had a late prehistoric origin (Great Woolden, Irby and Mellor) a transition from the Iron Age tradition of circular building to Roman-influenced rectangular building could be seen and follows a pattern already visible on many native sites in Cumbria and Northumberland (Higham 1986). Few palaeo-environmental remains have been excavated on these sites so it is still not possible to say with certainty what their economic base was, although mixed farming is indicated at Court Farm, Irby and Great W oolden. There is a particular lack of such remains from upland sites in the Mersey Basin, so as yet there is no evidence to support the other palaeo-environmental material which suggests an expansion of cereal agriculture in the 100m to 250mAOD zone (see above). However, the recently discovered site at Mellor, which lies a c. 220m AOD, may start to fill this gap. Even though the number of confirmed Romano-British rural farmsteads, and the extent of the excavation within these sites, remains low, the archaeological evidence hints at an expansion in settlement sites, and thus population, in the first two centuries of the Roman occupation. Seven sites appear to have only Roman activity. So far only Irby has produced evidence for occupation throughout the Roman period, the other five exclusively Roman sites appearing to fall out of use by the early third century AD. This expansion probably took place on the lighter soils of the region: the known and possible enclosures concentrate along the major river valleys of the Mersey Basin (the Bollin, Dee, Gowy, Irwell, Mersey, Sankey and Weaver) and along the sandstone ridges of the central Cheshire ridge and to the north of Warrington. Both Higham and

Whilst the ditched farmsteads can be seen as has having clear linkages with the late prehistoric settlement pattern, the large nucleated sites that emerged in the late first century AD in the Mersey Basin, and elsewhere in North West England, were new features of the landscape. These sites can be fitted into the settlement hierarchy seen elsewhere in the province of Britannia during the Roman period (Ringley 1989). The vici attached to the Roman forts of the Mersey Basin were, like their counterparts elsewhere in northern Britain, dependant on the military presence for their existence. Those at Castleshaw, Melandra and probably Northwich, for instance were dismantled when their parent forts were abandoned. Others such as Manchester and, further north in the region, 65

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Ribchester and Lancaster, remained occupied throughout the life of their forts, although in all three vici there was a decline in activity from the early third century onwards (Walker 1986). The relationship of these extra-mural settlements to the native rural population is unclear, but elsewhere some vici appear to have acted as a focus for the local population. Hints of such a relationship can be seen in the palaeo-environmental material from the Castleshaw valley, which indicates substantial improvement in the local upland pasture around the fort during the Roman occupation (Brayshay 1999; Redhead 1999).

produce evidence for Romanised civilian settlements between the farmstead and local centre level, i.e. settlements that covered c. 3ha to c. 6ha of a type which have been interpreted elsewhere in the province as large hamlets or villages (Ringley 1989). Nor have any potential villa sites yet been located beyond a radius of c. 25km from Chester, strongly suggesting that the influence of this type of farming was limited to the hinterland of the fortress. The auxilliary forts and their vici to the east and north of Chester may have occupied this apparent gap in the native settlement hierarchy. The obvious parallel is with Cumbria and Northumberland, where the needs of the local Roman garrisons probably stimulated growth in the local rural economy, which is visible archaeologically as an increase in the number of farmstead sites in the Roman period (Higham 1986).

The coastal site at Meols remains enigmatic but Matthews has argued persuasively that the site should be viewed not just as an ordinary agricultural settlement or even as a small port but as one of the series of late prehistoric and Romano-British emporia, or trading sites, known from the English Channel and the Irish Sea with strong trading links with the Iron Age tribes of Brittany and, later, the western Mediterranean (Matthews 1994: 16-8). Meols potentially had a unique role in the region as a facilitator of international trade throughout both periods and is one of only a handful of sites in North West England where fifth and sixth century settlement activity can be identified.

The limited nature of the impact of the Roman conquest and occupation on the Mersey Basin can be seen in the chronological spread of the introduction of new monument types. I will take the definitions offered by the Thesaurus of Monument Types published by English Heritage and the Royal Commission on Historic Monuments (England) (1996) as a point of reference. Of the 47 new monument types introduced into the area between c. 250 BC and c. AD 250, the vast majority were of Roman origin and were established between c. AD 70 and c. AD 125. This period marks the peak in the Roman garrison of North West England. Whilst such a short period of innovation might be the product of two centuries of research focussed almost exclusively on the Roman military presence in the region, it seems highly likely that there is a deeper link.

Of the larger nucleated Roman settlements at Chester, Heronbridge, Middlewich, Tilston and Wilderspool, only Chester and Middlewich started as vici. The Roman fort at Middlewich appears to have been abandoned by the end of the first century (Rogers 1995, 1996; Shorter 2000b), although the settlement there expanded rapidly to become, at c. 30ha, probably the second largest in the region after Chester, which itself covered an area of over 50 ha. The other sites appear to have developed as ribbon developments either at crossroads or river crossings and all appear to have been occupied into the fourth century AD. These sites have been well studied, in comparison with the native-style farmsteads of the Mersey Basin, and fit within the wider settlement pattern of the province. Most of these settlement can be characterised as local, Romanised, market and industrial settlements (Ringley 1989). The exception is the largest settlement in the region, Chester, with its close association with the legionary fortress. This settlement almost certainly had a wider regional administrative function, although it is unclear whether it reached colonia status in the late Roman period (Higham 1993; Mason 1987). Nevertheless, this idea has received slight additional support in recent years with the reinterpretation of two fourth-century inscriptions on lead brine pans from Nantwich which may now refer to a bishop, presumably resident at Chester (Mathews 1999; Petch 1987).

Conclusion

Barri Jones warned us to be careful of dismissing Roman North West England as a cultural backwater, simply because it lacks the large-scale sites and material culture characteristic of other parts of the Roman Empire (Jones 1999b). His own work in frontier regions such as northern Cumbria/south-western Scotland and Tripolitania in Africa demonstrates the unique character of these border areas, where the interplay between two different cultures was accentuated by the marginal nature of the local topography, climate and social structure. In this context, and in view of the evidence from Great Woolden Hall and Irby (Nevell 1999c; Philpott and Adams 1999), Matthews has argued against the assumption that the rural communities of the North West were too poor to purchase the new material culture of the Roman period, (Matthews 1996, 1999). What is clear is that the durable cultural remains (pottery and structures) are more limited in their scope and numbers than on many other rural sites of the later prehistoric period in northern Britain. Indeed, this trend continued in to the Romano-British period and the sub-Roman and early medieval periods (Higham 2000; Nevell 1999c: 59-61; Philpott and Adams 1999: 70-1; Cowell and Philpott 2000). With the late prehistoric finds

Despite the greatly expanded range of site types visible in the Roman period, there remain gaps in the settlement hierarchy which suggest that the underlying late prehistoric settlement pattern, and by implication the native social structure, was not substantially altered by the Roman presence. For instance, research in the region has yet to

66

Mike Nevell: The Edge of Empire

record dominated by a few pottery types (Nevell 1994), and with only sparse evidence for the use of metalwork and Roman pottery during the Romano-British era, there has been a temptation to assume a lower level of material culture than perhaps is warranted (Higham 1993; Kenyon 1991). What appears to be at fault is not the archaeology but our own explanations. Our models of cultural usage and exchange in the late prehistoric period, which have been based upon the concept of economic need and cultural imperialism during the Romano-British period (Higham 1993; Petch 1987; Thompson 1965), need to be re-evaluated. For instance, ethnographical parallels have shown how many small-scale pre-industrial societies used material culture as a means of constructing and reinforcing individuality rather than as an expression of economic need (Hodder 1992; Weissner 1984). Thus, if we look at the issue of the apparent 'paucity' of portable finds such as pottery from Great W oolden Hall and Irby, we find that anthropological models suggest two main types of exchange mechanism. One is subsistence exchange, often referred to as socially disembedded trade, which was concerned with everyday needs. Additionally, there is ceremonial or gift exchange, often termed socially embedded trade, which was concerned primarily with strengthening social ties through gift-partnership, exchange cycles, tribute, and diplomatic exchanges. In other words, exchange was often for reasons other than profit, and this may be expressed in the composition of a finds assemblage. The wider implication for the study of rural settlement in the North West is that we should be seeking models that address the issues raised by the regional evidence.

from an agrarian to an industrial society when the region became internationally important.

BIBLIOGRAPHY

Bartley, B. B., 1975. Pollen analytical evidence for prehistoric forest clearance in the upland area west of Rishworth, West Yorkshire. New Phytologist 74: 375-81. Beswick, P. and Coombs, D.G., 1986. Excavations at Portfield, hillfort, 1960, 1970 and 1972. In Manby, G. and Turnbull, P. (eds.) Archaeology of the Pennines, Studies in Honour of Arthur Raistrick. Oxford: British Archaeological Reports, British Series 158: 137-80. Brayshay B., 1999. Palaeoenvironmental evidence for marginality in the upper Mersey Basin. In Nevell (ed): 82-89. Birks, H.J. B., 1964. Chat Moss, Lancashire. Memoirs and Proceedings of the Manchester Literary and Philosophical Society 106: 1-24. Birks, H.J.B., 1965. Pollen analytical investigations at Holcroft Moss, Lanes, and Lindow Moss, Cheshire. Journal of Ecology 53: 299-314. Carrington, P., (ed.) From Flints to Flower Pots. Current Research in the Dee-Mersey Region. Chester: Cheshire City Council Archaeological Service Occasional Paper No 2 Collens, J., 1994. Recent Discoveries from the air in Cheshire. In Carrington, P., (ed.). 19-25. Collens, J., 1999. Flying on the Edge: Aerial Photography and Setlement Patterns in Cheshire and Merseyside. In Nevell, M. (ed.): 36-40. Coombs, D.G.,1982. Excavations at the hillfort of Castercliff, Nelson, Lancashire 1970-1. Transactions of the Lancashire and Cheshire Antiquarian Society 81 : 111-30. Coombs, D.G. and Thompson, F.H., 1979. Excavations of the Hillfort of Mam Tor, Derbyshire, 1965-69. Derbyshire Archaeological Journal 99: 7-51. Cowell, R. W. and Innes, J. B., 1994. The Wetlands of Merseyside. Lancaster: Lancaster University Archaeology Unit, Lancaster Imprints 2. Cowell, R. W. and Philpott, R. A., 2000. Prehistoric, Romano-British and Medieval Settlement in Lowland North West England. Archaeological excavations along the A5300 road corridor in Merseyside. Liverpool: National Museums and Galleries on Mersuside and English Heritage. Crowe, P.R., 1962. Climate. In Carter, C. F., (ed.). Manchester and its Region. Manchester: Manchester University Press: 17-46. Cunliffe, B., 1991. Iron Age Communities in Britain: An Account of England, Scotland And Walesfrom the Seventh Century BC until the Roman Conquest. Third Edition, London: Routledge. Dark, K. and Dark, P., 1997. The Landscape of Roman Britain. Frome: Sutton Publishing.

The research undertaken during the 1990s has demonstrated that in the late prehistoric and RomanoBritish period North West England was marginal in many ways. Geographically, climatic changes meant that the Pennine foothills were agriculturally marginal for much of the late prehistoric period, but not during the Roman era, whilst highly local conditions in the valleys could favour subsistence farming in most centuries. Politically, in the late prehistoric period the area was marginal in as much as it lay on the boundary between at least three tribal groupings; the Brigantes, Deceangli and Cornovii. This perhaps explains an emerging settlement hierarchy that lacks large central places as foci of power and status. In the Roman period the Mersey Basin and North West England remained politically marginal, with power concentrated in the south-eastern part of the province and in the heart of the Empire, Rome. Economically, however, it was not marginal, since there is growing evidence for an expansion in agriculture during this era in both the lowland and upland areas of the region. North West England is thus an area that historically has been a marginal or transitional region. An understanding of the interplay between the political, economic and geographical forces in this area during the Roman era has much to contribute to our understanding of such zones elsewhere in the Roman Empire, as well as during other periods of cultural transition within the region itself Such include the shift 67

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Ellis, P., (ed.), 1993. Beeston Castle, Cheshire. A Report on the Excavations 1968-85 by Laurence Keen and Peter Hough. London: Historic Buildings & Monuments Commission for England (English Heritage). Forde-Johnston, J., 1962. The Iron Age Hillforts of Lancashire and Cheshire. Transactions of the Lancashire and Cheshire Antiquarian Society 72: 9-46. Godwin, H. and Willis, E. H., 1960. Cambridge University natural radio-carbon measurements. II. American Journal of Science, Radiocarbon Supplement 2, 1960: 62-72. Hall, D., Wells, C. E. and Huckerby, E., 1995. The Wetlands of Greater Manchester. Lancaster: Lancaster University Archaeology Unit, Lancaster Imprints 3. Hart, C. R., 1984. The North Derbyshire Archaeological Survey to AD 1500. Sheffield: Derbyshire Archaeological Society. Higham, N. J., 1980. Rural settlement west of the Pennines. In Branigan, K., (ed.), Rome and the Brigantes. Sheffield: Sheffield University: 41-7. Higham, N.J., 1987. Brigantia Revisited. Northern History 23: 1-19. Higham, N. J., 1993. The Origins of Cheshire. Manchester: Manchester University Press. Higham, N.J., 2000. The Tatton Park Project, Part 2. The Medieval Estates, Settlements and Halls. Journal of the Chester Archaeological Society 75 (for 1998-99): 61-133. Higham, N.J. and Cane. T., 1999. The Tatton Park Project, Part 1: Prehistoric to sub-Roman settlement and landuse. Journal of the Chester Archaeological Society 74 (for 1996-7): 1-61. Higham, N. J. and Jones, G.D.B., 1985. The Carvetii. Gloucester: Alan Sutton. Ringley, R., 1989. Rural Settlement in Roman Britain. London: B A Seaby. Hodder, I. R, 1992, Theory and Practice in Archaeology. London: Routledge. Jones, G.D.B., 1979. The future of aerial photography in the North. In Higham N J, (ed.). The Changing Past. Manchester: Manchester University ExtraMural Department: 75-87. Jones, G. D. B., 1999. Conclusion: Marginality, Their Fault or Ours? A Warning from the Cumbrian Evidence. In Nevell, M. (ed.): 90-6. Kenyon, D., 1991. The Origins of Lancashire. Manchester: Manchester University Press. Kosse, K., 1990. Group size and societal complexity: thresholds in the long-term memory. Journal of Anthropological Archaeology 9: 275-303. Lamb, H.H., 1982. Climate, History and the Modern World. London: Methuen. Leah, M., Wells, C. E., Appleby, C. and Huckerby, E., 1997. The Wetlands of Cheshire. Lancaster: Lancaster University Archaeology Unit, Lancaster Imprints 5.

Longley, D. M. T., 1987. Prehistory. In Harris, B. E., (ed). The Victoria History of the County of Chester, vol 1. London: University of London Institute of Historical Research: 36-114. Mason, D. J. P., 1987. The prata legionis at Chester. Journal of the Chester Archaeological Society 69 (for 1986): 19-43. Matthews, K., 1994. Archaeology without artefacts: the Iron Age and sub-Roman period in Cheshire. In Carrington, P., (ed.): 51-62. Matthews, K., 1996. Iron Age sea-borne trade in Liverpool Bay. In Carrington, P., (ed.). Where Deva Spreads her Wizard Stream. Trade and the Port of Chester. Papers from a Seminar held at Chester, November 1995. Chester: Cheshire City Council Archaeological Service Occasional Paper No 3: 12-23. Matthews, K., 1999. Rural Settlement in Roman Cheshire: a Theoretical View. In Nevell, M. (ed.): 27-34. Millett, M., 1990. The Romanization of Britain: an essay in archaeological interpretation. Cambridge: Cambridge University Press. Nevell, M.D., 1989. Great Woolden Hall: Excavations on a Late Prehistoric/Romano-British Native Site. Greater Manchester Archaeological Journal 3 (for 1987-8): 35-44. Nevell, M .D., 1992. Settlement and Society in the Mersey Basin c. 2,000 B.C. to c. A.D. 400. A Landscape Study. Manchester: Unpublished PhD, Department of Archaeology, Manchester University. Nevell, M. D., 1994. Late prehistoric pottery types from the Mersey Basin. In Carrington, P., (ed: 33-41. Nevell, M. D., 1999a. Iron Age and Romano-British Rural Settlement in North West England: marginality, theory and settlement. In Nevell, M. (ed.): 14-26. Nevell, M. D., (ed.), 1999b. Living on the Edge of Empire: Models, Methodology and Marginality. LatePrehistoric and Romano-British Rural Settlement in North-West England. Manchester: CBA North West with the Field Archaeology Centre (University of Manchester) and Chester Archaeology. Nevell, M. D., 1999c. Great Woolden Hall: A Model for the Material Culture of Iron Age and RomanoBritish Rural Settlement in North West England? In Nevell, M. (ed.): 48-63. Ogle, M., Robinson, M. and Shimwell, D., 1997. A Palaeoenevironmental Assessment of a Small Basin Mire at Godley Hill, Hyde, Greater Manchester. Palaeoecological Research Unit, Manchester University: unpublished client report. Parry, M. L., 1975. Climatic Change, Agriculture and Settlement. Chatham: Dawson Archon Books. Petch, D., 1987. The Roman Period. In Harris, B. E., (ed.) The Victoria History of the Count of Chester, vol 1. London:University of London Institute of Historical Research: 115-236.

68

Mike Nevell: The Edge of Empire

Philpott, R., 1994. The implications oflrby. In Carrington, P., (ed.): 26-32. Philpott, R. and Adams, M., 1999. Excavations at an Iron Age and Romano-British Settlement at Irby, Wirral, 1987-96: An Interim Statement. In Nevell, M. (ed): 64-73. Pollard, S., 1997. Marginal Europe. The Contribution of Marginal Lands Since the Middle Ages. Oxford: Clarendon Press. Redhead, N., 1999. Edge of Empire: Extra-Mural Settlement in a Marginal Context: Roman Castleshaw. In Nevell, M. (ed): 74-81. Rogers, 1.,1995. Gifford Archaeological Service: Recent Work in the North West 1993-4. Archaeology North West 9 (Vol 2, Part III): 70-1. Rogers, I., 1996. The Conquest of Brigantia and the development of the Roman Road System in the North-West. Britannia 27: 365-8. Schoenwetter, J., 1982. Environmental archaeology of the Peckforton Hills. Cheshire Archaeological Bulletin 8: 10-2. Shorter, D.C.A., 2000a, Chester. The Evidence of Roman Coin Loss. Journal of the Chester Archaeological Society 75 (for 1998-99): 33-50. Shorter, D.C.A., 2000b, Middlewich. The Evidence of Roman Coin Loss. Journal of the Chester Archaeological Society 75 (for 1998-99): 51-60. Tallis, J. H. and McGuire, J., 1972. Central Rossendale: the evolution of an upland vegetation. Journal of Ecology 60: 721-37. Tallis, J. H. and Switsur, V. R., 1973. Studies on Southern Pennine Peats VI. A radio-carbon dated pollen diagram from Featherbed Moss, Derbyshire. Journal of Ecology 61: 743-51. Thompson, F.H., 1965. Roman Cheshire. Chester: Cheshire Community Council. Varley, W.J., 1976. A Summary of the Excavations at Castle Hill, Almondbury, 1939-72. In Harding, D.W., (ed.). Hillforts; Later Prehistoric Earthworks in Britain and Ireland. London: Academic Press:119-292. Walker, J. S. F., (ed.), 1986. Roman Manchester - A Frontier Settlement. Manchester: Greater Manchester Archaeological Unit. Webster, P. V., 1971. Melandra Castle Roman Fort: Excavations in the Civil Settlement, 1966-69. Derbyshire Archaeological Journal 91: 58-118. Weissner, P., 1984. Reconsidering the behavioural basis for style: a case study among the Kalahari San. Journal of Anthropological Archaeology 3: 190234.

69

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

5 .1 Beeston Castle, a Late Bronze Age to Middle Iron Age hillfort in the central Cheshire Ridge and the largest late prehistoric site in the Mersey Basin. Photograph: Dr N.J. Higham.

70

Mike Nevel!: The Edge of Empire

40

20

00

5.2: The Iron Age Settlements ofNorth West England Key: Circles= hillforts (enclosures above 3ha); squares= farmsteads (enclosures belwo 2.5ha in area). Open symbols= possible sites. (1) Castercliffe; (2) Portfield; (3) Dutton's Farm; (4) Castlesteads; (5) Rainsough; (6) Brookhouse; (7) Mellor; (8) Legh Oaks I; (9) Eddisbury; (10) Kelsborrow; (11) Maiden Castle. 0100m

~Okm

~uAfi)'

170

J F. Healey: 'Romans Always Conquer'

Healey, J. F., 1989. Were the Nabataeans Arabs? Aram 1: 3-44. Healey, J. F., 1993. The Nabataean Tomb Inscriptions of Mada 'in Salih. Oxford: Oxford University Press. Howard, G., 1981. The Teaching of Addai. Chico, California: Scholars Press. Hutchinson, J. and Smith, A D. (eds.), 1996. Ethnicity. Oxford: Oxford University Press. Ibrahim, M. (ed.), 1989. Arabian Studies in honour of Mahmoud Ghul. Wiesbaden: Harrassowitz. Jenkins, R., 1996. Social Identity. London: Routledge. Jobling, W. J., 1993. The 'Aqaba-Ma'an Archaeological and Epigraphic Survey 1988-1990. Syria 70: 244-48. Jobling, W. J. and Tanner, R. G., 1989. Zeno the Tribune. Studi Epigrafici e Linguistici 6: 135-39. Jones, G.D. B. 1997. From Brittunculi to Wounded Knee: a study in the development of ideas. In Mattingley (ed.): 185-200. Jones, S., 1997. The Archaeology of Ethnicity. London: Routledge. Knauf, E. A, 1985. Ismael. Untersuchungen zur

A D., 2000. The Nation in History. Historiographical Debates about Ethnicity and Nationalism. Oxford: Polity Press. Ross, S. K., 2001. Roman Edessa. London: Routledge. Smith,

Tanner, R. G., 1990.

Greek Epigraphy in South Jordan.

Zeitschri.ftjii,r Papyrologie und Epigraphik 83: 183-93. Wenning, R., 1993. Eine neuerstellte nabataischen Dynastie. Boreas 16: 25-38.

Geschichte Paliistinas und Nordarabiens im 1. Jahrthausend v. Chr. Wiesbaden: Harrassowitz. Knauf, E. A, 1986. Die Herkunft der Nabatiier. In Lindner (ed.): 74-86. Knauf, E. A, 1989. Nabataean Origins. In Ibrahim (ed.): 56-61. Langdon, S., 1927. The "Shalamians" of Arabia. Journal of the Royal Asiatic Society 1927: 529-33. Lindner, M. (ed.), 1986. Petra. Neue Ausgrabungen und Entdeckungen. Munich: Delp. Macdonald, M. C. A, 1991. Was the Nabataean Kingdom a "Bedouin State"? Zeitschri.ft des deutschen Paliistina-Vereins 107: 102-19. Mattingly, D. J. (ed.), 1997. Dialogues in Roman

Imperialism. Power, discourse, and discrepant experience in the Roman Empire. Portsmouth, Rhode Island: Society for the Promotion of Roman Studies(=

Journal of Roman Archaeology Supplementary Series no. 23). Milik, J. T., 1972. Dedicaces faites par des dieux (Palmyre, Hatra, Tyr). Paris: Geuthner. Millar, F., 1987. Empire, Community and Culture in the Roman Near East: Greeks, Syrians, Jews and Arabs. Journal of Jewish Studies 38: 143-64. Millar, F., 1993. The Roman Near East 31 BC - AD 337. Cambridge, Mass.: Harvard University Press. Negev, A, 1977. The Nabateans and the Provincia Arabia. Aufstieg und Niedergang der romischen Welt II Principat, 8, 520-686. Negev, A, 1991. Personal Names in the Nabatean Realm. Jerusalem: Institute of Archaeology of the Hebrew University. Parr, P. J., Harding, G. L., Dayton, J. E., with contributions by Beeston, A F. L., and Milik, J. T. 1971. Preliminary Survey in N. W. Arabia, 1968. Bulletin of

the Institute of Archaeology of the University of London 10: 23-61. Segal, J. B., 1970. Edessa, the Blessed City. Oxford: Oxford University Press.

171

Liste

der

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

16.1

The 'Romans always conquer' inscription (the late W. J. Jobling - by courtesy ofM. Bannigan)

172

17 Khirbet Khaw: a Roman town and fort in northern Jordan 1 David Kennedy

Traiana, just a little over ten kilometres to the north-northeast.

Abstract Khirbet Khaw was known to a succession of travellers in northern Jordan between 1876 and 1930. Then it was incorporated within the perimeter of the military base at Zerqa and effectively lost to scholarship. Early reports were vague and gave little real impression of the scale and character of the site. In 1994 examination of old air photographs revealed Khirbet Khaw as of more than usual significance: a small town of about 5.5 ha incorporating a fort and 'caravanserai'. Interpretation of the air photographs and preliminary fieldwork enables us now to add the site to the list of small Roman towns in northern Jordan and adds a previously unknown fort of, probably, c. AD 300.

The site today may be located (Illus.17 .2) by following Highway 30 north-eastwards out of Zerqa. About two kilometres after crossing the railway tracks there is a line of three small hills on the north side of the road. The ruins lie on the last of these, the most easterly. Despite its size, evident quality of preservation and a number of large structures, in 1994 it was scarcely known to scholars. Early Exploration: 1876-1930 The earliest notice of the site relates to the travels and explorations in the region of the American scholar Selah Merrill in 1876. In the book he published five years later (1881: 271), he reported that while at Zerqa: 'The keeper of the castle[= Kala'at Zerqa] says there is a large ruin to the east, half an hour or an hour distant, called Hau or Khau.' Later, while still at Zerqa, and in the context of information received about Azraq, Merrill added that: '... a young man ... said the only ruin he knew anything about, lying near the castle to the east [seemingly Qasr elAzraq rather than Qasr el-Hallabat], was one called Khau' (1881: 398).

Introduction Northern Jordan is rich in Roman military sites (Illus. 17.1)2. Well-known are the striking and well-preserved forts at places in the steppe and desert - Qasr el-Azraq, Deir el-Kahf and Qasr el-Hallabat are obvious examples. Now,too, we have evidence for forts in some of the ancient villages and small towns of the region - Kh. es-Samra, Umm el-Jimal and Umm el-Quttein are important and relatively recent discoveries. There are others and we may expect still more to be discovered. The purpose of this article is to revisit a site first recorded in the late nineteenth century. Despite contradictions and unsatisfactory impressions in these early records, it was described in terms from which one could infer extensive ruins and probable significance. It is a surprise, therefore, that it largely disappeared from even scholarly works and, to a large extent, even from sight. It is in fact still highly visible though not normally accessible.

Seventeen years later, in 1893, George Robinson Lees reported in an article in The Geographical Journal that after an hour of travelling - that is, c. four Roman miles, north from Zerqa: ' ... we saw a ruin to the south-east perched on one of the hills about an hour away, possibly the remains of some fortress - an outpost probably, the bottom part of which seemed vaulted, as half of it looked towards us dark and hollow. The hills to the east are in many places huddled close together, with very little space between them ... ' (1895: 12).

In 1994, during a systematic interpretation of air photographs taken in 1953, the remains ofa substantial site just north-east of Zerqa were noted and subsequently identified with the place-name 'Khirbat Khaw' on the 1: 50,000 map and 'Kh. Haw (Gadda)' on the Archaeological Map of Jordan. 3 Further examination of the remains as revealed on the photographs suggested a small town of c.5.5 ha. It is about the same size and has a similar degree of preservation as the well-explored contemporary site of Khirbet es-Samra on the Roman highway, the Via Nova

For Merrill and Lees, Khaw was a site mentioned but not seen, or seen only from a distance. The first recorded visitors to the site itself came five years later. In 1898 the two great German scholars Rudolph Briinnow and Alfred von Domaszewski explored the ruins and published some detailed observations of what they called 'Hau' (1905: 224-225): 'The remains of a fairly large town located on a slight rise; there are many cisterns around. Although crude, the walls of the houses are well-built (Figs 843, 845); here and there one sees squared blocks. On the highest point lies a betterpreserved building, a temple perhaps; occasional good quality pieces, cornices, door frames as well as an architrave piece and large column drums of almost 2m diameter are lying around. I have also noticed a millstone and a sarcophagus. Nearby is an underground grave cavern

1 A brief account of the site, based on an early draft of the present article, can be found in Kennedy 2000b, 95-6. 2 All the sites named here are discussed and illustrated fully in Kennedy 2000b: chs 7-10. 3 JADIS 2516.004; KHIRBET KHAU; Map 3254.III; UTM Zone 37; UTME 2294; UTMN 35543; PGE 257.100; PGN 166.800

173

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

with ornamented doors and a niche on the left. The back wall has two arched windows with a door between them which leads into a further chamber (Figs 846, 847). Similar caverns are constructed with pillars and walls; in another lies a simple architrave piece. These grave caverns have exactly the same form as the older grave chambers at Petra, only the smaller and coarser are well-worn; also one finds similar niches and loculi. As we shall see, therefore, an old Nabataean town is likely here - probably the town of Gadda of the Tabula Peutingeriana. At least it seems that the numeral XIII after Philadelfia refers to the distance to Gadda, which concurs with the mileage to Hau; ...' The relevant volume of Briinnow and von Domaszewski's Provincia Arabia was published in 1905 and their description of an exciting site was known to the members of the Princeton Expedition to Syria which passed that way in 1909: 'East of Amman we followed the Roman road, by milestones, to Kal' at Zerka, and, from then, to Khirbit isSamra. At the 37th milestone, numbering from Bosra, we turned eastward from the Roman road to the ruins of Khau, which were plainly visible on a hilltop to the right. Khau is a deserted ruin of little interest. It seems to have been completely rebuilt, in a poor manner, since the rise of Islam. It has been visited by a number of explorers, as may be seen by reference to Professor Briinnow's Provincia Arabia . ... Khau is all of basalt, Zerka is of limestone; the dividing line between the volcanic and calcareous rock lies somewhere between the two places' (PES 111.A.2:App. xiv-xv). In the Geography and Itinerary volume Butler is even more dismissive: 'After leaving Kal'at iz-Zerka, still travelling on the Roman road, we made a short detour near midday and ate our lunch at Khau, an uninteresting ruin on a hilltop to the left [sic]' (PES I: 85). As the ruin is far from uninteresting even today, it is tempting to ask if they were visiting the correct place. Certainly, in view of the important record they preserved of other such sites, it is a misfortune not to have a detailed description by them. Twenty more years passed without reference. Then, in 1929, Gerald Lankester Harding, later to be Curator (later still Director) of Antiquities of Jordan, visited the site and left notes in his unpublished records still preserved in Amman in the files of the Department of Antiquities. They were apparently written after a visit on 24 August: 'Kh. Khow (Khau). Ruined village, caves, tombs, cisterns. Identified with Gadda. Badly damaged The[n] to Khow which lies east of [the] railway on the edge of [the] desert. There is now a road running parallel with [the] railway line; so it is easy to reach. It is called Gadda on the map. It is a fair[l]y extensive ruin and seems to have been a military station and has a castellum at the north end. No remains of arches are visible. The walls are laid dry and the stones are roughly squared. Byzantine Pottery lies on the surface. There is an interesting

columbarium on the west side. Digging seems to have been done recently in the grave yard round it.' The final published account was made by C. C. McCown, Director of the American School of Oriental Research in Jerusalem, in 1930 in the course of the field trip which he conducted each year (McCown1930: 18). One may guess his choice of route was influenced by the reports made by Harding the previous year: 'Next morning, following advice from Major Howard of the [Trans Jordan] Frontier Force, we diverged from the road northward [from their camp at Zerqa] to Khirbet esSamra, and following Armoured Car Track X across the railroad came to Khirbet Khau, which Briinnow and von Domaszewski identify with the illusory Gadda of the Peutinger Tables (sic). In the shallow valley to the south through which the road runs are artificial catch basins for water with cisterns at the center. The khirbeh [= ruined site] lies on a low rocky hill and consists of numerous small buildings with rough walls of poorly shaped stones. Its chief peculiarity is the large number of caves in the rock fitted out as abodes for shepherds. They have certainly been made over since the city was occupied in Byzantine times. Nothing has ever been found to determine the name or precise date of the ruins .... The strongest argument against identifying Khirbet Khau with Gadda is that the ruins are those of a town too insignificant to find a place on a road map of the Roman military roads. Either Khirbet es-Samra or Qoseir el-Hallabat, which I visited later, would be much more suitable.' For half a century, until 1930, Khirbet Khaw was visited by travellers and explorers. There was some significant disagreement between these visitors, none prepared a plan and only two photographs and three small drawings were published (Briinnow and von Domaszewski 1905: 225, figs. 843-847). But then visits ended and knowledge of the site effectively disappeared. Khirbet Khaw does not appear in any of the numerous publications of Nelson Glueck, based on fieldwork which largely took place in the 1930s onwards, and the site was virtually unknown, even by name, to many archaeologists. It is one of the least known major sites in Jordan: it has only a brief entry in the JADIS database (2516.004), is omitted from The Archaeology of Jordan volumes (AJ) and not listed in the Index to the first 30 volumes of Annual of the Department of Antiquities of Jordan. 4 The explanation is simple: in the 1930s the army, which already had a large base to the south of the site, beyond the modem road, also took over the area of Kh. Khaw. It remained visible from the road but was fenced and inaccessible.

Parker 1976 and 1986 were dependent on published references to military sites - and Khaw is only said to include a fort in Harding's unpublished records (above). 4

174

David Kennedy: Khirbet Khaw: a Roman town and fort in northern Jordan

Yet, if we set aside the Princeton Expedition, collectively the reports quoted leave little doubt that the remains are those of a small town with several notable features: a suggested 'castellum', a possible temple, subterranean tombs, cisterns, architectural pieces. Suggested dates range from Nabataean to Byzantine.

Air Photographic Interpretation of the Remains

Rediscovery on old Air Photographs

The Fort: First, a large structure (A) in the north-east which dominates the site entirely; plainly the 'castellum' referred to by Harding (above). It is much to the credit of the latter - principally a prehistorian - that he identified this structure, the only visitor to have done so. Briinnow and von Domaszewski, whose massive enterprise was explicitly aimed at recording Roman military sites, referred only to a large structure on the summit, which they thought might be a temple. In his discussion of the forts of the 'Arabian Limes', Briinnow (1909: 65 and 67) omitted Khaw and marked it only as an 'old town' on his map. McCown noted no more than 'numerous small buildings'. In this respect, the tumbled stones of the ruins seem to have obscured the outline of this structure every bit as well as happened at Umm el-Jimal where the Roman fort was only detected amongst the far better known remains in 1981 (De Vries 1986; 1998: Figs 2a, 2b, 229-31 passim; Parker 1996: 26-27; cf Kennedy and Riley 1990: 183-185; Gregory 1996b: 2, 270; Kennedy 2000b: 80-86).

The shapes are quite well-defined by shadows implying walls still standing to a significant height. That is true even of internal walls. The ruins may be divided into several components.

There was always the potential for Kh. Khaw to have been viewed on old air photographs - the older the better. Unfortunately no photograph of the site is to be found amongst the hundreds taken by the German Air Force operating in Palestine and Transjordan during World War I and preserved now in Munich.5 Nor is it amongst the many glass negatives taken by the RAF during the 1920s and held now in London (Institute of Archaeology, University College). The later film taken by the RAF in the 1940s and early 1950s still held by their Air Photo Unit (RAF Brampton, Northamptonshire), includes a series of low-level verticals of the area around Zerqa but stopping short of Khaw. As the RAF must have overflown the site regularly, this is unexpected; moreover, in 1939 they were engaged in flying Sir Aurel Stein around Transjordan explicitly to photograph such sites. Stein's survey in northern Jordan was condensed and never included Kh. Khaw (cf Kennedy 2000a). Fortunately, the vertical survey of western Jordan carried out by Hunting Air Survey in 1953 does include this area and it was in the course of systematic analysis and interpretation of all 4000 frames that the site was first encountered. It stands out sharply as a place of obvious significance. Four frames show the site and two others cover the immediately adjacent territory (App. 2). The aerial view (Illus.17.3) suggests a flat-topped hill with relatively steep sides on at least the north, east and south. What it may have looked like before it was built over in antiquity may be gauged by reference to the smaller, but probably similar, hill just to the west which has a relatively flat top marked by a clearly defined scarp. The air photographs offer a marvellous alternate view of a site whose jumbled ruins and badly damaged character were found so off-putting by visitors (Illus. 17.3- 17.5). Although size and shape are the subject of little more than very broad indicators by visitors, both may be determined rapidly from the photographs. The ruins can be separated out into a series of structures and complexes, which collectively form a small but compact town. As measured off the air photographs the overall dimensions are approximately 262 m north-south by 210 m east-west, c. 5.5 ha/13.5 acres.

The shape is that of a distorted square, the north and south walls being roughly parallel, but the west is slightly oblique to them, and the east seems to strike out perpendicularly from the north-east comer but then takes a more sinuous course, probably following a contour of the hillside. The dimensions as measured from the air photograph are: c.100 (N), c. 100 (E), c.94 (S), and c. 88 (W) m. This gives an area of c.0.93 ha/ 2.3 acres. There is an opening in the middle of the west side and, almost certainly, another in the middle of the south wall. On the north and east sides, where the ground falls away sharply, there are no such traces and probably gates there would have been unnecessary. The two gateways appear to be flanked by projecting towers. Close examination of the wall circuit, and comparison with similar views of such sites, suggests there are projecting towers at the south-east, south-west and, perhaps, north-west angles; the north-east angle appears to have collapsed down the slope or been overbuilt. Between the south gate and the south-west angle there may be a small projecting tower. Comparisons suggest that we should expect other towers even where they are not apparent from the photograph. The exception in this case may be the north and east walls. The fort appears to be packed with buildings, some against the walls but most set out in orderly lines across the interior. The photographs do not permit these to be drawn with any clarity but one may note the alignment that extends across the fort from the west gate. One area in which detail is clearer is in the north-west quadrant where there is a building with a succession of paired rectangular rooms on either side of a central north-south spine wall.

5 Most of these photographs are in Munich in the Bayerisches Hauptstaatsarchiv, Abt. N Kriegsarchiv. See Kedar 1999 and cf. Kennedy forthcoming a.

175

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

There are eight, perhaps nine such pairs but it may rather be a block of five and one of three or four. In the north-east quadrant the layout seems very distorted on the air photograph. Two features stand out: a large square solid structure in the centre of this area measuring c. fifteen to eighteen metres square and standing in the middle of an open yard, and a long rectangular structure which seems to overlap the eastern end of the north wall and extends for some twenty five to thirty metres. The first of these may be what Briinnow and von Domaszewski thought might be a temple; the latter is similar to the rectangular projection noted by the Princeton Expedition at Deir el-Kahf (cf. Kennedy and Riley 1990: 178-181, fig. 125; Kennedy 2000b: 67-73) although the fort there is much smaller and of a very different character. Four black spots probably mark roofless cisterns inside the fort. There may be others whose roofs are still in place and do not show the deep shadows indicative of a steep-sided pit. That on the northern edge between the two quadrants appears larger and, on the more recent photograph of 1990 (Illus. 17.4; cf 17.5), it seems to have been cleared out and brought into use as an open rectangular reservoir, which is probably what it was in antiquity too (cf below). The Caravanserai: Just west of the fort lies a large rectangular building (B) whose eastern wall runs only some ten to fifteen metres from the west side of the fort. It seems to be an almost perfect rectangle, its dimensions measured on the air photograph being c. seventy x forty five metres. The character of the structure is quite clear: rooms are constructed against the interior of the circuit wall leaving a rectangular courtyard; indeed, two courtyards. The photographs give different impressions. Illustration 17.3 appears to show an entrance in the centre of the north wall leading into the west courtyard, and ranges of rooms two deep around both courtyards. Illustration 17.4, however, seems to have a double entrance and a mixture of two and one deep rooms around the courtyards. Although the details seem sharper on the latter, I suspect the clarity is due to stone robbing. On the east side, for example, the track between this structure and the fort seems to have been constructed over the robbed line of the external wall. I have preferred, therefore, to determine the layout from the older photograph and use the newer ones only to clarify details.

noted. Just south of the fort there seems to be the trace of a wall running on from the east wall of the fort (C). As no other traces can been seen, it is possibly no more than the clearly marked edge of the scarp which is visible too on the western hill. However, there is then another line, discontinuous, striking off from it westwards and suggesting either a walled annex to the castellum or an earlier enclosure over the northern two thirds of which the castellum was constructed. The houses are scattered around in a random fashion with both their own private space and probably public open spaces. There may, however, be a street dividing the eastern and western halves (D). At point E there seem to be two sides of a platform just above the scarp, the northern edge accentuated by a deep shadow. By 1990 (Illus. 17.4) this point seems to have been dug out, perhaps by stone robbers exposing a wall line. Next, there is the curious three-sided enclosure running down the slope north-east of the fort (F). Four tall structures within it cast unexpectedly long shadows for the time of day. By 1990 there was little to be seen of this enclosure wall and nothing of the structures. Tombs: some of the subterranean tombs referred to by early visitors may be what seem to be cave openings on the slope below the north-east angle of the fort (G) and perhaps also on the high ground north-west of the caravanserai (H) (Illus. 17.3). By 1990, however, the clearance of the entrances had exposed the presence of several tombs scattered across the town site itself. On illustration 17.4 may be seen features appearing as a series of parallel dark lines thus:

IIIIIIII In places the lines become progressively shorter on one side as overtaken by a triangular shadow. It seems likely that one is looking at the entrances to tombs which are approached down a long flight of stairs showing here through the shadows on individual steps but then, in cases where the sun is at an angle to the direction of the steps, the stair shaft becomes hidden in a long shadow. I have counted seven with apparent staircases. From the air photographs as a whole, from 1953 to 1998, I have counted 30 such openings, mostly with stairs though some may be cisterns.

