Exercitus Moesiae: The Roman Army in Moesia from Augustus to Severus Alexander 9781407314754, 9781407344522

Citation preview

________ Conor Whately is an associate professor of Classics at the University of Winnipeg. He has published book reviews, encyclopaedia entries, journal articles and book chapters on various aspects of Rome at war from Augustus to Heraclius. His first book examines battle narrative in Procopius.

BAR S2825 2016  WHATELY  EXERCITUS MOESIAE

This book is a military organisational history of the Roman Empire on the lower Danube from the emperor Augustus (r. 27 BC–AD 14) to the emperor Severus Alexander (r. AD 222–235). Using a diverse body of evidence, from Roman military diplomas to funerary inscriptions and literary sources, the book looks at changes in troop disposition involving the legions, auxiliary units, the vexillations and the naval units based in Moesia Superior and Inferior, and around the northern and western coasts of the Black Sea. The book also examines the emplacement of the region’s units, and contextualises both the disposition of troops and their emplacement in terms of regional strategy and the strategy of the empire as a whole. Besides the discussion and analysis, the book also includes detailed maps of the region and useful tables that summarise the results.

Exercitus Moesiae The Roman Army in Moesia from Augustus to Severus Alexander

Conor Whately

BAR International Series 2825 B A R

2016

Exercitus Moesiae The Roman Army in Moesia from Augustus to Severus Alexander

Conor Whately

BAR International Series 2825 2016

Published in 2016 by BAR Publishing, Oxford BAR International Series 2825 Exercitus Moesiae © Conor Whately 2016 Cover image The banks of the River Danube looking north from the area around Novae in ancient Moesia, modern day Bulgaria. The Author’s moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781407314754 paperback ISBN 9781407344522 e-format DOI https://doi.org/10.30861/9781407314754 A catalogue record for this book is available from the British Library

BAR titles are available from:

Email Phone Fa x

BAR Publishing 122 Banbury Rd, Oxford, ox2 7bp, uk [email protected] +4 4 (0)1865 310431 +4 4 (0)1865 316916 www.barpublishing.com

Table of Contents PREFACE AND ACKNOWLEDGEMENTS..............................................................................................v LIST OF ABBREVIATIONS......................................................................................................................vi INTRODUCTION ........................................................................................................................................1 The Topic...............................................................................................................................................1 The Physical Geography of the Moesias ...............................................................................................2 Methodology and Evidence ...................................................................................................................3 Methodology ..................................................................................................................................3 Evidence .........................................................................................................................................3 Literary Evidence ....................................................................................................................4 Material Evidence ...................................................................................................................4 Epigraphic Evidence ...............................................................................................................4 Unit Names..............................................................................................................................5 Modern Scholarship on the Roman Military .........................................................................................5 The Organisation of the Roman Military .......................................................................................6 The Roman Military and the Lower Danube..................................................................................6 Chapter Organisation .............................................................................................................................7 PART I: THE DATA CHAPTER 1: The Lower Danube from 27 BC to AD 81 .........................................................................11 Introduction .........................................................................................................................................11 Background: Military History..............................................................................................................11 Augustus.......................................................................................................................................11 Nero..............................................................................................................................................12 The Year of the Four Emperors and Vespasian............................................................................12 Troop Movements from Augustus to Titus..........................................................................................12 Legions .........................................................................................................................................12 Julio-Claudian Dynasty Auxiliae ..................................................................................................16 The Evidence from the Diplomata ...............................................................................................19 Poorly Attested Flavian Dynasty Auxilia .....................................................................................21 Concluding Thoughts ..........................................................................................................................21 CHAPTER 2: The Lower Danube from AD 81 to AD 161 .......................................................................23 Introduction .........................................................................................................................................23 Background: Military History..............................................................................................................23 Domitian.......................................................................................................................................23 Trajan ...........................................................................................................................................23 Hadrian and Antoninus Pius .........................................................................................................24 Legions from Domitian through Antoninus Pius.................................................................................24 Auxiliae: Moesia Superior ...................................................................................................................27 Domitian.......................................................................................................................................27 Trajan ...........................................................................................................................................28 Hadrian .........................................................................................................................................31 Antoninus Pius .............................................................................................................................32 Auxiliae: Moesia Inferior.....................................................................................................................32 Domitian.......................................................................................................................................32 Nerva ............................................................................................................................................34 Trajan ...........................................................................................................................................36 Hadrian .........................................................................................................................................39 Concluding Thoughts ..........................................................................................................................41 CHAPTER 3: The Lower Danube from AD 161 to AD 235 .....................................................................42 Introduction .........................................................................................................................................42 Background: Military History..............................................................................................................42 Marcus Aurelius ...........................................................................................................................42 Commodus ...................................................................................................................................42 iii

Exercitus Moesiae Septimius Severus to Severus Alexander ....................................................................................42 Troop Movements from Marcus Aurelius to Severus Alexander ........................................................43 Legions .........................................................................................................................................43 Auxiliae: Moesia Superior ............................................................................................................44 Auxiliae: Moesia Inferior..............................................................................................................48 Concluding Thoughts ..........................................................................................................................49 CHAPTER 4: Vexillations and the Black Sea ...........................................................................................50 Introduction .........................................................................................................................................50 Vexillations..........................................................................................................................................50 General .........................................................................................................................................50 The Lower Danube .......................................................................................................................51 The Black Sea......................................................................................................................................53 Moesian Land-Based Soldiers on the Black Sea ..........................................................................54 The First Century..........................................................................................................................54 Tyras and Olbia ............................................................................................................................56 Chersonesus and the Crimea ........................................................................................................57 The Moesian Navy .......................................................................................................................59 Concluding Thoughts ..........................................................................................................................60 CHAPTER 5: Troop Emplacement in the Moesias ...................................................................................63 Introduction .........................................................................................................................................63 Legions ................................................................................................................................................64 The Legions from Augustus to Titus ...................................................................................................64 The Legions from Domitian to Antoninus Pius...................................................................................64 The Legions from Marcus Aurelius to Severus Alexander..................................................................65 Auxiliaries ...........................................................................................................................................66 Introduction ..................................................................................................................................66 Moesia ..........................................................................................................................................66 Moesia Superior ...........................................................................................................................67 Moesia Inferior .............................................................................................................................68 Vexillations..........................................................................................................................................70 Troop Emplacement and Logistics ......................................................................................................71 Concluding Thoughts ..........................................................................................................................73 PART II: RESULTS AND DISCUSSION CHAPTER 6: Rome’s Strategy on the Lower Danube ..............................................................................77 Roman Strategy ...................................................................................................................................77 The Size and Strength of the Moesian Army................................................................................78 The Moesian Strategy...................................................................................................................82 Moesian Military Units and the Wider Roman Empire ................................................................85 Concluding Thoughts ..........................................................................................................................85 CONCLUSION: The Value of Traditional Roman Military History ........................................................86 APPENDIX 1: Summary Tables of Legions and Auxiliary Units..............................................................88 APPENDIX 2: List of all the Military Units Based in Moesia ..................................................................92 APPENDIX 3: List of Moesian Diplomas .................................................................................................95 APPENDIX 4: List of Ancient Sources Used.............................................................................................98 APPENDIX 5: The Numeri ......................................................................................................................107 APPENDIX 6: Maps.................................................................................................................................108 BIBLIOGRAPHY.....................................................................................................................................113 INDEX......................................................................................................................................................122

iv

Preface and Acknowledgements This book started life as an MA thesis, which was completed at McMaster University in September of 2005. After its completion, I pursued a PhD at Warwick in the UK on late antique things (Procopius). The original plan had been to revise the MA thesis for publication by the time I had finished that PhD. One thing led to another, however, and it would be over a decade before the revisions were finally complete. Although this has taken a lot longer than I would have liked, and this has meant that this project has been hanging over my head for quite a while, the upside is that I was able to take advantage of much important work that has been undertaken on the subject matter, the Roman military in Moesia, between the thesis’ completion and the day that I sat down to write this acknowledgement. I discuss some of that work in the introduction below. Perhaps two of the most important developments have been the success of academia.edu, which has given me access to a plethora of articles and chapters from hard-to-access collections, which in the past would have been very difficult to track down, especially here in Winnipeg; and, the Epigraphik-Datenbank Clauss-Slaby has made life much easier, and again for the same reasons – I can access thousands upon thousands of inscriptions from Winnipeg, many of which would have been darn-near impossible to get otherwise. When I sat down to research the thesis in the first place (2004), my gathering of data consisted of me parking myself in a cubicle in Mills library with physical copies of CIL and AE in front of me, combined with an occasional trip to Robarts and the ROM library in Toronto. The extra time and access to resources has improved this project immeasurably. One last benefit of the extra time has been that I was given the opportunity to visit the part of the world where much of this book is set. In 2012, the twenty-second Limes Congress was held in Bulgaria, and thanks to a start-up grant from the University of Winnipeg, I was able to attend. Ultimately, then, this extra time has been quite a boon. Admittedly, there is still more I wish I could have done – my comprehension of Bulgarian, Polish, Russian, and Serbian did not suddenly materialize between the submission of the thesis and the finishing of these revisions – but I am at least vaguely satisfied with the end result, even if the word “satisfied” seems a bit of an exaggeration. Since this is my second book, and given that in the first I managed to thank nearly everyone under the sun, I am going to keep this portion short. The subject of this book was suggested by my MA supervisor, Evan Haley, who took me to the Classics department’s copy of the Barrington Atlas and pointed out to me the maps devoted to the lower Danube and the myriad of fortifications. He also wrote the recommendation that led to this revised-version’s publication with BAR. To Evan Haley, then, is owed my biggest thanks – and for his insightful comments back then. I must also thank the examiners of the thesis, Alexandra Retzleff and Claude Eilers, as well as the other faculty members of the department at McMaster at the time for the help, whether direct or indirect. It goes without saying too that my graduate peers made the project just that much more bearable. Allison Cooley gave me some feedback as I was about to embark on the revisions for this book near the end of my PhD at Warwick, and Michael Whitby, my PhD supervisor, read the occasional tidbit here and there, and even though it fell well outside his area of interest and expertise. Aspects of this book were presented at the CAC in 2009 in Vancouver, which proved a boon less for the feedback on this material than for my career. Basically, it led to my employment by the University of Winnipeg – and that institution deserves my thanks for funding various activities associated with my research for this book. Not long after the CAC conference in Vancouver, I presented another portion of the work to some fellow Classicists here in Winnipeg from the Classics’ departments at the UofM and the UofW. Those who attended the session at Limes XXII where I presented some of this material deserve my thanks too, as does Matt Maher who read what became the published version for the Limes Congress proceedings. Jonathan Eaton and Matt Gibbs deserve special consideration for their feedback on the manuscript in its entirety. Gabe Moss at the Ancient World Mapping Center deserves thanks for providing the wonderful maps. At BAR, the editorial team have proved invaluable. Finally, I want to thank my family for their love and support. Hannah, Ella, and Don were with me at the end, and as always helped me in immeasurable ways. Given that my mom, dad, and sister were with me at the beginning (2003), along with my aunt Jan and Nan (my roomies for one year), however, I want to dedicate this book to them.

v

List of Abbreviations The abbreviations that I have used for ancient authors are those found in the Oxford Classical Dictionary (OCD), while those used for journal articles are those found in Année philologique. Most of the abbreviations below are those found in the Epigraphik-Datenbank Clauss-Slaby, with some exceptions. ADBulgar =E. Kalinka, Antike Denkmäler in Bulgarien, Vienna 1906 (ND Nendeln 1976) AE =L'Année Épigraphique AMN =Acta Musei Napocensis CBI =E. Schallmayer - K. Eibl - J. Ott - G. Preuss - E. Wittkopf, Der römische Weihebezirk von Osterburken I: Corpus der griechischen und lateinischen Beneficiarier-Inschriften des Römischen Reiches, Stuttgart 1990 CERom =C.C. Petolescu, Cronica epigrafica a României CIL =Corpus Inscriptionum Latinarum CIMRM =M.J. Vermaseren, Corpus Inscriptionum et Monumentorum Religionis Mithriacae, 2 Bd., Den Haag 19561960 GeA =C. Schmidt Heidenreich, Le glaive et l'autel. Camps et piété militaires sous le Haut-Empire romain, Rennes 2013 Iatrus =Iatrus-Krivina VI. Spätantike Befestigung und frühmittelalterliche Siedlung an der unteren Donau. Ergebnisse 1992-2000, Mainz 2007 IBulgarien =V. Besevliev, Spätgriechische und spätlateinische Inschriften aus Bulgarien, Berlin 1964 IDR =Inscriptiones Daciae Romanae, Bucharest 1975IDRE =C.C. Petolescu, Inscriptiones Daciae Romanae. Inscriptiones extra fines Daciae repertae, Bucharest 1996IGRR =R. Cagnat, Inscriptiones Graecae ad Res Romanas Pertinentes, Paris 1906/1927 (Bände 1, 3 und 4) IIFDR =E. Popescu, Inscriptiones intra fines Dacoromaniae repertae Graecae et Latinae anno CCLXXXIV recentiores, Bucharest 1976 ILBulg =B. Gerov, Inscriptiones Latinae in Bulgaria repertae, Sofia 1989 ILD =C.C. Petolescu, Inscriptii latine din Dacia (ILD) - Inscriptiones latinae Daciae, Bucharest 2005 ILJug =Inscriptiones Latinae quae in Iugoslavia ... repertae et editae sunt, Ljubljana 1963-1986 ILS =Inscriptiones Latinae Selectae IMS =Inscriptions de la Mésie Supérieure, Belgrade 1976InscrIT =Inscriptiones Italiae, Rome 1931IOSPE2 =Inscriptiones antiquae Orae Septentrionalis Ponti Euxini graecae et latinae IScM =Inscriptiones Scythiae Minoris Graecae et Latinae, Bucharest 1980Kayser KJ Krier

=F. Kayser, Recueil des inscriptions grecques et latines (non funéraires) d'Alexandrie impériale, Cairo 1994 =Kölner Jahrbuch für Vor- und Frühgeschichte =J. Krier, Die Treverer außerhalb ihrer Civitas. Mobilität und Aufstieg, Trier 1981 vi

List of Abbreviations LNCh RGZM RMD RMR SCI SCIVA

=E.I. Solomonik, Latinskie nadpisi Chersonesa tavriceskogo, Moscow 1983 =Römische Militärdiplome und Entlassungsurkunden in der Sammlung des Römisch-Germanischen Zentralmuseums, Mainz 2004 =Roman Military Diplomas =Roman Military Records =Scripta Classica Israelica =Studii si cercetari de istorie veche si arheologie

vii

forces that at that time still needed work. 2 Almost forty years later, Michael P. Speidel was asked to do the same and take stock of the current “state of the art” in Roman military studies. 3 Speidel noted that much of what Birley had called for had been accomplished, and in many cases by his own students. 4 Yet, he also acknowledged that there was still a lot of work remaining. And so Speidel decided to highlight those aspects of the organisation of the army that still need work. 5 After calling for the collation of the many disparate articles concerned with the army, and recognising the value of monographs, he provided a table with those topics “that will richly repay research, especially if they come with a catalogue of the sources” (Speidel 1992a: 16). The topic that Speidel directed the greatest amount of attention to was the provincial armies. Speidel said: Some topics … [include] a survey of the various provincial armies, such as the exercitus Moesiacus or Syriacus. What were they called? How were they looked at from Rome? What clout did they have there? How did they differ from each other in name, strength, structure, quality, weapons, equipment and fighting techniques, in recruitment, administration, religion, buildings, fortifications, and in their role in the field army? How was their uniformity maintained? How did they cooperate or even communicate with each other? Some of the answers are found in Tacitus, Suetonius and Vegetius, but the inscriptions and papyri would also tell much. 6

INTRODUCTION THE TOPIC Despite the popular image of a monolithic Imperial Roman Army, in the nearly 250 years from Augustus to Severus Alexander the largest institution on the state’s payroll underwent considerable change. What started out as a campaigning army of amateurs was transformed into a largely standing army of professionals by the end of that period. This development was gradual, and different parts of the Imperium Romanum witnessed this transformation of the army into a permanent force at different times. Indeed, scholars are increasingly emphasizing the regional character of the Roman armed forces. 1 The gist of the argument is that we should see the armed forces as Roman armies, a collective, composed of the various provincial forces, such as the British army, the Syrian army, and the Moesian army, rather than as a single Roman Army. It is with this view in mind that we concentrate on one corner of the Roman World, the Lower Danube and Crimea, a region which is reasonably well represented in the sources, the principal of which are the inscriptions. This will, in many ways, be a traditional book about the Roman military, with the organisation of military units, their movements, and the strategy/ies they employed the emphasis. This is not a book about the military as a social organisation, though that is what it was. As such, as important and interesting as the subject is, we leave it out of this discussion with one caveat: I hope that the information contained herein should go some way towards delineating the Roman military communities of the lower Danube, insofar as it will identify the units in the region, note when they were there, and identify where possible where they came from. This information, in turn, can then be used to delve into the wider world of the Roman military in Moesia as a social organisation, a topic worth pursuing.

Nearly a decade after the publication of Speidel’s article, James wrote a wide-ranging paper, ostensibly on Roman military archaeology in Britain, but in fact something much more substantial. 7 In that paper, like Birley and Speidel before him, James highlighted a number of problems with Roman military studies as they were conducted at that time, and highlights some future avenues for research. James was interested in a number of issues, including the following: how was the army constituted; and did the army exist at all? In the latter case, he drew attention to the fact that no word in Latin, the language of the Roman military, existed for the army as whole, though there were plenty of terms in use for specific armies, units, and

Although much important work has been done to catalogue the region’s units and set out their history, document and discuss some of the important warfare, and to explain the strategic situation in part or as a whole, rarely has there been work that does all of this. Thus, this book offers a synthesis, of sorts, and is more than a catalogue of units and their histories. Although that sort of discussion and material are a big and important part of this project, and though previous studies that have done just that have had a marked impact on this one, this book goes further, situating the changing disposition of Moesia in the context of empire-wide Roman military organisation and strategy.

2

This intriguing paper was republished as the opening chapter of Birley’s collection of essays, The Roman Army Papers 1929-1986 (1988). Pp 3-11. 3 It was originally published in 1989 (BIAL 26), and republished in Speidel’s second collection of essays, Roman Army Studies Volume 2 (1992a). 4 In an enlightening review article Lendon (2004: 441-449) made some similar pronouncements – although he did not specifically outline those particular topics that should garner more attention – in regard to the Roman army which for him, and rightfully so, was “essentially Roman”, “part of Roman society”, and “an aspect of Roman culture.” 5 Speidel 1992a: 13-20. The title of the paper is in fact, “Work to be done on the Organization of the Roman Army”. 6 Speidel 1992a: 16. 7 James 2002.

The impetus for this book is this. In 1952 at a conference for epigraphists Eric Birley presented a paper in which he set out those aspects of the armed 1

Speidel 1992: 16; James 2002.

1

Exercitus Moesiae collections of soldiers like exercitus, legiones, auxiliae, numeri, and milites. 8 Instead, and essentially following Speidel, James argues that attention should be drawn to provincial armies, and that what we had was not one, monolithic army, but several provincial armies, among other things. To get back to Speidel’s quote then, of those two armies singled out by Speidel, that is the Syrian and Moesian armies, it is the latter that, until fairly recently, had shown more signs of neglect, though more on this later.

people and whether we should consider it as an institution, albeit one that played an important role in the local landscape. 11 On the other hand, and again as noted above, it is hoped the information provided here could lay the foundation for such an approach, hints of which can be found in the work of Alexandrov and Boyanov, for instance. 12 As a parallel, where Alston included discussion of the garrison of Roman Egypt in a book largely concerned with the army’s social role, 13 this book only focuses on the garrison, as well as the concomitant issues of strategy and internal security, in the hopes that by doing so it could lay the foundation for something on that issue.

As noted, Speidel calls for more work than the size of this book permits. Thus, we must be selective. What follows is a history of the movement of the various legions and auxiliary units that passed through, and remained in, the Moesias from 29 BC to the demise of Severus Alexander nearly 270 years later in AD 235. In this book we will also cover in some detail troop emplacement, and there will be a survey of the battles both won and lost in the Moesian provinces. As a result, we should be able to address the questions: “What were they called?”; and to some extent, “How did they differ from each other in name, strength, structure?” Select aspects of the issues of “quality, weapons, equipment and fighting techniques” and the differences “in recruitment, administration, religion, buildings, and fortifications” 9 are beyond our scope, though we will have recourse to the topic of fighting techniques, administration, and fortifications in the last chapter.

While the focus is on troop dispositions and movements, this is not simply a catalogue of military units. Rather, the aim is to chart the changes in disposition over time with a view to uncovering, where possible, not only how the provincial army/ies changed, especially in light of changing dynamics in the Pontic-Danubian region, but also why. After setting out these changes, then, we will take a look at what it all means, which entails discussing the complex and contentious issue of strategy. Wilkes noted, “the movements and activities of Danubian legions throughout the early Empire pose many problems that cannot properly be examined in a local context”. 14 We will bear this in mind as we proceed: although the focus is on the Moesian army, we will aim to situate the Moesian army within its context, both in terms of the Roman military empire-wide (Pannonias, Thrace, Dacia, and beyond), and the foreign policy concerns of the Roman state (conflicts – Trajan’s Dacian wars, Parthian war, etc.).

The goal of this book is to set out the organisational details of the Moesian provincial army, and a number of theoretical ideas underline this approach: the army as an institution, military community, and organisation, all covered by James. 10 We will not be able to get into important aspects of those approaches, for instance, and as noted above, we will not be addressing whether we should see the Moesian army as comprised of

This introduction is organised as follows. We begin with an overview of the physical geography of the region. Then, after a brief note on what evidence we have for troop movements in the region, from the literary to the material and the epigraphic, we survey the relevant scholarship on everything from legionary and auxiliary organisation to the military on the lower Danube, before finishing off with an outline of the organisation of this book.

8

James 2002: 38-39; cf. Whittaker 1994: 68. That leaves the following questions: (1) “How were they looked at from Rome?”; (2) “What clout did they have there?”; (3) “How was their uniformity maintained?”; and (4) “How did they cooperate or even communicate with each other?” On the level of the provincial armies, these questions could all form the basis of future work (although this may not be possible for every provincial army based on the number of inscriptions, papyri, and the focus of the ancient authors). However, if we look at the army as a whole, there are a few works available that address those questions. In regard to (1), and I am equating the emperor with Rome, I would direct the reader to Campbell’s (1984), The Emperor and the Roman Army; and for the relationship between the army and the individual soldier Stäcker’s (2003), Principes und miles. For a look at how the provincials regarded the army in the east, see Isaac’s (1992), The Limits of Empire: the Roman army in the east. In regard to (2), there is work to be done, but again Campbell’s and Stäcker’s works might be good starting points. In regard to (3), I would direct the reader to Goldsworthy’s (1996), The Roman Army at War; Goldsworthy’s (1997), The Roman army as a Community and in particular his own article, “Community under pressure: the Roman army at the siege of Jerusalem”; Lendon’s article (2004), “The Roman army now”; and Lendon’s book (2005), Soldiers and Ghosts. Finally, we come to (4) and as with (2), there is work to be done, although in regard to communications I would direct the reader to Woolliscroft’s (2001), Roman Military Signalling. 10 James 2002. 9

THE PHYSICAL GEOGRAPHY OF THE MOESIAS One of the principal aspects of the physical geography of the Balkans, and the Moesias in particular, is the mountainous terrain. 15 Modern Bulgaria, the northern two-thirds of which correspond to Lower Moesia, is divided into four distinct areas: the northern plateau, the east-west range of the Balkan Mountains (the ancient Haemus Mountains), the central and coastal lowlands, and the south-west Rhodope Mountains. The 11

On these and similar matters see Pollard (1996, 2000) and James (2011). 12 Alexandrov 2012, Boyanov 2012. 13 Alston 1995: 13-38. 14 Wilkes 2005: 154. 15 For a more detailed treatment see Batty (2007: 55-76).

2

Introduction northern part of the country borders the river Danube and many of the steep banks of the Danube rise to the low, fertile Danubian plain. South of this region the land rises gently towards the fragmentary Rhodope Mountains. The highest point is Mt. Musala at 2925m, which defines the southern border of Lower Moesia. Upper Moesia is also quite mountainous. The highest peak of the former Yugoslav republic of Serbia, where most of ancient Upper Moesia is situated, is 2656 metres. The Balkan Mountains, which dominate the rugged and wild landscape of eastern Serbia, rise, at their peak, to approximately 1800 metres; moreover, they demarcate the border of the modern nations of Romania, Bulgaria, and Serbia.

METHODOLOGY AND EVIDENCE Methodology The focus for this book will be two of the most prominent parts of the Roman military, the legions (legiones) and auxiliary units (auxiliae). Throughout the imperial period, roughly 27 BC to AD 235, the number of legionaries and auxiliaries under arms were broadly similar. These two classes of soldiers were heavily involved in the empire’s wars and conflicts besides whatever other activities in the provinces and on the frontiers arose that required military involvement. Our focus will be on setting out and explaining their organisation within the province, later provinces, of Moesia, during the imperial period, which for our purposes will include the years from 27 BC to AD 235. We will also discuss their prospective bases, the forts and fortifications that we find across the middle and lower Danube, but we will save this discussion to near the end of this book due, in large part, to issues with the evidence, which we return to below. Examining the organisation of the Moesian army will entail looking at where particular units were at particular times where possible during the imperial period, and we will be exploring changes in unit disposition collectively to get a sense of the broader changes over time, and to set these variations into context, both in terms of the broader region in which we find them (the Balkans), and in terms of the empirewide military system. To carry this out, we need evidence that provides the names of legionary and auxiliary units and that is datable. If we find enough of this evidence, then we can chart changes over time. Evidence that we cannot date with any certainty, however, will be discarded because of our desire for chronological markers. We are fortunate in that some of this evidence does exist, though much of it dries up as we move into the third century. Indeed, it is for that reason that the cut-off date for this book is around AD 235. We lack detailed, datable evidence after that, which would have allowed us to extend our study into the fourth century. 18

Much of the Lower and Middle Danube run through plains, which have very gentle inclines. There are also stretches of terrain where the banks of the river make it very difficult for embarkation; in particular, there is a 300 kilometre stretch below the Iron Gates where the cliffs reach heights of 100 metres or more. 16 In addition, the water level of the river drops to only a few metres in some spots, thence becoming difficult to navigate. There are also many marshy areas along the banks of the Danube River, which disappear west of the ancient city of Oescus. Thus, although there was a Moesian fleet, the Danube could pose significant problems for water-borne traffic. As regards the climate of the region, there is considerable variation among the Black Sea Coast, inland along the Danube, and in the mountainous parts of the provinces. In southern Bulgaria, the sea has a moderating effect, which creates a Mediterranean climate for the coast with warm and dry summers, and mild and wet winters. As one heads inland, there are extremes of temperature, with hotter summers and colder winters. The mean temperatures in July in Sofia and Belgrade are 21 and 22 degrees Celsius respectively, whereas in January the means are -2 and 0 degrees Celsius. The average annual precipitation in Sofia is 635 millimetres, whereas in Belgrade it is 610 millimetres. Thus, both along the Danube and a little ways inland, the climates are similar; however, we should note too that there can be heavy snowfalls in the winter in the more mountainous regions. There is one last point to note about modern Bulgaria and Serbia: cereal growing is a significant part of both of their economies, and there is also grape farming; however, the major farming region of Serbia corresponds more with Lower Pannonia than with Upper Moesia. This suggests that the Roman provinces lay close to or occupied land that was suitable both for farming crops and raising animals. Furthermore, we must also note that the nature of the Danube and its at times precipitous banks made crossing the river with troops or otherwise at these points very difficult. 17

Evidence There is a range of evidence that will be brought to bear in this book to illuminate the various movements of the region’s military units. This includes the literary evidence comprised almost exclusively of ancient histories; the material evidence, which ranges from the physical remains of forts to assorted pieces of Roman military equipment; and the epigraphic evidence, which includes everything from military diplomas to tile stamps and epitaphs, and which arguably makes up the most important category of evidence for this book. Finally, we come to the papyri and related materials like wooden tablets, and they have not survived in any

16

Lee 1993: 99. In regard to the regions where cereal-growing would have been possible and on a significant scale, I am thinking of the Pannonian, Macedonian, and Thracian plains, as well as the Crimean Peninsula. 17

18

3

See Rocco (2012) and Whately (2015).

Exercitus Moesiae useful fashion (or at least have not yet been found), with one notable exception, Hunt’s Pridianum, which is discussed at various points in this book.

– so permanent structures would, of course, be lacking. Getting back to Wilkes, we do not have much in the way of physical evidence for the Roman military deployment on the Danube until the aftermath of Trajan’s Dacian wars. 21 When we turn to emplacement in chapter five, we will return to some of this evidence.

Literary Evidence As Speidel noted, some of the answers to the questions on the military that he posed are indeed found in the ancient authors. Those authors which have the most to tell us about troop movements include Strabo, Velleius Paterculus, Florus, Josephus, Suetonius, Tacitus, who is undoubtedly the best source, Ptolemy, Cassius Dio, Herodian, the author/s of the Antonine Itinerary, Aurelius Victor, Eutropius, and the writer of the Historia Augusta. None of those authors, however, provides straightforward accounts of troop movements.

The next significant group under the category of material evidence for us is the visual evidence, and the standout here is Trajan’s Column. When we turn to Trajan’s Dacian Wars, or at least their outline, we will base our account using not only what limited literary evidence remains, but also the visual evidence that comes in the form of the narrative frieze that surrounds the column itself. The assigning of particular scenes from Trajan’s Column to events in the Dacian wars is a practice that has been adopted before, and here we will be basing our use of this material on the assumptions of both Rossi and Le Bohec. 22 Although it is worth asking whether the scenes found on Trajan’s Column depict historical reality in any demonstrable way, and so whether the column should be used in this manner, 23 Stefan’s 24 forceful attempts to rehabilitate the historical value of the monument are attractive. 25 While the monument should not be used uncritically, it is invaluable to our understanding of Trajan’s Dacian Wars. 26

When it comes to the specifics of troop movements then, the ancient historians are not terribly useful, especially when it comes to auxiliaries. 19 There are some exceptions for legionary movements, found in digressions (if you can call them that) in the works of Tacitus and Cassius Dio, which we will discuss in due course in the relevant chapters (one and three) below. On the other hand, where the historians are lacking when it comes to Roman military organisational matters, they are extremely useful when it comes to contextual matters, like the outlines of particular wars or the political machinations of individual emperors. Even here, though, our coverage of the relevant events in the empire’s military and political history is spotty. Few of the works of ancient history are complete, and the few that are happen to be least useful for our purposes. So, regrettably, we have only part of Tacitus’ Annals and Histories, and Cassius Dio’s History. Indeed, the portions of Dio’s History that would be most useful for second century events survive in little more than fragments and excerpts from the works of medieval Byzantine historians, like Xiphilinus and Zonaras.

Epigraphic Evidence In his chapter in the Oxford Handbook of Roman Epigraphy, Speidel notes a variety of issues of interest to scholars of the Roman military. 27 Some of those issues constitute what he calls basic aspects of the military and they include: names, history, internal organisation of individual types of units; military camps, fortresses, and their garrisons; troop movements and the composition of provincial armies army in action, like wars, and the composition of field armies; the existence and official names of wars and military expeditions; and aspects of combat and warfare. 28 Given that those aspects describe, quite well, the subject matter of this book, it should come as no surprise that inscriptions are invaluable to this study. The movements of the auxiliary units, for instance, are decipherable from Vespasian through Antoninus Pius primarily because of the diplomas. When we get to the reign of Marcus Aurelius, stone inscriptions are essential, although they will play a significant role throughout the book. 29

Material Evidence Under the category of material evidence, there are the physical remains of the fortifications that housed Moesia’s legionaries and auxiliaries. Although comparatively few have been excavated, there are dozens of forts, fortlets, and towers along the Danube and various major routes in the interior. To get some idea of the current state of the physical remains for fortifications, we need only consult Wilkes’ overview of the Roman Danube. He notes that physical evidence for military activity in the region from Augustus to Vespasian remains elusive. 20 Some of this is the result of the materials used. Perishable materials like wood, for obvious reasons, do not survive. It also has to do with the nature of the Roman presence. Although, as we will see, the Roman military moved into the region in increasing numbers from Augustus on, their presence was not significant until the reign of Domitian

21

Wilkes 2005: 152. Rossi 1971: 130-212; Le Bohec 2000: XI-XXII. 23 Wheeler 2010: 1190-1195, 1221-1227. 24 Stefan’s 2005: 503-694. 25 cf. Coulston 1989; Charles 2002. 26 For a detailed analysis of that other famous column, Marcus Aurelius’, a different sort of monument with a different set of problems, see Beckmann (2011). 27 Speidel 2015: 320. 28 Speidel 2015: 320. 29 There is a range of inscriptions that I shall be using, particularly soldiers’ or veterans’ epitaphs, in addition to any honorary inscriptions that a soldier, or unit, may have erected. 22

19

On the bias of Roman historians against auxiliaries see Gilliver (1996). 20 Wilkes 2005: 149-151.

4

Introduction One of the key issues with inscriptions involves dating; with some exceptions, it is often difficult to date particular inscriptions with much in the way of precision. 30 Some of the most significant exceptions are the diplomas, which we can often date to a particular day, month, and year. With that said, they are usually limited to the period between AD 54 and 205, and the auxiliaries at that. Still, these are some of the most valuable pieces of evidence. Diplomas (a modern name) are bronze copies of constitutions stored in Rome that gave citizenship to auxiliaries, though we have them for other kinds of soldiers too, including legions (very rare), praetorians (less rare), and members of the fleet (still rare). We often have multiple copies of diplomas for a single constitution. 31 Besides giving us a precise date, diplomas usually list all those (auxiliary usually) units in a province that had a man eligible for discharge. This, then, allows us to pinpoint which units were in a province at a given time.

names of units are potentially complicating factors in this study, for as we will see, units are not always reported the same way in each piece of evidence that we find them. While this is less of an issue with the legions, whose names, composed of legio, a number like V, and its title (cognomen) like Macedonica, so legio V Macedonica, are generally consistent across the period in question, besides some variations in the manner in which they are abbreviated on inscriptions, this is not the case with the auxiliary units. Rather, not only are auxiliary unit names often longer, but how they are recorded varies considerably from document to document – and the evidence is almost entirely epigraphic. 34 The components of the titulature that tend to be consistent are the unit type and their number, so ala (cavalry wing) or cohors (infantry cohort), and I or II, while the rest, the title, any honorific element (pia fidelis), a title designating equipment and weaponry (scutata), a tactical component title (equitata), and a title that denotes size (milliaria), tend not to be, though this varies too. The first three components, then, the type, number, and title, are the most common, as in cohors II Mattiacorum. Sometimes, however, the title, which might come from the name of a past commander (Caesariani), a tribal name (Aravacorum), a provincial name (Moesiaca), or the name of an emperor (Antoniniana), might vary over time. The cohors I Aelia Dacorum based in Britain, for instance, assumed a host of new titles over the course of the third century, which reflected changes in emperor. It became the cohors I Aelia Dacorum Antoiniana, and then the cohors I Aelia Dacorum Gordiana, the cohors I Aelia Dacorum Postumiana, and the cohors I Aelia Dacorum Tetriciana successively. 35 In that case, changes over time are not difficult to recognise given that the titles match reigning emperors. In other cases, however, relevant components, like whether a cohort was mounted (equitata), or whether it was composed of archers (sagittaria/sagittariorum), are left out in most of the evidence that we have for a particular unit. While this might be less problematic when we are concerned with the movements of a particular unit and the core elements (type, number, title) are present in most inscriptions, when there are multiple units with the same or a similar name and we are trying to determine which is which, or when we turn to the strategic objectives of the military (what sorts of soldiers did they deploy, like cavalry or infantry), ideally we would like to know the full title in all contexts. Unfortunately, this is not the case, and so we need to be aware of this as we proceed.

The various other inscriptions at our disposal, the dedicatory inscriptions, assorted brick and tile stamps, and the countless epitaphs can usually only be dated approximately, and sometimes to no more than a particular century, or half of one. Indeed, as Meyer notes, the “dating of inscriptions is notoriously difficult. With the exception of a small minority of epigraphic texts that have consular dates or that preserve imperial titulature in whole or in part, the dates of inscriptions are determined in general by the formulae and abbreviations used, the type and style of its decoration and lettering, and external archaeological and historical evidence”. 32 Thus, these sorts of inscriptions are less useful when it comes to telling us when a particular unit was in a place at a given time. We have, for instance, dozens and dozens of inscriptions from Durostorum that list soldiers and veterans of the legio XI Claudia. Based on the quantity alone, there is little doubt that the legion was based in the city for a prolonged period of time. What those inscriptions do not tell us, however, is when the legion moved in or out of the city. For that sort of information, we often have to resort to other kinds of evidence. The epitaphs are even less helpful, for while an abundance of epitaphs naming deceased former members of legion x in location b certainly suggests that legion x was in location b at some point, it could also be that, when their term of service came to an end, legion x’s former soldiers tended to retire there. Thus, while these other kinds of inscription can be useful for all sorts of things, when it comes to troop movements they can only be used with considerable caution. 33

MODERN SCHOLARSHIP ON THE ROMAN MILITARY

Unit Names Before we turn to the scholarship, one outstanding issue bears consideration, namely unit names. The

We finish this introduction with a survey of the relevant scholarly literature on the Roman military’s

30

See Meyer 2013: 10 Holder 2007b. 32 Meyer 2013: 10. 33 Note Wilkes’ (2005: 173) cautionary remarks about the use of stamps to elucidate troop movements. 31

34 Cichorius 1893: 1224-1228; 1900: 232-234; Cheesman 1914: 4549; Holder 1980: 14-63; and Saddington 1982: 169-177 35 CIL VII.818, 819, 820, 823; Cheesman 1914: 47.

5

Exercitus Moesiae organisation, troop movements, and its place on the lower Danube. The volume of work that appears year over year about the army is staggering, and an exhaustive discussion is not realistic. Indeed, since the first iteration of this book, an MA thesis from McMaster University defended in the fall of 2005, the quantity of work that has been published has grown enormously. The number of new diplomas alone that have been published in some form or other that are relevant here in the past 10-15 years is impressive. So, to prevent this review from getting out of hand, we will limit ourselves to the most important – and pertinent – research. In particular, we will concentrate on the research on the organisation of the Roman military.

much broader chronological range than the previous three, and he covers social issues that pertain to the empire and its provinces as a whole. Knight and Holder, in a handful of articles and chapters, have looked at the deployment and movements of the auxiliaries throughout the imperial era. 45 Moving from the regular units to the irregular ones and beyond, numeri and vexillations have attracted some attention, though the quantity, and quite often the quality, of evidence we have available has limited discussion in no small measure. Saxer’s book on vexillations is the starting point for any discussion of these detachments, 46 though a handful of articles by Tully have expanded our knowledge, especially with respect to legionary vexillations and their commanders. 47 The so-called numeri are something of an enigma – the most complete account is Southern’s, 48 though Speidel and Le Roux, for instance, have made important contributions. 49 Although the praetorians play a very small role in our discussion, it is worth drawing attention to Bingham’s comprehensive study. 50 Finally we come to the Roman Empire’s assorted fleets, and the two most important studies are those of Starr and Reddé, though Spaul has written a near equivalent to his contributions on the auxiliae. 51

The Organisation of the Roman Military The starting point for those interested in the history of the Roman legions is Ritterling’s classic PaulyWissowa article. 36 Mention should also be made of Parker’s history of the legions. 37 Parker’s book is more synthetic than Ritterling’s, for Ritterling devoted significant attention to individual legions, while Parker tended to look at the legions as a whole. Ritterling’s book-length article was updated significantly in a twovolume work edited by Le Bohec and Wolff. 38 Although not every legion was included, most were; we will have recourse to many of that collection’s chapters in this book. Campbell’s entry on the legions in the Oxford Classical Dictionary, and Wesch-Klein’s in Wiley’s Encyclopedia of Ancient History are also Besides those more scholarly worth noting. 39 contributions, at least three popular books on legionary history have appeared, one by Farnum, which covers a great deal of material, but which is devoid of citations; another by Dando-Collins, which is richly illustrated, but, like Farnum’s, lacking in documentation; and Pollard and Berry’s book, which seems to strike a balance between a popular and an academic work. 40 Szilágyi’s wide-ranging article on the disposition of the legions also bears consideration. 41

The Roman Military and the Lower Danube We now turn to the research specifically concerned with the Roman military and the lower Danube. There are a handful of older books that deal either exclusively or partially with the military organisation on the lower Danube. 52 Wagner’s monograph concentrated on the auxiliaries on the lower Danube and the Rhine, with the Moesias making up a significant chunk. 53 Conversely, Filow’s short book focused on the legions in the Moesias. 54 Like Wagner, Kraft and Beneš do not limit themselves to the Moesias alone. Kraft’s book, ostensibly on auxiliary recruitment on both the Rhine and Danube, does delve into the military organisation of the related provinces. 55 Beneš’ monograph provides something of an update to Wagner’s, though his study is limited to the Moesias and Dacia. 56 Besides the larger studies on the military units that populated the Moesias, a number of scholars have looked at the region’s fortifications, a comparatively understudied

The counterpart to Ritterling’s entry on the legions was two entries by Cichorius in the Pauly-Wissowa: one on the alae, the other on the cohortes. 42 As with Ritterling, both entries have been updated, and in this case both by Spaul, who published two significant collections on the histories of individual Roman auxiliary units. 43 There have also been a number of wide-ranging studies on the auxiliaries, including the important works of Cheesman, Holder, Saddington, and most recently Haynes. 44 Haynes’ book covers a

45

Knight 1991; Holder 2003, 2006a. See too Holder’s (2005) paper on auxiliary cavalry in Pannonia and Moesia during the Flavian period. 46 Saxer 1967. 47 Tully l998, 2002, 2004. 48 Southern 1989. 49 Speidel 1975; Le Roux 1986. 50 Bingham 2013. 51 Starr 1960; Reddé 1986; Spaul 2002. 52 Gren’s (1941) monograph is primarily concerned with economic matters in the Balkans and Asia Minor, though he does touch on military matters (1941: 89-155). 53 Wagner 1938. 54 Filow 1906. I use the 1963 reprint of this book. 55 Kraft 1951. 56 Beneš 1978.

36

Ritterling 1925. Originally published in 1928, Parker’s book was updated in 1958 – and that is the edition used in this study. 38 Le Bohec and Wolff 2000a. 39 Campbell 2007; Wesch-Klein 2012. 40 Farnum 2005; Dando-Collins 2010; Pollard and Berry 2012. 41 Szilágyi 1954. 42 Cichorius 1893 (alae), 1900 (cohortes). 43 Spaul 1994, 2000. 44 Cheesman 1914; Holder 1980; Saddington 1982; Haynes 2013. 37

6

Introduction subject. 57 Biernacka-Lubanska, Gudea and Zahariade, and Gudea (alone) have written books on Moesian Karavas’ University of Durham fortifications. 58 dissertation concentrated on lower Danubian fortifications. 59 There are also a handful of edited volumes that deal largely with fortifications, including those on specific sites, like Novae and Durostorum, 60 the military presence in modern countries like Bulgaria, 61 and even urbanism in particular provinces that touch on military architecture. 62 Wilkes’ 2005 JRS article provides something of a bridge between the former, studies on military units, and the latter, studies on fortifications, and he includes some important bibliographic entries, not to mention a gazetteer of Danubian road stations and fortifications. 63

Chersonesus, 74 and one on Olbia and Tyras. 75 Bounegru and Zahariade wrote what purports to be a comprehensive account of the navy on the lower Danube and Black Sea. 76 Batty’s detailed study deals with the lower Danube and Black Sea as well, and he does delve into military concerns, though in this respect he is less interested in the military organisation of the region than he is in its strategic role, and the various interactions between Romans and nonRomans. 77 Stefan’s monumental book on Trajan’s Dacian wars is also deserving of mention; 78 though Stefan is concerned with a much narrower period of time and series of events than we are here, his book does contain much of value. Wheeler’s extended review of both Batty’s (2007) book and Stefan’s (2005), which also examines Roman strategy on the Danube, is also worth noting. 79

In the last decade or so, the one scholar who has made the most significant contribution to our understanding of the military organisation of the Moesias, and Moesia Inferior in particular (so far), is Matei-Popescu. MateiPopescu has made a number of notable contributions including: in conjunction with Tentea an update of Spaul’s work on the auxiliaries, at least insofar as it pertains to the Dacias and the Moesias; 64 a discussion of the auxiliaries in Moesia before the province was split; 65 an article on the auxiliary units of Moesia Superior; 66 on auxiliary units in Moesia Superior; articles on the participation of troops from the Moesias in Trajan’s wars; 67 a monograph on the army of Moesia Inferior, 68 and a number of other publications.

The last group of studies worth discussing comprises the various publications on military diplomas. The first major collection could be found in volume sixteen of the CIL. The number of published diplomas has increased, however, and seemingly exponentially. This development has had a significant impact on our understanding of the military organisation of the provinces, empire-wide. Most notable are the volumes of diplomas originally collected and edited by Roxan, with five volumes now in print. 80 Also worth noting is the compilation compiled and edited by Pferdehirt. 81 Besides those collections, there are the countless diplomas published year over year, and which have appeared in journals such as ZPE and Chiron. 82

When it comes to the Roman military presence in the Black Sea, some of the most important studies are the assorted papers of Sarnowski and Zubar. Among other things, Sarnowski has published on the command structure in and around the Black Sea, 69 written a detailed treatment of the military round the Black Sea, 70 examined the presence of the Ravennate fleet 71 and the role of the Moesian fleet in the Chersonesus, 72 and explored the existence of a little known war in the region, the Bellum Bosporanum. 73 Zubar has not always focused specifically on military issues, though he has contributed two noteworthy papers on the

CHAPTER ORGANISATION Having taken a look at the impetus for this book, the methodology, the ancient evidence, and the modern scholarly literature, it remains to set out the structure of this book. There are six chapters. The first three chapters are arranged roughly chronologically and cover, collectively, the legionaries and auxiliaries in the region from Augustus to Severus Alexander. They set out the changes in troop disposition in the provinces, the first from 27 BC to AD 81, the second chapter from AD 81 to AD 161, and the third from AD 161 to AD 235. All three chapters also provide some background; they lay out some of the key conflicts in

57

Wilkes 2005: 154. Biernacka-Lubanska 1982; Gudea and Zahariade 1997; Gudea 2001; Gudea 2005. 59 Karavas 2001. 60 Ivanov, Atanasov, and Donevski 2006; Vagalinski 2007; Derda, Dyczek, and Kolendo 2008. 61 Derda, Dyczek, and Kolendo 2008; Vagalinksi, Sharankov, and Torbatov 2012. 62 Mladenovic 2012. 63 Wilkes 2005. 64 Tentea and Matei-Popescu 2004. Those see now Matei-Popescu (2015) on the auxiliaries in Moesia Superior and Dacia. 65 Matei-Popescu 2013. 66 Matei-Popescu 2001/2002; Matei-Popescu 2006/2007. 67 Matei-Popescu and Tentea 2006; Matei-Popescu 2007b. 68 Matei-Popescu 2010a. 69 Sarnowski 1995b. 70 Sarnowski 1988. 71 Sarnowski 2006a. 72 Sarnowski 1991. 73 Sarnowski 2006a. 58

74

Zubar 1995, 2001. Zubar 2007. 76 Bounegru and Zahariade 1996. 77 Batty 2007. 78 Stefan 2005. 79 Wheeler 2010, 2011. 80 Roxan 1978, 1985, 1994; Roxan and Holder 2003; Holder 2006b. In this book I refer to the diplomas themselves by their volume and diploma number, and I include the abbreviation RMD, though if I am referring to the commentary I will refer to the relevant volume. 81 Pferdehirt 2004. These diplomas are usually abbreviated by their diploma number and RGZM. 82 Petolescu and Popescu 2004; Eck and Pangerl 2008; MihailescuBirliba 2008; Weiss 2008; Eck and Pangerl 2009; Schindel 2010; Eck and Pangerl 2013; Eck and Pangerl 2014a; Eck and Pangerl 2014b; Eck, Holder, Pangerl, and Weiss 2015. 75

7

Exercitus Moesiae the region during the respective periods. The fourth chapter focuses on vexillations and the Black Sea region, and it includes an overview of vexillations empire-wide before concentrating on their presence in the Moesias. The second half of the chapter turns to cities like Tyras, Olbia, and Chersonesus, and evaluates the related matter (to vexillations) of their respective garrisons. The fifth chapter concentrates on the emplacement of the various legions, auxiliary units, vexillations, and the fleet in the Moesias. The sixth and final chapter situates the discussion of military organisation into wider debates on Roman strategy, and looks at, among other things, comparative troop numbers and Roman imperial policy. The material discussed in this book is complex, and as a result the book is full of detail. Those who lack the fortitude to wade through the wealth of detail should head for chapter six, the results and discussion portion of this book. I should note, too, that summaries of much of the data that I discuss in chapter six can be found in summary tables reproduced in appendix one. Those tables are also found in their respective chapters (one through four). In addition, appendix two provides a summary of all the units that are found in the book. Appendix three lists the diplomas used in the book, while appendix four the epigraphic, literary, and papyrological evidence. Appendix five discusses the evidence for numeri in the region. Finally, appendix six provides maps that identify the places discussed in this book.

8

PART I: THE DATA

Chapter 1: The Lower Danube from 27 BC to AD 81

particularly the Triballi and the Dentheleti, who wanted to rid the region of Dacian control. The Triballi too were likely allies of Rome. 92 Crassus took on the Bastarnae, the Dacians, and the Thracians who had been Roman allies. 93 They were summarily crushed; after Crassus subdued both the Thracians and the remaining Moesian rebels, order was restored.

INTRODUCTION We start this chapter with a discussion of the military situation in Moesia from Augustus to the death of Titus. The first section sets the stage, outlining some of the key conflicts that occupied the Roman state at this time. We then turn to the legions, and discuss their movements. We finish with the auxiliary units, which we know less about before the advent of constitutions and diplomas. As it happens, much of our evidence for troop movements during this 100 or so year period comes from texts.

Dio reports some problems in Macedonia and Thrace in the year 16 BC caused by the wayward Scordisci and Dentheleti, those same people who thirteen years before had appealed to Rome for assistance. 94 Tiberius was sent to pacify the Scordisci. A number of problems plagued the region in 10 BC when the Danube froze and the Dacians crossed and plundered the neighbouring provinces. Both Dio and Augustus recorded the event, though Augustus gave it more prominence. 95 After chasing the invaders out, Augustus claims that he conquered the recalcitrant tribes. Mocsy plausibly suggests that the general in charge of this operation was M. Vinicius, who left an account of his exploits on an inscription erected in Tusculum. 96

BACKGROUND: MILITARY HISTORY Augustus We do not know precisely when Moesia first became a province. Sometime after the campaigns of M. Licinius Crassus in 29-27 BC and the conquest of the Moesians, 83 the civitates of Moesia fell under the jurisdiction of the province of Macedonia. 84 How long this situation persisted is also unclear. The first imperial legate of Moesia was M. Caecina Severus, whom Dio mentions while describing the events of AD 6 in the province. 85 Appian, on the other hand, says that “[the Mysians] were not subject to tribute in the time of Augustus, but by Tiberius”. 86 This is not a lot of evidence to go on. Suffice to say, the scholarly consensus is that Moesia likely became an independent (from Macedonia) provincia governed by an imperial consular legate around AD 6. 87

The next set of operations in the region occurred between 6 BC and AD 4, with the precise dates unclear. Florus, in his “Bellum Dacicum”, remarks that the Danube froze over which allowed King Cotiso to cross and make raids upon his neighbours. 97 Augustus then sent Lentulus to push them beyond the banks of the river. In addition, he also established garrisons on the nearer bank. 98 Florus also discusses the “Bellum Sarmaticum”. 99 Here “it was deemed sufficient to debar [them] from access to the Danube, and Lentulus was entrusted with this task also”. 100 On the other hand, according to Ovid, at some point early in the first century the Getae took the cities of Aegyssus and Troesmis. 101 Strabo briefly talks about Boerebistas, a Getan, who took control of the tribe and then became a menace to the Romans, as he would often cross the Danube and plunder Thrace, as well as Macedonia and even Illyria. 102 However, before the Romans could send a punitive expedition, he was deposed and his kingdom divided. Strabo also curiously states that Augustus made an expedition against the Getae, “only recently”. 103 This could be the events that he described a little earlier when Aelius Catus, a consul in AD 4, transplanted from the far side of the Danube onto the Roman side nearly fifty thousand Getae. He says that these people now live in Thrace and were henceforth

The earliest recorded military operation in the region involving Roman forces after Actium occurred in 29 BC, when M. Licinius Crassus engaged the Dacians and their Bastarnaean allies. 88 Dio says that the Bastarnae overran part of Moesia and part of Thrace, but entered territory that belonged to the Dentheleti, a tribe who were Roman allies. 89 Marcus Crassus was then sent to engage the Bastarnae and the Dacians in Macedonia. 90 Mocsy posits that the Romans undertook an offensive operation by attacking the Dacians under Cotiso. 91 This was unsuccessful, and as a result of his defeat Cotiso appealed to his erstwhile allies, the Bastarnae. The Bastarnae then attacked those tribes, 83 Note Wheeler (2011: 199) who says that the Getae (former inhabitants) became Moesi/Mysi in the eyes of the Romans. 84 Wilkes 1996: 993; CIL V.1838. 85 Cass. Dio 55.29.3. Cf. Syme 1934; Wilkes 1996. 86 App. Ill. 30, trans. White. 87 Syme 1934, 131ff; Zahariade 2013b. 88 Mocsy 1974: 23. This was not the first such encounter with Moesia’s inhabitants. Both Scribonius Curio and Lucullus defeated Moesians in the 70s BC (Zahariade 2013b). 89 Cass. Dio 51.23.3-4. This affair is also briefly described by Florus at 2.26. 90 Cass. Dio 51.23.2. 91 Mocsy 1974: 23.

Mocsy 1974: 24. Cass. Dio 51.25.1-27.1. 94 Cass Dio 54.20.3 ff. 95 RG 30; Cass. Dio 54.36.2. 96 ILS 8695; Mocsy 1974: 35. 97 Flor. 2.28. 98 Flor. 2.28: citra praesidia constituta. 99 Flor. 2.29. 100 Flor. 2.29. trans. Forster. 101 Ov. Pont. 1.8.11-16; 4.9.79-80. 102 Strabo 7.3.11. 103 Strabo 7.3.11. 92 93

11

Exercitus Moesiae called Moesians. 104 We also have a brief footnote in Tacitus’ Annals. 105 According to Tacitus, the year AD 25 saw the end of two great nobles, including Gnaeus Lentulus, who won triumphal decorations against the Getae. We cannot date the campaigns of Lentulus and the transplantation of Getae by Aelius Catus. The most plausible scenario is that these connected events occurred between AD 1 and AD 4. 106

Here we get the introduction of the Dacians, who “stormed the winter quarters of our auxiliary infantry and cavalry and put themselves in possession of both banks of the Danube”. 112 The Dacians, however, were prevented from causing any further injuries through the efforts of the Sixth legion. The last major event of significance occurred in the winter of 70 – 71, when the Sarmatians attacked. Josephus tells us that their crossing of the Danube caught the Romans unawares. 113 A great many of the Roman soldiers, deployed to protect against such an incursion, fell. What is more, the governor of Moesia, a certain Fonteius Agrippa who had advanced to meet them, was killed in battle. When Vespasian learned of this invasion, he sent Rubrius Gallus with reinforcements to push back the Sarmatian aggressors. He was successful and “the war in Moesia was thus speedily decided”. 114

The years AD 6 – 9 filled the Romans with fear, as the Pannonians revolted against their rule. A large force was required to quell this disturbance. Early on in this campaign there was an incursion in Moesia and A. Caecina Severus, the first known governor of Moesia, who had been operating in Pannonia, was forced to return to deal with the situation. These events occurred in AD 6 and involved the Dacians and Sarmatians. 107 Nero The next major military event does not come until the reign of Nero. Sometime around AD 60, or perhaps shortly thereafter, the Moesian governor Plautius Silvanus claims that he brought across the Danube more than 100,000 from the Transdanubian peoples. 108 This boastful inscription also describes the suppression of an insurrection of Sarmatians; positive diplomatic dealings with the Bastarnae and Rhoxolani, including the return of sons of the respective kings; and the successful raising of a siege of the city of Chersonesus in the Crimea, which had been assaulted by the Scythians. 109

TROOP MOVEMENTS FROM AUGUSTUS TO TITUS Legions We now turn to troop movements and start with the legions. The earliest record of the Moesian legions is spotty at best, and we sometimes have to deduce origins from legionary nomenclature. Beginning with the campaigns of Crassus, the legio V Macedonica might have earned its name from its service in Macedonia from 30 BC to AD 6, 115 though Mitchell believes that the legion, led by L. Calpurnius Piso, earned the title for meritorious service during the uprising in Thrace, noted below. 116 Not all scholars, however, are wholly convinced that the Fifth legion spent the duration of those years in Macedonia. Both Syme and Mitchell believe that the legion may have spent an indeterminate amount of time in the east. 117 Some too have suggested that the legion ought to be equated with the V Urbana. 118 The legio IIII Scythica might also have earned its title from service in the same war of M. Licinius Crassus. 119 This would give Macedonia – Moesia was not yet a provincia 120 – a garrison of two legions in this early phase of its history.

The next major incursion occurred at the beginning of the civil war. Sometime in the winter of AD 67 and AD 68, the Rhoxolani had reportedly massacred two Roman cohorts. 110 By the second half of AD 68, they had again invaded Moesia with a force 9000 strong. Despite the civil war, the Romans were prepared for battle and crushed their Sarmatian adversaries. The current governor of Moesia, Marcus Aponius, was rewarded for his victory with a triumphal statue; his three legionary commanders, Fulvius Aurelius, Julianus Tettius, and Numisius Lupus, were also rewarded with “the decorations of a consul”.

Sometime between 13 and 11 BC there was an uprising in Thrace. L. Calpurnius Piso was summoned from Pamphylia to deal with this problem, which persisted

The Year of the Four Emperors and Vespasian There were more problems on the lower Danube in October, or, at the very least, the autumn of AD 69. 111

112

Tac. Hist. 3.46, trans. Moore. Joseph BJ 7.4.3. 114 Joseph BJ 7.4.3, trans. Thackeray. 115 Parker 1958: 266; Campbell 1996a: 840; Keppie 1998: 207; Farnum 2005: 19; Dando-Collins 2010: 136. 116 Mitchell 1976: 307. 117 Syme 1933; Mitchell 1976. 118 Strobel 2002; Farnum 2005: 19. cf. Matei-Pescu 2010a: 35-36; Wesch-Klein 2012: 4. 119 Parker 1958: 266; Campbell 1996a: 840; Keppie 1998: 206; Speidel 2000: 327; Farnum 2005: 18; Wesch-Klein 2012: 4. 120 According to Syme (1933), the legions in the lower Danubian region would have been under the command of the proconsul of Macedonia prior to the transformation of Moesia into a province, or at the very least, some sort of a military command. 113

104

Strabo 7.3.10; Wheeler 2011: 201-202. Tac. Ann. 4.44. 106 Syme 1934: 131 ff.; Cf. Mocsy 1974: 35 – 37. 107 Cass. Dio 55.30.1 ff. 108 ILS 986. Cf. Berciu 1981: 271; Rice 1981: 290; Conole and Milne 1983; Wilkes 1983: 259-260. 109 On some of the problems with this inscription and discussion with full bibliography, see Wheeler (2010: 1218, n. 82, 1219, n. 86; 2011, 203-205). 110 Tac. Hist. 1.79. 111 Tac. Hist. 3.46; Filow 1963: 27; Stefan 2005: 393-395; Wheeler 2011: 206. 105

12

The Lower Danube from 27 BC to AD 81 for three years. 121 The two legions in Macedonia, the Fourth and Fifth legions, or some part thereof, were mobilised to deal with the threat. 122 Upon his arrival, Piso assumed command and won the day. A problem remains, however: did Piso bring and then leave any part of this Anatolian force behind? 123 Farnum suggests that the V Macedonica went to Galatia in 25 BC and that it later returned to the Balkans in AD 6. 124 If it had gone to Galatia in 25 BC, but returned for the Thracian revolt, this would provide a solution to our problem. But, is there evidence for this?

whom we should allocate those five legions: if we stick with two legions, then we would have the same number of legions under Caecina Severus that were left in the province as we do in AD 23. But, we do not know how many legions were deployed under Silvanus Plautius who was in Galatia before the Pannonian war. 131 During the year in which A. Caecina Severus was dispatched to Pannonia, there arose a disturbance in Moesia. He was forced to return to defend his province. Dio, who describes Severus’ movement, does not provide us with the strength of the force with which he returned. 132 What is more, we know from Dio that Silvanus Plautius continued on in the war, although which legions he took with him is unclear. 133 Syme suggested that Silvanus took with him the legions that had already been deployed in Moesia rather than those legions he had in the east. 134 In the absence of additional evidence, there is no definitive solution. At best, in AD 6 there were two legions in Moesia, which had been deployed there previously (it not immediately so), were the Fourth Scythian and the Fifth Macedonian. 135

We do have a funerary inscription from Galatia for a soldier who had served in the V Macedonica. 125 If the man had served in the legion while it was in Galatia that would make his age at death rather remarkable. There are a few other inscriptions from Turkey that list a former soldier of the V Macedonica, but there is nothing in them to suggest that the unit itself had been based in the region. 126 We know that in AD 23 there were two legions in Moesia, the Fourth Scythian and the Fifth Macedonian. 127 Prior to Tacitus’ digression on the disposition of the Roman armed forces during the reign of Tiberius, the earliest record that we have for the movement of units, if not to Moesia at least through Moesia, pertains to AD 7 and the Pannonian revolt. At the time, Tiberius had been stationed on the Danube preparing for a major assault upon the Marcomanni in Bohemia. 128 When news broke about the revolt of the Pannonians, and then the Dalmatians, Tiberius’ attention shifted. Dio tells us that while Tiberius marched to head off some of the attackers, Caecina Severus, the Moesian governor who had already been mobilised, hurried with an indeterminate force to save Sirmium from the Breucians, a Pannonian Additionally, tribe participating in the revolt. 129 Velleius outlined the size of the Roman force. He said that the two consulars, Aulus Caecina (the same governor of Moesia mentioned by Dio) and Marcus Silvanus Plautius, “were bringing five of our legions, together with the troops of our allies and the cavalry of the king (for Rhoemetalces, king of Thrace, in conjunction with the aforesaid generals was bringing with him a large body of Thracians as reinforcements for the war)”. 130 The problem is we do not know to

The legio VII is often considered part of the garrison of Moesia/Macedonia under Augustus; moreover, on at least five occasions the legion is found with the There are two cognomen, Macedonica. 136 interpretations of the origins of this title. On the one hand, this legion may have earned that name while stationed in Macedonia much like the Fifth legion. 137 In this case, it may very well have served in the campaigns of M. Licinius Crassus against the Bastarnae and Dacians in the years 29BC – 27BC. On the other hand, Mitchell posits that the legion earned this title from commendable service during the war in Thrace in which Piso participated. 138 He argues that since the Thracian war of 13 – 11 BC was so serious that Piso received the triumphal decorations for his victory, it would not be far fetched to imagine that the Seventh legion too earned due praise for its efforts. We now know that the Seventh legion was in Macedonia early in Augustus’ reign. Three of the four inscriptions from CIL III that name this unit, however, hail from the east, and more specifically from Pisidian Antioch and Iconium. 139 It is probably significant that in these inscriptions the legion no longer bears the

121

Vell. Pat. 2.98; Cass. Dio 54.34.5ff. Cf. Farnum 2005: 18. 123 The consensus among many scholars, including Ritterling, Filow, Syme, Szilágyi, and Parker, is that by AD 6 there were three legions in Moesia; however, while all agree that the Fourth Scythian and Fifth Macedonian legions were there at the time, uncertainty surrounds the identity of the third legion. See Ritterling 1925; Syme 1933; Szilágyi 1954; Parker 1958; Filow 1963. Parker (1958: 92) believes there were three or even four legions, while all the other scholars noted prefer a total of three. 124 Farnum 2005: 19, 81. Mitchell (1976: 307) considers this a possibility, and that it was stationed along with the legio VII. Cf. Matei-Popescu 2010a: 36. 125 AE 1984, 886. 126 AE 1946, 51; AE 1963, 3 (2006: 1553); AE 1990, 893, 896; AE 1973, 500. 127 Tac. Ann. 4.5. 128 Vell. Pat. 2.110. 129 Cass. Dio 55.30.4. 130 Vell. Pat. 2.112.4 trans. Shipley. 122

131

Farnum (2005: 81) claims the V Macedonica and the VII Macedonica were in Galatia from 25 BC. 132 Cass. Dio 55.30.1ff. 133 Cass. Dio 55.32.3 134 Syme 1933: 399. The inscription (ILS 921) that celebrates Silvanus’ success does not say a thing about the units he used. 135 Wheeler (2011: 202) argues for three legions in AD 6. 136 Ritterling 1925: 1362; Syme 1933: 31; Szilágyi: 124; Keppie 1998: 159; Laporte 2000: 560-561. The five inscriptions which bear the title Macedonica are: CIL III.7368; CIL X.1711, 4723, 8241; and AE 1938, 141. 137 Keppie 1998: 208. 138 Mitchell 1976: 302. 139 CIL III.6826, 6827; Cf. IGRR III.1476.

