The Frontiers of Imperial Rome 9781848844278, 1848844271

Table of contents :
List of Illustrations
List of Plates
Acknowledgements
Technical Matters
Abbreviations
Select List of Roman Emperors by Dynasties
Introduction
Part I: Sources
Chapter 1: The Frontiers
Chapter 2: An Overview of the Sources
Chapter 3: The Romans on Frontiers
Chapter 4: The Romans on Frontier Installations
Chapter 5: Regulations and Treaties
Chapter 6: The Building Blocks of Frontiers
Part II: The Frontiers
Chapter 7: Linear Barriers
Chapter 8: River Frontiers
Chapter 9: Desert Frontiers
Chapter 10: Mountain Frontiers
Chapter 11: Sea Frontiers
Chapter 12: Forests, Marshes and Swamps
Chapter 13: The Deep Frontier: Defence-in-Depth?
Part III: Interpretation
Chapter 14: The Development of Frontiers
Chapter 15: Military Deployment
Chapter 16: A Comparison of Frontiers
Chapter 17: Decision Making
Chapter 18: How Did Frontiers Work?
Chapter 19: The Purpose and Operation of Roman Frontiers
Chapter 20: Were Roman Frontiers Successful?
Conclusions
Further Reading
Sites to See
Notes
Index

Citation preview

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page i

The Frontiers of Imperial Rome

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page ii

For Bill Hanson and Lesley Macinnes

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page iii

The Frontiers of Imperial Rome David J. Breeze Maps prepared by CHC, University of Salzburg, Austria

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page iv

First published in Great Britain in 2011 by Pen & Sword Military an imprint of Pen & Sword Books Ltd 47 Church Street Barnsley South Yorkshire S70 2AS Copyright © David J. Breeze 2011 ISBN 978-1-84884-427-8 The right of David J. Breeze to be identified as Author of this Work has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988. A CIP catalogue record for this book is available from the British Library. All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording or by any information storage and retrieval system, without permission from the Publisher in writing. Typeset in 11pt Ehrhardt by Mac Style, Beverley, E. Yorkshire Printed and bound in the UK by CPI Pen & Sword Books Ltd incorporates the imprints of Pen & Sword Aviation, Pen & Sword Maritime, Pen & Sword Military, Wharncliffe Local History, Pen and Sword Select, Pen and Sword Military Classics and Leo Cooper. For a complete list of Pen & Sword titles please contact PEN & SWORD BOOKS LIMITED 47 Church Street, Barnsley, South Yorkshire, S70 2AS, England E-mail: [email protected] Website: www.pen-and-sword.co.uk

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page v

Contents

List of Illustrations ............................................................................................ vii List of Plates ..................................................................................................... ix Acknowledgements ............................................................................................... x Technical Matters ............................................................................................. xii Abbreviations ................................................................................................... xiv Select List of Roman Emperors by Dynasties ....................................................... xv Introduction .................................................................................................... xvii Part I: Sources ................................................................................................. 1 Chapter 1

The Frontiers ............................................................................. 3

Chapter 2

An Overview of the Sources ....................................................... 7

Chapter 3

The Romans on Frontiers ......................................................... 14

Chapter 4

The Romans on Frontier Installations ....................................... 25

Chapter 5

Regulations and Treaties ........................................................... 30

Chapter 6

The Building Blocks of Frontiers .............................................. 34

Part II: The Frontiers ................................................................................... 53 Chapter 7

Linear Barriers ......................................................................... 55

Chapter 8

River Frontiers ......................................................................... 92

Chapter 9

Desert Frontiers ..................................................................... 118

Chapter 10

Mountain Frontiers ................................................................ 133

Chapter 11

Sea Frontiers .......................................................................... 146

Chapter 12

Forests, Marshes and Swamps ................................................. 159

Chapter 13

The Deep Frontier: Defence-in-Depth? .................................. 161

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page vi

vi The Frontiers of Imperial Rome Part III: Interpretation ............................................................................... 165 Chapter 14

The Development of Frontiers ................................................ 167

Chapter 15

Military Deployment .............................................................. 172

Chapter 16

A Comparison of Frontiers ..................................................... 177

Chapter 17

Decision Making .................................................................... 181

Chapter 18

How Did Frontiers Work? ....................................................... 184

Chapter 19

The Purpose and Operation of Roman Frontiers ..................... 194

Chapter 20

Were Roman Frontiers Successful? ......................................... 206

Conclusions ..................................................................................................... 209 Further Reading .............................................................................................. 213 Sites to See ..................................................................................................... 223 Notes .............................................................................................................. 226 Index .............................................................................................................. 234

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page vii

List of Illustrations

Frontispiece. Map of the Roman empire ........................................................ xxii 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

A fort gate at Maryport carved on a stone .................................................... 9 The Rudge Cup .......................................................................................... 9 Geophysical survey of Maryport (survey by TimeScape) ............................ 11 Inscription of Commodus from Lower Pannonia ....................................... 27 Plan of the unfinished first-century legionary base at Inchtuthil (Scotland) ................................................................................................. 40 Plan of Housesteads fort on Hadrian’s Wall (Britain) ................................. 41 An artist’s impression of a fort. Drawn by Michael J. Moore ...................... 42 An artist’s impression of a fortlet. Drawn by Michael J. Moore .................. 42 An artist’s impression of a tower. Drawn by Michael J. Moore ................... 43 Plans of the late-first-century fort at Elginhaugh and the fourth-century fort at Altrip .............................................................................................. 45 Map of the frontier in Upper Germany and Raetia .................................... 57 Plan of the fort at Hesselbach on the frontier in the Odenwald (Germany) ... 60 An artist’s impression of a tower on the German frontier. Drawn by D. Baatz .................................................................................................... 60 Map of Hadrian’s Wall in its landscape setting ........................................... 62 Plans of a milecastle and a tower on Hadrian’s Wall ................................... 64 Plan of the fort at Chesters on Hadrians Wall ............................................ 66 An artist’s impression of Hadrian’s Wall. Drawn by Michael J. Moore ....... 69 Map of the Antonine Wall in its landscape setting ...................................... 72 Plan of the fort and annexe at Rough Castle on the Antonine Wall (Scotland) ................................................................................................. 74 An artist’s impression of the Antonine Wall. Drawn by Michael J. Moore ..... 75 Diagram illustrating the development of the limes in Upper Germany ....... 78 An artist’s impression of the frontiers in Upper Germany and Raetia ........ 79 Diagram illustrating the development of Hadrian’s Wall and the Antonine Wall .......................................................................................................... 81 Map of the frontier in Lower Germany ..................................................... 94 Map of the late Roman frontier on the Danube-Iller-Rhine line ................. 97 Plan of Divitia/Cologne (Germany) ........................................................... 98 An artist’s impression of a fortified landing place ....................................... 99 Map of the frontier in the Middle Danube region .................................... 101

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page viii

viii The Frontiers of Imperial Rome 29. Plan of the fortlets/fortified landing places at Ulcisa Castra/Dera Patak and Contra Florentiam/Dunafalva ............................................................. 107 30. Plan of the fort at Contra Aquincum/Budapest (Hungary) ........................ 108 31. Plan of the fort at Eining (Germany) ........................................................ 109 32. Plan of the late forts at Boiodorum/Passau (Germany) and Drobeta/ Turnu Severin (Romania) ........................................................................ 110 33. Plans of towers at Mains Rigg, Asperden, Budakalasz, Goldsborough ....... 111 34. Map of the northern section of the eastern frontier in Cappadocia and Syria ....................................................................................................... 113 35. Map of the southern sector of the eastern frontier in Syria and Arabia ..... 121 36. Plan of el-Lejjun (Jordan) ........................................................................ 123 37. Plan of Qasr Bshir (Jordan) ...................................................................... 124 38. Map of Tripolitania ................................................................................. 127 39. Map of the Nile Valley and the Eastern Desert of Egypt .......................... 130 40. Map of Dacia .......................................................................................... 134 41. Map of the towers and barrier to the north-west of the province of Dacia ....................................................................................................... 135 42. Map of the frontier in Mauretania Caesariensis ........................................ 140 43. Map of the eastern end of the Black Sea .................................................. 149 44. Map of the north-western corner of the Black Sea ................................... 151 45. Map of the Saxon Shore .......................................................................... 154 46. Plans of the Saxon Shore fort at Portchester ............................................ 156 47. Plan of the Saxon shore fort at Pevensey .................................................. 157 48. Plans of fortlets dating to the early empire in Britain and Germany .......... 168

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page ix

List of Plates

1. Coins of the emperors: a) Hadrian; b) Antoninus Pius; c) Septimius Severus; d) Constantine I and e) Valentinian I 2. The landscapes of frontiers: a) the River Euphrates in Turkey; b) the River Danube in Austria; c) the Carpathian Mountains in Romania; d) the Atlas Mountains in Morocco 3. A Roman fort depicted on Trajan’s Column 4. Legionaries building on Trajan’s Column 5. A tower on Trajan’s Column 6. A supply ship on Trajan’s Column 7. Auxiliary soldiers 8. A writing-tablet from Vindolanda 9. Posts in the German palisade 10. The posts of a timber tower at Utrecht 11. The German limes from the air 12. A reconstructed tower and stretch of palisade on the German frontier 13. A reconstructed Roman gate at the fort at the Saalburg (Germany) 14. Hadrian’s Wall at Cawfields 15. Hadrian’s Wall at Cuddy’s Crag 16. Pits on Hadrian’s Wall at Byker looking east 17. The Antonine Wall at Rough Castle (Scotland) 18. The Hutcheson Hill distance slab (Scotland) 19. The fort gate at Traismauer (Austria) 20. Eining (Germany) from the air 21. The legionary fortress at Satala (Turkey) 22. The fort at Qasr Bshir (Jordan) 23. The cistern at Qasr Bshir (Jordan) 24. The fort at Hân al-Manqoûra (Syria) from the air 25. The city at Dura-Europos (Syria) from the air 26. The headquarters building at Lambasesis (Algeria) 27. The south gate at Bu Ngem (Libya) 28. A fort in the Eastern Desert of Egypt

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page x

Acknowledgements

have been lucky to have worked intensively on two frontiers in Britain, Hadrian’s Wall and the Antonine Wall; fortunately, both retain the ability to surprise through new information obtained from excavation or as the result of research in the study. I have also visited the whole of the European frontier from Katwijk on the Dutch coast to the Black Sea with the exception of modern Serbia, much of the eastern frontier but only a small section in North Africa. In spite of several visits to some areas, I would not admit to having more than a limited knowledge and am grateful to the many friends and colleagues who have helped me visit frontiers and understand them better and who have discussed frontier problems with me over many years, and in particular those who have read and commented on earlier drafts of this book as well as colleagues who have helped with references. These include Dietwulf Baatz, Alan Beale, Paul Bidwell, James Bruhn, Walter Cockell, Orietta Cordovana, Brian Dobson, Beryl Elliott, Bill Errington, George Findlater, Phil Freeman, Erik Graafstal, Bill Hanson, Nick Hodgson, Sonja Jilek, Rebecca Jones, Valerie Maxfield, Ioana Oltean, Andrew Poulter, Boris Rankov, Alan Rushworth, Sebastian Sommer, Matthew Symonds, Andreas Thiel, Zsolt Visy and Willem Willems. Jackie Henrie has applied her considerable skills as a copy editor to the text, for which I am most grateful. I am also grateful to Phil Sidnell of Pen & Sword for inviting me to write this book. While I was working on this book I was invited by Professor Bill Hanson to give the Dalrymple lectures at the University of Glasgow. This was useful in helping me to shape and firm up many ideas which are detailed here and I am grateful to the members of the audience for their feedback. I should like to thank Richard Beleson for financial assistance with the illustrations, and the following for help with providing the illustrations: Jim Devine of the Hunterian Museum, Kurt Schaller of the University of Salzburg who prepared all the maps, Graeme Stobbs who provided several plans, and Andreas Thiel. I am grateful to the following publishers for permission to reproduce translations: Professor A. R. Birley; Professor Brian Campbell; Dr. J. S. Johnson; Gerald Duckworth & Co. Ltd (Bristol Classical Press); The publishers and Trustees of the Loeb Classical Library, Cambridge, Mass.: Harvard University Press, for the quotations from volumes 2, 111, 176, 249, 263 and 454, by the President and Fellows of Harvard College. Loeb Classical Library is a registered trademark of the

I

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xi

Acknowledgements xi President and Fellows of Harvard College; Liverpool University Press; Penguin Books Ltd.

Picture acknowledgements AirFotos, Newcastle upon Tyne pl. 14; Bayerisches Landesamt für Denkmalpflege (photograph Otto Braasch) pl. 20; Culture 2000 Project Frontiers of the Roman Empire Frontispiece; Deutsche Limeskommission pls. 9 and 11; Dietwulf Baatz figs. 12, 13, 21, 22; William S. Hanson fig. 10a; Historic Scotland fig. 19, pl. 17; Hunterian Museum pls. 1a–e, 18; Simon James pl. 25; Angus Lamb pls. 3–6; Michael Mackensen pl. 26; Valerie Maxfield 41, pl. 28; Michael J. Moore figs. 7–9, 17, 20; Belinda Ratnayake pl. 2d; Roman Society fig. 5; Römisch-Germanisches Zentralmuseum, Mainz pl. 7; Kurt Schaller/University of Salzburg figs. 11, 14, 18, 24, 25, 28, 34, 35, 38, 39, 40, 42–45; Society of Antiquaries of Newcastle upon Tyne figs. 15, 16; Graeme Stobbs figs. 10b, 26, 29–33, 37, 46, 47, 48; TimeScape fig. 3; Tyne and Wear Archives and Museums pl. 16; Utrecht pl. 10; Vindolanda Trust pl. 8. Other photographs are copyright of the author.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xii

Technical Matters

ll dates are within the CE (Common Era, AD as it used to be known) unless otherwise stated to be BCE (Before the Common Era, BC). In most cases, both the Roman and modern names of a site are given when known, e.g. Vindonissa/Windisch, the Roman being in italics. Some sites such as Alexandria, Petra and Vindolanda are so well known by their ancient names that these have not been italicised. Measurements are in metric with imperial (i.e. American/British) in brackets. Roman measurements are given when they are significant. Ancient sources are cited in translation in the text with the source of the translation referenced in the section on Further Reading; they are referenced by author, book, chapter and verse. Many translations are now available on-line. End notes have been provided for the citing of secondary sources, but I have not sought to reference every statement relating to books listed in the Further Reading. Publications which appear in the Further Reading are referenced in the end notes by author, date and page reference; other publications appearing in the end notes are cited in full. There is frequent mention of specific Roman emperors. Rather than give their dates every time, a list of the main emperors is provided below. There are different words for the edge of a state: border, boundary, frontier. I have generally used ‘frontier’ in connection with military works and ‘border’ or ‘boundary’ to describe the legal boundary. There should undoubtedly be more qualifying words such as ‘possibly’, ‘probably’ and ‘perhaps’ throughout the text, but I have sought to indicate the degree of uncertainty behind the statements made below.

A

Maps Kurt Schaller and I have tried to make the maps as accurate, clear and consistent as possible. Sometimes, these  aspirations clash. For example, although most of the forts on Hadrian’s Wall and the Antonine Wall were attached to the linear barrier, for cartographic ease they are shown depicted as if not attached. On the maps a triangle indicates a legionary fortress, a square an auxiliary fort and a circle a town; a black square indicates certain occupation, a white square probable/possible occupation, and a grey square occupation at a different time from the main period of the map. In order not to over-burden the maps, fortlets and towers are generally not marked. Where, on occasion, a different convention is used, an explanation is given in the caption. Only sites mentioned in the text are named. Main rivers

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xiii

Technical Matters xiii and seas are also named and significant roads shown. Provincial boundaries of the second century are marked, as on the map of the empire appearing as the frontispiece. There will inevitably be disagreement about the dates of occupation of some forts and it is all but certain that not all the forts shown on each map were occupied at the same time; we have done our best with the available evidence. As far as possible, maps (and plans) are reproduced at the same size for comparative purposes.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xiv

Abbreviations

AE BAR CIL IGR ILS IRT Limes RIB SEG O P

L’Année Epigraphique British Archaeological Reports (IS = International Series; SS = Supplementary Series) Corpus Inscriptionum Latinarum Inscriptiones Graecae ad res Romanas pertinentes Inscriptiones Latinae Selectae Inscriptiones Romaines d’Tripolitania The proceedings of the Congress of Roman Frontier Studies. The Roman Inscriptions of Britain Supplementum Epigraphicum Graecum Ostraca Papyri

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xv

Select List of Roman Emperors by Dynasties

The Julio-Claudians (family name either Julius or Claudius) Augustus 27 BCE-14 CE Tiberius 14–37 Gaius 37–41 Claudius 41–54 Nero 54–68 Flavians (family name Flavius) Vespasian 69–79 Titus 79–81 Domitian 81–96 The ‘good emperors’ and the Antonines Nerva 96–98 Trajan 98–117 Hadrian 117–138 Antoninus Pius 138–161 Marcus Aurelius 161–180 Lucius Verus 161–169 Commodus 180–192 The Severans (family name Severus) Septimius Severus 193–211 Caracalla 211–217 Elagabalus 218–222 Severus Alexander 222–235 The third century Maximinus Thrax Gordian III Philip Decius Gallus

235–238 238–244 244–249 249–251 251–253

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xvi

xvi Lost Legion Rediscovered

Valerian Gallienus Aurelian Probus

253–259 253–268 269–275 276–282

The Tetrarchy Diocletian Maximian Constantius I Chlorus

284–305 286–305 305–306

Galerius

305–311

Licinius

308–324

House of Constantine Constantine I Constantine II Constans

306–337 337–340 337–350

Constantius II

337–361

Julian

361–363

House of Valentinian Valentinian I

364–376

Valens I Gratian Valentinian II

364–378 367–383 375–392

House of Theodosius Theodosius I Honorius Arcadius

379–395 395–423 395–408

Emperor in the East Emperor in the West Caesar in the West 293–305, Emperor 305–306 Caesar in the East 293–305, Emperor 305–311 Emperor in the East

Emperor in the West Emperor in the centre and after 340 also the West Emperor in the East and after 350 all the empire Caesar 355–360 and then Emperor in the West 360–361

Senior Emperor and Emperor in the West Emperor in the East Emperor in the West Emperor in the centre and after 383 also the West

Emperor in the East Emperor in the West Emperor in the East

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xvii

Introduction

n the last day of 406, a large force of Alans, Suebi and Vandals crossed the frozen River Rhine and struck deep into Gaul and then on into Spain. Previous invasions had been defeated by the Roman army, but the army was powerless to halt this incursion. The attack – or rather its success – opened the last chapter of the western empire and seventy years later its emperor was deposed and sent into retirement. A few years earlier a similar event had occurred in the east when the Goths defeated a Roman army at the Battle of Adrianople in 378 and killed the emperor. On both occasions, the narrow band of military frontier installations proved to be ineffective against an attacking army, as they had been before. This brings into sharp relief the central question relating to Roman frontiers: were they constructed to protect the Roman Empire from invasion or was their intention more linked to frontier control, that is preventing illicit entry and raiding, small-scale threats? This distinction is with us today. Barriers such as the Berlin Wall and the Israeli fence and the installations which lie along the border between Mexico and the USA were all erected to control the movement of people not armies. While the operation of campaigning armies has changed, not least in the age of long-range missiles, the concern with controlling the movement of individuals across frontiers has not. The vital questions at the heart of this book are therefore: how did Roman frontiers operate and what was their purpose and role? This is no small task. The frontier works pass through twenty modern countries in three continents. The body of archaeological evidence is considerable. The chronological span is a minimum of four centuries. Innumerable Roman emperors were involved in decision making, to a greater or lesser degree. Some, like Augustus, were expansionist, others, of which Hadrian is the prime example, were more ‘defensive’ in their policies; all were governed by a determination to protect the interests of Rome. This book, however, is not about emperors, except insofar as they affected the frontier arrangements, nor about Roman politics or foreign policy, though most books purporting to be accounts of Roman frontiers actually concentrate on such aspects. Nor is it about the Roman army, though an understanding of the army is essential to a full appreciation of the frontiers. The focus is firmly on the frontier installations themselves. An appreciation of the nature of the military works which together formed the frontiers and their development is essential to a greater understanding of how the frontiers worked and what their purpose was. A further aim of this book is to help

O

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xviii

xviii The Frontiers of Imperial Rome readers and visitors understand one of the great monuments of the Roman Empire, its frontier. That frontier, over 7,500km (4,800 miles) long, defined the Roman Empire and is the single largest monument surviving from the Roman world. The evidence used to understand the frontier includes literary sources and other documents such as the records on papyri and the writing tablets, inscriptions, sculpture, the fruits of archaeological excavation and survey, and the frontier works themselves. Today, the most visible and prolific element of all these sources of evidence is the archaeological site which is the frontier. This, therefore, will be the focus of the book, all the other evidence being used to help build up a picture of its development, use and collapse. I have tried to avoid getting caught up in too much detail of fort names and discussion of vague evidence of foundation dates but have concentrated on broader patterns which in turn allow specific interpretations. Indeed, the archaeological evidence is essential to an understanding of Roman frontiers for it forms the single largest body of evidence. As we will see, literary sources are sparse on the details of frontiers. Ancient writers tell us that frontiers existed, but very little about details of military installations other than that there were forts and soldiers to protect the empire. Most books on Roman frontiers tend to be written by ancient historians who are not familiar with the archaeological evidence. Susan Mattern in her excellent Rome and the Enemy (1999) devotes only six pages to frontier installations. C. R. Whittaker, while making more use of the archaeological evidence, focuses more strongly on the literary evidence and wider interpretations. Benjamin Isaac’s The Limits of Empire. The Roman Army in the East (1990) is wider than its title suggests, offering wide-ranging views on frontiers, but the approach is also primarily from the literary sources. N. J. E. Austin and B. Rankov in Exploratio, Military and Political Intelligence in the Roman World from the Second Punic War to the Battle of Adrianople (1995) offer a far more integrated approach to the literary and archaeological evidence, thus providing the best discussion of how frontiers worked. We may also note that frontiers hardly figure in the collections of sources such as Brian Campbell’s The Roman Army, 31 BC–AD 337: A Source Book (1994). At times, one senses almost a feeling of despair among ancient historians that archaeologists have been surveying and excavating the frontiers of the Roman Empire for one hundred years or more and have not been able to interpret them. Mattern goes so far as to say, ‘recent scholarship has argued that the purpose of Roman frontiers is uncertain in all cases’, while Isaac offers a scenario of the changing nature of a hypothetical military site with the despondent conclusion that none of the changes could be recognized archaeologically and therefore the function of Roman forts similarly cannot be determined.1 The implication is that the archaeological evidence can be ignored, an attitude Mark Driessen has characterised as ‘limes denial’. In my view the statements by Mattern and Isaac are overly pessimistic; it is possible to determine the purpose of Roman frontiers. This can be achieved through study of the literary and epigraphic sources combined with analysis of the archaeological evidence. There are continuing and new discussions by many archaeologists and historians about the interpretation of the frontier installations, admittedly not all of whom are in

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xix

Introduction xix agreement, but most of this has taken place in learned journals and specialist books, not least the reports of the three-yearly meetings of the International Congress of Roman Frontier Studies and the handbooks which usually accompany these meetings, and no general synopses have appeared. The nearest approach is Stephen Johnson’s Late Roman Fortifications, published over twenty-five years ago but restricted to Europe and covering only a hundred years or so. One aim of this book, therefore, is to seek to redress the balance and offer the archaeological perspective, which, it is hoped, will illuminate the wider debate about the nature and purpose of Roman frontiers. To paraphrase Paul Kendrick, the mute stones can speak. The remains of Roman frontiers still survive in today’s landscapes. Many are presented and interpreted for visitors, museums display the artefacts found in these sites and guide-books abound. I hope that this book will encourage people to go to visit the sites and museums. The remains of Roman frontiers are best seen and understood in their topographical settings even though these have been modified over the last 2,000 years. It is impossible to write without consciously or unconsciously choosing language which supports one’s own views. So, I should declare my background and ‘prejudices’. Educated at Durham University under the tutelage of Eric Birley, Brian Dobson and John Mann, I have a particular view of the Roman army as a welltrained, well-disciplined and well-organized fighting force, used to fighting and winning in the field, with a strong infrastructure which enabled it to move about the country relatively freely. This is not to say that I do not recognize that the Roman army could come under severe pressure from its enemies, that it lost battles – though it rarely lost wars – and that its individual members could be corrupt; Juvenal in one of his Satires said that no civilian would dare to beat up a Roman soldier and if beaten himself would fail to obtain redress from a military court (Juvenal, Satire 16, The Advantages of Army Life).2 For eight centuries, the Roman army was the pre-eminent fighting force in Europe. It did not acquire that position by sitting behind walls. Thus, for me, the Roman army was an offensive fighting machine, and I interpret its remains with that in mind; forts, accordingly, were primarily bases from which troops moved out to fight, though they could also operate as guard points. The army was also very bureaucratic, maintaining detailed records and even known to issue a receipt in quadruplicate. This equally helps to colour my interpretation of the purpose of Roman frontiers. The span of time encompassed by this book runs from Augustus, Rome’s first emperor who ruled from 27 BCE to 14 CE, through to the end of the fourth century. Although the Western Empire staggered on until 476, the events of 406 are a useful terminal date for this consideration of frontiers. The Eastern Empire was to continue for about a thousand years, but the catastrophic defeat of the Roman army at the Battle of Adrianople in 378 is a convenient watershed, coinciding with the permanent split of the empire into two parts. The book is divided into three parts. The first is primarily concerned with the sources. It starts with a consideration of the range of literary references for the study

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xx

xx The Frontiers of Imperial Rome of Roman frontiers and then reviews the Romans’ own comments on frontiers before moving on to a description of the building blocks of frontiers, the military installations which together formed the frontiers and the army which built, manned and maintained them. The second part is the core of this book, the frontiers themselves. The vast amount of material has to be organized and presented in one way or another, not least to avoid repetition. The framework I have chosen is to divide the discussion into sections based on the type of frontier or the terrain through which frontiers ran, and within each section consider the evidence chronologically. Roman frontiers reached their peak of development in the middle years of the second century, from the early part of the reign of Hadrian, who introduced the linear barrier into the repertoires of frontiers, to the last years of his successor, Antoninus Pius, that is the period from about 120 to shortly after 160. This period of forty years, and the linear barriers built then, will therefore form the start point for the study of all Roman frontiers. In order to understand the location of the Hadrianic frontiers, it is necessary to consider the earlier history of conquest and occupation in the areas where they were built. This discussion runs forward to the end of the fourth century. The third and final part is an interpretation of the evidence already presented, together with additional documentary references, mainly inscriptions. The development of frontiers is discussed, together with military deployment. Consideration is given to questions such as how frontiers worked, the distinctions between the different frontiers, who took the relevant decisions before moving on to a consideration of wider issues and whether we can recognise a ‘grand strategy’ behind the frontiers of the Roman Empire. Archaeologists will appreciate that I have modelled the book on an excavation report: a review of existing knowledge; the evidence, with discussion at the end of each section; interpretation. As in an excavation report, illustrations are required. However, in this case, the individual sites are not important, the significance lies in the patterns of forts, fortlets and towers. The maps have been prepared with that in mind, though the text acknowledges individual fort names when the sites are significant. Sites of all types and periods are represented in the drawings which illustrate the best examples of each. It is necessary to acknowledge, also, what is not in this book. The frontiers could not have existed without an efficient method of supply of men, food, equipment, pottery and so on. With reluctance – because this is a subject in which I am particularly interested – I have resisted the temptation to explore this aspect of frontiers.3 The existence of so many men in these military installations resulted in the growth of cities, towns and villages along the frontiers.4 Rudyard Kipling did not exaggerate so much when he said: But the Wall is not more wonderful than the town behind it. Long ago there were great ramparts and ditches on the South side, and no one was allowed to build there. Now the ramparts are partly pulled down and built over, from end

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xxi

Introduction xxi to end of the Wall; making a thin town eighty miles long. Think of it! One roaring, rioting, cock-fighting, wolf-baiting, horse-racing town, from Ituna on the West to Segedunum on the cold eastern beach!5

With regret, this town and its counterparts around the empire are also ignored in these pages. This exclusion is also disappointing because the study of the different ethnicities of soldiers and civilians on the Roman frontier is one of the important areas of current research. The impact of frontiers on the indigenous people on both sides of the border, a subject which has attracted a lot of attention over the last two decades, is similarly important, but beyond the scope of this review of frontiers.6 I have also omitted discussion of the late defences in the Julian Alps, the Long Walls of Thrace and other, less well understood, defensive features such as those on the Great Hungarian Plain and at Galaţi in Romania.7 That said, there is certainly enough material to produce a full account of Roman frontiers, their nature, history, development and function. Finally, it must be acknowledged that this review is more heavily biased towards Europe than I would like. This is partly because of the intensive nature of the archaeological research which has been undertaken there over several centuries as well, generally, as the greater complexity of the military remains. This is underlined by the fact that each of the land frontiers, the German limes, Hadrian’s Wall and the Antonine Wall, is a large, complex monument in its own right and particularly in relation to the individual sites which combine to form the frontier along the rivers or in the deserts fringing the empire. In particular, the two British frontiers have a distinct advantage in that their various significant elements touch each other and therefore a chronological sequence can be determined.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xxii

Roman provinces and frontiers in the middle of the second century AD

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xxiii

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page xxiv

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 1

Part I

Sources

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 2

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 3

Chapter 1

The Frontiers

here is no such creature as a typical Roman frontier. Each section relates to the countryside through which it runs, the enemy, the builders, the materials of construction locally available, the date it was built, existing installations, and how it was surveyed and laid out. A hundred years ago, Lord Curzon, former viceroy of India, discussed many of these issues – and more – in a series of lectures on frontiers.1 He acknowledged that this subject had not been much considered before, but at that time was growing in importance. He noted that most of the treaties of his day between sovereign states concerned the definition of frontiers. Curzon characterized frontiers as being natural or artificial. In the former he included the sea, deserts, mountains, rivers and features such as forests, marshes and swamps. His description of artificial frontiers encompassed linear barriers, but also a broad zone of separation, ‘a razed or depopulated or devastated tract of country’, and buffer states. He also noted the role of the protectorate or, as we would call them in the Roman world, the client or friendly kingdom, on frontiers, and spheres of influence. Finally, he appreciated the concept of ‘hinterland’, that a new conquest has its own hinterland which forms part of the new acquisition. For Curzon, the landscape appears to have been the most significant criterion and so it was for the Romans. In many places, the boundary of their empire was defined by the sea. This is particularly noteworthy in North Africa and western Europe, where the Atlantic Ocean and the English Channel/La Manche provided a clear edge to the empire, but the Red Sea and the Black Sea were also significant in the East. Beyond the Channel lay two large islands, Britain and Ireland. Roman arms only extended to the former and here, again, Roman dominion was brought to the edges of the island, except in the north. A major river was adopted as the de facto imperial boundary in several places. The most important rivers were the Rhine, Danube and Euphrates which formed the greater part of the frontier around the whole of the northern half of the empire for the greater part of its life. In the 70s, the frontier in Germany was moved beyond the headwaters of the Rhine and Danube, but, when this sector was abandoned in the 260s, the new line then adopted followed a route along the rivers Danube, Iller and Rhine (pl. 1). In the Middle East and in North Africa, that is the southern half of the empire, deserts offered a defining boundary, except in Egypt as Nubia lay further up the Nile. No powerful state lay beyond the empire in the Sahara Desert, nor beyond

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 4

4 The Frontiers of Imperial Rome modern Jordan, that is Roman Arabia. But east of the northern section of the eastern frontier lay Parthia, the greatest enemy faced by Rome. Here the Roman frontier ebbed and flowed as Roman emperors captured one or other of the oasis cities or were forced to relinquish them by the Parthians or their successors, the Persians. Holding such cities was also the key to the control of further conquests as each formed a stepping stone to the next. Mountains could offer a strategic boundary, such as the Carpathians which enveloped the province of Dacia, modern Romania. But there are always passes through mountains and in Romania we have recently been discovering the ways in which the Roman army controlled these routes by the erection of barriers across the passes. Even in areas technically beyond the boundary of the empire, the army sought to control passes through mountains such as the Caucasus between the Black Sea and the Caspian Sea. In some places, there was no obvious boundary. Thus, in north Britain, Germany and parts of North Africa, artificial barriers were constructed. That in Germany ran for 550km (330 miles) and included a stretch of 43km (25 miles) where the River Main and its tributary the Neckar formed the frontier. In Britain, Hadrian’s Wall was only 130km (74 miles) long and the Antonine Wall in modern Scotland half that length, though in each case installations continued along the river banks to west and east. Each such artificial frontier took the form of a linear barrier, of timber, earth, turf or stone, which, in effect, defined the limit of the empire. The countryside, or at least its geology, dictated the materials used to build Roman frontiers. The German frontier was of timber and when that decayed it was replaced by either an earthen bank or a stone wall. The Fossatum Africae in modern Algeria was partly of mud brick and partly of stone, both being used in the construction of the forts. Hadrian’s Wall in Britain was built of both stone and turf, while the Antonine Wall was primarily of turf and timber but with some fort buildings of stone or at least on stone footings. One of the interesting aspects of Roman frontiers is that they are all different. The available materials of construction play a part in these differences, but only a part. A stretch of the Upper German frontier built under Antoninus Pius was mathematically straight for over 80km (48 miles), totally ignoring the changes in the landscape through which it passed; the adjacent section of frontier acknowledged the landscape. Hadrian’s Wall was erected in the 120s in relation to an existing series of military installations and this affected its location. On the line of the Antonine Wall, however, there were no such forts and the line was more sinuous and more closely related to the topography through which it ran. Recent research has demonstrated that these two British frontiers were laid out in very different ways.2 The relationship of a frontier to the ‘enemy’ is more difficult to determine. The great nineteenth-century German ancient historian Theodor Mommsen suggested that the two British frontiers appeared to be more defensive than their German counterparts because the enemy pressed more strongly upon the borders in Britain.3 In this book I will offer an alternative reason for these differences. Along the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 5

The Frontiers 5 Danube, the lightly held stretches of the frontier would appear to relate to the lack of enemy across the river, which itself related to the topography. The definition of what constitutes a ‘frontier’ is of some interest. The word ‘frontier’ is often used as if it is defined as a single line on a map. In fact, while this may have been the case along certain borders of the empire, for example along the major rivers of the empire, it was certainly not the case elsewhere. In several provinces there was a broader military frontier area than a simple, single line and each case was unique. Further, outposts existed beyond many frontiers. Moreover, in the first century at least, some land across the Rhine was retained for the use of the army. In the desert regions, forts were located where there was water and thus were distributed over a wide area. In North Africa, forts might be pushed as far out into the desert as possible along traditional routes, though again respecting oases. A line drawn on a map joining the outermost forts looks seductively like a frontier, but this is a mirage for the links of these forts went back to the settled areas not along a ‘frontier’ line. In these circumstances, it is difficult to know where the exact boundary of the empire lay. Did the military frontier coincide with the legal frontier? Even if we can assume that a river formed the border, Rome might seek to control activities immediately beyond. This can be seen in the treaties agreed in the 170s between Marcus Aurelius and the Marcomanni, Quadi and Iazyges to the north of the Danube whereby a cleared zone was established north of the river. Moreover, the Romans had a habit of seeing their allies beyond their formal border as being part of their empire, which is an additional complication. It is also important to acknowledge that the relationship between Rome and her neighbours was not one between equals, except in the case of Parthia and her successor Persia. Rome was a world power, in our terminology a super power, and she acted as such, crossing her frontiers to retaliate against attacks on her territory and her people and in some areas maintaining outposts beyond her boundaries and even a military presence on the territory of adjacent states. There is perhaps one further aspect to the definition of frontiers, the so-called ‘frontier zone’, an area in which the Romans and the indigenous population both within and without the empire were complicit in the maintenance of life in the frontier region. As Owen Lattimore noted, the degree of economic integration between conqueror/occupier and the native population is of paramount importance: upon that rests the ability of the occupying force to maintain its position.4 Yet, while accepting that view, my primary concern in this book lies with the military installations which lay on or close to the boundary of the Roman Empire, accepting both that these could form part of a wide military zone and also that the frontier works did not necessarily constitute the boundary of the empire. Some words on terminology are essential. The Romans themselves had various words for frontiers and their components. Inscriptions from North Africa and Germany use the word fines to denote a boundary. In both literature and epigraphy limes is used to designate a land boundary – not a frontier – of the empire, with ripa

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 6

6 The Frontiers of Imperial Rome designating the river boundary. Usage changed over the centuries. Limes, at first a road, by the beginning of the second century had come to be used to describe the boundary of the empire, and later a frontier district, such as the limes Tripolitanus, the Tripolitanian frontier. Both Hadrian’s Wall and the Antonine Wall were called vallum on inscriptions and murus in literature. In the early-third-century road book, the Antonine Itinerary, occurs the phrase, a limite, id est a vallo, ‘from the boundary, that is the Wall’, which does suggest that Hadrian’s Wall was the boundary of the province even though there were Roman forts beyond it. A river frontier was a ripa. The frontier area might be known as the praetentura, the forward area.5 The Romans obviously knew where the boundary of the empire lay, even if this is not clear to us. The Notitia Dignitatum, dating to about 400, recorded forts as being in barbarico, that is outside the empire. Descriptions of the contents of treaties or the actions of emperors and governors demonstrate that they knew the precise line between Rome and barbaricum, and when they crossed it. The individual types of installations had their own names: Tacitus called a legionary base and an auxiliary fort castra. In the Antonine Itinerary, castra is a fort and praesidium something smaller, perhaps a fortlet. In the Eastern Desert of Egypt castra is used for the main base and praesidium for the outposts, though praetensio was used for an outpost in Arabia. The Diocletianic inscriptions use praesidium for both forts and a fortlet, a garrison in fact, as used by Tacitus. A fort could also be a castellum (which was also used in Africa to describe a non-military civil settlement) or, in the late empire, a munitio. On inscriptions, burgus is either a tower or a fortlet, sometimes a tower sitting within an enclosure, while turrem appears on some inscriptions to describe a tower. Today, a camp is the term used to describe a temporary fortification: a marching camp occupied for a few nights; a labour camp used by a building party; a siege camp; or a practice camp. A fort is a permanent structure. A legionary base is often referred to as a fortress. A small fort held a small unit or a detachment drawn from a larger unit, and a fortlet usually no more than about eighty men.6 ‘Limes’ has also come into general use today, especially in Germany where it is the generic term for the Roman frontier. Elsewhere, it is used to describe sections of frontier such as the Limes Ponticus, the Black Sea coast, and the Limes Transalutanus, the barrier across, that is to the east of, the River Olt (Alutus in Latin) in modern Romania.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 7

Chapter 2

An Overview of the Sources

he frontiers of Rome do not feature much in the literary sources for the Roman Empire. There is only one clear reference each to the construction of Hadrian’s Wall and the Antonine Wall, and one slightly less clear reference to the German limes, but all were written some 200 years after the events to which they relate. Some statements are so opaque to us today that historians still argue about their meaning. The reference to the creation of a palisade under Hadrian, for example, is not directly linked to the description of his visit to Germany. Roman historians rarely provide an account of the destruction of military installations, merely that the frontier was crossed, and certainly no description of the nature of the frontier crossed. Significant accounts of frontiers do, however, survive. The governor of Cappadocia in the 130s, Arrian, has left a unique account of his tour of that part of his province fringing the eastern shore of the Black Sea in which he reported on his inspection of the forts as well as describing the actions he ordered for their better protection and the political affiliations of the local kings. Literary sources can inform us about how frontiers worked. Two references by the historian Tacitus writing in the late first and early second centuries described the regulations which operated on the frontier in Germany for those wishing to enter the empire. The History of Rome by Cassius Dio, written in the early third century, includes much useful information about treaties and other actions on the Danube frontier, particularly relating to the wars of Marcus Aurelius. They indicate that here at least the army sought to prevent and control settlement beyond the frontier. There is little surviving material relating to the third century, but for the second half of the fourth century we have the detailed accounts of Ammianus Marcellinus. He wrote about the actions of the emperors on frontiers and the measures they took to protect the empire. Other commentaries provide descriptions of the Roman army at work, such as Julius Caesar’s Gallic War, which provides an account of the building of the siege works at Alesia in the mid-first century BCE. This has been drawn on by archaeologists for centuries seeking to understand their own discoveries. Like all written statements, our ancient sources for Roman frontiers have to be analysed and not necessarily taken at face value. Not only did Caesar and Tacitus consciously write for a specific readership and to justify their own actions, or rather in the case of Tacitus the actions of his class, but the ancients’ understanding of the writing of history was very different from our own. All writers were – and are – subject to the influence of the times in which they lived. An essential fact about

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 8

8 The Frontiers of Imperial Rome Tacitus’ book on Agricola is that it was about his father-in-law, while it is important to acknowledge that Ammianus Marcellinus wrote during the reign of the Emperor Theodosius which means that we should be cautious about his description of the exploits of the emperor’s father, Count Theodosius. Certain documents provide considerable evidence about the deployment of the Roman army. On retirement, an auxiliary soldier could request and obtain a statement of his privileges. The statement took the form of a bronze tablet known today as a diploma. These are a valuable source of information because each includes a list of all units in the particular soldier’s province or army group discharging men at that time. In this way, the size and deployment of provincial armies can be traced from the late first century into the early third century. The diplomas thereby offer a comment on the scale of the local threat, though this might be internal rather than external. We also have an amazing survival from the end of the fourth century, the Notitia Dignitatum. This lists all military and civilian appointments in both the eastern and western parts of the empire. The arrangement of the document into these two parts indicates the date, which is about 400, though there is still discussion about how it was prepared and whether it includes earlier, out-of-date information.1 The inscriptions erected by the Roman army constitute a rich source of information. They can tell us who ordered the construction of a frontier or a fort, when the building work was undertaken, who built it, sometimes how much was built, and, occasionally, why it was built and, sometimes, the names of the individual elements, while concentrations of inscriptions of a particular date can indicate the interest of an emperor in an area. In 1853, a great project was commenced, to record all the inscriptions of the Roman Empire in a single series of volumes, the Corpus Inscriptionum Latinarum (CIL). This remains of incomparable value though long supplemented by annual reports of new discoveries in L’Année Epigraphique (AE), and by many volumes relating to provinces or modern countries. In certain parts of the empire, more vulnerable documents have survived written on papyrus, wooden tablets or stylus tablets. These provide an incomparable insight into the operation of the army, though rarely do they describe frontier installations themselves. The papyri are found in the eastern provinces of the empire, but over the last fifty years writing tablets have been found on the line of Hadrian’s Wall at Vindolanda and Carlisle and at Vindonissa/Windisch in Switzerland (pl. 8). These newly discovered documents have two significant elements: they not only inform us about local activities but also, by their broad similarity to the material from the East, allow us to use these documents on papyrus confidently as providing a general picture of life on all frontiers. In North Africa, ostraca, records or messages on fragments of pottery, are valuable in recording details of frontier life. Coins are of great value in two ways. They are pictures providing evidence of actions, such as an emperor visiting individual frontiers, which can then be related to other evidence. Their discovery in archaeological contexts also helps date the construction of individual sites and the changing nature of their occupation.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 9

An Overview of the Sources 9 Trajan’s Column and the Column of Marcus in Rome, both dating to the second century, illustrate the Roman army in action (pls. 3–6). They also illustrate buildings, including forts. Trajan’s Column is particularly clear in its depiction of forts with turf ramparts and timber towers over the gates. A fort gate is shown on a stone at Maryport in Britain (fig. 1) and a drawing of a fort incised on a stone at Bu Ngem in Libya, Fig. 1: A fort gate at Maryport carved on a for example. Mosaics show city walls, stone. while it is possible that a schematic view of Hadrian’s Wall is shown on the Rudge Cup (fig. 2). Several ceramic models and brooches of fort gates and towers also survive.2 One problem for those wishing to study the building, occupation and abandonment of frontiers is that the literary sources and inscriptions in particular provide information on the date of the building and restoration of frontiers, but they rarely inform us when frontiers were abandoned, nor of the physical nature of the frontier installations. It is archaeological research which provides such information. The study of Roman frontiers might be said to have commenced immediately after the fall of the Roman Empire. Medieval writers offered their own views on the date and purpose of frontiers, recently abandoned, but whose date of construction had already been forgotten. In the fifteenth and sixteenth centuries antiquarians started to visit frontiers and forts and record their discoveries. The eighteenth century saw the development of a more scientific approach, including the development of more sophisticated surveys of the surviving remains; the first guide-books to frontiers were produced too. Excavations also started about the same time, though these were extremely rudimentary by modern standards. In the eighteenth century, the Empress Maria Theresa, for example, ordered the investigation of the Great Baths in Aquincum beside Budapest. Such work continued elsewhere on frontiers throughout the nineteenth century, but it was the 1890s which saw the beginning of modern excavation across Europe. In 1890, the Glasgow Archaeological Society set out to determine if the Antonine Wall really had been built of turf. Two years later, the modern series of excavations Fig. 2: The Rudge Cup. The design is generally began on Hadrian’s Wall. In the same year, taken to be a depiction of Hadrian’s Wall.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 10

10 The Frontiers of Imperial Rome Theodor Mommsen founded the Reichslimeskommission with the purpose of surveying and recording the linear barriers from the Rhine to the Danube. A similar body was established in neighbouring Austria in 1897. The interests of these frontier scholars expanded from their home bases. Rudolph Brünnow and Alfred von Domaszewski examined the southern sector of the eastern frontier in 1897–8 and produced a seminal publication.3 The detailed survey of frontiers from the 1890s into the 1930s, coupled with selective excavation, provided the basis of much of our modern understanding of the dating and development of frontiers. More recently, the role of such individual surveyors and excavators has been continued by the foreign schools of archaeology. The study of the artefacts found on frontiers has become increasingly sophisticated. Coins, pottery and small finds – brooches, arms and armour, etc. – have been used to help date sites for over a hundred years. Now such material can be used to help us understand supply, movement, ethnicity, the manning of frontiers, and other elements of frontier life. Museums were established along the frontiers to house and present inscriptions, sculpture and the finds from excavations. The National Museum was founded at Budapest in Hungary in 1802, Leiden in the Netherlands in 1818, Bonn in Germany in 1820, Cluj in Romania in 1859 and Carnuntum in Austria in 1904. Kaiser Wilhelm II went further and funded the reconstruction of the perimeter walls and some internal buildings of the fort at the Saalburg near Frankfurt, establishing a museum in a rebuilt granary. Since that date, many towers have been rebuilt on the German frontier and other elements of forts there and in other countries. The twentieth century saw the development of a new tool to study frontiers, aerial photography. On the eastern frontier, from 1925 to 1932 Antoine Poidebard recorded the military remains of Syria from the air. His work inspired Marc Aurel Stein to undertake similar work in Iraq and Jordan in 1938–9. In the late 1940s, Jean Baradez revolutionized our knowledge and appreciation of the African frontier and coined a new phrase, the Fossatum Africae. Even on frontiers with a long tradition of survey and recording, new information came to light, in Britain particularly through the work of J. K. St Joseph and in Germany Otto Braasch. It was as a result of aerial survey that both fortlets and camps were found along the line of the Antonine Wall in Scotland in the late 1940s.4 The fall of the Berlin Wall in 1989 led to the opening of new possibilities in eastern Europe, for aerial survey was now possible. Some work had been done earlier, in particular by Zsolt Visy, in tracing the frontier installations in Hungary, but the removal of many restrictions led to more flying. New forts and camps were discovered; the observation of a Roman marching camp halfway across Slovakia leads to the real possibility that such camps may be found as far north as Poland. New discoveries continue to be made, not just because the area available for study expands, but also because the process requires repeat visits in different weather conditions and crops in order to achieve the best results. Such programmes are now supplemented by satellite images, which are particularly useful in the desert areas.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 11

An Overview of the Sources 11

Fig. 3: Geophysical survey of Maryport showing the fort bottom left and the civil settlement stretching north-eastwards (survey by TimeScape).

A further approach to the study of frontiers was developed from the middle of the twentieth century, geophysical survey. There are two main methods, resistivity which measures impulses sent into the ground, and magnetometry which records the magnetic signal. The result is that, to put it simply, ‘hard’ features have a different response to ‘soft’. In this way, roads, walls, ditches and other intrusions are planned. New information on fort plans has come to light, but perhaps more significant have been the results of surveying the areas outside the fort walls. Extensive civil settlements have been discovered, revolutionizing our understanding of the position of the fort in its immediate landscape.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 12

12 The Frontiers of Imperial Rome Metal detecting, too, can lead to the discovery of artefacts which help us to locate structures – or even more amorphous remains such as battlefield sites as has happened in Germany, particularly with the discovery of the putative site of the Varus disaster of 9. An important scientific tool of a different nature is dendrochronology, the study of tree rings. Individual timbers from forts or frontiers can be dated, sometimes with startling results. For example, a timber used in the construction of the Hadrianic palisade in Germany has been dated to 119/20, the year before Hadrian’s visit, while the construction of the Antonine palisade in the province of Raetia has been dated to the 160s, most of the dendrochronological dates falling into the years following the death of Antoninus Pius in 161.5 In general, scientific disciplines play an increasingly greater part in archaeological research, providing evidence on diet and disease, the environment of the frontier zones and land use both inside and outside forts, and even the ethnic origins of soldiers and their dependents. As a result of the study, recording, surveying and excavation of Roman military structures over several centuries, a considerable body of evidence has been accumulated about the structural remains themselves, their building, the nature of their occupation and their decay and destruction. It is possible to analyse the development of complex frontiers such as Hadrian’s Wall and the Antonine Wall, refine the date of the building of the German palisades, study the nature of the occupation of each frontier, record the various destructions of forts and towers, and seek to relate all to historical events recorded in the ancient literature. None of these activities take place in a vacuum. Each interpretation grows out of previous views which have been developed about Roman frontiers. These form, in the memorable phrase of Leo Rivet, ‘the opinions of modern scholars’.6 They are our interpretations of the operation and purpose of Roman frontiers and they change over time as we learn more both in the field and through research and discussion. Yet we are all bound by the past; it is difficult to forget or shrug off what we learned from our predecessors and start the thinking process afresh. We are in a real sense governed by the past. Fortunately, it is possible to review the evidence and reach new conclusions by continuing to study the reports of our predecessors, by participating in the challenge of discussion, and by visiting the remains of the frontiers. In fact, in order to understand frontiers, it is essential to visit them and see the position of frontiers, forts and towers in the landscape; there is no alternative, the surviving remains and the topography within which they sit have to be observed in person. This is doubly important because so often the surviving remains and their location in the landscape form the most direct evidence for their function. One forum for discussion on frontiers is the International Congress of Roman Frontier Studies founded by Eric Birley in 1949. This usually meets every three years, each meeting taking place on a different frontier, and its proceedings are published: they are listed at the end of this book. Over the last twenty-four years the programme of the Congress has been increasingly thematic and the themes

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 13

An Overview of the Sources 13 particularly relevant to this book are listed after the reference for each Congress volume. It is worth emphasizing both that new evidence continues to come to light about frontiers and that there are significant difficulties in creating a pattern of military deployment in every period. Many forts were situated in prime locations which are still occupied, making tracing their history difficult. This is especially true of the earliest foundations which are buried most deeply and therefore most difficult to locate and explore, not helped by the army’s habit in the first century of moving its forts around. The existence of some sites is only hinted at by the discovery of pottery, coins, artefacts and tombstones. Excavation and geophysical survey, it is to be hoped, will in time confirm such indications. It is also noteworthy that research on Roman frontiers does not take place in a uniform way across the empire. Interest in Roman military remains fluctuates within each modern country over time and is also difficult to undertake in some parts of the empire. Yet, as the world continues to open up, more research takes place and new techniques are increasingly applied across the whole empire.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 14

Chapter 3

The Romans on Frontiers

t would be incorrect to say that there were no frontiers to Rome’s empire during the Republic. Obviously, the territory governed by Rome had boundaries. But the Republic, for most of its existence, did not maintain a standing army to protect its empire. This changed in the first century BCE as the Republic, led by its ambitious generals, entered a new era of expansion. They not only conquered vast areas, such as Gaul, but brought many states under the control of the Republic, as did Pompey in the East.

I

Augustus and the first century The expansion of the late Republic was haphazard. Caesar invaded Gaul, leaving to his rear unpacified Alpine tribes, while north-west Spain also remained unconquered. It was his heir, Augustus, who demonstrated a long-term view, remarkable in view of his delicate health. In his testament, the Res Gestae, he listed his achievements. These included the following: I extended the territory of all those provinces of the Roman people on whose borders lay peoples not subject to our government. I brought peace to the Gallic and Spanish provinces as well as to Germany, throughout the area bordering on the Ocean from Cadiz to the mouth of the Elbe. I secured the pacification of the Alps from the district nearest the Adriatic to the Tuscan seas, yet without waging an unjust war on any people. My fleet sailed through the Ocean eastwards from the mouth of the Rhine to the territory of the Cimbri, a country which no Roman had visited either by land or sea, and the Cimbri, Charydes, Semnones and other German peoples sent ambassadors and sought my friendship and that of the Roman people. At my command and under my auspices two armies were led almost at the same time into Ethiopia and Arabia Felix; vast enemy forces of both peoples were cut down in battle and many towns captured. Ethiopia was penetrated as far as the town of Nabata, which adjoins Meroë; in Arabia the army advanced into the territory of the Sabaeans to the town of Mariba. I added Egypt to the empire of the Roman people. Greater Armenia I might have made a province after its king had been killed, but I preferred, following the model set by our ancestors, to hand over that kingdom to Tigranes. … When the same people later rebelled and went to war, I subdued them and handed them over to be ruled by Ariobarzanes. … The Pannonian peoples … were conquered … I brought them into the empire of the Roman

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 15

The Romans on Frontiers 15 people, and extended the frontier (fines) of Illyricum to the banks of the Danube. When an army of Dacians crossed the Danube, it was defeated and routed under my auspices, and later my army crossed the Danube and compelled the Dacian peoples to submit to the commands of the Roman people. Embassies from kings in India were frequently sent to me; never before had they been seen with any Roman commander. The Bastarnae, Scythians and the kings of the Sarmatians on either side of the River Don, and the kings of the Albanians, Iberians and the Medes sent embassies to seek our friendship. The following kings sought refuge with me as suppliants: Tiridates, King of Parthia, and later Phraates, son of King Phraates; Aravasdes, King of the Medes; Artaxares, King of the Abiabeni; Dumnobellaunus and Tincommius, Kings of the Britons; Maelo, King of the Sugambri; … rus, King of the Marcomanni and Suebi. Phraates, son of Orodes, King of Parthia sent all his sons and grandsons to me in Italy, not that he had been overcome in war, but because he sought our friendship by pledging his children. (Augustus, Res Gestae 26–7. Translation by V. Ehrenberg and A. H. M. Jones.) This statement not only illustrates the violent, military nature of Roman statecraft and foreign policy, but also that the Romans liked to fight a just war; whenever possible a war was portrayed as being started by the other side. Here also Augustus offered many examples of his actions which demonstrate different ways of protecting his empire: • • • • •

he sent out expeditions beyond the empire to explore; he provided examples of boundaries, the sea and rivers, the Danube and the Elbe; he secured control of foreign kingdoms through imposing his own ruler; he fostered diplomatic relations with his neighbours; he welcomed refugees to his empire who might be used later with advantage in their own countries; • he used hostages from foreign kingdoms as a means of ensuring peace on his borders; • he dealt expeditiously with any invasion of his empire; • and, significantly, he dealt with internal problems so that he could take action more comfortably on the frontier.

Few of these actions were new. During the third war against Mithridates of Pontus, Lucullus and Pompey had explored the East in the 70s and 60s BCE, the latter almost reaching the Caspian Sea. In the following decade, Caesar twice invaded Britain. He also acknowledged the role of rivers, stopping his conquest of Gaul at the Rhine. For decades, Rome had interfered with foreign kingdoms, supporting pro-Roman rulers, sometimes with an army, while welcoming their children in Rome to be educated. The difference with Augustus was his far-sighted approach. In particular, he dealt with the awkward problems ignored by his predecessors such as subduing

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 16

16 The Frontiers of Imperial Rome the Alpine tribes and completing the conquest of the Iberian peninsula in order to bring peace within his new boundaries.1 Considerable discussion has surrounded the wider aspects of Augustus’ policy.2 It has been argued that, after dealing with the ‘internal’ problems such as conquering the Alpine tribes, completing the conquest of Spain and bringing Roman arms to the Danube, Augustus sent out expeditions to explore the lands beyond the empire. Once he had assured himself that there was no threat – or land to conquer – in North Africa, up the Nile or across the Red Sea, and bearing in mind the existence of the strong state of Parthia on his eastern boundary, he decided on a policy of expansion in Europe. As a result Roman arms were brought to the River Elbe. But was that to be his final frontier in Europe, or merely the first step to a wider expansion? Certainly, from what we know of contemporary Roman maps, Augustus may have believed that it was possible to continue his advance and conquer the whole known world. The poet Virgil, writing under Augustus, offered the contemporary view of Roman expansionism, of Roman aims: ‘Rome will extend her empire to the ends of the earth and her heroism to the sky’ (Virgil, Aeneid VI, 756), in short have ‘power without end’ (Virgil, Aeneid I, 279). It is possible that Augustus himself did not know what his final intentions were but was feeling his way forward. It may be firmly stated that within this framework the building of solid frontiers such as were created in the second century did not exist. Indeed, it may be, as Tony Birley has suggested, that the long-term aim of Augustus ‘was to dispense with the necessity for any frontiers at all, to rule an empire without end, an imperium sine fine’.3 In that case, the frontiers – the subject of this book – are symbols of abdication and failure to complete the conquest of the known world rather than of the continuing strength and dominance of Rome.4 In effect, Augustus’ testament is not a statement about frontiers or even frontier policy, it is a statement of imperial expansion. Augustus added more territory to Rome’s empire than anyone before or later. All the more surprising is the codicil added to his testament, ‘to keep the empire within its boundaries’. This is recorded by both Tacitus and Dio, though in slightly different versions. To his version, quoted above, Tacitus added that, ‘he [Augustus] either feared dangers ahead, or he was jealous’ (Tacitus, Annals 1, 11). Dio’s version is a little more helpful. He stated that Augustus had recommended that his successors ‘be satisfied with present possessions and in no way to seek to increase the area of the empire’ (Dio 56, 33). Whittaker has discussed both versions of the statement. He has linked Dio’s version to two aspects of Augustus’ actions. Whittaker reminds us that ‘despite the many wars beyond the Rhine and Danube, Augustus never claims in the Res Gestae that he had added a single new province to the Roman Empire’. His achievement was to bring order to what he had inherited. Moreover, Augustus distinguished between the territory ruled directly by Rome and the wider empire which contained client states. His advice therefore could be interpreted as suggesting that the empire itself should be kept within its existing boundaries but that his successors could continue to extend Rome’s control over the peoples beyond the formal limits of the empire.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 17

The Romans on Frontiers 17 Whittaker links this to Augustus’ statement that he had never wished to acquire anything except when compelled to do so and interprets Augustus’ advice as indicating that he is justifying his own expansion of the empire by the argument that he only acquired new territory when forced to do so by the provocative actions of others and he recommended that this policy also was continued.5 It is within the framework of Augustus’ expansion of Roman power and territory that the references to bringing Roman arms to the banks of the Danube have to be interpreted. In his next statement Augustus recorded that his armies crossed the river to invade Dacia, forcing the Dacian peoples to submit to the commands of the Roman people. Augustus did not appear to believe that the Danube should reflect the limit of Roman power, though it may have been the limit of Roman territory at the time. The Elbe is clearly another river of consequence. Augustus refers to the mouth of the Elbe being at one end of his dominions. In the references to the conquest of Germany through his reign, the Elbe is frequently mentioned as the (current) limit of Roman arms. The river was reached by Drusus in 9 BCE. Subsequently, it was crossed and a treaty of friendship made with the peoples beyond. A frontier based on the Elbe would presumably have brought all the land between the Rhine, the North Sea, the Elbe and the Danube within the empire. And yet we have no knowledge that this would have been the limit of Augustus’ ambitions. Almost a hundred years after the death of Augustus, Tacitus wrote his history of Rome. In his account of the reign of Tiberius, describing the events of 23, Tacitus provides a description of the army, noting the different types of units, their location and the reasons behind their positioning. He states: Italy was guarded by two fleets, one on each of the seas on either side, at Misenum and Ravenna; the adjacent coast line of Gaul was protected by the warships which Augustus had captured in his victory at Actium and stationed at Forum Julii with a strong complement of rowers. Our principal force, however, was close to the river Rhine where eight legions were stationed as a defence against Germans or the Gauls. The Spanish provinces, recently subdued, were in the charge of three legions. King Juba had received the people of Mauretania as a gift of the Roman people. The rest of Africa was guarded by two legions, and Egypt by an equal number. Then all the great sweep of territory from the borders of Syria to the River Euphrates was kept under control by four legions; bordering on this are the Iberians and Albanians and other kings who were protected against external regimes by the greatness of our power. Rhoemetalces and the sons of Cotys ruled Thrace, while the bank of the Danube was guarded by two legions in Pannonia and two in Moesia. A further two legions were stationed in Dalmatia, geographically behind the other four, and could be summoned from close by if Italy needed rapid help. Rome, however, had its own garrison in the form of three urban and nine praetorian cohorts, recruited mainly from Etruria and Umbria or the old

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 18

18 The Frontiers of Imperial Rome territory of the Latins, and the ancient Roman colonies. In suitable locations in the provinces were stationed allied ships, auxiliary cavalry alae and infantry cohorts, virtually equivalent in strength to the legions. But it is difficult to trace them all since they were moved to different locations as circumstances demanded, and their numbers increased or decreased from time to time. (Tacitus, Annals 4, 5. Translation by B. Campbell.) It will be observed that there was a strong force in or adjacent to Italy. Elsewhere, not all legions were based on the frontier: internal stability was also essential. The weight of frontier dispositions lay along the Rhine, while in some areas border kingdoms played a part in defence. The legions were complemented by auxiliary regiments and the fleets, also organized on military lines. Many provinces contained no army units. In such cases, the soldiers to man the provincial staffs appear to have been provided by the troops in neighbouring provinces. The location of the legions helps us to understand which frontiers the Romans considered the most important, or vulnerable. Analysis of the deployment of auxiliary units through inscriptions or diplomas provides valuable additional information. The fewer the auxiliary units in a province, the more peaceful the province and the less threatened its frontier, and of course vice versa. The mentioning of rivers remains important in the jargon of frontiers. In his praise of Augustus, Tacitus noted that ‘the empire’s frontiers were on the ocean, or distant rivers’ (Tacitus, Annals 1, 9), and this is echoed down the centuries. The Jewish writer Josephus, compiling his history of the Jewish War of 68–70 at the end of the first century, includes in a speech a statement relating to the frontiers of the Roman empire: ‘for all Euphrates is not a sufficient boundary for them on the east side, nor the Danube on the north; and for their southern limit, Libya has been searched over by them, as far as countries uninhabited, as is Cadiz their limit on the west’ (Josephus, The Jewish War 11, 4). Here appear seas, rivers and deserts as the boundaries of the Roman Empire. In the middle of the second century, Appian saw the boundaries of the empire as being the three rivers, acknowledging that ‘going beyond these rivers in places they rule some of the Celts over the Rhine, and the Getae over the Danube, whom they call Dacians’ (Appian, History of Rome, Preface 6). The anonymous writer of a panegyric to an emperor written in the early fourth century refers to the Rhine and the Euphrates as the frontier (Pan. Lat. X (2) 7, 3–7). The Rhine and the Danube remain important in the history of Ammianus Marcellinus written in the late fourth century (for example, Ammianus 17, 5, 5). Rivers are, of course, strong features in the landscape and some of these references may be no more than an acknowledgement of that for it is clear that during the first half of the first century Rome maintained a presence in the territory of the Frisii north of the Rhine delta and further upstream allocated land on the right bank of the river for the use of soldiers in grazing their animals. Elsewhere, Roman army units were based beyond what appears to be the boundary of the empire, above the Nile’s First Cataract, in the Hejaz south of the province of Arabia,

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 19

The Romans on Frontiers 19 and south of the Atlas Mountains, for example. In the fourth century, bridgehead forts were established across both the Rhine and the Danube, though they lay immediately on the far bank of the river. Tacitus offers two somewhat different comments when writing about frontiers. On the one hand, he appears to assume that the boundaries of the empire would continue to expand – ‘the conquest of Germany is taking a long time’, he wrote in 98 (Tacitus, Germania 37). And while he does mention forts on the frontier, it is almost incidental to his wider discussion of the events he describes, such as the Batavian Revolt of 69–70 on the Lower Rhine. On the other hand, in the biography of his father-in-law, the governor of Britain, Julius Agricola, he makes a significant statement: If the spirit of the army and the glory of the Roman name had permitted it, a frontier had been found within Britain itself. For the Firths of Clyde and Forth, carried far inland by the tides of opposite seas, are separated by a narrow neck of land. This was now being securely held by garrisons and the whole sweep of country on the nearer side was secured: the enemy had been pushed back, as if into a different island. (Tacitus, Agricola 23. Translation by A. R. Birley.) It may be that Tacitus was acknowledging the possibility of a pause in the conquest of Britain rather than a complete halt. As it happened, neither the glory of the Roman name nor the spirit of the army would allow a frontier to be established in the island and soon the army swept forward to victory over the Caledonians at a battle fought somewhere beyond the isthmus; Britain, according to Tacitus, was conquered, though he goes on to say that it was let go (Histories 1, 2). But who actually decided what the spirit of the army or the glory of the Roman name would allow? Surely only the emperor could determine this.