It is not possible to identify how access was gained to the

eastern courtyard. One assumes it was from the western courtyard rather than from an independent external gate but there may be a break in the south wall. The dark spot on illustration 17.3 seems to be an umoofed cistern or small reservoir in the western courtyard; in illustrations 17.4 and 17.5 it has been considerably enlarged. The Town: Three quarters of the ruins lie beyond the walls of the two structures so far described, and seem to be those of a small town. There are clear traces of numerous individual structures and the dark spots of what may be identified as roofless cisterns. Only a few features need be

Water supply: a possible open reservoir has already been noted inside the fort and a probable second in the caravanserai (above); likewise, some of the dark spots visible on the site on the 1953 photograph may be roofless cisterns. In light of the immediately preceding discussion, however, one must consider that some of the staircases might lead down into cisterns rather than to tombs. Moreover, in addition to the shafts with clear staircases, there are at least a dozen more visible on the 1990 photograph (Illus. 17.4) where triangular shadows reveal a shaft without any obvious trace of stairs. More still can be seen on the air photographs of 1998 (Illus. 17.5). These 176

David Kennedy: Khirbet Khaw: a Roman town and fort in northern Jordan

could be either tombs or cisterns. The curving wall at the south-east comer of the site (J) looks to be a reservoir wall circuit - perhaps the 'catch-basin and cisterns' in the valley to the south noted by McCown.

Domaszewski meant by a 'temple' on the highest point. The 'street' dividing the site east - west is visible but distorted by collapsed material which in turn may have been from later re-use of the site. Rooms can be seen built against walls at the north centre and in the south-east; elsewhere ranges of rooms were seen, some paired (northwest quadrant) on either side of a common wall. A few architectural pieces were noted on the surface, including a cornice and a piece with elaborate floral decoration (Illus. 17.8).

The air photographs of 1998 (Illus. 17.5) reveal the pale limestone wall of a large structure in the north-east quadrant of the fort. It may be the temple referred to by Briinnow and von Domaszewski. Fieldwork at Khirbet Khaw Through the kindness of the Commanding Officer at the Jordan Army military base at Zerqa, permission was given in May 1997 for a 30 minute visit to the site. No photography was possible but the opportunity was taken to collect some pottery.

The supposed reservoir in the north has been remodelled and could not be confirmed. There is a tomb just west of the centre. The latter was approached down steep steps and comprised several recesses around a chamber. Numerous other tombs were noted across the site, most approached down steep flights of steps.

The outline of both fort and caravanserai were clearly identified. Despite the considerable infill of collapsed masonry, the walls of the fort were still in places over a metre in height and at the comers as much as 2 m. of rubble was standing above external ground level (Illus. 17.6). Several tomb entrances were noted but none could be entered. Throughout the site, structures are built from black basalt but here and there some blocks of limestone were visible.

Also noted elsewhere were cisterns and, at the bottom of the hill beside the modem road, a reservoir, part of whose wall as still well-built on the south side (Illus. 17.3). Finally, the caravanserai was planned in outline together with the position of the eastern courtyard and of the tomb in the western one. It is very regular in shape, rooms are located against the outer walls around two courtyards.

In 1998, three days were allowed by the Jordanian Army for some detailed planning inside the army perimeter fence. There was time for the main lines of the fort to be plotted and the outline of the caravanserai. 6 Pottery was collected more systematically and the site was explored more fully. General photography was not possible.

Pottery was thick across the site and pieces of terra sigillata were common. A reading of the pottery collected is appended below (App. 1). Discussion

The outcome of this brief groundwork may be seen in the plan plotted from the survey (Illus. 17.7). The outline emerges of a fort with walls about one metre thick and at least one comer tower (south-west). A recent spoil heap obscures part of the north-east comer but inspection reinforced the view that this quadrant had been remodelled after initial construction. A substantial wall beyond the centre of the north side suggests a structure projecting there which may have been original. However, there are difficulties with the north wall, as the angle drawn at the north-west comer indicates. There is a possibility the west wall and north wall have been remodelled and originally radiated from the comer shown. That would have the attraction of straightening the two halves of the west wall and making it as long in total as the north and south walls i.e. 100 m. A cistern just beyond the north-east angle may have required a postem there or even a projection to include it (cf. Deir el-Kahf: Kennedy and Riley 1990: 178181; Kennedy 2000b: 67-73). As the plan shows, however, the north-east quadrant as a whole consists of larger structures which, on the ground, appear as of a more sophisticated and more monumental build and constructed from limestone. This may be what Briinnow and von

The best-known ruined town in the region is Umm elJimal, thirty five kilometres to the north-east. However, comparison with Umm el-Jimal is inappropriate - at some 34 ha/ 85 acres, it is one of the largest towns of the Southern Hauran (De Vries 1998). Rather we should look at the smaller sites. For example, Kh. Khaw is about the same size as the town area at Umm er-Resas south of Amman (Kennedy and Riley 1990: 189-193; Kennedy 2000b: 129-132). For a still closer parallel, we should turn to Kh. es-Samra, c. eleven kilometres to the north-northeast (Humbert and Desreumaux 1998). There we see (Illus. 17.9) a compact town, c. 160 x 200-220 m, 3.2-3.5 ha/ 8 acres with a small fort in its midst (as measured from the sketch plan in Kennedy and Riley 1990: 198-200; Fig. 146). The fort was reused later for civilian purposes and includes a church. However, the town area at Samra is bounded by a wall, a feature not noted at Kh. Khaw. The fort at Kh. Khaw is a distorted square, 100 (N) x 109 (E) x 100 (S) x 102? (W), c. 1 ha/ 2.47 acres. There was a projecting rectangular tower at the south-west comer and possibly at the others too originally. There may have been some projecting interval towers on the south and west sides. Gates have been noted in the middle of the west side and just off-centre on the south. The interior is crammed with buildings whose collapsed rubble does not obscure walls still a metre or more in height. A road runs east from

I am grateful to George Findlater for carrying out this arduous task and to Bob Bewley who assisted. 6

177

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

the west gate and there are hints of narrower streets radiating at right angles to it. Rooms were built against the walls but also as free-standing blocks in the middle. Some of the latter have rooms in pairs on either side of a common central wall, reminiscent of barrack blocks everywhere in the region. The water supply was provided by a cistern inside the fort, a second just beyond the northeast angle (perhaps inside a projecting room at that point) and possibly by a rectangular cistern just inside the north wall. In size and character the fort is strongly reminiscent of at least three others, one not far distant. Site

Dimensions

Area

be cautious. On the other hand, all three close parallels have been the subject of excavation. As the table shows, Jimal and Da'ajaniya have been dated to, broadly, c. AD 300. Avdat 1 has been more controversial with suggestions ranging from Nabataean to Late Roman. Extensive recent excavation there has continued the controversy with the co-directors divided sharply on the date. One of them, Erickson-Gini, has argued powerfully in a forthcoming publication for a construction date of late third/ early fourth century AD. We may turn now to the pottery collected in and around the fort (App. 1 below). As always it is necessary to be very cautious about interpreting a random collection of surface sherds. The groups from and immediately beside the fort, however, have a common characteristic: pottery is overwhelmingly third and fourth century or sixth century onwards. In broad terms it bears out a construction date in the former period.

Date

Umm el-Jimal c.95 x c.112

c.1 ha/ 2.47 acres early 4th c

Da'ajaniya

c.100 x c.100

c.1 ha/ 2.47 acres late 3rd/ early 4th c

Avdat 1

c.l 00 x c.100

c.l ha/ 2.47 acres late 3rd/ early 4th c?

Kh.Khaw

c.100 x c. 102/9 c.l ha/ 2.47 acres

?

Table 1: Comparison of dimensions, areas, and dates for forts in the Arabia and Palaestina region (derived from Kennedy and Riley 1990: 172-175; 183-185; Parker 1991: 134-141). Not just size, but the plans of these forts, too, are strikingly similar (Illus. 17.10). All are irregular squares or very nearly so. Typologically they have been grouped before, by Kennedy and Riley (1990: 167-193 - 'Large Forts with External powers') and more narrowly now by Parker (1995: 253-257). Parker, however, deals only with forts in Jordan and, while noting the very similar sizes and appearance of Da'ajaniya and the 'Castellum' at Umm elJimal, omits Avdat 1. Gregory (1996a: especially Fig. 9) groups most of the same sites (preferring Avdat 2 to Avdat 1: cf. Kennedy and Riley 1990: 196-198), plus, on typological grounds, adding Khirbet es-Samra and Nessana. Certainly, despite its notably smaller size (c.60 x 65 m, 0.39 ha/ 0.95 acres), the fort at Khirbet es-Samra should be contrasted both because of its similarity in appearance and date, and its proximity, just over ten kilometres north of Khaw. Despite the similarities with the well-preserved fort at Da'ajaniya in southern Jordan, including there a rectangular reservoir inside the walls, the closest parallel is the so-called 'Castellum' at Umm elJimal. Seen from the air (Kennedy 2000: Figs 9.6 and 9.lOC) it looks very similar and, as fig. 10 shows, the plan there is likewise of a distorted square, with irregular projecting rectangular comer towers, and internal buildings laid out in an orderly but careless fashion.

The character of the fort remains imply it was re-used at a later date, probably for non-military purposes. As just noted, the second part of each pottery group is sixth century AD onwards. That would accord with the appearance of subterranean graves inside the wall circuit: an abandoned fort taken over for civil purposes including burial. It may have been now, too, that a monumental structure made of limestone was erected in the north-east quadrant and perhaps the north-east angle remodelled. Some of the structures represented in the tangle of walls inside the circuit may belong to this later civil use. In broad terms it resembles the process detected inside the fort walls ofKh. es-Samra to the north and Umm er-Resas south of Amman (Kennedy and Riley 1990: 198-199, 189193; Humbert and Desreumaux 1998; Kennedy 2000: 9798, 129-132). The caravanserai is a precise rectangle with external dimensions of 72 x 43.5 m, 0.31 ha/ 0. 77 acres. The measurements are just slightly less than 250 x 150 Roman Feet. Although it appears to be of a single construction phase, it can be interpreted as intended from the outset to consist of two parts divided in the ratio 2: 1. In the west is the larger part with the main entrance through the north wall. The courtyard is almost entirely surrounded by rooms, mainly two deep. There is a tomb beneath the courtyard. The remaining third of the structure to the east seems to be cut off by a wall from north to south forming a rectangular area 43.5 x 24 m, with its own smaller courtyard. Rooms surround the entire courtyard - on east, south and west they are one row deep; in the north perhaps three deep. This part of the caravanserai is reminiscent though larger - of the caravanserai at Kh. el-Khalde m southern Jordan which is c. 32 x 22 m (Illus. 17.11).

In the light of the work of Lander (1984) and, more recently and directed explicitly to the East, that of Gregory (1995-7: 1996b), one should avoid simplistic equations based on crude similarity alone. Gregory - assigning the fort at Khirbet Khaw to the 4-6th centuries AD - is right to

It is not possible to date the caravanserai from its

appearance alone. There are several examples in Jordan of Roman fort and caravanserai in close proximity and,

178

David Kennedy: Khirbet Khaw: a Roman town and fort in northern Jordan

arguably, contemporary, rather than the latter being a later Islamic addition to an abandoned Roman fort. 7 The obvious example which springs to mind is at Khirbet elKhalde (Kennedy forthc. b) and, perhaps, closer to Amman, at Umm el-Walid (Kennedy 2000b: 124-125). Here, however, there are some difficulties. First, the structure is very close to the west wall of the fort and blocks direct approach to the West Gate. Second, the pottery from the caravanserai, in contrast to the fort, is overwhelmingly Late Byzantine, sixth century onwards. Perhaps we should view the caravanserai as a later construction after the fort was abandoned. We might note, too, that there are two tombs beneath the 'road' between fort and caravanserai (Illus. 17.3).

site had been worked to be used as 'sherd tools'. Part of a single small, fourth (?) century Roman coin was found. Two features not noted so far are inscriptions and baked brick or tile.

At least three quarters of the site is given over to other buildings, all rectilinear in form. Most are broadly aligned with the fort and caravanserai but there is no formal layout to this part of the site. A road bisects the site from north to south, straight as it passes between caravanserai and fort but angling a little to the west thereafter. Houses appear to consist of an enclosure wall with rooms internally around part of the circuit leaving a courtyard. An example is clear at Lon illustration 17.3.

The starting point is the Peutinger Table which locates the place name Gadda XIII MP north of Philadelphia in the direction of Bostra. Khaw lies in this direction but about two MP east of the Roman highway to Bostra, the Via Nova Traiana. The nearest milestone on the Via Nova appears to be that for XIV MP from Philadelphia. Neither point is necessarily an impediment to identification. Next, the Notitia Dignitatum (Oriens 37. 20) records for about AD400: Equites sagittarii indigenae, Gadda. 'Native Horse Archers, at Gadda'

Pottery ranges widely in date from first century AD to the sixth century and beyond. There is also a little N abataean pottery on the highest part of the site - the earliest area occupied? - and perhaps some early Islamic from the fort.

Finally, the question of an identification. McCown rejected Briinnow and von Domaszewski's identification of Kh. Khaw with Gadda on the grounds it was too small. As we have seen, however, it is larger tham Kh. Samra which he preferred and his (still smaller) alternative of Qasr elHallabat. Plainly we need to look seriously at the identification, the more so as we now know - as Briinnow and von Domaszewski and McCown did not - that there is a significant fort on the site.

In short, there are positive reasons for identifying Kh. Khaw with Gadda - but others against it. A more obvious place to identify along this route would have been at the site of the medieval settlement in Zerqa itself The valley is broader and more fertile there and well-watered. It is also where the Roman highway crossed the R. Zerqa to travel thereafter along the west bank towards Philadelphia. A Latin inscription was found many years ago re-used in the medieval khan on its 'small isolated hill' beside the river at Zerqa. As the Princeton Expedition noted, this khan is on the edge of the desert(= steppe) but close by the perennial Wadi Zerqa. The inscription whose first lines are badly damaged (PES 111.A.1,16-17; App. 111.A.1,v-vi), reads:

Neighbouring sites are rich in churches, often first detected from a curving apse wall and frequently containing mosaic floors. Despite three days spent at Kh. Khaw, no obvious church building was observed and not a single tessera was found anywhere on the site. There are certainly numerous tombs at the site and Briinnow and von Domaszewski provided sketch plans of two (Illus. 17.12). At least one lies inside the fort and a large one with open air courtyard and rooms opening off it on two sides lies just west of the caravanserai. The number of openings is, however, suspiciously high and they are strewn throughout the living area rather than around the periphery as one might expect. Few can be entered. The alternatives seem to be: (a) Many were originally cisterns rather than tombs - clear examples of cisterns, a necessary feature of such a site, are relatively uncommon. (b) They are tombs whose locations are visible now but were hidden from sight as the settlement grew over an earlier cemetery. (c) They are tombs built in a later period in abandoned parts of the site. (d) The inhabitants were relaxed about having the dead interred in their midst.

[. .. O I I I ...I]mp(eratores) Augg(usti) tute[lae] gratia ex Palaes/[tina in Arabia]m tra[nstu]lerunt, [c]astra quoque (e)x solo oppo/[rtuno loco]exstruxerunt per Aurel[ium Theone]m leg(atum) Aug(ustorum)/ [. ...] 'The Emperors Augusti Valerian and Gallienus for the protection of the country transferred troops from Palestine into Arabia, also built this Fort in a suitable place through the agency of Aurelius Theo, Imperial Governor ... ' Date: AD 253/59 (the known dates of Aelius Aurelius Theo in Arabia: Sartre 1982: 92, no. 38) The Princeton Expedition has been followed in more recent times by commentators (e.g. Speidel 1977: 725) who have supposed the inscription orginated not at Qal'at Zerqa but at Qaryat el-Hadid, a little further south (Kennedy forthcoming a). That may be the case, but it is not necessarily so. We might equally see it originating at

A little glass was noted here and there and the ceramicist observed that several sherds found in various parts of the 7

I owe this observation to David Graf.

179

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Kh. Khaw or, the simplest solution, at the site of the khan

Caravanserai: The eight sherds were 6th c Byzantine.

itself In short, Gadda might better be located beneath the medieval site of Qala'at Zerqa or at Qaryat el-Hadid about 1.5 km to its south. Wherever it was, at the close of the fourth century it was a military base. Kh. Khaw offers a significant fort of perhaps AD 300 and later but we know too little of Qala'at Zerqa and Qaryat el-Hadid to dismiss them and cannot even be certain the Khaw fort was occupied c. AD 400.

1998: Fort: area around tomb in NW Quadrant (Bag 3) Imports: African Red Slip;"Jerash bowl" fragment; Hawran ware 3rd onwards * a few Late Roman * predominantly Late Byzantine 6th onwards * glass Late Byzantine Other: "sherd tools" - lot of "micro tools"

The future

Kh. Khaw merits close attention. A project should be established to carry out a detailed survey, photographing and drawing of the site as a whole and its various components. With adequate planning, the military authorities are likely to be as helpful as in the past. Testtrenching should allow dating of the circuit wall of the fort and of the caravanserai. Beyond the clear examples of walls for structures there are also earthworks on the west and north, in particular. These too should be explored on the ground and might well repay photographing from the air when there is vegetation on the site. Although the site is plainly 'Roman' in the broad sense of that term (first century BC to seventh century AD), exploration should obviously be sensitive to a Nabataean origin and continuity into the Islamic period, as hinted at by surface pottery. In the longer term there is the possibility the army will relinquish the site into the custody of the Department of Antiquities. That would ease access for investigation and scholarly visits but might leave the site vulnerable. At the moment, it is striking that the site protected by the army for over half a century has suffered very little damage despite its proximity to a major highway and the growing town of Zerqa.

Appendix 1 (by Ina Kehrberg)

Fort: spoil heap around concrete tower (Bag 4): Imports: none 1st BC/AD (possibly into 2nd AD) * half Roman 6th c. onwards * half Late Byz Other: a "sherd tool" Area just beyond N wall of fort (Bag 5): Imports: ARS * a few 1st BC/ AD sherds * a few Late Roman 3rd c. onwards * mostly Late Byzantine 6th c. onwards Other: Late Byzantine glass, modem glass and "sherd tools". Beyond N wall and NW quadrant (Bag 8): Imports: Nabataean; Hawran wares * Late Roman 3rd c. onwards * Late Byzantine 6th c. onwards Area around beduin graves N of caravanserai (Bag 2): Imports: Eastern Terra Sigillata; Hawran Streaky Burnished Byzantine ware; fragment ofa "Jerash" lamp * (one or two Iron Age) 2ndc. AD * very few Roman late 6th-7th * mostly Late Byzantine Other: Byzantine glass

Pottery was examined on a preliminary basis with the following results:

Caravanserai - area around tomb (Bag 6): Imports: 1 tiny Hawran ware fragment * very few Roman sherds * predominantly Late Byzantine 6th c. onwards Other: Late Byzantine glass

1997: Fort: 47 sherds, part of a thick brick, half of badly corroded tiny coin - possibly Late Roman(?). Many of the sherds were body sherds and not closely datable other than to 'Classical'. Several others were more informative, including: * sherds of I-2nd c; * sherds of 2nd c; * sherds of 3rd c.; * fragment ofa Jerash bowl (late 6th c.) * a possible Islamic sherd. * a sherd which had been used as a tool * brick piece may have been part of a heating flue system.

Area N ofbirkeh (Bag 1): Imports: Eastern Terra Sigillata; ARS * (one or two Iron Age) * a few 2nd AD * mostly Late Roman 3rd c. onwards * some Late Byzantine 6th c. onwards Other: "sherd tools"/ glass possibly Late Roman-Byzantine (very worn) Area beyond bank ofbirkeh (Bag 7): Imports: Eastern Terra Sigillata; Hawran wares; Arretine? (or a version of ARS) * a few 2nd c. AD otherwise * seems even mix of Late Roman 3rd 180

David Kennedy: Khirbet Khaw: a Roman town and fort in northern Jordan

* and Late Byzantine 6th c. onwards Other: Late Rom glass fragment; "sherd tools". General (Bag 9): Imports: Eastern Terra Sigillata * mix of Late Roman * and Late Byzantine

Erickson-Gini, F. 2001. Nabataean or Roman? Reconsidering the date of the army camp in Avdat in the light ofrecent excavations. In P. W. M. Freeman et al. (eds) Limes XVIII. Oxford: British Archaeological Reports, International Series, forthcoming. Gregory, S. 1996. Was there an eastern origin for the design of Late Roman fortifications? Some problems for research on forts of Rome's eastern frontier. In D. Kennedy (ed.) The Roman Army in the East. Ann Arbor: Journal of Roman Archaeology, Supplementary Series 18: 169-209. Gregory, S. 1995-97. Roman Military Architecture on the Eastern Frontier. Amsterdam: Hakkert, 3 vols Homiis-Fredericq, D. and Hennessy, J. B. (eds) 19861989. Archaeology of Jordan. 3 vols, Leuven: Peeters, Akkadica Supplementum II, VI, VII. Humbert, J.-B. and Desreumaux, A 1998. Khirbet esSamra, Jordanie. 1. Lavoie romaine; le cimetiere; les documents epigraphiques. Paris: BREPOLS Kennedy, D. L. 1982. Archaeological Explorations on the Roman Frontier in North East Jordan. The Roman and Byzantine military installations and road network on the ground and from the air. Oxford: British Archaeological Reports, International Series 132. Kennedy, D. L. 2000a. Relocating the Past: missing inscriptions from Qasr el-Hallabat and the air photographs of Sir Aurel Stein for Transjordan. Palestine Exploration Quarterly 132: 28-36 Kennedy, D. L. 2000b. The Roman Army in Jordan, London: CBRL/ British Academy Kennedy, D. L. forthcoming a. Qaryat el-Hadid: A 'lost' Roman military site in northern Jordan. Levant. In preparation. Kennedy, D. L. forthcoming b. Two Nabataean and Roman sites in southern Jordan: Khirbet el-Qirana and Khirbet el-Khalde. In press. Kennedy, D. L. and Riley, D. N. 1990. Rome's Desert Frontier from the Air. London: Batsford/ Routledge Lander, J. 1984. Roman Stone Fortifications. Oxford: British Archaeological Reports, International Series 206. Lees, G. R. 1895. Across Southern Bashan. The Geographical Journal 5, 1-27. MacAdam, H. I. 1986. Studies in the History of the Roman Province of Arabia. The Northern Sector. Oxford: British Archaeological Reports, International Series 295. McCown, C. C. 1930. Spring Field Trip, 1930. BASOR 39: 10-27. Merrill, S. 1881. East of the Jordan. A Record of Travel and Observation in the Countries of Moab, Gilead and Bashan. London: Bentley Palumbo, G. (ed) 1994. JADIS. The Jordan Antiquities Database and Information System, Amman: Department of Antiquities of Jordan/ American Center for Oriental Research. Parker, S. T. 1976. Archaeological survey of the Limes Arabicus: a preliminary report. ADAJ 21: 19-31.

3rd c. AD 6th c. onwards

Appendix 2: Photographs of the Site

* Briinnow and von Domaszewski published two small ground photographs (1905: Figs 843; 845). These were the only two taken by them (MacAdam 1986: 287 for the catalogue of their photographs now held at Princeton University). * Air photographs: APA 53/HAS 54.034; 035; (036?); 54.052; 051; (050?) APA90/Rl.18; R3.19 APA98/10.25-10.36 (colour slides) APA98/3.3-3.9 (B&w) AP A98/3.3 (Colour print) * Ground views: 1998: 3.29-3.36 (B&w) 1998: 5.22-32 (Colour slides) ABBREVIATIONS USED

AJ = D. Homiis-Fredericq and J. B. Hennessy (eds.),19861989. Archaeology of Jordan, 3 vols, Leuven: Peeters, Akkadica Supplementum II, VI, VII. JADIS = G. Palumbo (ed.), 1994. JADIS. The Jordan Antiquities Database and Information System, Amman: Department of Antiquities of Jordan/ American Center for Oriental Research. PES = H. C. Butler et al., 1907-1949. Publications of the Princeton University Archaeological Expeditions to Syria in 1904-05 and 1909. Leiden: Brill. 2 divisions in many parts.

BIBLIOGRAPHY

Bauzou, T. 1985. Les voies de communication dans le Hauran a l'epoque romaine, in J.-M. Dentzer (ed.) Hauran I. Paris: Geuthner: 137-165. Briinnow, R. 1909. Die Kastelle des arabischen limes, in G. Maspero (ed.) Florilegium ou recueuil de travaux d'erudition dedies a Monsieur le Marquis Melchior de Vogiie. Paris: Imprimerie Nationale: 65-77 Briinnow, R. E. and von Domaszewski, A. 1904-09. Die Provincia Arabia. Strassburg:Triibner, 3 vols. Butler, H. C. et al. 1907-1949. Publications of the Princeton University Archaeological Expeditions to Syria in 1904-05 and 1909. Leiden: Brill. 2 divisions in many parts. De Vries, B. (ed.) 1998. Umm el-Jimal. A Frontier Town and its Landscape In Northern Jordan, I. Portsmouth, RI: Journal of Roman Archaeology, Supplementary Volume 26.

181

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Parker, S. T. 1986. Romans and Saracens. A History of the Arabian Frontier. Winona Lake, In.: Eisenbrauns, ASOR Dissertation Series 6. Parker, S. T. 1995. The typology of Roman and Byzantine forts and fortresses in Jordan. In K. 'Amr, F. Zayadine and M. Zaghloul (eds) Studies in the History and Archaeology of Jordan V. Amman: Department of Antiquities: 251-260. Sartre, M. 1982. Trois etudes sur [ 'Arabie romaine et byzantine. Brussels: Latomus

182

David Kennedy: Khirbet Khaw: a Roman town and fort in northern Jordan

■ Umm el-Jimal

eJerash

■ Qasr

el-Hallabat

0

17.1

30Km

Sketch map of northern Jordan showing the location of Kh. Khaw and other sites in the area discussed in this essay.

Kh. Khaw



3km

17.2

Khirbet Khaw: location map (based on K737 Series 1: 50,000 sheet: 3254. III).

183

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

17.3 Khirbet Khaw: air photograph of 11 April 1953, taken at 11 am (APA53/HAS 56.034). For convenience of comparison with the description in the text, north is at the top. For greater clarity, however, tum the page through 180 degrees to bring the shadows nearest the viewer.

17.4

Khirbet Khaw: air photograph of 1990 (APA90/IGN R3.19). 184

David Kennedy: Khirbet Khaw: a Roman town and fort in northern Jordan

17.5

17.6

Khirbet Khaw: air photograph of 1998 (APA98/10.33)

Khirbet Khaw: the west gate of the fort in 1998 (1998/ 5.32).

185

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

\)omb

□ 50 m (approx,)

0

17.7

Khirbet Khaw: plan of the visible remains of the fort and outline of the caravanserai (George Findlater 1998).

17.8

Khirbet Khaw: two architectural pieces on the surface inside the fort in 1998.

186

David Kennedy: Khirbet Khaw: a Roman town and fort in northern Jordan

0 I

17.9

20 I

100M I

Kh. es-Samra: plan of the site drawn from the air photograph and showing the location and form of the fort (from Kennedy and Riley 1990: 200, Fig. 146).

187

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

C

50 m (approx.)

0

17.10

Comparison of the plans of 4th century AD forts: (a) Umm el-Jimal; (b) Da'ajaniya; (c) Avdat 1; (d) Kh. Khaw (after Kennedy and Riley 1990: passim). 0



0 17.11

30m

Kh. el-Khalde: plan of the caravanserai (after Kennedy forthc b)

17.12 Khirbet Khaw: tombs recorded by Briinnow and von Domaszewski

188

18

Philo ofByblos and the Phoenician History: Ethnicityand Culturein HadrianicLebanon Henry Innes MacAdam

came. In his memory I have prepared this study, and I hope in doing so it answers some of his questions.

A Remembrance of Barri Jones My first encounter with Barri Jones was not auspicious. I had wandered into the Department of Archaeology for a reconnaissance of its facilities shortly after beginning doctoral research with Alan Bowman (then [1974] at the University of Manchester). Barri charged out of his office and the two of us nearly collided in the hallway. He was in a hurry to get to a meeting, and not at all inclined to take a Yank on a tour of his own domain. But he did point to the departmental bulletin board, and in that staccato speech rhythm we learned to appreciate, said: 'There's always a lot going on here and we have visiting lecturers all the time. Find out what's corning this fall and join us for anything of interest.' He was gone before I could thank him.

Introduction: Past and Present Almost all of the literature produced by the Phoenicians has been lost (Drower 1975: 148; Albright 1975: 516-526)); consequently most of what we know about them derives from writers of other cultures, and its content is often sarcastic and sometimes envious and hostile. Harden (1963: 123) drew attention to the portrait of a Phoenician (more accurately, Punic) merchant in Plautus' comedy Poenulus: he (the tradesman) conceals his multilinguistic ability for purposes of deception and (ultimately) monetary gain:

Et is omnis linguas scit, sed dissimulat sciens Se scire: Poenus plane est: quid verb is opust? (Poen. 112-113)

Neither ofus knew then that Barri would supervise the final stages of my doctoral dissertation after Alan had moved to Oxford. It was a responsibility he undertook with graciousness and plenty of good humor, qualities as characteristic of him as his boundless energy. His advice during that final year at Manchester: 'Always have some small project to work on when you're engaged in heavy-duty research on a big project. It will give you some perspective-and a break!'

It doesn't matter that Plautus based this play on a now-lost Greek original from the fourth century; the sentiments expressed were common enough at the time (c. 200 B.C.?) that Plautus wrote for Romans (or at least for audiences which understood Latin). Had some comparable example of Phoenician literature survived, it is fair to assume that the table would have been turned and a similar stereotype of the Roman character would serve as the butt of Phoenician ethnic jokes.

That was sensible advice, and while completing a thesis on Roman Arabia I began making notes on Philo of Byblos (c. 70-140), the few remaining parts of his Phoenician History, and the related question of Phoenician origins. Barri was intrigued with that topic, partly because his own field work in Roman North Africa and partly because a study of Phoenician origins involved not only archaeology in Lebanon but also in the Persian Gulf

The classical world was full of stereotypes: the Relentless Roman, the Garrulous Greek and the Enigmatic Egyptian are only a few. The Crafty Phoenician and the Cruel Carthaginian were but two sides of the same ethnic coin. One danger is, of course, that the stereotype replaces reality, that the humorous image becomes an idee f,xe. The other extreme is consigning a group to insignificance, as did Prof Arnaldo Mornigliano for the Phoenicians/Carthaginians in his highly praised collection of essays, Alien Wisdom (1975), from reasons of ignorance: 'I would gladly talk about the ideas of the Carthaginians, if we only knew them. ..[u]nfortunately there is not enough evidence to make a coherent account of how Carthaginians and Greeks saw each other in the third and second centuries B.C ... ' (Mornigliano 1975: 4;6). Consequently the Phoenicians of the Levant went unmentioned, although he devoted an entire chapter to the Celts, for whom we have equally few literary remains (P. Trogus' Historiae Philippicae, and a small corpus of inscriptions). Yet Punic literature was extensive enough to be stored in libraries at Carthage before the destruction of that city in 146 B.C. Sznycer (1968/69) has offered an excellent overview of what survives and what was lost.

Since I had been resident in Lebanon for a decade, and wanted to return there when the dissertation was finished, Barri insisted that I continue researching 'your old friend, Philo' and 'find out how and when the Phoenicians got to Lebanon, before coming to Libya and the rest of the western Mediterranean.' Before I left Manchester, Barri took me on a tour of several of his 'car-park digs' in the Deansgate area of Manchester. Students under his supervision were uncovering Roman Mancunium, and Barri had a large grid-map on which he showed me the past, present and future sites that he wanted to explore. One of the recent finds was special: the first example of the ROTAS/SATOR wordsquare in Britain (see Herner 1978/79). When I revisited Manchester in the spring of 1991 Barri was unfortunately away and a further chance to visit him never

189

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

The archaeological record is still patchy but constantly improving--both inside and outside Lebanon--though its interpretation is not always easy, and consequently what can be extrapolated is not free from controversy. The end of the Lebanese civil war (1975-1991) has made it possible for excavations of Phoenician sites to resume (for instance, at Tyre: Illus. 18.1) or to be undertaken for the first time (e.g. the Roman/Byzantine forum and port area of Beirut--see MacAdam, 2001). The problem all along has been the inaccessibility of major cities along the Lebanese coast; only at smaller settlements (e.g. Byblos/ Jubayl in the north and Sarepta/Sarafand in the south) have large-scale, multi-era excavations been possible. The most recent history of Phoenicia (Markoe 2000) incorporates the latest archaeological material, but ends at 300 B.C. with no mention of a book devoted to Phoenicia in the Hellenistic era (Grainger 1991; MacAdam 1993). One might hope that an interested historian will one day take us beyond the Hellenistic period. Other disciplines offer limited help for the Hellenistic/Roman era. Phoenician and Punic epigraphy sheds some light on the culture which produced it, but the result is all to often limited to augmenting the onomasticon and thus of little help to historians. Papyrology has not produced many documents relating directly to Phoenicia, but the few which do are of considerable interest to social and economic researchers. Consider (e.g.) several in the third-century B.C. collection known as the 'Zenon Archive', or the registration, dated AD. 252, of the sale of a slave at Aurelia Tripolis in Phoenicia (P.Oxy. 3053). Then there is Philo of Byblos and his Phoenician History, a second-century AD. work which survives only in fragments quoted by later writers, particularly Eusebius of Caesaraea. Because we do not know the methodology of Eusebius, he may have quoted his source correctly, but he may have quoted him out of context, i.e. 'to suit his own purpose', as James Barr (1974/75: 40 note 2) suggested.

Arguments for Hadrian's personal, imperial contacts with Phoenicia are set out in a later portion of this contribution, which provides a context for the other, preceding segments. It is also an apologia for the relative absence of archaeological evidence in a volume dedicated to the career of Barri Jones as an explorer of the Roman archaeological world. However, archaeology does not go completely unnoticed. In a separate but companion piece to this study (MacAdam, 2002), I re-examine the related question of Phoenician origins (which the fragments of Philo's Phoenician History do not directly address). Those for whom that issue (in its ancient and/or modem context) evokes curiosity may consult with profit Salles 1993 (for the literary aspects) and Rollig, 1983 (for the archaeological and linguistic perspectives). That parallel study explore the works of classical authors, who discussed whence, and sometimes when, the Phoenicians came to the Levant, as well as the few sources from the Roman imperial period and Late Antiquity which assert that the Phoenicians were native to the region, rather than immigrants from further east. Material remains from the Persian/Arabian Gulf and early evidence for the purple-dye industry there and in the eastern Mediterranean provide intriguing new clues, if not the definitive answers that some would prefer. Philo of Byblos

Sometime late in the reign of Hadrian (117-138), or early in the reign of his successor Antoninus Pius (138-160), a Phoenician antiquarian of indeterminate age claimed that he translated from Phoenician into Greek a cosmological and mythological account of Creation and the subsequent origins of Phoenician culture. It is probable, as argued below, that he was inspired to undertake this 'ethnological' study of his homeland as the result of one (or more) of Hadrian's visits to Phoenicia, perhaps (as will also be argued) a visit to Philo's patria ofByblos (ancient Gubla, modem Jubayl).

The purpose of this study is to re-examine Philo and his Phoenician History (hereinafter PH). THis will focus not, as in so many earlier studies, on the elucidation of what Philo had to say about the religious beliefs of Phoenicians at some earlier period, but on what Philo had to say regarding Phoenician traditions in his own lifetime, particularly during the reign of Hadrian.

That much we know about Philo. The name was then so common that it required additional 'identification' to avoid confusion with other homonyrnic personages such as Philo (Judaeus) of Alexandria. Hence Philo Byblios ('the Byblian' or 'the Byblite':OCD 1168). Stem's mis-identification of Philo with 'his native Berytus' (1980: 138) is unfortunate. Philo's contemporary (perhaps a Phoenician?), the geographer Marinus of Tyre, also had to be identified by city to avoid confusion with another contemporary, Marinus the anatomist of Alexandria (MacAdam, 2000: 340 for both; OCD 924 for the latter only).

There is no reason to doubt that Hadrian was a philhellene, as Ronald Syme reminded us in a comprehensive essay (Syme 1985). However, there is every reason to believe that Hadrian's interest in culture other than Graeco-Roman brought him to Phoenicia on at least one of the several occasions he visited the Near East. That visit may have had a two-edged effect: enlightening the Emperor with regard to one of many subject peoples, and awakening ethnic and cultural pride in Philo and several of his Phoenician contemporaries.

On Philo's probable dates (65-145) see Attridge and Oden (1981: 22 & n. 1-4). A century later, Origen (Contra Ce/sum 1.15), and later yet in the sixth century loannes Lydus (De mensibus 4.154), refer to Philo ofByblos as Herennius Philo. This has prompted speculation as to whether they were one and the same person or if later tradition had confused (or conflated) the identity and date of two different persons.

190

Henry MacAdam: Philo of Byblos and the Phoenician History

Herennius simply records recognition of his Roman patron, the senator Herennius Severus, whom Birley (1997: 227) identified not only as a 'pen pal' of the younger Pliny, but also as one of the suffect consuls of AD. 129: 'Severus and Arrianus might then be two persons of cultured disposition', i.e. Philo's patronus and Hadrian's personal friend, the historian. Mention will be made below of two other literati, both younger contemporaries of Philo and both from Roman Berytus: Marcus Valerius Probus and Hermippus. Some of this has been reviewed in Baumgarten (1981: 33-4 and n. 910). The relevant passages ofboth Origen and loannes Lydus are reproduced (with an English translation) in Attridge and Oden (1981: 98-9 and 70-1, respectively).

Nor does it help establish his probable jloruit. A close comparison of Sanchuniathon with Moses in any capacity other than chronological is purely modem; no ancient source makes a judgment of their 'equality' from a moral, religious or cultural point of view; regarding the latter, Homer or Hesiod would be better suited as logotypes or genretypes. Philo's translation of Sanchuniathon's Phoenician History was a work of eight (Porphyry, de Abst. 2.56) or nine (Eusebius, PE 1. 9.23) 'books', not as we understand 'volumes' but numbered by the papyrus rolls used in making a complete copy. Arnaldo Momigliano (1980) suggested that Eusebius considered Philo's work Concerning the Jews to be an organic part of the Phoenician History, whereas Porphyry understood it as a discrete composition. What motivated Philo to render this ancient work (if indeed it was that) into the lingua franca of his own day? Eusebius (PE 1.19. 27) quotes him as wanting 'to understand Phoenician culture' based on 'sources' available to Sanchuniathon, plus his (Philo's) personal examination of 'much material' (po/le hule). By the latter did he mean related material from a time more recent than Sanchuniathon? Whatever the case, I hope to show below that Philo's motivation was a response to the Emperor Hadrian's personal interest in Phoenicia.

Philo ofByblos claimed he translated what is now designated the Phoenician History (Phoinikike Historia, or Phoinikika), most of which is now lost, except for excerpts preserved in later authors. Philo claimed it was based on 'chronicles' attributed to a certain Sanchuniathon, a native Phoenician whom Eusebius of Caesaraea, the fourth-century church historian, referred to as a 'theologian' (PE 1.9.22). Eusebius is our main source for information not only about Philo, but whatever we have of the original text of the Phoinikika. Attridge and Oden (1981) and Baumgarten (1981) rely on F. Jakoby, FGrH290 for the fragments of Philo in Eusebius' Praeparatio Evangelica.

Greek sources (e.g. Homer, Hesiod, Herodotus) were biased, at least according to Eusebius (PE 1.9.27--who quotes Philo on this point), and thus responsible for creating an image of Phoenicians and their culture which is 'self-contradictory and was composed by some people more [concerned] with polemics than with truth'. This same reasoning had already prompted Flavius Josephus (37-c.100) to compose in Greek his own account of Hebrew history, the Antiquities of the Jews. However, Josephus wrote in the melancholy aftermath of the disastrous helium ludaicum (66-73) in part of which he had himself participated.

Eusebius also noted that Sanchuniathon lived, 'as they say', before the time of the Trojan War (i.e. prior to c. 1,200 B.C.). He quoted the native Phoenician philosopher Porphyry of Tyre (jloruit c. 300) as suggesting that Sanchuniathon had been a contemporary of Moses who claimed Beirut as his birthplace (PE 1.9.21). Other ancient sources associate Sanchuniathon with Tyre: e.g. Athenaeus (Deipnosophistai 3.126, wherein the name is spelled 'Souniathon') and the Byzantine lexicon known as the Suda (s.v. 'Sanchuniathon').

Philo of Byblos' scholarly production was modest but wideranging. He is said to have authored Regarding the Acquisition and Selection of Books (twelve 'volumes'), Concerning Cities and the Illustrious Men Each Produced (thirty 'volumes', later condensed into three). an essay On the Reign of Hadrian and works of indeterminate length on topics such as Concerning Useful Knowledge, Concerning the Jews, On Semantic Differences, and Astounding History. As with his Phoenician History, only fragments survive, quoted in later writers (on these, see Attridge and Oden 1981: 2-3). See also Stem (1980: 138-145), who conveniently collected and discussed the remaining portions of Concerning the Jews, and Palmieri (1981: 47-80) for extant portions of Philo's On Semantic Differences (which, however, are presented in a critical, highly technical discussion). Also attributed to Philo are two essays (or perhaps chapters in a larger work or works on semantics): 'Concerning the Language of the Romans' and 'Forms of Verbs' (Gudemann 1913). No wonder then that the ancient Byzantine lexicon called the Suda refers to Philo as a grammatikos ('grammarian'). So thoroughly did his reputation as a scholar of semantics dominate the classical

Eusebius adds (PE 1.9.24) that Philo regarded Sanchuniathon as a 'well educated' (polymathes) and 'very inquisitive' (polypragm6n) person, attributes which he no doubt thought of regarding himself Both the name Sanchuniathon and his early date - indeed, his actual existence, like that of Moses-were long ago deemed suspect by the scholarly world. That led in one instance to the brilliant attempt by one scholar (see below) to recreate Philo's translation of his (Sanchuniathon's) ancient chronicles. The name, at very least, has been authenticated in Punic epigraphy (third century B.C.) as Sknytn (Sakkunyaton), a theophoric personal name meaning '[the god] Sakkun has given': see PNPPI s.v. SKN pp. 365-6, Barr (1974/75: 3637) and Baumgarten (1981: 42-44). It also is attested in the Phoenician personal name Pumayaton (graecized as Pygmalion), which means '[the god)] Pumay has given' (see PNPPI s.v. 'PMY': 391-2; on likely date(s) for Sanchuniathon see Baumgarten 1981: 262-4). A close Biblical analogy for Sanchuniathon would be the Hebrew name Jonathon, '[the god] Yahweh has given'. But even if this supports Sanchuniathon's historicity, it does not prove it.