13

Exercitus Moesiae cognomen, Macedonica. Prior to the earning of the title Claudia, the legion is referred to in most inscriptions as the Seventh Legion (VII legio). 140 Thus, the legion seems to have shifted to the east sometime after the episode with Crassus, or perhaps following the campaign against the Thracians in 13-11 BC. 141 Mitchell, against Syme, convincingly argues that the legion was stationed in Galatia, at least for a time, perhaps along with the Fifth Legion prior to the Pannonian Revolt. In fact, he thinks that Macedonia was only garrisoned by one legion prior to the Pannonian Revolt.

earned the title Scythica as a result of its exploits in that war. The Seventh Legion also seems to have participated in that war; at the very least, it was in Macedonia at the time, where it earned the epithet Following this war against the Macedonica. 146 Bastarnae and Dacians, the Fourth and Fifth remained, while the Seventh was transferred to Galatia where it stayed at least until the war of L. Calpurnius Piso. While in Galatia it dropped the title Macedonica. Around the conclusion of the Pannonian Revolt, Varus’ three legions were massacred north of the Teutoburg Forest in AD 9. With the subsequent gaps in the German Army, units were shifted from elsewhere, notably Spain and the Upper Danube. The Seventh legion might have moved to Dalmatia either initially with the arrival of Silvanus’ reinforcements, or some time later during the military reorganisation that followed the Teutoburg disaster, if not later. 147

We do have at least three inscriptions that give just the title of the Seventh Legion in Moesia: one hails from Naissus in Moesia Superior, one from modern Kosovo in Moesia Inferior, and one from Cupria in Moesia Superior. 142 The first two of those inscriptions are from soldiers, and so we must at least entertain the possibility that these soldiers were based in Moesia at some point. The bulk of the remaining inscriptions listing the legion from CIL III in the Balkans are from Dalmatia, which is where the legion is believed to have been deployed after AD 9. 143 We might speculate that the Seventh legion made a trip to Macedonia at some point prior to its return under Nero, but we cannot say the exact dates when – or even if – that may have been.

There were more changes to the legionary garrison of Moesia during the reign of Claudius. As late as AD 43 the two legions in Moesia were still the Fourth Scythian and Fifth Macedonian; 148 by AD 46, the legio VIII Augusta had been transferred from Illyricum and moved to Novae. 149 This is based on an inscription from Castulo in Spain that names a Quintus Cornelius Valerianus who was a prefect of a vexillation composed of the Fifth Macedonian and Eighth Augustan legions. 150 Reddé argues that the legion arrived in Moesia the previous year, AD 45, for the campaign of Didius Gallus against Mithridates, the king of the Bosporan kingdom. 151 Parker suggests that it may have come in AD 44. 152 The truth is, we cannot be certain. The book in Tacitus’ Annals that would have described the events is lost. 153 Instead, we have a passing reference to a campaign of Gallus from the later uprising involving Mithridates. Tacitus tells us that Didius Gallus had gone with the suitably vague “the strength or main body of the army [roburque exercitus]”, but that he had left behind a few cohorts under the Roman knight Julius Aquila. 154 We do not know the exact makeup of the Roman force, but a likely scenario is that it was composed largely of legionary vexillations and auxiliaries (the assumption being no one branch would be dispatched without the other) based on the information contained in the

There are two other legions that Syme suggested might have been stationed in Moesia at some point under Augustus, namely the Eighth Augustan Legion and the Eleventh Legion. 144 Unlike the other three legions, the Fourth Scythian, Fifth Macedonian, and Seventh legion, we have no evidence that either of those two legions (VIII Augusta and XI) were based in Macedonia. Their nomenclature does not tell us anything, unlike the corresponding names with the other three legions; in fact, there are no inscriptions for the Eleventh legion in Moesia without the epithet Claudia. 145 In the absence of any clear evidence, we are again left with the Seventh legion as the only other possible third legion in the region besides the IIII Scythica and V Macedonica. The most plausible reconstruction for the legionary garrison of Augustan Moesia is as follows. The Fourth Scythian Legion and Fifth Macedonian Legion were stationed in Macedonia following the civil war between Antony and Octavian and participated in the war of M. Licinius Crassus, with the Fourth Legion likely having

146 In this case it received the title for serving in Macedonia, not for an excellent performance in the war. 147 Wesch-Klein (2012) says that the legion moved in AD 14, though his medium (encyclopaedia article) has meant that he was not able to produce any evidence in support of this. 148 Tac. Ann. 4.5. 149 CIL XI.1835; Ritterling 1925: 1363; Parker 1958: 132; Filow 1963: 10; Mocsy 1974: 48; Reddé 2000: 121; Wesch-Klein 2012. 150 CIL II.3272. 151 Reddé 2000: 121. For the conflict see Sarnowski (1988: 62-67). 152 Parker 1958: 132. 153 What we have instead is an indication that the legion had mutinied along with the others from Pannonia during the eponymous revolt (Tac. Ann. 1.16.2, 23.5, 30.4). cf. Farnum 2005: 21; Wesch-Klein 2012. 154 Tac. Ann. 12.15.

140

CIL III.192, 195, 1674, 1813, 1818, 2033, 2048, 2071, 2709, 2710, 2716, 2717, 2882, 2908, 2913, 2914, 3200, 6826, 6827, 8487, 8488, 8493, 8544, 8687, 8723, 8763, 8767, 9711, 9712, 9733, 9734, 9736, 9737, 9738, 9741, 9742, 9832, 9864a, 9939, 9973, 12416, 12666, 12814. 141 Mitchell 1976: 307. 142 CIL III.1674, 12416, 12666. 143 Ritterling 1925: 1363; Szilágyi 1954: 141; Parker 1958: 119; Laporte 2000: 561. 144 Syme 1933: 31. 145 As with the Seventh Legion, the Eleventh Legion also was bestowed with the epithet Claudia subsequent to the revolt of Scribonianus in AD 42.

14

The Lower Danube from 27 BC to AD 81 inscription, which named the three legions, and the references in Tacitus. 155

above that around AD 45 or 46 the Eighth Augustan Legion was transferred to Moesia. The other legion which eventually went to the province was the legio VII Claudia, the aforementioned legio VII (Macedonica) whose name was changeable. It was transferred from Dalmatia sometime between AD 42 and 62. 165 We know that the Seventh and Eleventh legions were in Dalmatia in AD 42. Dio tells us that the legions received the titles “Claudian” and “Loyal and Faithful” following their loyalty in the revolt of Marcus Furius Camillus Scribonianus. 166 By now we know that there would have been three legions prior to the departure of the Fourth and Fifth, with the arrival of the Eighth a few years earlier; moreover, given the increased activity on the lower Danube, it is rather unlikely that Moesia would have been left with only one legion for too long. Thus, the Seventh Claudian Legion probably arrived sometime between the years of AD 57 and 59, just before or after the departure of the two legions for the east. 167

The next major movement of legions in Moesia took During Corbulo’s place around AD 56-58. 156 campaigns in the east, Tacitus tells us that “a legion from Germany was added with the cavalry wings and infantry cohorts.” 157 Tacitus’ statement is problematic: there is no independent evidence for the departure of a legion from Germany for the east around this time. Yet, the Fourth Scythian Legion had turned up in the east with Corbulo by AD 62, though it was never attested in Germany. 158 Thus, the Fourth Scythian Legion, which had been based in Moesia, is the legion likely referred to by Tacitus and from this time onward was stationed in Syria. 159 Around the time of the departure of the Scythian Legion, the Fifth Macedonian Legion also left the province. In those same campaigns against Vologeses in the east, Corbulo had been compelled to summon additional reinforcements, and in particular the Fifth Macedonian Legion. 160 The career inscription of Tiberius Plautius Silvanus, which dates to the 70s but details events in the 60s (and earlier), 161 provides us with a clue as to whether the legion departed for the east at the same time as the Fourth legion, or only a short time later. 162 In that inscription it says he had achieved success against the Sarmatians, “although he had sent a great part of his army for the expedition in Armenia.” We already know that by AD 46 there were three legions in the Moesian army. The question is whether Silvanus would have considered the loss of only one legion, from a force of three legions and several auxiliary units, a great part of his army. True, it would be unwise to take the boastful inscription at face-value, as the 100,000 transplanted transdanubian peoples is surely an exaggeration. 163 Yet, Silvanus could still aggrandise his exploits with minimal “artistic licence” by mentioning the great troop transfer, which likely included the movement of both the Fourth and Fifth legions, but omitting the subsequent arrival of another legion in compensation. Thus, halfway through Nero’s reign, the two stalwarts of the Moesian army, the Fourth and Fifth legions, had left. 164

The next major period of change on the Moesian front came at the end of Nero’s reign and the civil war of AD 68-69. In a speech in Josephus’ Jewish War, M. Julius Agrippa II says that the neighbours of Thrace, the Illyrians, were “kept in check by no more than two legions”. 168 Josephus says that the Illyrians inhabited the land between Dalmatia and the Danube. He also said these two legions were used to repel incursions of Dacians. The Dacians did not dwell in that part of the Danube opposite Pannonia. Instead they inhabited the land opposite Moesia. According to Tacitus, by the first half of AD 69, most likely by the last month or two of spring, the Third Gallic legion was stationed in Thus, by Illyria, Agrippa, or rather Moesia. 169 Josephus, meant Moesia, as it was held by two legions at this point: the Seventh Claudian legion and the Eighth Augustan Legion. 170 Moreover, Suetonius tells us that there were detachments from the three legions in the army that had been sent to help Otho, presumably early in AD 69. 171 Thus, the legion was transferred sometime between AD 66 and 69. In fact, the legion was transferred before the death of Nero. 172 This probably took place in the first half of AD 68, and perhaps - in light of the revolt of Vindex - in March of that year.

To compensate for the loss of the two legions, some other legions were transferred to the province. We saw

During the civil war, the Moesian legions, and perhaps by proxy the auxiliaries too, 173 declared their allegiance

155

165

CIL II.3272; Tac. Ann. 12.16. Most scholars date the changes to AD 58, but Speidel (2000: 329) suggests that they may have taken place in AD 56 or 57. 157 Tac. Ann. 13.35. 158 Tac. Ann. 15.6. 159 Speidel 2000: 329. cf. Farnum 2005: 19; Dando-Collins 2010: 132; Wesch-Klein 2012. 160 Tac. Ann.15.6. 161 Batty 2007: 406. cf. Milns 1983. 162 ILS 986. 163 Not surprisingly this inscription has attracted a great deal of attention. Note, for example, Milns (1983) and the comments of Batty (2007: 406-410). 164 Note Wheeler 2012b: 140.

Laporte 2000: 561. Cass. Dio 60.15.4. Campbell (2007) says that it moved in 56-57, Wesch-Klein (2012) that it moved in 55. Farnum’s (2005: 20) suggestion of a transfer in 62 is far too late. 168 Joseph BJ 2.16.4, trans. Thackeray. 169 Tac. Hist. 2.74. 170 See Filow (1963: 22-23) and Thackeray (1997: 467 n. h) in the Loeb version of Josephus’ Jewish War. 171 Suet. Vesp. 6. 172 Suet. Vesp. 6. 173 This is only a suggestion. Though part of the same provincial army, there was undoubtedly a considerable bit of autonomy amongst the respective troop types.

156

166 167

15

Exercitus Moesiae to Otho. 174 We saw above that Tacitus reported a force of only 9,000 for the invasion of the Rhoxolani that year. If we assume that Tacitus’ figures are correct, or at least reasonably so, these Sarmatians had apparently felt that a force that size was sufficient, and probably because of the Roman preoccupation with the civil war. We must then wonder whether the bulk of the Roman forces had already moved to the Italian theatre by this point. We know that early in AD 69 Otho’s force had included the armies of Dalmatia and Pannonia. 175 We also read in a speech, which Tacitus credits to Suetonius Paulinus for Otho, that the troops from Moesia were on their way to Italy; however, they were not marching together, as we later find out that the advanced forces had reached Otho while the rest of the army was only in Aquileia. 176 If we accept that many of the auxiliary troops would be more lightly armed than their legionary counterparts, or at least had greater mobility, 177 this advanced force might have been a contingent of auxiliaries. 178 Of course, we cannot verify this, though we know that at least two auxiliary cavalry units were involved. 179

was composed of infantry and cavalry soldiers. 185 Licinius Mucianus’ layover in Moesia was brief: he was there long enough to repel the Dacians who had attacked in the absence of the regular Moesian legions. Following the battle Fonteius Agrippa was sent as governor. It is then that Mucianus likely left. Moesia, however, was not abandoned as additional troops from the army of Vitellius were sent. 186 This included the legio I Italica and the V Alaudae. 187 When these legions were sent to Moesia is open to debate. The former had been organised by Nero and was to be used for an eastern expedition (to the Caspian sea). 188 Tacitus says that, following the defeat of the Vitellians in the Battle of Cremona, the defeated legions were dispersed through Illyricum. 189 On the other hand, he does not specify whether these troops arrived prior to or after the campaigns against the Dacians. These legions may have been sent to Illyricum before being sent to Moesia, and at the same time as Fonteius Agrippa’s transfer to Moesia from Asia. 190 The following winter the Sarmatians attacked. 191 Not long after that Josephus tells us that Titus, who had remained in the east to tackle the Jewish Revolt while his father headed to Italy for the civil war, dispatched the Fifth Legion to its former station of Moesia. 192 Originally, the scholarly consensus had been that the V Alaudae survived through the reign of Vespasian to the Dacian incursions of AD 85 and AD 86. 193 The Adamklissi monument was considered a tribute to the fallen soldiers of that legion with the death of either Oppius Sabinus, or Cornelius Fuscus. 194 However, it is now clear that the Adamklissi monument dates to some time during Trajan’s Dacian wars, and that the V Alaudae did not last beyond AD 71 at the latest. 195 Franke has convincingly argued that the legion was heavily depleted following the death of Fonteius Agrippa in AD 70; as a result, the surviving soldiers were re-distributed. 196 Thus, by AD 71 and through to AD 81, the legions stationed in Moesia were the I Italica, V Macedonica, and VII Claudia. 197

Later, probably in the summer of AD 69 and after the death of Otho, all three Moesian legions were in Cisalpine Gaul trying to persuade their Pannonian and Dalmatian counterparts to jump to Vespasian’s side. 180 It seems, however, that this did not represent the entire Moesian force because later Vespasian summoned Aponius Saturninus with the rest of the Moesian army (“cum exercitu Moesico”, “the army in Moesia”). 181 We do learn that when Saturninus arrived in Italy, he was in command of the Seventh Legion. 182 The Moesian garrison included some auxiliary units because, when the Dacians attacked in the autumn of AD 69, it was auxiliary units alone that were attacked. Furthermore, we do not hear of an attack on the former garrison of Moesia, which included the Third Gallic, the Seventh Claudian, and the Eighth Augustan legions. 183 Instead, it was the Sixth legion that proved to be Moesia’s salvation. Thus, Saturninus had probably left by the autumn of AD 69 prior to the attack of the Dacians.

Julio-Claudian Dynasty Auxiliae

With the absence of the Moesian garrison for the civil war, the province was supplemented by reinforcements from the east. While in Berytus campaigning during the Jewish War, Vespasian sent ahead an advance force under Mucianus to Italy. 184 This force included the legio VI Ferrata and a vexillation 13,000 strong, which

Evidence for troop movements in the early imperial history of the auxilia across the empire is hard to come by before the appearance of the first diplomas. 185

Joseph. BJ. 4.11.1; Tac. Hist. 2.83; Filow 1963: 26. Tac. Hist. 3.46. 187 Parker 1958: 148; Absil 2000: 227-228; Franke 2000: 43-46. 188 Suet. Nero 19.3; Matei-Popescu 2010a: 77; Wesch-Klein 2012. 189 Tac. Hist. 3.35. 190 Ritterling 1925: 1409, 1410, 1569; Parker 1958: 142; Filow 1963: 27; Absil 2000: 229; Farnum 2005: 19; Campbell 2007; Matei-Popescu 2010a: 77-78; Wesch-Klein 2012. 191 This was in AD 70 or AD 71. 192 Joseph. BJ. 7.5.3. 193 For example, see Rossi (1971: 22) and Mocsy (1974: 82). 194 See Stefan (2005: 562-568) and Turner (2013). 195 Campbell (1996a: 840) alluded to the possibility that the legion may not have survived too long beyond the civil war. Cf. Turner 2010: 181-188. 196 Franke 2000: 44-45. 197 Wilkes 1983: 265-266; Strobel 1989: 38. 186

174

Tac. Hist. 1.76. Tac. Hist. 2.11. 176 Tac. Hist. 2.32, 2.46. 177 This question of legionary versus auxiliary equipment is far more complicated than it might at first appear. Note the discussion of Bishop and Coulston (2006: 254-259). 178 Tac. Hist. 3.18. 179 Tac. Hist. 3.2. 180 Tac. Hist. 2.85-86. 181 Tac. Hist. 3.5. 182 Tac. Hist. 3.9; Filow 1963: 24, 25. 183 Tac. Hist. 3.46. 184 Joseph. BJ. 4.11.1, 5.1.6. 175

16

The Lower Danube from 27 BC to AD 81 recorded in Germania Inferior in the reign of Trajan. 210 As regards the Moesian unit, the nomenclature suggests that it may have been in the province, at least at first.

Nevertheless, there are some things we do know thanks to the limited literary and epigraphic evidence that does survive. As noted, ancient authors were not interested in recording the specifics of ancient auxiliaries. Case in point: Tacitus recorded the transfer of auxiliary units from Moesia in AD 21, but he did not name those units. 198 Fortunately, in other instances, and if we use a variety of evidence, we can be more certain. In the first decade of the first century AD, between the years AD 1 and AD 6, Spaul argues that there were at least four alae operating in Macedonia/Moesia. 199 These alae were the ala Atectorigiana, Mauretana, Thracum Herculiana, and the Scubulorum. The evidence for the first of those four alae, the ala Atectorigiana, which was named after a Gallic chieftain, Atectorix, 200 in Moesia before Vespasian is thin. 201 We do not have any evidence for the second ala, the ala Mauretana, in Moesia, though a comparable unit has left a trace in Egypt. 202 We also lack evidence for an ala Thracum Herculiana in Moesia, and so such a unit was likely never based in the province. There does seem to have been such a unit in Syria, however. 203 The last of those four alae, the ala Scubulorum, might have been transferred to Britain for the invasion of Claudius, 204 or it might have been sent to Pannonia immediately when it left. 205 We also know that an ala Bosporanum was transferred from Moesia to Syria sometime between the reigns of Tiberius and Nero. 206 There are at least two inscriptions that point to its time in Moesia. 207 Knight composed a list of auxiliary units moved to the Rhine after the Batavian revolt. 208 The list included the ala Sulpicia, ala Moesica, ala Afrorum veterana, and the cohors II Varcianorum. 209 He suggested that some of these units might have been in Moesia before the civil war (AD 69/70). The respective inscriptions do suggest that the Sulpician, African, and Varcianorian units were in Moesia prior to the accession of Vespasian. The latter, the cohors II Varcianorum, is

Some of the details that we know about the Thracian uprising in AD 26 and the Batavian revolt of AD 70 imply that auxiliary units were initially stationed close to their homeland. Tacitus, for instance, tells us that the reason for the revolt in AD 26 was the displeasure of the auxiliary units from Thrace, who believed that they were to be mixed with other tribesmen, sent to foreign lands, and led by Romans as opposed to their own chieftains. 211 Assuming that stationing auxiliary units close to their recruiting area was standard practice, we might also conjecture that the ala Moesica was posted in Moesia prior to its transfer. The Thracian revolt of AD 26 produces more evidence of the Tiberian troop disposition. Tacitus tells us that Poppaeus Sabinus, the imperial legate of Moesia, talked to the tribesmen to try and assuage their fears, and to delay any further eruptions of violence prior to the arrival of reinforcements from Moesia. 212 Pomponius Labeo then arrived with an unknown legion, and King Rhoemetalces arrived with auxiliaries. 213 These reinforcements were added to Sabinus’ troop contingent, which included the aforementioned auxiliary units. In the same context, Tacitus mentions a select group of archers (delectos sagittariorum) and a cohort of We can assume that these Sugambrians. 214 Sugambrians were part of the Moesian force, and that they were an auxiliary cohort and not a cavalry wing because there was a cohors I Sugambrorum veterana in Moesia in AD 75. 215 The epithet veterana might suggest that the unit had been in Moesia for a while, 216 and that this unit had some seniority over the other cohorts in the province. 217 That epithet also enables us to identify Tacitus’ cohort of Sugambrians with the cohors I Sugambrorum veterana, rather than the cohors I Sugambrorum tironum, which, as its name suggests (tironum, from tiro meaning “new recruit”), was a unit with a shorter history. Now, with respect to the group of archers, there was a cohors II Chalcidenorum (sagittaria) in Moesia by AD 75. 218 Spaul suggests that this unit might have been raised in Lebanon before

198

Tac. Ann. 3.39.1. cf. Matei-Popescu 2013: 209. Spaul 1994: 261. 200 See Holder (1980: 21-22), Speidel (1982), and Haynes (2013: 42-43) on units named after commanders. 201 Matei-Popescu 2013: 214-215. 202 The ala Thracum Mauretana: CIL III.14=CIL III.6581=Kayser 106 = ILS 2543; CIL III.13578; SCI 2012, 54=AE 2012, 1959. 203 CIL XVI.106=ILS 9057=IDRE II.349=AE 1900, 27=AE 1900, 58=AE 1972, 669; AE 2007, 1238. 204 Spaul 1994: 38. 205 Matei-Popescu 2013: 217. 206 CIL III.6707; AE 1925, 70; AE 1969/1970, 649; Cf. Knight 1991: 193. 207 CIL X.1258; ILBulg 137. Cf. Matei-Popescu 2013: 212. Holder (2005: 82) mentions the unit only in passing, even saying that the unit is first attested in the region during the reign of Trajan on the basis of RMD I.21. 208 Knight 1991: 194. 209 Knight (1991: 194) believes that the ala Sulpicia might have been in Moesia based on two inscriptions: CIL XIII.8311, 8312; that the ala Moesica may have been in Moesia based on its title; that the ala Afrorum veterana might have been in Moesia based on three inscriptions from CIL XIII: 8304, 8305, and 6223; and that the cohors II Varcianorum might have been in Moesia based on CIL XIII.7382. 199

210

See RMD IV.216 and KJ 2010, 181=AE 2010, 1865, for instance. Tac. Ann. 4.46-47. 212 Tac. Ann. 4.46. 213 We do not know which legion, and, Tacitus provides us with no clue. 214 Tac. Ann. 4.47. 215 RMD I.2. 216 On the meaning of the name veterana in the context of auxiliary units see Cheesman (1914: 48), Holder (1980: 18-19), and Saddington (1982: 174-175). 217 Matei-Popescu (2012: 209) suggests that the unit, which had been raised in Germania Inferior during the reign of Augustus, had been sent to the region not long after, a suggestion that supports my argument here. 218 RGZM 1. Cf. CIL XVI.45. 211

17

Exercitus Moesiae the Flavians took the throne. 219 Admittedly, we have no evidence for its presence in Moesia prior to AD 99.

Modestus, who was prefect of the cohors Thracum Syriaca. 232 This inscription can probably be dated to sometime between AD 50 and AD 68. 233 One further auxiliary unit to document for the Julio-Claudian period is the cohors I Cisipadensium, which a handful of scholars have suggested served with the Fourth Scythian legion in Moesia under Tiberius. 234

There is yet another cohort that might be equated with these Tacitean archers, namely the cohors I Cilicum sagittaria in Moesia in AD 78. 220 Wagner suggests that the unit was in Moesia by the reign of Claudius, 221 while more recently Matei-Popescu has favoured a date during the reign of Nero. 222 Interestingly, there is an inscription from Uxama in Spain that lists a certain M. Magius Antiquus who was both prefect of the cohors Cilicum, and military tribune of the legio IIII Scythica. 223 This inscription dates to Augustus and from that, and the reference to the Scythian legion, we can infer that the Cilician cohort was indeed in Moesia, at this point Macedonia, during the reign of Augustus. 224 Thus, we can probably argue that the archers referred to by Tacitus are none other than those hailing from this Cilician cohort.

A certain cohors I Montanorum may also have been in Moesia during the reigns of Claudius and Nero. Wagner, Kraft, and Szilágyi suggest that the cohort was in Moesia briefly during that period, though it may have originated in Noricum. 235 Wagner and Szilágyi based their suggestion on a funerary inscription found at Timacum Minus. 236 Wilkes, on the other hand, believed that this unit was part of the early JulioClaudian army in Dalmatia, which later moved to Moesia under Claudius. 237 The problem is that Wagner and Szilágyi seem to be referring to a different unit than Wilkes. Indeed, there are in fact at least two different Montani cohorts in the region, and it is this detail that has led to the differing theories. 238 In the end, the best that we can say is that one cohort of Montani was possibly in Moesia briefly and perhaps in between its postings in Dalmatia and Pannonia. 239

We are on firmer ground as we move further into the age of the Julio-Claudians. There are a handful of additional inscriptions that identify auxiliary units stationed in Moesia. Matei-Popescu argues that an ala Pansiana was based in Moesia by the reign of Claudius, if not earlier (Tiberius) on the basis of an inscription found at Oescus. 225 There is no record of this unit, however, and so it was either disbanded, amalgamated (into a unit we cannot identify), or destroyed. There is an ala Asturum in an inscription that probably dates to the reign of Claudius or Nero; 226 it seems, at least, to have been in Moesia by the reign of Vespasian. 227 The cohort of Cretans also seems to have been stationed in Moesia, and more specifically at Naissus, during the first century. 228 When it arrived in Moesia is unclear. There is a veteran of a cohort of Lusitanians who was based at Tomis, possibly during the reign of Nero. 229 Coincidentally, in a diploma for the province of Illyricum dated to AD 60, there is a cohors I Lusitanorum listed. 230 There is also a cohors I Lusitanorum stationed in Moesia in AD 75, which Spaul argues is the same unit. 231 Thus, it is reasonable to suppose that sometime before the civil war, if not shortly thereafter, this unit was transferred to Moesia from Illyricum. We also have an inscription from Timacum Minus in Moesia for a certain L. Vecilius

Legions (2) IIII Scythica, V Macedonica

Auxiliary Units (11) ala Asturum (?), ala Atectorigiana, ala Bosporanorum (?), ala Pansiana, ala Scubulorum, cohors I Cilicum sagittaria eq, cohors Cisipadensium (?), cohors Cretum, cohors Lusitanorum (?), cohors I Sugambrorum veterana eq, cohors II Varcianorum

Table 1.1. Dispositions of Legions and Auxiliary Units in Moesia in AD 40 232

CIL III.8261. Spaul (2000: 366) points out that evidence for this unit in Syria after AD 70 is lacking, which suggests that it was Moesia before that time; moreover, this unit was at Timacum Minus by the reign of Vespasian, as another inscription (CIL III.14375) proves. 234 Wagner 1938: 121; 464; Kraft 1951: 173; Roxan and Weiss 1998: 371-420; Spaul 2000: 464. 235 Wagner 1938: 170; Kraft 1951: 181; Szilágyi 1954: 152. 236 AE 1903, 289. 237 Wilkes 1969: 473. This move seems likely as the Seventh Claudian Legion moved from Dalmatia to Moesia some time after AD 42. 238 Tentea and Matei-Popescu 2004: 289. 239 CIL XVI.26 (Dalmatia); CIL XVI.31 (Pannonia Inferior). Even though a cohort of Montani is listed on both diplomas that might be considered one and the same unit, Spaul (2000: 292-294) believes that they are in fact different units. Wagner (1938: 168-171), moreover, also distinguished between the two units, describing, as did Spaul both a cohors I Montanorum, and a I Montanorum civium Romanorum. 233

219

Spaul 2000: 429. CIL XVI.22; RMD IV.209. 221 Wagner 1938: 119. Kraft (1951: 173) believes that it made its appearance in Moesia in AD 78. 222 Matei-Popescu 2013: 219. 223 ILS 8968. 224 Devijver 1987: 178. 225 Matei-Popescu 2013: 216-217. See ILBulg 50=Krier 59=AE 1960, 127. 226 AE 1988, 988. 227 IScM II.172, ILBug 305. Cf. Matei-Popescu 2013: 211, who discusses aspects of the unit’s officers’ early prosopographical history. 228 AE 1964, 262. At the very least it was in Upper Moesia by AD 94. cf. CIL XVI.39; Matei-Popescu 2013: 220. 229 AE 1957, 189. 230 CIL XVI.4. 231 RMD I.2. 220

18

The Lower Danube from 27 BC to AD 81 Hispanorum, Ubiorum, Tyriorum, and the Cilicum. Of those units, we already know that the cohors Cilicum was in the province. The I Cantabrorum may have been in Moesia by the reign of Claudius on the basis of an inscription found at Brixia. 251 Vespasian might have moved it to Moesia after the men were raised in Spain. 252 It has left little trace anywhere in the empire, 253 and it is not entirely clear what happened to the unit after AD 75. Perhaps it should be associated with the cohors II Cantabrorum that we find in Judea. 254 With respect to the cohors I Thracum Syriaca, there is little evidence for this unit in Syria after AD 70 and so it was probably transferred to Moesia early in the reign of Vespasian. 255 It might have come with Licinius Mucianus towards the end of the civil war with the legio VI Ferrata. 256 The cohors I Sugambrorum tironum is a distinct unit from the aforementioned cohors I Sugambrorum veterana, as both the name veterana from the other cohort and a later diploma prove. 257 Based on the title tironum, this unit probably came to Moesia after the veteran cohort of Sugambrians, and before AD 75, given that an earlier Moesian diploma lists the other Sugambrian unit as veterana. 258

The Evidence from the Diplomata The first Moesian constitution that lists auxiliaries is dated to April 28, AD 75. 240 It lists the following ten cohorts: I Antiochensium, I Sugambrorum veterana, I Raetorum, I Lusitanorum, the III, IV, V, VII, and VIII Gallorum, and the Cisipadensium. The cohors I Raetorum had arrived from Cappadocia, 241 and is later found in Moesia Inferior, if briefly. 242 It might have been posted to Raetia at the end of Trajan’s Dacian We saw earlier that the cohors I wars. 243 Sugambrorum veterana was in Moesia. This diploma also lists a considerable number of Gallic units. The Third cohort of Gauls might have been transferred to Moesia sometime in the second half of AD 74 since it is listed in a German diploma dated to May 21st of that year; but it is absent from the next two German diplomas. 244 Matei-Popescu notes that it was recorded in Germany in AD 42 at the latest. 245 The Fourth cohort of Gauls also seems to have arrived sometime before AD 75, for a veteran of the unit is recorded on an inscription from Oescus. 246 Using that inscription, Matei-Popescu postulates that the unit might have been based in the fort between the years AD 62 and 71. 247 It would not be unreasonable to suggest that it came to Moesia with the Third cohort, though not necessarily from the same place. 248 The same is probably true for the Fifth, Seventh, and Eighth units. The unit of Cisipadenses may have been stationed in Moesia from the reign of Tiberius; it remained in Moesia until the Finally, the cohors I death of Domitian. 249 Lusitanorum had been in Moesia since the latter half of Nero’s reign. The First cohort of Raetians might have been transferred to the Danube with Licinius Mucianus. The cohors VIII Gallorum has left no other trace beyond this handful of diplomas dating to AD 75, and so its possible that the unit was disbanded, or even destroyed during Domitian’s conflict with the Dacians in the mid 80s AD.

Once upon a time, the cohors I Bracaraugustanorum would have represented something of a problem, for prior to the publication of the diploma, RGZM 1, the earliest evidence we had for it in Moesia came from nearly a decade later. There was a cohors I Bracarum in Mauretania Tingitana on January 9, AD 88, and that unit had been identified with this Moesian unit. 259 Given the evidence of this earlier diploma it seems unlikely, if not necessarily impossible, that these two units, the cohors I Bracaraugustanorum from this diploma and the cohors I Bracarum from Mauretania, are one in the same. 260 We have, in fact, two distinct, if similarly named, units – and we will return to the cohors I Bracarorum in chapter two below. The cohors I Bracaraugustanorum found on the constitution of AD 75 remained in Moesia, and when the province was divided it was based in Moesia Inferior. 261 Before its arrival in Moesia, the unit seems to have been based in Dalmatia during the JulioClaudian era. 262

There is another constitution from April 28, AD 75, and three published diplomas. 250 The units listed consist of cohorts alone: the cohortes I Cantabrorum, I Thracum Syriaca, I Bracaraugustanorum, I Sugambrorum tironum, II Chalcidenorum, III 240

RMD I.2. See Christol and Drew-Bear 1995. 242 Popescu and Petolescu 2004. 243 Matei-Popescu 2010a: 228. 244 CIL XVI.20. It is absent from CIL XVI.36, which dates to AD 90 and is from Upper Germany. It is also absent from the next diploma from Lower Germany (RMD I.52). Of course, these two diplomas come from quite a bit later and the diploma from Lower Germany dates to AD 158. Still, there is no reason to suspect that we simply lack the evidence to prove its existence in Germany sometime earlier than AD 158, but later than May 21, AD 74. 245 Cf. Matei-Popescu 2013: 220. 246 CIL III.14417=ILBulg 61. 247 Matei-Popescu 2013: 220. 248 See Spaul 2000: 164-165. 249 Wagner 1938: 121; Holder 1980: 304; Saddington 1982: 166; Roxan and Weiss 1998: 371-420; Spaul 2000: 464; Matei-Popescu 2013: 219. 250 RGZM 1; Weiss 2008: 270-273; Eck and Pangerl 2009: 506-509. 241

251

Matei-Popescu 2013: 219. Cf. Holder 1980: 304. Spaul 2000: 69-71. 253 Note CIL V.4326; ILJug II.463; InscrIt 10-5, 737; AE 2008, 1728; and AE 2010, 1853, for instance. 254 CIL XVI.33; AE 2005, 1731. 255 Cf. Matei-Popescu 2013: 221-222. 256 Mocsy 1974: 81; Spaul 2000: 366. 257 RMD IV.222. 258 RMD I.2. The diploma names the cohors I Sugambrorum veterana. 259 CIL XVI.159. cf. Cichorius 1900: 255; Wagner 1938: 97-98; Kraft 1951: 170; Beneš 1978: 18; Strobel 1984: 125; Knight 1991: 200; Spaul 2000: 89; Tentea and Matei-Popescu 2004: 274. 260 The relative lack of diplomas for Mauretania complicates attempts at recovering the history of this unit in North Africa; regardless, such an enterprise is beyond the scope of this book. 261 Note, for instance, the evidence of Petolescu and Popescu 2004. 262 Alföldy (1987): 249-250. 252

19

Exercitus Moesiae The cohors II Chalcidenorum first shows up in Moesia in this diploma. 263 We can then rule out the possibility that Trajan brought the unit to the province. 264 Instead, like a number of other units, Vespasian probably raised the unit, and perhaps when he was active in the east. 265 The cohors I Tyriorum, judging by its nomenclature, was raised in Syria. This diploma marks its historical debut. Given the date of this constitution, the unit, like the II Chalcidenorum, may have been raised in the east during the civil wars and then sent to the Danube by Vespasian along with Licinius Mucianus. The cohors Ubiorum was transferred to the province as early as the reign of Vespasian, if not Claudius. 266 Finally, with respect to the cohors III Hispanorum, although there might have been an ala Hispanorum based in Moesia, there is no additional evidence for the presence of this cohort prior to the date of this constitution. At present, there is no evidence for the unit anywhere else – empire-wide – after this date, which suggests its disbandment, or even destruction, perhaps at the hands of the Dacians in mid-80s.

lack of a number associated with the unit name on an official document such as this suggests that this might have been the only such unit at this time. 273 Spaul, however, claims, “many units use two forms, one a simple ethnic name without numeral, and the other with numeral.” 274 No distinction is made, however, between different types of documents in his statement. As both documents listed here are diplomas, which purport to be official documents, and ones of considerable importance to their owners, one wonders why the lapicide might have been allowed to miss what is presumably an important part of a unit’s name. 275 On balance we must leave open the possibility that there were in fact two units. Both Wagner and Kraft distinguished between the cohors Mattiacorum and the cohors II Mattiacorum, as do Tentea and MateiPopescu. 276 Wagner even refers to two inscriptions that list, respectively, a L. Spurennius Rufus bucinator ch(o)r(tis) Mattiacorum, 277 and a L. Clodius Ingenuus praef(ectus) coh(ortis) Mattiacor(um); 278 however, he does raise the possibility that the cohors Mattiacorum developed into the cohors II Mattiacorum. Two later constitutions from Moesia Inferior that list a unit of Mattiaci date to August 14, AD 99. 279 They list a cohors II Mattiacorum. Although this might only be a mistake on the part of the lapicide, it is not unreasonable to suppose that the original unit, the cohors Mattiacorum, was destroyed when either Oppius Sabinus or Cornelius Fuscus suffered serious defeats, likely in the years AD 85-86. Thus, the cohors II Mattiacorum may be the creation of a levy subsequent to the destruction of the original unit of Mattiaci.

There are a number of diplomas from Moesia for AD 78, which are drawn from three different constitutions, many of which only consist of tiny fragments that add almost nothing to our knowledge of the province’s auxiliary unit deployment. There are a number of constitutions, which list cohorts alone and date to Feb. 7, AD 78. 267 Another constitution lists alae alone. 268 On these, the majority of the auxiliary units are found on at least one of the earlier diplomas, with some exceptions. There is, for instance, no evidence for the cohors II Lucensium before these diplomas. There is another unit listed in these diplomas of 78, as well as some of the earlier diplomas of 75, which I have not yet discussed: the cohors Mattiacorum. Spaul records the existence of only one of these units, that being the cohors II Mattiacorum. 269 There is no evidence, however, that Spaul’s unit is the one listed here. Upon inspection of tabella I for RMD IV.209, at least in the photograph provided on page 404, there is no room for the insertion of a “II” before the “Mattiacorum” in the unit’s name. 270 The numbers are clearly reproduced and visible for other units, such as the I Sugambrorum and the III Gallorum. 271 Moreover, RMD IV.208 was restored to correspond with CIL XVI.22. 272 Although CIL XVI does not provide a photograph of number 22, there again seems to be no indication that the “II” was intentionally omitted. The 263

Cf. Tentea and Matei-Popescu 2004: 278. This was the view of Strobel (1984: 126). 265 Wagner (1938, 118) suggested long ago that the unit was raised during the Flavian dynasty. 266 Matei-Popescu 2013: 222. 267 CIL XVI.22; RMD IV.208; Eck and Pangerl 2008: 319-321, 321323; Weiss 2008: 270-275; Eck and Pangerl 2010: 237-243. 268 RMD V.325; Eck and Pangerl 2008: 324-326. 269 Spaul 2000: 243-244. 270 Roxan and Holder 2003: 404. 271 Roxan and Holder 2003: 404. 272 Roxan and Holder 2003: 405 n. 4. 264

273

Devijver 1987: 178. Spaul 2000: 244. 275 Indeed, it is an issue to which we will return in chapter six below, whether the names on diplomas should be considered official. 276 Wagner 1938: 164-165; Kraft 1951: 180. Cf. Tentea and MateiPopescu 2004: 288-289, though see Matei-Popescu 2013: 221. 277 CIL III.12437. 278 CIL VI.37247. 279 CIL XVI.44, and 45. 274

20

The Lower Danube from 27 BC to AD 81 Dardanorum. 286 Both Holder and Matei-Popescu argue for an ala Gallica stationed in the province in the middle of the first century based on a tombstone from Following from that, Matei-Popescu Ratiaria. 287 argues, quite plausibly, that the ala Gallica is likely identical to the ala I Gallorum from this diploma. 288 In turn, this unit is associated with the ala I Claudia Gallorum Capitoniana later attested in Dacia Inferior under Hadrian. 289 The ala Claudia nova, listed on a German diploma dated to September 20, AD 82, 290 also seems to have been transferred to the province sometime under Vespasian. 291 This cavalry unit was in Germany on May 21, AD 74, and so was transferred to Moesia sometime in between. Perhaps it was transferred with the Gallic cohorts in the second half of AD 74, or early in AD 75. 292

Legions (3) I Italica, V Macedonica, VII Claudia

Auxiliary Units (21) ala Asturum, ala Atectorigiana, ala Claudia nova, ala Dardanorum, ala I Gallorum, ala Gallorum Flaviana, ala Hispanorum, ala I Vespasiana, cohors Antiochensium, cohors I Bracaraugustanorum, cohors I Cantabrorum, cohors Cilicum eq, cohors III Gallorum, cohors IV Gallorum, cohors V Gallorum, cohors VII Gallorum eq, cohors VIII Gallorum, cohors II Lucensium eq, cohors Mattiacorum, cohors I Sugambrorum tironum, cohors I Sugambrorum veterana eq, cohors I Thracum Syriaca eq Table 1.2. Dispositions of Legions and Auxiliary Units in Moesia in AD 78

An ala Hispanorum might have been based in Moesia at least as early as the year of the four emperors, though the evidence is slim. 293 Finally, both Holder and Matei-Popescu mention an ala Augusta as a unit based in Moesia, 294 the existence of which hinges on CIL III.12347, which reads: “C. OPPIUS VARUS EQUES ALAE AU[GUSTAE]”. What is more, Holder believes that it could be associated with the ala Augusta attested in Spain in AD 40, or the ala Augusta Germaniciana recorded in Galatia. Regardless of its identification with two other units, the evidence for its possible presence in Moesia is too thin; although plausible, it is not certain that “GUSTAE” is what should be restored after “AU”. Thus, until additional evidence appears, its presence in Moesia must remain a possibility and not a certainty, at best.

We end this discussion of auxiliary diplomas preDomitian with one dated to September 20, AD 80, from Germany, on which the ala Scubulorum is listed. This same cavalry unit had departed Moesia sometime after the civil war and was stationed in Germany by the 21st of May, AD 74. 280 That diploma also lists a future Moesian ala, namely the ala Claudia nova. This unit was part of the army of Upper Moesia in AD 93. 281 Three other cavalry units had departed Moesia by the 15th of April, AD 78: the ala Sulpicia, the ala Moesica, and the ala Afrorum veterana. 282 These three units also departed from the Balkans for Germany in the wake of the civil war or the Batavian revolt. The remaining five alae, the Asturum, the Atectorigiana, the I Gallorum, the Gallorum Flaviana, and the I Vespasiana Dardanorum all remained in Moesia through the reign of Trajan.

CONCLUDING THOUGHTS As we have seen, our evidence for troop movements in this region is thin, especially for the auxiliaries before we have diplomas. The legions are better served. This makes it hard to say too much about how troop movements changed from year to year, or even decade to decade. With those difficulties in mind, it is worth remembering the data presented in tables 1.1 and 1.2 found above, which provided two snapshots of the garrison of the province: one from AD 40 in the years

Poorly Attested Flavian Dynasty Auxilia There are a number of alae for which we have little evidence. As noted, it is more than likely that two of the four alae discussed above, namely the I Asturum and the Atectorigiana, remained posted in Moesia in the interim. 283 Holder and Matei-Popescu note the presence of an ala (Claudia Gallorum) Capitoniana in Moesia. 284 The evidence for this is a lone tombstone from Augusta. 285 We do have one Moesian diploma, dated to AD 75 or 78, which may have listed the following three unattested units: the ala I Gallorum, the ala Gallorum Flaviana, and the ala I Vespasiana

286

RMD IV.209. The inscription (AAntHung-1967-102) reads: “]nis f(ilius) Fuscus eq(ues) ala Gal(l)i/ca vixit an(nos) / XL meruit / an(nos) XXV h(ic) s(itus) e(st) / herede[s et(?)] / frater [3] / [3]MV[“. Holder 2005: 80; Matei-Popescu 2013: 213. Note Holder’s comments at 80, n. 18. 288 Matei-Popescu 2013: 213. 289 AMN 2006/2007, 186=SCIVA 2009, 318 =AE 2007, 1759; AMN 2006/2007, 190=SCIVA 2009, 319; AMN 2006/2007, 192. 290 CIL XVI.28. This diploma also names the Third cohort of Gauls. 291 The evidence for a transfer to Moesia before the reign of Domitian comes from a fragmentary diploma (Eck and Pangerl 2008: 324325). Cf. Matei-Popescu 2013: 212. 292 CIL XVI.20. 293 CIL III.12361, III.12378; Tentea and Matei-Popescu 2004: 268; Matei-Popescu 2013: 215-216. Cf. CIL V.4058. 294 Holder 2005: 80; Matei-Popescu 2013: 210-211. 287

280

CIL XVI.20. CIL XVI.39. CIL XVI 23, from Germany. 283 All three are recorded in later diplomas. 284 Holder 2005: 79-80; Matei-Popescu 2013: 214. Cf. AE 1941, 105; 1967, 425. 285 AE 1912, 187. 281 282

21

Exercitus Moesiae leading up to the annexation of Thrace, another from AD 78, as we near the end of Vespasian’s reign and the start of a period of significant activity on the lower Danube. In table 1.1, which corresponds to AD 40, based on the evidence that we have there were 2 legions and 12 auxiliary units in Moesia at that time. Nearly four decades later, in table 1.2, we saw that those totals had increased significantly. There were now 3 legions, and 24 known auxiliaries. This increase is likely down to the growing importance of the frontier and the specific challenges that it posed, which is evidenced, in part, in the increasing number of conflicts in this area and beyond. Stationing troops on the lower Danube provided a ready body of troops that could be dispatched to points east, west, and north. We will wait to chapter six below to discuss the implications of these changes.

22

Diurpaneus and the Dacians. 306 By the end of AD 86, Domitian had divided Moesia into two provinces: Moesia Inferior to the east and Moesia Superior to the west, with the boundary fixed originally on the Cebrus river, and later shifted west to the Almus (both in Bulgaria). 307 In AD 88 another campaign against the Dacians was launched, this time under the command of L. Tettius Iulianus. 308 Iulianus got as far as the Dacian capital Sarmizegethusa. 309 The war broke off during the winter of AD 88 – as it had during the winters of the previous campaigns – and by June and July of the following year, AD 89, peace negotiations had begun with Decebalus, the new Dacian king, who was in charge by AD 87. 310 By the end of July, a peace treaty had been signed and Dacia was now recognised as a client kingdom. 311

Chapter 2: The Lower Danube from AD 81 to AD 161 INTRODUCTION This chapter is concerned with the busiest period in Moesia’s history under Rome, at least before late antiquity. 295 We start, again, with an outline of the key conflicts, which includes the famous wars against Decebalus and the Dacians. Next we turn to the legions before finishing with the auxiliary units. In this case, the auxiliary garrison of the two new provinces, Upper and Lower Moesia, created in AD 85 or 86, are discussed separately. BACKGROUND: MILITARY HISTORY Domitian

Trajan 312

As early as AD 81 there seem to have been problems on the Moesian frontier. A two-province diploma for AD 82 (CIL XVI.28) and coin hoards found in the province, for instance, suggest instability. 296 Just a few years later, in AD 84, 297 or even June of AD 85, the Dacians under their king Diurpaneus (Decebalus’ predecessor) invaded what later became Moesia Inferior. 298 Some claim that Diurpaenus should be equated with Decebalus, 299 but there is no evidence to support such an assertion. 300 The impact was significant and immediate, for it was in the middle of this year that the governor C. Oppius Sabinus fell, 301 and we learn that the winter-quarters of the legions were threatened. 302 In addition, Jordanes tells us that “many fortresses [castella] and cities were seized”. 303 Domitian then prepared a counterattack and sent in Cornelius Fuscus against the marauding Dacians. This Roman force attacked Diurpaneus on the Danube. 304 However, by July of AD 86, Cornelius Fuscus had fallen. 305 At some point in September or October of that year, a certain M. Cornelius Nigrinus Curiatius Maternus launched another campaign against

Trajan launched two Dacian campaigns: the first one ran from AD 101 through AD 102, the second from AD 105 through AD 106. 313 It was in AD 106 that the province of Dacia was created. Before the wars began, Trajan made a treaty with the Iazyges. 314 This Sarmatian tribe then provided additional troops for Rome’s campaigning army. 315 In AD 101 Trajan marched as far as Tapae and engaged the encamped Dacians. 316 Before the Romans won, Decebalus sent envoys in the hopes of reaching a new treaty. Dio’s excerpter, however, says that Decebalus’ inability to make the plea in person helped ensure that no agreement would be reached. 317 Trajan then crossed the Danube and marched towards the interior of Dacia. 318 While marching through Dacia, “Trajan seized some fortified mountains…”. 319 This success led Decebalus to sue for peace again; Dio’s excerpter claims that this was so that he could recuperate from his reverses and strengthen his position. 320 At the conclusion of the first war, Decebalus “was ready to agree without exception to every demand that had been made.” 321 Trajan then left the camp at Sarmizegethusa,

295

306

Note the comments of Wilkes 2005: 151-152. Dušanić 2007, 82-85; and Wheeler 2011, 209-210, with additional bibliography. 297 Stefan 2005: 399-407. Cf. Wheeler 2011: 210. 298 My summary of the events of the reign of Domitian is based, in part, on the reconstruction of Strobel (1989) and Stefan (2005: 399424). For detailed discussion consult Stefan (2005: 399-424). 299 Fear 2010: 311, n. 140. 300 Orosius, who names Diurpaneus, relies for his brief summary of Domitian’s wars with the Dacians on Tacitus, an historian who strikes me as unlikely to have muddled the two names. With Jordanes, the source of the other reference to Diurpaneus, I think there is reason to suppose that he too used Tacitus. See Stefan 2005: 399-400, Wheeler 2010: 1216-1220. Cf. Zahariade 2012a. 301 Suet. Dom. 6; Cass. Dio 67.6.1; Eutr. 7.23; Jord. Get. 13.76. See Stefan (2005: 400-407). 302 Tac. Agr. 41. 303 Jord. Get. 13.76. 304 Suet Dom. 6; Cass. Dio 67.6.5; Jord. Get. 13.77; Cf. Strobel 1989: 116; Stefan 2005: 402-407. 305 Suet. Dom. 6; Orosius 7.10.4; Jord. Get. 13.77; Cf. Strobel 1989: 117; Stefan 2005: 404.

Strobel 1989: 117; Stefan 2005: 406. Strobel 1989: 117-118; Wilkes 2005: 140. Hadrian might have been the driver of this later reorganisation (Wilkes 2005: 139). 308 Stefan 2005: 411-402. 309 Strobel 1989: 118; Stefan 2005: 418-420. 310 Strobel 1989: 119; Stefan 2005: 425-437. 311 Mart. 6.76; Cf. Strobel 1989: 119; Stefan 2005: 425-437; Wheeler 2010: 1220, n. 90; Wheeler 2011: 210. 312 The reconstruction of the Dacian wars of Trajan is based on the literary evidence and is supplemented by the artistic evidence. 313 Detailed discussion of the wars, and everything from the armies involved to the topography of the campaigns, can be found in Stefan (2005: 503-704). See too Wheeler (2010, 2011). 314 Mocsy 1974: 91. 315 Note Stefan’s (2005: 624) comments about the peace of AD 102. 316 Cass. Dio 68.8.1-2; Stefan 2005: 550-554. 317 Cass. Dio 68.9.2-3; Trajan’s Column Plate 45. 318 Trajan’s Column Plates 4-5. 319 Cass. Dio 68.9.3, trans. Cary; Trajan’s Column Scenes 27-28, 30, 39-40, 43, 46, 52. See Stefan 2005: 572-594. 320 Cass. Dio 68.9.5; Trajan’s Column Plates 54-55. 321 Cass. Dio 68.8.3, trans. Cary. See Stefan 2005: 624-632.

296

307

23

Exercitus Moesiae where the treaty had been ratified, and returned to Italy thus ending the first Dacian war. 322

Although the reign of Antoninus Pius has in the past been seen as quiet and uneventful largely on the basis of the silence of the surviving literary sources, evaluations of other evidence have revealed something else. Varbanov, for instance, has found traces of conflict in Lower Moesia and Thrace from 138-192 (the latter of half of which we discuss in the next chapter). 338 Indeed, it seems that there is evidence of conflict at least twice during the reign of Antoninus Pius, during 143 and 156/157. 339 Some of this conflict also seems to have afflicted the settled regions along the northwestern shore of the Black Sea and the Chersonesus, though more on this in chapter four below. 340

By AD 105, “Decebalus was reported to him to be acting contrary to the treaty in many ways…” 323 As a result, “the senate declared him an enemy and Trajan once more conducted the war against him in person instead of entrusting it to others.” 324 By the summer of AD 105 the Romans had set out again. 325 Dio’s excerpter tells us that the war was running fairly smoothly for the Romans, and poorly for the Dacians. So, Decebalus attempted to assassinate Trajan through the use of deserters. 326 The Dacian king did manage to capture one of the Roman commanders, a certain Longinus, though the general later tricked Decebalus and Longinus committed suicide. 327 Trajan built his famous bridge across the Danube at this time, which the excerpter of Dio describes at great length. 328 After Trajan crossed the Danube he steadily marched towards Sarmizegethusa and “crushed the Dacians.” 329 Decebalus was then captured, and later committed suicide. 330 The war was over by the end of the summer of AD 106, and Trajan then set about reorganising the conquered territory into the new province of Dacia. 331

LEGIONS FROM DOMITIAN THROUGH ANTONINUS PIUS We now turn to troop movements. In AD 81 there were three legions stationed in Moesia: the I Italica, the V Macedonica, and the VII Claudia. Were any legions lost in Domitian’s Dacian conflicts? There were at least three possible times that this could have happened, with scholars generally focusing on the defeats in 85 (Oppius Sabinus), 86 (Cornelius Fuscus), and 92. 341 Regrettably, there is not a great deal of verifiable information about the various conflicts in and around Moesia. 342 Beginning with one of the latest accounts, the fourth century historian Eutropius referred to four campaigns during the reign of Domitian, three of which are pertinent to this discussion: one against the Sarmatians, and two against the Dacians. 343 These campaigns led to triumphs and a laurel crown, or so Eutropius, though along with them came a number of military disasters. He said, “in Sarmatia his [Domitian’s] legion was Suetonius said destroyed with its general”. 344 something similar, noting the destruction of a legion with its commander in Sarmatia. 345 Eutropius also referred to the defeats of (Oppius) Sabinus and Cornelius Fuscus, who were cut down with their great armies (cum magnis exercitibus) by the Dacians. 346 We do not know, however, what “great armies” means, exactly. Nor does Eutropius’ account give us any indication of the chronology. The loss of a legion seems to have been in AD 92, after the last of Domitian’s clashes with the Dacians had ended, that is with the aforementioned Sarmatian campaign; the loss of a legion might explain his celebration of a laurel crown alone. 347 Were any of the defeated troops,

Hadrian and Antoninus Pius Another major disturbance occurred during the reign of Hadrian. Hadrian’s biographer notes that trouble arose shortly after he ascended to the throne. “For those nations which Trajan had subjected rebelled”, including the Sarmatians, “who were waging war.” 332 Apparently, following the second war with the Dacians, Trajan had concluded a peace treaty with the For reasons that are unclear, the Rhoxolani. 333 Rhoxolani developed some sort of quarrel with Rome over this treaty and thus banded together with their neighbours to wage war. The governor of Dacia, C. Julius Quadratus Bassus, was killed. 334 When Hadrian heard news of the incursions of the Sarmatians and Rhoxolani and the death of Bassus, he set out for Moesia himself. 335 He put Marcius Turbo in charge of the Pannonian and Dacian command and eventually made peace with the Rhoxolani king. 336 This conflict lasted from AD 117 to AD 119. 337

322

Cass. Dio 68.9.7. Cass. Dio 68.10.3. trans. Cary. 324 Cass. Dio 68.10.4, trans. Cary. 325 Trajan’s Column Plates 59-60, 66. 326 Cass. Dio 68.11.3. 327 Cass. Dio 68.12.1-5; Trajan’s Column Plates 67-112; Stefan 2005: 649-655. 328 Cass. Dio 68.13.1-6; Cf. Aur Vict. Caes. 13. 329 Cass. Dio 68.14.1. 330 Trajan’s Column Plates 108-109; Stefan 2005: 663-665. 331 Wheeler 2010: 1227. 332 HA Hadr. 5.2. 333 Mocsy 1974: 100. 334 Mocsy 1974: 100. 335 HA Hadr. 6.6. 336 HA Hadr. 6.7-8. 337 Mocsy 1974: 99. 323

338

Varbanov 2007. Note the map on p. 155. Varbanov 2007: 153. Cf. HA Ant. Pius 5.4-5. 340 Varbanov 2007: 154. 341 Turner 2010: 185. 342 Stefan 2005: 404. 343 Eutr. 7.23. 344 Eutr. 7.23. 345 Suet. Dom. 6. 346 Eutr. 7.23. 347 Syme (1928: 45) suggested that it was the XXI Rapax that was annihilated by the Sarmatians in AD 92, but more recently Bérard 339

24

The Lower Danube from AD 81 to AD 161 campaign, 355 though they were not the first. The aforementioned prefect, Cornelius Fuscus, was sent to Dacia and died in 86 or 87; 356 he may have arrived as early as 84 or 85. 357 We do not know how many praetorians accompanied him, though Bingham argues that Martial’s commemoration of a fallen soldier is suggestive. 358 Getting back to Trajan, a number of praetorians along with their prefect Tiberius Claudius Livianus fought in the war with the participating legionaries and auxiliaries. 359 Some of the evidence for their participation comes from Trajan’s Column, as difficult as it can be to distinguish between unit types. 360 Incidentally, praetorians also seem to have fought against the Parthians and the Marcomanni during the reign of Marcus Aurelius. 361 Although we cannot be certain that they were ever in Moesia, and they were certainly never based there, their presence in the Dacian wars suggests that they were active in the region, if briefly.

including that legion which suffered in Sarmatia, based in Moesia? With respect to the latter, that depends in part on where the “Sarmatia” that Eutropius referred to was. 348 Scholars have speculated that the conflict was somewhere in the general region of the later provinces of Lower Moesia and Dacia. This would mean that there was a possibility that the Moesian army had suffered some losses. With respect to the Dacian conflicts, the later province was located north of Moesia and so there is every reason to believe that they would have been involved. We still do not know, however, the specific impact on the Moesian army. As noted, Eutropius’ evidence is rather late; we do have material that is a bit more contemporary, but it is just as problematic. Tacitus vaguely noted that “so many armies [tot exercitus] in Moesia and Dacia and Germany and Pannonia were lost with the thoughtlessness or through the laziness of their generals; so many military men [militares viri] with so many cohorts [cum tot cohortibus] were killed or captured.” 349 There was the aforementioned account of Suetonius, who had said that both Oppius Sabinus and Cornelius Fuscus, whom Domitian had entrusted with the conduct of the war, were defeated. 350 We also have the account of the excerpter of Dio who claimed that when Trajan took some of the Dacians’ fortified positions he “found the arms and the captured engines, as well as the standard which had been taken in the time of Fuscus.” 351 Despite being closer in time to the events in question, these brief accounts of the loss do not give specifics about the military units involved.

With regard to the possibility of identifying depleted units from the Moesian army, we are no better off. There is an additional piece of evidence that might provide some independent information: this is the monument at Adamklissi with the inscription that lists the names of close to 3800 Roman soldiers. 362 Unfortunately, the dedicatory portion of the inscription is missing so we cannot date it. Given that at least one of the two other monuments at Adamklissi is Trajanic in date (the Tropaeum Traiani), it is quite possible that this monument is too. What is more, the only reference to a particular unit that we find on it is to a cohors II Batavia; there is no indication whether any, or better how many, of the soldiers named on the inscription were legionaries. Additional evidence of destruction must be sought elsewhere, then.

In fact, the only clue we have lay in Dio’s comment that a standard from the defeat of Fuscus had been recovered, for Suetonius said that Cornelius Fuscus was the prefect of the praetorian guard. 352 This has, in part, led scholars to speculate that Fuscus’ army was a motley force composed of a number of vexillations, along with auxiliaries and some members of the praetorian guard. 353 It is possible that the standard in question had belonged to a blighted legion – the cohorts referred to by Tacitus may have belonged to just such a unit. On the other hand, the auxiliaries had cohorts, as did the praetorians, and given Fuscus’ position as praetorian prefect, it seems reasonable to suggest that the lost standard did belong to the praetorian guard. 354

The literary evidence is unanimous in its assertion of some significant destruction of the region’s military, even if scholars no longer believe that a legion had been lost, as they did previously. 363 There is some circumstantial evidence that, if nothing else, is suggestive. For one thing, Domitian’s eagerness to sign a treaty with the Dacians at the close of these campaigns might indicate that he was concerned about the current state of the military, i.e., it had suffered significant losses. There also happens to be evidence for the transfer of at least one legion to the region around the presumed time of the war, which in itself could be indicative of the scale of the losses.

Although the praetorian guard are not the focus of this study, they do seem to have been active on the lower Danube during Trajan’s Dacian wars. Bingham notes that a contingent of the guard accompanied Trajan on

355

Bingham 2013: 41. Cass. Dio 67.6.6, 68.9.3. cf. Juv. 4.110-112. 357 Roxan 1993: 72. 358 Bingham 2013: 149, n. 181; Mart. 6.76. 359 Bingham 2013: 41. 360 Rossi 1971: 100-102; Bingham 2013: 41. 361 Bingham 2013: 41-42. 362 CIL III.14214=ILS 9107. For a detailed discussion of this monument see Turner 2013. 363 See Rossi (1971: 22) and Mocsy (1974: 82) for the supposition that a legion was lost in combat. 356

(2000: 56) says that it may have existed at least until the reign of Trajan. 348 Sarmatia is generally understood to have been opposite the Pannonias, but Eutropius might not have known this. 349 Tac. Agr. 41. 350 Suet. Dom. 6. 351 Cass. Dio 68.9.3, trans. Cary. 352 Suet. Dom. 6. cf. Bingham 2013: 148-149, n. 181. 353 Stefan 2005: 403. cf. Strobel 1989: 53. 354 Wilkes 1983: 289, n. 42.

25

Exercitus Moesiae The IIII Flavia was a creation of Vespasian, as its name implies, 364 and it had been based in Dalmatia. The legion was later transferred to Upper Moesia, though we do not know when exactly. Most scholars believe that it had moved to Moesia Superior by AD 85 or 86, though there is no firm evidence for this. 365 This supposition, originally Ritterling’s, is based on the presumed date for one of Domitian’s campaigns against the Dacians, AD 86, and the date of the division of the province, also AD 86. The reasoning is sound enough: our evidence for the legion in the province tends be from Upper Moesia, and there is every reason to believe that additional troops would have been brought in for the aforementioned campaign. Although there is some speculation that the unit was based at Viminacium initially, its eventual headquarters was Singidunum. Interestingly enough, it was not constructed until at least the reign of Nerva; the earliest attested emperor on the coinage found on site dates to the reign of Nerva. 366 Thus, while a date of 86 for the transfer seems plausible, it is not certain; some would prefer a date after 89. 367 At the very least, the legion had arrived by 98.

career in the same area, then on the basis of the dispositions of these auxiliary units we ought to conclude that the I Adiutrix operated in Pannonia or Moesia during the reign of Domitian. 372 Of these two possibilities, that is a posting in Pannonia or Moesia, a Pannonian posting is preferable since we know that there is at least one attestation of each of those cohorts in Pannonia (rather than Moesia) during the wars. 373 A Pannonian base, however, need not preclude the legion from operating in the Moesias during the Dacian or Sarmatian wars. 374 In sum, by the end of the Flavian dynasty, AD 96, the garrison of the now two provinces of Moesia contained all three of the legions that had been in the province previously. Only now, the I Italica and the V Macedonica were in the new province of Moesia Inferior, while the VII Claudia was in the new province of Moesia Superior. The legio IIII Flavia, upon its arrival in the Moesias, possibly during the reign of Domitian, was based in Moesia Superior as well at Singidunum.

There is also some indication that the I Adiutrix was transferred there for the Sarmatian war, if not the Dacian war. 368 There is an inscription that suggests that a vexillation of the legion had participated in one of Domitian’s Balkan conflicts. 369 The legion certainly seems to have been in the region during the reign of Nerva as the Q. Attius Priscus, mentioned in CIL V.7425, received the dona donativa from that emperor for the war with the Suebi. 370 This Priscus, who was a tribune of the I Adiutrix, was also prefect of the cohors I Hispanorum, the cohors I Montanorum, and the cohors I Lusitanorum. There was a cohors I Hispanorum in Illyricum in AD 60. This unit had moved to Moesia Inferior by AD 99. 371 A different cohors I Flavia Hispanorum had arrived in Moesia sometime between AD 78 and AD 80. We do not know, however, which Spanish cohort is mentioned in Priscus’ inscription. There were two cohortes I Montanorum on the Danube in the last quarter of the first century. One of those units remained in Pannonia through the reign of Trajan. The other cohort had moved to Moesia Superior by AD 96. There was a cohors I Lusitanorum in Moesia by AD 75; there was also a cohort of Lusitanians in Pannonia around AD 84. If we assume that this officer spent the bulk of his

By the start of Trajan’s first Dacian war in 101, the disposition of the legions remained the same as it had been under Domitian. Unfortunately, we cannot trace the movements of the legions through the course of the Dacian wars. But we do have a good idea about which additional legions were transferred to the area for the war. There is an inscription from Rome that lists the names of the empire’s legions in order based on the province and the city where their camps were located. It dates to the middle of the second century, and perhaps to the beginning of the reign of Marcus Aurelius. 375 This inscription tells us that the legio IIII Flavia and the legio VII Claudia were based in Moesia Superior; and that the legio I Italica, the legio V Macedonica, and the legio XI Claudia were based in Moesia Inferior. The information contained therein, when used in conjunction with some additional evidence, indicates that one extra legion was added to the Moesias following the Dacian wars of Trajan, namely, the Eleventh Claudian Legion, though the specific year for the arrival of that legion is unknown. We know that at the start of Trajan’s reign (AD 98) the legion had been based in Upper Germany at Vindonissa. 376 In between its postings in Germania Superior and Moesia Inferior, the legion was posted in Pannonia. Campbell, Fellman, Pollard and Berry, and Wesch-Klein date the transfer to either AD 100 or AD 101. 377 The earliest evidence for the legion belonging to the army of Moesia Inferior comes from an inscription in Rome that lists a Lucius Paconius Proculus, who was a military tribune of the legio XI

364

Wilkes (1983: 279, n. 42) suggests that it is also possible that the legion was not a new unit, but in fact a reformed older unit, particularly the Fourth Macedonian legion. 365 Ritterling 1925: 1542. Cf. Syme 1928: 46; Parker 1958: 153; Filow 1963: 39-40, 46; Strobel 1989: 44, 71; Bojovic 1996: 54; Le Bohec and Wolff 2000b: 239; Farnum 2005: 19; Campbell 2007; Dando-Collins 2010: 130; Pollard and Berry 2012: 184; WeschKlein 2012: 3. 366 Bojovic 1996: 58-65. 367 Matei-Popescu 2006-2007: 33. Cf. Strobel 1989: 71. 368 CIL V.7425 (from Libarna in Regio IX). 369 ILS 9200. Cf. Saxer 1967: 22-23. 370 The inscription is dated to AD 97. 371 Cavenaile 1975: 184.