The second century Fifty years later, this attitude of continual expansion appears to have been abandoned. Aelius Aristides, a citizen of a Greek city on the western coast of Asia Minor, visited Rome during the reign of Antoninus Pius and wrote an oration in her praise. He acknowledged that the empire was protected by a well-trained army drawn from her provinces rather than Rome herself and based on her frontiers, allowing the empire itself to be largely demilitarized, and he continued: To place the walls around the city itself as if you were hiding her or fleeing from your subjects you considered ignoble and inconsistent with the rest of your concept, as if a master were to show fear of his own slaves. Nevertheless, you did not forget walls, but these you placed around the empire, not the city. And you erected walls splendid and worthy of you, as far away as possible, visible to those within the circuit, but, for one starting from the city, an outward journey of months and years if he wished to see them. Beyond the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 20

20 The Frontiers of Imperial Rome outermost ring of the civilised world, you drew a second line, quite as one does in walling a town, another circle, more widely curved and more easily guarded. Here you built the walls to defend you, and then erected towns bordering upon them, some in some parts, others elsewhere, filling them with colonists, giving these the comforts of arts and crafts, and in general establishing beautiful order. An encamped army, like a rampart, encloses the civilised world in a ring … from the settled areas of Aethopia to the Phasis, and from the Euphrates in the interior to the great outermost island towards the west; all this one can call a ring and circuit of walls. They have not been built with asphalt and baked brick nor do they stand there gleaming with stucco. Oh, but these ordinary works too exist at their individual places – yes, in very great number and, as Homer says of the palace wall ‘fitted close and accurately with stones, and boundless in size and gleaming more brilliantly than bronze’. But the ring, much greater and more impressive, in every way altogether unbreachable and indestructible, outshines them all, and in all time there has never been a wall so firm. For it is a barrier of men who have not acquired the habit of flight. It is they who defend these ordinary walls. … Such are the parallel harmonies or systems of defence which curve around you, that circle of fortifications at individual points, and that ring of those who keep watch over the whole world. (Aristides, Roman Orations 26, 80–4. Translation by J. H. Oliver.) Aristides was not alone in his praise of Rome. A contemporary, Appian, a Greek from Alexandria and author of a history of Rome, said: The emperors, in addition to the original provinces, have added further areas to their rule and have suppressed some which broke away. In general, possessing the best part of the earth and sea they have, on the whole, aimed to preserve their empire by the exercise of prudence, rather than to extend their sway indefinitely over poverty-stricken and profitless tribes of barbarians, some of whom I have seen at Rome offering themselves, by their ambassadors, as its subjects, but the emperor would not accept them because they would be of no use to him. For other people, the emperors appoint kings, not requiring them for the empire. On some of the provinces they spend more than they receive, thinking it shameful to give them up even though they make a loss. They surround the empire with a great circle of camps and guard so great an area of land and sea like an estate. (Appian, History of Rome, Preface 7. Translation by H. White.) Another Greek, Herodian, writing in the early third century, offered a commentary on Augustus’ actions; he remarked: when Augustus established his sole rule, he relieved Italians of their duties, and stripped them of their arms. In their place he established a defensive system

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 21

The Romans on Frontiers 21 of forts for the empire, in which were stationed mercenary troops on fixed rates of pay to act as a barricade for the Roman Empire. He also fortified the empire by hedging it round with major obstacles, rivers and trenches and mountains and deserted areas which were difficult to traverse. (Herodian, History of Rome 2, 11, 5. Translation by C. R. Whittaker.) This is hardly an accurate commentary on Augustus, but it does describe more correctly the actions of the second-century emperors Hadrian and Antoninus Pius, closer to Herodian’s time. The Greeks had a long history of building walls to protect their cities and it is not surprising that educated Greeks continued to think in terms of city walls, even though in these panegyrics the city was now the empire and the walls far removed from the capital. The Romans had a somewhat different perspective. Rome herself had long outgrown her Republican defences and the cities of the western empire were still mostly without walls. These statements all convey, as Valerie Maxfield has said, a ‘tone of total satisfaction with the status quo’.6 They may also contain an element of justification that the empire had stopped expanding, certainly in the statement of Appian. Two hundred years before, Strabo – another Greek – had stated that there was no purpose in Rome taking over Britain because it had nothing to offer. This did not stop Claudius invading the island a couple of generations later. These writers offer their own justification for the abandonment of the intention of the Romans to conquer the whole world: it was not worth having. While it is true that these statements were made at a time when the pace of conquest had slowed, almost to a halt, it is significant that they were made by Greeks who were not of the same expansionist mind as the Romans. The Romans continued to demonstrate by their actions a different mentality. It was said that Marcus Aurelius, had he lived, intended to add two new provinces to the empire, Marcomannia and Sarmatia, and the raising of two new legions may be related to such an action. The details of his treaties in relation to the Marcomanni, Quadi and Iazyges, however, might be thought to belie this.7 Yet the fact that it was said suggests that the possibility of such expansion was not regarded as fanciful. Septimius Severus, moreover, did conquer more territory beyond Rome’s eastern frontier, creating the new provinces of Osrhoene and Mesopotamia. He also set out to complete the conquest of Britain and he pushed forward in North Africa, building new forts beyond the existing frontier. Nor was Severus the last emperor with the ambition to extend Roman rule. A century after Severus, Constantine described himself as propagator imperii Romani, ‘extender of the Roman Empire’.

The late second and third centuries These years saw a rise in the number and scale of invasions of the empire. The first attacks in Europe started in the late 160s, the invasions of the Marcomanni and Quadi who struck as far south as Italy. They were only repelled and subdued with

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 22

22 The Frontiers of Imperial Rome considerable effort by Marcus Aurelius. The accounts of his actions, particularly his treaties with his enemies, are most useful in helping us understand Roman methods of control on frontiers and these will be discussed below. In the wider framework, Edward Gibbon recognized these invasions as the tipping point, starting his Decline and Fall of the Roman Empire here. Throughout the third century there are records of invasions by and wars with German tribes in western Europe, the Goths in eastern Europe and the new Persian Sassanian state also. Sometimes there are the briefest of mentions of actions by emperors to maintain the frontier works, but more evidence comes from inscriptions and through excavations. Emperors, however, in particular from the death of Severus Alexander in 235 to the accession of Diocletian in 284, were often more concerned with maintaining their position against rivals than defending the empire. Invasions penetrated far into the empire and some emperors were strong enough to take the fight to the enemy. Such a man was Probus who had a short reign from 276 to 282. The Life of Probus, admittedly not a very trustworthy source, records his actions in Germany: [Probus], after slaying about 400,000 who had seized upon Roman soil, drove all the rest back beyond the river Neckar and the district of Alba [the Swabian Alp]. … Opposite the Roman cities, moreover, he built forts on barbarian soil and in these he stationed troops. (Historia Augusta, Life of Probus 13. Translation by D. Magie.) [Probus] never ceased fighting until nine princes of different tribes came before him and prostrated themselves at his feet. From these he demanded, first hostages, which they gave him at once, then grain, and last of all their cows and their sheep. It is said, moreover, that he sharply ordered them not to use their swords, since now they might count on protection from Rome in case they must be defended against any foe. It appeared, however, that this could not be accomplished unless the Roman frontier was advanced and the whole of Germany turned into a province. … He took, besides, 16,000 recruits, all of whom he scattered through the various provinces, incorporating bodies of fifty or sixty in the detachments or among the soldiers along the frontier. (Historia Augusta, Life of Probus 14. Translation by D. Magie.) These quotations are interesting because the actions they describe could have taken place at any time over the whole of the history of the empire. They include the driving out of invaders, the occupation of territory, the building of forts beyond the frontier and the acceptance of responsibility for defence, the enforcement of peace through the taking of hostages, the requirement of tribute including food and men, and finally an acknowledgement of a wider strategic appreciation of the situation.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 23

The Romans on Frontiers 23

The fourth century By the fourth century, the Roman army had abandoned all territory beyond the Rhine and the Danube which now became again the frontiers of the empire. The pressure of external invasions and internal civil wars led to the necessity for renewal. Zosimus commented on the actions of Constantine: Constantine likewise took another measure, which gave the barbarians unhindered access into the lands subject to the Romans. For the Roman Empire was, by the foresight of Diocletian, everywhere protected on its frontiers … by towns and fortresses and towers, in which the entire army was stationed; it was thus impossible for the barbarians to cross over, there being everywhere a sufficient opposing force to repel their inroads. But Constantine destroyed that security by removing the greater part of the soldiers from the frontiers and stationing them in cities that did not require protection; thus he stripped those of protection who were harassed by the barbarians and brought ruin to peaceful cities at the hands of the soldiers, with the result that most have become deserted. He likewise softened the soldiers by exposing them to shows and luxuries. To speak plainly, he was the first to sow the seeds of the ruinous state of affairs that has lasted up to the present time. (Zosimus 2, 34. Translation by N. Lewis and M. Reinhold.) This is a splendid piece of hostile reporting. Diocletian may have renewed frontiers, but he did not invent them as the passage implies, nor did Constantine remove all soldiers; rather, he built on Diocletian’s creations, including the field armies. Towards the end of the fourth century, two important books were written. Ammianus Marcellinus was a serving soldier and wrote a history of the Roman Empire of which only his account of the fourth century survives. The other author is anonymous, his surviving text being about Roman war machines, de rebus bellicis, written, it is believed, about 368. He continued the earlier theme about the nature of Roman frontiers, though in this case noting the natural features which protected the enemies of Rome: First of all, it must be recognised that frenzied native tribes, yelping everywhere around, hem the Roman Empire in, and that treacherous barbarians, protected by natural defences, menace every stretch of our frontiers. For these peoples to whom I refer are for the most part either hidden by forests or lifted beyond our reach by mountains or kept from us by the snows; some, nomadic, are protected by deserts and the blazing sun. There are those who, defended by marshes and rivers, cannot even be located easily. (On war machines 28. Translation by R. Ireland.) The section of Ammianus’ history on the second half of the fourth century contains many references to the Rhine and Danube as the frontiers of the empire, and he

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 24

24 The Frontiers of Imperial Rome wrote of the actions of Julian in the 350s and Valentinian in the 360s and 370s on frontiers, recording the building of forts in particular, as we shall see below. He also offers a comment on the frontier policy of Valentinian. ‘It was’, he said, ‘a greater service for the emperor to keep the barbarians in check by frontier barriers than to defeat them in battle’ (Ammianus 29, 4). Finally, the Life of St Severinus written by Eugippius in 511 describes how frontiers ended, or at least the frontier in Noricum ended. Severinus was a churchman, but also leader in the province during the last days of the frontier. He arrived in Noricum in about 460 and lived there until his death at Faviana/Mautern in 482. During those years, he organized the defence of the province, negotiated with Germans to stop their raids, and secured the return of captured Romans. Only two forts are mentioned as retaining troops, Boiodurum/Passau and Faviana/ Mautern, though other forts appear to have been defended by civilians (fig. 28). When no money arrived to pay the soldiers, some of those based at Passau travelled to Italy to collect it but were killed on the journey by bandits. Eugippius describes the relentless pressure along the frontier, the gradual capture of Roman forts and the abandonment of others. Austura/Zeiselmauer, east of Faviana/Mautern, was lost in the 460s. Following the capture of Ioviacum/ Schlögen and Bodiorum/Passau to the west in the 470s, Severinus evacuated all towns upstream from the former legionary fort at Lauriacum, bringing his flock to live in the small sector between Lauriacum and Faviana/Mautern. Shortly after the death of Severinus in 482, this was lost and in 488, Odoacer, king of Italy, abandoned all territory between the Alps and the Danube, ordering its inhabitants to leave for Italy. Thus ended 500 years of Roman rule on the Danube.8

Conclusions We can see in these references different views of frontiers over time. The reign of Augustus was a period of expansion and it is boundaries and not frontiers which feature in these accounts, rivers being prominent as demarcation lines. Towards the end of the first century a significant distinction starts to emerge. Tacitus stated that the conquest of Germany was taking a long time, while also acknowledging that a halting place might be found within Britain. In the second century, the era of the linear barriers, there is a self-satisfied tone to the comments on frontiers: the Romans hold all that they need and have taken appropriate measures to protect their empire. During the following century, emperors and would-be emperors were preoccupied with their own positions and in the face of external pressure both Dacia and the land beyond the upper reaches of the Rhine and Danube were abandoned. Yet some emperors sought to create a new frontier line and, although the empire continued to be under external pressure in the fourth century, many emperors were active on frontiers, maintaining the existing installations and building new, as well as negotiating treaties with their neighbours. It is not possible to see a slow decline of the frontiers. On the contrary, they were maintained and strengthened while the empire still lived.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 25

Chapter 4

The Romans on Frontier Installations

s we have seen, Roman – and Greek – writers tended to restrict themselves to statements on what we might call foreign policy, and military activities, or simply retire into hyperbole. Few mention the frontier installations themselves. Tacitus stated that Agricola chose the sites of his camps and Dio that Caracalla did the same, but these were the stock phrases relating to the activities of good generals (Tacitus, Agricola 20; Dio 78, 13, 4). Tacitus certainly recorded the existence of forts for both legions and auxiliary units, and when they were attacked, for example, in the Batavian Revolt of the late 60s. Helpfully, he recorded that the base at Vetera/Xanten held two legions (Tacitus, Histories 4, 22). In discussing the events of the governorship of Agricola in Britain over a decade later, he mentioned the placing of garrisons across the Forth–Clyde isthmus, as we have seen. Amongst all our sources, Arrian, governor of Cappadocia in the 130s, is in a class of his own. He is not writing history but an account of his voyage along the east coast of the Black Sea, though he is aware of the literary associations of his report (fig. 43). In it he provides the information which we would welcome for all frontiers: the distances between each pair of forts; their locations; materials of construction; the size and nature of the force stationed at each site and their purpose and the measures he undertook to improve their protection; the navigability of the rivers which flowed into the Black Sea; and the political affiliations of the neighbouring states:

A

At Apsaros, where the five cohorts are stationed, I gave the army its pay and inspected its weapons, the walls, the trench, the sick and food supplies that were there. [At Phasis], the fort itself, in which 400 select troops are quartered, seemed to me, owing to the nature of its site, to be very secure, and to lie in the most convenient spot for the safety of those who sail this way. In addition, a double ditch has been put round the wall, each ditch as broad as the other. The wall used to be of earth, and wooden towers were set up about it; now both it and the towers are made of baked brick. And its foundations are firm, and war machines are installed, and in short, it is fully equipped to prevent any of the barbarians from even approaching it, let alone to protect the garrison there against the dangers of a siege. But since the mooring-place for the ships must also be secure, as well as the whole area outside the fort settled by veterans of the army, various merchants and others, I decided to construct another ditch

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 26

26 The Frontiers of Imperial Rome that surrounds the wall as far as the river, which would enclose both the harbour and the houses outside the wall. (Arrian, Circumnavigation of the Black Sea 6, 2; 9, 3–5. Translation by A. Liddle.) An important source of comment on the creation of the frontiers of the second century is the collection of the Lives of the emperors in the compilation known as the Historia Augusta. These were written anonymously in the late fourth century. They have to be treated with some caution because the authors may have been reflecting the conditions of their own time rather than the period they were writing about two centuries earlier. Nevertheless, they provide several significant statements: [Hadrian] was the first to build a wall in Britain eighty miles long from sea to sea to separate the Romans and the barbarians. (Life of Hadrian 11, 2.) During this period, and frequently at other times, in a great many places where the barbarians are separated off not by rivers but by frontier barriers (‘limites’), he set them apart by great stakes driven deep into the ground and fastened together in the manner of a palisade. (Life of Hadrian 12, 6.) Antoninus conquered Britain through his general Lollius Urbicus and, having driven back the barbarians, built a new wall, this time of turf. (Life of Antoninus 5, 4. Translations by A. R. Birley.) These three brief comments are the only references in ancient literature to the building of both Hadrian’s Wall and the Antonine Wall in Britain, and, it is generally presumed, to the erection of the palisade in Germany. The first also offers a reason for the construction of such linear barriers, to separate the Romans and the barbarians. One of the most important statements about the purpose of frontier installations appears on a series of inscriptions of the late second century in Lower Pannonia. They record that: Commodus fortified the whole of the stretch of the river bank with towers (burgi), built from the ground up, and with garrisons stationed at suitable points, to prevent surprise crossings by bands of brigands. (ILS 8913) It would be difficult to find a clearer and more straightforward purpose of Roman frontiers: to prevent raiders from crossing the frontier. Brief references in the third century indicate that in spite of the enormous energy emperors expended in maintaining their own position, some had a care for the frontier. Some, like Decius, had to respond to the threat caused by new enemies, in his case the Goths, and the references to the frontier are brief in the extreme.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 27

The Romans on Frontier Installations 27 Fig. 4: Inscription of Commodus from Lower Pannonia.

Ammianus in the fourth century wrote extensively about frontiers, or rather warfare on frontiers and foreign policy. He offers few details and he tells us why: such details are beneath the dignity of history (Ammianus 26, 1, 1). Nevertheless, he does provide much useful information, though there is a touch of exaggeration in his descriptions. He refers twice to the protection of the Rhine frontier: Valentinian, meditating important and useful plans, fortified the entire Rhine from the beginnings of Raetia as far as the strait of the Ocean with great earthworks, erecting high fortresses and forts, and towers at frequent intervals, in suitable and convenient places for the whole length of Gaul; in some places structures were also built even on the farther bank of the river, which flows by the lands of the savages. (Ammianus 28, 2, 1. Translation after J. C. Rolfe.) Valentinian strongly reinforced the army, and fortified the high ground on both banks of the Rhine with strongholds and forts, so that no attack on our territory could be launched unobserved … if any of the enemy made a move, he was seen from above from the watch-towers and overcome. (Ammianus 30, 5, 6; 29, 4, 1. Translation after J. C. Rolfe.) In Britain, following the ‘Barbarian Conspiracy’ of 367, his general, Count Theodosius, ‘restored the garrisons’ fortresses and protected the frontiers with sentries and forts’ (Ammianus 28, 3, 7). We also have an incomparable source for the late Roman army, the Notitia Dignitatum. This appears to have been compiled in about 400 and is a list of all the officials of the whole empire, though the document is divided into east and west. Here are listed the commanders of every military unit together with their bases. When we know the location of each Latin name, we can truly create maps of the frontiers. In this way, for example, we can determine the regiment based in nearly every fort along the line of Hadrian’s Wall in Britain at this date together with several of the forts to the south. The section for Hadrian’s Wall starts:

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 28

28 The Frontiers of Imperial Rome Also, along the line of the Wall – Tribune of the Fourth Cohort of Lingones, at Segedunum [Wallsend] Tribune of the First Cohort of Cornovii, at Pons Aelius [Newcastle] Prefect of the First Cavalry Regiment of Asturians, at Condercum [Benwell] (Notitia Dignitatum, Oc. XL, 32–5) While within the command of the Duke of Armenia at the other end of the empire are: The Mounted Archers, at Sabbu The Mounted Archers, at Domana Prefect of the Fifteenth Legion Apollinaris, at Satala [Kelkit] Prefect of the Twenty-second Legion Fulmin(ata), at Melitena [Malatya] In Pontus: Prefect of the First Legion Pontica, at Trapezus [Trabzon] (Notitia Dignitatum, Or. XXXVIII, 10–16) Inscriptions also provide supporting and complementary evidence to the literary sources on the dating of various elements of the frontier. They furnish detailed information, for example, on the construction of both Hadrian’s Wall and the Antonine Wall in Britain. Two centuries and more later, the literary references to the actions of Valentinian on frontiers are supported by the inscriptions of the early 370s recording the construction of towers on frontiers in both Europe and the Middle East. Several important military documents survive from the Byzantine Empire. An anonymous treatise on strategy of the sixth century, probably of the reign of Justinian (527–565), is closest in date to the end of our period and it offers useful statements which presumably reflect the operations of earlier centuries: One way of arranging a good defense is to station sentinels and troops in outposts, to light signal fires, and to set up fortified positions to give warning of the approach of the enemy. Some of the posts may be located on open and clear ground, some in wooded areas, others in swampy ones. The best observation posts are those with level, open ground in front of them, so the guards will not miss any movement of the enemy. If there is no open, clear ground but just hills, then the lower terrain between the hills should be checked out in case enemy troops could pass through there and elude the guards. Forts are used for several purposes: first, to observe the approach of the enemy; second, to receive deserters from the enemy; third, to hold back any fugitives from our own side. The fourth is to facilitate assembly for raids against outlying enemy territories. These are undertaken not so much for

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 29

The Romans on Frontier Installations 29 plunder as for finding out what the enemy are doing and what plans they are making against us. These forts should be erected near the frontier and not far from the route the enemy are expected to take, so that any hostile advance will not go undetected by the garrison. They should not be located too much out in the open. If they are, the enemy, taking advantage of the ground, could keep them under observation from very close up to a great distance and so prevent any of our men, if need arise, from entering the fort or from leaving it when they wish. (The Anonymous Byzantine Treatise on Strategy 6, 7 and 9. Translation by G. T. Dennis.) Here are many statements which we will observe in action on different frontiers of the empire.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 30

Chapter 5

Regulations and Treaties

ilitary installations were but one aspect of frontiers. The Romans sought to control their neighbours by treaties and their entry into the province by regulations. Practically every frontier has produced evidence, either in literature or inscriptions, of treaties between the empire and her neighbours and these are discussed below. Peace on the Eastern Frontier was normally obtained through treaties between Rome and Parthia. The details of the treaties between Marcus Aurelius and Commodus and the Marcomanni, Quadi and Iazyges in the 170s and 180s recorded by Dio are particularly informative about the nature of such agreements and about wider aspects of Roman frontier policy. Dio’s account starts in 170 when:

M

the Quadi asked for peace, which was granted them, both in the hope that they might be detached from the Marcomanni, and also because they gave him many horses and cattle and promised to surrender all deserters and the captives, besides – thirteen thousand at first, and later all the others as well. The right to attend the markets, however, was not granted to them, for fear that the Iazyges and the Marcomanni, whom they had sworn not to receive and not to allow to pass through their country, should mingle with them, and passing themselves off for Quadi, should reconnoitre Roman positions and purchase provisions. (Dio 72, 11. Translation by E. Cary.) Five years later, in 175: When the Marcomanni sent envoys to him, Marcus, in view of the fact that they had fulfilled all the conditions imposed upon them, albeit grudgingly and reluctantly, restored to them one-half of the neutral zone along their frontier, so that they might now settle within a distance of about 38 stades [8km = 5 miles] from the Ister [Danube]; and he established the places and days for their trading together (for these had not been previously fixed) and exchanged hostages with them. … The Iazyges were defeated and came to terms … and made the same compact as that to which the Quadi and Marcomanni had agreed, except that they were required to dwell twice as far away from the Ister [Danube] as those tribes. (Dio 72, 15–16. Translation by E. Cary.)

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 31

Regulations and Treaties 31 Three or four years later, Marcus: released [the Iazyges] from many of the restrictions that had been imposed upon them, in fact, from all save those affecting their assembling and trading together and the requirements that they should not use boats of their own and should keep away from the islands in the Ister [Danube]. And he permitted them to pass through Dacia in order to have dealings with the Rhoxolani, as often as the governor of Dacia should give them permission. (Dio 72, 19. Translation by E. Cary.) On the death of his father Marcus and his succession in 180, Commodus made terms with the Marcomanni: In addition to the conditions that his father had imposed upon them, he also demanded that they restore to him the deserters and the captives that they had taken in the meantime, and that they furnish annually a stipulated amount of grain, a demand from which he subsequently released them. Moreover, he obtained some arms from them and soldiers as well, thirteen thousand from the Quadi and a smaller number from the Marcomanni; and in return for these he relieved them of the requirement of the annual levy. However, he further commanded that they should not assemble often nor in many parts of the country, but only once each month and in one place, and in the presence of a Roman centurion; and, furthermore, that they should not make war on the Iazyges, the Buri or the Vandili. On these terms, then, he made peace, and abandoned all the outposts in their country beyond the strip along the frontier that had been neutralized. Commodus granted peace to the Buri, when they sent envoys. … he compelled the others to take an oath that they would never dwell in nor use for pasturage a 40–stade [about 8km = 5 miles] strip of their territory next to Dacia. (Dio 73, 2; 3, 1–2. Translation by E. Cary.) This succession of treaties is illuminating. The emperors sought to impose on their enemies a neutral zone beyond the frontier, but had difficulty enforcing their will. They moved from a 16km-wide (10–mile) strip, to half of that, and finally to a requirement not to use boats on the Danube or land on the islands in the river. This inability to control the situation is not unlike Rome’s earlier inability to prevent the occupation by the Frisii of land set aside for military units on the ‘barbarian’ side of the Rhine in 59 (Tacitus, Annals 13, 54). Dio also records how the Romans sought to divide their enemies and obtain the return of deserters. In addition, the emperors demanded men for service in their army as well as grain, while their exchange of hostages was a time-honoured method of seeking to ensure peace. Finally, the Romans insisted on the control of access to markets, stipulating that meetings should only take place once a month and in the presence of a Roman centurion.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 32

32 The Frontiers of Imperial Rome These regulations echo those listed by Tacitus at the beginning of the second century in relation to the Tencteri, a tribe living across the Rhine from Colonia Agrippinensis/Cologne. In 70, during the Batavian Revolt, they sent an embassy to Colonia Agrippinensis/Cologne to complain that: until today the Romans have closed rivers and lands, and in a fashion heaven itself, to keep us from meeting and conferring together. Or else – and this is a severer insult to men born to arms – to make us meet unarmed and almost naked, under guard and paying a price for the privilege. (Tacitus, Histories 4, 64. Translation by C. H. Moore.) North of the Danube lived the Hermunduri. Tacitus commented that: [Next] comes the state of the Hermunduri, which is loyal to the Romans. For this reason, they are the only Germans who trade not on the river-bank but deep inside the frontier and in the province of Raetia’s splendid colonia [the city of Augsburg]. They cross over at all points and without a guard. To other peoples we only show our arms and our forts; to them we open our town houses and country mansions. (Tacitus, Germania 41. Translation by A. R. Birley.) These statements by Tacitus and Dio are broadly similar. The references to the control of movement in the frontier zone are supported by the evidence provided by the passes for travellers found in the Eastern Desert of Egypt. One such pass states: Q. Accius Optatus, to the four officers (curators) of the posts (praesidia) of the Claudianus route. Let pass Asklepiades. (O. Claud. 48) It is a fair implication that as these regulations were imposed on these four frontiers, they were also in operation elsewhere and that normally people living outside the empire could only enter at specified places and thereafter could only proceed unarmed under military supervision. The treaties described by Dio and the regulations quoted by Tacitus imply that the exact boundary of the empire was recorded and known. Other literary sources refer to the frontier of the empire being crossed.1 This is what we might expect for the Romans were careful to mark out and record property boundaries. Several treatises survive describing how land should be surveyed and property markers remain too, as well as fragments of maps indicating boundaries and ownership.2 Colonies for retired soldiers were carefully surveyed and land allocated to each veteran. In the frontier areas, markers survive which denoted the boundary between the land assigned to a military unit and that of a neighbouring civilian community, between different civilian communities, between imperial and private land and

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 33

Regulations and Treaties 33 between different provinces. Sometimes an army officer, often a centurion, was given the task of supervising the definition of the precise boundary. It might be considered that the exact location of the frontier would be marked by stones. None are, however, known. It has been argued that this is because ‘the very concept of such a boundary had no relevance in antiquity’.3 On this basis, the only known stone, recording that boundaries were established between the Roman province of Osrhoene and the kingdom of Abgarus during the reign of Septimius Severus, is dismissed as being irrelevant as relating to the boundary between a province and a client kingdom, Rome viewing the latter as part of her empire. Yet, as we have seen, there are references on inscriptions and in documents to the boundary of the empire.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 34

Chapter 6

The Building Blocks of Frontiers

The Roman army he Roman army had many branches. In Rome were the Praetorian Guard, the Urban Cohorts and the Fire Brigade, which was organized on army lines, and the Emperor’s Mounted Bodyguard. In the provinces were the legions and auxiliary units as well as fleets, also organized on military lines. It is the provincial army which is our concern here. The provincial armies were divided into two main branches. The legions were the traditional fighting force of Rome (pl. 4). By the first century CE, each consisted of a little over 5,000 heavily armed, well-disciplined infantry men, with a small detachment of 120 cavalry. These soldiers were Roman citizens, though many had never seen Rome. In the Republic, Rome had turned to her allies for military support, and in particular for cavalry. This branch of the army became known as the auxilia, literally ‘helpers’. In the early empire, these came to be organized into units of several sizes and types, nominally either 500– or 1,000–strong, infantry or cavalry, or mixed (pl. 7). The infantry units were organized into centuries of eighty men and the cavalry into troops thirty-two-strong. The actual size of the unit depended upon the number of centuries and troops and therefore the mixed units tended to be larger than their nominal strength while the larger cavalry unit was markedly below 1,000. There is also evidence that regiments were often below their nominal strength. The small number of surviving strength reports indicates that these units were below their theoretical size, in one case by about 25 per cent. Soldiers might also be based away from the parent unit for some time, on outpost duty, at provincial headquarters, or serving at another fort. In addition, the records of the Twentieth Cohort of Palmyrenes based at Dura-Europos show that it had a completely different internal organization than any other unit in the army. The workhorse of the frontier was the 500–strong mixed unit of infantry and cavalry. This combination of infantry and cavalry also appears in the detachments outposted from the cohort stationed at Dura-Europos in the early third century. There were also smaller units, each called a numerus, which simply means ‘a unit’, the infantry version perhaps normally consisting of four centuries, presumably each eighty men strong. In the desert areas, there were units of camel riders, or small numbers of such soldiers attached to other units. Diocletian was responsible for a radical overhaul of the army. The number of regiments was increased, but all types of units appear to have been reduced in size,

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 35

The Building Blocks of Frontiers 35 though the actual numbers are still subject to debate. Legions may have been down to about 1,000–2,000 men each, and possibly normally nearer the smaller figure. The figures we have for auxiliary units suggest that they were about a quarter of their earlier size. Further, new field armies were created. In some instances this involved the withdrawal of soldiers from the frontier forces in order to form the field armies. The units on the frontiers became second-class troops in relation to the new field armies. The process which led to the creation of these new field armies started in the middle of the second century. By this time most army units had already been at their bases for many decades and it was not easy to move them. In the Moorish War of Antoninus Pius in the late 140s, for example, detachments from regiments based in other provinces rather than complete units were drafted in to help the small local army. The same process was adopted in the Marcomannic Wars of his successor Marcus Aurelius when occurs the last attested movement of a whole legion. Septimius Severus raised three new legions and based one near Rome together with his strengthened Praetorian Guard. These had the primary purpose of protecting the emperor as well as accompanying him on campaign. The civil wars of the third century gave impetus to the process. Gallienus created a small field army which was continued by Diocletian and regularized under Constantine when there were over 300 regiments in the field armies. Diocletian retained the field armies under direct imperial control, there being one for each emperor and his deputy, four altogether. This directs us to their primary purpose, to protect the person of the emperor. Later in the fourth century regional field armies were created, at Antioch, Augusta Treverorum/Trier, Singidunum/Belgrade and in North Africa. The process of creating new field armies continued to at least the end of the fourth century, by which time an additional six such forces had been established. While the creation of such field armies appears to have been a reaction to the fossilization of the armies on the frontier, we have little hard evidence to indicate how they worked, other than their locations. The implication is that the field army at Antioch was located to counter the Persian threat, with those near the Rhine and on the Danube supporting the frontier troops there. Augustus established fleets at Ravenna on the east coast of Italy, at Misenum on the west and also at Forum Iulii/Frejus in Narbonensis, though this was later disbanded. The purpose of the fleets was to eradicate the piracy which had become endemic during the civil wars of the first century BCE when even Julius Caesar was captured. In this they were successful and as a result had little purpose thereafter; some sailors were employed operating the awning in the Colosseum. At the eastern end of the Mediterranean there was a fleet at Selucia in Syria and another at Alexandria in Egypt. Fleets were also placed along the rivers and seas which formed the borders of the empire, though not all date to as early as Augustus. There were two fleets on the River Danube, one below and one above the Iron Gates in modern Serbia, one on the Rhine, with the Black Sea and the English Channel each having a fleet. In the middle of the second century a detachment of the Egyptian fleet was

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 36

36 The Frontiers of Imperial Rome based at Caesarea/Cherchel in Mauretania Caesariensis to help protect the western Mediterranean from the attacks of Moors. Although each fleet had its own base, their ships might be spread amongst several forts, as Tacitus records for the Rhine in 70 (Histories 4, 15) and as appears also to be attested in the Notitia Dignitatum as several forts are marked as fleet bases. The fleets were branches of the army and it might be assumed that their primary purpose was military, guarding and protecting the frontiers, as was specifically attested in Roman literature (Pan. Lat. VIII, 14, 12, 1); Tacitus also records Agricola’s use of the fleet in prosecuting the war against the Caledonians in 83 (Agricola 25). The periodic recurrence of piracy in the Mediterannean and Black Seas and along the western seaboard of the empire emphasizes the continuing importance of the fleet. It is often assumed, though with little direct evidence, that the fleet’s main role in peacetime was to supply the army units based along the river. This is possible because military documents record army units involved in supply, such as a detachment of the First Cohort of Spaniards which served on grain ships (P. Lond. 2851), which suggests that the soldiers of the fleet may have been similarly employed. Soldiers of the fleet were also involved in building activities (e.g. RIB 1340, 1944–5). All these references emphasize that the men of the fleet undertook the same range of duties as other soldiers (pl. 6). Throughout the empire, regiments were given colourful names. These might relate to the officer who founded the unit, the emperor in whose reign it was established, the region where it was originally raised, or its valour in battle. However, the practice of local recruitment quickly led to these titles becoming irrelevant. Recruits to auxiliary units were generally raised from the province in which the unit was based, or in a neighbouring province. The late third and the fourth centuries saw many new units being raised. Their names followed the previous practice, but soon became irrelevant again in the face of local recruitment. Sometimes, however, these names are interpreted literally today, that is that the soldiers of the unit of Tigris bargemen based at Arbeia really were, and continued to be, what their title implies. In support of this, a second unit of bargemen was based at Lancaster on the River Lune on the west coast of Britain (RIB 601). The Roman world had no separate police force and the army often undertook the role of civil law enforcer. Troops were maintained in areas where there was likely to be internal unrest, such as Alexandria and Judaea. They also had a useful role in maintaining control of mining areas. Hanson has suggested that troops were only in Wales after the early second century in order to supervise the mines, and the legion may have been based at Apulum/Alba Iulia in Dacia to perform such a role.1 Soldiers were frequently taken away from military duties to support the judicial system such as investigating crimes or escorting prisoners. They also helped with civilian building projects.2 The use of the army in a wide variety of roles caused Trajan to remonstrate about soldiers being taken away from the colours (The Letters of the Younger Pliny 10, 20).

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 37

The Building Blocks of Frontiers 37

Officers and men All the commanders in both the armies and the fleets were members of the Roman aristocracy. The senior officers in the legions were from the senatorial families, the higher group of aristocracy, while the equestrian order provided the commanding officers of the auxiliary units and some of the officers in the legions. Each group of officers had its own career structure and in the early empire it was not possible under normal circumstances for an equestrian officer to graduate into a senatorial post. At the top of the provincial hierarchy was the governor, often an ex-consul in the case of a frontier province. His status as governor depended largely upon the size of his army. Hence, up to the end of the second century, Britain, Syria and Upper Pannonia (modern Austria and Hungary) were the three most important provinces below the more prestigious, but largely honorific, appointments of Asia (western Turkey) and Africa (Tunisia, western Libya and eastern Algeria). Governors of the important military provinces had considerable experience. At the beginning of the Bar-Kokhba rebellion in Judaea in 132 the preferred general to take command was Julius Severus, governor of Britain; he was transferred across the whole empire in order to deal with this insurgency, and this action was not unique. In other emergencies, special appointment might be made covering more than one province so that the forces at the command of the general were appropriate for the task. In 171, Avidius Cassius, governor of Syria, was appointed commander of the whole Orient, charged with reorganizing the eastern frontier following the war with Parthia. The grading of officers is important because in turn that allows us to determine the hierarchy of the auxiliary units. The first appointment for an equestrian officer would be as commander of a 500–strong (quingenary) auxiliary cohort. This might be followed by either command of a 1,000–strong (milliary) cohort or the post of junior tribune in a legion. The third step was as prefect of a cavalry unit. For a special few, a final appointment was to the command of a 1,000–strong cavalry regiment. At each stage, unsatisfactory officers were weeded out. Probably most had reached their first appointment as a result of patronage, having previously served as city magistrates, but many were unsuitable for military life. Thus about a third of the prefects of quingenary units would fail to receive a new appointment and about half of the next level would retire too. For those who proved their value, promotion to a post at provincial level such as procurator or command of a fleet might follow. Below these senior officers in both legions and auxiliary units were the centurions who commanded the centuries and the decurions in charge of the cavalry troops. In the main these officers were soldiers who had risen from the ranks and been promoted, usually after at least fourteen years’ service, though it was possible to obtain a direct commission from civilian life to the post of centurion. The advantage of this route for a civilian was that a centurion had no retirement age, while the appointment of an auxiliary commanding officer lasted only three years and, as we have seen, he might receive no further commission; nevertheless his promotion prospects were greater.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 38

38 The Frontiers of Imperial Rome These long-serving officers were the professional core of the army. Caesar records how he brought them into his councils of war. In comparison, many of the more senior officers were inexperienced. The senior tribune in a legion was only about 20 years of age, the junior tribunes older but only on their second military postings. Of the most senior officers, only the prefect of the camp was an experienced soldier. He was a promoted centurion, having therefore twenty or more years’ service. His duties, as we will see below, included running the legionary base. The Roman army was composed of men who had, in the main, volunteered to join, though levies did occur from time to time. They mostly joined between the ages of 18 and 21 and many were the sons of soldiers, born ‘at the fort’ as their birth records stated. They underwent basic training, which, at least in theory, continued throughout their career, which lasted twenty-five years. Almost from the beginning of their career many sought advancement to a post such as a clerk, musician or engineer, some moving on to a series of such appointments which brought them to the centurionate. The Roman army was intensely bureaucratic. Each unit retained a file on every soldier; there was even a record of every horse. Duty rosters were prepared, daily reports issued, records kept of supplies, arms and armour and pay and finances. Receipts were issued – in one surviving case in quadruplicate (P. Amherst 107). Annual returns of units’ strength were made to Rome, of which fewer than ten copies survive for the whole life of the empire. On retirement, a soldier could claim a document which stated his privileges as a veteran. The volume of documents which resided in the records stores of the Twentieth Cohort of Palmyrenes based at Dura-Europos by the 250s was so great that it was thought worth using them to heighten the city walls preparatory to a siege! The serious wars of the 160s and 170s started to change the nature of the command structure with more men of the lower aristocracy promoted to the top military posts. This process was to increase sharply in the third century when the earlier career structure fell apart. Diocletian began to separate the civil and military appointments, a process completed by Constantine. Now the senior military posts were all held by professional soldiers. New titles were introduced, such as count and duke, from which our modern titles descend, though in the Roman world their status was the other way round. By this time men were not so keen to volunteer for the army. As a result, in the fourth century, military service became a hereditary obligation. It is doubtful if this would have had much effect on the frontier armies who were probably already hereditary to a significant degree. Diocletian also reorganized the provinces of the empire. The general effect was to divide every existing province into at least two new provinces thereby reducing the size of the military force at the command of every governor and restricting the range of his activities. This was in many ways a logical step along a long road of dividing provinces, part of the purpose of which was to reduce the number of troops available to governors and thereby restrict their chances of successfully challenging the emperor. However, to some extent, this process was balanced by the increased

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 39

The Building Blocks of Frontiers 39 activities by emperors on frontiers as can be graphically seen in the pages of Ammianus Marcellinus’ History of Rome. Here Julian and Valentinian in particular are depicted as not only energetically dealing with foreign invaders but also restoring and strengthening the frontier installations.

Forts, fortlets and towers Down to the time of the late Republic, new armies were raised for each campaign. Accordingly, individual units did not require a permanent home. On campaign an army was protected overnight by the construction of a temporary fortification. This camp was normally only defended by a rampart and ditch. Soldiers were provided with stakes which they stuck into the rampart to reinforce their protection and the camp entrances were also protected by a separate section of rampart and ditch.3 In the second century BCE, the Greek writer Polybius described the Roman camp, detailing the layout and specifying where each unit should be placed. As he said, ‘no matter where this is done, one simple formula for a camp is employed, which is adopted at all times and in all places’ (Polybius, The Rise of the Roman Empire 6, 26). In this way, irrespective of which army a regiment served with, its officers would know where within the camp to locate their unit, while if the army was attacked at night, each soldier would know his own way round the enclosure. Four centuries later, an anonymous writer also detailed the arrangements of the measuring out and disposition of troops within a temporary camp. Although writing in the fourth century, it is clear that he is using earlier material, perhaps mainly of the second century CE. Our writer, usually called Pseudo-Hyginus, is especially useful for he describes the layout of the tents and enumerates how many soldiers should be in each tent – ten (Pseudo-Hyginus, The Fortifications of the Camp 1, 2, 34). These two documents are important because they emphasize the care devoted by the army to such matters, which carried over to the arrangements of permanent forts. Caesar’s Gallic War is another useful source. While he described campaigns, he also touched on the arrangements in winter quarters. Here, the soldiers would live in thatched buildings, perhaps used only for one winter (Caesar, The Gallic War 8, 1). Similar actions might be undertaken when a force had to stay in one place for a time, such as a siege. The defences of the enclosure were often strengthened and the tents surrounded by low walls, as, for example, at Masada. As the army became a permanent force, that is a standing army, its soldiers required better accommodation. This took the form of forts (pl. 3). These usually had a very recognizable shape, a rectangle or square with rounded corners, known as the playing card shape. In Europe, forts were usually originally built of timber and protected by ramparts of turf or earth. In North Africa and on the eastern frontier, where stone was in plentiful supply but timber scarce, forts were often constructed in this material from the beginning, though mud-brick was an alternative material. The use of stone in the East often resulted in a different style of military structure, smaller, with square corners and often taller.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 40

40 The Frontiers of Imperial Rome Forts of the first century CE might have a relatively short life, being abandoned as the pace of conquest rolled forward. If they survived long enough, which became more usual as the forward momentum was lost, they would be rebuilt in stone, or another more durable material. Many turf and timber forts appear to have been rebuilt in stone in the early second century, in the reign of Trajan, not necessarily because the emperor ordered such action but because timbers required replacing and there was a tacit appreciation that frontiers were settling down; the contemporary new forts in Dacia were built in timber. Arrian records the process of rebuilding in his Circumnavigation of the Black Sea. At Phasis, ‘formerly the rampart

Fig. 5: Plan of the unfinished first-century legionary base at Inchtuthil (Scotland).

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 41

The Building Blocks of Frontiers 41 was of earth, and the towers placed on it of wood. Now both rampart and towers are made of brick’ (Arrian 9, 4). Where stone was not available locally, for example in the Lower Rhineland, turf and timber continued to be used, and new forts elsewhere in Europe might still be built of turf and timber, as we can see at the legionary base at Carpow on the Tay in Scotland, erected in the early third century. Here, the fort was provided with a stone wall and stone principal buildings, but the barrack-blocks were of timber.4 Such bases might be massive, holding two legions, over 10,000 men, though during the first century these were gradually phased out, leaving the largest base that for a single legion of about 5,000 men, covering about 20 hectares (ha) (50 acres) (fig. 5). Auxiliary units were smaller than legions and their forts were commensurate in size (figs. 6, 7 and 10). They usually varied between 1.2ha and 4ha (3 and 10 acres) depending upon the size of the regiment. Sometimes two (or more?) units might be brigaded together resulting in the construction of larger forts. Each fort normally contained a headquarters building, house for the commanding officer, granaries, barrack-blocks, stables, store-houses and a latrine. Larger forts were provided with a hospital and from the mid-first century onwards there was usually a bath-house.5 It has been assumed that the normal arrangement was for each fort to be occupied by a single unit, but archaeological excavations are casting doubt on this simplistic view. It is clear that the arrangements were more complicated, with many forts being occupied by more or less men than appropriate for a single unit.6 Further, units might be of different structural frameworks, or could have groups of soldiers detached for duty elsewhere. About the year 100 over half of the unit based at Vindolanda in north Britain was at Coria, though still administered from Vindolanda. In the second century, several forts elsewhere in northern Britain appear to have been planned with the acknowledgement that some soldiers would be permanently outposted. In the early third century, the Twentieth Cohort of

Fig. 6: Plan of Housesteads fort on Hadrian’s Wall (Britain).

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 42

42 The Frontiers of Imperial Rome

Fig. 7: An artist’s impression of a fort. Drawn by Michael J. Moore.

Fig. 8: An artist’s impression of a fortlet. Drawn by Michael J. Moore.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 43

The Building Blocks of Frontiers 43 Palmyrenes stationed at Dura-Europos on the Euphrates maintained as many as eight outposts each containing fewer than a hundred soldiers.7 There are forts whose size lay between the legionary base and the auxiliary station. It is difficult to determine the function of these bases as so few have been excavated. Some may have held a special force or been a supply base. In many instances, especially on the frontier, a smaller force than even an auxiliary unit might be required and then a small fort or a fortlet was constructed. A small fort usually contained four centuries or less in Germany at least and was the base of an ‘irregular’ unit, a numerus (fig. 12). A fortlet usually held no more than a century of eighty men, outposted from elsewhere. The internal buildings were generally arranged on each side of a central road or in a U-shape. On Hadrian’s Wall a special fortlet, known as a milecastle, was adopted. It usually only contained a single building interpreted as a barrack-block. The soldiers based here probably had the dual roles of guarding the gate through the Wall at this point and maintaining surveillance from a tower over the gate (figs. 15 and 47). The smallest military structure was the tower, built of timber or stone and often surrounded by a low rampart and one or two ditches (figs. 9, 13, 15, 33, pl. 5). They have been found as early as the time of Augustus. Timber towers tended to be about 3m (10ft) square, while stone towers were usually about 6m (20 ft) square externally. Many stone towers in Germany had the security of an entrance on an upper floor, but this was not the case on Hadrian’s Wall. Towers had two possible functions, observation and signalling. A number of inscriptions of the late second century in Lower Pannonia record that towers were erected on the bank of the Danube ‘to

Fig. 9: An artist’s impression of a tower. Drawn by Michael J. Moore.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 44

44 The Frontiers of Imperial Rome prevent the surprise crossings by bands of brigands’ (ILS 8913). These clearly state the purpose of the towers, observation in an area subject to infiltration. Many towers on frontiers were closely spaced in a regular system and we therefore presume that their primary purpose was observation. Soldiers based here can also be assumed to have sent signals to convey the results of their observations. The use of towers off the line of a frontier for signalling is more difficult to demonstrate as it depends on the recognition of a line of intervisible towers and these are hard to prove as the chain is rarely complete. Other towers found along roads are generally regarded as helping to maintain security on supply routes. These were the range of structures which the army used when building frontiers. In many circumstances they adapted existing types of installations to suit their requirements even inventing new styles of structures when necessary. It is frequently suggested that soldiers in the East were billeted or based in towns. There is little evidence for this, Dura-Europos being the only certain example, though, military compounds are sometimes known on the edge of cities, as at Palmyra. In this urbanized part of the empire, billeting in cities would not be surprising. This, for example, was how the Praetorian Guard was quartered in Rome until its own fort was erected under Tiberius. The earlier forts had been relatively lightly defended, as Tacitus remarked. None had especially elaborate defensive arrangements. They were usually protected by two ditches. At Wörth in Germany, during excavations in the 1880s, two sections of the fort wall were found collapsed into the ditch. In one place it had formerly stood 4.2m (14 Roman ft) high and in the other 4.8m (16 Roman ft), in each case to a rampart walk protected by a parapet 1.6m (5.4 Roman ft) wide. The height of the main wall was the same as Hadrian’s Wall. Most auxiliary forts had four gates and each might have two portals. Tacitus remarked on the lack of readiness of the legionary base at Vetera/Xanten on the Rhine when it was attacked: Augustus had believed that these winter quarters could keep the Germanies in hand and indeed in subjection, and had never thought of such a disaster as to have the Germans actually assail our legions; therefore nothing had been done to add to the strength of the position or of the fortifications: the armed force seemed sufficient. (Tacitus, Histories 4, 23. Translation by C. H. Moore.) This statement reflects the general Roman attitude. Theirs was a force used to fighting and winning in the field; wars were not won by armies who cowered behind walls. Nor were Roman soldiers well equipped with defensive weapons for fighting from wall tops. Each soldier had two spears or javelins and was trained in the use of the bow and arrow, but there is little evidence for its widespread use in the army; possibly such weapons were normally left in the hands of the soldiers in the specialist units of archers. On Trajan’s Column a telling scene shows what could happen when soldiers in a fort were under attack: they were reduced to throwing

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 45

The Building Blocks of Frontiers 45

Fig. 10: Plans of the late-first-century fort at Elginhaugh (top) and the fourth-century fort at Altrip.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 46

46 The Frontiers of Imperial Rome objects – rocks or even turves from the rampart – at the attackers (Scene CXXXIV). Forts continued along broadly the same lines into the third century, though with an increased use of stone. Thereafter there was a change (figs. 10, 31, 32 and 46). The start of this change appears to have been during the reign of Severus in the early third century. Some forts were now built in non-standard shapes. These were usually outposts making good use of the terrain for surveillance.8 It is, however, rather later in that century when more substantial changes appear and they gather pace at the time of major reconstruction work undertaken by Diocletian. The newstyle forts had thick, tall walls, with the extra protection of external projecting towers or bastions and small, well-defended gates, often only two single portal gates in comparison to the four double-portal entrances of earlier years: defence was clearly more important. The walls were usually about twice the width of their early empire predecessors and had no earthen backing: the lack of such a backing contributed to the extra thickness of the walls. The height of the walls at the fourthcentury fort at Rutupiae/Richborough on the English side of the Channel is 7.3m (24ft) and those at Anderita/Pevensey 7.5m (25ft) to a wall-walk. The bastions, a distinctive feature of these fourth-century forts, changed over time. Those built under Diocletian were mainly U-shaped, while along the Rhine Constantine introduced circular towers straddling the fort’s walls and Valentinian added splayed fan towers to earlier forts on the Danube. The new-style forts were often of irregular shape, being oval or triangular almost as often as square or rectangular. Sometimes barrack-blocks were built against the inside face of fort walls. It would be fair to say that these forts are more like medieval castles than earlier Roman forts. Many of the ‘new’ features had long been commonplace in the eastern provinces, where fort walls did not normally have earthen backing, but were occupied by buildings. The earliest known such structures in the eastern provinces appear to date to the late second century.9 In Europe, their locations were also often different from earlier forts. In the early empire, the characteristic position for a fort was on a low hill, the equivalent of an upturned pudding bowl, or on slightly raised ground. In the late empire, they were more usually placed on higher ground in good defensive situations. In many cases, smaller, more strongly defended structures were built within the circuit of earlier forts. The late fort at Abusina/Eining on the Danube occupies just a sixth of the area of its predecessor; it is, in effect, a fortlet (fig. 31, pl. 20). Throughout the empire, more use appears to have been made of fortlets. These were equally as strongly defended as forts, with strong walls and projecting corner towers; this type of fortlet is often referred to as a quadriburgium. A new type of military installation appears to have been devised in the late third or early fourth century, the fortified landing place. This was a generally rectangular enclosure, no more than 0.4ha (1 acre) in size, with walls on three sides, the fourth lying on the bank of a river or a sea. They were the fourth-century equivalent, in size at least, of the earlier fortlet. Indeed, some may have been fortlets and not

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 47

The Building Blocks of Frontiers 47 landing places as it is not always clear if the wall on the river side has been washed away (figs. 27 and 29). Towers changed too, becoming larger and stronger (fig. 33). They were now about 10m (30ft) or even 16m (50ft) square, with walls up to three times as thick as earlier towers, suggesting a greater height. They were so large that the upper floors often required support from internal columns. Many sat within an enclosing wall so that they resembled in size the earlier fortlet. Not surprisingly therefore the inscription from Ravenscar on the Yorkshire coast of Britain records the construction of a tower and ‘fort’, though its size is more that of a fortlet (RIB 721). Yet, these new-style fortifications were not erected on every frontier. There were no Saxon Shore type forts built on Hadrian’s Wall, for example, nor fortlets along the roads in the frontier zone, while the installations on the Rhine were a mixture of old and new. Perhaps this tells us something about the differential nature of the threats on these frontiers. In the face of the major invasions of the third and fourth centuries, towns too were protected by similar walls with small, highly defended gates. These created strongly defended points not just on the frontier but deep into the interior, which could be used by the army as well as civilians. The world had changed for the Romans; they were certainly living in a more threatening age. It is worth noting that Roman military installations were broadly similar no matter where they were built within the empire. This was as true of forts built in the desert as in the temperate climate of northern Europe. Faced with the plan of a second-century fort, it would not be easy even for a specialist to determine exactly where within the empire it had been constructed. The same pattern occurs in the late empire. A tower might be erected in Hungary to a similar plan to one built on the north-east coast of Yorkshire. The result is that forts with no other dating evidence than their plans are often assigned a date on the basis of comparative analysis, which can potentially lead to circular arguments. A further difficulty is that in some areas it is not possible to distinguish with ease between military and civilian sites. In the hotter parts of the empire, in the Middle East and North Africa, where there was a longer and more sophisticated building history, similar small structures were built both by soldiers and farmers over many centuries. A wall, two storeys high, protected a square enclosure in which the buildings were normally placed against the inside face of the surrounding walls leaving a courtyard in the centre. Protection was provided both against the sun and enemies and even when weapons are found at such structures they may have belonged to farmers determined to defend themselves rather than soldiers on frontier duty. The same difficulty applies to even simpler structures such as towers. They may be for military or civilian defence over a wide range of dates. Especially when ruined, a small square structure could be interpreted as overnight protection for a farmer or shepherd in a hostile terrain and climate rather than as an observation tower for soldiers.10

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 48

48 The Frontiers of Imperial Rome

The builders of frontiers Inscriptions demonstrate that these frontiers, forts, fortlets and towers and their linking roads were built by the soldiers of the Roman army (fig. 4, pl. 4). The inscriptions record the names of the units responsible for the building work. This was normally a legion or an auxiliary unit, though the soldiers of the fleet built as well, as the following inscriptions demonstrate: The Antonine Wall in Britain: [titles of Antoninus Pius dating to shortly after 142] the Second Legion built 3666½ paces… (RIB 2193) Risingham/Habitancum fort in Britain: [titles of Septimius Severus dating to 205–7] the First Cohort of Vangiones … restored from ground level this gate with its walls, which had fallen in through old age … (RIB 1234) Qasr el-Hallabat in Arabia: [titles of Caracalla dating to 213–14] the soldiers of [four cohorts] built the new fort (castellum) through the agency of the governor Furnius Iulianus (Kennedy 2000, 93) Ybbs in Noricum: [titles of Valentinian, Valens and Gratian dating to 370] Count Equitius ordered the construction of a tower which was carried out by the Lauriacensian auxiliaries … (ILS 774) The texts of these inscriptions demonstrate that soldiers continued to build through to the very end of the frontiers, and this is also recorded in the literary sources. The, admittedly suspect, Historia Augusta, Life of Probus records that in Egypt he ordered the construction of ‘bridges and temples, porticos and basilicas, all by the labour of the soldiers’ (Probus 9). Ammianus mentions two incidents relating to the activities of Valentinian in Germany in 369 which corroborate such activities by soldiers. The first records the task of protecting a fort on the bank of the River Neckar: When he [Valentinian] considered that a lofty and secure fortification (which he himself had built from its very foundations) since a river called the Neckar flowed at its foot could gradually be undermined by the immense force of the waters, he even thought of turning the course of the stream in a different direction; and after he had collected men skilled in hydraulic work, the difficult task was begun with a great force of soldiers. For during many days beams of oak were bound together and placed in the bed of the river; but although they were fastened again and again by great piles driven close to them on both sides, they were forced from their place by the rising waters, and finally were swept away by the force of the current and lost. Yet, finally the day was won by the efficient supervision of the emperor and the labour of his obedient soldiers, who as they worked were often sunk chin-deep in the water. And, at last, though not without danger to some of the men, the defensive works, relieved of the pressure of the snarling river, are now strong. (Ammianus 28, 2, 2–4. Translation by J. C. Rolfe.)

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 49

The Building Blocks of Frontiers 49 Following this operation, Valentinian ordered the construction of a fort across the Rhine, but, while being constructed ‘a band of barbarians … attacked our soldiers, who were half-nude and still carrying earth’ (Ammianus 28, 2, 8). The fourthcentury statesman and philosopher Themistius knew several emperors in the second half of the fourth century, including Valentinian and his brother Valens. He recorded the construction of a fort by Valens in Thrace (Bulgaria) about 370: The emperor was not blind to the topographical advantages of the terrain. He spotted a thin strip of land running out into the shallows and culminating in a high hill from where all the surrounding area was visible. Here he established a new fort, showing the faint traces of an old fort found to be of use by some previous emperor, but abandoned because of the difficulties. There was no stone nearby, no baked bricks, you could not easily achieve anything, everything had to be brought from miles away by thousands of pack animals. Who then could not excuse those who had formerly withdrawn from the site on the grounds that the design could not be carried out? … You would have said, though, that the stone seemed to bring itself to the site, that bricks appeared by magic, and that the fort began to rise without the help of architects or masons. Such was the extent of the soldiers’ discipline and such the imperial mastery of the site’s problems. (Themistius, Orationes 10, 137b. Quotation from Johnson 1983, 61.) This might have been written at any time during the life of the empire. The general’s choice of the site and the army’s indomitability in its work are themes which run through the period of 400 years. Yet, on a day-to-day basis, more mundane activities fell to the prefect of the camp, that is the legionary base, in the early empire. The fourth-century author Vegetius writing about the earlier army, recorded his duties: There was an officer called the prefect of the camp. He was junior in rank to the commanding officer, but had a post of considerable importance. He was concerned with the site of the camp, the rampart, and the ditch, and the tents or barracks of the soldiers and all the baggage were under his charge also. He was responsible for the care of the sick soldiers and the medical orderlies, who looked after them, and the expenses involved. It was also his duty to provide vehicles, pack-horses, and the metal tools, such as saws and axes, for cutting and felling timber, spades for digging ditches, turf-cutters for constructing ramparts, and others for the supply of water. He was also responsible for seeing that there was always an adequate supply of firewood, straw, as well as rams, onagers, ballistas, and the other types of artillery. The man chosen for this post was always one who had had a long and successful career, so that he could correctly instruct others in matters that he himself had done with distinction. (Vegetius 2, 10. Translation by N. P. Milner.)

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 50

50 The Frontiers of Imperial Rome Each unit retained on its complement soldiers trained in surveying and building. Obviously, each legion had a larger staff. Inscriptions record architect-engineers, surveyors, masons, carpenters; even a glazier is attested. The range of such soldiers is listed by Vegetius: The legion had in addition smiths, carpenters, builders, wagon-makers, blacksmiths, painters and other craftsmen trained to construct the buildings in a winter-camp and the wooden towers and the other equipment for attacking or defending a site, and men who could make or repair wagons, vehicles, and other siege engines. (Vegetius 2, 7. Translation by N. P. Milner.) The Digest of Roman Law (50, 6, 7) lists soldiers granted some immunity from general fatigues. The list includes, ‘ditchers, farriers, the architects … glassfitters, smiths, arrowsmiths, coppersmiths, … water engineers … plumbers, blacksmiths, stonecutters, limeburners, woodcutters and charcoal burners’. Each regiment had its own workshop, often placed within the fort. It was also assigned an area of land round its fort where raw materials could be won, tileries and potteries constructed and other such activities undertaken. The Vindolanda writingtablets refer to the soldiers undertaking building activities. One document dated to 25 April in a year to either side of 100, states that: in the workshops, 343 men. of these: shoemakers, 12 builders to the bath-house. 18 for lead … … to the kilns … for clay … plasterers … … for rubble … (P. Vindol. 155) A second dated to 7 March includes: sent with Marcus, the medical orderly, to build the residence, builders, number 30 to burn stone, number 19 (?) to produce clay for the wattle fences of the camp … (P. Vindol. 156) The building of forts, fortlets, towers and roads was undertaken by soldiers from all branches of the army. Inscriptions cited above record work by legionaries and auxiliaries. Others from Hadrian’s Wall attest work by the British fleet (RIB 1340,

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 51

The Building Blocks of Frontiers 51 1944–5), while the anonymous fourth-century writer of de munitionibus castrorum (24) stated that ‘all the soldiers of the fleet camp in the forward tenting area because they go out first to build roads’, though other sources indicate that auxiliaries and legionaries also built the roads. There is no evidence for civilian involvement in Roman military projects during this time, rather the reverse. The army possessed the skilled surveyors, architectengineers and craftsmen to undertake both small and large building projects and might be called upon to help with civilian schemes. This is demonstrated by the second-century career of Nonius Datus, surveyor of the Third Augustan Legion based in North Africa (AE 1931, 38). He was thrice called upon to help the city of Saldae who had mismanaged the construction of an aqueduct. Unusually, a handful of inscriptions found on Hadrian’s Wall do record work by some of the communities of the province, but this appears to have been later repair work and is undated (e.g. RIB 1673, 1844 and 1962). Land for new forts had to be acquired and, according to one reference, might involve compensation. Frontinus recorded that when Domitian ordered the construction of forts in the territory of the Cubii in Germany following the Chattan War of the mid 80s, he instructed that the farmers should be compensated for the loss of their crops (Frontinus, Stratagems 2, 11, 7). This offers a balance to Juvenal’s Satire 16 quoted in the introduction.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 52

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 53

Part II

The Frontiers

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 54

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 55

Chapter 7

Linear Barriers

The Emperor Hadrian he reign of Hadrian (pl. 2) is a good point at which to start a discussion of Roman frontiers not least because Roman writers recognized his unique personal interest in frontiers, the army and foreign affairs (Dio 5; 9; Historia Augusta, Life of Hadrian, 10). Hadrian is recorded as visiting the frontiers in the Middle East, on the Danube, in Germany, in Britain and in Africa. Part of his speech to the army in Africa survives on stone inscriptions. It is also recorded that he sought peace and obtained it through his policy of military preparedness. His policy of maintaining the frontier within its existing boundaries was enforced by his renewal of the stones marking the boundary of the city of Rome. Hadrian’s frontier policy in action can be observed from the beginning of his reign in 117 when he abandoned Trajan’s conquests in the East as well as part of the newly acquired land across the River Danube. He returned to Rome via Pannonia (modern Hungary and Austria) where he inspected his Horse Guards swimming fully armed across the Danube. In 121 he was in Germany and the following year in Britain where he was specifically credited with the building of a new frontier. His subsequent visit to Spain was interrupted by the necessity to go back to the East to negotiate with the Parthians in order to prevent war. Five years later he was in Africa, where he inspected the troops in the province of Proconsular Africa. Among the military sites he visited was Lambaesis, base of the Third Augustan Legion, and where the text of his speech to the regiment was recorded in stone. In 129, Hadrian visited Cappadocia and Syria on the eastern frontier. During his reign a new harbour for the Pontic fleet was constructed at Trapezus/Trabzon, possibly as a result of his inspection of the area. Many of these visits, and his addresses to the army, were recorded on his coinage. The effect of Hadrian’s actions was to define the limits of the empire in Europe, Africa and the East. In this way, Hadrian was following the advice and actions of a predecessor, Augustus, whom he admired and on whom he consciously modelled himself and his actions. Hadrian’s interest in architecture is recorded in both the primary sources for the reign of Hadrian, Cassius Dio’s History of Rome and the Historia Augusta. Dio, the nearest to Hadrian in time, stated that the emperor modelled and painted and implied that he designed the Temple of Venus and Roma, putting to death the architect Apollodorus because he pointed out the errors in Hadrian’s plan, though the veracity of this anecdote has been challenged. The Historia Augusta refers to his

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 56

56 The Frontiers of Imperial Rome many building works, including the villa at Tivoli, as recent studies have emphasized.1 Another writer states that he organized the various types of workers in the building trade on military lines (Epitome de Caesaribus 14, 5).