191

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

literary tradition that the Suda's entry on Philo relegates his Phoenician History to the catch-all category of 'other works'.

Worth noting here are ancient cultural and commercial connections, dating back to the Middle Kingdom, between Byblos and Egypt (see, e.g. the Story of Wen-Amon, c. 1100 B.C., in Brown 1969: 180). In that light, Philo's reference to Sanchuniathon's discovery of some 'ancient records' in the shrines of Ammon need not mean that these were archives in Egypt itself, but instead written texts in shrines dedicated to Ammon at Byblos. On that debate see Edwards (1991: 215 n.14; 217).

Philo's Phoinikika As outlined above, we have only a small portion of the PH, and what survives was excerpted for argumentative purposes by Porphyry and Eusebius. This amounts to fewer than twenty pages of printed text. This might have been more than enough for a detailed abstract of the contents if the excerpts had been representative of the work. Unfortunately, they are not; only portions of Philo's 'Preface' and limited selections from other books were used. It is, however, possible to identify some subjects or topics within even this very limited material. Attridge and Oden (1981:3) noted six headings: Cosmology, Cultural History, A History of Kronos, Accounts of Later Rulers, Human Sacrifice, and Serpents. Baumgarten (1981:38) extracts eight topics from 'Eusebius' discussion of Philo': the six noted directly above, plus 'Porphyry on Sanchuniathon' and 'Philo on Sanchuniathon'. The number of subdivisions are within the eye of the researcher, and we certainly do not know how many are missing. James Barr (1974/75: 22-23) has subdivided the central core of the PH into three sections: a Cosmogony (the creation of the universe), a Technogony (origins of knowledge) and a Theogony (account of the gods). Not all those who study Philo Byblios would agree with that tripartite arrangement. Below I give an outline of the main themes, without attempting to fit them into one or more of these patterns. Those who admire the energy and erudition of an earlier age on this same topic may find it profitable to read two comparative essays by Eduard Meyer: 'Mythen und Gottersagen' (Meyer 1931, Vol. 2:2: 176-178) as well as 'Phoenikische und Israelitische Kosmogonie' (ibid: 178186). What survives of Philo's Cosmogony is reminiscent of several other Near Eastern cosmogonies, notably the Old Testament's Book of Genesis, the Babylonian Enuma Elish, some of the theories known to us from Pre-Socratic Greek thinking, and Hesiod's famous Theogony. From a primordial darkness composed of a boundless mist as well as water, a dynamic that Philo termed 'Desire' generated the basic substance (variously spelled Mot or Mouth), of which all of Nature is composed. Heroic attempts have been made to establish what Philo meant by Mot/Mouth, the etymology of which signifies (in context) either 'primordial mud' or 'shining forth'. Baumgarten (1981: 111-113) favors an identification with the Ugaritic deity Mot, Lord of the Underworld, which is hardly the communis opinio in the field. Somehow he failed to discuss the attractive proposal by Oden (1978: 125-6 and n. 69) that Egyptian m~wt (which actually has the dual meanings noted above) lies behind the Philonic term. On this point see also Attridge and Oden (1981: 76-77 note 29). Ebach (1979: 40-46) suggests that Mot and Mouth represent an original Semitic and later Greek transliteration, respectively, of Canaanite or Ugartic Mot.

Through a succession of cosmic disturbances (including earthquakes, a feature of Phoenicia), levels of intelligent creatures emerge to begin populating the Earth. First mentioned are Aion and Protogonos (presumably of opposite sexes) and their offspring Genas ('Family') and Genea ('Posterity'). Baumgarten (1981: 149) maintains that Genos and Genea 'must be transliterations of some Phoenician names' which so far elude identification. That they might represent translations is equally possible. Sanchuniathon's cosmology, however Philo might or might not have modified it, was not the only 'homespun' version. We know of one other, by the Phoenician 'atomist' Mochos of Sidon, of uncertain date and background. His work (Moscati 1973: 5455; Millar 1983: 64; Strabo, Geo. 16.2.23-24) was transmitted via an excerpt preserved by the neo-Platonist Damascius (c.A.D. 500?). Attridge and Oden (1981: 103) draw attention to the passage, a portion of which is worth reproducing for comparison and contrast with a corresponding passage in a fragment of Philo's PH:

At first there was Aether and Aer, the twofirst principles, from which Oulomos, the god perceived by Intellect, was created .. They say that from him, when he had intercourse with himself, there was born Chousor, the 'First Opener', and then an Egg. The latter is the mind perceived by intellect, and the Opener Chousor is the Intellectual Power, in as much as it first distinguishes Indistinct Nature ... Chousor is the first order after that perceived by Mind. The Egg is heaven, for when it was broken in two, Heaven and Earth came into being ... (Damascius, de Principiis: 125). The fragments of Philo's PH make no mention of Aether/Aer, and the 'self-made' Chousor ofMochos becomes one of two brothers credited with the discovery of iron (Eusebius, PE 1.10.13 = Attridge and Oden, 1981: 45). 'Chousor', according to Philo, 'practiced verbal arts--including spells and prophecies. He is in fact Hephaestos, and he invented the hook, lure, line and raft, and was the first among men to sail'. On the 'Chousor' mentioned in both passages see Brown (1965: 201-203), who puts them into a more meaningful 'mythic' perspective. Whatever their etymology, Genos and Genea settled in Phoenicia and their descendants discovered fire, named great mountains, founded cities, hunted and fished, smelted iron, and fashioned ships which could sail the Mediterranean. The sun-god Ba 'alshamin was the deity worshipped by Genos and Genea as the source of all life. Worthy of note here is that Ba 'alshamin's major shrine in historic Phoenicia was the

192

Henry MacAdam: Philo of Byblos and the Phoenician History

cult-centre at Heliopolis/Baalbek. Apart from Ba'alshamin, other pre-eminent deities within this proto-Phoenician pantheon included Elioun (defined as hypsistos, 'most-high') and his consort Berouth (goddess of wells and springs) who, Philo adds with obvious pride, 'settled near Byblos'. That allusion must be to Berouth's ultimate identification as the eponymous deity of Beirut. Elioun's grandson El (Kronos) is credited with founding Byblos, 'the first of Phoenicia's cities'. These deities and their progeny, who are as manifold, diverse and complex as those of Rome or Greece, share the Cosmos with their human counterparts on Earth. They also share human emotions as well as human failings, but they don't appear to interact directly in human affairs (which if true contradicts what Greeks and Romans understood about Phoenician gods and goddesses). Philo's PH, had we the full text, no doubt retailed the loves, hates, jealousies and deceptions of the immortals, their drunken rages and their libidinous behavior. Surely Philo's deities wouldn't behave very differently from their counterparts elsewhere, who steal, quarrel, bargain and bluster. They kill, and are killed. They expect man to serve them, fear them, and to build shrines and temples dedicated to them. They are, very simply, much like humans.

that was later dedicated to Abibalos, King of Beirut. 'Abibalos of Beirut' is unknown from any other source, though there are homonymic kings of Arados, Byblos and Tyre (Baumgarten, 1981: 57 with citations). Philo attributed Sanchuniathon's information to 'Taautos [who] was the first to have invented "the letters" (ta grammata) and to have begun the keeping of records' (Baumgarten 1981: 63 and commentary, 68-73; Barr 1974/75: 37-40). These 'writings' had been discovered by Sanchuniathon, i.e. 'secret works procured from the sanctuaries, composed in the letters of the Ammoneans, which certainly were not familiar to everyone'. This reference to the 'Ammoneans' appears in an especially difficult passage, examined by Attridge and Oden (1981: 31 and n. 11) in detail. Could it be that Philo, or his source regardless of its antiquity - was aware that the Phoenician script was itself preceded by another form of writing? A script that had become so archaic by Sanchuniathon's time that it could be read (as was true of hieroglyphic Egyptian) by only the most learned? Philo's 'Ammoneans' may be associated with the Egyptian god Ammon. Philo is more specific about the name Taautos, telling us that it is the Phoenician rendering of Egyptian Touth or Thoth, identified by the Greeks as Hermes. The combination of 'Egyptian' elements in the passages is striking: an ancient scribe/seer with an Egyptian name (Taautos); long-forgotten 'texts' in an enigmatic script associated with sanctuaries dedicated to a major Egyptian deity (Ammon) - each of those alone would be noteworthy (Ebach 1979: 60-71).

Throughout the fragments of the PH there is a distinct euhemeristic or rationalizing aspect with regard to both the origins and to the characteristics of the Phoenician pantheon. (Attridge and Oden 1981: 7-8; Baumgarten 1981: 38-39; 8082). It would seem that Eusebius was already aware of that when he wrote an 'epilogue' to the segments of the PH which he excerpted. Here is part of what he so thoughtfully set down for us: The theology of the Phoenicians has this aspect. The salvific word proclaims that we [Christians] flee from this without turning back and search out the cure for ancient peoples' madness. For the myths and fictions of poets do not happen to contain some hidden theory in allegoricalform ... It should be evident that it is no longer necessary to track down forced explanations of natural phenomena in these matters, since the facts of themselves provide a clear refutation of such an endeavor. Such then, is the theology of the Phoenicians (Attridge and Oden 1981: 69).

Bronze Age literary sources, in concert with archaeological work at the site, confirm that the port of Gubla/Byblos was then Egypt's major outlet to central Canaan. The Story of Wen-Ammon (noted above) occurs precisely when Egyptian weakness following the invasion of Sea Peoples is evident. That is the transitional period (c. 1100 B.C.) when the Phoenicians as an identifiable culture first appear in historical and literary sources (Rollig 1983). Might it be that Philo's mysterious script, if not from Egypt's hieroglyphic system, was the enigmatic 'Byblos Syllabic' associated through archaeology with Byblos' Middle Bronze Age period (Mendenhall 1985)? This 'corpus' of just nine texts, some stamped on copper plates and several on spatulas, others incised on stones, dates probably to the mid- and late second millennium B.C. They include a royal decree, a marriage contract, a building inscription of some sort (a temple dedication?) and very short texts whose meaning is uncertain (perhaps because they are incomplete). Identifying Byblos syllabic with Philo's palaia grammata is obviously highly speculative. So, equally, is making Philo Byblios' Taautos the same semi-mythical person as Porphyry's Hierombalos, or assuming that Sanchuniathon's ancient sources had flowed from only one fountain (with royal patronage!).

Brave words, even inspired words, from the Bishop of Caesarea, who can look back on the pagan world from his post-Nicaean, friend-of-the-Emperor vantage point and lay out what he considers the short-comings of the 'non-salvific' realm. Nothing speaks more loudly or clearly than these words ofEusebius to my assumption (note below) that Philo had no interest in - probably no real understanding of - those who were Christian. His interest in Judaism was intellectual. But Sanchuniathon of Berytus wasn't the ultimate source of Philo's information. The excerpts from, or the later comments about, the PH demonstrate that Philo believed that Sanchuniathon gathered, sifted and borrowed from a collection of earlier written material and oral traditions (Barr 1974/75: 36-40; Baumgarten 1981: 41-61, 63-92). According to Eusebius (PE 1.9.21), Porphyry asserted that a certain priest named Hierombalos had composed a body ofliterature

We are here in the realm of tradition, which can (and does) embrace a multitude of sources enriching the flavor of the story. Philo was well aware of that. It would seem that wherever we look in the PH, and in the association of its transmitter with his home of Byblos, we find traces of 193

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Egyptian culture. This cannot be accidental. The awareness of the cultural legacy of Egypt, intertwined with the history of Byblos, which surfaces precisely during the reign of an emperor (Hadrian) so intensely interested in antiquarian matters (especially following an extended tour of Egypt?), is surely worth bearing in mind.

nineteenth century can simply be written off as bias based largely on ignorance and inspired by prejudice. Very little was known then about the Phoenicians or their culture apart from references in the Bible and from Greek and Latin sources--most of which are unflattering (the incident involving a Syro-Phoenician woman in The Gospel of Mark [7:24-30] is a notable exception). But with the advent of sound Phoenician scholarship a more reasoned and informed approach might be expected. Archaeological research, which until c. 1900 was often undisciplined and unreliable, soon began to add a contextual, scientific dimension to the process of research.

Philo and Modern Criticism

The Phoenician History, its 'author' Sanchuniathon of Beirut, and its 'translator' Philo of Byblos have undergone periodic, intense scrutiny since the late eighteenth century. Some of the opinion expressed has been outspokenly sceptical, with other views less so (Albright 1941 provides an excellent summary). Philo's reputation certainly wasn't enhanced through exposure of a literary hoax concocted by Friedrich Wagenfeld. Wagenfeld was a twenty-five year-old German scholar, student of both theology and philology at the University of Gottingen (1829-1832), who, in 1836, published what he claimed was a translation into German of extracts from 'the newly-discovered manuscript' of Philo's Phoenician History. This he claimed to have found 'complete' in a Portuguese monastery at Oporto. Wagenfeld's epitome, boldly entitled Sanchuniathons Urgeschichte der Phonizier (Hanover, 1836), was given a thirty-two page foreword by G.F. Grotefend, then Director of Hanover's prestigious Lyceum. Wagenfeld supplied an introduction to the 'translated' excerpts. This proved to be so popular that it was followed (a year later) by Wagenfeld's publication (with a Latin translation on the opposite pages) of the hitherto 'lost' Greek text (now 'recovered' in nine books), claimed to be Philo's original translation of the Phoenician History. What motivated Wagenfeld, other than the intense intellectual delight of duping his acclaimed academic superiors, still remains a mystery. W agenfeld gave this second volume the impressive, even grandiose, Latin title of Sanchuniathonis Historiarum Phoenicae Libras Novem Graecae versos a Philo Byblio, edidit Latineque versione donavit F. Wagenfeld Bremae [i.e. Bremen] 1837. That proved to be hubristic; both books were read by the preeminent classical philologist, Karl Ottfried Muller, who immediately exposed these 'sensational' works as fraudulent. (For a brief biography of Wagenfeld see Allgemeine Deutsche Biographie 1896, 40: 467-77 and Farrer 1907: 191201.) A recent sketch of Wagenfeld's escapade (in the context of other such hoaxes) is Haas (1985: 72-75--with a facsimile ofa segment ofWagenfeld's Greek text at 73). In spite of (perhaps because of?) Wagenfeld's notorious prank, the Phoenician History remained a topic of great interest through the later nineteenth century; James Barr could write, a quarter of a century ago, that 'it's probable that more people knew about him [Philo] a hundred years ago than know of him today' (Barr 1974/75: 18). Even so, cognizance does not necessarily imply confidence, and if Philo of Byblos enjoyed greater name-recognition then (at least among the academics and literati), few expressed support for his credibility. Criticism prior to the mid-

One avenue of enlightenment was to set Philo's Phoenician History against a solid background of linguistic and literary studies. The very same year (1837) that Wagenfeld's hoax was exposed by Muller, the first important work on Phoenician epigraphy--Wilhelm Gesenius' Scripturae Phoeniciae Monumenta--was published. That was followed, four years later, by the first installment of F.K. Movers' erudite but flawed five-volume monstrum on Phoenician history and culture, Das Phonizische Alterthum (1841-56). Movers was (initially) severely sceptical about Philo's Phoenician History as a source of useful, historically verifiable information, but in subsequent publications he modified those views. Movers' 'shift' of opinion gained some ground for Philophiles, but nothing prepared the small world of Phoenician studies for the scornful denunciation of George Rawlinson, whose one-volume History of Phoenicia appeared in 1889. In it, he had this to say about Sanchuniathon's champion:

'... the literary value of Phi/o's work [is] exceedingly small. His style is complicated and confused; his material, for the most part, worthless, and his mixture of Greek, Phoenician and Egyptian etymologies absurd. If we were bound to believe that he translated a real Phoenician original, and [that] original was a fair specimen of Phoenician literary talent, the only conclusion we could come to would be that the literature of the nation was beneath contempt. ' (Rawlinson 1889: 388-9). Rawlinson's opinion, condescending as it was, stopped just short of denouncing the Phoenician History as totally fraudulent. He was not aware that the heaviest blow against Philo had already been struck when History of Phoenicia was well into press. Otto Gruppe's first volume of a large study, Die griechischen Culte und Mythen in ihren Beziehungen zu den orientalischen Religionen (Vol. I, 1887) clearly delineated the extremist negative view, declaring Sanchuniathon 'a fake' (Gruppe ibid: 373-4). That total rejection of Philo's source was exceptional but at least it reflected the prevalent scepticism of the time far more clearly than did the enthusiastic support for Philo expressed a generation earlier by the French savant Ernest Renan (1858). Even if it could be proved that Sanchuniathon never existed, did that judgment include what Philo claimed were the ancient writings he translated? If not, was it possible that

194

Henry MacAdam: Philo of Byblos and the Phoenician History

Philo himself did not 'invent' those texts but simply translated material which he believed was genuinely ancient?

cause (Albright 1946: esp 69-71; 71-76; 81; 93-94 passim). This post-World War II rebirth of PhiloF orschungen reached a peak in the decade 1968-78, perhaps as a result ofan event and an 'issue' which coincided then. The event was the archaeological work undertaken at Sarafand (ancient Sarepta/Zarephath) between 1969-74 by the late James B. Pritchard, the first such investigation of an uninhabited Bronze Age/Iron Age urban settlement (situated between Sidon and Tyre) within the Phoenician homeland (Pritchard 1978). The 'issue' was a controversy during the same years regarding the authenticity of a Phoenician inscription (describing a journey of about the mid-sixth century B.C. This inscription was allegedly found and copied in the 1870s on the northeastern coast of Brazil (Van Den Branden 1968; Gordon 1968a and 1968b, 1972; 1973: 120-6; Cross 1968 inter alia; see Elenchus Bibliographicus 50 [1969]: 372).

There may be an instructive parallel to this in the tangled history of a literary work from the Severan era, namely the Ephemeris Belli Troiani, allegedly the memoirs of 'Dictys the Cretan', a companion of Idomeneus, leader of the Cretan contingent against Troy (Iliad 2. 645-52). This archaic 'original' was translated from Greek (i.e. primo conscripsit litteris Punicis) into Latin by L. Septimius (of uncertain origin), whom Champlin (1981) has tried to identify with Septimius Serenus (OCD 3 1390) and Serenus Sammonicus (OCD 3 1392). Champlin's study notes that literary forgeries were common throughout antiquity, just as they are today: 'It cannot be emphasized too often that the text of Dictys Cretensis is a Latin version [but not a translation] of a lost Greek original, [i.e.] we are dealing with a two-layered forgery... [Thus] Dictys is clearly and incontestably a fiction, but...is there any reason to suppose that Septimius was not precisely what he claims to be, a Roman scholar who chanced one day upon what he believed to be an authentic history?' (1981: 195). Much of the nineteenth century's skepticism and doubt concerning the late second millennium B.C. date of Sanchuniathon centered on the nearly complete lack of comparative written material from Phoenicia itself (the central and southern portion of biblical Canaan). None of the literary material (historical and/or mythological) from excavations at sites in Sumer, Babylonia, or Assyria shed light on the world of Sanchuniathon, nor did contemporary tum-of-the-century excavations in the Aegean area (notably Mycenae, Tiryns and Knossos). That absence of contemporary texts (other than Homer, Hesiod, and the Bible--all from the period after 1,200 B.C.) was rectified very dramatically in the 1930s. Then, French excavations yielded the royal archives from the Late Bronze Age Mediterranean commercial port of Ras Sharnra/Ugarit in northwestern Syria (see Barr 1974/75: 19-20; Baumgarten 1981: 2-6). For the first time scholars had at hand an abundant corpus of textual material, older than most of the Hebrew Bible and the extant archaic Phoenician epigraphy, and all from the ruins of a city (Ras lbn Hani) and its hinterland within the immediate vicinity of the northernmost edge of Phoenician territory. The epic myths of Ugarit, written with syllabic cuneiform and in a language with close affinities to Phoenician and Hebrew, commanded the attention of comparative linguists and students of Near Eastern religion. At first glance it seemed that this material, featuring a pantheon with familiar names such as El, Ba' al, Astarte, Mot and others, was expressed in a linguistic style and structure which distantly 'echoed' the extant fragments of Philo. This, therefore, might come closest to offering the parallels sought by some who claimed a Late Bronze Age date for Sanchuniathon. Ugarit brought back Philo with a vengeance (e.g. Eissfeldt 1952; 1956). A triumphant vindication of Philo and the Phoenician History seemed at hand when the American biblical scholar William Foxwell Albright championed their

The stone itself on which the enigmatic eight-line inscription was supposedly incised has never been recovered; even if it were found again it would not be in situ. Thus the handwritten copy, made just over a century and a quarter ago, which attests a storm/ shipwreck and survivors of a Sidonian expedition which departed from the Red Sea and circumnavigated Africa, only to be blown off course and reach landfall in what is now Brazil, is likely to remain (in spite of Guzzo 1988) an unsolved epigraphic mystery. In passing, it may however be worth noting that Herodotus (Histories 4.42) recounts a similar maritime feat only a halfcentury or so earlier. The very least that can be said is that renewed interest in Philo's Phoenician History coincided with both the excitement of Sarepta's excavation (unfortunately truncated by the Lebanese civil war), and the furor over the 'Paraiba Text' from Brazil. In no small measure, the re-emergence of scholarly Philophilia was due to an article by the former colleague of Professor Barri Jones, Professor James Barr (1974/75, noted above), who represents a growing number of scholars equally at home in the fields of Classics and Semitics. In the decade opened by Barr's study, no fewer than five new books on the same subject appeared. First was Troiani (1974), published almost simultaneously with Barr. Slightly later came Ebach (1979), then the volumes by Attridge and Oden (1981) and Baumgarten (1981), and lastly Schiffinan (1986). All of the last four acknowledge an enormous debt to the topicality, retrospective scope and interdisciplinary approach of Barr's scholarship. Attridge and Oden, and then Baumgarten, cast their nets very wide to bring a large segment of the ancient sources and modem criticism to bear on the career of Philo and the scope of his works, as does Lucio Troiani. Schiffinan's approach is wider yet, bringing into focus the spread of Phoenician culture far beyond the Levant, with illustrations of archaeological material to support the textual evidence. Philo is therefore little more than a sideline, given no more attention than short studies designed either for references (e.g. Sznycer, 1966) or critical review (e.g. 195

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Lipinski, 1983). Jurgen Ebach's Weltenstehung und Kulturentwicklung bei Philo van Byblos is primarily an attempt (see especially pp. 22-71 passim) to isolate Semiticisms other than proper names in the text of Philo's Phoenician History, followed by a detailed analysis of the fragments through a comparative study of their mythdevelopment with regard to biblical and related sources. Widespread as the interest in Philo was in the 1970s and 1980s, it failed to attract the attention of the editors of that never-ending Festschrift, Aufetieg und Niedergang der romischen Welt, which has so far devoted a total of twentyeight volumes to 'Religion' in the imperial period, and ten to 'Sprache und Literatur', several of which are devoted specifically to the Hadrianic era. Nor has Philo piqued interest among contributors to the ongoing series Studia Phoenicia, which to date has published about fourteen volumes. It remains to be seen if the pendulum will swing back to Philo Byblius at some point in this new century.

the famous law school in the Roman colonia of Berytus. Imperial interest in Phoenicia is also reflected in the successful petition of Tyre, through the good offices of the rhetor Paulus (a contemporary of Philo), to gain formal recognition as a metropolis. As several scholars have noted (e.g. Millar 1993: 288 andn.13; Birley 1997: 227-228), Tyre was already attested by that title in the 90s. It is not until Paulus headed an embassy to Hadrian that the city formally received that prestigious recognition. That would mean any earlier representation of Tyre as metropolis is to be taken as a literal reference to its ancient role as 'mother' of Phoenician colonies in Cyprus and North Africa. Millar resolves that anomaly thus: 'There is no real contradiction: it would fit perfectly with the nature of intercity competition if Tyre had first used this title, then had it challenged, and then secured it by an embassy to the Emperor' (Millar 1993: 289). Paulus' success with Hadrian may be commemorated in a special way. A headless imperial statue, identified as Hadrian, was found along the spectacular colonnaded entrance into Tyre during the 1950s:

As all studies noted above examine numerous facets of the PH (with variations of emphasis), and its main features have been noted, it would be worthwhile now to explore its context, i.e. the Phoenicia of Philo's day. In doing so, I hope to provide the motive for Philo's publication of the PH at this particular time, which is one aspect of PhiloForschungen which has been neglected.

Elle eut peut-etre pour but d'affirmer le loyalisme de Tyr envers l'Empereur [i.e. Hadrian], dont la politique economique, qui lui etait si profitable, se trouvait compromise par la revolte juive de Bar Kokeba (Chehab 1962: 33).

Philo and Hadrianic Phoenicia

If that identification is accurate, Tyre may have commissioned the statue both to honor Hadrian for officially designating the city as metropolis, and also to record his presence in the city. Tyre also boasted of being home to Marinus, the premiere geographer of the day and an ahnost exact contemporary (c.70-c.140) of Philo Byblius. That was the kind of company Hadrian liked to keep, and we may find even more compelling evidence for his association with Tyre.

Philo published the PH when Phoenicia was enjoying much imperial attention and many benefactions. Phoenicia hadn't experienced that since the founding of the veteran colonia at Berytus (including in its territory Heliopolis/Baalbek) under Augustus, and the parallel favor of the Herodian dynasty (Millar 1993: 264-295, incorporating Millar 1990 and Millar 1983). Philo may have become inspired by the personal and positive effects of Roman interest in Phoenicia (this process is summarized in Hall 1999, but without mention of Philo; see also Hall 2001). Phoenicia's identity did not come to its fullest fruition until the Severan dynasty at the end of the second century, when the new province of Syria Phoenice was created (194), from which the future emperor Severus Alexander originated. Nevertheless, I hope to show below that the inspiration for that 'ethnic' renaissance has much to do with Hadrian's personal and imperial interest in Phoenicia, and the response to Hadrian's benefactions by certain men of that time. This is precisely the time that many cities throughout the empire, from North Africa to the Syrian steppes, enjoyed imperial favors.

Yet another consideration is whether Hadrian, as his biographer in the Historia Augusta maintains, actually considered a sub-division of Syria into two provinces: Syriam a Phoenice separare [Hadrianus] voluerit (HA 14.1). Syme (1983) and Bowersock (1985: 87-88) argued (independently) that the statement retrojected the partition of the province (allegedly prompted by a profound imperial dislike of the Antiochenes) from a decision by Septimius Severus in 194 to one by Hadrian some 60-65 years earlier. But the HA entry, whether or not it is misleading, draws attention to something important. As Bowersock puts it, 'the passage alludes to a genuine concern of Hadrian' (1985: 87). That could only be the granting of metropolis status to Tyre and also to two cities of the Syrian interior (Damascus and Samosata), as we know from the evidence of coins (Syme 1983: 322 n.7). Hadrian's motive isn't clear, but raising the municipal status of certain cities is a feature of his reign. Antioch received other benefactions from the Emperor. Though Hadrian did not partition Syria, his granting of approval for two new metropoleis in the province may have prompted the decision later made by Septimius Severus.

Josephus fought against Rome in the war that ultimately destroyed Jerusalem and dismembered the Jewish homeland, and he wrote with a sense of foreboding that the Hebrew/Jewish culture of Palestine might not recover from the devastation wrought there in AD. 66-73. Philo by contrast (as is argued below) witnessed or at least reacted to, an imperial visit from Hadrian which established a forest preserve on Mount Lebanon and (perhaps) as I suggest elsewhere (MacAdam, 2001; summarized below) founded

196

Henry MacAdam: Philo of Byblos and the Phoenician History

The Roman law school at Beirut, often considered a Severan foundation, may in fact have been founded by Hadrian. Berytus was the home of two great grammarians, both younger contemporaries of Philo: Marcus Valerius Probus (OCD3 1580) and Hermippus (OCD3 692). Probus, a collector of old Latin texts and commentator on Plautus, Terence, Sallust and Vergil, was most probably a descendant of a family of military veterans from the earliest days of Roman Beirut (Hall 1999: 87-88). The Suda notes that Hermippus originated from 'a village near Beirut', i.e. most probably within its territorium. Born into slavery, he became a pupil of Phylo Biblius and later the author of several books now lost. Hadrian would have been well aware of Beirut's contribution to contemporary literature, and the city's century-and-a half-tradition of Romanitas. That, plus his own attested interest in jurisprudence (Gaius, Institutiones 1.7), and the Emperor's own rescriptio on the subject of juristic responsa (official opinions) is also known (on the Institutes, Gordon and Robinson 1988: 23; on Hadrian's responsa, Torrent 1983). Gaius' dates remain uncertain (c.120-c.180). He is thought to have been a teacher oflaw both in Rome and in an eastern city (Beirut is most likely, ocn3620) by the 160s. As his modem 'biographer' puts it: 'It is impossible to show a direct connection between Gaius and the law school of Berytus, but if a tradition connected him with it this would partly account for Justinian's use of [the expression] Gaius nosier: [i.e.] the founder of our most famous law school.' (Honore 1962: 126) The point is an important one; if Gaius had anything to do with the law school in Beirut, then its foundation-date must be pre-Severan, since it is improbable that he would still have been alive when the school was supposedly founded c. 200. At the very least, as Honore (1962: 126) would like to believe, Gaius brought to completion his study of one Roman law (Senatus Consultum O,jitianum) while living in Beirut. But Hadrian's founding of the Beirut school must remain, for the present, non liquet.

the city was granted colonial status c. 198] and to offer no reflection whatsoever of the existence of emperors.' It is just possible that Hadrian's connection with Phoenicia, i.e. with individual cities of that portion of Syria, goes back to his military service during Trajan's Parthian War (113116), followed by his brief governorship of Syria. Edward Dabrowa summarizes that episode thus:

[I]t is known that when Trajan was leaving Syria after the war with the Parthians at the turn of July and August of A.D. 117, [he appointed Hadrian] as legate of the province. [Hadrian] must have assumed that office a little earlier, probably about the middle of the year ...[he] governed the province for a few months only. After Trajan's death, when Hadrian took power as emperor, he appointed a new governor [L. Catilius Severus] of Syria before he left Antioch (Dabrowa 1998: 89 and the sources quoted there). By Hadrian's reign the unregulated exploitation of the forests of Phoenicia had become a subject of primary concern. The decision to create an 'imperial preserve' on Mount Lebanon was perhaps related to one or more of the four putative visits which Hadrian could have made to Phoenicia: as governor in 117; during the Parthian campaign of 123; during his imperial tour of the eastern provinces in 129/130; and again at the time of the outbreak of Second Jewish revolt only a few years afterward (132). Syme (1983: 328) asserts that Hadrian 'saw Antioch again [after his brief governorship] towards the end of his first journey abroad' (in 123 or 124: see also Syme 1988: 160-61; Birley 1997: 153). Apart from the hellenized Phoenician cities themselves, there were other attractions that would have beckoned to a Roman emperor, such as the cult-centre of Heliopolis and the colonia of Berytus, both thoroughly Romanized after 150 years. There is some compelling evidence that Hadrian visited both places, but probably not during a journey to the Parthian frontier to conduct a military campaign.

Just as tantalizing, and yet just as vexatious, is establishing the date at which Colonia Julia Augusta Felix Berytus was bestowed with ius Italicum. The first mention of it occurs in Justinian's Digest (50.15.7); the reference there is to the jurist Gaius. Once again, if that legal right was granted to Berytus during Gaius' lifetime, it precedes the Severan period. It would seem an appropriate honor for the city to enjoy at the time the law school was created, and Hadrian - as a benefactor of cities and provinces - might have been the most likely grantor of such a privilege.

Millar (1990: 19; 1993: 279-280) drew attention to the illustrious career of M. Licinius Pompenna Pititus Urbanus, honored by Berytus with a statue in his honor at Heliopolis. Only the base remains on which the dedicatory inscription registers the civic offices held by him (IGLS 7: 2791). Noteworthy is the fact that Urbanus had been granted the status of knight (equus publicus) by Hadrian, an honor that Birley (1997: 230) suggests (correctly, I think) was awarded him in person by the Emperor. Byblos was but a short trip up the coast. Three pieces of evidence indicate the emperor's physical presence in Byblos. One is a reference in the Suda to Aspasius ofByblos, who is reported to be 'a contemporary of [the orator and writer Aelius] Aristides and of Hadrian'. More importantly, Aspasius was credited with producing an encomium on Hadrian 'and some others' of the same period. Where Aspasius lived when he wrote about Hadrian is not known (the Suda's reference to him is reproduced in FGrH 792 Tl); he may have been resident in Byblos during an imperial visit, as suggested by the composition of the encomium (on that see Brown 1969: 153). The second is a

There are no commemorative coins, imperial or provincial, struck in honor of Hadrian's visit to any Phoenician city. That it itself has no negative connotation regarding visits he may have made to any of them. Tyrian coins are the best known of Phoenician cities', and, as Millar (1993: 289-290) points out, even when the provincial mint at Antioch took over the production of Tyrian silver shekels in 58/59, 'the bronze coins of Tyre continued to be resolutely "Tyrian" [i.e. until

197

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

fragmentary Latin inscription (CIL 3.6696) copied during the late eighteenth century (Rampoldi 1824: 454-455) and not rediscovered since. Rampoldi transcribed it from a re-used stone sopra la prima casa entrando dalla porta di [lbJrahim in Byblos The text is badly mangled but it clearly makes reference to the 'restoration' of something under Hadrian. What Rampoldi copied was reproduced without emendation by Mommsen from the 1814 edition (not available to me) edition of Rampoldi's journal of his travels in Syria during the year 1784:

more than just the boundaries of the imperial domains. They outline as well the limits of the territorium of Roman Berytus. Whether Byblos lay within that territory, which Breton asserts, is not so easy to determine - coins of Byblos would suggest otherwise. Certainly its nearness to Beirut and the forestry preserve help to explain the city's use of Latin not just regarding the inscription Rampoldi copied but also for CIL 3. 6697, the tombstone of a soldier, the only other inscription ofByblos included in that volume. Halfmann (1986: 194) suggested that following Hadrian's departure from Alexandria (late winter/spring 131), he undertook a 'Seereise entland der syrischen Kiiste'. Hadrian is not attested anywhere until he reached Phaselis (Lycia) that summer (TAM II 194); among several stops he must have made along the way it is not overly speculative to submit that the major Phoenician cities would have welcomed a visit from the Emperor. Surely Byblos would have been one of them.

D. TRAIANI.F. CONS.IV. ..MARIS VBIQ .. . OMNES. ..PORTv. .. RESTAVR .. ..

Mommsen's cryptic comments are: Admonuit de titulo Lumbrosus [i.e. amicus} noster (a reference to a colleague) followed by lnterpolata omnino. It is indeed difficult to know what Rampoldi, who displays no epigraphic skills, actually saw when he copied the inscription. His own description of this frammento d'inscrizione omits mention of how many lines, the amount of space between the word breaks, or the size of the letters. CONS must be a mistake for COS, IV for Ill (Hadrian held only three consulships, the last in 119). PORTV[M} is possible, as Rampoldi concluded (sembra che questo porto sia stato ristaurato dall'imperadore Adriano), but the letters could just as easily be expanded as PORTIC[OS). Could MARIS VBIQ be the garbled remnants of some variation on MARMORIBVSQUE? We do not know what or who was the subject of the verb restauro, and therefore whether it should be rendered as singular or plural.

Admittedly those three items are only circumstantially related. However, collectively they suggest that Hadrian actually visited Byblos, and formulated a plan to survey, define, and regulate the harvest from the forest of Mt. Lebanon. A part of that plan could have been the restoration of Byblos' harbor to facilitate the shipment of timber, a view that is taken for granted (on the basis of CIL 3.6696) by Chapot (1907: 337) and later Rey-Coquais (1978: 53). The former saw the harbour at Byblos in a military context: a strategic port on the Via Maris connecting Antioch and Ptolemais. The only other opportunity would have been when Hadrian came east at the onset of the Second Jewish War in 132. The emperor is attested in Athens during the winter of 131/32. Halfmann (1986: 209-210) tried unconvincingly to show that Hadrian ignored events in Judaea, and returned directly to Rome from Athens. Syme took issue with that, noting that 'Hadrian ...did not make for [Rome] when he left Athens in 132. He spent the winter of 132/33 somewhere within reach of Palestine, making dispositions for the next year when (one assumes) he supervised the operations ofhis generals' (Syme 1988: 167). The location of Hadrian's headquarters for that winter is as yet an uncertainty; both Antioch (OCd: 485) and Gerasa (Kindler 1983: 37 n. 19) have been proposed. Neither seems a very suitable place from which to conduct a war in Palestine. Caesaraea or Tyre would be more likely, either city allowing the emperor plenty of opportunity to visit sites (including Tyre/Beirut/Byblos) along the Phoenician coast. Siege-engines would have required a large supply of timber; we know that Hadrian had need of them (Schurer 1973: 550).

But unless PORTV[M} is correct, and refers to a restoration of the city harbor, one might read instead OMNES [COLUMNAS} ET PORTIC[OS} RESTAVR[AVIT}?? or RESTA VR[A VERUNTJ?? (a suggestion which I owe to Brent Shaw). Perhaps this was a project initiated by the provincial governor after damage by storm or earthquake, or repairs undertaken by Byblos as the result of the Emperor's visit. Rey-Coquais (1972: 99 note 4; 1978: 53) saw the restoration to be of the port itself, 'sans doute pour permettre !'exploitation de ces forets'. Thus the inscription may be coeval with those more famous inscribed boundary stones (almost 190 are so far known) marking the definitio of Hadrian's forest preserved in the hills above and to the east and west ofByblos (Breton 1980: 27-35; Brown 1969: 140 212, esp. 152-155). A typical inscription (ILS 9385) reads: IMP(eratoris HAD(riani) AVG(usti) D(e)F(initio) S(ilvarum) A(rborum) G(enera) IV C(etera) P(rivata) These inscriptions must be dated no earlier than 123, the year that Hadrian added the honorific title Augustus to his titulature. There was a theory that some or all of the inscriptions (cut into living rock) were created during Hadrian's twentieth year; Breton (1980: 12) demolished that idea. This unique project must be linked to an imperial sojourn in Phoenicia--perhaps during the 'provincial tour' of 129/131.

Philo's own response to the visit might have been his essay On the Reign of Hadrian (since lost), but it had a far greater effect than that. Philo's extra inspiration could have been twofold. On the one hand he had an opportunity to provide a comprehensive, cultural history of the Phoenicians as an ancient and worthy people who participated in the civilizing of the Mediterranean world, on the other he was an observer of the desolation that came upon a neighboring culture. The

When the forestry inscriptions are plotted on a map of Lebanon, as Breton (1980: 27) acutely noted, they delineate

198

Henry MacAdam: Philo of Byblos and the Phoenician History

utter defeat of the Jews a second time came, of course, at the hands of the same Romans who had founded Colonia Julia Felix where the ruins of ancient Beirut had stood (destroyed by Tryphon in the late first century B.C.). The same Romans, also, who brought to completion the massive and magnificent temple complex they devoted to luppiter Optimus Maximus at Baalbek high in the Lebanese central valley (Ragette 1980). But in the collective memory of at least some Phoenicians, perhaps of Philo Byblius, was the recollection of Rome's systematic destruction of mighty African Carthage. While it is possible that Josephus' passionate and nationalistic account of his culture and history may have inspired Philo, neither the Antiquities of the Jews nor the Phoenician History was written for local readership. The Greek of both works was intended for a wider, Hellenized audience, one for whom the panorama of Hebrew and Phoenician history was either imperfectly understood, or completely unknown. In that sense, Philo's PH must have been larger and vastly more comprehensive than the surviving fragments would demonstrate. More than twenty years ago Giovanni Brizzi (1980) suggested that a combination of factors, centred on Hadrian's antiquarian interests and his curiosity about eastern cultures, prompted Philo to produce an apologia for his own people. Brizzi believes that Hadrian's avid interest in a well-governed (and self-contained) oikoumene, reduced in size following Trajan's recent ambitious territorial expansion - an imperial domain in which ethnic diversity might well flourish - struck a responsive chord in Philo (on Hadrian's unofficial policy of empire-wide 'containment', see Birley 1997: 1, 133-4, 304). The alternative to the enlightened acceptance of Roman hegemony was evident to anyone alive in the fourth decade of the second century. Hadrian's harsh reprisals against the Jews (132-35), enacted within a province contiguous to Phoenicia, was, for Brizzi, the crucial and defining event that led Philo to write both On the Reign of Hadrian (now lost), and an account of Phoenician culture based to an unknown extent upon Sanchuniathon's indigenous, ancient sources. Several aspects of the PH can be deemed certain. Much of the source material is genuinely Phoenician, even though the dates of several component parts are debatable. Many of the sources, in origin, were in the form of poetry. Oden (1978) argued strongly that Philo's source(s) should be regarded confidently as no earlier than the Hellenistic period. His reasoning is that certain 'motifs' or 'tendencies' within the PH are transparently Hellenistic features. Among them he notes a euhemeristic attitude toward divinities, the 'universal' scope of the work, a 'patriotic' flavour and its logical concomitant, an anti-Greek bias and, lastly, claims that the source or sources used are older and better-authenticated than those used by rival cultural historians or antiquarians. The force of this detailed argument is blunted by the fact that the surviving fragments of Philo's PH are not known to us as the result of a serendipitous modem discovery, parallel to the discoveries in Egypt of literary papyri of various periods within that province's Graeco-Roman history. Rather they exist today because Porphyry and Eusebius made specific selections, for the purpose of refutation or illustration, from

the full text which presumably was available to them. That 'editorializing' aspect alone should make us cautious when we attempt to reconstruct the outline of the PH and/or identify which features are specifically Hellenistic. Yet it is worth noting, in support of Oden's argument, that Renan (1858) dated Sanchuniathon (whom he considered historical) no earlier than the Seleucid era. That is at least a more moderate stance than several others would allow (e.g. Nautin 1949: 573). A more balanced view than any of the above is Fergus Millar's brief assessment of Philo and the PH in his comprehensive volume dealing with the hellenization of the cultures of the indigenous people of the eastern Mediterranean: 'It is in fact inconceivable that what Philo wrote (or copied, or interpreted) could be earlier than the Hellenistic period; and it may well be original to himself ..his work does contain a series of specific, and largely reliable, interpretations of Phoenician names and concepts...[but Phi/o's PH} cannot be taken as a factual report on established communal beliefs; it is an erratic, pseudoscholarly attempt to combine myth, adventurous etymologies (often relating to Greek names) and aspects of local history... What Philo wrote is first of all a reflection of the long-standing prominence of Phoenician deities in Greek culture. But it does reveal genuine connections between Phoenician cults of a much earlier period and those of the Roman Empire; and it does suggest that at least the meanings of individual words in Phoenician were still understood. (Millar, 1993: 278). Thus Millar ahnost, but not quite, ascribes the PH to Philo alone. Few modem scholars have taken the extreme position, given voice in the nineteenth century somewhat hesitantly by F.C. Movers (Barr 1974/75: 19) and more forcefully by Otto Gruppe (Attridge and Oden 1981:1 n.1), that Philo himself concocted the PH and 'invented' Sanchuniathon as his primary source. There is still no compelling reason to reject the notion that the PH is what Philo (or his excerptors) said it was: a work of great antiquity preserved until his own day in Phoenician and then translated for wider consumption. Its purpose was to correct (or at least, to minimize) misinterpretations/misrepresentations in Greek or Roman literary sources regarding the origin of Phoenician culture. Conclusions

Philo may have held the Phoenician cosmogony and epic legend, which he learned and undertook to transmit, to be the source and not the derivative of Hesiod's Theogony composed c. 700 B.C. Since we do not have the entire text of the PH, we cannot be quite certain of this. Neither should we accept, for the same reason, Oden's (or Millar's) conclusion that Philo's PH is 'typically Hellenistic'. The danger of analyzing the fragments of the PH should be transparent by now, and it is quite possible that Philo consciously imitated the style of earlier, Hellenistic authors, in rendering his own translation.