372

Syme 1928: 46; Lorincz 2000: 154-155. Pollard and Berry 2012: 192; Wesch-Klein 2012: 1. Parker 1958: 153; Strobel 1989: 72. 375 CIL VI.3492. 376 CIL XIII.6298, AE 1946, 258, 260; AE 1991, 1260. 377 Campbell 1996a: 841; Fellman 2000: 129; Pollard and Berry 2012: 178; Wesch-Klein 2012: 7. 373 374

26

The Lower Danube from AD 81 to AD 161 Claudia. 378 The inscription, which also indicates that he was a prefect of a vexillation from Moesia Inferior, dates to some point between AD 114 and AD 117. This means that the legion had likely moved by AD 114. In fact, some scholars suggest that the legion had arrived by the end of the second war in AD 106 or AD 107, which is probably not far from the truth. 379

Domitian We begin with the auxiliary garrison of Moesia Superior. After the division of the former province of Moesia, the first constitution that we have that hails from Moesia Superior dates to September 16, AD 94. 384 On it we find the following cavalry wings: the ala II Pannoniorum, the ala Claudia nova, and the ala Praetoria. The ala Claudia nova had arrived in Moesia by AD 80, as we saw above, and remained in the province for many years to come. 385 As regards the ala II Pannoniorum, this is the earliest record of its presence in Moesia Superior. It is mentioned in a diploma dated to November 7, AD 88 from Syria. 386 Thus, it presumably was transferred to Moesia Superior at some point in the intervening five years. The ala Praetoria was in Pannonia on September 3, AD 84. 387 It does not appear on any diplomas in the interim; as a result, it likely came to Moesia Superior directly from Pannonia without any other sojourns; it might have moved to Moesia Superior in the context of the Dacian wars of Domitian. 388

The Fourth Flavian Legion, which had arrived by the death of Nerva, had left the province in the midst of the second Dacian campaign. When the new province of Dacia was set-up, two legions were sent: the IIII Flavia, and the XIII Gemina. Following the death of Bassus in the Sarmatian conflict of the years AD 117119, Hadrian’s biographer said, “after he sent ahead the armies he made for Moesia.” 380 We do not know how big those armies were that Hadrian sent ahead with Marcius Turbo. Nevertheless, although the writer’s use of exercitus is hardly precise, that he uses the term for an army rather than a legion might be significant. On this basis, we can probably place the return of the legio IIII Flavia to Moesia Superior to the Sarmation War of AD 117-119. 381

With respect to the cohorts listed on the diploma, four were already stationed in Moesia: the cohors I Cilicum may have been in Moesia by the reign of Claudius; and the cohors I Antiochensium, the cohors V Gallorum, and the cohors I Cisipadensium are all listed in a diploma dated to April 28, AD 75 discussed in chapter one above. 389 That leaves five cohorts unaccounted for: the I Cretum, the I Flavia Hispanorum milliaria, the V Hispanorum, the II Gallorum Macedonica, and the cohors IIII Raetorum. The two Spanish units, namely the cohors I Flavia Hispanorum and the cohors V Hispanorum, are attested in Germany during the reign of Vespasian, but not in a German diploma dated to AD 80. 390 Thus, the units probably had moved to Moesia Superior between AD 78 and AD 80. Both the cohors I Cretum and the cohors II Gallorum Macedonica 391 are hitherto unmentioned units. They were likely raised between the reigns of Nero and Vespasian. The cohors IIII Raetorum is unattested in any earlier diploma and so its history probably began in Moesia Superior. 392

Over the middle decades of the second century the legions became increasingly stationary. In fact, the return of the legio IIII Flavia Felix marks the end of the changes in the legionary disposition of the Moesias for the years from Domitian to Marcus Aurelius. And so, by 161, the legiones I Italica, V Macedonica, and XI Claudia were based in Moesia Inferior; and the legiones IIII Flavia and VII Claudia were based in Moesia Superior. 382 As a result, when troops were needed elsewhere detachments of troops, called vexillations, were sent to the various fronts rather than entire legions. We will turn to these detachments in chapter four below. AUXILIAE: MOESIA SUPERIOR Our discussion of auxiliaries in the rest of this chapter will follow the dates of the diplomas, with the dates and diplomas of Moesia Superior coming first, Moesia Inferior second. There are two constitutions that date to the years just before Domitian split the province, and the fragments we have contain more alae than cohortes, but all the units listed were already in the province. 383

384

CIL XVI.39, RMD V.335, Weiss 2008: 279-280. It was still around in AD 100 (CIL XVI.46), AD 132 (RMD IV.247), and AD 159 (CIL XVI.111). 386 CIL XVI.35. 387 CIL XVI.30. 388 Strobel 1984: 116. 389 RMD I.2. 390 The cohors I Flavia Hispanorum milliaria was in Germany on April 15, AD 78 (CIL XVI.23); the cohors V Hispanorum was in Germany on May 21, 74 (CIL XVI.20). In CIL XVI.158, dated to AD 80 and from Germany, neither Spanish unit, or any other Spanish unit for that matter is mentioned. 391 This unit had been transferred to Dacia by October 14, AD 109 (RMD III.148). It likely went over as part of Trajan’s expeditionary force. Indeed, it is listed to diplomas for Moesia Superior dated to May of AD 100 (Eck and Pangerl 2008: 279, 286, 326). 392 Cichorius 1900: 327; Wagner 1938: 180; Strobel 1984: 141; Spaul 2000: 282. 385

378

CIL VI.32933. Ritterling (1925: 1690) postulates a date of AD 106. Filow (1963: 72) merely states that the legion was in the province by AD 107. Parker (1958: 158), Campbell (1996a: 841), and Fellman (2000: 130) are vague about the unit’s transfer. Cf. Szilágyi 1954: 171; and Strobel 1984: 93-95; Farnum 2005: 22; Pollard and Berry 2012: 180. 380 HA Hadr. 6.6. 381 Piso 2000: 211. 382 Ritterling 1925: 1365; Szilágyi 1954: 179; Parker 1958: 163. 383 CIL XVI.28; Weiss 2008: 275-279. 379

27

Exercitus Moesiae On the constitution dated to July 12, AD 96 there is only one ala listed, the ala Praetoria, which we saw was already in the province. 393 All of the remaining ten units listed are cohorts. Of those ten cohorts, the cohors I Cretum, the cohors I Cilicum, the cohors I Flavia Hispanorum milliaria, the cohors IIII Raetorum, and the cohors V Hispanorum were already in Moesia Superior. 394 There was a cohors I Lusitanorum in AD 92 and AD 99 in Moesia Inferior, but this latter cohors Lusitanorum from Upper Moesia is probably not that unit in Lower Moesia as we will argue below. This cohors Lusitanorum should probably be equated with that unit listed in Moesia in AD 75. 395

Legions (2)

Auxiliary Units (11) IIII Flavia, VII ala Claudia nova, ala II Claudia Pannoniorum, ala Praetoria, cohors I Antiochensium, cohors I Cilicum milliaria eq, cohors I Cisipadensium, cohors I Cretum, cohors I Flavia Hispanorum milliaria, cohors V Gallorum, cohors II Gallorum Macedonica, cohors V Hispanorum Table 2.1. Dispositions of Legions and Auxiliary Units in Moesia Superior in AD 93

The other recently arrived cohorts include the I Montanorum, the II Flavia Commagenorum, the IIII Raetorum, the VI Thracum, and the VII Breucorum civium Romanorum. Based on the nomenclature and the date of the diploma, the cohort of Commageni was likely raised by Vespasian. It may have come west with Vespasian as he marched from the east in the wake of the turmoil around AD 69-70, and consequently remained in the province. 396 The name of the cohors VII Breucorum civium Romanorum also provides us with some clues about its origins. The title civium Romanorum is usually bestowed on a unit for meritorious service. There is a cohors VII Breucorum mentioned in the Pannonian diploma dated to September 5, AD 85. 397 For that reason, the regiment was probably transferred to Moesia Superior in light of the Dacian incursion of that year, participated in the Roman counterattacks, and fought well enough to merit the honorary epithet civium Romanorum.

Trajan The next constitution from Moesia Superior is dated to May 8, AD 100, survives in a number of copies, 400 and lists three cavalry wings: the ala I praetoria, the ala I Claudia nova, and the ala II Pannoniorum. All three units were already stationed in Upper Moesia. While there is a paucity of cavalry units, the same cannot be said about the number of infantry cohorts. There are twenty-one cohorts listed, six of which were not previously recorded in the province: the cohors I Flavia Bessorum, the cohors I Thracum civium Romanorum, the cohors I Vindelicorum milliaria civium Romanorum, the cohors II Hispanorum, the cohors II Brittonum milliaria civium Romanorum pia fidelis, and the cohors III Brittonum. The cohors I Flavia Bessorum may have been in Moesia Inferior as early as AD 97, before being transferred to Moesia Superior. 401 Regardless, it was in a Moesian province by the end of the century.

The cohors I Montanorum was in Pannonia prior to its arrival in Moesia Superior. In CIL XVI.31, a diploma from Pannonia, and dated to September 5, AD 85, there are two cohortes I Montanorum. On the one hand, one of those two cohorts remained in Pannonia, and is listed in diplomas dated from AD 102 and later. 398 On the other hand, the other cohors I Montanorum moved to Moesia Superior and first appears in this diploma. On September 3, AD 84, and September 5, AD 85, the cohors VI Thracum was based in Pannonia. 399 This Thracian unit is first recorded in Moesia Superior in AD 96.

The cohors I Thracum civium Romanorum named in this diploma presents us with a whole host of problems. First, there was a cohors I Thracum based out of Upper Germany on October 27, AD 90. 402 Wagner and Beneš both believe that the Thracian cohort based out of Germania Superior is to be equated with the Thracian cohort from this diploma from Moesia Superior. 403 Rather conveniently, we do not have any record of the cohors I Thracum’s presence in Germany at any point between AD 90 and AD 116. What is more, when the unit does “resurface” in Upper Germany on September 8, AD 116, it has the full title cohors I Thracum civium Romanorum. Trouble arises, however, when we look at two separate diplomas not available to Wagner or Benes: one dated to February 17, AD 110 and from Dacia; the other dated to July 2, AD 110 and from

393

RMD I.6. The cohors IIII Raetorum was transferred Cappadocia a few decades later, however, along with, it seems, the cohors I Raetorum that had been based in Moesia Inferior. Both cohorts are mentioned at the beginning of Arrian’s Expedition Against the Alans (Arr. Ekt. 1). 395 RMD I.2. 396 Spaul 2000: 401. 397 CIL XVI.31. 398 CIL XVI.47 dated to November 19, AD 102 through to RMD III.181 dated to AD 166/168. 399 CIL XVI.30 and CIL XVI.31. 394

400

CIL XVI.46; Eck and Pangerl 2008: 326-329, 340-347; Eck and Pangerl 2009: 562-566. 401 Weiss 1997: 234. 402 CIL XVI.36. 403 Wagner 1938: 198; Beneš 1978: 52.

28

The Lower Danube from AD 81 to AD 161 Pannonia Inferior. 404 One solution would be to suggest that the unit was operating in Dacia during the Dacian wars, where it remained following the inception of the province through the first quarter of AD 110, only to return west to Upper Germany, with a brief layover in Pannonia Inferior later in AD 110. 405 Indeed, we have another diploma from Germania Superior dated to AD 129/130 that lists a cohors I Thracum civium Romanorum. 406

observations we can conjecture that this unit of Thracians was created during the Flavian period, was immediately based in what later became Moesia Superior, and that it was rewarded for its efforts in the Dacian wars of Domitian. This cohors I Thracum civium Romanorum remained in Moesia Superior after the war, then moved to Dacia following Trajan’s Dacian war, before finally being transferred to the Pannonias, where it remained.

The second problem is as follows. After its posting in Upper Germany, this Thracian unit returned to the middle Danube where it was recorded in a diploma dated to May 19, AD 135 from Pannonia Inferior; 407 however, this same diploma lists two cohortes I Thracum civium Romanorum. So, by AD 135 there were two cohortes I Thracum civium Romanorum operating in the middle Danube region. How do we disentangle this mess? Spaul’s solution is to suggest two units: one of the units, which was first attested in the Upper Moesian diploma of AD 100, was active in Dacia in AD 110 before being transferred to Pannonia Inferior in AD 110, and flip-flopping back and forth between Pannonia Inferior and Superior – which may reflect the flexible nature of provincial boundaries more than a busy unit – for the rest of its existence; 408 the second unit was in Germany - and then Upper Germany - from its inception through AD 129/130 until being transferred to Pannonia Inferior, where it spent the rest of its days. 409

Let us turn to the remaining cohorts, beginning with the cohors I Vindelicorum milliaria civium Romanorum. Most scholars suggest that this unit started out in Germany on the basis of a funerary inscription from Köln which names the unit. 414 Wagner goes further, suggesting that the unit was based in Pannonia in the interim; 415 however, his argument is suspect, particularly in light of the discovery of a diploma from Lower Germany. 416 That diploma, dated to February 20, AD 98, lists a cohors I Vindelicorum civium Romanorum milliaria. This new piece of evidence confirms the theories of Beneš and Spaul, and modifies the theory of Wagner, thereby placing the transfer of the unit sometime between AD 98 and AD 100. The cohors II Hispanorum was stationed in Pannonia as late as September 5, AD 85. 417 It was probably transferred to Moesia Superior later that year or perhaps in the following year. It is first recorded in the province in a fragmentary diploma dated to August 14, AD 99, in which the recipient served in the cohors II Hispanorum. 418

A problem remains: we are still left with the issue of trying to make sense of the title civium Romanorum from the Moesian diploma for a Thracian unit that does not seem to have been in Moesia Superior prior to AD 100. 410 The majority of the citizenship grants under Trajan were granted after the Dacian wars. 411 So it could be that the unit earned the epithet before Trajan, and more specifically during the Flavian dynasty. There were not many opportunities for a unit to earn praise for its performance in combat, realistically the only avenue that a unit had to distinguish itself. So, we must turn to the wars on the Danube. Knight has singled out Moesia as a region of the empire which experienced a massive build-up of troops; and he argued that the chief architect of this was Domitian. 412 Indeed, our discussion so far has supported Knight’s assertions. In addition, he pointed out the regional nature of most auxiliary transfers: troops in Germany would most often be sent to Britain whereas troops in Pannonia would most often be sent to Moesia, and then from Moesia to Dacia. 413 And so, based on these

The final two cohorts both have British origins. In regard to the first of those two units, the cohors II Brittonum milliaria civium Romanorum pia fidelis, there is a cohors II Britonnum listed on a diploma from Germania Inferior dated to February 28, AD 98. 419 This German-based unit is probably to be equated with the Moesian unit. 420 It was transferred sometime between AD 98 and AD 100. Its stay in Moesia Superior was brief, as it later shows up in Dacia. 421 The cohors III Brittonum seems to have been a relatively new creation. It is not documented in any other province prior to AD 100; what is more, it seems to have spent all of its time in Upper Moesia. Thus, the unit probably arrived in Moesia Superior not long after it was recruited.

414

Wagner 1938: 196-197; Beneš 1978: 54; Spaul 2000: 289; CIL XIII.8320. 415 Wagner 1938: 196-197. 416 RMD IV.216. 417 CIL XVI.31. 418 RMD I.7. 419 RMD IV.216. 420 This was the conjecture of Wagner (1938: 110), Strobel (1984: 125) and Spaul (2000: 198), none of whom had access to this diploma. Both Wagner and Spaul based their arguments on inscriptional evidence which document the unit’s presence and assistance in building construction in Lower Germany. 421 It is listed on a Dacian diploma dated to October 14, AD 109 (RMD III.148).

404

CIL XVI.57 (Dacia); CIL XVI.164 (Pannonia Inferior). Though see the comments of Matei-Popescu (2015: 410-411, 413). 406 RMD II.90. 407 RMD IV.251. 408 Spaul 2000: 361. 409 Spaul 2000: 364. 410 Spaul 2000: 362. 411 Holder 1980: 34. 412 Knight 1991: 207. 413 Knight 1991: 207. 405

29

Exercitus Moesiae

Legions (2)

This constitution also marks the debut of the cohors III campestris civium Romanorum in the history of Moesia Superior. It is not attested on any earlier documents. Spaul suggests that the fourth Cyprian cohort of Roman citizens was raised in a levy made by Claudius, and that it was based in the Chersonesus and served in support of the Bosporan kingdom – more on this in chapter four below. 427 It seems to have been in Moesia Superior only briefly, because it turns up in a Dacian diploma dated to February 17, AD 110, along with the It is aforementioned cohors III campestris. 428 unattested after that point, though it may have returned to the Crimea. 429 The Eighth Raetian cohort of Roman citizens was based in Pannonia Inferior on September 5, AD 85. 430 With no record of its presence in Pannonia Inferior after that date, it was likely transferred to Upper Moesia in AD 86, that is, in the wake of Domitian’s Dacian wars. Like the cohors IIII Cypria civium Romanorum, this unit was later part of the Dacian occupying force. 431

Auxiliary Units (24)

IIII Flavia, VII Claudia

ala I Claudia nova, ala II Pannoniorum, ala I praetoria, cohors I Antiochensium, cohors VII Breucorum c R, cohors II Brittonum milliaria c R p f, cohors III Brittonum, cohors I Cilicum milliaria eq, cohors I Cisipadensium, cohors I Cretum, cohors I Flavia Bessorum, cohors II Flavia Commagenorum, cohors I Flavia Hispanorum, cohors II Gallorum Macedonica, cohors V Gallorum, cohors II Hispanorum, cohors V Hispanorum, cohors I Lusitanorum, cohors I Montanorum c R, cohors IIII Raetorum, cohors I Thracum c R, cohors I Thracum Syriaca eq, cohors VI Thracum, cohors I Vindelicorum milliaria c R Table 2.2. Dispositions of Legions and Auxiliary Units in Moesia Superior in AD 100

There is a constitution, dated to AD 112, which lists at least one ala and seven cohortes, and which survives in a fragmentary diploma. 432 The listed ala, the I Hispanorum Campagonum civium Romanorum, would seem to make its debut in Moesia Superior here, for there is no other record of Campagones elsewhere in the Moesias, though we do find them later in Dacia, and Dacia Superior in particular. 433 Matei-Popescu and Tentea have argued that the unit was transferred to Dacia in the wake of Trajan’s Dacian wars. 434 There is another fragment, possibly of the same constitution and so that dates to the same year, which might also point to this ala’s presence in Moesia Superior; however, the fragments are meagre. 435 The unit came from Britian, and before its move to Dacia it passed through Pannonia Inferior. 436 Shifting to the cohorts, most of them were already based in the province, though this document adds to our evidence for the cohors I Antiochensium, which we know was in the province from AD 94, 437 but which, aside from this diploma, has left few other traces of its presence before AD 132/133. 438

We have a constitution that dates to one year later, May 16, AD 101, and it lists 1 ala and 12 cohortes. 422 All of those units are previously attested. The next substantial constitution, however, dated to AD 103/105 and from Moesia Superior, lists five new units. 423 All of these units are cohorts, and they are the cohors I Brittonum milliaria, the cohors Britannica milliaria civium Romanorum, the cohors III campestris civium Romanorum, the cohors IIII Cypria civium Romanorum, and the cohors VIII Raetorum civium Romanorum. The cohors I Brittonum milliaria was in Pannonia on September 5, AD 85, and was likely transferred to Moesia Superior in the context of the Dacian incursions. 424 The cohors Britannica milliaria civium Romanorum first appears in Moesia Superior in this diploma. Prior to its arrival in the province, it had been based out of Pannonia, where it is recorded on September 5, AD 85. 425 This unit moved on to Dacia following Trajan’s war, and became part of the initial occupying force of the new province. 426 The cohort was probably among the many auxiliary units transferred to Moesia in the wake of Domitian’s Dacian troubles.

The one cohort that stands out is the cohors III Augusta Nerviana Brittonum, which has left very very little evidence of its existence, at least in the diplomata, and especially in the form that Eck and Pangerl have 427

Spaul 2000: 389. CIL XVI.57. 429 Wagner 1938: 127-128; Spaul 2000: 389. 430 CIL XVI.31. 431 It is found in two Dacian diplomas: one, RMD III.148 dated to October 14, AD 109; the other, CIL XVI.57, dated to February 17, AD 110. 432 Eck and Pangerl 2008: 355-360. 433 RMD V.384; CIL XVI.107, 108. It might even have spent time in Pannonia Inferior. See Spaul 1994: 74; Eck and Pangerl 2008: 356. 434 Matei-Popescu and Tentea 2006: 132. 435 Eck and Pangerl 2008: 360. 436 RMD III.153. See Matei-Popescu 2015: 408-409. 437 CIL XVI.39, RMD V.335, Weiss 2008: 279-280. 438 Matei-Popescu and Tentea 2006: 134. 428

422

Eck and Pangerl 2008: 329-337. CIL XVI.54, RGZM 13. 424 CIL XVI.31; Cf. Wagner 1938: 106; Beneš 1978: 20; Strobel 1984: 124; Spaul 2000: 195-197. 425 CIL XVI.31. 426 Cf. CIL XVI.160 dated to August 11, AD 106. 423

30

The Lower Danube from AD 81 to AD 161 reconstructed here. 439 We have a diploma and constitution from 109 where the following has been reconstructed: “III AUG(USTA) NERVIANA] BRITTONU[M”, but some of the relevant pieces are If Eck and Pangerl’s admittedly missing. 440 reconstruction is right, that might make the diploma from 109 the earliest record of the unit in Moesia Superior. There is a cohors III Brittonum listed on a diploma from almost a decade earlier, which also happens to have been its first appearance. 441 If it had been a new creation at that time, as we posited above, and if it is to be equated with this unit, that would make some sense of the “NERVIANA” (it was created during the reign of Nerva), though we might have expected that name to appear on the earlier constitution – and it does not account for the name “AUGUSTA”. Again, if we assume it was the same unit, then it would seem that it stayed in Moesia Superior for some time after this, though thereafter it is listed as the cohors III Brittonum veterana – and the “VETERANA” epithet might somehow be explained in terms of its service in Parthia. 442 Eck and Pangerl, on the other hand, argue that this unit is not to be equated with the aforementioned cohors III Brittonum; rather, they argue that this is the first appearance of this unit, 443 which also shows up in a diploma that details aspects of Trajan’s troop transfers to the east for his Parthian wars. 444 Since it does not appear again, at least in this form, cohors III Augusta Nerviana Brittonum, Eck and Pangerl suggest that the unit might have been destroyed in the war. 445 Given that we already have evidence for other additional British cohorts that are similarly named, 446 the cohors I and II Augusta Nerviana, it could well be that a third unit was recruited at the same time as those other units. It is worth pointing out that one of the surviving diplomas that records the cohors III Brittonum in AD 105 is nearly complete – and both the inner and outer faces for that matter: there is no space where the abbreviations for Augusta and Nerviana could be slotted in. 447 Ultimately, then, Eck and Pangerl are probably right that this particular third cohort of Britons, the cohors III Augusta Nerviana Brittonum, is not to be equated with the earlier unit, the cohors III Brittonum (later veterana).

Parthian expedition. 448 It is no longer recorded in Moesias after this, for it seems to have stayed in east afterwards, if Arrian’s apparent reference to unit in his Expedition Against the Alans is to believed. 449

the the the be

Hadrian A constitution, which dates to September 9, AD 132, 450 also happens to be of one of the last documents from the period under review from which we can deduce any changes in the auxiliary garrison of the province before the reign of Marcus Aurelius. 451 The auxiliae worthy of note are all cohortes. There are three cohorts listed in this diploma that had been in Moesia Superior, were transferred to Dacia during and after Trajan’s second Dacian war, and then returned to Moesia Superior. The first of these is the cohors I Cretum sagittariorum. When it was last recorded in Moesia Superior, it was called the cohors I Cretum. 452 The difference in title could be because the unit was now comprised of archers, which it had not been before. On the other hand, we do not find the epithet sagittaria or sagittariorum on Moesian diplomas before the second century, and so this could simply reflect a change in epigraphic habit. 453 This unit participated in Trajan’s Dacian wars and then stayed when the new province first emerged, which is when it was first listed as the cohors I Cretum sagittariorum. 454 It is recorded in Dacia as late as May of AD 114 with the same titles. 455 Thus, this Cretan unit returned to Moesia Superior during the reign of Hadrian, though at what point we cannot say. The other two units are the cohors I Montanorum and the cohors III Campestris. These two units were last attested in Moesia Superior in AD 103/105. 456 Both are then listed on a Dacian diploma dating to October 14, AD 109. 457 The cohors I Montanorum was based in Dacia at least as late as the second or third of May in AD 114; 458 the cohors III Campestris as late as the third or fourth of May in AD 114. 459 It is conceivable that these two units, the cohors I Montanorum and the cohors III Campestris, were transferred back to Moesia Superior around the same time as the cohors I Cretum, perhaps early in the reign of Hadrian.

Before we move on to Hadrian, there is one last unit to note. The cohors IIII Raetorum, noted above, had appeared in Moesia Superior in the build-up to Trajan’s Dacian wars. We later find it on diplomas dated to the summer of AD 115, when it participated on Trajan’s

448

Eck and Pangerl 2009: 363-370; Eck and Pangerl 2015: 229-230. Arr. Ekt. 1. 450 RMD IV.247. 451 The next diploma, CIL XVI.111 which dates to AD 159/160, lists the same units. 452 CIL XVI.46. 453 Note, for instance, RGZM 10, which is the earliest account we have of a diploma for a Moesian province (here Moesia Inferior) that lists a unit of archers and which includes an appropriate epithet (here for the cohors I Tyriorum sagittaria). 454 Cf. CIL XVI.163, which dates to July 2, AD 110. 455 RMD IV.226. 456 CIL XVI.54. 457 RMD III.148. 458 RMD IV.225. 459 RMD IV.226. 449

439

Eck and Pangerl 2008: 356-357. Eck and Pangerl 2015: 226-229. 441 CIL XVI.46; Eck and Pangerl 2008: 326-329, 340-347; Eck and Pangerl 2009: 562-566. 442 Cf. RGZM 37 and RMD V.418, for instance, where we find a form of cohors III Brittonum veterana. 443 Eck and Pangerl 2008: 358. Cf. Matei-Popescu and Tentea 2006: 140. 444 Eck and Pangerl 2008: 370. 445 Eck and Pangerl 2008: 358. 446 RGZM 10. 447 Eck and Pangerl 2008: 326-329. 440

31

Exercitus Moesiae Gallorum Macedonica and the cohors II Gallorum Pannonica. 468 The cohors II Gallorum Pannonica was still in Dacia Superior in AD 138, 469 and we find it on another diploma dated to February 23, AD 144, 470 which is the last known date of the unit in the province. It moved to Moesia Superior at some point after that, and before AD 157. As for the cohors II Gallorum Macedonica, it had returned to Moesia Superior by September 9, AD 132. 471

Legions (2) IIII Flavia, VII Claudia

Auxiliary Units (12) ala I Claudia nova miscellanea, ala I Gallorum Flaviana, cohors I Antiochensium, cohors III Brittonum, cohors III Campestris, cohors I Cretum sagittariorum, cohors II Gallorum Macedonica, cohors V Gallorum, cohors V Hispanorum, cohors I Lusitanorum, cohors I Montanorum, cohors I Pannoniorum Table 2.3. Disposition of Legions and Auxiliary Units in Moesia Superior in AD 138

AUXILIAE: MOESIA INFERIOR Domitian We shift now to Lower Moesia, and the first significant constitution for Moesia Inferior is dated to May 21, AD 92, and it lists seven alae. 472 Of those seven units the II Claudia Gallorum seems out of place for two reasons. The first problem is that there is no independent record of any second wing of Gauls anywhere and at any time in the Balkans. There was, however, an ala II Gallorum in Spain in the first century before it was transferred to Cappadocia. 473 Cichorius and Spaul suggested that the unit may have stopped in Moesia, perhaps for Domitian’s Dacian war, on its way east. 474 This conjecture becomes a little more plausible when we realise that this unit does not show up in any diplomas for the Moesias, or anywhere else for that matter, before or after AD 92. It is conceivable that the unit was in the region solely for the Dacian wars before heading, or continuing, east.

Antoninus Pius The last constitutions worth discussing date to January 20, AD 151, 460 and April 23, AD 157, 461 and both survive in a few copies. Two units deserve further discussion: the cohors II Gallorum Macedonica and the cohors II Gallorum Pannonica. We discussed the introduction of the cohors II Gallorum Macedonica to Moesia above, and we argue in the next chapter that the unit remained in Moesia Superior well into the third century. The difficulty is that for some these two names, which likely denote two distinct units, should be understood as one unit, variously described. Thus, the cohors II Gallorum Macedonica that we find on RGZM 31 is simply another form of the cohors II Gallorum Pannonica we find on RGZM 37. Holder is of the opinion that the former, with Macedonica, is the correct form, while Pannonica, based on the current amount of evidence, is inexplicable, and is possibly due to lapicidal error. 462 It seems, rather, that we in fact have two distinct units, and the evidence is conclusive. If we look at the cohors II Gallorum Macedonica for the moment, the unit had been transferred to Dacia in the wake of Trajan’s Dacian wars. 463 The unit was in the later province of Dacia Superior, 464 before getting transferred to Dacia Porolissensis and, 465 eventually, sent back to Moesia Superior, which is the transfer referred to in RGZM 31. 466 RGZM 37 marks the introduction of the cohors II Gallorum Pannonica to Moesia Superior, after a transfer from Dacia. 467 In fact, we have a diploma, which dates to November 24, AD 124 and concerns Dacia Superior, that lists both units: the cohors II

The second problem centres on the name Claudia; this rules out the possibility that the ala II Claudia Gallorum is a previously unknown unit that may have been raised specifically for the Dacian wars of Domitian. Had Domitian raised a new unit of Gauls, it is unlikely that he would have called it the II Claudia Gallorum and not the II Flavia Gallorum. The inscription which places an ala II Gallorum in Spain, reads “…ALAE II GALLOR…”. 475 Claudia is not part of the unit’s name in this inscription. The same is true for an inscription naming the unit from Caesarea. 476 It is not unusual to find units honoured for meritorious service with imperial cognomina or some other title. As a result, and with these problems in mind, there is one further possible identity for this unit worth considering. In a diploma from Moesia Inferior, dated to May 13, AD 105, we find an ala I Claudia Gallorum. 477 It is also mentioned in a diploma, from 468

Eck and Pangerl 2010: 248; AE 2010, 1857. RMD V.384. 470 CIL XVI.90. 471 RMD IV.247. 472 Petolescu and Popescu 2004: 269-273. 473 Cf. Cichorius 1893: 1246-1247; Le Roux 1982: 86; Spaul 1994: 130. 474 Arr. Ekt. 2, 9; CIL IX.3610. 475 CIL IX.3610. 476 AE 1925, 44. 477 CIL XVI.50.

460

469

RGZM 31; Eck and Pangerl 2008: 372-376. 461 RGZM 37; RMD V.418, 419 (?); Eck and Pangerl 2008: 383-386; Eck, MacDonald, and Pangerl 2008; and Weiss 2008: 309-312. 462 Holder 2006b: 840, n. 3; 841, n. 2. 463 Matei-Popescu 2006: 127, 137. 464 RMD III.148; CIL XVI.57, 163; RMD IV.225. 465 RGZM 22. 466 Spaul 2000: 160. 467 RMD V.419 seems to be a copy of the constitution recorded in RGZM 37.

32

The Lower Danube from AD 81 to AD 161 the same province, dated to September 25, AD 111. 478 It might be that the lapicide simply made a mistake in writing “II” and not “I”. 479 This would be the simplest solution, and without further proof for the existence of the II Gallorum, or another II Gallorum in Moesia for that matter, the most plausible. 480 Now, if we assume that the diploma should read ala I Claudia Gallorum, then perhaps we might conjecture that this title, ala I Claudia Gallorum, is what should be restored in the intus of tabella I of RMD IV.209 from Moesia, where we find “]…ET I GAL[…”. 481 Thus, the unit listed in the AD 92 diploma may really be the I Claudia Gallorum, and, therefore, may have been part of the garrison of Moesia since the reign of Vespasian, if not earlier. 482

92, the ala Gallorum Flaviana, and the ala I Vespasiana Dardanorum had both been in Moesia previously, 486 whereas the I Flavia Gaetulorum, the I Pannoniorum, and the Hispanorum all came to Moesia during the Flavian dynasty. 487 Moving on to the infantry cohorts listed on the AD 92 diploma, there are fifteen listed. Of those fifteen cohorts, there are seven which are previously unattested in the Moesias: the I Lusitanorum Cyrenaica, the I Flavia Commagenorum, the [II Chal]cidenorum, the II Bra[caraug]ustanorum, the II Flavia Bessorum, the [II] Gallorum, and the Ubiorum. One of those seven, the cohors I Lusitanorum Cyrenaica may be the exception, though we argued above that the cohors Ubiorum was likely already in the province, even if direct evidence is lacking. 488 In RMD I.2 from Moesia, dated to April 28, AD 75, a cohors I Lusitanorum is mentioned without the name Cyrenaica. If these two units were the same we are left to wonder under what circumstances the unit would have left the province for Cyrene – the most plausible explanation for the appearance of the name Cyrenaica before returning in AD 92. During that span of seventeen years Moesia experienced a major war with the Dacians. Because of its obvious importance for the war effort, this means that we can probably rule out a mistake on the part of the lapicide: the cohors I Lusitanorum Cyrenaica was either a different unit which had recently entered the province, or did not have any soldiers eligible for discharge any earlier than AD 92. In three later diplomas dated to AD 99 489 and AD 105, 490 the cohors I Lusitanorum Cyrenaica is still in the province. Yet, a cohors I Lusitanorum shows up in Pannonia on September 3, AD 84. 491 This same unit, however, is also listed for a diploma from Illyricum dated to July 2, AD 60. 492 And so, with no obvious reason for the cohors I Lusitanorum to have left Illyricum in the mid seventies, only to return to Pannonia by AD 84, we must entertain the possibility that there were two cohorts of Lusitanians in Moesia; hence the Cyrenaica. The fact that on Aug 20, AD 127 there was a cohors I Lusitanorum listed on a Moesian diploma, and, only ten years later a cohors I Lusitanorum Cyrenaica listed in Moesia Inferior, confirms this. 493

The other six cavalry wings are less problematic. Above we followed the suggestion of Spaul in placing an ala Atectorigiana in Moesia from the Julio-Claudian dynasty on. This diploma is, however, the first such attestation of the unit in the province. Prior to its publication, the earliest record of an ala Atectorigiana was September 18, AD 97, and it was listed without the title Gallorum that appears in this diploma. 483 Thus, on the one hand, it is quite possible that the “]…ET I GAL[…” from RMD IV.209 dated to AD 75 or AD 79, might be referring to the ala I Gallorum Atectorigiana. On the other hand, the earlier inscriptions are more likely to leave off the unit number; thus, we must also entertain the possibility that in this case the diploma is referring to some other wing of Gauls. In fact, the subsequent mention of the ala Atectorigiana is from a diploma, from Moesia Inferior, dated to August 20, AD 127, and there it is called the ala Gallorum Atectorigiana. 484 It is not until around AD 157 that the unit is found with the title ala I Gallorum Atectorigiana; 485 moreover, the outer face of tabella I from RMD IV.241 is nearly complete, particularly for the unit lists. Despite all of these discrepancies in the names of the unit, and the problem of the official status of the diplomas themselves, there is no reason to suppose that we are dealing with multiple units. Of the remaining alae from the diploma dated to May 21, AD 478

RMD IV.222. For some further examples of possible errors on the part of the lapicide as regards the number of the units in diplomas see, for example, RMD I.47 (= CIL XVI.110), RMD I.53, RMD III.185, RMD IV.239, RMD IV.278, and RMD IV.287. We should keep in mind, however, that in some of these examples the incorrect number is recorded only on the intus of the tabella; in some cases all we have is the fragment from the intus that names the unit and we do not have the number from the extrinsecus. 480 I have not seen the diploma myself and thus if the “II” is a reconstruction, I cannot say whether it is a mistake on the part of Petolescu or Popescu or the lapicide. The unit is, nevertheless, listed on the outer face of the tablet, which supports the accuracy of “II” (Spaul 1994: 11; Spaul 2000: 8). Thus, the possibility remains that this is a previously unknown auxiliary unit. 481 This is the suggestion of Roxan and Holder (2003: 407, n. 4). 482 Matei-Popescu 2010a: 181. 483 MacDonald and Mihaylovich 2002: 225–228. However, we should keep in mind that with this diploma the text is from the inner face. 484 RMD IV.241. 485 RMD I.50. 479

486

RMD IV.209. For the ala I Flavia Gaetulorum, see: Cichorius 1893: 1243; Wagner 1938: 38; Strobel 1984: 111; Spaul 1994: 124-125; and Matei-Popescu 2010a: 172-178. For the ala I Pannoniorum see Wagner 1938: 58; Kraft 1951: 155; Strobel 1984: 113; Spaul 1994: 167-17; and Matei-Popescu 2010a: 191-192. For the ala Hispanorum see Wagner 1938: 44; Kraft 1951: 150; Beneš 1978: 10; Strobel 1984: 111-112; Spaul 1994: 145; Holder 2005: 79; and Matei-Popescu 2010a: 186-188. 488 The date of this diploma means that the unit was raised some twenty or so years earlier. 489 CIL XVI.44 and XVI.45. 490 CIL XVI.50. 491 CIL XVI.30. 492 CIL XVI.4. 493 RMD IV.241; CIL XVI.83. 487

33

Exercitus Moesiae Of the remaining six cohorts, at least two were Flavian creations: the I Flavia Commagenorum and the II Flavia Bessorum. The cohors II Chalcidenorum first shows up in Moesia in AD 75. It too remained in One Moesia Inferior for quite some time. 494 surprisingly problematic cohort is the cohors II Gallorum, a point recognised by Matei-Popescu. 495 The cohors II Gallorum might have arrived in the Moesias as early as Vespasian, 496 though it was certainly in Moesia Inferior by AD 92, the date of the diploma under consideration here. According to Spaul, however, by November 24, AD 107, the unit had moved to Mauretania Caesariensis. 497 It then returned to Moesia Inferior sometime before AD 114, prior to setting out for Britain most likely between AD 119 and June 17, AD 122. 498 It seems likely, however, that those cohortes II Gallorum listed in Mauretania and Britain were different from the second cohort of Gauls found here in Moesia Inferior. 499 It seems to have been transferred to Dacia, perhaps in the wake of Hadrian’s reorganisation of the province. 500 Indeed, for all we know it might be one of the cohortes II Gallorum later recorded in Moesia Superior, though this is speculation. The cohors Ubiorum arrived in Moesia sometime after AD 70 – Tacitus records its presence in Germany during the revolt of Civilis – and perhaps in AD 86 as Beneš suggests. 501

or Trajan to dispatch the unit to the Danube to help counter the Dacian incursions. As noted, this unit was “back” in Thrace by AD 114. It is not, however, listed in a subsequent Thracian constitution that is dated to Oct. 10, AD 138. 506 Indeed, it is recorded in Moesia Inferior in a diploma dated to AD 136. 507 The cohort had moved back to Moesia Inferior by this point; on the other hand, it could also mean that the unit did not have any soldiers eligible for discharge at this time. With that said, its location, roughly halfway between the Aegean Sea and Danube River, suggests the boundary between Moesia Inferior and Thrace shifted southwards in the second century around 136. 508 Legions (2)

Auxiliary Units (23?) I Italica, V ala I Asturum, ala I Macedonica Claudia Gallorum, ala Dardanorum, ala I Flavia Gaetulorum, ala Gallorum Flaviana, ala Gallorum Atectorigiana, ala Hispanorum, ala I Pannoniorum, ala Vespasiana, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq, cohors II Chalcidenorum, cohors II Flavia Bessorum, cohors I Flavia Commagenorum eq, cohors II Gallorum, cohors IIII Gallorum, cohors VII Gallorum eq, cohors II Lucensium eq, cohors I Lusitanorum Cyrenaica, cohors Raetorum, cohors I Sugambrorum tironum, cohors I Sugambrorum veterana, cohors Ubiorum eq Table 2.4. Dispositions of Legions and Auxiliary Units in Moesia Inferior in AD 93

At last we come to the cohors II Bracaraugustanorum. Paunov and Roxan suggest that the unit may have moved to Thrace, but was recruited from Spain, following the creation of the province during the reign of Claudius in AD 46, along with its namesake, the The unit was cohors I Bracaraugustanorum. 502 recorded in Thrace on June 9, AD 114. 503 Since the Thracian diploma of AD 114 is the earliest diploma that we have for that province, that would not be an unreasonable suggestion. On the basis of a bilingual inscription found in modern Shipka in northern Thrace, the unit was possibly guarding a pass through the Haemus Mountains. 504 Prior to the publication of this Moesian diploma dated to AD 92, this unit was not attested in Moesia Inferior, or Superior for that matter, anytime before AD 136. 505 It does seem, however, based on this diploma, that it was active in Moesia at least as early as the Dacian wars of Domitian, if not also the wars of Trajan. With a posting so close to Moesia Inferior, it would have been easy for Domitian

Nerva The earliest of our next set of constitutions from Moesia Inferior is dated to September 9, AD 97. 509 On

494 Cheesman (1914: 156) placed it in Moesia Inferior through AD 161. It is, for example, listed in a diploma dated to AD 157 (RMD I.50). 495 Matei-Popescu 2010a: 208. See too Cichorius 1900: 288. 496 Strobel 1984: 130. 497 Spaul 2000: 157-158. See CIL XVI.56. 498 CIL XVI.69. 499 Matei-Popescu 2010a: 208. See CIL XVI.58. 500 See RMD V.376; RMD V.380; RMD I.39. 501 Tac. Hist. 4.18, 77; Beneš 1978: 54; Strobel 1984: 145. 502 Paunov and Roxan 1997: 275; Spaul 2000: 91. 503 RMD IV.227 (= RMD I.14). 504 AE 1965, 347; Paunov and Roxan 1997: 275. Indeed, the Shipka Pass was the site of a number of engagements in the Russo-Turkish war of 1877-1878. 505 Schindel 2010.

506 RMD IV.260/V.385; cf. Paunov and Roxan 1997: 269-279; Roxan and Weiss 1998: 371-420. 507 Schindel 2010; cf. Matei-Popescu 2010a: 196. 508 Gerov 1979: 218-219; Velkov 1990: 250. On the basis of two inscriptions from Sipka in Thrace, Matei-Popescu (2010a: 196) argues that the unit was equitate (cohors II Bracaraugustanorum equitata). See IGBR III.2.1741=IDRE II.350 and IGBR III.2.1741a=AE 1965, 347. The inscriptions in question list a Celsus Marius, who was an eques of the cohort. 509 Weiss (1997: 233-238) originally dated the diploma to January 17, 97, but Holder (RMD V.338) has since revised the date of the diploma.

34

The Lower Danube from AD 81 to AD 161 it we find the following: the [I Claudi]a Gallorum, the II Ara[vacorum], the [Flavi]ana Gallorum, the Hispanorum […c.11…]a, the I Hispanorum [veterana], the [Sugambroru]m [tir]onum, the I Fla[…c.7…], the I Flavia Commagenorum, 510 the [I Fla]via Bessorum, the [II Lu]censium, the IIII Gal[lo]rum, and the Ubiorum. We can determine which of those units listed were alae, and which were cohortes, with reasonable certainty. An ala Gallorum Flaviana, and an ala Hispanorum are in the earlier Moesian diploma dating to May 21, AD 92. 511 There is also an ala I Claudia Gallorum listed in that same diploma. Even the fact that this unit comes first in the list ensures us that we are dealing with a cavalry wing. Indeed, we argued above that the diploma of AD 92 should read ala I Claudia Gallorum, which is the most probable reading here. Turning to the II Aravacorum unit, if we look at the situation in Pannonia nearly fifteen years earlier, we can track it down. In CIL XVI.30 and CIL XVI.31 we find an ala II Arvacorum (sic). 512 This unit does not show up in any further diplomas in Pannonia, and thus was probably sent to Moesia Inferior in or shortly after AD 85, and can be identified with this unit. Until recently, this unit was only found with the double ethnic name in diplomas hailing from Moesia. However, this diploma eliminates any doubt that this unit was transferred to Moesia from Pannonia and allows us to claim with confidence that the II ARA recorded in this diploma is in fact the ala II Aravacorum (et Hispanorum). 513 As Matei-Popescu suggests, the addition of et Hispanorum was probably due to the arrival of recruits from Spain. 514 Although auxiliary recruits were sometimes local in origin, for the period under review (Augustus to Severus Alexander), and so far as we can tell, recruits were regularly brought into the frontier from much further afield. 515

amount of clout over the other, newer cohorts in the province. Yet, we cannot completely rule out the possibility that there was another Spanish unit operating in the province at the same time and that this cohort took the epithet veterana to distinguish it. It most likely participated in the Dacian wars, after which it remained in the new province of Dacia, 518 before moving to the yet newer province of Dacia Superior some time before AD 119. 519 It is perhaps as a result of its faithful service under the new emperor Hadrian that it was renamed the cohors I Hispanorum pia fidelis. A second diploma from Moesia Inferior dates to September 9, AD 97, 520 and lists five alae, all of which were in the province by the date in question. 521 There are also nine cohorts listed in this new diploma. With one possible exception, none of the cohorts listed were new to the province. The problematic cohort is the cohors I Flavia ---rum, 522 and trying to identify it is no mean feat. 523 The problem is trying to find a suitable cohort that bears the name “FLAVIA” and ends with the suffix “RUM”. The two most plausible suggestions, namely the cohors I Flavia Commagenorum, and the cohors II Flavia Bessorum, are listed in this diploma, as they are in Weiss’ diploma. If we go back to the diploma of AD 92, we find seven cohorts of the fifteen listed that are not listed on this new diploma. 524 They are the cohors I Raetorum, the cohors I Bracaraugustanorum, the cohors I Lusitanorum Cyrenaica, the cohors [II Chal]cidenorum, the cohors II Bra[caraug]ustanorum, the cohors [II] Gallorum, and the cohors VII Gallorum. Based on the number of characters that Weiss believed were lost, we can eliminate the cohors I Bracaraugustanorum, the cohors I Lusitanorum Cyrenaica, the cohors [II Chal]cidenorum, and the cohors II Bra[caraug]ustanorum. 525 Yet, of those three remaining, none bear the epithet Flavia. If we look ahead two years, we find a cohors II Flavia Brittonum recorded on two identical diplomas from Moesia Inferior. 526 This cohort is not attested any earlier than August 14, AD 99. There are two problems with this possibility: the “II” before the epithet “FLAVIA”, and the “NUM” as opposed to

The remaining units listed on this diploma were certainly cohorts. Of these, all but one were already stationed in Moesia Inferior, namely the cohors I Hispanorum veterana. This unit is listed in two diplomas from Moesia Inferior, which both date to AD 99. 516 It also seems to be listed in an Illyrian diploma. 517 That diploma, dated to July 2, AD 60, lists a cohors I Hispanorum. The title veterana is not merely a distinguishing epithet; it is also a mark of seniority. Since this unit was probably active under Claudius, if not earlier, this unit may have had a certain

518

CIL XVI.57 dated to February 17, AD 110, and CIL XVI.163 dated to July 2, AD 110; Eck and Pangerl 2009: 514-519. cf. Spaul 2000: 109-111. 519 Matei-Popescu 2010a: 187. 520 Eck and Pangerl 2005: 185-192. 521 Eck and Pangerl 2005: 192. 522 In the article from which this diploma hails, Eck and Pangerl discuss three previous diplomas, namely RMD III.140 dated to AD 97, the diploma published by Weiss (1997=RMD V.338) discussed above, and the diploma published by MacDonald and Mihaylovich (2002=RMD V.337) discussed below. 523 This unit did appear in the diploma discussed above. Weiss 1997: 233-238. Cf. Eck and Pangerl 2009: 510-517. 524 Petolescu and Popescu 2004: 269-273. 525 In the article Weiss (1997: 238) gives the following reconstruction: “I Fla[via (?) - ca. 7 - ]rum”. Significantly, this line is recorded on the extrinsecus. 526 CIL XVI.44 and CIL XVI.45. These diplomas are discussed below.

510

This unit later appears in Dacia: RMD III.148 (Oct. 14, AD 109). Petolescu and Popescu 2004: 269-273. One constitution lists 4 alae and 10 cohortes (RMD V.337; AE 2005, 1705), while the other (RMD V.338; AE 2005, 1332; Eck and Pangerl 2009: 510-512) lists 5 alae and 10 cohortes. 512 CIL XVI.30 is dated to September 3, AD 84; CIL XVI.31 to September 5, AD 85. 513 The only uncertainty that remains now surrounds the discrepancy between ala II Arvacorum in the later Pannonian diplomas on the one hand, and ala II Hispanorum et Arvacorum from an earlier Illyrian diploma, and the later Moesian diplomas, on the other hand. 514 Matei-Popescu 2010a: 188. 515 Haynes 2013: 121-134. 516 CIL XVI.44 and CIL XVI.45, both dating to August 14, AD 99. 517 CIL XVI.4. 511

35

Exercitus Moesiae “RUM” for the ethnic component of the name. There are units entitled “Britannorum”, but that name seems to be too long for this fragment. There is one further possibility, and the length of the ethnic name fits well with the known number of missing characters. There was a cohors I Flavia Numidarum that Beneš and Spaul suggested arrived in Moesia Inferior in the midsecond century. 527 More recently published diplomas, however, such as those alluded to here, have shown that the unit was in Moesia by the first quarter of the second century, if not earlier. There is a cohors I Numidarum listed on a diploma from Syria dated to November 7, AD 88. 528 Roxan was of the frame of mind that there were in fact two series of Numidian cohorts: the Flavian series, and the rest. 529 There is a cohors I Flavia Numidarum recorded on a diploma, from Moesia Inferior, dated to May 13, AD 105. 530 In this case we have six characters, only one shy of Weiss’ conjecture: “NUMIDA”. This is the most likely of the available options, and so the unknown cohort is probably the cohors I Flavia Numidarum. 531 If this unit is the Numidian cohort listed in the Syrian diploma dated to November 11, AD 88, 532 in which the cognomen “Flavia” is not found, the unit might have moved to Moesia Inferior shortly thereafter. MateiPopescu, on the other hand, argues against this noting the presence of a cohors I Numidarum amongst Arrian’s forces. 533 Arrian’s account, however, is not straightforward, and though a brief excursion to Cappadocia from Moesia Inferior is not perhaps surprising, it is by no means a certainty.

transferred to Moesia Inferior. The honorary title civium Romanorum suggests that the unit distinguished itself in combat at some point during those seventeen years. Thus, the unit was probably transferred to Moesia Inferior during Domitian’s Dacian war, where it might have earned this honorary title for commendable service. Trajan We have a number of copies of two imperial constitutions dated to August 14, AD 99. 537 Of the six cavalry wings recorded, all of which were already stationed in the province, only the ala II Hispanorum et Arvacorum requires further discussion. There is a discrepancy between the name ala II Aravacorum found in Moesia Inferior two years earlier, and ala II Hispanorum et Arvacorum found in this diploma. The Arevaci were an Hispanic tribe and the earlier diplomas tell us that the unit was originally called the ala II Arvacorum (sic). 538 Usually, in regard to alae, when two ethnic names are used in the genitive, this is a result of a conflation of two different ethnic groups into one auxiliary unit, which need not be the case with the cohortes. 539 Perhaps what happened in this case is that in the years, and perhaps months, leading up to the first Dacian offensive, Trajan may have had cause to replenish any potentially depleted ranks with new recruits from their original recruiting grounds. For reasons that we may never know, there may not have been enough suitable recruits from among the Arevaci; thus, Trajan would have had recourse to collect additional troops from elsewhere in Spain, and thus we see the appearance of the “new” name. The change could not have simply been a distinguishing feature, as there was no other Arevacian unit in the province, or even on the Danube at this time, that might have caused some confusion.

The other known constitution which dates to September 9, AD 97 lists four cavalry units which were already based in the province. 534 The published diploma also claims that it will list ten cohorts; however, only the names of three can be restored. These three are the cohors I Lusitanorum Cyrenaica, the cohors I Tyriorum, and the cohors I Lepidiana civium Romanorum. As we saw above, the cohors I Lusitanorum Cyrenaica was already based in Moesia Inferior. On June 13, AD 80, the cohors I Lepidiana was posted in Pannonia. 535 This diploma marks its debut in Moesia Inferior. 536 As noted in chapter one above, the cohors I Tyriorum had been transferred to the province at least during the reign of Vespasian, if not earlier. At some point between the seventeen years separating that diploma and this one, the unit was

There are thirteen cohorts listed on some of the diplomas of August 14, AD 99. Two of these cohorts are previously unattested in Lower Moesia: the II Mattiacorum and the II Flavia Brittonum. The cohors II Mattiacorum appears for the first time in this diploma. 540 Some scholars have argued that the cohors Mattiacorum listed in an earlier diploma is to be identified with this unit; 541 for the moment, we shall 537 CIL XVI.44 and CIL XVI.45. There is a fragmentary diploma, RMD IV.217, from Moesia Inferior that is dated to AD 99 and lists the following three units: the Gallorum Flaviana?, the I Sugambrorum veterana VEL tironum, and the II Flavia Brittonum. The first unit is an ala while the other two are cohortes. There are at least three others which list only three alae and seven cohortes: RGZM 8; the fragment published by Mihailescu-Bîrliba (2008); and the more complete copy published by Eck and Pangerl (2012: 295301). Cf. AE 2006, 1862; AE 2008, 1195; Eck and Pangerl 2009: 512-514. 538 Saddington 1982: 73; Spaul 1994: 35. There is an Illyrian diploma (RMD IV.202), which dates to July 2, AD 61 that lists a cohors II Hispanorum et Aravacorum. 539 Holder 1980: 22. 540 Cf. Matei-Popescu 2010a: 222. 541 Wagner 1938: 164-165.

527 Cichorius (1900: 320), Wagner (1938: 110), Kraft (1951: 172), and Strobel (1984) fail to discuss the presence of this unit in Moesia Inferior; none of these four scholars were privy to the evidence now available. Cf. Beneš 1978: 47; and Spaul 2000: 473. 528 CIL XVI.35. 529 Roxan and Holder 2003: 433-434, n. 5. 530 RGZM 10. 531 Cf. Matei-Popescu 2010a: 225. 532 CIL XVI.35. 533 Arr. Ekt. 3-4. Matei-Popescu 2010: 225. Cf. Roxan 1997: 292293. 534 RMD V.337=MacDonald and Mihaylovich 2002: 221-228. 535 CIL XVI.26. 536 Mihailescu-Bîrliba 2008: 205; Matei-Popescu 2010a: 218.

36

The Lower Danube from AD 81 to AD 161 stick with our earlier statement and assume that this is a distinct, and new, unit. 542 As regards the cohors II Flavia Brittonum, the cognomen Flavia suggests that the unit was raised, or rewarded, either by Vespasian, or Domitian. 543 This auxiliary unit was not previously attested; therefore, it may have been created around AD 74.

evidence for some of these other units, such as the cohors II Nerviorum 550 and the cohors I Nerviorum, 551 dates to about the same time as the Moesian unit, we could well imagine that they were recruited amongst the same people as part of a single recruitment drive. With that said, when a name refers to a people rather than a place it is usually in the genitive plural, which likely rules out option three. As regards option two, the association with Nerva is itself problematic, for some of the reasons outlined by Spaul. 552 That brings us back to option one, and Holder’s identification with Gloucester. Not definitive, but the most plausible of the lot.

There are two complete diplomas, representing two different constitutions, that date to May 13, AD 105. 544 The first of the two, RGZM 10, lists the ala I Pannoniorum, the ala Hispanorum, and the ala Atectorigiana, while the second, RGZM 11, lists the ala I Flavia Gaetulorum, the ala I Asturum, and the ala II Hispanorum et Arvacorum. All six are previously attested in the province. With respect to the cohorts, there are two which make their historical debut here, and they are found on RGZM 10: the cohors I Augusta Nerviana Pacensis milliaria Brittonum and the cohors II Brittonum Augusta Nerviana Pacensis (milliaria). 545

Some doubt the very existence of the cohors I Augusta Nerviana Pacensis milliaria Brittonum. 553 On the one hand, this unit has been identified with the aforementioned cohors I Augusta Nerviana velox found, if briefly, in Mauretania Caesariensis. On the other hand, unlike the cohors II Brittonum Augusta Nerviana Pacensis (milliaria) 554 and the cohors II Flavia Brittonum, 555 Matei-Popescu has argued that the cohors I milliaria Brittonum in one diploma dated to AD 111 556 and another two dated to AD 116, 557 both from Moesia Inferior, is probably to be identified with this unit, the cohors I Augusta Nerviana Pacensis milliaria Brittonum. There is still the matter of determining its relationship to the cohors I Brittonum milliaria Ulpia torquata civium Romanorum, which is recorded in diplomas from Dacia dating to AD 106, 109, and 110 respectively. 558 The nomenclature, and the “Ulpia” in particular, would seem to support the opposite view: they are different units. In the diploma from AD 105 (RGZM 10) the full title appears on the exterior and the interior (tabella I): “I AUGUSTA NERVIANA PACENSIS ∞ BRITTONUM”. Roxan and Holder claim that it was recorded in Dacia Superior in AD 119, though the source of this information is unknown. 559 A British Ulpian cohort is also mentioned in Dacia Inferior (the province was later divided) nearly two decades later in a diploma dated to AD 129/130, 560 and another dated to July 2, AD 133. 561 By December 13, AD 140, a cohors I Augusta Nervia Pacensis Brittonum milliaria is listed on a diploma from Dacia Inferior. 562 Nearly twenty years later, however, we are back to a diploma, in this case from Dacia Porolissensis (more divisions and so another provincial garrison change) dated to Sept. 27, AD 159, with a cohors I Ulpia Brittonum milliaria

A shared problem with both cohorts is their names. Regarding the nomenclature, the two most problematic elements are the “Nerviana” and the “Pacensis”. With respect to “Pacensis”, its origins are obscure, and no one has ventured a plausible identification. The “Nerviana” too has generated debate. One of the problems is that some form of “Nerv” has appeared in the titulature of a number of auxiliary units, and there is considerable variation. Sticking with the Moesian unit, Holder suggested that the unit owed this name to the place whence its men were originally recruited: colonia Nerviana Glevensium (Gloucester). 546 MateiPopescu discussed the other auxiliary units with “Nerviana” as part of their title, including two auxiliary units with that name in Mauretania Caesariensis, the ala Nerviana Augusta fidelis milliaria and the cohors I Augusta Nerviana velox. 547 As a result of this he argued that it is much more likely that the “Nerviana” (or “Nervia”) referred to the reigning emperor at the time of their creation, namely Nerva. 548 The third option is that “Nerviana” refers to the Belgic Nervii. There are a number of Nervian auxiliary units attested in Britain, and there is good reason to believe that they were so named because of their initial recruitment among the Nervii. 549 Given too that our earliest 542

There are two earlier Moesian diplomas, both dated to AD 78, that list a cohors Mattiacorum: CIL XVI.22 and RMD IV.208. For a fuller discussion of the possibility that these two units may be one and the same unit, see above. 543 According to Holder (1980: 16), Titus did not bestow the title on any auxiliary units; given the brevity of his rule and the fairly stable conditions across the empire, this point is undoubtedly correct. 544 RGZM 10, 11. 545 Note, however, that there is a cohors II Flavia Brittonum listed on RGZM 11, which was probably a different unit (Matei-Popescu 2010a: 198). Matei-Popescu has discussed the difficulties with these units in a variety of places: Tentea and Matei-Popescu 2004: 276277; Matei-Popescu 2010a: 197-198; 2010b. 546 Holder 1980: 40-41. Cf. Matei-Popescu 2010b: 396. 547 CIL XVI.56. cf. Spaul 2000: 217. 548 Matei-Popescu 2010b: 397. 549 Note, for example, CIL XVI.69, 70; RMD III.168, 184. Cf. Spaul 2000: 218; Cuff 2010: 261, n. 61.