The Upper German-Raetian limes In 120 or 121 Hadrian started on a tour of the western provinces. He visited Gaul en route to Germany, and thence travelled to Britain. Crossing back to Gaul, he journeyed to Narbonensis (modern Provence) before turning westwards to Spain. His visit to Germany received little attention in the ancient sources. The historian Cassius Dio referred to the ‘German snows’, implying that he was there over winter (Dio 9, 4). The Historia Augusta, written 200 years later, refers to his training of the army rather than to his construction of the frontier. However, in the Historia Augusta’s Life of Hadrian (12, 6), at the end of the account of his tour of these provinces, appears a significant passage: ‘during this period, and frequently at other times, in a great many places where the barbarians are separated off not by rivers but by frontier-barriers, he set them apart by great posts driven deep into the ground and joined together like a palisade’. Although not specifically referring to Germany, this is an accurate description of the German frontier and is generally taken as such. In 1892, the Reichslimeskommission was founded in order to record the land frontier in Upper Germany and Raetia through survey and excavation. The remains of a palisade were discovered. It was indeed formed of ‘great posts’ for the timbers were of oak placed next to each other in a trench about 1m (3ft) deep. Most of the posts were split into two with the flat surface facing outwards, but round tree trunks and posts about 30cm (12in) square were also used at intervals. These larger logs were often held upright by chocking stones. Excavation also produced evidence for nails, which again supports the statement in the Historia Augusta. It is difficult to know how high the palisade stood, but, bearing in mind its depth below the ground and the diameter of the tree trunk, a height of 3m (10ft) may not seem out of place. The palisade was not accompanied by a ditch (figs. 21 and 22, pls. 9 and 12). This barrier was found to run from the River Rhine, a little upstream from Bonn, through the Taunus Mountains, sometimes along the crest of the ridge, encircling the Wetterau area north of modern Frankfurt until it reached the River Main (fig. 11). For a number of miles the river formed the frontier, but, where the river turned eastwards, the barrier recommenced and ran southwards through the Odenwald, taking an irregular course through the rough and wooded terrain. Where it met the River Neckar, the barrier again ended. The purpose of this relatively small section of barrier was to link the two river sections of frontier and, unlike other sections where there was only a path, a road ran along behind the barrier. The frontier path continued along the Neckar, though not slavishly following the river but taking a more direct line and therefore crossing the loops of the river. To the east, in the province of Raetia, the barrier does not appear to have been completed at this time. The eastern sector of the Raetian limes was erected under Hadrian, but dendrochronology has indicated that the western sector was not built until the 160s

Fig. 11: Map of the frontier in Upper Germany and Raetia: the Hadrianic forts abandoned in the Antonine period are shown as half-filled squares while late forts are open squares.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 57

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 58

58 The Frontiers of Imperial Rome during the reign of the next emperor, Antoninus Pius. The eastern end of the palisade was at the River Danube close to the fort at Abusina/Eining. In places the line of the barrier was adapted to the countryside through which it ran, elsewhere it ran straight ignoring the topography. The reason(s) for these differences is described below (pages 85–6). The route of the palisade was chosen because it related to an existing series of military installations. In the time of Claudius the frontier had more or less followed the rivers Rhine and Danube. A generation later, Vespasian started the process, which continued well into the reign of his younger son Domitian, of occupying the triangular area of land between the headwaters of the rivers, an area called for reasons we do not known, the Agri Decumates. He had a road, protected by fortlets, built from Argentorate/Strasbourg on the Rhine to the Danube and he also built forts in the Wetterau, the rich farmland north-east of Mogontiacum/Mainz. In 83, his younger son, Domitian, campaigned against the Chatti, a German tribe living north-east of the Taunus Mountains, celebrating victory at the end of the season. The war may have resulted from the provocation caused to the local tribes by the Romans building forts in the Wetterau. Certainly, following the war, the Wetterau was protected by the construction of a series of towers through the Taunus Mountains which formed its western and northern edge. In Upper Germany, these towers lay immediately behind a path, perhaps little more than a track, while in Raetia, they lay in front of the path. The construction of this road has been linked to a statement by Julius Frontinus. In his book Stratagems (1, 3, 10) he states that ‘the Emperor Domitian, by advancing the limites through 120 miles … brought the enemy under control by exposing their hiding places’. The word limites in this context is thought to mean clearings through the forest rather than frontiers and therefore the comment is interpreted as being a description of the creation of fields of vision to control enemy movement in the Taunus Mountains. The excavation of many sites over the last 120 years has allowed us to appreciate the several phases in the construction of the various frontier installations. At first, fortlets were erected in the passes through the mountains, such as the two found below the later fort at the Saalburg, and on other routes. Some of these, such as the earlier fortlet at the Saalburg, may have been erected during the Chattan War. Presumably the intention was that these should control movement into the empire. The exact date of the construction of the towers is not certain. Although they have traditionally been dated to the years immediately following the Chattan War, recent re-examination of the archaeological evidence has suggested a date about twenty years later, during the reign of Trajan. Further, the finds suggest a slightly later date for the section through the Odenwald, between the Main and the Neckar, than that in the Taunus Mountains. The towers in Raetia are probably about the same date for it appears to have been only in the 90s that the forts were moved north from the early Flavian line on the Danube and through the Swabian Alps to a new line on the north side of these hills.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 59

Linear Barriers 59 The proposed new date for the construction of these two lines of towers is 105–15. Trajan’s officers were certainly active in the area because a new road was constructed between Mogontiacum/Mainz on the Rhine and Augusta Vindelicorum/ Augsburg on the Lech, a tributary of the Danube, in about 100. It may not be coincidental that the erection of the towers occurred shortly afterwards. Further, between the forts of Grinario/Köngen and Donnstetten to the southeast, a short stretch of palisade fronted by two ditches has been recorded: it is known as the Sibyllenspur or the Lautertal-limes. Grinario/Köngen is the most southerly fort on the eastern frontier of Upper Germany, while Donnstetten lies on the line of forts forming the north-west boundary of Raetia. The purpose of the barrier may have been to cover that gap, or even extend further south-eastwards to the Danube. The excavation of a section of the barrier and an adjacent fortlet has led to the suggestion of a date for occupation about 100, that is the beginning of the reign of Trajan. This may thus be the earliest known linear barrier, the first move to close the gap between the headwaters of the Rhine and Danube basins with a physical obstacle. It may be noted that there are no towers known on this barrier.2 A suggested chronology may therefore start with the cutting of clearings through the forests, which Frontinus appears to relate to the campaigning during the Chattan War of 83–5. The building of fortlets to control the passes presumably took place during or immediately after the war. The erection of the Sibyllenspur/Lautertallimes may have been part of a plan to control a weak point. The erection of the towers was the next step, and finally the construction of the palisade. The towers were placed about 500–600m (about a third of a mile) apart and linked, as we have seen, by a track. Those in the Taunus Mountains were about 2.5–3m (8–10ft) square and formed of four posts; each sat within a ditch. The slightly later towers in the Odenwald sector were also of timber but seated on a lower storey of stone. Uniquely on the German frontier, these towers were decorated with inscriptions, sculpture and architectural ornaments. It is generally presumed that there was a balcony for observation on the second floor like the towers on the nearcontemporary Trajan’s Column in Rome. The discovery of millstones, pottery and hearths makes it likely that soldiers slept here, but there were probably only four or five men at each tower. The towers did not stand alone. Small forts constructed of earth and timber, the bases of forces no greater 160 men, were erected immediately behind the line of the towers. Most were apparently placed to control routes across the frontier. Along the Neckar and between that river and the Danube there was greater spacing between the small forts and this may have related to the lack of German settlements in the vicinity of the frontier hereabouts. At first no forts for auxiliary regiments appear to have been built on this line. Presumably the troops to man the towers, and perhaps the fortlets, were provided by the forts behind the line in the Wetterau and in the Rhine valley. Over the following years, however, some forts were built, of earth and timber, on the frontier line, usually, like the small forts, at points of access, though at the more significant routeways. The

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 60

60 The Frontiers of Imperial Rome construction of each new fort led to the abandonment of a fort behind the frontier and the passing over to civilians of the land thus released from military control. This process continued until nearly all the forts behind the frontier were abandoned and the army concentrated in a narrow zone along the boundary of the empire. These forts and small forts on the frontier line were supplemented by fortlets, many capable of holding only about twenty to thirty men. The presumed purpose of the towers was observation. They are too close together to be for signalling along a line and too slight to be concerned with defence of the province. A Fig. 12: Plan of the fort at Hesselbach distinction can be made between the role of the on the frontier in the Odenwald soldiers in the towers – to observe anyone (Germany). approaching the frontier – from that of the soldiers in the forts, small forts and fortlets which, we may assume, was to intercept any who tried to cross. The towers were generally well placed to send messages to soldiers in the forts and this was very necessary as the forts were not intervisible.3

Fig. 13: An artist’s impression of a tower on the Odenwald sector of the German frontier.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 61

Linear Barriers 61 Only one fort lay behind the frontier, at Friedberg in the centre of the Wetterau. This fort had probably been placed here originally because it lay on a strategic route across the frontier. After the Augustan campaigns, at the latest after the Chattan War of Domitian, Friedberg became a hinterland fort, controlling the road into the province. When the limes was created, a 1,000–strong mixed infantry and cavalry unit of archers, the largest regiment in the provincial army after the legions, was based here. Such a unit could maintain the earlier role of controlling the route, but its nodal position in the road network of the region suggests that it will have acquired a new function, that of providing support for the units on the arc of the frontier round the northern tip of the Taunus Mountains. The legions themselves remained on the Rhine, at Mogontiacum/Mainz and Argentorate/Strasbourg, retained here presumably because of the advantage of supply along the rivers. A connection between the towers, the Chattan War and the statement of Frontinus is attractive. A date for their construction under Trajan is more difficult to explain. This emperor was more concerned about dealing with the long-running problem of the Dacian kingdom and the glory of winning victories over the Parthians. His interest in Germany therefore may have been in holding the existing situation and protecting the western provinces from attack. Perhaps that is the framework within which we should see the creation of this new style of frontier, a simple linear system of towers concerned with frontier control. It is certainly in relation to Germany that we learn about Rome’s concern to control access to her empire. Writing at the end of the first century and beginning of the second, the historian Tacitus describes Rome’s relations with two German tribes (see page 32). His description makes clear that many people living outside the empire were only allowed access to specified places and were forced to travel under escort and unarmed and after payment of a fee. The last element to be added to these frontier installations was the palisade. A date obtained by dendrochronology for a single timber in the Hadrianic palisade shows that it was felled in the winter of 119/20, shortly before Hadrian’s visit. On face value, this would suggest that the construction of the palisade started before the emperor’s visit. This is not beyond the bounds of possibility. Hadrian had previously served in the province and would have been conversant with the military situation in Germany, so it is possible that he ordered construction of the palisade before his visit, inspecting the work during his tour there in 121. If this was the case, we should perhaps see the frontier more as a developing institution rather than a structural innovation of Hadrian. However, it is but one date and more are required; this particular tree, for example, may have been felled for a different purpose.

Hadrian’s Wall From Germany, Hadrian travelled to Britain where, according to the Historia Augusta (11, 2), ‘he put many things to right and was the first to build a wall 80 miles long to separate the barbarians from the Romans’. This was written 200 years after the event but is the only certain ancient reference to Hadrian’s association with the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 62

62 The Frontiers of Imperial Rome Wall. We can determine that it is correct in that we do know that Hadrian was the first to build a frontier wall in Britain and that it was 80 Roman miles (130km) long. As in Germany, Hadrian’s Wall related to existing military installations. A generation before, in about 105, under Trajan, it would appear that the last of the forts north of the Tyne–Solway isthmus were abandoned, completing the retreat from Agricola’s victory at Mons Graupius 20 years before. Some forts already existed

Fig. 14: Map of Hadrian’s Wall in its landscape setting.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 63

Linear Barriers 63 across the isthmus from Corbridge, the lowest bridging point across the River Tyne, to Luguvalium/Carlisle in a similar position on the Solway estuary on the west coast. These forts were linked by a road known by its medieval name, the Stanegate. Additional forts were now built, reducing the distance between each pair to about 11km (7 miles), half-a-day’s march. In one 24km (15 mile) sector, from Vindolanda to Brampton, the forts were supplemented by the erection of at least three fortlets and four towers (fig. 33a), the significant elements found on all frontiers. Hadrian’s Wall was constructed from Segedunum/Wallsend on the River Tyne in the east to Maia/Bowness-on-Solway in the west. Excavation has demonstrated that the original plan was modified in a number of ways during the building process. One of these changes, it has hitherto been believed, was the extension of the Wall from the previously preferred original eastern end at Pons Aelius/Newcastle 6km (4 miles) down river to Segedunum/Wallsend. Existing evidence, however, will allow the proposition that it was always planned for the Wall to end at Segedunum/Wallsend, though the construction of this sector was undertaken later. The first stage in the building process will have been to lay out the line of the Wall. In a recent analysis, John Poulter has suggested that Hadrian’s Wall was surveyed from each end towards the middle, in convenient sectors up to 32km (20 miles) long. This work was presumably carried out by military surveyors. The task of building the Wall was allotted to the three legions based in the province, the Second, the Sixth and the Twentieth, each, at least in the early stages, apparently assigned a length of 5 Roman miles (8km) to build. It is possible that one of these legions, the Sixth, was brought to Britain from Lower Germany at this time, possibly with the specific intention of providing additional troops for the new building project. The first plan was for a stone wall 10 Roman feet (3m) wide from the Tyne for 49 Roman miles (80km) to the River Irthing and a turf rampart 20 Roman feet (6m) wide for the 30 Roman miles (48km) thence to Maia/Bowness (pl. 14). In front lay a wide and deep ditch, except where the terrain rendered it unnecessary. The width of the wall – the stone sector in particular – has led to the general assumption that there was a walkway along its top. This has been challenged on the basis that such a feature was manifestly not present on the Germany palisade and in any case was not a requirement for a Roman frontier. Paul Bidwell’s review of the evidence, however, suggests that there probably was a wall-walk.4 A gate through the wall occurred at every mile and each was protected by a small fortlet which appears to have provided barrack accommodation for a small number of men, perhaps eight or ten, though some of these milecastles had larger internal buildings (figs. 15 and 17). Between each pair of milecastles were two towers, termed turrets on Hadrian’s Wall. We presume that a tower also stood over the north gate of each milecastle so as to complete the series of such towers at one-third of a Roman mile intervals all along the Wall. The role of the soldiers in the milecastles was to guard the gate, maintain observation from its tower, and, probably, provide a base for the soldiers at the towers on either side: the lack of evidence for furniture and box fittings at the towers suggests that soldiers were not permanently based there.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 64

64 The Frontiers of Imperial Rome

Fig. 15: Plans of a milecastle and a tower on Hadrian’s Wall.

The system of fortlets and towers was carried about 40km (25 miles) down the Cumbrian coast from the end of the Wall at Maia/Bowness, though without a linear barrier. These milecastles are a new feature, not previously found on any other frontier in this form, and only replicated on the later Antonine Wall. Towers were certainly not a new feature. As we have seen, some already existed on the isthmus, and they were of stone: two were incorporated into the Wall. Earlier towers, such as those on the Gask Ridge in Scotland, probably dating to the late first century, were of timber. At various locations in the eastern 17km (11 miles) of the Wall, pits have been found on the berm, the space between the wall and the ditch (pl. 16). At Byker between Segedunum/Wallsend and Pons Aelius/Newcastle, for example, there were three rows and each held two stout posts on average 180mm (7in) in diameter, perhaps tree trunks to which were still attached their branches, sharpened to make them a formidable obstacle, as was described by several Roman writers.5 The clean fill of the pits suggested that they were contemporary with the construction of the Wall. At the presumed site of one turret further west the pits appear to have turned towards the turret as if the berm was narrowing. Certainly, narrowing of the berm is visible at one other turret and the phenomenon was noted on the Turf Wall in the 1930s. The pits would have helped to make Hadrian’s Wall a formidable obstacle, but, as yet, their distribution is localized and excavation elsewhere on the berm has

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 65

Linear Barriers 65 failed to find them. Even so, the pits were thought important in the areas where they have been discovered for those at Byker were replaced. There were, at this stage, no forts on the Wall line. Existing forts behind the Wall along the Stanegate and to the west of Luguvalium/Carlisle were retained, though the smaller forts appear to have been abandoned. North of the western sector of the Wall, three outpost forts were built. There is some evidence to suggest that they were part of the first plan for the Wall. This first plan for Hadrian’s Wall is important for it allows us to see a distinction between the linear barrier and its structural elements on the one hand and the forts to the south on the other. The forts held the troops whose duty was defence of the province, though they may also have provided the soldiers to man the structures on the barrier. The barrier, with its many observation towers, we might presume, was erected to help prevent raiding from the north; some have seen its purpose as defence, or at least holding up an invasion until reinforcements arrived. Its construction would have also aided the enforcement of the regulations governing entry to the empire. The provision of so many gates indicates that movement across the barrier was envisaged, though perhaps the gates were provided primarily for the army rather than civilians.6 The relationship between the soldiers on the barrier and those in the forts behind has been emphasized by the recent work by Poulter on how the Wall was originally surveyed. He has pointed out that when the base line had to move to negotiate a feature in the landscape, it moved to the south. This was presumably to allow the soldiers to remain in contact with the existing forts to the south. It was for this reason that the Wall was placed on the south side of the ridges along which it ran; from the point of view of the soldiers in the forts, the Wall stood on the skyline. Poulter’s analysis corroborates the work of David Woolliscroft who has shown some milecastles and turrets were moved slightly out of position, apparently to ensure communication with the existing military sites to the south.7 This would not have been necessary if there had been forts on the Wall line. Poulter goes on to suggest that if defence had been a primary concern for the builders of the Wall, the Wall would have been placed in a slightly different location. It would have been erected a little further north, on the northern edge of the ridge along which the Wall runs, and would have been more sinuous as it followed the landforms. The fact that the Wall did not use the landscape in this way suggests that its primary purpose was not defence. The need to keep in touch with the forts to the south resulted in the Wall sometimes being placed in a location with relatively restricted views to the north, emphasizing the lack of concern for defence. There is, however, a problem with the proposition that towers on the Wall were specifically located to aid communication to installations behind the frontier line: this is that there are no known contemporary forts behind the Wall between Pons Aelius/Newcastle and Corbridge (the only known, but undated, fort at Washingwell, south of Newcastle, faces east not north), though it could be argued that the intention to build some had been overtaken by the decision to place forts on the Wall

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 66

66 The Frontiers of Imperial Rome

Fig. 16: Plan of the fort at Chesters, placed astride Hadrian’s Wall.

line. It may be noted that the forts added at the eastern end of the Wall appear to have been the earliest to have been built. The work on this plan was far from complete when it was decided to amend it and build forts actually on the Wall line. These forts were placed 7⅓ miles (11.7km) apart, that is half-a-day’s march. The position of the new forts is unique within the Roman empire for, where the topography allowed, they were erected astride the Wall (fig. 16). This involved the demolition in places of erected lengths of Wall, turrets and, in one case, a milecastle, as well as the infilling of sections of the Wall ditch, so the decision was not taken lightly. The forts were placed so as to allow three main gates – unusually each double-portal – to open north of the Wall and one to the south. In addition, the forts were provided with two extra gates, one at each end of the main road across the fort behind the central range. This greatly increased mobility in the Wall area. Previously, troops wishing to move north of the Wall had to pass through a single-portal milecastle gateway. Now, the equivalent of six milecastle gateways were provided north of the Wall and four to the south. This helps us understand the purpose of the milecastle gates too; they are more likely to have had a military function than to have been provided primarily for civilian traffic.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 67

Linear Barriers 67 It is interesting to note that neither of the later forts to be built on the Wall was placed astride the frontier: each was placed to the south, though still attached to the Wall. It was as if it were realized that once forts were on the line of the Wall it was not necessary for them to project to the north. The forts were unusual in another manner. With the exception of tiny Congavata/ Drumburgh, all were of a size to have held a complete auxiliary regiment. In earlier years, many forts seem to have held either detachments or composite garrisons, and the same is also true of the later Antonine Wall. More stone was used in the forts, at least those on the stone sector of Hadrian’s Wall, than was normal with newly built forts which usually had turf ramparts and timber buildings. Another innovation occurred about the same time, the construction of a great earthwork behind the Wall (pl. 15). This consisted of a ditch with a mound set back equidistant on each side: it offered no advantage to either side. This might be seen as the Roman equivalent of barbed wire, controlling access to the military zone through which the Wall ran. Access was certainly controlled because the number of crossing points was now reduced from an original number of eighty or so – one at each milecastle and the two points where the roads north passed through the Wall – to about sixteen, one at each fort and the road crossings. At each fort, a causeway was left across the ditch and on it was erected a gate, opened from the fort side. This earthwork has been known since the time of the Venerable Bede, 1,200 years ago, as the Vallum. It is clearly later than the forts because, throughout most of its length, it lies so close to the Wall that it had to be diverted round them. Again the work of John Poulter is valuable because it shows that the Vallum was laid out from the forts. While the general purpose of the Vallum is clear – control of some kind in the frontier zone – the reason why this was thought necessary is another matter. So far as we can see, no civilian buildings were allowed to be built between the Wall and the Vallum. After the placing of the forts on the Wall line, further changes were made. The stone Wall, where not completed, was narrowed from an original 10 Roman feet to 8 Roman feet or less (3m to 2.5m) – widths of 6ft (2m) have been recorded. There was also a reduction in the quality of craftsmanship employed in building the Wall. These two changes may be the result of a desire to hasten the completion of this vast building project. There is certainly evidence that the process of building the Wall took a long time. For example, nearly half a metre of peat and debris had accumulated in the bottom of the Wall ditch at Cilurnum/Chesters before the fort builders arrived. At Peel Gap, work was abandoned on the construction of the wall and when it recommenced, the foundations had become overgrown leading to the vegetation being cut down and burned on site. However, there were also additions to the original plan, including an extra fort at Brocolitia/Carrawburgh to break an over-long gap between two original forts. The various changes produced overlapping actions in that the first plan was not completed before the second scheme – the moving of forts onto the line of the Wall and the construction of the Vallum – was started. The emphasis appears to have moved from building the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 68

68 The Frontiers of Imperial Rome barrier to erecting the forts, with the result that several were completed before the Wall. Yet, before Hadrian’s death, a start had been made on rebuilding the turf sector in stone, though only the eastern 8km (5 miles) was completed before he died. Some parts of this great building project can be dated. It is highly likely that Hadrian arrived in Britain in the weeks before 17 July 122 when a document records the presence in the province of the new governor, A. Platorius Nepos. Nepos came from Lower Germany, as did Hadrian (and the Sixth Legion), so there is a strong possibility that they travelled together. Our assumption is that the building of the Wall started that summer. The name of Nepos appears on several milecastle inscriptions, though all date to later in the building programme. The implementation of the second scheme certainly started under Nepos as his name also appears on two inscriptions found at forts. Thereafter, there is evidence of a noticeable gap before some forts were built, while an inscription from Aesica/Great Chesters, demonstrably late in the building programme, gave Hadrian the title ‘Father of his Country’ which he did not formally take until 128. There are even later inscriptions recording building work at the fort at Magna/Carvoran in 136–7. Indeed, it is possible that not all forts were fully operational before the Wall was abandoned in the early 140s as Banna/Birdoswald appears to have been unfinished. Sometime during the reign of Hadrian there was a British expedition (ILS 2726 and 2735). The career of T. Pontius Sabinus, the leader of the 3,000–strong detachment which took part in the expedition, suggests that the event occurred in the second half of the 120s. If there was warfare in Britain during the building of the Wall, it could have led to a suspension of building work and, when tools were taken up again, a decision to cut back on the original specifications in order to complete the task more quickly. Hadrian’s Wall is very different from its contemporary frontier in Germany and is unique in several ways: the milecastles, the great use of stone, and the sheer size of the stone wall of the first plan and the forts astride the Wall and the Vallum of the second scheme, the regularity of planning displayed in both. It has been argued that some of these features result from the intervention of Hadrian. It has seemed unlikely that Hadrian could have been responsible for the unique elements in both the first and the second plans for the Wall, but, in a detailed analysis of the construction of the Wall, Peter Hill has shown that little may have been built before the decision was taken to add forts to the Wall line. The laying of the lowest courses of stone in the first sections of Wall to be built had not even been completed, and, except at some structures, nowhere does the original broad stone wall stand higher than 2m (6ft). Accordingly, if Hadrian had stayed in Britain for three or four months, it would have been possible for him to have been responsible for the unique features of both stages of the building process.8 Yet, not all decisions may be attributed to the emperor. The relative lack of timber in northern Britain may have resulted in the use of other building materials. Turf would normally have been the next choice and was used in the western sector. However, the poor turf in the central sector coupled with the emphasis on

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 69

Linear Barriers 69

Fig. 17: An artist’s impression of Hadrian’s Wall. Drawn by Michael J. Moore.

agriculture further east – plough marks have been discovered below many parts of the Wall – may have led to building in stone. Structures built in stone were more difficult to change than those in turf or timber and accordingly work in stone may have affected the mindset of the builders, deciding upon a rigid system of gates and towers rather than a more pragmatic approach. The building of Hadrian’s Wall transformed the military landscape of northern Britain. A continuous barrier and its associated high concentration of soldiers, together with the continuing occupation of the forts on the Stanegate immediately to the south, acted as a magnet for civilians. Troops were brought here from elsewhere in northern Britain and the Welsh peninsula, changing the pattern of military deployment across the province. There is one element of the German frontier which is not known on Hadrian’s Wall in its Hadrianic phase – a track or road. A road already existed across the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 70

70 The Frontiers of Imperial Rome Tyne–Solway isthmus, but it only ran from Corbridge to Luguvalium/Carlisle, though a road leading further westwards may date to this time. In some areas, a path has been recorded close to Hadrian’s Wall, but it is undated. One further comparison may be made with the frontier in Germany. The actions of Domitian and Trajan had been to incorporate the rich farmlands of the Wetterau into the empire: Hadrian merely acknowledged the status quo and erected his barrier on the line chosen by his predecessors. In Britain, he did the same, placing the Wall immediately in front of the existing military installations. In neither case did Hadrian undertake a basic review to examine the best locations for the empire’s frontiers but accepted the situation he had inherited and strengthened the existing installations.9

The Emperor Antoninus Pius Hadrian died on 10 July 138 after a long illness. It would appear that he had decided on the young Marcus Aurelius, probably a distant relative, as his eventual successor. But Marcus was only fourteen and too young to succeed. Hadrian firstly cleared two of his own close relatives from the scene by murdering them. He then chose a man to succeed him who was not expected to live long, but he died before Hadrian. His choice finally fell on the uncle-by-marriage of Marcus, the man known to history as Antoninus Pius (pl. 2). Pius not only would keep the throne warm for Marcus, but he was relatively old, being only ten years younger than Hadrian. Ironically, he was to live into his seventies and have the longest reign of any emperor over the united empire after Augustus. Unlike his predecessor, the man who succeeded Hadrian had no military experience. This was a disadvantage for a ruler whose authority ultimately depended upon his control over the army. In such circumstances, Antoninus Pius, like Claudius a hundred years before, might have needed a triumph to help secure his position. He certainly ordered an advance of the frontier in Britain for which he was acclaimed Imperator, Conqueror, the only time during his long reign.10 The personal involvement of the emperor is emphasized in a contemporary source, a speech in the Senate in Rome. The speaker was Cornelius Fronto, who was not only consul but a member of the imperial household and tutor to the young Marcus Aurelius. Only a fragment of the speech survives. Fronto declared that ‘although he [Antoninus] had committed the conduct of the war to others, while sitting at home himself in the Palace at Rome, yet like the helmsman at the tiller of a ship of war, the glory of the whole navigation and voyage belonged to him’ (Fronto, From the Speech of the War in Britain). The occasion is also recorded for us by the anonymous writer of the Historia Augusta in his Life of Antoninus Pius (5, 4): ‘he conquered Britain through his legate Lollius Urbicus, and, having driven back the barbarians, built a new wall, this time of turf ’. Emperor and legate are recorded on several building inscriptions. Two are from Corbridge on Dere Street, the road leading north into Scotland; they date to 139 and 140 and indicate that planning for the move north came very soon after the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 71

Linear Barriers 71 accession of Pius, emphasizing the political rather than the military reason behind the advance. A third records building at Bremenium/High Rochester further north along the road, while the remaining two inscriptions, both fragmentary, were found at Balmuildy on the Antonine Wall. The name of Lollius Urbicus does not appear on any other of the many inscriptions on the new Wall and it would appear that he left the province before it was completed.

The Antonine Wall The line chosen for the Antonine Wall did not relate to any existing military installations; there have been many attempts to associate the new forts with their Agricolan predecessors of sixty years before, but these had been long abandoned anyway. As a result the Wall was placed in the best position in relation to its landscape. For much of its length it was erected on the southern edge of the Midland Valley of Scotland, overlooking the River Carron which flowed eastwards into the Firth of Forth, and the River Kelvin which flowed in the opposite direction into the River Clyde (fig. 18). The Midland Valley was accordingly a great trough in front of the Wall, probably very boggy at that time, as it is today. The eastern sector of the Wall used the raised beach to the south of the Firth of Forth while to the west, where there was no strong linear feature, the Wall jumped from high point to high point, utilizing the hillocks – drumlins – of the Clyde valley before it reached the river a little above the lowest fording point. To the north of the Wall, a line of forts, four in number, lay along the road leading north to the River Tay. The wall of turf started by Lollius Urbicus was the latest, state-of-the-art frontier. In its detail, it was an advance on Hadrian’s Wall even though it was constructed of different materials. The ‘wall’ was actually a turf rampart, but placed on a stone base 4.4m (about 15 Roman ft) wide (pl. 17). This may have been to improve drainage across the line of the Wall – it would appear that drainage across Hadrian’s Wall had been a problem for in its later phases of construction drains were more regularly provided than earlier – or allow the construction of a more compact rampart to the same height as its Hadrianic predecessor. It is not known how high the rampart stood, nor how it was finished off at the top, but excavations have revealed evidence for up to 20 turf lines suggesting a minimum height of least 3m (10ft). It was fronted by a wide and deep ditch, the material from which was tipped out onto the north side to form a low, wide upcast mound, and, as on Hadrian’s Wall, in certain locations, pits have been found on the berm. The ditch is a very substantial obstacle in the central sector where it was 12m (40ft) wide and 4m (13ft) deep, though elsewhere it is often about half that width. These three linear elements – rampart, ditch and upcast mound – ran from modern Bo’ness on the Firth of Forth to Old Kilpatrick on the River Clyde, a distance of 60km (40 Roman miles); the new Wall was therefore almost exactly half the length of Hadrian’s Wall. The original intention was to place six forts along the line of the Wall about 11–13km (7–8 miles) apart, and it is probable that there was also to be a fortlet, like a Hadrian’s Wall milecastle, at every mile in between, though the series is far from

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 72

72 The Frontiers of Imperial Rome

Fig. 18: Map of the Antonine Wall in its landscape setting; north of the Wall, the earlier firstcentury forts are shown as half-filled squares (or triangle in the case of Inchtuthil) and thirdcentury Carpow as an open square.

complete and so we cannot be certain about this point (fig. 20). The main pattern therefore appears to have been similar to that of the abandoned Hadrian’s Wall and here presumably lay the basis of the plan for the new frontier. On the Antonine Wall the forts did not lie astride the barrier but were placed behind, though usually still attached to, the Wall. If we pursue this comparison, we might expect there to have been towers on the Antonine Wall. None, however, have been found, though, if they were of timber and only about 3m (10ft) square, they could be hidden within the width of the stone base. The Antonine Wall, on the other hand, contained features not found on its predecessor. In three places, pairs of ‘expansions’ have been identified. Two pairs stand on each side of Rough Castle and face north to the outpost forts, while the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 73

Linear Barriers 73 pair on the west side of Croy Hill has a wide prospect to the south, to the fort at Bothwellhaugh in Clydesdale. Excavation at one of the ‘expansions’ revealed that it was a turf platform erected on a stone base, attached to and apparently contemporary with the turf rampart of the Wall. Burnt turf and wood and fragments of pottery suggested that soldiers had been on duty here and, it has been argued, lit fires on top of the platform. Below the platform was a pit from which gravel had been extracted, presumably to help build the adjacent Military Way, thus suggesting that the road was part of the original plan for the Wall. The third innovation was the construction of small enclosures. Only three are known, in the vicinity of the fortlet at Wilderness Plantation. Excavation of one has revealed a small ditch surrounding a turf rampart which enclosed an area 5.5m (18ft) square, about the same size as a turret on Hadrian’s Wall, and contemporary with the Antonine Wall, but it provided no evidence of its function. The spacing of the enclosures is, however, closer than the turrets on Hadrian’s Wall, being roughly one-sixth of a mile apart, though the variations are too great to confirm an intention for such a spacing. This Wall, like its predecessor, was built by the three legions of Britain, though in this case they recorded their work through the carving of wonderful inscriptions decorated with scenes of religious rituals and fighting (pl. 18). The stones record the lengths of Wall built by each legion: 4⅔ miles, 3⅔ miles, 3 miles, and then a stretch of 2 miles (3km) in shorter lengths of feet. The distances are specified to half a pace in some cases; this is precision planning. Analysis of the lengths by Mark Hassall has led to the suggestion that the 20–Roman-mile long (30km) central sector of the Wall was planned to be built first, probably followed by the eastern sector and then, finally, the western 2 miles (3km).11 We can also see that the sites of the structures, the forts and fortlets, were identified first, and probably in most cases built, before the line of the Wall was erected between them. This plan was not finished before a significant change was made. This was the addition of more forts to the line of the Wall (fig.19). At least ten new forts were constructed, reducing the average gap between the forts from around 13km (8 miles) to 3km (a little over 2 miles). Some forts replaced existing fortlets; elsewhere they appear to have been erected in new positions. While all the original forts were large enough to have held a complete regiment, many of the secondary forts were smaller and could only have held a detachment. Some of these detachments appear to have been formed of legionaries, an unusual feature on the British frontier and perhaps suggesting that the construction of the Antonine Wall overstretched the manpower resources of the army of Britain. This has been reinforced by a review of the interiors of forts by Lawrence Keppie.12 He points out that several may not have been fully occupied. The forts on the Antonine Wall were unlike those on Hadrian’s Wall in that many did not hold whole units, and the arrangement of their internal buildings was more closely related to the frontier than those in the Hadrian’s Wall forts which give the appearance of being merely on the line of the linear barrier for convenience. The

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 74

74 The Frontiers of Imperial Rome

Fig. 19: Plan of the fort and annexe at Rough Castle on the Antonine Wall (Scotland).

Antonine Wall forts tended to have turf ramparts, stone main buildings and timber barracks, stables and storehouses. Several were especially well defended with ditches, while the north front of Rough Castle was protected by ten rows of pits known as lilia. The addition of so many forts led to a change in the function of the fortlets. In the few cases investigated, the internal buildings were demolished and the interior covered with a layer of cobbles or gravel, though the enclosures appear to have continued in use in some form. In one case, the north gate went out of use and it may be then that the causeways in front of the gates were removed. A further amendment to the original plan was the construction of annexes beside some forts. We can see that annexes were an addition to the Wall because they were added to the side of the primary forts at Mumrills and Castlecary where the fort ditches were infilled when the changes were made. Annexes also occur at some of the secondary forts. However, at Bearsden, a secondary fort, the area of the fort was subdivided into fort and annexe during building work, which is the best indication we have for the place of the annexes within the building programme. There is also some evidence for the date of the construction of the annexes for pottery dating to the second half of the 150s was found in the infilled ditches at Mumrills. This change appears to have come late in the building programme for we believe that the Antonine Wall was abandoned in the 160s. The reason for its abandonment is not known, but may have related to a review of military commitments which may

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 75

Linear Barriers 75 have included the necessity to provide soldiers from Britain to help deal with trouble in Germany. The Antonine Wall possesses features in common with Hadrian’s Wall: the linear barrier, ditch, pits on the berm, forts on the barrier and fortlets in between. It does, however, have distinctive elements. These include the expansions, generally interpreted as beacon-platforms, and the small enclosures in the vicinity of Wilderness Plantation, closer together than the turrets of Hadrian’s Wall. These appear to have been part of the original plan for the Antonine Wall. During the course of construction of the frontier, new forts were added. The reason for this is not known, but it does suggest some opposition to the building of the Wall. The fact that the two new types of structures were part of the original plan may indicate that opposition was anticipated from the beginning. This receives some support from the dense cluster of forts and fortlets in the hinterland of the frontier line, an area more tightly controlled than earlier. The road along the Antonine Wall appears to have been part of the original plan; the existing Stanegate was presumably considered sufficient on Hadrian’s Wall.

Fig. 20: An artist’s impression of the Antonine Wall. Drawn by Michael J. Moore.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 76

76 The Frontiers of Imperial Rome There was no Vallum on the Antonine Wall, but many forts had an annexe attached to one side. We may conclude that the existence of the Vallum on Hadrian’s Wall rendered annexes superfluous there, and, as the Vallum appears to have had a military purpose, there is a good reason to suppose that the northern annexes were provided for army rather than civilian use. Several contained the unit’s bath-house, but otherwise excavation has revealed little about structures within them.

The Outer Limes in Germany Throughout most of the reign of Antoninus Pius, the Hadrianic frontier in Germany was maintained. As timber towers decayed, they were replaced in stone on an adjacent location; for example, such work was in progress in 146 when tower 10/33 on the Odenwald limes was rebuilt.13 Unusually, some of the new towers were not square but hexagonal. Inscriptions suggest that this frontier continued in use until at least 158 or 159. At the end of the reign of Antoninus Pius, there was a significant change when the frontier was moved forward about 30km (18 miles) in both Upper Germany and Raetia (fig. 11). This move cannot be dated in Upper Germany other than that it appears to have occurred later than 158 and 159. Dendrochronology undertaken in Raetia, however, has recently provided new evidence for the dating of this action. Dates from timbers in the newly aligned palisade range from 160 to 165.14 These dates all come from the section of the frontier not completed under Hadrian. Indeed, at the present time, there appears to be little evidence that any of the palisade in Raetia can be dated to the reign of Hadrian. The reason for the forward move of the frontier in these two provinces is not known. The move brought into the empire rich farmland, which was soon developed, and it appears that it brought the frontier up to the edge of the forest within which there are few known settlements. The advance, however, was at about the time that the Chatti invaded this frontier region: the event is listed in the disturbances occurring at the beginning of the reign of Marcus Aurelius and Lucius Verus and the special command given to Gaius Popilius Pedo in Upper Germany about 155 may indicate earlier trouble (Historia Augusta, Life of Marcus 8). The move may therefore have related in some way to these events. Whatever the reason, it was certainly a serious construction project to build a new palisade with nearly twenty forts and about 250 watch-towers. One significant effect of the new frontier was that it closed the gap between the palisade in Upper Germany and that in the eastern part of Raetia. The new frontier was a timber palisade like its predecessor. The timber was taken from oak trees between 100 and 200 years old. Some 700 such trees were required for each kilometre of the frontier and many had to be transported some distance to the building site. As in the earlier palisade, the logs were halved, with their smooth sides forming the outer face of the continuous palisade and with larger posts at irregular intervals. The foundation trench was of a similar depth to that of the Hadrianic palisade, from 1–1.5m (3–5ft), and again suggests a fence standing up to

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 77

Linear Barriers 77 3m (10ft) high. Also, as in the Hadrianic palisade, cross braces helped to hold the upright timbers in place. The palisade in Upper Germany is unique in that it ran straight for 81km (50 miles) without any consideration for the terrain, and with a deviation of only 1m (3ft). It is as if a line had been drawn on a map. The line in Raetia to the east followed a more normal pattern. As before, towers were placed on the frontier. They ranged from a little under 300–800m (900–2,400ft) apart and were located to ensure intervisibility. In some areas the towers were noticeably closer together. It has been suggested that this might relate to the need to control movement, perhaps associated with trading, though few settlements are known beyond the frontier in this area. The towers were not of a regular size, but ranged from 4m to 8m (13ft to 26ft) square. They were of stone and plastered, with red lines painted to simulate masonry, so that they will have shown as white structures in the landscape. Roofs appear to have been of oak shingles. There were gaps in the palisade, though not protected by fortlets as on Hadrian’s Wall and the Antonine Wall. However, there appears to have been a tower beside each gap. Some fortlets are known on this line, positioned to guard passes through the hills. The forts on the old line were abandoned and their regiments moved up to new stations; they were placed about 11km (7 miles) apart. Woolliscroft has noted that the forts were all well placed to observe the line of the barrier, with the towers equally carefully located to maintain contact with the troops in the forts, even where the lie of the land is unhelpful.15 In some instances, two auxiliary regiments were placed together. At Echzell in the Taunus Mountains, for example, two units occupied the one fort. At Neckarburken, further south on the Hadrianic line, there were two forts, one occupied by an auxiliary unit of 500 men and the other a numerus. When the fort was abandoned for a new location at Osterburken on the Outer Limes, the two units moved closer together, the numerus occupying the annexe of the fort where the auxiliary regiment was based. At Welzheim at the southern end of the Outer Limes, the main fort was occupied by a 500–strong cavalry unit and the other by two numeri. There is a similar situation in Raetia, though the distances between the ‘twinned forts’ are greater. At Rainau-Buch close to the frontier is a mixed 500–strong infantry and cavalry unit, while the province’s 1,000–strong cavalry regiment is located a little further back at Aalen. Ellingen was probably the base of a numerus, while at Weissenburg there was a cavalry unit. Finally, towards the east of the province, a numerus probably occupied Böhming and a mixed infantry and cavalry unit the nearby fort at Vetoniana/Pfünz. In each case, the lower grade unit was placed closer to the linear barrier. Dietwulf Baatz sees the two types of units having different functions, the extra auxiliary regiments being placed in these strategic locations in order to act as a reserve ready to move along the line of the frontier whenever required without interrupting the routine duties of frontier watch undertaken by the other units. In particular, he draws attention to the placing of two cavalry regiments close to the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 78

78 The Frontiers of Imperial Rome boundary between Upper Germany and Raetia, at Welzheim and Aalen.16 They faced, across the frontier, country cut by deep ravines and with dense forests – still in existence today – which were unsuitable terrain for cavalry, thereby suggesting that their primary line of movement lay along and not across the frontier. The new frontier required maintenance and it is likely that it was when the palisade was sufficiently decayed that it was replaced in Upper Germany by a bank and ditch erected behind the timber fence (figs. 21 and 22). The material from the ditch was used to create the bank, which sat on the inner lip of the ditch. There is no evidence for a parapet along the top of the bank. A coin of 194/5 found in the bank to the north of the Saalburg is the best dating evidence for its construction. In the Taunus Mountains a dry-stone wall was erected in some places probably because the hardness of the rock would have made ditch digging difficult, and elsewhere there were short gaps, usually where swamps occurred. In Raetia, the palisade was replaced by a stone wall, 1.2m (4ft) wide, built of coursed rubble on a shallow foundation, though not before some sections of the palisade appear to have been replaced by a wattle fence. There was no ditch in front of this wall. The stone wall and the bank and ditch met at the provincial boundary in the Rotenbach valley where an altar was erected to the goddess of the fines. The construction of the wall is traditionally dated to the early third century, and this is supported by new evidence indicating that it was erected in 206. Therefore the improvements to the frontier might relate to the growing threat from the German tribes which led Caracalla to campaign against them in 213. The linear barrier through Upper Germany and Raetia remained the frontier of the Roman empire until the 260s, its forts undergoing periodic repair. The frontier installations survived the attacks which appeared to start about 230, evidence for

Fig. 21: Diagram illustrating the development of the limes in Upper Germany: 1. the path and timber towers; 2. the palisade added; 3. the timber towers replaced in stone; 4. the palisade replaced by an earthen bank.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 79

Linear Barriers 79

Fig. 22: An artist’s impression of the later wall and stone tower in Raetia and the frontier in Upper Germany showing the palisade and its replacement earthen bank, though both were not in use at the same time, and a stone tower.

destruction being recorded at many places. The province of Raetia was strengthened by the emplacement of a legion, the Third Italica, at Castra Regina/Regensburg in the 170s. Yet, such measures failed and in about 260, during the reign of Gallienus, all the territory north of the Danube and east of the Rhine was evacuated. The empire’s boundary returned to where it had been before the advances of the 70s, and became, in effect, a river frontier and will therefore be discussed below.

Hadrian’s Wall reoccupied The exact date of the abandonment of the Antonine Wall is uncertain. While the Roman army erected inscriptions to record the construction of buildings, they did not record their abandonment. The pottery found on the Antonine Wall suggests a date of abandonment no more than a few years after 160. The latest coin found in a fort on the Wall, Old Kilpatrick, is a worn issue of the Empress Lucilla struck in

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 80

80 The Frontiers of Imperial Rome 164. However, a now lost inscription from Hadrian’s Wall recorded rebuilding work on the curtain in 158. It may be that the decision to abandon the Antonine Wall and reoccupy Hadrian’s Wall was taken in, or shortly before, 158, the process then taking some years to complete. It would not be surprising if the task of recommissioning Hadrian’s Wall took a little time to complete. When the frontier had been abandoned, the gates were removed from the milecastles and the Vallum was slighted by creating gaps in the mounds and a crossing over the ditch every 41m (135ft), thus rendering the barrier open to north–south movement. Now this had to be undone. The order was clearly given to fill in the gaps in the Vallum mounds and remove the crossings, though equally clearly the work was never completed, just as the earlier order to slight the earthwork had not been finished. Either then or later the Vallum ditch was cleaned out with the material dumped on the lip of the ditch, usually on the south side. The milecastle gates were replaced and roadways resurfaced. It is possibly now that the causeways over the ditch in front of milecastles were removed. Turrets were brought back into commission. It is unclear how much had to be done to forts, as several have produced evidence that they continued in use during the period the Antonine Wall was occupied. In general, however, the actions suggest that the intention was to return the frontier to its state before abandonment twenty years before. One specific addition was the construction of a road, the Military Way, immediately to the rear of the Wall. It is particularly interesting to note what was not done. There was no attempt to reduce the gaps between the forts to bring Hadrian’s Wall into line with the 3km (2 miles) spacing between the Antonine Wall forts by building new forts. No annexes were built and no expansions were added to the Wall. In that way, the Antonine Wall may be seen as the end of a line of development, starting with the first plan for Hadrian’s Wall on which there were to be no forts, then the placing of forts on the Wall line, followed by a similar scheme for the Antonine Wall which was superseded by the addition of at least ten new forts to the barrier. The Antonine Wall marked the furthest swing of the pendulum and the reoccupied Hadrian’s Wall returned to a more normal pattern (fig. 23). It took some years before significant changes were made to Hadrian’s Wall. At some date – or dates – in the later second century, several turrets, particularly in the central sector, were abandoned. This was merely recognizing reality; many turrets had become superfluous immediately forts were built on the Wall line. Many milecastle gates were narrowed, and it was perhaps at this time that most of the causeways across the Wall ditch in front of the milecastle north gates were removed. The Vallum appears to have gradually fallen out of use, though this is only recognized in the vicinity of forts where civilian buildings were allowed to encroach on the earthworks and approach the fort walls. It would be attractive to relate these changes to known historical events, but this is not possible. One event, however, may have led to recognizable changes. In about 180, ‘the tribes in the island crossed the Wall that separated them from the Roman

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 81

Linear Barriers 81

Fig. 23: Diagram illustrating the development of Hadrian’s Wall and the Antonine Wall: a. the first plan for Hadrian’s Wall; b. the second plan for Hadrian’s Wall; c. the first plan for the Antonine Wall; d. the second plan for the Antonine Wall; e. Hadrian’s Wall in the later second century

legions and killed a general and his troops’ (Dio 72, 8, 2). It appears to have taken the Romans three or four years to recover the situation. One result may have been the changes in the units based at forts, which occurred some time in the later second century, though many of the changes cannot be recognized until inscriptions were erected in the early third century. The placing of an extra unit or units at some forts may also date to this time. In a way, this reflects the placing of extra forts on the Antonine Wall, though now, rather than build additional forts, the number of men based at some forts was increased, though it remains unclear where they lived. We can also recognize that this was part of a wider pattern. North of the Wall it appears that large mixed units of infantry and cavalry were based at the four outpost forts. At each of the two easterly forts the unit was supplemented by two smaller units, one being scouts, while one of the westerly forts, Netherby, was called Castra Exploratorum, ‘the fort of the scouts’. On the Wall, the cavalry regiments were clustered around the two main roads leading north through the Wall, at Portgate in

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 82

82 The Frontiers of Imperial Rome the east and Stanwix in the west. South of the Wall there also appears to have been a pattern, with the cavalry units based on the roads leading northwards towards these two main gates. Perhaps it was only at that point that a sensible and balanced frontier pattern of deployment was created. As far as we can determine, this pattern continued into the late third century when it would appear that some of the regiments in the forts behind Hadrian’s Wall were withdrawn for service elsewhere, perhaps in response to peaceful conditions on the northern frontier. The rise of the Picts in the fourth century created a new threat to which we will return.

Africa As we have seen, Hadrian visited North Africa in 128. His visit is chiefly notable because part of the text of the speech he gave to the army of Numidia survives.17 The speech inscribed in stone at Lambaesis, the base of the Third Augustan Legion, records that he inspected the building of a wall and the digging of a ditch by the Second Cohort of Spaniards, useful skills for soldiers who constructed frontiers. He observed training exercises by the soldiers of the legion and noted the difficulties created by the absence of legionaries serving elsewhere, both outside the province and on outpost duty within. On 7 July 128, the emperor inspected other units on his journey around the province and gave one a financial reward for its performance. All this fits well with the descriptions by the ancient authors of Hadrian’s interest in the army and its welfare. What is more difficult to demonstrate is his involvement in the frontier. By the time of Hadrian, the frontier of Numidia had been pushed westwards to embrace the Aurès Mountains which in effect formed the southern boundary of the province (fig. 42). One line of forts ran along the northern fringes of the mountain range and another to the south. To the north-west lay the Hodna Mountains. Hadrian’s contribution to the frontier works in North Africa is believed to have been the construction of a series of barriers, of different lengths, in a zone to the west and south-west of the Aurès Mountains, extending north-westwards to the Hodna Mountains. Together, they have been designated the Fossatum Africae. They were surveyed, from the air as well as terrestrially, and subsequently partially excavated by Jean Baradez and published in 1949. It was Baradez who gave the name Fossatum Africae to this system of military installations. The name is not based securely on an ancient source; an edict of 409 in the Theodosian Code stipulates the maintenance of a fossatum in a North African frontier zone, but its location is not certain (Cod. Theod. 7, 15, 1). The work of Baradez and his interpretations form the basis for all modern studies of the remains. The extent to which Baradez studied the archaeological remains was variable and his interpretations have not all survived the test of time. He identified four separate linear barriers at Mesarfelta, Seguia Bent el-Krass, Ad Maiores and round the Hodna Mountains. Detailed coverage was provided of the first two sectors. The barrier at Ad Maiores has been reinterpreted as a road, while considerable uncertainty surrounds the Hodna barrier.18

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 83

Linear Barriers 83 The Seguia Bent el-Krass sector lay south-west of the Aurès Mountains, to the south of the fort at Gemellae. The barrier, 60km (37 miles) long, consisted of a mudbrick wall 1.5–3m (5–10ft) wide and at least 2.5m (8ft) high fronted by a ditch varying in size but generally 4–6m (13–20ft) wide and 3m (10ft) deep. Some gates and towers have been found on the line. The towers measure about 4.5m by 5.5m (15ft by 18ft) and the gates are single-portal, with doors front and back, flanked by towers and with a passage about 1.9m (6ft) wide, a causeway crossing the ditch in front of the gate. Baradez argued that there was a gate every mile with a tower in between, but the known sequence is irregular. Further towers were recorded both in front of and behind the barrier, and elsewhere, where their purpose may have related to observation of road junctions and river crossings. Control was extended east and west of the barrier through the placing of forts in appropriate locations. The construction of the headquarters building in the fort at Gemellae, also built of mudbrick, is dated by an inscription of 123–33, that is about the time of Hadrian’s visit. The fort is large for an auxiliary unit, covering 2.9 ha (7.25 acres). North of Gemellae and west of the Aurès Mountains is another barrier. This runs for about 45km (28 miles) in a north-westerly direction from the fort at Mesarfelta to Thubunae and consists of a wall or mound with a ditch, gateways and towers. Its northern flank is protected by forts. The barrier runs parallel to the road and is diverted round the villages along the road. The third sector lies further north, encircling much of the eastern end of the Hodna Mountains. This sector is 140km (87 miles) long. Baradez provided little information on this barrier, his map not being supported by clear photographs of the remains. Such a barrier encircling the mountain range would have been unique in the Roman empire and, especially now that the Ad Maiores section has been reinterpreted, doubt must surround its existence, at least in the form described by Baradez. Baradez acknowledged that the barrier was often placed on a reverse slope. It could be argued that the towers erected in front of the barrier provided for observation over the forward ridge, but too little is known about these to be able to offer any firm conclusions. Moreover, the barriers meander across the countryside rather than take a direct line, which might be thought more appropriate if their purpose was defensive. The Gemellae barrier follows the line between the desert and the sown, the cultivated land, but the Mesarfalta–Thuburnae sector runs parallel to the road as if protecting traffic passing along it. The linear barriers and the forts, together with the topography, appear to be positioned to funnel traffic through particular points. Traffic heading northwards towards the empire in the region of Gemellae would be forced into the gap between the western end of this section of the Fossatum and the Zab Mountains where they would have been monitored by the troops based at the forts at Gemellae and Vescera. To the north, two sections similarly lie on either side of a gap leading north-eastwards from the Chott el-Hodna and in particular into the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 84

84 The Frontiers of Imperial Rome fertile basin centred on Constantine/Cirta. Zarai, on the northern side of this gap, was a Roman customs post. The plentiful provision of gates emphasizes that the aim was not to prevent the movement of people – and, presumably, animals – but control it. Whether they protected the irrigated land from raiders or from nomads practising transhumance is not possible to determine. Fentress has drawn attention to the differences between the narrow gates and the broad gaps. She suggests that although the ‘passages are fairly small, only 1.90m [6 ft 3in] wide, … while short-distance movement of flocks across it [the Fossatum] would not necessarily be substantially hampered, caravan traffic might have been diverted to run through the major centres.’ She further suggests that ‘the fossatum, by re-routing trade, served as a customs barrier through which the movement of flocks, trade, and people could be carefully overseen.’19 Baradez argued for a Hadrianic date for these linear barriers. He likened the towers to those in the fort at Gemellae and to those on Hadrian’s Wall, and he compared the profile of the ditch to that on Hadrian’s Wall. Pottery and coins from his excavations in the vicinity of Gemellae also indicate a Hadrianic date for the gates and towers. Whether the emperor himself was responsible for the construction of the barriers is another matter. Under Hadrian two forts are known to have been built further west in Mauretania Caesariensis, but they date to 119 and 122, that is before the emperor’s visit.

Dacia The fiercely independent – and threatening – kingdom of Dacia was defeated and conquered at the second attempt by Trajan (101–3 and 105–6). Trajan created a new province of Dacia in Transylvania, but also incorporated Wallachia and southern Moldavia into the province of Lower Moesia. Several forts are known to have been constructed there, probably under Trajan, and this has been interpreted as suggesting that Lower Moesia had been extended into this area at the end of the first Dacian War. Hadrian, however, abandoned eastern Wallachia and southern Moldavia, restricting Trajan’s acquisitions to Transylvania and the land to the south and south-west which linked the core of the province to the empire. This decision seems to have been taken in the early months of his reign, presumably when he travelled through the region on his return from Antioch to Rome. The forts already built in eastern Wallachia were abandoned.20 One of the main routes across Wallachia between the province of Lower Moesia and Dacia was the valley of the Olt, which flows southwards from the Carpathians into the Danube. Communication along this strategic route was protected by forts. Across the Wallachian plain, the forts were irregularly placed, up to 35–40km (22–24 miles) apart, but they were closer together in the mountains and there more use was made of small forts and fortlets (fig 40.) Thirty to forty kilometres (18–24 miles) to the east of the River Olt is an earthen bank with its own line of forts linked by a road, generally known as the Limes Transalutanus. Where the road crossed the plain the forts were no more than about 5–15km (3–10 miles) apart, though these

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 85

Linear Barriers 85 tended to be small installations. Further north, the distance increased to about 15–25km (10–15 miles). Some towers have been identified on this line. The bank was not continuous across the Wallachian plain. It ran north from the fort at Flaminda on the Danube for about 50km (30 miles) to Gresia on the River Vedea. To the north the forts lay on the western side of the river, the earthen bank presumably being considered unnecessary here. North of Urluieni, a short length of bank has been recorded on the east side of the river, and further north, between Albota and Pitesti, there is a section where the frontier crossed the col between two rivers. Excavations have demonstrated that the bank is of two phases. The earlier appears to have been of earth and timber, perhaps in the form of a box rampart, and it was replaced by a more substantial earth bank which encompassed its predecessor. In some areas the bank still survives up to 7m (23ft) high. The preferred interpretation is that the bank marked the provincial boundary and controlled movement in the areas where there was no natural physical boundary to the province. There is no dating evidence from the bank. Assigning its construction to the reign of Hadrian depends upon the date of the forts along the frontier and there is little evidence for this; they are usually regarded as Hadrianic because of their square shape, typical of most forts on this line.

Discussion As many as six different artificial frontiers were constructed in Germany, Britain, Africa and Dacia within the relatively short period of sixty years. While they share some characteristics, there are many differences. The location of these barriers within their landscapes is of considerable interest. The position of the Hadrianic German palisade, Hadrian’s Wall in Britain and the Fossatum Africae related to existing sites, the new barriers being placed on the outer side of the earlier installations. While excellent use was made of the topography of the area in choosing a line for Hadrian’s Wall, it resulted in the linear barrier not being placed in the best position if defence was the primary criterion. In several places, the outlook from the Wall is restricted. Nor do the sections of the Fossatum Africae always occupy good defensive positions; rather, they lie on reverse slopes. This, particularly in the Gemellae sector, may have been because the barrier followed closely the line of the earlier road. The Antonine Wall in Britain, the Outer Limes in Germany and the Limes Transalutanus, on the other hand, were not built in relation to any other structures. Otherwise, there was no similarity between them. The Antonine Wall was planned with a close eye to the landscape in which it was placed, its frequent changes in alignment ensuring that it achieved the best defensive position. The Outer Limes, however, totally ignored its landscape, being driven across it in a straight line. David Woolliscroft has offered an explanation for this phenomenon. He has suggested that: There may be not one, but two recurring frontier designs. The first runs the line essentially parallel to the principal landscape features (river valleys,

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 86

86 The Frontiers of Imperial Rome coastlines, etc) and might be called a ‘terrain-following’ system. These tend to have their forts set back from the line for a better view along these features and have direct or near direct signalling systems. The second type might be called ‘terrain-crossing’ systems as the line runs roughly at right-angles to the topography, often running far straighter and taking far less notice of tactical possibilities. These tend more towards linear signalling and their forts are usually much closer to the line, as there is less to be gained from setting them back. Their forts are also frequently set down in river valleys, closer to water and transportation corridors, often guarding the easiest routes into and out of the Roman Empire. (Woolliscroft 2001, 152) The Raetian section of the Outer Limes and the original plan for Hadrian’s Wall are good examples of ‘terrain-following’ systems while the Upper German section of the Outer Limes is a clear instance of the ‘terrain-crossing’ arrangement. Woolliscroft concludes that: only where nothing else was possible do the Romans seem to have adopted the terrain-crossing approach, and in the case of the [Outer Limes in Upper Germany] nothing else would have been possible without running the frontier very much further to the north. (Woolliscroft 2001, 152) Woolliscroft’s model appears to work in that examples of both types of linear barrier can be identified while his analysis offers an explanation for the different types of frontiers and their location within the landscape. Several frontiers used both linear barriers and rivers. There was a break in the Hadrianic barrier in Germany where the River Main formed the boundary, and similarly in Dacia the barrier across the Wallachian Plain stopped when the River Vedea was reached. West of Hadrian’s Wall the milefortlets and towers continued for about 40km (25 miles) down the Cumbrian coast but with no barrier and beyond the eastern end of the Wall the River Tyne formed the frontier. Similarly, on the Antonine Wall, there were forts to the east along the southern shore of the Firth of Forth and fortlets to the west overlooking the Clyde estuary. The materials of construction varied. This may have related to what was available. Timber was readily accessible in Germany, turf in western Britain, earth in Dacia, stone and earth for mud-brick in Africa. Hadrian’s Wall is the odd one out owing to the copious use of stone when building in turf or timber might have been expected. We have already noted that this may relate to the lack of availability of turf and timber. There was usually no ditch in front of the linear barriers on any of the German frontiers – only the bank replacing the palisade in Upper Germany had a ditch and that may be in part because it served as a source of soil for the bank. Both Hadrian’s Wall and the Antonine Wall were fronted by a wide and deep ditch, as was the Fossatum Africae. Discoveries over the last decade have demonstrated the existence

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 87

Linear Barriers 87 of pits holding obstacles in various sectors of the berms on both the British Walls, but no such features have been found in Germany. Towers appear on all the frontiers, apart from the Antonine Wall in Scotland. This difference has encouraged archaeologists to believe that there are towers yet to be found on the North British frontier. On Hadrian’s Wall, the towers are very regular in size, but on the German frontier they range from 4m to 8m (13ft to 26ft) square. They are roughly the same distance apart on both frontiers, 500m, that is ⅓ Roman mile. The provision of gates through the frontiers was different. Both Walls in Britain had fortlets with defended gates provided at regular intervals, 1 Roman mile (1.5km) on Hadrian’s Wall and probably the same interval on its more northerly neighbour. This type of fortlet is not found in Germany. The invention of the milecastle on Hadrian’s Wall may have been partly a response to a perceived necessity to protect these gates as well as to provide accommodation for the soldiers on frontier duties. The gates through the German barriers appear to have been provided only where they were considered necessary, and were usually overseen from a tower not a fortlet. On the two British frontiers the forts were attached to the line of the Wall; in Germany and Africa they always lay a little behind. Forts appear at roughly the same distance apart on some frontiers, that is Hadrian’s Wall, the first plan for the Antonine Wall, and the Outer Limes. This distance, 11km (7 miles), is about half-aday’s march. But this was not an immutable rule and on the German frontier it was more usual for forts and fortlets to be placed at more irregular intervals to control crossing points. On the Wallachian Plain, the small forts tend to be 5–15km (3–9 miles) apart. The provision of additional forts on the Antonine Wall, reducing the spacing to 3 km (a little over 2 miles) is unusual, as is the close spacing of fortlet and small enclosures in one sector of the frontier. Small units known as numeri are recorded on the German frontiers throughout the second century, some having been recruited from Britain, but none appears on Hadrian’s Wall until the early third century, and none is known on the Antonine Wall. Annexes lay beside many of the forts of the Antonine Wall. While their purpose remains a matter of debate, no one argues that they held a unit such as occurred at Osterburken on the Outer Limes. No other annexes are known on the German frontier and none at all on Hadrian’s Wall, though here the Vallum might have been regarded as providing the equivalent of an annexe along the whole length of the frontier. Distinctions between frontiers can be extended to the military landscapes within which the frontiers sat. In Germany nearly all the forts sat beside the linear barrier. Only the two legionary fortresses were left behind, perhaps because the rivers offered good supply routes, and a single auxiliary fort at Friedberg on a major route into the province, and also in a good topographical position to provide support to units along that section of the frontier. In Britain, on the other hand, both Hadrian’s Wall and the Antonine Wall sat within a deep military zone which stretched from

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 88

88 The Frontiers of Imperial Rome outpost forts, past the barriers south to hinterland forts and the legionary bases (see pages 161–63). These, like those in Germany, lay on navigable rivers. Even within a single province there were distinctions between different sectors of the frontier. There were small forts on the Odenwald section, but not along the River Neckar to the south. Distances varied between forts and small forts from one sector to another. Communications along frontiers were important and therefore roads usually a regular feature. There was a track along the German frontier and in some areas a road, and the track was reproduced on the Outer Limes. There was no road close to Hadrian’s Wall in the first plan, perhaps because the Stanegate already existed across most of the isthmus. A road, however, appears to have been part of the original plan for the Antonine Wall. It was only when the army returned to Hadrian’s Wall in the 160s that a road was constructed along the frontier. The differences extend to how the builders recorded their work. On the line of Hadrian’s Wall, the legions and their constituent parts, the cohorts and centuries, marked the end of the sectors they built by the erection of simple stones, each bearing an abbreviated statement, sufficient help, perhaps, for an inspector in his duties. The building of forts was marked in the more normal way with dedicatory inscriptions, as were those constructed on the Antonine Wall and on German frontiers. Each end of the different building sectors on the Antonine Wall, both north and south of the barrier, was, however, recorded by the erection of an elaborately carved stone which recorded the names of the emperor, the legion and the length constructed. It is difficult to avoid the conclusion that the erection of the Antonine Wall distance slabs was intimately connected to the reason for the building of the Wall itself, the need for Antoninus Pius to gain a military triumph, for they record the successful execution of the emperor’s orders and the support of the gods in his venture. The lack of such items on the Outer Limes in Germany in turn may suggest a different, more mundane reason for the forward move there. These differences between the frontiers suggest that while there was an overall – let us say, imperial – policy to create a frontier, it was largely left to the individual governor to decide the details, perhaps encouraged by the emperors’ distrust of too much communication between their governors. This would account for the rebuilding of the Upper German frontier as an earthen bank and ditch and that in the adjacent province of Raetia as a stone wall. In the case of Hadrian’s Wall, I would argue, the involvement of the emperor led to the unusual, indeed unique, nature of that frontier. This will not explain the broadly similar plan for the Antonine Wall. Here the best answer is still that offered by John Gillam that the more northerly frontier was built as an improved replica of its predecessor.21 There may have been links of an altogether different nature between frontiers. In 155–8 troops from the three legions of Britain served in Germany (RIB 1322). Was this in relation to warfare in Germany and/or the movement of the frontier? Was the Antonine Wall abandoned in order to provide more men for Germany? It would be interesting to understand such links better.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 89

Linear Barriers 89 A further difference between the frontiers is of particular importance. The various elements of the African, Dacian and German frontiers tend to be separate and not physically connected, while those on the British frontiers are mostly linked. It is therefore easier to distinguish the changes during the building processes and these changes require interpretation. It can therefore be seen that on both Hadrian’s Wall and the Antonine Wall during the building process more troops came to be placed on the linear barrier itself. This may suggest an increasing concern with control, mobility of the army, or reactions in the face of hostility. The construction of the Vallum behind Hadrian’s Wall restricted the movement of the army, but increased the level of control in the frontier area. The lack of troops on the Wall in the first plan for Hadrian’s Wall allows us to differentiate between the different military roles – defence, the prevention of raiding, and frontier control. The requirement for mobility is emphasized by the placing of the forts astride the linear barrier. In the first scheme for Hadrian’s Wall, the apparent need of the soldiers based in the milecastles and turrets to maintain communication with those in the forts and fortlets to the south appears to have led to the Wall not taking the best defensive line, an important indicator that these needs were not paramount to the army. This in turn can be related to the position of other linear barriers which did not take the best line if defence was the main criterion. Yet, Hadrian’s Wall and the Antonine Wall were each made into an effective obstacle by a substantial ditch and, in places, stakes and branches, while the Wall itself was a substantial stone or turf structure.