199

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

One might compare the preface of Josephus' Antiquities (Proem 1-4) with the 'statements of purpose' attributed to Philo (Josephus, of course, had both the Hebrew and Greek versions of the Old Testament at hand). It seems fair to suggest that Philo of Byblos worked with what he had: sources that he believed were ancient and authentic, but which could not be read by most of his own countrymen. The hellenization of Phoenicia had produced, at least in the cities, an educated class for which a reading knowledge of literary Phoenician was no longer either a necessity or even a possibility.

Recent studies exemplify this problem by allowing the fragments to take on a life of their own, as if they had not been excised from a larger work for either instructional or polemical illustration. There is a background sound, distant though distinct, of which we should be aware during this review of the evidence: it is the sound of two axes grinding. One axe is held by Porphyry, and the other by Eusebius. Once we understand that sound, we can proceed cautiously. Whether Philo's sources or his own plan for the work included some account of Phoenician history beyond its beginnings in some misty, mythical antiquity is unknown. Porphyry is quoted by Eusebius (PE 1.9.21) to the effect that Sanchuniathon 'collected and wrote up in the Phoenician language (Phoinike dialektos) the complete ancient history from records in each city and from texts in the temples ...' 'Complete ancient history' leaves open the question of how the term palaia is to be understood. We need to look briefly at how Philo, himself an educated and hellenized Phoenician, might have seen himself, and the culture from which he came, at or about the time of Hadrian. Philo's interests seem far more antiquarian than historical in the modem sense. He must have been aware of the stereotypes which had evolved about the Phoenician/Punic peoples. Some of those went back to Homer and Herodotus. But Philo also knew about Kadmos, and Dido, who were essential to the myths of the Greeks and Romans. He also must have known that Herodotus retailed other tales of the Phoenicians that were far closer to reality: the master mariners who commanded an important contingent of the fleet outfitted by the Great King of Persia, and the engineers who cut a canal across the promontory of Mt. Athos. Herodotus had also learned of Phoenicians who circumnavigated Africa for Pharaoh Necho II of Egypt c. 600 BC.

Added to the hellenization process was a new religious factor--by Hadrian's reign Christianity had been taking root in Phoenicia for a century. We hear little about it beyond the references in the Book of Acts (e.g. Acts 11:19; 21:3-6; 27:3) to missionary visits to Phoenicia or to Christian communities in Tyre and Sidon. However, the very fact that the gospels and associated Christian literature were composed and circulated in Greek meant that 'christianization' in Phoenicia could be fostered only through the knowledge of that language. The Gospel of Mark, with its many latinisms and (in some passages) a latinate sentence structure, indicates that it may have come into being among a Christian community in a heavily Romanized city-such as the Augustan colonia at Berytus (on this see MacAdam, 2001: Appendix). However that might be argued, it was the Greek language which made the primary impact on Phoenicia. As Philip Hitti (1962: 190) summarized almost 40 years ago: 'The change under Rome embraced language. In the second century Phoenician legends on coins vanish; Greek persists on coins and in inscriptions. The native tongue, along with the native institutions becomes extinct ... Roman rule tended to denationalize provincials, to obliterate the characteristicfeatures of their national life. '

Had Philo read more of Greek literature, he might have been charmed (as some ofus are today) by Socrates' discourse to his friends in Plato's Phaedo. There the stereotypes are stripped away by Socrates, who describes (for his friends, who will shortly grieve his death), his conception of the world of their own time: 'We live round the sea (from the Pillar of Hercules [the Straits of Gibraltar] to the River Phasis [now in the Georgian Republic]', says Socrates, 'like ants and frogs around a pond'. That put his world, his polis, and (I would like to think) his culture, into its proper perspective.

Millar (1993: 293) also draws attention to the same aspect of the hellenization process in Phoenicia: 'The combined evidence of the coinsfrom Phoenicia and of Philo of Byblos provides us with no more than individual words or phrases in Phoenician; not a single connected sentence is attested as being generated in Phoenician after the one Phoenician inscription of the Augustan period from Aradus, and the one possibly Imperial one from Sidon. ' But Millar is also aware that absence of evidence isn't necessarily evidence of absence; Phoenicia (like most of the coast communities around the Mediterranean) hasn't yielded perishable documentation. Had Philo not been excerpted, we would have nothing from his pen. Whatever the case, Philo responded to renewed imperial interest in Phoenicia (in Hadrian's reign) by delving into the distant past of his culture. In doing so he identified those specific features which he understood as being characteristically Phoenician. That would serve two purposes: first to impress non-native readers with Phoenicia's remarkable antiquity and noteworthy achievements, and second, to remind Phoenicians of their own venerable heritage. Philo's PH, if I am not mistaken in its intention, was also to put the case clearly for the origin

We don't know if Philo tried to do that for his own people and his culture. Certainly he would have heard, as a boy growing up among the city-states of the Phoenician littoral, tales of the sort that Herodotus had heard from temple priests at Tyre. Those stories, of a distant past that gave Phoenicians a heritage as ancient as any Mediterranean people, were what he found, and translated, in 'the complete ancient history' of Sanchuniathon--from the age of Moses. It meant at the very least that he wished Phoenician culture to be thought as old or older than that of Greeks, the Romans, and the Jews of Palestine.

200

Henry MacAdam: Philo of Byblos and the Phoenician History

FGrH= Fragmente der Griechischen Historiker (F. Jacoby) IGLS= Inscriptions Grecques et Latines de la Syrie ILS= Inscriptiones Latinae Selectae OCD= Oxford Classical Dictionary (3rd ed., 1996, unless otherwise specified) PE= Praeparatio Evangelica PNPPI= Personal Names in Phoenician and Punic Inscriptions RE= (Pauly-Wissova's) Real-Encylopadie (der classischen Altertumswissenschafl RHT= Revue d'Histoire des Textes Suda= Suidae Lexicon (ed. A Adler; Leipzig, 1828-38) TAM= TituliAsiae Minoris

and development of native religious tradition. In that sense he was not only following Josephus' lead, he was also indebted to the Hellenistic tradition of the Babylonian scholar Berosus, and the Egyptian priest and antiquarian Manetho of the early third century B.C. (OCD3 s.v.; on Manetho as author of 'native history' [patria historia] see Dillery 1999). This seems clear enough from the fragments of the PH which serve as a 'Preface' to it; these are conveniently collected (with Greek and English on facing pages) in Attridge and Oden (1981: 29-35). This may have been as much to 'clarify' the false assumptions of others with regard to Phoenician religious belief as it was to offer a counter-balance to the encroachment of Christianity into Phoenician culture during the Antonine period.

BIBLIOGRAPHY Philo's curiosity about Judaism may have embraced Christianity; in his youth, and perhaps even later in life, he might understood the latter to be just one more Jewish sect. The relevant fragments of Philo's treatise On the Jews have been conveniently collected and discussed in Stem (II, 1980: 138-145). Note that in the very first fragment (Eusebius, PE 1.9.20-21 = #323, p. 139) treated by Stem, Philo is there distinguished as 'Biblios, not Hebraios', which is simply Eusebius' manner of distinguishing him from Philo Judaeus. There is no clear evidence in the fragments that Philo exhibited an anti-Jewish attitude. Certainly it is tempting to 'read into' them that bias (as did Troiani 1974: 23-39 passim) in the knowledge of Hadrian's specific distaste for certain Jewish practices. That must be what prompted a leading Roman historian to write: 'It is likely enough that Hadrian met the sage of Byblos [and if not] would have certainly inspected [Philo's] literary productions--the Phoenician History displays a certain hostility to Jews' (Birley 1997: 228). If Philo was concerned with what had happened to the Jews of the neighboring province to the south (the very name Iudaea was now officially replaced with Syria Palaestina under Hadrian's direct orders), he may have dedicated himself to ensuring that Phoenician culture would survive. Hence he went back to basics and resurrected the oldest surviving accounts of which he knew--the works of 'Sanchuniathon of Berytus'. That material, for Philo, may have represented what the Bible did for Jews: a Book around which a national identity had evolved. In what survives of his work, certainly in the Phoenician History and Concerning the Jews, there may not be great sympathy expressed for the Jews, but there is clearly no animosity.

Albright, W.F., 1941. Review of Clemen (1939). Journal of Biblical Literature 60: 208-212. Albright, W.F ., 1946. Archaeology and the Religion of Israel. Baltimore: The Johns Hopkins University Press. Albright, W.F., 1961. The Role of the Canaanites in the History of Civilization. In Wright (ed.) 1961: 328-362. Albright, W.F., 1975. Syria, The Philistines and Phoenicia. Cambridge Ancient History. 3rd ed. 22: 507-536. Attridge, H.W. and Oden, RA, 1981. Philo of Byblos: The Phoenician History: Introduction, Critical Text, Translation, Notes. Washington, DC: Catholic Biblical Association of America. Barr, J., 1974/75. Philo of Byblos and his Phoenician History. Bulletin of the John Rylands University Library 57: 17-68. Baumgarten, A.I., 1981. The Phoenician History of Philo of Byblos: A Commentary, Leiden, E.J. Brill. Birley, AR., 1997. Hadrian: The Restless Emperor. London and New York, Routledge. Bowersock, G. W., 1985. Hadrian and Metropolis. Bonner Historia-Augusta-Colloquium, 1983/1984. Bonn, Rudolf Habelt Gmbh.: 75-88. Breton, J.-F., 1980. Les Inscriptions Forestieres d'Hadrien dans le Mont Liban (=IGLS 8.3). Paris, Paul Geuthner. Brizzi, G., 1980. 11 ''Nazionalismo Fenicio" de Filone de Byblos e la Politica Ecumenica di Adriano. Oriens Antiquus 19: 117-131. Brown, J.P., 1965. Kothar, Kinyras, and Kythereia. Journal of Semitic Studies 10: 197-219. Brown, J. P., 1969. The Lebanon and Phoenicia: Ancient Texts Illustrating their Physical Geography and Native Industries. Vol. 1: The Physical Setting and the Forest Beirut: American University of Beirut. Byrne, S. and Cueva. E. (eds.), 1999. Veritatis Amicitiaeque Causa: Essays in Honor of Anna Lydia Motto and John R. Clark. Chicago: Bolchazy-Carducci. Champlin, E., 1981. Serenus Sammonicus. Harvard Studies in Classical Philology 85: 189-212. Chapot, V., 1907. La Frontiere de l'Euphrate, de Pompee a la Conquete Arabe. Paris: Libraire des Ecoles Frarn;aises. Chehab, M., 1962. Tyre a l'Epoque Romaine: Aspects de la Cite a la Lumiere des Textes et des Fouilles. Melanges de l'Universite Saint-Joseph 38: 13-40.

ACKNOWLEDGEMENT Thanks are due to Brent Shaw for reading part of this paper in an earlier draft, and to the library staffs at Princeton Theological Seminary and Washington Township, NJ for their helpful assistance. ABBREVIATIONS AHR= American Historical Review ASNSP= Annali de/la Scuola Normale Superiore di Pisa de Abst. = de Abstinentia BW ANT= Beitrage zur Wissenschafl vom A/ten und Neuen Testament

201

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Hitti, P. K., 1962. Lebanon in History. 2nd edition. New York: Macmillan. Honore, A. M., 1962. Gaius. Oxford: The Clarendon Press. Kindler, A., 1983. The Coinage of Bostra. London: Aris & Philips. Lipinski, E., 1983. The Phoenician History of Philo of Byblos. Bibliotheca Orienta/is 40: cols. 305-310. (Critical review of Baumgarten, 1981) MacAdam, H. I., 1993. Phoenicians at Home, Phoenicians Abroad. Topoi 3: 321-344. MacAdam, H. I., 2000. Marinus of Tyre and Scientific Cartography: The Mediterranean, the Orient and Africa in Early Maps. Graeco-Arabica 7/8: 339-346. MacAdam, H. I., 2001. Studia et Circenses: Leontius of Beirut and the Roman Law School, ARAM 13: forthcoming. MacAdam, H. I., 2002. Literary, Historical and Archaeological Evidence Concerning the Origins of the Phoenicians in the Persian/ Arabian Gulf GraecoArabica 9/10: forthcoming. Markoe, G. E., 2000. The Phoenicians. Berkeley & Los Angeles: University of California Press. Mendenhall, G. E., 1985. The Syllabic Inscriptions from Byblos Beirut: American University of Beirut). Meyer, E., 1931. Geschichte des Alterthums5 9 Vols. Stuttgart & Berlin: J.G. Cotta'sche Buchhandlung Nachfolger. Millar, F., 1983. The Phoenician Cities: A Case Study of Hellenisation. Proceedings of the Cambridge Philosophical Society 209: 55-71. Millar, F., 1990. The Roman Coloniae of the Near East: a Study of Cultural Relations (in Solin & Kajava, 1990: 758). Millar, F., 1993. The Roman Near East: 31 BC-AD 337. Cambridge, MA: Harvard University Press. Momigliano, A., 1980. Interpretazioni Minime: il libro sugli Ebrei di Filone de Byblos. ASNSP 10.2: 1227-1231. Momigliano, A., 1975. Alien Wisdom: The Limits of Hellenization. Cambridge: Cambridge University Press. Moscati, S., 1975. Studi Fenici 1-3. Revista di Studi Fenici 3: 7-13. Moscati, S., 1973. The World of the Phoenicians. London: Sphere Books. Moscati, S., (ed.) 1988. I Fenici. Milano: Bompiani. Nautin, P., 1949. La Valeur Documentaire de l' "Histoire Phenicienne." Revue Biblique 56: 573-578. Oden, R.A., 1978. Philo of Byblos and Hellenistic Historiography. Palestine Exploration Quarterly 110: 115-126. Palmieri, V., 1981. "Eranius" Philo, de Differentia Significationis: La tradizione manoscritta di "Eranio" Filone, RHT 11: 4 7-80. Pritchard, J. B., 1978. Recovering Sarepta, A Phoenician

Clemen, C., 1939. Die phonizische Religion nach Philo van Byblos. Mitteilungen der vorderasiatisch-aegyptischen Gesellschaft. Band 42 Heft 3. Leipzig, J.C. Hinrichs. Cross, F. M., 1968. The Phoenician Inscription from Brazil: A Nineteenth Century Forgery. Orientalia 37: 437460. Dabrowa, E., 1998. The Governors of Roman Syria from Augustus to Septimius Severus. Bonn, Rudolf Habelt GMBH. Dillery, J., 1999. The First Egyptian Narrative History: Manetho and Greek Historiography. Zeitschrift filr Papyrologie und Epigraphik 127: 93-116. Drower, M., 1975. Canaanite Religion and Literature. Cambridge Ancient History. 3rd ed. Vol. 2:2: 148-160. Ebach, J., 1979. Weltentstehung und Kulturentwicklung bei

Philo van Byblos: Ein Beitrag zur Uberlieferung der biblischen Urgeschichte im Rahmen de altorientalischen und antiken Schopfungsglaubens. BW ANT # 8. Stuttgart-Berlin-Koln-Mainz, W. Kohlhammer. Edwards, M.J., 1991. Philo or Sanchuniathon? A Phoenicean [sic] Cosmogony. Classical Quarterly 41: 213-220. Eissfeldt, 0., 1952. Sanchunjaton von Berut und Ilumilku von Ugarit. Beitrage zur Religionsgeschichte des Altertums. Heft 5. Halle [Salle]: P.M. Niemeyer. Eissfeldt, 0., 1956. Art und Aufbau der phonizischen Geschichte des Philo von Byblos. Syria 33: 88-98. Farrer, J.A., 1907. Literary Forgeries. London: Longmans, Green&Co. Gordon, C. H., 1968. The Canaanite Text from Brazil. Orientalia 37: 425-436. Gordon, C. H.,1968. Reply to Professor Cross. Orientalia 37: 461-463 Gordon, C.H., 1972. Riddles of the Wise, Berytus 21: 17-38. Gordon, C. H., 1973. America Before Columbus: Links Between the Old World and Ancient America. New York: Crown Publishers. Gordon, W.M. and. Robinson. O.F.(eds. & trans.), 1988. The Institutes ofGaius. London; Duckworth. Grainger, J.D., 1991. Hellenistic Phoenicia. Oxford: Clarendon Press. Gudemann, A., 1913. Herennius Philon von Byblos. RE 8: cols. 650-661. Guzzo, M.G.A., 1988. I Fenici in America? In Moscati 1988: 570-572. Haas, V., 1985. Die junge Wissenschaft Assyriologie in der Schonen Literatur des 19. und 20. Jahrhunderts. In Schuller 1985: 71-104. Halfinann, H., 1986. Itinera Principum. Stuttgart, Franz Steiner. Hall, L. J. 1999. Latinitas in the Late Antique Greek East: Cultural Assimilation and Ethnic Distinctions. In Byrne and Cueva (eds.), 1999: 85-111). Hall, L. J., 2001. Classical Beirut Through the Texts: From Colonia to Civitas. ARAM 13: forthcoming. Harden, D., 1963. The Phoenicians. 2nd ed. London: Thames and Hudson. Herner, C. J., 1978/79. The Manchester Rotas-Sator Square. Faith & Thought 105: 36-40.

City: Excavations at Sarafand, Lebanon, 1969-1974, by the University Museum of the University of Pennsylvania. Princeton: Princeton University Press. Ragette, F., 1980. Baalbek. London: Chatto and Windus. Rampoldi, G.B., 1824. Annali Musulmani Vol. 7. Milano: Felice Rusconi.

202

Henry MacAdam: Philo of Byblos and the Phoenician History

Rawlinson, G., 1889. History of Phoenicia. London: Unwin. Renan, E., 1858. Memoir du l'origine et le caractere veritable de l'histoire phenicienne qui porte le nom de Sanchoniathon. Memoirs de l'Institut Imperial de France: Academie des Inscriptions et Belles-Lettres 23: 241-334. Rey-Coquais, J-P. 1978. Syrie romaine, de Pompee a Diocletien. Syria 68: 44-73. Rollig, W. 1983. On the Origin of the Phoenicians. Berytus 31: 79-93. Salles, J.-F. 1993. Les Pheniciens de la Mer Erythree. Arabian Archaeology and Epigraphy 4: 170-209. Schiffinan, I. 1986. Phonizisch-Punische Mythologie und geschichtliche Uberlieferung in der Widerspiegelung der antiken Geschichtsschreibung. Roma: Consiglio Nazionale delle Richerche. Schuller, W. (ed.), 1985. Antike in der Moderne. Konstanz: Universitatsverlag. Schurer, E., 1973. A History of the Jewish People in the Age of Jesus Christ Vol. 1 Revised edition by Vennes, G. and Millar, F., Edinburgh: T. & T. Clark. Solin, H. and Kajava, M. (eds.), 1990. Roman Eastern Policy and Other Studies in Roman History. Helsinki: Societas Scientiarum Fennica Commentationes Humanarum Litterum. Stem, M., 1980. Greek and Latin Authors on Jews and Judaism II. Jerusalem:Israel Academy of Sciences and Humanities. Syme, R., 1988. Journeys of Hadrian. Zeitschrift far Papyrologie und Epigraphik 73: 159-170. Syme, R., 1985. Hadrian as Philhellene: Neglected Aspects. Bonner Historia-Augusta-Colloquium 1982/1983. Bonn: RudolfHabelt Gmbh.: 341-362. Syme, R., 1983. Hadrian and Antioch. Bonner HistoriaAugusta Colloquium 1979/1981. Bonn: Rudolf Habelt Gmbh.: 321-331. Sznycer, M., 1966. Philon de Byblos. Supplement au Dictionnaire de la Bible 7: cols. 1351-1354. Sznycer, M., 1968/9. Punic Literature. Archaeologia Viva 1: 141-148. Torrent, A., 1983. La ordinatio edicti en la politica juridica de Adriano. Annuario de Historia de! Derecho Espanol 53: 17-44. Troiani, L., 1974. L'opera Storiografica di Fi/one da Byblos. Pisa: Golliardica. Van Den Branden, A., 1968. L'Inscription Phenicienne de Paraiba (Bresil), Me/to: Recherches Orienta/es 4: 55-75. Wright, G. E., 1961. (ed.). The Bible and the Ancient Near East: Essays in Honor of William Foxwell Albright. New York: Doubleday & Co.

203

Archaeology of the Roman Empire: a tribute to the life and works of Professor Barri Jones

This material originated on the Interactive Ancient Mediterranean Web slte (http://iam.classics.unc.edu) It has been copied, reused or redistributed under the terms of JAM's fair use policy. Copyright 1998, Interactive Ancient Mediterranean.

18.1 Byblos, Phoenicia and the eastern Mediterranean

204

19 Floodwater Farming, Roman Imperialism, and Desertification in Libya and Jordan Graeme Barker geographers or historians for or against the long-term role of people (and commonly, their animals) in desertification, and the confident assertions about causation in popular treatments of desert archaeology in the media, there have been remarkably few studies of desertification history in which long-term processes of natural and human agency have been investigated with real rigour. One problem has been that historians and archaeologists working with good information about settlement and/or land use have generally lacked reliable environmental data, whilst geographers with good information about past environments, including sometimes indicators of the impact of people on those environments, have lacked reliable information on the nature and scale of human activity. It has rarely if ever been possible to compare the implications of similarly high quality data sets collected by the separate disciplines from the same region. Modem studies indicate that many dryland environments can in fact be remarkably resilient, recovering surprisingly quickly from over-intensive exploitation (Beaumont 1993). Indeed, there are examples of degraded landscapes that have been restored in recent decades by comparatively simple measures such as building terrace walls against soil erosion (Mortimore 1998; Tiffen et al. 1994). We need, therefore, to learn from the past about in what circumstances the actions of past societies really impacted on a desertic landscape in a devastating way to the 'point ofno return', and when drylands were able to recover from overexploitation. What solutions did people develop for living in arid lands, and how successful were they in the long as well as the short term? Why did they take the choices they took? In what circumstances were people able to cope with climatic change or humanly-induced land degradation, and in what circumstances were they simply overwhelmed by such deterioration?

Archaeology and Desertification

Many dryland regions of the world that are now sparsely occupied by mobile or semi-mobile pastoralists have archaeological remains suggesting that these same arid and degraded landscapes must once have supported settlement of a very different order of intensity. This is nowhere more so than in the semi-arid littorals and interior deserts of North Africa and the Levant. From the early nineteenth century onwards, European travellers and explorers have left accounts describing spectacular ancient ruins, most of which could be ascribed to the centuries of the Roman empire (Illus. 19.1). Such writers often speculated on what might have happened to create the barren landscapes around them from, it had to be assumed, the green fields of antiquity. The debate has continued ever since in the scholarly literature of archaeology, ancient history, and historical geography. Was the 'greening of the desert' essentially a matter of new know-how, the expansion of the Roman frontier into these regions taking with it soldiers and farmers with sophisticated technologies and cropping systems? Or perhaps a key factor had been climate change, more abundant rainfall allowing the desert to be sown by Roman farmers, a return to aridity then causing the land to be abandoned. Had people sowed the seeds of their own destruction through their greed and stupidity, stripping the landscape for building timber and fuelwood and allowing their animals to overgraze the vegetation (Hughes with Thirgood 1982)? Perhaps factors such as moral or economic decline caused societies to abandon maintaining their land in good order. Perhaps unsuitable or over-intensive systems of irrigation had ruined soil fertility through salinization. Such speculations about the greening of the desert in Roman Africa and Arabia can be seen as part of a wider scholarly debate on desertification and its causes (Beaumont 1993; Fantechi and Margaris 1986; Millington and Pye 1994; Mortimore 1998; Thomas and Middleton 1994). The term was coined by Aubreville in 1949 in a study of African vegetation, but its meaning has developed through time. Some authors have used it to mean land degradation generally in arid and semi-arid lands, whether caused by climate or people. More recently, most scholars have tended to restrict the term to the human side of the equation (Millington and Pye 1994; Thomas and Middleton 1994: 9-10), though natural and cultural agencies clearly work in tandem to create dryland deterioration, as may be the case with global warming today. Understanding the past what are now deserts is (Barker and Gilbertson despite the frequency

Modem landscape archaeology uses inter-disciplinary methodologies linking archaeology with the social and environmental sciences to investigate settlement and landscape change, and potential interactions between the two, at the regional scale over long timescales. It therefore has the potential to contribute strongly to desertification debates, because it can provide high quality information about the kind of questions mentioned above, as well as a deep-time perspective on them. This paper therefore compares the results of two regional case studies in landscape archaeology with which I have been closely involved, one in Libya (the Tripolitanian Predesert) and one in Jordan (the Wadi Faynan). Both regions are now inhabited by sparse populations of mobile or semi-mobile pastoralists practising small-scale farming, yet they have abundant evidence of past, especially Roman-period, settlement associated with field systems implying very different forms of settlement and land use, and past levels of population. Both projects involved inter-disciplinary

history of human settlement in obviously critical to such debates 2000; Spooner 1989). However, of comment by archaeologists, 205

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

teams of archaeologists, environmental scientists and historians (many of the personnel contributing to both projects, in fact) working together in the investigation of long-term settlement and landscape change. Both have been characterized by detailed studies of climatic regimes, palaeoenvironments, land use systems, and their interrelationships. Many of the methodologies they have used have been the same. Given all these similarities, what are the lessons from their findings for the desertification debate?

camped in the Predesert in the summer and wintered to the south much further into the Sahara. These people were highly mobile, relying on their flocks and herds and on cultivating small patches of cereals on the wadi floors. They broadcast their seed-com on the floors of wadis after autumn rainstorms. Further rain too soon might wash away much of the seed before it was established, and not enough rain would mean the crop would not ripen, so it was a very precarious enterprise sustainable only because people were few and mobile, relying on their animals and harvesting whatever patches of cereals they could find that had ripened properly as they moved north in the spring. For drinking water they had access to a few deep wells and they stored rainwater in cisterns. The ephemeral archaeology they created, of tents and simple enclosures for animals (usually a few stones in a circle topped by thorn bushes), contrasted strikingly with the hundreds of imposing stone gsur and their (apparently) associated field systems that we had come to study.

The UNESCO Libyan Valleys Survey In the late 1970s the Libyan Department of Antiquities funded UNESCO to initiate archaeological fieldwork with a series of international teams in the 'Predesert', the region between the settled coastal zone and the Sahara, where it was known that there were hundreds of stone 'castles' (gsur in Arabic) or fortified farms of the Roman period. Dry-stone walls laid out along and across the beds of the wadis near these sites seemed to be evidence of kinds of irrigation systems. The Libyan government was interested to know what these remains implied in terms of ancient climate, environment, and land use, to inform their plans for agricultural development in the same region. Three major projects were planned by UNESCO within what was termed their 'Libyan Valleys Survey', to be led by British archaeologists in Tripolitania in the west, Italian archaeologists in Cyrenaica in the east, and French archaeologists in Sirte in between. For the British project, UNESCO approached the British Academy, the principal British institution fostering British archaeological research overseas, and the Academy consulted one of its senior Fellows, John Ward-Perkins, who had worked extensively in Libya (especially Tripolitania) in the years following the Second World War. Ward-Perkins proposed that the British project should be directed jointly by Barri Jones and myself. He knew both of us well because we had both been research scholars at the British School at Rome when he was Director, and Barri was a Romanist with experience and active involvement in Libyan archaeology, whereas I was a prehistorian with a particular research interest in past agricultural systems and field experience in intensive ground survey. In the event, the Anglo-Libyan project was the only one of the three envisaged at the outset that was able to run its full course.

The first season of fieldwork in 1979 was the beginning of a close working relationship, and friendship, between Barri and myself that lasted until his untimely death in 1999. It was a friendship which survived the many stresses and strains of two totally different personalities being pitched together for weeks on end, often in the same tent (I remember one spectacular row at Ghirza in the first season of fieldwork, but that is another story!). Five seasons of fieldwork were planned, the progress of which was much complicated by the difficult relations between Britain and Libya during the 1980s, especially after the breaking of diplomatic relations in 1984. The fieldwork was undertaken in 1979, 1980, 1981, 1984, and 1989. After the first season of fieldwork David Gilbertson joined the project to coordinate work on geomorphology and palaeoecology to underpin the studies of settlement and land use. David Mattingly, an undergraduate student of Barri's at the time of the inception of the project in 1979 (and a member of the field team then), then wrote his PhD with Barri on Roman olive oil farming in North Africa, which emerged as an important aspect of the agricultural systems we were studying; and as well as overseeing the bulk of the work on the survey archive, he went on to become the fourth member of the management team overseeing the project's development and publication. A wide range of specialisms was involved as the project developed, including archaeobotany and archaeozoology, archaeological survey, architectural studies, epigraphy, geomorphology, pollen analysis, remote sensing, and GIS. Some 30 papers on the project's fieldwork and subsequent laboratory studies were published from 1979 onwards in Libyan Studies, the final report being published in 1996 as two volumes, one of synthesis and one of specialist accompanying data (Barker et al., 1996a).

Our assigned study area focussed on the Sofeggin and ZemZem wadi systems, between the modem town of Beni Ulid to the north and the settlement of Gheriat el-Garbia (a Roman oasis fort) to the south (Illus. 19.2). The coastal zone today has more than 200 mm of rainfall, the lower limit for Mediterranean-style tree and crop farming, and the northern edge of our study area around Beni Ulid is well-watered enough for a version of this tree and crop farming to be practised. Further south, the annual rainfall declines rapidly to 50 mm or less. Before the oil revolution transformed rural settlement in Libya, the study area was occupied mainly by transhumant pastoralists who

Goodchild (1950, 1952-53) and Goodchild and WardPerkins (1949) had originally argued that the Tripolitanian gsur were a system of defence in depth occupied by Roman soldier-farmers (limitanei) to protect the coastal

206

Graeme Barker: Floodwater Farming, Roman Imperialism, and Desertification in Libya and Jordan

cities and villas from inland desert tribes. Certainly they were fortified farms, but we were able to demonstrate from inscriptions at several gsur and associated funerary sites that their occupants were indigenous Libyans, 'RomanoLibyans' as we termed them (Mattingly 1983, 1987). The pottery we collected from the gsur indicated that their construction began generally in the third century AD, much as Goodchild had thought, but we also established that they had been preceded by unfortified or open courtyard farms, often built in a recognisably Roman style (opus Africanus), this first phase of Romano-Libyan settlement beginning in the late first century AD. Before then, settlement seems to have been small-scale and mobile, much as in recent centuries. The heyday of the gsur appeared to be the third and fourth centuries AD, but many gsur continued to the sixth and seventh centuries AD, whilst in the northern wadis around Beni Ulid their occupation has continued to recent times.

the principal animals kept were sheep and goats, their dung indicating that they were stalled part of the year on the farms. Their water and pasture needs would have been competing with those of the people and crops, and the abundance of wild animals such as gazelle in the farm middens suggests that they were hunted as much for their meat as to protect the vulnerable crops on the wadi floors. Integrating all these various strands of the project, we concluded that indigenous Libyan elites in the Predesert moved swiftly from subsistence to cash-crop farming in response to the opportunities of Romanization in the first century AD (Barker et al. 1996b; Mattingly 1996). The shift from unfortified to fortified farms in the third century AD was primarily an indicator of internal processes of competition, not a defence system against desert raiders. The decline of the system was also a gradual process. Gsur abandonment started earliest in the southernmost, most arid, parts of the study area, suggesting that people returned to traditional lifestyles soonest in these most marginal environments. Gsur farming continued apace further north, but there was no simple end to it coinciding with the decline of Roman markets - in certain wadis local warlords appear to have retained control over their peasant farmers and maintained their floodwater-farming systems until the Arab conquest and beyond. The environmental studies found evidence for erosion through these centuries that might reflect the pressures of overgrazing and fuelwood collection, but there was nothing to suggest environmental degradation on such a scale as to have been a factor in either the shift from open to fortified farms, or the gradual decline of the system (Gilbertson 1996).

Our geomorphological and palaeoecological studies were able to establish that, whilst significantly wetter climates than today developed in the first half of the Holocene, the climate over the past 3000 years has been essentially the same as today. The only discontinuity derives from indications of a wetter episode in the medieval period, perhaps equating with the Little Ice Age in temperate Europe (Gilbertson and Hunt with Smithson 1996). There was no evidence, therefore, that Romano-Libyan settlement had developed in the context of a more favorable environmental regime than today. Our investigations of the field systems associated with the farms indicated that Romano-Libyan farmers practised sophisticated floodwater farming, building walls to divert surface runoff after seasonal storms, and channeling it down from the plateau edges into the fields that they laid out on the wadi floors (Illus. 19.3, 19.4). By channeling the surface runoff from a large area into a small area of protected fields, they were able to grow crops that needed much more water than was available in terms of annual rainfall levels.

The Wadi Faynan Landscape Survey

The Wadi Faynan is situated about 40 kilometres from Petra, the capital of the Nabatean kingdom before its incorporation within the Roman province of Arabia in AD 106. The catchment of the wadi forms a transect about fifteen kilometres long (west/east) and about five kilometres wide (north/south), running from the rim of the Jordanian plateau c.1100 metres above sea level to the Wadi Araba rift valley at about sea level (Illus.19.5). The plateau receives more than 200 mm of rainfall a year, so the villages here practise Mediterranean-style farming, whereas the Wadi Faynan a few hours' walk down the escarpment is rainless for most of the year, sparsely vegetated, and used largely by Bedouin pastoralists. Some of the latter live in the Wadi Araba all year, whereas the more prosperous transhume between the Wadi Araba and the plateau.

Carbonized and desiccated plant remains demonstrated that these crops included three cereals (six-row hulled barley, durum or hard wheat, and bread wheat), three pulse crops (lentil, pea, and grass pea), four oil plants (olive, safflower, linseed or flax, and castor), five Mediterranean fruits (grape, fig, pomegranate, almond, peach), five African fruits (date palm, water melon, wild pistachio, the Christthom, and sumach), and three herbs (purslane, dill, and celery) (van der Veen et al. 1996). Many farms have stone pressing structures for making oil and wine, and our calculations of vat capacities indicated that most were designed for producing a surplus well beyond the needs of the inhabitants. These indicators of surplus production chimed with ostraka evidence from forts such as Bu Ngem and images on tombs at Ghirza indicating that the Predesert Romano-Libyan farmers were producing agricultural commodities such as olive oil to supply to the military (and perhaps to the coastal populations as well). Pastoralism, by contrast, appears to have been small scale:

Like the Libyan Predesert, the Wadi Faynan has abundant archaeological evidence indicative of patterns of settlement and land use in the past very different from those of the modem pastoralists. There is one principal archaeological monument, long known to early travellers (Glueck 1935; Musil 1907), called the Khirbet Faynan, the 'Ruin of Faynan' (Illus. 19.6). This is a major settlement of Nabatean, Roman and late Roman (Byzantine) date, 207

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

located about a kilometre west of the escarpment edge at the point where three tributary wadis (the Dana, Ghuwayr, and Shayqar) come together to form the main channel. Nearby is an aqueduct, reservoir and water mill of Roman/Byzantine date. The hillslopes around the Khirbet Faynan complex are black with slag, the residues of ancient smelting, and the surrounding hills are rich in mineral deposits and honeycombed with ancient mine shafts. Investigations by a team from Bochum Mining Museum have revealed a rich prehistory and history of copper mining and smelting, from Chalcolithic to Byzantine times (Hauptmann 1989, 1992, 2000; Hauptmann et al. 1992). Khirbet Faynan is generally assumed to be the site of the settlement of Phaino mentioned in ancient sources as a centre of copper and lead mining industry in Roman times - there is a reference to Christians of Palestine and Egypt being transported there as slave labour in the third and fourth centuries AD.

which the fields had functioned? Had climatic change and/or humanly-induced degradation had any role in their development, use, and eventual abandonment? Equally, what were the social and economic contexts in which they were built, maintained, and abandoned? When and why had the field systems gone out of use and the present-day systems of pastoral-dominated settlement developed? Given the implications of its abundant archaeological remains - settlements, cemeteries, field systems, and mining residues - for a complex history of settlement and land use very different from those of today, the Wadi Faynan seemed an ideal location for investigating the long term 'archaeological history' of interactions between a desertic landscape and its human inhabitants, as a contribution to desertification debates. The Wadi Faynan Landscape Survey was therefore established to investigate the landscape evolution of the wadi from prehistoric times to the present day. Five seasons of fieldwork were projected, and have taken place (1996-2000), the results of four of which have been reported in Levant (Barker et al. 1997, 1998, 1999, 2000), and the team is now preparing a two-volume final publication of the project along the lines of the ULVS reports.

Extending for some five kilometres to the west of Khirbet Faynan is a substantial field system of rubble walls, its surface pottery indicating primary use contemporary with that of the Khirbet Faynan settlement (Illuss. 19.6, 19.7). I first visited the Wadi Faynan in early spring 1995 as a member of a review committee of the British Academy. We were reviewing the facilities and activities of the British Institute at Amman for Archaeology and History, so BIAAH arranged visits to areas of its current fieldwork. The northern side of the wadi lies within the Dana Nature Reserve of Jordan's Royal Society for the Conservation of Nature, and shortly before our visit a small team from BIAAH had undertaken a reconnaissance survey to provide an initial archaeological record of the wadi's archaeology to help the RSCN develop its management plan for the Reserve. This work had located a variety of prehistoric and classical sites both within and outside the field system (Ruben et al. 1997). BIAAH was also coordinating a programme of rescue excavation, in collaboration with Y armouk University, of a large cemetery of Byzantine date (the 'South Cemetery') about a kilometre from Khirbet Faynan that was suffering from extensive grave-robbing. Two Neolithic settlements were being investigated by other teams (al-Najjar et al. 1990; Simmons and al-Najjar 1996).

The first three seasons of fieldwork concentrated on mapping and dating the ancient field systems west of Khirbet Faynan. These extend over some 250 hectares, and are estimated to contain about 1000 individual fields. The surface material within these fields was collected systematically. In many fields the finds on the surface were extraordinarily dense, so a system was devised whereby all material on a metre-wide transect was collected from one in every four transects, the latter ten metres apart, the material in the intervening transects being recorded using clicker counters to give an assessment of density of pottery and lithics to compare with the collected material. We also recorded every wall according to a typology we developed in a pilot exercise in the first season, and identified any structures indicating water control, building on our experiences of the stone sluices and baffles used by floodwater farmers in Roman Tripolitania (Gilbertson and Hunt 1996). In Tripolitania the mapping exercise had been extraordinarily difficult because of the dearth of accurate maps, but we were greatly aided in the survey of the Faynan field systems by access to a photogrammetric map prepared in the UK from vertical air photographs, though our fieldwork added considerable detail to this. It quickly became apparent from this survey that my first impressions that the Faynan field systems were like those of Tripolitania were entirely wrong, in terms of both their history of construction and the ways in which they had operated.