550

CIL XVI.43. CIL XVI.51. Spaul 2000: 218. 553 Spaul 2000: 201, n.1. 554 RGZM 10. 555 RGZM 11. 556 RMD IV.222. 557 Eck and Pangerl 2006: 99-102; Eck and Pangerl 2009: 525-530. 558 CIL XVI.160: August 11, AD 106; RMD III.148: October 14, AD 109; CIL XVI.163: July 2, AD 110. 559 Roxan and Holder 2003: 434, n. 7. 560 RMD V.376; Weiss 1997: 244. 561 RMD I.35. 562 RMD I.39. 551 552

37

Exercitus Moesiae listed. 563 Either we have a host of different units, or the name could be recorded in a number of different ways. Indeed, some of these changes could be explained away as carelessness or lapicidal error, especially given the mistakes that appear on diplomas from the second century, at least before AD 153. 564 Of those diplomas noted above, “RITT ∞” alone on the exterior of RMD V.376 is preserved, there are two forms of the name preserved on RMD I.35 (AD 133) (on the interior (tabella I) “I ULP BRITT ∞”, and the exterior (tabella I) “I ULP BRITTON ∞”); likewise two forms of the name are preserved on RMD I.37 (AD 140) (on the interior (tabella I) “I AUG PAC BRITT ∞”, and the exterior (tabella I) “I AUG NERV PAC BRITT ∞”); and the name “BRITT ∞” on RMD I.47 (AD 159), the exterior of the diploma. The names used vary, sometimes on the same diploma. The only consistent elements in the name are the “I”, the “BRITT”, and the “∞ “, though the order for the second and third elements vary considerably. If these many diplomas with the different names all refer to the same unit, it is easy enough to explain why it changes from a form with “AUGUSTA NERVIANA PACENSIS”, to one with “I BRITTONUM ∞ ULPIA TORQUATA C R”, to one with “I MILLIARIA BRITTONUM”. Such a unit could have earned the names “Ulpia Torquata Civium Romanorum” during the Dacian wars, and for one reason or another it might have made sense to strengthen its numbers (up to milliaria in size) while stationed, if briefly, in Dacia in the years immediately following the conquest. Why the “Ulpia” or even the “Nerviana Pacensis” would later return is more difficult to explain. On the other hand, with tradition being as important as it is in the Roman military, it is easy enough to imagine a scenario in which a unit might feel compelled to emphasise its storied history; moreover, given the flexibility in titulature on diplomas in the second century before AD 154, the use of a particular name really might be a question of preference, or changing fads amongst the powers that be. A modern parallel illustrates a similar series of changes: the Canadian air force was originally (1920) entitled the “Canadian Air Force”, became the “Royal Canadian Air Force” in 1924, was amalgamated with other aspects of the Canadian military in 1968 to become the “Air Command” in the “Canadian Forces”, and then more recently, in 2011, was renamed the “Royal Canadian Air Force”. Getting back to this Roman infantry cohort, it seems likely that by AD 159 the state had settled on the simpler form, “I BRITTONUM ∞ “, which is how it appears on all later diplomas. 565 In the end, what we have is an auxiliary infantry cohort raised in Britain, sent to Moesia Inferior, then moved back and forth between Moesia and Dacia over the course of the Dacian wars, especially between 105 and 116, before finally settling in Dacia, first Inferior (by AD 129), later Porolissensis (by AD 159). As it turns out, there is a cohors II

Brittonum Augusta Nerviana Pacensis (milliaria) on a number of diplomas from the Pannonias (Inferior in particular), some as early as the beginning of September of 114, 566 and so it likely moved to Pannonia Inferior before AD 114. 567 Legions (2) I Italica, V Macedonica

Auxiliary Units (25) ala I Asturum, ala I Flavia Gaetulorum, ala Gallorum Atectorigiana, ala I Gallorum Flaviana, ala Hispanorum, ala II Hispanorum et Aravacorum, ala I Pannoniorum, ala I Vespasiana Dardanorum, cohors I Bracarorum, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq, cohors II Chalcidenorum, cohors I Cilicum milliaria sagittaria eq, cohors II Flavia Brittonum, cohors I Flavia Commagenorum equitata sagittaria, cohors I Flavia Numidarum eq, cohors II Gallorum, cohors VII Gallorum eq, cohors I Hispanorum veterana eq, cohors I Lepidiana eq c R, cohors I Lusitanorum Cyrenaica, cohors II Mattiacorum milliaria eq, cohors I Sugambrorum veterana eq, cohors I Tyriorum, cohors Ubiorum eq Table 2.5. Dispositions of Legions and Auxiliary Units in Moesia Inferior in AD 100

In the years following Trajan’s Dacian wars, the number of Moesian diplomas indicating new units drops. There is a diploma dated to AD 107 that lists an otherwise unaccounted for auxiliary unit, at least in Moesia Inferior: the cohors II Flavia Numidarum. 568 The unit was still in the province a few years later, in AD 116. 569 Not long after that (July 17, AD 122), however, we find it in Dacia Inferior, as part of the new garrison of the province. 570 Given that it is not found on inscriptions from Moesia Inferior that date later than 566

CIL XVI.61; RMD II.87, III.152, IV.228. Matei-Popescu 2010b: 396. 568 Eck and Pangerl 2009: 514-519. See too Matei-Popescu 2010a: 226. 569 Eck and Pangerl 2006: 99-102. 570 RGZM 20. 567

563

CIL XVI.110. cf. RMD I.47. Cooley 2012: 174. 565 Note, for example, RMD I.63 and 64. 564

38

The Lower Danube from AD 81 to AD 161 slightly later constitution dated to 157. 580 Ultimately, the Sugambrian unit from the constitution of AD 134 might have been the veteran unit, 581 and it is entirely likely that the cohors I Sugambrorum tironum went east in the context of Trajan’s Parthian expedition. 582

116, and certainly not in any diplomas, it likely was transferred to Dacia not long after that, presumably between 118 or 119 in the wake of the reorganisation of Dacia under Hadrian. 571 The next diploma that provides evidence of change in the auxiliary garrison of Moesia Inferior is dated to September 25, AD 111. 572 The alae listed were already recorded in the province. Of the seven cohorts, one of them requires further discussion: the cohors I Claudia Sugambrorum tironum. The cohors I Claudia Sugambrorum tironum is unusual because of the imperial cognomen Claudia. On the one hand, none of the earlier diplomas which list this unit add the cognomen Claudia. 573 On the other hand, the much better documented cohors I Sugambrorum veterana, mentioned in the diploma of AD 111, also seems to have been called cohors I Claudia Sugambrorum veterana. 574 Roxan suggested that the epithets tironum and veterana indicate that both cohortes Sugambrorum bore the cognomen Claudia in their full titles. 575 This is certainly plausible if we consider that the title Claudia would have been sufficient to distinguish between the two units, had only one of the two used that title. Unfortunately, with the exception of this diploma, all other cases in which the title Claudia appears do not include the epithets tironum or veterana. 576 Thus, trying to decipher which cohort is which can be quite problematic. We saw above that this particular cohort of Sugambrians may have been in Moesia sometime before AD 75. This evidence shows that the unit remained in the province well into the second century. In this light, the unit (I Claudia Sugambrorum) mentioned on a Syrian diploma dated to AD 156/157 should probably be identified with this one, the cohors I Claudia Sugambrorum tironum. 577 We can only speculate, however, about when the unit departed for the east, with the two most probable occasions being the Parthian war of Trajan and the second Jewish war under Hadrian. It was still in Moesia Inferior in AD 116, 578 and there is a fragmentary diploma from April 2, AD 134 that lists a cohors I Sugambrorum, though we cannot be certain whether it was the veterana or tironum unit, for all that survives of relevance is “SUGAMBR”. 579 There are, however, some later diplomas that attest to the presence of the cohors I Sugambrorum veterana in the years just before the publication of the Syrian diploma. RMD V.414, a fragmentary diploma for a constitution dated to 155, lists a cohors I Sugambrorum, as does a

Hadrian One of the first constitutions from the reign of Hadrian dates to October 19, AD 120, and it provides evidence of some significant change. 583 All of the alae, of which there are five, listed on the diploma are previously attested in the province. The same cannot be said for the cohortes, however. One of the eight cohorts listed might make its debut here: the cohors I Bracarorum civium Romanorum. This unit is not to be confused with the already mentioned cohors I Bracaraugustanorum. 584 The next attestation of this unit comes five years later; it surfaces in RMD IV.235, dated to June 1, AD 125. Prior to its arrival here, the unit had been based in Mauretania Tingitana, as noted in chapter one above, where it is attested in at least two inscriptions, the earliest dated to AD 88, 585 the other to AD 103-104. 586 As for the timing of the move, that might have come as early as AD 105. There is a fragmentary diploma, RMD V.369, which lists a “I BRA”, which could be the cohors I Bracarorum, and which could date to 105 or 127; on the other hand, the listed unit might be the previously discussed cohors I Bracaraugustanorum, 587 though Matei-Popescu prefers the former reading. 588 The seemingly definitive piece to the puzzle, and the one missing from RMD V.369, is the presence of some form or abbreviation of civium Romanorum (“cR” or “CIV ROM”), which seems to have been an instrumental part of the name of the cohors I Bracarorum – and which is missing from the inscriptions that list the cohors I Bracaugustanorum. Regardless of the timing of its arrival, the unit seems to have stayed in the province for quite some time, and might be attested in Moesia Inferior on a diploma dated as late as 157. 589 It is certainly attested in the 140s. 590

580

Weiss 2008: 312, reconstructed on the basis of RMD I.50. Though note the comments of Matei-Popescu 2010a: 231. 582 Indeed, scholars like Eck and Pangerl (2008), Bennett (2010), and Matei-Popescu (2015) have noted the considerable use of Danubian units in Trajans Parthian War. 583 Eck and Pangerl 2009. 584 I made this very mistake in an earlier iteration of this book (the MA thesis on which this is based: Whately 2005: 71-72), as did Roxan and Holder (2003: 461-462) in RMD IV. Note the comments of Matei-Popescu (2010a: 195). Cf. Knight 1991: 200. 585 CIL XVI.159. 586 Papi 2004. 587 The editor (Holder 2006b: 770, n. 3) admits that it could be either. 588 Matei-Popescu 2010a: 195. 589 RMD 1.50. Again, there is some dispute as to whether we should read Bracarorum or Bracaraugustanorum, with the editor of the diploma Roxan (1978: 72) preferring the latter, and Matei-Popescu (2010a: 195) again preferring the former. 590 RMD III.165/V.399; Weiss 1999: 279; Weiss 2008: 315-316; Petolescu and Popescu 2007: 150-151. The cohors I Bracarorum in Weiss 2001 is a restoration. 581

571

Schmitz 2012. RMD IV.222. Cf. CIL XVI.58. 573 CIL XVI.22 (February 7, AD 78): I SUGAMBRORUM TIRONUM; Weiss 1997: I SUGAMBRORUM TIRONUM. 574 Note Matei-Popescu 2010a: 228-230. 575 Roxan 1978: 73, n. 2. 576 Some of these diplomas, however, are fragmentary, such as RMD III.165. 577 Cf. CIL XVI.106. This is the view of Roxan (Roxan and Holder: 434 n. 6). 578 Eck and Pangerl 2006: 100. 579 Weiss 2008: 300-301. 572

39

Exercitus Moesiae Another, slightly later, constitution from Moesia Inferior records yet more change. 591 This constitution, dated to June 1, AD 125, lists one unit which previously had not appeared on a diploma from the province. The cohors I Thracum Syriaca was last recorded in a diploma from Moesia Superior that dates to May 8, AD 100. 592 It is not recorded in any of the subsequent diplomas from Moesia Superior, including the next substantial one which dates to AD 103/107.593 Thus, we might surmise that the unit was transferred to Moesia Inferior perhaps as early as AD 103. Yet, we have a diploma from Moesia Inferior that dates to September 25, AD 111, and the cohort is not listed. 594 We seem, then, to have a twenty-five year gap in which the unit is missing in action. 595 Perhaps this is because the unit may not have had any soldiers who were eligible for discharge during that period. If that was the case, and given the length of the interval, 596 then it seems that the unit would have experienced the recruitment of an almost completely new complement of soldiers around AD 100; but this is hard to rationalise. With this scenario, we may very well suppose that the cohort experienced some sort of catastrophic defeat, perhaps in Trajan’s first Dacian war. 597 Still, this is pure speculation and so it is hoped that new evidence of the unit’s whereabouts between AD 100 and AD 125 surfaces in the future. 598

from a merger between the ala I Claudia Gallorum and the ala I Pannoniorum. 603 Matei-Popescu, however, pointed out the problems with identifying two separate alae named ala I Claudia Gallorum. 604 The next problem rises because there is an ala Pannoniorum et Gallorum based in Dacia Porolissensis, perhaps as early as September 27, AD 154. 605 In this instance it is recorded as “…GALL ET PANN…” with the two ethnic names inverted. The same abbreviations are found in CIL XVI.110, dated to September 27, AD 159. In a later diploma from Dacia Porolissensis dated to July 21, AD 164, we find the ala “…II GALLOR ET PANNO…” with the name again inverted. 606 What the epigraphic evidence for Dacia implies is that the ala I Pannoniorum et Gallorum from Moesia Inferior, in which the Gallic ethnic name is listed first, never left the province long enough to be listed in Dacia Porolissensis. In that first Moesian diploma it is recorded as follows: “…I PANN ET GALL…”. 607 For Spaul, if there was only one Gallic and Pannonian unit in a province, such as Dacia Porolissensis, then there may not have been a need to specify the unit number in any official documentation. 608 It remained in Moesia Inferior at least as late as AD 157, 609 and at some stage it gained the epithet catafracta, for an Italian inscription, dated to the reign of Antoninus Pius, refers to it as the ala I Gallorum et Pannoniorum catafracta. 610 When the unit became comprised of cataphracts we cannot say, though the second century seems to have been the time when the Roman military started to make use of heavily-armed cavalry. 611

We have now only to examine the last set of changes in Moesia Inferior. Two constitutions, one dated to AD 121, 599 another dated to August 20, AD 127, 600 list the ala I Pannoniorum et Gallorum. This unit first appears in the 120s, and two principal questions emerge: when was it raised; and when, if ever, was it transferred to Dacia? Spaul suggested that this cavalry wing may have been raised around AD 109; however, in light of this new evidence it seems that it may, instead, have been raised sometime between the end of the first Dacian war of Trajan and the beginning of the second. 601 In fact, if we assume that the diploma of AD 121 records the first recruit who was eligible for discharge, then we might conjecture that this unit was raised in preparation for the second of Trajan’s Dacian wars. 602 Gayet argued that the unit had been formed

The diploma dated to August AD 121 also lists a cohors I Germanorum. This is the first attestation of the unit, which seems to have remained in the province through the reign of Marcus Aurelius. 612 This diploma dated to AD 127 also lists the cohors II Lucensium. 613 That cohort no longer appears, however, on diplomas from Moesia Inferior that postdate this one. Moreover, a diploma from Thrace that dates to some point between September and December of AD 138 lists the Second cohort of Lucenses. 614 The cohort then appears on three subsequent diplomas from the province of Thrace. It seems that at some point between the 20th of August, AD 127, and September of AD 138, the unit moved south. Velkov has argued quite convincingly for AD 136 as the date for the

591

RMD IV.235, V.364, V.375; Weiss 2009: 241-242; Eck and Pangerl 2009: 537-538; Eck and Pangerl 2014a: 245-249. 592 CIL XVI.46. 593 CIL XVI.54. 594 RMD IV.222. 595 It is not found in any other province in the interval. Its absence could be due to a lack of diplomas from either province during that period. 596 Twenty-five years is the standard tour of duty of an auxiliary soldier. 597 This is possible, as the date for the mass recruitment roughly falls within the time frame of the war. The capture and death of Longinus might provide just such a circumstance. 598 Petrović (1995: 44) had suggested that the unit was transferred to Moesia Inferior in the middle of the second century, but RMD IV.235 proves that the unit was transferred well before that. 599 Weiss 2008: 297-300. 600 RMD IV.241. 601 Spaul 1994: 83. 602 Cf. Matei-Popescu 2010a: 185.

603

Gayet 2006: 80-82. Matei-Popescu 2010a: 185-186. 605 RMD I.47. 606 RMD I.64. 607 RMD IV.241. 608 Spaul 1994: 83. The fact that the Gallic and Pannonian elements of the name are often reversed has no bearing on either view, that is, that the singular unit moved to Dacia, or that it never left. 609 RMD I.50. 610 CIL XI.5632; Matei-Popescu 2010a: 185. 611 See Whately 2012. 612 It is recorded in RMD I.47 dated to AD 127, RMD III.165/V.399 dated to AD 145, RMD IV.270 dated to AD 146, and from RMD I.50 dated to AD 157. Cf. Weiss 2008: 299. 613 RMD IV.241. 614 RMD IV.260. 604

40

The Lower Danube from AD 81 to AD 161 transfer of the cohort to Thrace. 615 There is an inscription from the modern Kabyle in Thrace that lists a prefect of the cohort and it is dated to AD 136. 616 What is more, Velkov notes that the administrative boundary between the two provinces of Moesia Inferior and Thrace was changed in AD 136. 617 As such, Velkov’s argument is convincing.

cohors II Chalcidenorum, cohors I Cilicum milliaria sagittaria eq, cohors I Claudia Sugambrorum veterana eq, cohors II Flavia Brittonum, cohors I Flavia Numidarum eq, cohors I Germanorum, cohors I Lusitanorum, cohors II Mattiacorum milliaria eq, cohors I Thracum Syriaca eq Table 2.6. Disposition of Legions and Auxiliary Units in Moesia Inferior in AD 138

That same diploma seems to be the last record of the cohors I Lepidiana in Moesia Inferior. Indeed, Spaul suspected that the cohors I Lepidiana was transferred to the Middle East at some point in the second century because it was active at Melik Cherif on the Euphrates under Septimius Severus. 618 The inscription naming the I Lepidiana includes the titles equitata and bis torquata, neither of which have appeared in other inscriptions. Some have argued that the cohort was transferred earlier than that, however, and possibly during the reign of Hadrian. 619 The cohors VII Gallorum equitata was also transferred to the east in the second century, perhaps during the reign of Hadrian. It is last attested in Moesia Inferior around AD 114, 620 and is later recorded in Syria in AD 156/157. 621

CONCLUDING THOUGHTS As we close this chapter, it is time to take stock of the changes in troop deployment, at least where possible. Although we are much better informed about the movements in these years, which run from Domitian to Marcus Aurelius, significant gaps remain. Still, it is much easier, thanks in large part to the abundance of military diplomas. We will again rely on the snapshots provided by the tables presented above, which themselves focus on three years: AD 93, near the end of Domitian’s reign and in the aftermath of the most significant of his Dacian problems; AD 100, in the years – even months – just before Trajan’s two Dacian wars; and AD 138, at the end of Hadrian’s reign, one characterised by consolidation. In that first year, there were two legions each in Moesia Inferior and Superior. That makes four total for the former province. On the other hand, there might have been as many as 23 auxiliary units in Inferior, and there were 11 in Superior. Less than a decade later, in AD 100, although the number of legions remained the same, there had been a huge increase in the number of auxiliary units in Moesia Superior: 24, from an earlier total of 11. A few decades later, in AD 138, when the dust had settled, Moesia Inferior had an additional legion, three instead of two, while Moesia Superior still had two legions. Both provinces had seen a significant reduction in their respective auxiliary garrisons: from 23 to 18 for Inferior, and from 24 to 12 for Superior. The change in troop numbers that we looked at in this chapter reflects the significant role that the provinces now played in Roman strategic thinking. This also reflects the considerable amount of conflict that beset the region. As with chapter one, we discuss the implications of these changes in chapter six.

One last point: as noted above, the cohors I Cilicum had been based in Moesia Superior since it arrived in Moesia. At some point during the reign of Hadrian, however, it was transferred to Moesia Inferior. In a constitution dated to April 2, AD 134, the cohort is listed in Lower Moesia. 622 Its last known presence in Moesia Superior was in the summer of AD 115, when it was recorded on a few diplomas. 623 Those same diplomas record units dispatched eastwards for Trajan’s Parthian War, and so if the unit had been transferred to Moesia Inferior upon its return, that would have been AD 117 or 118 at the earliest.

Legions (3) I Italica, V Macedonica, XI Claudia

Auxiliary Units (17) ala I Flavia Gaetulorum, ala Gallorum Atectorigiana, ala II Hispanorum Aravacorum, ala I Pannoniorum et Gallorum catafracta, ala I Vespasiana Dardanorum, cohors I Bracarorum, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq,

615

Velkov 1990: 250. Cf. Spaul 2000: 83-84. AE 1999, 1370=GeA 471 = AE 1991, 1402; Velkov 1990: 250. 617 Velkov 1990: 250. 618 Spaul (2000: 156) on the basis of AE 1908, 22. 619 Speidel 2007: 82-83; Matei-Popescu 2010a: 218-219. 620 CIL XVI.58. 621 CIL XVI.103, 106. Matei-Popescu 2010a: 212. 622 CIL XVI.78. 623 Eck and Pangerl 2009: 363-370; Eck and Pangerl 2015: 229-230. 616

41

not provide us with any clue as to where this may have happened, or when, although it does fall before Avidius became the commander of the Syrian army. Thus, if it is authentic it may fall sometime early in the reign of Marcus Aurelius. 6 The last possible instance of warfare occurred around AD 178 when, “the Scythian situation again called for him [Marcus Aurelius]”. 7 However, that is all we learn about this.

Chapter 3: The Lower Danube from AD 161 to AD 235 INTRODUCTION This chapter covers the last period that we will review in this book in any detail, and it runs from the reign of Marcus Aurelius to the death of Severus Alexander, that is just before the so-called “Third Century Crisis”. The epigraphic habit falls off as we move into the third century, and the auxiliary diplomas, so valuable for the period covered by the previous chapter, all but disappear. We start with the military background. We turn next to the legions before finishing with the auxiliary units.

Commodus There are more hints of trouble in the vicinity of the Moesias in the beginning of Commodus’ reign. Dio’s excerpter refers to some problems with 12,000 neighbouring Dacians. They were apparently on the point of attack when they were dissuaded by a certain Sabinianus, who promised them land in Roman Dacia. 8 There are also reports of a war with barbarians purported to have been from the region beyond Dacia, and this conflict was led on the Roman side by both Albinus and Niger. 9 Dio claims that the two later pretenders to the throne “won fame” 10 in those encounters. The HA may be referring to the same event when it says that, “at this time in Sarmatia the noteworthy accomplishments belonging to the two other generals were attributed by Perennis to his son.” 11 Herodian tells us that, at least at first, Commodus was sensible enough, “allocating the control of the Danube campaign to trustworthy commanders, with orders to check the incursions of barbarians”. 12 Regrettably, the sparse detail cannot be corroborated with further materials and so we do not know the scale of the operations. Unfortunately, all three authors provide us with three different and equally vague descriptions of the location of the operation: Dio’s excerpter says that it was beyond Dacia; the Historia Augusta claims that it was in Sarmatia; and Herodian alleges that it was on the Danube. It could very well be that this operation was little more than a clean-up expedition following Roman success in the Second Marcomannic war. It is also possible that these three authors are referring to three different events.

BACKGROUND: MILITARY HISTORY Marcus Aurelius The writer of the Historia Augusta refers to the Marcommanic War as one “which surpassed any in the memory of man”. 1 Fortunately for the inhabitants of the provinces of Lower and Upper Moesia, the focal point for this horrific war was around the upper and middle Danube; in fact, indirectly or otherwise, the sources do not tell us whether any fighting from the wars spilled over into the Moesias. 2 There are, however, signs of some activity in Dacia and in Moesia Inferior. 3 At some point in AD 169 or AD 170 some tribes crossed into Dacia who demanded money and threatened war, though this need not have occasioned assistance from the Lower or Upper Moesian administration. 4 There is a vague reference in the Historia Augusta to Avidius Cassius severely punishing a group of auxiliaries under his command, who had slaughtered 3000 Sarmatians. 5 This reference is in the midst of a discussion of Avidius’ reputation as a strict disciplinarian; moreover, it is reminiscent of Tacitus’ descriptions of Corbulo during the eastern campaign under Nero. The Historia Augusta also does

Septimius Severus to Severus Alexander

1

HA Marc. 17.2. trans. Magie. 2 A good summary of the war and many of its problems is found in Birley’s (1993: 249-255) biography of Marcus Aurelius. He goes into more detail about the war on pp 159-183. 3 The comment above does not mean that the Moesian armies were not active in the wars, or that there was no activity in the area. Mocsy (1974: 187), for example, suggests that Verus was active in both Dacia and the Moesias. At the same time, Birley (1993: 160161) has said that in the opening phase, that is around AD 169, much of the fighting was confined to the Hungarian plain, which was part of the barbarian territory bordered by Lower Pannonia in the west, Upper Moesia in the south, and Dacia in the east. What is more, later during the war (c. AD 170) the Costoboci passed through Thrace and Macedonia before reaching as far as Eleusis in Greece. Unless they travelled by sea, they presumably would have passed through Moesia Inferior on their way south and west. Moesia Superior (Birley 1993: 164ff) may have had its fair share of activity from AD 170 onwards. 4 Cass. Dio 72.11.1-2. 5 HA Avid. Cass. 4.6.

Generally, following the reign of Commodus and prior to the chaos of the middle of the third century, we have very few specific details about military operations that took place in the Moesias. What we do know is that 6

Assuming that Marcus Aurelius’ Parthian war started in AD 162, that would put Avidius’ command to some time before then. Regrettably, Scriptor (the unknown author of the Historia Augusta) also does not give us any idea as to what army he may have been leading: “cum exercitum duceret” (HA Avid. Cass. 4.6). 7 Cass. Dio 72.33.1. 8 Cass. Dio 73.3.3. 9 Cass. Dio 73.8.1. 10 Cass Dio 73.8.1. trans. Cary. 11 HA Comm. 6.1. 12 Hdn. 1.6.8. trans. Whittaker.

42

The Lower Danube from AD 161 to AD 235 the region was almost beset with war in AD 196. Dio’s excerpter alleges that, “the Scythians were in a mood for fighting at this time”. 13 Mother Nature, however, managed to kill their desire to act upon this. 14

Claudia were in Moesia Inferior; the legio IIII Flavia and the legio VII Claudia were in Moesia Superior. The first sign of movement came several years later. In book 55 of his History, following a discussion between Livia and Augustus, Cassius Dio includes a digression on the legions and their financing. In this passage Dio said, “Twenty-three, or, as others say, twenty-five legions of citizen soldiers were being supported at this time. At present only nineteen of them still exist, as follows”. 23 The “now”, or “at present” (νυν), that Dio refers to is some point between the years of AD 211 and AD 229. 24 There are two additional points of interest. Firstly, Dio provided the names of the nineteen remaining legions and their provinces. Those relevant here are: “the Fifth Macedonica in Dacia 25…and the Seventh which is in Upper Moesia, and which is most certainly called Claudia 26…the Eleventh in Lower Moesia, the Claudian; for two legions were thus named after Claudius because they did not wage war against him in the revolt of Camillus”. 27 Secondly, he updated the list, so referring to those that had been created since the days of Augustus: “I shall speak of the other legions also which exist today and tell of their enlistment by emperors subsequent to Augustus”. 28 Among this supplementary list are two legions relevant to this discussion, namely the legio I Italica, and the legio IIII Flavia. Dio has this to say: “For Nero named the first one the Italica and it winters in Lower Moesia 29…Vespasian…the Fourth Flavian in Upper Moesia”. 30 Thus, the legionary disposition of the two Moesias in Dio’s lifetime was as follows: the legio I Italica and the legio XI Claudia were based in Moesia Inferior; and the legio IIII Flavia and the legio VII Claudia were based in Moesia Superior. The legion missing from that list is the legio V Macedonica, which had moved from Moesia Inferior to Dacia. 31

During the reign of Caracalla in about AD 214, the HA alleges that Caracalla conquered some Goths on his way east. 15 This story, however, may be based on the knowledge that Caracalla had murdered his brother Geta, and the similarity between the names Geta and Getae. For he says: “ ‘Add besides, please, Geticus Maximus,’ because he had killed his brother Geta, and the Goths are called Getae”. 16 Around AD 218, the Dacians ravaged portions of Dacia; after getting back hostages the raiding stopped. 17 Towards the end of Severus Alexander’s reign, perhaps between AD 231 and AD 235, Herodian tells us that, “the Germans were on the march across the Rhine and Danube, devastating the Roman Empire, over-running the garrisons on the river banks, and also the cities and villages”. 18 Herodian is likely referring to the Upper Danube, 19 though that does not rule out the possibility of problems along other stretches of the river. We would like to say something about the Bellum Bosporanum, a little known conflict that is recorded solely in an inscription from Preslav, which dates to the first half of the third century. 20 The lack of additional evidence, however, means there is little else for us to say, including even to what degree the Romans themselves were involved. 21 TROOP MOVEMENTS FROM MARCUS AURELIUS TO SEVERUS ALEXANDER Legions As with the previous two chapters, we start with the legions. We have an inscription from the early years of the reign of Marcus Aurelius, found in Rome, which lists the legions of the empire in geographical order from west to east. 22 Based on that legionary list, and the reconstruction presented at the end of the last chapter, the legionary disposition in the Moesias early in Marcus Aurelius’ reign was as follows: the legio I Italica, the legio V Macedonica, and the legio XI

When did the Fifth Macedonian Legion leave Lower Moesia? Many scholars, such as Forni, Parker, Filow, Gudea, Campbell, Piso, and Farnum favour a date of AD 167 or 168. 32 Others, such as Dando-Collins, Pollard and Berry, and Wesch-Klein are more vague, 23

Cass. Dio 50.23.2, trans. Cary. These are the years when Cassius Dio is believed to have been writing. 25 Cass. Dio 55.23.3. 26 Cass. Dio 55.23.3. 27 Cass. Dio 55.23.4. 28 Cass. Dio 55.24.1. trans. Cary. 29 Cass. Dio 55.24.1. The text reads: “Vespasian [organised] the Second Adiutrix which is in Lower Pannonia and the Fourth Flavian in Upper Moesia”. 30 Cass. Dio 55.24.3. 31 Cass. Dio 55.23.3. 32 Forni 1958: 202 - AD 167; Parker, 1958: 167-168 – he does not specify the date; Filow 1963: 77-78 – AD 167 or AD 168; Gudea 1979: 84 – AD 167; Campbell 1996a: 840 – AD 167 or AD 168; Piso 2000: 215 – AD 168; Farnum 2005: 19. Gudea (1979: 84) notes that “few units were transferred to Dacia after 120; such arrivals are recorded in Dacia Inferior only.” In fact, the two major waves of troop movement in Dacia are in the years immediately following the conquest, and between the years of AD 110 and AD 120 (Gudea 1979: 84). 24

13

Cass. Dio 75.3.1, trans. Cary. The excerpter reports: “The Scythians were in a mood for fighting at this time; but while they were consulting together, thunderings and lightnings, accompanied by rain, suddenly broke over them, and thunderbolts fell, killing their three chief men, and this restrained them” (75.3.1, trans. Cary). 15 HA M. Ant. 10.6. 16 HA M. Ant. 10.6. 17 Cass. Dio 79.27.5. 18 Herodian. 6.7.2. trans. Whittaker. 19 This is suggested by his choice of the term “Germans”, and by his later claim that, “[the Germans were] putting the Illyrians who bordered Italy as neighbours in considerable risk” (6.7.2, trans. Whittaker). 20 AE 1991, 1378=AE 2011, 1142. 21 See Sarnowski 1991. 22 CIL VI.3492=ILS 2288. See Campbell (1994: 84-85). 14

43

Exercitus Moesiae though they suggest that the transfer came at the end of Lucius Verus’ Parthian war. 33 The last firmly dated inscriptions from Moesia Inferior that name the legion date no later than AD 167. These inscriptions, of which there are two, are both from Troesmis. This first one dates to AD 162 and records a centurion of the legio V Macedonica. 34 The second one dates to some point between AD 162 and AD 167, 35 and records a Valerius Valens, a soldier of the Fifth Macedonian legion who died in the Parthian campaign. 36 The transfer likely did take place after AD 167. Unfortunately, the earliest Dacian inscription that refers to the legion dates to AD 195. 37 Thus, we should date the change in disposition to some point between the years AD 167 and AD 195.

Fortunately, we have yet more evidence. There is a rather lengthy inscription that preserves the names of soldiers who enlisted in AD 169 for the legio VII Claudia and were discharged in AD 195, 42 and Mirković has published a fragment that includes a military roster. 43 There were around 270 names of soldiers on it ready for discharge in AD 195. Although we have at least two other military rosters that list a greater number of soldiers, 44 we have another inscription that names soldiers recruited around the same time (AD 168) from Egypt for the legio II Traiana. 45 By comparison, there are only around 120 soldiers named on the Egyptian list. Also, we have another roster for the legio VII Claudia. 46 In this case the soldiers were enlisted in AD 134 or 135, and they were discharged around AD 160. Significantly, while there are still several soldiers listed, 239 to be exact, this number is still lower than the inscription of AD 195. In other words, there is a demonstrable increase in recruits that we might well attribute to the needs of the Marcommanic wars. 47

Can we narrow down the date any further? Well, there is an inscription, which probably dates to AD 170, that records a Tiberius Valerius, a veteran of the legio V Macedonica. 38 This man served in the Oriental (Parthian) and the German (that is Marcomannic) campaigns of Marcus Aurelius’ reign. Significantly, this Valerius was honourably discharged from Dacia, “H(ONESTA) MISSIONE IN DACIA”, and not Moesia Inferior, despite the fact that the inscription itself was found at Troesmis (Lower Moesia). Thus, we might well assume that the legion was based in Dacia when the inscription was put up, 39 and we can now narrow down the date range to the years AD 167 and AD 170.

If extra soldiers were called into service and enlisted in the legio VII Claudia, might a similar recruiting drive have taken place in the other legions? Tiberius Valerius and the rest of his comrades in the legio V Macedonica were certainly called into action by AD 170 (at the latest) for the second major Roman offensive; it follows that by that time the legion was based in Dacia. 48 Therefore, it is possible that once news of the invasion had reached the Roman forces in the east and the dust had settled from that conflict, the Danubian troops were sent back post-haste. As a result, the legion had been transferred by the end of AD 167, or before the first Roman attack in AD 168. 49 And so, by AD 168 the legiones IIII Flavia and legio VII Claudia were in Moesia Superior; the legiones I Italica and legion XI Claudia were in Moesia Inferior, all of which remained in the region well into the third century and beyond, for all four legions are fond in the Notitia Dignitatum. 50

As it turns out, we can be even more precise, for select details of the Marcommanic wars allow us to pin down the transfer further. 40 We know that the first incursion was in AD 167. The Roman counter-offensive, however, did not begin before AD 168; it was cut short, thanks, in no small part, to the death of Lucius Verus in AD 169. 41 There was an offensive in the spring of AD 170, and by AD 171 the Romans had expelled the invaders. But, there was another offensive again in AD 172. In the end, it is the offensive of AD 170 that should be equated with the expeditio Germanica from the aforementioned inscription, CIL III.7505.

Auxiliae: Moesia Superior

42

CIL III.14507 (from Viminacium). Mirković 2004. 44 See the chart on p 24 of Mirković’s article. There, an inscription from Troesmis (IScM V.137) and one from Lambaesis (CIL VIII.18068) both list around 300 soldiers. 45 CIL III.6580. 46 IMS II.52. 47 The legio II Traiana did not participate in the conflict. 48 The Romans were not ignorant of the problems that their tribal neighbours, such as the Sarmatians, could pose. Conversely, the Sarmatians and Rome’s other neighbours were generally well aware of Rome’s affairs as the first Sarmatian incursion seems to have occurred towards the end of the Parthian war, or at least before the Roman troops from the Danube had returned. 49 Note the comments of Matei-Popescu 2010a: 57. 50 All four legions are listed in the Notitia Dignitatum (Not. Dign. or. 40.30, 33 (I Italica and XI Claudia), 41.30, 31 (IIII Flavia and VII Claudia)), though they would have been quite different from the legions as discussed here. See Whately (2015: 10-11). 43

33

Dando-Collins 2010: 137; Pollard and Berry 2012: 201; WeschKlein 2012. 34 CIL III.6169. 35 These are the dates for the Parthian campaign led by Lucius Verus. 36 CIL III.6189. 37 CIL III.905. 38 CIL III.7505. Part of the inscription reads (lines 4-9): ...ex|[pedi]t(ione) Orientali sub St|[at(io) Pri]sco, Iul(io) Severo, M[art(io)| Vero] c(larissimis) v(iris), item Germ(anica) sub|[Cal]pur(nio) Agricola, Cl(audio) Fronto|[n]e c(larissimis) v(iris)... Cf. Ritterling 1925: 1579; Parker 1958: 167-168; Filow 1963: 77-78; and Piso 2000: 214-215; IScM V.160. 39 This is not to say that the province of Dacia did not feel the effects of the war, in AD 167 (Birley 1993: 252) the gold mines in Dacia were attacked. 40 For the dates and events of the Marcommanic wars I am relying on the reconstruction of Birley (1993: 249-255). 41 The plague also hampered Roman efforts in AD 168.

44

The Lower Danube from AD 161 to AD 235 We now turn to the auxiliary units posted in the two provinces during this last period. Unfortunately, our last diploma that gives us a list of units dates to February 8, AD 161, 51 which is about a month before Marcus Aurelius took the throne. What is more, the ancient authors are, if anything, even less informative than they were for earlier periods regarding the auxiliary units. Thus, we are almost totally dependent on what stone inscriptions we have for the period from AD 161 to AD 235.

Dardanian region of Moesia Superior was close to Dalmatia, and it is easy enough to imagine some overlap, and confusion, with respect to the borders of the two. Beneš and Spaul correctly restrict its emplacement to Moesia Superior. 59 Both the cohors I Aurelia Dardanorum and the cohors II Aurelia Dardanorum 60 remained in Moesia Superior well into the third century. 61 These Dardanian units may not have been the only new units raised in the province, for there may have been a handful of other Moesian units raised by Marcus Aurelius. Mirković and Dusanić argued that four additional ones were raised in AD 169: the cohors II Aurelia nova, the cohors II Aurelia nova milliaria civium Romanorum, the cohors I Aurelia Pasinatum civium Romanorum milliaria, and the cohors II Aurelia nova Sacorum. 62 Not everyone agrees, however, and the names of these units themselves are problematic.

There is more evidence for change in the garrison of Moesia Superior from the reign of Marcus Aurelius on than there is for Moesia Inferior. The HA, in a discussion of the extreme measures taken by Marcus Aurelius in light of the eastern plague and the troubles with the Marcomanni, said: “He even made the bandits of Dalmatia and Dardania soldiers.” 52 Strabo, Pliny, and Ptolemy identify Dardania with Moesia Superior. 53 From the list of known auxiliary units in Moesia Superior before Marcus Aurelius, none had a clear Dardanian origin. 54 There was a cavalry wing, the ala I Vespasiana Dardanorum, which was based in Moesia Inferior, but this unit was never part of the army of Moesia Superior. If we turn to the inscriptions that we can date from Moesia Superior, we find a few new units, which could be identified with those noted by the HA. Both the cohors I Aurelia Dardanorum and the cohors II Aurelia Dardanorum first appear during the reign of Marcus Aurelius; moreover, the unit’s imperial cognomina point towards a connection with that emperor. Their ethnic name is even more indicative. In this case, the creation of the Dardanian units, the HA was not ill-informed.

Beginning with the cohortes II Aureliae novae, the most noticeable feature of these units is their nomenclature. All three, for example, are numbered “II”, bear the imperial cognomen Aurelia, and are referred to as new, or reformed – nova. The first question, then, is whether we actually are dealing with three distinct units. Of those three, the two that are the best candidates for being considered one unit are the cohors II Aurelia nova and the cohors II Aurelia nova milliaria civium Romanorum. The epithet milliaria reflects the size of the unit: the standard unit was quingenary and would have had a paper-strength of about 500 soldiers whereas the larger milliary unit would have been about twice the paper-strength of the quingenary unit with around 1000 men. So, the epithet milliaria marks a unit out as larger than the standard quingenary unit, but there need not have been another similarly named quingenary unit to necessitate the use of this epithet. As noted earlier, the epithet civium Romanorum was bestowed on a unit for meritorious service, and was not usually a title given to a unit when it was created. Based on this evidence alone, upon this single unit’s creation it may have borne the title cohors The earliest securely dated II Aurelia nova. 63 inscription that we have dates to AD 179 and the unit is called the cohors II Aurelia nova milliaria equitata on

There is one more matter to discuss with respect to the Dardanian units. Cichorius dated the formation of the cohors I Aurelia Dardanorum to the reigns of either Antoninus Pius or Marcus Aurelius. 55 He correctly limits its emplacement to the province of Moesia Superior. Wagner and Kraft, however, do not. Wagner erroneously suggests that the unit was first stationed in Dalmatia in the first half of the second century. 56 Kraft rightly dates its foundation to the reign of Marcus Aurelius, though he suggests that the unit may have been in Dalmatia before moving to Naissus in Moesia Superior. 57 Their argument is based on an inscription from Dalmatia, the existence of which does not prove that the unit was ever based in the province. 58 The

below, there are a handful of cohorts that bear the imperial cognomen “Aurelia”. 59 Beneš 1978: 30; Spaul 2000: 349. 60 There are no such scholarly discrepancies surrounding the garrison of the cohors II Aurelia Dardanorum, although Cichorius (1900) was not aware of its existence. 61 The cohors II Aurelia Dardanorum is recorded in Moesia Superior as late as AD 242 (AE 1952, 191). Unfortunately, the latest datable inscription for the cohors I Aurelia Dardanorum dates to sometime after the accession of Septimius Severus (CIL III.8251). 62 Mirković & Dusanić 1976: 104. Mirković’s conviction seems to be supported as regards the period of recruitment, as is the case with the Dardanian units, by their nomenclature. 63 Whether the title nova was included in the original name of the unit is debatable; it depends on the translation accepted for this particular unit, that is ‘new’ or ‘reformed’.

51

Marcus Aurelius ascended to the throne on March 7, AD 161 (Birley 1996: 220). 52 HA Marc. 21.7. 53 See Strabo 7.5.7; Plin. HN 3.26.149; and Ptol. Geog 3.9.1-4. 54 The units are listed in RMD IV.247, which dates to September 9, AD 132; CIL XVI.111, which dates to AD 159/160; and RMD I.55, which dates to February 8, AD 161 – that is about a month before the accession of Marcus Aurelius. 55 Cichorius 1900: 280. 56 Wagner 1938: 130-131. 57 Kraft 1951: 175. 58 CIL III.14700. This inscription records a Surus Victor, a soldier of the “COH I MIL AUREL”. There is no indication from those abbreviations that we should associate this particular cohort with the cohors I Aurelia Dardanorum. As we saw above and shall see

45

Exercitus Moesiae that inscription. 64 This suggests that the unit was milliary from its inception, or shortly thereafter. 65 With the passage of time and following meritorious service the unit’s name was probably expanded to include civium Romanorum. 66 It seems most likely, then, that we do have just the one unit, which likely remained in Moesia Superior past the chronological terminus of this book.

Mongolia in the east. 75 The Sacae/Sakai, Saka, or Sakas which they are sometimes called, were, generally speaking, eastern Scythians. 76 They lived around the Caspian Sea in the region north of the Jaxartes River. 77 They were neighbours to the Alans, 78 whom Arrian, as governor of Cappadocia, came up against in the middle third of the second century AD. These Scythians were apparently able to hold sway over a vast area until things fell apart in the second or first century BC. 79 After this, McGovern says that they became scattered over a host of “various isolated regions”, among which he includes the Crimean peninsula and the Dobrudja. 80 Their recorded history, at least from a Greek and Persian perspective, stretches back to the sixth century BC. 81 The Sacae are, however, still discussed at the end of the first century BC, 82 and at some point in the second century. 83 Perhaps significantly, McGovern says that the Sacae were remarkable warriors. 84

We now turn to the cohors II Aurelia nova Sacorum. The existence of this unit rests on one lone inscription. 67 That inscription records an Aurelius Victor, a soldier of the “C(OHORS) II AUR(ELIA) N(OVA) SACOR(UM)”. Although this unit contains the familiar II Aurelia nova, the SACOR is much less so. Wagner wants to identify the Saci with a Dacian tribe that lived south of Apulum, 68 and he denies the equation of these Saci with the Scythian Sacae. 69 Beneš, on the other hand, believed that this cohort was recruited from the Saci, but claims that we do not know anything about this tribe, or where they might have lived. 70 He does suggest, however, that they may have been from somewhere around the Dobrudja where there was a city named Sacidava. 71 Although the equating of the tribe with the city in the Dobrudja is dubious, the Scythian connection and the possible association with the Sacae need to be explored further.

Thus, we have an identifiable confirmation for a Scythian presence in the Dobrudja, and we know that the Sacae, at least from a Roman perspective, were active as late as the second century. There is one further possibility, for McGovern says: “the decay and eventual downfall of the Scythians was due almost entirely to invasion by their distant kinsmen, the Sarmatians.” 85 Wilkes has also noted that the historical sources tended to confuse “the Sarmatians” and “Scythians”. 86 Thus, we might tentatively identify the SACOR from the cohors II Aurelia nova Sacorum with the Sarmatians recruited as part of the treaty with the Iazyges. The Romans of Moesia Superior may had some vague knowledge of the Sacae from Scythia Minor and confused them with the new Sarmatian recruits. 87

Dio’s excerpter tells us that around AD 175 the Iazyges came to terms with the Romans. 72 Among the terms discussed, which included the return of 100,000 captives, the excerpter also says: “they promptly furnished as their contribution to the alliance eight thousand cavalry, fifty-five hundred of whom he sent to Britain.” 73 The Iazyges were a Sarmatian people who lived more or less between Pannonia Inferior and Roman Dacia. 74 Conversely, the Scythians inhabited an area that spanned from the expanse of the Hungarian plain and the foot of the Carpathians in the west, to

The last potential unit to discuss is the cohors I Aurelia nova Pasinatum. There are two distinct schools of thought with regard to the cohors I Aurelia nova Pasinatum. On the one hand, we have one group that acknowledges the existence of this unit. Wagner says

64

CIL III.14537. Wagner (1938: 91-92) believes that this was the unit’s official name. 65 It may have increased in size with the needs of the Marcommanic wars. 66 Kraft (1951: 168) only identifies one unit, namely the cohors II Aurelia nova milliaria equitata. The same is true of Spaul (2000: 484). Nowhere do either Cichorius (1900) or Beneš (1978) discuss the possibility of one cohors II Aurelia nova, let alone two. 67 CIL III.14217-6 (AE 1901, 21; IMS I.119). I have been unable to find any other inscriptions from the Moesias that refer to this alleged unit. 68 Wagner 1938: 182. 69 From a linguistic perspective, Wagner’s decision to renounce the Sacae as a possibility has its merits. The genitive plural of Sacae is Sacarum and not Sacorum. Of course, if we were to equate the SACOR from this inscription with the Sacae, this might be a mistake on the part of the stonecutter. 70 Beneš 1978: 50. 71 Beneš 1978: 50. Beneš’ hypothetical tribe from Dobrudja is probably the Scythian tribe referred to by Wagner. 72 Cass. Dio 72.16.1-2. On Marcus Aurelius’ relationship with the Iazyges, especially in light of the hostilities on the Danube, see Batty (2007: 437-441). 73 Cass. Dio 72.16.2. trans. Cary. 74 For a good overview see Wilkes (1983: 255-289) and Batty (2007: 225-236). Cf. Rice 1981: 281-293.

75

Rolle 1989: 16; Batty 2007: 205. For an overview of the Scythians see Batty (2007: 204-214). 76 Rolle 1989: 37. Cf. Batty 2007: 489; Kim 2013: 10-13. 77 McGovern 1939: 40. 78 McGovern 1939: 40. 79 McGovern 1939: 38. 80 McGovern 1939: 38. 81 Rolle 1989: 47. Cf. McGovern 1939: 38; Hdt. 1.153, 6.113, 7.96, 7.184, 8.113, 9.31, 9.71, 9.113. At 3.93, Herodotus lumps the Sacae together with “the Caspians”: he is discussing the tribute-paying peoples that are under the Persian yoke. At 7.64 Herodotus has a brief digression about the Sacae. 82 Strabo 11.8.2. 83 Ptol. Geog. 6.13. cf. Ael. Arist. Or. 26.29-30. 84 McGovern 1939: 40. “These Sakan regiments were among the most famed of all the fighting forces of Asia.” 85 McGovern 1939: 40. 86 Wilkes 1983: 255. 87 If the learned Roman authors confused the two peoples, there is little reason to see why the same thing might not have happened among poorly educated Roman soldiers. Still, the preceding argument is not without its problems. For example, why was the unit numbered “II” and not “I” when no other unit is known? Moreover, Xiphilinus refers to cavalry when he notes the Sarmatian military contribution.

46

The Lower Danube from AD 161 to AD 235 cohors I Pannoniorum. 96 It is recorded in the province as late as February 18, AD 165, 97 and so probably stayed in the province well into the third century. However, that does not mean that we can negate the presence of the other units.

that Marcus Aurelius raised the cohors I Aurelia Pasinatum civium Romanorum, but leaves open the possibility that it was either at the beginning of the Marcomannic wars, or in the year AD 175. 88 Curiously, Kraft refers to both a cohors I Aurelia nova Pasinatum civium Romanorum and a cohors II Aurelia nova Pasinatum civium Romanorum, 89 and does not explain the discrepancy. Beneš agrees that the unit was probably raised in the second century, though he does not specifically state that it was Marcus Aurelius who raised them. 90 On the other hand, we have another group that does not acknowledge the unit’s existence. Cichorius did not refer to the unit in his important history of the infantry cohorts, nor does Spaul in his follow-up to Cichorius. Mocsy seems to associate this unit with the cohors II Aurelia nova. 91 Which of these two possibilities – it existed or it did not – is more likely?

The ala I Claudia nova miscellanea 98 remained in Moesia Superior into the third century; the same can be said for the ala I Gallorum Flaviana. 99 The cohors V Gallorum stayed in Moesia Superior at least until the beginning of Marcus’ reign. By April 1, AD 179, the unit had been transferred to Dacia Superior. 100 At the same time, the cohors III campestris was in Upper Moesia as late as February 8, AD 161. 101 By April 1, AD 179, the cohors III campestris had moved to Upper Dacia. 102 Perhaps both units had moved to Dacia as part of the push to defeat the Marcomanni in the last of the northern wars. The cohors V Hispanorum likely spent the rest of its days in Moesia Superior. 103 With that said, it might have been transferred to Numidia in the wake of some unrest, at least if an inscription found at Lambiridi is anything to go by, though we could not say how long it would have gone for. 104 We also do not know when that might have happened, besides some point in the third century – and so possibly after the period we are concerned with. There is some indication too that the unit was involved in combat based on the phrase: “DESIDERATO IN ACIE” (“died in battle”). 105 The cohors I Montanorum likely remained in Moesia Superior; the same can probably be said for the cohors I Cretum. 106 The cohors II Gallorum likely stayed in Moesia Superior into the third century, as did the cohors III Brittonum. The cohors I Lusitanorum was in Moesia Superior at the accession of Marcus Aurelius. Only six years later on May 5, AD 167, this particular cohort of Lusitanians is recorded in Pannonia Inferior. 107 This might to be a bit early to be considered part of the gathering of troops for the Marcomannic wars, and could must have been some other reason for its transfer. 108 Given how long it would take to prepare for such a war, and bearing in

As with the previous unit, the epigraphic evidence is limited. We have a lone inscription from CIL that reads as follows: [I](OVI) O(PTIMO) M(AXIMO) [S]CRIBONIUS FAUSTUS V(ETERANUS) E AURE(L)IA(?) PASINA(?) V[---]CRCX [&. It is from this fragmentary reference that the existence of an entire unit, the cohors I Aurelia nova Pasinatum civium Romanorum milliaria, is deduced. 92 If we try to identify the puzzling “Pasina” in the inscription we do not get very far. Pliny refers to a tribe called the Pasini, who had a city named Aenona on the northeast Adriatic. 93 That is, however, all we hear about them. What is more, the inscription does not specify what kind of unit the veteran Scribonius Faustus was (legionary or auxiliary). Indeed, if anything the Aurelia Pasina might refer to an individual and not a unit. Given this limited evidence for the unit’s existence and the ambiguity surrounding the identity of the “Pasina”, it is easy to understand why Cichorius and Spaul excluded it. 94 In fact, in the absence of any corroborating evidence this unit ought to be stricken from the historical record. We must now discuss the whereabouts of the remaining auxiliary units in Moesia Superior. Of the twelve units in the province in AD 161, only one is found on an inscription which we can firmly date to Moesia Superior in this period. 95 That unit is the

96

CIL III.6302 records the cohort and is dated to some point between and including the years AD 161 and AD 180. CIL III.8162 records the cohort and is dated to late in the second century or early in the third century. 97 CIL XVI.120. 98 See Cichorius 1893: 1238; Wagner 1938: 28-29; Kraft 1951: 144-145; Beneš 1978: 7; Spaul 1994: 90. 99 See Beneš 1978: 9; Spaul 1994: 115. 100 RMD II.123. 101 RMD I.55. 102 RMD II.123. 103 Cf. Wagner 1938: 155-156; Spaul 2000: 135. 104 CIL VIII.4116. 105 See Bertolazzi 2015. 106 Spaul (2000: 385) highlights some of the confusion surrounding the emplacement of this unit. He notes that the unit put up what was probably an official inscription in Dacia (AE 1968, 453), though it was recorded in a Moesian diploma (RMD I.55). Indeed, provincial boundaries are known to have fluctuated and so while a unit might have stayed at the same base, the base itself may have flip-flopped between two neighbouring provinces. 107 CIL XVI.123. 108 This could be another instance of border shifting. Spaul (2000: 62) believes that this might have been the case.

88

Wagner 1938: 179. Kraft 1951: 101, n. 4; 183. It is possible, or rather probable that the reference to a cohors II Aurelia nova Pasinatum civium Romanorum is an editorial mistake. 90 Beneš 1978: 48. 91 Mocsy 1974: 195, n. 71. 92 CIL III.14545. 93 Plin. HN 3.21.140. 94 Spaul (1994 and 2000) tends to identify fewer units rather than more, and there certainly is much to be said for this tendency. In the words of Lendon (2004: 443), “Spaul is a pessimist about the accurate use of full unit names on inscriptions, and so a minimalist when it comes to counting units.” 95 In fact, most of the datable inscriptions are for those “Aurelian” units raised by Marcus Aurelius. 89

47

Exercitus Moesiae mind that the conflict erupted in the mid 160s, an earlier transfer might well be connected with this.

cohors I Pannoniorum

The last auxiliary unit from Moesia Superior to discuss is the cohors I Antiochensium, another unit for which there had been, at least until fairly recently, a great deal of uncertainty. 109 Due to what had been limited evidence for the unit in the latter part of the second century and start of the third century, Spaul had wondered whether the unit was reorganised following the Marcommanic wars. In fact, Spaul had suggested that the unit was renamed the cohors I Hemesenorum; much like the Antiochenes, the Hemeseni are also from Syria. 110 On the other hand, there was a cohors I Hemesenorum based in Pannonia Inferior from at least AD 178 on. 111 We also now have a handful of diplomata which record the cohors I Antiochensium, and which date to the second half of the second century. 112 Thus, while evidence from the end of the second and beginning of the third might be slim, as we have seen, in this regard it does not differ a great deal from some of the auxiliary units mentioned above. In other words, this would seem to be the result of changing epigraphic habits rather than any potentially catastrophic loss. Indeed, even if the unit had moved, such a dramatic name change seems unlikely, and so Spaul’s suggestion should be discounted. 113 Legions (2) IIII Flavia, VII Claudia

Table 3.1. Disposition of Legionaries and Auxiliaries in Moesia Superior in AD 170 Auxiliae: Moesia Inferior Of the eighteen units that were based in Moesia Inferior at Marcus Aurelius’ accession, only eight units have left a datable inscription recording their presence during the reigns of Commodus and the Severan emperors, and for many only one of those is datable. A soldier from the cohors I Cilicum sagittaria left his mark at Chersonesus at some point at the end of the second century or beginning of the third century. 114 A veteran of the cohors I Claudia Sugambrorum veterana was in the Chersonesus at the end of the second century or beginning of the third century. 115 We have a decurion of the ala Gallorum Atectorigiana at Tomis during the reign of Caracalla. 116 A prefect of the cohors II Mattiacorum was recorded in Lometz in AD 198. 117 On the other hand, the ala II Hispanorum Aravacorum left the greatest mark in this period in Moesia Inferior. We have a number of inscriptions that record its presence, the earliest of which dates between AD 161 and AD 169, 118 the latest of which dates to the beginning of the third century. 119 The ala I Vespasiana Dardanorum was found in Cerna at the beginning of the third century. 120 We have a prefect of the ala I Flavia Gaetulorum at Tomis between AD 238 and AD 244. 121 Finally, soldiers of the cohors II Flavia Brittonum were recorded working on the bathhouse at Sexaginta Prista in AD 230. 122

Auxiliary Units (15) ala I Claudia nova miscellanea, ala I Gallorum Flaviana, cohors I Antiochensium, cohors I Aurelia Dardanorum, cohors II Aurelia Dardanorum, cohors I Aurelia nova Pasinatum, cohors II Aurelia nova milliaria eq, cohors II Aurelia nova Sacorum, cohors III Brittonum, cohors I Cretum, cohors II Gallorum Macedonica, II Gallorum Pannonica, cohors V Hispanorum, cohors I Montanorum,

In the previous chapter we demonstrated that the ala I Pannoniorum et Gallorum catafracta had never been transferred to any part of Dacia, despite the earlier contention of scholars. 123 In fact, it is not recorded on any Dacian diploma that falls within the chronological limits of this chapter. 124 As a result, the unit likely

109

114

Spaul 2000: 424. The Hemeseni are the people from the area around Emesa. 111 CIL XVI.131. 112 See, for example, RGZM 31, RGZM 37, and RMD V.419. 113 Even if, by chance, the Antiochene cohort found itself in Pannonia Inferior as a result of a boundary change, there seems to be no real justification for such a dramatic name change: there was at the time no other Antiochene unit based in Pannonia Inferior (or anywhere else for that matter). Had there been another such unit, it would have been extremely unusual for the incoming unit, or the pre-existing one for that matter, to make such drastic changes in its nomenclature. For example, earlier we saw that when there were two distinct units of Sugambrians in Moesia Inferior, the Moesian command-structure opted for the following two names: the cohors I Sugambrorum [t]ironum,and the cohors I Sugambrorum veterana.

CIL III.13751. AE 2000, 1276. We discuss their presence in the Chersonesus in greater detail in chapter four below. 116 IScM II.93. 117 CIL III.14428. 118 IScM V.23. 119 CIL III.14214-22. 120 IScM V.218. 121 IScM II.106. 122 CIL III.7473. 123 Cichorius 1893: 1245-1246; Wagner 1938: 38-39; Spaul 1994: 83. 124 We have at least seven, one of which is from Dacia Superior while the other six are from Dacia Porolissensis. Significantly, Dacia Porolissensis is the province that was purported to have been the new

110

115

48

The Lower Danube from AD 161 to AD 235 remained in Moesia Inferior during this period. The cohors I Lusitanorum presumably remained in the province through the third century. 125 The cohors I Flavia Numidarum left the province of Lower Moesia sometime after the accession of Marcus Aurelius. It is listed on a diploma from Lycia and Pamphylia dated to August 23, AD 167, 126 and on another diploma from the same province dated to March 23, AD 178. 127

Thanks to changes in our quantity and quality of evidence, as well as changes in general Roman troop movement patterns, we cannot realistically offer the same number of snapshots of the garrison of the two provinces here as we did in the previous two chapters. Units became even more stationary than they had been in the reigns of Hadrian and Antoninus Pius; moreover, vexillations, which we discuss in the next chapter, were used increasingly often. There were a few scattered hints of transfers, especially in light of the conflicts of the reign of Marcus Aurelius, but if there were others we have virtually no evidence. On the other hand, some new auxiliary units were created, at least in Moesia Superior. In AD 170, the garrison for the two provinces had changed little from the last snapshot, AD 138. Moesia Inferior’s legionary garrison had dropped slightly to two from three, while Superior’s continued to be two-strong. In terms of their auxiliary garrisons, Inferior’s dropped slightly again, going from 18 to 17, while Moesia Superior’s increased slightly, from 12 to 15. We keep with our usual pattern and save discussion of the implications of these changes for chapter six.

The cohors I Thracum Syriaca likely stayed in the province throughout its history. Beneš suggested that towards the end of the second century the unit, or a vexillation of the unit, served in the Chesonesus. 128 This is possible, but not certain. 129 The cohors I Germanorum presumably remained in Moesia Inferior The three units of into the third century. 130 Bracaraugustani, the cohors I Bracarorum, the cohors I Bracaraugustanorum and the cohors II Bracaraugustanorum, also probably stayed in Moesia Inferior well into the third century. We bring our discussion of the auxiliary units of Moesia Inferior to a close with the cohors II Chalcidenorum. Yet again, the evidence is lacking and thus we can assume that the unit remained in Moesia Inferior in the third century. Legions (2)

Auxiliary Units (16)

I Italica, XI Claudia

ala I Flavia Gaetulorum, ala Gallorum Atectorigiana, ala II Hispanorum Aravacorum, ala I Pannoniorum et Gallorum catafracta, ala I Vespasiana Dardanorum, cohors I Bracarorum, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq, cohors II Chalcidenorum, cohors I Cilicum milliaria sagittaria eq, cohors I Claudia Sugambrorum veterana eq, cohors II Flavia Brittonum, cohors I Germanorum, cohors I Lusitanorum, cohors II Mattiacorum milliaria eq, cohors I Thracum Syriaca eq

Table 3.1. Disposition of Legionaries and Auxiliaries in Moesia Inferior in AD 170

CONCLUDING THOUGHTS

home of this ala. The latest dated diplomas, however, record the ala II Gallorum et Pannoniorum (CIL XVI.185 and RMD II.116). 125 Beneš 1978: 44-45; Spaul 2000: 60. Cf. Wagner 1938: 163-164. 126 RMD I.67. It turns out that a number of units from the Moesias seem to have left traces in Lycia and Pamphylia. See Bennett (2010). 127 CIL XVI.128. It is conceivable that this Numidian cohort marched east with Lucius Verus as he set out against the Parthians. 128 Beneš 1978: 53. 129 More on this in chapter four below. 130 In the words of Spaul (2000: 256): “Its service seems to have been based at Capidava on the banks of the Danube west of Constantza. There is little else to say.” Regrettably, Spaul’s last statement can be applied to most auxiliary units.

49

half of the second century and into the third century because it became difficult to transfer entire units. 3 As soldiers became settled in the province in which they were based, and as recruits became increasingly local, finances and stability became issues. The cost of moving entire legions – and auxiliary units for that matter – may have become prohibitive, except in the most dire of circumstances. 4 At the same time, soldiers with families were increasingly unwilling to uproot and abandon them for a few years, or longer, to serve the state in a foreign land. As a result, the thinking is that the Romans increasingly sent those soldiers with few ties in the belief that they would have had less to keep them back, whereas the remaining soldiers would fight better while at home if put to the test while their comrades were away.

Chapter 4: Vexillations and the Black Sea INTRODUCTION The subjects of this chapter are Moesian vexillations and the Roman troops along the northwest shore of the Black Sea, two not unrelated topics. Vexillation use in the Moesias roughly corresponds to what we know about Black Sea garrisons. In other words, the known Roman troops in the region seem to have been drawn from the Moesias, and Moesia Inferior in particular; that province seems to have been the source of the bulk (majority, even the sum total) of those cities’ vexillations. We also take a look at the Roman navy, which played a significant role in the functioning of the land-based Moesian soldiers. Although the evidence for this chapter, largely comprised of inscriptions and on occasion the physical remains of the region’s fortifications, is limited, there is more than enough material to get some sense of the military disposition. We begin with the vexillations, and then turn to the Black Sea, its garrison, and the region’s naval forces.

We do not know the size of vexillations, though it is probably safe to assume that the size of the detachment depended upon the exigency of the moment, as well as the needs of the province which was supplying the troops. Saxer stressed that the number of units detached depended on the size of the provincial army. 5 With respect to the number of soldiers in each vexillation, scholars have looked for some uniformity. Two of the numbers used are 1,000 and 2,000; this is because Tacitus, Suetonius, and Josephus, when discussing the size of detachments, tend to give figures in units of 1,000. 6 Luttwak, Holder, Keppie, and Le Bohec all use 1,000 or 2,000 as the typical size of a legionary vexillation. 7 Saxer himself identified the references which claim sizes of 1,000 and 2,000. 8 On the other hand, Tully argued for about 500 (or 480) as the standard size of the quingenary legionary vexillation. 9 His argument is based on PseudoHyginus’ treatise on camps. 10 Despite the suggestion for a standard size, Tully is quick to point out that this was not always the case. Even if 500 or thereabouts was the standard size, this does not mean that combat vexillations were necessarily created from pre-existing legionary cohorts. 11

VEXILLATIONS Vexillations have attracted limited attention with some noteworthy exceptions including the monograph of This is Saxer and the articles of Tully. 1 understandable: by their very nature they are ephemeral, or at least they were meant to be, and they have left behind few traces of their existence. Yet, evidence does exist, and the difficulty in undertaking any sort of meaningful analysis should not be a deterrent. Our focus here is vexillations from, and serving in, the Moesias (Inferior and Superior), including the northwest coast of the Black Sea and the Crimea. For the moment, we will leave out potential vexillation fortifications for the next chapter, when we turn to troop emplacement. We begin with an overview of vexillations before looking at their presence and activities in and beyond the Moesias.

3

Such an explanation can be found in any survey of the Roman army such as Goldsworthy (2007: 106) and Rankov (2007: 72-73). 4 This is a topic beyond the scope of this book, which deserves attention. 5 Saxer 1967: 118-119. 6 See, for example: Tac. Ann. 6.41, Tac. Ann. 15.10, Tac. Hist. 2.18; ; Suet. Vesp. 6; Joseph. BJ. 2.18.9, Joseph. BJ. 2.16.4. 7 Luttwak 1976: 124; Holder 1982: 40; Keppie 1998: 197; Le Bohec 2000: 30. 8 Saxer 1967: 1000, n. 6, 10, 12, 25, 36, 47; 2000, 12, 17, 26, 29. 9 Tully 2002: 133-134. 10 Le Bohec (2000: 30) does in fact allude to the same figure when he states that: “A provincial army could send to the scene of operations a whole legion and a detachment from each of its legions, or only send the equivalent of two or four cohorts per unit, i.e. 1,000 or 2,000 men each time.” 11 Tully 2002: 133. Indeed, the nature of vexillations is such that they were created from soldiers from many different legions and auxiliary units. Tully (2002: 134) also concludes that the cohorts of combat vexillations were probably organised just as a regular legionary cohort, that is, with six centuries of 80 men, or about 480 men at full strength.

General The vexillum, which was a military standard employed by both legionary and auxiliary units, was the standard that a vexillatio would rally around, and from where a vexillation drew its name. 2 These vexillations could be sent both to different parts of a province, or to different parts of the empire, and they could be used for a variety of different reasons, from combat assistance to administrative duties. The scholarly consensus is that these vexillations were used increasingly in the second 1 2

Saxer 1967; Tully 1998, 2002, 2004. Campbell 1996: 1594; Webster 1998: 138; Le Bohec 2000: 30.

50

Vexillations and the Black Sea Vexillations had been in use before Marcus Aurelius: the earliest account that we have of a vexillation comes from Velleius’ description of the Pannonian uprising from AD 6 to AD 9. 12 There are a few references to the use of vexillations that date to the first century; however, that need not mean that they were frequently used. Indeed, current thinking has it that vexillations only became widely used in the second century, with them well established by the reign of Marcus Aurelius. 13

“Thracia”, or Thrace. What we should restore in the square brackets, however, is anyone’s guess. Saxer, who adds a “XV” following the “TRACHIA”, believes that it should be restored as followed: “[NUMERORUM HONOR A LEG IV SCYTHICA A LEG V MACE]”. 15 Domaszewski on the other hand, shows a little more caution in believing that it should read, “[A LEG V MACE]”. 16 We know that both the VIII Augusta and V Macedonia were stationed in Moesia during the reign of Claudius, and as early as AD 46. The Fourth Scythian legion was also based in the province during that period, but we should hesitate to restore “A LEG IV SCYTHICA”, because we do not know how many letters are missing. The Fourth Scythian legion was in Moesia at the same time as the other two legions, but that need not mean that an expeditionary force was composed of men from all three legions, even if there is some precedent for this. Since, the Fifth Macedonian legion had departed by AD 58, the vexillation mentioned was active between AD 46 and AD 58. It is possible that this vexillation was sent to Thrace around the time of the creation of the province in AD 46.

There are dozens of inscriptions that identify vexillations in operation in the Moesias and the Black Sea. They also go by a number of different names: vexillatio XII Catafractariorum, vexillatio equitum scutariorum, vexillatio legionis V Macedonicae, vexillatio legionis I Italicae et legionis II Herculiae, vexillatio legionum I Italicae V Macedonicae XI Claudiae et cohortium, vexillatio legionis VII Claudiae, vexillatio legionis XI Claudiae, vexillatio legionum I Italicae V Macedonicae et VII ad Tropaeum Traoiani, vexillatio legionum V Macedonicae XI Claudiae, vexillatio legionum I Italicae XI Claudiae, vexillatio legionum I Italicae XI Claudiae classis Flavia Moesicae, vexillatio classis Ravennatis, vexillatio exercitus Moesiae inferioris, vexillatio exercitus Moesiae, vexillatio Moesiae inferioris, vexillatio Capidavensium, vexillatio Chersonissitanae, vexillatio Ponticis aput Scythia et Tauricam, vexillatio equitum Moesiae inferioris et Daciae, vexillatio expeditionis per Asiam et Lyciam Pamphyliam, vexillatio per Germaniam et Raetiam et Noricum et Pannoniam et Moesiam, vexillatio Egisseis Valerius (?). This remarkable variety in nomenclature is in part down to the ephemeral nature of these units, though it might also reflect the many different ways that these vexillations were used. Some vexillations were named for the units of which they were comprised (legions, auxiliary units, fleets), some their location (city, province, region), which might in part point towards the duration of their posting (if the location refers not to the location where they were drawn from and ratherto the location where they were posted), and expeditions on which they participated. As we will see, however, the presence of a form of the term vexillatio is not the only way to find the existence of a vexillation, although it is, by far, the most common way.