The role of Hadrian One of the attractions of studying a period when there is both literary and archaeological evidence is to try to relate the archaeological remains to the events in literature. This is treacherous ground. Many actions have been ascribed to Hadrian, and to other figures mentioned in the literary sources, merely because of their appearance in such sources. Sometimes, it appears that the history of the empire hiccuped along in relation to these references, the only actions occurring when a figure is mentioned in literature. Archaeology can demonstrate that actions occurred between the events recorded in literature. To Hadrian has been ascribed the honour of creating the artificial frontiers of the Roman empire. Yet, we now have a short length of a linear barrier apparently constructed in Germany during the reign of his predecessor and a date for a timber from the palisade in Germany, the construction of which he is generally believed to have ordered, a year or two before he arrived in the province. It has even been argued that Hadrian’s Wall was started before the arrival of the emperor, the second plan being instigated as a result of the intervention of the emperor.22 It is impossible to reconcile the various strands of evidence without new discoveries. However, we could acknowledge that under Hadrian many cities were honoured by grants raising their status – more than under his predecessor and or his successor – while the literary sources stress his interest in his army, frontiers and

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 90

90 The Frontiers of Imperial Rome the protection of the – his – empire. It is within that general framework that we should see the great changes in the frontiers that we can assign to his reign.23 There would appear to be several similarities between the first plan for Hadrian’s Wall and the Fossatum Africae. These include the position of the barriers within the landscape, the provision of gates and towers, and the location of forts some distance to the rear. On the other hand, the frontiers in Germany and Britain are not alike in detail. The German frontier was a relatively simply fence, albeit a stout palisade, with no ditch. Gaps through it were simple, while towers and forts were placed behind the barrier. Hadrian’s Wall, however, was a substantial wall of stone and turf, with obstacles on the berm, at least in part of the eastern sector, defended gates at every mile, two towers between. Later forts were placed astride the barrier and a protective earthwork, the Vallum, behind. Some of the differences may relate to the availability of building materials in Germany and Britain. It has been argued, however, that the British frontiers were so substantial because of the requirement to protect the provincials against the continual pressure from the unconquered peoples north of the Walls. Yet another view is that in Britain Hadrian decided to build a frontier in the style with which he was familiar, that is the Greek city walls. This might account for the size of the stone sector of Hadrian’s Wall and the possibility of a wall-walk. Further, the various unique elements of Hadrian’s Wall – the use of stone and the fortlets of the first plan and the forts astride the Wall and the Vallum of the second – could both have been the result of his involvement if he stayed on the island for just three or four months. Perhaps the emperor considered the frontier in Germany too extensive for such treatment; perhaps work had commenced before he arrived and he let it continue; perhaps his disappointment in that frontier led to an over-reaction in Britain.24 In spite of the well-attested interest of Hadrian in the army and in frontiers, his frontier policy did not constitute a root-and-branch review of what was needed to protect best the empire and its inhabitants. Rather, it was an acceptance of the status quo and a consolidation and development of the existing frontiers. This is perhaps most clearly seen in Germany where the palisade was simply placed immediately in front of the existing towers. In Britain, there were also forts across the Tyne–Solway isthmus, supplemented in one 24–km (15-mile) stretch by three small forts and several towers, an embryonic frontier. The Wall was again built to the north of these existing installations and, from analysis of the line of the Wall and the location of the installations on it, was intended to operate in conjunction with these existing forts. Antoninus Pius was little different. His advance in Britain was not with the intention of dealing with the problem of the northern frontier in Britain and completing the conquest of the island, but a relatively short step forward, taking in territory which had been conquered and abandoned sixty years ago and had presumably been under Roman surveillance since. This is explicable if it is accepted that the reason for the advance was to give the new emperor a triumph, military

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 91

Linear Barriers 91 prestige. The advance of the frontier in Germany to the Outer Limes was an even shorter step forward, taking into the empire good farmland which had certainly been under Roman surveillance. But while Antoninus Pius moved the frontier, he did not abandon the concept of the linear barrier; he simply ordered the construction of new barriers on the new imperial boundaries.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 92

Chapter 8

River Frontiers

ivers have been seen by many as forming good frontiers.1 At least they have the advantage that they are visible. They can, however, freeze over, as several Roman writers described. I started this book with reference to the freezing of the Rhine in 406 and its consequence for the Roman Empire (Orosius 7, 38, 3–4). This was by no means the first time that the rivers bounding the Roman Empire had been crossed in this way. In the 30s BCE and again in 10 BCE the Dacians crossed the frozen Danube (e.g. Florus 2, 28). In 89 the Rhine froze during the rebellion of the governor Saturninus. Fortunately, it thawed before his allies, the Chatti, could cross the river from Germany to join his army (Suetonius, Domitian 6, 2). The Romans successfully defeated the Iazyges in a battle on the frozen River Danube in 173/4 (Dio 72, 7, 1). In addition, the water level might fall, rendering a river easier to cross. This caused concern during the revolt of Civilis in 69/70 (Tacitus, Histories 4, 26). The anonymous writer of an early-fourth-century panegyric also mentions a lower water level resulting in fear (Pan. Lat. X (2), 7, 3–7). Rivers can act as good conduits for communication. People who live on river banks can usually sail or row and the river is an easier mode of travel than land. River basins therefore often form cultural zones. Nevertheless, to the Romans rivers formed a convenient marker, as we have seen. Caesar regarded the Rhine as the boundary between the Gauls and the Germans, though some tribes occupied land on each side of the river (Caesar, The Gallic War 4, 1; 4, 4; Tacitus, Histories 4, 12). It may be that Caesar’s statement was his justification both for pushing forward his conquests to the Rhine and for stopping at that point. Later, Drusus, the stepson of Augustus, was recorded as reaching the River Elbe and Lucius Domitius Ahenobarbus as crossing the river (Dio 54, 33, 1; 55, 1, 2). Perhaps that was seen as the boundary of the new province of Germania before the Varus disaster of 9. Tacitus, in his discussion of the death of Augustus in 14, remarked that ‘the empire’s frontiers were on the ocean, or distant rivers’ (Tacitus, Annals 1, 9). The River Danube also formed a convenient boundary for Rome. It was the northern boundary of the kingdom of Noricum (modern Austria), for example, and ancient authors recorded that Augustus brought Roman arms to the river bank. In the 70s, the governor of Moesia on the Lower Danube listed his activities, which included operations beyond the Danube, yet within the statement is the phrase ‘he brought kings hitherto known or hostile to the Roman people to the bank of the river which he was guarding’, thereby indicating the nature of the river as a boundary, and a defended boundary at that (CIL XIV 3608 = ILS 986).

R

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 93

River Frontiers 93

The River Rhine Caesar may have conquered Gaul in the 50s BCE, but it was not pacified. Uprisings and invasions punctuate the first half of the reign of Augustus; one invasion involved the loss of a legion. Troops were therefore needed both within the province and on its boundary, the Rhine. Sometime before 13 BCE, it would appear, all legions were moved to the Rhine in preparation for the campaign of Drusus into Germany. This took place over the following years until Drusus died following a fall from his horse in 9 BCE. At this time, there were six legionary bases on the Lower Rhine, from Noviomagus/Nijmegen at the start of the delta upstream to Colonia Agrippinensis/ Cologne. Three were occupied by two legions, Noviomagus/Nijmegen, Vetera/ Xanten and Novaesium/Neuss. Vetera/Xanten and Mogontiacum/Mainz were placed astride the two main invasion routes into Germany. We know little of the location of the auxiliary regiments at this time but the information that we have suggests that forts were generally about 40–50km (25–30 miles) apart. At least two forts were occupied north of the Rhine delta including one beyond the Ijsselmeer (formerly the Zuyder Zee). Over the period from 9 BCE until 9 CE, Germany was organized into a province. Roads were built, cities and markets established, and forts, including a legionary base, were constructed, several on the major line of penetration into Germany, the River Lippe which flowed into the Rhine at Vetera/Xanten. The eastern boundary of the province was probably the River Elbe, but no Roman bases have been recognized there and, indeed, there is no necessity for there to have been any so far east. By 6 CE, it would appear that Augustus regarded Germany as pacified and he appointed the first non-military governor and moved on to the next step in his expansionist policy. However, firstly rebellion in Dalmatia and then in 9 the massacre of the governor of Germany, Quinctilius Varus, and his army of three legions and nine auxiliary units brought all to an end. A younger Augustus would have bounced back, but by then the emperor was in his 70s and had lost his resilience. Germany was evacuated. In the wake of the Varus disaster, all surviving troops were withdrawn to the Rhine and there they settled down. Two legions occupied each of the fortresses at Vetera/Xanten, Colonia Agrippinensis/Cologne and Mogontiacum/Mainz waiting for a return to more positive days and for a successful re-invasion of Germany. That day never came. Germanicus campaigned in Germany in 14, 15 and 16, but Roman resources were stretched and Tiberius brought the campaigns to a halt. Roman intentions remained expansionist, even if Germany was generally seen to be difficult to conquer. Gaius – better known as Caligula – raised two legions, presumably because he intended to undertake an invasion, perhaps of Germany, possibly of Britain. There were later campaigns, but, even after Domitian’s limited move forward in 83, Tacitus could remark, in 98, ‘the conquest of Germany is taking such a long time’ (Germania 37). In fact, it had ended eighty years before. The tacit acknowledgement of the abandonment of the intention to conquer Germany can be recognized on the ground, not least in the breaking up of the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 94

94 The Frontiers of Imperial Rome

Fig. 24: Map of the frontier in Lower Germany. The late forts are shown as white squares.

double legionary fortresses, Colonia Agrippinensis/Cologne in 35, Vetera/Xanten in 69 and Mogontiacum/Mainz in 89. This process in turn gave an impetus to the development of the frontier. These six legions were now given new bases along the Rhine, or moved to the Danube, and the gaps gradually came to be occupied by auxiliary forts. Each of these newly constructed forts tended to be placed at a point where a tributary flowed into the Rhine. This may have been because movement between the tributary and the main river could be better monitored, but these points offered good locations for harbours sheltered from the strong flow of the main river.2 The fleet was an important element in the defence of the empire. The German fleet was founded under Augustus and from about 30 was based at Alteburg, a little

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 95

River Frontiers 95 upstream from the legion at Colonia Agrippinensis/Cologne. The primary purpose of the fleet was presumably to help the army maintain watch over the frontier; indeed, the fleet was regarded as a branch of the army. Tacitus’ description of the demise of the Rhine fleet during the Batavian Revolt implies that the twenty-four vessels were divided among several forts (Tacitus, Histories 4, 16). By the time of the Notitia Dignitatum fleets were based at many sites along the river. At the time of Claudius, there were changes on the Rhine, mainly observable in the delta (fig. 24). In the reign of Tiberius, there were few forts in the delta, perhaps only Fectio/Vechten, strategically placed near to the junction of the Old Rhine and the Vecht which flowed into the Ijsselmeer. The Frisii, north of the Rhine delta, remained allied to Rome and were taxed by the Romans at least until 28. A fort remained in their territory at Velsen north of the Old Rhine and beside the North Sea. Claudius withdrew all units from the lands of the Frisii. This had an unfortunate effect for it encouraged their neighbours the Chauci to invade the province in 47. Corbulo was sent to retrieve the situation, but appears to have exceeded his authority and was ordered by the emperor to withdraw across the Rhine. It may not be coincidental that now some auxiliary forts appear to have been built along the bank of the Old Rhine downstream from Fectio/Vechten and it is possible that the newly discovered timber towers in the same area date to that time; they were later linked by a new road (pl. 10). The forts were much closer together than normal, being about 7–8km (4½–5 miles) apart and smaller than usual, often covering only 1.5 ha (3.7 acres). Further upstream some new forts were built but overall the pattern was not changed, with forts generally about 20km (13 miles) or so apart, that is a day’s march, and larger than those in the delta. Later in the first century, the distance between forts was reduced to about 15km (9–10 miles) from Fectio/Vechten upstream to Vetera/Xanten. Fectio/Vechten retained its importance with the only cavalry unit known to have been stationed in the delta based there in the second century. The other cavalry units in the province were located on each side of the legionary base at Vetera/Xanten, reflecting an earlier disposition of units at the base of the invasion route into Germany. The occupation of land beyond the Lower Rhine under the Julio-Claudians was mirrored upstream where from the 40s some forts were maintained across the Rhine in the vicinity of Mogontiacum/Mainz, again demonstrating that the Romans did not see the river as the firm boundary. Further, the land across the river was kept free of Germans and assigned to the legions as part of their territory. Unfortunately, it was not used productively and in 58 the tribe of the Ampsivarii complained that, while they were without land, these tracts, intended for grazing horses and pack animals, were not in fact used by the army and were lying waste (Tacitus, Annals 13, 54). From the time of Claudius, the Upper Rhine gradually ceased to be the frontier, though a military presence was maintained by the fleet. In 69/70, a serious uprising broke out on the Lower Rhine, the Batavian Revolt. Rome suffered a severe shock when Roman forts were attacked and destroyed. Following crushing of the rebellion, changes were instituted. Some, such as the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 96

96 The Frontiers of Imperial Rome move from the area of the disaffected troops do not concern us here, but others affected deployment on the frontier. Noviomagus/Nijmegen was reoccupied by a legion in order to help watch over the Batavian region, while other forts were rebuilt along the whole frontier. The most significant change thereafter was the gradual shift of the focus of military activity from the Rhine to the Danube. This was in part a recognition of the abandonment of the intention to conquer Germany and partly the result of the pressure of the independent kingdom of Dacia on the Lower Danubian provinces, notably in the 80s. Domitian tacitly recognized the abandonment of expansionist ambitions in Germany by the conversion of the military commands into two new provinces in the frontier area, Upper Germany and Lower Germany. The eight legions based on the Rhine under Augustus were reduced to four, two within each province. The 150 years after the Batavian Revolt appear to have been largely peaceful along the lower Rhine, though there were raids and minor invasions, including that of the Chauci defeated by Didius Iulianus in the late 160s (Historia Augusta, Life of Didius Iulianus 1). In the early third century, however, the frontier regions were troubled by the aggressive activities of the German tribes, which brought the emperors Caracalla and Severus Alexander to the frontier. Continuing pressure on the lower Rhine and the land between the upper reaches of the Rhine and the Danube caused both areas to be abandoned in about 260. On the Lower Rhine, Noviomagus/Nijmegen now became the lowest held fort on the river, with a new installation built on the Valkhof, the site of the earlier legionary fortress and where Charlemagne later had a palace. A line of timber fortlets along the road from Colonia Agrippinensis/Cologne through Maastricht to Bagacum/Bavai was constructed, perhaps in the 260s and 270s, and two towers are also known on this line. This, however, was not a frontier, but rather a road, protected because of its importance as a communication route between the frontier and its hinterland, in the manner that such roads had been protected from the early empire. The area between the Rhine, the Maas and the Scheldt thus abandoned was soon occupied by the Franks. Upstream, new stone fortlets were built in some locations beside the river. The abandonment of the land between the upper reaches of the Rhine and Danube brought the frontier back to the rivers, more or less back to the line it had been under Claudius 200 years before. The Alamanni and other tribes maintained their pressure, invading the empire at least three times during the 270s. However, emperors were taking steps to protect the empire. The walls of the early first-century legionary fortress at Vindonissa/Windisch on a tributary of the Rhine in modern Switzerland were rebuilt in 260. New forts were erected along the river as far as Lake Constance/the Bodensee, and thence along the south side of the lake to Bregenz. Some of these date to the reign of Diocletian for Vitudurum/Oberwinterthur has yielded an inscription of 294. He was also probably responsible for the placing of a newly raised legion, the First Martia, at Castrum Rauracense/Kaiseraugst, a strategic location at the bend of the Upper Rhine.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 97

River Frontiers 97

Fig. 25: Map of the late Roman frontier on the Danube-Iller-Rhine line. The late-fourthcentury towers along the frontier are indicated.

Diocletian was also responsible for the formation of new units of ripenses for the frontier. Constantine continued this building work, but was also responsible for a striking new initiative (pl. 2). Hitherto, so far as we can see, there were no permanent bridges over the Rhine. Caesar bridged the river, but it was merely to illustrate his power and when his objective was achieved, he ordered its demolition. Tacitus records the existence of a bridge at Vetera/Xanten during the campaigns of Germanicus in Germany between 14 and 16. The fainthearted sought its destruction in 15, but Germanicus’ wife Agrippina forbade it. This is the only reference to the bridge. Other bridges across the upper reaches of the Rhine or the Main lay within Roman territory. This was to change under Constantine. In 310, he had a bridge built across the Rhine at Colonia Agrippinensis/Cologne, linking to an existing fort at Divitia/ Deutz on the far bank (fig. 26). He then replaced the fort in 312–15 (CIL XIII 8502). The new fort, ‘built in the presence of the emperor’, had thick walls, bastions, two single-portal gates protected by projecting towers, and contained sixteen buildings, most apparently barrack-blocks. The fort opposite Mogontiacum/Mainz also appears to have been rebuilt about this time, replacing a third-century predecessor. The bridge here is undated.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 98

98 The Frontiers of Imperial Rome Fig. 26: Plan of Divitia/Cologne.

Constantine was active elsewhere along the Rhine. Upstream from Colonia Agrippinensis/Cologne lie a series of early-fourth-century forts, all constructed in the new style. They lie at Bonna/Bonn, Rigomagus/Remagen, Confluentes/Koblenz, Bodobriga/Boppard and Bingium/Bingen and are roughly evenly spaced at intervals of 20km (12 miles) or so (fig. 11). Few are securely dated and have generally been assigned to this period on account of their archaeological dating evidence roughly coinciding with the known activity of Constantine. More rigorous consideration of this evidence, however, suggests that Bodobriga/Boppard, for example, is more likely to have been built later under Julian. Smaller forts, such as Haus Bürgel, of a size to be categorized earlier as fortlets, were also erected under Constantine. Other defended sites, large enough to be forts or small enough to be fortlets, are less easy to categorize as so many civilian settlements were now defended. Behind the frontier line, movement along roads was now more guarded by forts and fortlets. These installations protected links, for example between Augusta Treverorum/Trier, then one of the chief cities of the empire and a significant seat of government, and the frontier to north and east, controlled passes and sought to ensure the safe movement of supplies. In some areas, these fortified sites are up to 100km (60 miles) behind the frontier. Their foundation dates are often uncertain, ranging from the late third century well into the fourth, some not being built until the reign of Valentinian. Following Constantine, the next activities recorded in literature date to the middle of the fourth century. In 355, Julian, cousin of Constantius II, was appointed

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 99

River Frontiers 99 Caesar in the West and undertook an energetic series of campaigns against the Alamanni and other Germanic enemies of Rome, and repaired and rebuilt forts on the frontier. In 357, he repaired Tres Tabernae on the Rhine: which the enemy had recently destroyed by a determined attack. Its rebuilding guaranteed that the Germans could no longer penetrate, as they had been used to do, into the heart of Gaul. He finished the work sooner than expected, and stocked it with a whole year’s provision for the garrison which was to be stationed there. (Ammianus 16, 11, 11). The last comment is reminiscent of the statement by Tacitus that Agricola always ensured that his forts contained a year’s supply of food (Tacitus, Agricola 22). In the following year, Julian repaired three forts on the River Meuse ‘to place the security of the provinces on a firm basis’ (Ammianus 17, 9, 1). This work of protecting the Rhine frontier was continued by Valentinian (Ammianus 28, 2, 1). An inscription records the erection of a tower near Confluentes/Koblenz in the early 370s. Along some stretches of the upper Rhine, the towers were as close as 1–2km (⅔-1⅓ miles) apart. In the fourth century, some fortified landing places were erected on the opposite bank rather in the style of a ferry terminus (figs. 27 and 29). They take the form of a three-sided enclosure with the river forming the fourth side. On the side opposite

Fig. 27: An artist’s impression of a fortified landing place.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 100

100 The Frontiers of Imperial Rome the river sits a rectangular enclosure with a tower on the landward side and the gate below and, usually, small square or round projecting towers at the corners. Sizes fall within the range of 50m by 30m (165ft by 100ft) up to 60m by 50m (190ft by 165ft). Some are protected by a ditch. Three are known on the Rhine at Rheinbrohl near Andernach, Engers opposite Confluentes/Koblenz, and Mannheim-Neckerau opposite Altrip, with another two possible sites at Niederlahnstein between Confluentes/Koblenz and Bodobriga/Boppard and at Wiesbaden-Biebrich. Dating evidence is slight, but some is consistent with a date for their construction, or at least occupation, during the reign of Valentinian.

The River Danube It is not easy to characterize the Danube river and its valley. At times it flows through a relatively wide valley, elsewhere through mountains or gorges. It might be sluggish, not least in the delta, or rushing through a narrow defile at the cataracts known as the Iron Gates in modern Serbia where river traffic was impossible. Such aspects had an effect on military deployment along its banks. It was Augustus who sought to bring the intransigent peoples of the Balkans under Roman rule. His generals campaigned both inland and along the Danube. By 6, Roman arms had been brought to the river down to the delta and campaigns were launched across the river. It was from the Danube, probably Raetia, that Domitius Ahenobarbus set out to march to the River Elbe sometime between 6 and 1 BCE. In 6, Rome was poised to invade the Marcomanni of Bohemia until thwarted by the uprising in Illyricum and then the Varus disaster. Throughout this period, the main army bases lay in the interior of each province. That at Siscia/Sisak consisted of twenty-three cohorts in the 30s BCE and they were later replaced by a legion. Other legions controlled population centres and lines of communication. As these newly acquired territories became peaceful, movement of the legions became possible. They were generally transferred to the Danube. There is some evidence for a military presence under Augustus at several sites on the river, including Mursa/Osijek, Sirmium/Mitrovica in Pannonia and Oescus/Gigen in Moesia (fig. 28). In north-west Pannonia, the earliest known Roman military installation was a small square stone building, possibly a tower, across the Danube on the great rock at Devin near Bratislava in modern Slovakia. This was an excellent observation point overlooking the confluence of the Danube and the Morava. Under Claudius a legion was established at the more convenient site at Carnuntum/Petronell a few kilometres to the west and it was probably at this time that Devin was abandoned by the army. At Carnuntum/Petronell the legion was carefully placed to control the point where the Amber Route entered the empire. Elsewhere, along the river some auxiliary units are attested. Those are Arrabona/Győr and Brigetio/Szôny which guarded routes north leading to the territory of the Suebi. A cavalry regiment was placed at Aquincum/Budapest where a route through Pannonia crossed the frontier onto the Great Hungarian Plain, with units also at Lussonius/Dunakömlőd and Lugio/Dunaszekcső where there were

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 101

Fig. 28: Map of the frontier in the Middle Danube region.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 102

102 The Frontiers of Imperial Rome important river crossings. Elsewhere along the Danube, the Claudian dispositions are uncertain. Only four forts are known in Noricum, for example, two in the west and two in the east of the province, each holding an auxiliary unit, though by 69 there were nine such units in the province. The abandonment of forts in the interior and the movement of troops onto the line of the river may be an acknowledgement that Claudius’ only expansionist interests lay in Britain, and that on the Danube, as on the Rhine, he was ensuring security for the empire. There are sites that emphasize our lack of knowledge about this early period in the development of frontiers. At Nersingen and Burlafingen on the upper Danube near Guntia/Günzburg in Raetia, Michael Mackensen has excavated two fortlets (fig. 48). These seem to have been established in the late 30s, continuing in occupation for about fifty years. They appear to have been located to guard minor crossing points on the river. Further downstream in Raetia, similar fortlets have been located, while chance finds point to the existence of others. Together with the forts, they form a chain of military installations along the river.3 The Danube, like the Rhine, had a fleet based on its bank, or rather two fleets as the Iron Gates prevented traffic from passing along the whole river. On the upper part of the river the Pannonian fleet was based at Taurunum/Zamum near modern Belgrade and on the lower reaches the Moesian fleet headquarters lay at Noviodunum/Isaccea in Romania, though it clearly also had a base at Sexaginta Prista – ‘Sixty Ships’ – much closer to Novae/Svistov (figs. 40 and 44). The fleets were established at least as early as Tiberius, though subsequently reorganized. Tiberius was also responsible for ordering the construction of roads to link the frontier to the heartlands of the empire, and for helping to make the Danube more navigable. It would appear that rocks were removed from the river so that barges could be towed upstream and a wooden towing path created, attached to the side of the rock face along the side of the river. This required regular repair, which certainly occurred under Domitian and then Trajan, who had the road cut into the cliff itself. It was military pressure from across the Lower Danube in the 70s and 80s which led to changes in the deployment of troops along the river. During the Civil War of 68–9, in the absence of the two legions of Moesia, the Dacians crossed the Danube, captured some forts and threatened the legionary bases (Tacitus, Histories 3, 46). Following his victory, Vespasian strengthened the frontier by transferring a legion to Moesia to supplement the two legions already based there. Domitian moved more auxiliary units to the provinces of Pannonia and Moesia, which covered the river from Vindobona/Vienna to the delta. The main threat on the Lower Danube was from the powerful kingdom of Dacia located in modern Transylvania. The challenges posed by this state had been recognized by the Romans since the time of Julius Caesar. Now, a new king, Decebalus, was in power and intent on causing grief to the Romans. He invaded the empire in late 85 or early 86, to the surprise of the Romans, defeating the provincial army, killing the governor of Moesia and destroying a legion, probably the one most

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 103

River Frontiers 103 recently moved there. The emperor himself came to supervise the Roman reaction, but that did not prevent another defeat – following Domitian’s return to Rome – in which the praetorian prefect was killed. Domitian returned to the Danube and his army was successful in 88 and peace concluded in the following year. The war had an effect on both the administration and military deployment in the area. While on the Danube in 86, the emperor took various actions. Moesia was divided into two provinces. Upper Moesia ran downstream to a little east of Ratiaria/Archar with Lower Moesia occupying the Lower Danube (fig. 40). Each province was governed by a consular, emphasizing the importance of this frontier. Two more legions were transferred to the area, bringing the total to five or six; perhaps the intention was to have three in each province. Surprisingly, two legions seem to have occupied the same fortress at Viminacium/Kostolac. Domitian’s troubles on the Danube continued. He invaded the territory of the Marcomanni and Quadi between the victory over the Dacians at Tapae in 88 and the conclusion of peace the following year to chastise them for not providing troops to support his army, but this escalated into a war with the Suebians and Sarmatians. Another legion suffered defeat. This brought Domitian to Pannonia for eight months in 92 or 93 and a victory of sorts, though trouble rumbled on into the next reign. It would appear from the troop movements which we know about that Domitian appreciated the fragile nature of the Pannonian frontier through the 80s. The number of legions was increased from two to four, while auxiliary units were also transferred to the province. As a result, Upper Pannonia became one of the most important military provinces of the empire, and the military balance of the empire shifted decisively from the Rhine to the Danube. This was all the result of the threatening activities of states to the north of the empire in the long stretch from Noricum to the Danube delta. The peace treaty negotiated by Domitian with the Dacians was spurned by Trajan who was soon preparing to invade the kingdom. It took two wars, from 101–2 and 105–6, to defeat the Dacians. A major expansion of the empire took place and a new Roman province was created (fig. 40). The boundaries of this province were formed by the arc of the Carpathian Mountains and are dealt with elsewhere. Trajan had a bridge constructed across the Danube to provide a link with his new province, but his successor Hadrian removed the superstructure. Dio subsequently wrote about this: the bridge is no benefit to us now, but its piers stand useless without a roadway, as though they were built for a single purpose of showing that there is nothing that human cleverness cannot do. For Trajan was fearful that if ever the Danube froze hostilities would break out against the Romans on the far bank, and thus built the bridge so that reinforcements could reach them quickly across it. But his successor Hadrian feared that it would give them an easy crossing into Moesia, and removed the superstructure. (Dio 68, 13, 5–6).

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 104

104 The Frontiers of Imperial Rome This is a nice anecdote illustrating different attitudes, but its veracity has been challenged and the suggestion made that it is a later interpolation into Dio’s text.4 It remains uncertain whether this bridge or others remained in use. The absorption of Dacia into the empire had an effect on the existing frontier along the Danube. Trajan’s conquests included the territory east of the Carpathians and north of the Danube, that is north of Moesia, modern southern Moldavia and eastern Wallachia. It was presumably to help control that territory that two legions were now moved to the Lower Danube. Even though Hadrian abandoned this land north of the river, the legions remained at their new bases. Upstream, some forts were retained along the south bank of the river to control the routes into Dacia. Otherwise, the auxiliary units appear to have been mainly removed from the river in Lower Moesia, which was in effect demilitarized. Rather more auxiliary units seem to have remained in the forts in the part of the river where Upper Moesia met the south-west corner of Dacia. These lay astride one of the main routes into Dacia. There were also some units based in the interior of the province, again probably mainly to guard routes though others probably protected the mines. By the early second century a pattern of military deployment had been established along the Danube which was to survive largely unchanged for the following two centuries. The units were not distributed evenly along the river. Rather, they related to the topography and to the location of potentially hostile populations. The legions were based in the four Pannonian and Moesian provinces. Those at Vindobona/Vienna, Carnuntum/Petronell and Brigetio/Szôny in Upper Pannonia faced the Suebi, Quadi and Marcomanni, while the three at Aquincum/Budapest, Singidunum/Belgrade and Viminacium/Kostolac faced the Sarmatians. In Lower Moesia there were a further two legions ready to support the army in Dacia. These legions were supported by about half of all the auxiliary units in the whole Roman army. The pattern of military deployment varied in detail. The whole stretch of the Norican frontier from Castra Batava/Passau at the junction of the Danube and the Inn down to Krems was densely wooded and almost uninhabited (pl. 1b). Within this stretch were two sectors where the river flowed through steep-sided valleys, for 60km (36 miles) east of Castra Batava/Passau and in the Wachau north of Melk. In the sector east of Passau there was only one small fort and a single fortlet, and there are no known forts in the Wachau. East of here, however, the land changes. The open plain providing rich farmland to the north also offered easy access to potential invaders such as the Marcomanni. As a result five forts were placed in this 50–km (30–mile) stretch. These forts formed the bases for a substantial force of two cavalry units, two mixed units of infantry and cavalry and a 1,000–strong infantry cohort. A similar pattern can be observed at the Danube bend in modern Hungary where there is a close concentration of forts about 8km (5 miles) apart. West and south of the bend, on the other hand, the forts tend to be more evenly distributed, about 22km (14 miles) apart. To the south, this would appear to be because they sit on the high western bank of the river which forms a kind of plateau facing the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 105

River Frontiers 105 monotonously flat Great Hungarian Plain. In these locations they were both far removed from the river – though excellently placed for observation – and subject to erosion; a new fort had to be built at Intercisa further back from the river bank. As on the Rhine, here and elsewhere along the Danube forts were often located at the confluence of a tributary with the main river. In a recent review of military deployment along the River Danube, Sebastian Sommer has considered the location of forts in relation to the river itself.5 He has noted the different patterns of the river, sometimes cutting through hilly terrain, elsewhere meandering across a broad flood plain. Sommer argued that the latter topography would have been difficult to control, surveillance almost impossible and the army at a disadvantage in not wishing to place its forts where they might get flooded. Hence forts were located at ‘the beginnings and ends of large floodplains of the river, and the beginnings and ends of mountainous regions. From here it was easy to control the traffic and sometimes, where necessary, to cross the river.’ Such a place was the Iron Gates. The only known forts in this 130–km (80–mile) stretch are at each end with one in between. Locations were preferred where the banks were stable and where movement across and along the river could be monitored. This accounts, for example, for the location of the legionary fortress at Novae/Sistov in modern Bulgaria, the advantage of the position heightened by the presence of a tributary flowing into the Danube at that point providing a safe harbour and by its location at the north end of a route leading into Thrace. Peace on this frontier was disturbed in the 160s, not on the Lower Danube, now protected by Dacia, but in Pannonia and Noricum. The force of the invasion was such that it led to the presence of the emperor, Marcus Aurelius. He campaigned against the Marcomanni and Quadi for several seasons before dying in Vindobona/ Vienna in 180. During his campaigns, the frontier in this stretch of the river was strengthened. Two new legions were raised, the Second and the Third Italica Legions, and based in the provinces of Noricum and Raetia. The Second Legion was at first based in south-east Noricum, at Ločica, where it could guard the route into Italy, but in the 180s it was moved up to the river, at first based at Albing but then moving to its long-term home at Lauriacum/Lorch. This site was carefully chosen as it guarded a significant river crossing and lay at the north end of an important road into the empire. Like so many legionary bases, it was located at the junction between the Danube and a tributary, in this case the Enns. During the warfare, the army was strengthened by the arrival of mounted archers from Africa and Syria, presumably to help the local troops fight the enemy cavalry. Some of these units remained in Pannonia after the war. At about the same time a fort was also erected across the Danube from the legionary base at Brigetio/Szôny. Celemantia/Iza was a turf and timber fort, built in the 170s when Brigetio/Szôny was one of the main bases for the campaigns against the Quadi. Coins suggest that the life of this fort came to an end soon after 179, that is at the close of the campaigns. It was, however, rebuilt in stone shortly after, an expensive operation as all the building materials had to be ferried across

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 106

106 The Frontiers of Imperial Rome the river.6 This fort continued in occupation until the middle of the third century when it was destroyed by fire. It was rebuilt and continued in use throughout the fourth century. Trajan’s Column depicts towers, presumably standing on the banks of the Danube (pl. 5). Few towers have been located along the Danube, and most of those discovered date to later years. Some timber towers dating to the second century have been located west of the Danube bend in Hungary. In the reign of Commodus there was clearly an intention to build more towers – and forts – probably of stone, south of the bend. Some towers were erected in 184/5. Several stone inscriptions were found at Intercisa, having been removed from the towers when Commodus was damned following his murder in 192 (fig. 4). The inscriptions specifically state that the towers and forts were erected to stop the clandestine crossings of ‘brigands’; in effect, what is meant is that the army wished to exercise control over those nefarious elements seeking to move into the empire (ILS 8913). The abandonment of the land between the headwaters of the Rhine and Danube shortly after 260 led to a reorganization of the frontier (fig. 25). The earliest known fort is at Vemania/Bettmauer, built shortly before 280. It is a typical late fort, on a high point, irregular in shape with thick walls, a single gate and external bastions – the pattern for the future. Most of the evidence for this area, however, points to major activity under Diocletian. Many of the new forts probably held no more than 300 men and many considerably less. The spacing between the forts was greater than before, though this may reflect the lack of river crossings requiring control.7 Severe pressure on the lower Danube region caused the provinces of Dacia to be abandoned shortly before or after 270. The legions were transferred to Oescus/ Gigen and Ratiaria/Archar, while new forts were constructed along this section of the river. Diocletian visited the area in 291 and again in 294 and was active, building in 298/9 at Durostorum/Silistra, Transmarisca/Tutrakan and Sexaginta Prista in 298–9 and Halmyris in 301–2, for example. The six inscriptions of his reign from the Iron Gates in Serbia to the Danube delta form a unique record, not paralleled during the reign of any other third- or fourth-century emperor.8 A significant development at this time was the construction of some forts on the far banks of the rivers and the construction of bridges over the rivers. As we have seen, a fort had existed at Celemantia/Iza opposite the legionary fortress at Brigetio/ Szôny since the 170s. Then it was supplemented by others. In 294, ‘forts were constructed in Sarmatia opposite Aquincum [Budapest] and Bononia [Vidin]’, both legionary fortresses (Constantinopolitan Consular Lists). In the Notitia Dignitatum these are stated as being in barbarico (Occ. XXXIII, 48), which led Mócsy to suggest that they lay not on the left bank of the river but deep into enemy territory.9 Several other fortified sites on the left bank of the river may date to this time. One, called Contra Florentiam/Dunafalva, sat opposite Lugio/Dunaszekcső, guarding access to an important river crossing whence a route led towards Dacia. This continued in use into the early fifth century. As elsewhere, there was usually a pair of fortlets, one on each bank.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 107

River Frontiers 107

Fig. 29: Plan of the fortlets/fortified landing places at Ulcisia Castra/Dera Patak and Contra Florentiam/Dunafalva (Hungary).

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 108

108 The Frontiers of Imperial Rome Fig. 30: Plan of the fort at Contra Aquincum/Budapest (Hungary).

Constantine also had a bridge constructed across the Danube from Oescus/Gigen to Sucidava/Celei in 328, that is, after he had won control of the eastern part of the empire. This was the longest stone bridge across the river. The forts at Whylen opposite Castrum Rauracense/Kaiseraugst, on the left bank of the Danube at Contra Aquincum/Budapest and at Daphne opposite Transmarisca/Tutrakan have also been tentatively dated to the reign of Constantine (figs. 27, 29 and 30). Forts were repaired and strengthened from the late third century throughout the fourth century along the whole length of the Danube, some as late as Valentinian (fig. 32, pl. 2). They include surviving visible sites such as Capidava on the Lower Danube, the walls and projecting towers of which remain impressive today. Even so, there are likely to have been fewer than 500 men based in this fort. Further upstream, Noricum (Austria) also contains some remarkable survivals of this time, fort towers and even a gate still standing to wall-head height (pl. 19). In some areas, the process of filling in gaps continued. The Danube below Sirmium/Mitrovica and the bend between Estergom and Aquincum/Budapest were two such areas. One site is Pilismarót near Estergom, west of the Danube bend. This was oval in shape, following the shape of the hill on which it sat, with projecting towers. It was erected between 294 and 297. A little to the east is Visegrád, triangular in shape, also placed on a hill and dating to the latter part of the reign of Constantine. Even later, in the same area at Visegrád is a fortlet with fan-shaped towers at the corners, built during the reign of Constantius II. Many of the modifications to existing forts are undated. The style can, however, be recognized. It was in the fourth century that fan-shaped towers were added to existing forts, though the dating evidence is scanty. Some forts had one or more of

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 109

River Frontiers 109 Fig. 31: Plan of the fort at Eining (Germany). In the south-west corner of the fort sits a later fortlet.

their gates closed by a curving wall, thus creating a strong point. At others, smaller, well-defended fortlets were erected in the corner of the earlier fort, such as at Eining in the early fourth century, Austura/Zeiselmauer in the later fourth century and Gerulata/Rusovce under Valentinian (fig. 31, pl. 20). The fort ditches tended to be wider than before, with the berm, the space between the fort wall and the inner ditch, also wider. These forts along the river were supported to the south by new fortified sites, both military and civilian. As on the Rhine, these extend as far back as 100km (60 miles) from the Danube, and presumably served similar purposes. These forts in the hinterland were probably mainly concerned with the protection of supplies to the frontier, including the products of the state factories in northern Italy. Amongst the recognizable buildings are granaries or store-houses. Veldidana/ Innsbruck-Wilten is only unusual in being dominated by three large buildings of this type. It may have acted as a collection and distribution point for the army of this section of the frontier. Under Valentinian, a fortified supply base was erected at Tokod, a little to the west of the Danube bend and a few kilometres south of the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 110

110 The Frontiers of Imperial Rome

Fig. 32: Plan of the late forts at Boiodorum/Passau (Germany) and Drobeta/Turnu Severin (Romania).

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 111

River Frontiers 111

(a) (b)

(c) Fig. 33: Plans of towers at: a. Mains Rigg (Britain) b. Asperden (Germany) c. Budahelasz (Hungary) d. Goldsborough (Britain). (d)

river. It has been suggested that these supply bases relate to a policy of defence in depth, but in reality they are little different from the depots which the Roman army had been constructing in the frontier areas since the time of Augustus onwards. As far as we can see, the Danube frontier was quiet throughout much of the fourth century, though fort defences continued to be improved. In 337–40, for example, additional fortifications were erected at Troesmis/Iglita (IGR 238). It was

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 112

112 The Frontiers of Imperial Rome in the 370s, however, that a major programme of work commenced. In 364, Valentinian ordered the local military commander to construct towers to improve the surveillance of the crossing points on the frontier (Codex Theod. 15, 1, 13). This work was extended to almost the whole length of the Danube and is well attested by inscriptions as well as the structures themselves. Towers were also built on the upper Danube along the frontier on the Danube–Iller–Rhine line both east and west of Lake Constance/Bodensee (ILS 8949 is one of five such inscriptions). These towers were generally larger than the earlier timber and stone structures, ranging from 7m to 17m (23ft to 56ft) square, but with walls so thick that the interior could be less than 4m (13ft) square (fig. 33). Most towers were too far apart to signal easily to one another. Their spacing perhaps related to local requirements. Some fortlets have also been discovered on this line. They are a similar size to earlier structures but are now much more strongly defended, with corner towers as well as strong walls. It was about this time, as indicated by dendrochronology, that a new waterfront was constructed for the fleet based at Brecantia/Bregenz on Lake Constance/Bodensee.10 Most of the known towers along the Danube west and south of the bend in modern Austria and Hungary also date to this time. Inscriptions date those west of the bend to 371 (e.g. ILS 774 and 775). Different groups of towers can be recognized on the basis of size, enclosure walls and the shape and number of the ditches. These differences may reflect not just the chronological span of their construction but the activities of different army groups. The Valentinianic towers at the Danube bend were all stone, but in the sector around Intercisa/Dunaújvávaros and Annamatia/Baracs they were of timber. One group of stone towers near Ulcisia Castra/Szentrendre just north of Aquincum/Budapest, for example, are 8m (26ft) square, though the usual dimensions are 10m (30ft). The Valentinianic towers are marked out in another way; they were surrounded by an enclosure wall and a ditch. The resulting enclosure, often about 25–26m (75–78ft) across resembles a fortlet. Even these were not the latest structures on this section of the Danube for further towers appear to have been erected around the year 400. Valentinian also erected a fort across the Danube in the territory of the Quadi, whom he treated as subject to Rome; this excited resentment (Ammianus 29, 6, 2). This has been identified with an unfinished fort at GödBócsaújtelep across the Danube from Szentendre.11 In addition, a tower has been found 60km (38 miles) beyond the Danube within the land of the Sarmatians. There was also action behind the frontier with the defences of towns strengthened. The Notitia Dignitatum provides a record of the frontier units about 400, though in many cases these had changed little during the last century. Some units had names reflecting the early empire while others were probably creations of Diocletian and Constantine. Now units were equally spread along the frontier, with a good balance between infantry and cavalry units and fleet detachments located at many forts. A reference to the Danube fleets also occurs in an almost contemporary law of the emperors Theodosius II and Honorius. This required the provision of new patrol boats and the repair of old vessels on the Lower Danube, together with measures for the continuing maintenance of the boats (Codex Theod. 7, 17, 1).

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 113

River Frontiers 113

Fig. 34: Map of the northern section of the eastern frontier in Cappadocia and Syria.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 114

114 The Frontiers of Imperial Rome

The River Euphrates To turn from the Occident to the Orient is to enter a new world. Here, Rome took over long-established kingdoms and inherited their frontiers. Equally significantly, she faced a dangerous foe, the only other major power which bordered the empire, Parthia. Much of Rome’s relationship with her neighbour was governed by peace treaties and a wary understanding of each other’s power. The route to Parthia lay across and down the River Euphrates, and for most of the first and second centuries this river formed the boundary between the two states (pl. 1a, fig. 34). The point where the two empires were conjoined was comparatively narrow. In the time of Augustus and for the next 200 years, Rome and Parthia met for a distance of only about 260km (155 miles) along the River Euphrates from Armenia to the north to the Syrian Desert to the south. In the desert there was little room for manoeuvre so the frontier remained static, but to the north, there was a continuous tussle to control the strategically located kingdom of Armenia lying to the north of the Parthian empire and east of the Roman province of Cappadocia. Following the conquest of the Seleucid kingdom based on Syria and later Egypt, legions were maintained in the new provinces. Troops stationed in Judaea were concerned with internal security rather than with any external threat, as in Egypt. Syria was a different matter. It was the bulwark against Parthia. Augustus based four legions, supported by auxiliary units, within the province, including at the capital Antioch, athwart the main route into – and out of – Parthia. The area between Roman Syria and the Black Sea was left to be controlled by the resident client kings who undertook the work of defence on behalf of the empire. They defended not only their section of the frontier but, on demand, supplied troops to operate with the Roman legions, as attested from Augustus to Vespasian. These client kingdoms, however, were gradually incorporated into the empire as their kings died or Roman emperors decided. By the time of Trajan the process was complete. The legions were all now on the River Euphrates, and spread more evenly than in previous years along the frontier, yet still controlling the access routes to Parthia. Two legions lay in Syria, at Zeugma/Belkis and Samosata/Samsat facing Parthia and two in Cappadocia to the north, at Melitene/Malatya and Satala/Kelkit opposite Armenia. The southern three guarded routes across the river, routes which allowed passage not only for caravans but also for armies, while the fourth lay at the southern end of the Zigana Pass leading to the Black Sea (fig. 43). While the Euphrates might have formed the frontier, it cut through the mountains of Cilicia and in places the resulting gorge was 3,000m (10,000ft) deep, a formidable obstacle in its own right. There is little more that can be said about this frontier. At least fifteen auxiliary regiments are known to have been based in Cappadocia in the middle of the second century and about twice that number in Syria, but the bases of these units are generally not known. This frontier came under significant threat from the Parthians in the second century, as will be discussed below. The Upper Euphrates from Melitene/Malatya to Satala/Kelkit and thence northwards to the Black Sea remained the frontier, but

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 115

River Frontiers 115 further south Rome pushed forward into the desert, and this will also be considered below (pages 118–20). The Euphrates remained a significant line of communication and on its southern bank sat the great oasis city, and Roman garrison, of DuraEuropos (pl. 25). Excavations between World Wars I and II led to the identification of the military quarter in the city, the only eastern site where this arrangement has been recognized and planned.

Discussion The histories of each of the river frontiers were different. From the time of Augustus, possibly before, legions were based on the Rhine waiting to move forward. The units in the Danube provinces tended to be deployed internally, but had moved up to the river by the late first century. On the Euphrates the situation was different again with the legions lying astride potential invasion routes and therefore in essentially defensive positions. Gradually units were spread out along the river frontiers and, as the decades passed, the number of such units increased. By the late second century, every frontier province in Europe from the North Sea to the Black Sea contained at least one legion, in addition to many auxiliary units. The legions were generally placed strategically, to control routes used by the army, river crossings or potential invasion routes. The auxiliary units were spread along the rivers. In some areas, such as along the long stretch of the Danube through Lower Pannonia facing the Great Hungarian Plain, the forts were more or less equally placed, about a day’s march apart, that is 22km (14 miles), elsewhere their locations related closely to the local terrain. The control of routes remained important for the disposition of the auxiliary units. It can be no coincidence that the cavalry units based in Lower Germany lay to each side of the legion which itself was strategically placed at the start of one of the major routes into Germany, or that one of the two cavalry units in Upper Germany lay on another route into Germany. The same held for the frontier on the Euphrates: each main line of movement over the border was controlled by a legion. The act of distributing forts along the rivers was a long process. It can be most closely observed in Europe. Here, it started after the Varus disaster of 9, received an impetus as the two-legion fortresses were broken up and as a result of the proactive, but restricted, expansionist policy of Claudius which had a knock-on effect on other frontiers. The wars of 68–70 led to further movement, though it is difficult to know whether the transfer of the units based in the interior of provinces up to the Rhine and the Danube was linked to those events. Building in stone, which mainly started under Trajan, helped to solidify the current positions. Few fortlets and towers have been noted on the river frontiers in Europe, in spite of the appearance of towers on Trajan’s Column apparently placed on the bank of the Danube and the inscriptions of Commodus in Lower Pannonia. Fortlets appear on the Middle Danube under Tiberius and towers on the Lower Rhine under Claudius but generally they are found in isolated groups. While it is possible that they were only placed in locations where they were required rather than in a more

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 116

116 The Frontiers of Imperial Rome patterned formation as on the artificial frontiers, it is more likely that others await discovery, or have been destroyed by the movement of the river. The finding of new minor installations along the rivers in the last thirty years suggests that more have yet to be discovered. All the river and artificial frontiers of continental Europe share a common feature; with very few exceptions all forts lay on, or close to, the frontier line itself, that is the river bank or the linear barrier. In Upper Germany, two legions remained behind the frontier on the river, perhaps because they could be better supplied by water. On the two great European rivers were based the imperial fleets. While their primary purpose was probably defense – the evidence is sparse – like the soldiers of the army, they presumably helped supply the frontier forces. A greater understanding of the activities of the fleets would aid our appreciation of the operation of the frontiers. The river frontiers of the empire also all shared one fact in common: they were rarely bridged. When Roman armies crossed the Rhine, Danube or Euphrates, in any century, bridges had to be erected. The implication that the Romans regarded the rivers as frontiers and that they served a useful defensive role may still receive some support from the anecdote cited above about the different attitudes of Trajan and Hadrian to the Bridge of Apollodorus erected over the Danube at Drobeta/Turnu Severin under Trajan. Similarly, the Euphrates does not appear to have been bridged. When Corbulo crossed the river in 62 to advance against a threatened Parthian invasion, he erected a pontoon bridge at Zeugma/Belkis, as did Avidius Cassius a hundred years later. Constantine had stone bridges constructed over both the Rhine and the Danube. In each case, the bridge crossed the river at a legionary fortress and was guarded by a fort on the far bank. Ammianus, however, mentions bridges at Zeugma/Belcis and Capersana being broken down as defensive measures in the 360s. In 357, Julian erected a bridge over the river at Mogontiacum/Mainz to allow him to invade Germany. The following year, his cousin, Constantius II, crossed the Danube on a bridge of boats to attack the Sarmatians. Another innovation of the late third or early fourth century was the construction of fortified landing places or fortlets on the river banks. There is little good dating evidence for these and dates ranging from Diocletian to Valentinian have been proposed. There are usually two, one opposite the other on the Rhine or the Danube, and all are similar in form, a small defended enclosure with a gate surmounted by a tower. The approach in the later empire appears to be more flexible than before. Forts were of irregular size and shape and we know little about their internal arrangements. Few can be dated accurately and thereby fitted into a chronological framework, though many attempts have been made to relate them to the known building activities of emperors such as Diocletian, Constantine, Julian and Valentinian. The permanent bridges, bridge-head forts and fortified landing places indicate a new desire to relate more closely to the enemy beyond the empire. It has been suggested that ‘Constantine’s bridges reflect the strategic thinking behind his

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 117

River Frontiers 117 expansion of the mobile field army, which shifted the emphasis from reliance on static defence at the frontier line to dynamic reaction and punitive pursuit’.12 Finally, it is worth emphasizing that rivers throughout the life of the empire were seen as useful, very obvious markers. They were also exceedingly helpful towards the efficient functioning of the army, providing a ready-made line of communication between the forts along the river and a route for the efficient distribution of supplies. This was recognized by the Romans as we can see in the effort put into improving movement along the Danube and the careful placing, and retention, of forts on rivers. Indeed, rivers were so useful in all these ways that their value as the empire’s boundary created a hindrance to further forward movement; in short, they emphasized the forces of inertia, though, of course, they did not impede Rome’s more energetic emperors.

Plates_Layout 1 15/07/2011 11:02 Page 1

(a)

(b)

(d)

(c)

1a–e. Emperors who were particularly involved in frontiers, as depicted on their coinage: a) Hadrian, b) Antoninus Pius, c) Septimius Severus, d) Constantine I and e) Valentinian I. (e)

Plates_Layout 1 15/07/2011 11:02 Page 2

(a)

(b)

Plates_Layout 1 15/07/2011 11:02 Page 3

(c)

(d) 2a–d. The landscapes of frontiers: a) the River Euphrates in Turkey; b) the River Danube in Austria; c) the Carpathian Mountains in Romania; d) the Atlas Mountains in Morocco.

Plates_Layout 1 15/07/2011 11:02 Page 4

3. (left) A Roman fort depicted on Trajan’s Column with a timber tower. 4. (below) Legionaries building on Trajan’s Column. The soldiers are working in armour with their weapons, shields and helmets to hand. A row of turres lies in the foreground, while a soldier is relieved of a turf block from his shoulders where it was held on by a strap. The earth is carried from the ditch in baskets. 5. (bottom) A supply ship on Trajan’s Column. It appears to be loaded with tents.

Plates_Layout 1 15/07/2011 11:02 Page 5

6. (right) A timber tower shown on Trajan’s Column. 7. (below) Auxiliary soldiers. To the right is an infantry man, while a cavalryman stands to the left; his ornate helmet and longer sword are typical. 8. (bottom) A writing-tablet from Vindolanda. Towards the bottom left frumen [ti] (corn) can be read.

Plates_Layout 1 15/07/2011 11:02 Page 6

9. Posts in the German palisade, uncovered during excavations in 1894.

10. The posts of a timber tower at Utrecht excavated in the 2000s.

Plates_Layout 1 15/07/2011 11:02 Page 7

11. As seen from the air, the German limes still survives as a strong feature in the landscape.

Plates_Layout 1 15/07/2011 11:02 Page 8

12. A reconstructed tower and stretch of palisade on the German frontier.

13. A reconstructed Roman gate at the fort at the Saalburg (Germany).

Plates_Layout 1 15/07/2011 11:02 Page 9

14. Hadrian’s Wall at Cawfields. The wall runs along the crags to the left with milecastle 42 (Cawfields) in the foreground while the Vallum takes the lower ground to the right.

Plates_Layout 1 15/07/2011 11:02 Page 10

15. Hadrian’s Wall at Cuddy’s Crag, looking east.

16. Pits on Hadrian’s Wall at Byker, looking east. The wall base lies to the right and the ditch to the left.

Plates_Layout 1 15/07/2011 11:02 Page 11

17. The Antonine Wall at Rough Castle (Scotland), looking east.

18. The Hutcheson Hill distance slab (Scotland). The goddess Victory places a laurel wreath in the beak of an eagle, watched by two bound, captive and kneeling defeated enemies.

Plates_Layout 1 15/07/2011 11:02 Page 12

19. The fort gate at Traismauer (Austria). The gate dates to the fourth century, though the roofs, windows and mortar rendering are modern.

20. Eining (Germany) from the air. The headquarters building lies in the centre of the fort and the fourth-century fortlet in the top left corner.

Plates_Layout 1 15/07/2011 11:02 Page 13

21. The legionary fortress at Satala (Turkey).

22. The fort at Qasr Bshir (Jordan). The inscription over the entrance records construction of the fort between 293 and 305.

Plates_Layout 1 15/07/2011 11:02 Page 14

23. The cistern at Qasr Bshir (Jordan).

24. The fort at Hân al-Manqoûra (Syria) from the air.

Plates_Layout 1 15/07/2011 11:02 Page 15

25. The city at Dura-Europos (Syria) from the air.

26. The headquarters building at Lambasesis (Algeria).

Plates_Layout 1 15/07/2011 11:02 Page 16

27. The south gate at Bu Ngem (Libya). Drawn by G.F. Lyon in 1819.

28. Al-Heita, a fort in the Eastern Desert of Egypt.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 118

Chapter 9

Desert Frontiers

he frontiers in the desert areas were entirely different from land or river frontiers. Here water was also important, though not in the same way. Rainfall governed the extent of farming and settlement and therefore the boundary of the empire and the location of Roman forts. The location of the forts in Syria and Arabia closely related to the line of the 200mm Isohyet. When forts were built in the desert, their location was determined by the position of oases or the presence of sub-surface water which could be reached by wells. The resulting pattern is rather different from that on a land or river frontier. The placing of the fort at an oasis not only had the advantage of ensuring that there was a water supply for the troops but also enabled the soldiers to supervise the civilians living there or using the oasis while travelling as well as denying its use by an enemy.

T

The Eastern Frontier Syria As Rome sought to move eastwards from her frontier on the River Euphrates, control of the oases in the desert became more important. The second century was punctuated by wars between Rome and Parthia. They occurred under Trajan, Marcus Aurelius and Lucius Verus, and Septimius Severus, while there were ‘scares’ at other times, under Hadrian and Antoninus Pius. The net result was to extend Roman bases and Roman influence eastwards. Although Hadrian abandoned Trajan’s newly acquired provinces in 117, the balance of power on the frontier had tilted in Rome’s favour, as his nominee remained as king of Armenia and the kingdom of Osrhoene became a client state of Rome. Parthia remained unhappy at this turn of events and made threatening gestures in the 120s, serious enough to bring Hadrian hurrying eastwards. But in 162 she invaded the empire, destroying a legion together with the governor of Cappadocia. Rome’s reaction was powerful and, when the dust settled, her nominees remained in Armenia and Osrhoene. More significantly, there were now Roman forces based even further east, including at Dura-Europos on the Euphrates. The situation remained precarious. The king of Armenia was only retained on his throne through Roman support in the form of a detachment drawn from the Cappadocian legions based at his capital of Kainepolis. Additional infrastructure in the form of the construction of a road between the legionary base at Satala/Kelkit and the capital helped secure the Roman presence.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 119

Desert Frontiers 119 As so often happened, a Roman civil war led to disturbances on the frontier, and this was the case following the murder of Commodus in 192. The client states on the eastern border took the opportunity to seek independence. The new emperor, Septimius Severus, however, was energetic and took active steps to restore order. He created a new province of Osrhoene east of the Euphrates, but left the former ruler at Edessa. Severus next attacked Parthia, sacking Ctesiphon and trying twice – each time failing – to capture Hatra. Following the war, he pushed the frontier forward from the Euphrates almost to the River Tigris, creating a new province of Mesopotamia east of Osrhoene. Here, he placed his newly raised legions at oasis cities between the rivers, perhaps originally at Nisibis and at Singara, the most easterly point of the new province (fig. 34). These became, in effect, heavily defended fortresses, designed to withstand attack and siege. Auxiliary units were brought in to defend the province. Again, the creation of the new network of military sites was accompanied by road building. This new military deployment did not follow a line on a map, but acknowledged the geography of the desert, the army being based in the oasis cities. In such a situation, the holding of each base – each city – was important. Each was not only a stepping stone forward, but an important link in the lines of supply. As emperors such as Trajan and Severus sought to push ever eastwards they created more attenuated lines of communication, as well as stretching the military resources of the empire. As Dio noted: Severus used to declare that he had added a vast territory to the empire and had made it a bulwark of Syria. On the contrary, it is shown by the facts themselves that this conquest has been a source of constant wars and great expense to us. For it yields very little and uses up vast sums; and now that we have extended our frontiers to the neighbours of the Medes and Parthians, and we are constantly so to speak at war in their defence. (Dio 75, 3. Translation by E. Cary.) In such circumstances, it is not surprising that Hadrian had abandoned Trajan’s eastern conquests. Yet, what was the reason for Severus’ conquests? John Mann argued that it was to strengthen Roman control of the main caravan route between Rome and Parthia rather than to act as a military bulwark.1 We might also ask whether there was a logical reason behind the actions bearing in mind Severus’ professed love of fighting (Herodian 3, 14, 2). The eastern advance of the frontier by Severus led to changes in Syria, now far to the rear. The province itself was divided into two and legionary detachments were moved from the bases at Zeugma/Belcis and Samosata/Samsat to sites further east. The Roman army did not sit in its major eastern bases, the cities, and wait for attack but employed outposts across the desert. The Twentieth Cohort of Palmyrenes at Dura-Europos provided soldiers for at least eight outposts. The number of soldiers at each ranged from three or four to ninety-three, though the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 120

120 The Frontiers of Imperial Rome numbers are a minimum in view of the fragmentary state of the rosters of 219 and 222 (P. Dura 100 and 101). We do not know the location of most of the outposts but one lay 150km (nearly 100 miles) downstream from Dura-Europos. The establishment of a new dynasty, the Sassanians, in Persia in 224 presented Rome with an energetic enemy on their Eastern Frontier. Under their dynamic king, Shapur I, they defeated several Roman armies and in 253 penetrated as far into the empire as Antioch, which they sacked. The Persian records state that thirty-six castles and cities were taken in these campaigns and 60,000 soldiers killed. These wars of attrition continued throughout the third century. Dura-Europos fell in 256, and was never reoccupied. The Emperor Valerian was captured in 260. This, however, was the high point for the Persians and Rome thereafter was able to maintain the status quo. Indeed, she did more than that, for a Roman force sacked Seleucia and Ctesiphon in 283 and Roman dominance in Armenia was reasserted. Diocletian was as active here as elsewhere. He formed new legions and established them in the eastern provinces from Trapezus/Trabzon southwards, often grouping them in pairs. The cavalry component of the frontier armies was also increased. A new road was built, the strata Diocletiana, from Azrak in Arabia, where it connected with Trajan’s new road, to Palmyra, where a legion was established in a fort built beside the city wall, and thence on northwards to Soura/Souriya on the Euphrates. The route from Damascus to Palmyra was dictated by the lie of the mountains, which ran in a south-west to north-east direction and the road was constructed to the south-east of this line. All along the road were fortlets constructed in the new style. They lay at intervals of about 20–25km (12–15 miles), the exact location of each being determined by the availability of water. A note of caution must be sounded, however, for the vast majority of sites along the road are undated. The most significant changes on the frontier occurred as a result of the death of Julian in 363 while campaigning against Persia and the humiliating peace signed by his successor, Jovian. Eastern Mesopotamia was surrendered, but the frontier was still far in advance of the pre-Severan line, with the border on the River Tigris. The late Roman military arrangements on the Eastern Frontier are recorded in the Notitia Dignitatum. In the face of the Sassanian threat, there was a return to earlier principles. A large mobile force was stationed at Antioch, where there had been a significant legionary base under Augustus, and serving the same purpose. ‘The old frontier forces have now been reduced to the functions performed under Augustus by the client-kings, the detailed containment of small-scale hostilities, and the control of traffic across the frontier line.’2

Arabia South of Syria lay Arabia, mostly now lying within modern Jordan. Most of this area was divided between various client kings in the first century CE, many well-known to us through the Bible. These kingdoms were absorbed into the empire in the first century, Arabia being the last when it was taken over by Trajan in 106. A legion was eventually established at Bostra/Busra in the north of the province (in southern

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 121

Fig. 35: Map of the southern sector of the eastern frontier in Syria and Arabia.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 122

122 The Frontiers of Imperial Rome modern-day Syria), well-placed to be supplied from a fertile hinterland, but also able to support the legions in Syria to the north in case of an invasion from Parthia. This was the Third Cyrenaica and it was still at Bostra/Busra, according to the Notitia Dignitatum, 300 years later. A base of a size to hold the complete legion was erected, but there are many inscriptions of soldiers of the legion throughout the province and it is possible that detachments were regularly away from base on outpost duty. Building work is recorded in the year following annexation in a letter written on 26 March by a legionary, Julius Apollinaris, to his father in Alexandria (P. Mich. 466). He states that his colleagues are cutting building stones. This may have been for a new road, Trajan’s new road, the via nova Traiana, constructed from Bostra/ Busra past Petra to Aila/Elat at the head of the Gulf of Aqaba, the eastern branch of the Red Sea. Milestones record that work on the road had begun by 111. This road was an important line of communication from north to south along the Eastern Frontier following an old caravan route. It lay towards the outer edge of the area where there was sufficient rainfall to allow the maintenance of Roman forts but still within the settled area: when forts came to be erected along this section of the frontier under Diocletian, they generally lay east of the road. Ten to twelve auxiliary units, as well as the legion, are known to have been stationed in the province, but their locations are mostly unknown. A legionary detachment was posted to Petra, at least in the early years following annexation. An early-second-century fort lay at Humayma, halfway between Petra and Aila/Elat, also the base of a detachment of the Legion Three Cyrenaica, and there was therefore probably a fort at Aila/Elat. It is likely that these forts lay at points on the road previously guarded by the Nabataeans.To the north of the province, an inscription of Marcus Aurelius indicates the existence of a fort at Umm el-Djemal (CIL III 141492). There are hints at other fort locations. A dedication of the late second or early third century at the oasis of Dumata/Jauf, in modern Saudi Arabia, 370km (230 miles) south-east of Azraq along the Wadi Sirhan, by a centurion of the Legion Three Cyrenaica suggests a local military presence, if only an outpost. This in turn suggests a second-century fort at Azraq, though the earliest attestation is under Severus. Azraq lay east of Trajan’s new road, but a link connected it to the forts to the north. Several milestones of the last years of the reign of Severus attest work on this road. Also along the road are towers to aid communication, though none are dated. They are 6–7m (20–23ft) square. Communication to the west was aided by a fort at Qasr el-Hallabat erected, or rebuilt, under Caracalla.3 Another record of Roman activity beyond the frontier in the late second century is an inscription at Rawaffa, also in Saudi Arabia, 225km (135 miles) south-east of Aila/Elat. This records the construction of a temple to the emperors in about 166 by the Thamudic confederacy and mentions the intervention of the provincial governor in a local dispute. This area had previously been part of the Nabataean kingdom and was incorporated into the province of Arabia. Even further to the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 123

Desert Frontiers 123

Fig. 36: Plan (partially reconstructed) of el-Lejjun.

south-east, at Medain Saleh, another 225km (135 miles) beyond Rawaffa, are two Roman inscriptions, one an altar dedicated by a soldier of the legion at Bostra/Busra and the other graffiti scratched on a rock by Roman soldiers. These soldiers presumably were helping to protect the caravan route through the Hejaz and it is not surprising to find mention of soldiers of a regiment of camel riders at one site.4 Inscriptions and excavations demonstrate considerable building activity under Diocletian. Substantial new forts were built at Udruh and el-Lejjun, both probably with legionary troops in occupation (fig. 36). The fort at el-Lejjun covers 4.7ha (nearly 12 acres). It has thick walls with external bastions and small gates, a typical late-empire fortification. Excavations by Thomas Parker have demonstrated that it was constructed about 300 and revealed considerable information about the interior arrangements, including the headquarters building and many barrack-blocks.5 Major rebuilding occurred after an earthquake of 363. The fort continued in occupation for nearly another 200 years. Udruh is of a similar size and plan. These legionary bases were supplemented by forts and fortlets. Their spacing along Trajan’s new road was about 20km (12 miles), that is about a day’s march, apart. The new forts were not usually erected immediately beside the road, but on conveniently elevated ground in the vicinity and close to a source of water; each was normally accompanied by a great cistern for the collection of water (pls. 22 and 23, fig. 37). Forts and fortlets in the south include Qasr Bshir and Da’janiya. At Qasr Bshir the corner towers still stand three storeys high with the walls in between two storeys, and the building inscription of 293–305 over the gate (CIL III 14149). Inside, the rooms on the lower floor were for animals, as the mangers in the walls demonstrate, and the upper rooms for the soldiers. The walls of Da’janiya were built with black basalt blocks which today are framed by white mortar creating an incredible chequerboard pattern. Both are remarkable survivals.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 124

124 The Frontiers of Imperial Rome Fig. 37: Plan of Qasr Bshir.

To the north, a denser network of forts and towers protected access to the cities and settled land of the western part of the province. Deir el-Kahf was erected in 306, with further work in 348/9 and 367–75, and Qasr el-Hallabat presumably rebuilt or enlarged about the same period. There was also work at Azraq at this time, but it is difficult to determine the size of the Roman fort owing to its continuing use as a fortification – up to World War I when T. E. Lawrence used it as his base. An inscription erected under Diocletian at Azraq attests work by legionaries on the road built or repaired by Severus a hundred years before.6 The forts have similar plans. They have high walls, a single gate, square towers at each corner and the buildings placed against the inside of the surrounding wall. Many are small covering about 0.3ha (0.74 acres) or less. Da’janiya occupies 1ha (2.5 acres) and is better described as a fort; its courtyard is occupied by buildings. Several of these sites appear to have continued in use into the sixth century. The considerable building programme under Diocletian may have been a response to the rise of the Saracens. They are first recorded by Ptolemy of Alexandria about 140. In the 290s, Diocletian campaigned against them. Ammianus Marcellinus described their activities: The Saracens, however, whom we have never found desirable either as friends or as enemies, ranging up and down the country, in a brief space of time laid

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 125

Desert Frontiers 125 waste whatever they could find, like rapacious kites which, whenever they have caught sight of any prey from on high, seize it with swift swoop, and directly they have seized it make off. (Ammianus 14, 4, 1. Translation J. C. Rolfe.) Ammianus in describing the events of 363 on the Eastern Frontier mentions one by name, Podosaces, ‘a notorious brigand, who with every kind of cruelty had long raided our territories’, and he wryly acknowledged when Saracen princes offered their support that they were ‘gladly received since they were adapted for guerrilla warfare’ (Ammianus 24, 2, 4; 23, 3, 8). Later in the fourth century, Valentinian, who ordered the construction of so many towers on the Rhine and Danube frontiers, was also active here. An inscription records the construction of a tower at Umm el-Djemal in 371 by the Ninth Cavalry Regiment of Dalmatians (ILS 773). Fortified enclosures behind the Roman frontier in modern Israel have also been linked to this type of threat. Mordechai Gichon coined the term ‘courtyard castellum’ or ‘courtyard fort’ to describe this small square enclosure, and has emphasized its long history in the area, being used by both farmers and soldiers.7 He has plotted the distribution of such enclosures, demonstrating how many guarded passes and oases, and has argued that they protected the fertile lands of Palestine from raiders living in the Negev and Sinai. This line from Gaza to the southern end of the Dead Sea and beyond certainly becomes important in the fourth century when several cavalry units were stationed along it. Towards its eastern end, to the south-west of the Dead Sea, a military context has been argued for two late small ‘forts’ at Upper Zohar and En Boqeq, though this has been challenged.8 It may be concluded that it is not clear that all ‘courtyard forts’, and the towers which also appear in this area, were occupied by soldiers, rather than that they were the defensive dwellings of farmers. Some similar enclosures in Arabia are regarded as being farms. There has been considerable discussion about the role of the forts along Trajan’s new road. The lack of evidence for any military posts along this line in the second and third centuries has led to the suggestion that those built there in the fourth century were erected in response to a particular situation. This, it has been proposed, was the raiding which appears to have increased in the late third century. But are the forts protecting traffic on the road or the people in the province behind? Trajan’s new road was built on an old caravan route and the fourth-century forts along its line were about a day’s march apart, the normal distance between forts on a frontier, though, of course, shorter distances do occur. On the other hand, the road lies approximately on the boundary between the desert and the sown, the cultivated land which the Roman authorities would seek to protect from raiding. The existence of towers does not affect the argument either way because such look-out towers would be required for either function.9 The crucial point is surely the reason for the construction of these forts and that is likely to have been the increased raiding. Traffic on the road would certainly have

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 126

126 The Frontiers of Imperial Rome needed protecting and the forts would have provided a convenient protected location for caravans to camp each night. Yet, the soldiers based in these forts would also have been able to protect the farmers of the interior and, like so many issues, there may have been interlinked reasons for the building of these forts along the route between Aila/Elat and Bostra/Busra and on to Palmyra and the Euphrates.