The field systems I drove through that day en route to Khirbet Faynan posed obvious questions to me in the light of the UNESCO Libyan Valleys Survey. Did they represent similar solutions by ancient farmers in Jordan to coping with a similarly arid environment? When had they been built, and how long had they remained in use? Did the almost five kilometres of the main field system (termed WF4 in the BIAAH survey catalogue) represent something that had grown organically and piecemeal over a long period, or was it a planned system that had been laid out and maintained more or less as an integral system, or was it a combination of the two? How had the system or systems functioned? Given our debates about the Libyan field systems, what were the environmental conditions in

Whilst the main phase of use was Roman and Byzantine, roughly contemporary with the main phase of intensive farming in Tripolitania, the field systems constructed at this time in Wadi Faynan represented the culmination of some 4000 years of floodwater farming technologies on the site. Our surveys, and excavations by other teams,

208

Graeme Barker: Floodwater Farming, Roman Imperialism, and Desertification in Libya and Jordan

have demonstrated that farming in fact began in the wadi some 10,000 years ago, as early as the first farming anywhere in the Near East. A Pre-Pottery Neolithic A community living by a mix of hunting, gathering, and plant cultivation, camped by a spring where the Wadi Ghuwayr exits from the mountains (Finlayson and Mithen 1998). A thousand or so years later there was a substantial PrePottery Neolithic B village nearby, a collection of wellbuilt stone houses inhabited by people who practised mixed farming (Simmons and al-Najjar 1996). By the sixth millennium BC there were farming villages out on the main floor of the wadi west of Khirbet Faynan, in the area of the field system (al Najjar et al. 1990). Our geomorphological investigations show that all of these early farmers were operating in a climate quite different from today, much wetter - the Faynan was a perennial stream at that time, flowing through a well--vegetated landscape. The conditions allowed people to grow their crops on seasonally-flooded soils beside the springs and streams. As aridification then began to develop in the fifth millennium BC, Early Bronze Age farmers started to develop floodwater farming strategies for coping with water shortages, building circular storage cisterns on hillslopes with guidance walls to collect rainwater from the surrounding catchments and simple terrace walls across the floors of low-gradient wadis to collect and trap seasonal floodwaters and the sediments being moved by them. Specialized pastoralism may also have developed at this time, the flocks being kept at a distance from the arable fields on the evidence of numerous enclosures in the hills.

channels to carry floodwater through the system demonstrated the same engineering skills in moving water relatively long distances over gentle gradients that were displayed by the Roman engineers who designed the rockcut feeder channel that brought water several kilometres from the Wadi Ghuwayr spring to the aqueduct feeding the reservoir and ore-crushing mill near Khirbet Faynan. Presumably this was a highly organized imperial estate designed to feed not only the agricultural population but also the large numbers of industrial workers (some slaves, certainly, but many others probably paid labourers and craft specialists) operating the Faynan mines and smelting installations, all controlled by the Khirbet Faynan garrison. This sequence of agricultural development took place within a changing landscape. As already mentioned, the initial development of floodwater farming techniques was in the context of aridification, though this does not of course provide a simple deus ex machina explanation. The Early Bronze Age was a period of immense social change in the Levant, characterized by the development of metallurgy, long-distance trade networks, and, in some regions, complex polities with quasi-urban settlements. In Faynan, the Early Bronze Age witnessed the first systematic mining and processing of copper ores, production and trade probably being controlled by local elites. We have used EDMA (Energy Dispersive X-ray Micro-Analysis) to measure the amount of environmental pollution caused by metal-working, and the indications from analysis of pollutants in Early Bronze Age sediments are that this first main phase of metallurgy caused smallscale localized pollution (Gilbertson et al. 1999). Pollen analysis also produced evidence for the impact of human activities on the landscape at this time in terms of vegetation clearance, perhaps for fuelwood for smelting, though it is conceivable that intensification in systems of cultivation and/or pastoralism might have played a part, and there are some indicators for soil erosion in the geomorphological record. However, the evidence suggests that the scale of human impact was relatively small, in that the landscape seems to have recovered in the second millennium BC, a period when Faynan was probably only sparsely settled.

By the early first millennium BC, the Edomite Iron Age, more sophisticated systems of floodwater farming were being practised in Wadi Faynan involving the construction of long sinuous walls upstream of field systems designed to collect floodwaters from extensive areas ofhillslope and divert them into the fields, which were generally in the same low-angle locations as those of Early Bronze Age farmers (Illus. 19.8). Nabatean farmers used the same principles but applied them much more effectively, concentrating their efforts on hillslopes where check-dams linked to short but well-designed diversion walls could trap water as it issued through gaps in the hills and divert it along the contour of the slope. At that point it could be channeled in a controlled manner through simple sluices (gaps) and spillways (step structures) onto a succession of terraced fields below (Illuss. 19.6, 19.9). The GIS analysis indicates a series of discrete agricultural estates along the southern slopes of the wadi dating to the Nabatean period.

By contrast, there is then very strong evidence for tremendous landscape transformations in the Nabatean and more especially the Roman/Byzantine periods, coinciding with the massive expansion of the mining industry. We were able to core a series of sediments that had built up behind a barrage wall constructed at the foot of Khirbet Faynan, a sequence that we believe is a more or less continuous record from Nabatean to recent times. The initial analysis of the pollen (Hunt and Mohammed 1998) indicates two distinct vegetational phases: in the lower levels there was a steppic landscape in which cereal and olive cultivation was practised. However, at a relatively early stage in the sequence this developed into the extremely degraded and desertic landscape that endures today. Correlating with this, studies of charcoal from mining sites show that Edomite and Nabatean miners were

In the Roman period these farms were abandoned and their estates were incorporated within a single huge field system that extended westwards from the Khirbet Faynan settlement over upper and lower slopes for some five kilometres, the field system area mapped in Illus. 19.7. Systems of long parallel walls were built to carry floodwaters from the main channel throughout the field system (Illus. 19.10), the internal arrangements demonstrating that water resources were shared down the length of the field system. The construction of these 209

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

able to collect local fuelwood for their smelting, whereas by the Roman period timber had to be brought down from the plateau because local supplies had been exhausted (Engel 1993; Hauptmann 1992). Stripping the land of vegetation made it vulnerable to erosion, of course, and we have found widespread evidence that Roman farmers were trying to combat the effects of wadi-downcutting in their alterations to the diversion wall systems, though ultimately these were unsuccessful - the main wadi has now cut a channel at least five metres below the Roman fields. The EDMA geochemical studies of the heavy metals in the barrage sediments indicate extraordinarily high levels of air and water pollution in Roman and Byzantine times, levels several times lethal in terms of modem criteria for safe working conditions. It is likely that this pollution severely affected crop fertility, because today biomass and cover values increase in barley plants with distance from the contamination 'hot spots' - perhaps the extraordinary quantities of Roman/Byzantine pottery and other settlement rubbish on the fields are evidence of increasingly desperate attempts to redress falling yields with manuring, though this simply served to spread slag onto arable land at a distance from Khirbet Faynan. There are significant levels of pollution in the modem flora and fauna of the barrage locality, and the inescapable conclusion from the environmental science would seem to be that the Roman and Byzantine miners of Wadi Faynan created a desert for themselves, together with a legacy of pollution here for future generations to come.

developed and adapted extensive field systems laid out by the Nabateans prior to the region's annexation by Rome, which themselves incorporated vestiges of earlier systems whose history reached back to the fifth millennium BC, yet the field system that resulted was an entirely new imperial landscape of control, an integrated agricultural and hydraulic operation geared to feeding the workforce of the mines. In Tripolitania, Roman-Libyan floodwater farming continued in places well beyond the collapse of the Roman markets and the Arab conquest, and there is no evidence that it impacted on the landscape in terms of significant vegetation clearance and erosion and vegetation. In the Wadi Faynan, by contrast, we can see a landscape that had recovered from small scale floodwater farming and industrial activity in the Early Bronze Age, which was then devastated by the processes of imperialism, leaving a legacy of pollutants that still poses a threat to the health of the present-day inhabitants and their livestock. As David Gilbertson and I conclude in a survey of the archaeology of dryland farming systems (Barker and Gilbertson 2000: 8), the building bricks of floodwater farming are commonly the same - walls to trap soil and divert or stem water flow, channels to divert water, and so on. However, the archaeological record demonstrates that dryland farming was never a simple matter of commonsense observation by ancient farmers of what was the 'best fit' to particular environmental or economic circumstances: 'people took choices, and not always the right ones, within a complex mix of factors, including perception of risk, the need or desire for economic advantage, and the institutional and regional context in which they were operating'. As these case studies illustrate, modem desertification theory can learn much from the complexity that is emerging from the landscape archaeologies of drylands: there has been no simple evolutionary development of land use systems from simple to complex or from extensive to intensive, and no simple correlations between particular human activities and particular pathways of desertification. Dryland agricultural and pastoral histories have been 'formed and changed within specific, place-bound, social, historical, and ecological contexts' (Widgren 2000: 262).

Conclusion

Both of the dryland regions discussed in this paper, the Tripolitanian Predesert in Libya and the Wadi Faynan in southern Jordan, were transformed by Roman imperialism, one aspect of which in both was the development of intensive systems of floodwater farming. Both are desertic, degraded, landscapes used in recent times by small-scale and largely mobile populations heavily dependent on pastoralism for their livelihood, but with abundant archaeological evidence for dense and largely sedentary societies in the Roman period. In both regions, detailed geomorphological and palaeoecological research indicates that such settlement did not develop in a climatic regime fundamentally different from that of today. In both regions, the populations of the Roman period were sustained by floodwater farming systems which were in their fundamental principles similar to those still practised by the modem-day inhabitants, though with much greater investment in wall building and maintenance, and with a greater degree of organization and spatial coordination of activity. Yet at this point the stories diverge significantly. In Tripolitania, the sudden and dramatic development of sedentary floodwater farming geared to producing cash crops, in what was a climatically marginal region of the Roman frontier, was probably in the context of local elites seizing the new economic opportunities of Roman military and urban markets. In Wadi Faynan in southern Jordan, the systems of floodwater farming practised in Roman times

ACKNOWLEDGEMENTS It is my pleasure first to acknowledge the support of the many funding bodies whose generous contributions financed the fieldwork and laboratory studies on which this paper is based: for the UNESCO Libyan Valleys Survey, notably the British Academy, the Leverhulme Trust, the Society for Libyan Studies, UNESCO, and the Universities of Leicester, Manchester, and Sheffield; and for the Wadi Faynan Landscape Survey notably the Arts and Humanities Research Board, the British Academy, the British Institute at Amman for Archaeology and History (now part of the Council for British Research in Levant), the Humanities Research Board, the Society of Antiquaries, and the Universities of Aberystwyth and Leicester. This chapter

210

Graeme Barker: Floodwater Farming, Roman Imperialism, and Desertification in Libya and Jordan

draws on the research of the many scholars in the two projects, but it draws in particular on the work of Russ Adams, John Dore, Crispin Flower, David Gilbertson, Annie Grant, John Grattan, Chris Hunt, Barri Jones, David Mattingly, Hwedi Mohammed, Paul Newson, Brian Pyatt, Tim Reynolds, Roberta Tomber, and Marijke van der Veen. Barri Jones introduced David Gilbertson, David Mattingly and myself to desert archaeology in Libya, and infected us all with his passionate commitment to its importance. His wisdom and encouragement supported us in the planning of the Wadi Faynan Landscape Survey. In dedicating The Archaeology of Dry/ands to his memory, David Gilbertson and I wrote that 'he was a frenetic personality, who was both enchanting and exasperating to work with - he was notorious for doing too many things at once, mostly whilst nominally in control of a Landrover! Amongst his many talents, though, he had an extraordinary topographical eye: he was liable to get us lost in some desert waste to visit an archaeological site once noted by a traveller a hundred years ago, and thrash the vehicle in the process, but he was also by far the best person to be with to get safely back to camp again'. I am enormously indebted to the generosity of his support and friendship.

Palmer, C., Pyatt, F.B., Reynolds, T.E.G. and Tomber, R. (1999) Environment and land use in the Wadi Faynan, southern Jordan: the third season of geoarchaeology and landscape archaeology (1998). Levant 31: 255-92. Barker, G., Adams, R., Creighton, O.H., Daly, P., Gilbertson, D.D., Grattan, J.P., Hunt, C.O., Mattingly, D.J., McLaren, S.J., Newson, P., Palmer, C., Pyatt, F.B., Reynolds, T.E.G., Smith, H., Tomber, R. and Truscott, A. J., 2000. Archaeology and desertification in the Wadi Faynan: the fourth (1999) season of the Wadi Faynan Landscape Survey. Levant 32: 27-52. Beaumont, P., 1993. Dry/ands: Environmental Management and Development. London: Routledge. Engel, T., 1993. Charcoal remains from an iron age copper smelting slag heap at Feinan, Wadi Arabah (Jordan). Vegetation History and Archaeobotany 2: 205-211. Fantechi, R., and Margaris, N. S. (eds), 1986. Desertification in Europe. Proceedings of the Information Symposium in the EEC Programme on Climatology Held in Mytilene, Greece, 15-18 April 1984. Dordrecht: D. Reidel Publishing Company. Finlayson, B. and Mithen, S., 1998. The Dana-Faynan (South Jordan) Epipalaeolithic Project: report on reconnaissance survey, 14-22 April 1996. Levant 30: 27-32. Gilbertson, D. D., 1996. Explanations: environment as agency. In G. Barker, D. Gilbertson, B. Jones and D. Mattingly, Volume One: Synthesis: 291-317. Gilbertson, D. and Hunt, C.O., 1996. Romano-Libyan agriculture: walls and floodwater farming. In G. Barker, D. Gilbertson, B. Jones, and D. Mattingly, Volume One: Synthesis: 191-225. Gilbertson, D. and Hunt, C.O., with Smithson, P.A.,1996. Quaternary geomorphology and palaeoecology. In G. Barker, D. Gilbertson, B. Jones, and D. Mattingly Volume One: Synthesis: 49-82 Gilbertson, D., Grattan, J. and Pyatt, B., 1999. Environmental impacts of ancient mining and smelting activities: initial EDMA studies. Pp. 262-9 in G. Barker, R. Adams, 0. H. Creighton, D. Crook, D.D. Gilbertson, J. P. Grattan, C.O. Hunt, D. J. Mattingly, S. J. McLaren, H. A. Mohammed, P. Newson, C. Palmer, F. B. Pyatt, T.E.G. Reynolds and R. Tomber, Environment and land use in the Wadi Faynan, southern Jordan: the third season of geoarchaeology and landscape archaeology (1998). Levant 31: 255-92. Glueck, N. (1935) Explorations in Eastern Palestine II. New Haven, Conn.: American Schools of Oriental Research. Goodchild, R. G., 1950. The Limes Tripolitanus IL Journal of Roman Studies 90: 30-8. Goodchild, R. G., 1952-53. Farming in Roman Libya. The Geographical Magazine 25: 70-80. Goodchild, R. G. and Ward-Perkins, J. B., 1949. The Limes Tripolitanus in the light of recent discoveries. Journal of Roman Studies 39: 81-95. Hauptmann, A., 1989. The earliest periods of copper metallurgy in Feinan, Jordan. In A.Hauptmann, E.Pernicka and G.A.Wagner (eds.) Archaemetallurgie

BIBLIOGRAPHY

Aubreville, A., 1949. Climats, Forats et Desertification de l'Afrique Tropicale. Paris: Societe d'Editions Geographiques Maritimes et Coloniales. Barker, G. and Gilbertson, D., 2000. Living at the margin: themes in the archaeology of drylands. In G. Barker and D. Gilbertson (eds.), The Archaeology of Dry/ands: Living at the Margin London: Routledge, One World Archaeology 39: 3-18. Barker, G., Gilbertson, D., Jones, B. and Mattingly, D., 1996a. Farming the Desert: the UNESCO Libyan Valleys Archaeological Survey. Volume One: Synthesis. Volume Two: Gazetteer and Pottery. London: Society for Libyan Studies; Paris: UNESCO Publishing. Barker, G. and Gilbertson, D., with Hunt, C.O. and Mattingly, D., 1996b. Romano-Libyan agriculture: integrated models. In G. Barker, D. Gilbertson, B. Jones, and D. Mattingly, Volume One: Synthesis: 26590 Barker, G., Creighton, O.H., Gilbertson, D.D., Hunt, C.O., Mattingly, D.J., McLaren, S.J. and Thomas, D.C., 1997. The Wadi Faynan Project, southern Jordan: a preliminary report on geomorphology and landscape archaeology. Levant 29: 19-40. Barker, G., Adams, R., Creighton, O.H., Gilbertson, D.D., Grattan, J.P., Hunt, C.O., Mattingly, D.J., McLaren, S.J., Mohammed, H.A., Newson, P., Reynolds, T.E.G. and Thomas, D.C., 1998. Environment and land use in the Wadi Faynan, southern Jordan: the second season of geoarchaeology and landscape archaeology (1997). Levant 30: 5-26. Barker, G., Adams, R., Creighton, O.H., Crook, D., Gilbertson, D.D., Grattan, J.P., Hunt, C.O., Mattingly, D.J., McLaren, S.J., Mohammed, H.A., Newson, P., 211

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

det Alten Weltt/Old World Archaeometallurgy: 119-36. Bochum: Deutsches Bergbau-Museum, Der Anschnitt, Beiheft 7. Hauptmann, A, 1992. Feinan/Wadi Feinan. American Journal of Archaeology 96: 510-12. Hauptmann, A, 2000. Zur fruhen Metallurgie des Kupfers in Fenan/Jordanien. Bochum: Deutsches BergbauMuseum, Der Anschnitt, Beiheft 11. Hauptmann, A, Begemann, F., Heitkemper, E., Pernicka, E. and Schmitt-Strecker, S., 1992. Early copper produced at F einan, Wadi Araba, Jordan: the composition of ores and copper. Archaeomaterials 6: 1-33. Hughes, J.D., with Thirgood, J.V., 1982. Deforestation, erosion and forest management in ancient Greece and Rome. Journal of Forest History 26: 60-75. Hunt, C.O. and Mohammed, H. A, 1998. Palynology of the Khirbat barrage reservoir sediments: preliminary results. Pp. 21-23 in G. Barker, R. Adams, O.H. Creighton, D.D. Gilbertson, J.P. Grattan, C.O. Hunt, D.J. Mattingly, S.J. McLaren, H.A. Mohammed, P. Newson, T.E.G. Reynolds and D.C. Thomas, Environment and land use in the Wadi Faynan, southern Jordan: the second season of geoarchaeology and landscape archaeology (1997). Levant 30: 5-26. Mattingly, D. J., 1983. The Laguatan: a Libyan tribal confederation in the late Roman empire. Libyan Studies 14: 96-108. Mattingly, D.J., 1987. Libyans and the 'limes': culture and society in Roman Tripolitania. Antiquites Africaines 23: 71-94. Mattingly, D. J., 1996. Explanations: people as agency. In G. Barker, D. Gilbertson, B. Jones, and D. Mattingly,. Volume One: Synthesis: 318-42. Millington, A C. and Pye, K. (eds.), 1994. Environmental Change in Drylands: Beogeographical and Geomorphological Perspectives. Chichester: Wiley. Mortimore, M., 1998. Roots in the African Dust: Sustaining the Drylands. Cambridge: Cambridge University Press. Musil, A, 1907. Arabia Petraea II. Edom. Vienna: Alfred Holder. al-Najjar, M., Abu Dayyeh, A es-S., Suleiman, E., Weisgerber, G. and Hauptmann, A, 1990. Tell Wadi Feinan: a new pottery neolithic tell in southern Jordan. Annual of the Department of Antiquities of Jordan 34: 27-56. Palmer, E. H., 1872. The Desert of the Exodus. New York: Harper and Bros. Rubin, I., Barnes, R.H. and Kana'an, R., 1997. Mapping and preliminary survey in Wadi Faynan, south Jordan. Annual of the Department of Antiquities of Jordan 41: 433-52. Simmons, A.H. and al-Najjar, M., 1996. Test excavations at Ghwair I, a neolithic settlement in the Wadi Feinan. ACOR Newsletter 8.2: 7-8. Spooner, B., 1989. Desertification: the historical significance. In R. Huss-Ashmore, and S. H. Katz (eds) African Food Systems in Crisis. Part One:

Micro-Perspectives: 111-62. New York: Gordon and Breach. Thomas, D.S.G. and Middleton, N.J., 1994. Desertification: Exploding the Myth. Chichester: John Wiley and Sons. Tiffen, M., Mortimore, M. and Gichuki, F., 1994. More People, Less Erosion: Environmental Recovery in Kenya. Chichester: John Wiley and Sons. van der Veen, M., Grant, A and Barker, G. (1996) Romano-Libyan agriculture: crops and animals. In G. Barker, D. Gilbertson, B. Jones, and D. Mattingly, Volume One: Synthesis: 227-63. Widgren, M., 2000. Islands of intensive agriculture in African drylands: towards an explanatory framework. In G. Barker and D. Gilbertson (eds) The Archaeology of Dry lands: Living at the Margin: 252-67. London: Routledge, One World Archaeology 39.

212

Graeme Barker: Floodwater Farming, Roman Imperialism, and Desertification in Libya and Jordan

19.1

Palmer's pen-and-ink sketch of the ancient town ofShivta in the central Negev desert, Israel, a typical nineteenthcentury image contrasting past glories with present-day barrenness. (After Palmer, 1872: 314) N 250 kms

200mm

NIGER MAU

0

CHAD

lOO

~ 00mm

0

v( ,';>

0

'(,

Fs1R'\ GRAND

!~---,.__~-

ERG

ORIENTAL

.-----~!-(/_ BaNJem

C

50mm

------

-1

19.2 Tripolitania, northwest Libya, showing the principal landforms and settlements, and the location of the UNESCO Libyan Valleys Survey. The 200 mm, 100 mm, 50 mm, and 25 mm dashed lines are rainfall isohyets. The contours are in metres.

213

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

limeHooe

'tr;:::r-

"""' =====

19.3 Romano-Libyan floodwater farming systems in the Wadi Gobbeen in the Libyan Predesert.

19.4 From right to left, Barri Jones, Bob Bewley and Peter Hayes examining the sediments exposed by a cut through wall 1250 in Figure 19.3. (Photograph: G. Barker)

214

Graeme Barker: Floodwater Farming, Roman Imperialism, and Desertification in Libya and Jordan

Lake Tiber'

,t.

/,'

.,

.,.

\

,.--,.

f"'t-

----

...

.,,,,

,.., I

New~

20.16: Euesperides (Benghazi). Updated plan ofEuesperides, showing locations of Areas P, Q and R and transects 1 and 2.

235

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

C E

M Tombs E ---1973 T E R I E

s

HARBOUR

I

E

s INNER

HARBOUR

Ancient Quarries

~

20.17: Euesperides (Benghazi). Sketch plan of Benghazi showing the location ofEuesperides, the later Hellenistic settlement and the possible location of the later harbour. (After J. Lloyd).

236

21 Romano-Libyan Agriculture in the Tripolitanian Predesert: Mediterranean Perspectives Chris Hunt and David Gilbertson parallels from extant floodwater farming systems around the Mediterranean to examine aspects of the functioning of the Predesert systems. This approach might be termed crudely ethno-archaeological.

Abstract Romano-Libyan agriculture in the Tripolitanian Predesert has many aspects in common with modern and recent Mediterranean farming systems, including the range of crops grown and a number of agricultural practises. Thus, some problematical features of the archaeological record of Romano-Libyan agriculture in the Predesert may, potentially, be explained by comparison with these Mediterranean systems. Similarities may exist, either because of common inheritance or because only a limited range of possibilities are available for farmers to address certain problems. Analogies, however, are only analogies: there are inevitably problems of equifinality (different causes leading to the same phenomenon) and multicausality in this approach.

The nature and quality of evidence from observational research in the Predesert The evidence from the Predesert was largely summarised in the two-volume report of the UNESCO Libyan Valleys Survey (Barker et al. 1996 a & b ). The survey was a combined enterprise between the Department of Antiquities in Tripoli and a group of archaeologists and landscape specialists from the Universities of Leicester, Huddersfield, Manchester, Sheffield and Bournemouth. The archaeological evidence is summarised in a Gazetteer, based upon 2437 site records (Barker et al. 1996b). Many sites in the Gazetteer are complex, thus a single entry, for example, deals with several complex networks of wadi walls over a 10 km stretch of the Wadi umm el-Kharab (Barker et al. 1996a). Subsequent research is presented in Gilbertson et al. (2000) and Hunt et al. (in press). This paper explores problematical areas in the survey results by drawing ethno-archaeological parallels between contemporary or recent Mediterranean floodwater farming systems and the evidence found in the Tripolitanian Predesert.

Introduction The Libyan Valleys Survey (Barker et al. 1996a, b) has resolved much of the character and pattern of ancient settlement in the Tripolitanian Predesert, which, during and for many years after the Romano-Libyan period, revolved around the large-scale application of floodwater farming technology. In the first century AD, the floodwaterfarming systems were established for the production of significant agricultural surpluses, notably of olive oil, which were traded with the great coastal cities of Tripolitania and thence into the Roman world. This necessitated the indigenous Macae pastoralists developing a sedentary system, which then persisted for over a thousand years. After the end of the Roman trading system, during the Byzantine and Arab periods, the settlements and their floodwater farming systems persisted in many areas but they were worked largely for subsistence.

One of the problems with work of the type done by projects such as the Libyan Valleys Survey is that multicausality and equifinality are rife in the real world. Occam's razor, in the context of observational science, may be more a lethal weapon than a useful tool, particularly because causality may be counter-intuitive. Competing explanations of phenomena, although mutually contradictory, may be all more-or-less correct locally, in certain given sets of circumstances. Many potential explanations probably exceed the strength of the present evidence and cannot be investigated effectively with the methods available to the project. It is extremely difficult to separate cause from effect; pre-disposing or maintaining factors from proximate causes; or to disentangle feedback or synergistic effects in any observational project. Indeed, sustained, intensive and systematic study was largely impossible for the UNESCO Libyan Valleys Survey.

Many issues relating to the initiation, growth, stability and eventual decline of farming and settlement in the Predesert still remain remarkably unclear, despite two decades of research by a large multidisciplinary team (Gilbertson et al. 2000). Analyses of the farming systems in the Predesert have, so far, been mostly from archaeological, palaeoenvironmental and geomorphological perspectives. It is clear that the floodwater farming technology was indigenous and sometimes extremely sophisticated (Gilbertson and Hunt 1996). Recent research (Hunt et al. in press; Gilbertson et al. 2000) suggests that the walls of the floodwater farming technology are simply the most visible part of a sophisticated wider strategy for land management in which animal husbandry was a critical component. Most of the details of this wider land use strategy are still remarkably obscure. Here, we adduce

Aspects of this indeterminacy were articulated early in the UNESCO Libyan Valleys Survey project. Thus, Gilbertson et al. (1984) observed that walls built by the Romano-Libyan agriculturists might have different functions - either at the same time, or at different times. Nevertheless, it is clear that hydrological considerations were of critical importance for the design of much of the

237

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Local, neo-Punic-speaking Macae people (herein referred to as Romano-Libyan) seem to have built most of these structures. The Macae were transhumant pastoralists before the extension of Roman influence to Tripolitania. Under favourable trading conditions brought about by Roman control of the coastlands, they rapidly developed a robust, long-lasting, mixed farming economy, which later endured turbulence caused by the end of Roman influence, wars during the Byzantine period, and the Arab and Turkish invasions. This dramatic transformation of Predesert settlement was associated with a rise in population to about 20,000, with perhaps 2,000 farms in the study region during the first century AD and an estimated 1,000 gsur from the third century AD (Flower and Mattingly 1995, Mattingly with Flower 1996).

ground plan of many of the wall systems investigated (Gilbertson et al. 1984, Hunt et al. 1986; Gilbertson and Hunt 1996). Other aspects of the systems are not hydrologically plausible, for instance, many walls in the supposed spillway area in the Wadi el-Amud complex (Gilbertson and Hunt 1996). Information is still very limited in scope and reliability. Many biases are evident - for example, the difficulties of access by off-road vehicle in the rugged and inhospitable terrain meant that a concentration of survey upon the more accessible wadis was inescapable. Vast areas of difficult ground, especially on the plateaux between wadis, were not surveyed. This survey pattern is likely to have underrepresented whole classes of features. Some of these biases were detected through analysis of Landsat images (Dorsett et al. 1984). Detailed surveys on foot were done only in the Wadi Umm el-Kharab, and sectors of the Wadis Mansur, el-Amud, Gobbeen, and Mimoun (Barker et al. 1996a). Some walls on the wadi-floor were no doubt removed in antiquity by erosion or because their stones were needed elsewhere. Some walls are likely to be completely buried by sediment, as was demonstrated by finding walls invisible on the surface in rare gully exposures (Hunt et al. 1986). Other buried walls were sometimes visible as lines of bushes (Gilbertson et al. 1994). Elsewhere in the Predesert, our understanding is mostly based upon vehicle traverses which generated field sketches and simple field maps (Gilbertson et al. 1984) - a consequence of the lack of air photographs and even accurate base maps during the survey. There are similar uncertainties with the palaeoeconomic information, which is derived from excavations at middens or in buildings at only four open farms and six gsur (Barker and Gilbertson 1996: Van der Veen et al. 1996).

During the Roman period, surpluses of olive oil and cereals were produced and traded with the great coastal cities of the Tripolis, whence they were exported into the huge market of the Empire. Products produced elsewhere in the Empire were traded southward into the Predesert, alongside 'exotic' food such as deep-water fish. Less is known about activities that leave little trace in the landscape, such as stock-keeping, but clearly domestic animals were kept. Evidence from one farm in the south of the area suggests that stock were stalled at least part of the time and thus may not have been allowed to roam and graze widely, as do goats in the area today (Hunt et al. in press). As Roman influence waned in the Predesert during the sixth to seventh centuries AD, quantities of olive oil sent to the coast declined substantially and the rather dispersed network of farming systems were consolidated into substantial 'agricultural estates' controlled by powerful local elites (Mattingly 1996). These 'estates' are indicated by a reordering of walls and gsur, without any clear topographic or hydrological causality (Mattingly 1996; Gilbertson and Chisholm 1996). At this time, GIS-based analyses suggest both a general 'thinning' of settlement and a slight northward and westward shift. During the Islamic period Mattingly with Flower (1996) distinguished a notable shift northwards and a trend towards clustering of settlement. Today, the only remaining significant floodwater-farming is associated with the large settlement at Beni Ulid, located in the northern part of the Predesert. Elsewhere, transhumant pastoralism characterises the human geography of the modem landscape.

Predesert agriculture

The scale of ancient agricultural act1v1ty in the Tripolitanian Predesert was considerable. It took place in a desert environment broadly similar to that of today (Hunt et al. 1987; Gilbertson et al. 1994, 1996). The wall systems, which seem to have been the main tools of floodwater-farming based agriculture occupy the floors of all but the largest wadis (Gilbertson & Hunt 1996). The walls were initially associated with large undefended farms and farmsteads attributed to the first to the third centuries A.D., and barn-like, imposing, possibly fortified gsur, mainly attributed to the third to fourth centuries A.D, though activity occurred at some dates as late as the sixteenth century A.D. (Gilbertson and Hunt 1988). These must have provided massive secure storage of food as a buffer against sequences of adverse drought years. A number of studies of floodwater farming have commented on the importance of risk-spreading and risk-minimisation by storage strategies (Hunt et al. 1998; Gilbertson et al. 2000).

Stock Control

One of the key areas of indeterminacy is the role of livestock in the ancient economy of the Tripolitanian Predesert. Studies of animal remains suggest that livestock management was one of the key elements leading to the success of agricultural systems in the arid Mediterranean lands during the Roman and post-Roman periods (Van der Veen et al. 1996; Gilbertson et al. 2000;

238

Chris Hunt and David Gilbertson: Romano-Libyan Agriculture in the Tripolitanian Predesert

Rosen 2000). Current models (Barker and Gilbertson 1996; Gilbertson et al. 2000) suggest that the manure produced by livestock may have been important in maintaining longterm fertility in these agro-ecosystems. Many aspects of this stock management are still largely unknown.

possibly interpret past processes and patterns. Modem analogues in the Tripolitanian Predesert are rare, because only in the Wadi Merdum near Beni Ulid are there partlyoperating floodwater farming systems with large, but now very damaged cross-wadi walls (Gilbertson et al. 1994). Cross-wadi walls promote sedimentation and infiltration, which raises soil moisture, and promotes shrub growth on their upstream side. Downstream from walls, flood scour and sapping occurs. Here, during the passage of a storm pulse, the roots of tree crops - olives and date palms - bind the floodplain alluvium locally. Between the trees, scour can produce gullies over a metre deep. Broadcast barley is grown on recently deposited flood-loams, especially on sediments confined by cross walls in the side wadis. This is in many ways a poor analogue for the ancient farming, since land management today appears haphazard, walls are not maintained and the range of crops grown is comparatively narrow.

Recent palynological research (Hunt et al. in press) on pollen assemblages from goat droppings at the LM4 farmstead in the Wadi el-Amud suggests that livestock were penned and fed on crop residues. Geomorphological evidence at the same site is consistent with this interpretation (Gilbertson and Hunt 1990). Large quantities of goat droppings in middens at the Medieval Gasr Adzam (Gilbertson and Hunt 1988) also point to stock penning. Similar evidence is still lacking for the rest of the Predesert, but indications may be gained from geomorphological studies and event correlations in the Wadis Merdum and Mansur (Hunt et al. 1986; Gilbertson and Hunt 1996). There, the major wall systems were built on what appears to have remained a relatively stable land surface from the time of wall construction, most probably in the first few centuries AD until a period of major sediment flux, probably during the sixteenth century AD. Palaeoecological studies of the post-Roman abandonment phase in the Wadi Mimoun (Hunt et al. 1987) and at LM4 (Hunt et al. in press) suggest that the landscape prior to the sixteenth century AD was steppic in aspect, with abundant grassy and shrub vegetation. In contrast, a single sample from a wall infill probably equivalent stratigraphically to the sixteenth century floodloams in the Wadi Mansur gave an indication of a substantially degraded steppe, with comparatively high incidences of taxa typical of open, eroding habitats. Goat dung is abundant locally lying on the old land surface at the base of the floodloams (Hunt et al. 1986). Within the floodloams in the Mansur, and in other wadis, a series of small cairns were found lying on a minor palaeosoil. These were hypothesised to be placed around trees to protect their roots from erosion (Hunt et al. 1986).

The crops and animals once grown in the Predesert are overwhelmingly of Mediterranean origin (Table 1: Van der Veen et al. 1996) so it is arguably valid to search for parallels in floodwater farming in other Mediterranean countries. The present archaeobotanical evidence suggests that there were no fundamental differences between the agricultural economies of the Romano-Libyan, Byzantine and Islamic periods. In general, hunted as opposed to herded animals became increasingly more important further south into the desert. Table 1: Crops and animals in the Tripolitanian Predesert, First to Fifteenth Century AD. Century AD. CROPS 1-5 10-16 Hordeum vulgare (hulled six rowed barley) + + Triticum (wheat) + + Pisum sativum (field pea) + + Lens culinaris (lentil) + + other pulses + + Ficus carica (fig) + + V itis vinifera (grape) + + Phoenix dactylifera (date) + + Olea europea (olive) + + Prunus amygdalus (almond) + + Pistacia atlantica (wild pistacio) +

Here, we run into the problems of equifinality and multicausality. It is possible to regard the signs of landscape degradation in the sixteenth century as evidence for a change in land management and the end of the Romano-Libyan practise of livestock penning (Gilbertson et al. 2000). It is equally possible that a major arid phase may have destabilised the landscape by reducing vegetation cover, thus causing aggradation of floodloams. Such was reported further east in the margins of the arid zone during the 'Little Ice Age' and may have been felt as far west as Tunisia (Hunt et al., in preparation). Equally, it is possible that such an arid event may have caused the final abandonment of livestock penning, and other aspects of the Romano-Libyan system.

ANIMALS sheep/goat gazelle bovid pig canid camel hare/rabbit equid antelope

+ + + + + + + + +

+ + + + + + + +

Mediterranean parallels

Extant floodwater-farming systems are present in arid Mediterranean countries including the Maltese Islands (Jones and Hunt 1994; Jones et al. 1998), Almeria, Spain

To subvert Hutton's dictum, the present can offer useful analogues and models with which to compare, evaluate and 239

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

(Thompson and Scoging 1995; Hunt & Gilbertson 1998), and the Balearic Islands (unpublished data). None of these places are as arid as the Tripolitanian Predesert, and social and economic conditions are markedly different from Roman Tripolitania. The people, their customs and traditions are different, too. Nevertheless, they deal with similar problems, if at different scales, and they share in the rich heritage of Mediterranean agriculture.

below the checkdams in the lower part of the system. Livestock are usually stalled. Goats are sometimes grazed on the interfluves between floodwater-farming systems, but they are excluded from the systems. This system of polyculture is very similar to that hypothesised for the Predesert (Gilbertson et al. 1994: 238). Tree crops, in this case almonds, are also grown behind checkdams of another modem floodwater-farming system, which was actually rebuilt around 1990, at Gergal adjacent to the nineteenthcentury system described by Hunt and Gilbertson (1998).

Maltese floodwater farming (Jones and Hunt 1994; Jones et al. 1998) is on a very small scale and is fed by springs and runoff, which is often collected into small rock-cut cisterns and then used for hand-watering. It is characterised by immense care and attention to detail, with small areas of crops separated one from another by lines of stones or low earth banks and with intricate channel networks for irrigation (Jones et al. 1998). Today agriculture on Malta is mainly horticultural, but in the recent past, grain and cotton were important crops. Olives are cultivated without irrigation. In the Maltese system, livestock are stalled. An informant on the Island of Gozo maintained that the livestock needed shelter from the summer sun and were therefore stalled. They would also be difficult to control on the steep terraced slopes and would eat off or trample all the tender plants. This parallels the hypothesised stalling of livestock in the Predesert. In the Balearic Islands, floodwater farming systems are widely distributed. These provide a number of useful parallels with the Predesert. On the Island of Formentera, rock mulches are used at the base of fig trees, partly to anchor laths around the trunk of the trees to prevent animal damage (Illus. 21.1 ). Large numbers of molluscs colonise the space between the tree trunks and the laths. This is a close parallel for the stone mounds described by Hunt et al. (1986) from the sixteenth century floodloams in the Wadi Mansur, which were full of snail shells. The stone mounds in the Predesert were thus probably used, with laths or similar material, to protect trees from damage by stock. The age of stone mounds in other wadis thus becomes important for the evaluation of the hypothesis of stalled animals prior to the sixteenth century. Although no Roman period mounds were located in the Wadi Mansur, absence of evidence cannot be taken as strong evidence of absence. Formenteran agriculture, like that on Malta, is also characterised by careful demarcation of crops by lines of stones. These parallel stone lines in the floodwater farming systems of Tripolitania which were hypothesised to have demarcated crops (Gilbertson and Hunt 1996).

Discussion and conclusion Careful landscape management appears to have characterised Predesert farming. This is very evident in the thoughtful construction of some Predesert wadi wall systems (Gilbertson and Hunt 1996). Other features of Predesert agriculture, notably the stone lines, seem to reflect similar small scale, detailed care and patterns of land management not dissimilar to those still present on Malta and the Balearic Islands. This may imply that in some areas horticulture - or horticultural-style agriculture was of importance. In general, the role of tree crops seems extremely important in Almeria today and extant systems there offer close parallels for the pattern of land management hypothesised for the Predesert by Gilbertson et al. (1994). This coincides with the important evidence for olive oil production in the Predesert (Mattingly 1996). As part of that pattern of careful management which is characteristic of many extant floodwater farming systems, livestock were a critical component of the Romano-Libyan agro-ecosystem. It is suggested that livestock were largely stalled, as they often are today in arid Mediterranean lands, rather than free-ranging. This enabled farmers to maintain long term landscape stability - for well over a millennium in some areas. Landscape stability was precious to farmers - rapidly aggrading or incising wadi floors would have necessitated wholesale rebuilding and remodelling of systems and a consequent high labour requirement, which would have made farming less profitable and possibly nonviable. Gilbertson et al. (1993) and Barker and Gilbertson (1996) suggested dung from livestock would have been critical as a source of manure to maintain long-term fertility. The midden excavations at LM4 (Gilbertson and Hunt, 1990) and Gasr Adzam (Gilbertson and Hunt 1988) show, however, that this important resource was not always used. The feeding of crop residues to stock or controlled grazing on purposely-grown swards is, however, consistent with a high value being placed upon livestock. Hunt et al. (in press) have also cited evidence of the symbolic importance of livestock in the Predesert.

In the Provincia di Almeria, Spain, large floodwater farming systems approaching the Tripolitanian systems in scale are still operational (Thompson and Scoging 1995; Hunt and Gilbertson 1998), for instance near Turre. Hunt and Gilbertson (1998) describe hydrological control by a series of checkdams and sluices. Tree crops, principally almonds, are grown in checkdammed fields in the upper part of the system, cereal fields lie behind checkdams in the lower part of the system. Old olive trees are grown just

The appearance of stone mounds associated with wadi alluviation in the sixteenth century may reflect a shift from stalled to free-ranging livestock, which may signal the effective end of the Romano-Libyan land-management

240

Chris Hunt and David Gilbertson: Romano-Libyan Agriculture in the Tripolitanian Predesert

system and the inception of land degradation. The role of climate change in these events is still obscure, but it is clear that simple climate change on its own was unlikely to have caused the end of Romano-Libyan floodwater farming patterns. The enduring resilience of floodwater farming technology is clear.

Barker, G. W. W., Gilbertson, D. D., Jones, B. & Mattingly, D. J. (eds.) Farming the Desert: The UNESCO Libyan Valleys Survey. Paris: UNESCO: 91225. Gilbertson, D. D., Hunt, C.O., Fieller, N.R.J. and Barker, G.W.W., 1994 The environmental consequences and context of ancient floodwater farming in the Tripolitanian Predesert. In Millington, A. C. and Pye, K.E. (eds.). Environmental Change and Geomorphic Processes in Arid Lands. Chichester: John Wiley: 229 - 251. Gilbertson, D. D., Hunt, C. 0. and Smithson, P.A. 1996 Quaternary Geomorphology and Palaeoecology. in Barker, G. W. W., Gilbertson, D. D., Jones, B. and Mattingly, D. J. (eds.) Farming the Desert: The UNESCO Libyan Valleys Survey. Paris: UNESCO: 4982. Gilbertson, D. D., Hunt, C.O., and Gillmore, G.K., 2000. Success, longevity and failure. Romano-Libyan floodwater farming in the Tripolitanian Pre-desert. In Barker, G.W.W. and Gilbertson, D. D. (eds.). The Archaeology of Drylands: living on the margins. London and New York: Routledge: 137-159. Hunt, C. 0. and Gilbertson, D. D., 1998 Context and impacts of ancient catchment management in Mediterranean countries: implications for sustainable resource use. In Wheater, D. and Kirby, C., (eds.) Hydrology in a changing environment. Volume IL Chichester: John Wiley: 473-483. Hunt, C. 0., Gilbertson, D. D., Van de Veen, M., Jenkinson, R. D. S., Yates, G. and Buckland, P. C., 1987 The UNESCO Libyan Valleys Survey XVII: the palaeoecology and agriculture of the abandonment Phase at Gasr Mml0, Wadi Mimoun in the Tripolitanian Predesert. Libyan Studies 18: 1-14. Hunt, C. 0., Mattingly, D. J., Gilbertson, D. D., Barker, G. W. W., Dore, J. N., Bums, J. R., Fleming, A. M. and Van der Veen, M., 1986. The UNESCO Libyan Valleys Survey XIII: interdisciplinary approaches to ancient farming in the Wadi Mansur, Tripolitania. Libyan Studies 17: 7-47. Hunt, C. 0., Rushworth, G., Gilbertson, D. D. and Mattingly, D. J., in press. Romano-Libyan dryland animal husbandry and landscape: pollen and palynofacies analyses of coprolites from a farm in the Wadi el-Amud, Tripolitania. Journal of Archaeological Science. Jones, A. M. & Hunt, C. 0., 1994. Walls, wells and water supply: aspects of the cultural landscape of Gozo, Maltese Islands. Landscape Issues 11, 1: 24-29. Jones, A. M., Hunt, C. 0. & Crook, D. S., 1998. Traditional irrigation strategies and their implications for sustainable livelihoods in semi-arid areas. in Wheater, D. and Kirby, C. (eds.) Hydrology in a changing environment. Volume IL Chichester, John Wiley: 485-493. Mattingly, D. J., 1996. Explanations: people as agency. In Barker, G.W.W., Gilbertson, D. D., Jones, G.D.B. and Mattingly, D. M. (eds.), Farming of the Desert: The UNESCO Libyan Valleys Archaeological Survey.