A decade or two later, we find more evidence of vexillations in the region. Suetonius says that: “Two thousand soldiers of the three legions that made up the army in Moesia had been sent to help Otho.” 17 There is also the case of Licinius Mucianus, whom we discussed in a previous chapter. Vespasian had sent this commander west from Berytus in the east. Mucianus marched westward with, “the Sixth legion and thirteen thousand soldiers from the vexillations were following [Mucianus] in a huge column.” 18 These troops were only temporarily stationed in Moesia. They engaged the Dacians in battle, and then moved on. Finally, mention should also be made of a diploma, found near Novae, which lists units from both Germania Superior and Moesia and which dates to September 20, AD 82. 19 It suggests a joint force, comprised of troops from both provinces, 20 was put together in light of some military necessity in the Whether the participating units were region. 21 necessarily vexillations is another matter, though the document at least provides evidence of troop transfers into Moesia, if on a short term basis. The movement of vexillations also went the other way, for the Danubian provinces were important as sources of reinforcements for eastern campaigns. 22 We have, in fact, an inscription of Trajanic date from Rome which alludes to the use of Moesian troops in his

The Lower Danube Our earliest evidence for detachments from the Moesian army comes as early as the reign of Claudius. There is an inscription from Castulo in Hispania Tarraconensis that says the following: “…PRAEF VEXILLARIORUM IN TRACHIA [ - - - ] DONICA A LEG VIII...”. 14 The “Trachia” is presumably

15

Saxer 1967: 9. Domaszewski 1981: 200-201. 17 Suet. Vesp. 6. trans. Rolfe. 18 Tac. Hist. 2.83. Cf. Joseph. BJ. 4.11.1, 5.1.6. 19 CIL XVI.28. 20 The diploma lists five alae and nine cohortes in Germania Superior, and one ala and two cohortes in Moesia (ala Claudia nova, cohors III Gallorum, cohors V Hispanorum). 21 Note Wheeler 2011: 209. 22 See, for example, Adams (1976: 16-17, 19, 21-23). 16

12

Vell. Pat. 2.110.6. Landelle 2015: 1066. 14 CIL II.3272. 13

51

Exercitus Moesiae Parthian campaign. 23 A certain Lucius Paconius Proculus is listed in the inscription. Proculus was a military tribune in the legio XI Claudia pia fidelis. He was also a prefect of the, “VEXILLATION(EM) EQ(UITUM) MOESIAE INFER(IORIS) ET DACIAE EUNTI IN EXPEDITIONE PARTHIC(A)”. We have another inscription from Berytus that names a Caius Valerius Rufus who was sent, “CUM VEXILLO AB IMP(ERATORE) NERVA TRAIANO OPTUMO AUG(USTO) GERM(ANICO) DACICO PARTH(ICO) CYPRUM IN EXPEDITIONUM…”. 24 This Rufus was also a military tribune of the legio VII Claudia, which was based in Moesia Superior during the reign of Trajan. Therefore, it is probably safe to assume that the vexillation referred to in this inscription was at least partially composed of the Seventh Claudian legion. 25

Dacia, Moesia Superior, Pannonia Superior, and Pannonia Inferior – in the minds of Roman commanders preparing for eastern campaigns, though a lack of evidence means that any such conclusions are speculative. 34 Still, some of the other vexillations mentioned by Saxer that served in the Parthian wars come from provinces other than those around the Danube. 35 Another important consideration, as regards vexillations sent eastwards, is the tendency of inscriptions listing vexillations from the Danube to be restricted to the wars of Trajan and Antoninus Pius, though this could just as well be an issue with the evidence, and the lack thereof. Their deployment for eastern wars aside, there are a few other inscriptions from the Moesias that bear on this discussion. One such inscription dates to around AD 155 and was found near Montana, 36 and it records a vexillation of the Eleventh Claudian legion. During the civil war which followed the murder of Commodus, we have evidence of another vexillation from the legio XI Claudia, and it comes in the form of two inscriptions, both of which are from Italy and list a certain Marcus Aquilius Felix. 37 This Aquilius Felix, who is attested at two different points in the Historia Augusta and who was sent to assassinate Septimius Severus, was the prefect of a vexillation active in Italy. 38 The abbreviations “AGENTIUM IN ITAL” in AE 1945, 80 allow us to identify the Marcus Aquilius Felix in both inscriptions with the Aquilius mentioned in the lives of Didius Julianus and Pescennius in the Aquilius Felix was also a Historia Augusta. 39 centurion of the legio XI Claudia; yet another clue that the two men are one in the same. 40 There is another inscription, which dates to the middle of the second century, from Montana that lists a vexillation. 41 In this instance, the vexillation is much more varied in its composition, styled as it is as a vexillatio legionum I Italicae XI Claudiae classis Flaviae Moesicae. Its duties, collecting animals for imperial hunts, “OB VENATIONEM CAESARIANAM…URSIS ET VISONTIBUS”, also attest to the variety of tasks to which these units were assigned.

We find evidence for the movement of Moesian troops to the east, particularly for Trajan’s Parthian War, in the diplomas as well. 26 In chapter two we looked at the cohors III Augusta Nerviana Brittonum, which appears in a diploma dated to AD 112. As noted, Eck and Pangerl argued that the unit was sent east – not, strictly speaking, a vexillation then – and quite possibly destroyed in the process. 27 From that same article, Eck and Pangerl discuss a diploma dated to August of 115 that is specifically concerned with Trajan’s Parthian War, 28 a diploma for which an earlier fragment had been published a few years earlier. 29 In this instance, however, as in the previous one, the indication is that entire units were dispatched to the east, rather than detachments from those units. 30 In other words, vexillations were not involved. Some form of the term vexillatio does not appear; rather, we find “TRANSLATIS IN EXPEDITIONE”, 31 a phrase which also appears on the earlier diploma. 32 Unfortunately, we have no further evidence from the Moesias concerning the use of vexillations that were sent east. We do, however, have a handful from the Pannonias and Dacia. 33 Although a discussion of the vexillations from the Pannonias and Dacia sent to the east is outside of the scope of this book, nevertheless, a cursory glance at the evidence for those provinces from Saxer’s monograph suggests no discernible preference for any of those five provinces – Moesia Inferior,

There is an inscription in Saxer’s monograph that poses a few questions and merits some detailed discussion. Saxer has dated this inscription, n. 266 in his corpus, to

23

CIL VI.32933. Saxer (1967: 26) rightly dates this inscription to AD 114-117, the dates of Trajan’s Parthian war. 24 ILS 9491. 25 The full text of the inscription is as follows: C(aio) Valerio T(iti) f(ilio) Fab(ia) Rufo – praef(ecto) coh(ortis) VI praetor(iae) tr(ibuno) mil(itum) leg(ionis) VII Cl(audiae) p(iae) f(idelis) misso cum vexillo ab imp(eratore) Nerva Traiano optumo Aug(usto) Germ(anico) Dacico Parth(ico) Cyprum in expeditionem - . 26 See Bennett 2010. 27 Eck and Pangerl 2008: 358. 28 Eck and Pangerl 2008: 363-370. 29 Eck and Pangerl 2005. The new fragments presented in this more recent article expand upon that earlier discussion. 30 Eck and Pangerl 2008: 367. 31 Eck and Pangerl 2008: 370. 32 Eck and Pangerl 2005: 51. 33 See Saxer 1967, nos. 48, 49, and 52.

34

In other words, troops from Lower Pannonia were not necessarily chosen over those from the four other provinces listed. 35 See Saxer 1967: 25-35. 36 CIL III.7449; Saxer 1967: 89-90. 37 CIL X.6657 is from Antium in Italy; and AE 1945, 80 is from Cannes in Italy. 38 The two references from the Historia Augusta, which are nearly identical, hail from the life of Didius Julianus (5.8) and the life of Pescennius Niger (2.6). 39 See Saxer 1967: 43. Cf. Pflaum 1960/1961: II.598. 40 HA. Did. Iul. 5.8: “In addition he sent the centurion Aquilius, known as the killer of senators, to kill Severus.” HA Pesc. Nig. 2.6: “Finally he had sent even Aquilius the centurion known as the killer of commanders, as if so great an emperor could be killed by a centurion.” 41 AE 1987, 867=AE 1999, 1327=AE 2001, 120=AE 2011, 1120.

52

Vexillations and the Black Sea AD 167/168. 42 It reads “FALCO[NI - - - - - - ] VEXIL[- - -] II[- - - - - - ] SUB CURA [- - -] LEG V MA[CEDONICAE]”. Although it is unclear what follows the “II”, the “SUB CURA … LEG V MA”, suggests that the vexillation was from Moesia Inferior. 43 It might be that the “II” is part of “IIII”, and thus the Fourth Flavian legion which was based in Moesia Superior during the start of Marcus Aurelius’ reign. The problem is the “LEG” or “L” signifying legio usually precedes both the number of the legion and its cognomen. Ignoring a mistake on the part of the lapicide, there is no room for the “LEG” or “L” before the “II”. It would be highly unusual for the stone cutter to have inscribed IIII LEG FF. For the sake of convenience, it would also be much easier to send vexillations from the same province, or neighbouring Dacia for that matter, rather than to have summoned a vexillation from a legion considerably further west in Moesia Superior. Based on the information provided in inscription 266 from his corpus, Saxer conjectures that we should read, “VEXIL[ARII] II[I LEGIONUM?]”. 44 The problem with this reconstruction is that, unlike the diplomas that routinely list the number of auxiliary units in a given province, we have no other evidence for vexillation inscriptions that do this, assuming that is what we have here. While keeping in mind that there does not seem to be an official format, and that inscriptions were prone to the preferences of the individual stonecutter and his or her patron, Saxer’s restoration is questionable. The inscription must remain unresolved unless additional evidence comes to light.

(scutariorii), were later, late antique, creations, 51 for while there is so little evidence for the former unit in the region and none for the later, there are plenty of references to these sorts of units in the Notitia Indeed, although we know that Dignitatum. 52 cataphracts were part of the Roman military in this period and at least one is attested in Moesia Inferior as the ala Gallorum et Pannoniorum catafracta, 53 units of cataphracts are much more common from the third century onwards. 54 Obviously we could not say for certain whether the vexillatio Capidavensium was an earlier imperial or later imperial vexillation. The fact that this vexillation was named after a place might suggest a degree of permanence – the unit was based in the town for some time. That, in turn, might point towards a later date for the vexillatio Capidavensium. As it happens, there was an equestrian unit based at Capidava in late antiquity, but it was not a vexillatio; rather, it was a cuneus equitum Solensium. 55 In another inscription, from Numidia that dates to around AD 177, we read of a M. Valerius Maximianus, a procurator of Moesia Inferior who, at the same time (eodem in tempore), was in charge of some vexillations, presumably from Moesia Inferior, in some operations in Macedonia and Thrace. 56 There are a handful of inscriptions that point to the existence of vexillations with some sort of Moesian connection and which date to before the Tetrarchy. There is evidence for a vexillation from the legio VII Claudia from an inscription, 57 found at Lederata in Moesia Superior, which dates to the early third century. We have a record of a vexillatio legionis I Italicae et legionis XI Claudiae Piae Fidelis in Moesia Inferior that might also be late second century or later in date. 58 Finally, there is also the vague reference to vexillationes – no additional names given – in an inscription from Salsovia in Moesia Inferior that might date to the second century. 59 In sum, there is some evidence for the vexillations comprised of Moesian units, though we are less certain about their size or their duties, or even where they were operating. Indeed, to really see Moesian vexillations in action we need to turn to the northwest coast of the Black Sea just beyond the borders of Moesia Inferior.

There are a few other indications of the activity of vexillations from the Moesias, Black Sea region excluded. Two inscriptions from Histria record a vexillatio XII Catafractariorum, 45 a unit that has left little evidence for its presence in the Moesias. We also find two other inscriptions, one from Carsium, 46 the other from Garno Harsovo, 47 that simply record: “DE NOMERO (!) CATAFRACTARIORUM”. These are, along with the aforementioned ala I Gallorum et Pannoniorum discussed in chapter two above, our lone examples of cataphracts in the Moesias, though we also know of at least one unit referred to on a Dacian diploma. 48 There is evidence for a vexillatio equitum scutariorum at Capidava, or at least a veteran. 49 The mysterious vexillatio Capidavensium, obviously from the same site, might allude to the same detachment – and its relative permanence. 50 It seems likely that many of these units, the cataphracts and shieldmen

THE BLACK SEA The previous section essentially sets up our discussion here, which will concentrate on the presence of Roman 51

Zahariade 2007: 10. Note, for instance, the cuneus equitum scutariorum under the Dux Scythiae listed Not. dign. or. 39.12, and the cuneus equitum catafractorium also under the Dux Scythiae listed at Not. dign. or. 39.16. 53 See CIL XI.5632 and XVI.110. 54 Eadie 1967; Whately 2012. 55 Not. dign.or. 39.2. 56 AE 1956, 124. 57 CIL III.1643 = 8099. 58 AE 1988, 9992=AE 1989, 640=AE 2003, 1550. 59 IScM V.290b=IIFDR 271.

42

52

AE 1934, 112. 43 It could also have been from Dacia. The interpretation depends upon what year we accept both for this inscription and what year we accept for the legio V Macedonica’s transfer to Dacia. 44 Saxer 1967: 90, n. 266. 45 AE 1919, 18; IIFDR 110. 46 AE 1912, 192. 47 IBulgarien 52. 48 CIL XVI.110. 49 AE 1976, 634=IIFDR 221. 50 AE 1935, 171=IIFDR 220; AE 2004, 1278.

53

Exercitus Moesiae soldiers on the northwestern shore of the Black Sea and the Crimea, not untrodden ground. And indeed, we should not be surprised to discover that we have far more evidence for vexillations operating outside of the Moesias than we do for those inside it. Although vexillations will ostensibly remain a prime focus of our analysis, a separate issue arises: should we consider the units active around the Black Sea region vexillations, at least in the conventional sense, or something else? As we just saw, vexillations tended to be detachments employed for a particular task for a specific period of time, at least in the first few decades of the imperial era. Does the term apply if we find evidence for their deployment on a more permanent, or at least regular, basis? And what does this tell us about the nature of Roman administrative boundaries and military commands in the lower Danube region? Thus, although what on the surface might seem to be a peripheral issue to the topic under discussion in this book, it has wider implications not only for the book as a whole, but for our understanding of the Roman military and administration in the wider world as a whole.

is an additional point. Chersonesus had originally been a Greek colony, and for some it maintained, albeit with difficulty, 63 its independence in the Crimea throughout the first three centuries AD. 64 It might be apocryphal to say this, but to my mind the evidence for the incorporation of Olbia and Tyras as well as the western Crimea into the province of Moesia Inferior is insufficient to make any definitive claims. As we will see, however, there undoubtedly was a military presence in most of these places. And, while it might originally have been temporary, it seems to have become permanent. Given too that vexillations in the region were drawn from Lower Moesia, whether the region was officially considered part of the province or not, the Roman government based in Moesia Inferior certainly had some authority over the region.

Moesian Land-Based Soldiers on the Black Sea The First Century There is a not inconsiderable body of evidence for a Roman military presence in the Crimea and along the northwestern shore of the Black Sea area. Much of the evidence is epigraphic, some is material, and some is even textual. Beginning with the latter, M. Julius Agrippa II, in Josephus’ Jewish War to dissuade the Jews from further war, says: Do I have to speak of the Heniochi and the Colchians and the race of the Taurians, Bosporans and the nation dwelling around the Pontus and the Maeotis? According to them, before they knew no native master, but now [they] are placed under 3000 armed-soldiers (΄οπλίταις), and forty long ships keep the peace before the unapproachable and furious sea. 65 It is unclear from that statement alone whence the troop allocation came. 66 Although Josephus is referring to, at least in part, the Chersonesus, he also refers to the Colchians and the Heniochi. Saxer suggests that some of the three thousand soldiers were in fact from Cappadocia. 67 He also holds that the region was garrisoned by troops from the Lower Moesian legions. 68 To determine this it would help to locate the “groups” listed by Josephus. The Taurians might be the people who inhabit the Crimea, and the same could

Our evidence for a Roman military presence on the northwest Black Sea coast comes from a combination of the archaeological evidence (the remains of fortifications and assorted materials) and inscriptions, with, as noted above, vexillations comprising a significant part of that total; how significant is open to interpretation. The primary sites that we will consider include Tyras, Olbia, Chersonesus, and Charax, and we will discuss them in that order moving geographically from west to east. We will begin with our evidence for the presence of the Roman military that does not explicitly indicate the use of vexillations. Then we will look at the evidence for vexillations. First, a few comments on the boundaries of Moesia and Roman provincial administration more generally, as there is some debate over whether the north coast of the Black Sea and the Crimea were ever incorporated into Roman territory, and much of the debate surrounds the Crimea, and in particular the area around Chersonesus, less so the area around Tyras and Olbia. Most scholars argue that at least the eastern portion of the Crimea remained under the control of the Bosporan Kingdom round Panticapaeum throughout the period under consideration here, regardless of the possible presence of the odd Roman soldier. 60 At the same time, some believe that when the region around Tyras and Olbia became incorporated into the province of Moesia Inferior, 61 so was Chersonesus, even if this was a later development (late second century AD). 62 There

Inferior, it is not hard to see why the Romans may have wanted some sort of association with the region. It was known for its fertility in antiquity, and would have given the Romans another base for their Pontic fleet. Thus, for both economic and strategic reasons it would have been beneficial. In fact, the region remained important to Rome, and in particular its later capital of Constantinople, well into Late Antiquity. 63 See Plin. HN 4.84; Fr. Gr. Hist. 257; IOSPE2 I.362. 64 Haensch 2009: 211. 65 Joseph. BJ 2.16.4 66 See Sarnowski 1988: 63. 67 Saxer 1967: 91. Saxer calls these troops legionaries, though this is far from certain. 68 Saxer 1967: 91ff.

60

Braund 1996a: 254. Rice (2001: 289), like Braund, also believes that the Romans garrisoned the Chersonesus. She does not, however, tells us whether she thinks that the western portion of the Crimea was a client kingdom garrisoned by Roman troops. 61 Wheeler (2011: 213) argues it took place in the wake of Trajan’s organisation of the newly conquered province of Dacia. 62 Zahariade and Gudea 1997: 29. If we look at the Crimea in isolation, regardless of its relationship with the rest of Moesia

54

Vexillations and the Black Sea be said for the Bosporans. 69 The other two peoples, and in particular the Heniochi, are from the southeastern coast of the Black Sea. 70 Saxer suggested that the force quoted by Josephus was in part supplied by the forces of Cappadocia, or some force based in northeastern Asia Minor, a view largely followed by Wheeler, who argues for Pontus and Colchis as the bases for these troops. 71 Indeed, Wheeler suggested that Josephus’ 3000 hoplites could be explained in two other ways: first, as part of Plautius Silvanus’ relief of a Scythian siege of Chersonesus, 72 and secondly as part of Nero’s planned Caucasian expedition. 73 But, he doubts that Plautius went to the Crimea in the 60s, the approximate timeframe of this siege, and discounts the presence of a Moesian vexillation. 74

such an invasion was in the works, it seems that the planned invasion route would have passed through Anatolia to the south, and not along the north shore of the Black Sea. 85 A few years later, during the reign of Vespasian, there was an expedition to the south into Iberia (in the Caucasus) led by one M. Hirrius Fronto Neratius. 86 The impression, then, is that the southern coast of the Black Sea was the scene of a considerable amount of military activity from the reign of Nero to Vespasian. So where did the Roman component come from? Tacitus refers to an uprising in Pontus that occurred at some point after the kingdom became a province during the reign of Vespasian. 87 This uprising was led by Anicetus, a freedman of Polemo II. 88 At the head of a band (manus), Anicetus attacked Trapezus and massacred a cohort, which had been supplied by the king (Polemo II). 89 To anticipate some of our discussion of navies below, Tacitus also tells us that the sea was unpatrolled because the fleet had been at Byzantium at the time. Pliny referred to a number of castella along the southern and eastern shore of the Black Sea, 90 although he did not detail the size of their garrisons or who supplied them. Braund believes that they were garrisoned by Roman troops. 91 At the time that Josephus and Pliny were writing, Pontus and Cappadocia were two separate entities. 92 What is more, from a military standpoint, the Cappadocian component of the eastern frontier did not begin to become developed until the reign of Vespasian. 93 Thus, it seems that garrisons of the south and eastern shores of the Black Sea were probably not significant in size. There were no legions based in the region when Josephus was writing, or later when Arrian

Josephus aside, there is additional evidence for Roman military activity along both its northern and southern shores of the Black Sea. As we noted in chapter one above, Didius Gallus and some Roman forces seem to have been engaged in a conflict with King Mithridates of the Bosporan kingdom. 75 Some believe that this force was comprised of troops from Moesia, others that it was troops from Asia Minor. 76 There is debate too over what kind of troops were involved, legionaries or auxiliaries. 77 Ultimately, Sarnowski’s suggestion that the forces involved were likely mixed (Moesian and eastern) is sensible. 78 Jumping ahead a couple of decades, we know, for instance, that Ptolemo II’s kingdom of Pontus was annexed by Nero in AD 64. 79 Furthermore, Rice, following the inscription of Plautius Silvanus, notes that at some point during the reign of Nero, perhaps around AD 62, 80 the Scyths besieged Chersonesus, 81 the defence of which was, in Wheeler’s eyes, the responsibility of the king of the Bosporan kingdom, Cotys I. 82 Rice argues that Nero desired to annex the Crimean kingdom so that it might be used as a bridgehead for an invasion of the Caucasus. 83 At the same time, she admits that this course of action – the plan to annex the Chersonesus – was not followed by any of Nero’s successors. 84 What is more, assuming

85

See Braund 1994: 175-176. Cf. Suet. Ner. 18-19, Plin. HN 6.159. Mitchell 1993: 119. 87 Tac. Hist. 3.47. 88 Tac. Hist. 3.47. 89 Tac. Hist. 3.47 (caesa ibi cohors, regium auxilium olim). 90 Plin. HN 6.4: …Tripolis castellum et fluvius, item Philocalia et sine fluvio item Liviopolis…flumen Absarrum cum castello cognomina…castellum Sebastopolis a Phaside… The first three castella are fairly close together. Pliny, perhaps significantly, does not specify the distance, which he does frequently before and after listing these castella. Absarrus (or Apsaros) is 240 Roman miles from Liviopolis, while Sebastopolis is another 100 Roman miles from the city of Phasis. Pliny does not specify the distance from Phasis to Absarrus, or even Trapezus. The first three fortress sites lie roughly within the ancient boundaries of Pontus, whereas the other two, Absarrus and Sebastopolis, lie roughly within the boundaries of ancient Colchis. Arrian (Arr. Peripl. M. Eux 6.1-2) tells us that in his day, perhaps around AD 132, Apsaros (Absarrus) had five cohorts (πέντε σπεῖραί) stationed there. Arrian (Arr. Peripl. M. Eux 9.3) also tells us that “400 select troops (τετρακόσιοι στρατιῶται ἐπίλεκτοι)” were quartered at Phasis. In regard to Sebastopolis, Arrian (Arr. Peripl. M. Eux 10.3) does not specify the type of troops that are there, nor does he give us any indication of their numbers. 91 Braund 1994: 178. 92 Broughton and Spawforth 1996: 288. It was not until the reign of Trajan that the two provinces were united. 93 Broughton and Spawforth 1996: 289; Mitchell 1993: 118-119. Cf. Suet. Vesp. 8.4: Cappadociae propter adsiduos barbarorum incursus legiones addidit consularemque rectorem imposuit pro eq. R.. “Because of the persistent incursions of barbarians he added legions to Cappadocia and placed it under a consular governor instead of a Roman knight.” 86

69

This assumption is based on the names of the two tribes. See The Barrington Atlas of the Greek and Roman World, and map 87 in particular for the location of the Heniochi and Colchians. 71 Saxer 1967: 91; Wheeler 2012b: 137. 72 ILS 986: Scytharum quoque regem a Cherronensi quae est ultra Borustenen opsidione summto. 73 Wheeler 2012b: 135. 74 Wheeler 2012b: 136-137. Contra see Zubar 2001: 734-735. 75 Tac. Ann. 12.15; CIL XI.6163; ILS 9197; Sarnowski 1988: 61-67. 76 Sarnowski 1988: 64. 77 Sarnowski 1988: 63-65. 78 Sarnowski 1988: 65. 79 Braund 1994: 175. 80 Rice 1981: 290; ILS 986. 81 For an overview of the Scyths in classical antiquity see Batty (2007: 204-214). 82 Wheeler 2012b: 135. 83 Rice 1981: 290. Batty (2007: 432), who alludes to these desires, notes too that Roman troops were never stationed there permanently. For a comprehensive discussion of Nero’s plans in relation to regional troop movements, see Wheeler (2012: 137-141). 84 Rice 1981: 290. 70

55

Exercitus Moesiae was. 94 The only known legionaries in the region were members of a vexillation of the legio XV Apollinaris, who were there at least from the reign of Hadrian. 95 The legio XV Apollinaris was based at Satala, which is between 200km and 300km north of the Euphrates frontier in eastern Anatolia; 96 furthermore, the legion did not arrive at Satala before the reign of Trajan. 97 Thus, the troops mentioned by Josephus that were based around Colchis were not from Cappadocia. 98 The troops in the Caucasus could have been from mixed legionary and/or auxiliary vexillations; they also may have been provided by allied kingdoms. 99

involved. 104 And given what we argued about the nature of the 3,000 hoplite of Josephus discussed above, those troops should not be connected with whatever military actions are mentioned. Ultimately, Sarnowski doubts any Roman troops were operating in the sort of capacity alluded to in Silvanus’ inscription in the Crimea before the second century. Indeed, he notes that the Moesian fleet likely did not start transporting troops to the Crimea before that time. 105 Zubar, however, argues that the Ravennate fleet was responsible for ferrying troops to the Crimea at this time. 106 He also notes the plethora of views about the nature of Roman involvement, and tends to think that there was some sort of Roman military operation (i.e., not simply a diplomatic one), even if we do not know its size, character, or scope, aside from the allusions to the Ravenna fleet. 107 The evidence, however, is thin, and the reality is that we have to wait until the second century before we find clear evidence of Roman military activity on the northwest shore of the Black Sea.

Returning to the northwest Black Sea and the Crimea, there was no legion based in the Chersonesus at this point, and so the troops that made up the aforementioned vexillation found in Josephus, regardless of whether they were auxiliary or legionary, did not hail from the northwest coast of the Black Sea. Ultimately, a significant percentage of Josephus’ τρισχιλίοις ὁπλίταις must have been based along the southern and eastern shore of the Black Sea, and those troops were likely a combination of auxiliary Roman troops, and troops supplied by Rome’s allies.

Tyras and Olbia Shifting to the second century, we lack evidence for a Roman military presence in Tyras, Olbia, or the Crimea before at least the reign of Trajan, if not Hadrian. 108 Beginning with Tyras, the first datable inscription, to AD 115, was erected, “PER VEXIL(LATIONEM) L[EG(IONIS) V MAC(EDONICAE)]”. 109 Another inscription lists the legio I Italica, the legio V Macedonica, and the legio XI Claudia, along with an unknown unit of auxiliaries: it too may date to c. 115. 110 We have an inscription, found at Tyras, that records a handful of men who seem to have served in the legio V Macedonica, and which might date to AD 116-117. 111 A fourth inscription records a vexillation composed of those same legions minus the auxiliaries. 112 Then there are some inscriptions that likely date to the second century, though when, exactly, we cannot say. 113 There are some brick stamps, for instance, which record “LE(GIO) V M(ACEDONICA”. 114 Although we do not know when the detachment from the Fifth Macedonian Legion arrived, 115 it seems likely that they had left by AD 167 or so in light of the legion’s transfer from Moesia Inferior to Dacia between 167

In chapter one above, we discussed, in connection with first century military operations and legionary transfers, the famed career inscription from Tibur of Tiberius Plautius Silvanus. 100 It is worth looking back at that inscription because, at least on the surface, it would seem to offer us some insight into Roman military operations in the Crimea during the first century. On that rather lengthy inscription we learn that, in addition to his remarkable – and fantastical – transfer of 100,000 people across the Danube, 101 Tiberius expelled the king of the Scythians from the Borysthenic Chersonesus (Scytharum quoque regem a Cherronensi, quae est ultra Borystenen, opsidione summoto). Much debate has centred on identifying the conflict referred to in that passage, and how, exactly, the Roman military fits into this situation. 102 Following Sarnowski, there are a few Greek inscriptions that hint at military action as well as a Roman presence that might reflect Silvanus’ claims, 103 though they are undated and fragmentary. On the other hand, there is evidence for some destruction on the western shore of the Crimea in the first century, though we do not know whether Roman troops were

104

Sarnowski 2006: 3. Sarnowski 2006: 7-8. Cf. Wilkes 2005: 154; Wheeler 2012b: 123-127. 106 Zubar 2001: 734. 107 Zubar 2001: 735. 108 Wheeler 2012b: 137. 109 AE 1990, 868. Cf. Sarnowski 1988: 71-72. 110 Saxer 1967: 90, n. 266. 111 Sarnowski 1988: 72. 112 AE 1925, 78. 113 Our focus, of course, is on inscriptions with clear military content, which means we are excluding those that are more ambiguous. Note, for instance, the additional inscriptions collected by Sarnowski (1988: 74), and not included in this discussion. 114 AE 1925, 77. 115 Zubar 2001: 737-739.

94

105

See Braund 1994: 181ff. 95 Braund 1994: 198 96 For a survey of the frontier of Anatolia see Mitchell (1993: 118142). 97 The date of its departure from Carnuntum to Satala is in fact a controversial point; it was there by AD 135 (Wheeler 2000: 259308). Wheeler convincingly argues that there is no evidence for the presence of the legio XV Apollinaris in Pannonia after AD 106. 98 Contra Saxer 1967: 91. 99 See Speidel (1985: 97-102) on the role of Bithynian troops in the Bosporan Kingdom. 100 CIL XIV.3608=ILS 986. 101 Note Batty’s (2007: 407) comments. 102 Sarnowski 2006: 3. 103 IOSPE2 I.355, 369, 420.

56

Vexillations and the Black Sea and 195. 116 It might have left long before that, however. There is another inscription, which dates to the end of the second century or beginning of the third, that records a centurion of the legio I Italica, an unnamed vexillation, and a medicus of the Moesian fleet (classis Flaviae Moesica). 117

doubted. 126 While there is every reason to doubt the veracity of the Historia Augusta’s claim, given what we know about troop movements into the area it is not wholly unreasonable. Indeed, several inscriptions survive that date to the second half of the second century and which point to a possible Roman military presence of some sort. Sarnowski listed an inscription that allegedly records a vexillation of the legiones I Italica, V Macedonica, and XI Claudia, as well as an as yet undiscussed cohors VI Asturum or Maurorum. 127 A more recent reading, however, sees the inscription as listing not an as yet unidentified cohort, but six undisclosed cohorts under a certain Malicus. 128 There is an inscription, set up by a Galerius Montanus for his mother, which likely dates to the second half of the second century and records an armatura of the legio XI Claudiae. 129 There are two third century inscriptions: one records a deceased soldier from the legio I Italica, 130 while the other lists a soldier of the cohors I Cilicum Decianae. 131 Indeed, the third century emperors Philip and Decius both seem to have been embroiled in conflicts that affected Lower Moesia and Thrace. 132 While they both date to some point after the period under consideration here – the inscription naming the cohort date to the middle of the third century because of the “DECIANAE”, which refers to the Emperor Decius (r. 249-251) – the units mentioned might have been in the region for some time. In the end, however, the evidence for military units in Olbia, like Tyras, is slim, though we need not doubt some sort of Roman military presence in the region. The units that have left some evidence of activity include the legiones I Italica, V Macedonica, and the XI Claudia. 133

We have a career inscription, which dates to sometime between the Marcomannic wars and the reign of Commodus, 118 found in Africa proconsularis, which seems to refer to a northwest Black Sea coast vexillation. 119 This inscription, which comes from Mactar in Numidia, lists the various offices of a Tiberius Plautius Felix Ferruntianus who was a praepositus of a vexillation of Pontus near Scythia and It reads: “PRAEPOSITUS Taurica. 120 VEXILLATIONIBUS PONTICIS APUT SCYTHIA ET TAURICAM”. Taurica is frequently found in conjunction with the Chersonesus of the Crimea in order to distinguish it from the Thracian Chersonesus. As we saw above, Scythia in this period is generally the eastern portion of Moesia Inferior along the Black Sea that also runs northwards into the Dobrudja and beyond, and perhaps as far as the Don. 121 Thus, we might assume that when “APUT SCYTHIA ET TAURICAM” was inscribed for Felix Ferruntianus, what we should understand is the region around Tyras and Olbia, both cities along the northwest coast of the Black Sea, which of course is the Pontus referred to in the inscription. 122 Moving from Tyras to Olbia in the second century and beyond, Zubar notes that Dio Chysostom does not mention any Roman military presence in the city. 123 In AD 106-111, however, we get our first hint of encroaching Roman forces, an inscribed grave of a certain Athenocles points to the presence of an auxiliary unit. 124 The consensus, however, is that military activity in Olbia was restricted in the years from AD 100-130.

Chersonesus and the Crimea We now move a bit further east, and in particular the sites in the western Crimea, such as Chersonesus, Charax, and the modern Balaklava (ancient Sybmolon). There is much more evidence of Roman military activity in this part of the region than there is for Tyras and Olbia, and our first strong evidence for the presence of detachments from the Roman military dates to the first half of the second century. 134 Three inscriptions, dating from 139-161 and found in modern Balaklava, provide evidence of Roman soldiers in the Crimea. One inscription lists a centurion of the legio I Italicae. 135 The second inscription records a military tribune of a vexillation of an army, “EXERC(ITUS)”,

We do not know when a garrison might have been established in the city, though the author of the Historia Augusta claims that Antoninus Pius sent troops (auxilia) to Olbia to help the residents out against the Tauroscythians, 125 a claim that some have

116

See chapter three above. AE 1995, 1350. 118 Haensch (2009: 217) argues that its terminus ante quem is AD 176. 119 CIL VIII.619. Cf. Saxer 1967: 42-43. Tiberius Plautius Felix Ferruntianus was also a praepositus of the Third Augustan legion among the Marcomanni (APUT MARCOMMAN(N)OS), or rather, during the Marcomannic war. 120 For a detailed discussion of this vexillation and its commander Plautius Ferruntianus see Sarnowski (1988: 91). 121 Braund 1996b: 1374. 122 Sarnowski (1988: 75) also discusses the Greek inscription, from the third century, found on a shieldboss from Dura Europos, which refers to some Black Sea locales, such as Tyras and Chersonesus. 123 Zubar 2007: 174. 124 Zubar 2007: 174. 125 HA Ant. Pius 9.9. See Sarnowski 1988: 76. 117

126

Zubar 2007: 175. IOSPE I2.322. The inscription reads: “C(OHORTIS) VI M(ILLIARIAE) A(STURUM?)/MA(URORUM?)”. 128 AE 1995, 1348. “C(OHORTIUM) VI MA[3]LICO”. 129 AE 1909, 16. 130 AE 2002, 1252. 131 AE 2004, 1289. 132 Dexippus 18; Zos. 1.20.1; Jord. Get. 16.90-93; Zubar 2007: 177. 133 See Sarnowski 1995b. 134 See Zubar 2001: 736-738. 135 AE 1998, 1154. 127

57

Exercitus Moesiae who is unidentified. 136 The editor of the inscription suggests that we should read some sort of abbreviation of Moesia Inferior following the exercitus. The third also records a vexillation, presumably too of the exercitus Moesiae inferioris, and it mentions a military tribune, seemingly of the legio I Italica. 137 It seems likely, then, that troops from the legio I Italica were active in the region by the AD 130s. There are a number of other inscriptions from Balaklava that list military units, though we cannot date them with the same precision as these other three. Two inscriptions list centurions of that same legion and the XI Claudia. 138 The latter, at least, dates to the late second/early third century. There are additional references too to the aforementioned vexillatio exercitus Moesiae inferioris, 139 though also, simply, a vexillatio exercitus Moesiae, 140 which come by way of a number of brick stamps. Of those 71 stamps, some simply read VEX, 141 while others LEXICL (legio XI Claudia). 142

was found in the modern town of Kadikoi in Moesia Inferior. 148 The inscription records a vexillation of the First Italian legion. Besides the references to vexillations of specific legions, there are a couple of named vexillations that include a form of Chersonesus as part of their title. The next inscription from the Crimea probably dates to the reign of Commodus, and, more precisely, the years AD 185/186. 149 This lengthy inscription does not provide us with any indication of the units from which the vexillation may be drawing. Instead, it records a “VEXILLATIONE CHERSONESSITANA”, much like a later inscription, which dates to around AD 250. 150 That later inscription also names a centurion of the legio I Italicae [Decianae]. One last point: we have two inscriptions that allude to the presence of vexillations of the legio I Italica 151 and the legio XI Claudia 152 at Charax. Our discussion to this point has concentrated on the evidence for legionaries along the northwestern shore of the Black Sea. It remains to discuss the evidence for auxiliaries, for which there is far less information. For instance, we have limited evidence for auxiliaries operating before the second century. Spaul suggests that the fourth Cyprian cohort of Roman citizens was raised in a levy made by Claudius, and that it was based in the Chersonesus and served in support of the Bosporan kingdom. 153 It seems to have been in Moesia Superior only briefly, because it turns up in a Dacian diploma dated to February 17, AD 110. 154 It is unattested after that point, though it may have returned to the Crimea. 155 There is a little more evidence for auxiliary units operating in the Crimea as we near the end of the second century. Beneš suggested that towards the end of the second century the cohors I Thracum Syriaca, or a vexillation of the unit, served in the Chersonesus. 156 Ivantchik argued that the Thracian cohort was active in the Bosporan Kingdom on the basis of an inscription found at Tanais in the delta of the River Don. 157 The inscription does not, however, prove that a unit was based in that location; it only suggests that a Roman soldier was buried there, and he might have moved after his term of service came to an end. 158 On the other hand, it does point towards a

Besides Balaklava, we have considerable evidence for a Roman military presence in Chersonesus. 143 Scores of roof tiles stamped with LEGMNC, LEGM, and LEGMNCV have been recovered at Chersonesus that might date to the 150s – and no later than the late 160s (when the legion moved from Moesia Inferior to Dacia). 144 There are a number of inscriptions, dating from the end of the second century, or beginning of the third, which attest to a Roman military presence in Chersonesus. We have an inscription that records a vexillation of the legio I Italica. 145 Regrettably, we cannot date this inscription precisely, though it might well date to at least the second century, and perhaps the second half at that, based on our knowledge of the wider context, among other things. One inscription names both the Eleventh Claudian legion, with the epithet Severianae, as well as a vexillation. 146 We have an inscription from Chersonesus, which dates to some point between AD 222 and AD 234, which records a vexillation of the legio XI Claudia Severiana. 147 There is one last inscription to mention that names a legion from the Black Sea region, which may date to sometime during the Marcomannic wars, and which 136

AE 1998, 1155. AE 1998, 1156. 138 AE 1998, 1158, 1161. 139 AE 1998, 1163a. 140 AE 1998, 1163b. 141 AE 1998, 1163c. 142 AE 1998, 1163d. 143 Zubar (2001: 736-739) argues that there were two stages to the Roman military presence: the first ended in the 130s (the first troops appearing at the end of the first/beginning of the second century AD); the second after 138 – and even later than the 140s – after a gap of a decade or two. For Zubar, the presence of military forces is to be connected with Chersonesus’ attempts at obtaining the designation eleutheria on two occasions. While I do not dispute the desire for such a status (note Haensch 2009, for instance), I am wary about necessarily connecting it with a transfer of troops. 144 For the abundant roof tiles see Zubar (2001: 140). For the transfer of the legion, see chapter three above. 145 AE 1984, 805. 146 AE 2000, 1274. 147 AE 2000, 1274. 137

148

CIL III.14433. CIL III.13750. See Saxer 1967: 91-92. 150 AE 1996, 1358. 151 CIL III.14215-4=AE 1900, 200=AE 1903, 3. See Sarnowski and Zubar 1996: 234. 152 CBI 661; AE 1997, 1332. For the latter inscription see Sarnowski and Zubar (1996: 230). 153 Spaul 2000: 389. Though note Wheeler’s (2012: 144-145) sage comments. 154 CIL XVI.57. 155 Wagner 1938: 127-128; Spaul 2000: 389. 156 Beneš 1978: 53. 157 Ivantchik 2014. Ivantchik connects this with the transfer of the legio I Italica in the late second century. 158 Note, however, that Ivantchik (2014: 175) recognises that the inscription need not be proof that cohort was based in the town. On the other hand, Ivantchik’s case for associating the Thracians named on the inscription with the cohort is convincing. 149

58

Vexillations and the Black Sea noticeable presence in the kingdom, whatever form it might have taken, 159 and it is worth noting that Ivantchik dates the inscription to around AD 230, which is at the end of the period that we are discussing in this book. 160 There is also some speculation that all or some part (a vexillation) of the cohors I Bracaraugustanorum or the cohors I Bracarum was stationed in the Crimea during the middle of the second century. 161 A soldier from the cohors I Cilicum sagittaria left his mark at Chersonesus at some point at the end of the second century or beginning of the third century. 162 The same is true of a veteran of the cohors I Claudia Sugambrorum veterana, who was in the Chersonesus at the end of the second century or beginning of the third century. 163 We also have evidence for the cohors II Lucensium, which, depending on how we date the epigraphic evidence, might have been in the Crimea as early as AD 136. 164 Taken together, this is limited evidence indeed for Roman auxiliary military activity in the region, though the lack of additional support does not preclude a more significant presence for the auxiliaries than is evident at present. At some point after the reign of Maximinus Thrax (r. 235-238), the Romans abandoned the Crimea, and presumably parts further west, with one exception around AD 250. 165 While we might question the exact administrative designation of the Black Sea coast, there is no reason to doubt whether there were Roman troops operating in some capacity in the region. 166 Indeed, the evidence that we have is not inconsiderable. 167 In sum, on the Crimea we have evidence for soldiers from the legiones I Italica, V Macedonica, and the XI Claudia, as well as contingents from the Moesian fleet and, quite possibly, a handful of auxiliary units, with only the cohors I Cilicum sagittaria, cohors I Bracaraugustanorum, the cohors I Claudia Sugambrorum veterana, and the cohors II Lucensium leaving some trace of their presence.

The Moesian Navy The first traces of a fleet operating on the lower Danube comes from the first two decades of Augustus’ reign during the conquest of the Balkans. Thus, a fleet was operational prior to the creation of the province. 168 Starr believes that it was not long after the formation of the province of Moesia that the fleet received its title, Classis Moesica. 169 He also believes that at some point during the reign of Claudius the fleet began patrolling the north shore of the Black Sea; Reddé, however, is a little more sceptical. 170 What limited evidence we do have for any members of the fleet along the Black Sea tends to date to the second century; moreover, that evidence is almost entirely limited to the Chersonesus. 171 There are a handful of inscriptions from Chersonesus that date to the second century. 172 The lone inscription that dates to the first century which names (or might name) a member of the Moesian fleet and was found around the Black Sea is AE 1967, 429, which lists a soldier (miles) of the cl(assis) Fl(avia) Misi(ca). So, although it is possible that the fleet was operating in the Black Sea in the first century, the sparse epigraphic evidence seems to suggest otherwise. The fleet was reorganised under Vespasian, and then renamed the classis Flavia Moesica either during the reign of that same emperor (Vespasian), 173 or even Domitian. 174 It kept the honorary title of Flavia through the third century. Before we look at the bases for the fleet, it is worth noting that the classis Flavia Moesica might not have been the only fleet operating in the region. Vegetius claimed that both the Ravenna and Misenum fleets operated in the Black Sea. 175 Indeed, on the basis of a handful of inscriptions from Charax, the case has been made that the Ravennate fleet was active along the northwest coast of the Black Sea, at least on occasion. Those few stamps from Charax (Aj-Tor) and Yalta read: “VEX(ILLATIO) C(LASSIS) RAV(ENNATIS) Not everyone is S(UMPTU) P(UBLICO)”. 176 convinced, however. Sarnowski, for instance, has argued that we should read on these stamps some variation of “VEX(ILLARII or ILLATIO) ) G. RAV(ONIO ?) SP(ERATO?) – that is with some form of Gaius Ravonius Speratus. 177 To my mind, however, the anecdote from Vegetius is suggestive and the restoration of some form of G. Rav. Speratus far too speculative – we have no other evidence for this individual, and so, in the absence of

159

It is hard to make sense of the “square/rectangle of hoplites” (plinthiou tou hopleitwn) mentioned in the inscription (Ivantchik 2014: 168), though I suspect it is likely a Greek form of legio. Note Ivantchik’s comments (2014: 175). If the inscription is a Hellenised Latin epitaph, it seems more than likely that the square is a legion, and its inclusion could even be part of the purported career of the named individual that we find in many Latin inscriptions. The inscription is fragmentary, however, and certainty is not possible. 160 Ivantchik 2014: 190. 161 IOSPE2 I.553; ILS 9160=LNCh 33; Wagner 1938: 99. On the identification of COH I BRA with the Bracarum cohort see Zubar 2001: 741. 162 CIL III.13751. 163 AE 2000, 1276. 164 AE 1909, 166=AE 2005, 1339; Zubar 2001: 737. 165 Wheeler 2012b: 147. 166 Sarnowski (1988: 87-95) provides a detailed history of the presence of Roman troops in Tyras, Olbia, and the Crimea in the second and third centuries, which includes incorporation of the evidence for individual units into what we know about north Black Sea military and political history, amongst other things. Some of the content veers later in time (middle of the third century) than our discussion here. 167 We have not discussed the physical evidence for the military in the north of Black Sea. On that matter see Sarnowski (1988: 81-87), a bit older, but still very useful.

168

Starr 1960: 130; Bounegru and Zahariade 1996: 7. Starr 1960: 131. 170 Starr 1960: 131; Reddé 1986: 263-264. 171 See AE 1903, 1 which dates to AD 185 and names a Tiberius Aurelius Secundus, a trierarch. 172 AE 1967, 428, 431, 432, AE 1984, 806. 173 Starr 1960: 132; Wheeler 2012b: 123. 174 Bounegru and Zahariade 1996: 9-10. 175 Veg. Mil. 4.31.6; Wheeler 2012b: 121. 176 CIL III.14215-5a, b, c; AE 1903, 2. See Reddé (1986: 263) and Bounegru and Zahariade (1996: 12). 177 Sarnowski 2006: 258. 169

59

Exercitus Moesiae additional evidence, the communis opinio seems the most likely of the two. When they might have been there is another matter: Zubar argues that the Ravennate fleet played an important role in Silvanus’ expedition in the middle of the first century, certainly a possibility. 178 He also dates those stamps to the first century. 179

Yet, given the propensity to load the Danubian crossing-points with military installations and troops, one wonders how great that role would have been, unless the legionaries or auxiliaries had left on a sortie. Moreover, on the lone stretch of the Danube with a limited number of fortifications that was patrolled primarily by the fleet, it would have been very difficult for invaders to cross. 191 This area, between Salmorus and Noviodunum, was in the Dobrudja and was primarily marshland. Thus, it seems likely that the fleet’s principal military function was to patrol the river and to carry out surveillance, in conjunction with the string of forts posted along the Danube and into the Dobrudja. 192 The fleet seems to have acted largely independently of the legionaries and auxiliaries, with the exception of times of war when the fleet was a key logistical aid.

We now move to the bases of the fleet. As with the majority of the land-based forces, there is no evidence of the fleet at any sites during the Julio-Claudian period. During the Flavian period, Margum and Egeta were the two bases in Moesia Superior. 180 By the middle of the second century, the number of bases had increased. Taliata, Transdierna, Diana, and Aquae were added. 181 As regards Moesia Inferior, only one suspected base has been suggested for the Flavian period, namely Aegyssus. 182 We are, however, a little better informed about the middle of the second century with bases at Dinogetia, Barbosi, Troesmis, Axiopolis, and Noviodunum, of which the latter may have been the headquarters for the fleet in the second century. 183 There also seems to have been a base at Chersonesus, in addition to Tyras and Charax. 184 Though, as noted, there is limited evidence for the Ravennate fleet at Charax, it seems unlikely that there was a base for the fleet at this Crimean locale. 185

CONCLUDING THOUGHTS So, what can we conclude from all this, if anything? Moesian troops were occasionally sent to the east for major campaigns, a practice which is in keeping with what we know about imperial strategy and troop deployment patterns, and a topic we will return to in chapter six. For all the individuality of the provincial armies noted earlier, each respective army was still an integral component in the wider military system with reinforcements usually sent from the closest available region. Thus Moesian, and Danubian, troops would be sent to the east, and in their place the Balkan provinces could usually expect troops from Germany. With regard to movement involving the Moesias themselves, there is only limited evidence of activity in the western portion of Moesia, what later became Moesia Superior. There is much more evidence for the presence of vexillations, which were often semi-permanent, in Moesia Inferior, and it seems that the fluid character of the province’s boundaries, as well as its close ties may have dictated the use of vexillations on the Black Sea coast.

There is one last point worth discussing, notably the role of the fleet in the provisioning of the empire’s legions and auxiliary units. The Moesian pridianum lists an indeterminate number of soldiers from the cohors I Hispanorum veterana who were “given to the Classis Flavia Moesica”. 186 The papyrus mentions grain ships, but separately from the reference to the fleet. 187 Thus, those grain ships do not appear to have been a regular part of the fleet, 188 though that is not to downplay, or obviate, the role of the fleet in the provisioning of supplies. Quite the contrary. That was presumably its primary duty, and more so in times of war. 189 As regards the fleet’s military role, Starr suggests that the fleet served in conjunction with the land forces, particularly to cut off fleeing invaders. 190

As we have just seen, Roman soldiers had a marked presence along the northwest shore of the Black Sea and into the Crimea, and so it remains to discuss the nature of their presence. We found evidence of quite a number of soldiers operating in the region, and, as noted above, the vexillations come with a range of different names. Some of those vexillations named individual units: the vexillatio XII Catafractariorum, the vexillatio legionis V Macedonicae, and the vexillatio legionis XI Claudiae. Some named more than one unit: the vexillatio legionum I Italicae XI Claudiae, the vexillatio legionum I Italicae XI Claudiae classis Flavia Moesicae, and the vexillatio legionum I Italicae V Macedonicae XI Claudiae et cohortium. The names of others implied an even more varied contingent: the vexillatio exercitus Moesiae, the

178

Zubar 2001: 735. Zubar 2001: 735. 180 Karavas 2001: 73 (Margum), 97 (Egeta). 181 Karavas 2001: 86 (Taliata), 90 (Transdierna), 93 (Diana), 99 (Aquae). 182 Karavas 2001: 119. 183 Bounegru and Zahariade 1996: 11 (Troesmis), 11 (Barbosi and Dinogetia), Karavas 2001: 120 (Noviodunum), 121 (Dinogetia), 122 (Barbosi), 125 (Troesmis), 127 (Axiopolis). See too Reddé 1986: 305. 184 As we saw above, a handful of inscriptions listing the fleet have been found there. See Starr 1960: 136; Bounegru and Zahariade 1996: 12; Wilkes 2005: 157. 185 Wheeler 2012b: 122. 186 Col ii, line 4. 187 Col ii, line 33. 188 Reddé (1986: 372) agrees. 189 Reddé (1986: 370). A quick glance at Trajan’s Column supports this. Where ships are depicted, they serve to transport both men and supplies, and do not engage the Dacian forces. See plates 3, 4, 5, 23, 24, 25, 33, 34, 59, 60, 61, and 65. 190 Starr 1960: 137. 179

191 192

60

Karavas 2001: 246. See Bounegru and Zahariade 1996: 93ff.

Vexillations and the Black Sea vexillatio Moesiae inferioris, the vexillatio exercitus Moesiae inferioris, the vexillatio Chersonissitanae, and the vexillatio Ponticis aput Scythia et Tauricam. One thing we might like to do is divide the different vexillation name types into chronological periods with the idea being that we would then be able to see changes in nomenclature and military organisation. Sadly the evidence is insufficient. Suffice to say, many of the vexillations that name more than one – or an interminable number of – unit/s date to at least the second half of the second century. On the other hand, that also seems to be when most of the vexillations as a whole were active.

in other contexts, why would there be an exception in this case? This could be the product of the nature of the threats faced by Chersonesus and its territory, whether that came from murky barbarians beyond, or the neighbours of the residents of Chersonesus Taurica in the Kingdom of the Bosporus. Indeed, Haensch argues that the military threat that they faced necessitated a larger than usual vexillation, and one led by a more highly ranked than usual commander, a tribunus militum. 194 Of course, we do not actually know the size of the vexillation referred to in that inscription, and so whether it actually was any larger is speculation.

Looking at the particular places where we find these vexillations, Wilkes, for instance, has argued that Chersonesus was the base for their operations, and that the soldiers found at Tyras, Olbia, and Charax were detachments from that force on the Crimea. 193 There is no doubt that there was a sizable concentration of soldiers, but given we are dealing almost exclusively with vexillations, it is nothing like the deployment we found along the Danube, even if the evidence is far less abundant than it is for that along the lower Danube. Does this suggest that the region was not a great a concern to Roman authorities, at least in the same way that the Balkans seem to have been?

Ultimately, what we can say is that there is no doubt that vexillations from the Moesias, and Moesia Inferior in particular, were used increasingly in the second century in keeping with the pattern across the empire. In the case of Moesia Inferior, these vexillations were often sent to the Black Sea, and for an indeterminable amount of time. The fact that the distinction vexillatio was so common suggests to me that these soldiers operating in the region were always, in some sense at least, temporary in nature, even if in reality it meant that the powers that be never quite got around to assigning those units names that reflected their permanence. In other words, on the surface they were seen as a temporary, makeshift solution, even if the reality was something else. 195

That most of the units that we have evidence for in the region are vexillations is suggestive. To my mind, at least, it suggests that the region, though connected to Moesia Inferior, was never fully considered part of the province. The existence of inscriptions that list some form of vexillatio exercitus Moesiae/Moesiae Inferioris seems to imply that the army of Moesia Inferior was in some way distinct from this region; we do not find such a designation in what we might call Moesia Inferior proper. As noted above, vexillations tend to be used for specific, exceptional tasks (so far as we know), and the creation of vexillations from the legions and auxiliary units already based in the province for duties within said province would seem to be superfluous: the mandate for legions and auxiliaries meant operating within their province without the need for independent vexillations. This is best illustrated by Hunt’s pridianum. Indeed, the presence of troops from Moesia Inferior was the most logical of options, geographically speaking, for it was the closest province with an abundance of troops to those sites on the northwest coast of the Black Sea. To my mind, then, from the perspective of military administration, places like Tyras, Olbia, and the assorted sites on the Crimea seem to have been outside of Moesia Inferior proper.

List of Legions Operating Around the Black Sea Location Legions (vexillations) Tyras I Italica, V Macedonica, XI Claudia Olbia I Italica, V Macedonica, XI Claudia Chersonesus I Italica, V Macedonica, XI Claudia Symbolon Charax

There is one caveat: in all other parts of the empire all known legions were based within the Roman provinces themselves. While it is also likely that they operated outside of those provinces and beyond into the barbaricum, we lack the evidence for this. If units were always based within the boundaries of a province 193

194

I Italica, XI Claudia I Italica, XI Claudia

Haensch 2009: 217. He is referring to the military tribune of the First Italian Legion referred to in CIL III.14214. 195 In the fifth volume of his fictional Warrior of Rome series, Sidebottom (2013) characterises the detachment in Olbia in the middle of the third century as permanent and established, even if middling in quality.

Wilkes 2005: 153.

61

Exercitus Moesiae List of Auxiliary Units Operating Around the Black Sea Location Tyras Olbia Chersonesus

Symbolon Charax

Auxiliary Units (vexillations) cohors I Cilicum cohors IIII Cypria, cohors I Thracum Syriaca, cohors I Bracarum or Bracaraugustanorum, cohors I Claudia Sugambrorum veterana, cohors II Lucensium

List of Other Units Operating Around the Black Sea Location Tyras

Olbia Chersonesus

Symbolon

Charax

Other vexillatio ponticis aput Scythia et Tauricam, classis Flavia Moesica vexillatio Chersonessitana, classis Flavia Moesica vexillatio exercitus Moesiae inferioris, vexillatio exercitus Moesiae Ravennate fleet (?), classis Flavia Moesica

62

garrison of a great number of bases is significant, and it is a point we will come back to when we turn to the auxiliaries. Ostensibly this is less of an issue for the legionary bases than it is for the auxiliary bases. Dozens of smaller fortifications have been identified in the Moesias, but we often do not know the garrison, and our evidence is invariably paltry. Often a lone inscription provides our clue to the identity of a site’s military unit. In other cases, we have a few inscriptions, but each one names a different unit, so complicating the matter further. Epitaphs, which are found throughout the Moesias, are less useful in identifying the garrison of a fort, as we will see.

Chapter 5: Troop Emplacement in the Moesias INTRODUCTION 1 Half the strategic puzzle for the Roman Empire involves evaluating the changes in troop disposition. The other half, at least for our purposes, centres on where in the provinces those soldiers were. With this in mind, we turn to troop emplacement, and we will begin with the legions and finish with the auxiliaries, largely because the evidence for the bases of the legions is much easier to find than that for the auxiliaries. Before we turn to the specifics in the Moesias, an outline, of sorts, would not go amiss. 2

Ostensibly, brick and tile stamps offer a reasonable means of identifying a unit with the assumption being that they are, by some measures, official. On the other hand, the presence of a stamp or two does not provide unequivocal evidence for the presence of a unit. 10 A unit might well have been in a temporary location, perhaps aiding in the construction of said fortification, but this is not certain. Given that this is often all we have to go on (the fortifications on the lower Danube are understudied, after all), however, this class of evidence should be considered, even if this means that results are unlikely to be as clear-cut as we would like. Still, we will bear in mind Kurzmann’s three points concerning what the presence of stamps of a unit might mean: firstly, a unit (or a detachment) was garrisoned at the location for a period of time; second, the unit was not garrisoned there, and instead delivered bricks to assist those present; and third, a detachment came to the location temporarily to help with stamp production. 11 Another problem is stamp reuse: we cannot determine when or if stamps were reused, and if so whether this was done so by different units from those named in any stamp. 12 If we move from the general to the specific, while we might be fairly well informed about stamp production in some parts of the empire, this is not the case for the lower Danube. 13 For instance, brick stamps that mention Dierna have been found in at least 12 different locations, which might mean that they came from one workshop, but we lack suitable criteria to date them, let alone use them to determine the garrison of particular places, something a central workshop would call into question anyway. 14 Indeed, in general much more work is needed on the stamps on the region, 15 a fact which complicates immensely attempts at determining the garrison of particular fortifications.

We turn to Wilkes’ overview to sketch, in broad strokes, the pattern in fortification placing on the Danube. The fortifications started to become permanent in the second half of the first century, and in the case of the Balkans Wilkes notes that the arrangement that emerged came down to the “prolonged warfare along the Danube under the Flavians and Trajan”. 3 When Dacia was annexed, this had a marked impact on a number of fortifications in the Moesias. We know, for instance, that some forts were evacuated, while other portions of the frontier, especially the eastern end near the Black Sea, were reinforced. 4 Mobile auxiliaries were dispatched to higher ground where possible, and in places that afforded careful observation of important plains, parts of the river, and crossings. 5 It also seems that auxiliaries tended to be dispersed in smaller forts from the middle of the second century, with the ultimate effect being the creation of a continuous land route between the River Inn and the Black Sea by the end of the century. 6 Although a significant number of the known fortifications were along east-west routes, both on the Danube itself and further south between Montana and Odessus, there were also a number of important routes, lined with fortifications, so to speak, We also find through the Haemus mountains. 7 evidence for fortifications clustered around major mining centres. 8 Before we turn to the specifics, there a few more points worth stressing. At this point, many of the forts that lined the Danube and beyond have not been thoroughly investigated, archaeologically and otherwise, for a variety of reasons. 9 That we know little about the

Identifying the garrisons of the provinces is not, then, a straightforward exercise, with the obvious exception of the primary legionary headquarters. Caution, then, is

1

All place names cited are ancient names unless otherwise specified. All of the locations named in this chapter can be found in the maps located in appendix six below. 3 Wilkes 2005: 151. 4 Wilkes 2005: 153. 5 See Lemke 2015. 6 Wilkes (2005: 156-157) collects the evidence. 7 Wilkes 2005: 158. 8 Note the discussion in chapter three above on the auxiliaries posted to Moesia Superior created during the reign of Marcus Aurelius. 9 Wilkes 2005: 155. 2

10

Kurzmann 2006: 4. Kurzmann 2006: 5. 12 Note Kurzmann’s (2006: 141) comments on how little the problem of reuse has been discussed in studies of military stamps in southeast Europe. 13 Sarnowski 1995a; Kurzmann 2006: 115. 14 Benea 1977; Kurzmann 2006: 127. 15 Kurzmann 2006: 140-141. 11

63

Exercitus Moesiae needed. In most instances, we will not be able to determine, with absolute certainty, when – or even if – a given unit was garrisoning a particular fortification, and even if we think a unit is in a given spot, we invariably do not know when this was.

Viminacium upon its arrival in the province between AD 57 and AD 59. 25 The legio III Gallica was in Moesia too briefly for us to determine its location. Similarly, the legio V Alaudae was also in Moesia briefly, and we do not know where it stayed. When the Fifth Macedonian legion was transferred back to Moesia around AD 71, it returned to Oescus where it remained until some point before or after Trajan’s two Dacian wars. 26 The First Italian legion was based in Novae when it arrived in Moesia, and remained there through the second century. 27

LEGIONS The Legions from Augustus to Titus We start with the legions, and our focus in this discussion will be on their primary residences – their headquarters, so to speak. The evidence is always much more abundant, and definitive, for legionary bases than it is for the auxiliary bases. The one problematic issue is the possible dispersal of detachments of legions into smaller forts, something which we will come back to when we look at auxiliary garrisons.

The Legions from Domitian to Antoninus Pius For the next period of legionary emplacement we have Ptolemy’s Geography, which details the legionary garrisons of cities. 28 Ptolemy, who was writing between about AD 146 and AD 170, lists Singidunum in Moesia Superior along with the legio IV Flavia, and by the middle of the second century, the legion was based in that city. When did the legion arrive? Ritterling thinks that when the legion initially arrived in Upper Moesia, it was probably based at Viminacium, the winter quarters of the Legio VII Claudia. 29 With its sojourn in Dacia after Trajan’s Dacian wars, Ritterling posits that the unit then moved to Singidunum during the reign of Hadrian. 30 Mocsy also suggests Viminacium as the possible first base of the legion, but does not rule out the possibility that the legion was based initially at Singidunum, or even Ratiaria. 31

In the early history of Moesia, the legio V Macedonica was stationed from the reign of Tiberius at Oescus. 16 Although we do not know when the legion was moved to Oescus on the Danube, Florus states that Lentulus created garrison posts on the near bank of the river. 17 Thus, we should probably associate those garrison posts with the creation of a new legionary camp at Oescus for the Fifth Macedonian legion, as Mocsy does. 18 The base of the legio IIII Scythica may have been close to Macedonia at Scupi, or even Naissus. 19 M. A. Speidel admits that we know little more concerning the early history of the legion than Ritterling did. 20 The unit was involved in road construction between Ratiaria and Viminacium in AD 33/34, but so was the Fifth Macedonian legion. 21 Gospodin Vir, the modern location of the roadwork, is a considerable distance from Oescus; thus, if the legio V Macedonica travelled some ways to participate, it would not be unreasonable to assume the same from the legio IIII Scythica. But, this need not mean that the legion was stationed in Viminacium or Ratiaria, or even near by. 22 The question of the legion’s location remains unanswered.

On the other hand, Strobel, Bojović, Campbell, and Le Bohec and Wolff say rather unequivocally that the Legio VII Claudia spent the duration of its time in Moesia Superior at Singidunum. 32 The difficulty with identifying the legion’s first Moesian base is the disjointed picture presented by the high number of inscriptions from both Singidunum and Viminacium, the difficulty in dating many of the inscriptions from both sites, and the lack of precise, independent With no definitive evidence from Viminacium. 33 evidence for Viminacium as its base, unlike Singidunum, it would be best to maintain that the legion spent its first years in Moesia at Singidunum,

When the legio VIII Augusta arrived in Moesia, it was stationed at Novae. 23 Traces of a Julio-Claudian camp have been found at the site. We also have an inscription that attests its presence in the region between the years AD 45 and AD 69. 24 The legion does not seem to have changed its base while in Moesia. The Seventh Claudian legion garrisoned

25

Mirković 1986: 36; Campbell 1996a: 841; Strobel 2000: 528. Campbell 1996a: 840; Zahariade and Gudea 1997: 44. 27 See Filow (1963: 63), Campbell (1996: 839), Zahariade and Gudea (1997: 43-45), and Absil (2000: 228-232). 28 The edition of Ptolemy’s Geography used for this book is Mullerus’ (1883). 29 Ritterling 1925: 1543. 30 Ritterling 1925: 1543. 31 Mocsy 1974: 82-83. In Ptolemy’s city list for Moesia Superior he does mention Ratiaria. Significantly, he does not associate a legion with that city, although he does call it a colonia. 32 Strobel 1984: 88; Bojović 1996: 53-68; Campbell 1996a: 840; Le Bohec and Wolff 2000b: 239. 33 For a brief discussion of the epigraphic habits of the legion see Le Bohec and Wolff (2000b, 239-240). Cf. Mirković and Duŝanić 1976: 28-29. 26

16

Mocsy 1974: 43. See too Gerov (1980: 2ff, 149ff). Flor. 2.28: citra praesidia constituta. 18 Mocsy 1974: 43. 19 Mocsy 1974: 44. 20 Speidel 1998: 163; Speidel 2000: 327. 21 CIL III.1698. 22 Parker 1958: 126. 23 Reddé 2000: 121; Karavas 2001: 132. 24 ILBulg 300. 17

64

Troop Emplacement in the Moesias which is where it returned after its time in Dacia. 34 With respect to Viminacium, Ptolemy has noted the presence of a legion; however, he merely recorded λεγίων. 35 Scholarly opinion, however, favours the legio VII Claudia as the occupying legion. 36 There are scores of inscriptions that attest to the legion’s presence in this city. 37

The Legions from Marcus Aurelius to Severus Alexander At the beginning of this last period, in AD 161, the legions were emplaced in the following camps: the legio IIII Flavia was at Singidunum, the legio VII Claudia was at Viminacium; 45 the legio I Italica was at Novae, the legio V Macedonica was at Troesmis, and the legio XI Claudia was at Durostorum. 46 To determine changes in unit emplacement, in the next period we have the evidence of the Antonine Itinerary, a list of settlements and distances, which dates to some point in the third century. It lists the following legions and camps: the legio XIII Gemina at Ratiaria, 47 the legio V Macedonica at Oescus, 48 the legio I Italica at Novae, 49 the legio XI Claudia at Durostorum, 50 the legio I Iovia at Troesmis, 51 and the legio II Herculia at Noviodunum. 52 Of those legions mentioned, it is the First Italian and Eleventh Claudian legions that remained stationary. The Fifth Macedonian legion had moved to Dacia around AD 168 and, by somewhere between AD 211 and AD 229, it had not yet returned. The legion likely returned to the Moesias following the abandonment of Dacia in AD 271. When it did return to Moesia Inferior, however, it did not go back to its former base at Troesmis, but rather to its initial Moesian base at Oescus.