North Africa The provinces of North Africa tended to be restricted to the narrow coast area bordering the Mediterranean or, in the case of Egypt, to the equally narrow valley of the River Nile. Beyond all the provinces of North Africa except Egypt lay the Sahara Desert. This was, however, not uniformally desert, nor was it deserted. In Cyrenaica and Tripolitania to the east (both in modern Libya), the Sahara approached close to the coastal cities, and these areas were therefore largely devoid of troops, but further west the several ranges of the Atlas Mountains lay between the empire and the desert. This area will therefore be considered in the section on mountain frontiers. Tripolitania Under Augustus and his immediate successors, the Garamantes of the modern Fezzan were a significant threat to the stability of the coastal cities. Cornelius Balbus campaigned against them in 20 BCE, for which he was awarded a triumph, and there were subsequent expeditions. During the civil war of 68–9, the city of Oea obtained the help of the Garamantes against their neighbours in Leptis, but the subsequent Roman retaliation appears to have cowed them into submission. Nothing more is heard of these people, described by Tacitus as much given to raiding (Tacitus, Histories 4, 50). In the second century some forts and fortlets were established in western Tripolitania around the edge of the Great Eastern Erg. It is possible that Remada/ Tillibari is as early as Hadrian though the fortlets at Tisavar/Ksar Rhilane and at Bezereos/Bir Rhezene date to the time of Commodus at the earliest. Major developments, however, took place under Severus (pl. 2, fig. 38). He established new forts at oases in the northern Sahara Desert in the early third century. These were at Gholaia/Bu Ngem, Gheriat el-Garbia and Cidamus/Ghadames (pl. 27). The first two lay on the southern edge of the area of cereal cultivation in favoured wadis and therefore at the limit of ancient settlement. Cidamus/Ghadames, however, was placed at an important oasis 300km (200 miles) into the desert. All were manned by detachments of the Third Augustan Legion.10 It is not clear why Severus had these forts built so far south. Perhaps there was perceived to be a need to protect the expanding agricultural lands of the coastal zone, and in that respect his biographer records that Severus secured the safety of Tripolitania by defeating war-like tribes (Historia Augusta, Life of Severus 18, 3). Goodchild, however, suggested a different reason for the placing of the fort at Gholaia/Bu Ngem. He drew attention to the size of the fort, its monumental gateways and the well-dressed stone-work and suggested that it may have been

Fig. 38: Map of Tripolitania.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 127

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 128

128 The Frontiers of Imperial Rome constructed for propaganda reasons: ‘it was perhaps intended that the traveller on the caravan routes should become immediately aware that he had entered Romancontrolled territory, and that his liberty of action was now under definite restrictions.’11 Gholaia/Bu Ngem continued in occupation until at least the 250s, and during its fifty years’ existence appears to have provided soldiers for at least four outposts. In the third century a new style of small fort appeared in Tripolitania and in Cyrenaica to the east. It is not always clear whether they were occupied by soldiers or civilians; whoever lived there would have required the same protection from the heat and dust as much from any threat of attack. One such site, Qasr Duib, is dated. It was 16m (53ft) square, built in 244–6 on the Tripolitanian frontier, erected on the orders of the governor to ‘close a route used by raiders’ and described on its building inscription as a centenarium (IRT 880). This fortlet lay on the limes [Ten]theiranus which had its own officer, the praepositus limitis. These new fortifications, which continued to the end of the century, may, in part at least, be a reaction to the raiding and invasions which are known to have afflicted the African frontier through the middle years of the third century.12 The Notitia Dignitatum (Oc. XXXI) reveals that regional officers still existed, grouped into the command of the duke of Tripolitania. In Tripolitania there are also short lengths of mortared walls and banks together with ditches termed clausurae. Some walls survive up to 3m (10ft) high; one short length is twice that height. These barriers are shorter than those further west in Numidia though, in conjunction with the hills and escarpments, they created a longer, continuous barrier. In some places gates and towers have been noted. The gates consisted of a passage 2.5m (8ft) wide between two towers. Dating evidence is sparse, but finds indicate a second century origin with use continuing into the third century and possibly the fourth. These barriers were often placed on the outer edge of agricultural land, and it has been suggested that their purpose, like that of the Fossatum Africae, was ‘to regulate transhumance, seasonal labour and trading movements between the predominantly pastoral and the predominantly sedentary zones’.13 Raiding by the Laguatan in the late third century brought Maximian to the region, and his actions appear to have been successful for the area seems to have remained peaceful into the late fourth century. In the 360s, the Austurians raided the coastal cities of Tripolitania, achieving more success than they ought owing to the corrupt Roman administration (Ammianus 28, 6). The size of the army grew in the fourth century, achieving a strength of nineteen cavalry units and twelve infantry regiments according to the Notitia Dignitatum (Oc. VII, 141–52; 180–98). The high numbers of cavalry probably related to the nature of the threat. Little is known of the military bases of this time and it is possible that most troops were based in the cities though some fortlets have been recorded.14

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 129

Desert Frontiers 129

Cyrenaica Only a single auxiliary mixed cohort is known in Cyrenaica during the early empire, no doubt reflecting the proximity of the desert and the lack of threat from that area. The gravest threat to peace came from unrest in the urban centres and the army of the province proved incapable of preventing the Jewish revolt of 115. At the very end of the fourth century the Austurians extended their raiding from Tripolitania into Cyrenaica. In earlier years, however, it would appear that fortlets and towers had been placed across the approaches from the south-west, the region where the Austurians were later to live. It is not surprising that the same sites came to be occupied by Italian forts in the twentieth century. Other fortlets were built in the fourth century, but, as elsewhere, it is difficult to distinguish between military posts and civilian farms, while some sites are Byzantine rather than Roman in date. Egypt Egypt, at the eastern end of North Africa was something of a special case. Legions were retained here to maintain law and order, and particularly to seek to ensure peace between the Egyptians and the Jews in the more populated areas. By the time of Hadrian, only one legion remained, in Alexandria, with the other troops spread through the province, mainly along the River Nile. Egypt, however, did have a frontier for to the south lay Nubia. Under Augustus, this kingdom originally had friendly relations with Rome, but in 25 BCE it invaded the empire, sacking forts in the frontier region. The Roman governor responded, marching over 600km (370 miles) south to capture and destroy one of the Nubian capitals and establish a Roman garrison 200km (120 miles) south of the Roman frontier at Primis/Qasr Ibrim beside the river. Subsequently, Augustus annexed most of this part of the Nubian kingdom, abandoning any pretensions further south. The southern boundary of Egypt was established at Syene/Aswan, with outposts stretching south along the river for 110km (66 miles) to Hierasykaminos/ Maharraqa. One such outpost, the fort at Pselkis/Dakka, held a force of legionaries and auxiliaries guarding the end of a route across the Eastern Desert. The Roman army therefore acted exactly the same as on other frontiers. And the arrangements were good for the frontier remained largely peaceful for 300 years, allowing for the agricultural exploitation of the land between the first and second cataracts. Following serious raids and invasions in the third century, Diocletian allowed the Nobades to occupy the land south of Syene/Aswan on condition that they helped protect the frontier. There was also a concentration of small forts and fortlets along the roads through the Eastern Desert between the Nile and the Red Sea (pl. 28). Over seventy are known, though they may not all have been occupied at the same time. They are mostly small in size, many only 30m (100ft) square, and only three exceeding half a hectare (less than 1 acre) in size, with commensurately small garrisons. Each had its own cistern to collect water: here there are no oases. This is not an area conducive

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 130

Fig. 39: Map of the Nile Valley and the Eastern Desert of Egypt. The small forts and fortlets are indicated by small squares.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 131

Desert Frontiers 131 to settlement and, as little has happened in the area over the last 1,500 years, many of the small forts and fortlets can still be seen.15 The major routes radiated out from Kainepolis/Qena and Coptos/Qift on the Nile. A cavalry unit was based at Coptos/Qift presumably to provide the isolated desert outposts with speedy support in the case of an emergency. Auxiliary units were placed at the other points where roads across the desert reached the Nile.16 The northern route from Kainopolis/Qena soon branched into two roads. One led to the imperial quarries at Mons Porphyrites, which produced the purple stone beloved of emperors, and was extended under Diocletian to Abu Sha’ar on the western arm of the Red Sea. There Diocletian had a fort erected. The other branch took a different route to the same place but passed the quarries at Mons Claudianus. Outside the military posts on these routes lay animal lines and watering troughs, both presumably for the draught animals pulling the heavy stones from the quarries. There is also evidence that this road was paved. The southern road from Coptos/Qift also split into two, one branch heading east to Myos Hormos/Quseir where there was a port with a military presence. Several of the posts along this road conformed to the same plan, suggesting contemporaneous construction. They are square, with projecting external towers and a courtyard containing a well. The route here was not paved, which may imply its main use by caravans of camels. The other road led south-east, crossing the desert for 420km (260 miles) to Berenice, a major port of entry for luxury goods from further east, and also a military base. The supervision of the southern network of roads appears to have been under the command of the prefect based at Berenice. This pattern continued into the fourth century when the more southerly forts, including Berenike, appear to have been abandoned. The soldiers attested at these posts mainly came from units based on the adjacent stretch of the Nile. The number at each site will have been small; only 60 of the 920 people recorded at Mons Claudianus in one document were soldiers.17 The senior officer was accordingly often of a fairly low rank, not even a centurion or decurion. The question is, what was the purpose of this tight military control? Two reasons have been offered. Firstly, that they formed a frontier, almost a limes, to protect Egypt against raiding from the tribes living beyond the frontier. In the desert and along the coast lived several tribes, the Blemmyes, Trogodytes and the Nobades. There are records of frontier raids in the late first and in the second centuries, in one of which a successful Roman counter-attack led to the recovery of booty. The roads through the Eastern Desert were also important routes for traders bringing goods from the Red Sea ports of Berenice and Myos Hormos/Quseir into the empire and their caravans required protecting, as did the watering places along the routes. A secondary purpose of the military presence was more certainly to supervise work in the quarries and mines, for this region was rich in gold and emeralds as well as high-quality building stone. The quarries included the only one in the empire which produced porphyry, the purple stone favoured by emperors for their own statues and busts and their own buildings. No doubt, the quarry workers and miners

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 132

132 The Frontiers of Imperial Rome themselves needed protection from attack as well as supervision, so the forts may have served a dual purpose. Surviving military records from this area allow us to appreciate the high level of control exercised by the army on the movement of civilians and the protection they accorded caravans and travellers, and these will be considered below.

Discussion There are considerable differences between the so-called desert frontiers. There are, however, two important constants, the extent of Roman rule related to the rainfall and the area of cultivable land. On all frontiers, the distances between forts are greater than on the land and river frontiers of Europe which suggests that the lines of forts which can be drawn on a map do not relate to the same type of frontiers. A strong argument has been made that the forts constructed under Diocletian and later in Arabia were to protect travellers on the caravan route leading from the Red Sea northwards. Their construction followed the early Saracen raids and it remains possible that the forts also protected the local provincials from such raids. There is evidence for increased agricultural production in several frontier areas which could have encouraged raiding. In Arabia and in North Africa, outposts were established at considerable distances beyond the presumed frontier line, or rather beyond the main line or group of forts. In North Africa, these almost appear to be part of a continuous forward movement of the frontier in Numidia through the second century into the early third. This move, however, stopped in the early third century after the actions of Severus. One purpose of his forts may have been to guard caravan routes into the empire, and this seems also to have been the situation in Arabia where the outposts sat astride routes into the province. The outposts there may have had an additional role of maintaining contact with people beyond the frontier. One unusual feature in both Arabia and Numidia was the use of legionaries to man forts and outposts. The reason for this is not known. It possibly reflects the relative lack of auxiliary troops in these provinces. Possibly the legionaries here were otherwise underemployed as there were few threats. Diocletian was at work on all frontiers, as demonstrated by the construction of forts in the new style of architecture. In North Africa, ironically, the last major threat came with an invasion from across the sea, by the Vandals who sacked Carthage in 439. Roman rule ended here and in the Middle East with the Arab invasions of the seventh century.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 133

Chapter 10

Mountain Frontiers

e have already seen that rivers can flow through mountainous terrain and the resulting gorges, as on the Middle Danube or the Euphrates, can act as a severe impediment to movement, and settlement. Often, the army seemed to consider that little extra protection was required in such circumstances. Usually, however, they sought to control passes on every frontier, even on the land frontiers – every pass in the hilly country of the Odenwald in Germany, for example, was controlled by a fort or fortlet. In northern Britain, a network of fortlets guarded the passes through the Southern Uplands in the Antonine period (fig. 47d).

W

Dacia The frontiers of the province of Dacia, formed from the kingdom conquered by Trajan in the opening decade of the second century, were mainly defined by mountains, the Carpathian Mountains (fig. 40, pl. 1c). These offered a boundary as well as an obstacle to attack. Their form helped create a uniquely defended province, a useful reminder that the Romans could adapt to special and different circumstances. Dacia was an unusual province. It projected beyond the Lower Danube in a great bulge, its frontier following the mountain ranges which circumscribed and identified the core of the previous Dacian kingdom. The bowl which is Transylvania is surrounded to the north-east, east and south by the Carpathian Mountains, with lower ranges to the west. Transylvania is, as a result, a natural fortress, the mountains creating an amphitheatre. To the west, however, there are two significant weak points. These are the gaps caused by rivers cutting through the mountain ranges. One lies to the north-west, between the western hills and the northern Carpathians where the River Somes and its tributaries cut through the mountains. The other, more or less in the centre of the western boundary of the province, is where the River Mureş cuts through the mountains and here a major route across the Great Hungarian Plain enters the province from Lower Pannonia. This Transylvanian heartland was separated from the empire by the Wallachian plain, 150km (90 miles) wide, which lies between the Danube and the mountains. The unusual nature of the provincial boundary and its defensive needs was acknowledged in the division of Dacia into three provinces. To the north lay Dacia Porolissensis, to the south-east Lower Dacia, while Upper Dacia straddled the centre from the Carpathians in the east to the Danube in the west. The shape of Dacia helped to create its own unique military deployment. To the north and east, the outer shell lay in the mountains. The auxiliary units there were

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 134

134 The Frontiers of Imperial Rome

Fig. 40: Map of Dacia.

mainly infantry. The main pass to the east was strongly guarded with additional units being based there. Some towers have also been recorded in this sector but no pattern has yet emerged. A similar pattern pertained in the west where several auxiliary forts, some containing 1,000–strong mixed infantry and cavalry regiments, protected the access route along the MureşValley. The cavalry component may have intended to ‘counterbalance the powerful Iazyges cavalry’.1

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 135

Mountain Frontiers 135 There was a noticeable concentration of forts in the north-west sector, in the gap between the western and northern Carpathians. There were as many as thirteen auxiliary units here, though the forts may not all have been occupied at the same time. The arrangement tended to be infantry units on the outer line, tactically placed on the inside of the passes, with cavalry units stationed further back. The major fort in this area lay at Porolissum/Moigrad. It covered 7ha (17 acres), together with a smaller fort of only 0.66ha (1.65 acres), the two together housing a force of about 2,000 soldiers. The placing of a larger and smaller installation in close

Fig. 41: Map of the towers and barrier to the north-west of the province of Dacia.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 136

136 The Frontiers of Imperial Rome proximity has been noted elsewhere in Dacia, both in the western sector and to the south along the River Olt and on the Limes Transalutanus. The usual arrangement appears to have been a fort with a smaller fort for a numerus or a fortlet nearby.2 Porolissum/Moigrad and its adjacent fort were protected by a section of bank about 4km (2.5 miles) long, constructed of both earth and stone (fig. 41). Two fortlets and several stone towers are known on this line. This was but part of a wider framework of linear barriers and towers in this north-western sector of the Dacian frontier. To the south of Porolissum/Moigrad some sixty-six towers and eight fortlets are known so far in a 75-km (45-mile) stretch, while in places passes were blocked by the construction of banks and ditches. Along the northern rim of the province, the line of towers and fortlets, but, as yet, no barriers, extended for about 180km (110 miles) east of Porolissum/Moigrad. Elsewhere in the Carpathians, individual fortlets, towers and barriers have been identified and it is not impossible that the whole province was ringed by such military installations. The purpose of the towers was to maintain observation over the access routes into the province, the role of the banks to hinder unauthorized access. In the words of Bill Hanson and Ian Haynes, ‘the dispositions … indicate a considerable concern for frontier security in the difficult mountainous terrain.’3 While the purpose of the arrangement of forts, fortlets, towers and barriers seems clear, to control access to the province, the operational details of this system are unclear. This includes the relationship of the men in the towers to those in the fortlets and forts. Presumably arrangements were in place to relay messages back to the troops in the forts behind the line of towers. In the centre of the Transylvanian amphitheatre were located one and later two legions, with good lines of communication so that they could move in any direction to support threatened sections of the frontier. The legion at Apulum/Alba Iulia was also placed close enough to the main gold mining areas to police the mines. Rather surprisingly there were no forts west of the legionary base along what might be characterized as a frontier line. Perhaps the legionary presence was considered sufficient. Besides the legion, a number of auxiliary units, mostly cavalry, were placed in the interior of the province, usually located at road junctions, presumably strategically placed to move in any direction to support any threatened troops on the frontier. It was not until the late 160s that a second legion was moved into Dacia. It was stationed at Potaissa/Turda, about halfway between Apulum/Alba Iulia and Porolissum/Moigrad. Its transfer here has been linked to actual or threatened attacks at the beginning of the Marcommanic War. The south-western sector of the frontier, the area now known as the Banat, does not appear to have been heavily protected. This was the southern part of the Great Hungarian Plain and through it ran the River Tisza on its journey to join the Danube at the south-eastern corner of Lower Pannonia (fig. 28). The river was fed by numerous streams, creating a series of marshes. Perhaps the army regarded the marshes as sufficient protection, with the units based across the Danube in Lower Pannonia and Upper Moesia able to maintain watch over the area.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 137

Mountain Frontiers 137 This arrangement was conceived under Trajan, amended under Hadrian and continued thereafter with little change until the abandonment of the province in about 260. Another aspect of the Dacian frontier is worth noting. The province remained a great salient sticking out to the north of the Danube. There was no attempt to smooth out the kinks in this stretch of the Northern Frontier. This could have been achieved by bringing Wallachia and southern Moldavia to the east into the empire and thereby running the frontier from the eastern Carpathians to the Danube bend at the start of the delta and by incorporating the Great Hungarian Plain between Lower Pannonia and Dacia to the west. A road was constructed across the latter, but otherwise relations with the Roxolani to the east and the Sarmatians to the west of Dacia were governed by treaties.

The North-east Frontier The northern section of the Eastern Frontier between the Black Sea and the Euphrates also ran through mountains, the Pontic Mountains. The north–south line of communication between the river and the sea was the Zigana Pass. The southern approaches to the pass were guarded by a legion based at Satala/Kelcit, as we have seen (fig. 43). Corbulo carefully controlled the passes through the Pontic Mountains during his campaign against Tiridates of Armenia in 58 to ensure that supplies from the Black Sea region continued to reach his army (Tacitus, Annals 13, 39). The Caucasus Mountains ran roughly north-west to south-east between the Black Sea and the Caspian Sea. Passes through this range of mountains, which rose to 5,000m (15,000ft) were few. In the centre was the Darial Pass and, by the Caspian Sea, the Derbend Pass. In 75, under Vespasian, Roman soldiers helped the king of Iberia strengthen his fortifications at Harmozica at the southern end of the Darial Pass. The reign of his son, Domitian, saw a soldier of the Twelfth Legion based at Melitene/Malatya on the Euphrates cutting an inscription on a rock beside the Caspian Sea at the southern end of the Derbend Pass (AE 1951, 263). In the third century, the Romans paid for a Parthian force to mount guard over the Darial Pass and 200 years later, in the face of the threat from the Huns, the arrangement was renewed. Control of these passes was important to both the Romans and the Parthians because it was through them that invaders came from the north. Prominent among these were the Alans. So serious was the threat posed by these that in 75 the Parthian king, Vologeses I, suggested a joint expedition to Vespasian. Two generations later, the Roman governor of Cappadocia, Arrian, took the field against the Alans, leaving an invaluable record of his actions in his book, The Expedition against the Alans. North Africa The Atlas Mountains stretch from Carthage in modern Tunisia for about 2,200 km (1,300 miles) in a west-south-westerly direction before reaching the Atlantic Ocean south of modern Marrakech in Morocco (pl. 2d). In places, they rise to 4,000m (over

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 138

138 The Frontiers of Imperial Rome 13,000ft). The northern point of the High Atlas, known as the Rif, effectively divided Mauretania Caesariensis from Mauretania Tingitana. While the coastal area was predominantly agricultural, nomadic pastoralism dominated in the intermediate zone between the coastal fringe and the Sahara Desert. It is important to note that this region had strong similarities in its political, social and administrative infrastructure with the Eastern Frontier area. Carthage, situated in modern Tunisia, had been a formidable foe to the Roman Republic and the adjacent kingdom of Numidia (eastern Algeria) had a long history continuing into the early empire. Rome had acquired its North African provinces at different times. Proconsular Africa was formed into a province following the final defeat and destruction of Carthage in 146 BCE; the province was later extended along the coast to embrace Tripolitania (modern western Libya). The cities of Cyrenaica had been bequeathed to Rome in 96 BCE by the Ptolemies of Egypt. Egypt itself was absorbed into the empire by Augustus following his defeat of Cleopatra and Mark Anthony in 31 BCE. Numidia and Mauretania became client kingdoms. The former was abolished by Caesar in 46 BCE and incorporated into Africa. Mauretania remained until Ptolemy, grandson of Cleopatra and Mark Anthony, was killed by his cousin, the Emperor Gaius, in 40. Annexation met with opposition, which was suppressed, and early in the reign of Claudius the former kingdom was divided into two provinces, Tingitana in the west and Caesariensis in the east. Beyond the coastal areas, in the Sahara, there was no significant enemy as Augustus had discovered when he sent Cornelius Balbus to explore the area shortly before 19 BCE. Balbus penetrated as far as the kingdom of the Garamantes in the Fezzan province of modern Libya. The Garamantes would occasionally cause trouble for the Romans, but they were not a major threat. About 42, the governors Suetonius Paulinus and Hosidius Geta chased invading Moors across the Atlas Mountains into the Sahara Desert, recording that the place was uninhabitable because of the heat. There is no record of a subsequent expedition into the desert. Under Augustus and Tiberius there had been serious fighting in Numidia on two occasions, but, following the defeat of the Gaetuli and Musulamii in 24, this area was generally peaceful, being only disturbed by raiding by the tribes occupying the desert fringes, raiding which continued throughout the life of the empire. This led to a small number of troops being maintained in the North African provinces, no more than 30,000, including a single legion, about half the size of the army of Britain to cover a far larger area. In 68–9, the combined army of the two Mauretanias was only five cavalry regiments and nineteen cohorts. The army in North Africa remained relatively small until Diocletian increased it considerably. It now consisted of eight legions together with the auxiliary regiments, all, though, of smaller size than before. The frontier command was divided between three senior officers, with thirty regional commanders below them. In addition, there was a field army.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 139

Mountain Frontiers 139

Numidia Proconsular Africa at first consisted of the hinterland of Carthage. Numidia lay to the west, originally mainly the region around Constantine/Cirta with the Aurès Mountains forming the southern boundary. Under Augustus, the army wintered in the province of Africa, but about the time of his death in 14, the Third Augustan Legion was established at Ammaedara/Haïdra just over the border into Numidia at the north-east corner of the Aurès Mountains. About 75, the legion was moved south-westwards thence to Theveste under Vespasian. Just a few years later, in 81, it built a fort at Lambaesis near the north-west corner of the Aurès Mountains, and the whole legion moved there, probably under Trajan. The movement of the legion was accompanied by the building of new auxiliary forts. Under Vespasian these were constructed along the northern fringes of the Aurès Mountains as far as Lambaesis. Trajan was certainly building at Ad Maiores south of the Aurès Mountains in 104–5 and possibly completed the whole line of forts on this line as far west as Thubunae in the gap between the Aurès and Hodna Mountains. This southern line of forts brought the Roman military presence to the edge of the desert (fig. 42). The construction of the separate lengths of the Fossatum Africae, probably by Hadrian, was the next step in the establishment of this network of forts, for they were located in the gap between the Aurès and Hodna Mountains (above pages 82–4). However, the separate sections of this ‘system’ were but part of a wider framework of military deployment. As many as thirty forts lay in the vicinity of the Fossatum, but, as Fentress has remarked, it is unlikely that all were occupied at once.4 The difficulty lies in trying to date the forts in an area which has seen little recent archaeological research. Some, filling in over-long gaps on the main routes to north and south of the Aurès Mountains, probably date to the reign of Hadrian, but inscriptions attesting work on the road network under Antoninus Pius and Commodus indicate that activity continued for some time. Under Commodus, towers were also built. One, erected in 188 to the rear of the Mesarfelta-Thubunae section of the Fossatum Africae in western Numidia, was described as ‘an observation post, Commodianum, to be established between two roads as a new protection for the safety of travellers’ (CIL VIII 2495). An important development under Antoninus Pius was the construction, in 148–9, of a fort near Medjedel – to the south of the High Plains on the northern slopes of the Saharan Atlas, controlling access to a pass through the mountains. Medjedel lies about 150km (100 miles) south-west of the Hodna Mountains and south of the neighbouring province of Mauretania Caesariensis. It may have been supported by other forts, but we know of none of that date for certain. Under Marcus Aurelius an inscription cut in 174 recorded a Roman military presence further to the south-west at Agneb at a point 400km (250 miles) south of the frontier (CIL VIII 21567). No fort has been located here. Under Septimius Severus there was considerable activity in Mauretania Caesariensis and Numidia. Much, perhaps most, of the building of this time seems

Fig. 42: Map of the frontier in Mauretania Caesariensis; the half-filled squares indicate forts of the first and second centuries, the black squares indicate the Severan forts.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 140

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 141

Mountain Frontiers 141 to relate to a special command given to Anicius Faustus, governor of Numidia from 197 to 201. This covered an enormous area, over 1,500km (1,000 miles) long. Faustus was active building forts in the desert south of Tripolitania as well as in Mauretania Caesariensis. He continued the process of the previous fifty years, extending the construction of forts to the south-west. One of these forts, Castellum Dimmidi/Messad in the Wadi Djedl, was built in 198 (AE 1948, 214). The fort was linked to the province by a number of fortlets. Many of these military installations were as much as 50km (30 miles) apart and the number of troops based in each was small. Why these forts were built so far south is, however, another matter. Possibly they reflected continuation of the general imperial expansionist policy, clearly exemplified by Severus. Possibly there were specific reasons, such as the need to maintain closer observation of potentially troublesome tribes beyond the frontier, such as the people living in the Chott el-Hodna, and of caravan routes. Some forts, such as those in the Saharan Atlas, were relatively small, with space for about a hundred men, and may have served as outposts and the bases for forward patrolling. They were certainly too small to deter or repel an invasion; they were of a size only to deter raiding. As Charles Daniels pointed out, this indicated both considerable Roman self-confidence and little tribal opposition.5 The occupation of the area south of the Saharan Atlas by the army of Numidia rather than by the army of Mauretania Caesariensis to the north of the mountains appears, at first glance, to be strange. However, the links of this region had long been towards Numidia, whose army had maintained watch over this approach to the empire and the people who lived there. Further, it was easier to sustain and support troops in the oases of this area rather than across the arid steppes of the High Plains. Nor did the army of Mauretania Caesariensis have the manpower to take on the occupation of this area. As it was, many of these forts were maintained by the soldiers from the Third Augustan Legion based at Lambaesis. There was a disruption in the occupation pattern in the late 230s and the 240s. This may relate to the temporary disbanding of the Third Augustan Legion following its opposition to Gordian III in 238. Castellum Dimmidi/Messad and possibly other outposts in and beyond the Saharan Atlas were abandoned at this time. It has also been suggested that the abandonment reflected the strain on the provincial army caused by the occupation of so many forts. Yet, other forts were built or rebuilt in the frontier area in the 240s, which implies that the main reason for their abandonment was the disbanding of the legion. There was serious warfare on the frontier in the second half of the third century, but, apparently, no significant loss of territory. Diocletian divided the province into two and increased the size of the army. Southern Numidia saw the construction of new small forts from about 300 onwards, reinforcing the existing installations. These were of the standard type, roughly square, with the internal buildings erected against the enclosure wall and surrounding a courtyard. As in Arabia, it is difficult to determine whether these were military or civilian in function. The North African frontier continued to

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 142

142 The Frontiers of Imperial Rome operate into the fifth century, perhaps ending with the Vandal invasion in the 430s, though Roman North Africa was to survive until the Arab invasions of the seventh century.

Mauretania Caesariensis The province of Mauretania Caesariensis consisted mainly of the cities along the Mediterranean Sea which occupied a narrow coastal plain together with their hinterland as far south as the Atlas Mountains, the higher reaches of which were densely forested. The mountains consisted of a series of ranges running parallel to the coast. They were extremely rugged, but well watered. Between the two northern ranges lie some fertile areas such as the valley of the River Chélif. To the south are the high arid steppes of the High Plains, fringed on its southern edge by the Saharan Atlas. The rainfall was sufficient in the Saharan Atlas to allow for some farming supporting significant populations. Little is known about early military deployment though the army was probably based in the area of the capital, Caesarea/Cherchel. Under Claudius the first steps were taken to disperse the army when new forts were established in the fertile Chélif valley some 50km (30 miles) south of the capital. In its lower reaches, the valley ran east–west and this was the line followed by the forts extended under the Flavians and by Trajan westwards to the Mediterranean at Siga and eastwards to Auzia where one branch ran to Saldae/Bejaia on the coast, while the other continued round the north side of the Hodna Mountains to Zarai on the border with Numidia. The coastal cities were now protected by a single line of forts linked by a road, but, apart from in the Chélif valley, following no natural line. In the western sector, infantry and cavalry units were alternated, presumably not a haphazard arrangement. Nevertheless, it is difficult to see the forts as forming a frontier line. Rather, they are located in the best positions on routes within the mountainous landscape. Inscriptions attest the building of forts along this line under Hadrian, at Albulae at the western end of the province in 119 and at Rapidum and Thanaramusa Castra on the eastern sector of the road leading to Numidia in 122. As might be expected in a mountainous terrain, there appears to be no standard spacing between forts. Distances of 30km and 45km (18 and 26 miles) occur in this sector. There may also have been towers along this road at this date for later building under Commodus refers to ‘old towers’. It is noticeable that most towers were erected in the more rugged, mountainous regions, such as that where Nonius Datus of the Third Augustan Legion was robbed in 152 while journeying to Saldae/Bajaia to try to help its citizens build a new aqueduct (CIL VIII 2728 = ILS 5795). In the following century, Gargilius Martialis, based at Auzia, undertook decisive action against a rebel chieftain, while the governor Aurelius Litua celebrated a victory in Caesarea/Cherchel over a group of barbarians called the Transtagnenses (CIL VIII 9047 = ILS 2767; CIL VIII 9324). In Mauretania Caesariensis a new line of forts supplemented by some fortlets and towers, linked by a road, was established south of the Flavian–Trajanic–Hadrianic

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 143

Mountain Frontiers 143 line and south of the main coastal mountain ranges; it was termed ‘the new frontier zone’ (Nova Praetentura) (e.g. CIL VIII 22602 = ILS 5850; CIL VIII 22611). Fort and road building is attested in 198 and in 201. Many of the earlier forts were abandoned; Auzia, at a nodal point in the road network, was retained. Valleys were used to provide both east–west and north–south links where possible with some forts carefully positioned to control passes leading south, routes used by modern roads.6 The new east–west road lay roughly on the southern border of the province, though there were areas of settlement beyond it. The road also lay on the edge of the mountains, a line which corresponded to the southernmost limit of the dry farming of cereals. Some of the new forts were described as winter quarters, which suggests that in the summer months the army was on patrol beyond the provincial boundary.

Mauretania Tingitana On the south-western fringe of the empire lay the province of Mauretania Tingitana. This was not only small, being restricted to the relatively fertile coastal plain south of the Straits of Gibraltar, but also separated from its neighbouring province to the east at the point where the Rif Mountains overlook the Mediterranean coast. The Atlas Mountains provided an effective eastern boundary to the province. There is good evidence for the relationship between the Romans and the local tribes, especially the Baquates, south of the province being controlled by treaties through the second and third centuries (e.g. AE 1920, 44; 1921, 23; 1957, 202–4). Yet the province held a force little smaller in size than Mauretania Caesariensis. The forts appear to be spread across the province, with a concentration in the area of Volubilis. Through the second century, however, they were gradually dispersed, mainly along a line from Thamusida on the coast to Volubilis. These were supported by towers. The most south-westerly point of the province was at Sala, near modern Rabat, where there is a mere 15-km (10–mile) gap between the foothills of the mountains and the sea. The town was walled under Antoninus Pius and a fort subsequently placed here. Extra defence was provided in the form of a ditch, in places supplemented by a wall, located some 6km (3¾ miles) south of the town. Eleven kilometres (7 miles) long, this reached from the sea to the River Salat, which ran along the edge of the higher ground, and barred access to the province from this direction. The date of the ditch and wall can only be assigned loosely to the second century and, while it is generally assumed to have had a military purpose, a role as a municipal boundary has also been suggested.7 The late third century saw serious fighting in Mauretania Tingitana and at the end of the century, under Diocletian, most of the southern part of the province, including Volubilis, was abandoned, though Sala was held into the fourth century and is recorded in the Notitia Dignitatum. The rump was now associated with the diocese of Spain.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 144

144 The Frontiers of Imperial Rome

The Red Sea Mountains The western coast of the Red Sea was skirted by a range of mountains through which passed the roads linking the Red Sea ports to the Nile (see pages 129–33). These routes naturally made use of the passes through the mountains. Along the roads lay small forts linked by towers carefully located on high points on the edges of the twisting ravines so as to provide for both observation and communication. Britain In the 80s the advance of Roman arms halted at the edge of the Highlands (fig. 18). There was planted a line of forts. These included a legion at Inchtuthil on the north side of the River Tay at the entrance to a major glen which today carries the main road north. An auxiliary fort was placed in the mouth of each of the glens leading into – and out of – the Highland massif. The purpose of these, however, is a matter of dispute.8 One interpretation is that they were so placed to prevent raiding or even invasions down the glens. One difficulty with this interpretation is that archaeological research has suggested that relatively few people lived in the Highlands at that time. Another view is that they were springboards for advances up the glens, a forward movement which did not happen because troops were withdrawn from Britain soon after Agricola’s victory over the Caledonians at Mons Graupius in late 83 to support the hard-pressed armies on the Lower Danube. In this interpretation, the pattern of military deployment which survives is merely part of a wider network which was planned but never completed. Moreover, this interpretation might be thought to fit well with the judgement of Tacitus, ‘Britain was conquered and immediately abandoned’ (Tacitus, Histories 1, 2), though it could equally well reflect the abandonment of Inchtuthil and all forts north of the Cheviot Hills (roughly the modern Anglo-Scottish border) just four years after Mons Graupius. It might be argued that it would not have made sense for the Romans to try to occupy the Highlands, but these mountains were no more difficult terrain for the army than other areas the Romans mastered, while the Hanoverian army in the eighteenth century demonstrated how the task could be achieved. They established a system of military deployment very similar to those used by the Romans elsewhere with large forts, often holding two regiments, in the main valleys and fortlets between, all linked by roads. Further, the presence of a legion on the edge of the Highlands under Agricola and the construction of the legionary base at Carpow, also on the Tay, by Severus, who, according to Dio, intended to complete the conquest of the island, might be considered as giving support to the possibility of Agricola (or his successor) having the same intention. Discussion The Roman attitude to passes comes into the category of ‘control of routes’. Valleys were always important lines of communication. Forts were carefully placed in Dacia to watch over passes through the Carpathians and similarly in the Caucasus

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 145

Mountain Frontiers 145 Mountains. Where there were breaks in the high plateaux beside the River Euphrates which allowed for a route across, there was established a legionary fortress. In Germany, each pass through the Odenwald was guarded by a fort. The only mountains which appear to have been treated differently were the Atlas Mountains in North Africa, but this may relate in part to our poor knowledge of the details of the frontier installations.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 146

Chapter 11

Sea Frontiers

he ultimate frontier was the sea. Such a frontier was achieved in the west where Rome’s armies reached the Atlantic in Mauretania Tingitana, Spain, Gaul and through much of Britain. The conquest of Spain took 200 years and throughout this period there is no sense of the Romans seeking to bring their arms to the sea so as to solve the problem of the frontier. Yet, once the sea had been reached and the various states of the peninsula subdued, it was then possible to remove most of the army, leaving only one legion and some auxiliary units, whose main duties may have been to supervise the important silver mines. Gaul, too, once its conquest had been completed and the new provinces pacified under Augustus were largely demilitarized with only an urban cohort based at Lyon, the site of a mint. Bringing the boundary of the empire to the sea accordingly had advantages in terms of stability and manpower. The elimination of piracy by Augustus had a similar effect in the Mediterranean Sea. Yet, while the Mediterranean basin, including the Black Sea, was under Roman control, any relaxation of that control might have severe consequences (Tacitus, Histories 3, 49). Pirates took advantage of Rome’s Civil War in 68–9 to raid towns round the Black Sea. At the end of the Jewish War of 66–73, some dissidents turned to piracy and caused havoc along the coast from Syria to Egypt (Josephus, The Jewish War 3, 9, 2). About 100 years later, the Costoboci sailed out of the Black Sea to attack towns round the Adriatic, perhaps taking advantage of Rome’s preoccupation with the Marcomannic Wars and the instability on the eastern frontier (Pausanias, Description of Greece 10, 34, 5). Vigilance was always necessary. And not just in the Mediterranean. To the west, in the second century Moors sailed across the Straits of Gibraltar to raid Spain (Historia Augusta, Life of Marcus, 21; Life of Severus 2) while in the third and fourth centuries the Scots and the Attacotti from Ireland harried the western coast of Britain and the Saxons and Franks attacked the east coast of Britain and the west coasts of Gaul (Panegyric of Constantius 11, 1; Ammianus 20, 1, 1; 26, 4, 5; 27, 8, 1). The Romans took steps to deal with each of these threats.1 The sea offered important lines of communication as well as cheaper transport than travel overland. Some scattered outposts of empire were only reached by sea, including the cities and forts around the Black Sea and Essaouira. This city lay 400km (250 miles) south-west of Sala in Mauretania Tingitana along the Atlantic coast. It was founded by the Carthaginians and lay at the end of a caravan route into

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 147

Sea Frontiers 147 the interior of Africa. Juba II of Mauretania established a factory here in the time of Augustus to extract purple dye from the molluscs found along the coast. The discovery of a Roman villa on the Purple Isles suggests continuing Roman interest in the place, though no military presence has been detected.

Britain Roman intentions for Britain were as vague as those for Spain. Claudius arrived in 43 in the wake of the successful invasion and after participating in the conquest of Colchester and indulging in some diplomacy, he returned to Rome asking his governor to ‘subdue the remaining areas’ (Dio 60, 21, 5). This, the governor, Plautius, and his successors proceeded to do. During the course of the advance, the sea was reached in a number of places where there was still unconquered land beyond. In the mid-first century, two fortlets were built on the north Devon coast to maintain watch in case of attack from the Silures of south Wales across the Bristol Channel (fig. 47). The extension of Hadrian’s Wall down the Cumbrian coast is also usually ascribed to the need to maintain watch over the Solway Estuary, the lands beyond which lay outside the empire. After the Civil War of 68–9 and the subsequent Batavian Revolt, a new advance began under Vespasian, who had taken part in the invasion of Britain twenty-five years earlier. The vigorous actions of his governors suggest that the intention was to complete the conquest of the island. Vespasian’s third governor, Agricola, pushed Roman arms further north than any of his predecessors, but in the course of his northern advance he paused and considered halting rather than completing the conquest of the island. In relation to his fourth season, 80, Tacitus stated: if the courage of the armies and the glory of the Roman name had allowed, a halting place would have been found within Britain itself for the Clyde and the Forth, which run far inland on the tides of the two seas, are separated by a narrow stretch of land. This was at that time fortified with garrisons and the whole area to the south made fast, the enemy being removed, as it were, into another island. (Tacitus, Agricola 23. Translation by A. R. Birley.) This is an important statement, implying that Rome might not proceed to the conquest of the whole island. Few certain Agricolan sites have been on the isthmus – two forts and a small fort – so our understanding of the arrangements set in hand by Agricola is sketchy. However, neither the glory of the Roman name nor the courage of the army allowed the halt to be permanent and two years later Agricola continued his northern advance. His pause coincided with the reign of Titus (79–81), his move forward with the start of the reign of his successor, Domitian, also opening campaigning in Germany at the same time. It was during the pause in campaigning in Britain that Agricola reputedly considered the invasion and conquest of Ireland. If that had been achieved, it would

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 148

148 The Frontiers of Imperial Rome have brought the empire to the Atlantic Ocean and removed all threat of invasion of the western seaboard. Agricola’s final campaign ended with victory over the Caledonians at the Battle of Mons Graupius in 84. Whether he or his successor intended to complete the conquest of the island by occupying the Highlands cannot be known for wars elsewhere conspired against any such intentions. The events elsewhere were the Dacian invasions of the provinces of the lower Danube in the second half of the 80s. Troops were needed for these wars and they were withdrawn from Britain, which was, after all, much further from Rome and the empire’s core provinces than the theatre of war. Perhaps a quarter of the army was transferred from Britain to the Continent, leading to the abandonment of many northern forts, and the chance of completing the conquest of Britain being lost for that generation. In fact, the cause was lost for longer. Trajan directed his attentions elsewhere, Hadrian did not favour expansion and Antoninus Pius, who did expand in Britain, had limited aims halting his advance at the River Tay. It was not until the dynamic and militaristic emperor Septimius Severus was on the throne that another opportunity to complete the conquest of the island occurred. In 208, Severus came to Britain (pl. 1c). Various reasons are offered by the historians Cassius Dio and Herodian for his campaigns against the Caledonians and Maeatae, but the former explicitly states that the emperor’s aim was to conquer the rest [of the island] and to that end he reached the end of the island where he took readings of the stars (Dio 77, 13). A new legionary base was built on the River Tay at Carpow. The untimely death of Severus at York on 4 February 211 brought an end to his ambition. His sons ‘made treaties with the enemy, evacuated their territory, and abandoned the forts’ (Dio 78, 1, 1). Britain throughout the third century remained peaceful, or at least our sources quiet, and the emperors were consumed with saving their own skins rather than expanding the empire which each held so briefly. In the new world Diocletian created, Constantius I campaigned against the Picts in 305, as did his son and grandson, but the tide of history was against expansion of the empire in Britain and the completion of the conquest of the island. The conquest of Spain took the Romans 200 years, admittedly in the haphazard world of the late Republic. Two hundred years from 43 would take Britain into the middle of the turbulent years of ‘The Anarchy’. At that stage, there was certainly no chance that the whole island would be conquered.

The Black Sea The clearest evidence for the protection of a sea frontier lies within the pages of Arrian’s Circumnavigation of the Black Sea, written in the 130s. Arrian describes the units based at four forts around the south-eastern edge of the Black Sea, and their purpose; although he nowhere explicitly states the purpose of the forts, this is implicit from his account (fig. 43). He notes larger garrisons at certain places and he describes the adjacent peoples and their political affiliations. One purpose of the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 149

Sea Frontiers 149

Fig. 43: Map of the eastern end of the Black Sea.

forts would appear to have been to keep watch over these tribes. He also describes the physical location of the forts, recording which river mouth each fort sat beside and guarded. It was presumably no coincidence that the largest fort of the series, at Phasis/Poti, lay beside the largest river flowing into the south-eastern corner of the Black Sea. Arrian specifically mentioned pirates on the Black Sea and stated that they needed dealing with (Arrian 11, 1–2). This was particularly important because the north Black Sea coast provided corn for the army of Cappadocia and the supply lines needed guarding from pirates, who had previously been a threat in the region. Tacitus acknowledged the importance of the provisions from this area when he

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 150

150 The Frontiers of Imperial Rome described the war of Corbulo against Armenia in 58 (Tacitus, Histories 13, 39). Tacitus also described what happened when the fleet was not available to control the waterways, in this case because Rome was preoccupied with a civil war: There was a sudden armed uprising in Pontus led by a barbarian slave who had once been prefect of the royal fleet … At the head of a band, which was far from being negligible, he suddenly attacked Trabzon, a city of ancient fame, founded by Greeks at the extreme end of the coast of Pontus. There he massacred a cohort, which originally consisted of auxiliaries furnished by the king. Later its members had been granted Roman citizenship and had adopted Roman standards and arms, but retained the indolence and license of the Greeks. He also set fire to the fleet and escaped by sea, which was unpatrolled since Mucianus had concentrated the best light galleys and all the marines at Byzantium. Moreover, the barbarians had hastily built vessels and now roamed the sea at will, despising the power of Rome. (Tacitus, Histories 3, 47. Translation by C. H. Moore.) We can see from Arrian that the forts running along the Black Sea coast eastwards from Trapezus/Trabzon had a dual role in that they faced both to sea in order to provide bases for the fleet seeking to keep the seaways open and protect traffic from pirates and inland to keep watch over the peoples of Colchis. The more easterly forts lay close to the kingdoms of Albania, Iberia and Armenia, as well as the Alans north of the Caucasus. There were as many as five cohorts stationed at Apsaros, presumably because of its strategic position in relation both to the client kingdom of Iberia and to the Dariel Pass. Speidel has argued that this was the base for Arrian’s expedition against the Alans in 135.2 After Arrian’s governorship, the chain of forts was extended further round the eastern coast of the Black Sea with a new fort being established at Pityos where a detachment of the legion based at Satala/Kelkit was outposted. Forts are still visible at two of the sites described by Arrian, but from their design they date to the late empire. They demonstrate continuing Roman interest in the area. The forts along the southern and eastern fringes of the Black Sea were supported by the fleet based at Trapezus/Trabzon located at the north-east corner of the province of Cappadocia and the point from which Arrian sailed. We may assume that the fleet patrolled the Black Sea and ensured the safe transport of supplies to the forts. The Romans had considerable knowledge of the Black Sea coast and about half of Arrian’s account of his circumnavigation is a description of that part of the coast that was not in Roman hands. Arrian does not mention Roman military deployment along the northern shore of the Black Sea, presumably because it lay beyond his jurisdiction. Several forts are known between the Danube delta and the Crimea, some probably serving as fleet bases (fig. 44). One base lay at Chersonesus Taurica/Sevastopol. It was apparently occupied from the first through to the third century. Inscriptions attest activities

Fig. 44: Map of the north-western corner of the Black Sea.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 151

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 152

152 The Frontiers of Imperial Rome there by three legions and a similar number of auxiliary units as well as the Moesian fleet. A detachment of the Ravenna fleet is even recorded at another fort in the Crimea, Charax/Ai-Todor, near Yalta. Most inscriptions at these sites date to the time of Antoninus Pius and Marcus Aurelius. Piracy is recorded after the time of Arrian, but it was Roman weakness in the second half of the third century which resulted in a more serious resumption of piracy in the Black Sea. The Borani from north of the Black Sea obtained ships from the kingdom of the Bosphorus, whose affinity to Rome had slackened during the century, and ravaged the east coast of the Black Sea. In their second raid, they attacked Phasis/Poti and sacked Pityos and Trapezus/Trabzon. Their example was followed by the Goths who ravaged the western seaboard of the Black Sea. Unsurprisingly, Rome continued to maintain her military bases round the eastern end of the Black Sea, at least up to the end of the fourth century for several are listed in the Notitia Dignitatum (Or. XXXVIII)

The Red Sea The roads from the Nile across the Eastern Desert and Red Sea Mountains reached the western coast of the Red Sea at several locations (fig. 39, pl. 28). Travellers along these routes were protected by soldiers based in fortlets, as we have seen (pages 129–32). There seems to have been a military presence at two of these termini, Myos Hormos/Quseir and Berenice, and, later, a fort at Abu Sha’ar. In 137, Hadrian created a new road, the via Hadriana. It left the Nile at Antinoopolis, reaching the Red Sea at Abu Sha’ar, and continuing southwards to Berenice. Although provided with ‘abundant cisterns, way-stations and military posts’, little is later heard of this road and it may never have become a significant line of communication. It is possible that the army maintained a military presence on the east side of the Red Sea. A description survives of sites along this coast in an account of a voyage, The Circumnavigation of the Red Sea, undertaken in the middle of the first century when the area was still under the control of the kings of Nabataea. This mentions Leuke Kome/White Village: a port and garrison, from which there is a road to Petra. … It serves as a market-town for the small vessels sent there from Arabia, and so a centurion is stationed there as customs officer collecting one-quarter of the merchandise imported, with troops as a garrison (Circumnavigation of the Red Sea 19). The Romans may have continued to maintain a military presence at Leuke Kome/White Village. There is no evidence from the time of Augustus that the Arabs across the Red Sea had a significant maritime presence. However, as Rome’s trade with the East grew, so did piracy. It is possible that Hadrian’s construction of a new road may have been intended to improve contact between the interior of the province and the fleet.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 153

Sea Frontiers 153

The North Sea Piracy became a serious problem at the western end of the empire in the late third century. A special command was created and given to Carausius. Eutropius explains what happened: At this time too Carausius, although of very humble birth, had achieved an outstanding reputation in a vigorous career. He had been given the responsibility throughout the Belgic and Amorican areas, with his headquarters at Boulogne, of clearing the sea, which was infested by Franks and Saxons. On many occasions he captured large numbers of barbarians but he failed either to return all the booty to the provincials or to send it to the emperor. A suspicion grew up that he was letting in the barbarians on purpose so that he could catch them as they passed with their booty and grow rich on the proceeds. So Maximian ordered him to be put to death, whereupon he declared himself emperor and seized Britain. (Eutropius 9, 21. Translation by H. W. Bird.) Carausius, and his successor Allectus, held power in Britain for nine years from 287 before the empire in the person of the Caesar Constantius Chlorus was able to recover the island. That part of the story does not concern us here; our interest is in the piracy and the measures which Rome took to deal with it. Nevertheless, we should note that Carausius’ command related only to Gaul and the implication is that it was this area which was suffering from the attention of the German raiders. There had been a British fleet probably from the time of the conquest. It was based at Gesoriacum/Boulogne in Gaul. There were also bases in Britain, including the fort at Dubris/Dover which dates to the early second century. The latest reference to the British fleet dates to the reign of Philip, 244–9. An anonymous writer of the early third century stated that the purpose of the fleet was to protect Gaul (Pap. Lat. VIII, 14, 12, 1), and Eutropius also recorded that it was the coast of Gaul that was attacked by the Franks and Saxons (Eutropius 9, 21). The fort at Dubris/Dover was slightly unusual in plan, but recognizably a Roman fort. There are two other coastal bases, stations which give the appearance of being normal early-empire-style forts, apparently built in the third century. These are Regulbium/Reculver on the north Kent coast and Branodunum/Brancaster on the north Norfolk coast; both also have attested at them early-empire-style auxiliary cohorts. There is, however, another group of forts along the east and south coast of Britain which are very different. These are Gariannonum/Burgh Castle, Walton Castle and Othona/Bradwell, north of the River Thames, Rutupiae/Richborough, a new fort at Dubris/Dover, and forts at Lemanis/Lympne, Portus Adurni/Portchester and Anderita/Pevensey south of the river (figs. 45 and 46). They are all built in the style of the late empire with high, thick walls, small gates and external bastions. All of

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 154

154 The Frontiers of Imperial Rome

Fig. 45: Map of the Saxon Shore in Britain and the forts along the opposite coast.

these are square or rectangular in plan, with the exception of Pevensey which is oval. The dating evidence for the construction of these forts is slight. Stephen Johnson has argued on the basis of their architectural style that they were probably built within the fifty years from 250 to 300, and that Gariannonum/Burgh, Walton Castle, Branodunum/Bradwell in East Anglia and Rutupiae/Richborough, Dubris/Dover and Lemanis/Lympne ‘should fall within the decade 275–85, with Portus Adurni/Portchester and even Anderita/Pevensey following in reasonably close sequence.’3 Johnson’s date of 275 comes from Rutupiae/Richborough where ‘abundant finds’ indicate this date for the construction of the new fort. Coins have

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 155

Sea Frontiers 155 been taken to indicate a date between 286 and 290 for the construction of Portus Adurni/Portchester, while a single coin of 330 is usually taken to provide a terminus post quem for Anderita/Pevensey. Nine of these forts are listed in the Notitia Dignitatum (Oc. 28) as being in the command of the Count of the Saxon Shore; the missing fort is Walton Castle. In all but two cases the unit recorded at each fort is one of the new fourth-century-style units. It is also noteworthy that they are military units not fleets. Rather surprisingly, the forts are not listed in geographical order, but they do appear to be in pairs. John Mann drew attention to the fact that each pair seems to include one fort which had a harbour and one with a beach for landing. Thus, Dubris/Dover had a harbour and Lemanis/Lympne a beach.4 One other element of this command is worth pointing out. This is the lack of roads connecting any of the forts with its neighbour. It is probably a safe assumption that the units at the forts communicated with each other by sea. The way in which the late-fourth-century command known as the Saxon Shore operated has long puzzled archaeologists. It appears to grow out of the earlier British fleet, but no naval units are listed in the command. Little is known about the interior of the forts. There is no known military hinterland to the forts and little known of communications with the interior of the province by which reinforcements could be moved forward. However, before considering the operation of the command, the military installations on the other side of the Channel should be reviewed. It is unfortunate that less is known of the area of Carausius’ command in Gaul. A hundred years later, the duke of the Armorican and Nervican coast held a command which extended through five late Roman provinces along the north and west coasts of modern France and Belgium. Not all the forts in the command can be identified and few have been extensively examined archaeologically. The known sites extend from Nantes in the south to Rouen in the north. As across the Channel, none of the units in the coastal forts identified in the Notitia Dignitatum has a title indicating that it had a naval component. The earlier fort at Gesoriacum/Boulogne dates to the second century. It is large at 12ha (30 acres), but conforms to the general plan of the period. A little more is known of the later forts and they too have the characteristics of forts built in the late third or fourth century. The occupation of this fort came to an end shortly after 269 when it was replaced by a smaller station, a date comparable to the forts across the Channel. The operation of the Saxon Shore has to be reconstructed from the scarce literary and archaeological evidence. Stephen Johnson described it thus: The tactical purpose of the Saxon Shore forts was threefold. They were strongholds and naval bases for the sailors of individual flotillas whose duty was to control pirate raids; they were garrisoned and defended bases where a body of mobile troops was ready to combat pirate landings; and they were an active discouragement, sited on the principal river estuaries, to penetration by pirates into the inland areas of both Gaul and Britain. (Johnson 1976, 126.)

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 156

156 The Frontiers of Imperial Rome

Fig. 46: Plan of the Saxon Shore fort at Portus Adurni Portchester.

Johnson emphasized that the system of forts faced primarily towards the north-east, towards the area where the Saxons and Franks lived. As they sailed southwards, they were prevented from landing by the forts and flotillas on patrol and therefore pushed on towards the Channel Straits where a concentration of defence was waiting for them. There they would be caught and any which slipped through the trap could be apprehended by the sailors based in the forts to the south and west.5 This scenario has, however, been challenged. John Cotterill, on the basis of an assumption that the main role of the British fleet had been the supply of men and materials, has argued that the Saxon Shore forts served as defended collection centres for state supplies prior to their trans-shipment elsewhere.6 This proposal has certain advantages in that it helps to explain the gradual development of the network, the fact that most forts appear to have been constructed before the first reference to piracy in the North Sea, which relates to Gaul not to Britain, and the unusual senior status of the officer in command of the Saxon Shore.

Fig. 47: Plan of the Saxon Shore fort at Anderita/Pevensey.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 157

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 158

158 The Frontiers of Imperial Rome We may note that there are other late forts along the west coast of Britain, at Cardiff and Lancaster. At Caernarfon there is a fortlet and, on the adjacent island of Anglesey, a fortified landing place.7 These may be connected as such sites often appear in pairs on the rivers Rhine and Danube though in this case the two sites are not opposite each other. There is also evidence for the reoccupation of some sites on the Cumbrian coast further north. It is difficult to know whether these are connected with the protection of supplies or simply the western seaboard generally. Whichever explanation is preferred, they imply a disturbed situation in the Irish Sea. The Irish are described as the enemies of Rome in a passage relating to events in 297 (Eumenius, Panegyric of Constantius 11, 3) and the Scots and Attacotti of Ireland were troublesome in the later fourth century. The last sea frontier of Britain lay in the north-east where a series of five towers has been recorded along the coast of Yorkshire. The towers are of the usual late Roman type, being about 16m (50ft) square with internal supports for floors, within an enclosure about 30m (100ft) square with external towers or bastions at each corner and a surrounding ditch. An inscription from one site records the construction of a tower and ‘fort’ (RIB 721). Pottery evidence suggests a date after 367, which coincides with the architectural details. Construction under Valentinian who is recorded erecting towers on both the Northern and Eastern Frontiers would not be out-of-place. These towers were intended to be tall and, it would appear, intervisible, presumably so that the observations from one group of soldiers could be passed down the line. Where, however, the message would go is not clear as there is little evidence for the necessary support troops which would have been required to meet any maritime threat, though it is not impossible that one or more fleet bases have been lost to coastal erosion.

Discussion The Romans may not at first have been sailors, but during the empire they developed their fleets to ensure that the seaways were as open as land routes as well as using them as part of the armoury of protection on the frontier lines. Rome was normally able to control activities within the Mediterranean basin, but the western seaboard was more difficult. Defensive measures were taken against attacks from the Franks and Saxons, but, so far as we know, no proactive military expeditions were launched to stop the problem at source, which would have been difficult to achieve anyway. Fleets required land bases. Many of these are known on the river and sea frontiers of the empire. Unfortunately, we have little idea of the relationship of the sea fleets to their land bases. Nor do we have any greater information on how the land bases, forts and towers related to the wider military infrastructure of the provinces within which they sat.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 159

Chapter 12

Forests, Marshes and Swamps

ord Curzon included forests, marshes and swamps as helping to define frontiers. Many ancient writers refer to forests and marshes in describing the campaigns of Roman generals. Both Agricola in the late first century and Septimius Severus in the early third were said to have encountered these in north Britain, even though pollen analysis suggests that the areas through which they marched were largely devoid of trees, having been farmed for nearly 4,000 years.1 Tacitus mentions the Frisii ‘moving their fighting men over marshes and through woods near the Rhine in order to occupy land assigned to Roman soldiers but not used’ (Tacitus, Annals 13, 54). In discussing the campaigns of Constantius II on the Danube in 358, Ammianus records that the marshes around the junction of the Danube and the Tisza could not be crossed without local knowledge, thus helping protect the local people from both the Romans and the barbarians (Ammianus 17, 13). Earlier, this area had lain in a corner between the provinces of Lower Pannonia, Upper Moesia and Dacia, the area of modern Banat. There were few forts in this south-western sector of the Dacian frontier and it has been argued that this is because the Roman army regarded the marshes of the River Tisza and its tributaries as sufficient protection, though we should be cautious as the lack of forts may rather reflect our lack of knowledge. Marshes and swamps of course existed in many areas up to the great era of drainage in the eighteenth century and it has been argued that their existence in north Britain affected the precise line of Hadrian’s Wall and the Antonine Wall. In places, the line of the Antonine Wall plunges down into the valley which lies to the north, ignoring slightly raised ground to the south. John Poulter has argued that the army may have considered that it was better to place the Wall on the edge of the marsh, leaving no solid ground between the Wall and the marsh. It is noteworthy, too, that there is little cavalry based in the Antonine Wall forts and that may be related to the boggy ground to the north of much of the Wall which would not have been good cavalry country.2 In Upper Germany, the frontier was pushed forward to what is regarded as the edge of the forest under Antoninus Pius in the 160s and that belief is underlined by the lack of artefacts in the land beyond the frontier. In Noricum, few forts were placed in the western sector of the frontier along the Danube where there was dense forest to the north. It would appear that the Romans did pay attention to the location of forests in relation to their frontiers.

L

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 160

160 The Frontiers of Imperial Rome The anonymous writer of de rebus bellicus recorded that the native tribes beyond the empire were either hidden by forests or lifted beyond our reach by mountains or kept from us by the snows; some, nomadic, are protected by deserts and the blazing sun. There are those who, defended by marshes and rivers, cannot even be located easily. (de rebus bellicus. Translation by R. Ireland.)

Discussion It seems that forests and marshes played only a small and localized part in the location of frontiers. In Germany, the spacing of forts on that section of the frontier facing the Black Forest was no different from elsewhere. There also appears to have been no difference in the distribution of forts along the Danube in the area of the great marshes at the junction with the River Tisza than elsewhere on the river. The presence of the marshes has, however, been suggested as a reason for the fewer forts in the adjacent section of the Dacian frontier.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 161

Chapter 13

The Deep Frontier: Defence-in-Depth?

n most parts of the Roman world, the frontier appears to have been a thin line of military installations. In fact, the military zone was much wider as the army drew on the surrounding area for supplies, for example, and might interfere in the affairs of states beyond their border, sometimes to the extent of basing military units there. This is what might be termed the frontier zone. At some times and in some areas, the army maintained forts behind the frontier. Legionary bases might be left behind as the frontier advanced, as in Britain, Upper Germany and Syria. In Mauretania Caesariensis some early forts remained in use when Severus moved the frontier forward, notably Auzia at a significant road junction. In the fourth century, for example, as we have seen, forts and supply bases were occupied behind the frontier on the Upper Danube. The line of stations along the road leading into the empire behind the Lower Rhine may be seen in the same light. Under Valentinian, however, it has been suggested that some of the forts behind the frontier were abandoned as new posts were built on or across the Rhine and the Danube. Dacia may be seen as a special case. Here and in Britain the patterns of military deployment are different from many other frontiers and certainly most of continental Europe. Dacia has already been considered. In Britain the forts are spread across the landscape with the linear barrier part of a wider network.

I

Britain The frontier arrangements in Britain are very different from those elsewhere in Europe in two ways. Both linear barriers, Hadrian’s Wall and the Antonine Wall, are, as we have seen, stronger in every way than the artificial frontiers in Germany. Their placement within the military landscape is also different. In Britain the two linear barriers lay within a broad military landscape (figs. 14 and 18). To their north lay outpost forts, to the south forts in their hinterland. Even further south were the legionary bases, Eburacum/York to the east of the Pennines and Deva/Chester on the west coast. The continuing occupation of these bases was probably simply the result of inertia. The failure to conquer the whole island of Britain left the legions in the locations which were once in the vanguard of conquest but now far to the rear. This is underlined by the position of the Second Augustan Legion at Isca/Caerleon in south Wales. The pattern of military deployment in northern Britain can be traced for 300 years from the late first century. At that time, before the construction of Hadrian’s

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 162

162 The Frontiers of Imperial Rome Wall, the network of forts extended northwards as far as the edge of the Highlands. The general arrangement was for the major forts to sit in river valleys, presumably because that is where most people lived, with smaller posts in between, all linked by roads. The unusual element at this time is a line of timber towers and fortlets along the main road leading north from the River Forth to the River Tay (see fig. 18). These have been seen as a frontier, similar to the broadly contemporary line of towers in Germany. But the towers in Germany stand alone, whereas those in Britain are part of a wider framework of military deployment. In that way, they are only part of the pattern of military control and thus are rather similar to the position held later by Hadrian’s Wall and the Antonine Wall in their military landscapes.1 The tribes in Wales and the Brigantes in northern England were both held within a network of forts linked by roads. As they became more peaceful, or as priorities changed, forts might be abandoned and units moved elsewhere. After the construction of Hadrian’s Wall, there were few units left in Wales and those that remained were probably responsible for supervising the mines.2 In the third century, from 211 to 305, we know of no warfare on the Northern Frontier. While this may reflect our lack of knowledge, there is some supporting evidence. The appearance of new-style units in at least twelve of the northern forts in the Notitia Dignitatum led John Mann to argue that the units previously based at these sites had been disbanded or moved elsewhere during the third century, suggesting peaceful conditions. It was the rise of the Picts in the fourth century which led to more units being sent to the island and the earlier pattern of occupation renewed.3 Three explanations have been advanced to explain why the same basic network of military deployment was retained in northern Britain for over three hundred years: that the tribes beyond the frontier remained troublesome throughout the Roman period; that the Brigantes within the province remained subversive; and that the pattern largely related to the geography of Britain. Our literary sources provide many examples of disturbances on the northern frontier of Britain from the 80s through to the reign of Severus in the early third century (pages 80–81, 148 and 199), and certainly a strong army was retained on the island. It may be, however, that the substantial army was retained because the Caledonians were unconquered rather than that they were unconquerable. The location of the cavalry regiments in the area of the two main routes north from Hadrian’s Wall, the placing of other principal units on the main roads to the south together with the reoccupation of earlier forts and the construction of new forts in the vicinity of the frontier following the return to Hadrian’s Wall in the 160s all lend support to the proposition that the main concern of the Roman army was the protection of the province from attack. Most of the new-style units sent to Britain in the fourth century were also, where we know their locations, based on the main roads leading north from Eburacum/York to the frontier. The supposedly continuing opposition of the Brigantes to Roman rule depends not upon any statement in ancient literature, but upon the continuing occupation of

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 163

The Deep Frontier: Defence-in-Depth? 163 forts in their territory: this comes close to a circular argument. We may assume that if the Brigantes remained troublesome, it was those living in the Pennine Hills who caused the disturbances as it would have been easier to control the farmers of the lowland areas. Yet the units based in the forts within the hill country of the north were nearly all smaller than those placed on the main routes leading north to the frontier, each only of a size for the smallest auxiliary unit. We may surmise that the role of the soldiers stationed in the forts on the cross-routes was more likely to have been the maintenance of communications along the road network rather than the control of unruly locals. Some forts in the area have annexes lying to one side, but their function is not certain.4 Brian Dobson has suggested that it is the geography of Britain which led to the particular nature of the military deployment.5 It was the failure to complete the conquest of the island, coupled, it would appear, with the continuing intention to complete the task, which led to the retention of the large number of regiments in the island, though, in addition, the fact that Britain was an island and its army therefore more difficult to reinforce, may have played a part. But the narrowness of the island at the two isthmuses would not allow the employment of all the army units along this line and this forced the substantial provincial army to be spread across a deep zone to the south. This is not ‘defence-in-depth’, merely adaptation to the local geographical situation.