BIBLIOGRAPHY

Barker, G., 1996. Prehistoric settlement. In Barker, G.W.W., Gilbertson, D. D., Jones, G.D.B. and Mattingly, D. M. Farming of the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis. Paris, UNESCO Publishing: 83110.

Barker, G. and Gilbertson, D. D. (with Hunt, C. 0. and Mattingly, D. J.), 1996. Romano-Libyan agriculture: integrated models. In Barker, G.W.W., Gilbertson, D. D., Jones, G.D.B. and Mattingly, D. M., Farming of the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis. Paris, UNESCO Publishing: 265-290. Barker, G.W.W., Gilbertson, D. D., Jones, G.D.B. and Mattingly, D. M., 1996a Farming of the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis. Paris, UNESCO Publishing, (Ed. G.W.W. Barker). Barker, G.W.W., Gilbertson, D. D., Jones, G.D.B. and Mattingly, D. M., 1996b. Farming the Desert: the UNESCO Libyan Valleys Archaeological Survey. Volume 2. Gazetteer and Pottery. Paris, UNESCO Publishing, (Ed. D. M. Mattingly). Dorsett, J. E., Gilbertson, D. D., Hunt, C. 0. and Barker, G. W. W., 1984. The UNESCO Libyan Valleys Survey IX: Image analysis of Landsat data and its application to environmental and archaeological Surveys. Libyan Studies 15: 71-80. Flower, C.P. and Mattingly, D.J., 1995. ULVS XXVII: Mapping and spatial analysis of the Libyan Valleys Data using GIS. Libyan Studies 26: 49 - 78. Gilbertson, D.D. and Chisholm, N.W.T., 1996. ULVS XXVIII: manipulating the desert environment: ancient walls, floodwater farming and territoriality in the Tripolitanian pre-desert of Libya. Libyan Studies 27: 17-52. Gilbertson, D.D. and Hunt, C. 0., 1988. The UNESCO Libyan Valleys Survey XIX: The Cenozoic geomorphology of the Wadi Merdum, Beni Ulid, in the Libyan Pre-desert. Libyan Studies 19; 95-121. Gilbertson, D. D. and Hunt, C.O., 1990 The UNESCO Libyan Valleys Survey XXI: geomorphological studies of the Romano-Libyan Farm, its floodwater control structures and weathered building stone at site LM4, at the confluence of the Wadi el Amud and the Wadi Umm el Bagel in the Libyan Predesert. Libyan Studies 21: 25 - 42. Gilbertson, D. D. and Hunt, C. 0., 1996b Romano-Libyan Agriculture: Walls and Floodwater Farming. m 241

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Volume 1. Synthesis. Paris, UNESCO Publishing: 319-342. Mattingly, D. J. with Flower, C., 1996. Romano-Libyan settlement: site distribution and trends. in Barker, G.W.W., Gilbertson, D.D., Jones, G.D.B. and Mattingly, D.M., (eds.) Farming of the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis. Paris, UNESCO Publishing, (Ed. G.W.W. Barker): 159-190. Rosen, S.A., 2000. The decline of desert agriculture: a view from the Classical Period Negev. In Barker, G.W.W. and Gilbertson, D. D. (eds.), The Archaeology of Drylands: living on the margins. London and New York: Routledge: 45-62. Thompson, D. A. and Scoging, H., 1995. Agricultural terrace degradation in SE Spain: modelling and management studies. In McGregor, D. F. and Thompson, D. A. (eds.) Geomorphology and land management in a changing environment. Chichester: John Wiley: 153-177. Van der Veen, M., Grant, A. and Barker, G., 1996. Romano-Libyan agriculture: crops and animals. In Barker, G.W.W., Gilbertson, D. D., Jones, G.D.B. and Mattingly, D. M. 1996a. Farming of the Desert: The UNESCO Libyan Valleys Archaeological Survey. Volume 1. Synthesis. Paris, UNESCO Publishing; 227-264.

242

Chris Hunt and David Gilbertson: Romano-Libyan Agriculture in the Tripolitanian Predesert

21. 1

Stone mulch and laths protecting a young fig from animal damage, Formentera. (Photo D. D. Gilbertson)

243

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

244

22 Farming the Desert: two contrasting Libyan case-studies David Mattingly irrigated systems of the oases (Birebent 1962; Shaw 1984; 1991; Slim 1992; Trousset 1986; 1987).

Introduction This paper combines three themes that were prominent features of the career of Barri Jones: Libya, Roman frontier zones and rural settlement. I shall review the work of two archaeological surveys in the Libyan desert. The UNESCO Libyan Valleys Survey (ULVS) was launched at the behest of Colonel Ghaddafi to investigate the abundant traces of Roman period farming in the Libyan pre-desert and was co-directed by Barri Jones and Graeme Barker. I was fortunate enough to be included in their team, investigating the wadi systems of the Sofeggin and Zem-Zem basins c.100-300 km south of Tripoli (a zone some 50,000 sq. km in extent) across five seasons (1979-1981, 1984, 1989). The Fezzan Project (FP) has been directed by myself and between 1997-2001 has examined the archaeology of an oasis belt deep in the Sahara, over 1000 km south of Tripoli (Illus. 22.1 ). This was the heartlands of the Garamantes, a Saharan people presented by the Roman sources as the archetype of barbarian nomads (though, as we shall see, archaeological research indicates different conclusions). The archaeological stories that have emerged in both cases have been extraordinarily rich in detail and quality of data, yet are fundamentally different in many respects. 'Farming the Desert' was not just about knowhow and technology. I shall argue that the divergence between Predesert and full desert farming regimes is illustrative equally of environmental and social differences between the zones.

In examining each of these systems we find archaeological corroboration of what the literary and epigraphic sources clearly attest; Africa was brought to prosperity not because of its natural abundance, but through intensive and timeconsuming investment in hydraulic systems. The availability and relative abundance of water at a given location were critical factors in its selection for settlement. If such factors were important even in the better-watered parts of the Maghreb, they were even more critical in the Sahara. (ii) political factors

In assessing the evidence for ancient farming in my two case-study areas, I have considered a series of factors that may have contributed to the precise solutions that emerged:

The decisions by the Roman state that lay behind the vast areas of centuriated lands in the area of present day Tunisia, represent one end of the spectrum of political interference in pre-Roman landholding (Mattingly 1997a; 1998). The Roman domination of the Maghreb clearly brought with it major changes in the way land was held, and used (and by whom). The scale of sedentary farming increased dramatically and in some areas of the high steppe and Predesert developed from practically nothing. The evidence of Roman delimitation of tribal lands demonstrates that there was generally an element of political control involved in this development. We cannot presume that indigenous development of land took place in a vacuum, even on the desert margins. Negotiation between the Roman State and her subjects, both collectively and individually, was a determining factor in the allocation and confirmation of land. The key point to bear in mind is that the Predesert zone surveyed by the ULVS, though not centuriated, still, under Roman rule, comprised a landscape of imperialism. On the other hand, the Garamantes lay outside the Roman empire, though not entirely beyond its military reach, and the evolution of farming there should be seen as an essentially indigenous process.

(i) Environmental factors

(iii) social factors

Farming in ancient Africa depended on soil quality and water availability, with the latter being the more critical factor. The exploitation and management of the scant water resources of North Africa has always been linked to the economic and demographic growth of the region (Mattingly and Hitchner 1995: 187-88). There were several different water regimes in North Africa, each with its own archaeology. These relate, for instance, to the exploitation of perennial springs and seasonal rainfall, to the runoff farming and water technology of the Predesert, to the

Who were the people involved in settling new areas of farmland in the Predesert? At one extreme, they could have been colonists from Italy setting up farms in the deep hinterland of one of the coastal cities. At the other, they were African pastoral tribes becoming sedentarised, whether willingly or under a degree of persuasion. These are crude stereotypes. In reality the possible social groups involved in the development of new areas of farming in the Roman frontier zone were more numerous and more complex. There are vital questions to ask. Were the farmers Africans? Were they free? Is there evidence of

Barri was responsible for my North African interests (and much else in my academic formation and career) and this paper is a small tribute to him. What factors constrain rural settlement patterns?

245

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

significant social differentiation? Did settlement patterns change over time in line with social developments?

et al. 1996a: 291-342). From this, it is clear, for instance, that climatic change does not provide a unitary explanation for the demise of wadi farming, though environmental changes brought about by a range of human and natural factors can be seen to have contributed to a very complex equation. Overall, it is the socio-political factors that now seem to have been most critical in opening up the area for sedentary farming and in contributing to its later decline.

Garamantian agriculture also raises interesting social questions, notably about the extent to which it may have depended on the use of forced or slave labour. It is also appropriate to question to what extent a perceived shrinkage in the area of irrigated agriculture can be linked to a decline in the kingdom's power at the regional scale.

The settlement hierarchy was dominated by elite farms, initially undefended, but, from the third century AD, increasingly replaced or augmented by fortified farms, known as gasr (singular), gsur (plural). There were large numbers of less substantial farms and farmsteads, small settlements, huts and tent bases (occurring singly and in groups), along with a more restricted number of hilltop villages (which may represent an element of pre-Roman settlement continuing into the Roman period). Olive production played a significant part in the regional economy to judge from over 60 presses located, though mostly in single units and with no more than two presses known at any site (Barker et al. 1996a: 112-58; Jones 1985).

(iv) Technological diffusion The processes by which technological mastery of desert environments was achieved and the technology diffused are elusive. What is clear from the studies I shall present is that two adjacent areas could adopt entirely different technical solutions to the question of water exploitation. In much previous scholarship the assumption has been that the Mediterranean and its dominant powers and civilisations (Greeks, Phoenicians and Romans) formed the key conduit for technological diffusion in classical antiquity. One of the interesting points to emerge from this Libyan research is the identification of significant inland desert routes along which innovations in crops and technology may have passed.

It is now abundantly clear that the wadi farmers were

indigenous Libyans, Romanised and/or Punicised to a varying degree, and not settlers from outside the region or overseas (Mattingly 1995: 160-70). The economic success of a sizeable elite group amongst them, despite the harshness of the Predesert climate, is indicated by more than seventy mausolea and by the architectural pretension of the major sites (exemplifying also the fine preservation of all categories of site). It is presumed that many of the minor structures were held in some sort of dependency relationship by the upper echelon settlements (Barker et al. 1996a: 321-4). The farming settlements are believed to represent the sedentarisation within the frontier zone of Libyan tribes who were probably transhumant pastoralists in the pre-Roman period (Mattingly 1998: 168-78; Grahame 1998). Questions remain as to whether this was a process internal to the Libyan tribal groups or more directly attributable to the incorporation of this area into the Roman frontier zone of the province.

The UNESCO Libyan Valleys Survey

The dossier of information generated by the ULVS is vast. It comprises more than 2,500 sites recorded (Illus.22.2), over 55,000 sherds of pottery collected and processed, over 30 preliminary publications and a final two-volume report published (Barker et al. 1996a/b; most of the interim reports are published in Libyan Studies from volume 11 (1980) onwards). It is obvious from a glance at the map that the settlement is clustered along the dry river beds (wadis) that are a key feature of the Predesert landscape. The intervening landforms comprise a range of barren gravel plains and rocky plateau, which offered only very limited possibilities for cultivation. Relief carvings of agricultural and pastoral scenes from Roman-period tombs, combined with botanical and faunal data from excavated sites, show that a mixed farming economy flourished here, contrary to the indications of modem land-use and settlement (primarily transhumant bedouin). A major aspect of the ULVS work was the examination of the field systems, water catchment and water control features (Illus. 22.3), demonstrating that the 'greening' of the Predesert was the result of sophisticated floodwater farming, maximising the impact of the limited and erratic rainfall (Barker et al. 1996a: 191-225, 265-90). The range of data collected on the geomorphology, sedimentology, erosion patterns, hydrology, palynology and palaeoecology is also substantial (Barker et al. 1996a; 21-82, 227-63). As a result, we can evaluate the reasons for the rise and decline of the run-off farming systems. In the final report of the project, two chapters were devoted to discussing 'Environment as Agency' and 'People as Agency' (Barker

In preparing the final report of the survey, we used a GIS to facilitate the mapping and analysis of the spatial data, with some interesting results (Flower and Mattingly 1995; Barker et al. 1996a: 159-90). The distribution of early Romano-Libyan sites (first - early second century AD) demonstrates that the introduction of the Romano-Libyan farming system was the result neither of a gradual diffusion of people and/or ideas, nor of a process of military imposition in a few locations followed by outward spread into surrounding areas. Rather, the process appears to have been one of rapid and wholesale change, with settled agricultural sites achieving their maximum geographical coverage within the early Romano-Libyan period.

246

David Mattingly: Farming the Desert: two contrasting Libyan case-studies

Furthermore, there is very little change in the overall distribution of sites with the transition from open farm to gasr-based systems in the third century and later, even though there is evidence for some expansion or intensification in the exploitation of individual wadis in this period (Barker et al. 1991). Perhaps, then, the apparent general continuity in the geographical extent of settlement conceals spatial changes at a larger scale, but if so, what was the nature of these changes, and how might they be identified and analysed?

hypotheses and observations on the complete data set of the ULVS using the spatial analysis capabilities of the GIS. In order to evaluate the hypotheses outlined above, a cost surface analysis of distributions of distance from wadi junctions and headwaters was compiled for the following classes of sites: all farms/farmsteads present in the early Romano-Libyan period (first and second centuries AD), all gsur, all Romano-Libyan cemeteries, and, as a form of comparative control, all sites recorded by the Valleys Survey (the method is described in detail in Flower and Mattingly 1995: 65-76). The vast majority of sites fall within four kilometres of a major wadi junction or headwater, with numbers falling off rapidly beyond this distance. Site densities for gsur, early Romano-Libyan farms and cemeteries show firstly that site densities are greatest within one kilometre of junctions, and secondly that the distributions of gsur, early farms and cemeteries are different.

The fact that almost all Romano-Libyan settlement remains are found on or near wadis demonstrates that the water catchment potential of the wadis was the primary factor affecting settlement location. However, catchment potential also varies along wadis, governed both by natural topography and also by the presence of other settlements, with their associated wadi walls and irrigation systems. The prime locations, those with the maximum water catchment potential and the least susceptibility to the effects of upstream settlements, are the wadi junctions and headwaters (Barker 1985; Barker et al. 1996a; 191-225; Gilbertson and Chisholm 1996; Gilbertson et al. 1984).

When we compared the shapes of these distributions, we saw clearly that gsur, early farms and cemeteries show different spatial distributions in relation to wadi junctions and headwaters. Cemeteries and early farms, in particular, seem to deviate from the 'control' distribution of all sites, clustering close to wadi junctions and headwaters. In terms of site density, early Romano-Libyan farms are overrepresented within four kilometres of wadi junctions and headwaters, and they are under-represented above this distance. What stands out in the gasr distribution is that it rises above the average at four-six kilometres distance from junctions and headwaters. This may reasonably be interpreted as evidence in support of the view that some part of gasr-building involved an infilling of the earlier settlement and wadi cultivation pattern, with a general spread outward from the locations most favourable for settlement and agriculture. The modem appearance of near continuous settlement and wadi walls along the wadis is thus deceptive, being the product of progressive extension of the area under cultivation up- and down-stream from the early Romano-Libyan 'estates' that may have been initially separated by some distance one from the next.

Intuitive reasoning might therefore suggest that settlements in the early Roman period would tend to have occupied these locations (Illus. 22.4, Model A). Any growth in the system may have led to the expansion of existing farming populations onto land immediately upstream or downstream, or to exploitation of less favourable areas. Thus, we might envisage a second stage to our initial model, with a much more infilled wadi landscape and the gradual replacement of open farms with gsur (Illus. 22.4, Model B). Indeed, it was noted during fieldwork and preliminary analysis of the data from individual wadis such as Mansur and Khanafes (Barker et al. 1996b: 150-61, 234-4 7) that just such patterns seemed to be occurring. In addition, it was apparent that the preferred locations of cemeteries throughout the Romano-Libyan period may also have been related to these topographical features, possibly reflecting a role in expressing territoriality or status in the landscape (Illus. 22.4, Model A-B). However, an alternative model for the sedentarisation of the wadi systems could be devised, based either on the assumption that all parts of the wadis were equally attractive to settlement, or that the division of the available land was carried out at a single moment by a political authority with sufficient power (such as Rome) to achieve the simultaneous settlement in equal allotments of entire wadi systems (Illus. 22.4, Model C). Such a model would predict that settlement would be relatively evenly spread from early times along the length of individual wadi systems with little tendency for clustering. Equally, in time, the replacement of some of the farms with gsur could lead to the creation of the sort of landscape represented in Illus. 22.4, Model B. Although such a view did not conform to the impression gained from a number of more detailed local studies, it could not be ruled out when considering the overall database. We tested these

The Fezzan Project and the Garamantes

The Garamantes of Fezzan were an important early civilisation of the Sahara (broadly dating to the period 900 BC - AD 500). Survey and excavation by my team have amplified important results obtained in the 1960s and 1970s by Charles Daniels (Daniels 1970; 1971; 1989; Mattingly et al. 1997; 1998; 1999; 2000). At the time of his premature death in 1996, much of Daniels' work was unpublished, but, thanks to major research grants from the Leverhulme Trust and the British Academy, this has been prepared for the press alongside the renewed British fieldwork (Edwards et al. 1999; Mattingly et al. Forthcoming a/b/c). For much of the Roman period the Garamantes thrived on a combination of intensive oasis agriculture (using

247

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

sophisticated irrigation systems) and trade. Their tribal capital Garama took on a strong urban character and their use of material culture and writing marks them out as a Saharan civilisation of considerable magnitude - a polity rather than a tribe (Mattingly 2000a/b ).

15). The bulk of the population appear to have lived in these nucleated settlements (Illus. 22. 7). A recent extension of the cultivated area, and the growth of modem villages that has accompanied it, has particularly affected the preservation of one of the most important and enigmatic classes of monument, the foggaras. These are underground irrigation canals, similar to the Persian qanat or the Arabianfalaj or aflaj, which tapped into an aquifer below the foot of the escarpment and led flowing water out into the oasis proper (Goblot 1979; Mattingly et al. 1997:. 20-22; 1998: 137-42; 1999: 139-42). They are readily identifiable at the surface, where traces survive, from the regularly spaced vertical shafts that were dug to facilitate construction and maintenance of the channels, though they must have added hugely to the labour involved. The shafts can be up to twenty metres deep, gradually diminishing in depth until the channel emerges at the surface, from which point surface channels will have distributed the water into the irrigated plots. The available dating evidence shows that the foggara system was introduced to Fezzan during the Garamantian period and their use probably extended into the early Islamic period. It is clear that these structures were a key to ancient irrigation in the region, though evidently they have been dry for many centuries now. There are many hundreds of these structures visible on the air-photographs, and most were several kilometres in length. The labour involved in their construction and maintenance was on a significant scale. Settlement density, the number and scale of the cemeteries and the foggara systems all combine to highlight the early centuries AD as a period of peak population and oasis cultivation.

The heartlands of the Garamantes lay in the Wadi el-Agial, a sinuous depression c.150 km long by three-five kilomtres broad running broadly west to east (Illus. 22.5). It is sandwiched between a towering sand sea to the north and a cliff-like rock escarpment backed by a barren rock plateau (hamada) to the south. Annual rainfall in this region is negligible - less than ten millimetres on average - but frequently with none whatever for several years at a stretch. It is a very unpromising environment at first sight. However, water could be found in antiquity at a shallow depth below the bottom of the valley, permitting intensive cultivation of a narrow oasis belt in the valley floor. The Garamantian capital Garama (Old Genna) was a longlived urban centre, providing us with an 'urban' sequence going back 2,500 years. This can be extended still further when the nearby proto-urban hillfort called Zinchecra is added to the picture. The primary phase of occupation at Zinchecra was from c.900 - 500 BC. Daniels' excavations here produced evidence not only of the huts and shelters terraced into the hill, but of animal bones and plant remains. The faunal record, perhaps unsurprisingly was dominated by sheep and goat bones, but the botanical record revealed an astonishing picture. From a series of contexts dated by C 14 samples to the first half of the first millennium BC, came a range of agricultural produce (wheat, barley, grapes, the date palm), while all the weed species present were types indicative of a hyper-arid climate much as today. The clear conclusion is that the early Garamantes were already advanced agriculturalists long before they had contact with the Greco-Roman world. They were practising irrigation in a region of negligible rainfall, where subterranean aquifers are the only significant source of water (Daniels 1989: 51; Mattingly 2000b: 171-72; van der Veen 1992). Accompanying this technological and agricultural revolution, Zinchecra marks the earliest stages of settlement nucleation and new forms of social differentiation, which were continued and accelerated when a true urban centre developed after 500 BC at Genna, a few kilomtres away in the centre of the plain. In the Genna region, then, we have an effective urban sequence of nearly 3,000 years length, with the Garamantes being identified as the people responsible for bringing about a series of momentous socio-economic changes.

Garamantian civilisation was thus the result of raised population levels in the northern Sahara following the development of advanced irrigation systems. The concentration of tens of thousands of people in the largest of these oases allowed them to dominate a large expanse of the Sahara - launching military expeditions and trading in equal measure to all points of the compass. They were at times a thorn in the side of the Roman empire, but Classical sources also speak of the Garamantes hunting the troglodytae and 'Ethiopians', a strong hint of slave raiding against neighbouring peoples (Herodotus 4.183; Ptolemy 1.8). Quite apart from the possibility of selling-on such captives north across the Sahara, the intensive irrigated cultivation and the dangerous task of constructing underground irrigation canals (foggaras) could have absorbed large numbers of slaves. Comparisons and Conclusions Three key distinctions can be made between the farming regimes that have been identified by the two surveys. The first concerns the nature of water exploitation and the hydraulic technology developed to facilitate this. In the Predesert, rainfall is very low and irregular, but careful use of water catchment features and wadi farming systems allowed the periodic floods to be exploited to maximum

Systematic fieldwalking since 1997 has revealed that the Garamantian settlement pattern was far denser than previously suspected (Illus. 22.6). There were numerous satellite villages all around Old Genna and at least two other sites that were of true urban character (Daniels 1989: 49; Edwards et al. 1999: 113-14; Mattingly et al. 1997: 13, 19-20; 1998,: 131, 133; 1999: 135-38; 2000: 108-10, 114-

248

David Mattingly: Farming the Desert: two contrasting Libyan case-studies

effect. There were few springs or wells in this system, but numerous rain-fed cisterns. On the other hand, the situation in Fezzan was focused on the exploitation of subterranean fossil water sources, in the absence of appreciable rainfall. The obvious corollary of this is that in the Predesert, the hydraulic technology was all concerned with catching, controlling, directing and storing rain and floodwater, whereas in Fezzan the technology is all directed to getting access to and augmenting the extraction from aquifers.

these elite sites a network of lesser farms and farmsteads. There is nothing in the archaeological record from the UL VS to suggest other than that the minor farms and farmsteads were dependant tenant farms of the larger estates. On the other hand, the Garamantian settlement pattern was dominated by large nucleated sites (towns and villages) and, although there are abundant indications of the existence of a Garamantian elite class, the scale of organisation of land and social space suggests state involvement. For instance, the massive investment in time and labour in the construction of the foggara of the Garamantian heartlands is suggestive of state organisation and an abundance of cheap and disposable labour (probably slaves) to dig them.

The political dimension of the two farming regimes also highlights differences between the regions. The remarkable expansion of sedentary farming in the pre-desert zone of the later first century AD was at the very least an officially condoned and encouraged activity. But it was clearly not over-regulated. An important point to draw from this study is that power networks in Roman imperial landscapes were not simply exploitative and repressive, they also empowered some of the subjects and provided opportunities for aggrandisement and growth (Mattingly 1997a). Initially, farms were established on a wide spread basis, rather than intensively along just a few wadis. The first settlements showed a preference for certain types of location over others - in part reflecting their perception of the hydraulic system, perhaps also a choice to live near headwaters and junctions that lay on route-ways through the region. The technology of floodwater farming can be shown to have a pre-Roman pedigree in North Africa broadly we might think of it as representative of the arid Maghreb - but its greatest diffusion is linked to the period of Roman dominance (Shaw 1984; 1991).

Although the environmental constraints of water availability and cultivable soils had an important bearing on the location of sedentary populations in both the Predesert and Fezzan, it is thus clear that the evolution of desert settlement in Libya was also profoundly influenced by social and political factors. The two scenarios of 'farming the desert' outlined here reflect in part fundamental differences in hydrology, but they also indicate the very different trajectories taken by people and territory within and outside the Roman empire. The creation of the Roman frontier zone played a crucial role in defining the shape of rural development in the Predesert zone, whilst the remote heartlands of the Garamantes were undoubtedly affected by contact and trade with the Roman empire (Mattingly 1992). An ostracon found at the Roman fort at Bu Njem on the south-eastern edge of the UL VS area, is a message from an outpost reporting the approach of a Garamantian donkey caravan bearing a consignment of barley (Mattingly 1995: 74-75). The Roman frontier in Africa was truly a meeting place of the Mediterranean and Saharan worlds.

The development of Garamantian agriculture pre-dated contact with Romans (and Greeks and Phoenicians for that matter) and must be understood as an internal African adaptation to the onset of desert conditions. It is possible, however, that the intensification of production and the evident demographic peak in the early centuries AD was in part a product of contact with the Roman world. It is in this period that we have the clearest evidence for the existence of a Garamantian state, able to dominate a large part of the central Sahara and certainly conducting slaving raids against the 'Ethiopian' populations of Niger and Chad. Garamantian agriculture was well developed long before that, with technology that had probably travelled west across the Sahara from the oases in the Western Desert of Egypt, close to the Nile Valley, where foggara systems are also known. If floodwater farming is a Maghrebian system, the foggara irrigation is essentially Saharan.

ACKNOWLEDGEMENTS This paper was written during my tenure of a British Academy Research Readership, focused on the Garamantes of Fezzan. The Leverhulme Trust have also provided substantial support for the preparation of Charles Daniels' research for publication. Barri Jones worked in Fezzan with Daniels in the 1960s and first stimulated my interest in the Garamantes. He also took me to Africa and opened my eyes to its archaeology. This article (and the research and detailed publications that lie behind it) is a token of my gratitude.

BIBLIOGRAPHY

The final point to make is that both farming systems were operated by native Libyan people, though organised at different scales. The dispersed settlement pattern of the Predesert farming system seems to have been based on elite estates controlling stretches of wadi, rather than on individual small free-holdings existing side by side, or territory organised on a tribal basis. The occupants of the elite sites were undoubtedly Libyans from the region (Brogan and Smith 1984; Mattingly 1995: 162-68, 205-09; 1999), but there are strong reasons for associating with

Argoud, G., Marangou, L. Panayotopoulos, V. and VillainGandossi, C. (eds), 1992. L'eau et les hommes en Mediterranee et en Mer Noire dans l'antiquite de l'epoque Mycenienne au regne de Justinien. Athens: Centre National de Recherches Sociales. Barker, G.W., 1985. The UNESCO Libyan Valleys Survey: developing methodologies for investigating

249

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

ancient floodwater farming. In Buck and Mattingly (eds.): 291-307. Barker, G.W. and Gilbertson, D.D. (eds), 2000. The Archaeology of Drylands. Living at the margin, London: Routledge Barker, G.W., Gilbertson, D.D., Jones, G.D.B. and Welsby, D.A., 1991. ULVS XXIII: the 1989 season. Libyan Studies 22: 31-60. Barker, G.W.W., Gilbertson, D.D., Jones, G.D.B. and Mattingly, D.J., 1996a. Farming the Desert The UNESCO Libyan Valleys Archaeological Survey. Volume 1 Synthesis. (principal editor, G. Barker), Paris/London: UNESCO, Society for Libyan Studies. Barker, G.W.W., Gilbertson, D.D., Jones, G.D.B. and Mattingly, D.J., 1996b. Farming the Desert The UNESCO Libyan Valleys Archaeological Survey. Volume 2, Gazetteer and Pottery (principal editor, D.J. Mattingly), Paris/London: UNESCO, Society for Libyan Studies. Birebent, J., 1962., Aquae Romanae. Recherches d'hydraulique romaine dans l'est Algerien. Alger: Service des Antiquites de l'Algerie. Brogan, 0. and Smith, D.J., 1984. Ghirza: a RomanoLibyan Settlement in Tripolitania. Tripoli: Libyan Antiquities Series 1. Buck, D.J. and Mattingly, D.J. (eds.), 1985. Town and Country in Roman Tripolitania. Papers in honour of Olwen Hackett. Oxford: British Archaeological Reports international series 274. Daniels, C. M., 1970. The Garamantes of Southern Libya. London: Oleander. Daniels, C. M., 1971. The Garamantes of Fezzan. In Gadallah (ed.): 261-85. Daniels, C.M., 1989. Excavation and fieldwork amongst the Garamantes. Libyan Studies: 20: 45-61. Edwards, D.N., Hawthorne, J., Dore, J.N. and Mattingly, D.J., 1999. The Garamantes of Fezzan revisited: publishing the C. M. Daniels archive. Libyan Studies 30: 109-27. Ferguson, B. and Whitehead, N. (eds), 1992. War in the Tribal Zone. Expanding States and Indigenous Waifare. Santa Fe: School of American Research Advanced Seminar Series. Flower, C. and Mattingly, D.J., 1995. ULVS XXVII: the mapping and spatial analysis of the Libyan Valleys data using GIS. Libyan Studies 26: 49-78. Gadallah, F.F. (ed.), 1971. Libya in History. Proceedings of a conference held at the Faculty of Arts, University of Libya 1968. Benghazi: GarYunis University. Gilbertson, D.D. and Chisholm, N.W.T., 1996. ULVS XXVIII: manipulating the desert environment: ancient walls, floodwater farming and territoriality in the Tripolitanian pre-desert of Libya. Libyan Studies 27: 17-52. Gilbertson, D.D., Hayes, P.P., Barker, G.W.W. and Hunt, C.O., 1984. The UNESCO Libyan Valleys Survey VII: an interim classification and functional analysis of ancient wall technology and land use. Libyan Studies 15: 45-70.

Goblot, H., 1979. Les qanats: une technique d'acquisition de l'eau. Paris and New York: Industrie et artisanat, 9. Grahame, M., 1998. Rome without Romanization: cultural change in the pre-desert of Tripolitania (first-third centuries AD). Oxford Journal of Archaeology 17.1: 93-111. Hodge, A.T. (ed.), 1991. Future Currents in Aqueduct Studies. Leeds: Leeds University Press. Jones, B. 1985. The development of the settlement survey. In Buck and Mattingly (eds.): 263-89. Lancel, S., (ed.), 1999. Numismatique, langues, ecritures et arts du livre, specificite des arts figures. VIie colloque sur l'histoire et archeologie de l'Afrique du nord. Paris: Comite des Travaux Historiques et Scientifiques. Mattingly, D.J., 1992. War and peace in Roman Africa. Some observations and models of State/Tribe interaction. In Ferguson and Whitehead (eds.): 31-60. Mattingly, D.J., 1995. Tripolitania. London: Batsford. Mattingly, D.J., 1997a. Africa, a landscape ofopportunity? In Mattingly 1997b: 115-38. Mattingly D.J. (ed.), 1997b. Dialogues in Imperialism. Power, discourse and discrepant experience in the Roman empire. Portsmouth, RI: Journal of Roman Archaeology Supplementary volume 23. Mattingly, D.J., 1998. Landscapes of imperialism in Roman Tripolitania. L'Africa romana 12 (1998): 16379 Mattingly, D.J., 1999. The art of the unexpected: Ghirza in the Libyan pre-desert. In Lancel (ed): 383-405. Mattingly, D.J., 2000a. Making the desert bloom. The Garamantian capital and its underground water system. Archaeology Odyssey March/April 2000: 30-37. Mattingly, D.J., 2000b. Twelve thousand years of human adaptation in Fezzan (Libyan Sahara). In Barker and Gilbertson (eds.): 160-79. Mattingly, D.J. and Hitchner, R.B., 1995. Roman Africa: an archaeological review. Journal of Roman Studies 85: 165-213. Mattingly, D.J., al-Mashai, M., Balcombe, P., Chapman, S., Coddington, H., Davison, J., Kenyon, D., Wilson, A.I. and Witcher, R., 1997. The Fezzan Project 1997: methodologies and results of the first season. Libyan Studies 28: 11-25. Mattingly, D.J., al-Mashai, M., Aburgheba, H., Balcombe, P., Eastaugh, E., Gillings, M., Leone, A., McLaren, S., Owen, P., Pelling, R., Reynolds, T., Stirling, L., Thomas, D., Watson, D., Wilson, A.I. and White, K., 1998. The Fezzan Project 1998: preliminary report on the second season of work. Libyan Studies 29: 115--44. Mattingly, D.J., al-Mashai, M., Balcombe, P., Drake, N., Knight, S., McLaren, S., Pelling, R. Reynolds, T. Thomas, D., Wilson, A. and White, K., 1999. The Fezzan Project 1999: preliminary report on the third season of work Libyan Studies 30: 129-45. Mattingly, D.J., al-Mashai, M., Balcombe, P., Barnett, T., Brooks, N., Cole, F., Dore, J., Drake, N., Edwards, D., Hawthorne, J., Helm, R., Knight, S., Leone, A., McLaren, S., Pelling, R., Preston, J., Reynolds, T.

250

David Mattingly: Farming the Desert: two contrasting Libyan case-studies

Townsend, A., Wilson, A. and White, K., 2000. The Fezzan Project 2000: preliminary report on the fourth season of work. Libyan Studies 31: 103-120. Mattingly, D.J. Daniels, C.M. Dore, J.N., Edwards, D. and Hawthorne, J., forthcoming a. The Archaeology of Fezzan. Volume 1, Synthesis. London. Mattingly, D.J. Daniels, C.M. Dore, J.N., Edwards, D. and Hawthorne, J., forthcoming b. The Archaeology of Fezzan. Volume 2, Gazetteer, Pottery and other finds. London. Mattingly, D.J. Daniels, C.M. Dore, J.N., Edwards, D. and Hawthorne, J., forthcoming c. The Archaeology of Fezzan. Volume 3, Excavations carried out by C.M Daniels. London. Shaw, B.D., 1984. Water and society in the ancient Maghrib: technology, property and development. Antiquites africaines 20: 121-73. Shaw, B.D., 1991. The noblest monuments and the smallest things: wells, walls and aqueducts in the making of Roman Africa. In Hodge (ed.): 63-91. Slim, H. 1992. Maitrise de l'eau en Tunisie. In Argoud et al. (eds.): 513-32. Trousset, P. 1986. Les oasis presahariennes dans l'Antiquite: partage de l'eau et division du temps. Antiquites africaines 22: 161-91. Trousset, P. 1987. De la montagne au desert: limes et maitrise de l'eau. Revue de [ 'Occident Musulmane et de la Mediterranee 41-42: 90-115. Van der Veen, M., 1992. Garamantian agriculture: the plant remains from Zinchecra, Fezzan. Libyan Studies 23: 7-39.

251

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

OASES CENTRES OF THE NORTHERN SAHA}{A

SAND

HIGH 0

SEA

RELIEF

,oo

100

22.1 The Libyan Desert, showing the location of the ULVS and the FP survey areas.

'~',,.,:;.,.,31t.-'(""'~-C:-j,'')

.

-

""'1,ir •

·:;--\}¥~,-;~{

'·-~

,_

'o·•,

I

.-:...........

··~· ,,.,,.,.,,.

//K,-

·?::"

":\:\,i,:\,,,, ' '~i---,~~, ........ ~'lJ

/

t "

ri All ULVSSites

N

t

50

Km

22.2 The ULVS survey area: distribution of all sites recorded in the Wadi Sofeggin/ZemZem region.

252

David Mattingly: Farming the Desert: two contrasting Libyan case-studies

.i...-.i,

-

..;.--t

~s

a~

§

26~® .

0

?'::~,., 17 _,./)

'.t

2029

K,h\4

·:···;·:.:,_ 24 .'

:

·....···

N

I --

200m

0

d

100,v~\·•. ·.·. ·.

\,•Khl9

· Kh40

.'f

··@ .•·

.

/ f/:y

@ .•· ./~o

tc.h4l

·\~\

~::~~:··

': ..__ i;::::::=::=:=j

-0···•. ' / ~b,;·• ..t:J.• //i~·i ·.···•ontifical Institute of Christian Archaeology. Downes, J., and Pollard, T. (eds.,), 1999. The Loved Body's Corruption: Archaeological contributions to the Study of Human Mortality. Glasgow: Cruithne Press. Drinkwater, J., and Elton, H. (eds.), 1992. Fifth-Century Gaul: a Crisis of Identity? Cambridge: Cambridge University Press. Duval, N., 1977. La 'Basilique de Zana' (Diana Veteranorum): une nouvelle eglise a deux absides ou on monument a auges?, Melanges de !'Ecole Francaise de Rome: Antiquite, 89: 847-73. Duval, N. (ed), 1989. Actes du XI Congres Internat. D 'Archeologie Chretienne. Paris/Rome: Ecole

ACKNOWLEDGEMENTS

This paper is a revised, enlarged, and updated version of one first given as the Frend Lecture to the Society of Antiquaries of London in October 1994, and subsequently, in April 2000, as one of the Millennial Lectures organised by the British Archaeological Association. As the text and bibliography make clear, I have been much helped in this general endeavour by Kenneth Painter, who kindly read earlier versions of this paper. I have also received help and encouragement from a large number of other distinguished experts in the field, not only British, but also French, German, Italian, and Swiss: I must single out for special thanks Professor Noel Duval and his colleagues in Paris. At Manchester, Dr John Peter Wild has provided wise advice and generous hospitality at various times over the past thirty years or so. Nick Higham offered some judicious comments on the text.

BIBLIOGRAHY

Archibald, M., Brown M. and Webster, L., 1997. Heirs of Rome: the shaping of Britain AD400-900. In Webster and Brown (eds.), 208-12. Bassett, S., 1989. Churches in Worcester before and after the Conversion of the Anglo-Saxons. Antiquaries Journal 69: 225-56. Bidwell, P. (ed.), 1999. Hadrian's Wall, 1989-99. Carlisle: Cumberland and Westmorland Antiquarian and Archaeological Society/Society of Antiquaries of Newcastle upon Tyne Birch, D., 1998. Pilgrimage to Rome in the Middle Ages. Woodbridge: Boydell Press. Blair, W.J., 1992. Anglo-Saxon Minsters: a Topographical Review. In Blair and Sharpe (eds.): 226-66. Blair, W.J. and Pyrah, C. (eds), 1996. Church Archaeology: Research Directions for the Future. York: Council for British Archaeology Research Report 104. Blair, W.J. and Sharpe, R. (eds.), 1992. Pastoral care before the Parish. Leicester: Leicester University Press Bonnet, C. 1993. Les Fouilles de l'Ancien Groupe episcopal de Geneve (1976-93). Geneva.

325

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Howlett, D.R. 1997. British Books in Biblical Style: The Competence and Craftsmanship of Cambra-Latin Compositions. Dublin: Four Courts Press. James, E., 1993. Review of Duval et al. (eds.), 1991. Medieval Archaeology 37: 340-1. Johns, C.M., 1981. The Risley Park Silver Lanx: a Lost Antiquity from Roman Britain, Antiquaries Journal 61: 53-72. Johns, C.M. and Painter, K., 1995. The Risley Park Lanx Re-discovered: Bauge, Buch, or Britain?, in Orbis Romanus Christianusque ab Diocletiani aetate usque ad Heraclium. (Festschrift N. Duval). Paris: 175-89. Jones, G.D.B., and Mattingly, D.J., 1990. An Atlas of Roman Britain. Oxford: Blackwell. Jones, M.J., 1994. St. Paul-in-the-Bail, Lincoln: Britain in Europe? In Painter (ed.): 325-47. Jones, M.J., 1996. Early Church Groups in the British Isles. Antiquite Tardive 4: 211-15. Jones, M.J., 1998. Recent Discoveries in Britain and Ireland. In Cambi and Marin (eds.), vol. III: 395-420. Jones, M.J., forthcoming. Recent Research in Britain, in Pillinger (ed.). Jones, M.J. and Vince, A.G., forthcoming. Lincoln: an Archaeological Synthesis. Karivieri, A, 1994. The So-called Library of Hadrian and the Tetraconch Church in Athens. In Castren (ed.): 89-114. Kemick, P., 1986. Excavations at Sabratha 1948-51. London: Journal of Roman Studies Monograph 2. Knight, J. K., 1999. The End of Antiquity: Archaeology, Society and Religion AD235-700. Stroud: Tempus. Le Maho, J., 1993. Le Groupe episcopal de Rouen du IV au X Siecle. In Stratford (ed.): 20-30. Loseby, S., 1992. Bishops and cathedrals: order and diversity in the fifth-century urban landscape of southern Gaul. In Drinkwater and Elton (eds.): 14455. Macmullen, R., 1984. Christianising the Roman Empire. New Haven and London: Yale University Press. Mango, C., 1995. The Pilgrim's Motivation. In Dassmann and Engemann (eds.): If£ Markus, R.A., 1990. The End of Ancient Christianity. Cambridge: Cambridge University Press. Matejcic, I., and Chevalier, P., 1998. Nouvelle du complexe episcopal "pre-Euphrasien" de Poree. Antiquite Tardive 6: 355-65. Mawer, C.F., 1995. Evidence for Christianity in Roman Britain; the Small Finds. Oxford: British Archaeological Reports, British Series, 243. Menis, G. C., 1996. La liturgia battesimale ad Aquileia nel complesso episcopale del IV secolo. Antiquite Tardive 4: 61-74. Mitchell, S., 1985. Archaeology in Asia Minor 1979-84. Journal of Hellenic Studies 105: 70-105. Niblett, R., 1999. The Excavation of a Ceremonial Site at Folly Lane, Verulamium. London: Britannia Monograph Series, 14.