The first city that Ptolemy mentions in Moesia Inferior with a legion is Dourostoron (Durostorum), the alleged base of the legio I Italica. 38 However, Ptolemy was probably mistaken, since the legion was likely still based at Novae, to which Ptolemy does not assign a legion. 39 Troesmis is the next major centre assigned a legion by Ptolemy. Ptolemy claims that the legio V Macedonica was in Troesmis. 40 It seems that the legion had moved from its earlier base at Oescus, which is where it had been based since its return to Moesia from the east during the reign of Vespasian. 41 Then, at some point during or after the second Dacian war of Trajan, the legio V Macedonica moved to a new base at Troesmis. 42 Unfortunately, Ptolemy does not list any other towns with garrisons; yet, we know that the legio XI Claudia had moved to the province by this period. As it turns out he had mentioned the city that housed the legion, namely Durostorum; however, he improperly assigned the legio I Italica to the city. 43 In fact, since the legion’s arrival in the province its base had been Durostorum, which is where it remained for the duration of its time in Moesia Inferior. 44

The Fourth Flavian legion and the Seventh Claudian legion are not found in the Antonine Itinerary in the section pertaining to the Moesias. What is more, the writer/s does not attach any special significance to the two Upper Moesian legionary sites of Singidunum 53 and Viminacium. Bojović has demonstrated, however, that the legio IIII Flavia was based at Singidunum at least until the reign of the emperor Constantine the Great. 54 With regard to Viminacium, we have a few inscriptions that name soldiers of the legion that date to the end of the second century and beginning of the third century. 55 As such, there is little reason to believe that the legio VII Claudia left its base at Viminacium before the end of Severus Alexander’s reign.

34

Gudea 2001: 45-50. Ptol. Geog. 3.9.3. 36 Filow 1963: 62-63; Mocsy 1974: 82; Campbell 1996a: 841; Le Bohec and Wolff 2000b, 242; Gudea 2000: 53-55. 37 See, for example, AE 1903, 294; AE 1903, 295; AE 1905, 160; CIL III.6324a, CIL III.8103=IMS II.5=AE 1889,95; CIL III.8119 = IMS II.124. 38 Ptol. Geog. 3.10.5. 39 For the assigning of the legio I Italica to a base at Novae see Filow (1963: 63), Campbell (1996a: 839), Zahariade and Gudea (1997: 43-45), Absil (2000: 228-232), and Gudea (2005: 421-425). Although this legion does not seem to have been quite as prolific as some other legions, there are at least a handful of inscriptions that attest its presence at Novae at this time: AE 1999, 1332; ILBulg 301; ILBulg 306; ILBulg 307; ILBulg 329. 40 Ptol. Geog. 3.10.5. 41 CIL III.12348; CIL III.14415. For the legion’s tenure at Oescus see Ritterling (1925: 1575-1576), Filow (1963: 64), and Zahariade and Gudea (1997: 44). 42 Ritterling 1925: 1576-1578; Filow 1963: 64; Campbell 1996a: 840; Zahariade and Gudea 1997: 44; Gudea 2005: 452-453. See also CIL III.6166; CIL III.6168; and CIL III.6169. 43 Zahariade and Gudea (1997: 44). 44 Note, for instance, ADBulgar 458; AE 1902, 133a; AE 2005, 1335; CIL III.7475=CIMRM II.2273=IScM IV.107; CIL III.7474; CIL III.7619a. See also Ritterling 1925: 1698; Filow 1963: 65; Campbell 1996a: 841; Zahariade and Gudea 1997: 44; Fellman 2000: 130; Gudea 2005: 434-438. 35

45

Both sites are in Moesia Superior. All three sites are in Moesia Inferior. 47 It. Ant. 219.3. The legion was not transferred to the province before AD 271. 48 It. Ant. 220.5. 49 It. Ant. 221.4. 50 It. Ant. 223.4. 51 It. Ant. 225.2. We do not know when this legion was raised, though presumably it was in the second half of the third century, and possibly during the reign of Diocletian. This fact and the notice of the return of the legio V Macedonica point towards a Diocletianic date for the document, though this ignores the possibility for later amendments to the text. 52 It. Ant. 226.1. 53 In fact, Singidunum is conspicuously absent. 54 Bojović 1996: 66ff. 55 IMS II.12, dated to c. AD 204; IMS IV.44, probably dated to c. AD 193; CIL III.14507, AD 195; CIL III.14509, probably post AD 211 (after the death of Septimius Severus). We have a mass of other inscriptions from Viminacium that name the Seventh Claudian legion, but we cannot date them with any precision. 46

65

Exercitus Moesiae

Introduction

the case under the Julio-Claudians, it was not long before permanent auxiliary fortifications were established, at least in other parts of the empire. Excavations at Valkenburg (Netherlands), for instance, indicate that by AD 40 this fort had been established with the express purpose of housing an auxiliary unit alone. 58 Regrettably, and as noted in the introductory section of this chapter, many fortifications in Moesia have not received the same amount of attention as those in other parts of the empire, so we cannot say whether this is the case there too. It is hard to believe, however, that once the province was stable and established, that the Romans did not employ some of the same practices here as well. With this in mind, in the subsequent discussion we move beyond the legionary bases associated with auxiliaries to specifically auxiliary ones, at least where possible. In the following paragraphs, discussion will, for the most part, proceed with fortifications running from west to east, where possible; this presupposes that we start, both before the division of the province and after, with Moesia Superior (or the area that became Moesia Superior). We will, where possible, use the ancient names.

AUXILIARIES

We now turn to the auxiliaries, and in nearly every case we are far less certain about their locations. As noted above, there are a number of issues with the auxiliaries and the known forts from the region. We know – or we think we know – the names of all of the region’s auxiliary units, and we have identified a number of fortresses, but there is often a disconnect between the two categories: the evidence for auxiliary emplacement is much harder to come by than that for legionary emplacement, as we will see. Indeed, when it comes to the auxiliaries, we are largely in the dark, although the size of a fortification might seem to be a good barometer of the approximate size and character of its inhabitants. As we saw above in our discussion of vexillations, not all of the soldiers from a given unit necessarily stayed together for the duration of their service: vexillations were not only used temporarily, for they were permanently, or semi-permanently, based in disparate locations, like the northwest coast of the Black Sea. As a result, what might seem to be an auxiliary base because of its size, might in fact be the base for a detachment of legionaries. It might even house both groups of soldiers, if not others too (numeri, for instance). The evidence for places like Chersonesus, incomplete though it may be, indicates that disparate groups of soldiers were often housed together, and often far later in time than we might have expected. In our discussion we should also bear in mind that the evidence rarely, if ever, indicates whether all of unit garrisoned a particular place, or only a detachment.

Mocsy says that auxiliary units were likely stationed on the road through the Morava river valley in pre-Flavian times. 59 The same seems to have been true for the Timok river valley, for there is evidence of a preFlavian fort at Timacum Minus. 60 Indeed, the cohors I Thracum Syriaca was based at Timacum Minus in the years from AD 70 – AD 106. 61 Karavas argues that Saldum was the home of the cohors I Antiochensium, the cohors I Cisipadensium, and the cohors I Raetorum. 62 Mocsy tells us that there was a fort at Taliata in Moesia Superior during the early years of Vespasian. 63 The fort’s size suggests that an auxiliary unit might have been its garrison. In fact, Karavas argues that by the last quarter of the first century the cohors I Raetorum had moved to Taliata. 64 The ala Claudia Nova may have been based at, or at least close to, Ratiaria. 65 The cohors II Lucensium might have been stationed at Montana when it came to Moesia. 66 The cohors Cilicum might have been stationed at Naissus under the Flavians. 67

In sum, all of these issues and more pose considerable challenges to any attempts to identify the garrison of the region’s fortifications particularly in the absence of overwhelming evidence. We must bear all these points in mind as we delve into auxiliary troop emplacement, and the related issue of vexillation emplacement. Indeed, a great deal of the discussion will be speculative, though in the absence of additional evidence, it is better to say something rather than nothing.

Shifting to the central sector of the frontier, the cohors Mattiacorum might have served at Sexaginta Prista in

Moesia Initially, it seems that auxiliary units were based in or around the fortifications that housed the legions. Mocsy notes, for instance, that the pre-Flavian tombstones of auxiliaries that we have in Moesia tended to be restricted to places in and around Naissus and perhaps Ratiaria. 56 He also acknowledges the presence of auxiliary units around the base of the legio V Macedonica at Oescus. 57 Whether this was always

58

Haynes 2013: 53. Mocsy 1974: 51. Pre-Flavian tombstones have been found around Naissus and Ratiaria. 60 CIL III.8261; Cf. Mocsy 1974: 51. 61 Karavas 2001: 102. 62 Karavas 2001: 81. 63 Mocsy 1974: 81. 64 Karavas 2001: 86. This is based on the discovery of a diploma on the site. See too Gudea 2001: 70. 65 Spaul 1994: 90. See CIL III.14500. The funerary stone does not prove the presence of the unit at this later colony; but there is a possibility, however remote, that a fort was attached to the colony. 66 Spaul 2000: 84. See too Velkov 1990: 247-256 67 Spaul 2000: 398. 59

56

Mocsy 1974: 43-44. Mocsy 1974: 43. Cf. Gerov 1980: 149ff. This is based on the presence of tombstones from some soldiers. 57

66

Troop Emplacement in the Moesias the early Flavian period. 68 The cohors I Cisipadensium was probably stationed at modern Lomets. 69 There is some evidence for the ala Scubulorum at Nicopolis, which is where it may have been from the reign of Tiberius. 70 The ala Atectorigiana may have been stationed at the modern town of Ryahovo (ancient Appiaria). 71 The ala Bosporanorum may have been at Securisca. 72

cohors I Cantabrorum 84 and the cohors III Campestris 85 might both have been based at modern Prahovo (Aquae). 86 Several auxiliary units were also active in the vicinity of the two main legionary bases, Viminacium and Singidunum. We have evidence that the cohors III Campestris, the cohors VII Breucorum, and the cohors I Flavia Hispanorum were all active in the area around Viminacium at the end of the first century and beginning of the second. 87 One location put forward as a possible base for a couple of different auxiliary units, and which illustrates well the problems we have in identifying the garrison of a given place, is Cuppae, just east of Viminacium along the Danube. Mirković argued that by the reign of Hadrian the cohors III Campestris was based at Cuppae. 88 For Karavas, also there at some stage was the cohors V Hispanorum. 89 One unit that has not be assigned to Cuppae, however, is the cohors I Flavia Hispanorum milliaria. 90 None of the inscriptions can be dated securely enough to posit changes in the garrison, so it seems we could not say that the Campestrian unit was there first, followed by one of the Hispanic units. 91 It also seems unlikely that more than one unit would ever have been housed at the site at any given time, which makes it difficult to make sense of what inscriptions we have. The fact is that the evidence for all three units is extremely slim: one inscription for the first, Campestris, one inscription for the third, I Flavia Hispanorum, and an inscription that simply records “COH V”, and which may be the V Hispanorum. Admittedly, the inscription for the Campestris cohort is an epitaph, while those for the Hispanic units are brick stamps. The former implies that a soldier was at least at Cuppae at the end of his life; the latter that those units were involved in construction, at least, at the site. None give conclusive evidence of a permanent garrison for the units.

We finish with the Dobrudja and the Black Sea coast. The ala Dardanorum unit might have been based at Troesmis, 73 or it may have been at Arrubium. 74 The ala I Asturum may have been stationed at Tomi; 75 the ala I Asturum may also have been based at modern Sofia, quite a ways to the west. 76 Finally, when the cohors VII Gallorum moved to Moesia it was probably stationed at Tomi. 77 Moesia Superior We begin with the interior of Moesia Superior. There is scattered evidence for the postings of auxiliary units. Petrović says that when the cohors I Thracum Syriaca arrived in the province it garrisoned Timacum Minus. 78 It stayed there until it was transferred to Transmarisca. 79 Following the cohors I Thracum Syriaca’s transfer, the cohors II Aurelia Dardanorum replaced it at Timacum Minus. 80 The cohors I Montanorum might also have been at Timacum Minus, if briefly. 81 Towards the end of the first century the cohors I Cilicum and the cohors I Cretum were both based at Naissus. 82 By AD 134, however, the cohors I Cilicum had moved to Moesia Inferior. 83 During their brief sojourns in Moesia and Moesia Superior, the

68

Spaul 2000: 244; Karavas 2001: 131. Cf. AE 1957, 307, which is a dedicatory stone (to Vespasian). 69 Spaul 2000: 464. This is based on the presence of a dedicatory stone, albeit a later one (CIL III.14429). 70 Spaul 1994: 325. See too Gerov 1980: 155-163; and AE 1967, 426 which is a funerary stone with an invocation. 71 Spaul 1994: 326. See CIL III.12452, which is a funerary stone with an invocation. 72 Gerov 1980: 163; Spaul 1994: 326; Karavas 2001: 134. Karavas noted that there was epigraphic evidence discovered at the site during excavations. Cf. AE 1925, 70, which is a funerary stone with an invocation. 73 Spaul 1994: 324. See. CIL III.7504, which is a funerary stone with an invocation. 74 Spaul 1994: 325. See CIL III.7512, which is a building inscription. 75 IScM II.172=AE 1988, 998. 76 Spaul 1994: 37, 326. 77 Spaul 2000: 171. See CIL III.7548, which is a funerary stone. 78 Petrović 1995: 44; Gudea 2001: 95-96. Cf. Mocsy 1974: 81; Spaul 2000: 366. 79 Petrović 1995: 44; Gudea 2005: 431. 80 Cf. Gudea 2001: 96. 81 Petrović 1995: 44. Cf. AE 1903, 289, which is a funerary stone. 82 Petrović 1979: 31. For the cohors I Cretum see AE 1963, 262 (a funerary stone); for the cohors I Cilicum see CIL III.8250 (funerary stone). Naissus is a curious omission from Gudea’s (2001) study of the frontier in Moesia Superior, even if his ostensible focus is the northern portion of the frontier. After all, Timacum Minus is included. 83 Petrović 1979: 31.

The cohors I Montanorum, which had been based in the interior, was now based at Novae while the cohors V Gallorum was based at Transdierna. 92 When it returned to Moesia Superior in the second century from Dacia, it may have been based in the modern town of Pojejena. 93 Karavas believes that the cohors IX voluntariorum, which is not mentioned in any 84

ILJug II.463. ILJug II.461=AE 1971, 424. 86 Gudea 2001: 89. 87 Mirković 1986: 38; Gudea 2001: 55. 88 Mirković 1986: 38; Gudea 2001: 62. See too Radnoti 1959: 142ff. 89 Karavas 2001: 84-85. CIL III.1702. Cf. Gudea 2001: 62. 90 AE 1910, 85. 91 Note the comments of Gudea (2001: 62). 92 Strobel 1984: 131; Strobel 1984: 140. Mirković 1986: 38; Karavas 2001: 79 (Novae), 90 (Cuppae). For the cohors V Gallorum see AE 1963, 165 (altar), AE 1972, 490 (altar), and Gudea (2001: 73). The cohors I Montanorum seems to have been based at Novae while it was in Moesia Superior at the end of the first century and beginning of the second century. It returned to Novae when it came back to Moesia Superior during Hadrian’s reign. See Gudea 2001: 63. 93 Strobel 1984: 131; Gudea 2001: 59. 85

67

Exercitus Moesiae diplomas, was based at Transdierna in the first half of the second century. 94 Tricornium may have been a base for the cohors I Flavia Bessorum from AD 100 to AD 120. 95 Later in the second century, Tricornium seems to have been the headquarters for the cohors I Pannoniorum. 96 Karavas suspects that in the last decade of the first century the cohors VI Thracum was based in Diana. 97 The ala Praetoria may have been based at Teutoburgium while in the Balkans. Spaul bases this supposition on what we know about the provinces where the unit is known to have served, namely Pannonia Inferior, Moesia Superior, and then Pannonia Inferior. 98 His suggestion is that the unit only changed provinces with the corresponding changes in the provincial boundaries. Most scholars assign Teutoburgium to Pannonia Inferior, however, which casts doubt on any connection between this fort and its service in Moesia Superior. We know that the cohors I Hispanorum was operating around Stobi in AD 105, which is in Macedonia. 99 This could have been a temporary measure – its province of origin was Moesia Superior – in the context of Trajan’s second Dacian War. Spaul argued that the cohors I Cisipadensium was stationed in modern Lomets (in what was Moesia Inferior), 100 though given that the unit seems to have spent the duration of its posting in Moesia in Moesia Superior, Gudea’s suggestion of Bononia is preferred. 101

to the former and those on the north bank to the latter, as the evidence for Drobeta demonstrates. Turning to the last period under review, Mirković claims that the cohors II Aurelia nova garrisoned the modern city of Kosmaj after AD 169. 106 Dusanić says that the cohors I Aurelia Dardanorum was probably stationed at Naissus, whereas the cohors II Aurelia Dardanorum was stationed at Timacum Minus. 107 The cohors II Aurelia nova was also located in the north of Moesia Superior at the fort in the modern town of Stojnik. 108 Karavas also thinks that a vexillation of the cohors I Ulpia Pannoniorum was based at Stojnik. 109 The whereabouts of the cohors II Aurelia nova Sacorum, poorly documented as it is, remains unknown. 110 Mirković has noted the presence of the cohors I Montanorum at Novae, where it probably remained in the third century. 111 The cohors V Gallorum probably remained at Transdierna before its transfer. Spaul suggests that the ala I Claudia nova miscellanea may have been stationed at a fort near the colony of Ratiaria. 112 The ala I Gallorum Flaviana may have been stationed near Sirmium, which may have been part of Moesia Superior. 113 Gudea, on the other hand, argues for the presence of the ala II Pannoniorum in Sirmium. 114 The cohors III Brittonum was at Pontes from the reign of Marcus Aurelius onwards. 115

The cohors III Brittonum veterana was possibly at Pontes at the end of the first century and into the second century. 102 The same might be true of the cohors I Cretum and the cohors II Hispanorum on the basis of some brick and tile stamps. 103 The cohors IV Cypria cR may have been based, or at least launched its campaigns during Trajan’s Dacian wars, from Drobeta prior to its move to Dacia. 104 The evidence, however, is limited. 105 A complicating factor, however, would seem to be determining which province we should associate Drobeta with. Conventional wisdom suggests it should be Dacia – it is across the Danube from Pontes. But, inscriptions that name all sorts of different units, legionary and auxiliary alike, have been found at the site, and many of those were active in Moesia Superior. Plus, the boundary between Moesia Superior and Dacia was not necessarily the Danube, with the all the sites found on the south bank belonging

Moesia Inferior Turning to the auxiliary units of Moesia Inferior, plenty of evidence has been found from Trimammium, in the form of tile stamps, for the presence of the cohors I Bracarorum. 116 The cohors II Flavia Brittonum replaced the cohors II Mattiacorum at Sexaginta Prista between AD 145 and AD 151, and likely remained

106

Mirković 1976: 104-105. Kosmaj is in the northwest of the province just south of Singidunum. 107 Dušanić 2000: 349. Cf. Petrović 1995: 44. This supposition is supported by the inscriptional evidence: we have several inscriptions recording the unit’s presence in these towns, particularly for the second cohort. For Naissus, see for example CIL III.8251 and IMS IV.94. For Timacum Minus see for example AE 1904, 92, AE 1952, 191, and AE 1976, 610. 108 Dušanić 2000: 349; Gudea 2001: 95. 109 Karavas 2001: 103. Cf. Gudea 2001: 95. 110 Karavas (2001: 103) suspects that the cohors II Aurelia nova Sacorum may have been based around the modern town of Stojnik. 111 Mirković 1986: 38. 112 Spaul 1994: 90. Spaul notes that the funerary stone (CIL III.14500) found at Ratiaria does not mean that the unit was there, although he notes that there are instances of forts attached to colonies in Britain. From this, he suggests that there may have been a fort in the vicinity of Ratiaria. It is only speculation however. 113 Spaul (1994: 115) conjectures that the town may have been part of Moesia Superior at one point, even though the site (ancient Sirmium) was part of Pannonia Inferior. There is an inscription which lists a P. Helvius Pertinax from the unit from Sirmium. As the unit was never based in the Pannonias, Spaul’s suggestion must remain a possibility, however remote. 114 Gudea 2001: 47. 115 Karavas 2001: 94. 116 AE 2011, 1133a1-35, b1-6. See Matei-Popescu 2010: 195.

94

Karavas 2001: 90. See too Beneš 1978: 56; Strobel 1984: 146; Spaul 2000: 38; Gudea 2001: 73. The possible evidence comes from a brick stamp: AE 1977, 740. 95 Gudea 2001: 52; Karavas 2001: 71. 96 Gudea 2001: 52; Karavas 2001: 71. Karavas’ suggestion is based on the presence of brick stamps naming the unit and a fragmentary diploma. 97 Karavas 2001: 93. See too Gudea (2001: 77). 98 Spaul 1994: 188. 99 RMR 63. 100 Spaul 2000: 464. See CIL III.14429. 101 Gudea 2001: 92. 102 Strobel 1984: 125; Gudea 2001: 79. 103 Gudea 2001: 79. 104 Strobel 1984: 129. 105 See IDR II.179 and Gudea (2001: 81-85).

68

Troop Emplacement in the Moesias there well into the third century. 117 It is possible that the ala I Vespasiana Dardanorum was based at (Nova) Cerna at the beginning of the third century, though this supposition hinges on a single inscription. 118 The cohors I Thracum Syriaca may have stayed at modern Tutrakan, ancient Transmarisca. 119 Spaul and Velkov claimed that the cohors II Lucensium was based at Montana at the start of the reign of Domitian before its transfer to Abrittus around AD 86, 120 but the evidence for the cohort’s presence in either site is slim. 121 When the cohors I Thracum Syriaca returned to Moesia Inferior, it was probably based at Transmarisca. 122

II Gallorum might have been there from AD 99-112, 130 though it might also have been the cohors IIII Gallorum. 131 We close with the Dobrudja and the Black Sea. The cohors I Lusitanorum Cyrenaica may have been at Cius in the first half of the second century. 132 Karavas suspects that the ala I Vespasiana Dardanorum was at Arrubium in the second century. 133 The ala I Pannoniorum may have been based at Troesmis, though only briefly around AD 101/102. 134 The cohors I Ubiorum was likely at Capidava in the first half of the second century, at least until around AD 143. 135 Thereafter, the cohors I Germanorum cR was probably based at Capidava during its stay in Moesia Inferior. 136 Both Spaul and Karavas allege that the ala II Hispanorum et Aravacorum was based at Carsium when it came to Moesia Inferior late in the first century 137 on the basis of two building dedications were found at Carsium, one of which is dated to AD 102-103, but is fragmentary. 138 There are a couple of other inscriptions from Carsium, 139 and some limited evidence for the unit in other Lower Moesian sites. 140 On balance, however, Carsium was likely the unit’s base. 141 The ala Gallorum Flaviana might also have been based at Carsium for a time. 142 The cohors II Mattiacorum likely garrisoned Barbosi in the middle of the second century and beyond. 143

Spaul, Zahariade, and Gudea asserted that early in the third century the cohors I Lusitanorum may have moved to Nigrinianis-Candidiana from Cius. As is often the case, the evidence is slight: there is but one published inscription from the site and it does not name the unit. 123 Gudea has characterised the site as an auxiliary fort, and argued that, at least in the third century, the cohort of Lusitanians was based there. 124 There is a lone fragmentary inscription, found in modern Lomets, that mentions a legate of the cohors I Lusitanorum, and it has been restored to include “MAXIMINIANAE”, suggesting a third century date. 125 Lomets is some distance from NigrinianisCandidiana, however, and so the attribution must be dismissed until further evidence comes to light. Staying with Lomets, Zahariade and Gudea argue that by AD 198 the cohors II Mattiacorum had moved its headquarters from Barbosi to Troianhissar, presumably near the village of Lomets, which would make for quite a significant move. 126 We have little clear evidence of this, however.

Finishing with the coast, the ala Gallorum Flaviana may have been based at Tomi when it was in Moesia Inferior. 144 The ala Gallorum Atectorigiana might also have been based at Tomi around AD 224, and possibly before then. 145 But, the decurion who set up the career inscription alluded to in the previous sentence may

The cohors I Cilicum remained at Sacidava into the third century. 127 the cohors I Sugambrorum equitata veterana was based at Sucidava, though perhaps only briefly. 128 That same unit seems to have been based at Montana by some point in the first quarter of the second century. 129 Sacidava was probably the home of two cohorts in the second century: the cohors I Cilicum equitata milliaria was there for about 150 years beginning in the early second century; the cohors

130

Zahariade and Gudea 1997: 46. AE 1981, 745=IScM IV.169. 132 Karavas 2001: 125. See CIL III.12480=IScM V.118. 133 Karavas 2001: 123-124. This supposition is based on the lack of evidence for the unit anywhere else. 134 Strobel 1984: 113. 135 Karavas 2001: 126. AE 1950, 46=AE 1960, 360=ISsM V.24; AE 1997, 1353. 136 Radnoti 1959: 149; Covacef 2000; Spaul 2000: 256; Karavas 2001: 126. See AE 1939, 87=IScM V.16; and AE 1950, 76=IScM V.36; AE 1997, 1329. 137 Spaul 1994: 35; Karavas 2001: 126. 138 The partial inscription is AE 1980, 814 (=IScM V.94) whereas the inscription that names the unit is CIL III.7603a (=IScM V.95a). The second inscription, however, dates to AD 200. 139 AE 1980, 815 and AE 1960, 333. The second inscription names a veteran. 140 CIL III.6218=IScM V.253 (Arrubium); CIL III.12359=ILBulg 120 (Gauren); CIL III.14214-22=IScM V.117=AE 1977, 763 (Cius); CIL III.14214-29=IScM II.225 (Tomi); IScM V.23=AE 1934, 106 (Capidava). 141 Strobel (1984: 112, 188-189) merely states that the unit was involved in construction at Carsium around the end of the 1st Dacian war. Cf. Radnoti 1959: 146. 142 AE 1992, 1496=CERom XI.566; CERom XVII.745a-b. 143 Radnoti 1959: 149; Zahariade and Gudea 1997: 46; Karavas 2001: 121. Karavas says that the cohort garrisoned the site sometime after AD 145. See IScM V.308=CIL III.7620; AE 1974, 562b. 144 Spaul 1994: 115. Cf. CIL III.7557 (fragmentary inscription). 145 Spaul 1994: 49. See CIL III.6154. 131

117

Zahariade and Gudea 1997: 47; Karavas 2001: 131. IScM V.218=AE 1980, 822. 119 Eck and Roxan 1997: 197. This suggestion is based on the presence of an altar set up by a prefect of the unit (AE 1939, 101). 120 Spaul 2000: 83-84; Velkov 1990: 247-256. 121 There is a single epitaph that names a former singularius of the cohors II Lucensium, which was found at Abrittus (AE 1925, 66). 122 Radnoti 1959: 149; Eck and Roxan 1997: 197. Cf. AE 1939, 101 (an altar). 123 Spaul 2000: 60; Zahariade and Gudea 1997: 47; Gudea 2005: 432. IScM IV.154=AE 1972, 504. Note the comments of Gudea (2005: 353). 124 Gudea 2005: 432. 125 AE 1964, 180. Cf. Matei-Popescu 2010a: 221, n. 1950. 126 Zahariade and Gudea 1997: 47; Gudea 2005: 394. See CIL III.14428. 127 Zahariade and Gudea 1997: 46. Cf. Scorpan (1980: 98-102) who cites some inscriptions naming the unit found during excavations in 1979. 128 Zahariade and Gudea 1997: 46. 129 Rankov 1983: 42; Strobel 1984: 143. 118

69

Exercitus Moesiae which names an unattributed veteran. 152 Some have speculated that Lederata housed a few auxiliary units, including the ala I Claudia, the ala II Pannoniorum, and the cohors II Hispanorum. 153 Far more of the inscriptions that we have suggest that, if anything, a vexillation of the legio VII Claudia garrisoned the site, however. 154 There is evidence, at least, of late antique occupation. 155 There is also evidence for auxiliaries at Translederata, 156 for there are hints of the ala II Pannoniorum, 157 the cohors I Cretum, 158 and the cohors II Hispanorum. 159 Indeed, the evidence for these units, and the legiones IIII Flavia 160 and VII Claudia, 161 suggest that the fort was part of Moesia Superior, not Dacia. Next we come to the fortlet at Boljetin. We have one inscription from the site, and it names both the legio IIII Scythica and the legio V Macedonica, which suggests that vexillations of those legions might have operating from this location. 162 Bononia, a 20ha fort that housed the cuneus equitum Dalmatarum Fortensium in late antiquity has left virtually no evidence for a garrison earlier than that with one possible exception, noted above. Indeed, it might well be a late antique fort. 163 A lone inscription, a diploma, was found at the site that dates to AD 159/160. 164 But it is fragmentary, and it gives us no indication which unit the soldier who owned it belonged to. Egeta might have been garrisoned either by a vexillation or an auxiliary unit, at least on the basis of the surviving inscriptions. We have two inscriptions that name the cohors I Cretum, and two that name the legio VII Claudia. 165 Of those, one each for both the cohort 166 and legion 167 might be considered official in any capacity.

have done it in the patron’s hometown, and not necessarily the unit’s station. 146 VEXILLATIONS In our discussion of legionary emplacement in the Flavian dynasty and beyond, a new problem manifested itself, namely the problem of vexillations and their bases. For the earlier period, and to a certain degree later periods, we often think of legions based in one central, larger fort. We also know, however, that there is an abundance of smaller forts across the empire, and while many of them surely did house auxiliary units, it seems likely that they might have housed vexillations too. If a detachment from a legion, or auxiliary unit for that matter, is operating in its own province does it still count as a vexillation? And if it is based somewhere else (i.e., not in the primary legionary headquarters), does this still count? Can we say when legionaries were split up? Is the evidence secure enough (with respect to the date of particular inscriptions or the date of the phases of construction of particular fortifications) to pinpoint, or even give a reasonable estimate of, when a unit moved into these various spaces? To give but one example, there is evidence from Dacia for vexillations from the legio II Adiutrix and the legio VI Ferrata operating as early as the reign of Trajan at Sarmizegethusa Regia. 147 Bearing these questions in mind, in this section we turn to those forts and fortifications that we have not yet discussed, which includes those that might be possible bases for vexillations (legionary in particular), auxiliary units or some combination of the two. As alluded to above with respect to Dacia, a number of socalled vexillation forts have been identified. 148 We will limit ourselves, however, to those that at least offer some evidence of their garrison. 149 While, on the surface, it would seem that a smaller fort should be associated with smaller units like auxiliae, there is no evidence for a direct link between fort size/plans and unit type. 150 As with the previous section, we will start in the west (Moesia Superior) and work our ways east.

We now turn to Moesia Inferior, and as above in many instances the evidence is too paltry for any firm conclusions about garrison to be drawn. Take, for instance, the fort of Utus: we have a lone inscription that names a veteran of the legio V Macedonica, 168 which is insufficient for identifying the presence of a vexillation. In other cases we are more fortunate. Moving east, a handful of traces have been left at Iatrus for both legionaries and auxiliaries. Inscriptions naming centurions, soldiers (miles/milites), a signifer, and veteran of the legio I Italicae have been found

In our discussion above, we covered the majority of the known fortifications in Moesia Superior; there are only a few outstanding examples that we have not yet discussed. We begin with Aureus Mons, which we know was garrisoned by a cunues equitum Dalmatarum in late antiquity, 151 but has left no other clear trace of any possible earlier unit. We have a lone inscription

152

AE 2008, 1177. Gudea 2001: 56; Jeczmienowski 2012: 39. 154 Note, in particular, CIL III.1643. See too AE 1912, 79; CIL III.8275-1b; CIL III.1644. 155 Mladeonovic 2012: 181-182. 156 Gudea 2001: 57. 157 CIL III.8074-5b=IDR III.1.5=AE 2005, 1297. 158 IDR III.1.6=AE 1912, 78. 159 CIL III.8074-20=IDR III.1.7a. 160 IDR III.1.7b. 161 CIL III.8071c-d=IDR III.1.7c. 162 CIL III.1698. See too Gudea 2001: 67-69. 163 Not. dign. or. 42.13; Jeczmienowski 2012: 37. 164 CIL XVI.111. 165 Legion: CIL III.6324d; CIL III.12676. Auxiliary unit: AE 1981, 737; IlJug II.466=AE 1966, 366=AE 1968, 453. See Gudea 2001: 89. 166 IlJug II.466=AE 1966, 366=AE 1968, 453. 167 CIL III.6324d. 168 IBulg 128=AE 1935, 74. Cf. Gudea 2005: 416. 153

146

This is the assertion of Spaul. Stefan 2005: 349-355; Wheeler 2010: 1201. 148 Bishop 2012: 140. 149 That means that all those forts, fortifications, and fortlets that have left no indication of their garrison have been excluded. That also means that I have excluded those sites where our only evidence for the presence of the military comes in the form of an epitaph, which is suggestive, but hardly conclusive. 150 Haynes 2013: 53, n. 10. Cf. Richardson 2002. 151 Not. dign. or. 41.15. 147

70

Troop Emplacement in the Moesias there, and in significant numbers. 169 There is good reason to believe, then, that a vexillatio of the legio I Italicae was operating out of Iatrus. 170 There is also sparse evidence for the cohors I Cretum, but the inscription in question is an epitaph, and so should not be understood as direct evidence for an auxiliary garrison. 171

Moesica. 188 The last place that we come to is Arrubium, where a detachment of the legio V Macedonica was active. 189 None of the results of the previous discussion should surprise. We have long known that Roman soldiers were active in areas that often extended well beyond established bases; moreover, all the soldiers in a unit were rarely housed in the same structure. Indeed, it is worth remembering that this discussion has been limited to those fortifications where we have evidence for individual units. Just because we lack that evidence in other sites, that does not mean that units or detachments, whether they be of legionaries or auxiliaries, were not used in those spots. A glance at either Jeczmienowski’s paper on Upper Moesia fortifications or Gudea’s catalogue of Lower Moesia fortification shows just how many fortifications there were in those two provinces, and how few we have discussed here. 190 Fortifications of various types, whether larger legion-scale forts, smaller auxiliary or vexillation-scale forts, or even smaller fortlets and towers, were found all along the Danube, though also too along a few major routes in the interior. 191 And Roman legionaries and auxiliaries, among others, would have peopled those fortifications. Now we need to discuss why the assorted legionary and auxiliary units were based where they were. To answer that, or at least to try, we will take a look at the role of logistics in determining location.

Above we discussed Sexaginta Prista in the context of its potential auxiliary garrison. It seems likely that legionaries were operating out of the site too in the form of a vexillation of the Eleventh Claudian Appiaria too might have housed a legion. 172 vexillation of the Eleventh Claudian for a period of time. 173 There is plenty of evidence for vexillations operating out of Sucidava. 174 Some were comprised of the legio IIII Flavia, 175 some of the legio V Macedonica, 176 some of the legio VII Claudia, 177 some of the legio X Gemina (a legion based in Dacia), 178 and others of the legio XIII Gemina (another legion based in Dacia). 179 Moving a bit farther to the east, we find some evidence for legionary vexillations operating out of Sacidava too. 180 In this case, if they were present they likely consisted of vexillations drawn either from the legio Italica 181 or the legio XI Claudia. 182 The next destination is Capidava, where there is some evidence for the presence of vexillations 183 of the legio V Macedonica 184 and the legio XI Claudia. 185 We also find there the aforementioned vexillatio Capadavensium. 186 As we noted in chapter four, we do not know where the soldiers that comprised the unit were drawn from, though the presence of inscriptions that point to the activities of the legiones V Macedonica and XI Claudia might point towards a solution. A bit farther north on the Danube was Carsium, where a vexillation of the legio I Italica might have been active. 187 This would have been along with the handful of auxiliary units noted above, not to mention a detachment from the classis Flavia

TROOP EMPLACEMENT AND LOGISTICS A possible link between the disposition of troops and the construction of the road-network is likely to be found in the need of the Roman high command to supply troops for war. 192 Tacitus, in his discussion of the turmoil in the Bosporus during the reign of Claudius, points out in regard to the terrain of the region, that “he [Claudius] would be undertaking a war in a roadless country [avio itinere] and upon a harbourless sea.” 193 Indeed, Adams found a correlation between road-construction and impending warfare. 194

169 AE 1985, 762; IBulg 338; IBulg 343; AE 2003, 1540=Iatrus 13; Iatrus 17; Iatrus 52-61, 93-98. 170 See too Gudea (2005: 426), who suggests that a numerus Syrorum was also active there. 171 AE 2004, 1252=Iatrus 16. 172 AE 1944, 6-9; AE 2005, 1333; AE 2010, 1418. Note Gudea (2005: 428). 173 CIL III.1257=IScM IV.143g. 174 Gudea 2005: 441-442. 175 AE 1966, 325. 176 AE 1939, 90; AE 1944, 66; AE 1976, 582a-b; AE 2003, 1527; AE 2006, 1180a; CIL III.8066b-c; IIFDR 280, 282, 283a-b, 285, 286. 177 AE 1976, 582d; CERom II.105; CERom VI.388b; CERom XVII.733a; CERom XVII.733b; IIFDR 287; ILD 120, 122. 178 IDR II.237. 179 AE 1950, 75a; AE 1976, 582c; AE 2006, 1180b; AE 1944, 65; IIFDR 288. 180 Gudea 2005: 443-444. 181 IScM IV.200; 182 IScM IV.201a-b, 203 183 Gudea 2005: 448. 184 IScM V.53. 185 IScM V.54a-b. 186 AE 2004, 1278=CERom XXV.1139. 187 IScM V.113; AE 1998, 1147=CERom XIX/XX.909; Gudea 2005: 449.

Logistics, of course, is a big topic, and we cannot hope to discuss the issue in any depth here. 195 Instead, we will limit ourselves to one element of the supply chain, namely the food needs of the military. This is not something we will evaluate in detail; rather, we will highlight certain aspects of food supply for the 188

AE 1998, 1146=CERom XIX/XX.908. IScM V.254a-b; Gudea 2005: 453. 190 Gudea 2005: 411- 505; Jeczmienowski 2012. This does make me something of a minimalist. While I obviously do not deny the presence of soldiers in those other sites, in the absence of clear-cut evidence, or close to it, I hesitate to make too many speculations. 191 Gudea 2001; Karavas 2001; Gudea 2002; Jeremic 2007; Ivanov 2012; Jeczmienowski 2012. 192 Bishop 2012: 38. 193 Tac. Ann. 12.20. trans. John Jackson. 194 Adams 1976 195 Important studies include Adams 1976, Kissel 1995, Junkelmann 1997, Roth 1999, and Stalibrass and Thomas 2008. 189

71

Exercitus Moesiae purposes of illustrating the general supply issues that military units faced with a view to using this insight to gain some understanding of Roman decision making, at least with respect to troop emplacement.

their goods on their own, without the assistance of civilians in times of war. Hunt’s Pridianum is unique because it gives us not only a list of absentee soldiers, but it also gives us some indication of the sorts of activities that those very soldiers might have been involved in. Yet, it is not the only document that we have in which soldiers are dispatched a considerable distance from their station. We also have a strength report from Vindolanda for the First Cohort of Tungrians. 202 Although this remarkable document does not provide us with the sorts of activities that Hunt’s Pridianum does, it too lists the number of soldiers absent. This includes the number of soldiers who were in London, the number who were in Coria [Corbridge], and the number who were at the office of a certain Ferox, whom Bowman thinks is a high-ranking officer that may have been based in York. 203 We may never know the reason for the posting of those soldiers so far afield; it would not be too great a stretch, however, to imagine that those soldiers dispatched to London, the provincial capital, were there to pick up supplies as per those from the Moesian unit sent to Gaul. 204 So, we can say with some confidence that it was not unusual for soldiers to be dispatched considerable distance from their stations to procure and/or transport supplies.

The consensus among scholars is that the staples were grain, meat, wine, and oil; however, wine (vinum) and grain (frumentum) could be substituted by sour-wine (acetum) and biscuit (bucellatum). 196 Roth further divides the components of the soldier’s diet into the grain ration (frumentum) and the ration of other food, namely the cibaria. Under cibaria, the non-grain ration, he includes five further categories: meat, and particularly salt-pork or bacon; vegetables, and particularly lentils and beans; cheese; salt; and sourwine. 197 A soldier’s diet, particularly an officer’s, was not necessarily restricted to those items; moreover, there could be considerable variation within those categories, depending on the local conditions and the cultural-background of a soldier. Interestingly enough, at some sites, at least, both officers and foot soldiers seem to have had relatively equal access to the meat available. 198 While we know what soldiers were eating, some details about the nature of its supply continue to pose problems, despite the great deal of attention that the topic has received. 199 As it stands, the documentary evidence provides some clues. We are fortunate to have the aforementioned Hunt’s Pridianum, which gives us some insight into the operations of one auxiliary cohort, the cohors I Hispanorum Veterana. 200 On this papyrus, dated to AD 100 – 105, we find a list of absentees, including soldiers active both in the province and abroad. Some of the soldiers were in Gaul to get clothing, presumably in the same area to get what Fink identified as grain, across the Erar river to get horses, and in Dardania at the mines. 201 Those soldiers still on the Danube, but absent from the fort, are described as being across the Danube on an expedition, likewise (item) to defend the crops (annona[m]), at the grain-ships (ad naves frumentarias), to the Haemus [mountains] to bring cattle (ad armenta adducenda), and on guard over draft animals. Thus, we have here a snapshot of army-life. Soldiers, though obviously under the direction of higher-ranking officers, seem able to attend to many of their supply needs through their own efforts. It must be noted, however, that this document is concurrent with Trajan’s First Dacian War, and so it may not be representative of the army’s usual practice. Regardless of that fact, the evidence of Hunt’s Pridianum still suggests that the soldiers were capable of procuring

The supply needs of the Roman military were undoubtedly determining factors in the disposition of the units and the expansion of the transportation network, and there were three ways that the military usually obtained these: by importing what they needed from afar; by obtaining what they needed locally; or by some combination of the previous two. Knowing something about the agricultural sustainability of the region goes some ways towards answering this. We saw in the introduction above that the Moesias – at least the modern countries that occupy the same land – are capable of producing the food needed to sustain the military. Does this mean that they necessarily did this on all occasions? There are a number of things we would need to know to answer this satisfactorily, and arguably one of the most important is whether the region’s fortifications would have had the manpower required for producing goods locally, and if not, whether they could rely on their neighbours, particularly the civilian communities of the region. 205 202

Tab. Vindol. 1.154. Despite those soldiers dispatched at Coria, London, and the office of Ferox, there are four further lines which scholars have not been able to restore, but originally did identify the whereabouts of several other absentee soldiers. See: Bowman 1994: 105. 204 By around AD 90, the Roman capital of Britain was London. This document is dated to AD 92 – 97 (Bowman 1994: 22), which was not a period of instability in Britain. This then precludes the possibility that the situation presented in this document would be one unique to times of war. That conjecture is also supported by the document itself, which lists only 31 sick or wounded out of the total of 752 (Bowman 1994: 23; Tab. Vindol. 1.154). 205 Here, I am thinking of the villages (vici), townships (vici and canabae), and other various settlements (municipia and so on) that may have dotted the landscape both near the various forts and fortresses, and beyond. 203

196 See Roth 1999, Goldsworthy 1996, Le Roux, 1995, Davies 1989, Jones 1964: 628-629 and 1261 note 44. 197 Roth 1999: 26. Cf. Davies 1989. 198 Toplyn 2006: 474. 199 See, for instance, Adams 1995, Le Roux 1995, and Whittaker 2004. 200 Fink RMR 63. 201 Fink (RMR 63) notes that the Erar river has not yet been identified.

72

Troop Emplacement in the Moesias It is worth stressing too that the logistical needs of the Roman military were not only significant in quantity, but also quite varied in content. Soldiers would need everything from food and water to military equipment, weaponry, and, of course, building construction/maintenance (stone, wood, etc.) and cooking materials (firewood, oil, utensils, etc.). 206 Consequently, when considering the logistical needs of the military, the full gamut of supplies should be considered.

CONCLUDING THOUGHTS While it was not always the case, in most circumstances careful consideration seems to have gone into the positioning of Roman military fortifications along the lower Danube, and their concomitant soldiers, at least based on what we know about where they were. There is little doubt that a sound awareness of Moesian geography was required to construct the frontier military network in the one, later two, province/s. In his book on rivers, Campbell stressed the importance to commanders of knowledge of local topography and the geographical area. 214 He notes too that locations were chosen to ease movement, protect river crossings, secure communications, and ensure a stable water supply. 215 With respect to the Danube, Campbell notes, the locations were connected to concerns over hostile peoples, and that they were put in spots where Roman forces would control movements – at the beginning and end of floodplains and mountainous regions. 216 Indeed, in general, though particularly from the Iron Gorge to the Black Sea, the southern bank of the Danube, the Moesian side, is often steep and high, while the southern bank is low-lying and prone to floods, and filled with marshes and swamps. 217 It always made more sense, then, to concentrate soldiers on the southern, or right, bank, and in particular at important crossing points, such as where various north-south tributaries emptied into the Danube. We can go further, for Lemke, who has focused specifically on Moesia Inferior, noted that of the majority of the fifty or so fortified sites, nearly all were on situated on communication routes. 218 The ease with which they could be supplied was also an important factor. 219 It is also worth stressing that many of these fortifications were established in locations regularly used by invasion forces crossing the river, and so were there to check these sorts of movements. A good example of this is Novae. 220 Ultimately, more often than not a great deal of thought determined where the Romans situated their fortifications.

What does all this discussion of logistics have to do with the Moesian army? Undoubtedly a significant consideration when choosing locations for military bases was the ease with which a site could be supplied, for what the varied evidence shows is that the Romans were very concerned with the logistical needs of their troops – and eager to ensure that they could maintain the necessary levels of supplies regularly. 207 While it seems likely that the Moesias could supply a good deal of this material eventually, 208 a significant portion would have had to be imported, and there were two principal means of transporting goods in the Roman Empire, by land and by sea. It is no coincidence that the bulk of the provinces’ bases were located along the Danube, or some other major transportation route (other river, road). Rivers played a fundamental role in military supply. 209 Travel by water was often cheaper, safer, and quicker, though there are exceptions. 210 With that said, the Romans constructed a road that connected the fortresses along the Danube – so giving them options – as well as a number of others, some which ran parallel, others which ran northsouth. 211 Indeed, the major road along the Danube joined up with the older, and more established, road which ran along the Black Sea coast, where the region’s oldest Graeco-Roman settlements happened to be as well. 212 This riverine and road connectivity would enable the military units stationed on the frontier, where the bulk of Moesia’s soldiers happened to be, to access with comparative ease the goods that they needed – and to move around, and communicate with those in, 213 their environment without too much difficulty.

Now that we have looked at one important reason why soldiers might have been based where they were, it is time to take a look at some of the other considerations, namely those pertaining to strategic interests. For while logistical concerns undoubtedly influenced Roman decision making, there were other matters that warranted careful consideration. Indeed, in his overview of Roman legionary bases, Bishop argued that there were three points that determined the location of legionary bases: strategic concerns, tactical concerns, and topographic concerns. 221 The military threat posed by the region’s multitudinous nomadic

206

Note the comments of Matthews (2015: 611). Note the comments of Bishop (2012: 38), who is specifically concerned with legionary bases. 208 Note Matthews (2015), who argues that while the military communities in the Dobrudja could support quite a lot of the food needs of the soldiery, this only amounted to about 2/3 of their needs. Thus, the rest would need to be imported. 209 Campbell 2012: 177-180. 210 Batty 2007: 104-105. On the costs and benefits associated with land transport, at least in Egypt, see Adams (2007). 211 Panaite 2015: 593-596. 212 Panaite 2015: 593. 213 Campbell 2012: 172. 207

214

Campbell 2012: 161. Campbell 2012: 172. 216 Campbell 2012: 176. 217 Batty 2007: 72; Lemke 2015: 845. 218 Lemke 2015: 846. 219 Lemke 2015: 846. 220 See Lemke 2008. 221 Bishop 2012: 38. 215

73

Exercitus Moesiae peoples had a significant impact on deployment decisions in the Moesias, as we will see in the next chapter. 222

222

Note Batty’s (2007: 442) comments.

74

PART II: RESULTS AND DISCUSSION

Strategy of the Roman Empire. He has, of course, been challenged by, among others, Mann, Isaac, and Whittaker. 4 Some have focused on the nature of the frontier itself, 5 preferring to characterise it, for a variety of reasons, as a zone of interaction rather than a straight line comprised of military units. Others have focused on the nature of the operations of the units based on those frontiers, 6 regardless of the varied frontiers’ character, and have preferred to see the soldiers engaged in the control of movement as well as pseudo-policing duties. On the other hand, not everyone is convinced by such exceptions, at least not entirely, with Wheeler one of the chief dissenters – and so supporters of Luttwak. 7

Chapter 6: Rome’s Strategy on the Lower Danube In chapters one to three, and to a lesser degree four, above, we were able to evaluate in a number of instances the troop disposition of the Moesian provinces. Although the evidence is inconsistent, and irregular, we can get a sense of how that disposition changed over time for a portion of the period under review, and in particular between the years 40 and 170. Between those years, the evidence is significant enough for us to look at both legionary and auxiliary changes. In the introductory section of part three of his detailed study of the Pontic-Danubian realm in antiquity, Batty questions the value of a strategic overview of the region, and instead advocates a multifaceted approach. 1 As this is a book about the military organisation of the Moesian portion of that region, it behoves us to discuss Roman strategy in the region, especially considering the relationship between troop disposition and strategy. In this final chapter we do just that, and in particular focus on the following issues: the size and strength of the Moesian army/ies; Roman strategy in the Moesias; and the relationship between the Moesian armies and the wider empire.

Essentially Luttwak characterised Roman strategy on the frontiers as follows. 8 During the Julio-Claudian dynasty, there was little in the way of a defensive perimeter; the legions and auxiliaries were deployed to the peripheries, and client kingdoms were an important part of Roman policy. 9 From the Flavians to the Antonines Roman strategy could be characterised by: 10 “the maintenance of frontier security against lowintensity threats, the major business of the Roman army during much of the second century, called for lighter forces trained and equipped for guard, patrol, and escort duties as well as highly mobile but small-scale warfare. It is not surprising, therefore, that the proportion of auxiliary troops in the army seems to have increased during the second century”. 11 Finally, from the Severans and into the third century, the key feature of Roman policy on the frontiers was the employment of a defence-in-depth strategy, in which central reserves worked in support of troops based on the frontiers. 12 Admittedly, this third aspect deals with a period beyond the scope of this book.

ROMAN STRATEGY The Lower Danube was the site of significant expansion of the Roman imperium during the Principate. 2 When Trajan conquered Dacia early in the second century, the frontier Moesias became, at least in part, interior provinces of the Roman Empire. Ostensibly Dacia now filled the role of frontier province, while the Pannonias maintained their frontier identity. Similar changes to those in the Moesias have occurred on other frontiers. In Britain, too (in part), the limits of Roman control were extended while the troops that had been at those earlier borders now found themselves some distance from the new frontier. And yet, as Wheeler notes, on the lower Danube at least, we know that this push to the river led to regular confrontation with a host of different peoples: 3 military action was a regular, or at least semi-regular, occurrence.

There are additional views about strategy relevant to this discussion, and some which focus specifically on the Balkans. Wilkes, notably, identified three phases in Roman troop deployment in the Balkans. 13 The first ran from Augustus to Hadrian, and could be divided into four smaller stages: 14 the first involved military movements and the like in the wake of Augustus’ wars of conquest; the second involved the positioning of units in bases on the roads along the Danube and at major crossing points; the third involved the placing of those units (legionary and auxiliary) in camps along the

Much of the scholarly discussion of expansion has been tied to the prevalence – lo the existence – of longterm planning, however long-term it might have been. Scholars have put forth a number of views, many of which we will focus on here. The most influential work to have appeared on Roman frontiers and strategy over the past four decades or so is Luttwak’s Grand

4

Mann (1979), Isaac (1990), Whittaker (1994). Although not concerned with the army and strategy, Millar’s (1977) book on the emperor is of some relevance. Therein he advocates an ad hoc approach to imperial policy. 5 Whittaker 1994, for instance. 6 Isaac 1990. 7 Wheeler 1993a, b; 2010, 201. See too Ferrill 1991 8 For a concise overview of Luttwak’s position see Arguin (2015: 939). 9 Luttwak 1976: 7-50. 10 Luttwak 1976: 51-126. 11 Luttwak 1976, 121-122. 12 Luttwak 1976: 127-190. 13 Wilkes 2005, 148-168. 14 Wilkes 2005: 148-149.

1

Batty 2007: 350. The second part of Wheeler’s (2011: 192, and passim) extended discussion of Rome’s Dacian wars is effectively a response to Batty. We will discuss Wheeler’s (2011) arguments where relevant in this chapter. 2 On the changing nature of the term imperium in Roman thinking from republic to empire see Richardson (2008). 3 Wheeler 2011: 193.

77

Exercitus Moesiae Danube; and the fourth involved the extension of control into Dacia and the resulting changes elsewhere that went along with that expansion. Wilkes characterised the second phrase, from Hadrian through the third century, as static with defence the emphasis. 15 In the third phase, from the third century into the fourth and beyond, Roman strategy along the Danube was characterised by modification until the frontier’s disintegration late in the fourth century. 16

Romans were practical: changing policy does not indicate an absence of policy and, as noted elsewhere, bad strategy or the wrong strategy likewise does not indicate an absence. Armies can serve more than one function: frontier defense and internal security need not be contradictory. 20 The Size and Strength of the Moesian Army

There is one additional approach to strategy useful to our discussion, and it comes from Kagan’s broad discussion of Roman strategy in which she argued that we ought to understand grand strategy as “the setting of a state’s objectives and of priorities among those objectives, allocating resources among them, and choosing the best policy instruments to pursue them”. 17 To that end she said: We can focus our research on resource allocation. The epigraphic, archaeological, and narrative records combined show, with a high degree of resolution, what military units the emperors had and where they were deployed. Narrative sources and inscriptions allow us to interpret how these forces were used. We can trace major troop movements over time. There is no doubt that we have enough information to speak authoritatively about the allocation of the military resources of the state, a vital component of grand strategy. A study of the allocation of military resources, furthermore, can help to illustrate the emperors’ de facto priorities, if not their abstract goals. It can reveal their perceptions of changing threats. It can even clarify how, generally, they set out to ensure the security and, sometimes, expansion, of their state. 18 Kagan’s suggestions here provide us with a means of addressing the issue of long-term planning – that is strategy – and the duties of the soldiers that are the focus of this book. They also allow us to address and evaluate the aforementioned views of Luttwak and of Wilkes. 19 Indeed, after summarizing the troop totals for the Moesias, we will test the claims of Luttwak and Wilkes, and to a lesser extent Kagan.

In an attempt to determine something of Roman strategy, in this section we look at the potential size of the Moesian army over the centuries under review here using established paper strengths for both the legions and auxiliary units, and for this I have relied on the figures provided by Roth. Although these numbers should be used cautiously, they will help us to visualise changes in the size of the army. They will also provide some insight into the preferences amongst leading Roman military officials, especially with respect to the number of troops each province required, as well as the types of unit best suited to the particular conditions. Starting with the legions, according to Roth there were about 6600 persons in each, which included about 5280 soldiers. 21 With respect to the auxiliary units, Roth postulated that there were 512 men in a quingenarian ala, 22 1024 in a milliarian ala, 23 480 men in a quingenarian cohors, 24 960 in a milliarian cohors, 25 600 in a quingenarian mounted (equitata) cohors (480 infantry, 120 cavalry), 26 and 1040 in a milliarian mounted (equitata) cohors (800 infantry, 240 cavalry). 27 It is worth stressing, however, that even though we will use Roth’s figures, it is highly unlikely that any individual unit, whether legionary or auxiliary, ever had precisely that many soldiers enrolled. In what follows, we use the data provided by the tables found in chapters one, two, and three (and repeated in appendix one) that listed all those units in a given province in a particular year. Those years were AD 40, 78, 93, 100, 138, and 170. This does mean that we will leave out a few auxiliary units – in other words, some were not in Moesia during those years. On the other hand, providing a year-by-year account is not only impossible, but also impractical. So by using these snapshots, we should still get a representative picture of the changes in troop dispositions in the province, later provinces, so enabling us to look at changes over both the short and long term. In our discussion, we will be looking at quite a lot of information: the number of units per province and how this changed over time; the

In the article, Kagan focused on legionaries active at the time of the Trajanic wars as her resources. Here we will do the same with the notable addition of auxiliaries. We will also use a spattering of figures based primarily on our discussion in chapters one to four above. Ultimately, at least with respect to the lower Danube, we will address the much-debated issue of Roman imperial strategy. As we proceed, it is worth heeding Wheeler’s admonitions:

20

Wheeler 2011: 194. Roth 1994: 361. 22 Roth 1999: 336. 23 Roth 1999, 337. 24 Roth 1999, 337. 25 Roth 1999, 337. 26 Roth 1999, 338. 27 Roth 1999, 338. Cf. Cupcea and Marcu (2006), who agree with Roth’s calculations about the legions, but not about the auxiliary units. 21

15

Wilkes 2005: 149. Wilkes 2005: 149. 17 Kagan 2006: 348. 18 Kagan 2006: 346. 19 Cf. Kagan’s (2006: legionaries. 16

355-360) discussion of the Moesian

78

Rome’s Strategy on the Lower Danube number of soldiers per unit type (legionary versus auxiliary) and how this changed over time; and the differences in numbers between infantry and cavalry (of varying types) and how this changed over time. The tables will also include percentages. 28 Below are the tables that illustrate these changes in terms of unit totals and ideal paper strength soldier totals. Year

Legions Number of Units

40

2

78

3

Year

12 (6 ala, 5 coh, 1 coh eq)

Number of Units 2

100

2

138

2

170

2

Legions Number of Soldiers 10,560

93

Auxiliary Units Number of Soldiers

100

6072 (3072 ala, 2400 coh, 600 coh eq) 22 (9 ala, 15,840 11,448 8 coh, 5 (4608 ala, coh eq) 3840 coh, 3000 coh eq) Table 6.1. The Number of Units and Soldiers in Moesia 29

Legions

93

Auxiliary Units Numbers of Units

Year

Auxiliary Units Numbers of Units 11 (3 ala, 6 coh, 1 coh mil eq, 1 coh mil) 24 (3 ala, 18 coh, 1 coh mil eq, 1 coh eq, 1 coh mil)

Legions Number of Soldiers 10,560

138

Auxiliary Units Number of Soldiers

170

6416 (1536 ala, 2880 coh, 1040 coh mil eq, 960 coh mil) 10,560 12,776 (1536 ala, 8640 coh, 1040 coh mil eq, 600 coh eq, 960 coh mil) 12 (2 ala, 10,560 5824 (1024 10 coh) ala, 4800 coh) 15 (2 ala, 10,560 7824 (1024 12 coh, 1 ala, 5760 coh mil coh, 1040 eq) mil eq) Table 6.2 The Number of Units and Soldiers in Moesia Superior

Legions

Auxiliary Units Numbers of Units

Legions

Auxiliar y Units Number Number Number of Units of of Soldiers Soldiers 2 23 (9 ala, 10,560 11928 9 coh, 5 (4608 coh eq) ala, 4320 coh, 3000 coh eq) 2 25 (8 ala, 10,560 14,336 7 coh, 2 (4096 coh mil ala, 3360 eq, 8 coh coh, eq) 2080 coh mil eq, 4800 coh eq) 3 17 (5 ala, 15,840 13280 13 coh, 2 (2560 coh mil ala, 6240 eq, 4 coh coh, eq) 2080 coh mil eq, 2400 coh eq) 2 16 (5 ala, 10,560 10880 8 coh, 2 (2560 coh mil ala, 3840 eq, 3 coh coh, eq) 2080 coh mil eq, 1800 coh eq) Table 6.3. The Number of Units and Soldiers in Moesia Inferior

In some ways, the inclusion of AD 40 as our first snapshot is misleading, as our knowledge of the auxiliary units present at that time is incomplete. Nevertheless, it provides us with a starting point, and if we look to the second snapshot, AD 78, it should come as little surprise to see that there were far more attested units in the province by the end of the reign of Vespasian. While some of the increase is undoubtedly due to the gaps in our evidence, some of it must also be due to the increasing importance of the province in the eyes of the Roman state. After the split into two provinces, Superior and Inferior, the number of soldiers continued to increase, and the jump to AD 93 should likely be understood in terms of the military activity in the region that erupted under Domitian. The high numbers betray the growing importance and instability in the region, in this instance likely attributable to external forces, notably the problems with the Dacians. The civilian population of the Moesias, at least in those areas that housed soldiers, was too low at this stage for internal threats to be much of a threat. 30 The number

28

Colombo (2009), whose paper concerns some similar issues to those touched on here, also uses percentages (cf. Whately 2005). While the use of percentages based on ideal and theoretical troop numbers is not optimal, I do think they provide a good visual sense of the changes being considered. Given, too, that we are not concerned with absolutes at specific times, presence of snapshot dates aside, this should not have a significant impact on our results. 29 ala=ala; coh=cohors; eq=equitata; mil=milliaria.

30

79

See Mrozewicz 1982 and Mladenovic 2012.

Exercitus Moesiae of legions remained fairly consistent within the two provinces from AD 93 through AD 170 and beyond, with some exceptions.

Year

Moesia Moesia Inferior Superior c. 43% c. 57% 93 c. 39,464 c. 16,976 c. 22,488 c. 50% c. 50% 100 c. 47,120 c. 23,336 c. 23,784 c. 36% c. 64% 138 c. 46,104 c. 16,384 c. 29,720 c. 46% c. 54% 170 c. 39,824 c. 18,384 c. 21,440 Table 6.4. Comparison of Troop Totals in Moesia Superior vs. Inferior

One point that this data reveals is that when additional troops were needed, they were nearly always auxiliary units. Indeed, our evidence for auxiliary unit dispositions allows us to see that the Romans were preparing for an expedition against Dacia well before the first war under Trajan actually erupted. 31 This is how we should understand the huge increase in auxiliary troops in both Moesia Superior and Inferior that we can see took place between AD 78 and AD 93, and later AD 100. That the number of troops for Moesia Superior drops much more dramatically than it does for Moesia Inferior after Trajan’s Dacian wars, 32 and most of those units end up in Dacia, 33 lends support to earlier claims that Moesia Superior served as the launching point for Trajan’s wars. 34 It is striking too that quite a large number of these auxiliaries had been deployed to Moesia Superior by the mid 90s AD, which is well before Trajan’s expedition, but also after Domitian’s Dacian troubles. 35 In other words, Trajan’s wars were not ad hoc affairs; rather, the expeditions were prepared well in advance, as clear evidence as any of long-term strategic planning.

With the exception of AD 100 in the wake of Trajan’s Dacian wars, the garrison of Moesia Inferior was sufficiently higher than that of Superior throughout the history of the two provinces, at least for the period under review. Although parallels do not quite work, a good part of modern Serbia (perhaps 1/2 to 2/3) corresponds to Moesia Superior, and a good part of Bulgaria (perhaps 2/3, plus a chunk of Romania) corresponds to Moesia Inferior. Using this crude parallel, the current area of Serbia is 88,361 km2, while Bulgaria’s is 110,994 km2. The frontier portion of Moesia Inferior, which corresponds roughly to the length of the Danube River from modern Vidin to the delta and the Black Sea, is also much more extensive than that of Superior. The point is, to effectively monitor the frontier effectively more troops would be needed in Moesia Inferior than in Moesia Superior, and our evidence supports such an interpretation. Indeed, what this discussion brings to light is the relative worth of legionaries and auxiliaries.

What is, perhaps, most surprising is that troop increases – and all those noted thereafter – nearly always involved adding a large number of smaller units, not fewer larger ones. As noted, it was almost always auxiliary units, and to increase troop totals in the Moesias the Romans opted to bring in quingenary auxiliary units, not milliary ones, or even legions. Indeed, the number of milliary units operational in the Moesias remained small throughout, and fairly consistent. This is particularly striking in the case of the increase in the garrison of Moesia Superior in AD 100. What this suggests is that the Romans saw some value in using auxiliary forces rather than legionary ones to bolster their major campaigns, though whether this was a cost issue or a tactical one is not something we can answer at this moment, though we will return to it below, at least to some degree.

In the next set of tables, we compare the use of legionaries to the use of auxiliaries. In keeping with much of the research carried out in the past few decades, the auxiliaries should not be undervalued. 36 Year

The data in the next table compares the total garrison of the two provinces.

Legions

Auxiliary Units

40

c. 63%

c. 37%

78

c. 57%

c. 43%

Table 6.5. Proportion of Legionary Soldiers to Auxiliary Soldiers in Moesia 31

See too the discussion of Stefan (2005: 526-532). Though note Petolescu and Matei-Popescu’s (2008). 33 Take a look at the list provided in appendix two. 34 Matei-Popescu and Tentea 2006; Matei-Popescu 2015. 35 Possible examples include: ala Praetoria, ala II Pannoniorum, cohors I Cretum sagittaria, cohors I Flavia Bessorum, cohors I Montanorum c R, cohors I Thracum c R, cohors I Vindelicorum milliaria c R, cohors II Brittonum milliaria c R p f, cohors II Flavia Commagenorum, cohors II Gallorum Macedonica, cohors II Hispanorum, cohors III Brittonum, cohors IIII Raetorum, cohors VI Thracum, cohors VII Breucorum c R, and the cohors VIII Raetorum c R. 32

36

See, for example, Gilliver (1996), Haynes (2013), and Meyer (2013).