Discussion Britain was a special case, the size of the army being determined by the failure to complete the conquest of the island and the intransigence of the enemy, while the army was perhaps kept at an unnecessarily high strength owing to the difficulty of reinforcing it in an emergency. It was manifestly impossible to place all these troops on the narrow frontier line and they were therefore spread across the countryside. Britain being relatively narrow, the zone of military deployment stretched far south from Hadrian’s Wall.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 164

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 165

Part III

Interpretation

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 166

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 167

Chapter 14

The Development of Frontiers

s we have seen, in the Republic and indeed under Augustus, military frontiers to the Roman empire did not exist. Under Augustus, some legions were based within the interior of provinces, as resistance to Rome often continued for many years. Caesar may have boasted that he had conquered Gaul. Augustus knew better. His henchman Agrippa was sent to Gaul as governor in 39 BCE and he campaigned in the north-east and south-west, and there was further campaigning in the south-west ten years later. Legions continued to be based within Gaul for at least another forty years. It appears that it often took some time for the new provincials to settle peacefully into their new role. It was frequently the first generation after the conquest that made one last effort to throw off the Roman yoke. In Numidia, there was an uprising soon after annexation. In Britain, the Roman armies came closest to defeat under Boudica in 60/61, eighteen years after Claudius’ invasion. A similar situation occurred in Thrace following its annexation by the same emperor. It was only when the new province was pacified that the army was moved from the interior onto the frontier line. This process took nearly a hundred years to complete in Noricum and Pannonia. It is not always easy to know whether this process took so many years because of genuine internal disturbances, or if the reason related to inertia. It is clear that emperors varied in their interest in provincial policy and in frontiers and even if an emperor was interested, a crisis on one frontier may have led to a lack of action elsewhere. In such circumstances the internal affairs of some provinces might be left to languish. The Varus disaster of the year 9 was of enormous significance for the Roman Empire. Nearly all troops were pulled back across the Rhine and settled into bases to wait for the order to move forward again and reoccupy Germany, an order which never came. Gradually the large army groups were broken up and units spread along the river. This was a long-drawn-out process. It was not until 89 that the last grouping of two legions in one fortress was abandoned along the Rhine, but that was nothing to do with frontier control. The rebellion of Saturninus was funded by the savings of the two legions and the separation of the two legions was to help prevent such an event happening again. During this process, the spacing between forts along the rivers Rhine and Danube was reduced from an earlier erratic framework to about 30–40km (18–24 miles). On other frontiers, deserts or mountains, the distance between forts was often greater, though long gaps were usually broken by fortlets and towers. In some places, gaps remained because the army perceived no need to place forts where they were

A

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 168

168 The Frontiers of Imperial Rome

(a)

(b)

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 169

The Development of Frontiers 169 Fig. 48: Plans of fortlets dating to the early empire in Britain and Germany, a. Burlafingen (Germany), b. Martinhoe (Britain), c. Rötelsee (Germany), d. Barbugh Mill (Britain).

(c)

(d)

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 170

170 The Frontiers of Imperial Rome unnecessary. Topography such as a deep gorge, or a flood plain, the lack of threat from across the frontier, or a thickly wooded landscape might explain these gaps. But over the following decades, they were filled, as in Noricum. There were other areas where there was a greater concentration of troops. These were generally because of an actual or perceived threat from beyond the frontier, the necessity to control a route or people living close to the frontier or a particularly fertile region. We generally assume that the spreading of units along the border was in order to help maintain the empire’s security, but another reason may have played a part, supply.1 It may have been easier for regiments to gather local supplies if they were more evenly spread across the landscape. This factor may have been partly instrumental in the breaking up of the large army groups of the time of Augustus and Tiberius and the distribution of units along the rivers (figs. 24 and 28). The next most significant change was the greater use of towers and fortlets on the frontier. Towers appear in Upper Germany under Augustus, and along the lower Rhine under Claudius (pl. 10). Tiberian fortlets are known on the upper Danube (fig. 47a). The mere existence of these structures suggests that there are more to be found. Our first evidence for their use in a more concentrated form on frontiers is in Britain in the late first century and shortly afterwards in Germany. In Britain there is still controversy about the precise function of such towers, but in Germany it is more clear that their purpose was observation along or across a frontier line. Towers appear on Trajan’s Column, but none of such early date have been identified on the banks of the Danube (pl. 5). The creation of linear barriers was the next innovation (figs. 11, 14 and 18; pls. 11, 14, 15 and 17). This is normally attributed to Hadrian (pl. 2), but a short length of barrier constructed in Germany appears to date to the reign of his predecessor. It should also be noted that linear barriers were in the Roman military portfolio: for example, Crassus sought to control Spartacus by the construction of a wall and, Caesar had constructed a linear barrier to channel movement in southern Gaul.2 Nevertheless, it would appear that it was Hadrian who developed this frontier element. Such barriers were constructed in Germany and Britain and when both were abandoned for a new forward line, they were replaced with similar structures. The barrier in Germany was repaired and rebuilt, but continued in use until that part of the empire was abandoned about 260; the linear barrier in Britain continued in use to the end of the empire. As these land frontiers were established, the distance between forts tended to be reduced to about 11km (7 miles). It is certainly true that there are many variations, but, where all factors are equal, that distance frequently occurs. This distance is half-a-day’s march and one can see how the shorter distance arose. In northern Britain, many forts are 22km (14 miles) apart, but on the frontiers this distance is reduced. When the distance between forts in the frontier line is less, the number of troops based at each is generally smaller. In Germany, typically, small forts were often only 5km (3 miles) or so apart, while a similar situation pertained on the Limes Transalutanus in Dacia. On the Antonine Wall, the distance between most forts was reduced to 3km (just over 2 miles), but this was unusual and not repeated elsewhere.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 171

The Development of Frontiers 171 The Antonine Wall was perhaps the most developed frontier, with its linear barrier, close spacing of forts, fortlets and small enclosures, and unique expansions/ beacon-platforms. Its complexity was not repeated on another frontier. Later in the second century, however, it is noteworthy that there are other developments. Under Commodus towers were erected on the Danube and fortlets in North Africa, both concerned with the protection of the frontier areas from raiders. Severus stepped beyond the frontiers he had inherited (pl. 1c). He sought, and obtained, new conquests in the East, though not all he wished. He advanced the line of forts in Mauretania Caesariensis, constructing a ‘new frontier zone’, and built new forts along the northern edge of the Sahara Desert in both Numidia and Tripolitania (figs. 38 and 42; pl. 27). In Britain he set out to complete the conquest of the island, according to the contemporary historian Cassius Dio, and was only prevented from achieving his ambition through his early death. His son abandoned both territory and forts, one of which was on the River Tay, some distance downstream from the first-century legionary base at Inchtuthil. His death in 211, or perhaps the end of his dynasty in 235, marked a watershed and Rome was never in the same position to take such offensive operations across her borders. Both the literary sources and the archaeological evidence, however, demonstrate that Rome did not withdraw within her frontiers never to venture beyond them again. The construction of forts and fortified landing places on the far side of the Rhine and the Danube as well as the construction of stone bridges across both rivers and the erection of bridge-head forts indicate a desire to intervene beyond the empire, but whether they were provision for counter-attack or further conquest is more difficult to determine. Literary sources certainly show a willingness to take the war to the enemy whenever it was strong enough to achieve one or both. Sometime towards the end of the third century, new types of forts came to be constructed (figs. 10, 32 and 46). They are characterized by high and thick walls, external bastions and small, well-defended gates. The reason for the building of forts with high and thick walls to protect the army against an enemy which mostly did not possess the equipment for siege warfare remains problematic. Many of these forts are smaller than their predecessors, often markedly smaller, and were frequently further apart. Their appearance coincided with the construction of new walls round the city of Rome under Aurelian in the 270s and this may not be coincidental. It is, however, not until the reign of Diocletian that we can see a thorough campaign of improving frontier installations and the wholesale adoption of this style of building. On many frontiers this entailed adding to the defences of existing forts as well as building new structures, but on the Eastern Frontier the work appears to have extended to the erection of a string of forts and fortlets for 800km (500 miles) from the Gulf of Aqaba to the Euphrates. Some of Diocletian’s successors such as Constantius II, Julian and Valentinian were energetic emperors and undertook extensive building programmes. The tools were as before, forts, fortlets and towers. Noteworthy was Valentinian (pl. 1b) who ordered an extensive programme of tower building from Britain to the eastern frontier in the 370s.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 172

Chapter 15

Military Deployment

he installations on the frontiers, their size, location, type, spacing, distribution and the units based there, have much to tell us about how frontiers operated. All these factors are better interpreted when related to the landscape in which the frontiers sat. Indeed, the landscape was the major element in the distribution of soldiers and the nature of the frontier works. The landscape affected where people could live, where they could produce food, and the need to maintain observation of such people and places was a major factor in the location of forts. Particularly in the early empire we can see that there were no forts where there were no people. People also travelled and therefore the control of routeways was important. As we have seen, a significant factor determining the location of the legions was the existence of major routes across the frontier. Legions were placed so as to be in a good position to repel invasions such as the legions grouped round Antioch under Augustus and again in the fourth century, or guard significant routeways, such as the Amber Route from the Baltic Sea which came into the empire at Carnuntum/Petronell. In some places, cavalry units were bunched round these legionary bases. Five lay up or down the Rhine from Vetera/Xanten at the end of one of the major lines of penetration into ‘Free Germany’ along the River Lippe. On the Danube, Brigetio/Szôny was flanked to the east by one cavalry unit, while a second lay across the river. In those provinces where there was no legion, or the legion lay well back from the frontier line, a cavalry unit would appear to have undertaken the same role. In Noricum, a 1,000–strong cavalry unit served in place of a legion, until one was stationed there in the 170s. In Upper Germany, a cavalry unit was placed at Echzell, where a main route from Mogontiacum/Mainz into Germany crossed the frontier. It was supported by a small infantry unit and a 1,000–strong mixed unit a little behind the frontier at Friedberg, though the fort lay at a nodal point in the road network, the units therefore being able to support the soldiers in the forts in the arc to west, north and east (fig. 11). It is possible that these three units were so placed so as to be able to operate as a single task force. In Britain, we can see from the several inscriptions dating to the early third century that several cavalry units were focused on the two main routes which led northwards through Hadrian’s Wall, on Dere Street in the east and at Stanwix to the west, where the 1,000–strong cavalry unit was based.

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 173

Military Deployment 173 The work horse of the frontier was the mixed infantry and cavalry unit, 480 infantry and 120 cavalry. This was the most common unit of the Roman army, which must have seen the combination of infantry and cavalry as particularly useful. The location of the other units, that is the cavalry regiments, infantry cohorts and the 1,000–strong units, requires explanation. Some have been offered above. Other factors, however, might affect the uses of infantry rather than cavalry. In the Rhine delta and on the Antonine Wall the reason for the lack of cavalry may lie in the wet nature of the terrain which was not good cavalry country. The desert required a different mount and accordingly here there were units of camel riders and sometimes camel riders attached to other units. In several provinces, an auxiliary unit was placed close to a legionary base. This occurred at Inchtuthil in north Britain and at Carnuntum/Petronell in Upper Pannonia, for example (figs. 18 and 28). This might imply two distinct functions for the two types of units, the legion having a more strategic role with the auxiliary unit deployed tactically. Dietwulf Baatz has suggested a similar arrangement on the Outer Limes in Upper Germany.1 He has argued that when two units are based together, albeit in adjacent forts, the role of one was frontier control, while the other was more strategically based with the facility to operate more widely as might be required in the frontier area. A similar doubling of units has been recognized in Dacia, though there the smaller structure tended to be a fortlet rather than a numerus fort. This arrangement has to date been little recognized and requires further research. In Britain the same separation of functions can be observed in the first plan for Hadrian’s Wall where the forts remained at some distance from the linear barrier (fig. 23). Mountains particularly focused attention on routes, in this case through passes. In Dacia and the land frontiers of Germany, soldiers were located to ensure Roman control of the passes (fig. 41). This concern extended to mountain ranges beyond the empire, for Rome always showed a keen interest in the control of the routes through the Caucasus Mountains in order to protect the eastern provinces from the depredations of the Alans and other peoples to the north. In Mauretania Caesariensis, the strong east–west lines of the Atlas Mountains together with the narrowness of the province forced a parallel east–west distribution of the single line of forts (fig. 42). Military deployment in the Atlas Mountains is also a useful reminder that forts had to be placed within the most sensible locations in such terrain (pl. 1d). This was not on the tops of mountains, but rather in the valleys. The line of forts therefore did not in itself constitute the frontier. In the desert regions, wadis served as communications routes and these required controlling. Sometimes, a single fort or outpost might be sufficient, but in Tripolitania and Numidia barriers were erected across lines of movement, presumably with the same purpose as barriers elsewhere. On the river frontiers, the nature of the river valleys as well as the location of tributaries affected the positioning of forts. Tributaries formed routes which needed guarding, while their mouths offered safer anchorages than the main river or the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 174

174 The Frontiers of Imperial Rome open sea. Accordingly, on river frontiers, spacing tended to be more irregular. Rivers did not flow in an even way but could pass through gorges or along flattish stretches. Within gorges, there tended to be no forts, at least in the early period, but forts were located at each end of such sectors and at each end of the sluggish sectors also. On the Euphrates, the legions were based at points where access across the gorge was possible. Water also affected the location of forts in the more arid parts of the empire. Forts in deserts tended to be placed at oases, both for the water available there as well as to deny their use to anyone else and maintain surveillance over the people living there. Where there were no oases, digging of wells and the creation of cisterns allowed a different pattern of military deployment. Water was still the basic requirement and along the eastern frontier and in North Africa the outermost forts tend to lie on the furthermost points of settlement allowed by rainfall. Along the line of the road stretching northwards from the Gulf of Aqaba to Soura/Souriya on the Euphrates the normal distance between the forts and fortlets was 20–25km (12–15 miles), the precise location of each site depending upon the local availability of water (figs. 34 and 35). Although there has been much discussion about the purpose of these forts and fortlets, their spacing is of little help for it is the same distribution as on most frontiers. Most archaeologists studying the Fossatum Africae in Numidia, which consisted of separate lengths of rampart and ditch backed by forts, fortlets and towers, agree that their purpose was to control transhumance. The location of military installations in these circumstances related to a specific point or line, that between the desert and the cultivated land. This interpretation of the military installations has been extended to the frontier in the neighbouring province of Mauretania Caesariensis. Here the construction of a line of forts had the purpose, it has been suggested, of maintaining watch over the nomads and helping to control their movement onto the grazing lands within the province, an ingenious argument which may bear little relationship to reality. Food to feed the army was probably also a factor in the location of units. In northern Britain, it has been noted that forts were generally located close to good farmland.2 This of course is also where people lived so the argument could easily become circular. But the cost and difficulty of transporting food in bulk will have led to a wish to undertake local purchases and the wider distribution of troops across the landscape will have aided those actions. The reverse of this pattern of deployment is the lack of forts in certain sectors. This has been linked above to the paucity of settlements in these areas or the absence of a threat from beyond the frontier. I have, however, previously offered an alternative possible explanation for the greater concentration of troops on both Hadrian’s Wall and the Antonine Wall in those sectors where we believe that few people lived north of the frontier.3 This is that normally, we can presume, local chiefs would have kept the Roman army informed of disaffected elements and of incoming malcontents for fear of Rome’s retribution. This may have reduced the necessity for the army to undertake patrols north of the frontier in the populated

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 175

Military Deployment 175 areas. In the sparsely populated countryside though, only the army could maintain watch and ward and this may have accounted for their greater numbers facing the open spaces of Spadeadam Waste north of Hadrian’s Wall and the Campsie Fells north of the Antonine Wall. Forts are only one element in the array of military installations found on Roman frontiers. Detachments might be placed in small forts and in fortlets. Structures which might be called small forts are used frequently on the German frontier and are also found in Britain. In Germany, small forts were often used to guard passes. In Britain fortlets occupied by eighty men or less were placed along roads between forts, presumably to help maintain order and protect movement along the road, and in positions where they could maintain surveillance, for example on the southern shore of the Bristol Channel looking across to the unconquered southern Welsh tribes. In Egypt, fortlets are placed along the main trade routes between the Nile and the Red Sea (fig. 39, pl. 28), while in Mauretania Caesariensis, Rushworth has suggested, they were located in areas were extra watch was required to protect travellers against possible hostilities.4 Fortlets were adapted to local circumstances, and one special type of fortlet is the milecastle on Hadrian’s Wall and the similarly sized structures on the Antonine Wall. On nearly all frontiers, interplay between forts, fortlets and towers can be seen. The land frontiers tend towards a regularity of dispositions: forts at about 11-km (7–mile) intervals, fortlets in between, towers 500m (1,500ft) or so apart. Yet, the same broad principles can apply elsewhere, such as Dacia, where field investigations are producing evidence for a line of fortlets and towers along the frontier running through the Carpathian Mountains, with short lengths of barrier in the passes. The pattern of fort, fortlet and tower continued through to the end of frontiers. It can be seen in the arrangements put in place on the Upper Danube following the abandonment of the land beyond the headwaters of the Rhine and Danube shortly after 260. On Hadrian’s Wall, the absence of items of furniture and box fittings from the turrets has led to the proposal that the soldiers lived in the towers for short periods, their main bases being in adjacent forts or fortlets, and it is possible that the same arrangements operated in other parts of the empire. The work of David Woolliscroft has certainly indicated an intimate relationship between the soldiers stationed in fortlets and towers and the units in nearby forts.5 The forts were linked by roads or tracks in order to provide for speedy communication; in fact, the frontier road formed a ring round the whole empire. Only in the east does it appear that a road, or a route, already existed along the line adopted for the frontier; elsewhere the roads were along new lines. Sometimes the major arterial road was supplemented by paths. Andreas Thiel has drawn attention to the existence of a road as well as a path on the section of the German frontier running through the Odenwald between the rivers Main and Neckar, describing it as ‘a large, guarded traffic route’, thereby emphasizing the importance of roads for the effective functioning of frontiers.6 On Hadrian’s Wall, too, there were local paths as well as roads along the frontier.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 176

176 The Frontiers of Imperial Rome A further common element is the existence of outposts beyond what appears to be the basic frontier line. These occur in North Africa, the Eastern Frontier and in Britain, though not mainland Europe, and in all periods. Even in the late fourth century, Valentinian was seeking to establish an outpost fort in Germany. The outposts appear to have had two purposes, to maintain watch over routes, especially potential invasion lines, and over groups of people. To a modern-day observer, many outposts appear to be vulnerable, but they indicate the confidence of the Roman army. Do they also suggest a preparedness to sacrifice soldiers based in such isolated locations? In summary, we can see that in every way, Roman forts and frontiers were positioned in order to control people and their movements. Yet, while we have evidence for the tight control of movement in the frontier zone, the Romans appear to have been prepared to allow the construction of their frontiers to drag on over many years. It seems that Hadrian’s Wall took six years or more to build, long enough for parts of it to become overgrown before the builders returned to complete the task. In that period, many sections stood unfinished. So, too, in Germany, long stretches of the palisade remained uncompleted for many years, perhaps decades. This appears to be at odds with a wish to create a framework for a 24–hour watch on the frontier.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 177

Chapter 16

A Comparison of Frontiers

his review of frontiers has brought certain factors into focus, though they do not always point in the same direction. On the one hand, Roman military installations appear to be broadly similar at the same date across the whole empire. It is somewhat surprising to find the plan of a fort built in North Africa or the Middle East to be so close to one in Germany or Britain. Towers, by their very nature, are more or less the same no matter where erected, though their plans change over time. Fortlets and small forts, however, show greater variety. Although there is a basic fortlet type, it appears to have been frequently adapted to local circumstances, as exemplified by the milecastles and milefortlets attached to the linear barriers in Britain. I have argued that this British fortlet was an innovation of Hadrian. The broad similarity of the style of forts continues into the late empire when many military structures are dated on their architectural details alone through comparison with dated structures elsewhere. There tends to be a broadly similar pattern of spacing across the empire. A ‘basic’ distance between forts is 22–32km (14–20 miles), though the spacing can be greater or less depending on the topography. On all frontiers, the longer distances are usually broken by fortlets and towers. The nature of the terrain generally determined the spacing and it is only on some of the land frontiers that a ‘standard’ distance 11km (7 miles) between forts occurs. On all frontiers there is flexibility in the deployment of troops. While there were forts for complete legions and auxiliary units, soldiers might be withdrawn for outpost duty. The size of these outposts varied considerably: one might expect a careful consideration of the military requirement of each site. Rather surprisingly in view of the strategic role of the legions discussed above, there is plentiful evidence for legionaries serving away from base. The Third Legion based at Bosra/Busra in Arabia and its counterpart in North Africa appears to have serviced many outposts (fig. 35). In northern Britain several forts seem to have had legionaries based at them. Perhaps these arrangements occurred because the army was overstretched, or possibly some legions did not have enough duties to keep their troops occupied. The depth of the frontier zone is worth exploring. In most areas, the norm seems to have consisted mostly of a single line. Forts might be held behind that line, such as the legions of Upper Germany or forts at nodal points in Mauretania Caesariensis. Some frontiers were very different, but in each case the distinction related to the terrain: Dacia, Britain and Severus’ province of Mesopotamia. Dacia

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 178

178 The Frontiers of Imperial Rome was unique in its pattern of military deployment. The passes through the Carpathian Mountains were guarded, mostly by infantry, with cavalry stationed behind and the legions based in the centre of the province (fig. 40). The arrangement suited the terrain. In Britain, it can be argued that the deep frontier zone reflected the narrowness of the island and the size of the army, which was related to the unconquered nature of the enemy to the north (figs. 14 and 18). On that section of the Eastern Frontier facing Parthia there was no close network of forts or a frontier distinguished by a line of forts, fortlets and towers. There certainly were legions placed in bases along the Euphrates, each guarding an east–west route. But when more territory was acquired further east by Severus, the pattern changed because now the army had to be based in the great oasis cities between the rivers Euphrates and Tigris. This was not a ‘frontier’ in the manner of most of the rest of the empire, nor defence-in-depth; the pattern of military deployment was a straightforward reaction to the landscape. The arrangements in Mauretania Tingitana, it has been suggested, were different from those in other provinces, relations between Rome and the neighbouring tribes being governed by treaties which in turn led to a special pattern of military deployment. However, we know little about the latter, while the many inscriptions recording treaties may simply be a matter of survival. The land barriers erected at about the same time were each distinctive. The German frontier built under Hadrian was a timber fence, though a very stout one. Hadrian’s Wall in Britain was of stone and turf, the Antonine Wall of turf, albeit on a stone base. These differences, it has been argued, are closely related to the availability of building materials and this manifests itself in the forts as well. Both the British Walls were provided with a ditch, but only the earthen bank which replaced the palisade in Germany was fronted by a ditch, which probably served mainly as a quarry for the earth to create the bank. Hadrian’s Wall and the Antonine Wall in Britain were surveyed differently. The Outer Limes in Germany was unique in being straight for 81km (50 miles). In the late empire, most frontiers were strengthened by the construction of forts in a new, more defensive style, while others were not so improved. So ubiquitous is the new style of architecture that oldfashioned frontier forts, as in north Britain, stand out as being unusual. David Woolliscroft has been able to explain the differences, and similarities, between different frontiers through his characterization of frontier lines as being either ‘terrain-following’ or ‘terrain-crossing’. Through this, the subtlety of military thinking can be observed. Care was taken to choose an appropriate style for each landscape so that the most effective arrangements for observation in the frontier zone could be put in place. The army also adopted what might be termed a mix-and-match approach. The land barrier in Germany was not continuous. The River Main separated the northern sector round the Taunus from the southern which firstly ran through the Odenwald and later was moved to the Outer Line (fig. 11). The Odenwald sector was shorter than its replacement for it used the River Neckar in its southern section.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 179

A Comparison of Frontiers 179 To the west of Hadrian’s Wall in Britain forts, fortlets and towers continued along the coast while to the east the River Tyne appears to have been considered sufficient barrier (fig. 14). Fortlets overlooked the Clyde estuary to the west of the Antonine Wall, and, to the east, forts lay on the south shore of the Firth of Forth (fig. 18). In Dacia the earthen bank of the Limes Transalutanus was only erected where there was no convenient river to form the boundary (fig. 40). The impression is that the army considered linear barriers to be only necessary when there was no other clear marker available. Finally, we should note those areas where there are no frontiers. These include the Libyan desert to the west of Egypt, Cyrenaica and Tripolitania throughout most of the life of the empire. The eastern frontier of Arabia through the early empire appears to have retained few troops. In some ways these long gaps existed for the same reason as the smaller gaps on the river frontiers: there was no need for forts where there were no threats from beyond the province. Can we identify the characteristic elements of a Roman frontier? In broad principles, ‘yes’, but so broad as to be almost meaningless because the variety is so great as the army responded to the wide range of factors which governed the disposition of their military forces and the types of structures which they were obliged to erect. There is therefore no one set of characteristics. Perhaps the clearest example of a ‘simple’ frontier is the Outer Limes in Germany: a linear barrier; towers beside the barrier; forts regularly placed immediately behind with no forts in advance nor to the rear. The Danube frontier from Castra Regina/Regensburg to the delta is broadly similar, though, of course, without the linear barrier. Elsewhere the frontier ‘line’ can be blurred by the use of advance forts or outposts or hinterland forts or land held beyond the presumed main frontier line. Where there was no land frontier or useful river, can a frontier be recognized? Nick Hodgson has drawn attention to the contrast between the Eastern and the Western frontiers and suggests that it relates to the underlying political and economic infrastructure of the two areas, and in particular the advanced urbanization of the eastern empire: ‘in Arabia, Palestine and Syria … the troops lived in cities, the region being urbanised enough to make it natural for them to do so.’2 There were no independent cities in the west in which to base troops, while ‘the points of diffusion were more numerous and less predictable than on the edge of the Syrian or Arabian desert.’ This led to the creation of artificial cities in the west, forts, which ‘supervised border lines with elaborate barriers to prevent movement in either direction, except at sanctioned places.’ Hodgson sees the eastern arrangements as natural and the western as the oddity. In the West: the troops ended up patrolling a limes which protected routes and communities that had to be protected because the troops relied on them. In the west the army essentially protected itself. The arrangement was a sort of artificial microcosm of the eastern system, in which, however, the troops policed real cities and their trade arteries rather than forts and their supply routes. In

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 180

180 The Frontiers of Imperial Rome Britain and Germany Rome engaged in large scale military building on the frontiers because these provinces lacked an urban heart and the main pattern of control lay away from the centre. The Hellenized provinces of the East attracted Roman forces to their urban centres. The north-west frontier provinces had most Roman soldiers in their outer shell, the frontier zone. Thus the frontiers became more elaborate. (Hodgson 1989, 180.) Hodgson’s conclusion, with which he does not now agree, is that what the frontiers of the West ‘are more often concerned with is local control.’3 In support of his argument that soldiers were based in the cities of the East, Hodgson drew attention to the detachments based at Petra, Gerasa/Jerash and Philadelphia/Amman, with legions in the cities of Bostra/Busra, Zeugma/Belcis and Samosata/Samsat and auxiliary units at Palmyra and Dura-Europos. The difficulties of interpreting the evidence have, however, been emphasized. A distinction must be drawn between soldiers who were based at a site and those who were merely passing through and left a record of their visit.4 Further, while part of the city of Dura-Europos was specially walled off for military use and the army at Palmyra established a base in the city, at other places there were ‘normal’ forts. The bases of the legions at Bostra/Busra and Satala/Kelkit were such forts. The major problem is that at most cities physical evidence of military occupation has not been found; no fort has been located at Samosata/Samsat or Zeugma/Belcis, for example. Nor is it easy to explain away this lack of evidence as resulting from the billeting of soldiers upon the civilian populations for there is no evidence of that either. In short, the argument cannot be pushed too far, not least because there are many different types of frontiers and here but two examples are being compared. Nevertheless, Hodgson offers a valid comparison. The great cities of the East, especially in the province of Syria, required protection and troops were either based within them or beside them: the troops followed the civilians. Across Europe, however, the reverse is the case. Here frontiers were established and the civilians followed.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 181

Chapter 17

Decision Making

odgson offers an argument for one difference between frontier arrangements in the eastern and western parts of the empire. Within such a wide framework, other reasons may operate. We have already considered the significance of topography, settlement patterns, available building materials and the personal involvement of the emperor. We now need to consider wider aspects of such personal decision making, starting with the emperor himself. The direct involvement of emperors in the details of frontier operations has already been discussed. Hadrian is the prime example, but other emperors, including Severus, Diocletian, Constantine, Julian and Valentinian, took close personal interest in frontiers (pl. 1). Lander has argued that planning of fortifications was undertaken at the imperial level, or at least that they sprang from imperial decisions relating to the frontiers.1 Emperors, however, could not be everywhere at once, nor were all as interested. How were matters normally ordered? There is ample evidence that emperors provided instructions to their provincial governors. Mandata, instructions, were issued to each governor as he took up office, though the level of detail is not known. His disregard of the instructions given to him by Tiberius formed a central part of the case against Calpurnius Piso, governor of Syria, at his trial for treason in 20 (Tacitus, Annals 3, 10). The emperor might subsequently send letters to a governor ordering a specific course of action as did both Tiberius and Claudius.2 In 139, the new governor of Britain was provided with specific orders to advance the frontier. The much later biography of Antoninus Pius stated that he ‘conquered Britain through his legate Lollius Urbicus, and, having driven back the barbarians, built a new Wall, this time of turf ’ (Historia Augusta, Life of Antoninus Pius 5, 4). A more contemporary record is the speech given by Cornelius Fronto in the Senate, almost certainly when he was consul in 142: ‘although he [Antoninus] had committed the conduct of the campaign to others, while sitting at home himself in the Palace at Rome, yet like the helmsman at the tiller of a ship of war, the glory of the whole navigation and voyage belonged to him.’ (Fronto, From the Speech of the War in Britain). This illustrates the personal element of Roman rule. An anecdote of the reign of Hadrian underlies this aspect of imperial rule. Dio recorded that a woman made a request as he passed by on a journey and when he said that he did not have time to deal with her, she called out, ‘cease being emperor then’, at which point he paused to listen (Dio 69, 6, 3). When Tiberius died, Vitellius, governor of Syria, stopped

H

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 182

182 The Frontiers of Imperial Rome campaigning against Aretas, king of the Nabataeans, who had invaded the empire, and placed his army in winter quarters because he had not yet received any instructions from the new emperor, Gaius (Josephus, Antiquities of the Jews 18, 124). When Nero died, Vespasian stopped campaigning in Judaea and sent to the new emperor for instructions, his operational orders having died with the last emperor (Josephus, The Jewish War 4, 498). Emperors called governors to heel when necessary; Claudius, for example, pulled Corbulo back across the Rhine, apparently when he exceeded his authority. In turn, governors reported their actions to the emperor, as did Agricola after winning the Battle of Mons Graupius in northern Britain in 83 (Tacitus, Agricola 39). More significant is the flow of letters between Trajan and Pliny when he was governor of Bithynia and Pontus and the detailed report which Arrian, governor of Cappadocia, made to Hadrian in the 130s. No doubt much depended upon the nature of the emperor and the confidence of the governor. Pliny referred the most trivial matters to his emperor (Pliny, Letters 10). Emperors, we can see, instructed governors, and in turn received reports from these officials, yet the levels of decision making are not clear. How much lee way was allowed to each governor? There is a difference of opinion between scholars. Millar, citing the events of Agricola’s career in Britain, has suggested that ‘Imperial “policy” could often consist of allowing legati to follow their own presumptions, until external factors, a major crisis or their own excessive activity, compelled intervention’, while others emphasize that ‘the wise governor consulted the emperor’ when considering moving beyond his instructions.3 The frontiers may go some towards helping answer this question. Frontier installations varied from one province to another, even adjacent provinces. This can be seen most clearly in Upper Germany and Raetia. In the former, the earlier palisade was replaced by an earthen bank, while in the latter by a stone wall. In North Africa, there also appears to have been separate development of frontier installations in the neighbouring provinces of Numidia and Mauretania Caesariensis with little attempt to integrate actions. It would appear that governors had a certain freedom within a broad policy framework, and perhaps were too frightened of imperial reaction if they tried to coordinate their actions. We can also discern a division at the next level down, that is, within each province. Individual legions possessed their own officers in charge of building operations, surveyors, architect-engineers and so on. The stone section of Hadrian’s Wall was planned to be 10 Roman feet (3m) wide, the turf section 20 Roman feet (6m) wide, the berm probably 20 Roman feet (6m) wide, and the ditch possibly 30 Roman feet (9m) wide. Gates protected by small fortlets were placed at every mile (milecastles) with two towers between. The assumption must be that there was an overall directive for the construction of Hadrian’s Wall. It is also clear that within these instructions the legions drew up their own plans. Differences between milecastles and turrets and even sectors of the Stone Wall erected by each legion have been noted, though these differences are slight.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 183

Decision Making 183 It might be expected that each governor would have an archive on which he could draw to help him take decisions. There are several examples of copies of letters surviving, as at Vindolanda (pl. 8), and Roman writers quote from official letters or dispatches. It would therefore be a reasonable assumption that Arrian, for example, would have retained a copy of his tour of inspection of the Black Sea for future reference. The special report which Arrian prepared for Hadrian contained not only a list of the stations he inspected but also the distances between them. This element of the document is paralleled in the road lists which survive. These are essentially travel maps not area maps in today’s style. In 357, the Caesar Julian crossed the Rhine and built a fort in the territory of the Alamanni. Ammianus states that he repaired the fort which had originally been built by Trajan (Ammianus 17, 1, 4). If correct, the fort had been constructed 250 years previously, yet knowledge of it remained, presumably in the Roman archives. Each governor possessed a large staff of adjutants and clerks, part of whose duties was to manage his archive. Each unit in his command maintained a file on every soldier – and horse – in the regiment, and submitted a report on its manpower every year. Each governor therefore had available an accurate statement of all the men in his province and it is possible also information on individual soldiers. He certainly would have information on the officers for he was responsible for appointing some of them, the emperor being responsible for the senior officers. Amongst the many men on his staff were interpreters and men at his command, whom we might call supernumeraries, who could undertake a range of duties on the frontier. His staff, also, might include officers who had previous service in the province and would have useful information about past activities, now placed at the governor’s disposal. Agricola’s, unusually, served three times in Britain, but many senior officers served twice. These would be useful members of the governor’s council. In summary, we might conclude that governors had considerable discretion in relation to the details of frontier construction, though they operated within an overarching framework of imperial policy and the inclinations – and whims – of particular emperors as well as the architectural conventions of the time.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 184

Chapter 18

How Did Frontiers Work?

e have now reviewed the nature of frontiers and the evidence for military deployment on frontiers and drawn specific conclusions. We have also considered some of the literary and epigraphic evidence for frontiers. It is therefore now time to investigate how frontiers worked.

W

Across the frontier The frontiers did not stand alone. Throughout her history, Rome carried out a vigorous dialogue with the states beyond her borders. This might include sending out expeditions to explore these lands. Pompey in the 70s BCE explored the East, reputedly turning back when he was only three days’ march from the Caspian Sea. Later, a soldier carved an inscription on a rock on the west shore of that sea (AE 1951, 263). Augustus, as we have seen, sent expeditions across the Red Sea, up the Nile and also into North Africa. At the time that Agricola was attempting to conquer the Caledonians in the early 80s, Demetrius of Tarsus was sent by Domitian to explore the islands off the coast of Britain (Plutarch, On the disuse of oracles 18; Tacitus, Agricola 10). This complemented the actions of the invading army, but elsewhere such expeditions were carried out before the campaign. These include those ordered by Caesar before his invasion of Britain in 55 BCE and Augustus before the eastern expedition of his grandson Gaius. This information, together with the memoirs of early commanders, would have provided governors with useful information on the people beyond their provinces.1 Practically every frontier has produced evidence, either in literature or inscriptions, of treaties between the empire and her neighbours. Peace on the Eastern Frontier was normally obtained through treaties between Rome and Parthia. A series of inscriptions in Mauretania Tingitana recorded the ordering of relations between the Romans and their neighbours, the Baquates. The details of those between Marcus Aurelius and Commodus and the Marcomanni, Quadi and Iazyges in the 170s and 180s are particularly informative about the nature of such agreements (see pages 30–32). In Britain, in 197, it was recorded that ‘the Caledonians did not keep their promises and made ready to assist the Maeatae’ (Dio 75, 5, 4). In 360, the Scots and the Picts broke their promises to keep the peace, presumably a reference to treaties between Rome and their northern neighbours (Ammianus 20, 1, 1). Treaties might be for mutual support. In 69, the kings of the Quadi provided troops to support Vespasian in his bid for the throne (Tacitus, Histories 3, 5; 21). When the Quadi and the Marcommani refused to provide troops to support the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 185

How Did Frontiers Work? 185 Romans against the Dacians in the 80s, Domitian attacked them and killed their ambassadors (Dio 67, 7, 1). This occurred between the Roman victory over the Dacians in 88 and the peace agreement of the following year. The Sarmatians did provide troops for Trajan’s Dacian wars twelve years later. Rome frequently intervened to ensure that neighbouring states were ruled by friendly kings. The buffer state of Armenia was the subject of frequent intervention by both Roman emperors and Parthian kings as each sought to have his nominee on the throne. Successive emperors in the first century used members of Herod’s family to rule minor client kingdoms in the East, installing and removing them at will. Many of these kings had been brought up in Rome and were on good terms with the imperial family, not that this ensured that they retained either their thrones or their heads. Nevertheless, over the period of a century members of the family ruled over Judaea, Armenia, Cilicia and Chalcis, while its female members married into the ruling dynasties of Cappadocia, Commagene, Nabataea, Pontus and Emesa. In the second century, Antoninus Pius gave a king to the Armenians and also to the Quadi who lived in modern Slovakia. This is marked by a coin issue which shows the emperor and the king as the same size and shaking hands.2 Rome might also support her friends financially or militarily. Sometimes, she placed her own regiments in the lands of her allies to provide support. Roman troops – legionaries and a cavalry unit – were based in Armenia in the middle of the first century (Tacitus, Annals 12, 45) and under Marcus and Lucius in the 160s a Roman force was placed at Kainepolis remaining there for at least twenty years (ILS 394 and 9117; cf. CIL III 13527a). Vespasian placed a unit in Iberia (modern Georgia) and Hadrian provided the king with a cohort to help protect his kingdom against the Alans who were disrupting the peace on the frontier (SEG XX 112). On the north shore of the Black Sea, Rome also maintained friendly relations with the kings of the Crimea to the point of attaching troops from Lower Moesia for duties at Chersonesus/Sevastapol. Arrian breaks off his account of his journey of inspection of the forts along the south-east coast of the Black Sea to describe the remainder of the coast because: when I heard that Cotys, king of the so-called Kimmerian-Bosporus, had died, I decided that it was my duty to explain the voyage as far as the Bosporus to you, so that, if you were planning something with regard to the Bosporus, you would be able to plan it without being ignorant of the voyage. (Arrian, Circumnavigation of the Black Sea 17, 3. Translation by A. Liddle.) Individual governors indulged in dialogue with kings beyond the frontier. Under Marcus Aurelius, for example, a bilingual inscription records the erection of a temple at Rawaffa, beyond the southern border of Arabia, by a confederation of the Thamadeni, encouraged by the Roman governor, who ‘made peace among them’. The range of actions of governors on and across their frontiers is encapsulated in one inscription erected in the 70s at Tivoli. It recorded the achievements of Tiberius

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 186

186 The Frontiers of Imperial Rome Plautius Silvanus Aelianus when governor of Moesia a decade before. It emphasizes how the governor maintained peace by taking positive actions across the frontier: To this province [Moesia] he transplanted – and forced to pay tribute – more than 100,000 Transdanubians with their wives and children, chiefs or kings. He suppressed an incipient disturbance of the Sarmatians, although he had sent a great part of his army to the expedition against Armenia. Kings hitherto unknown or hostile to the Roman people he brought to the bank which he guarded, to honour the Roman standards. He restored to the kings of the Bastarnians and Rhoxolanians their sons, and to the king of the Dacians his brothers, who had been captured or rescued from the enemy. From other kings he received hostages. By these measures he both strengthened and advanced the peace of the province. He also dislodged the king of the Scythians from the siege of Chersonesus, which is beyond the Dneiper river. He was the first to add to the grain supply of the Roman people a great quantity of wheat from that province [Moesia]. (ILS 986. Translation by N. Lewis and M. Reinhold.) One method of maintaining friendly relations with neighbouring states does not appear to have been used by Rome, and that is marriage. One reason is presumably that a Roman citizen could not ‘contract a legally valid marriage with a foreign woman’.3 Unusually, Caracalla sought the hand of the daughter of the Parthian king Artanaus in marriage and used refusal as an excuse to invade the country. Rome rather sought to maintain the support and obedience of foreign kings in different ways, not just by subsidies, but by using them and their families as hostages. The list of foreign kings, princes and princesses who spent time in Rome includes many of the scions of the eastern kingdoms, including Armenia, Judaea, Iberia and Parthia. This form of diplomacy was certainly an important part of the frontier policy of Rome. Roman support might be gained or reinforced by subsidies, or bribes depending on your point-of-view. There were generally three reasons for this: to support or even buy an alliance for military help against an enemy; to obtain immunity from attack; to foment divisions within Rome’s enemies. Subsidies had a long history. Caesar gave gifts to Ariovistus to guarantee his neutrality. Tiberius gave subsidies to Marobuduus the powerful king of the Marcomanni, and Claudius to Italicus, king of the Cherusci, nephew of Arminius who had humiliated Rome at the Battle of the Teutoburg Forest in 9. In order to maintain peace with the Dacians in the 80s, following two severe defeats, Domitian provided subsidies and technical aid. Trajan gave money to the neighbouring Roxolani on the Black Sea coast. Hadrian ‘purchased peace from many rulers’. In the 170s Marcus Aurelius used subsidies to encourage tribes beyond the frontier to attack Rome’s enemies, that is to divide and rule. In 197, Virius Lupus, governor of Britain, did not have the military resources to face a threat from the Caledonians and Maeatae beyond the province, ‘so he had no choice but to buy peace from the Maeatae for a considerable sum, receiving back

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 187

How Did Frontiers Work? 187 a few captives’ (Dio 75, 5, 4). Several coin hoards north of the Antonine Wall may have derived from this, or similar, payments. Dio stated that gold is paid each year to the Armenians by the Romans. As the empire was weakened through internal strife in the third century, such payments were used more as a device to obtain peace for the empire from her dangerous enemies. This procedure continued to the last years of the empire. Gratian and Theodosius both gave ‘gifts’ to the Goths, the latter on an annual basis. Even Attila took gold from Rome.4 Rome might go further and enforce her own nominee as king over a neighbouring state. A new king of the Quadi coming to power following a coup did not trouble to obtain the approval of Marcus Aurelius and was captured, deposed and exiled to Alexandria. Caracalla also executed Gaiobomarus, king of the Quadi and in extremis Valentinian sought to assassinate enemy kings, such as Vithicabius of the Alamanni (Dio 78, 20, 3; Ammianus 27, 10, 3). When all else failed, Rome might intervene militarily. Julian’s soldiers killed everyone they found, men and women alike, without distinction of age, when they surprised the enemy on an island in the Rhine, returning laden with booty. Later the same season, 357, they attacked the Alamanni, plundering and burning farms (Ammianus 16, 11, 9; 17, 1, 7). These actions are within the framework of normal Roman policy towards those who resisted them. If the enemy submitted, they would be treated fairly. If they then rebelled, they would suffer severe treatment, the extermination of the Ordovices mentioned by Tacitus, of the Caledonians ordered by Severus, of the Alamanni by Caracalla (Tacitus, Agricola 18; Dio, 76, 15, 1; Acta fratrum Arvalium, 197). Such actions are seen on the Column of Marcus. Scenes show the burning of enemy villages, though in this case during warfare deep into the enemy’s territory.5 It would be unwise, however, to envisage the Roman world as one in which there were relations between equal sovereign states. That was only true for the Eastern Frontier, and even there Rome saw herself as the superior power. Elsewhere, Rome intervened across the frontier at will. She appointed and deposed kings, or even incorporated their domains into her empire. She sought to force her neighbours to agree to neutral or empty zones beyond her own frontier where they could not settle. She maintained land for the use of the army beyond the Rhine, which she saw as the boundary of her empire.

Control on the frontier and in the frontier zone Having reviewed the archaeological and literary evidence, can we determine how frontiers worked? It seems sensible to tackle this within a tripartite framework, looking at the frontier line itself, then the area in front of the frontier, and finally the hinterland. The frontier One duty of the soldiers based on the frontier was to supervise movement across the boundary of the empire and within the frontier zone. There was no police force

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 188

188 The Frontiers of Imperial Rome in the Roman world and so many duties which today would fall to the civilian authorities to deal with were then undertaken by the army. The statements of Tacitus in relation to the Tencteri and the Hermunduri in the late first century suggest that normally people living outside the empire could only enter at specified places and thereafter could only proceed unarmed under military escort (see page 32). The treaties between Rome and the Marcomanni, Quadi and Iazyges in the 170s and 180s are remarkably similar. The conditions imposed by the Romans included restrictions on the right to attend markets, the location and regularity of assemblies and that such meetings should be supervised by a centurion (Dio 72; 73; see pages 30–31). The citing of regulations affecting both the Rhine and the Danube frontier suggests that is likely that Rome sought to impose such regulations on all frontiers. These statements receive support for the role of the army in the control of movement within a military zone from inscriptions and ostraca found in the Eastern Desert of Egypt. An inscription erected in 90 by the local army commander on the order of the governor of the province, sets out the tolls payable by travellers (IGR I 1183). Among the people listed on the inscription are sailors, sailors’ women and prostitutes, who had to pay the highest tariff. Excavations at the quarries of Mons Claudianus in the same area have led to the discovery of passes for travellers. They were for individuals or groups, including children and animals, and were instructions to the army officers along the route: ‘Q. Accius Optatus, to the four curators of the praesidia of the Claudianus route. Let pass Asklepiades.’ Sometimes the officer to whom the pass is addressed is named: ‘Valvennius Priscus, centurion, to Iulius, curator of Raima, let pass…’ (O. Claud. 48 and 49). The praesidia were army posts manned by a few auxiliary soldiers generally under the charge of a junior officer below the rank of centurion and often termed a stationarius. The known passes have been found at their place of issue and are therefore presumably copies of those given to travellers. The taking of copies appears to have been commonplace in the Roman army, as demonstrated by the Vindolanda writing-tablets (pl. 8) and the issuing of diplomas.6 The construction of observation towers helped the soldiers to maintain watch over the frontier and ensure that such regulations were followed. Towers were a significant element on all frontiers through four centuries. The inscriptions cut in Lower Pannonia in the late second century specifically state that ‘Commodus fortified the whole of the stretch of the river bank with towers built from the ground up, and with garrisons stationed at suitable points, to prevent surprise crossings by bands of brigands’ (ILS 8913). At the same time, fortlets were established on the main road through Mauretania Caesariensis ‘for border guards, between the two routes, for the safety of the travellers’ (CILVIII 2495). In 244–6 a fortlet was erected on the frontier in Tripolitania on the orders of the governor to ‘close a route used by raiders’ (IRT 880). In 256, a tower was built at Montana in Lower Moesia, 45km (28 miles) south of the Danube, for the protection of both soldiers and civilians (CIL III 12376). Trajan’s Column depicts towers on the banks of the Danube, though

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 189

How Did Frontiers Work? 189 none have been found dating to the early second century. Nevertheless, it is clear that towers were seen as advantageous in keeping watch over the river and their construction continued into the late fourth century, when, in the early 370s, Valentinian had another building campaign along the banks of the Danube, on the Eastern Frontier and probably in Britain. One source specifically states that their purpose was to improve surveillance of the crossing points on the frontier (Cod. Theod. XV, 1, 13). The need to keep watch over the river is recorded at other times. During the Batavian Revolt of 70, a drought had reduced the level of the Rhine and as a result ‘detachments were posted all along the bank of the Rhine to keep the Germans from fording it. … The ignorant regarded even the low water as a prodigy, as if the very rivers, the ancient defences of our empire, were failing us’ (Tacitus, Histories 4, 26). Nearly 300 years later, in 359, Constantius II had to place his troops in tents along the banks of the Danube when it froze over in order to watch over the Limigantes, who ‘had been planning to invade and devastate Pannonia, while the river could still be crossed anywhere because spring had not melted the snows, and while the frost would make it difficult for our troops to endure long periods in the open’ (Ammianus 19, 11, 4). The ostraca from Egypt provide vivid evidence of life in a frontier area (O. Krok. 87). A document of the reign of Hadrian records an attack by sixty barbarians on a military post at Patkoua. They attacked at 14.00 hours and the soldiers of the fort defended themselves until nightfall. The attack resumed the next day. On the first day one soldier was killed and a woman and two children taken by the barbarians – it would appear that the civilians were not allowed into the fort during the fighting. A soldier at the fort reported the incident to his superior, who passed it on to the officers in charge of the other stations along the road, presumably as a warning. Another document is a letter from a centurion to the curatores of military posts asking them to help soldiers involved in a chase, presumably of hostile elements (O. Claud. 357). These are among the most striking accounts of life on the frontier, but there are many references from elsewere in the empire to disturbed conditions in various border regions at different times. The comment by Tacitus on the propensity of the Garamantes in North Africa for raiding has already been mentioned. In the 250s, the Bavares and their allies caused trouble in Mauretania Caesariensis. This led to Gargilius Martialis being given command of a special force of two army units in the territory of the fort of Auzia in order to track down one of the rebel chiefs, Faraxen (fig. 42). The capture of Faraxen in 260 did not end the fighting: his captor was ambushed and killed (CIL VIII 9047 = ILS 2767). At the end of the same century, following the successful conclusion of military operations, Aurelius Litua rebuilt a bridge at Auzia ‘destroyed by the savagery of war’ (CIL VIII 99042 = ILS 627). It required the intervention of the emperor to end nearly fifty years of incipient warfare.7 Further east, in Tripolitania, the Austoriani raided the province in 363, ostensibly to avenge the execution of one of their fellow countrymen, Stachao, who

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 190

190 The Frontiers of Imperial Rome had been causing trouble. They twice attacked Leptis, killed local people and made off with great quantities of booty (Ammianus 28, 6). Ammianus’ descriptions of the raids of the Saracens on the Eastern Frontier have already been cited (page 124–5). The frustration felt by the Romans is clear in these accounts (e.g. Ammianus 14, 4, 1; 23, 3, 6; 24, 2, 4). In Germany, there were raids across the Rhine in the 160s. Two hundred years later, in 354, Constantius II ‘set out from Arelate/Arles for Valentia to make war upon the brothers Gundomadus and Vadomarius, kings of the Alamanni, whose frequent raids were devastating that part of Gaul which adjoined their frontiers.’ His intention in turn was to attack the lands of the Alamanni, though in the end he concluded a treaty with them (Ammianus 14, 10). In the following year he resumed campaigning against a different group of Alamanni, moving from eastern Gaul into Raetia (Ammianus 15, 4). In 358 he was in Pannonia leading a counter-attack following raiding of the Lower Danube provinces by the Quadi and the Sarmatians (Ammianus 17, 12). There was further raiding in the 370s in response to the building of Roman forts on the territory of the Quadi (Ammianus 29, 6; 30, 5, 11–30, 6, 6). On the far north-west frontier of the empire, at Corbridge beside Hadrian’s Wall, Quintus Calpurnius Concessinius, a prefect of cavalry, gave thanks after slaughtering a band of Corionototae (RIB 1142). At an unknown date, but probably in the fourth century, a soldier was ‘killed in the fort by enemies’ (RIB 3218). The fort in question was Ambleside, well to the south of Hadrian’s Wall. All frontiers suffered from raiding at various times and the Romans needed to be on the alert. Perhaps the most important work undertaken on the operation of frontiers over the last two decades has been by David Woolliscroft. He has demonstrated that frontiers were more carefully planned than previously appreciated. On Hadrian’s Wall, he has shown that in the first plan when forts remained to the south on the Stanegate, some milecastles and towers were moved slightly out of position in order, it would appear, to maintain better communication with these forts. While this relationship may be at its most clear on Hadrian’s Wall, Woolliscroft has recognized a similar close relationship between forts, fortlets and towers on the German frontiers and in Romania. This intimate relationship suggests a wish to maintain a twenty-four-hour watch over the frontier and be ever ready for an emergency. This offers us a different perspective on frontiers. It implies that they were carefully planned, with particular aims and operational objectives in mind. How long such a plan worked and how the soldiers dealt with the darkness of night, mist, or heavy rain is, however, another matter.8

In advance of the frontier Rome maintained a military watch over the land beyond the frontier. Outposts existed beyond many frontiers in all periods. The records of the Twentieth Cohort of Palmyrenes stationed at Dura-Europos in the first half of the third century are particularly informative. Two rosters survive for the years 219 and 222 (P. Dura 100

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 191

How Did Frontiers Work? 191 and 101). Although fragmentary, they show that many soldiers were on duty at as many as eight outposts, one apparently 150km (100 miles) distant from the parent base. All detachments were small, none consisting of more than a hundred men but most contained both infantry and cavalry, and several men stayed for at least three years. Several outposts lay beyond the frontier of Arabia. These included Medain Saleh, 450km (270 miles) south-east of Aila/Elat, the fort at the head of the Gulf of Aqaba, and Dumata/Jauf, 370km (230 miles) south-east of the fort at Azraq (fig. 35). These are enormous distances which partly reflect the terrain and the few oases in the Arabian peninsula. Under Septimius Severus the Syrian frontier was pushed eastwards, one city now garrisoned being Singara (fig. 34). Ammianus specifically states that this was occupied to give advance warning of attack, but the city suffered from the disadvantage of being isolated with the surrounding country lacking water (Ammianus 20, 6, 9). An inscription recorded 120km (80 miles) to the east, at an important river crossing on the River Tigris, bore a sculpture of an eagle and the words ‘the eyes of the legions’, implying that this was an outpost. Septimius Severus also established several forts beyond the frontier in the Sahara Desert; that at Cidamus/Ghadames was placed 300km (200 miles) south of the provincial boundary (fig. 38). Another lay at Gholaia/Bu Ngem. Many ostraca have been found here. These mention soldiers serving at an outpost and in watch-towers and another with the Garamantes, the tribe living south of the empire in modern Fezzan. One ostracon refers to spies and another to the monitoring of the caravan trade by the soldiers based in the statio camellariorum at Gholaia/Bu Njem (REL 53 (1973) 285). Also six outposts are mentioned on these documents. Soldiers based here presumably monitored oases and routes. Another inscription in North Africa recorded that a legionary detachment was ‘waiting in readiness’ (CIL VIII 2465 = 17953). Soldiers from the Twentieth Cohort of Palmyrenes based at Dura-Europos might be outposted for at least three years, while a cavalryman of the Vocontian Cavalry Regiment had a five-month posting at a site in the Eastern Desert of Egypt and another one of eighteen months (P. Dura 100 and 1010; P. Mich. III 203; ILS 9142 = AE 1911, 121). It is difficult to define an ‘outpost’. Distance from the main base plays a part and also size; perhaps an outpost might normally be seen as consisting of something less than a unit at full strength. On this basis, the so-called ‘outpost forts’ north of Hadrian’s Wall are not outposts; Mark Corby has called them ‘advance forts’ and this might be a better term.9 Two of these advance forts held units of scouts, indicating patrolling even further north: two inscriptions have been found 30km (18 miles) beyond the most northerly fort, hinting at the range of the scouts or perhaps the location of a more northerly outpost. The role of such men is recorded by Ammianus Marcellinus in relation to the events of 367 on the northern frontier of Britain: During these outstanding events the areani, who had gradually become corrupt, were removed by him from their positions. This was an organization founded in

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 192

192 The Frontiers of Imperial Rome early times, of which I have already said something in the history of Constans. It was clearly proved against them that they had been bribed with quantitites of plunder, or promises of it, to reveal to the enemy from time to time what was happening on our side. Their official duty was to range backwards and forwards over long distances with information for our generals about disturbances among neighbouring nations. (Ammianus 28, 3, 7. Translation by J. C. Mann and R. G. Penman.) What was the purpose of outposts? They were probably various, as an anonymous Byzantine author noted (see pages 28–9). The maintenance of watch over the approaches to the main frontier line, the boundary of the empire, was certainly one. The control of routes through the placing of outposts at oases in the desert is certainly another. The maintenance of contact with friendly kings beyond the frontier may have played a part. Discussion of outposts leads on to a consideration of patrolling. There is a general assumption that the Roman army maintained a system of patrols in the lands beyond the empire, at least in Europe. There is some evidence for this in the form of the existence of units of scouts based in frontier forts. Yet, Austin and Rankov have urged some caution in automatically assuming that patrols of Roman soldiers penetrated far beyond the boundaries of the empire. They suggest that a patrol, which would normally contain both infantry and cavalry, could only travel about 22km (14 miles) a day, 11km (7 miles) out and a similar distance back. But forts were based about that distance apart along the frontier and that would have made it difficult to cover any depth of territory beyond the frontier in detail. Further, if soldiers were to spend a night away, security would become a problem. Soldiers away from base would normally protect themselves by the construction of a camp. If regular patrolling took place, it might be expected that such temporary camps would survive, yet none are known. It is possible that the soldiers stayed overnight in friendly villages, but they would have been vulnerable in such circumstances. Austin and Rankov conclude that ‘this makes regular deep reconnaissance beyond the river frontiers inherently unlikely’.10 It is possible, however, that Austin and Rankov are being unduly cautious in view of the considerable distances some outposts lay beyond the empire’s frontiers.

The hinterland Across the whole of continental Europe, the Roman frontier in the main consisted of a single line of military installations. In Upper Germany two legions and an auxiliary unit lay behind the frontier, but the former probably stayed on the Rhine because of the advantage of the river in aiding supply. The auxiliary unit, however, was strategically placed on a route into Germany where its 1,000–strong mixed infantry and cavalry unit could move forward, help repel an invasion and give support to the units in the region. Everywhere, as far as we can see, civilian

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 193

How Did Frontiers Work? 193 administration extended up to the frontier itself. Dacia and Britain were both special cases each with their own unique military deployment. The invasions of the later third and fourth centuries prompted not only the greater defence of forts, but also of towns. Communities whether military or civilians who found themselves within the reach of the invaders protected themselves. In some frontier areas, new forts and military supply bases were built behind the frontier, but this cannot be characterized as defence-in-depth. It was simply doing what the army had done for centuries. There was a military supply base at Rödgen in the Wetterau built under Augustus apparently to support the campaigns into Germany. At Arbeia/South Shields at the mouth of the River Tyne in north-east England was another supply base, this one built under Severus, probably not just to support his campaigns against the Caledonians and Maeatae fought between 208 and 211, but more probably the subsequent occupation of their territory which he planned. The fortified granaries erected at Veldidena/InnsbruckWilten in the mid-fourth century is another manifestation of the same action. However, there was certainly a greater need in the late third and fourth centuries to ensure the protection of the army’s supply lines, especially as Rome’s enemies might, at times, roam freely within the empire. As a result, more protection was given to the movement of supplies along roads, through, for example, the construction of fortlets and towers behind the frontier in the Lower Rhineland and in the upper reaches of the Danube. In the desert areas, the location of rivers and oases governed military deployment. This might give the appearance of defence-in-depth, for example to the zone of contact between Rome and Parthia/Persia, but it was no such thing; it was merely the obvious response to the topography.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 194

Chapter 19

The Purpose and Operation of Roman Frontiers

any explanations have been offered for the function of Roman frontiers. Mommsen drew a distinction between the German limes and the British frontiers, suggesting that while the form of the construction of the former suggested that its purpose was frontier control, the nature of the latter indicated that defence was more in the minds of the builders.1 Some archaeologists maintain a defensive military purpose for Hadrian’s Wall, a view reinforced by the discovery of pits for obstacles on the berm of both British frontiers. Historians and archaeologists working in Arabia debate the function of the forts there; was their purpose policing and road-security or defence? A review of the military remains in North Africa has led to the proposal that ‘the Romans built the fossatum mainly to facilitate the collection of taxes on goods brought north towards the Tell’.2 Others have seen one purpose of frontiers as being to prevent people leaving the empire. Having reviewed the literary and archaeological evidence for all frontiers, we can now turn to consider the various possible functions of Roman frontiers. The first purpose to be considered is defence. The range of comments in Roman literature from Augustus to Valentinian demonstrates that Roman emperors were concerned about the protection of their empire. The location of the legions is equally indicative. Under Augustus and Tiberius several legions were located on the Rhine ‘for protection against the Germans or Gauls’ (Tacitus, Annals 4, 5). As the main threat moved eastwards to the Danube region, the legions moved too, and new legions were raised to increase the army’s defensive (and perhaps also offensive) capacity in the region in the face of Dacian aggression. Yet, on the Rhine, there remained a strong force beside one of the main routes into and out of Germany, along the River Lippe. Here there was a legion supported by cavalry units. In the early empire, legions were placed at Antioch astride the route into and out of Parthia. Although they were later spread along the Euphrates, in the late empire a new mobile force was re-established at Antioch. In some ways, this was part of a wider process of creating new mobile field armies in the early fourth century as the army had succumbed to inertia and fossilization. Although these field armies were first associated with the emperors and their deputies, these dignitaries no longer resided in Rome but in locations close to the main threats to the empire, so the field armies too were sensibly placed in relation to defence.