Frarn;aise de Rome/Pontifical Institute of Christian Archaeology. Duval, N., Fontaine, J., Fevrier, P-A., Picard., J-C. and Barruol, G. (eds.), 1991. Naissance des Arts Chretiens. Atlas des Monuments Paleochretiens de la France. Paris: Imprimerie Nationale. Edwards, N., 1996. Identifying the Archaeology of the Early Church in Wales and Cornwall. In Blair and Pyrah (eds.): 49-62. Esmonde-Cleary, S., Jones, M. and Wood, J., 1998. The Late Roman Defences at St. Bertrand-de-Comminges (Haute Garonne): Interim Report, Journal of Roman Archaeology 11: 344-54. Fevrier, P-A., 1989. Une Archeologie Chretienne pour 1986. In Duval (ed.):lxxx-xcix. Fixot, M., Guyon, J., Pelletier, J-P. and Rivet, L., 1985. Les Fouilles de la Cour de l'Archeveche. Aix-enProvence: Ville d' Aix en Provence/Direction des Antiquites Historiques Provence-Alpes-Cote d'Azur. Frend, W.H.C., 1992. Pagans, Christians, and the Barbarian Conspiracy of AD367 in Roman Britain. Britannia 23: 121-32. Frend, W.H.C., 1994. The Archaeology of the Early Church. In Painter (ed.): 1-16. Frend, W.H.C., 1996. The Archaeology of Early Christianity: a History. London: Geoffrey Chapman. Frend, W.H.C., 1998. The failure of Christianity in Roman Britain: Some Recent Evidence. In Cambi and Marin (eds.), vol.III: 287-300. Glaser, F., 1996. Eglises doubles ou familles d'eglises: les cinq eglises du Hemmaberg (Mont Sainte Hemma). Antiquite Tardive 4: 142-48. Guyon, J., 1991. Fragments d'Archeologie Chretienne. St. Bertrand-de Comminges: Conseil General de la HauteGaronne. Halsall, G., 1995. Early Medieval Cemeteries: an Introduction to Burial Archaeology in the Post-Roman West. Glasgow: Cruithne Press. Hamerow, H., and MacGregor, A (eds.), forthcoming. Essays on the Archaeology of Early Medieval Britain in Honour of Rosemary Cramp. Hawkes, J. and Mills, S. (eds.), 1999. Northumbria 's Golden Age. Stroud: Sutton Publishing. Higham, N. J., 1997 The Convert Kings: Power and religious affiliation in early Anglo-Saxon England. Manchester: Manchester University Press. Hill, P., 1997. Whithorn and St. Ninian: Excavations of a Monastic town. Stroud: Sutton Publishing. Hines, J., 1997. Religion: the Limits of Knowledge. In Hines (ed.): 375-410. Hines, J. (ed.). 1997. The Anglo-Saxons from the Migration Period to the Eighth Century. Woodbridge: Boydell & Brewer. Hopkins, K. 1999. A World Full of Gods: Pagans, Jews, and Christians in the Roman Empire. London: W eidenfeld and Nicholson. Howlett, D.R. 1995. The Celtic Latin Tradition of Biblical Style. Dublin: Four Courts Press.

326

Michael J. Jones: Early Christian Archaeology in Europe: Some Recent Research

O'Carragain, E., 1995. Rome pilgrimage, Roman liturgy, and the Ruthwell Cross. In Dassmann and Engemann (eds.): 630-9. O'Sullivan, D., forthcoming. Space, Silence, and Shortage on Lindisfame. In Hamerow and MacGregor (eds.). O'Sullivan, D., and Young, R., 1995. Lindisfarne: Holy Island. London: English Heritage/Batsford. Painter, K., 1994. Review of Duval et al., 1991. Bonner Jahrbucher 194: 647-50. Painter, K., (ed.), 1994. 'Churches Built in Ancient Times': Recent Studies in Early Christian Archaeology. London: Society of Antiquaries of London/ Accordia Centre. Painter, K., 1997. Silver Hoards from Britain in their Late Roman Context. Antiquite Tardive 5: 93-110. Painter, K., 1999. The Water Newton Silver Hoard: Votive or Liturgical? Journal of the British Archaeological Association. 152: 1-23. Parker Pearson, M., 1999. The Archaeology of Death and Burial. Stroud; Sutton Publishing. Pillinger, R. (ed.), forthcoming. Acta XIV Congressus Archaeol. Christianae. Vienna/Rome. Potter, T.W., 1995. Towns in Late Antiquity: fol Caesarea and its Context. Sheffield: University of Sheffield. Rahtz, P.A., and Watts, L., 1986. The Archaeologist on the Road to Lourdes and Santiago de Compostela. In Butler and Morris (eds.): 51-74. Reece, R., 1999. The Later Roman Empire: an Archaeology, AD150-600. London: Tempus Publishing. Sawyer, P.H. 1998. Anglo-Saxon Lincolnshire. Lincoln: History of Lincolnshire Committee, History of Lincolnshire, vol.III. Sodini, J.-P., and Kolokotsas, K., 1984. Aliki. II: la basilique double. Athens: Ecole Fran9aise d' Athenes. Stocker, D., and Went, D., 1995. The Evidence for a PreViking Church adjacent to the Anglo-Saxon Barrow at Taplow, Bucks. Archaeological Journal 152: 441-54 Stopford, J., 1994. Social Approaches to the Archaeology of Christian Pilgrimage, World Archaeology 26. l: 5772. Stratford, J. (ed.), 1993. Medieval Art, Architecture and Archaeology at Rauen. London: British Archaeological Association. Tarlow, S., 1999. Bereavement and Commemoration: An Archaeology of Mortality. Oxford: Blackwell. Thomas, AC., 1998. Christian Celts: Messages and Images. London: Tempus Publishing. Urbancyk, P., 1998. Christianisation of Early Medieval Societies: an Anthropological Perspective. In Crawford (ed.): 129-33. Ward-Perkins, B., 1996. Urban Continuity? In Christie and Loseby (eds.): 4-17. Watts, D., 1998. Religion in Late Roman Britain: Forces of Change. London: Routledge. Webster, L., and Brown, M. (eds.), 1997. The Transformation of the Roman World, AD400-900. London: British Museum. Weidemann, M., 1995. ltinerare des Westlichen Raumes. In Dassmann and Engemann (eds.): 389-451.

Wilmott, A., 1997. Birdoswald: Excavations of a Roman fort on Hadrian's Wall and its Successor Settlements 1987-92. London: English Heritage Archaeological Reports 14. Wilmott, A., and Wilson, P.R., (eds.), 2000. The Late Roman Transition in the North. British Archaeological Reports, British Series 299.

327

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

31. 1 A weekend foray from Manchester; investigating the Roman road at Delamere Fore st, Cheshire, 1970, watched by enthusiastic cows. John Williams (left) and John Manley (centre) try to look unimpressed as Barri Jones holds forth.

31.2 The sixth-century baptistery at Albenga, Liguria, N Italy; external view.

328

Michael J. Jones: Early Christian Archaeology in Europe: Some Recent Research

31. 3 The sixth century baptistery at Albenga, Liguria, N Italy; internal view.

31.4 The sixth-century ivory seat (cathedra) of Bishop Maximian (Ravenna cathedral museum).

LJUBLJANA

• Split - Salona Roma Ravenna Roma Aix-en-Provence Ravenna Trier Barcelona Roma Thessa/oniki Lyon - Grenoble - Geneve - Aosta Bonn

SPLIT

31.5

. POREC

1894 1900 1932 1938 1954 1962 1965 1969 .1975 1980 1986 1991

1994

vrnrn.1,

ANCmlA.

e

PESCAAA

~RO:

e

ORTO\IA

e

Map from the programme of the 1994 congress at Split and Poree, Croatia, listing the venues of earlier congresses.

329

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

31.6

Aerial view of the excavated church (fifth century and later), in the lower city at St. Bertrand-de-Comminges, Haute-Garonne, France (from Guyon 1991).

N

=

-----------------------

------~--~ , , I' I' I

fl

D D

D D i ,:_-_-_- _::::: _ --- :_ ___ - - -- - - -\I

D C

n

---= ------=-----

l

:

D

It

l: 11

1/ 11

: :::.:::::::~!

~----·------------JC-

____ :JIL_J

I~====~--

cl

[J

-

___J

D



D

D

'.]

D

D

D

D

D

D

D

D

D

1-,_-_:-_-:_-_--r-----: :~ : •

D D

D D CJ

I 31.7

,:

: l

:

t

: : J:

i::: iiI: i: i: ::

: :

:'-----J~--·

11

::

1:

:r--, 1-----1:

::

II

I I

ti

I I

::

: :

::

D □

i:_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-_:-.::.::.:: ___ ;] ii !i ii □ :: :

1

: :

-------------------------------------:-1-t.~~::::::,:

11

::

11 t 1

1 :

::



11

0

1,

lL______________________________________

,___------ - - . - --- - ---- ----- -- --------0

------------': ------ -------- - - - - -- -- -

--===---==:::i

20m

Plan of the earliest churches in the forum at Lincoln. 330

D

Michael J. Jones: Early Christian Archaeology in Europe: Some Recent Research

Cathedrale nord

Cathedrale sud





Troisiemecathedrale

31.8

31.9

The sequence of the episcopal group at Geneva (from Bonnet 1993).

The more southerly of the double churches of the episcopal group at Salona, near Split; the baptistery lies to the left, between the two churches.

331

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

31.10 The early church at Cimiez, Nice, involving an adaptation of part of a baths-building.

31.11 The reconstruction of the cella memoriae, excavated beneath Bonn cathedral, in the Rheinisches Landesmuseum, Bonn.

332

Michael J. Jones: Early Christian Archaeology in Europe: Some Recent Research

31.12 Part of the mosaic in the main apse of the sixth century basilica at Poree, showing (centre) Bishop Eufrasius with a model of his cathedral.

333

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

334

32 Bacchus triumphs on the North Downs: the Polesden Lacey sarcophagus Tom Rasmussen

Barri was expert in so many areas: among them, the archaeology of Britain and of Italy; Africa and especially Roman North Africa; and the Romans generally in the further flung parts of their empire. I think the item under discussion here would have intrigued him and given him pause for thought, in the way that it combines, even if only obliquely, a number of these interests.

hole at the bottom of the left side towards the back comer, under the griffin, which seems to have been plugged with darker stone. The marble of the Polesden sarcophagus has not been scientifically analysed, but the carving was almost certainly done at Rome. Recent studies have shown that such sarcophagi were often transported to Rome in blocked out form from quarries far afield - in the case of the Baltimore sarcophagus (see below), from Thasos in the northern Aegean (Ward-Perkins 1992).

The present setting Polesden Lacey is a stately home high on the North Downs near Guildford, Surrey, now owned and maintained by the National Trust. I visited it for the first time in 1991, when I came across a Roman sarcophagus with reliefs on three sides, which was then outside the house under the west portico, where I photographed it (Illus. 32.1 ). It is unpublished, acquired too late for the principal publications of ancient marbles in English country houses that were produced in the nineteenth century. It was the possession of Mrs Margaret Greville (nee McEwan), brewing heiress and collector both of works of art and also of royalty at her dinner tables, who lived at Polesden as her country residence from 1906 until her death in 1942. In 1995 the sarcophagus was moved into the interior of the house to the Picture Corridor, where archive photographs show it in position in Mrs Greville's time. After her death most of Margaret Grevilles's personal papers were destroyed on her orders. So, why and how the sarcophagus was acquired, to augment a collection which is best known for its Dutch seventeenth-century paintings and Italian maiolica, are unanswerable questions. Mrs Greville bought at various London auction houses, and may also have been in touch with the (wheeler-)dealer Joseph Duveen, if only indirectly (Rowell 1999: 62).

The relief on the front (Illuss. 32.2, 32.3, 32.4), showing the Triumph of Bacchus/Dionysus, is closely similar in composition to that of a sarcophagus in the Walters Art Gallery, Baltimore, which has been been described in detail (Lehmann-Hartleben and Olsen 1942: 12-14, 26-33; Matz 1968: 231-3). The description here can therefore be brief (left to right). Dionysus stands in a chariot (given little definition) drawn by two panthers ridden by two male child outriders. Behind him is a winged Victory, in front of her a dancing maenad. Behind the panthers is an elephant ridden by a satyr. Beyond the laurel (?) tree is a woman holding a bird in a nest(?), then a cloven-footed Pan with a curved stick, and behind him the head of another figure in profile. In front: two elephants with riders; on the rear of the larger one a seated bearded male figure facing frontally, beside him a large vessel on its side. Between the front legs of the elephants runs a small quadruped. Beyond the oak tree (with acorns) a maenad holds up a tambourine; to the front of her an elderly silenus (?) clutching a thickstemmed thyrsus (?). Then a giraffe and a lion, its face close to a basket (liknon) from which a snake emerges. A male figure bends over the lion, behind him a standing female. The procession is headed by two women, one frontal, the other in back view; behind them another oak tree in which sit two birds.

The sarcophagus The dimensions are (in cm.): length 217, height 66, width 77.5. The back of the sarcophagus has in the past come away from the rest, and there are repairs at either comer with large L-shaped brackets which are deeply embedded. There is also a bracket repair to the centre of the back where it has split. The front panel has worn right through in three places. The largest hole is above the left elephant's forehead, there is a tiny one below the same animal's head, and a semi-circular one above the maenad's tambourine. In each case the damage originates from the interior, where around the holes fresher whiter marble has been exposed.

The side-panels (which differ from those of the Baltimore sarcophagus) show the same single motif: a pantherheaded griffin facing towards the front (Illus. 32.5). Iconography and the Baltimore sarcophagus The Baltimore sarcophagus, a far more careful piece and complete with its lid, displays what must rank as one of the most sustained programmes of Dionysiac imagery in Roman sculpture (Illus. 32.6). At the front comers of the lid are two satyr masks; between them a long relief showing: the death of Dionysus' mother Semele; the birth of Dionysus from Zeus's thigh; the baby Dionysus nursed by nymphs. The side panels of the container (one of them unfinished) show heraldically confronted griffins with their front paws on a ram's head.

There are various other slots, grooves and holes along the top edge of the sarcophagus, some, but not all, probably connected with the fixing of a lid. In the post-Roman era the sarcophagus may have been used for a time as a waterbasin; suggesting this is what looks like a roughly circular 335

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

The Baltimore front panel displays three human figures which the Polesden omits: a grown male between the two child outriders, and a profile head between the elephant group and the maenad with the tambourine while behind the frontal bearded seated figure on elephant-back, who here wears a close-fitting cap on the top of his head, is the back of a second seated figure with head bowed, all in very low relief. There are also additional snakes. One crawls out of a rock in front of the panthers; emerging from a hollow in the trunk of the right-hand oak, another snake nips the tail of a lizard climbing higher up the trunk.

lion, as also panthers/leopards, occur(red) well as in Asia.

elsewhere as

Matz (1968: 233), keen to stress the eastern connections, describes the prisoners on elephant-back as Indians, as too the boy outriders on the panthers. But the latter have tightly curled hair and negroid facial features (Snowden 1970: 149; 1986: no. 41); at any rate they do at Baltimore, though at Polesden they look more Caucasian. Here at least there is an African element in the frieze; and there is another near the head of the procession - the giraffe, to which we will have to return.

There are other differences too on the Baltimore sarcophagus. The yoke of the panthers pulling Dionysus has the form of two dolphins, their tails intertwined; at Polesden this is translated into something much plainer and more cumbersome and the zoomorphism is lost. The woman by the laurel tree (or quince: Matz) holds an incense-burner (thymiaterion) rather than, at Polesden, a bird in a nest. At the front of the parade there seems to be a narrative within the narrative: the man bent over the lion looks to be about to strike the animal with a club, and is restrained by the figure second from the right (here male). The liknon by the lion is lidded, and the birds on the oak tree are here eagles or eaglets with wings outstretched.

Death and the mysteries of Dionysus

The Baltimore sarcophagus is of exceptional size: 98.5 cm wide as opposed to the more normal 77.5 cm of the Polesden. Perhaps it was for the burial of a married couple (Lehmann-Hartleben and Olsen 1942: 14). It is one often sarcophagi discovered in 1885 in two tombs by the Via Salaria at Rome (Notizie degli Scavi 1885: 42-45, 74-77). Five of these have mythological scenes on the fronts, four of them with Dionysiac narratives of various kinds: childhood of the god, Ariadne, thiasos, as well as the triumph (three are in Baltimore, the thiasos in Rome). It is clear that the Baltimore Triumph sarcophagus, with its exceptionally detailed portrayal of Dionsyiac myth and cult, was no ordinary commission. Lehmann-Hartleben and Olsen make out a case for it having been made for a patron who was initiated into the mysteries of Dionysus, in particular to that branch of them that was presided over by Sabazios, in origin a Thracian-Anatolian god who came to be identified with Dionysus himself The telling feature that connects with this god is the motif of the snake (Burkert 1993: 265), not only issuing from the liknon or cista - an iconographic detail to be found in contexts of most other ancient mystery cults - but occurring in all four times: three on the main relief, and one on the extreme right of the lid where it creeps from a hole in the rock closely observed by a satyr/Pan. Snakes had special prominence in initiation rituals of the Sabazios cult, when they were drawn through the draperies of the participating women as if they were mating with them (Burkert 1987: 106).

It is held by just about all who write about the Baltimore

sarcophagus that it is the Indian triumph of Dionysus that is depicted. It is a story that became popular after Alexander's Indian campaigns, which were compared by historians such as Arrian to those of the god. Certainly this interpretation would make best sense of the seated figure (Polesden) or figures (Baltimore), who sit with their arms tied behind them on the back of the elephant, for in mythology prisoners were taken on Dionysus' campaign, together with booty (represented here by the krater and tusk strapped to the elephant's back). Yet Scullard (1974: plate 20), while recognising the Indian campaign, captions his illustration of the central part of the Baltimore parade as 'Dionysos ... on an elephant', presumably referring to the bearded frontally facing seated figure. The god is, however, the figure drawn by the panthers: youthful and wearing an effeminate sleeved chiton, as so often in Roman art, and clutching his thyrsus. What else about the iconography might have Indian associations? I have so far called the animals pulling Dionysus 'panthers', as do McCann (1978: 89) and others, but Matz ( 1968: 231) takes them for tigers, as he does similar draught animals on other sarcophagi; this is how Roman poets, including Ovid and Vergil, also describe the god's chariot. The animals would then have a more specifically Asiatic ring about them, though only the missing painted surface could prove the matter one way or the other. Dionysus was associated with both beasts from early on, as Hellenistic mosaics from Delos show (Dunbabin 1999: 32). Then there are the elephants, which look to be more like the smaller-eared Indian variety. The

However, not all Dionysiac sarcophagi need to have been made for discerning initiates. The Polesden sarcophagus, though dependent on the Baltimore for iconography, is a far more modest piece, glossing over many details, misunderstanding others (see above) - such as, for example, the dolphin-yoke (another animal connected with the mythology of Dionysus). On such examples it was enough that Dionysus could simply be seen triumphing, and one did not have to be an initiate to know the simple message imparted. Dionysus is death-defying, whether, at the start of life, blasted by lightning, or (according to the Orphics) tom into tiny shreds. He is death-defying in another sense, too, in that he goes down into Hades and comes up again escorting his mother. Guettel Cole (1993:

336

Tom Rasmussen: Bacchus triumphs on the North Downs: the Polesden Lacey sarcophagus

280) reminds us that Plutarch (and his wife) was a Bacchic initiate and that he tells us (Moralia 61 ld-e) that the chief benefit of such initiation was to dispel the fear of death, for, according to the doctrine, the soul is released from the body even though the body itself is destroyed. Now these are mysteries, but if the essential mystery was the notion of the survival of the soul, then Plutarch would have been careful not to mention it, for he was no subversive iconoclast. What was secret, and so what Plutarch didn't divulge, was how you achieved this survival; but that Dionysus was a god who could assist you towards that goal was common knowledge.

exactly the same happens to Dionysus in Nonnus. At 35.262-340 Hera has sent him mad, and in recompense Zeus persuades her to adopt Dionysus and offer him her breast. Nonnus was a Greek writing in Egypt, and his sources for Dionysus are Greek of the Hellenistic period and earlier. It looks more than likely that behind the Etruscan mirror is some lost Greek account.

The subject of the Baltimore sarcophagus and the origin of its iconography We have strayed far in both directions from the date of the Baltimore sarcophagus, which according to Turcan (1966: 359) should be 190-200, according to Matz 170-80. Probably the date is some time in the reign of Marcus Aurelius or Commodus. In this case the Polesden should be later, into the third century, like the Woburn Abbey example and many others. Turcan dubbs the Baltimore artist 'un sculpteur de talent' (p. 224), which has to be an understatement. The Polesden sarcophagus is average, runof-the-mill in quality, whereas the Baltimore, by comparison, is arguably the finest of all the hundreds of Dionysiac sarcophagi.

A good many of the subjects of the reliefs of Roman sarcophagi were chosen because of their 'comfort value' to the deceased and their survivors. Achilles, whose exploits are very popular in this context, on one level may have been chosen because his life of physical valour might have been thought worth comparing with that of the dead person. But on another level, according to some strands of mythology that were not known to Homer, Achilles was thought to have broken free from death and Hades and to have been granted a life of perpetual happiness on the White Island. Herakles too - the subject of as many hundreds of sarcophagi as Dionysus - became immortal, only in his case he achieved the status of an Olympian equal to Dionysus and the other gods.

Of these hundreds a good many show Dionysus triumphing, but comparatively few are of the type incorporating bound prisoners who might be specifically associated with the Indian campaign. Fewer still are those with the god standing in a panther- or tiger-chariot. How did this type originate? Matz (1968: 220) drew up a stemma for this particular type in which the Baltimore relief plays an important role, as there are other reliefs that are derived directly from it. The example at Woburn Abbey is one stage further removed from the Baltimore, but all are ultimately derived, he argues, from some lost Hellenistic original, presumably a painting (see also WardPerkins 1992: 52).

There are indeed many links between Dionysus and Herakles, and in the context of the Indian triumph on sarcophagi they are sometimes shown together, as on an example in Woburn Abbey (Matz 1968: no. 100). One of these links is a historical one concerning the all-important figure of Alexander (who, incidentally, took Achilles as his exemplar in life): after his death, if not before, Alexander became identified with Dionysus because of their similar exploits in the East. Much later Mark Antony, returning from eastern campaigns, would proclaim a similar link with the god. During his life, however, Alexander associated himself very closely with Herakles, from whom he claimed descent. Other links are mythological. Both Dionysus and Herakles were sons of Zeus, and both endured painful early years due to the wrath of Hera. Herakles achieved immortality through his labours, and so in a sense does Dionysus in the fullest account of his career in Nonnus' epic poem Dionysiaca (written in the second half of the fifth century AD), where Zeus says at 7.96-8: 'This my son [Dionysus], after struggling on earth ... after the Indian campaign, will be received by the bright heaven to shine beside Zeus'. Only then will Dionysus change his status from demi-god with limited powers to Olympian god.

A striking figure in both the Polesden and Baltimore examples is the giraffe. On sarcophagi giraffes are confined to scenes of the Indian triumph, but there is only one other certain example, in Rome (Matz 1968: no. 96), and a couple of dubious examples that are more likely to be camels. Camels are certainly depicted on some of the sarcophagi with this scene (e.g. Matz 1968: no. 97), though one cannot tell whether they are single-humped African or Asiatic dromedaries. The giraffe is of course African, a fact that worries Toynbee (1973: 142) who writes that the 'erroneous notion that giraffes came from India appears again on the Baltimore sarcophagus depicting Dionysus' Indian triumph'. We can, however, stress the Indian-ness of Dionysus' triumph too strongly. The first hint of his eastern adventures comes in Euripides' Bacchae, lines 1317, where Lydia, Phrygia, Persia, Bactria, Media, Arabia and 'all of Asia' are mentioned. In the Indian campaigns described by Nonnus, many of the events unfold in Asia Minor and Syria as well as further east. But in Diodorus Siculus 3. 62 ff., and other writers, we learn that Dionysus had also conquered Egypt and Ethiopia, and in the latter

Herakles' elevation is even more dramatic: from mere hero to Olympian deity. In Etruscan art his promotion is shown symbolically with him suckling at Hera's breast (notably on an engraved mirror in Florence: Brendel 1995: fig. 285), and sometimes it has been assumed that this notion of adult adoption by suckling is a purely Italian one. But

337

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

country had found the warmest reception. Hence the Ethiopian boy outriders in front of the Baltimore chariot, who wear fawnskins, are likely to be young companions of the god.

There were plenty of opportunities for Romans to see African beasts at Rome, in the arenas and in the Emperor's military triumphs. Caesar exhibited a giraffe in the Games of 46 BC, and giraffes occasionally took part in the bloody events of the amphitheatre. The emperor Commodus in person went up to a giraffe in the arena and killed it (Dio Cassius 73.10.3). This seems unfair: a century before, the elder Pliny had correctly noted (8.27.69) that 'the giraffe is more remarkable for its appearance than its ferocity'. A century after Commodus, we are told in the Historia Augusta that Aurelian, having defeated Zenobia in Syria, triumphed at Rome, and the triumph included 20 elephants and other animals including giraffes (SHA: Divus Aurelianus 33.4). But before Commodus, Marcus Aurelius had triumphed in Rome in AD 176 (SHA: Marcus Antoninus 27.4-5), and he also staged 'marvellous games' (spectacula mirifica). No doubt animals were involved, but we are not told whether giraffes were included. As for elephants at Rome, the sources are much fuller, and throughout the period of the Empire there was a prosperous trade in the animals and their ivory, as the mosaic in the Foro delle Corporazioni at Ostia suggests (Scullard 1974: plate 18).

For the tone and content of the Baltimore procession one can turn to a description of a real festival held in Alexandria around 280 BC. The grand procession of Ptolemy Philadelphus was designed to extol both Dionysus, from whom the Ptolemies claimed descent, and Alexander whose popularity in Alexandria as a figure of cult was extreme. The parade is described by the Alexandrian historian Kallixeinos (quoted in Athenaeus), and among its many components was a procession with a statue of Dionysus accompanied by Nikai (Victories) with golden wings, and by many satyrs and maenads. There then followed (Athenaeus, 5.200D) a tableau of 'the return of Dionysus from India (ex Indon)'. This had a statue of the god reclining on an elephant, and a train that included 24 elephant quadrigae, with little boys mounted on them, Indian women and others dressed as prisoners drawn on carts, Ethiopian tribute-bearers carrying elephant tusks and kraters full of pieces of gold, and then a varied and very large menagerie including fourteen leopards, one Ethiopian rhinoceros and one giraffe.

One other feature of the Baltimore relief also suggests inspiration from the contemporary Roman world, as opposed to Hellenistic art. Pace Matz (see above), the front-facing prisoner on elephant-back has no particular 'Indian' look about him, but his longish hair and beard, hollowed cheeks, and loose leggings, are all familiar from the depiction of northern barbarians on the two spiral columns in Rome (on the Column of Marcus Aurelius one might compare, for example, Caprino et al. 1955: fig. 76, and - for barbarians wearing berets or caps - fig. 61).

From indications in Kallixeinos it is clear that the festival procession was a recurring event (Rice 1983: 186), though no doubt the one he describes was the grandest of all. It is also clear from Athenaeus (14.654C) that the Ptolemies kept collections of animals around their palaces, though the only species mentioned are exotic fowl. At any rate, it could be argued that some such Ptolemaic festival as this may well have provided the inspiration for the original work of art from which the sarcophagus reliefs derived.

It still seems likely that a Hellenistic original does lie behind the sarcophagus scenes, but that it does not account for the whole composition. It is improbable in any case that the original would have been of the long narrow shape to suit a sarcophagus front, more likely a squarish shape such as we see in Roman paintings that look to be based on Greek originals. Included in the original might have been the chariot group, the Nike and a maenad with flowing drapery. Even this kernel could be modified as necessary, as we see not only on the sarcophagi but also in mosaics where the same composition, but slightly simplified, was current in the third century (Dunbabin 1971). Here Dionysus is pulled along by two tigers, but a more ambitious version at Sousse in Tunisia doubles the number to four.

On the other hand all the sarcophagi discussed so far were carved in Italy, probably in Rome, and the Baltimore sculptor certainly seems to have seen a giraffe (Latin: camelopardus) at first hand. Although - true to the Latin name - he has incised the animal with a leopard's spots, he has captured the flattened head and elongated nostril very convincingly. True, there are no horns, but then the head is crammed beneath the upper moulding of the container; the lack of horns and the neck carried remarkably erect may also be indications of a juvenile (Guggisberg 1969: 67). Much closer to the time of the Alexandria procession is an attempt at a giraffe seen in a Ptolemaic tomb in Palestine (Peters and Thiersch 1905: plate 8), where clearly the artist had been nowhere near an actual camelopardus. More realistic, and probably the earliest rendering from Italy, is the pair of giraffes to be seen in the famous Palestrina mosaic, one stretching up to eat, the other bending down (Meyboom 1995: 119, fig. 11; Whitehouse 1976: 14, fig. Sa). They are labelled, in Greek, kamelopardali (sic), and the mosaic should date to the first or late second century BC. In the columbarium of the Villa Pamfili outside Rome (Reinach 1922: fig. 255.6) a giraffe is led along on a lead, a bell around its neck, as if it were a pet.

Matz's schema of progressive linear derivations from one sarcophagus to the next seems unlikely. Much more probable is the periodic publication of cartoons or patternbooks which could be adapted at the whim of sculptor or patron. A similar phenomenon should also explain the wide dissemination of themes and compositions in the field of painting (Ling 1991: 217-9). It seems inconceivable that at the height of sarcophagus production in the Antonine

338

Tom Rasmussen: Bacchus triumphs on the North Downs: the Polesden Lacey sarcophagus

period there were not new cartoons produced, instead of a reliance entirely on old Hellenistic compositions. So the Baltimore sculptor, one of the leading lights of this industry, took a basic Hellenistic core, added figures and trees and created a new design. The cartoon was made known, its influence is to be seen on other sarcophagi, and the Polesden sculptor attempted (unsuccessfully) to copy it exactly.

Guettel Cole, S., 1993. Dionysus and the dead. In Carpenter and Faraone (eds.): 276-95 Guggisberg, C.AW., 1969. Giraffes. London and New York: Arthur Barker Ltd and Golden Press. Lehmann-Hartleben, K., and Olsen, E.C., 1942. Dionysiac sarcophagi in Baltimore. Baltimore: Trustees of the Walters Art Gallery. Ling, R., 1991. Roman Painting. Cambridge: Cambridge University Press. Matz, F. 1968. Die dionysischen Sarcophage vol 2. Berlin: Gebr. Mann Verlag. McCann, AM., 1978. Roman Sarcophagi in the Metropolitan Museum of Art. New York: Metropolitan Museum of Art. Meyboom, P.G.P., 1995. The Nile Mosaic of Palestrina. Leiden: E.J. Brill. Peters, J.P. and Thiersch, H., 1905. Painted Tombs in the Necropolis of Marissa. London: Palestine Exploration Fund. Reinach, S., 1922. Repertoire de peintures grecques et romaines. Paris: Editions Ernest Leroux. Rice, E.E, 1983. The Grand Procession of Ptolemy Philadelphus. Oxford: Oxford University Press. Rowell, C., 1999. Polesden Lacey. London: The National Trust. Scullard, H.H., 1974. The Elephant in the Greek and Roman World. Cambridge: Thames and Hudson. Snowden, F.M., 1970. Blacks in Antiquity. Ethiopians in the Greco-Roman Experience. Cambridge, Mass.: Harvard University Press. Snowden, F.M. 1986. Aithiopes. In Lexicon Iconographicum Mythologiae Classicae 1: 413-9. Zurich and Munich: Artemis Verlag. Toynbee, J.M.C., 1973. Animals in Roman Life and Art. London: Thames and Hudson. Turcan, R, 1966. Les sarcophages romains a representations dionysiaques. Paris: Editions de Boccard. Ward-Perkins, J. B., 1992. Taste and technology: the Baltimore sarcophagi. In H. Dodge and B. WardPerkins (eds.), Marble in Antiquity: Collected Papers of J. B. Ward-Perkins. London: The British School at Rome: 39-54. Whitehouse, H., 1976. The Dal Pozzo Copies of the Palestrina Mosaic (British Archaeological Reports Supplementary Series 12). Oxford: British Archaeological Reports.

The Indian triumph of this type, then, is only loosely 'Indian', and reflects the god's triumph in a wider, more universal sense. The reliefs illustrating it are in spirit and composition a quintessentially Greco-Roman creation. The triumph is that of the Greek god, and it is he who crashes through the woods with his noisy retinue, but there are also conscious echoes of the clamour and tumult of numberless struggles in the arena and of a whole succession of triumphs of Roman emperors. ACKNOWLEDGEMENTS

A version of this paper was read at a research seminar in the School of Art History and Archaeology, University of Manchester in October 2000, and I am grateful for the comments of those present, especially Roger Ling and Frank Salmon. For courteous help with information and for permissions for photographs offered here as illustrations 32.2-4, I am very grateful to Christopher Rowell, Regional Director, National Trust (Polesden Lacey), and to his assistant Judith Mills. I am equally grateful to the Baltimore Art Museum, who have kindly provided and given permission for reproduction of illus. 32.6 herein. ABBREVIATIONS

SHA

Scriptores Historiae Augustae (Loeb ed.)

BIBLIOGRAPHY

Brendel, 0., 1995. Etruscan Art. New Haven and London: Yale University Press, 2nd edition. Burkert, W., 1993. Bacchic teletai. In Carpenter and Faraone (eds.): 259-75. Burkert, W., 1997. Ancient Mystery Cults. Cambridge, Mass.: Harvard University Press. Caprino, C., Colini, AM., Gatti, G., Pallottino, M., and Romanelli, P., 1955. La Colonna di Marco Aurelio. Rome: "L'Erma" di Bretschneider. Carpenter, T.H. and Faraone, C.A (eds.), 1993. Masks of Dionysus. Ithaca and London: Cornell University Press. Dunbabin, K.M.D., 1971. The triumph of Dionysus on mosaics in North Africa. Papers of the British School at Rome 39: 52-65 Dunbabin, K.M.D., 1999. Mosaics of the Greek and Roman World. Cambridge: Cambridge University Press.

339

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

32.1 The sarcophagus in the west portico at Polesden Lacey (photograph: Tom Rasmussen).

340

Tom Rasmussen: Bacchus triumphs on the North Downs: the Polesden Lacey sarcophagus

32.2, 3, 4.

Details of the front of the Polesden Lacey sarcophagus (photographs: National Trust).

341

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

32.5 Left end of the Polesden Lacey sarcophagus (photograph: Tom Rasmussen).

32.6.

Sarcophagus in the Baltimore Art Museum, inv. 23.37 (photograph: Baltimore Art Museum).

342

33 Image and Symbol in Severan State Portrait Cameos Valerie Worden 'Truth did not come into the world naked, but it came in types and images. One will not receive the truth in any other way .... The Bridegroom must enter through the image into the truth.' (Gospel of St.Philip c.2nd/3rd AD.: Robinson 1988)

Roman engraved gems provide us with the most spectacular examples of 'sacred' iconography when executed on a specific variety of agate, the Indian sardonyx. Witness, for example, the enigmatic narrative expressed in artefacts produced during the Ptolemaic period, such as the great serpent-handled Ptolemaic cup or the complex imagery of the Tazza Farnese.

up as if with arms spread wide, into the immensity of space.' The bowl's main potency however is its sacred 'inscription' 'Kristo', (identified as B.XRISTO R.S. XXX. according to a later published version of the inscription by the Jesuit scholar Lambecius) that lies beneath the surface and appears as 'if by a miracle' in the natural veining of the stone only when the right amount of light strikes the bowl. It has been said that the inscription is only 'visually accessible' to those who are spiritually initiate.

The superb gems executed during the Imperial period, such as the famous Gemma Augustea produced during the Golden Age of the Augustan workshop, would also seem to promote a far more potent message than the defined 'propaganda of Apotheosis'.

To the Romans, the bowl had an inherent sacredness. Because of the phenomenon of the inscription, Distelberger (1991: 112) argued that it was believed that, 'Nature had dedicated this inestimably precious stone of outstanding size to Christ her God and Creator' and surmises further that it may have been venerated as the Holy Grail, so seen as the key to the soul's salvation. It is possible that the origins of such beliefs stem from early RigVedic ideas, for Gonda (1969: 12) identifies references in the Vedic texts relating directly to the potency of gazing on this type of artefact in order to gain mystical union with the godhead.

This paper will focus on an analysis of image and symbol of the most ambitious of the later state cameos, those of Septimius Severus' reign. This period saw a great renaissance of the gem-cutter's art, during an epoch ruled by an Imperial family whose religious aspirations focused on allegiance to eastern Mystery cult, which revolved around belief in a solar deity. The author's intention is to reveal that the deeper potency identified with the agate in Roman culture made it the most prized stone for expressing in a specifically contrived esoteric manner, the images and symbols aligned to their religious beliefs. The paper will also suggest the genesis of such beliefs. This essay is offered in affectionate acknowledgement of the support given by Barri Jones to the author in respect of her esoteric approach to her thesis, and his much-valued assistance with its completion.

That agate was seen by the Romans as a potent medium for explaining sacred narrative is established by Pliny, in his Natural History: his description of orbicular agate (37.149) voices the stone's links with the solar deity and identifies that deity with an eastern genesis. Pliny also infers that the agate's potency is enhanced when it forms the phenomenon of the 'eye': 'Beli oculis albicans pupillam cingit nigram e medio aureo colore fulgentum et propter speciem sacratissimo Assyriorem deo dicatur.' ('The eye of Bel is whitish around a dark pupil which shines from the middle with a golden hue; because of its appearance it is dedicated to the most Holy God of the Assyrians.')

In the Roman period, the 'science' of stones (or lithology) was taken very seriously. While there was a belief that stones had very specific inherent qualities, which could be harnessed for healing or amuletic purposes, it was also understood that they could be used in a much more profound way, as a medium for 'mystical metamorphosis' (metempsychosis).

Hall (1974: 118) argued that, in the history of occult symbolism, the image of the eye was always seen as the symbol of god before he came to be represented in human form. Agate in Roman culture therefore should be viewed as a prized stone considered to have metaphysical associations with eastern solar cult, with which cultadherents readily identified for conveying the esoteric nature of a belief system which was infused with Mystery cult ideas of an eastern genesis.

Perhaps the most important example of the hardstone being used in this way is the immense bowl cut from a single block of agate, which had a history of Imperial ownership. This bowl is purported to be the largest gemmoglyptic bowl in the world - it has a 76cm. span. Distelberger (1991: 111-14) suggests that it had an almost 'magical fascination' for the Romans, 'the form of the bowl being broad, self contained and chthonic, simultaneously opening

343

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Agate's ability to reflect light and its potent connection with the sun aligned it closely with the Roman concept of 'lux perpetua'. Intrigued by its natural markings, particularly when of the banded or orbicular type, the stone was viewed as a reflection of the greater macrocosm. Agate's inherent sacredness in Roman belief is bound up with ideas concerning the origin of matter as first set down by Plato in the Timaeus (Lee 1977), based on an 'alchemical' process of creation from four basic elements. Stone was construed as formed from the elements by earth filtering through water, compressed by air, this latter process involving the element of heat, i.e. 'Fire'. It was Platonic theory that formed the basis of ideas expressed on the subject of Lithology. Evans (1922: 15) noted that such influential writers as Theophrastus, whose magnum opus, 'On Stones' was viewed as a definitive source ofreference, devoted his entire opening passage to such beliefs. Agate was considered a product of this 'alchemical' transmutation of matter; the stone's 'fiery' nature is further enhanced by its origins in volcanic magma.

The properties of this material freed the gem-cutter from the strictures of effecting meticulous realism, or 'verism', which Roman artists, particularly sculptors, were trying to effect in portraiture. Inspiration came from the principles of Platonizing ideals that Art and Nature alike impose a structure on matter in accordance with an inward vision of archetypal form, and that form (beauty) is imposed on matter by the Principal or Supreme Being. The gem-cutter was able to reveal, on this 'mystical medium', the inner spiritual realities of his subject. The emphasis of Platonizing philosophy on the spiritual and non-material aspects of reality connects with eastern ideas especially as regards the ancient concepts of cosmology and metaphysics. As the earliest attempts at 'portraiture' stem from eastern artistic tradition, inspiration for obtaining ' inner spiritual reality' may also have derived from a similar source. This eastern emphasis is not surprising when we consider Henig's observation (1983: 155) that, more often than not, the Roman gem-cutter, or his father before him, had been trained in an eastern court. That the art of gem-cutting, and the 'science of gems', Ratnasastra, have an eastern genesis (ultimately from India) reinforces the impression of orientalising influence. Thus the pre-occupation with projecting spirituality seems to stem from an ingestion of eastern philosophical ideas specifically concerned with the concepts of the afterlife and spiritual immortality. The very nature of the material they were using to project such profound ideas also has a specifically eastern genesis.

Thus the stone was appropriately associated with the esoteric concept of the 'alchemical' process in man, a Hermetic 'science', which purportedly held the secret of the transmutation process of the 'physical' into the 'spiritual' through initiation by 'fire', a concept that was both understood and expressed by the Alchemists of the Alexandrian School which flourished from 200 B.C. to A.D. 700 (Oxford Classical Dictionary 2nd ed. 1970 (henceforth OCD): 37). Of Egyptian extract, the legendary Hermes Trismegistos was seen as a protector of Osiris and helper of the dead. As identifed with the deity Thoth, he was associated also with the founding of learning and wisdom; his cosmic connections were largely of a solar nature; he is described by Lurker (1980: 121) as the tongue and the heart of Re. Adopted by both the Greek and Roman Pantheons, (Hermes and Mercury respectively) he remained the divine messenger and transporter of souls on their journey to the afterlife. The imagery of the upward spiralling entwined serpents as expressed on his attribute, the 'Caduceus', might arguably be an intended reference to the 'fire' or spiritual energy in man.