80

Rome’s Strategy on the Lower Danube Moesia Superior Year Legions

Auxiliary Units

That kind of epigraphic evidence is often limited, however. As a result, I have adopted a minimalist approach, which means that my list of mounted cohorts (actual and/or potential) is smaller than that proposed by other scholars. 41

Moesia Inferior Legions

Auxiliary Units

c. 62% c. 38% c. 47% c. 53% 93 c. 45% c. 55% c. 44% c. 56% 100 c. 64% c. 36% c. 53% c. 47% 138 c. 57% c. 43% c. 49% c. 51% 170 Table 6.6 Proportion of Legionary Soldiers to Auxiliary Soldiers in Moesia Superior and Inferior

Year Infantry Cavalry c. 12,960 c. 3672 40 c. 20,192 c. 7608 78 Table 6.7. Number of Infantry Soldiers vs. Cavalry Soldiers in Moesia Year 93 100 138 170

There were many more legionary soldiers operating in Moesia at first, though the relative balance had shifted between the legionaries and auxiliaries as we move from AD 40 to AD 78 (63% legionary to 57% legionary). After the creation of the new provinces, Moesia Superior ended up with a far greater proportion of legionaries than Moesia Inferior, 62% to 47%. In fact, for the first three snapshots, AD 93, AD 100, and AD 138, Moesia Inferior consistently had a higher number of auxiliary soldiers than legionary soldiers, while the opposite was true for Moesia Superior, save for AD 100, the year when the troop build-up for Trajan’s Dacian wars become apparent. Additionally, the proportion of legionaries to auxiliaries is not always, if ever, 50/50, the purported proportion empire-wide, 37 in any one province. Rather, if the Moesias are any indication, the Romans apportioned legionaries and auxiliaries in a province on a case-bycase basis, and in light of whatever needs a province might have.

Infantry c. 14,400 c. 20,160 c. 15,360 c. 16,320

Cavalry c. 2576 c. 3176 c. 1024 c. 2060

Table 6.8. Number of Infantry Soldiers vs. Cavalry Soldiers in Moesia Superior Year 93 100 138 170

Infantry c. 17,580 c. 12,808 c. 22, 680 c. 15,000

Cavalry c. 4908 c. 10,976 c. 7040 c. 6440

Table 6.9. Number of Infantry Soldiers vs. Cavalry Soldiers in Moesia Inferior Year Infantry Cavalry c. 78% c. 22% 40 c. 73% c. 27% 78 Table 6.10. Percentage of Infantry Soldiers vs. Cavalry Soldiers in Moesia

The last collection of data to discuss concerns the respective differences between the number of infantry soldiers and cavalry in the province, later two provinces, over time. To simplify matters (and hopefully complicate things only slightly), under the category of cavalry soldiers I have included all those soldiers in cavalry wings (alae) and mounted cohorts, so excluding the cavalry attached to legions and the infantry component of mounted cohorts, which would have been substantial. 38

Year Infantry Cavalry c. 85% c. 15% 93 c. 86% c. 14% 100 c. 94% c. 6% 138 c. 89% c. 11% 170 Table 6.11. Percentage of Infantry Soldiers vs. Cavalry Soldiers in Moesia Superior

Before I present the data, a few points of clarification on mounted cohorts (cohortes equitatae) are in order. As it happens, we have very little in the way of epigraphic evidence in the Moesias for cohortes equitatae, at least in terms of the titles of such units, a problem that exists empire-wide. 39 The title rarely appears in the diplomata; instead, we usually to have to rely on whether a soldier named on an epitaph that might refer to a mounted cohort is called an eques. 40

Year Infantry Cavalry c. 78% c. 22% 93 c. 54% c. 46% 100 c. 76% c. 24% 138 c. 70% c. 30% 170 Table 6.12. Percentage of Infantry Soldiers vs. Cavalry Soldiers in Moesia Inferior

37

See, for example, Rankov (2007: 71), who estimates that there were 160,000 legionaries in the empire in the second century (AD), and about 156,000 auxiliaries. 38 Again, these are not exact numbers, but approximations – they are meant to be representative, not exact. 39 I plan to explore the frequency of mounted cohorts across the empire in the future. 40 Saddington 1982: 174.

41

Compare the mounted cohorts listed in appendix two below with those found in Matei-Popescu’s (2010a: 193-236) history of the cohorts of Moesia Inferior. I list fewer mounted cohorts than he does.

81

Exercitus Moesiae only were more cavalry units operating in Moesia Inferior, but there was comparative variety: the Romans deployed comparable numbers of auxiliary cavalry wings (alae) and mounted cohorts in Moesia Inferior.

Year

Infantry Cavalry Units Units 8 6 40 11 14 70 Table 6.13. Number of Infantry Units vs. Cavalry Units in Moesia

Besides these standard units, there were also a few exceptional units operating in the Moesias. For instance, the ala I Pannoniorum et Gallorum operating out of Moesia Inferior gained at some undetermined time the epithet catafracta, which suggests that the unit became heavily armoured. It represents one of the earliest examples we have of Roman cataphracts. On the other hand, while none of the cavalry found in the Moesias are ever recorded as archers, we do find a handful of cohorts with the epithet sagittaria or sagittariorum, and they are predominantly found in Moesia Superior. The cohors I Antiochensium sagittaria (Moesia Superior), the cohors I Cilicum sagittaria (Moesia Superior), and the cohors I Cretum sagittaria (post war deployed to Dacia) were deployed in Moesia Superior, while the cohors I Flavia Commagenorum equitata sagittaria (post war deployed to Dacia), the cohors I Tyriorum sagittaria (later deployed to Dacia), and the cohors II Chalcidenorum sagittaria were deployed in Moesia Inferior. Half of those units were in the Moesias on the short term so to speak – there was a greater desire for archers in the Dacias than the Moesias, at least on the basis of the evidence.

Year

Infantry Cavalry Units Units 9 4 93 21 5 100 12 2 138 14 3 170 Table 6.14. Number of Infantry Units vs. Cavalry Units in Moesia Superior Year

Infantry Cavalry Units Units 11 14 93 9 18 100 16 11 138 10 10 170 Table 6.15. Number of Infantry Units vs. Cavalry Units in Moesia Inferior

We can conclude that cavalry were deployed much more regularly in Moesia Inferior than Superior, and that there was some variety in the kind of cavalry units deployed (alae versus cohortes equitatae, not to mention the unit of cataphracts). There was also a slight preference for archers in Moesia Superior over Inferior. So why the comparative differences in deployment over time and between provinces? In the next section we attempt to explain these variations.

Based on the data provided by these tables, the infantry always made up the highest proportion of troops used in the province, later provinces, which comes as no surprise. What does come as a bit of a surprise, however, is that, at least in these provinces, the increase in the number of cavalry troops being used over time is comparatively small and progressive. 42 If we assume that AD 100 is an anomaly because of Trajan’s Dacian wars, in Moesia Superior the proportion of soldiers goes from 85% (AD 93) to 94% (AD 138) to 89% (AD 170) infantry, and in Moesia Inferior from 78% to 76% to 70% infantry. As we might expect, the Romans have not applied a uniform approach. Even in this fairly confined area, there were significant variations in the deployment of cavalry in Moesia Superior and Moesia Inferior. There were far more cavalry soldiers and units deployed in Moesia Inferior than there were in Moesia Superior, and the numbers vary significantly. For instance, there were 10 auxiliary units in Moesia Inferior in AD 170, and only 3 in Moesia Superior, which corresponds to about 6440 cavalry soldiers to 2040. In AD 170 in Moesia Inferior, of those ten units (of sixteen total), five were alae, two were cohortes milliarie equitatae, and three were cohortes equitatae. Conversely, at the same time in Superior, of those three units (of fifteen total) two were alae and one was a cohors milliaria equitata. Not

The Moesian Strategy One of the things that we saw above was a significant increase in the number of troops deployed to the Moesias in the wake of, and in advance of, major wars against the Dacians (AD 80s and mid-100s). Indeed, the most obvious thing that the Roman soldiers were doing in those contexts was participate in combat situations against foreign opponents – and evidently with the express purpose of keeping them out, or at least in check. 43 There were some significant shifts in manpower over the period in question, both in terms of location, broadly speaking, and in terms of troop type. It also seems to be the case that the Romans were keen on shifting their strategy as the dynamics of these provinces changed. As we shift from the first century into the second century, units tended to be moved for

42

Thus, although we are only concerned with the Moesias, these results echo those of Colombo (2009).

43

82

See Goldsworthy (1996), for the army’s main purpose.

Rome’s Strategy on the Lower Danube the long-term, which therefore provides indirect evidence of long-term planning.

its approximate date, third century (and roughly between AD 222 and 235), 46 thanks to the beginning of the inscription, which reads, “LEG(IONIS) I ITAL(ICAE) [AL]EX[ANDRIANAE]”, and its name, which we have already noted (bellum bosponarum). 47 The inscription itself was found in Moesia Inferior. The dedicatee of the inscription claims to have suffered during the conflict, but survived nonetheless: “EIUS MULTIS PERICULIS IN BARBARICO LIBERATUS SIT”. Its location is hinted at in its name, Bosporanum, which would seem to allude to the Bosporan kingdom (and we know vexillations from the legio I Italica were operating out of the Crimea), and the reference to the barbaricum. 48 The barbaricum refers to the vague and little known area beyond the Roman Empire in southeastern Europe. It could well be that there were many more conflicts of this sort, which we know nothing about. That this particular bellum survives in only one inscription (no literary references) implies that the war (bellum) was unlikely to have been very significant, at least to the centre of power and the person of the emperor.

How does the evidence for the Moesias fit in with what scholars like Luttwak and Wilkes have argued? Beginning with Luttwak, we do have some evidence that units were dispatched to the periphery during the Julio-Claudian dynasty, though that evidence is insufficient. Our evidence is best for the next period that Luttwak discussed, which ranged from the Flavians to the Antonines, and it seems that he was right to argue for an increasing number of more lightly armed forces overall. On the other hand, some modification is needed, for while the numbers did increase, so too did the number of legionaries, and as we saw the biggest proportion of the more lightly armed auxiliary forces was found in Moesia Inferior. In other words, the Romans adopted their strategy to suit the context, or at least that is what the second century evidence tells us. Getting back to Wilkes, his suggestion that the units and the wider system in the Danube were static in the second century needs modification: while they do seem to have become static eventually, this was not until the last third, or even last quarter, of the second century. It is also worth getting back to Kagan’s suggestion. For the movement of troops also points towards active attempts by the Roman state to allocate resources, here in the form of military units, to pursue various priorities. From the AD 80s through the 90s, the priority was military action against the Dacians, and the evidence for movements reveals that the Romans had some sort of strategy to deal with them. After those wars, it was Moesia Inferior that saw the biggest increase in, or at least the biggest share of, troops in the region (the former province of Moesia). This likely reflects changing priorities too: Moesia Superior became an interior province, at least in part, while a good part of Moesia Inferior remained on the frontier – and it was the eastern half, where the bulk of the soldiers seemed to have been based. The fact that cavalry units were much more common in Moesia Inferior than Superior points towards mobility as a priority in the second century in this part of the region. That the cavalry component was a mixture of ordinary auxiliary cavalry wings and mounted cohorts also suggests that the desired mobility might have been in response to more than one need: external threats and internal security.

Another indication of the sorts of low-intensity operations that the Roman military might have been engaged in comes from the aforementioned Hunt’s Pridianum. While that papyrus details the events of one auxiliary cohort, the cohors I Hispanorum, presumably during a major war (Trajan’s Dacian wars), some of the activities described in the papyrus are of the sort that we would expect Roman soldiers to be participating in on a more regular basis. Some soldiers seem to have been killed in combat, for example. 49 Though I suggest below that brigandage might have been the cause, it might be the case that the references to guarding crops and draft animals are allusions to instances of small-scale conflict. 50 The comment, “trans danuuium in expeditionem”, might refer to one of Trajan’s Dacian wars, though it might be a case of low-intensity raiding. 51 Regardless of the nature of these encounters, it seems likely that casualties were part and parcel of this sort of activity. Although we do not have physical evidence for the violence suffered by Roman soldiers in the early and high imperial era, we do have evidence from the fourth century from the former province of Moesia, Viminacium in particular, for a soldier who had died in combat. 52 It is entirely possible that much more evidence for this sort of violence exists, only it has yet to be recovered. We now shift from external threats to internal ones. Indeed, one of the most obvious objectives for the soldiery was to suppress unrest in a general sense,

While large-scale wars were a part of the Roman military’s responsibilities, a significant amount of their attention would have been directed towards lowintensity threats, as suggested by Luttwak. 44 The sources, however, had little or no interest in conflict on a small scale, with only a few exceptions. One is the Bellum Bosporanum noted in chapter three above. 45 The only thing we know about this particular conflict is 44 45

46

See Sarnowski 1991: 140. AE 1991, 1378. 48 The most detailed discussion of the inscription and conflict is provided by Sarnowski (1991). 49 RMR 63.ii.11. 50 RMR 63.ii.31, 36. 51 RMR 63.ii.29. 52 Golubovic, Mrdjic, and Speal 2009. 47

See too Hoffman (2013) and Anders (2015). Batty 2007: 482.

83

Exercitus Moesiae whether that came from within or without. 53 Besides the low-intensity external threats, a good deal of the attention of the soldiers deployed in the Moesias would have been directed towards matters of internal security. The jurist Ulpian, for example, notes that it was the mandate of a governor of a province to purge said province of evil men, and to ensure that the province was pacified and quiet. 54 Some of the sorts of men he had in mind include those who committed sacrilege, bandits, kidnappers, and thieves (sacrilegos latrones plagiarios fures). 55 That governor had to keep order, and a good deal of his time was spent ensuring just this, with the exchanges between Pliny and Trajan a good indication of these goals, even if Bithynia is quite a different place from both Upper and Lower Moesia. 56 This is somewhat in keeping with the views of Isaac alluded to above. 57

that the peasantry of the region was largely comprised of cattle ranchers, shepherds, and even cattle rustlers, 65 then the conditions were ripe for significant internal instability of the sort that was likely to breed the kind of banditry that we find in other parts of the empire, like North Africa and the Near East. 66 As noted, we do have some epigraphic evidence for banditry in the Moesias, spotty though it may be. We have a few epitaphs from Moesia Superior, likely all dating to the second century, which record the victims of bandits. There was a certain Scerulaedus Sitaes, whose death came at the hands of bandits “INTERFECUS A LATRONIBUS”, 67 a Flavius Kapitonus, also killed by bandits “A LATRONIBUS ATROCISSIMA(M) MORTEM”, 68 and a Valerius Marcus, apparently killed by bandits “A LATRONIBUS INTERFECTUS”. 69 We also find latrunculi referred to in two inscriptions from Moesia Inferior. 70 Before we turn to the one pertinent papyrus, it should be stressed that while this might seem to be a small number of inscriptions, the three inscriptions that refer to a form of latrones represent a significant chunk of the total empire-wide (22 refer to latronibus), which is also the case for the two inscriptions that record latrunculi (the total is 15).

The Moesias, however, were quite different from the near east: the population density and population total, not to mention the level of urbanisation, were far lower in the Moesias. In Moesia Superior, for example, legions and auxiliary units were not usually based at sites of pre-conquest settlement; indeed, if anything settlement tended to follow the stationing of troops: civilians tended to abandon older settlements and move to military, or military-related, sites, such as vici, military roads, and former bases such as Ratiaria. 58 On the other hand, once these new settlements were established there is every reason to believe that the soldiers could have played a role in the maintenance of order.

As noted, Hunt’s Pridianum also records brigandage, for we read of a soldier killed by bandits: occisus a latron[i[bus. 71 The references to defending the crops and guarding over draft animals in that papyrus might We have also be an allusion to brigandage. 72 legislation that is designed to counter this sort of activity. 73 Indeed, as the papyrus makes clear, it would be up to the soldiers to enforce this legislation. The problem of banditry might well explain why a considerable number of soldiers were tasked with watching over the mines. Soldiers could be dispatched to watch over the movement of people and goods in regions of value to the emperor. 74 The mines at Mons Claudianus in Egypt, for example, were watched over by detachments of soldiers. 75 Some of the soldiers in Upper Moesia undertook similar activities. The mines of the province were important sources of materials, 76 and we find soldiers at sites throughout Upper Moesia, even after the creation of Dacia, in the interior of the The province and in significant quantities. 77 deployment of mounted cohorts throughout Moesia Inferior might also be connected with a desire to deal

And yet, banditry was a genuine concern, and it is a topic that Batty has devoted considerable attention to. 59 The truth is we have considerable evidence for the activity of bandits on the lower Danube, and Batty goes so far as to claim that it was, for all intents and purposes, rampant. 60 Our primary evidence for this comes from inscriptions, though the much discussed Hunt’s Pridianum hints at it too. 61 We even have the occasional literary reference, for Cyprian alludes to the perils of travel in the Moesias in the 250s, and Batty argues that what was true in Cyprian’s day was likely true earlier too. 62 Batty made a number of other salient points. He notes, for instance, Roman brutality during the conquest of Thrace, and the resentment that this is likely to have engendered. 63 That conquest, coupled with the conquest of Moesia, would likely have resulted in 1000s of slaves, and many disgruntled peoples left behind. 64 If Batty is right too in arguing

65

Batty 2007: 492-496. See too RMR 63.ii.35. Shaw 1984; Isaac 1990. 67 CIL III.8242=ILJug III.1434=AE 2010, 1400. 68 AE 1934, 209. 69 CIL III.14587=IMS III.2, 93=ILJug III.1343=ILS 8504=AE 1901, 19. 70 CIL III.12376=AE 1895, 60; CIL III.12483=IIFDR 238=ILS 724=AE 1891, 148. 71 RMR 63.ii.10. 72 RMR 63.ii.31, 36. 73 Dig. 47.8.2, 74.14.1-3. 74 Fuhrmann 2012: 208. 75 Fuhrmann 2012: 208. Cf. Bagnall 1977, Cuvigny 2005. 76 Note the following: Dig. 48.19.8-10. 77 Dusanic 2000. 66

53

Breeze 2010: 194-205; James 2011: 198-204. Ulp. Dig. 1.18.13.pr. Cf. Fuhrmann 2012: 151, 182. 55 Ulp. Dig. 1.18.13.pr. 56 Millar 1977: 114-115, 315-316, 325-328. 57 Isaac 1990. 58 Mladenovic 2012. 59 Batty 2007: 480-506. 60 Batty 2007: 480. 61 Batty 2007: 482. 62 Cyprian Ep. 68.3.3; Batty 2007: 484. 63 Batty 2007: 486. 64 Batty 2007: 486. 54

84

Rome’s Strategy on the Lower Danube with bandits, and the parallel that Davies’ drew between the Roman mounted cohorts and the earliest incarnation of the RCMP (Royal Canadian Mounted Police), earlier the North West Mounted Police, would then seem to be apt. 78

and troops from Moesia regularly transferred to Dacia and the east, as we have seen. 82 That these units could be moved wholesale with little apparent concern over their effectiveness in these new theatres illustrates this well. Focusing on the auxiliaries alone, the ala I Claudia nova miscellanea, the cohors I Flavia Hispanorum milliaria, the cohors I Thracum c R, the cohors I Vindelicorum milliaria c R, the cohors II Brittonum milliaria c R p f, the cohors III Gallorum, and the cohors V Hispanorum were all transferred from Germany to the Moesias, which might also be true of the cohors IV Gallorum, cohors V Gallorum, the cohors VII Gallorum equitata, and the cohors VIII Gallorum. Along the same lines, the cohors I Cantabrorum, the cohors I Lepidiana c R equitata, the cohors I Sugambrorum tironum, the cohors IIII Raetorum, and the cohors VII Gallorum equitata were all dispatched to the east, to say nothing of all those auxiliary units which comprised, in part, the garrison of Dacia, and which were transferred in the wake of Trajan’s wars and Hadrian’s reorganisation. Thus, for all that we must bear the provincial character of the Moesian army in mind, we must not forget that the units that comprised the Moesian army/ies remained part of a larger empire-wide military system.

Some have argued that the threats facing the Romans on the lower Danube were all low intensity: the soldiers and their fortifications were there to protect the Moesian provinces from small-scale raids rather than large-scale invasions. 79 Besides, the position of much of Lower Moesia, and to a lesser degree Upper Moesia, in relation to Dacia would seem to support this; the latter is likely to have been the front line in such cases from Hadrian to Commodus anyway. While we should not forget the impact of potential barbarian invasions on a large scale, a significant portion of the day-to-day military operations on the lower Danube would have been concerned with low-intensity threats and brigandage. Moesian Military Units and the Wider Roman Empire As we noted in chapter four above, troops could be used for operations in other theatres. From Vespasian to Commodus, there is some movement of legions and auxiliary units, with the reigns of Trajan and Hadrian the periods of greatest change. On the other hand, the reasons for these shifts are not always immediately clear, at least on the basis of the surviving evidence. The cohors I Flavia Numidarum in Lower Moesia moved to Lycia and Pamphylia, for example, though we do not know why. 80 In some instances units were raised and deployed to suit pressing needs such as trouble in Dacia, trouble with the Sarmatians, or Marcus Aurelius’ Marcomannic wars to name but a few. As noted just a bit earlier, that same emperor, Marcus Aurelius, saw the need for troops to monitor his mining operations in Dardania; 81 there were also a number of auxiliary units based in Moesia Superior that were created during his reign.

CONCLUDING THOUGHTS In closing, there are a number of ways that troop movements can be used to elucidate wider-reaching issues about strategy, imperial policy, and operations. Luttwak’s comments about an increase in auxiliary use over the second century need qualification. Also, although it is a question of scale, the disposition of soldiers in the Moesias was not entirely static – as Wilkes suggested – during the second century; some changes took place, even if they decreased in number after Marcus Aurelius. Getting back to Kagan, and to some degree Wheeler, it is also the case that the evidence of troop movements reveals a great deal about long-term strategic planning. While it is entirely likely that some movements were ad hoc, a great deal of consideration went into many of the transfers. It is also the case that the Romans gave due consideration to which types of units would be deployed in particular regions: it was not a one-size-fits-all mentality.

And yet, the fact is that for all the provincial character of the Moesian military (and that of other provincial armies), the units and their attendant soldiers remained part of a larger empire-wide military system. Up until the widespread use of vexillations later in the second century, entire units were regularly transferred from one part of the empire to another, with troops from Germany regularly transferred to Pannonia and Moesia,

Major wars aside, much of the Moesian army’s focus was directed towards smaller-scale conflicts, which admittedly have left little trace. On the other hand, we have a little more evidence for the military’s efforts at managing internal issues, such as brigandage. Some comments in Hunt’s Pridianum and a handful of inscriptions reveal that brigandage was a significant issue, and it seems likely the soldiers spent much of their time dealing with this problem.

78

See Davies 1989. Karavas 2001: 255ff. 80 RMD I.67, CIL XVI.128. See Matei-Popescu 2010a: 225. Spaul (2000: 156) suggests that the cohors I Lepidiana might have moved to Arabia, but this is conjecture. 81 HA Marc. 21.7. See Strabo 7.5.7; Plin. HN 3.26.149; and Ptol. Geog 3.9.1-4. The units are listed in RMD IV.247, which dates to September 9, AD 132; CIL XVI.111, which dates to AD 159/160; and RMD I.55, which dates to February 8, AD 161 – that is about a month before the accession of Marcus Aurelius. Cf. Cichorius 1900: 280; Beneš 1978: 30; Spaul 2000: 349; Hirt 2010: 166-201. 79

82

Note the discussion of Knight (1991), Eck and Pangerl (2005), Matei-Popescu and Tentea (2006), Bennett (2010), Juntenen (2013), Eck and Pangerl (2015: 230), and Matei-Popescu (2015).

85

were others we have virtually no evidence. On the other hand, some new auxiliary units were created, at least in Moesia Superior. In AD 170, the garrison for the two provinces had changed little from the last snapshot, AD 138. Moesia Inferior’s legionary garrison had dropped slightly to two from three, while Superior’s continued to be two-strong. In terms of their auxiliary garrisons, Inferior’s dropped slightly again, going from 18 to 17, while Moesia Superior’s increased slightly, from 12 to 15.

Conclusion: The Value of Traditional Roman Military History In this book we have examined the military organisation of the Roman province of Moesia, later Moesia Superior and Inferior, from the reign of Augustus to that of Severus Alexander. As we have seen, our evidence for troop movements in this region is thin before the Flavian dynasty, especially in the case of the auxiliaries. As a rule, the legions are much better documented. This makes it hard to say too much about how troop movements changed from year to year, or even decade to decade. Shifting to the individual snapshots, in AD 40 we found evidence for 2 legions and 12 auxiliary units in Moesia at that time. Nearly four decades later, there were now 3 legions, and 24 known auxiliaries. This increase is likely down to the growing importance of the frontier and the specific challenges that it posed, which is evidenced, in part, in the increasing number of conflicts in this area and beyond. Stationing troops on the lower Danube provided a ready body of troops that could be dispatched to points east, west, and north.

In chapter four we noted that Moesian troops were occasionally sent to the east for major campaigns, a practice which is in keeping with what we know about imperial strategy and troop deployment patterns. Indeed, for all the individuality of the provincial armies noted earlier, each respective army was still an integral component in the wider military system with reinforcements usually sent from the closest available region. Turning to vexillations, there is much more evidence for the presence of vexillations in Moesia Inferior than Superior, and it seems that the fluid character of the province’s boundaries as well as its close ties may have dictated the use of vexillations on the Black Sea coast. Indeed, Roman soldiers had a marked presence along the northwest shore of the Black Sea and into the Crimea. Ultimately, there is no doubt that vexillations from the Moesias, and Moesia Inferior in particular, were used increasingly in the second century in keeping with the pattern across the empire. In the case of Moesia Inferior, these vexillations were often sent to the Black Sea, and for an indeterminable amount of time.

We are much better informed about the movements in the years that run from Domitian to Marcus Aurelius, even if significant gaps remain, thanks in large part to the abundance of military diplomas. In our first snapshot, which corresponds to the year AD 93, there were two legions each in Moesia Inferior and Superior, which makes for a total of two for the former province. On the other hand, there might have been as many as 23 auxiliary units in Inferior, and there were 11 in Superior. Less than a decade later, in AD 100, although the number of legions remained the same, there had been a huge increase in the number of auxiliary units in Moesia Superior: 24, from an earlier total of 11. By AD 138, Moesia Inferior had three legions, while Moesia Superior still had two. Both provinces had seen a significant reduction in their respective auxiliary garrisons as well by this time: from 24 to 12 for Superior, and from 23 to 18 for Inferior. The change in troop numbers reflects the significant role that the provinces now played in Roman strategic thinking. This also reflects the considerable amount of conflict that beset the region.

While it was not always the case, in most circumstances careful consideration seems to have gone into the positioning of Roman military fortifications along the lower Danube, and their concomitant soldiers, at least based on what we know about where they were. There is little doubt that a sound awareness of Moesian geography was required to construct the frontier military network in the one, later two, province/s. The majority of the known fortified sites were situated on communication routes. The ease with which they could be supplied was also an important factor in their location; moreover, many of these fortifications were established in locations regularly used by invasion forces crossing the river, and so were there to check these sorts of movements. More often than not a great deal of thought determined where the Romans situated their fortifications. While logistical concerns undoubtedly influenced Roman decision making, there were other matters that warranted careful consideration. The military threat posed by the region’s multitudinous nomadic peoples had a significant impact on deployment decisions in the Moesias.

Thanks to changes in our quantity and quality of evidence and changes in general Roman troop movement patterns, we were not able to offer the same number of snapshots of the garrison of the two provinces in chapter three as we had done in the previous two chapters, one and two. Units became even more stationary than they had been in the reigns of Hadrian and Antoninus Pius; moreover, vexillations were used increasingly often. There were a few scattered hints of transfers, especially in light of the conflicts of the reign of Marcus Aurelius, but if there

In closing, there are a number of ways that troop movements, a very traditional topic in the wider category of Roman military studies, can be used to elucidate wider-reaching issues about strategy, imperial 86

The Value of Traditional Roman Military History policy, and operations. Indeed, the detailing of specific movements is important. For example, the Fourth Flavian Legion, which was one of those legions of Moesia Superior that had been in the province between 100 and 138, actually left for nearly a decade.

Ultimately, these sorts of details can tell us a lot about imperial planning. In the end, in the absence of commentarii and the like, focusing on troops movements can tell us a great deal about planning and strategy during the early and high imperial periods.

87

Appendix 1: Summary Tables of Legions and Auxiliary Units These are the same tables found throughout chapters one to three, reproduced here in one place for the sake of convenience. Because these tables represent snapshots, not all auxiliary units discussed in this book make it into the tables, though most do. Legions (2) IIII Scythica, V Macedonica

MOESIA Auxiliary Units (11) ala Asturum (?), ala Atectorigiana, ala Bosporanorum (?), ala Pansiana, ala Scubulorum, cohors I Cilicum sagittaria eq, cohors Cisipadensium (?), cohors Cretum, cohors Lusitanorum (?), cohors I Sugambrorum veterana eq, cohors II Varcianorum

Dispositions of Legions and Auxiliary Units in Moesia in AD 40 Legions (3) I Italica, V Macedonica, VII Claudia

Auxiliary Units (21) ala Asturum, ala Atectorigiana, ala Claudia nova, ala Dardanorum, ala I Gallorum, ala Gallorum Flaviana, ala Hispanorum, ala I Vespasiana, cohors Antiochensium, cohors I Bracaraugustanorum, cohors I Cantabrorum, cohors Cilicum eq, cohors III Gallorum, cohors IV Gallorum, cohors V Gallorum, cohors VII Gallorum eq, cohors VIII Gallorum, cohors II Lucensium eq, cohors Mattiacorum, cohors I Sugambrorum tironum, cohors I Sugambrorum veterana eq, cohors I Thracum Syriaca eq

Dispositions of Legions and Auxiliary Units in Moesia in AD 78

88

Summary Tables of Legions and Auxiliary Units

Legions (2) IIII Flavia, VII Claudia

MOESIA SUPERIOR Auxiliary Units (11) ala Claudia nova, ala II Pannoniorum, ala Praetoria, cohors I Antiochensium, cohors I Cilicum milliaria eq, cohors I Cisipadensium, cohors I Cretum, cohors I Flavia Hispanorum milliaria, cohors V Gallorum, cohors II Gallorum Macedonica, cohors V Hispanorum

Dispositions of Legions and Auxiliary Units in Moesia Superior in AD 93 Legions (2)

Auxiliary Units (24)

IIII Flavia, VII Claudia

ala I Claudia nova, ala II Pannoniorum, ala I praetoria, cohors I Antiochensium, cohors VII Breucorum c R, cohors II Brittonum milliaria c R p f, cohors III Brittonum, cohors I Cilicum milliaria eq, cohors I Cisipadensium, cohors I Cretum, cohors I Flavia Bessorum, cohors II Flavia Commagenorum, cohors I Flavia Hispanorum, cohors II Gallorum Macedonica, cohors V Gallorum, cohors II Hispanorum, cohors V Hispanorum, cohors I Lusitanorum, cohors I Montanorum c R, cohors IIII Raetorum, cohors I Thracum c R, cohors I Thracum Syriaca eq, cohors VI Thracum, cohors I Vindelicorum milliaria c R Dispositions of Legions and Auxiliary Units in Moesia Superior in AD 100

Legions (2) IIII Flavia, VII Claudia

Auxiliary Units (12) ala I Claudia nova miscellanea, ala I Gallorum Flaviana, cohors I Antiochensium, cohors III Brittonum, cohors III Campestris, cohors I Cretum sagittariorum, cohors II Gallorum Macedonica, cohors V Gallorum, cohors V Hispanorum, cohors I Lusitanorum, cohors I Montanorum, cohors I Pannoniorum Disposition of Legions and Auxiliary Units in Moesia Superior in AD 138

Legions (2) IIII Flavia, VII Claudia

Auxiliary Units (15) ala I Claudia nova miscellanea, ala I Gallorum Flaviana, cohors I Antiochensium, cohors I Aurelia Dardanorum, cohors II Aurelia Dardanorum, cohors I Aurelia nova Pasinatum, cohors II Aurelia nova milliaria eq, cohors II Aurelia nova Sacorum, cohors III Brittonum, cohors I Cretum, cohors II Gallorum Macedonica, II Gallorum Pannonica, cohors V Hispanorum, cohors I Montanorum, cohors I Pannoniorum Disposition of Legionaries and Auxiliaries in Moesia Superior in AD 170

89

Exercitus Moesiae MOESIA INFERIOR Auxiliary Units (23?) ala I Asturum, ala I Claudia Gallorum, ala Dardanorum, ala I Flavia Gaetulorum, ala Gallorum Flaviana, ala Gallorum Atectorigiana, ala Hispanorum, ala I Pannoniorum, ala Vespasiana, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq, cohors II Chalcidenorum, cohors II Flavia Bessorum, cohors I Flavia Commagenorum eq, cohors II Gallorum, cohors IIII Gallorum, cohors VII Gallorum eq, cohors II Lucensium eq, cohors I Lusitanorum Cyrenaica, cohors Raetorum, cohors I Sugambrorum tironum, cohors I Sugambrorum veterana, cohors Ubiorum eq

Legions (2) I Italica, V Macedonica

Dispositions of Legions and Auxiliary Units in Moesia Inferior in AD 93 Legions (2) I Italica, V Macedonica

Auxiliary Units (25) ala I Asturum, ala I Flavia Gaetulorum, ala Gallorum Atectorigiana, ala I Gallorum Flaviana, ala Hispanorum, ala II Hispanorum et Aravacorum, ala I Pannoniorum, ala I Vespasiana Dardanorum, cohors I Bracarorum, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq, cohors II Chalcidenorum, cohors I Cilicum milliaria sagittaria eq, cohors II Flavia Brittonum, cohors I Flavia Commagenorum equitata sagittaria, cohors I Flavia Numidarum eq, cohors II Gallorum, cohors VII Gallorum eq, cohors I Hispanorum veterana eq, cohors I Lepidiana eq c R, cohors I Lusitanorum Cyrenaica, cohors II Mattiacorum milliaria eq, cohors I Sugambrorum veterana eq, cohors I Tyriorum, cohors Ubiorum eq Dispositions of Legions and Auxiliary Units in Moesia Inferior in AD 100

Legions (3) I Italica, V Macedonica, XI Claudia

Auxiliary Units (17) ala I Flavia Gaetulorum, ala Gallorum Atectorigiana, ala II Hispanorum Aravacorum, ala I Pannoniorum et Gallorum catafracta, ala I Vespasiana Dardanorum, cohors I Bracarorum, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq, cohors II Chalcidenorum, cohors I Cilicum milliaria sagittaria eq, cohors I Claudia Sugambrorum veterana eq, cohors II Flavia Brittonum, cohors I Flavia Numidarum eq, cohors I Germanorum, cohors I Lusitanorum, cohors II Mattiacorum milliaria eq, cohors I Thracum Syriaca eq Disposition of Legions and Auxiliary Units in Moesia Inferior in AD 138 90

Summary Tables of Legions and Auxiliary Units Legions (2) I Italica, XI Claudia

Auxiliary Units (16) ala I Flavia Gaetulorum, ala Gallorum Atectorigiana, ala II Hispanorum Aravacorum, ala I Pannoniorum et Gallorum catafracta, ala I Vespasiana Dardanorum, cohors I Bracarorum, cohors I Bracaraugustanorum, cohors II Bracaraugustanorum eq, cohors II Chalcidenorum, cohors I Cilicum milliaria sagittaria eq, cohors I Claudia Sugambrorum veterana eq, cohors II Flavia Brittonum, cohors I Germanorum, cohors I Lusitanorum, cohors II Mattiacorum milliaria eq, cohors I Thracum Syriaca eq Disposition of Legionaries and Auxiliaries in Moesia Inferior in AD 170

91

Appendix 2: List of all the Military Units Based in Moesia In this list we provide, where possible, the original location of a unit, its date of transfer to Moesia (where applicable), its full name, and its year of transfer and final location (again, where applicable). Where there is a blank, we do not know – though it could be that the unit started its history in the Moesias upon recruitment. This is particularly the case with the auxiliaries. In most instances the dates are approximations. Where more than one date or date range is listed, this means that the unit returned to the Moesias NOTE (province of unit): M=Moesia, MI=Moesia Inferior, MS=Moesia Original Location

Year of Arrival in Moesia

Unit

East (via Illyricum?)

AD 69/70

Legio I Italica (MI)

East

AD 64-69

Dalmatia

AD 85/86, 117-119 29 BC

AD

Year of Departure from Moesia

Final Location

Legio III Gallica (M)

AD 69/70

Cisalpine Gaul

Legio IIII Flavia (MS)

AD 102-106

Dacia/East

Legio IIII Scythica (M)

AD 58?

Syria

East (via Illyricum?)

AD 69?

Legio V Alaudae (M)

AD 70

Destruction

Macedonia/Galatia?, East Dalmatia? Illyricum Upper Germany

29 BC and AD 70, AD 68-70? AD 57-59 AD 46 AD 106/107

Legio V Macedonica (MI)

AD 58?, AD 167-168

East, Dacia

AD 69/70

Cisalpine Gaul

AD 78

Germany

AD 125/126

Dacia Inferior

Tiberius and Nero?

Syria

Julio-Claudian Julio-Claudian/AD 92? Germany

ala I Claudia nova miscellanea (MS)

Claudius?

ala I Claudia Gallorum Capitoniana (MI) ala I Flavia Gaetulorum (MI) ala Gallorum Flaviana (MI)

AD 118/119130

Dacia Inferior

AD 125-132

Moesia Superior

ala Hispanorum (MI) ala Hispanorum (Campagonum c R) (MS) ala Moesica (M) ala I Pannoniorum (MI) ala I Pannoniorum et Gallorum catafracta (MI) ala Pansiana (M)

AD 119 AD 114

Dacia Superior Lower Pannonia

AD 78

Germany

AD 92 AD 112 Vespasian AD 120 Claudius?

Pannonia

ala Bosporanorum (M)

AD 74/75?

Vespasian AD 75/78-92 Pannonia Britain

Legio VII Claudia (MS) Legio VIII Augusta (M) Legio XI Claudia (MI) ala Afrorum veterana (M) ala Asturum (MI) ala Gallorum Atectorigiana (MI)

AD 94

ala Praetoria (MS) ala Scubulorum (M) Ala Sulpicia (M) 92

Before 78

AD

Destroyed?

Claudius ?

Britain

AD 78

Germany

List of all the Military Units Based in Moesia AD 75/78

ala I Vespasiana Dardanorum (MI)

AD 97/99

Flavian?

Judea?

Dalmatia

Claudius/Nero

Mauretania Tingitana Pannonia

AD 105-122 AD 101-103/105

Pannonia

AD 101-103/105

Spain?

Vespasian

ala II Hispanorum et Aravacorum (MI) ala II Pannoniorum (MS) cohors I Antiochensium sagittaria (MS) cohors I Augusta Nerviana Pacensis milliaria Brittonum (MI) cohors I Aurelia Dardanorum (MS) cohors I Bracaraugustanorum (MI) cohors I Bracarum c R (MI) cohors I Britannica milliaria c R (MS) cohors I Brittonum milliaria (MS) cohors I Cantabrorum (M)

Nero

cohors I Cilicum sagittaria (MS)

AD 115-134

Moesia Inferior

AD 75 AD 94, AD 117138 AD 96/97

cohors I Cisipadensium (MS) cohors I Cretum sagittaria (MS)

AD 110

Dacia

AD 109

Dacia

AD 110

Dacia

AD 127

East

AD 85/86? AD 109

Destruction? Dacia

Trajan

Raetia? Syria

AD 107, AD 125 AD 130 AD 144

Dacia, Moesia Inferior Dacia Inferior Dacia Superior

Syria

AD 94 AD 75 AD 105 AD 161-180

Syria

Vespasian

Germany

AD 100

cohors I Flavia Bessorum (MIMS?) cohors I Flavia Commagenorum equitata sagittaria (MI) cohors I Flavia Hispanorum milliaria (MS) cohors I Flavia Numidarum equitata (MI) cohors I Germanorum (MI) cohors I Hispanorum veterana equitata (MI) cohors I Lepidiana c R equitata (MI) Cohors Lusitanorum (MS) cohors Lusitanorum Cyrenaica (MI) cohors Mattiacorum (M) cohors I Montanorum c R (MS) cohors I Pannoniorum (MS) cohors I Raetorum (MI) cohors I Sugambrorum tironum (MI) cohors I Claudia Sugambrorum veterana equitata (MI) cohors I Thracum Syriaca equitata (MI) cohors I Thracum c R (MS) (MI)

Syria

Vespasian AD 70-86 AD 98-100

cohors I Tyriorum sagittaria (MI) cohors I Ubiorum equitata (MI) cohors I Vindelicorum milliaria c

Flavian Germany

AD 78-80

Syria

AD 97

Illyricum

AD 121 AD 97

Pannonia

AD 97

Illyricum Illyricum

AD 60 AD 92

Pannonia

AD 85-96 AD 103/105 AD 75

Cappadocia

Vespasian

Lower Germany

93

Exercitus Moesiae

AD 105 AD 161-180 AD 161-180 AD 161-180 AD 92, AD 136 Lower Germany

AD 98-100

East

AD 75 AD 92 AD 99 AD 96 AD 107

Upper Dacia Pannonia

AD 92 AD 94, AD 124132 AD 144-157 AD 85-99 Vespasian Vespasian

AD 109 AD 100 Pontus and Bithynia Germany

Germany?

AD 62? AD 94 AD 62? AD 78-80 AD 84-96 AD 96 AD 62?

Germany?

AD 62?

Pannonia

AD 85-103

cohors VIII Raetorum 1 c R (MS)

Germany? Germany Pannonia

AD 114

Pannonia

AD 114

Thrace

AD 129

Dacia Inferior

AD 109

Dacia

AD 116-122

Dacia Inferior

AD 114-130 AD 109

Dacia Inferior Dacia

AD 127-138

Thrace

AD 70-98

Germania Inferior

cohors III Augusta Nerviana Brittonum (MS) cohors III Brittonum (MS) cohors III Campestris (MS) cohors III Gallorum (M) cohors III Hispanorum (M) cohors IIII Cypria c R MS) cohors IV Gallorum (M) cohors IIII Raetorum (MS) cohors V Gallorum (MS) cohors V Hispanorum (MS) cohors VI Thracum (MS) cohors VII Breucorum c R (MS) cohors VII Gallorum equitata (MI) cohors VIII Gallorum (M)

Pontus and Bithynia Germany?

AD 103/105 AD 74 AD 75

R (MS) cohors II Augusta Nerviana Pacensis milliaria (MI) cohors II Aurelia Dardanorum (MS) cohors II Aurelia nova milliaria equitata (MS) cohors II Aurelia nova Sacorum (MS) cohors II Bracaraugustanorum equitata (MI) cohors II Brittonum milliaria c R p f (MS) cohors II Chalcidenorum sagittaria (MI) cohors II Flavia Bessorum (MI) cohors II Flavia Brittonum (MI) cohors II Flavia Commagenorum (MS) cohors II Flavia Numidarum equitata (MI) cohors II Gallorum (MI) cohors II Gallorum Macedonica (MS) cohors II Gallorum Pannonica (MS) cohors II Hispanorum (MS) cohors II Lucensium equitata (MI) cohors II Mattiacorum milliaria equitata (MI) cohors II Varcianorum (?) (M)

1

Destruction?

AD 109

Dacia

AD 109

Destruction? Dacia

AD 115 AD 179

Cappadocia Dacia Superior

AD 110 Trajan AD 114156/157 Flavian? Domitian? AD 103-109

Dacia Dacia Syria Destruction? Dacia

One unit not found in this list, and in the text above, is the cohors IX Batavorum milliaria equitata, which Matei-Popescu (2007a) has argued was never in Moesia Inferior to begin with.

94

Appendix 3: List of Moesian Diplomas This list is based on Eck & Pangerl (2009), and it has been updated for this book, in the process incorporating all diplomas published up to December 31, 2015 (where possible). The numbers in brackets at the end of each entry indicate the number of auxiliary alae (first number) and cohortes (second number). For the abbreviations, please see the list at the front of the book, and the bibliography at the end. Given that many of the diplomas are fragmentary, it is not always possible to identify the correct constitution, and so some of this is conjecture. The fragmentary state of some diplomas is why in some cases below a question mark follows the title for the diploma and not specific numbers of alae or cohortes. MOESIA 1. 73 – Eck and Pangerl 2006: 50 (fleet) 2. 28.4.75 – RMD I.2 (0, 10) 3. 28.4.75 – RGZM 1; Weiss 2008: 270-273; Eck and Pangerl 2009: 506-509 (0, 7) 4. 7.2.78 – CIL XVI.22; RMD IV.208; Weiss 2008: 273-275; Eck and Pangerl 2008: 319-321; RMD IV.209; Eck and Pangerl 2010: 237-243 (0, 8) 5. 7.2.78 – RMD V.325; Eck and Pangerl 2008: 321-323 (0, 6) 6. 78 – RMD IV.209; Eck and Pangerl 2008: 324-325 ((4), 0) 7. 20.9.82 – CIL XVI.28 (1, 2) 8. 82/83 – Weiss 2008: 275-279 (6, (2)) MOESIA SUPERIOR 1. 16.9.94 – CIL XVI.39; RMD V.335; Weiss 2008: 279-280 (3, 9) 2. 12.7.96 – RMD I.6 (1, 10) 3. Jan 97 – CIL XVI.41? (-, -) 4. Jul/Aug 97 (14.8.97?) – Eck and Pangerl 2005: 231; AE 2005, 1709; Eck and Pangerl 2013: 275-279 ((2), 19) 5. 14.8.99 – RMD I.7 (-, (1)) 6. 96/100 – RMD IV.218 (-, (3)) 7. 8.5.100 – CIL XVI.46; Eck and Pangerl 2008: 326-329; Eck and Pangerl 2014b: 218-220 (3, 21) 8. 100 – Eck and Pangerl 38: 340-342; Eck and Pangerl 2009: 562-566 (3, 19) 9. 16.5.101 – RMD III.143; Eck and Pangerl 2008: 329-337; Eck and Pangerl 2014b: 220-222 (1, 12) 10. 102 – Eck and Pangerl 2008: 348-353 (1) 11. 12.1.105 – RMD V.339 (-, (1)) 12. c. 103/105 – CIL XVI.54; RGZM 13; Eck and Pangerl 2009: 567-569; Eck and Pangerl 2015: 224-226 (2, 11) 13. 27.7.108 – Eck and Pangerl 2004: 103-115 (-, (1)) 14. 109 – Eck and Pangerl 2015: 226-229 (2, 11) 15. Trajan or 112 – Eck and Pangerl 2008: 354 ((1), (1)) 16. 112 – Eck and Pangerl 2008: 355-360, 360-363 (2, 13) 17. 05.07.115 – AE 2005, 1710; AE 2005, 1723=AE 2008, 1740; Eck and Pangerl 2014b: 225-227; Eck and Pangerl 2015: 229-230 18. 15.07.115 – Eck and Pangerl 2015: 229-230 (-, (4)) (1, 9 – units on an expedition) 19. Aug 115 – Eck and Pangerl Chiron 2005: 49-66; Eck and Pangerl 2008: 363-370 (1, 4+9) 20. Trajanic? – Eck and Pangerl 2009: 570-571 (-, (4)) 21. 1.7.126 – RMD V.366 (?); AE 2008, 1717; Eck and Pangerl 2014b: 227; Eck and Pangerl 2015: 231-236 (1, 3) 22. Dec 125/126 – Eck and Pangerl 2006: 102-104; Weiss 2008: 280-286 (1, 3); AE 2006, 1864 (?) 23. 9.9.132/133 – RMD IV.247 (2, 10) 24. 137 – Eck and Pangerl 2015: 236-239 (2, 9) 25. near 140? – Eck and Pangerl 2009: 572-574 ((1), -) 26. 144/146 – RMD V.402 (-, -) 27. 11.20.146 – Eck, Holder, Pangerl, and Weiss 2015: 230 (5, 11) 28. 20.1.151 – RGZM 31; Eck and Pangerl 2008: 372-376 (2, 9) 29. c. 151 – RMD V.405 (-, -) 30. 5.3.153 – Eck and Pangerl 2008: 377-380 (-, -) 31. 145/154 – Eck and Pangerl 2008: 371-372 (2, 10) 32. 23.4.157 – RGZM.37; RMD V.418, 419; Weiss 2008: 286-290; Eck and Pangerl 2008: 380-381, 381-382, 382, 383-384; Eck, MacDonald, and Pangerl 2008: 237-239 (2, 10); AE 2008, 1718 33. Jan/Feb 160 – RGZM 40; CIL XVI.111; Weiss 2008: 290-291 (2, 10) 34. 8.2.161 – RMD I.55 (2, 10) 35. 138/161 – CIL XVI.114 (-, -) 36. 18.2.165 – CIL XVI.120 (-, -) 95

Exercitus Moesiae 37. ? – Eck and Pangerl 2009: 574-575 (-, (2)) MOESIA INFERIOR 1. 14.6.92 – CIL XVI.37 (fleet) 2. 14.6.92 – Petolescu and Popescu 2004: 269; AE 2005, 1706 (7, 15) 3. 9.9.97 – RMD V.337; AE 2005, 1705 (4, 10) 4. 9.9.97 – RMD V.338; AE 2005, 1332; AE 2005, 1704; Eck and Pangerl 2009: 510-512 (5, 9) 5. 96 or 97 – Weiss 2005: 213-217; ((3), (3)) 6. 14.8.99 – CIL XVI.44 (3, 6) 7. 14.8.99 – CIL XVI.45; RGZM 8; Eck and Pangerl 2006: 97-99; Mihailescu-Bîrliba 2008: 199-210; AE 2006, 1862; Eck and Pangerl 2012: 295-301 (3, 7) 8. 99 – Eck and Pangerl 2014b: 215-218 (3, 6) 9. Sept/Oct 99 – RMD IV.217 (-, (3)) 10. 99/100 – Eck and Pangerl 2009: 512-514 ((1), (2)) 11. 13.5.105 – CIL XVI.50 (3, 7) 12. 13.5.105 – RGZM 10 (3, 7) 13. 13.5.105 – RGZM 11; AE 2004, 1256; Weiss 2008: 292-293 (3, 7) 14. 24.11.107 – AE 2009, 1803 and 1804 (?) (3, 7) 15. 24.11.107 – RGZM 14 (3, 7) 16. 24.11.107 – Eck and Pangerl 2014b: 222-225 (?) 17. Sept/Dec 107 – Eck and Pangerl 2009: 514-519, 519-522 (3, 7) 18. May/Aug 109 – RMD IV.219 (-, (1)) 19. 99/110 – RMD IV.221 (1, (3)) 20. 25.9.111 – RMD IV.222 (3, 7) 21. [25.9.] 111? – CIL XVI.58 (3, 7) 22. Jan/Mar 112 – RMD V.344; Eck, Pangerl, and Weber 2005: 247-254 (fleet) 23. 113? – RMD IV.224 ((1), (1) + (1)) 24. c. 113 – Weiss 2008: 293-296 (?, 9) 25. 113? – Eck and Pangerl 2009: 522-524 (-, (1)) 26. c. 112-116 – AE 2008, 1721 (?) 27. 102/114 – RGZM 65? ((2), (2)) 28. Jan/Mar 116 – Eck and Pangerl 2006: 99-102; Eck and Pangerl 2009: 530 (2, 5) 29. Sept 118 - /Oct 119 – RMD V.349, 350 (6, -) 30. 10.12.118-9.12.119 – Eck and Pangerl 2009: 530-533 (5, 8) 31. 19.10.120 – RMD V.356; Eck and Pangerl 2009: 533-537 (5, 8) 32. May/Dec 121 – Weiss 2008: 296-300 33. 1.6.125 – RMD IV.235, V 364, V.375; Weiss 2009: 241-242; Eck and Pangerl 2009: 537-538; Eck and Pangerl 2014a: 245-249 (2, 5) 34. 105/127 – RMD V.369 (-, 3) 35. 121/125 or 128 – Eck and Pangerl 2009: 538-541 (2, 5) 36. 20.8.127 – RMD IV.241; RGZM 23 (5, 10) 37. 2.4.134 – CIL XVI.78; Weiss 2008: 300-302 (2, 5) 38. 135 – Eck and Pangerl 2009: 541-542 (2, 6) 39. 136 – Schindel 2010: 259-263 (-, 5) 40. 28.2.138 – CIL XVI.83; RMD IV.253 (3, 5) 41. Hadrianic? – Eck and Pangerl 2009: 543-547 (-, (2)) 42. Jan/Nov 140 – Weiss 2008: 302-307 (5, 11) 43. 138/142? – RMD IV.265 (1, 3) 44. 7.4.145 – RMD III.165/V.399; Eck and Pangerl 2009: 548-550, 550-552 (5, 11) 45. Jan/Feb 146 – RMD IV.270; Petolescu 2007: 149-151; Weiss 2008: 307-309; Eck and Pangerl 2009: 553556 (4, 11) 46. 11.10.146 – Eck, Holder, Pangerl, Weiss 2015 (5, 11) 47. c. 151 – RMD V.405 (-, -) 48. 148/153 – AE 1997, 1778; RMD V.412 (-, (1)) 49. 27.9.154 – AE 2001, 2160; RMD V.414; Eck and Pangerl 2009: 557-562 (5, 11) 50. c. 157 – Weiss 2008: 309-312 (5, 11) 51. 8.2.157 – Ivantchik-Krapivina 2007: 219-239 (5, 11) 52. 158? – RMD I.50 (5, 11?) 53. Sept 23 160 – Weiss 2009: 247-250 (-, (1)) 54. Mar/Dec 161 – RMD II.111 ((1), -) TOTALS 96

List of Moesian Diplomas Moesia – 8 constitutions, 17 diplomas Moesia Superior – 37 constitutions, 67 diplomas Moesia Inferior – 54 constitutions, 84 diplomas

97

Appendix 4: List of Ancient Sources Used I have included in this list all the sources referred to in the book and in Whately (2015), which discusses the military organisation in the Moesias in late antiquity, with two exceptions: those diplomas published that at the time of writing were published only in journal articles (references to which can be found in appendix three above), and any visual evidence (particularly references to scenes from Trajan’s Column). EPIGRAPHIC EVIDENCE 1950, 75a 1950, 76 1952, 191 1956, 124 1957, 307 1960, 127 1960, 333 1960, 360 1963, 3 1963, 262 1964, 180 1964, 262 1965, 165 1965, 347 1966, 325 1967, 425 1967, 426 1967, 428 1967, 431 1967, 432 1968, 453 1969/1970, 649 1971, 424 1972, 490 1972, 504 1972, 669 1973, 500 1974, 562b 1975, 582a-b 1976, 582c 1976, 582d 1976, 610 1976, 634 1977, 740 1978, 716 1980, 814 1980, 815 1980, 822 1981, 737 1981, 745 1984, 805 1984, 806 1984, 886 1985, 762 1987, 867 1988, 988 1988, 998 1988, 9992 1989, 640 1990, 868 1990, 893 1990, 896

ADBulgar 458 AE (check 2003 and up) 1889, 95 1891, 148 1895, 60 1900, 27 1900, 58 1900, 200 1901, 19 1901, 21 1902, 133a 1903, 1 1903, 2 1903, 3 1903, 289 1903, 294 1903, 295 1904, 92 1905, 160 1908, 22 1909, 16 1909, 166 1910, 85 1912, 78 1912, 187 1919, 18 1919, 192 1925, 44 1925, 66 1925, 70 1925, 77 1925, 78 1934, 106 1934, 112 1934, 209 1935, 74 1935, 171 1938, 141 1939, 87 1939, 90 1939, 101 1941, 105 1944, 6-9 1944, 65 1944, 66 1945, 80 1946, 51 1946, 258 1946, 260 1950, 46 98

List of Ancient Sources Used 1991, 1260 1991, 1378 1991, 1402 1992, 1496 1995, 1348 1995, 1350 1996, 1358 1997, 1329 1997, 1332 1997, 1353 1997, 1778 1998, 1146 1998, 1147 1998, 1154 1998, 1155 1998, 1156 1998, 1161 1998, 1163a 1998, 1163b 1998, 1163c 1998, 1163d 1999, 1327 1999, 1332 1999, 1370 2000, 1274 2000, 1276 2001, 120 2002, 1252 2002, 1289 2003, 1527 2003, 1540 2003, 1550 2004, 1252 2004, 1256 2004, 1278 2005, 1297 2005, 1332 2005, 1333 2005, 1335 2005, 1339 2005, 1704 2005, 1705 2005, 1706 2005, 1709 2005, 1710 2005, 1723 2005, 1731 2006, 1180a 2006, 1553 2006, 1862 2006, 1864 2007, 1238 2007, 1759 2008, 1177 2008, 1195 2008, 1717 2008, 1718 2008, 1721 2008, 1728 2008, 1740 2009, 1803 2009, 1804

2010, 1418 2010, 1853 2010, 1857 2010, 1865 2011, 1120 2011, 1133a1-35 2011, 1133b1-6 2011, 1142 2012, 1959 AMN 2006/2007, 186 2006/2007, 190 2006/2007, 192 CBI 661 CERom II.105 VI.388b XI.566 XVII.733a XVII.733b XVII.745a-b XIX/XX.908 XIX/XX.909 XXV.1139 CIL II 3272 CIL III 14 192 195 905 1257 1643 1644 1674 1698 1702 1813 1818 2033 2048 2071 2709 2710 2716 2717 2882 2908 2913 2914 3200 6154 6166 6168 6169 6189 99

Exercitus Moesiae 6302 6324a 6324d 6580 6581 6707 6826 6827 7368 7449 7473 7474 7475 7494 7504 7505 7512 7548 7557 7603a 7619a 7620 8066b-c 8071c-d 8074-5b 8074-20 8099 8103 8119 8162 8242 8250 8251 8261 8275 8487 8488 8493 8544 8687 8723 8763 8767 9711 9712 9733 9734 9736 9737 9741 9742 9832 9864a 9939 9973 12348 12359 12361 12376 12378 12416 12437

12480 12483 12666 12676 12814 13578 13750 13751 14214 14214-22 14214-29 14215-4 14215-5a 14215-5b 14215-5c 14217-6 14375 14415 14417 14428 14429 14433 14500 14507 14509 14537 14545 14587 14700 CIL V 1838 4058 4326 7425 CIL VI 3492 32933 37247 CIL VII 818 819 820 823 CIL VIII 619 4116 18068 CIL IX 3610 CIL X 1258 1711 4723 6657 8241 100

List of Ancient Sources Used 160 161 163 164 165 183 185

CIL XI 1835 5632 6163 CIL XIII 6223 6298 7382 8304 8305 8311 8312 8320

CIMRM II.2273 D 2543 8504 9057 9160

CIL XIV 3608

GeA 471

CIL XVI 4 20 22 23 26 28 30 31 33 35 36 39 43 44 45 46 47 50 51 54 56 57 58 61 69 70 78 83 90 94 103 106 107 108 110 111 117 120 123 128 131 158 159

Iatrus 13 16 17 52-61 93-98 IBulgarien 52 IDR II.179 II.237 III.1.5 III.1.6 III.1.7a III.1.7b III.1.7c IDRE II.349 II.350 IGRR III.1476 III.2.1741 III.2.1741a IIFDR 110 220 221 238 271 280 282 283a-b 285 286 287 288 101

Exercitus Moesiae II.172 II.225 IV.107 IV.143g IV.154 IV.169 IV.200 IV.201a-b IV.203 V.16 V.23 V.24 V.36 V.53 V.54a-b V.94 V.95a V.113 V.117 V.118 V.137 V.218 V.253 V.254a-b V.290b V.308

ILBulg 50 61 120 137 300 301 305 306 307 329 ILD 120 122 ILJug II.461 II.463 II.466 III.1287 III.1343 III.1434 ILS 724 921 986 2288 2543 8504 8695 9057 9107 9160 9197 9200 9491

Kayser 106 KJ 2010, 181 Krier 59 LNCh 33

IMS I.119 II.5 II.12 II.52 II.124 III.2.22 IV.44

RGZM 1 8 10 11 13 20 22 31 37

InscrIT 10-5, 737

RMD I.2 I.6 I.7 I.14 I.21 I.35 I.39 I.47 I.50 I.52 I.53

IOSPE2 I.322 I.355 I.362 I.369 I.420 III.2 IScM II.93 II.106 102

List of Ancient Sources Used I.55 I.63 I.64 I.67 II.87 II.90 II.116 II.123 III.140 III.148 III.152 III.153 III.165 III.168 III.181 III.184 III.185 IV.202 IV.208 IV.209 IV.216 IV.217 IV.222 IV.225 IV.226 IV.227 IV.228 IV.235 IV.239 IV.241

IV.247 IV.251 IV.260 IV.270 IV.278 IV.287 V.325 V.335 V.337 V.338 V.339 V.364 V.375 V.376 V.380 V.384 V.385 V.418 V.419

SCI 2012, 54 SCIVA 2009, 318 2009, 319 DOCUMENTARY EVIDENCE or. 28.30 or. 28.31 or. 28.32 or. 28.33 or. 28.34 or. 31 or. 31.33 or. 31.37 or. 31.38 or. 32 or. 32.25 or. 32.30 or. 32.34 or. 32.38 or. 33 or. 33.17 or. 33.19 or. 33.27 or. 33.30 or. 33.36 or. 34 or. 34.16 or. 34.19 or. 34.22 or. 34.23 or. 34.24 or. 35 or. 35.16 or. 35.18

Theodosian Code 7.20.4 12.1.56 Digest 1.18.13pr 47.8.2 48.19.8-10 74.14.1-3 Notitia Dignitatum occ. 40 or. 7.53 or. 8.27 or. 8.32 or. 8.36 or. 8.45 or. 8.46 or. 25.20 or. 26.12 or. 28 or. 28.20 or. 28.21 or. 28.24 or. 28.25 or. 28.26 or. 28.27 or. 28.28 or. 28.29 103

Exercitus Moesiae or. 35.19 or. 36.20 or. 36.23 or. 36.24 or. 37.15 or. 37.18 or. 37.19 or. 37.28 or. 37.29 or. 37.30 or. 38.14 or. 38.15 or. 38.18 or. 38.19 or. 38.32 or. 38.33 or. 38.37 or. 39.2 or. 39.12 or. 39.14 or. 39.15 or. 39.16 or. 39.17 or. 39.21 or. 39.22 or. 40.11 or. 40.12 or. 40.14 or. 40.16 or. 40.17 or. 40.19 or. 40.21

or. 40.30 or. 40.31 or. 40.32 or. 40.33 or. 40.46 or. 40.48 or. 40.49 or. 41.13 or. 41.14 or. 41.15 or. 41.16 or. 41.17 or. 41.19 or. 41.20 or. 41.23 or. 41.25 or. 41.30 or. 41.31 or. 42 or. 42.13 or. 42.23 or. 42.26 or. 42.27 P. Oxy 43 RMR 63 Vindolanda Tablets 1.154 LITERARY EVIDENCE Ekt. 3 Ekt. 4 Ekt. 9 Peripl. M. Eux 6.1 Peripl. M. Eux. 6.2 Peripl. M. Eux. 9.3 Peripl. M. Eux. 10.3

Aelius Aristides Or. 26.29-30 Ammianus Marcellinus 13.18-19 14.11.15 18.9.3 24.1.2 24.1.9 26.5.1-3 30.2.6

Augustus RG 30 Aurelius Victor Caes. 13

Antonine Itinerary 219.3 220.5 221.4 223.4 225.2 226.1

Caesar BC 1.73 BG 3.9 BG 4.7 Cassius Dio 50.23.2 51.23.2-4 51.25.1-27.1 54.20.3 54.34.5 54.36.2

Appian Ill. 30 Arrian Ekt. 1 Ekt. 2 104

List of Ancient Sources Used 55.23.3 55.23.4 55.24.1 55.24.3 55.29.3 55.30.1 55.30.4 55.32.3 60.15.4 67.6.1 67.6.5 67.6.6 68.8.1 68.8.2 68.8.3 68.9.2 68.9.3 68.9.5 68.9.7 68.10.3 68.10.4 68.11.3 68.12-1-5 68.13.1-6 68.14.1 72.11.1 72.11.2 72.16.1 72.16.2 72.33.1 73.3.3 73.8.1 75.3.1 79.27.5

7.64 7.96 7.184 8.113 9.32 9.71 9.113 Historia Augusta Hadr. 5.2 Hadr. 6.6 Hadr. 6.7-8 Ant. Pius 5.4-5 Ant. Pius 9.9 Marc. 17.2 Marc. 21.7 Avid. Cass. 4.6 Comm. 6.1 Did. Iul. 5.8 Pesc. Nig. 2.6 Pesc. Nig. 7.7 M. Ant. 10.6 Sev. Alex. 47.1 Sev. Alex. 58.4 Gord. 28.2 Tyr. Trig. 18.6-9 Probus 14.7 Aurelianus 38.4 John Lydus Mag. 2.10 Jordanes Get. 13.76 Get. 13.77 Get. 16.90 Get. 16.91 Get. 16.92 Get. 16.93

Cyprian Ep. 68.3.3 Dexippus 18

Josephus BJ 2.16.4 BJ 2.18.9 BJ 4.11.1 BJ 5.1.6 BJ 7.4.3 BJ 7.5.3

Eutropius 7.23 9.15 Florus 2.26 2.28 2.29

Juvenal 4.110-112

Fragments of the Greek Historians 257

Latin Panegyrics 4.18.4

Herodian 1.6.8 3.7.1 6.7.2

Malalas 12.40 Martial 6.76

Herodotus 1.153 3.93 6.113

Onasander 6.14 105

Exercitus Moesiae Hist. 2.83 Hist. 2.85 Hist. 2.86 Hist. 3.18 Hist. 3.2 Hist. 3.35 Hist. 3.46 Hist. 3.47 Hist. 3.5 Hist. 3.9 Hist. 4.18 Hist. 4.58 Hist. 4.77

Orosius 7.10.4 Ovid Pont. 1.8.11-16 Pont. 4.9.79-80 Pliny HN 3.21.140 HN 3.26.149 HN 4.84 HN 6.4 HN 6.159

Themistius Or. 10.136d-8b

Ptolemy Geog. 3.9.1 Geog. 3.9.2 Geog. 3.9.3 Geog. 3.9.4 Geog. 3.10.4 Geog. 6.13

Vegetius Mil. 1.57 Mil. 2.10 Mil. 3.3 Mil. 4.7 Mil. 4.31.6

Strabo 3.4.20 7.3.10 7.3.11 7.5.7 11.8.2

Velleius Paterculus 2.98 2.110 2.110.6 2.112.4

Suetonius Ner. 18 Ner. 19 Vesp. 6 Vesp. 8.4 Dom. 6

Zosimus 1.20.1 2.34 4.3.4 4.3.5

Tacitus Agr. 19.4 Agr. 22.2 Agr. 41 Ann. 1.16.2 Ann. 1.23.5 Ann. 1.30.4 Ann. 3.39.1 Ann. 4.46 Ann. 4.47 Ann. 4.5 Ann. 6.41 Ann. 12.15 Ann. 12.16 Ann. 12.20 Ann. 13.35 Ann. 15.6 Ann. 15.10 Hist. 1.76 Hist. 1.79 Hist. 2.11 Hist. 2.18 Hist. 2.32 Hist. 2.46 Hist. 2.74 106

Appendix 5: The Numeri 1 The numeri are perhaps the most elusive type of unit employed by the Roman military during the imperial period. In fact, beyond the existence of a comparably small number of units, we know very little about their size and composition. 2 Regardless, there are a few points that we can discuss. The word numerus itself is a vague term. It could be used to replace the name of the type of unit, such as cohors. This is how the term seems to be used in at least one inscription. 3 Numerus could be used to represent a “number” of legionaries, auxiliaries, or some combination of both. The term could be used as a generic term that is probably best translated as unit. So, in Tacitus we find ‘sparsi per provinciam numeri’, which might be translated as ‘units spread through the province’. 4 According to Southern, the term could also be used to refer to those types of soldiers who were neither legionaries, auxiliaries, or ethnic units. 5 Numerus can refer to a particular type of unit, like an ala, a cohors, or a legio. In this last sense, these units are also called national numeri. 6 The most distinctive part of the nomenclature of these units were their ethnic names, which more often than not would stand alone without the term numerus. 7 These are the units that scholars had traditionally regarded as the replacements for the increasingly civilised, and as scholars believed, less effective auxiliaries. 8 This new class of unit was believed to have first appeared early in the second century, or, more specifically, during the reign of Hadrian. 9 These views are, however, unfounded. 10 Southern suggests that these units may have been recruited from those units with which Rome had a peace treaty; unfortunately, evidence is lacking and this must remain speculation. 11 Their size is perhaps one of the least known aspects. 12 Southern is probably right to hypothesise that there was great variation in their size. 13 Thus, it may have ranged from 100 to perhaps as many as 1000. Numerus, itself a general term, seems to point to the vagueness of these units’ size. Now, as regards the numeri and the Moesias, the nearest known units come from Dacia. 14 The only possibilities for the Moesias, and in this case Moesia Inferior, are the peculiar N C R. 15 The traditional expansion for these abbreviations is N(UMERUS) C(IVIUM) R(OMANORUM). Speidel, however, has challenged this reconstruction. 16 He notes that “no auxiliary regiments of Roman citizens were raised after Vespasian’s reign…”. 17 Instead, he suggests that what we should be reading is N(UMERUS) C(OLLECTUS) R(EGIONARIORUM). Regionarii were soldiers collected to police a particular area. 18 Thus, these Moesian numeri were a collection of troops gathered for some unspecified duty. They were not the national numeri discussed above.