M

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 195

The Purpose and Operation of Roman Frontiers 195 It is clear that defence of the empire was a primary concern of the army. Invasions, however, could only be defeated by a strong force and the creation of frontiers was counterproductive to the maintenance of such a force, a problem faced by other states into the twentieth century. Soldiers were increasingly spread along frontiers and, when a major invasion occurred, they could either be swept aside or the forts picked off individually; both are attested. Another purpose must therefore be sought for the frontiers and their installations. The evidence presented above indicates that raiding and banditry were problems on many frontiers. Sometimes, this could be so serious that the local fort commander was killed, but generally it was probably on a smaller scale. For the inhabitants of the frontier regions, such raids were, on a day-to-day basis, presumably more significant than large-scale invasions. The Roman attempts to hinder and prevent such raids would form a strong reason for the construction of the panoply of forts, fortlets, towers and barriers which characterize frontiers, a detailed and often integrated framework to protect the provincials in the frontier provinces and such an explanation is of course supported by the inscriptions in Lower Pannonia recording the creation of towers and forts to stop the clandestine crossings of brigands. We have also seen that the Romans sought to control access to their empire. There is considerable evidence that they sought to enforce regulations which governed entry to their empire and movement in the frontier zone. Tacitus provides literary evidence in relation to the German frontier and Dio on the Danube, while ostraca from North Africa provide examples of control of movement in the frontier zone. These duties were undertaken by soldiers who were based in forts, fortlets and towers; there was no other known force available. The significance for the Romans of this aspect is underlined by the bureaucratic nature of the Roman army; we are discussing an organization capable of issuing receipts in quadruplicate (P. Amherst 107). This single fact is one of the most important items of evidence for the purpose of Roman frontiers. The archaeological evidence may help us determine the weight to be offered to each of these proposals. It may be of little help in distinguishing between the prevention of raiding and the enforcement of regulations, but it should be possible to distinguish between military defence and local control. At first, modern commentators saw all frontiers as defensive. This was so obvious, it was rarely stated in print. John Collingwood Bruce, the great interpreter of Hadrian’s Wall in the second half of the nineteenth century, offered a useful comment: ‘the Roman Wall … is a great fortification intended to act not only as a fence against a northern enemy, but to be used as the basis of military operations against a foe on either side of it’. Bruce was acknowledging a separation of functions, local protection and more significant military activities.3 The great step forward, however, came two decades later. The publication of Theodor Mommsen’s The Provinces of the Roman Empire from Caesar to Diocletian in 1885 was to change the earlier simplistic approaches. Mommsen discussed the reasons for the differences

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 196

196 The Frontiers of Imperial Rome between the German and British frontiers and between adjacent German frontiers, making several significant observations. He considered the German frontier, stating that ‘occasionally the structure may have been used directly for purposes of war. But the proper and immediate object of the structure was to prevent the crossing of the frontier’. He suggested that ‘the fact … that watchposts and forts were erected, not on the Raetian but on the upper Germanic frontier, is explained by their different relations to the neighbours, in the former case to the Hermunduri, in the latter to the Chatti’. In comparing the German and British frontiers, Mommsen said, ‘These extensive military structures [i.e. the German frontier installations] had not, like the Britannic wall, the object of checking the invasion of the enemy. … The Romans in upper Germany did not confront their neighbours as they confronted the Highlanders of Britain, in whose presence the province was always in a state of siege.’4 In Britain, a major challenge to the defensive nature of Hadrian’s Wall came from R. G. Collingwood in 1921.5 Collingwood used both the literary sources and the newly emerging archaeological evidence to argue his case. He offered several arguments to support his case that Hadrian’s Wall could not have been defensive in purpose and these are worth listing as they are relevant to all Roman frontiers. • The Roman army was an offensive not defensive force and it preferred to fight in the open; when Hadrian’s Wall came under attack, its soldiers would march out to confront the enemy in the field. • Its soldiers were not appropriately equipped with defensive weapons: ‘the auxiliaries were armed in the ordinary Roman fashion, with the pilum and gladius; and though a heavy throwing-spear like the pilum would be useful to throw off a wall at a Caledonian, each man was only issued with two of them, and their use was to give a kind of “preparation” for a charge with the short sword. A man on the top of the Roman Wall who had thrown his two pila and was armed with nothing but a short sword would be simply put out of action.’ • The Wall was not built in a manner to aid a defensive purpose: the rampart walk offered ‘a very narrow fighting-front’, thereby also creating difficulty of movement, while the turrets provided insufficient access points. • The depiction of turrets on Trajan’s Column ‘makes it quite certain that turrets of this type were signalling-stations and not emplacements for artillery’. • It is unlikely that the Wall as a defensive obstacle could have been adequately manned by the number of soldiers available. Collingwood concluded that: the continuous wall, or fence, or ditch … in its origin … serves a different purpose from the line of forts; for while the latter contain bodies of troops intended to cope with armed enemy forces, the continuous line was at first designed to serve simply as a mark to show where the Roman territory ended.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 197

The Purpose and Operation of Roman Frontiers 197 With this primary function was combined the secondary function, not always emphasized by the character of the work of being an obstacle to smugglers, or robbers, or other undesirables. A man from beyond the frontier, found on the Roman side of the line, could not plead ignorance or innocent intentions if the line was clearly marked; and if, in addition, it was a slight obstacle, crossing it otherwise than at the authorized and controlled gateways was proof of a sinister purpose. All this is abundantly clear from the actual character of the Germany frontiers of the Roman Empire, which are closely analogous to the British. (Collingwood 1921.) In spite of his view that Hadrian’s Wall was not defensive and soldiers would not fight from its top, Collingwood maintained his belief in the existence of wall-walk, which he saw as an elevated sentry-walk. It was left to John Mann to argue that a wall-walk was not a necessary element of Hadrian’s Wall.6 This was partly on the basis that, as the German palisade did not have such a feature, it was patently unnecessary on other linear barriers. But Mann’s argument was also a logical extension of Collingwood’s critique of the wall top as a fighting platform. Collingwood’s view that Roman frontiers marked the edge of the empire and had an additional purpose of frontier control has not gained universal acceptance. Modern soldiers have acknowledged that Hadrian’s Wall was not defensive,7 but not all archaeologists agree and the discovery of obstacles on the berms of both Hadrian’s Wall and the Antonine Wall in Britain in the last decade has reopened the debate. Again, it is worth stating the arguments: • Dietwulf Baatz has argued from his investigation of Roman city walls that a walk along the top of Hadrian’s Wall would have been wide enough to serve as a fighting platform.8 • Paul Bidwell has reviewed the archaeological evidence for a wall-walk on Hadrian’s Wall. He has argued that the width of the wall, the blocking of the recesses of abandoned turrets, the relationship of the replacement stone wall to the turf wall turrets, the existence of chamfered stones from a string course and the nature of the bridges all suggest that there was a wall-walk.9 • In relation to weapons, Bidwell has stated that, ‘the army had weaponry which was defensive in purpose and was trained in its use, and the weaponry is found in the turrets and milecastles, some types apparently in larger quantities than in the forts.’ Jon Coulston, however, has acknowledged that while, ‘archery was widely practised by military units’ and that soldiers were trained in the use of the bow for mural defense, yet ‘it would be an exaggeration to to suggest that virtually all soldiers in the Roman army could have practised archery at some point in their careers’, while ‘weapons stores in forts must be envisaged as having had a few bows, bundles of arrows, bundles of light javelins, perhaps shaped stones for throwing and dropping’.10

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 198

198 The Frontiers of Imperial Rome • Bidwell has pointed out that other access points to the Wall top may have existed but have not yet been discovered, while suggesting that ‘it might have been safer to have confined access to the Wall-walk to easily defensible points such as turrets and milecastles.’ Yet security does not appear to have been a concern on Hadrian’s Wall. The doors to turrets were at ground level, unlike the towers on the German frontier where they were higher, while at two turrets on Hadrian’s Wall the internal hearth partly overlay the door threshold suggesting that the door may not normally have been closed.11 • Finally, Bidwell, in his review of the new evidence provided by the obstacles on the berms of both Hadrian’s Wall and the Antonine Wall, has stated that, ‘they were intended for defence, strengthening the Walls and holding up an attack so that reinforcements could be brought up. They would incidentally have served to hinder smaller-scale or individual incursions.’12 No one appears to have challenged Collingwood’s statements that the Roman army would have preferred to fight in the open rather than defend itself from walls nor that there were insufficient troops based on the frontier to provide an effective fighting force to defend the barrier itself. The discussions about the nature of Roman frontiers, the fighting methods of the Roman army, and in particular both aspects in relation to Hadrian’s Wall has been worth setting out in some detail because they are relevant to the wider consideration of the purpose of Roman frontiers. There seems to be little disagreement within this wider framework that the purpose of the linear barriers in Germany was frontier control. It is Hadrian’s Wall and the Antonine Wall which are unusual and therefore, it is argued, had a defensive intent. My reason for their distinctiveness, however, is simply that Hadrian had a strong hand in the design of his frontier in Britain, creating a wall in the style to which he was accustomed, and that the Antonine Wall was a copy of its predecessor.13 If that is accepted, it becomes difficult to base any argument about Roman frontier policy on the nature of Hadrian’s Wall. Yet, it is clear that the builders of both British frontiers sought to make it difficult for anyone to cross the barriers. Each wall, substantial in its own right, was fronted by a ditch and in certain places the pits, which we presume contained obstacles, have been found. It seems likely that these obstacles are no more than the traps and trip wires which were added to the Berlin Wall, to name but one modern parallel, the purpose of which was quite clearly the control of the movement of people – individuals or small groups – not armies. They are another element in the unusual nature of these two frontiers, but in themselves do not prove that their purpose was different from other frontiers, nor that the enemy in Britain was more warlike than the hostile forces which Rome faced elsewhere.14 A real difficulty in interpreting a massive barrier such as Hadrian’s Wall, or indeed many of the high walls of the late forts, is that Rome’s enemies were not equipped with siege machines, nor indeed with hand weapons for attacking such structures. In Britain, for example, there is evidence for spears in Iron Age contexts,

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 199

The Purpose and Operation of Roman Frontiers 199 but not bows and arrows, though they are known earlier and do appear on the Continent at this time. However, what damage could an attacker do while standing on the upcast mound 15m (50ft) from the face of Hadrian’s Wall and 5m (16ft) below its top? Nevertheless, a powerful case can be made for pressure on the northern frontiers of Britain. There are many references both to warfare and more vague troubles. They include the tombstone at Vindolanda recording the death of a centurion in warfare sometime during the reign of Trajan or Hadrian, the decoration of a unit for meritorious conduct in the reign of Trajan, trouble at the beginning of the reign of Hadrian, a military expedition to Britain during his reign and many military deaths, troubles also at the start of the reigns of Marcus, Commodus and Severus together with warfare under Severus. The size of the Roman army may also be cited as suggesting an unsettled frontier, though it may also reflect unfinished business on the British frontier or the fact that Britain was an island and therefore difficult to reinforce. To these we can add other factors, in particular the building history of both Hadrian’s Wall and the Antonine Wall. During the building of Hadrian’s Wall, the construction of the Vallum reduced the number of crossing points from about eighty to about fourteen which led to increased military control along the frontier. It is not impossible that the construction of the Vallum was a response to warfare. The addition of at least ten forts to the Antonine Wall while it was being built also points to local opposition to its construction. Then there is the sheer size of each barrier, especially in comparison to the German palisade and its successor bank and wall, and the very substantial ditches on both Walls. The strong impression is that the army was very concerned about security. Yet, the basic point was that made by Collingwood: in the event of an invasion, the army would open the gates and march forth to fight in the Roman manner. In such circumstances, why should the Roman army wish to build massive defence works to hide behind?15 This is not the only conundrum. It is widely acknowledged – and discussed above – that the linear barriers are often not in the best defensive location, but this is of no consequence if their primary purpose was not defence against major invasions. A counter argument might be that the strength of the army allowed these barriers to be erected in non-defensive locations. It would, however, be difficult to reconcile a view which deems pits and ditches as being required against a strong foe with the linear barrier itself not being erected in the best defensive position. While it is not possible to answer all questions, the basic points seem straightforward. In my view we need to acknowledge the different roles of the army on frontiers. These were frontier defence, which was exercised by the regiments based in the forts along the frontiers and in their vicinity, and the hindrance and prevention of raiding which was also carried out by soldiers but in this case spread along the frontiers, often based in small structures such as fortlets and towers. The presence of the army on the frontier led its soldiers to acquire a new function, the monitoring and control of people in the frontier area generally as indicated by the very existence and nature of the linear barriers and the inscriptions and literary

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 200

200 The Frontiers of Imperial Rome sources which indicate their function. This separation of function can be best seen on the linear barriers, especially in the first plan for Hadrian’s Wall and the Outer Limes in Germany. These barriers served the broadly similar role of seas, rivers, mountains and deserts elsewhere. They manifestly prevented free movement in the frontier zone and channelled movement through gates where the army could impose the regulations concerning the entry of people into their empire. In Britain, the barriers were more substantial, but this may primarily reflect the involvement of Hadrian; the obstacles merely a firm intention to enforce the control of movement. Some archaeologists have seen a particular military function for both Hadrian’s Wall and the German limes. In 1940, Ian Richmond suggested that in the first plan for Hadrian’s Wall the milecastles served as sally-points.16 He saw the Roman soldiers waiting until an enemy force had come close to the Wall before soldiers issued from different milecastle gates in a pincer action to surround the attackers and pin them against the Wall and its ditch, and this interpretation was subsequently adapted for the German frontier. There are, however, problems with this interpretation. The scenario depends on coordination by the two forces undertaking the pincer movement. It also depends on the enemy remaining close to the linear barrier when discovered, and indeed being prepared to allow themselves to get into a similar situation on a subsequent occasion. Further, Roman writers do not recommend this course of action. Frontinus offered several examples of surrounded soldiers being allowed to flee, which rendered them easier to kill (Frontinus, Stratagems 2, 6), while Vegetius wrote: most people ignorant of military matters believe the victory will be more complete if they surround the enemy in a confined place or with large numbers of soldiers, so they can find no way of escape. But trapped men draw extra courage from desperation, and when there is no hope, fear takes up arms. Men who know without a doubt that they are going to die will gladly die in company. For this reason Scipio’s axiom has won praise, when he said that a way should be built for the enemy to flee by. For when an escape-route is revealed, the minds of all are united on turning their backs, and they are slaughtered unavenged, like sheep. (Vegetius, On military matters 3, 21. Translation by N. P. Milner.) The use of the linear barrier as an obstacle against which an enemy could be rolled up and then cut down should therefore be rejected. It might be considered that if defence was the primary concern, there would have been more tidying up of the location of the boundary of the empire by taking in additional territory which would shorten the borders. This did occur under the Flavians in Germany. The frontier between the headwaters of the Rhine and Danube was gradually pushed forward, with the result that the linear barrier was built across the land between the rivers. Under Antoninus Pius, the frontier was moved forward to the edge of the forest, embracing the fertile region of the Neckar valley. Under pressure in the third century, however, this whole area had to be abandoned.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 201

The Purpose and Operation of Roman Frontiers 201 The pattern is different on the Danube frontier. The Great Hungarian Plain, lying between the river and the province of Dacia, is an obvious example. It survived for 150 years as a great salient penetrating the empire. This question takes us into a discussion of how the Romans understood their world geographically. No map in a modern sense survives, only road maps. Yet the Romans did have an understanding of their world and knew, to offer a simplistic example, that a road built across the Great Hungarian Plain linking Dacia and Lower Pannonia produced a shorter route than following the frontier round. It seems likely that the Romans had concluded that the Iazyges who occupied this land could be controlled by other means, specifically treaties, as well as the latent power of the Roman army. The discussions of the detailed evidence from various frontiers offered above help illuminate the purpose of the frontiers. The first plan for Hadrian’s Wall was for a linear barrier with a minimal military presence and the army units based some distance behind the barrier. We can presume that the primary purpose of the units based behind the Wall was the defence of the province, with the role of the soldiers on the Wall being the prevention of raiding and the control of the movement of people across the frontier. When the forts were moved up onto the Wall line, the distinction became blurred. Even so, the forts give the appearance of being placed on the Wall for convenience, unlike those on the Antonine Wall which are closely related to the linear barrier. Dietwulf Baatz has argued that extra units were provided at some locations on the German frontier to act as a reserve capable of moving along the frontier line to provide support wherever necessary, leaving the task of routine watch on the frontier line to the regiments strung out along the linear barrier. A similar pattern of ‘double’ sites has been recognized in Dacia, the subsidiary installation usually being a numerus fort or a fortlet. We may see here again a distinction between the wider duty of military defence and a more local role. The same distinction between defence and local control operates today. The Berlin Wall, the Morice Line between Algeria and Morocco, the West Bank Barrier, the Belfast barriers and the USA/Mexican border restrictions were or are not about preventing the movement of armies but of individuals or small groups of people. They are the best comparanda for Roman frontiers. Their aim is to separate ‘them’ from ‘us’, as Hadrian’s biographer stated, and to help to maintain security, in short, to control access to ‘our’ space. In some areas, there is still much discussion about the purpose of the forts along the frontier. Rushworth has argued that the purpose of the forts in Mauretania Caesariensis was to protect travellers moving along the frontier zone rather than across it. The point here is that travellers needed protection, and that was not just from local bandits, who are attested in the province, but in addition from raiding parties from outside the empire, who are also recorded. In Arabia, in the second century there are few forts known along Trajan’s new road, the road which is believed to run approximately along the boundary of the province. Several military installations, including two bases for legions, were built here during the late empire. One argument is that the purpose of these forts was to protect the farmers in the province behind the frontier from the raiding which is first attested in the late third

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 202

202 The Frontiers of Imperial Rome century. The alternative view is that these additional forts represented increased policing and enhanced road security in the frontier area. It seems difficult to avoid the conclusion that the raiding which is recorded on the Arabian frontier in the third century was the reason for the construction of these new forts. Traffic in the frontier zone needed protecting, as did the people within the province. The army could undertake both duties. The desert frontiers of the empire are different from those in temperate climes because they run along the edge of the land capable of sustaining agriculture in both the Middle East and North Africa. This has given rise to debate about their relationship to nomads and transhumance. Interpretations of the role of the desert frontiers have also developed, often in relation to changing views on colonialism. French archaeologists of the earlier twentieth century saw the nomad as requiring to be civilized and, when that failed to happen, expelled from the province. Modern long-distance transhumance has been projected into the past as being undertaken during the Roman period, but certainly cannot be proved. A new model has been proposed in the form of short-distance transhumance being part of a mixed but settled agricultural regime. This new model gains some support from analysis of the tariffs recorded at Zarai and Lambaesis. Those charged on livestock were lower than those imposed on trade commodities. The role of frontiers in relation to nomads and transhumance needs further study.17 Cherry has argued that ‘the army is unlikely to have built linear barriers whose sole (or even main) purpose was to protect civilian settlements and that the fossatum was probably not intended to protect agriculturalists (or, for that matter, anyone else), against raiding activities; the Romans must have known that determined bandits could easily cross it.’ He continued, ‘the Roman frontier in Algeria and its various fortifications had two main functions: to provide for the security of the soldiers themselves; and to enable the army, and, by extension, the imperial government, to tax the products of pastoralism.’ In conclusion, ‘the different segments of the fossatum were designed to direct transhumant traffic towards the places where it could more readily be taxed.’ The proposition is supported by citing two early-thirdcentury inscriptions which record that soldiers collected customs dues. The real purpose of the frontier therefore was to help the army pay for itself.18 Cherry’s argument that the frontiers were built to protect the army is close to that of Hodgson cited above (pages 179–80), and is supported by Lander, who commented that ‘Roman stone fortifications protected Roman soldiers whose vulnerability was largely the result of their government’s frugality and inefficiency’.19 Cherry’s argument that one important purpose of the frontier, at least in North Africa, related to the collection of customs dues, was acknowledged by Eric Birley. In a significant contribution to the purpose of Roman frontiers he argued that Hadrian’s ‘artificial frontiers were not designed as fighting-positions’ and that ‘the life of a Hadrianic frontier centred more on customs and passport control’. His conclusion followed a study of the frontiers particularly associated with Hadrian.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 203

The Purpose and Operation of Roman Frontiers 203 His position was that: Hadrian’s frontier-policy was based on the view that peace could be maintained by an ambitious policy of economic development along the frontiers of the empire, backed up by the intelligent deployment of alae and cohorts according to the military needs of each frontier-province, with the legions ready to provide support on occasions of full-warfare. (Birley 1956, 32.) Birley pointed to the development of towns and cities along the frontiers of the empire, and of farms and villages to their rear. He suggested that Hadrian’s policy would have encouraged the creation of a reservoir of recruits used to the local environment in the frontier zone and aided supply to the units based on the frontiers, as well as providing an exemplar to the empire’s neighbours of a peaceful and settled way of life. Certainly, towns grew up along frontiers; they were the natural concomitant of the placing of so many well-paid soldiers there. Since the army enforced peace, agriculture blossomed, farms and villages grew. But whether these were the purpose behind Hadrian’s frontier policy is more difficult to determine. They may have been the natural result of the creation of the frontiers and their armies. Further, it is going too far to suggest that the purpose of frontiers related to the collection of customs duties; such a role for the army is more likely to have been an effect of frontiers rather than their cause. In any case, it is far from clear that soldiers were involved more than occasionally in the collection of customs duties. Lewis and Reinhold, in commenting upon one of the lists of customs duties cited by Cherry, the list of customs duties at Zarai ‘instituted after the departure of the cohort’, state that: The reason for the posting of a new tariff after the removal of the garrison from this post is not clear. There is no evidence that the customs were ever collected by the military. Because of the special needs of the soldiers – who were, moreover, exempt from customs dues on articles purchased for their personal needs – their departure probably rendered the previous tariff schedule obsolete. (Lewis and Reinhold 1966, 146, n. 137 commenting on CIL VIII 4508.) It is also difficult to justify the construction of major engineering works such as the land frontiers of Europe on the basis that their primary purpose related to the collection of customs: surely there were simpler ways of collecting customs? Furthermore, it is even more difficult to demonstrate the existence of trade across the major ‘defensive’ frontiers of northern Britain. One remarkable aspect of the area is the relative lack of Roman goods north of the province. Erdrich has gone further and argued that for long periods the German frontier was closed, with few goods penetrating into the areas beyond.20 In this respect we may note a Roman law of the third century which forbade the export of grain, salt, iron and whetstones to those who had formally been declared enemies (Digest 39, 4, 11). This implies that

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 204

204 The Frontiers of Imperial Rome export was possible to friendly states and the enormous quantities of Roman goods discovered beyond the frontier emphasizes that there was a strong flow through the frontiers. Acceptance of the Erdrich position, and the law, would suggest that the purpose of Roman frontiers was not only to control the movement of people into the empire, but also the flow of goods out of it.21 Frontiers created a point – or rather many points – of friction between the peoples inside and those outside, as we have seen. One purpose of the army in the frontier area can therefore be seen as keeping the peace in areas where the army had created volatility. It has, for example, been suggested that the very act of building the forts along Trajan’s new road in Arabia in the early fourth century may itself have contributed to raiding along the frontier by nomads disgruntled by the interference of Rome.22 It has also been proposed that the Chattan War of 83 may have been sparked by the construction of forts in the Wetterau. It is occasionally suggested that one purpose of Roman frontiers was to keep people in the empire as well as restrict access. One example cited is the line of forts built in the late first century along the edge of the Highlands of Scotland. It has to be said that there is no evidence for groups of people wanting to leave the Roman Empire, as opposed to states or peoples wishing to throw off the Roman yoke such as Arminius in Germany in 9. The most that can be said is that there is evidence for Roman army deserters. These are mentioned, for example, several times by Dio in relation to the wars of Marcus Aurelius in Germany (Dio 72, 11, 2; 73, 2, 2). One final proposal for the function of frontiers is that they were symbols of Roman might intended to intimidate their enemies. The very size of Hadrian’s Wall led John Mann to suggest that: Hadrian wished to make a statement to the people of northern Britain. He wished to state that Rome is invincible: there is no point in making pin-prick attacks on its invulnerability. Look at this magnificent structure – defended gateways every mile, and turrets with equally imposing regularity. This is pure rhetoric. This is a simple statement of the might of Rome. What hope have you of challenging it?23 The mortar rendering which resulted in forts, walls and towers in the east and on the German limes appearing white, and the possibility of Hadrian’s Wall having been similarly treated, has been taken as support for the contention that frontiers were symbolic in nature. This, however, appears to have been the normal way of finishing many masonry walls in the Roman Empire and thus frontier installations are not unique in this.24 Luttwak summarily dismissed any notion that Roman frontiers were ‘merely “symbolic”’. He continued, ‘this reduced their function to that of mere boundary markers. If that were so, their construction would have been wildly irrational, given the vast effort needed to build them.’25 Yet, if not symbolic in that way, the construction of so many frontier works under Hadrian may reflect a particular problem, perhaps appreciated by the emperor

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 205

The Purpose and Operation of Roman Frontiers 205 himself, namely a shortage of manpower. This may have played a part in his decision to abandon Trajan’s eastern conquests as well as some land beyond the Danube. A generation later, the basing of legionaries in some of the Antonine Wall forts may have been another facet of the same problem. The abandonment of the Antonine Wall, Tony Birley has suggested, may have followed a review of military commitments. Hadrian’s understanding of the tensions within his empire may thus have played a significant part in his construction of frontiers. In conclusion, we might interpret the frontier arrangements as reflecting several levels of activity. The first concern was the defence of the empire. This was undertaken primarily by the legions and later the field armies. In some places, most notably astride significant routes leading to the empire, there were groups of units, usually including a strong cavalry component. Individual units were probably sufficient to deal with small attacking forces. At the next level, there were the forts, small forts, fortlets, towers and linear barriers whose function would appear to have been to hinder and prevent raiding across the frontier, as stated on the Commodan inscriptions from Hungary, for example. The third level was the attempt to control movement into, and out of, the empire. The existence of such regulations is attested in the literary sources and they – and other forms of evidence – also indicate that the army was responsible for operating the regulations. The building of the frontier installations, and particularly the linear barriers, would have facilitated the enforcement of these regulations. Perhaps the purpose of Roman frontiers can be summed up in one word: security.26 The very existence of this security would have allowed the peaceful economic exploitation of the adjacent countryside within the empire, but that does not carry with it the assumption that the frontiers were specifically created for that purpose. Nor was the construction of these myriad frontier installations necessary to protect the empire’s soldiers; the soldiers of Rome were amply capable of defending themselves and, as Collingwood emphasized, preferred to fight in the open where their superior training, discipline and weapons usually gave them victory.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 206

Chapter 20

Were Roman Frontiers Successful?

here is plentiful evidence that every Roman frontier was at one time or another and often many times, crossed by invading armies and these instances do not need to be repeated here. There is also widespread evidence for raiding. Sometimes evidence for the destruction invaders wrought is recorded in archaeological excavations.1 There are occasional references to the frontier army successfully repelling an invasion. In 166 or 167, a force of ‘6,000 Lombards and Obii crossed the Danube, but the cavalry under Vindex issued forth and the infantry commanded by Candidus arrived, so that the barbarians were completely routed’ (Dio 71, 3, 1a). The barbarians sued for peace. It may not be coincidental that the invasion occurred when the armies of the Danubian provinces had been denuded of troops to fight in the Parthian War. A complete legion had been withdrawn from Lower Pannonia and a second from Lower Moesia, probably as well as detachments from other legions of Upper and Lower Pannonia. This raid seems to have been an exploratory expedition and the main invasion followed, with the attackers reaching Aquileia at the head of the Adriatic Sea. It may also not be a coincidence that there is a reference to piracy in the Black Sea in the 170s when Roman arms were stretched by warfare on both the Eastern and Northern Frontiers. It is worth considering further what happened when Rome was distracted and failed to maintain her strong watch over her frontiers. During the civil war of 68–9 forts on the Lower Rhine were attacked and some destroyed, while piracy was resumed in the Black Sea. During the second half of the third century, when emperors and usurpers were seeking the throne and fighting between each other, the frontiers were stripped of troops to engage in these civil wars. Previously friendly kings, seeing the weakness of Rome, became more neutral in their attitude to the empire, sometimes even downright hostile. Is it surprising that the third century was a time of heightened raiding and invasions on all the frontiers of the empire? Did invasions, whether successful or not, affect frontiers? This is not necessarily easy to answer. András Mócsy argued that as a result of ‘experiences on the Danube frontier in the Year of the Four Emperors … the occupation of the frontier areas by stronger forces and the building of garrison-posts were regarded as urgent tasks.’2 The ‘experience’ was the invasion of the Lower Danube provinces by the Dacians in 69–70 when the Romans were otherwise engaged with their own civil war and the defeat and death of the provincial governor. Vespasian and his younger son Domitian progressively strengthened the Lower Danube frontier as a reaction to this event

T

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 207

Were Roman Frontiers Successful? 207 and the later threatening activities of the Dacians into the 80s and beyond. The legionary strength of the provinces of Moesia, divided into two by Domitian, was increased from two to six. The number of legions in Pannonia was increased from two to four and the military balance of the empire swung from the Rhine frontier to the Danube. But the threat on the Lower Danube was only removed by the conquest of the Dacian kingdom. The Batavian Revolt of 69–71 certainly had an effect on military deployment, but only to the extent that Batavian regiments were sent for service elsewhere in the empire. This probably did not last long before the army resumed local recruiting. It can be seen that the major and continuous invasions of the third century affected frontiers. The frontier armies were strengthened, more emperors were created each with his own field army, and new styles of military architecture were introduced to help prevent forts and towns falling to attack. Energetic emperors, as we have seen, went on the offensive when opportunities arose. Were frontiers successful in preventing raiding? This question is more difficult to answer because we have no detailed ancient accounts. The evidence from Egypt demonstrates that raiders or small groups might have felt strong enough to attack military posts. The creation of special forces to deal with brigands, such as that commanded by Gargilius Martialis in Mauretania Caesariensis in 260, illustrates the measures taken in response to border disturbances. The erection of towers and then more towers on the banks of the Danube suggests that there was a regular breakdown in the system of watch and ward. Yet, the Romans clearly believed in the efficacy of the measures – or had run out of other ideas – for they continued to build forts, fortlets and towers on their frontiers. Were the regulations that the Romans sought to impose on movement into the empire successful? This is even harder to answer. We can point to the same approach operating over a period of one hundred years from the late first century, but can only guess at their efficacy. We might presume, based on our own experience, that an individual would always be able to slip through the net. Yet, by using wide rivers, arid wastes or strong barriers as their boundaries, the Romans certainly made it difficult even for individuals. In one aspect, there is a strong hint that the Romans could not achieve success in imposing their regulations on their neighbours. In the 170s and early 180s, Marcus and Commodus signed a series of treaties with the Marcomanni, Quadi and Iazyges. At first the Marcomanni were required to retain a distance of 16km (10 miles) between themselves and the Danube. Then this was reduced, ‘grudgingly and reluctantly’, to half that distance. The Iazyges went through the same process, and indeed a step further for all the restrictions on their movement north of the Danube were removed, save using their own boats on the river or landing on the islands in the river. The progressive reduction of the conditions suggests that the Romans did not have it all their own way. One way of judging the success of frontiers is to examine the life of civilians in the frontier zone. Along every frontier, villages, towns and cities sprang up. Their

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 208

208 The Frontiers of Imperial Rome role was to service the soldiers, but also to exploit the soldiers. They survived as long as soldiers lived on the frontiers. Beyond these urban centres, agriculture developed, doubtless to enable the soldiers and civilians to be fed. The growth of agriculture under the beneficent umbrella of Rome can be seen in the hinterland of every frontier. The distribution of an army along a long frontier may have resulted in the broadly successful control of access to Rome’s space, but it was not a good military tactic. In the event of an invasion, it would have taken longer to assemble the army. The allocation of troops to many small posts was a problem from Varus in 9 to Julian in 357, and was the reason why Trajan insisted that soldiers should stay with the colours. The re-invention by the fourth-century emperors of the Augustan strategy for the protection of the Eastern Frontier is another example. The best form of protection was the placing of a strong mobile force at Antioch astride the main east–west route between Rome and Persia and also in the best location to move north or south in the event of trouble elsewhere. The Romans may have had to adjust and even re-invent their frontier arrangements, military deployment and tactics on occasions, but Roman frontiers had a long run of success lasting nearly 400 years. Some have argued that the construction of the military installations along frontiers led to a reaction from Rome’s enemies in the form of more intensive raiding or even warfare. While this may have occurred, a more persuasive case has been made that the very existence of the Roman Empire and the strength of her armies led to the increasing amalgamation of the peoples beyond the frontier until they were strong enough to challenge the might of Rome.3 In that situation, the existence and success of the frontiers played a part not just in defending the empire but in helping create the explosion which eventually occurred. Ultimately, of course, Roman frontiers were not successful. The reason for this has little to do with the frontiers themselves, but rather with the failure of the Roman Empire to expand and bring all peoples under her sway. In that way, Roman frontiers are, as John Mann pointed out, symbols of abdication and failure to complete the conquest of the known world rather than the continuing strength and dominance of Rome.4

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 209

Conclusions

t the start of this book, I commented on the sense of despair among ancient historians about Roman frontiers and about the lack of synoptic treatments of frontiers. I hope that this book has gone some way to state and prove the reverse. By studying frontiers in detail we can understand how they worked. In particular, I hope that I have been able to sweep away some of the pessimism mentioned in the introduction. While it is manifestly not always possible to elucidate the changing role of a single fort in a rapidly evolving situation, the development of Roman frontiers, on the whole, moved more slowly and close observation of the evolving pattern can allow specific conclusions to be reached. Separate roles for soldiers clearly existed – springboards for advance, defence from invasions and from raiding, frontier control, support, control of routes – and in some instances it is possible to observe the changing function of a single fort. I trust that a coherent picture has been drawn and that it is not possible again to ignore the contribution of archaeology to aiding our understanding of the development, operation and function of Roman frontiers. I have sought to demonstrate that Roman frontiers had two principal functions, separate, conflicting yet intertwined. Troops were located to help defend the empire from major invasions and from raiding across the border. The first function required the retention of major army groups while the latter led to the spreading of soldiers in small numbers along the frontier line. As a result, soldiers were able to undertake another duty, that of enforcing the regulations which governed the movement of people into the empire and through the frontier region. There is increasing evidence to suggest that the frontiers were planned in detail to allow for a strict watch on the empire’s boundary. This includes the analysis by Dietwulf Baatz of the operation of the land frontier in Germany and the implications of Woolliscroft’s conclusions about the positioning of milecastles and turrets on Hadrian’s Wall. How strictly such a regime was maintained is, however, another matter. Here we dip into the realm of Ramsay MacMullen’s interpretations of the actions of soldiers on frontiers.1 In his analysis of military deployment on the German frontier, Baatz stated that:

A

the view presented here contrasts with the conventional notion of ‘the static lining up of units along the limes’ which is said to have impeded field operation of the Roman army. In reality, the organization of the limes provided flexibility of deployment. (Baatz 1997, 17.)

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 210

210 The Frontiers of Imperial Rome That same flexibility of approach can be recognised on Hadrian’s Wall and the Antonine Wall. As the frontiers were being built there were several changes of plan. We are, I think, at liberty to assume that these were in reaction to problems perceived as the Walls were being constructed, or, perhaps, in reaction to local opposition, or, of course, both. At any rate, they reflect a willingness to change plans in order to create a better product. The senior army officers who grappled with their frontier problems appear to have had relatively open minds in seeking to provide the most effective solutions. I suspect that I have assumed, probably too readily, that the Romans responded logically to their frontier problems. Yet we know, sometimes only too clearly, that we do not always act logically today so why should we assume that the Romans always did? Emperors might act out of a desire for glory or because they liked fighting.2 Activities might take place on certain frontiers in order to free emperors for actions elsewhere as appeared to be the case under both Claudius and Trajan. Hadrian stands out not only because of his long-term interest in frontiers, but also because of the spectacular nature of his creation in Britain, Hadrian’s Wall. However we judge that frontier, it is larger and grander than what was strictly necessary and I think there is good reason for believing that this was due to the involvement of the emperor himself.3 If that is the case, it emphasizes the element of personal whim in the creation of Roman frontiers, as in all aspects of life. The discussion of the role of frontiers is part of a wider debate about the nature of the Roman Empire and Roman rule. Some archaeologists have denied that the frontier installations were built to protect the peoples of the empire, a function which others consider so obvious that it is not worth stating. Some see the role of army units in the frontier zone as protecting travellers, others that their construction was a reaction to raiding. The argument becomes more complicated when it is suggested that the raiding is the result of the act of building the frontiers. Ideas on Roman frontiers certainly continue to develop. It has become fashionable to see Roman frontiers as being symbolic. This view has a long history. Hadrian’s activities in erecting linear barriers, for example, have been seen as symbolic of his non-expansionist policy. Attention is now drawn to the way the towers in Germany were painted or the possibility that Hadrian’s Wall might have been whitewashed. Both have been interpreted as offering a statement to the barbarian of Rome’s might; indeed the very act of building a massive frontier such as Hadrian’s Wall or the Antonine Wall has been interpreted as the Roman equivalent of ‘look on my works … and despair’ (Shelley, Ozymandias). I hope that I have been able to show that there was a detail of planning and execution, of development and yet consistency, which suggests that Roman frontiers had a purpose more related to practical considerations than symbolism. I have also avoided entering the debate on why the empire stopped expanding. In that discussion one element seems to have been ignored, the role of water. The great rivers, the Rhine, Danube and Euphrates, formed too substantial features in the

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 211

Conclusions 211 landscape to be ignored and elsewhere in the northern half of the empire the seas were also significant as boundaries.4 In the southern half, water was equally important, but in a different way. Here the limit of the empire was governed by the availability of water to sustain life, or at least agriculture. In both halves of the empire, water was strongly instrumental in determining the location of frontiers. I end at the point where this book began. Roman frontiers were not designed to stop a major invasion. The stringing out of soldiers along a 7,500–km (4,800–mile) frontier could not achieve that aim, as was demonstrated many times during the life of the empire. There was a tension between static defence and mobility which can be seen throughout history. The Romans steadily moved towards building frontier installations in order to prevent ‘the secret activities of bandits’ and ensure that their frontier regulations were enforced. They were the ancient equivalent of a preoccupation which concerns us to this day, the attempt to control access to ‘our’ space. The concentration of troops on the frontier effectively created a stasis in the Roman army which emperors subsequently tried to counter by creating new field armies, the only force which could defend the empire by facing and defeating invaders in the field. It was the defeat of those mobile forces which led to the military collapse of the Roman Empire. One of the conclusions to be drawn from this review of Roman frontiers is the serious degree of continuity. All Roman emperors were concerned to protect Rome’s interests and several emperors who have received a bad press, such as Nero and Domitian, were very concerned to protect their frontiers in the way they thought best. Many sought to expand the empire; some achieved that. Most sought, when they had available resources, to take the fight to the enemy. They used their army to protect their civilians through to the late fourth century. All negotiated with their neighbours. When Rome was powerful, especially in the early empire, they sought to have their own nominees on the thrones of these kingdoms. They encouraged, at all times, the children and relations of their neighbouring kings to visit Rome, to be educated there, to become Romanized. They also used them as hostages. The foci of the threats might vary. But the crucial point is that the Roman methods did not vary. Indeed, their actions were in large measure forced on them by the topography within which they operated. In the Middle East and North Africa, oases had to be controlled, caravans protected, routes had to be monitored everywhere, fertile areas where people lived guarded and, if outside the empire, kept under surveillance. In short, there is hardly an element of frontier and foreign policy which cannot be exampled at any time from the first century BCE to the fifth century CE, and later. This is the Grand Strategy of the Roman Empire and Rome maintained it throughout five centuries as indeed do many modern states today. Edward Gibbon’s The Decline and Fall of the Roman Empire still casts a long shadow. There is an implication that the empire was on a slow, inexorable decline from its high point – under Augustus? Antoninus Pius? – to the end, which occurred in 476 in the West and, of course, a millennium later in the East. But, up until the end of the fourth century, the empire was still firing on all cylinders. Constantine

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 212

212 The Frontiers of Imperial Rome erected bridges across the northern rivers, protected on the opposite banks by bridge-head forts. Valentinian operated in the same way as many of his forebears. He took aggressive actions. When he wanted close contact with the enemy, he erected a bridge across the river and invaded the territory of his enemies. When he wanted to strengthen the frontier, he built new forts and towers, and not just in one area but across the empire from Britain to Arabia. His general, Count Theodosius, drove out the barbarians and restored the fortified places on the frontier of Britain. When the existing army proved to be insufficient to protect the island, a new field army was sent there about 400. It is difficult to see a slow decline here, rather a continuum of purpose and actions. The writing of a book is a personal voyage of discovery. This was certainly the case with this volume. In particular, I have been impressed by the ways in which archaeological research is improving our knowledge of Roman frontiers and archaeologists sharpening our interpretations. As the literary sources are unlikely to increase except minimally, and the new documents detail life on the frontiers but not their development, archaeology is the main source of any new understanding. Refinement of existing information, continuing excavations and the results of new techniques have led to significant changes in interpretations. Nevertheless, these interpretations remain embedded in an appreciation of the landscape within which the forts and frontiers sat. History may be about chaps and geography about maps, but the two are intertwined when the chaps are Roman soldiers for where they served and how they served related intimately to the geography in which they lived. At the same time, it seems to me axiomatic that the Romans sought to impose broadly similar solutions on every one of their frontiers. It is the tension between the general and the particular which is so fascinating.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 213

Further Reading

Introduction There is an extensive literature on Roman frontiers and the Roman army. The following, mainly books, offer an introduction to frontiers and foreign policy. Austin, N. J. E. and Rankov, B. 1995. Exploratio: Military and political intelligence in the Roman world, London Barrett, J. C., Fitzpatrick, A. P. and Macinnes, L. 1989. Barbarians and Romans in North-West Europe, BAR IS 471, Oxford Birley, A. R. 1974. ‘Roman Frontiers and Roman Frontier Policy: Some Reflections on Roman Imperialism’, Transactions of the Architectural and Archaeological Society of Durham and Northumberland, n. s. 3, 13–25 Birley, E. 1956. ‘Hadrianic Frontier Policy’, in Limes 2, 25–33 Braund, D. 1984. Rome and the Friendly King: The character of the client kingship, London Breeze, D. J. and Dobson, B. 1993. Roman Officers and Frontiers, Stuttgart. Curzon, Lord 1907. Frontiers, The Romanes Lectures, Oxford Drummond, S. K. and Nelson, L. H. 1994. The Western Frontiers of Imperial Rome, Armonk Dyson, S. 1985. The Creation of the Roman Frontier, Princeton Elton, H. 1996. Frontiers of the Roman Empire, London Ferrill, A. 1986. The Fall of the Roman Empire: The Military Explanation, London and New York Ferris, I. M. 2000. Enemies of Rome. Barbarians through Roman Eyes, Stroud Hanson, W. S. 1989. ‘The nature and function of Roman frontiers’, in Barrett, J. C., Fitzpatrick, A. P. and Macinnes, L., Barbarians and Romans in North-West Europe, BAR IS 471, Oxford, 55–63 Hanson, W. S. (ed.) 2009. The Army and Frontiers of Rome, JRA Supp. Ser. 74, Portsmouth, Rhode Island Isaac, B. 1990. The Limits of Empire. The Roman Army in the East, Oxford Luttwak, E. N. 1976. The Grand Strategy of the Roman Empire, Baltimore Mann, J. C. 1974. ‘The frontiers of the Principate’, in H. Temporini (ed), Aufstieg und Niedergang der römischen Welt, II 1, Berlin and New York, 508–33 = Mann 1996, 58–83 Mann, J. C. 1996. Britain and the Roman Empire, Aldershot Mattern, S. P. 1999. Rome and the Enemy. Imperial Strategy in the Principate, Berkeley, Los Angeles and London Maxfield, V. A. 1987. ‘The Frontiers: Mainland Europe’, in Wacher, J. (ed), The Roman World, vol. I, London, 139–93 Maxfield, V. A. 1990. ‘Hadrian’s Wall in its imperial setting’, Archaeologia Aeliana5 18, 1–27 Millar, F. 1966. The Roman Empire and its Neighbours, London Millar, F. 1982. ‘Emperors, Frontiers and Foreign Relations, 31 BC to AD 378’, Britannia 13, 1–23

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 214

214 The Frontiers of Imperial Rome Mommsen, T. 1968. The Provinces of the Roman Empire, The European Provinces, T. R. S. Broughton (ed.), Chicago and London Napoli, J. 1997. Recherches sur les Fortifications Linéaires Romaines, Collection de l’École française de Rome 229, Rome Whittaker, C. R. 1994. Frontiers of the Roman Empire. A social and economic study, Baltimore Whittaker, C. R. 2004. Rome and its Frontiers, The Dynamics of Empire, London Woolliscroft, D. J. 2001. Roman Military Signalling, London Part I: Sources Ancient writers Many of the ancient sources are available in translation. I provide details of these where appropriate. Some of these are available on-line, including on the Bill Thayer Lacus Curtius website. Aelius Aristides, Roman Oration (written in the middle of the second century) Translation: Oliver, J. H., The Ruling Power: A Study of the Roman empire in the Second Century after Christ Through the Roman Oration of Aelius Aristides, Transactions of the American Philosophical Society NS 43, 1953 Ammianus Marcellinus, History of the Roman Empire (written in the late fourth century) Translations: Hamilton, W., Ammianus Marcellinus, The Later Roman Empire (A.D. 354–378), Penguin, Harmondsworth 1986 Rolfe, J. C., Ammianus Marcellinus, History, Loeb, London 1950 Appian, Roman History (written in the middle of the second century) Translation: White, H., Appian’s Roman History, volume 1, Loeb, London 1912 Anonymous, de rebus bellicis (probably written about 368) Translation: Ireland, R., De rebus bellicis, BAR IS 63, Oxford 1979 Arrian, Circumnavigation of the Black Sea (written in the 130s) Translation: Liddle, A., Arrian, Periplus Ponti Euxini, Duckworth, London 2003 Arrian, The Expedition against the Alans (written in the 130s) Translation: DeVoto, J. G., FLAVIUS ARRIANUS, TAXNH TAKTIKA (Tactical Handbook) And EKTAΞIΣ KATA AΛANΩN (The Expedition against the Alans), Ares Publishers, Chicago 1993 The Achievements of the Divine Augustus (written shortly before 14) Translations: Ehrenberg, V. and Jones, A. H. M., Documents Illustrating the Reigns of Augustus and Tiberius, Oxford 1955, is reproduced in Brunt, P. A. and Moore, J. M. (eds.), Res Gestae Divi Augusti, The Achievements of the Divine Augustus, Oxford 1967 Aurelius Victor, History of the Caesars (written in the middle of the fourth century) Translation: Bird, H. W., Aurelius Victor: De Caesaribus, Liverpool University Press, Liverpool 1994 Caesar, The Gallic War (describing the conquest of Gaul in the 50s BCE) Translations: Edwards, H. J., Caesar, The Gallic War, Loeb, London 1917 Handford, S. A., Caesar, The Conquest of Gaul, Penguin, Harmondsworth 1951 Cassius Dio, A History of Rome (written in the first quarter of the third century) Translation: Cary, E., Dio Cassius, Roman History, Loeb, London 1925 Eutropius, Abridgement of Roman History (written in the late fourth century) Frontinus, Stratagems (written in the late first century) Translation: Bennett, C. E., Frontinus, Stratagems and Aqueducts, Loeb, London 1925

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 215

Further Reading 215 Herodian, A History of Rome (written about 248) Translation: Whittaker, C. R., Herodian, History of the Empire, Loeb, London 1969 Historia Augusta, Lives of the Later Emperors (written in the fourth century) Translations: Birley, A. R., Lives of the Later Caesars, Penguin, Harmondsworth 1976 Magie, D., Historia Augusta, Loeb, London, 3 vols., 1921–32 Josephus, The Jewish War (written in 75–79) Translations: Thackerey, H. St J., Josephus, The Jewish War, 4 vols, Loeb, London 1925–8 Williamson, G. A., Josephus, The Jewish War, Penguin, Harmondsworth, 1969 Tacitus, The Life of Agricola (published in 98) Translations: Birley, A. R., Tacitus, Agricola and Germany, Penguin, Harmondsworth 1999 Page, T. E., Tacitus, Dialogues, Agricola and Germania, Loeb, London, 1914 Tacitus, The Annals (published in 116) Translations: Grant, E., Tacitus, The Annals of Imperial Rome, Penguin, Harmondsworth 1959 Jackson, J., Annals, Books 1–3, in The Histories, Loeb, London 1931 Jackson, J., Tacitus, Annals, Vol. IV, Books 4–6, 11–12, Loeb, London 1937 Jackson, J., Tacitus, Annals, Vol. V, Books 13–16, Loeb, London 1937 Tacitus, Germania (published in 98) Translations: see under Agricola Tacitus, The Histories (written in 104–109) Translations: Moore, C. H., Tacitus, The Histories, Books I–III, Loeb, London 1925; Books IV–V, Loeb, London, 1931 Wellesley, K., Tacitus, The Histories, Penguin, Harmondsworth, 1964 Three Byzantine Military Treatises (dating to the sixth and the tenth centuries) Translation: Dennis, G. T., Three Byzantine Military Treatises, Dumbarton Oaks 1985 Vegetius, Epitoma rei militaris (written in the fourth century) Translation: Milner, N. P., Vegetius: Epitome of Military Science, Liverpool University Press, Liverpool 1993 Zosimus, A New History (written in about 500) Translation: Lewis, N. and Reinhold, M., Roman Civilization, Sourcebook II: The Empire, New York 1955, 475–6 Collections of sources Campbell, B. 1994. The Roman Army, 31 BC–AD 337, A Sourcebook, London Lewis, N. and Reinhold, M. 1955. Roman Civilization, Sourcebook II: The Empire, New York There are two useful collections of material relating to Britain: Dobson, B. and Maxfield, V. A. (eds.) 2006. Inscriptions of Roman Britain, 4th edition, London = LACTOR 4 Mann, J. C. and Penman, R. (eds.) 1985. The Literary Sources for Roman Britain, London = LACTOR 11 Military documents Bowman, A. K. and Thomas, J. D. 1994. The Vindolanda Writing-tablets (Tabulae Vindolandenses II), London Bowman, A. K. and Thomas, J. D. 2003. The Vindolanda Writing-tablets (Tabulae Vindolandenses) Volume III, London Fink, R. O. 1971. Roman Military Records on Papyrus, Cape Western Reserve University

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 216

216 The Frontiers of Imperial Rome Goodburn, R. and Bartholomew, P. 1976. Aspects of the Notitia Dignitatum, BAR SS 15, Oxford Seeck, O. (ed.) 1876. Notitia Dignitatum, Berlin Tomlin, R. S. O. 1998. ‘Roman manuscripts from Carlisle: the ink-written tablets’, Britannia 29, 31–84 The building blocks of frontiers Bidwell, P. 2007. Roman Forts in Britain, London Brewer, R. J. 2000. Roman Fortresses and their Legions, London Campbell, J. B. 1984. The Emperor and the Roman Army 31 BC–AD 235, Oxford Campbell, B. 2002. War and Society in Imperial Rome 31 BC–AD 284, London and New York Coello, T. 1996. Unit Sizes in the Late Roman Army, BAR IS 645, Oxford Connolly, P. 1981. Greece and Rome at War, London Davies, R. W. 1989. Service in the Roman Army, D. J. Breeze and Valerie Maxfield (eds.), Edinburgh Farnum, J. H. 2005. The Positioning of the Roman Imperial Legions, BAR IS 1458, Oxford Hanson, W. S. 1978. ‘The Roman Military Timber Supply’, Britannia 9, 293–305 Goldsworthy, A. 2003. The Complete Roman Army, London Goldsworthy, A. and Haynes, I. (eds.) 1999. The Roman Army as a Community in Peace and War, JRA Suppl. Ser., Portsmouth, Rhode Island Gregory, S. 1989. ‘Not “Why not playing-cards?” but “Why playing-cards in the first place?”’, in D. H. French and C. S. Lightfoot (eds.), The Eastern Frontier of the Roman Empire, BAR Int. Ser. 553, Oxford Johnson, A. 1983. Roman Forts, London Johnson, S. 1983. Late Roman Fortifications, London Jones, R. H. 2011. Roman Camps in Britain, Stroud Keppie, L. 1984. The Making of the Roman Army, London Lander, J. 1984. Roman Stone Fortifications. Variations and Change from the First Century A.D. to the Fourth, BAR IS 207, Oxford Le Bohec, Y. 1994. The Imperial Roman Army, London and New York Parker, H. M. D. 1928. The Roman Legions, London Petrikovits, H. von 1971. ‘Fortifications in the North-Western Roman Empire from the Third to the Fifth Centuries A.D.’, Journal of Roman Studies 61, 178–218 Reddé, M. 1986. Mare Nostrum, Paris Roth, J. P. 1999. The Logistics of the Roman Army at War (264 BC–AD 235), Leiden Saddington, D. 1982. The Development of Roman Auxiliary Forces from Caesar to Vespasian, Harare Speidel, M. P. 2006. Emperor Hadrian’s Speeches to the African Army – a New Text, Mainz Starr, C. G. 1960. The Roman Imperial Navy, 31 B.C.–A.D. 324, 2nd ed. Cambridge Symonds, M. 2007. The Design and Purpose of Roman Fortlets in the North-Western Frontier Provinces of the Empire, unpublished Oxford DPhil thesis Watson, G. R. 1969. The Roman Soldier, 2nd ed., London Webster, G. 1985. The Roman Imperial Army, 3rd ed., London Part II: The frontiers Linear barriers Germany Baatz, D. 1997. ‘Keeping watch over the Limes’, Archaeologia Aeliana5 25, 1–20 Baatz, D. and Herrmann, F.-R. 1982. Der Römer in Hessen, Stuttgart

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 217

Further Reading 217 Baatz, D. 2000. Der römische Limes, Archäologische Ausflüge zwischen Rhein und Donau, 4th ed. Berlin Creighton, J. D. and Wilson, R. J. A. 1999. Roman Germany. Studies in Cultural Interaction, JRA SS 32, Portsmouth, Rhode Island Czysz, W., Dietz, K., Fischer, T. and Kellner, H.-J. 1991. Die Römer in Bayern, Stuttgart Filtzinger, P., Planck, D. and Cämmerer, B. (eds.) 1986. Der Römer in Baden-Württemberg, Stuttgart Horn, H. G. (ed.) 1987. Der Römer in Nordrhein-Westfalen, Stuttgart Schönberger, H. 1969. ‘The Roman Frontier in Germany: An Archaeological Survey’, Journal of Roman Studies 59, 144–97 Schönberger, H. 1985. Die römischen Truppenlager der frühen und mittleren Kaiserzeit zwischen Nordee und Inn, Frankfurt am Main Sommer, C. S. 1999. ‘From conquered territory to Roman province: recent discoveries and debate on the Roman occupation of SW Germany’, in Creighton and Wilson 1999, 160–98 Thiel, A. (ed.) 2007. Forschungen zur Funktion des Limes, Stuttgart Thiel, A. (ed.) 2008. Neue Forschungen am Limes, Stuttgart Hadrian’s Wall Bidwell, P. (ed.) 1999. Hadrian’s Wall 1989–1999, Kendal Bidwell, P. 2005. ‘The system of obstacles on Hadrian’s Wall: their extent, date and purpose’, Arbeia Journal 8, 53–75 Bidwell, P. (ed.) 2008. Understanding Hadrian’s Wall, The Arbeia Society, South Shields Birley, R. 2009. Vindolanda, a Roman Frontier Fort on Hadrian’s Wall, Amberley Breeze, D. J. 2006. J. Collingwood Bruce’s Handbook to the Roman Wall, 14th ed. Newcastle upon Tyne Breeze, D. J. 2009. ‘Did Hadrian design Hadrian’s Wall?’, Archaeologia Aeliana5 38, 87–103 Breeze, D. J. and Dobson, B. 2000. Hadrian’s Wall, 4th ed. London Collingwood, R. G. 1921. ‘The purpose of the Roman Wall’, Vasculum 8, 1, 4–9 Hill, P. 2006. The Construction of Hadrian’s Wall, Stroud Hodgson, N. (compiler) 2009. Hadrian’s Wall 1999–2009, Kendal Poulter, J. 2009. Surveying Roman Military Landscapes across Northern Britain, BAR 492, Oxford Woolliscroft, D. J. 1989. ‘Signalling and the Design of Hadrian’s Wall’, Archaeologia Aeliana5 17, 5–20 The Antonine Wall Breeze, D. J. 2006. The Antonine Wall, Edinburgh Hanson, W. S. and Maxwell, G. S. 1986. Rome’s North-West Frontier, The Antonine Wall, Edinburgh Keppie, L. 1998. Roman Inscribed and Sculptured Stones in the Hunterian Museum, University of Glasgow, London Keppie, L. 2004. The Legacy of Rome: Scotland’s Roman Remains, Edinburgh Robertson, A. S. 2001. The Antonine Wall, A handbook to the surviving remains, L. Keppie (ed.), Glasgow Africa: The Fossatum Africae Baradez, J. 1949. Vue-aerienne de l’organisation Romaine dans le Sud-Algerien, Fossatum Africae, Paris Cherry, D. 1998. Power and Society in Roman North Africa, Oxford

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 218

218 The Frontiers of Imperial Rome Fentress, E. W. B. 1979. Numidia and the Roman Army. Social, Military and Economic Aspects of the Frontier Zone, BAR Int. Ser. 53, Oxford Dacia Bogdan Cătăniciu, I. O. M. 1981. Evolution of the System of the Defence Works in Roman Dacia, BAR IS 116, Oxford Bogdan Cătăniciu, I. O. M. 1997. Wallachia in the Defensive System of the Roman Empire, 1st–3th centuries A.D., Alexandria River Frontiers The Rhine Bogaers, J. E. and Rüger, C. B. 1974. Der Niedergermanische Limes, Köln Brulet, R., Léva, C., Mertens, J., Plumier, J. and Thollard, P. 1995. Forts Romains de la Route Bavay-Tongres, Louvain-la-Neuve Schönberger, H. 1969. ‘The Roman Frontier in Germany: An Archaeological Survey’, Journal of Roman Studies 59, 144–97 Schönberger, H. 1985. Die römischen Truppenlager der frühen und mittleren Kaiserzeit zwischen Nordee und Inn, Frankfurt am Main The Danube Alföldy, G. 1974. Noricum, London Bondoc, D. 2009. The Roman Rule to the North of the Lower Danube during the Late Roman and Early Byzantine Period, Cluj-Napoca Bülow, G. von and Milčeva, A. (eds.) 1999. Der Limes an der unteren Donau von Diokletian bis Heraklios, Sofia Gassner, V., Jilek, S. and Stuppner, A. 1997. Der römische Limes in Österreich, Vienna Ivanov, R. 1997. Das römische Verteidigungssystem an der unteren Donau zwischen Dorticum und Durostorum (Bulgarien) von Augustus bis Maurikios, in: Ber. RGK 78, 467–640 Mackensen, M. 1999. ‘Late Roman fortifications and building programmes in the province of Raetia: the evidence of recent excavations and some new reflections’, in J. D. Creighton and R. J. A. Wilson, Roman Germany. Studies in Cultural Interaction, JRA SS 32, Portsmouth, Rhode Island, 199–244 Mirković, M. 2003. Römer an der mittleren Donau. Römische Strassen und Festungen von Singidunum bis Aquae, Belgrade Mocsy, A. 1974. Pannonia and Upper Moesia, London Petrović, P. (ed.) 1996. Roman Limes on the Middle and Lower Danube, Belgrade Pinterović, D. 1968. Limesstudien in der Baranja und in Slawonien, Arch. Iugoslavica 9, Belgrade, 5–83 Pitts, L. F. 1989. ‘Relations between Rome and the German “Kings” on the Middle Danube in the First to Fourth Centuries’, Journal of Roman Studies 79, 45–58 Poulter, A. G. (ed.) 2007. The Transition to Late Antiquity: on the Danube and beyond, Oxford Vagalinski, L. (ed.) 2007. The Lower Danube in Antiquity (VI C BC–VI C AD), Sofia Visy, Zs. 2003. The Roman Army in Pannonia, Pécs Visy, Zs. 2003. The Ripa Pannonica in Hungary, Budapest Wilkes, J. J. 2005. ‘The Roman Danube: An Archaeological Survey’, Journal of Roman Studies 95, 124–225 Zahariade, M. 1997. The Fortifications of Lower Moesia (A.D. 86–275), Amsterdam Zahariade, M. 2006. Scythia Minor, A History of a Later Roman Province (284–681), Amsterdam

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 219

Further Reading 219 Desert frontiers The Eastern Frontier Bowersock, G. W. 1983. Roman Arabia, Harvard Brünnow, R. E. and Domaszewski, A. von 1904–9. Die Provincia Arabia, Strasbourg Freeman, P. and Kennedy. D. (eds.) 1986. The Defence of the Roman and Byzantine East, BAR Int. Ser. 297, Oxford French, D. H. and Lightfoot, C. S. (eds.) 1989. The Eastern Frontier of the Roman Empire, BAR Int. Ser. 553, Oxford 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. L. Kennedy (ed.), The Roman Army in the East, JRA SS 18, Ann Arbor, 169–209 Gregory, S. and Kennedy, D. L. (eds.) 1985. Sir Aurel Stein’s limes report, BAR IS 272, Oxford Hodgson, N. 1989. ‘The East as part of the Wider Roman Imperial Frontier Policy’, in French and Lightfoot 1989, 177–89 Jackson, R. B. 2002. At Empire’s Edge. Exploring Rome’s Egyptian Frontier, Yale Kennedy, D. L. (ed.) 1996. The Roman Army in the East, JRA SS 18, Ann Arbor Kennedy, D. 2000. The Roman Army in Jordan, Amman Kennedy, D. and Riley, D. 1990. Rome’s Desert Frontier from the Air, London Parker, S. T. 1986. Rome and Saracens, A History of the Arabian Frontier, Philadelphia Parker, S. T. 2006. The Roman Frontier in Central Jordan. Final Report of the Limes Arabicus Project 1980–1989, Dumbarton Oaks Studies XL, Washington DC Poidebard, A. 1934. Le trace de Rome dans le désert de Syrie, Paris Sheldon, R. M. 2010. Rome’s Wars in Parthia, London and Portland, Oregon North Africa Barker, G. and Mattingly, D. (eds.) 1996. Farming the Desert. The Unesco Libyan Valleys Archaeological Survey I–II, London Cuvigny, H. (ed.) 2003. La route de Myos Hormos. L’armée romaine dans le desert oriental d’Egypte. Praesidia du désert de Bérénice. Fouilles Institut français d’archéologie orientale 48, 1–2, Paris Daniels, C. M. 1987. ‘The Frontiers: Africa’, in J. Wacher (ed.), The Roman World, vol. I, London 1987, 223–65 Goodchild, R. G. 1976. Libyan Studies, London Lesquier, J. 1918. L’armée romaine d’Égypte d’Auguste à Dioclétien, Cairo Mattingly, D. 1995. Roman Tripolitania, London Maxfield, V. A. 2000. ‘The Deployment of the Roman Auxilia in Upper Egypt and the Eastern Desert’, in G. Alföldy, B. Dobson and W. Eck, Kaiser, Heer und Gesellschrift in der Römischen Welt, Gedenkschrift für Eric Birley, Stuttgart, 407–42 Maxfield, V. A. 2005. ‘Organisation of a desert limes: the case of Egypt’, in Limes 19, 201–10 Maxfield, V. A. 2009. ‘“Where did they put the men?” An enquiry into the accommodation of soldiers in Roman Egypt’, in Hanson 2009, 63–82 Sidebotham, S. E., Hense, M. and Nouvens, H. M. 2008. The Red Land. The Illustrated Archaeology of Egypt’s Eastern Desert, Cairo/New York Trousset, P. 1974. Recherches sur le Limes Tripolitanus, Paris

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 220

220 The Frontiers of Imperial Rome Mountain frontiers Dacia Bogdan Cătăniciu, I. O. M. 1981. Evolution of the System of the Defence Works in Roman Dacia, BAR IS 116, Oxford Bogdan Cătăniciu, I. O. M. 1997. Wallachia in the Defensive System of the Roman Empire, 1st–3th centuries A.D., Alexandria Gudea, N. 1979. ’The defensive system of Roman Dacia’, Britannia 10, 63–87 Gudea, N. 1997. Römer und Barbaren an den Grenzen des römischen Daciens, Zalau Hanson, W. S. and Haynes, I. P. (eds.) 2004. Roman Dacia, The Making of a Provincial Society, JRA SS 56, Portsmouth, Rhode Island North Africa Daniels, C. M. 1987. ‘The Frontiers: Africa’, in J. Wacher (ed.), The Roman World, vol. I, London, 223–65 Euzennat, M. 1989. Le Limes de Tingitane, La Frontière Méridionale, Ètudes d’Antiquités Africaines, Paris Rushworth, A. 1996. ‘North African deserts and mountains: comparisons and insights’ in Kennedy 1996, 297–313 Salama, P. 1977. ‘Les deplacements successifs de Limes en Mauretanie Césarienne (Essai de synthèse)’, in Limes XI, 577–95 Sea frontiers Britain Hanson, W. S. 1987. Agricola and the Conquest of the North, London Woolliscroft, D. J. and Hoffman, B. 2006. Rome’s First Frontier: The Flavian Occupation of Northern Scotland, London The Black Sea Braund, D. 1994. Georgia in Antiquity. A History of Colchis and Transcaucasian Iberia, 550 BC–AD 562, Oxford The Red Sea Tomber, R. 2008. Indo-Roman Trade, From Pots to Pepper, London The North Sea Cotterill, J. 1993. ‘The Late Roman Coastal Forts’, Britannia 24, 227–39 Johnston, D. E. (ed.) 1977. The Saxon Shore, CBA Research Report 18, London Johnson, S. 1976. The Roman Forts of the Saxon Shore, London Maxfield, V. A. (ed.) 1989. The Saxon Shore, A Handbook, Exeter Pearson, A. 2002. The Roman Saxon Shore Forts, Stroud Forests, marshes and swamps Poulter, J. 2009. Surveying Roman Military Landscapes across Northern Britain, BAR 492, Oxford

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 221

Further Reading 221 The Deep Frontier: Defence in Depth? Dobson, B. 1970. ‘Roman Durham’, Transactions of the Architectural and Archaeological Society of Durham and Northumberland 2, 31–43 Woolliscroft, D. J. and Hoffman, B. 2006. Rome’s First Frontier: The Flavian Occupation of Northern Scotland, London Part III Decision making Breeze, D. J. 1988. ‘Why did the Roman army fail to conquer Scotland?’, Proc. Soc. Antiq. Scotland 118, 3–22 = Breeze and Dobson 1993, 65–84. Millar, F. 1982. ‘Emperors, Frontiers and Foreign Relations, 31 BC to AD 378’, Britannia 13, 1–23 Potter, D. 1996. ‘Emperors, their borders and their neighbours: the scope of imperial mandata’, in D. L. Kennedy (ed.) 1996. The Roman Army in the East, JRA SS 18, Ann Arbor, 49–65 Proceedings of the Congress of Roman Frontier Studies (Limes Congress) The volumes are referred to as Limes 1–20 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

15. 16.

17.