While agate occurs on a small scale on the Italian peninsula, the size and type of sardonyx used by the Roman gem-cutter for executing imperial state portrait cameos (or indeed the great cups and tazza) came from large nodules imported from India. More specifically the orbicular or 'eye' agate type has its main provenance in India. While Pliny reiterates the stone's Indian origins (Natural History 37. 86), contemporary literary sources are even more specific: Ptolemy's Geography (Stevenson 1991), written in 140 AD, actually cites a set of coordinates for the Sardonyx mountains where the raw gemstone was found (7 .1). Despite the contraction evident in Ptolemy's southern latitudes of the subcontinent, the location can be identified with the area south of the Indus Valley, between the river systems of the Gulfs of Kutch and Cambay. The dividing range is formed by the Aravalli Mountains and these may be readily equated with the co-ordinates given by Ptolemy, namely 121 degrees 30mins by 21 degrees latitude, the latter being the more reliable locator. Moreover trading stations along the two gulfs are attested in the Periplus of the Erythraen Sea (Casson 1979: 48.16.15.) the date of which arguably lies sometime prior to A.D.50. According to Casson, (1979), there is no doubt that the passage relates to the export of precious stone, notably agate and sardonyx, exported principally through the Emporium of Barygaza (later Broach). This trade seems to have carried ideas about agate's properties and other eastern religious ideas over a

Plato's Timaeus is influenced by Pythagorean ideas. Coxon (O.C.D. 1997: 903-4) argues that this known devotee of the solar god Apollo was a staunch believer in 'metempsychosis'. He believed that though 'the soul was condemned to a cycle of reincarnation, it could win release by cultivation of an Apolline purity'. Thus he advocated a way of life in which the investigation of Nature became a religion, advocating such disciplines as celibacy, purification, silence and self- knowledge. Pythagorean cosmology was conceived as an astronomical system which pre-supposed a central fire around which circled the celestial bodies, including Sun, Earth and 'Counter Earth'. Agate, with its inherent 'metaphysical' qualities and divine connotation, was thus the supreme medium for projecting an image of 'spiritual realism' into Imperial portraiture.

344

Valerie Worden: Image and Symbol in Severan State Portrait Cameos

long period (Worden 1992: 45-56). One is forced to reflect on the early genesis of such beliefs.

deity, such as Apollo or Helios-Sol Mithraic cult).

(Sol-Invictus of

The Pan Orientalists, Cumont (1956: viii, 10) amongst them, proposed the theory that Rome derived from the east much of its intellectual and artistic ideas, including powerful religious influences. Indeed, infiltration of eastern influence started early in Roman culture, with its absorption of indigenous culture of the Etruscans, heavy with Orientalising overtones, followed by influences from the Graeco-Oriental cities of Asia Minor and as far as the Asiatic hinterland. The doctrine of resurrection and immortality, which began to surface from the end of the Republic onwards, came as a direct result of the infiltration of Oriental Mystery religions (Cumont 156: 29-30) and Pythagoreanism based on Orphism.

The motivation for projecting Septimius via the symbolism of solar cult is appropriate to an emperor who is notorious for his long-standing pre-occupation with eastern mystery religion. Throughout his life, as Turton observed (1974: 7), Septimius Severus expressed an inordinate interest in religious ideas and the rites via which they found expression. His Lepcitane background introduced Severus to Punic religious influence (Birley 1971: 21, 28). When he left for Rome (in 162 AD), the ancestral Punic Gods still retained a role but the joint patrons of Lepcis, Melquart and Shadrapa, had become Hercules and Liber (Bacchus) (Birley 1971: 62). Hercules identifies with the concept of 'one who is initiate', Bacchus with Dionesian Mystery cult. As legate of IV Scythia, in N.Syria, he arguably visited Tyre and Sidon and he did travel someway south (Cassius Dio 78.8.6) to consult the Oracle of Zeus Belos, the local Ba'al, at Apamea. Birley noted that the Oracle (quoting from Iliad 2.478-9) alludes to the 'divine nature' of Septimius, and suggests that since he visited Apamea, he may well have journeyed on to Emesa, which had a much more famous Temple to the God of a Solar Cult. The God worshipped there was Elagabalus, his cosmic connections made clear by the Greek and Roman rendering of his name, Heliogabalus, 'Helios' meaning, 'of the sun.' Such was the religious prestige of this shrine, that according to Herodian: 'All the neighbouring Princes and rulers sent generous and expensive gifts there every year.' Emesa's prestige lay in its main religious attraction, a huge black conical stone which the people of the city revered because of its cosmic identity with their solar god.

Agate (in its 'eye' form) is specifically linked with the dying and revifying god of solar cult in Euphratean culture (Worden 1997: viii), with, for example, specific reference to eye stones in Mesopotamian myth. The final passage of the Descent of Ishtar into the Underworld links these stones, identified as banded agate by Dalley (1991: 162), with the goddess' consort, Damuzzi, and his sister Belili. Dalley (1991: 154) suggests that the latter part of this text relates to a ritual enacted at the annual 'New Year' festival in honour of his seasonal resurrection. As in early ancient eastern culture, the Romans believed that stones improved communications between man and god. Thus the Agate, with its inherent translucency, was greatly prized, enabling the artisan to enhance the 'other worldliness' of imperial figures in scenes of so-called 'Apotheosis.' This stone was deemed as the most appropriate for imparting symbolic inference of the Emperor's immortal state and the ethereal nature of the soul. The gem-cutter would have fully appreciated also the intensely 'occult' nature of the stone, exploiting its natural layering as a perfect vehicle for projecting the more personal philosophical tendencies of his patron, subtly intimating such beliefs under the aegis of esoteric image and symbol. Such practice was arguably directly influenced by Oriental ideas (Cumont 1956: 17).

The worship of conical stones also has an ancient eastern provenance. In Rig-Vedic culture the conical stones which have been variously described as 'lingam ', inferring some sort of sexual connotation, originate from solar cult imagery. We are informed by Swami Sankarananda (1943: 1.125 -32.) that as emblems of 'Siva' god of the sun (ascending aspect), the divine teacher who revealed the absolute truth, the reality behind world illusion, the lingam was symbolic of the stump of the sacred tree.

One of the earliest of the fine state portrait cameos of the Severan era is that of the Severan family. (Illus. 33.1) Executed in three layered sardonyx, c. 198-209 A.D. (now housed in Paris, in the Cabinet des Medailles, Bibliotheque Nationale), this carved gem depicts portrait busts of Septimius, wearing a crown, with his second wife, Julia Domna and their sons Geta and Caracalla. The image of the crown is central to our understanding of its deeper meanings. Hall (1974: 79) interprets it as an emblem of sovereignty, both divine and earthly, but Fontana (1993: 71) presents a more esoteric vision. He proposes that its circularity represents perfection and infinity, and, if executed in gold, connotes solar power (see below, p. 348). The 'radiate' form of the crown arguably alludes to the radiate nature of the sun, and thus identifies with a solar

During a period in his career where he was either ousted from his legionary command, or required to retire form public life, c. 182 A.D. (the reason is not made clear in Historia Augusta: Severus 3.7), Septimius is known to have taken off on 'the grand tour.' Besides the obvious cultural attractions of Athens, such as 'monument spotting' and the inevitable viewing of antiquities, we are informed by the Augustan History that his journey had a deeper religious motive. The raison d'etre of this sacrorum causa is rather more obscure. Birley (1971: 119), however, suspects that Septimius expressed a desire 'to be initiated into the Mysteries of Eleusis.' While this pre-occupation

345

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

with the cult of Eleusis can be interpreted in political terms - to emulate four previous Emperors - it is equally likely that it may have been an entirely 'personal journey.' Severus' ongoing religious interests are indicated by his election into the elite Roman Priestly College, the 'quindecimviri sacris faciundis ', an event which may have pre-dated his imperial status (Birley 1971: 225). To the 'quindecimviri' was entrusted care of the Syballine oracles and of foreign religious rites of which the worship was licensed at Rome. This included the cult of Serapis and Isis. The oracles under their guardianship were those carefully vetted by Augustus and housed in the Temple of Palatine Apollo. One aspect of the duty of the Priestly College was, by tradition, the overseeing of rites carried out at the inaugeration of the Saecular Games, a momentous event which symbolised the inaugeration of a new Golden Age. Whether or not Septimius and his two sons Geta and Caracalla would have performed all the traditional rites of sacrifice and rites involved throughout the entirety of the games held in honour of his reign in 204 A.D. remains open to question. The final fames in the Circus would have been preceded by the Hymn Carmen Saeculaere, composed by Horace, whose words included an invocation to Apollo and Diana, a reference to Bacchus and the golden fields. The mention of Bacchus is significant, for the God of Wine identifies with Liber Pater, one of the two guardian deities ofLepcis Magna.

Significantly, Birley (1974: 236) equates Septimius with the solar deity, for the statue of the Sun, which was centrally placed in the Nymphaeum, perhaps depicted Severus himself - looking south towards Africa. He apparently designed his Septizodium with some astrological purpose in mind, particularly as Dio remarked that his embellishment of the Imperial Palace also had astrological overtones and he is also known to have equipped the imperial villa on the Aventine with a Mithraum, a further indication of his allegiance to Mystery religion revolving around a solar cult. Septimius also had ambitious building plans for the embellishment of Lepcis Magna, his native city. Of high import is his choice of architects and sculptors, for their genesis was eastern. Philostratus informs us in his Lives of the Sophists, that the first African Emperor contracted architects and sculptors from Asia Minor to take on the huge new restoration and development programme. While Cumont (1956: 8) points to a tradition of employing Oriental artisans because of 'the superiority of their technical knowledge and inventive genius', the Emperor may have been more intent on recruiting men who could properly understand the sacred significance of the di auspices and thus interpret such meaning. When we consider the esoteric nature of the embellishment of 'the excessively large Temple to Bacchus and Heracles' which Dio records his erecting, his motive for using Oriental artisans becomes clearer. A direct reference is made to the trials of the initiate, Hercules in the narrative of the sculpted reliefs on the decorated pedestals (Ward-Perkins 1993: 44. fig. 21.) Described by Squarciapino (1974: 1) as scenes of Gigantomachy, it may be that the image of the protagonist Hercules could be identified more specifically with his esoteric role. Could he be, equally, symbolically subjugating the lustful element in man? For the 'enemy' is defined in the imagery with 'serpent appendages'- a symbolic allusion to the area below the waist, which in Vedic religious discipline (SB. 1.3, 1.13: see Gonda 1969: 27) is defined as 'impure'. Platonic discipline speaks of the Titanic nature of man as a proverbial saying in the sense of his evil nature; the body is evil, the soul is divine (OCD: 759). Is it possible also that the circular motif with a central raised 'dot' on spandrel blocks from an arch of a colonnaded street at Lepcis (Ward-Perkins, 1993: pl.36 d.) could be an intended allusion to the image of the sun? As an eastern motif it would be entirely relevant to the cult as expressed in the religious environment in the emperor's native city.

Thus however traditional the prayers and rituals may have been, there was clearly some Severan modification involved in honour of the di patri of Septimius, Hercules and Liber (Birley 1971: 228). The coins issued in commemoration of these games convey the impression that the two Gods of Lepcis were the presiding deities of the whole occasion. The importance attached by Septimius to the di auspices of the North African province is indeed indicated by the exceptional prominence given to such religious imagery on coinage during his reign (ibid: 176). After Septimius's victories at Cyzicus, the ancestral Gods of Lepcis Magna, Hercules and Father Liber (Bacchus) appear for the first time on Roman coinage, alongside images of the Emperor's victory and the 'spirit of the Roman people'. During his improvement of the City of Rome, Septimius erected buildings specifically linked to his personal religious requirements. While the rebuilding of the Forum of Peace, the restoration of the Pantheon and the erection of the Severan Arch as an entrance to the Forum Boarium were all part of the imperial plan, he is known to have erected a magnificent building, The Septizodium, no trace of which survives (ibid: 236). Built at the comer of The Palatine that faced the Appian Way, it was designed to resemble a Nymphaeum, or theatrical scaenae frons. This huge building (nearly 100 feet high, 300+ feet long) betrays his interest in astrology, for it is thought to have contained statues of the seven planetary Gods, Saturn, Jupiter, Mars, Venus, Mercury, the Moon and the Sun.

Severus' pre-occupation with the inherent power of the Cosmos, as expressed in Mystery cult symbolism, was therefore an ethos totally infused into his life. His belief in astrology, his regular consultation of seers, his following of Mystery Cult religions, and his interest in the 'supernatural' became even further entrenched with his second marriage to Julia Domna, whose horoscope apparently predicted marriage to a king (Turton 1974: 4).

346

Valerie Worden: Image and Symbol in Severan State Portrait Cameos

Indeed, a large share of the credit for the intellectual and religious fertility of the reign is due to the influence of Severus' second wife. No doubt revered by her subjects as a Roman Gloriana, fitted by her rank to preside over the arts and sciences, her main contribution to the Severan reign was her philosophical input. Julia's views are revealed in a book, the Life of Apollonius written at her request and under her inspiration by one of her closest friends, Flavius Philostratus, a leading member of the Sophist circle, a literary and scientific society over which she presided. Philostratus expounds the ideas on religion and philosophy held by himself and Julia. They held a mutual belief that due respect should be paid to all religions alike, that (Turton 1974: 58): 'divinity has its place in the human soul rather than in the Temples.' Julia believed that the spiritual journey was a personal one; hers was a belief system derived from the cult of Mystery religion.

ancient tauromachia, to the spectacle of the Spanish bull fight. This is heightened if we follow Gonda's thesis (1972), that Mitra was a Vedic god and that the origins of Mithraism lie in Vedic rather than Iranian beliefs. The 'bull of heaven' of Vedic texts (Rig Veda 1.160.3: 1.79.3) is also connected with the life force, seasonal fertility and the natural cycle of death and rebirth; the life blood of the bull slaughtered by Mitra represents immortality. In zodiacal terms, Taurus appears in the earliest records, marking the beginning of the equinoctal year, whence bull ritual was woven into New Year festivities celebrating the dying and revivifying god (Davidson and Aldersmith 1924: 1,245). Even Vergil remarked that 'the white bull with the golden horns opens the year.'

In depth analysis of the symbolism expressed in the iconography on the sardonyx cameo of Julia Domna (Illus. 33.2), executed third century AD (now in the British Museum), indicates a specific intention to communicate her religious beliefs via the esoteric imagery of the narrative. This engraved gem, which depicts Julia with a firebrand in her left hand, riding in a chariot drawn by two bulls, has been described by Henig (1983: 156) as Julia in the character of the Goddess Luna, here perhaps equated with Dea Syria. Cumont (1956: 103-4) identifies this principal Syrian goddess as Atargatis, the first Semitic deity to enter Italy. That her cult included mysteries is clear from Lucan's De Dea Syria and there is in addition a third century 'creed' from Britain (RIB 1791) that accepts Dea Syria as one of several names or manifestations of the universal goddess. Henig's equation of Dea Syria with Luna suggests a connection with Mithraic mystery cult, since she is considered the inseparable companion of SolSanctissimus (Strong 1915: 226).

Mystery cult influence is also intimated by the firebrand brandished by Julia Domna, intimating a stage in spiritual initiation through the element of fire. There is a connection with Mithraic ritual, for fire transformed the 'lion grade' initiate (the Lion having a solar connotation) into a 'new man', one who is sanctified (Vermaseren 1963:135.) The Vedic parallels are interesting, for the texts (Gonda 1972) make direct reference to the links between Mitra and Agni, the Vedic god of fire. The fire symbolised the 'life force within the human body', which could be raised in Mystery cult initiation by a psycho-physical process (see below p. 350).

Thus it is possible that we are witnessing in the narrative of this cameo a deep astronomico-religious context connected directly with the tauromachia of the Mithraic Mysteries.

In the Roman belief system, the most profound insight into the esoteric meaning of the element of 'fire' is provided by the comments of the Neo-Platonic philosopher, lamblichus (Toynbee 1971: 194), on the metaphysical aspect of fire in his De Mysteriis, 12, (quoted in Strong 1915: 194). This work is concerned with defence of ritualistic magic as follows: 'Fire destroys the material part of sacrifice, it purifies all things that are brought near it, releasing them from the bonds of matter and in virtue of the purity of its nature making them meet for communion with the gods. So too it releases us from the bondage of corruption, it likens us to the gods, it makes us meet for their friendship and it converts one's material nature into an immaterial.' Stoic philosophy also fostered belief in the reabsorption of the soul into the 'fiery ether' after its separation from the body (ibid 193).

That there is a definite intention to project a deeper religious message in the narrative, one that is heavy with eastern mysticism, is further sustained by an esoteric interpretation of the symbols of the bull and the firebrand. Bull imagery is not seen as a stock motif for driving the celestial chariot (identified in Vedic terms by Gonda (1969: 56) as a means of ascending the celestial regions). The image may, therefore, have a deeper meaning. The bull has long been identified 'as an object of worship in primitive religions for its strength and fertilizing power' (Hall 1974: 54), but this vision might be extended: in both Babylonian legend and Vedic beliefs there are positive indications that in this 'aspect of fertility' the bull connotes the constant cycle of rebirth /reincarnation into this world. Only by the symbolic slaughtering of the bull can man escape this constant wheel of karma and rise to the immortal state (a process achieved by a series of initiations). It is arguably this deeper concept that defines the long history of images relating to bull-slaying, from

The suitability of sardonyx to express such a profound message is reinforced by contemporary perceptions of its associations with fire, formed as it was from the earth's own fiery magma. Julia Domna's links with the male-dominated cult of Mithras could appear tenuous until we consider Tertullian's testimonies (quoted in Vermaseren 1963: 164) concerning the 'female input.' He informs us that the

347

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Mithraic cult had 'virgines et continentes,' women who took on celibacy as an act of dedication to the god. Given her husband's construction of a Mithraeum in their imperial villa in Rome, Julia's interest seems wellsubstantiated. The image on the cameo reinforces this perception, therefore, but her interest had deeper roots than this. Emesa, her birthplace, was an important shrine for Semitic religion, the influence of which on Julia was profound, if only because her father, Julius Bassianus, was high priest of Ba'al at the Temple. The god Elagabalus was seen as a creator god. Various interpretations of the God's name inform us of his high status (see the arguments of Turton 1974: 11; Cumont 1956: 123). Such was the religious prestige of this holy shrine where the creator god, the solar deity, was believed to have been manifest, that pilgrims flocked there not only from the Roman province but from as far afield as Parthian territory.

yogic traditions is referred to as the 'ajna chakra' or the 'third eye' of the subtle etheric body. Further symbolism relating to this spiritual concept is provided by the crown or diadem adorning the head of the Emperor. When interpreted at an esoteric level the crown takes on a particular potency. Fontana (1993: 183) defines the symbol in alchemical terms, suggesting that, when the initiate has reached the spiritual aspect of the androgyne (hermaphrodite), he symbolically wears the 'crown of perfection.' The crown worn by Severus, probably of gold and set with three cameos, which would appear to be executed in sardonyx, is worn on the highest and most significant of the seven chakras, actually referred to as the 'crown chakra', in yogic discipline (Worden 1992: 7). Connected to the pineal gland, the entry point for various energies flowing through the body, it is the area of spiritual opening where one obtains personal identification with the Infinite, oneness with the godhead. Gonda (1980: 145) suggests that Vedic ritual states that golden regalia can be used for purification, for the sake of auspicious progress.

Julius Bassianus's naming of his daughter, according to Turton (1974:4), implies an intention to relate to the role of consort of Elagabel. He suggests that her divine aspect is corroborated by an inscription on a seal found at Cyzicus, for domna appears as the title of a Goddess of the Underworld. That it was seen as a custom for a Syrian God to have a female consort supports the theory established earlier, that Julia Domna as Luna was seen as a fitting consort of the solar deity (and the Emperor). The Severans presided over a Roman world experiencing profound religious change towards more personal conceptions of religion via eastern cults (Turton 1974). Through participation in the re-enactment of rites, with their inherent religious symbolism, initiates could literally enter/experience the truth in a process of mystical metamorphosis. Julia Domna thus followed the example of many leading thinkers of the day who attributed to sacred tradition a symbolic meaning. Interpretation of the esoteric symbolism was the method used by eminent philosophers to extract spiritual edification from such works as the Iliad, and with equal ingenuity by the Christian fathers, Origen especially (prob.185 - 254 AD) to elucidate refractory texts in the Bible. Indeed it is analysis of the image by this process that provides us with a deeper insight into the painted narrative expressed on the 'Berlin' Tondo, a wooden roundel from Egypt dated c.199-201 A.D. depicting the Severan family, which illustrates Julia Domna's connection with mystery cult practice (Illus. 33.3). This rare example of an imperial painted portrait reveals that the foreheads of both Julia Domna and the Emperor bore cult initiation marks, which, although there seems to have been an attempt made to obliterate them, are still visible.

Since painting of the Tondo (c. 199 AD) coincides with the Severan journey to the Nile in that year, it may be that this symbolism had a specifically eastern inspiration. The Severans clearly took full advantage of their political forays into the eastern provinces to further their religious interests. During the invasion of southern Mesopotamia, c.197 AD, the army reached Babylon, a city famous for its astrology and to which Julia's native religion arguably owed great a debt. Philostratus accompanied Julia and recorded the visit in the Life of Apollonius. Though now deserted and abandoned, Babylon was not wholly derelict; its famous Temple of Jupiter Belus, the main attraction for Julia, was still intact. Jupiter Belus was the Romanized name of the chief of the Babylonian Gods, Marduk (a solar god, as he represented the 'shadow in the sun', the 'eye' of the solar logos, his attribute was the serpent-dragon). Political input apart, it was largely the Severan preoccupation with mystery religion that also initiated their journey to Egypt. The Augustan History bears out this view: 'subsequently he [Septimius] always indicated he enjoyed this trip, because he had taken part in the worship of the God, Serapis.' Serapis was a God more or less invented by Ptolemy I Soter (c. 367/6-283/2 BC), to unite his Greek and Egyptian subjects. Thus Serapis had attributes borrowed from both Greek and Egyptian mythology. This composite God assimilated the characters of Osiris, the Apis Bull of Memphis and the Hellenistic elements of Zeus, Aesculapius and Dionysus. He was also God of the Underworld and his main cult shrine was the great Serapeum Temple (bull shrine) at Alexandria. Severus' enthusiasm reflected his deep religious views. However, as Ruler of Egypt, he was also regarded as divine, and thus the most auspicious overseer of the 'solemne sacrum ' which took place annually at Philae (Aswan). He would arguably have officiated at the May ceremony on the Upper Nile at which he, as Divine Ruler,

A papyrus text from Florence (Vermaseren 1963: 131) relates that: 'in order to be recognised as one who is initiate' one has to be symbolically marked either by tattoos on the hand or forehead. As Gonda acknowledges (1969: 70), it is this area of the forehead that in Indian

348

Valerie Worden: Image and Symbol in Severan State Portrait Cameos

would have cast gifts of gold and silver into a rock cave down near the river (Birley 1971: 205). The ritual was a daily re-enactment of the stages necessary to the god's continual life on earth.

sun god popular in sepulchral imagery (Strong 1915: 181), and Sankarananda (1943: 1). 110-24) identifies such imagery in early Rig Vedic culture. However this concept of power and victory can be interpreted in a more esoteric way. The bird was seen as a messenger of the sun. Winged images were symbols of the soul's victory over death and its transportation to celestial realms, intimating belief in the concept of life after death; man's inherent immortality. Thus the image of the winged victory, seen as a messenger of the gods, could also identify with this concept.

Julia was already familiar with the elaborate nature of such ritual, and was perhaps more interested in underlying ideas, especially those relating to Isis (Turton 1974: 69). Well-versed in contemporary literature, she will have grasped the spiritual meaning given to the rites by Apuleius in the last book of The Golden Ass (Graves 1990: 19.196) - the sense of divine companionship that rewards the believer initiated into the Mysteries. The Emperor's interest in this Rite of Serapis, and the underlying theme of the Egyptian Mystery Cult duo, Isis and Osiris, is confirmed by his immediate foundation of a private collection of as many books on esoteric doctrine as he could find (Dio 75: 13.2), many extracted unwillingly from the Alexandrine library.

The standard interpretation of the 'gorgone' is primarily as a protective talisman. As recorded in myth (Met. 4: 769803), the head of the Medusa would seem on the surface to be stock narrative for promoting fear. However, it may have a more profound meaning, its inherent benevolence linked with the sun, for myth also visualises this image connected with Zeus, the great sky god, and the blood from the 'gorgon's' body was used to revive the dead. In the moment of her death, she is said to have given birth to the winged horse Pegasus, who became a symbol of immortality.

Severus appears to have been fascinated particularly by enigmatic effigies. He is known to have visited the Pyramid complex with Julia Domna, on their long journey down the Nile in the early summer of c.199 AD. A particularly fine aureus struck during the Severan reign (ibid: 177) reinforces this view: witness the re-emergence of the saeculum frugiferum (a highly significant image as it represents the Latin name of Ba'al Ammon) and the solar deity, depicted wearing a fez-like head-dress, seated on a throne guarded by sacred sphinxes. Birley suggests that the figure closely resembles a Punic relief found in the shrine ofBa'al Ammon at Hadrumetum. The sphinx would thus appear to have some specific link with the solar god and the concept of creation. Its complex imagery informs of such a link, the body being leonine links in astrological terms with the sun, its head takes on human form symbolising man's solar allegiance. Literary sources hint at its divine aspect (Hesiod, Theog. 326).

That solar symbolism is intrinsically linked with the 'gorgone' during the Severan Age is sustained by the iconography of a spectacular Roman intaglio in the Cabinet des Medailles in Paris. Cut in green plasma, the gem shows a central motif of a winged Medusa's head surrounded by the twelve Zodiacal sigils. Plantzos (1998: 42) proposes that this intaglio was executed in the second century AD. On Roman gems, the 'gorgone' is often featured as winged, and it is a winged 'gorgone' that is featured on Caracalla's aegis. Thus the Medusa in her winged aspect arguably represents the concept of immortality. The image of the 'winged gorgone' also appears on Severan architecture, adorning the spandrel medallions decorating the portico arcades of the Forum at Lepcis Magna (WardPerkins, 1993, fig.6. and pl.9.b.), whether or not it was intended to project the same esoteric message.

Such was the intellectual and religious world in which the Severan heirs were raised. It is to the reign of Caracalla that we now turn to examine the imagery of one of the finest of the Severan imperial state cameos executed in Indian sardonyx. Esoteric iconography is well expressed on this superb imperial state cameo entitled 'the Apotheosis of Caracalla.' (Illus. 33.4). Executed c.217 AD, it is now housed in the Bibliotheque Municipale, Nancy. The narrative of this carved gem has Caracalla wearing the Aegis, with its image of the 'gorgone', carrying a winged victory in his right hand, a Cornucopia bearing orbicular 'eye' symbols on the upper rim, in his left, being raised up by the eagle. Such images surely project a far more esoteric message than the stock interpretation reflecting the imperial propaganda of deification of the Emperor.

The snake's importance in Roman religious ritual cannot be over-stressed. The image is always seen as connoting the soul in some way. It was connected with the exalted mysticism of Pythagorean philosophy, its various poses having specifically esoteric meanings. For example, the upward gliding/twining snakes of Roman sepulchral narrative seem to connote the progress of the soul towards celestial regions (Strong 1915:193). The serpent has long been worshipped as a potent symbol, for example in sacred narrative in both Ancient Near Eastern and early Rig Vedic culture, from where it seems to have been absorbed into the Greek and Roman ouevre. In all of these sacred disciplines, the image of the serpent is esoterically linked with the sun and with the symbol of the 'eye' (Mundkur 1983: 71). The cult of the serpent, with its accompanying purification rites is known to have existed at Taxila, the Indian outpost visited by Alexander, with whom Severus

The Eagle was adopted as an Imperial image because of ancient association with power and victory. Sacred to Jupiter, it could also be read by this date as an image of the

349

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

ostensibly associated himself. The link between eye and the sun is stipulated in early Rig Vedic Culture (Swami Sankaranada 1944: 2, 55), the centre of the sun being referred to as the 'eye of the sun', behind which is the solar logos, the source of divine intelligence. Vedic texts refer to the sun inter alia as the eye of Mitra and Agni (Gonda 1969: 41), the latter being a fire god. Thus Agni, the sun, is linked with the eye of gods and men. He is the eye and protector of the great rta (cosmos), and inevitably is linked with the inner eye, the ajna chakra, which Gonda remarks as 'situated on the forehead at the junction of the eyebrows .. the eye of Bodhic insight.' In Vedic sacred ritual, communication with divine intelligence is via contemplation of the solar phenomena through the inner eye (Gonda 1969: 11).

as integral to the complex iconography of the interior, which he identifies as esoteric narrative expressing Hermetic discipline. Previously, the 'gorgone' had been considered independent. Dwyer now proposes that the image has some solar aspect linked with the interior figural decoration, which consists of seven figures plus 'the sphinx'. While identifying this complex narrative as a temporal allegory of the Nile, he intuitively recognises a deeper layer of symbolism, and reads the iconography as an astronomical map and astrological document representing an allegory of the Creation of Man, and as an illustration of the text of the Corpus Hermeticum. Dwyer therefore makes two astrological connections between the 'gorgone' and the interior iconography. First (1992: 256), the pseudo-Eratosthenic Catasterisms associate it with the she-goat, Capella, nurse of Zeus (the sky god) when identified with the star alpha-Aurigae, which determines the vertical axis of the composition of the interior. Secondly, there is his proposal 'that the Gorgon represents the Sun in the Underworld, just at the point of appearing at the solstice point at the moment of sunrise.' Significantly, Dwyer identifies Alexandria as the source of the Tazza, perhaps made in the first century BC as a libation bowl (1992: 261). It was certainly thereafter considered a potent object, being prized by Augustus (Suetonius Aug. 71), who was supposed to have taken just this single piece of booty on his capture of the Palace at Alexandria. Augustus's contribution to Roman culture is legendary. As Diva Filias he was noted particularly for the revival of ancient solar oriented cult. That the iconography held specific importance for the Emperor is indicated by Pliny, (Natural History 37.10), but it is not specifically known to have come into Severan ownership.

In Egyptian narrative we see the udjat (eye) combined with the cobra and the solar disc. Such a combination of motifs is seen in Pharoanic regalia, largely worn as a filet. Significantly it is worn over this confluence on the forehead where Kundalini energy manifests itself While the 'mystic eye of Osiris' was worn as a protection against magic, the 'eye of Horus' was regarded as a symbol of power and substance. In Egyptian thought, likewise therefore, the eye was also linked with the power of the God of Light, the sun (Lurker 1980: 128). Marduk, the Babylonian solar god, is similarly linked with the solar logos, the eye of the sun, and has the attribute of the serpent/dragon. 'Eye' symbolism is also linked with Bel. In a passage relating to'the Descent of Ishtar' eye stones play an integral role in ritual concerning the dying and revivifying god (Dalley 1991: 160). Ishtar's links with Bel have been established by Groneborg (1986: 17): in her role as 'hermaphrodite,' her upper parts are Bel, her lower parts are Ninlil (consort of the sky god). The Greek cult of Zeus Ammon also identifies with eye symbolism. Cook's drawing of the homed god's torso, represented by the body of a snake, has each of the gastrostages defined by 'eye' symbols (Mundkur 1980: 71, fig.37b). This composite God is identified in Stoic philosophy with its highest principle, 'fire'.

The use of agate as a sacred medium for depicting esoteric narrative in the Ptolemaic court at Alexandria is perhaps linked with Hermetic beliefs expounded by the Alexandrian school (Dwyer 1991). A special allusion is made to this hardstone in combination with the symbol of the 'eye' in the narrative of the magnificent third-century BC 'Spirit of Alexandria' mosaic. (Illus. 33.5). This personification of Alexandria, as so-called 'mistress of the sea' shows her 'crowned' with a head-dress said to emulate 'the shape of a ship's prow', but the prow is decorated with an 'eye' symbol. Though this 'gli occhi di nettuno' could be a conventional 'protective' symbol for seagoing vessels, the image may project a deeper message when combined with the wide open gaze of Alexandria. Additionally, this imagery combines with the symbolism expressed in agate, for the lower portion of her diadem (to her right) is decorated with bands of colour reminiscent exactly of the natural bands of colour inherent in this sacred stone.

In Vedic philosophy, the combination of fire with the snake symbol represents the ability of fire to suddenly flare up, as if via the spinal system from a relaxed and coiled to an erect state; in Vedic terms, such is termed 'raising the Kundalini' and symbolises the same capabilities in man raising energy via the chakra system of the spine for spiritual transcendence and communication with the divine (the solar logos). Interpretation of the 'winged Medusa with Serpent locks' as a solar symbol, connoting spiritual transcendence and representative of the Mysteries aligns with research carried out by Dwyer (1992: 249-282) on the figural iconography of the Tazza Farnese - perhaps the most famous artefact ever to have been carved from sardonyx. He sees the 'gorgone' (winged) engraved on the obverse of the Tazza

Esoteric interpretation of the symbol of the Cornucopia offers even deeper insight into the sacred context of the narrative expressed in the imperial state cameo of the Emperor Caracalla. Hall (1974: 75) interprets this symbol as 'the horn of plenty.' A large horn with its mouth overflowing with the abundance of the earth's fruit, is the

350

Valerie Worden: Image and Symbol in Severan State Portrait Cameos

attribute of many virtues, beneficient personifications as well as Gods and Goddesses. However Hall surmises (primarily from Ovid Fastii 5.121-4; Metamorphoses 9. 85-92) that 'its true origin is probably to be found in the ancient belief that power and fertility resided in the goat's or bull's horn'. The author supports the ancient belief, enlarged upon by Fontana (1993, 124) that the Cornucopia or horn of plenty symbolizes the union of male and female , being both phallic and empty. Thus it can read as the symbol of the 'hermaphrodite', which in the spiritual sense represents a crucial stage in the process of spiritual initiation. This unification of male and female principles is common to numerous myths and legends, from Isis and Osiris to Orpheus and Eurydice, and is expressed at a deeper level in the Indian cult of Tantra. Fontana (1993: 32) argues that the major spiritual and occult traditions (particularly those of an eastern genesis) have taught that completion of the spiritual journey can be achieved only internally by the union of the female and male principles that we each carry within us - symbolising the active and the passive. The Cornucopia held by Caracalla in his left hand features 'eyes' on the Cornucopia's rim, thus combining these two powerful symbols with the sacred connotations implicit in the agate as a medium. This symbology may have been intended as a projection of Caracalla's allegiance to the practice of a Mystery cult which expounds the belief in immortality. If his religious affinities centred around an allegiance to the Mithraic cult, he would have been particularly conscious of the esoteric concept of this symbol, for the eye is the symbol of 'Pater', the seventh stage of initiation. Such symbolism is depicted in mosaics dating from the second half of the third century AD from the Mitreo de Felicissimo, Ostia (Vermaseren 1956-60: 1, CIMRM 299). Caracalla therefore grew up in an 'overtly religious' climate and inherited his parents' determination to develop the cult of the solar god in Rome - a dream that he no doubt intended to communicate via the medium of the imperial state cameo as a Severan dynastic metaphor. This commitment to Severan religious policy was ongoing, therefore, as the durability of the Cult of Elagabel was subsequently to demonstrate

Throughout, this discussion has been conducted on two levels. Initially a diachronic investigation of the religious background of both Septimius and Julia Domna established their allegiance to eastern Mystery cult. Secondly an art historical description of the cameos themselves was followed by an investigation of religious and/or metaphysical symbolism detectable via the complex imagery. This second area of discussion has it is hoped established the case for regarding much of the symbolic content as deriving from the ideologies or mystery cult religions of the east.

Conclusion

Birley, A., 1971. Septimius Severus, the African Emperor. London: Eyre and Spottiswoode. Boissevain, U.P., (ed.), 1898-1931. Dio: Cassii Dionis Cocceiani Historiarum Romanarum Quae Supersunt. Berlin: W eidmannos. Casson, L., 1979. The Periplus Maris Erythraei. New Jersey: Princeton University Press. Cumont, F., 1956 (Engl. Trans.) The Oriental Religions in Roman Paganism. New York: Dover. Dalley, S., 1991. Myths from Mesopotamia. Oxford: Oxford University Press. Davidson, D. and Aldersmith, H., 1924. The Great Pyramid. Its Divine Message. An original coordination of historical documents and archaeological evidences. Vol 1. London: Williams and Northgate.

Exploration of the sacred nature of the narrative expressed on these gems has also suggested fresh insight into the iconography expressed on the 'Berlin'Tondo, the rare Egyptian portrait roundel depicting the Severan family. Given that a mystical, as well as geological, connection with the east has been posited in published studies of individual objets d'art, such as the Tazza Farnese, this paper reinforces Dwyer's analysis of the Hermetic influence on this enigmatic plate and the attribution he makes to its manufacture at the Ptolemaic court in Alexandria. Whilst the Pan Orientalists have established that Eastern philosophical ideas had considerable influence on Roman culture, research into oriental origins has tended in the past to overlook the possible influences of Vedic philosophy on western Mystery cult tradition. The 'science of gems' (ratnasastra) derives from India, as did banded/orbicular agate,and sardonyx. It is suggested here that esoteric symbols and images associated with Vedic belief travelled westwards with them over a considerable period. The symbols of the 'homed god'(symbol of the heavenly bull), the uncoiling snake, and the potent image of the 'eye' as connected with the solar phenomenon, and the concept of metempsychosis, would all appear to have their earliest genesis in Rig-Vedic culture.

BIBLIOGRAPHY

This paper has attempted to identify and expose the concept that narrative expressed in Severan state portrait cameos, when executed on the medium of sardonyx, informs us of an inherent pre-occupation with the expression of sacred iconography relating directly to Roman religious beliefs infused with eastern overtones. The value of the gemstone in Roman ideology is apparent from literary, archaeological and art historical evidence. Agate was particularly used symbolically, in all of its banded, orbicular and related sardonyx forms. This gemstone had a special religious and metaphysical significance when carved into iconographic masterpieces in the worlds of the Ptolemies and Caesars.

351

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

Distelberger, R., 1991. Illustrated Guide to the Secular and Ecclesiastical Treasures of the Kunsthistoriches Museum, Vienna. Vienna: Kunthistorisches Museum. Dwyer, E.J., 1992. The Temporal Allegory of the Tazza Farnese. The American Journal of Archaeology. 96. pt. I: 249-82. Evans, J., 1922. Magical Jewels of the Middle Ages and Renaissance. Oxford: Clarendon Press. Fontana, D., 1993. The Secret Language of Symbols. London: Duncan Baird. Gonda, J., 1969. Eye and Gaze in the Veda. Verhandelingen der Akademie van Wetens. Nieuwe Reeks - Deel LXXV. Amsterdam-London: North Holland Publishing Co. Gonda, J., 1972. The Vedic God Mitra. Leiden: Brill. Gonda, J., 1980. Vedic Ritual. Leiden: Brill. Graves, R., (trans) 1990. Apuleius, The Golden Ass. London: Penguin. Groneberg, B., 1986. 'Die Sumerische - Akkadische Inanna/Istar: Hermaphroditos? Die Welt des Orients. 17. Hall, J., 1974. Dictionary of Subjects and Symbols in Art. London: John Murray. Henig, M., 1983. A Handbook of Roman Art. Oxford: Phaidon. Hohl, E., (ed.), 1965. The Augustan History. Leipzig: Tenbeur. Lee, D., (trans. and ed.) 1977. Plato. Timaeus and Critias. 3rd • ed. London: Penguin Lurker, M., 1980. The Gods and Symbols of Ancient Egypt. London: Thames and Hudson. Mundkur, B.,1983. The Cult of the Serpent: An interdisciplinary survey of its manifestations and origins. Albany: State University Press, New York. Plantzos, D., 1998. 'medius liquidus astris' Capricorn, Aquarius and Pisces on Graeco-Roman Amulets. Jewellery Studies, 8: 37-48. Pliny the Elder. Natural History. X.37. ed.D.E.Eichholz (1989). Cambridge, Mass: Harvard University Press. Robinson, J.M., (ed.), 1988. The Naghammadi Library in English 3rd • ed. Leiden: Brill. Sankarananda, Swami, 1943/44. Rig Vedic Culture of the Prehistoric Indus. 2 vols. Calcutta: Abhedananda Academy of Culture. Squarciapino, M.F., 1974. Sculture del Foro Severiano di Leptis Magna. Rome: Monografie de Archaeologica Libica X. Stevenson, E.L. (ed. and trans.), 1991. Ptolemy, The Geography. London: Constable. Strong, E., 1915. Apotheosis and Afterlife. London: Constable. Toynbee, J.M.C., 1971. Death and Burial in the Roman World. London: Thames and Hudson. Turton, G., 1974. The Syrian Princesses. The Women who ruled Rome A.D. 193-235. London: Cassell. Vermaseren, M.J., 1956-60. Corpus Inscriptionum et Monumentorum Religionis Mithriacaei 2 vols. Hagae: Nijhof£

Vermaseren, M.J., 1963. Mithras the Secret God. London: Chatto and Windus. Ward-Perkins, J.B. 1993. The Severan Buildings of Lepcis Magna.An Architectural Survey. London: Society for Libyan Studies Monograph. 2. Worden, V., 1992. 'Symbolism in Ancient Near Eastern Jewellery'. Unpubl. BA dissertation, University of Manchester. Worden, V., 1997. The Eye of the Agate'. A Diachronic Investigation into the Eastern origins and Sacred Symbolism of Agate in the Classical World. Unpublished MA thesis, Univerisity of Manchester.

352

Valerie Worden: Image and Symbol in Severan State Portrait Cameos

33.1 Portrait Cameo of the Severan Family, 198-209 AD (Photograph courtesy of the Cliche Bibliotheque Nationale de France.)

33.2 Portrait Cameo of Julia Domna, early third century AD (Photograph courtesy of the British Museum)

353

Archaeology of the Roman Empire: A tribute to the life and works of Professor Barri Jones

33.3 The Berlin Tondo portrait of the Severan family c. 199-201 AD (photograph courtesy ofBildarchiv Preussischer Kulturbesitz, Berlin)

33.4 The Apotheosis of Caracalla c. 217 AD (photograph courtesy of the Bibliotheque Municipale, Nancy)

354

Valerie Worden: Image and Symbol in Severan State Portrait Cameos

33.5 The Spirit of Alexandria Mosaic, third century BC, in the Greco-Roman Museum, Alexandria (photograph Barri Jones)

355