1

There have been a handful of studies that have examined the numeri of the Roman army, including Rowell 1937: 1327-1341; Rowell 1939: 73-108; Mann 1954: 501-506; Speidel 1975: 202-231; Le Roux 1986: 347-374; and Southern 1989: 81-140. Southern’s study is the most useful of those noted, and much of this survey is based on that article. 2 This is in comparison to the multitudinous auxilia, and the very well documented legiones. 3 CIL III.12257. 4 Tac. Agr. 18.2. 5 Southern 1989: 84. 6 This is the distinction made by Southern (1989: 84). 7 Southern 1989: 85. 8 See for example, the pronouncements of Cheesman (1914: 88-89). 9 The early second century dating applies to Cheesman (1914: 85ff); the Hadrianic dating applies to Rowell (1937: 1329ff). The fact of the matter is, we do not know when the national numeri first appeared. Southern speculates that the provincial militia, and the nationes and symmiacharii of Hyginus were the forerunners of these units. 10 The auxiliaries, although they were increasingly recruited locally and thus presumably more Romanised, continued to be a crucial element of the Roman armed forces throughout the second century and into the third. The fact that they continued to make up such a huge percentage of the forces posted in the many provinces bears this out. 11 Southern 1989: 88. 12 See Southern (1989: 103-106) for a full discussion of the problems associated with their size. 13 Southern 1989: 104. 14 See Southern 1989: 124-126. 15 I have found four such attestations: AE 1957, 338, which dates to AD 235 and is from Montana; AE 1975, 743, and 750, which date to the third century and are from Montana; and AE 1987, 884, which dates to the end of the second century or beginning of the third century and is also from Montana. 16 Speidel 1992b: 140-144. 17 Speidel 1992b: 143. 18 Speidel 1992b: 140.

107

Appendix 6: Maps

108

Maps

109

Exercitus Moesiae

110

Maps

111

Exercitus Moesiae

112

Bibliography Absil, M. 2000. “Legio I Italica,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 227238. Paris: Diffusion de Boccard. Adams, J. P. 1976. Logistics of the Roman Imperial Army: Major Campaigns on the Eastern Front in the First Three Centuries A.D. Unpublished PhD dissertation. New Haven: Yale University. Adams, C. 1995. “Supplying the Roman Army: O. Petr. 245,” ZPE 109: 119-126. Adams, C. 2007. Land Transport in Roman Egypt. Oxford: Oxford University Press. Alexandrov, O. 2012. “Ethnic and Social Composition of the Roman Army in Moesia Inferior: soldiers from Asia Minor and the eastern provinces of the Roman Empire,” in Armées grecques et romaines dans le nord des Balkans, ed. A. Rufin Solas, 199-207. Gdansk: Akanthina. Alföldy, G. 1987. “Die Auxiliartruppen der Provinz Dalmatien,” in Römische Heeregeschichte: Beiträge 1962-1985, 239-297. Amsterdam: J. C. Gieben. Anders, A. 2015. “The Face of Roman Skirmishing,” Historia 64: 263-300. Arguin, A. R. M. 2015. “Strategy: Principate,” in The Encyclopedia of the Roman Army, eds. Brizzi, G., DeschlerErb, E., Greatrex, G., Le Bohec, Y., Rankov, B., and M. Reddé, 939-940. Malden, MA: Wiley. Bagnall, R. 1977. “Army and Police in Roman Upper Egypt,” JARCE 14: 67-86. Barlow, J. and P. Brennan. 2001. “Tribuni Scholarum Palatinarum c. A.D. 353–64: Ammianus Marcellinus and the Notitia Dignitatum,” CQ 51: 237-254. Batty, R. 2007. Rome and the Nomads. The Pontic-Danubian Realm in Antiquity. Oxford: Oxford University Press. Beckmann, M. 2011. The Column of Marcus Aurelius. Chapel Hill: University of North Carolina Press. Benea, D. 1977. “Einige Aspekte der Römerherrschaft im Gebiet des Eisernen Tores nach Aurelians Rückzug. Die Militärziegelei von Dierna,” in Limes. Akten des XI. Internationalen Limeskongresses ed. J. Fitz, 321-328. Budapest: Akadémiai Kiadó. Beneš, J. 1978. Auxilia Romana in Moesia atque in Dacia. Prague: Academia. Bennett, J. 2010. “Auxiliary Deployment during Trajan’s Parthian War: Some Neglected Evidence from Asia Minor,” in Studies in Latin Literature and Roman History XV, ed. C. Deroux, 423-445. Brussels: Éditions Latomus. Bérard, F. 2000. “La légion XXIe Rapax,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 49-67. Paris: Diffusion de Boccard. Berciu, D. 1981. “The Dacians in the First Century AD: the Roman Conquest,” in The Roman Empire and its Neighbours, ed. F. Millar, 270-280. London: Weidenfeld and Nicholson. Bertolazzi, R. 2015. “A New Military Inscription from Numidia, Moesiaci Milites at Lambaesis, and Some Observations on the Phrase Desideratus in Acie,” in The Many Faces of War in the Ancient World, eds. W. Heckel, S. Müller, and G. Wrightson, 302-314. Newcastle: Cambridge Scholars Publishing. Biernacka-Lubanska, M. 1982. The Roman and Early-Byzantine Fortifications of Lower Moesia and Northern Thrace. Wroclaw: Zakład Narodowy im. Ossolińskich. Bingham, S. 2013. The Praetorian Guard: A History of Rome’s Elite Special Forces. Waco, Tx: Baylor University Press. Birley, E. 1988. “The Epigraphy of the Roman Army,” in The Roman Army Papers 1929-1986, 3-11. Leiden: Brill. Birley, A. R. 1993. Marcus Aurelius, A Biography. London: Routledge. Birley, A. R. 1986. “Aurelius, Marcus,” OCD, 3rd edition, 219-221. Oxford: Oxford University Press. Bishop, M. C. 2012. Handbook to Roman Legionary Fortresses. Barnsley: Pen & Sword Books. Bishop, M. C., and J. Coulston. 2006. Roman Military Equipment: from the Punic Wars to the Fall of Rome. Oxford: Oxbow. Bojović, D. 1996. “Le Camp de la Légion IV Flavia à Singidunum,” in Roman Limes on the Middle and Lower Danube, ed. P. Petrović, 53-68. Belgrade: Archaeological Institute. Borza, E. N. 1996. “Chersonesus (1),” OCD, 3rd 320. Oxford: Oxford University Press. Bounegru, O. and M. Zahariade. 1996. Les forces navales du Bas Danube et de la Mer Noire aux 1er – Vie siècles. Oxford: Oxbow Books. Bowman, A. 1994. Life and Letters on the Roman Frontier. New York: Routledge. Boyanov, I. 2012. “Veterans and Society in Lower Moesia and Thrace during the Principate,” in The Lower Danube Roman Limes, eds.Vagalinski, L., Sharankov, N., and S. Torbatov, 251-269. Sofia: National Archaeological Institute with Museum. Braund, D. 1984. Rome and the Friendly King: the Character of Client Kingship. London: Croom Helm. Braund, D. 1994. Georgia in Antiquity: A History of Colchis and Transcaucasian Iberia 550 BC – AD 562. Oxford: Oxford University Press. Braund, D. 1996a. “Bosporus (2), the Cimmerian,” OCD, 3rd edition, 254. Oxford: Oxford University Press. Braund, D. 1996b. “Chersonesus (2),” OCD, 3rd edition, 320-321. Oxford: Oxford University Press. Braund, D. 1996c. “Chersonesus (3),” OCD, 3rd edition, 321. Oxford: Oxford University Press. Breeze, D. 2010. The Frontiers of Imperial Rome. Barnsley: Pen & Swords Books.

113

Exercitus Moesiae Brennan, P. 1980. “Combined Legionary Detachments as Artillery Units in Late-Roman Danubian Bridgehead Dispositions,” Chiron 10: 553-567. Brennan, P. 1998. “Divide and fall: the separation of legionary cavalry and the fragmentation of the Roman Empire,” in Ancient History in a Modern University, eds. T. W. Hillard, R. A. Kearsley, C. V. E. Nixon, and A.M. Nobbs, 238-44. Cambridge: William Eerdmans. Brennan, P. 2007. “Zosimos 2.34.1 and ‘the Constantinian reform’; using Johannes Lydos to expose an insidious fabrication,” in The Late Roman Army in the Near East from Diocletian to the Arab Conquest, eds. A. Lewin and P. Pellegrini, 211-18. Oxford, British Archaeological Reports. Broughton, T. R. S. and A. J. S. Spawforth. 1996. “Cappadocia,” OCD, 3rd edition, 288-289. Oxford: Oxford University Press. Campbell, B. 1984. The Emperor and the Roman Army. Oxford: Oxford University Press. Campbell, B. 1996a. “legion,” OCD, 3rd edition, 839-842. Oxford: Oxford University Press. Campbell, B. 1996b. "vexillum," OCD, 3rd edition, 1594. Oxford: Oxford University Press. Campbell, B. 2007. “Legion,” OCD, 4th online edition. Oxford: Oxford University Press. Campbell, B. 2012. Rivers and the Power of Ancient Rome. Chapel Hill: University of North Carolina Press. Charles, M. B. 2002. “The Flavio-Trajanic Miles: the Appearance of Citizen Infantry on Trajan’s Column,” Latomus 61 (2002): 666-695. Cavenaile, R. 1975. “Cohors I Hispanorum Equitata et Cohors I Hispanorum Veterana,” ZPE 18: 179-191. Cheesman, G. L. 1914. The Auxilia of the Roman Imperial Army. Oxford: Oxford University Press. Chevalier, R. 1976. Roman Roads. Trans. N. H. Field. London: Batsford. Christol, M., and T. Drew-Bear. 1995. “Inscriptions militaires d'Aulutrene et d'Apamée de Phrygie,” in La hiérarchie de l’armeé romaine, ed. Y. Le Bohec, 57-92. Paris: Diffusion de Boccard. Cichorius, C. 1893. “Ala,” RE 1: 1223-1269. Cichorius, C. 1900. “Cohors,” RE 4: 231-355. Colombo, M. 2009. “La forza numerica e la composizione degli eserciti campali durante l’alto impero: legion e auxilia da Cesare Augusto a Traiano,” Historia 58: 96-117. Conole, P. and R. D. Milns. 1983. “Neronian Frontier Policy in the Balkans: the Career of Ti. Plautius Silvanus,” Historia 33: 183-200. Cooley, A. 2012. The Cambridge Manual of Latin Epigraphy. Cambridge: Cambridge University Press. Cope, C. R. 1999. “Faunal Remains and Butchery Practices from Byzantine and Islamic Contexts (1993-1994 seasons),” in Caesarea Papers volume 2, eds. Holum, K., Raban, A. and J. Patrich, 405-417. Portsmouth, RI: Journal of Roman Archaeology. Coulston, J. 1989. “The Value of Trajan’s Column as a Source for Miltiary Equipment,” in Roman Military Equipment: the Sources of Evidence. Proceedings of the Fifth Roman Military Equipment Conference, ed. C. van Dreiel-Murray, 31-44. Oxford: BAR. Covacef, Z. 2000. “Cohors I Germanorum à Capidava,” in Army and Urban Development in the Danubian Provinces of the Roman Empire, eds. H. Ciugudean, and V. Moga, 285-291. Alba Iulia: Muzeul National al Unirii Alba Iulia. Cuff, B. 2010. The Auxilia in Roman Britain and the Two Germanies from Augustus to Caracalla: Family, Religion and ‘Romanization’. Unpublished PhD thesis. Toronto: University of Toronto. Cupcea, G. and F. Marcu. 2006. “The Size and Organization of the Roman Army and the Case of Dacia Under Trajan,” Dacia 50: 175-194. Cuvigny, H. 2005. Ostraca de Krokodilô: La correspondence militaire et sa circulation. Cairo: Institut Français d’archéologie orientale. Dabrowa, E. 2000. “Legio III Gallica,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 309-315. Paris: Diffusion de Boccard, 2000. Dando-Collins, S. 2010. Legions of Rome: The Definitive History of Every Imperial Roman Legion. London: Quercus. Davies, R. W. 1989. “Cohortes equitatae,” in Service in the Roman Army, eds. Breeze, D. and V. Maxfield, 141-151. Edinburgh: Edinburgh University Press. Derda, T., Dyczek, P., and J. Kolendo (eds.). 2008. Novae: Legionary Fortress and Late Antique Town. Warsaw: Center for Research on the Antiquity of Southeastern Europe University of Warsaw. Devijver, H. 1987. “Cohortes Cilicum in the Service of Rome,” ZPE 47: 173-183. Domaszewski, A. 1967. Die Rangordnung des Römischen Heeres, 2nd edition. Cologne: Böhlau. Dušanić, S. 2000. “Army and Mining in Moesia Superior,” in Kaiser, Heer und Gesellschaft in der Römischen Kaiserzeit, eds. G. Alföldy, G., B. Dobson, and W. Eck, 343-363. Stuttgart: Franz Steiner Verlag. Dušanić, S. 2007. “Three Sidelights on the Early Diplomata Militaria,” in Militärdiplome: Die Forschungsbeiträge der Berner Gespräche von 2004, eds. Speidel, M. A. and H. Lieb, 55-85. Stuttgart: Franz Steiner Verlag. Eadie, J. W. 1967. “The development of Roman mailed cavalry,” JRS 57: 161-173. Eck, W., Holder, P. Pangerl, A., and P. Weiss. 2015. “Fünf Bürgerrechtskonstitutionen für die Auxiliareinheiten von Moesia Superior aus Traianisch-Hadrianischer Zeit,” ZPE 194: 223-240. 114

Bibliography Eck, W., MacDonald, D. and A. Pangerl. 2008. “Ein weiteres Diploma aus der Konstitution des Antoninus Pius für die Truppen von Moesia Superior vom 23. April 157,” ZPE 165: 237-239. Eck, W. and A. Pangerl. 2005. “Zwei Konstitutionen für die Truppen Niedermösiens vom 9. September 97,” ZPE 151: 185-192. Eck, W. and A. Pangerl. 2006. “Neue Diplome für die Auxiliartruppen in den Mösischen Provinzen von Vespasian bis Hadrian,” Dacia 50: 93-108. Eck, W. and A. Pangerl. 2008. “Moesia und seine Truppen. Neue Diplome für Moesia und Moesia superior,” Chiron 38: 317-394. Eck, W. and A. Pangerl. 2009. “Moesia und seine Truppen II: Neue Diplome für Moesia, Moesia inferior und Moesia superior,” Chiron 39: 507-589. Eck, W. and A. Pangerl. 2010. “Ein weiteres Diplom der Konstitution Vespasians für die Truppen Mösiens vom 7 Februar 78 N. Chr.,” ZPE 173: 237-243. Eck, W. and A. Pangerl. 2012. “Ein weiteres Diploma us einer Konstitution für die Truppen von Moesia Inferior vom 14. August 99 N. Chr.,” ZPE 180: 295-306. Eck, W. and A. Pangerl. 2013. “Neue Diplome mit den Namen von Konsuln und Statthaltern,” ZPE 187: 273-294. Eck, W. and A. Pangerl. 2014a. “Eine Konstitution für das Heer von Moesia Inferior vom 1. Juni 125 in Fünf Diplomen,” ZPE 188: 245-249. Eck, W. and A. Pangerl. 2014b. “Neue Diplome für die Truppen von Moesia Superior und Inferior,” ZPE 192: 215237. Eck, W. and M. M. Roxan. 1997. "A Diploma of Moesia Inferior: 125 Iun. 1," ZPE 116: 193-203. Elton, H. 1996. Warfare in Roman Europe AD 350-425. Oxford: Oxford University Press. Elton, H. 2007. “Military Forces,” in The Cambridge History of Greek and Roman Warfare. Volume II: Rome from the Late Republic to the Late Empire, eds. Sabin, P. Van Wees, H., and M. Whitby, 270-309. Cambridge: Cambridge, University Press. Farnum, J. H. 2005. The Positioning of the Roman Imperial Legions. Oxford: BAR. Fear, A. (trans.). 2010. Orosius: Seven Books of History against the Pagans. Liverpool: Liverpool University Press. Fellman, R. 2000. “Die 11. Legion Claudia Pia Fidelis,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 127-131. Paris: Diffusion de Boccard. Ferrill, A. 1991. Roman Imperial Grand Strategy. Lanham: University Press of America. Filow, B. 1963. Die Legionen der Provinz Moesia von Augustus bis auf Diokletian. Wiesbaden: Aalen Scientia Verlag. Forni, G. 1958. “Contributo alla Storia della Dacia Romana,” Athaneum 36: 183-218. Franke, T. 2000. “Legio V Alaudae,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 39-48. Paris: Diffusion de Boccard. Fuhrmann, C. 2012. Policing the Roman Empire: Soldiers, Administration, and Public Order. New York: Oxford University Press. Gayet, F. 2006. “Les unités auxiliaires gauloises sous le Haut-Empire romain,” Historia 55: 64-105. Gencheva, E. 2012. “Roman Military Equipment from the Bulgarian Lands (1st – 3rd c. AD),” in The Lower Danube Roman Limes (1st – 6th C. AD), eds. Vagalinski, L., Sharankov, N. and S. Torbatov, 83-119. Sofia, National Achaeological Institute with Museum. Gerov, B. 1979. “Die Grenzen der römischen Provinz Thracia bis zur Gründung des Aurelianischen Dakien,” ANRW II.7: 212-240. Gerov, B. Beiträge zur Geschichte der Römischen Provinzen Moesien und Thrakien. Amsterdam: Geiben, 1980. Gilliver, C. M. 1996. “Mons Graupius and the role of auxiliaries in battle,” G & R 43: 54-67. Goldsworthy, A. 1996. The Roman Army at War 100 B.C. – A.D. 200. Oxford: Oxford University Press, 1996. Goldsworthy, A. (ed.). 1997. The Roman Army as a Community. Portsmouth, RI: Journal of Roman Archaeology. Goldsworthy, A. 2007. “War,” in The Cambridge History of Greek and Roman Warfare Volume II: Rome from the Late Republic to the Late Empire, eds. Sabin, P., Van Wees, H., and M. Whitby, 76-121. Cambridge: Cambridge University Press. Golubovic, S., Mrdjic, N. and S. Speal. 2009. “Killed by the Arrow: grave No. 152 from Viminacium,” in Waffen in Aktion: Akten der 16. Internationalen Roman Military Equipment Conference, eds. Busch, A. W., and H.-J. Schalles, 55-63. Mainz am Rhein: Verlag Philipp von Zabern. Gren, E. 1941. Kleinasien und der Ostbalkan in der Wirtschäftlichen Entwicklung der Römischen Kaiserzeit. Uppsala: Lundequist. Gudea, N. 1979. “The Defensive System of Roman Dacia,” Britannia 10: 63-87. Gudea, N. 2001. Die Nordgrenze der Römischen Provinz Obermoesien: Materialien zu ihrer Geschichte (86-275 N. Chr.). Mainz: Jahrbuch des Römisch-Germanischen Zentralmuseums. Gudea, N. 2005. Der Untermoesische Donaulimes und die Verteidigung der Moesischen Nord-Und Westküste der Schwarzen Meeres: Limes et Litus Moesiae Inferioris (86-275 n. Chr.). Mainz: Jahrbuch des RömischGermanischen Zentralmuseums.

115

Exercitus Moesiae Haensch, R. 2009. “Not Official, but Permanent: Roman Presence in Allied States. The Examples of Chersonesus Taurica and the Bosporan Kingdom,” in Diplomats and Diplomacy in the Roman World, ed. C. Eilers, 209229. Leiden: Brill. Haynes, I. 2013. Blood of the Provinces: The Roman Auxilia and the Making of Provincial Society from Augustus to the Severans. Oxford: Oxford University Press. Hirt, A. M. 2010. Imperial Mines and Quarries in the Roman World: Organizational Aspedcts 27 BC – AD 235. Oxford: Oxford University Press. Hoffman, B. 2013. “Soldiers and Civilians – A New Look at Asymmetric Warfare in the Eastern Roman Empire in the 1st to 3rd century AD,” in Warfare and Society in the Ancient Eastern Mediterranean, eds. O’Brien, S. and D. Boatright, 43-48. Oxford: BAR. Hoffmann, D. 1969/1970. Das spätrömische Bewegungsheer und die ‘Notitia Dignitatum’. Düsseldorf: Rheinland Verlag. Holder, P. 1980. Studies in the Auxilia of the Roman Army from Augustus to Trajan. Oxford: BAR. Holder, P. 2003. “Auxiliary Deployment in the Reign of Hadrian,” in Documenting the Roman Army, ed. J. J. Wilkes, 101-145. London: Institute of Classical Studies. Holder, P. 2005. “Alae in Pannonia and Moesia in the Flavian Period,” in Limes XIX, ed. Z Visy, 79-83. Pécs: University of Pécs. Holder, P. 2006a. “Auxiliary Deployment in the Reign of Trajan,” Dacia 50: 141-174. Holder, P. 2006b. Roman Military Diplomas Volume V. London: Institute of Classical Studies. Holder, P. 2007a. “Observations on Auxiliary Diplomas from Vespasian to Commodus,” in Militärdiplome: Die Forschungsbeiträge der Berner Gespräche von 2004, eds. Speidel, M. A. and H. Lieb, 105-163. Stuttgart: Franz Steiner Verlag. Holder, P. 2007b. “Observations on Multiple Copies of Auxiliary Diplomas,” in Militärdiplome: Die Forschungsbeiträge der Berner Gespräche von 2004, eds. Speidel, M. A. and H. Lieb, 165-186. Stuttgart: Franz Steiner Verlag. Isaac, B. 1988. “The Meaning of the Terms Limes and Limintanei,” JRS 78: 125-47. Isaac, B. 1990. Limits of Empire. Oxford: Oxford University Press. Ivanov, R. 2012. “The Roman Limes in Bulgaria (1st – 6th C. AD),” in The Lower Danube Roman Limes (1st – 6th C. AD), eds. Vagalinski, L., Sharankov, N. and S. Torbatov, 23-42. Sofia, National Achaeological Institute with Museum. Ivanov, R., Atanasov, G., and P. Donevski. 2006. The Ancient Durostorum. Silistra: the Municipality of Silistra. Ivantchik, A. 2014. “Roman Troops in the Bosporus. Old Problem in the Light of a New Inscription Found in Tanais,” Ancient Civilizations from Scythia to Siberia 20: 165-194. James, S. 2002. “Writing the Legions: the Past, Present, and Future of Roman Military Studies in Britain,” AJ 159: 158. James, S. 2001. Rome and the Sword. London: Thames and Hudson. Jeczmienowski, E. 2012. “The Fortifications of the Upper Moesian Limes. Topography, Forms, Garrison Types,” Swiatowit 10: 31-57. Jeremic, G. 2007. “Watchtowers and Signal Towers on the Middle Danube,” in The Lower Danube in Antiquity (VI C BC – VI C AD), ed. L. Vagalinski, 305-314. Sofia: National Institute of Archaeology and Museum. Jones, A. H. M. 1964. The Later Roman Empire 284-602. Oxford: Oxford University Press. Junkelmann, M. 1997. Panis Militaris: Die Ernährung des römischen Soldaten oder der Grundstoff der Macht. Mainz am Rhein: Verlag Philip Von Zabern. Juntenen, K. 2013. “The Lost Books of Cassius Dio,” Chiron 43: 459-486. Kagan, K. 2006. “Redefining Roman Grand Strategy,” JMH 70: 333-362. Kaiser, A. M. 2015. “Egyptian Units and the Reliability of the Notitia Dignitatum, pars Oriens,” Historia 64: 243261. Karavas, J. 2001. The evolution of Roman frontier fortification systems in the lower Danube provinces, 1st-2nd centuries A.D. Unpublished PhD dissertation. Durham: The University of Durham. Keppie, L. 1998. The Making of the Roman Army from Republic to Empire. Norman, OK: University of Oklahoma Press. Kim, H. J. 2013. The Huns, Rome and the Birth of Europe. Cambridge: Cambridge University Press. King, A. 1999a. “Animals and the Roman Army: The Evidence of Animal Bones,” in The Roman Army as a Community, eds. Goldsworthy, A., and I. Haynes, 139-150. Portsmouth, RI: Journal of Roman Archaeology. King, A. 1999b. “Diet in the Roman World: A Regional Inter-site Comparison of the Mammal Bones,” JRA 12: 168202. Kissel, T. 1995. Untersuchungen zur Logistik des römischen Heeres in den Provinzen des griechischen Ostens (27 v. Chr.-235 n. Chr.). St. Katharinen: Scripta Mercaturae Verlag. Knight, D. J. 1991. “The Movements of the Auxilia from Augustus to Hadrian,” ZPE 85: 189-208. Kraft, K. 1951. Zur Rekrutierung der Alen und Kohorten an Rhein und Donau. Bern: A. Francke. Kurzmann, R. 2006. Roman Military Brick Stamps: A Comparison of Methodology. Oxford: BAR. 116

Bibliography Landelle, M. 2015. “vexillatio,” in Wiley Encyclopedia of the Roman Army, eds. Le Bohec, Y., Brizzi, G., DeschlerErb, Greatrex, G., Rankov, B., and M. Reddé, 1066. Malden, MA: Wiley. Laporte, J.-P. 2000. “La Legio VIIa et la deduction des colonies augustéenes de Césarienne,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 555-579. Paris: Diffusion de Boccard. Le Bohec, Y. 2000. The Imperial Roman Army. Trans. R. Bate. London: Routledge. Le Bohec, Y., and C. Wolff (eds.). 2000a. Les Légions de Rome sous le Haut-Empire. Paris: Diffusion de Boccard. Le Bohec, Y. and C. Wolff. 2000b. “Legiones Moesiae Superioris,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 239-245. Paris: Diffusion de Boccard. Lee, D. 1993. Information and Frontiers. Cambridge: Cambridge University Press. Lemke, M. 2008. “The Location of the Roman Legionary Fortress at Novae and the Crossing of the Danube by Russian Forces in 1877,” in Novae: Legionary Fortress and Late Antique Town, eds. Derda, T., Dyczek, P., and J. Kolendo, 291-297. Warsaw: Center for Research on the Antiquity of Southeastern Europe University of Warsaw. Lemke, M. 2015. “Towards a Military Geography of Moesia Inferior,” in Limes XXII: Roman Frontier Studies, eds. Vagalinski, L., and N. Sharankov, 845-852. Sofia: National Institute of Archaeology. Lendon, J. E. 2004.“The Roman army now,” CJ 99: 441-449. Lendon, J. E. 2005. Soldiers and Ghosts. New Haven: Yale University Press. Lenski, N. 1997. “Initium mali Romano imperio: Contemporary Reactions to the Battle of Adrianople,” TAPhA 127: 129-68. Lenski, N. 2002. Failure of Empire: Valens and the Roman State in the 4th c. A.D. Berkeley: University of California Press. Le Roux, P. 1982. L’Armée romaine et l’organisation des provinces ibériques d’Auguste a l’invasion de 409. Paris: Diffusion de Boccard. Le Roux, P. 1986. “Les diplôme militaires et l’évolution de l’armée romaine de Claude à Septime Sévère: auxilia, numeri et nations,” in Heer und Integrationspolitik: die römischen Militärdiplome als historische Quelle, eds. Eck, W. and H. Wolff, 347-374. Cologne: Böhlau Verlag. Le Roux, P. 1995. “Le ravitaillement des armée romaines sous l’Empire,” in Du Latifundium au Latifondo, ed. R. Etienne, 403-424. Paris: Diffusion de Boccard. Liebeschuetz, W. 2007. “The Lower Danube Under Pressure: from Valens to Heraclius,” in The Transition to Late Antiquity on the Danube and Beyond, ed. A. Poulter, 101-134. Oxford: Oxford University Press. Lörincz, B. 2000. “Legio I Adiutrix,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 151-158. Paris: Diffusion de Boccard. Luttwak, E. N. 1976. The Grand Strategy of the Roman Empire from the First Century A.D. to the Third. Baltimore: Johns Hopkins University Press. MacDonald, D. and A. Mihaylovich. 2002. “A New Moesia Inferior Diploma of 97,” ZPE 138: 225-228. Mann, J. C. 1954. “A Note on the Numeri,” Hermes 82: 501-506. Matei-Popescu, F. 2001/2002. “Trupele auxiliare romane din Moesia Inferior,” SCIVA 52-53: 173-242. Matei-Popescu, F. 2006/2007. “The Auxiliary Units from Moesia Superior in Domitian’s Time and the Problem of CIL XVI 41,” Ephemeris Napocensis 16-17: 31-48. Matei-Popescu, F. 2007. “Participarea legiunlilor din Moesia Inferior la expeditiile Dacice ale lui Traian,” in Dacia Felix, 290-300. Cluj-Napoca: Editura Tribuna. Matei-Popescu, F. 2010a. The Roman Army in Moesia Inferior. Bucharest: Conphys Publishing House. Matei-Popescu, F. 2010b. “Cohortes Augustae Nervianae Pacenses Brittonum,” in Identitati Culturale Locale Si Regionale in Context European, eds. Pop, H., Bejinariu, I., Bacuet-Crisan, S., and D. Bacuet-Crisan, 395-398. Cluj-Napoca: Editura Mega. Matei-Popescu, F. 2010/2011. “The Roman Auxiliary Units of Moesia,” Il Mar Nero 8: 207-230. Matei-Popescu, F. 2013. “The Roman Auxiliary Units of Moesia,” Il Mar Nero 8: 207-230. Matei-Popescu, F. 2015. “The Auxiliary Units in Moesia Superior and Dacia. A Reviw and an Update,” in Limes XXII: Roman Frontier Studies, eds. Vagalinski, L., and N. Sharankov, 407-418. Sofia: National Institute of Archaeology. Matei-Popescu, F. and O. Tentea. 2006. “Participation of the Auxiliary Troops from Moesia Superior in Trajan’s Dacian Wars,” Dacia 50: 127-140. Mattern, S. P. 1999. Rome and the Enemy: Imperial Strategy in the Principate. Berkeley: University of California Press. Matthews, S. 2015. “Supplying the Limes Garrison in the Dobrogea,” in Limes XXII: Roman Frontier Studies, eds. Vagalinski, L., and N. Sharankov, 839-844. Sofia: National Institute of Archaeology. McGovern, W. M. 1939. The Early Empires of Central Asia. Chapel Hill: University of North Carolina Press. Meyer, A. 2013. The Creation, Composition, Service and Settlement of Roman Auxiliary Units Raised on the Iberian Peninsula. Oxford: BAR. Mihailescu-Birliba, L. 2008. “Un nouveau diplôme militaire de Mésie Inférieure,” Dacia 52: 199-210. Millar, F. 1977. The Emperor in the Roman World. London: Duckworth. 117

Exercitus Moesiae Mirković, M. 1986. Inscriptions de la Mésie Supérieure II. Belgrade: Centre d'études épigraphiques et numismatiques de la Faculté de philosophie de l'Université de Beograd Mirković, M. 2004. “The Roster of the VII Claudia Legion,” ZPE 146: 211-220. Mirković, M. and S. Duŝanić. 1976. Inscriptions de la Mésie Supérieure I. Belgrade: Centre d'études épigraphiques et numismatiques de la Faculté de philosophie de l'Université de Beograd. Mitchell, S. 1976. “Legio VII and the Garrison of Augustan Galatia,” CQ 26: 298-308. Mitchell, S. 1993. Anatolia: Land, Men, and Gods in Asia Minor – Volume 1, The Celts and the Impact of Roman Rule. Oxford: Oxford University Press. Mladenovic, D. 2012. Urbanism and Settlement in the Roman Province of Moesia Superior. Oxford: BAR. Mocsy, A. 1974. Pannonia and Upper Moesia: a History of the Middle Danube Provinces of the Roman Empire. London: Routledge. Mrozewicz, L. 1982. “Roman Military Settlements in Lower Moesia (1st – 3rd century),” Archaeologia 33: 79-105. Mugnai, N. 2014. “A Military Diploma for Antoninus Pius for the Troops of Moesia Inferior,” ZPE 190: 305-310. Neumann, A. 1958. “Vexillatio,” RE 16: 2442-2446. Nicasie M. J. 1998. Twilight of Empire: The Roman Army from the Reign of Diocletian until the Battle of Adrianople. Amsterdam: J. C. Geiben. Panaite, A. 2015. “Roman Roads in Moesia Inferior: Archaeological and Epigraphic Evidence,” in Limes XXII: Roman Frontier Studies, eds. Vagalinski, L., and N. Sharankov, 593-600. Sofia: National Institute of Archaeology. Papi, E. 2004. “Diploma militare da Thamusida (Mauretania Tingitana): 103/104,” ZPE 146: 255-258. Parker, H. M. D. 1958. The Roman Legions. Oxford: Oxford University Press. Paunov, E. I. and M. M. Roxan. 1997. “The Earliest Extant Diploma of Thrace, A.D. 114 (=RMD I 14),” ZPE 119: 269-279. Petolescu, C. C. and F. Matei-Popescu. 2008. “The Presence of the Roman Army from Moesia Inferior at the North of the Danube and the Making of the Dacia Inferior Province,” in Die Römischen Provinzen. Begri und Gründung, ed. I. Piso, 357-367. Cluj-Napoca: Editura Tribuna. Petolescu, C. C. and A. T. Popescu. 2004. “Ein Neues Militärdiplom für die Provinz Moesia Inferior,” ZPE 148: 269273. Petrović, P. 1979. Inscriptions de la Mésie Supérieure IV. Belgrade: Archaeological Institute. Petrović, P. 1995. Inscriptions de la Mésie Supérieure III/2. Belgrade: Archaeological Institute. Pferdehirt, B. 2004. Römische Militärdiplome und Entlassungsurkunden in der Sammlung des Römisch-Germanischen Zentral-museums. Mainz: Verlag des Römisch-Germanischen Zentralmuseums and Rudolf Habelt. Pflaum, H. G. 1960/1961. Les carriers procuratoriennes équestres sous le Haut-Empire romain. Paris: Institut français d’Archéologie de Beyrouth. Piso, I. 2000. “Les légions dans la province de Dacie,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 205-225. Paris: Diffusion de Boccard. Pollard, N. 1996. “The Roman Army as 'Total Institution' in the Near-East?' A Case Study from Dura-Europos,” in The Roman Army in the East, ed. D. Kennedy, 211-227. Portsmouth, RI: Journal of Roman Archaeology Supplement. Pollard, N. 2000. Soldiers, Cities, and Civilians in Roman Syria. Ann Arbor: University of Michigan Press. Pollard, N. 2013. “Imperatores castra dedicaverunt: Security, Army Bases, and Military Dispositions in Later Roman Egypt (Later Third-Fourth Century),” JLA 6: 3-36. Pollard, N. and J. Berry. 2012. The Complete Roman Legions. London: Thames and Hudson. Radnoti, A. 1959. “Zur Dislokation der Auxiliartruppen in den Donauprovinzen,” in Limes-Studien: Vorträge der 3. Internationalen Limes-Kongresses in Rheinfelden-Basel 1957, 134-151. Basel: Institut für Ur- und Frühgeschichte der Schweiz Rance, P. 2012. “The Third Equites Stablesiani at Cyrrhus,” Chiron 42: 345-58. Rankov, N. B. 1983. “A Contribution to the Military and Administrative History of Montana,” in Ancient Bulgaria, ed. A. Poulter, 40-73. Nottingham: University of Nottingham. Rankov, B. 2007. “Military Forces [the Late Republic and the Principate],” in The Cambridge History of Greek and Roman Warfare Volume II: Rome from the late Republic to the late Empire, eds. Sabin, P., Van Wees, H., and M. Whitby, 30-74. Cambridge: Cambridge University Press. Reddé, M. 2000. “Legio VIII Augusta,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 119-126. Paris: Diffusion de Boccard. Reddé, M. 1986. Mare Nostrum. Rome: Bibliothèque des Écoles françaises d’Athènes et de Rome. Richardot, P. 2005. La Fin de l’Armée Romaine 284-476. Paris: Economica, 2005. Richardson, A. 2002. “Camps and Forts of Units and Formations of the Roman Army,” Oxford Journal of Archaeology 21: 93-107. Rice, T. T. 1981. “The Scytho-Sarmatian Tribes of South-Eastern Europe,” in The Roman Empire and its Neighbours, 2nd edition, ed. F. Millar, 281-293. London: Duckworth. Ritterling, E. 1925. “Legio,” RE 12: 1186-1837. 118

Bibliography Rocco, M. 2012. L’esercito romano tardoantico: persistenze e cesure dai Severi a Teodosio I. Studi e progetti. Padova: libreriauniversitaria.it edizioni. Rolle, R. 1989. The World of the Scythians. Berkeley: University of California Press. Rossi, L. 1971. Trajan’s Column and the Dacian wars. Ithaca: Cornell University Press. Roth, J. 1994. “The Size and Organization of the Roman Imperial Legion,” Historia 43: 346-362. Roth, J. 1999. The Logistics of the Roman Army at War (264 B.C. – A.D. 235). Leiden: Brill. Rowell, H. T. 1937. “Numerus,” RE 17: 1327-1341. Rowell, H. 1939. “The Honesta Missio from the Numeri of the Roman Imperial Army,” YCS 6: 73-107. Roxan, M.M. 1978. Roman Military Diplomas 1954-1977. London: Institute for Classical Studies. Roxan, M. M. 1985. Roman Military Diplomas 1978 – 1984. London: Institute for Classical Studies. Roxan, M. M. 1994. Roman Military Diplomas 1985 – 1993. London: Institute for Classical Studies. Roxan, M.M. and P. Holder. 2003. Roman Military Diplomas IV. London: Institute for Classical Studies. Roxan, M. M. and P. Weiss. 1998. “Die Auxiliartruppen der Provinz Thracia. Neue Militärdiplome der Antoninenzeit,” Chiron 28: 371-420. Saddington, D.B. 1982. The Development of the Roman Auxiliary Forces from Caesar to Vespasian 49 B.C. – A.D. 79. Harare: University of Zimbabwe. Sarnowski, T. 1988. “Das Römische Heer im Norden des Schwarzen Meeres,” Archeologia 38: 61-98. Sarnowski, T. 1991. “Barbaricum und ein bellum Bosporanum in einer Inschrift aus Preslav,” ZPE 87: 137-144. Sarnowski, T. 1995a. “Legionsziegel an militärischen und zivilen Bauplätzen der Prinzipatszeit in Niedermösien,” in Roman Frontier Studies 1995, eds. Groenman van Waateringe, W. van Bek, B. L. ,Willems, W. J. H., and S. L. Wynis, 497-501. Oxford: Oxbow. Sarnowski, T. 1995b. “L'organisation hiérarchique des vexillationes Ponticae au miroir des trouvailles épigraphiques récentes,” La hiérarchie de l’armeé romaine, ed. Y. Le Bohec, 323-329. Paris: Diffusion de Boccard. Sarnowski, T. 2006a. “The Phantom Squadron of the Ravennate Fleet on the Black Sea in the 1st Century AD,” ZPE 157: 256-260. Sarnowski, T. 2006b. “Ti. Plautius Silvanus, Tauric Chersonesos and Classis Moesica,” Dacia 50: 85-92. Sarnowski, T., and V. Zubar. 1996. “Römische Besatzungstruppen aufder Südkrim und eine Bauinschrift aus dem Kastell Charax,” ZPE 112: 229-234. Sauer, E. 2000. “Alchester, a Claudian « vexillation fortress » near the western boundary of the Catuvellauni,” AJ 157: 1-78. Saxer, R. 1967. Untersuchungen zu den Vexillationen des Römischen Kaiserheeres von Augustus bis Diokletian. Köln: Böhlau Verlag. Schindel, N. 2010. “Ein Militärdiplomfragment aus Moesia Inferior (136 n. Chr.),” ZPE 174: 259-263. Schmitz, M. 2012. “Dacia”, in Wiley Encyclopedia of Ancient History, eds. Bagnall, R., Brodersen, K., Champion, C., Erskine, A., and S. Huebner. Malden: Wiley. Scorpan, C. 1981. “Cohors I Cilicum at Sacidava and Scythia Minor,” JRS 71: 98-102. Shaw, B. 1984. “Bandits in the Roman Empire,” P&P 105: 3-52. Sidebottom, H. 2013. The Amber Road. London: Penguin. Southern, P. 1989. “The Numeri of the Roman Imperial Army,” Britannia 20: 81-140. Spaul, J. 1994. Ala2. Andover: Nectoreca Press. Spaul, J. 2000. Cohors2. Oxford: BAR. Spaul, J. 2002. Classes Imperii Romani. Andover: Nectoreca Press. Speidel, M. P. 1975. “The Rise of Ethnic Units in the Roman Imperial Army,” ANRW II.3: 202-231. Speidel, M. P. 1982. “Auxiliary Units Named After Their Commanders: four new examples from Egypt,”Aeg 62: 165-172. Speidel, M. P. 1985. “Bithynian troops in the Kingdom of the Bosporus,” EA 6: 97-102. Speidel, M. P. 1992a. “Work to be done on the Organization of the Roman Army,” in Roman Army Studies Volume 2, 13-20. Stuttgart: Franz Steiner Verlag. Speidel, M. P. 1992b. “Regionarii in Lower Moesia,” in Roman Army Studies Volume 2, 140-144. Stuttgart: Franz Steiner Verlag. Speidel, M. A. 1998. “Legio IV Scythica, Its Movements and Men,” in The Twin Towns of Zeugma on the Euphrates: Rescue Work and Historical Studies, ed. D. Kennedy, 163-204. Portsmouth, RI: Journal of Roman Archaeology. Speidel, M. A. 2000. “Legio IV Scythica,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 327-337. Paris: Diffusion de Boccard. Speidel, M. A. 2015. “The Roman Army,” in The Oxford Handbook of Roman Epigraphy, eds. Bruun, C. and J. Edmondson, 319-344. Oxford: Oxford University Press. Stäcker, J. 2003. Princeps und miles. Hildesheim: Georg Olms Verlag. Stallibrass S. and R. Thomas (eds.). 2008. Feeding the Roman Army: The Archaeology of Production and Supply in NW Europe. Oxford: Oxbow. Starr, C. 1960. The Roman Imperial Navy: 31 BC to AD 324. Cambridge: W. Heffer. 119

Exercitus Moesiae Stefan, A. S. 2005. Les guerres daciques de Domitien et de Trajan: architecture militaire, topographie, images et histoire. Rome: Collection de l’école Française de Rome. Strobel, K. 1984. Untersuchungen zu den Dakerkriegen Trajans: Studien zur Geschichte des mittleren und unteren Donauraumes in der Hohen Kaiserzeit. Bonn: Dr. Rudolf Habelt. Strobel, K. 1989. Die Donaukriege Domitians. Bonn: Dr. Rudolf Habelt. Strobel, K. 2000. “Zur Geschicte der Legiones V (Macedonica) und VII (Claudia pia fidelis) in der frühen Kaiserzeit und zur Stellung der Provinz Galatia,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 515-528. Paris: Diffusion de Boccard. Syme, R. 1928. “Rhine and Danube Legions under Domitian,” JRS 18: 41-54. Syme, R. 1933. “Some Notes on the Legions Under Augustus,” JRS 23: 14-33. Syme, R. 1934. “Lentulus and the Origin of Moesia,” JRS 24: 113-137. Szilágyi, J. 1954. “Les Variations des Centres de Prépondérance Militaire dans les Provinces Frontières de L’Empire Romain,” AantHung 2: 117-219. Tentea, O. and F. Matei-Popescu. 2004. “Alae et Cohortes Daciae et Moesiae,” AMN 39/40: 259-296. Tomlin, R. 1972. “Seniores-Iuniores in the late Roman field army,” AJP 93: 253-78. Tomlin, R. 2000. “The legions in the late empire,” in Roman Frontiers and Their Fortresses, ed. R. J. Brewer, 159-78. London/Cardiff: Society of Antiquaries of London. Tomlin, R. 2008. “Jones and the army of the 4th c.,” in A. H. M. Jones and the Later Roman Empire, ed. D. Gwyer, 143-165. Leiden: Brill. Toplyn M. 2006. “Livestock and Limitanei: the zooarchaeological evidence”, in The Roman Frontier in Central Jordan: Final Report on the Limes Arabicus Project, 1980-1989, ed. S. T. Parker, 463-507. Washington DC: Dumbarton Oaks. Tully, G. D. 1998. “The stratarkhes of Legio VI Ferrata and the employment of camp prefects as vexillation commanders,” ZPE 120: 226-232. Tully, G. D. 2002. “The Vexillarii in Pseudo-Hyginus’ De Munitionibus Castrorum and Some Observations on the Strengths and Structure of Legionary Vexillations,” AHB 16: 127-134. Tully, G. D. 2004. “Did centurions lead detachments of their legions in wartime?” AClass 47: 139-150. Turner, B. 2010. Military Defeats, Casualties of War and the Success of Rome. Unpublished PhD thesis. Chapel Hill: University of North Carolina. Turner, B. 2013. “War Losses and Worldview: Re-examining the Roman Funerary Altar at Adamclisi,” AJP 134: 277-304. Vagalinski, L. 2007. The Lower Danube in Antiquity (VI C BC – VI C AD). Sofia: National Institute of Archaeology and Museum. Vagalinski, L., Sharankov, N., and S. Torbatov (eds.). 2012. The Lower Danube Roman Limes (1st – 6th C. AD). Sofia: National Institute of Archaeology and Museum. Vannesse, M. 2012. “Ammianus Marcellinus on Julian’s Persian Expedition (363 A.D.): A Note on the Supply Chain,” in Studies in Latin Literature and Roman History XVI, ed. C. Deroux, 639-645. Brussels: Éditions Latomus. Varbanov, V. 2007. “Barbarian Invasions in the Provinces of Moesia Inferior and Thracia between AD 138-192 (According to the Numismatic Data),” in The Lower Danube in Antiquity (VI C BC – VI C AD), ed. L. Vagalinski, 153-170. Sofia: National Institute of Archaeology and Museum. Velkov, V. 1990. “Cohors II Lucensium Equitata in Moesia and Thracia,” AantHung 41: 247-256. Wagner, W. 1938. Die Dislokation der römischen Auxiliarformationen in den Provinzen Noricum, Pannonien, Moesien und Dakien von Augustus bis Gallienus. Berlin: Junker und Dünnhaupt. Webster, G. 1988. The Roman Imperial Army of the first and second centuries A.D., 3rd edition. Norman, OK: University of Oklahoma Press. Weiss, P. 1997. “Neue Militärdiplome,” ZPE 117: 227-268. Weiss, P. 1999. “Ein Diplom des Antoninus Pius für Moesia Inferior von Dez. 145/Dez. 146,” ZPE 124: 279-286. Weiss, P. 2001. “Das Konsulnpaar vom 7. April 145 n. Chr. Noch einmal zu den Auxilien von Moesia inferior und Thracia unter Antoninus Pius,” ZPE 134: 261-266. Weiss, P. 2008. “Militärdiplome für Moesia (Moesia, Moesia superior, Moesia inferior),” Chiron 38: 267-316. Weiss, P. 2009. “Statthalter und Konsulndaten in Neuen Militärdiplomen,” ZPE 171: 231-252. Wesch-Klein, G. 2012. “Legions, history and location of,” in Wiley Encyclopedia of Ancient History, eds. Bagnall, R., Brodersen, K., Champion, C., Erskine, A., and S. Huebner. Malden: Wiley. Whately, C. 2005. The Movement and Emplacement of the Legions and Auxiliary Units of the Roman Army in Moesia from 29 BC to AD 235. Unpublished MA thesis. Hamilton: McMaster University. Whately, C. 2012. “Kataphractoi,” in Wiley Encyclopedia of Ancient History, eds. Bagnall, R., Brodersen, K., Champion, C., Erskine, A., and S. Huebner. Malden: Wiley. Whately, C. 2015. “Making Sense of the Frontier Armies in Late Antiquity: an Historian’s Perspective”, in Roman Military Architecture on the Frontiers: Armies and Their Architecture in Late Antiquity, eds. Collins, R. and M. Weber, 6-17. Oxford: Oxbow. Wheeler, E. 1979. “The legion as a phalanx,” Chiron 9: 303- 318. 120

Bibliography Wheeler, E. 1993a. “Methodological Limits and the Mirage of Roman Strategy: Part I,” JMH 57: 7-41. Wheeler, E. 1993b. “Methodological Limits and the Mirage of Roman Strategy: Part II,” JMH 57: 215-240. Wheeler, E. 2000. “Legio XV Apollinaris: From Carnuntum to Sartala – and beyond,” in Les Légions de Rome sous le Haut-Empire, eds. Y. Le Bohec, and C. Wolff, 259-308. Paris: Diffusion de Boccard. Wheeler, E. 2004a. “The Legion as a Phalanx in the Late Empire (1),” in Les Légions de Rome sous le Haut-Empire, eds. Le Bohec Y., and C. Wolff, 309-358. Paris: Diffusion de Boccard. Wheeler, E. 2004b. “The Legion as a Phalanx in the Late Empire, Part II,” Réma 1: 147-175. Wheeler, E. 2010. “Rome’s Dacian Wars: Domitian, Trajan, and Strategy on the Danube, Part I,” JMH 74: 11851227. Wheeler, E. 2011. “Rome’s Dacian Wars: Domitian, Trajan, and Strategy on the Danube, Part II,” JMH 75: 191-219. Wheeler, E. 2012a. “Notitia Dignitatum, Or. 38 and Roman Deployment in Colchis: Assessing Recent Views,” in Visions de l’Occident romain. Hommages à Yann Le Bohec, eds. Cabouret, B., Groslambert, A., and C. Wolff, 621-676. Paris: Diffusion de Boccard. Wheeler, E. 2012b. “Roman Fleets in the Black Sea: Mysteries of the ‘Classis Pontica’,” Acta Classica 55: 119-154. Whitby, M. 2007. “The Late Roman Army and the Defence of the Balkans,” in The Transition to Late Antiquity on the Danube and Beyond, ed. A. Poulter, 135-161. Oxford: Oxford University Press. Whittaker, C. R. 1994. Frontiers of the Roman Empire: A Social and Economic Study. Baltimore: Johns Hopkins. Whittaker, C. R. 2004. Rome and its Frontiers: the Dynamics of Empire. London: Routledge. Wilkes, J. J. 1969. Dalmatia. London: Routledge. Wilkes, J. J. 1983. “Romans, Dacians and Sarmatians in the First and Early Second Centuries,” in Rome and her Northern Provinces, eds. B. Hartley and J. Wacher, 255-289. Gloucester: Alan Sutton. Wilkes, J. J. 1996. “Moesia,” OCD, 3rd edition, 993. Oxford: Oxford University Press. Wilkes, J. J. 2005. “The Roman Danube: An Archaeological Survey,” JRS 95: 124-225. Woolliscroft, D. I. 2001. Roman Military Signalling. Stroud: Tempus. Zahariade, M. 1988. Moesia Secunda, Scythia şi Notitia Dignitatum. Bucharest: Editura Academiei Republicii Socialiste România. Zahariade, M. 20007. “The Tetrarchic Reorganization of the Limes Scythicus,” in The Lower Danube in Antiquity (VI C BC – VI C AD), ed. L. Vagalinski, 1-2. Sofia: National Institute of Archaeology and Museum. Zahariade, M. 2012a. “Decebalus,” in Wiley Encyclopedia of Ancient History, eds. Bagnall, R., Brodersen, K., Champion, C., Erskine, A., and S. Huebner. Malden: Wiley. Zahariade, M. 2012b. “Moesia,” in Wiley Encyclopedia of Ancient History, eds. Bagnall, R., Brodersen, K., Champion, C., Erskine, A., and S. Huebner. Malden: Wiley. Zahariade, M. M. and N. Gudea. 1997. The Fortifications of Lower Moesia (AD 86-275). Amsterdam: Adolf M. Hakkert. Zubar, V. 1995. “Zur römischen Militärorganisation auf der Taurike in der zweiten Hälfte des 2. und am Anfang des 3. Jahrhunderts,” Historia 44: 192-203. Zubar, V. 2001. “Tauric Chersonesus and the Roman Empire,” in Ancient Greek Colonies on the Black Sea 2,eds. D. Grammenos and E. Petropoulos, 729-788. Oxford: BAR. Zubar, V. 2007. “Roman Military Units in Olbia,” in Classical Olbia and the Scythian World from the Sixth Century BC to the Second Century AD, eds. Braund, D. and S. Kryzhiskiy, 173-178. Oxford: Oxford University Press, 2007.

121

Index Abrittus 69, 111 Aegyssus 11, 60, 111, 112 Alae Asturum 18, 21, 34, 37, 38, 57, 67, 92 Gallorum Atectorigiana 17, 18, 21, 33, 34, 37, 38, 41, 48, 49, 67, 69, 92 Bosporanorum 17, 92 I Claudia nova miscellanea 32, 47, 48, 68, 85, 92 I Flavia Gaetulorum 33, 34, 37, 38, 41, 48, 49, 92 Gallorum Flaviana 21, 32, 33, 34, 35, 36, 38, 47, 48, 68, 69, 92 Hispanorum 21, 33, 34, 35, 37, 38, 92 Hispanorum (Campagonum c R) 30, 92 Moesica 17, 21, 92 I Pannoniorum 33, 34, 37, 38, 69, 92 I Pannoniorum et Gallorum catafracta 40, 41, 48, 49, 53, 82, 92 Pansiana 18, 92 Praetoria 27, 28, 30, 68, 80, 92 Scubulorum 17, 18, 21, 67, 92 Sulpicia 17, 21, 92 I Vespasiana Dardanorum 21, 33, 34, 38, 41, 49, 67, 69, 93 II Hispanorum et Aravacorum 35, 36, 38, 41, 48, 49, 69, 93 II Pannoniorum 27, 28, 30, 68, 93 Appiaria 67, 68, 111 Aquae 60, 67, 110 Arrubium 67, 69, 71, 111, 112 Augustus 4, 11, 13, 14, 43, 59, 77 Aureus Mons 70, 110 Barbosi 60, 69, 111, 112 Boljetin 70, 110 Bononia 68, 70, 110 Capidava 49, 53, 69, 71, 112 Carsium 53, 69, 71, 111, 112 Cerna 48, 69, 111 Cius 69, 139, 112 Cohortes I Antiochensium sagittaria 19, 21, 27, 28, 30, 32, 47, 48, 66, 82, 88, 89, 93 I Augusta Nerviana Pacensis milliaria Brittonum 37, 38, 93 I Aurelia Dardanorum 45, 48, 68, 118 I Bracaraugustanorum 19, 21, 34, 35, 38, 39, 41, 49, 59, 62, 93 I Bracarum c R 19, 59, 62, 93 I Britannica milliaria c R 30, 93 I Brittonum milliaria 30, 37, 38, 93 I Cantabrorum 19, 21, 67, 85, 93 I Cilicum sagittaria 18, 19, 21, 27, 28, 30, 38, 41, 48, 49, 57, 62, 66, 67, 69, 82, 93 I Cisipadensium 18, 19, 27, 28, 30, 66, 67, 68, 93 I Cretum sagittaria 18, 27, 28, 30, 31, 32, 47, 48, 67, 68, 70, 71, 80, 82, 93 I Flavia Bessorum 218, 30, 35, 68, 80, 89, 93 I Flavia Commagenorum equitata sagittaria 33, 34, 35, 38, 82, 93 I Flavia Hispanorum milliaria 27, 28, 30, 67, 93 I Flavia Numidarum equitata 36, 38, 41, 48, 85, 93 I Germanorum 40, 41, 49, 69, 90, 93 I Hispanorum veterana equitata 37, 38, 60, 67, 93 I Lepidiana c R equitata 36, 38, 41, 85, 93 Lusitanorum 18, 19, 26, 30, 32, 47, 48, 69, 93 Lusitanorum Cyrenaica 28, 33, 34, 35, 36, 38, 41, 49, 93 Mattiacorum 20, 21, 36, 37, 66, 93 Montanorum 18, 26, 28, 30, 31, 32, 57, 48, 67, 68, 80, 93 I Pannoniorum 32, 33, 47, 48, 68, 93 I Raetorum 19, 34, 35, 66, 93 122

Index I Sugambrorum tironum 19, 21, 34, 36, 39, 85, 93 I Claudia Sugambrorum veterana equitata 17, 18, 19, 21, 34, 36, 38, 39, 41, 48, 49, 59, 62, 69, 93 I Thracum Syriaca equitata 18, 19, 21, 30, 40, 41, 49, 58, 62, 66, 67, 69, 93 I Thracum c R 28, 29, 30, 72, 93 I Tyriorum sagittaria 19, 20, 31, 36, 38, 82, 93 I Ubiorum equitata 19, 20, 33, 34, 35, 38, 69, 93 I Vindelicorum milliaria c R 28, 29, 30, 80, 93 II Augusta Nerviana Pacensis milliaria 31, 37, 38, 94 II Aurelia Dardanorum 45, 48, 67, 94 II Aurelia nova milliaria equitata 45, 46, 48, 94 II Aurelia nova sacorum 45, 46, 48, 68, 94 II Bracaraugustanorum equitata 34, 38, 41, 49, 94 II Brittonum milliaria c R p f 28, 29, 30, 35, 80, 85, 94 II Chalcidenorum sagittaria 17, 19, 20, 34, 38, 41, 49, 82, 94 II Flavia Bessorum 33, 34, 35, , 94 II Flavia Brittonum 35, 36, 37, 38, 41, 48, 49, 68, 72, 94 II Flavia Commagenorum 28, 30, 33, 94 II Flavia Numidarum equitata 38, 94 II Gallorum 33, 34, 35, 38, 47, 69, 94 II Gallorum Macedonica 27, 28, 30, 32, 47, 48, 80, 94 II Gallorum Pannonica 32, 48, 94 II Hispanorum 28, 29, 30, 68, 94 II Lucensium equitata 20, 21, 34, 40, 59, 62, 66, 69, 94 II Mattiacorum milliaria equitata 5, 20, 36, 38, 41, 48, 49, 68, 69, 94 II Varcianorum 17, 18, 94 III Augusta Nerviana Brittonum 30, 31, 52, 94 III Brittonum 29, 31, 32, 47,48, 68, 80, 94 III Campestris 30, 31, 32, 47, 67, 94 III Gallorum 19, 20, 21, 34, 51, 85, 94 III Hispanorum 19, 20, 94 IIII Cypria c R 30, 58, 62, 68, 84, 94 IV Gallorum 19, 21, 34, 69, 85, 94 IIII Raetorum 27, 28, 30, 31, 80, 85, 94 V Gallorum 19, 21, 27, 28, 30, 32, 65, 68, 85, 94 V Hispanorum 27, 28, 30, 32, 47, 48, 51, 67, 85, 94 VI Thracum 28, 30, 68, 94 VII Breucorum c R 28, 30, 67, 80, 94 VII Gallorum equitata 19, 21, 34, 35, 38, 67, 85, 94 VIII Gallorum 19, 21, 85, 94 VIII Raetorum 30, 80, 94 Cuppae 75, 110 Diana 60, 68, 110 Dinogetia 60, 111, 112 Dobrudja 46, 57, 60, 67, 69, 73, 111, 112 Domitian 54, 19, 23, 24, 25, 26, 27, 29, 32, 34, 41 Drobeta 68, 110 Durostorum 5, 7, 65, 111 Egeta 60, 70, 110 Hadrian 24, 27, 31, 39, 41, 49, 85 Histria 53, 112 Iatrus 70, 71, 111 Kosmaj 68, 110 Legiones I Italica 16, 21, 24, 26, 27, 34, 58, 41, 43, 44, 49, 56, 57, 58, 59, 61, 65, 70, 71, 92 III Gallica 64, 92 IIII Flavia 26, 27, 28, 30, 32, 43, 44, 48, 64, 65, 92 V Alaudae 16, 64, 92 V Macedonica 5, 12, 13, 14, 16, 18, 21, 24, 26, 27, 34, 38, 41, 43, 44, 56, 57, 61, 64, 65, 66, 70, 71, 92 VII Claudia 14, 15, 16, 21, 24, 26, 27, 28, 30, 32, 43, 44, 53, 64, 65, 70, 71, 92 VIII Augusta 14, 15, 16, 51, 64, 92 XI Claudia 5, 14, 26, 27, 43, 44, 52, 53, 56, 57, 58, 61, 65, 71, 92 Lederata 53, 70, 110 123

Exercitus Moesiae Lomets 67, 68, 69, 111 Marcus Aurelius 42, 43, 45, 46, 47, 48, 85 Naissus 14, 18, 45, 64, 66, 67, 68, 110 Nicopolis 67, 111 Nigrinianis-Candidiana 69, 111 Novae 7, 14, 45, 51, 64, 65, 67, 68, 73, 111 Noviodunum 60, 65, 111, 112 Oescus 3, 18, 19, 64, 65, 66, 111 Pojejena 67, 110 Pontes 68, 110 Ratiaria 21, 64, 65, 66, 68, 84, 110 Sacidava 46, 69, 71, 111 Saldum 66, 110 Scupi 64 Securisca 67, 111 Serdica 111 Sexaginta Prista 48, 66, 68, 71, 111 Singidunum 26, 64, 65, 67, 68, 110 Sirmium 13, 68 Stojnik 68, 110 Sucidava 69, 71, 111 Taliata 60, 66, 110 Teutoburgium 110 Timacum Minus 18, 66, 67, 68, 110 Timok River Valley 66 (see Timacum Minus) Tomi/Tomis 18, 48, 67, 69, 110, 111 Trajan 4, 7, 23, 24, 25, 26, 29, 31, 36, 39, 40, 52, 80 Transdierna 60, 67, 68, 110 Translederata 70, 110 Transmarisca 67, 69, 111 Trimammium 68, 111 Tricornium 68, 110 Troesmis 11, 44, 60, 65, 67, 69, 111, 112 Troianhissar 69 (see Lomets) Utus 70 Valkenburg 66 Vespasian 16, 19, 20, 26, 27, 28, 51, 55, 69 Vexillationes Capidavensium 51, 53 XII Catafractariorum 51, 53, 60 Chersonissitanae 51, 61 classis Ravennatis 51 Egisseis Valerius (?) 51 equitum Moesiae inferioris et Daciae 51 equitum scutariorum 51, 53 exercitus Moesiae 51, 58 exercitus Moesiae inferioris 51, 58 expeditionis per Asiam et Lyciam Pamphyliam 51 legionis I Italicae et legionis II Herculiae 51 legionum I Italicae V Macedonicae et VII ad Tropaeum Traoiani 51 legionum I Italicae V Macedonicae XI Claudiae et cohortium 51 legionum I Italicae XI Claudiae 51, 53 legionum I Italicae XI Claudiae classis Flavia Moesicae 51, 52 legionis V Macedonicae 51, 56 legionum V Macedonicae XI Claudiae 51 legionis VII Claudiae 51 legionis XI Claudiae 51, 58 Moesiae inferioris 51 per Germaniam et Raetiam et Noricum et Pannoniam et Moesiam 51 Ponticis aput Scythia et Tauricam 51, 57, 61 Viminacium 26, 44, 64, 65, 67, 83, 110 124