Birley, E. (ed.) 1952. The Congress of Roman Frontier Studies 1949, Durham Swoboda, E. (ed.) 1956. Carnuntina, Graz/Köln Laur-Belart, R. (ed.) 1959. Limes-Studien, Basel Not published Novak, G. (ed.) 1963. Quintus congressus internationalis limitis Romani studiosorum, Zagreb 1967. Studien zu den militärgrenzen Roms, Graz/Köln Applebaum, S. (ed.) 1971. Roman Frontier Studies, 1967, Tel Aviv Birley, E., Dobson, B. and Jarrett, M. G. (eds.) 1974. Roman Frontier Studies, 1969, Cardiff Pippidi, D. M. (ed.) 1974. Actes du XIe congrès international d’études sur les frontières romaines, Bucureşti Haupt, D. and Horn, H. G. (eds.) 1977. Studien zu den Militärgrenzen Roms, II, Köln/Bonn Fitz, J. (ed.) 1978. Akten des XI. Internationalen Limes Kongresses, Budapest Hanson, W. S. and Keppie, L. J. F. (eds.) 1980. Roman Frontier Studies 1979, BAR IS 71, Oxford Unz, C. (ed.) 1986. Studien zu den Militärgrenzen Roms, III, Stuttgart Vetters, H. and Kandler, M. (eds.) 1990. Akten des 14. Internationalen Limeskongresses 1986 in Carnuntum, Wien [Theme: the impact of the Roman army on indigenous populations] Maxfield, V. A. and Dobson, M. J. (eds.) 1991. Roman Frontier Studies 1989, Exeter [Themes: Roman and Native; realities of frontier life; the problems of desert frontiers] Groenman-van Waateringe, W., van Beek, B. L., Willems, W. J. H. and Wynia, S. L. (eds.) 1997. Roman Frontier Studies 1995, Oxbow Monograph 91, Oxford [Themes: problems of river frontiers versus artificial frontiers; problems of late defence; across the frontier; resources and supply] Gudea, N. (ed.) 1999. Roman Frontier Studies 1997, Proceedings of the XVIIth International Congress of Roman Frontier Studies, Zalau

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 222

222 The Frontiers of Imperial Rome 18. Freeman, P., Bennett, J., Fiema, Z. T. and Hoffmann, B. (eds.) 2002. Limes XVIII, Proceedings of the XVIIIth International Congress of Roman Frontier Studies held in Amman, Jordan (September 2000), BAR IS 1984, Oxford 19. Visy, Zs. (ed.) 2005. Limes XIX, Proceedings of the XIXth International Congress of Roman Frontier Studies, Pécs, Hungary, September 2003, Pécs [Themes: how did frontiers actually work?; Roman frontiers – barbarians] 20. Morillo, A., Hanel, N. and Martín, E. (eds.) 2009. Limes XX, Estudios Sobre la Frontera Roman, Roman Frontier Studies, Anejos de Gladius 13, Madrid [Themes: the internal frontiers; the end of frontiers; the walled towns and military fortifications in late Roman period; the early development of frontiers] Theiss guides Bechert, T. and Willem, W. J. H. (eds.) 1995. Die römische Reichsgrenze von der Mosel bis zur Nordseeküste, Stuttgart Beck, W. and Planck, D. 1980. Der Limes in Südwestdeutschland, Stuttgart Klee, M. 1989. Der Limes zwischen Rhein und Main, Stuttgart Rupp, V. and Birley, H. 2005. Wanderungen am Wetteraulimes, Wiesbaden Schallmayer, E. 2010. Der Odenwald Limes, Entlang der römischen Grenze zwischen Main und Neckar, Stuttgart Ulbert, G. and Fischer, T. 1983. Der Limes in Bayern, Stuttgatt Visy, Zs. 1989. Der pannonische Limes in Ungarn, Budapest The Frontiers of the Roman Empire series of popular guides Breeze, D. J. 2009. The Antonine Wall, Edinburgh Breeze, D. J. 2011. Hadrian’s Wall, Hexham Dyczek, P. 2008. The Lower Danube Limes in Bulgaria, Warsaw and Vienna Harmadyová, K., Rajtár, J. and Schmidtová, J. 2008. Slovakia, Nitra Jilek, S. 2009. The Danube Limes, A Roman River Frontier, Warsaw Visy, Zs. 2008. The Roman Limes in Hungary, Pécs World Heritage Sites guides Klee, M., Der römische Limes in Hessen, Geschichte und Schauplätze des UNESCO-Welterbes, Regensburg Fischer, T. and Riedmeier-Fischer, E. 2008. Der römische Limes in Bayern. Geschichte und Schauplätze entlang des UNESCO-Welterbes, Regensburg Thiel, A. 2005. Wege am Limes, Stuttgart Report of the Culture 2000 Frontiers of the Roman Empire project Breeze, D. J. and Jilek, S. 2008. Frontiers of the Roman Empire, The European Dimension of a World Heritage Site, Edinburgh DVD series: Frontiers of the Roman Empire Hadrian’s Wall – Antonine Wall (2004) The Limes in Germany (2005) The Danube – Rome’s great river frontier (2008) The DVDs are available in Britain from www.northern-heritage.co.uk and in Germany from www.theiss.de

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 223

Sites to See

I have only included reference to the Roman name when it is in general use. Britain Hadrian’s Wall and the Cumbrian coast: the forts and museums at South Shields, Segedunum/Wallsend, Chesters, Corbridge, Housesteads, Vindolanda and Birdoswald and Maryport; the Wall between Housesteads and Gilsland; the milefortlet at Swarthy Hill. The Antonine Wall: the fortlet and museum at Kinneil; the forts at Rough Castle and Bar Hill; the bath-house at Bearsden; the Wall from Castlecary to Bar Hill. The Saxon Shore: the forts at Burgh Castle, Richborough, Pevensey and Portchester. Yorkshire coast: the late tower at Scarborough. Wales: the remains and museum at Caerleon; the forts at Segontium/Carnarfon and Neath; the fortified landing/fortlet at Caer Gybi; the museum and reconstructed fort at Cardiff. England: the legionary bases at Chester and York; the forts at Risingham, High Rochester, Lanchester, Piercebridge, Old Carlisle, Hardknott, Moresby, Bowes; the fortlet at Maiden Castle; the bath-house at Ravenglass. Scotland: the legionary base at Inchtuthil; the forts at Ardoch and Birrens; the fortlets at Castle Greg, Durisdeer, Kaimes Castle and Redshaw Burn. The Netherlands The fort at Zwammerdam; the museum at Nijmegen. German limes The Rhine: the fort at Boppard; the museum, city wall, late tower at Cologne and the bridgehead fort at Cologne-Deutz. The Rhine to the Main: the forts at Holzhausen, the Saalburg (and museum, civil settlement and reconstructed palisade and bank), Feldberg, Kapersburg and Grosskrotzenburg; the bath-houses at Kapersburg and Rückingen; reconstructed towers at Rheinbrohl (1/1), Oberbieber (1/37), Bendorf (1/54), Hillscheid (1/68), Arzbaach (1/84), Bad Ems ( 2/1), Orlen (3/15), Pfaffenwiesbach (4/16), Butzbach (4/33), Pohlheim (4/49), reconstructed section of the frontier between Limeshain and Rommelshausen (4/102 and 4/103). The Odenwald: the fort at Neckarburken; the bath-houses at Neckarburken and Würzberg. The Outer Limes: the forts at Haselburg, Osterburken (and museum and annexe), Mainhardt, and Welzheim; the fortlets at Rötelsee and Dalkingen; the bath-house at Walldürn; reconstructed towers at Geisselhardt (9/64), Grosserlach-Grab (9/83), Murrhardt (9/96); reconstructed section of frontier at Walldürn and Grosserlach-Grab. Raetia: the gates of the legionary fortress and museum at Regensburg; the forts at Aalen (and museum), Rainau-Buch (and civilian houses), Eining, Ellingen, Weissenburg, Pfünz; the fortlets at Dalkingen and Harlach; the bath-houses at Buch, Theilenhofen and Weissenburg; the reconstructed towers at Lorch (12/44), Rainau-Buch (12/77),

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 224

224 The Frontiers of Imperial Rome Mönchsroth (13/2), Grüb-Dühren (13/24), Rittern (14/17), Burgsalach (14/48), TittingErkertshofen (14/63), Kipfenberg (14/78), Hieheim; the section of wall at Unterschwaningen-Dennenlohe (13/38 and 13/39). Switzerland The fortress and museum at Windisch; the late fort at Augusta Raurica/Raiseraugst; the late fort at Zurzach. Austria The forts and museums at Mautern, Traismauer, Tulln, Zeiselmauer; the tower at Bacharnsdorf; the town and museum at Carnuntum. Slovakia The fort at Celemantia/Iza; the late Roman tower and museum at Gerulata in Bratislava. Hungary The legionary fortress, town and museum at Aquincum/Budapest; the late forts at Contra Aquincum/Budapest, Tokod, Visegrád; the towers at Leányfalu and Visegrád. Croatia Remains at Mursa; the late fortification at Siscia. Serbia Remains of the imperial palace at Sirmium/Sremska Mitrovica; the fortress at Viminacium/ Kostolac. Romania The forts at Porolissum, Buciumi, Capidava and Dinogetia; the monument at Adamklissi. Bulgaria The forts at Bononia/Vidin, Lom, Oescus/Gigen, Novae/Svištov, Ruse (Sexanginta Prista and the museum), Tutraken and Durostrorum/Silistr; the late fortlet at Kula. Turkey The fortresses at Satala and Meletine/Maletya. Georgia The fort at Apsaros/Gonio. Syria The forts at Ad-Diyatheh, Hân Aneybeh, Hân el-Hallabat, Hân al-Manqoûra and Sa’neh; the legionary base at Bostra; the cities at Dura-Europos, Resafa and Zenobia. Jordan The forts at Azraq, Deir el-Kahf, Umm el-Jimal, Qasr el-Hallabat, Da’ajaniya and Qasr Bsher; the late legionary fortresses at el-Lejjun and Udruh. Israel The fortlet at Upper Zohar.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 225

Sites to See 225 Egypt The forts and fortlets of the Eastern desert, especially those on the road from Qift to Quseir; Mons Claudianus; Abu Sha’ar. Libya The forts at Bu Ngem and Gheria el-Garbia. Tunisia The forts at Tisavar/Ksar, Ghilane and Bezereos/Bir Rhezan; the fortlets at Benia/Guedah Essede and Tibulici/Ksar Tarcine; the clausura at Bir Oum Ali. Algeria The fortress at Lambaesis; the forts at Ad Majores and Gemellae. Morocco The fort and ditch at Sala, the forts at Thamusida, Tocolosida/Bled Takourat and Sidi Moussa.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 226

Notes

Introduction (pages xvii–xxi) 1. Mattern 1999, 112; Isaac 1990, 410. Z. Visy, ‘Similarities and differences on the late defense system on the European and Eastern frontiers’, Limes 18, 74 makes the same point: ‘Isaac basically relies on written historic sources and epigraphic materials while the archaeological finds have got a secondary rôle in his train of thoughts.’ 2. R. MacMullen, Corruption and the Decline of Rome, New Haven and London 1988. 3. e.g. J. P. Roth, The Logistics of the Roman Army at War (264 BC–AD 235), Leiden 1999. 4. e.g. S. K. Drummond and L. H. Nelson, The Western Frontiers of Imperial Rome, New York 1994, 127–71. 5. R. Kipling, Puck of Pook’s Hill, 1906; the quotation is from the chapter entitled ‘On the Great Wall’. 6. Whittaker 1994. 7. J. Crow, ‘The Anastasian Wall and the Danube frontier in the sixth century’, in L. Vagalinski (ed.), The Lower Danube in Antiquity (the fifth century BC – the beginning of the seventh century AD), Tutrakan, Bulgaria 2007, 397–401; O. Ţentea and I. A. Oltean, ‘The Lower Danube Roman Limes at Galaţi (Romania). Recent Results from Excavation and Aerial Photographic Interpretation’, in Limes 20, 1515–23. The Frontiers (pages 3–6) 1. Lord Curzon 1907. 2. Poulter 2009. 3. T. Mommsen, The Provinces of the Roman Empire, The European Provinces, T. R. S. Broughton (ed.), Chicago and London 1968. 4. O. Lattimore, Studies in Frontier History, London 1962. 5. B. Isaac, ‘The meaning of “Limes” and “Limitanei” in Ancient Sources’, Journal of Roman Studies 78 (1988) 125–47. 6. For a discussion of terms see Isaac 1990, 172–84; Rushworth 1996, 312–3; Z. Visy, ‘Praesidia et Burgi in the Early Roman Empire’, Limes 20, 989–96. An Overview of the Sources (pages 7–13) 1. Seeck, 1876. J. C. Mann, ‘What was the Notitia Dignitatum for?’, in R. Goodburn and P. Bartholomew (eds.), Aspects of the Notitia Dignitatum, BAR SS 15, Oxford (1976) 1–9 = Mann 1996, 226–34; ‘The Notitia Dignitatum – Dating and Survival’, Britannia 22 (1991) 215–20 = Mann 1996, 249–53. 2. C. Flügel and J Obmann, ‘Fibule architettoniche Romane. Un contributo per l’architecttura militare Romana’, Quaderni Friulani di Archeologia anno 18, 1 (Dicembre 2008) 145–53. 3. Brünnow and von Domaszewski 1904–09.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 227

Notes 227 4. Poidebard 1934; Gregory and Kennedy 1985; Baradez 1949; R. W. Feachem, ‘Six Roman camps near the Antonine Wall’, Proc. Soc. Antiq. Scot. 89 (1956) 329–39. For problems of definition on the Eastern Frontier see Gregory 1996. 5. W. Czysz and F. Hervig, ‘Neue Dendrodaten von der Limespalisade in Raetien’, in Theil 2008, 183–93 6. A. L. F. Rivet, Town and Country in Roman Britain, London 1958, 28–30. The Romans on Frontiers (pages 14–24) 1. R. Sherk, ‘Roman Geographical Exploration and Military Maps’, in H. Temporini (ed.), Aufstieg und Niedergang der römischen Welt, II, Pt. l, Berlin and New York 1974, 534–62. 2. e.g. C. Wells, The German Policy of Augustus, Oxford 1972, 2–13; Birley 1974, 13–5. 3. Birley 1974, 15. 4. Mann 1974, 508 = Mann 1996, 58. 5. Whittaker 2004, 40–2. 6. Maxfield 1990, 2. 7. Birley 1974, 19–22. 8. Alföldy 1974, 220–6; Eugippius, Life of St Severinus, 11, 20, 24–31, 44. Regulations and Treaties (pages 30–33) 1. e.g. Herodian 3, 14, 10: ‘after the army had crossed the rivers and fortifications which marked the borders of the empire’ (translation by C. R. Whittaker). 2. B. Campbell, The Writings of the Roman Land Surveyors: introduction, text, translation and commentary, London 2000; O. A. W. Dilke, The Roman Land Surveyors, An Introduction to the Agrimensores, Newton Abbot 1971. 3. Isaac 1990, 396. The Building Blocks of Frontiers (pages 34–52) 1. W. S. Hanson, ‘Rome, the Cornovii and the Ordovices’, in Limes 13, 47–52. 2. Davies 1989, 33–68. 3. Jones 2011. 4. W. S. Hanson, ‘Building the forts and the frontiers’, in Hanson 2009, 33–43. 5. P. Bidwell, ‘The earliest occurrences of baths at auxiliary forts’, in Hanson 2009, 54–62. 6. V. A. Maxfield, ‘Pre-Flavian Forts and their Garrisons’, Britannia 17 (1986) 59–72. 7. P. Vindol. 154; P. Dura 100 and 101; D. J. Breeze, ‘The Garrisoning of Roman Fortlets’, in Limes 10, 1–6 = Breeze and Dobson 1993, 505–10. 8. Lander 1984, 307–9. 9. Johnson 1983, 27–30. 10. On the problems of differentiating between Roman and other period structures in the East see Gregory 1989 and 1996. Linear Barriers (pages 55–91) 1. M. T. Boatwright, Hadrian and the Cities of the Roman Empire, Princeton 2003; T. F. Fraser, Hadrian as Builder and Benefactor in the Western Provinces, BAR IS 1484, Oxford 2006; Dio 68, 13, 5–6. 2. A. Thiel, ‘Der Limes wird Geschlossen. Zum Begin Durchgehender Grenzenbefestigungen in Südwestdeutschland’, in Limes 20, 977–87; Sommer 1999, 177–8. 3. Woolliscroft 2001, 115. 4. J. C. Mann quoted by D. J. Breeze and B. Dobson, ‘Hadrian’s Wall: Some Problems’, Britannia 4 (1973) 187 = Breeze and Dobson 1993, 409; P. Bidwell, ‘Did Hadrian’s Wall

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 228

228 The Frontiers of Imperial Rome

5. 6.

7. 8. 9.

10.

11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

have a Wall-walk?’, in Bidwell 2008, 140. A further possibility offered by Peter Hill is that there may have been a wall-walk in the first phase, abandoned when the wall was narrowed. P. Bidwell, ‘The system of obstacles on Hadrian’s Wall: their extent, date and purpose’, Arbeia Journal 8 (2005). For treatment of the two main opposing views see B. Dobson, ‘The function of Hadrian’s Wall’, Archaeologia Aeliana5 14 (1986) 21–4 (raiding), and P. Bidwell, ‘The system of obstacles on Hadrian’s Wall: their extent, date and purpose’, Arbeia Journal 8 (2005) 74 (defence and raiding). Poulter 2009; D. J. Woolliscroft, ‘Signalling and the Design of Hadrian’s Wall’, Archaeologia Aeliana5 17 (1989) 5–20. Hill 2006, 146. D. J. Breeze, ‘Why was Hadrian’s Wall built across the Tyne–Solway isthmus?’, “Eine Ganz Normale Inschrift” … und Ähnliches zum Geburtstag von Ekkehard Weber, (eds. F. Beutler and W. Hameter), Vienna 2005, 13–166. Suetonius, Claudius 17; Birley 1974; D. J. Breeze, ‘The abandonment of the Antonine Wall, its and implications’, Scottish Archaeological Forum 7 (1975) 67–80 = Breeze and Dobson 1993, 351–64. M. Hassall, ‘The Building of the Antonine Wall’, Britannia 14 (1983) 262–4. L. Keppie, ‘The garrison of the Antonine Wall: endangered species or disappearing asset?’, in Limes 20, 1135–64. Baatz 2000, 194–5. W. Czysz and F. Herzig, ‘Neue Dendrodaten von der Limespalisade in Raetien’, in Thiel 2008, 183–95. Woolliscroft 2001, 143–51. Baatz 1997, 14–7. Speidel 2006. Daniels 1976, 242–6; P. Trousset, ‘Les milliaires de Chebika (Sud Tunisien)’, Antiquités Africaines 15 (1980) 135–54. Fentress 1979, 114; Cherry 1998, 58–66. M. Zahariade and D. Lichiardpol, ‘The Roman Garrison in N. Wallachia in A.D. 101–118: its composition, size and structure’, in Hanson 2009, 173–83. J. P. Gillam, ‘Possible changes in plan in the course of the construction of the Antonine Wall’, Scottish Archaeological Forum 7 (1976) 51–6. C. E. Stevens, The Building of Hadrian’s Wall, Kendal 1966, 86. Cf. n.1 above. Mommsen 1964, 165; J. Crow, ‘A review of recent research on the turrets and curtain of Hadrian’s Wall’, Britannia 22 (1991) 57–8.

River Frontiers (pages 92–117) 1. B. Rankov, ‘Do rivers make good frontiers?’, in Limes 19, 175–81; Austen and Rankov 1995, 173–89. 2. W. A. M. Hessing, ‘Das Niederlandische Küstengebeit’, in Bechert and Willems 1995, 89–101. 3. M. Mackensen, ‘Raetia: late Roman fortifications and building programmes’, in Creighton and Wilson 1999, 199–244. 4. Bogdan Cataniciu 1981, 22. 5. C. S. Sommer, ‘Why there? The positioning of fort along the riverine frontiers of the Roman empire’, in Hanson 2009, 103–14.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 229

Notes 229 6. M. Hajnalová and J. Rajtár, ‘Supply and subsistence: the Roman fort at Iža’, in Hanson 2009, 194–207 7. M. Mackensen, ‘Raetia: late Roman fortifications and building programmes’, in Creighton and Wilson 1999, 199–244. 8. T. Sarnowski, ‘Die Anfänge der spätrőmischen Militärorganisation des unteren Donaurraums’, in Limes 14, 855–61; A. Poulter, The Lower Danubian frontier in Late Antiquity: evolution and dramatic change in the frontier zone, c. 296–600, Vienna forthcoming. 9. Mócsy 1974, 269. 10. M. Mackensen, ‘Raetia: late Roman fortifications and building programmes’, in Creighton and Wilson 1999, 199–244. 11. Z. Mráv, ‘Quadian policy of Valentinian I and the never-finished late Roman fortress at Göd-Bócsaújtelep’, in Limes 19, 773–84. 12. Austin and Rankov 1995, 236. Desert Frontiers (pages 118–132) 1. Mann 1974, 525 = Mann 1996, 75. 2. Mann 1974, 524 = Mann 1996, 75. 3. Kennedy 2000, 55–6; 93; 208. 4. Isaac 1990, 127; Kennedy 2000, 37–8. 5. Parker 2006. 6. Kennedy 2000, 67–73; 90–6; 55–6. 7. M. Gichon, ‘The courtyard pattern castellum on the Limes Palaestinae, strategic and tactical features’, in Limes 14, 193–214. 8. Isaac 1990, 191–3. 9. To support his argument that the purpose of the posts was to protect travellers, Isaac argues that the words praetorium, which appears on the dedicatory inscription at Qasr Bshir, and praesidia have particular meanings in relation to this function. It is doubtful, however, if the names can be used so specifically. 10. Goodchild 1976, 7; 46–58. 11. Goodchild 1976, 57. 12. D. J. Smith, ‘The Centenaria of Tripolitania and their Antecedents’, in Libya in History, Benghazi 1972, 299–318. 13. Mattingly 1995, 79; 106–15. 14. Mattingly 1995, 186–201. 15. Maxfield 2000; Maxfield in Limes 19, 201–10. 16. Maxfield in Limes 20, 73–84. 17. Maxfield in Limes 19, 201–10. Mountain Frontiers (pages 133–145) 1. Zahariade 1997, 607. 2. F. Marcu, ‘“Geminari Castra” in Dacia’, in Limes 19, 703–11. 3. Hanson and Haynes 2004, 26. 4. Fentress 1979, 110. 5. Daniels 1987, 253. 6. C. R. Whittaker, ‘Trade and Frontiers of the Roman Empire, in P. Garnsey and C. R. Whittaker (eds.), Trade and Famine in Classical Antiquity, Cambridge 1983, 110–27, has suggested that the two lines of forts were held at the same time and that the purpose of this was to enable the troops in the more southerly to keep watch over the nomads

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 230

230 The Frontiers of Imperial Rome grazing their cattle in these areas while those in the more northerly forts regulated their movement into the agricultural areas when the animals needed to graze on the stubble in high summer. This concept of a ‘Waiting Zone’ does not survive the abandonment of the proposal that the earlier line of forts continued in occupation. 7. J. Spaul, ‘The Roman “Frontier” in Morocco’, University College London Institute of Archaeology Bulletin 30 (1994) 108. 8. For a review of the present positions: D. J. Breeze, L. M. Thoms and D. W. Hall (eds.), First Contact, Rome and Northern Britain, Perth 2009. Sea Frontiers (pages 146–158) 1. AE 1948, 201, from Rhodes contains a reference to piracy, perhaps in the second century. 2. M. P. Speidel, ‘The Caucasus Frontier. Second Century Garrisons at Apsarus, Petra and Phasis’, in Limes 13, 657–60. 3. Johnson 1979, 44. 4. J. C. Mann, ‘The Historical Development of the Saxon Shore’, in Maxfield 1989, 10–11 = Mann 1996, 212–3. 5. Johnson 1976, 131. 6. J. Cotterill, ‘The Late Roman Coastal Forts’, Britannia 24 (1993), 227–39. 7. B. C. Burnham and J. L. Davies (eds.), Roman Frontiers in Wales and the Marches, Cardiff 2010, 216–7; 220–3. Forests, Marshes and Swamps (pages 159–160) 1. Tacitus Agricola 20 and 25; Dio 76, 13.1. R. Tipping, ‘The form and fate of Scotland’s woodlands’, Proceedings of the Society of Antiquaries of Scotland 124 (1994) 31–2. 2. Poulter 2009, 116; Breeze and Dobson 2000, 112. The Deep Frontier: Defence-in-Depth? (pages 161–163) 1. D. J. Breeze, L. M. Thoms and D. W. Hall (eds.), First Contact, Rome and Northern Britain, Perth 2009, includes papers summarising present views. 2. W. S. Hanson, ‘Rome, the Cornovii and the Ordovices’, in Limes 13, 47–52. 3. Mann 1996, 217–25 = ‘The Northern Frontier after A.D. 369’, Glasgow Archaeological Journal 3 (1974) 34–42. 4. P. Bidwell and N. Hodgson, The Roman Army in Northern England, Kendal 2009. 5. Dobson 1970. The Development of Frontiers (pages 167–171) 1. B. Dobson, ‘The role of the fort’ in Hanson 2009, 31. 2. Plutarch, Life of Crassus 10, 4–6; Appian, Roman History, The Civil Wars, 14, 118–20: I owe these references to Alan Beale. Caesar, The Conquest of Gaul 2, 1. Military Deployment (pages 172–176) 1. Baatz 1997, 14–7. 2. N. J. Higham, ‘The Roman impact upon rural settlement in Cumbria’, in P. Clack and S. Haselgrove, Rural Settlement in the Roman North, Durham 1982, 105–22. 3. D. J. Breeze, Roman forces and native populations’, Proc. Soc. Antiq. Scot. 115 (1985) 223–8 = Breeze and Dobson 1993, 385–90. 4. Rushworth 1996, 303.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 231

Notes 231 5. L. Allason-Jones, ‘Small finds from Turrets on Hadrian’s Wall’, in J. C. Coulston (ed.), Military Equipment and the Identity of Roman Soldiers, Proceedings of the Fourth Roman Military Equipment Conference, Oxford 1988, 197–233; Woolliscroft 2003. 6. A. Thiel, ‘The Odenwald limes and its relation to the Antonine Wall’, in Hanson 2009, 134–41. A Comparison of Frontiers (pages 177–180) 1. Little is known about military installations in the western desert: cf. M. Reddé, ‘A l’ouest du Nil: une Frontière sans soldets, des soldats sans Frontière’, Limes 15, 485–93. 2. Hodgson 1989, 177–89. 3. Hodgson 1989, 182. 4. Isaac 1990, 125. Decision Making (pages 181–183) 1. Lander 1984, 307–9. 2. D. Potter, ‘Emperors, their borders and their neighbours: the scope of imperial mandata’, in Kennedy 1996, 52. 3. Millar 1982, 9. How Did Frontiers Work? (pages 184–193) 1. R. Sherk, ‘Roman Geographical Exploration and Military Maps’, in H. Temporini (ed.), Aufstieg und Niedergang der römischen Welt, II, Pt l (1974) 534–62; M. Cary and E. H. Warmington, Ancient Explorers, Harmondsworth 1973. 2. L. F. Pitts, ‘Relations between Rome and the German ‘Kings’ on the Middle Danube in the First to Fourth Centuries A.D.’, Journal of Roman Studies 79 (1989) 45–58. 3. R. Syme, The Roman Revolution, Oxford 1939, 280, n. 3. 4. Austin and Rankov 1995, 147–9. 5. Alföldi 1952 for a discussion of the relations between Rome and her neighbours in northern Europe. 6. Maxfield in Limes 19, 201–10; Bowman and Thomas 1994, 42. 7. Fentress 108–110. 8. Woolliscroft 2003. 9. M. Corby, ‘Hadrian’s Wall and the defence of North Britain’, Archaeologia Aeliana5 39 (2010) 9–13. 10. Austin and Rankov 1995, 179–80. The Purpose and Operation of Roman Frontiers (pages 194–205) 1. Mommsen 1968, 165. 2. Cherry 1998, 66. 3. J. C. Bruce, The Wallet-book of the Roman Wall, Newcastle upon Tyne 1863, 16. 4. Mommsen 1968, 165, translation by T. R. S. Broughton. 5. R. G. Collingwood, ‘The purpose of the Roman Wall’, Vasculum 8, 1 (1921) 4–9. 6. J. C. Mann quoted by D. J. Breeze and B. Dobson, ‘Hadrian’s Wall: Some Problems’, Britannia 4 (1973) 187 = Breeze and Dobson 1993, 409. 7. e.g. M. Corby, ‘Hadrian’s Wall and the defence of north Britain’, Archaeologia Aeliana5 39 (2010) 9–13; Richard Holmes (pers. com.). Cf. G. H. Donaldson, ‘A military appreciation of the design of Hadrian’s Wall’, Archaeologia Aeliana5 16 (1988) 132: ‘It is blindingly obvious from a comparison of the length of the Wall with the manpower pool

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 232

232 The Frontiers of Imperial Rome

8. 9. 10.

11. 12. 13.

14.

15. 16.

17. 18. 19. 20. 21.

22. 23. 24.

25. 26.

available to Hadrian in Britain that the Wall can never have been visualized as a continuous fighting platform.’ D. Baatz, review of D. J. Breeze and B. Dobson, Hadrian’s Wall, Harmondsworth 1976, in Bonner Jahrbücher 177 (1977) 773–5; cf. Bidwell 2009, 130–3. P. Bidwell, ‘Did Hadrian’s Wall have a Wall-walk?’ in Bidwell 2008, 129–43. For some opposing arguments see: Breeze 2006, 109–10 and Breeze 2009, 100–1. J. Coulston, ‘Roman archery equipment’, in M. C. Bishop (ed.), The production and distribution of Roman military equipment. Proceedings of the second Roman military equipment research seminar, BAR IS 275, Oxford 1985, 283. P. Bidwell, op. cit. 141. Turrets on Hadrian’s Wall: Breeze 2006, 164 and 175. P. Bidwell, The systems of obstacles on Hadrian’s Wall: their extent, date and purpose’, Arbeia Journal 8 (2005) 74. J. Crow, ‘A review of recent research on the turrets and curtain of Hadrian’s Wall’, Britannia 22 (1991) 57–8; Breeze 2009, 87–103; J. P. Gillam, ‘Possible changes in plan in the course of the construction of the Antonine Wall’, Scottish Archaeological Forum 7 (1976) 51–6. It is too easy to over-play the defensive quality of the barrier itself; as a younger man, I was able to scale the reconstruction of Hadrian’s Wall at Vindolanda by climbing on a friend’s shoulders. Breeze and Dobson 2000, 120. I. A. Richmond, ‘Ancient Rome and Northern England’, Antiquity 14 (1940) 294–300; see also, D. J. Breeze and B. Dobson, ‘Hadrian’s Wall: Some Problems’, Britannia 3 (1972) 183–93 = Breeze and Dobson 1993, 405–15; B. Dobson, ‘The Function of Hadrian’s Wall’, Archaeologia Aeliana5 14 (1986) 1–30 = Breeze and Dobson 1993, 431–60. J. H. W. G. Liebeschutz, ‘The defences of Syria in the sixth century’, Limes 10, 487–99. Cherry 1998, 59–66. Lander 1984, 309. M. Erdrich, Rom und de Barbaren, Mainz 2001. M. Galestin, ‘Barriers for Barbarians?’, Limes 19, 221–5 cites evidence for trading across the Lower Rhine while Pitts 1989, 54–8, provides a summary of the position north of the Middle Danube and Harmadyová, Rajtár and Schmidtová 2008, 48 and 86–9, provide a map and discussion. Isaac 1990, 214–5. J. C. Mann, ‘The Function of Hadrian’s Wall’, Archaeologia Aeliana5 18 (1990) 535 = Mann 1996, 167. Oxyrhynchus Papyri LV, 3793, line 10 states that the walls of the fort at Psobthis require lime-washing: the documnt dates to 340. D. P. S. Peacock and V. A. Maxfield, Survey and Excavation. Mons claudianus 1987–1993, vol. 1. Topography and Quarries, Cairo 1997, 38, records stucco on the outside of the gate-towers. I owe these references to Walter Cockle. Luttwak 1976, 68. As emphasized to me by Bill Hanson.

Were Roman Frontiers Successful? (pages 206–208) 1. See the papers brought together by S. Jilek in Carnunturn Jahrbach (2005), Akten der 14 Internationalen Roman Military Equipment Conference (ROMEC) Wien, 27–31 August 2003. 2. Mócsy 1974, 80.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 233

Notes 233 3. J. C. Mann, ‘The Northern Frontier after A.D. 369’, Glasgow Archaeological Journal 3 (1974) 41 = Mann 1996, 224; P. Heather, The Fall of the Roman Empire, A New History, London 2005, 450–9. 4. Mann 1974, 508 = Mann 1996, 58. Conclusions (pages 209–212) 1. R. MacMullen, Soldier and Civilian in the Later Roman Empire, Harvard 1963. 2. Cf. the comments of Herodian and Dio on Severus’ reasons for his campaigning in Britain in the early third century: Severus liked glory and wanted to win some victories in Britain; he was anxious to get his sons away from the fleshpots of Rome; the army was becoming slack through inactivity; and that there had been an invasion which required more troops or the presence of the emperor (Herodian 3, 7, 11). A. R. Birley has argued that the request is so similar to other statements in Herodian’s History that it should be regarded as a topos (A. R. Birley, ‘Virius Lupus’, Archaeologia Aeliana4 50 (1972) 187–8. 3. Breeze 2009, 87–103. 4. Mann 1974, 513 = Mann 1996, 63.

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 234

Index

Page numbers in italics indicate illustrations Aalen, 57, 77 Abu-Sha’ar, port, 130, 131, 152, 225 Abusina/Eining, fort, 46, 57, 58, 109, 223, pl. 20 Ad Maiores, fort, 82–4, 139, 140, 225 Aelius Aristides, writer, 19–20 Africa, province, xxiii, 37, 55, 137 frontier, 3–5, 8, 16–17, 21 Agneb, outpost, 139, 140 Agricola, Julius, see Julius Agricola Ai-Todor, see Charax Aila/Elat, fort, 121, 122, 126, 191 Alamanni, people, 96, 186–7, 190 Alans, people, xvii, 137, 150, 173, 185 Alba Iulia, see Apulum Albania, kingdom, 17, 150 Albota, fort, 85, 134 Albulae, fort, 140, 142 Altrip, fort, 45, 57, 100 Ammianus Marcellinus, historian, 7, 8, 18, 23–4, 27, 39, 99, 116, 124–5, 128, 146, 159, 183, 187, 189–92 Anderita/Pevensey, fort, 46, 153–5, 154, 157 Annamatia/Baracs, fort and towers, 101, 112 annexes, 74, 87 Antioch, city, 35, 84, 120, 172, 194, 208 Antonine Wall, xxi, 4, 6–7, 10, 12, 26, 28, 48, 67, 70–7, 72, 74, 75, 79, 80, 81–2, 81, 82, 85–9, 90, 159, 162, 170, 173–5, 178, 197, 199, 201, 210, 223, pls. 17 and 18 Antoninus Pius, emperor, xv, xx, 4, 12, 19, 21, 26, 35, 48, 58, 70–1, 76, 88, 90, 139, 148, 152, 159, 181, 185, 200, 211, pl. 1b Appian, historian, 18, 20 Apsaros, fort, 25, 149, 224 Apulum/Alba Iulia, legionary base, xxiii, 36, 134, 136

Aquincum, legionary base, xxiii, 9, 100, 101, 104, 106, 107, 108, 112, 224 Arabia, province, 4, 6, 18, 48, 120–3, 125, 132, 141, 152, 177, 179, 191, 194, 202 Arbeia/South Shields, fort, 36, 62, 193, 223 Archar, see Ratiaria Argentorate/Strasbourg, legionary base, xxii, 57, 58, 61, 97 Armenia, kingdom/province, 14, 28, 114, 118, 120, 137, 149, 150, 185–7 army, Roman, duties of, 36, 48–1, 130–1, 183, 187–8, 195, 201–2 organization of, 34–9, 41 recruitment to, 36 training, 55, 82 (see also, auxiliary regiments, fleets, legions, officers) Arrian, governor and writer, 7, 25–6, 40, 137, 148–50, 152, 182–3, 185 Aswan, see Syene Atlantic Ocean, 3, 146–7 Atlas Mountains, 19, 126, 137, 139, 141–3, 145, pl. 2d Augsburg, see Augusta Vindelicorum Augusta Treverorum/Trier, 35, 57, 94, 98 Augusta Vindelicorum/Augsburg, city, 32, 59, 97 Augustus, emperor, xv, xvii, xix, 14–8, 20–1, 24, 35, 43–4, 93–4, 96, 100, 114, 120, 129, 137–8, 146–7, 167, 170, 172, 184, 194, 211 Aurelian, emperor, xvi, 171 Aurès Mountains, Arica, 82–3, 137 Austura/Zeiselmauer, fort, 24, 109 auxiliary units, 28, 34–9, 42–3, 67, 73, 77, 81–2, 103–5, 114–15, 119, 125, 128–9,

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 235

Index 235 134–6, 139, 142, 163, 172–3, 177–8, 180, 191–2 Auzia, fort, 140, 142, 143, 161, 189 Azrak, fort, 121, 122, 124, 191, 224 Baatz, Dietwulf, archaeologist, 77, 78, 79, 173, 197, 201, 209 Bagacum/Bavai, fortlet, 96, 97 Bejaia, see Saldae Banna/Birdoswald, fort, 62, 68, 223 Baradez, Jean, archaeologist, 10, 82–4, Baracs, see Annamatia Batavian Revolt, 25, 32, 92, 95–6, 147, 189, 207 Bavai, see Bagacum Bearsden, fort, 74, 223 Belkis, see Zeugma Belgrade, see Singidunum Berenice, port, 130, 131, 152 Berlin Wall, xvii, 10, 198, 201 Bettmauer, see Vemania Bezereos/Bir Rhezene, fortlet, 126, 127, 224 Bidwell, Paul, archaeologist, 63, 197–8 Bingen, see Bingium Bingium/Bingen, fort, 57, 94, 98 Birdoswald, see Banna Birley, Eric, archaeologist, xix, 12, 202–3 Birley, Tony, ancient historian, 16, 19, 26, 32, 147, 205 Black Sea, xxiii, 3–4, 6–7, 25–6, 35–6, 40, 114–15, 137, 146, 148–52, 149, 151, 184–5, 206 Bodensee/Lake Constance, 96, 97, 112 Boiodurum/Passau, fort, 24, 57, 101, 110 Bodobriga/Boppard, fort, 57, 94, 98, 100, 223 Bonn, see Bonna Bonna/Bonn, legionary base, xxii, 56, 57, 94 Boppard, see Bodobriga Bostra/Busra, legionary base, xxiii, 120–2, 121, 126, 177, 180, 223 Boulogne, see Gesoriacum boundary/ies, 5, 33, 92, 95, 114, 117 Bowness-on-Solway, see Maia Bradwell, see Othona Brancaster, see Branodunum Branodunum/Brancaster, fort, 153, 154, 154

Bremenium/High Rochester, fort, 71, 72, 223 bribes, see subsidies bridges, 97, 103, 116–7, 171, 197 Brigetio/Szôny, legionary base, xxiii, 100, 101, 104, 105, 106, 172 Britain, xxii, 19, 21, 37, 55, 146–8, 153–7, 161–3, 167, 170, 173, 177, 182, 189, 193 (see also Antonine Wall, Hadrian’s Wall, Saxon Shore) Budapest, see Aquincum buffer states, 3, 185–7 building materials, 4, 39–40, 46, 48–9, 56, 68–71, 77–8, 83–6, 105, 115, 136, 178, 181–2, 204, 210 Bu Ngem, see Gholaia bureaucracy, xix, 38, Burgh Castle, see Gariannonum burgi, 26–7 Burlafingen, fortlet, 102, 168 Busra, see Bostra Caesarea/Cherchel, city, 36, 140, 142 Caledonians, people, 19, 36, 144, 148, 162, 184, 186, 187, 193, 196 Capidava, fort, 108, 224 camps, 10, 39, Cappadocia, province, xxiii, 7, 25, 55, 114, 118, 137, 149, 150, 182, 185 Caracalla, emperor, xv, 78, 96, 186, 187 Carausius, usurper, 153, 155 Cardiff, fort, 154, 158, 223 Carlisle, see Luguvalium Carnuntum/Petronell, legionary base, xxii, 100, 101, 104, 172, 224 Carpathian Mountains, 4, 103, 133, 134, 136, 144, 175, 178, pl. 2c Carpow, legionary base, 41, 72, 144, 148 Carvoran, see Magna Caspian Sea, 4, 15, 137, 184 Cassius Dio, historian, 7, 16, 30–3, 55–6, 92, 103–4, 119, 144, 147–8, 171, 181, 184–5, 187–8, 195, 204, 206 Castellum Dimmidi/Messad, fort, 140, 141 Castra Batava/Passau, fort, 104 Castra Exploratorum/Netherby, fort, 62, 81 Castra Rauracense/Kaiseraugst, legionary base, 96, 108, 224

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 236

236 The Frontiers of Imperial Rome Castra Regina/Regensburg, legionary base, 57, 79, 179, 223 Caucasus Mountains, 4, 137, 144–5, 149, 150, 173 cavalry, 34, 37, 61, 67, 77–8, 95, 100, 104, 115, 134–6, 139, 142, 159, 162, 172–3, 178, 185, 191–2, 194 Celemantia/Iza, fort, 101, 105–6, 224 Charax/Ai-Todor, fort, 151, 152 Chatti, tribe/Chattan War, 58–9, 61, 76, 204 Chauci, people, 95, 96 Chélif, river, 140, 142 Cherchel, see Caesarea Chersonesus Taurica/Sevastopol, fleet base, 150, 151, 185, 186 Cherry, John, archaeologist, 202, 203 Chester, see Deva Chesters, see Cilurnum Cidamus/Ghadames, fort, 124, 126, 191 Cilurnum/Chesters, fort, 62, 66, 66, 67, 223, pl. 16 Cirta, see Constantine civilian settlements, 11, 25–6, 47, 125, 129, 174 Civil War of 68–9, 115, 147, 150, 206 Claudius, emperor, xv, 21, 58, 70, 95, 115, 138, 142, 147,167, 170, 181–2, 186, 210 Clyde, river, 19, 25, 71, 72, 86, 179 cohorts, 25, 34, 36, 38, 42–3, 119, 125, 128, 139, 190–1 coins, 8, 10, 55, 78, 79, 105, 154–5 Collingwood, R. G., archaeologist, 196–9, 205 Cologne, see Colonia Agrippinensis Colonia Agrippinensis/Cologne, city, 32, 93, 94, 94, 95, 96, 97, 223 Commodus, emperor, xv, 26–7, 27, 30–1, 106, 115, 119, 126, 139, 171, 184, 188, 199, 205, 207 Confluentes/Koblenz, fort, 57, 94, 98, 99, 100 Constance, Lake, see Bodensee Constans, emperor, xvi Constantine/Cirta, city, 84, 140 Constantine I, emperor, xvi, 21, 23, 38, 46, 97, 98, 108, 116, 181, pl. 1d Constantine II, emperor, xvi, Constantius I, emperor, xvi, 146, 148, 153, 158

Constantius II, emperor, xvi, 98, 108, 159, 171, 189, 190 Contra Florentiam/Dunafalva, landing place, 106, 107 Coptos/Qift, fort, 130, 131 Corbridge, fort, 62, 63, 65, 70, 190, 223 Cornelius Fronto, senator, 70, 181 Cornelius Tacitus, see Tacitus, Cornelius Ctesiphon, 119, 120 Curzon, Lord, 3, 159 Cyrenaica, province, xxiii, 126, 129, 138, 179 Dacia, kingdom/province, xxiii, 4, 15–16, 18, 24, 31, 36, 40, 61, 84–5, 92, 96, 102–4, 106, 133, 144, 159–61, 170–1, 173, 175, 177, 179, 185–6, 193–4, 201, 207 Da’janiya, fort, 121, 123, 224 Danube, river, 3, 5, 10, 15–19, 23–4, 26, 30–2, 35, 46, 55, 58–9, 59, 79, 92, 96, 100–12, 101, 115–7, 133, 134, 137, 148, 150, 151, 159, 161, 167, 171, 175, 189–90, 193–4, 201, 206, 210, pl. 2b Darial Pass, 137, 150 Decius, emperor, xv, 26 Deir el-Kahf, fort, 124, 224 dendrochronology, 12, 56, 61, 76, 89 desert frontiers, 3, 23, 114–15, 118–32, 171, 174, 202 Deutz, see Divitia Deva/Chester, legionary base, xxii, 62, 161, 223 Dio, see Cassius Dio Diocletian, emperor, xvi, 22–3, 34–5, 38, 46, 97, 106, 116, 120, 122, 124, 129, 132, 139, 141, 143, 148, 171, 181 Divitia/Deutz, fort, 94, 97, 98, 98, 223 Dobson, Brian, archaeologist, xix, 163 Domitian, emperor, xv, 58–9, 61, 70, 96, 102–3, 137, 147, 184–6, 206–7, 211 Domitius Corbulo, general, 95, 137, 150, 182 Dover, see Dubris Drobeta/Turnu Severin, fort, 110, 116, 134 Drumburgh, see Congavata Dubris/Dover, fort, 153, 154, 155 Dumata/Jauf, outpost, 121, 122 Dunafalva, see Contra Florentiam

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:16 Page 237

Index 237 Dunakömlőd, see Lussonium Dunaszekcső, see Lugio Dunaújvávaros, see Intercisa Dura-Europos, city and military base, 34, 38, 44, 113, 115, 118–20, 180, 190, 224, pl. 25 Durostorum/Silistra, fort, xxiii, 106, 151 Eburacum/York, legionary base, xxii, 161, 223 Eckzell, fort, 57, 77, 172 Edessa, city, 113, 119 Egypt, province, xxiii, 3, 6, 14, 32, 114, 126, 129–32, 130, 146, 175, 179, 188–9, pl. 28 Elat, see Aila Elbe, river, 13, 16, 17, 92, 100 Elginhaugh, fort, 45, 72 Ellingen, fort, 57, 77 Euphrates, river, 3, 17, 18, 20, 114–5, 118–9, 126, 133, 145, 174, 178, 194, 210 ‘expansions’, 72, 75, 75, 171 farming, 31, 84, 95, 125–6, 128, 139, 174, 187, 201–3, 207–8 Faviana/Mautern, fort, 24, 101, 224 Fectio/Vechten, fort, 94, 95 Fentress, Elizabeth, archaeologist, 84, 139 Flaminda, fort, 85, 134 fleets, 14, 17, 34–6, 50–1, 94, 102, 116, 150, 153, 155, 158 forests, 3, 23, 159–60 Forth, Firth of, 19, 25, 71, 72, 86, 162, 179 fortlets, 42–3, 46, 58–61, 63–4, 71–5, 75, 86, 87, 98, 102, 109, 115, 120, 128, 130, 133, 136, 147, 152, 170–1, 174–5, 177, 179, 188, 193, 195, 201 (see also milecastles) forts, 25–6, 39–43, 42, 46–51, 58–61, 66, 73, 75, 80, 84, 88, 96–9, 106–8, 109, 122, 124, 129, 141–2, 147, 152, 168–9, 171–7, 179, 195–7, 223–5 function of, 28–9, 61, 65, 77, 106, 149, 163, 173, 176, 201 late forts, 46, 96–9, 106–111, 116, 124, 153–7, 171 spacing of, 63, 66, 71, 73, 80, 84–5, 87, 93, 95, 104, 115, 142, 160, 167–71, 174, 175, 177, 179, 192

Fossatum Africae, frontier, 4, 10, 82–6, 128, 139, 140, 174, 194 Franks, people, 96, 146, 153, 156 Frejus, see Forum Iulii Friedberg, fort, 57, 61, 87, 172 Frisii, people, 18, 31, 95, 159 frontier zones, 3, 161–3, 177–8 frontiers, development of, 58–60, 66–8, 73–5, 78, 80–1, 167–71 function of, 59–61, 65, 77, 83–4, 194–205, 209 location within the landscape, 58, 65, 71, 76, 82–8, 118–9, 126, 129–31, 133–5, 137, 141–4, 150, 159, 161–3, 172–5, 177–9, 191 modern, xvii, 129, 198, 201 Frontinus, see Julius Frontinus Fronto, see Cornelius Fronto Gaius, emperor, xv, 93, 138, 182 Gallienus, emperor, xvi, 35, 79 Garamantes, people, 126, 138, 189, 191 Gargilius Martialis, commander, 142, 189, 207 Gariannonum/Burgh Castle, fort, 153, 154, 154, 223 gates through frontiers, 63–4, 67, 72–4, 80, 83, 87 Gask Ridge, Scotland, 162, 170 Gemellae, fort, 83, 84, 140, 225 Germany, province(s), xxii, 63, 96 conquest, 14, 19, 22, 24, 93 frontier, xxi, 3, 4, 12, 19, 26, 48, 55–61, 70, 75–9, 85–9, 159, 170, 177, 182, 200–1, pls. 9, 11, 12 Gerulata/Rusovce, fort, 101, 109, 224 Gesoriacum/Boulogne, fleet base, 153, 154, 154, 156 Ghadames, see Cidamus Gheriat el-Garbia, fort, 126, 127, 225 Gholaia/Bu Ngem, fort, 9, 126, 127, 128, 191, 225, pl. 27 Gibbon, Edward, writer, 22, 211 Gigen, see Oescus Goths, people, 22, 26 Gratian, emperor, xvi, 48, 187 Great Chesters, see Aesica Great Hungarian Plain, 105, 115, 137, 201

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 238

238 The Frontiers of Imperial Rome Habitancum/Risingham, fort, 48, 62, 72, 223 Hadrian, emperor, xv, xvii, xx, 7, 12, 21, 26, 55–6, 61, 68, 70, 76, 82, 85, 89–90, 103–4, 116, 118, 137, 142, 148, 152, 170, 177–8, 181, 185, 198–9, 203, pl.1a Hadrian’s Wall, xxi, 4, 6–7, 9, 12, 26–8, 41, 43, 48, 50–1, 61–70, 62, 64, 66, 69, 72, 75, 77, 80–2, 82, 85–90, 159, 162–3, 172, 174–5, 176, 178–9, 182, 190–1, 194–5, 197–9, 200, 204, 208–9, 223, pls. 14–16 Hanson, W S, archaeologist, 36, 136 Hermunduri, people 32, 188 Herodian, historian, 20–1, 119, 148 Hesselbach, fort, 6, 57, 60 Highlands of Scotland, 144, 162, 204 High Rochester, see Bremenium Historia Augusta, book, 26, 55–6, 61, 70, 76, 126, 146, 181 Hodgson, Nick, archaeologist, 179–81, 202 Hodna Mountains, Africa, 82–3, 139, 142 Honorius, emperor, xvi, 112 Horse Guards, 34, 55 Housesteads, see Vercovicium Hyginus, see Pseudo-Hyginus Iazyges, people, 5, 21, 30–1, 92, 101, 184, 187, 207 Iberia, kingdom, 17, 150, 185–6 Iglita, see Troesmis Iller, river, 3, 112 Inchtuthil, legionary base, 40, 72, 171, 173, 223 Innsbruck-Wilten, see Veldidana inscriptions, 8, 26–7, 47, 50–1, 55, 68, 70, 73, 80, 82, 84, 88, 97, 106, 111–12, 122–5, 128, 141–3, 158, 184–6, 188–9, 191, 203 Intercisa, fort and tower, 101, 112 invasions, xvii, 14–15, 22–3, 76, 80–1, 93, 96, 99–100, 102–3, 105, 115, 118–20, 129, 136, 195, 206 Iron Gates, 100, 102, 105–6 Iza, see Celemantia Jauf, see Dumata Johnson, Stephen, archaeologist, xix, 49, 154–6 Josephus, historian, 18, 146, 182

Judaea, province, 36–7, 114, 186 Julian, emperor, xv, 24, 39, 98–9, 116, 120, 171, 181, 186–7, 208 Julius Agricola, governor of Britain, 36, 62, 144, 147, 159, 182–3 Julius Caesar, 7, 15, 35, 38–9, 92–3, 102, 138, 167, 170, 186 Julius Frontinus, writer, 51, 58, 61, 200 Juvenal, writer, xix, 51 Kainepolis, Armenia, 118, 185 Kainepolis/Qena, fort, 130, 131 Kaiseraugst, see Castra Rauracense Kelcit, see Satala Koblenz, see Confluentes Köngen, see Grinario Kostolac, see Viminacium Lake Constance, see Bodensee Lambaesis, legionary base, xxii, 82, 139, 140, 141, 202, pl. 26 Lancaster, fort, 36, 158 landing places, 46–7, 99, 99–100, 107, 116, 171 Lauriacum/Lorch, legionary base, 24, 101, 105 Lautertal-limes, 59, 170 legions, 17–8, 28, 34–5, 38, 41, 51, 63, 73, 79, 81–2, 87–8, 96, 102–5, 122–3, 126, 132, 137–9, 141, 161, 167, 177, 182, 185, 191–2, 194 Lejjun, legionary base, 121, 123, 224 Lemanis/Lymne, fort, 153, 154, 154, 155 linear barrier, 55–91, 143, 161, 170, 173, 178, 195–6, 199–200, 210, 223–5 Lippe, River, 93, 172, 194 literary sources, 6–8, 14–32, 36, 39, 44, 48–9, 55–8, 61, 70, 76, 79–81, 89, 92–3, 96, 99, 102–3, 106, 112, 119, 126, 137, 144, 146, 148–9, 152–3, 158–60, 171, 181–91, 200–1 Lorch, see Lauriacum Lollius, Q., Urbicus, governor, 70, Lugio/Dunaszekcső, fort, 100, 101, 106 Luguvalium/Carlisle, fort and town, 62, 63, 70 Lussonium/Dunakömlőd, fort, 100, 101 Lymne, see Lemanis

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 239

Index 239 Magna/Carvoran, fort, 62, 68 Maia/Bowness-on-Solway, fort, 62–4 Main, river, 4, 56, 57, 178 Mainz, see Mogontiacum Malatya, see Melitene Mann, John, ancient historian, xix, 119, 155, 162, 197, 208 Marcomanni, people, 5, 15, 21, 24–5, 30–1, 35, 100, 101, 103, 105, 136, 146, 184, 186–7, 207 Marcus Aurelius, emperor, xv, 5, 7, 9, 21–2, 30–1, 35, 70, 76, 122, 152, 184–7, 199, 204 marshes, 3, 136, 159–60 Maryport, fort, 9,11 Mauretania, kingdom, 17, 138, 147 Mauretania Caesariensis, province, xxii, 36, 84, 138, 140, 141, 161, 171, 173–5, 177, 182, 185, 188–9, 201, 207 Mauretania Tingitana, province, xxii, 138, 143–4, 178, 184 Mautern, see Faviana Melitene/Malatya, legionary base, xxiii, 28, 113, 114, 137, 224 Mesarfelta, fort, 82, 83, 139, 140 Messad, see Dimmidi milecastles, 43, 63–9, 64, 69, 71, 80, 177, 190, 200 function of, 43, 63, 190, 200 Mitrovica, see Sirmium Moesia, province(s), xxiii, 17, 84, 92, 100, 102–4, 136, 159, 185–6, 188, 206 Mogontiacum/Mainz, legionary base, xxii, 57, 58–9, 93–5, 97, 116, 172 Moigrad, see Porolissum Moldavia, 84, 137 Mommsen, Theodor, historian, 4, 10, 195–6 Moore, Michael J., artist, 42, 43, 69, 75 mountain frontiers, 3, 133–45 Mureş, river, 133–4, 134 Mursa/Osijek, legionary base, 100, 101 Myos Hormos/Quseir, port, 130, 131, 152 Nabataean kingdom, 122, 152, 185 Neckar, river, 4, 22, 48, 57, 88, 97, 178, 200 Neckarburken, fort, 57, 77 Nero, emperor, xv, 182, 211 Netherby, see Castra Exploratorum

Neuss, see Novaesium Newcastle, see Pons Aelius Nijmegen, see Noviomagus Nile, river, 3, 16, 18, 126, 129, 130, 144, 152, 184 Noricum, province, xxii, 24, 92, 102–3, 105, 108, 159, 167, 170, 172 North Sea, xxii, 17, 153–8 Notitia Dignitatum, document, 6, 8, 26, 28, 36, 95, 106, 120, 122, 128, 143, 152, 156, 162 Novae/Sistov, legionary base, xxiii, 102, 105, 134, 224 Novaesium/Neuss, legionary base, 93, 84 Noviodunum/Isaccea, fleet base, 102, 151 Noviomagus/ Nijmegen, legionary base, 93, 94, 96 numerus, army unit, 34, 36, 43, 77, 87, 173 Numidia, province, xxii, 132, 138, 142, 167, 171, 174, 182 oases, 4, 118, 125, 174, 178, 211 Oberwinterthur, see Vitudurum Odenwald, Germany, 58–60, 76, 88, 133, 145, 178 Oescus/Gigen, legionary base, 100, 106, 109, 134, 224 officers, Roman, 37–8, 128, 131, 188–9, 210 Old Kilpatrick, fort, 71, 72, 79 Olt, river, 6, 84, 134, 136 Osijek, see Mursa Osrhoene, kingdom/province, 21, 33, 113, 118, 119 Osterburken, fort, 57, 77, 87, 223 ostraca, 8, 188, 189, 190, 191, 195 Othona/Bradwell, fort, 153, 154 Outer Limes, Germany, 76–9, 79, 85–90, 173, 178–9, 200, 223–4 outposts, 65, 72, 81, 122–3, 126, 129, 139, 141, 176–7, 190–2 palisade, 56–9, 61, 76–7, 85, 89 Palmyra, city and military base, 44, 113, 120, 126 Pannonia, province(s), xxii–xxiii, 14, 26–7, 37, 55, 100, 103–5, 115, 136, 167, 173, 188, 190, 195, 201, 206 papyri, 38, 120, 122, 190–1, 195

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 240

240 The Frontiers of Imperial Rome Parthia/Parthians, 4, 15, 24, 114, 118, 137, 178, 186, 193–4 Passau, see Bodiorum and Castra Batava Persia/Persians, 4, 5, 22, 120, 193 Petra, city, 121, 122, 180 Petronell, see Carnuntum Pevensey, see Anderita Pfünz, see Vetoniana Phasis/Poti, fort, 20, 25, 40, 148, 149, 152 Philip, emperor, xv, 153 Picts, people, 82, 148, 162, 184 piracy/pirates, 35, 146, 149, 150–2, 156 Pons Aelius/Newcastle, fort, 28, 62, 63–5 Porolissum/Moigrad, fort, 134, 135, 135, 136, 224 Portchester, see Portus Adurni Portus Adurni/Portchester, fort, 153–5, 154, 156, 223 Potaissa/Turda, legionary base, 134, 136 Poti, see Phasis Poulter, John, surveyor, 63, 65, 67, 159 Praetorian Guard, 17, 34, 35, 44 Primis/Qasr Ibrim, fort, 129, 130 Probus, emperor, xvi, 22, 48 Pselkis/Dakka, fort, 129, 130 Pseudo-Hyginus, writer, 39, 51 Qasr Bshir, fort, 121, 123, 124, 124, 224, pls. 22 and 23 Qasr el-Hallabat, fort, 48, 121, 122, 124, 224 Qift, see Coptos Quadi, people, 5, 21, 24, 30, 101, 103, 105, 112, 184–5, 187–8, 190, 207 Queseir, see Myos Hormos Raetia, province, xxii, 12, 27, 32, 56, 58–9, 76–9, 86, 88, 102, 182 raiding, 43–4, 96, 106, 120, 124–5, 128–30, 132, 146, 152, 188–90, 195, 201–2, 206 Rainau-Buch, fort, 57, 77, 224 Rapidum, fort, 140, 142 Ratiaria/Archar, legionary base, 103, 106, 134 Ravenscar, tower, 47, 158 Rawaffa, 121,122–3, 130, 185 Reculver, see Regulbium Reichslimeskommission, 10, 56 Red Sea, xxiii, 3, 16, 129, 130, 132, 144, 152, 175, 184

Regensburg, see Castra Regina regulations, frontier, 32, 61, 65, 188, 195, 207, 211 Regulbium/Reculver, fort, 153, 154 Remada/Tillibari, fort, 126, 127 Rhine, river, xvii, 10, 14–17, 19, 23–4, 27, 31, 35, 46–7, 56, 57, 58–9, 79, 92, 94, 97, 112, 158–9, 161, 167, 170, 173, 175, 189, 194, 206, 210 Richborough, see Rutupiae Risingham, see Habitancum river frontiers, 3, 18, 86, 92–117, 125, 174 (see also individual rivers by name) roads, 56, 65, 69–70, 73, 75, 78, 80, 82, 88, 93, 105, 120, 123, 125, 129, 139, 142–4, 152, 161–3, 172, 190, 193, 201, 204 Rome, 34–5, 55, 70, 84, 171 Rough Castle, fort, 72, 74, 94 routes, 59, 61, 81–4, 93,100, 104–6, 114–15, 123, 141, 144, 172, 194, 201–2 Rushworth, Alan, archaeologist, 175, 201 Rusovce, see Gerulata Rutupiae/Richborough, fort, 46, 153, 154, 154 Saalburg, fort, 10, 57, 58, 78, 224, pl. 13 Sahara Desert, 3, 126, 139, 141, 171 Sala, fort, 143, 146, 225 Saldae/Bejaia, city, 51, 140, 142 Samosata/Samsat, legionary base, xxiii, 113, 114, 119, 180 Samsat, see Samosata Saracens, people, 124–5 Sarmatians, people, 104, 137, 186, 190 see Iazyges Satala/Kelkit, legionary base, xxiii, 28, 114, 118, 137, 149, 150, 180, 224, pl. 21 Saxons, people, 146, 153, 156 Saxon Shore, 47–8, 153–7, 223 sea frontiers, 3, 146–58 Segedunum/Wallsend, fort, xxi, 62–4, 223 Seguia Bent el-Krass, 82–4 Septimius Severus, emperor, xv, 21, 35, 48, 119, 120, 122, 126, 132, 139, 141, 144, 148, 161, 162, 171, 177, 181, 191, 193, 199, pl. 1c Sevastopol, see Chersonesus Taurica Severus Alexander, emperor, xv, 22, 96 Sexaginta Prista/’Sixty Ships’, fleet base, 102, 106, 134, 224

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 241

Index 241 signalling, 72–3 Silistra, see Durostorum Singara, city, 113, 119, 191 Singidunum/Belgrade, legionary base, 101, 105 Sirmium/Mitrovica, legionary base, 35, 100, 108, 224 Soura/Souriya, fort, 113, 174 Souriya, see Soura South Shields, see Arbeia Spain, 14–6, 56 Stanegate, road, 65, 69, 75, 88, 190 Strasbourg, see Argentorate subsidies, 186–7 Sucidava/Celei, fort, 106, 134 Suebi, people, xvii, 15, 104 Svistov, see Novae swamps, 3, 159–60 Sybyllenspur, see Lautertal-limes Syene/Aswan, fort, 129 Syria, province, xxiii, 17, 35, 37, 55, 105, 114, 119, 161, 179–81 Szentrendre, see Ulcisia castra Szôny, see Brigetio Tacitus, Cornelius, historian, 6, 7, 8, 16–19, 24–5, 31–3, 36, 44, 61, 92–3, 95, 97, 99, 102, 126, 137, 144, 146–7, 149–50, 159, 181–2, 184–5, 187, 189, 195 Taunus Mountains, Germany, 56, 58–9, 77, 178 Taurunum/Zamum, fleet base, 101, 102 Tay, river, 71, 72, 148, 161, 171 Thamusida, fort, 143, 225 Thanaramusa Castra, fort, 140, 142 Theodosius, count, 8, 27, 212 Theodosius I, emperor, xvi, 8, 187 Thubunae, fort, 139, 140 Tiberius, emperor, xv, 17, 44, 95, 102, 115, 139, 170, 181, 185–6, 194 Tigris, river, 119–20, 178, 191 Tillibari, see Remada Tisavar/Ksar Rhilane, fortlet, 126, 225 Tisza, river, 101, 136, 159, 160 Titus, emperor, xv, 148 towers, 26, 28, 43, 43–4, 47, 58–9, 63–6, 64, 72, 76, 77–9, 80, 83, 86–7, 97, 100, 106, 111–12, 111, 115, 122, 134, 135, 136, 139, 142–3, 158, 161, 170–1, 175, 177,

179, 188, 190–1, 193, 195, 196–8, 208, 223–4 function of, 43–4, 59, 63, 106, 112, 122, 136, 139, 158, 188 Trabzon, see Trapezus trade, 77, 84, 119, 202–4 Trajan, emperor, xv, 36, 40, 55, 58–9, 62, 70, 84, 102–3, 116, 118–20, 123, 125, 133, 137, 148, 150, 182–3, 185–6, 199, 201, 204–5, 210 Trajan’s Column, 9, 44–6, 59, 106, 115, 170, 188, 196, pls. 2–6 Transalutanus, limes, Romania, 84–5, 134,136, 170, 179 Transmariscus/Tutraken, fort, 106, 108 Transylvania, Romania, 84, 102, 133, 136 Trapezus/Trabzon, fort, 28, 55, 120, 149, 150, 152 treaties, 5, 30–2, 137, 178, 184–6, 207 Trier, see Augusta Treverorum Tripolitania, 6, 126–8, 127, 139, 179, 188 Troesmis/Iglita, fort, xxiii, 101, 111 Turda, see Potaissa Turnu Severin, see Drobeta turrets, see towers Tutraken, see Transmariscus Tyne, River, 63, 86, 179, 193 Udruh, legionary base, 121, 123, 224 Ulcisia castra/Szentendre, fort, landing place and towers, 101, 107, 112 Umm el-Djemel, city, 121, 125, 224 Valens, emperor, xvi, 48–9 Valentinian I, emperor, xvi, 24, 27–8, 39, 48–9, 98, 100, 109, 112, 115, 125, 158, 161, 171, 176, 181, 189, 194, 212, pl. 1e Valerian, emperor, xvi, 120 Vallum, 67–8, 75, 80, 89, 199 Vandals, people, xvii Varus disaster, 12, 93, 115, 167, 208 Vechten, see Fectio Vegetius, writer, 49–50, 200 Veldidana/Innsbruck-Wilton, fort, 97, 109 Velsen, fort, 94, 95 Vemania/Bettmauer, fort, 106, Vercovicium/Housesteads, fort, 41, 62, 223 Vespasian, emperor, xv, 58, 102, 137, 139, 147, 184–5, 206

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:24 Page 242

242 The Frontiers of Imperial Rome Vetera/Xanten, legionary base, xxii, 25, 44, 93, 94–5, 94, 97, 172 Vetoniana/Pfünz, fort, 57, 77 Vienna, see Vindobona Viminacium/Kostolac, legionary base, xxiii, 105, 134, 224 Vindobona/Vienna, legionary base, xxii–xxiii, 101, 102, 105, Vindolanda, fort, 62, 63, 199, 223 writing tablets, 8, 41, 50, 183, 188, pl. 8 Vindonissa/Windisch, legionary base, 8, 96, 97, 124 Visegrád, fort, 101, 108 Vitudurum/Oberwinterthur, fort, 96, 97 Wales, 69, 147, 161, 223 Wallachia, Romania, 84, 86, 87, 133, 137 Wallsend, see Segedunum Walton Castle, fort, 153, 154, 155–6

water, affecting the location of frontiers, 16–7, 92, 118, 146, 174, 192–3, 210–1 Wetterau, Germany, 56, 58–9, 61, 70, 204 Whittaker, C R, ancient historian, xviii, 16–7, 21 Woolliscroft, David, archaeologist, 65, 77, 85–6, 175, 178, 190, 209 Xanten, see Vetera York, see Eburacum Yorkshire coast towers, 47, 111, 158 Zamum, see Taurunum Zarai, fort, 84, 140, 142, 202–3 Zeiselmauer, see Austura Zeugma/Belcis, legionary base, xxiii, 113, 114, 116, 119, 180 Zigana Pass, 114, 137

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 243

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 244

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 245

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 246

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 247

The Frontiers of Imperial Rome_Layout 1 15/07/2011 09:17 Page 248